The Biology, Epidemiology and Control of Liverwort Infestation of Nursery Plant Containers

The biology, epidemiology and control of liverwort infestation of nursery plant containers

Jill Elizabeth England BSc (Hons)

A thesis submitted to Imperial College London in partial fulfillment of the requirements governing the award of the degree of Doctor of Philosophy.

Division of Biology Imperial College London

University of London

October 2007 Declaration

I declare that this thesis is my own work, and has not been submitted in any previous application for a degree. The data presented within are from experiments and analyses that I carried out, except where other sources have been acknowledged.

Jill England

2 Abstract

Liverwort (Marchantia po/ymorpha) growing on media surface is a major weed problem in plant nurseries; control using chemicals is inadequate and hand removal expensive. This project investigated liverwort biology, life cycle and epidemiology, and considered non- pesticidal control methods: fungal antagonists and glucosinolate hydrolysis products.

Variation of temperature (25 °C, 15 °C and 10 °C) and light level (800 pmol m-2 st and 400 pmol m-2 s-1) produced greater liverwort growth (radial expansion, fresh and dry weights) and development (gemma cup production) under high temperature than low, and under low light than high light; less liverwort growth was observed when grown under the extreme light levels. Endogenous plant growth regulator lunularic acid (LNA) levels increased under long day lengths.

The number of liverwort gemmae (vegetative propagules) dispersed during overhead irrigation was related to water drop size, not quantity. More gemmae were dispersed using overhead irrigation than drip or capillary matting, and with increased irrigation frequency.

Bioassays evaluating potential fungal antagonists found Fusarium equiseti and Phaeodothis winter' to be more aggressive than Bryoscyphus atromarginatus, Penicillium ve/utinum and Trichoderma harzianum on liverwort thalli. However, F equiseti and P. winteri did not provide control in glasshouse experiments.

The glucosinolate glucolimnanthin and its hydrolysis product (3-methoxybenzyl isothiocyanate) were extracted from Limnanthes alba seed meal and tissue. Bioassays indicated a herbicidal effect of 3-methoxybenzyl, phenylethyl, benzyl and allyl isothiocyanates (ITC), and herbicides Metazachlor and Lenacil on liverwort gemmae, 3- methoxybenzyl ITC having the lowest ED50. These ITCs did not inhibit cress (Lepidium sativum) seed germination or radicle elongation. Root exudates of Diplotaxis tenuifolia, grown hydroponically, were harvested and analysed for ITCs, however these were not found.

Practical implications of this research are that provision of shade, exposure to high light levels and use of sub-irrigation and improved irrigation scheduling could reduce liverwort infestation. Control using glucosinolate hydrolysis products requires further development.

3 Acknowledgements I would like to offer my thanks to the following people for their help and support throughout this project:

Professor Mike Jeger for developing the project, for his support, well-judged supervision and encouragement. Dr John Rossiter and Dr Peter Buckley for their support throughout.

The Horticultural Development Council for funding the project, and HDC Technical Manager Lindrea Latham and Scott Raffle for their advice with the project reports. Professor Asakawa of Tokushima Bunri University, Japan for the sample of lunularic acid he kindly donated, and Access Irrigation for the gift of irrigation nozzles. Chris Atkinson and Roger Payne for their help with the irrigation experiment based at East Mailing Research. John Atwood (ADAS) for providing dying liverwort bearing fungal antagonists. Bob Chapman and Frank for providing help and space for the shading experiment at Palmstead Nursery Wye.

Practical help, advice and much appreciated general support, all given with unfailing generosity, were provided throughout the project by numerous Imperial College at Wye staff members: Janet Spicer, Sue Farris, John Haines, Valerie Elliott, Wendy Byrne, Norma Wilson and Mark Bennett, and in the glasshouses Adrian Russell, Jo Browning, Peter and Diana. Professor John Mansfield, Dr Glen Powell and Dr Simon Hodge for providing help, advice, equipment and space. Dr Fabiana Gordon for her help with the statistical analysis. Dr John Rossiter, Mark H Bennett, Eleana Kazana and Ian Brown for their much needed help given to a laboratory novice. Ana Woods for her invaluable help in setting up and collecting results from the glasshouse fungal inoculation experiment.

Fellow students Dr Jo Sharp, Lucinda Denness, Naomi Young, Dr Marcelo Kern, Lucinda Warner and Kathy Mitchell for coffee and chat. Dr Tom Pope and Dr Jenny Barker for their friendship and their by now very well developed listening and advisory skills. Finally I would like to thank my husband Bruce, sons James and Richard, my parents and parents-in-law for their unfailing support and patience throughout eight years of study.

4 Table of contents

Chapter 1 Introduction and literature review 21 1.1 Introduction 21 1.1.1 Main aims and objectives 22 1.2 Liverwort biology 23 1.3 Marchantia polymorpha life cycle 26 1.4 Environmental effects on liverwort growth 33 1.5 Lunularic acid 36 1.6 Liverwort control 39 1.6.1 Cultural and environmental controls 40 1.6.2 Chemical controls 42 1.6.3 Non-herbicidal controls 45 Chapter 2 General procedures and method development 68 2.1 General procedures 68 2.1.1 Identification of liverwort species infesting nurseries 68 2.1.2 Glasshouse liverwort culture 69 2.1.3 In vitro liverwort culture 70 2.1.4 Autoclave procedures 76 2.1.5 Data analysis 76 2.2 Analytical methods 77 2.2.1 Thin layer chromatography 77 2.2.2 Gas chromatography (GC-MS) 78 2.2.3 High performance liquid chromatography (HPLC) system 79 Chapter 3 The effects of environment on liverwort establishment, growth and development 81 3.1 Introduction 81 3.2 Experimental work 83 3.2.1 The effect of light level and temperature on the growth and development of Marchantia polymorpha (25 °C and 15 °C) 83 3.2.2 The effect of light level and temperature on the growth and development of male and female Marchantia polymorpha at 15 °C and 10 °C 95

3.2.3 The effect of lunularic acid on the growth and development of M. polymorpha - HPLC optimisation. 100 3.2.5 Investigation into compounds suitable for use as an internal standard 106 3.2.6 Preliminary investigation into simplifying LNA extraction 109 3.2.7 LNA recovery rates using the simplified extraction method 111 3.2.8 Method development for extraction techniques using TissueLyser 112 3.2.9 The effect of day length and environment on lunularic acid content of liverwort grown in nursery conditions. 116 3.2.10 The effect of shading on liverwort establishment and growth 120 3.3 Discussion 130 Chapter 4 Liverwort epidemiology 134 4.1 Introduction 134 4.2 Experimental section 137 4.2.1 The effect of nozzle, water pressure and nozzle height on gemmae dispersal using an overhead sprinkler system 137 4.2.3 Clumping of gemmae 159 4.2.4 Characterisation of gemma dispersal in nursery irrigation systems 167 4.3 Discussion 181 Chapter 5 The use of glucosinolate hydrolysis products as herbicides 185 5.1 Introduction 185 5.2 General methods 191 5.2.1 Hydroponic growth of plant species 191 5.2.2 Buffer preparation 193 5.3 Experimental work 194 5.3.1 Glucosinolate extraction from L. a/ba seed meal 194 5.3.2 Glucosinolate extraction from L. a/ba plant tissue 195 5.3.3 Extraction and identification of glucolimnanthin hydrolysis products from seed meal using a GSL hydrolysis time course assay, preliminary experiment 199 5.3.4 Optimisation of glucolimnanthin hydrolysis products extraction 201 5.3.6 Scale up of ITC extraction and purificaton 205 5.3.7 Preliminary bioassay to investigate the effect of ITCs on liverwort gemma growth 207 5.3.8 Herbicide bioassays 210 5.3.9 Cress ITC bioassays using Petri dishes 213 5.3.10 Preliminary investigation into ITC collection using Amberlite resin 215

5.3.11 Collection of root exudates 218 5.3.12 Collection of root exudates from plants grown in glass containers 221 5.3.13 Glucosinolate extraction from D. tenuifolia root tissue 222 5.4 Discussion 223 5.4.1 Phytotoxicity 224 5.4.2 Limnanthes a/ba 225 5.4.3 Hydroponics 225 Chapter 6 Fungal antagonists 227 6.1 Introduction 227 6.1.1 Btyoscyphus atromarginatus 228 6.1.2 Phaeodothis winteri 229 6.1.3 Peniclllium vefutinum 230 6.1.4 Fusarium equiseti 231 6.1.5 Trichoderma harzianum 234 6.2 General methods.. 235 6.2.1 Fungal culture on agar media 235 6.2.2 Spore suspension preparation 236 6.2.3 Incubation chamber preparation 236 6.2.4 Liverwort tissue preparation for microscopic examination 236 6.3 Experimental work 237 6.3.1 Preparation of fungal pathogens from dying liverwort samples 237 6.3.2 Preliminary spore germination bioassay 239 6.3.3 Preliminary liverwort inoculation with fungal spores 240 6.3.4 Liverwort inoculation with fungal spores 243 6.3.5 Preliminary liverwort inoculation with fungal agar plugs 245 6.3.6 Liverwort inoculation with fungal plugs 248 6.3.7 Fungal glasshouse experiment 250 Chapter 7 Discussion 278 Chapter 8 Reference list 288 Appendix 1 Glossary 306 Appendix 2 Abbreviations 308 Appendix 3 Suppliers 309 Appendix 4 Anova tables relating to Section 6.3.7.3 310

7 List of figures

Figure 1-1. Barrel pore of Marchantia po/ymorpha. x400 24 Figure 1-2. Rhizoids of Marchantia polymorpha (a) tuberculate with invaginations clearly visible (b) smooth. 25 Figure 1-3. Life cycle of Marchantia polymorpha 26 Figure 1-4. Marchantia po/ymorpha antheridiophores (d) 28 Figure 1-5. Marchantia polymorpha archegoniophores (?) 29 Figure 1-6. (a) Liverwort thallus bearing gemma cups (b) gemma cup bearing numerous gemmae surrounded by mucus 32 Figure 1-7 Structure of lunularic acid 36 Figure 1-8 Molecular diagram of metham sodium 50 Figure 1-9 Structure of glucosinolate. R represents the variable side chain (Mithen, 2001) 50 Figure 1-10 General structure of GSLs and their major hydrolysis products. Adapted from (Vaughn et al., 1996). 51 Figure 1-11 L. alba growing in a hydroponic system 54 Figure 1-12 Ascocarps produced by members of Ascomycetes. (a) Perithecium (b) Apothecium and (c) Cleistothecium. Adapted from (Alexopoulos, 1979) 63 Figure 2-1. Lunularia cruciata found at Hadlow College, Kent (a) growing on a wall (b) detail of crescent-shaped gemma cups (Carey, 2005) 69 Figure 2-2 Chromatography plate layout 78 Figure 2-3 Calculation of the retention factor 78 Figure 3-1. Example of the layout within each Fitotron cabinet. Shelves containing high light treatment (800 pmol m-2 s-1) were positioned higher (nearer to the fluorescent tubes) than those containing the low light treatment levels (400 pmol m-2 s-1). M and F refer to male and female liverwort gemmae respectively 84 Figure 3-2. Gemmaling growth (In) at (a) 15 °C and (b) 25 °C. High light = 800 pmol m-2 s-1, low light = 400 pmol m-2 s-1. Growth (mm) data was transformed using natural logs. 86 Figure 3-3. Growth (In) of male and female gemmalings at temperatures of (a) 15 °C and (b) 25 °C. High light = 800 pmol m-2 s-1, low light = 400 pmol rn-2 s-1. Growth (mm) data was transformed using natural logs. 87

8 Figure 3-4. Relative growth (radial expansion) rate of gemmalings in (a) 15 °C (b) 25 °C. High light = 800 pmol m-2 st, low light = 400 pmol m-2 s-1 88 Figure 3-5 Comparison of liverwort gemmalings after 6 weeks 89 Figure 3-6. Fresh weight of (a) female and (b) male gemmalings at 15 °C and 25 °C. High light = 800 pmol m-2 s-1, low light = 400 pmol m-2 st 91 Figure 3-7. Dry weight of gemmalings after 6 weeks of growth. HL=high light (800 pmol m-2 S-1), LL=low light (400 pmol m-2 St) Temperature=15 °C & 10 °C 92 Figure 3-8. No. of gemma cups produced at temperatures of (a) 15 °C (b) 25 °C. High light = 800 pmol m-2 st, low light = 400 pmol m-2 st 93 Figure 3-9. Growth (radial expansion) of gemmalings after 4 and 6 weeks of growth at 10 °C and 15 °C. HL=high light (800 pmol m-2 s'), LL= low light (400 pmol m-2 s-1) 97 Figure 3-10. Dry weight of gemmalings after 4 and 6 weeks of growth. HL=high light (800 pmol m-2 st), LL=low light (400 pmol m-2 st) Tennperature=15°C & 10 °C 97 Figure 3-11. Fresh weight of gemmalings after 4 and 6 weeks of growth. HL=high light (800 pmol m-2 s-1), LL = low light (400 pmol m-2 st) Temperature=15 °C & 10 °C 98 Figure 3-12. No. of gemma cups present on gemmalings after 4 and 6 weeks of growth. HL=high light (800 pmol m-2 st), LL = low light (400 pmol m-2 s-1) Temperature = 15 °C & 10 °C 99 Figure 3-13. No. of gemma cups mm-2 after 4 and 6 weeks of growth. HL=high light (800 pmol m-2 s-1), LL = low light (400 pmol m-2 st) Temperature = 15 °C & 10 °C.. 100 Figure 3-14. HPLC analysis of lunularic acid using fluorescence detection. Excitation spectrum produced at 400 nm wavelength emission. Emission spectrum produced at 200 nm wavelength excitation. 102 Figure 3-15 HPLC chromatogram produced by lunularic acid 102 Figure 3-16. HPLC chromatograms produced by lunularic acid 103 Figure 3-17. Summary of lunularic acid extraction method 105 Figure 3-18. Structure of 3-hydroxy-2-naphthoic acid (HNA) 106 Figure 3-19. Calibration curve produced for lunularic acid 118 Figure 3-20. Analysis of lunularic acid content of liverwort tissue. Results are an average of three replicates 119 Figure 3-21 Experiment layout. Each treatment is replicated twice within the polytunnel and outside, and provides 73%, 44% and 0 shading. 121 Figure 3-22. Construction of shading structures. Galvanised steel hoops were forced into the ground and clad with shading fabric. 122 Figure 3-23. Shade tunnels (a) outside (b) inside the polytunnel 122

9 Figure 3-24. Examples of liverwort growth showing (a) very dry compost (I-1-73) (b) an area of liverwort die back with new growth (I-0-0) (c) many small, congested plants (I-0-73) (d) sparse liverwort establishment (I-0-0) (e) vigorous growth, (I-I- 44) (f) large gemma cups with many gemmae (I-0-0). Scale bars = 10 cm 124 Figure 3-25. Temperature, relative humidity and light level readings taken from each shade tunnel during the experiment. 125 Figure 3-26. The effect of light level on (a) liverwort growth, expressed as pot coverage, (b) establishment. (c) The number of pots in each treatment containing gametophore-bearing liverwort. I = inside, 0 = outside. 0, 44 and 73 = % shade. 126 Figure 3-27. % pot coverage compared to light levels for inside and outside treatments. Light levels are means of weekly readings for each treatment, taken over the full term of the experiment. I = inside, 0 = outside. 0, 44 and 73 = % shade. . 127 Figure 4-1. Arrangement of collection pots, with 16 pots were arranged in each line 139 Figure 4-2. Equipment layout 139 Figure 4-3. Experiment layout 140 Figure 4-4. Example of liverwort used as gemmae source. 141 Figure 4-5. Dispersal gradients obtained operating nozzles at 3 bar. 144 Figure 4-6. Dispersal gradients obtained operating nozzles at 2.5 bar. 145 Figure 4-7. Dispersal gradients obtained operating nozzles at 2 bar. 146 Figure 4-8. Dispersal gradients obtained operating nozzles at 1.5 bar. 147 Figure 4-9. Water sensitive paper showing droplet stains for the blue nozzle (105 L hr-

1), 2 m height and 2 bar water pressure treatment. 148 Figure 4-10. Droplet size distribution graphs for the blue nozzle (105 L hr 1), 1 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar 150 Figure 4-11. Droplet size distribution graphs for the blue nozzle (105 L hi-1), 2 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar 151 Figure 4-12. Droplet size distribution graphs for the grey nozzle (60 L hr-'), 1 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar 152 Figure 4-13. Droplet size distribution graphs for grey nozzle (60 L hr-'), 2 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar 153 Figure 4-14. Droplet size distribution graphs for the brown nozzle (160 L hr-1) (a) 1 m 2 bar (b) 1 m 2.5 bar (c) 1 m 3 bar (d) 2 m 2 bar (e) 2 m 2.5 bar (f) 2 m 3 bar 154 Figure 4-15. Total number of gemmae collected, excluding initial gemmae count 156 Figure 4-16. Gemma cup replenishment; 3-day and weekly treatments and 4-weekly. 12th May was the initial collection 157

10 Figure 4-17. Linear regression of number of gemmae vs gemmae weight. 95% confidence interval indicates a 95% certainty that the population mean will fall between the limits. 161 Figure 4-18. Fresh weights of small (S) = 26-50 gemmae, medium (M) = 101-125 gemmae, large (L) = 176-200 gemmae. 'Clumps' = groups of gemmae found in a tightly packed mass and dispersed intact. 'Groups' = clumps of gemmae that have been separated and then dispersed 164 Figure 4-19. Dry weights of small (S) = 26-50 gemmae, medium (M) = 101-125 gemmae, large (L) = 176-200 gemmae. 'Clumps' = groups of gemmae found in a tightly packed mass and dispersed intact. 'Groups' = clumps of gemmae that have been separated and then dispersed 165 Figure 4-20. Establishment of small (S) = 26-50 gemmae, medium (M) = 101-125 gemmae, and large (L) = 176-200 liverwort clumps and groups. 'Clumps' = groups of gemmae found in a tightly packed mass and dispersed intact. 'Groups' = clumps of gemmae that have been separated and then dispersed. 166 Figure 4-21. Example drip irrigation arrangement with microtubes connecting the feed pipe with a dripper held in position in each pot with a stake 169 Figure 4-22. Capillary bed construction (a) Layers comprising the base: 1-plywood, 2- black plastic waterproof membrane, 3-capillary matting, 4-perforated plastic (b) Layout of Netafim tape, positioned between layers 2 and 3, and water supply 170 Figure 4-23. Spanish tunnel construction: (a) side view (b) treatment area with pots protected by plastic screens (c) screen attachments using elasticated tarpaulin ties and balls (d) view of general layout 171 Figure 4-24. Layout of treatments showing irrigation design. T=timer, N=nozzle, C=control, CM=capillary bed, D=drip, TWD=twice daily, 2D=2-daily. Dotted lines indicate pipes connecting mains water supply, timers, sprinklers and drippers and capillary beds. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread). 173 Figure 4-25. Layout of treatments showing irrigation design. T=timer, N=nozzle, C=control, CM=capillary bed, D=drip, TWD=twice daily, 2D=2-daily. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread). Dotted lines indicate position of clear plastic screens. The solid line to the right indicates the position of the neighbouring polytunnel. Screens were omitted between treatments where overhead irrigation was not used and was therefore unlikely to affect adjacent plots 174

11 Figure 4-26. Liverwort infestation after 56 days: (a) Liverwort establishment, measured as the number of pots with more than one gemma present. Results for each treatment are the average of forty pots arranged in two blocks. (b) Number of gemmae dispersed. Results for each treatment are an average of the number of gemmae collected in pots during one two day irrigation cycle N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread), D = drip irrigation, C = control (no irrigation), CM = capillary bed irrigation. 2D = irrigation every two days, TWD = irrigation twice daily 176 Figure 4-27. Liverwort infestation after 56 days. (a) Liverwort area. (b) No. of liverwort colonies. Results for each treatment are the average of forty pots arranged in two blocks. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread), D = drip irrigation, C = control (no irrigation), CM = capillary bed irrigation. 2D = irrigation every two days, TWD = irrigation twice daily. 178 Figure 4-28. Characteristic size distributions of droplets produced by the MP Rotator model 1000 (nozzle 1) and Dan modular 180° spread (nozzle 2) nozzles. 180 Figure 4-29. Characteristic size distributions of droplets produced by nozzles. N1 = MP Rotator model 1000 nozzle, N2 = Dan modular 180° spread nozzle, I = block 1, II — block 2, 2D = irrigated every two days, TWD = irrigated twice daily. 181 Figure 5-1. Chemical structure of glucolimnanthin. 185 Figure 5-2. Plant species used in GSL experiments (a) Limnanthes a/ba (b) Diplotaxis tenuifolla (c) Sisymbrium orientale (d) Brassica juncea. 187 Figure 5-3. Chemical structures of ITCs used in bioassays 189 Figure 5-4. Chemical structures of herbicides used in bioassays 189 Figure 5-5. Diagram of hydroponic system. Nutrient solution is pumped from the reservoir by the submersible pump to the end of the NFT channel, flows down the channel over the plant roots and back into the reservoir in the direction of the blue arrows. 192 Figure 5-6. Hydroponics system set up: a) two NFT channels with their nutrient reservoirs, b) NFT channel containing a Brassica juncea seedling in its rockwool block, c) a Limnanthes alba seedling with its block and NFT channel covered in light omitting plastic 192 Figure 5-7. (a) Chromatogram and (b) mass spectrum (EI) of 3-methoxybenzyllTC. Extracted from 50 mg L. a/ba seed meal in 1 ml DCM at pH 7.0 200 Figure 5-8. 3-MethoxybenzyllTC 200

12 Figure 5-9. Preliminary glucolimnanthin degradation curve produced during a time course assay at hourly intervals, pH 7.0, extracted with DCM (1 mL) 201 Figure 5-10. Glucolimnanthin degradation curve produced during a time course assay, samples taken at 20 min intervals, pH 7.0, extracted with DCM (1 mL) 202 Figure 5-11. Degradation curves for 3-methoxybenzyllTC extracted from L. alba seed meal at pH 5.0, 6.0, 7.0, 8.0, 9.0, at hourly time intervals. 50 mg seed meal extracted with DCM (1 mL). Samples were analysed by GC-MS, 50-1 split method 203 Figure 5-12. 3-Methoxybenzylamine. 203 Figure 5-13. Chromatogram of L. a/ba GSL hydrolysis products extracted from seed meal at pH 8.0 for 4 hrs. 50 mg seed meal was extracted with DCM (1 mL). Analysed by GC-MS, 50-1 split method (Section 2.5.2) 204 Figure 5-14. Mass spectrogram (IE) of 3-methoxybenzylamine, GSL hydrolysis product extracted from L. a/ba seed meal at pH 8.0 after 4 hrs, analysed by GC-MS, 50-1 split (Section 2.5.2). 50 mg seed meal was extracted with DCM (1 mL) 204 Figure 5-15. Preliminary investigation into the effect of ethanol on gemmaling growth (radial expansion). Gemmaling areas shown are a percentage of the control. 208 Figure 5-16. Dose-response curves of ITC concentrations and gemmaling growth (radial expansion). Control: treatment with no isthiocyanate or alcohol 210 Figure 5-17. Dose-response curves of herbicides a) lenacil and b) metazachlor and liverwort gemma growth (radial expansion). Control: treatment with no solvent (DMSO or alcohol) or herbicide. 212 Figure 5-18. Dose-response curves of ITCs and cress (Lepidium sativum) radicle elongation. Control: treatment with no alcohol or ITC. 214 Figure 5-19. Dose-response curves of ITCs and cress seed (Lepidium sativum) germination (%). Control: treatment with no alcohol or ITC. 215 Figure 5-20. (a)Chromatogram of the four standard ITCs used and their mass spectra (EI): (b) benzyllTC (c) 2-propenyllTC (d) 3-methoxybenzyllTC (e) 2-phenylethylITC217 Figure 5-21. In-line filter construction 218 Figure 5-22. ITC extraction equipment 219 Figure 5-23. Diplotaxis tenuifolia root exudates chromatograph, analysed by GC-MS, 3- 1 split method (Chapter 2.5.2) 220 Figure 5-24. Analysis of GSL content of Diplotaxls tenuifolia root tissue: chromatographs analysed by (a) total ion current (90-400 atmospheric mass units) and (b) detected by UV absorption at 230 nm wavelength. Chromatogram peaks and mass

13 spectra (EI) refer to GSLs: (i) 4-methylsuiphinylbutyl (ii) p-hydroxybenzyl (iii) 4- methylthiobutyl (iv) 4-methoxy-3-indolemethyl. 223 Figure 6-1. Btyoscyphus atromarginatus, holotype (a) asci with ascospores (b) ascospores in water (c) detail of ascus apical apparatus (scale bars — 10 pm). Adapted from Verkley, 1997. 228 Figure 6-2. Ascospores and ascus tips (a) Didymosphaeria marchantiae (b) Phaeosphaerella marchantiae, neotype B (c) Didymosphaeria schroeteria (d) D. thalictri (e) D. petrakiana . b-f have been reassigned as P. winteri. Adapted from Aptroot, 1995. 229 Figure 6-3 Eupenicillium stolklae (CSIR 1041) (A) Penicilli, (B) conidia, (C) asci producing singly on ascogenous hyphae, (D) asci containing different numbers of ascospores, (E) Ascospores (de Scott, 1975) 230 Figure 6-4. F equiseti (a) conidia (b) conidiophores and (c) chlamydospores (Booth, 1971) 232 Figure 6-5 Trichoderma conidiophore bearing conidia. Adapted from Agrios (1988) 234 Figure 6-6. Incubation chamber containing inoculated liverworts. 236 Figure 6-7. Samples of dying liverwort provided by John Atwood (ADAS) 237 Figure 6-8. Fungal cultures isolated from dying liverwort samples. (a) T harzianum (b) F. Equiseti (c) P. velutinum 238 Figure 6-9. Fungal specimens sourced from CBS: (a) P. winteri CBS 102466 (b) P. winteri CBS 102483 (c) P. winteri CBS 162.31 (d) P. winter/ CBS 182.58 (e) P. winteri CBS 429.96 (f) P. winteri CBS 551.63 (g) B. atromarginatus. 239 Figure 6-10. Liverwort infected with P. velutinum (a) thallus (b) rhizoid. 242 Figure 6-11(a) Liverwort infected with F. equiseti with orange sporodochia (b) Sporodochium releasing macroconidia, found on liverwort inoculated with F. equiseti. (c) F. equiseti macroconidia. Magnification x 400. 242 Figure 6-12(a) Liverwort infected with F. equiseti (b) fungus reisolated from liverwort inoculated with F equiseti 243 Figure 6-13. P. winteri CBS 551.63 (a) Mycelium infecting liverwort thallus. (b) Pycnidia releasing thousands of conidia, (c) Pycnidia on the thallus surface. (a) and (b) Stain = lactophenol tryphan blue. Magnification x 400 246 Figure 6-14. Experimental layout of randomly assigned controls and treatments, each 12 tray set contained within one growth chamber. Block 1=glasshouse 1; block 2=glasshouse 2; C1=control 1, unsterilised compost; C2=control 2, sterilised compost

14 only; C3=control 3, sterilised compost with formulation only; F=fungal treatment applied in formulation. 254 Figure 6-15. Growth cabinet arrangement 254 Figure 6-16 Pre-emergence treatments (a) II-pre pw—f (b) II-pre-fe-f, 5 days post- inoculation. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P. winter', f = fungal treatment. 256 Figure 6-17 Liverwort thallus colonised with F. equiseti mycelium. Pre-emergence treatment (H-pre-fe-f) 14 days post-inoculation. II = block 2, pre = pre-emergence, fe = F. equiseti, f = fungal treatment. 256 Figure 6-18 Microscopic analysis of infected liverwort tissue 5 weeks post inoculation. F. equiseti (a) macroconidia x400 (b) mycelium, x200 and P. winter' (c) and (d) mycelium wrapped around rhizoids x400 (e) spores x400 (f) an unidentified nematode among liverwort rhizoids x400 257 Figure 6-19. Liverwort growth after 11 weeks (a) II-pre fe-C3 (b) II-post pw-C3 (c) area of contamination, I-post pw-C2. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P. winter', f = fungal treatment, C2=control 2 sterilised compost only, C3=sterilised compost with formulation. 258 Figure 6-20 Liverworts with gemma cups infested with fungal mycelium (a) II-post pw- C3 (b) I-pre fe-f (c) II-post pw-f. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P. winter', f = fungal treatment, C3 = control 3, sterilised compost with formulation, 258 Figure 6-21. Areas of liverwort treated pre-emergence with P. winter" 260 Figure 6-22. Areas of pre-emergence treated liverwort after 11 weeks. Control 1 = unsterilised compost only. Control 2 = sterilised compost only. Control 3 = sterilised compost with formulation. Pre Fe = pre-emergence F. equiseti treatment. Pre Pw = pre-emergence P. wintevitreatment. 261 Figure 6-23. Areas of liverwort treated pre-emergence with F. equiseti. Total and healthy liverwort areas represented with two outliers removed 262 Figure 6-24. Growth curve of total liverwort area of liverwort subjected to (a) one and (b) two applications of F. equiseti 264 Figure 6-25. Growth curve of healthy liverwort subjected to (a) one and (b) two applications of F. equiseti 265 Figure 6-26. Growth curve of liverwort dieback, subjected to (a) one and (b) two applications of F equiseti 266

15 Figure 6-27. Total area of liverwort subjected to (a) one and (b) two applications of fungal inoculum, after eleven weeks. Fe=Fusarium equiseti, Pw=Phaeodothis winter': Control 1 — unsterilised compost only, control 2 - sterilised compost only, control 3 sterilised compost with formulation. 267 Figure 6-28. Area of healthy liverwort subjected to (a) one and (b) two applications of fungal inoculum, after eleven weeks. Fe=Fusarium equiseti, Pw=Phaeodothis winter'', Control 1 — unsterilised compost only, control 2 - sterilised compost only, control 3 sterilised compost with formulation. 268 Figure 6-29. Area of dieback of liverwort subjected to (a) one and (b) two applications of fungal inoculum, after eleven weeks. Fe=Fusarium equiseti, Pw=Phaeodothis winteri, Control 1 — unsterilised compost only, control 2 - sterilised compost only, control 3 sterilised compost with formulation. 269 Figure 6-30. Growth curve of total liverwort area of liverwort subjected to (a) one and (b) two applications of P. winteri 271 Figure 6-31. Growth curve of healthy liverwort subjected to (a) one and (b) two applications of P. winteri 272 Figure 6-32. Growth curve of liverwort dieback subjected to (a) one application (b) two applications of P. winteri 273

16 List of tables

Table 1-1. Summary of the epidemiology of Marchantia po/ymorpha 27 Table 1-2 Description of lunularic acid by Valio (1969) 36 Table 1-3 Taxonomy of L. a/ba 54 Table 1-4 ITC content of members of Limnanthaceae (Miller et a/., 1964) 56 Table 1-5 Ascocarp grouping of Ascomycetes (Agrios, 1997) 64 Table 2-1. 'Imperial College' compost media components 69 Table 2-2. Peat content of SHL Professional Potting Compost 70 Table 2-3 Glasshouse environmental conditions 70 Table 2-4 MS and MSMC media preparation 72 Table 2-5 In vitro liverwort cultivation media components 73 Table 2-6. The effect of sterilization treatments on gemma size and survival. 75 Table 2-7 Standard growth cabinet conditions 75 Table 2-8. Autoclave procedures (Haines, 2006; Kandasamy et al., 1996). 76 Table 2-9. Polarity of solvents (Murov, 2007) 77 Table 2-10. Gas chromatography equipment 79 Table 2-11. Gas chromatography methods 80 Table 3-1. Statistical analysis of gemmaling growth by applying the Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interections; data transformed to natural logs. Significant results only are presented. R2 = 0.865 (Adjusted R2 = 0.854) 85 Table 3-2. Statistical analysis of gemmaling fresh weight after 6 weeks analysed by applying a Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interactions. Significant results only are presented. R2 = 0.746 (Adjusted R2 = 0.711) 89 Table 3-3. Statistical analysis of gemmaling dry weight after 6 weeks by applying a Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interactions. Significant results only are presented. R2 = 0.642. 90 Table 3-4. Statistical analysis of no. of gemma cups by applying the Negative Binomial Regression Model, using STATA 8.2, showing effects of treatments and interactions. Significant results only are presented. 92 Table 3-5. Statistical analysis of results after 6 weeks analysed by applying a Univariate General Linear Model, using SPSS 11.0. Only significant results are presented. 96 Table 3-6. UV absorbance spectra for 3-hydroxy-2-naphthoic acid (HNA) 108

Table 3-7. HPLC using 3-hydroxy-2-naphthoic acid (HNA) 109 Table 3-8. HPLC analysis of liverwort gemmae, LNA and SA standards detected using fluorescence 110 Table 3-9. HPLC analysis of liverwort gemmae detected using fluorescence 110 Table 3-10. LNA quantities obtained with five extractions from 25 mg liverwort thallus. LNA was still detected after 5 extractions using this method. Figures relate to peak areas produced by LNA during HPLC analysis, detected by fluorescence. 111 Table 3-11. Appearance of extracts of liverwort thallus, comparing solvent and solvent concentration. G = green, LG = light green, Y = yellow, C = clear 114 Table 3-12. HPLC analysis of LNA extracted from liverwort thallus in methanol and acetonitrile (ACN) 114 Table 3-13. Recoveries for LNA tissue extracted five times with 60% ACN, 0.1% acetic acid 115 Table 3-14. Comparison of HPLC analysis results obtained from selected samples analysed on 16th and 17th November. 115 Table 3-15. Molecular extinction coefficients of LNA in neutral ethanol at given UV wavelengths 117 Table 3-16. Absorbances of 0.1 mg m1-1 LNA in ethanol 117 Table 3-17. Comparison of light levels, temperature and humidity inside the polytunnel and outside for each treatment. Figures are averages of 12 readings with block I and II figures amalgamated. 0 = outside, I = inside, 0, 44 and 73 = % shade 124 Table 3-18. Analysis of variance for inside treatments using logit transformation. 128 Table 4-1. Average distance travelled (cm) by splash droplets containing red dye using the blue nozzle at two different heights 141 Table 4-2. Results obtained from gemma dispersal experiment analysed using log linear regression, providing intercepts and regression coefficients. Nozzles: grey = 60 L hr-1, blue = 105 L hr-1, brown = 160 L hr-1. R2 values relate to transformed data (In). 142 Table 4-3. Maximum distances travelled by gemmae (cm). The brown nozzle (160 L hr-

1) did not operate at 1.5 bar. 143 Table 4-4. Total no. of liverwort gemmae dispersed 148 Table 4-5. Comparison of number of gemmae dispersed with no. of droplets and droplet size. 149 Table 4-6. Nozzle flow rates (L hr 1) at different water pressures 149

18 Table 4-7. Position of replicates. R = red, B = blue, C = 4-weekly. Numbers refer to replicates of each treatment. 155 Table 4-8. Initial gemma collection (12th May) 156 Table 4-9. Analysis of variance comparing gemma cup replenishment of 3-day, weekly and 4-weekly treatments 158 Table 4-10. Analysis of variance comparing gemma cup replenishment of 3-day and 4- weekly treatments. 158 Table 4-11. Weights and predicted number of gemmae in small, medium and large size classes 160 Table 4-12. Experimental design showing size classes and treatment structure. 162 Table 4-13. Completely randomised design of pots on glasshouse bench. 162 Table 4-14. Analysis of variance of fresh weights of gemmae distributed by different application methods Cclumps' and 'groups') in different size classes. 163 Table 4-15. Analysis of variance of dry weights of gemmae distributed by different application methods Cclumps' and 'groups') in different size classes. 164 Table 4-16. Fresh and dry weights and percentage establishment of small (S) = 26-50 gemmae, medium (M) = 101-125 gemmae, large (L) = 176-200 gemmae. 'Clumps' = groups of gemmae found in a tightly packed mass and dispersed intact. 'Groups' = clumps of gemmae that have been separated and then dispersed 165 Table 4-17. Experimental design: all treatments were replicated once in each block, with drip and overhead irrigation applied using two time schedules, every two days or twice daily. 168 Table 4-18. Analysis of variance results of area of liverwort 177 Table 4-19. Mean droplet sizes (pm) produced by nozzle 1 (MP Rotator model 1000) and nozzle 2 (Dan modular 180° spread). 179 Table 5-1. Details of selected plant species with root GSLs listed in order of abundance. The root GSL profile for L. a/ba is unknown (Fahey et a/., 2001; Kirkegaard, 1998). 188 Table 5-2. Genesis Formula nutrient solution components 193 Table 5-3. Fresh and dry weights of L. a/ba tissue. 196 Table 5-4 Solvent gradient conditions used for HPLC analysis. Solvent A = water, Solvent B = 20% acetonitrile 198 Table 5-5. Glucolimnanthin content of L. a/ba tissue. All values given are an average of three samples. Retention time (Re) of glucolimnanthin standard was 17.25 min. 198 Table 5-6. Movement of extract in various solvent ratios during TLC. 206 Table 5-7. Gemmaling diameter (mm) when grown at 25 °C, 400 p mol m-2 s-1 207

19 Table 5-8. Treatments used in preliminary experiment 208 Table 5-9. The effect of ethanol on liverwort gemma growth after 14 days 208 Table 5-10. Estimated ED50s and standard errors of ITCs applied to liverwort gemmae, obtained using probit analysis. ED50 is the effective dose where 50% of the gemma population has an area <1.5 mm2. 209 Table 5-11. Estimated ED50s and standard errors of herbicides applied to liverwort gemmae, obtained using probit analysis. ED50 is the effective dose where 50% of the gemma population has an area <1.5 mm2. 211 Table 5-12. ITC weights used for standards 219 Table 5-13. Quantities of ITCs extracted from root exudates o D. tenuifolia 220 Table 6-1. Taxonomic details of fungal specimens (Index Fungorum, 2004) 233 Table 6-2 Components of media used for fungal culture. Supplier: Sigma 235 Table 6-3. Fungal strains sourced from Centraalbureau voor Schimmelcultures 238

Table 6-4. Germination bioassay of spores suspended in water and PD broth. ✓ = spores germinated; x = germination failed 240 Table 6-5. Concentration of fungal spores used in liverwort inoculation 241 Table 6-6. Results of spore germination bioassay. 243 Table 6-7. Spore densities applied to liverworts 244 Table 6-8. Spore germination bioassay inoculation. 244 Table 6-9. Levels of collapse of liverwort 21 days post inoculation.. 244 Table 6-10. The effect of fungal plug inoculation of liverwort sections, 8 days post inoculation. Figures refer to the number of plugs affected 247 Table 6-11 The effect of fungal plug inoculation of liverwort sections, 11 days post inoculation. Figures refer to the number of plugs affected. 249 Table 6-12 Controls and fungal treatments 250 Table 6-13. Fungal species used for pre and post-emergence treatment; inoculum quantities used for pre-emergence treatment 251 Table 6-14 Post-emergence treatment fungal inoculum concentration 252 Table 6-15. Post-emergence treatment fungal inoculum concentration 252 Table 6-16 Record of successful reisolation of fungi from fungal treatments. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, Fe = F. equiset4 Pw — P. winter': 256

20 Introduction and literature review Chapter 1

Chapter 1 Introduction and literature review

1.1 Introduction Weeds, liverworts and mosses growing on the surface of container plant compost are a major problem to the horticultural industry. The definition of a weed is any kind of plant growing in an undesirable place and the main difficulties weeds pose to the horticulture industry are competition with the crop plant for resources, they can act as alternative hosts for pests and diseases and they negatively affect the aesthetics of container plants, reducing their saleability (Adams et al., 1998). The removal of liverwort, moss and weeds from pots is estimated to cost the horticultural industry 4% of total production costs, predominately in labour costs and equating to £13 million annually, (Scott and Hutchinson, 2001). These figures, combined with zero tolerance by the increasing number of accreditation schemes which demand good quality, weed- free plants for sale, and government withdrawal of chemical approvals, are putting huge pressure on the industry to find an alternative method of control.

The attributes of successful weeds, ranging through algae, mosses, liverworts and higher plants, include their ability to out-compete both crop plants and native vegetation. They often germinate faster and over longer periods, and seeds, spores or vegetative propagules are often small and produced in prolific numbers. Many also

21 Introduction and literature review Chapter 1 reproduce by vegetative propagation, and these are often particularly difficult to eradicate (Adams et al., 1998).

Liverworts (Hepaticae) and mosses (Musci) are both members of the division Bryophyta and form low spreading mats of vegetation that affect both protected and hardy nursery stock. Liverworts are most prevalent in the warm, moist conditions provided in propagation systems, where plants are at their most vulnerable. Liverworts compete with the crop for water, nutrients and light, prevent irrigation water penetration, and manual removal can cause damage to crop roots (Adams et al., 1998). Chemicals currently in use have a short-lived effect, multiple applications are necessary, and they are anathema for those striving to achieve greener, eco-friendly lifestyles, growers and customers alike (Scott and Hutchinson, 2001).

Liverworts are commonly referred to as being 'thalloid' or 'leafy' types, the two extremes of liverwort form. 'Thalloid' liverworts tend to have a flat dorsiventral 'ribbon- shaped, dichotomising' form whilst 'leafy' liverworts have a 'cylindrical axis' or stem with associated `dorsiventral appendages' or leaves. The majority of leafy liverworts are classed as members of the order Jungermanniales (Watson, 1971).

1.1.1 Main aims and objectives

Overall aim The overall aim of the research described in this thesis is to provide new biological knowledge on Marchantia polymorpha, the most common liverwort infesting plant nurseries and to investigate novel control methods using natural products, with an emphasis on identifying areas that could be used to target future research and improve current control systems. It was assumed that all material used was M. polymorpha subsp. ruderalls.

Objectives ■ Review current knowledge of the biology, life cycle and epidemiology of liverwort. • Investigate aspects of liverwort biology, life cycle and epidemiology currently lacking and the effects of environment on liverwort growth and development.

22 Introduction and literature review Chapter 1

• Characterise the epidemiology of liverworts, considering how they spread to infest new areas, identifying areas where changes in common nursery practice could reduce or prevent spread. ■ Investigate natural control methods: GSL hydrolysis products and fungal antagonists. ■ Use the knowledge gained to perform large scale experiments to aid the design of an integrated liverwort control program.

1.2 Liverwort biology In this section the structure, life cycle and epidemiology of M. po/ymorpha is briefly described. A wealth of detailed information is available from studies carried out at the end of the 19th century and early 20th century as M. po/ymorpha formed an important part of classroom botany studies (Durand, 1908).

Structure The thallus has a two-dimensional dorsiventral form with a broad laminar surface for maximum interception of light, approximately 30 cells thick at the midrib and 10 at the edges (Raven et al., 1999). The upper (dorsal) region is rich in chlorophyll and the lower ventral region colourless, bearing rhizoids and usually two rows of ventral scales. These scales are a series of small 'leaves', one cell thick, on either side of the mid-line and frequently deep purple. They are found close behind the apex and lie over and protect the apex, which contains groups of apical initials from which new cells are produced (Watson, 1971).

There is a hexagonal pattern on the upper (dorsal) surface, each 'mesh' of which has a central pore. In Marchantia these are 'barrel pores' which open into underlying photosynthetic chambers, the supporting walls of which form the hexagonal mesh structure seen on the surface (Round, 1969). Barrel pores (Figure 1-1) are open to the air, which penetrates the epidermis and have a cutinised and water repellent rim (Proctor, 1984). These structures are air chambers that permit the exchange of gases, and are lined with chains of oval-shaped chlorophyll-rich photosynthetic cells rising from the base and culminating in colourless flask-shaped cells, arranged so that at least one surface of every photosynthetic cell is exposed to the air (Watson, 1971). The tissue below the photosynthetic zone has few chloroplasts (Round, 1969). Specialised cells surround each pore, arranged in 4-5 circular layers of four cells each

23 Introduction and literature review Chapter 1 which usually protrude into the chamber, but become juxtaposed when dry, reducing the rate of water loss, and then separate in moist conditions (Raven etas., 1999).

Carbon dioxide availability is increased by direct exposure of the photosynthetic cells in the barrel pores to airspaces within the thallus, reducing the length of the relatively slow aqueous diffusion path. Each cell contains a large vacuole, so the chloroplasts are located in the thin zone of cytosol pressed against the plasma membrane, again reducing the aqueous diffusion path (Hopkins, 1999).

Figure 1-1. Barrel pore of Marchantia polymorpha. x400

Barrel pores create problems regarding water loss as they are permanently open and can therefore only survive in moist conditions. Water loss is reduced by a water impermeable cuticle; rhizoids also play a role, absorbing water and thereby stabilising the water potential of cells (Hopkins, 1999).

Bryophytes, unlike vascular plants have developed the ability to undergo severe drying out and recover immediately water is applied, an adaptation to intermittent water supply (Watson, 1971). According to Schonherr and Ziegler (1975) it is commonly assumed that members of Bryophyta do not have a cuticle, however the surface is difficult to wet due to wax globules and a non-cellular layer, which can be stained with Sudan III, can just be detected on the surfaces of thalloid liverworts, suggesting the presence of lipids (Martin and Juniper, 1970).

Marchantia lives in damp conditions, reducing the risk of desiccation that could occur due to the lack of a vascular system and therefore has a requirement for water; however, too much water entering the thallus must be prevented. Barrel pores are an

24 Introduction and literature review Chapter 1 entry point and have developed hydrophobic ledges which reduce the entrance size in a similar manner to an iris shutter, restricting entry of liquids into the air pore (Schonherr and Ziegler, 1975). Growth habit also reduces water loss as the overlapping of thalli introduces mutual shading, the trade-off for which is a decrease in photosynthetic efficiency. The prostrate growth habit also reduces water loss above and around the structure compared with a plant held above the ground and exposed to air movement, and also leaves the liverwort in close contact with the moist ground (Hopkins, 1999).

Figure 1-2. Rhizoids of Marchantia polymorpha (a) tuberculate with invaginations clearly visible (b) smooth.

Members of the Marchantiales have two types of rhizoids, smooth- and tuberculate- walled (Figure 1-2). Smooth-walled rhizoids penetrate the soil to anchor the plant and absorb water from it. Tuberculate rhizoids are lined with invaginations or pegs and are of a similar diameter as smooth-walled rhizoids and lie parallel to the underside of the thallus. Their external walls form a capillary conducting system, transmitting water to the absorptive tissues (Smith, 1955); although the ventral surfaces of the thallus are wettable, only the rhizoid walls (smooth and tuberculate) absorb water (McConaha, 1941), however the significance of the projections on the inner surface is unknown (Smith, 1955). Duckett et a/ (2000) suggests that smooth rhizoids act as an entry point to the inner thallus cells for endophytic fungi. M. po/ymorpha differs from other liverworts as the rhizoids are in localised areas over the whole ventral surface of the thallus (McConaha, 1941), and also grow down the grooves in the gametophore stalks (Bell, 1992).

25 Introduction and literature review Chapter 1

1.3 Marchantia polymorpha life cycle M. polymorpha is a dioecious hepatic that has developed a complex life cycle (Figure 1-3) with a variety of mechanisms to ensure its continued spread and establishment, providing temporal, environmental and spatial solutions to potential limitations (Table

1-1)

ASEXUAL SEXUAL REPRODUCTION REPRODUCTION Gemma Cups <=1 <=3 Dioecious Gametophyte (n)

Vegetative Fragmentation Spores Propagules II (n) Archegoniophores Antheridiophores Gemmae Dioecious (n) (n) (n) Gametophyte Meiosis (n) Spore Bearing Layer (2n) Archegonia Antheridia (n) (n)

8 Gamete (n) Gamete (n) Capsule 9 (2n) Zygote (2n) I Fertilisation Sporophyte (2n)

Figure 1-3. Life cycle of Marchantia polymorpha.

Sexual reproduction M. polymorpha is dioecious with gametogenous antheridia (d) and archegonia (y) on stalked antheridiophores (Figure 1-4) and archegoniophores (Figure 1-5), respectively. Fertilisation occurs prior to the elongation of the archegoniophores stalk and is followed by sporophyte formation (Rook, 1999). Gametophores of both sexes are highly modified thalli and, although different shapes, have the same composition with different cell types found only in the archegonia and antheridia.

Introduction and literature review Chapter 1

Table 1-1. Summary of the epidemiology of Marchantia polymorpha

Gemmae are produced in larger numbers during short day conditions Time of Year (Voth, 1943). Gametophores are produced during long days (Wann, 1925). Gemmae proliferate in the cool, moist Temporal conditions prevailing during winter months, their dispersal aided by water. Environment Spores are dispersed during warmer, dryer weather, with the aid of hygroscopic elaters that propel spores into the air (Watson, 1971). Fragmentation increases the number of plants in the immediate vicinity of the parent plant (Cavers, 1903a). Gemmae are vegetative propagules Asexual Spatial produced by gemma cups (splash (Reproduction Methods) cups) and dispersed by water droplets over distances up to 60cm (Brodie, 1951). Spores produced provide a long Sexual distance dispersal mechanism.

The antheridiophore Antheridia are simple sac-like structures with an outer sterile wall one cell thick, and a stalk consisting of a few cells and contain antherozoids, mother cells that mature into male gametes (Round, 1969). When approaching maturity antheridia contain small cubical cells (spermatocytes) which develop into spermatids and subsequently biflagellate gametes (spermatozoids). This development involves the spermatid ultimately becoming elongated and coiled, with a blepharoplast appearing in the cytoplasm from which two flagella grow out, providing the sperm with a degree of motility (Watson, 1971).

Introduction and literature review Chapter 1

1 .4--- Antheridiophore

Figure 1-4. Marchantia polymorpha antheridiophores (d)

Antheridia occur in sunken pits on the top of the antheridiophores. When mature the wall cells of the antheridium become distended and their pressure on the chamber walls pushes on the mass of spermatozoids within. A large amount of mucilage is formed, the cap cell bursts and the male gametes are released, although still enclosed by delicate membranes of their mother cells (Watson, 1971). The spermatozoids exude in a milky mass from canals of the antheridial chambers onto the top disc of the antheridiophore. They are free swimming once released, and positively chemotactic (Smith, 1955). Water droplets then disperse the sperm. The disc at the top of the antheridiophore is flat and saucer-like with edges flared upwards, forming a splash dispersal mechanism comparable to the cup-shaped gemma cups (Brodie, 1951), which increases the spore dispersal distance; distances up to 49.6 cm have been observed (Wyatt and Anderson, 1984). Maturing sperm can be seen when the disk is about 3 mm in diameter with a reddish-brown area in the centre indicating maturity (Anthony, 1962).

Sperm can move short distances using their flagella in a film of water that may bridge between the ripe antheridia to the region of the archegonia, but only very small distances. Sperm can remain motile for up to 6 hrs, and the distance they can swim in that time is usually less than 1cm (Wyatt and Anderson, 1984). Sperm may also may also be transferred with microfauna. Mites, springtails, flies, aphids, spiders have been observed having mucilage with large numbers of motile sperm on their bodies after visiting the male heads of ferns (Wyatt and Anderson, 1984), mosses and liverworts (Watson, 1971).

28 Introduction and literature review Chapter 1

The archegoniophore The archegoniophore develops by the upward bending of a branch of the thallus. Species of Marchantia have two furrows with scales and rhizoids on the ventral side of the stalk, which divides into two at the swollen apex, and then divides again to form rays on a bilaterally symmetrical rosette-like 8-lobed disc, and this is diagnostic of members of the genus. The apex of the archegoniophore distinguishes different genera and has similar morphology to the thallus with pores, assimilating chambers and scales, with rhizoids beneath (Smith, 1955).

Fertilisation of the mature archegonium occurs when it is only slightly raised above the thallus. The stalk continues to elongate and the central area of the disc swells upwards, inverting the edges of the disk so the archegonia necks point downwards and older archegonia are found at the edge of the disk. A tissue then develops on both sides of each receptacle and each group of archegonia becomes surrounded by a 2- part curtain-like involucre (Smith. 1955).

Archegoniophore Sporophyte bearing yellow spores

Stalk

Figure 1-5. Marchantia polymorpha archegoniophores (y)

A dorsal furrow acts as a capillary tube, with enough water to allow the antherozoids to swim down the neck canal of the archegonium to the egg. Several may enter but only one fertilises the egg (Smith, 1955).

Archegonia are formed from superficial cells behind the apex and are usually protected by an upgrowth of surrounding tissue (Round, 1969). They are flask-shaped with a wide, rounded venter containing the egg and ventral canal cell and a long, narrow neck enclosing a core of protoplasmic neck canal cells. There is a single, non-motile

29 Introduction and literature review Chapter 1 female gamete. When archegonia are ready for fertilisation the neck canal cells and ventral canal cell degenerate into a slimy mass and the secretion of chemicals attracts motile male gametes down the neck canal to the egg (Watson, 1971). In liverworts the attractant is the protein albumin (Doyle, 1973). The diameter of the egg nucleus far exceeds that of the male gamete. A number of spermatazoids enter each archegonium (Watson, 1971).

The embryo divides periclinally to produce a capsule differentiated into amphithecium and endothecium (outer and inner layers, respectively). The amphithecium is one cell thick and develops ring-like thickenings on the inner walls of cells so it splits longitudinally into segments at maturity (Smith, 1955).

The sporophyte The sporophyte is comprised of the foot (forming the base of the sporophyte), the seta (stalk) and the capsule (containing spores and elaters). The foot provides anchorage and absorbs nutrients from the gametophyte (Watson, 1971). Chloroplasts occur in the capsule wall cells, so they do not depend entirely on the gametophyte for nutrients (Bold, 1938). The seta holds the spores proud of the archegoniophores to facilitate dispersal. The capsule is oval shaped with a cuticle to prevent excessive water loss (Watson, 1971). Archegoniophores are ultimately taller than antheridiophores, holding the sporophyte clear of liverwort structures, where it is exposed to air currents for dispersal (Anthony, 1962).

For protection against desiccation and damage, members of Marchantia have a calyptra, perianth and involucres (pseudoperianth), with an individual perianth around each archegonium which divides into 4 to 16 strips which sometimes remain joined at the tip forming a slit through which spores disperse (Watson, 1971).

The endothecium divides to form sporogenous tissue (archesporum), often becoming sterile at the apex. A proportion of sporogenous cells are vertically elongate and divide to form cubical sporocytes, while others elongate into elaters with spiral wall thickenings; there are approximately 128 spores to 1 elater. The average number of spores is 300,000 per capsule for Marchantia (Smith. 1955), with 7 million spores produced per plant (24 capsules, each producing 300,000 spores) (O'Hanlon, 1926).

30 Introduction and literature review Chapter 1

Spores of Marchantia are actively discharged due to strong hygroscopic movement of elaters. The capsule is suspended below the archegoniophores on a short seta and the wall of the capsule splits at the apex, rolls back and exposes the yellow spores and elaters, which are straight at this point. These then begin to dry, become distorted, coiled, and entangled until the movements produced expel all spores into the air, until only the elaters remain (Ingold, 1974).

Once dispersed and in favourable conditions the spore germinates to form a generally small, inconspicuous and short-lived filamentous protonema. The spore wall splits to expose the contents and following cell division forms a mass of 6-8 cells, possibly two broad at the apex, of undifferentiated parenchymatous tissue. A cell at the edge differentiates to form an apical meristematic cell, which in turn divides to give two faces and eventually produce a row of apical initials (Smith, 1955). Subsequent divisions form the recognisable dorsiventral gametophyte thallus (Rook, 1999; Round, 1969) and differential growth of these cells causes notches to appear in the apical region when the gametophyte consists of 30-40 cells (O'Hanlon, 1926).

In this study the epidemiological work concentrated on vegetative reproduction. The liverworts grown in the glasshouse did not produce gametophores early in the project, and little detailed information was available relating to vegetative reproduction in relation to the effect of glasshouse irrigation on gemma dispersal. Sexual reproduction is important in nursery systems as a means of providing an epidemiological mechanism during hot summer conditions when gemma production is reduced. Higher temperatures within glasshouses compared to outside during the winter can extend the time frame during which sexual reproduction may occur, thus this is an area requiring further research.

Vegetative reproduction There are three methods of vegetative reproduction used by M. polymorpha: fragmentation (older portions of thallus die back leaving separate pieces of younger tissues), adventitious shoots (the separation of buds from the main axis of thallus) rarely seen in Marchantia, and vegetative propagules (gemmae) (Cavers, 1903a).

According to Cavers (1903b) regeneration of M. polymorpha was first described by Necker in 'Physiologia Muscorum 1774. M. polymorpha produces outgrowths from

31 Introduction and literature review Chapter 1 plant fragments capable of producing a new plant and in favourable conditions each cell of the liverwort is capable of producing a complete new plant, except for antheridal cells (Cavers, 1903b).

Liverworts reproduce asexually by vegetative propagules (gemmae)(Figure 1-6) produced by gemma cups (circular structures found on the upper surface of the liverwort thallus), and are released when water droplets splash into the cup, and break up into droplets which transport gemmae away from the parent plant. The gemma cups of M. po/ymorpha have sides at an angle of 60-70° to the horizontal and together with lentil-shaped gemmae they form an efficient splash-cup mechanism. Water droplets falling into the gemma cup displace water, thrusting gemmae upwards along the cup sides. Dispersal distances of up to two feet by small raindrops have been observed (Brodie, 1951).

Figure 1-6. (a) Liverwort thallus bearing gemma cups (b) gemma cup bearing numerous gemmae surrounded by mucus

Gemmae are lentil-shaped, bilaterally symmetrical and several cells thick in the centre with two shallow indentations containing a row of apical cells from which new plants are formed (Smith, 1955). Each gemma develops from a single cell at the base of the gemma cup and when mature is attached by a single-celled stalk. Mucilaginous glands (long club-shaped hairs) grow up from the base of the cup between the gemmae, (Cavers, 1903a) which are thought to imbibe water and swell breaking the gemmae from their stalks and forcing them out of the base of the cup into the upper section in preparation for dispersal (Round, 1969).

Both sides of the gemmae are morphologically and physiologically identical. During germination rhizoids appear first, on the underside, and growth of the thallus is then

32 Introduction and literature review Chapter 1 initiated at two meristematic apical initials on the gemma, forming a dichotomous branching pattern (Taren, 1958). Water is required for germination but not nutrients. Gemmae do not usually germinate in the dark unless previously sensitised to light (during experimentation, by adding auxin to the substrate). High temperature promotes germination (Taren, 1958). There are conflicting reports of the effect of pH on liverwort growth. When comparing gemma growth under pH conditions between 1.65 and 7.30, Taren (1958) found poor growth below pH 3.5, but above that (and below 7.30) no effect was observed. However, Clemens et at (1991) found that by increasing compost media pH to approximately 6.5 using dolomite (10g/1) liverwort coverage was reduced from 85-100% in the control to 65%.

Available literature describing Marchantia reproduction and establishment is often concerned with bryophytes in their natural habitats rather than the artificial environment found within plant nurseries, where liverworts are exposed to overhead irrigation systems, designed to provide even distribution of water to the whole bed or tunnel. Characterisation of the effect of overhead irrigation on gemma dispersal would provide an insight into the effective design and use of irrigation regimes as part of an integrated management system to reduce liverwort infestation.

1.4 Environmental effects on liverwort growth There is limited literature concerning the effect of light intensity on liverwort growth and development. Terui (1981) found that medium light intensity (3500 lux) produced maximum vegetative growth, but high light (5000 lux) suppressed growth, and suggests that photosynthesis and growth were saturated at 3500 lux, and inhibited by an excess of light at 5000 lux (Terui, 1981).

With Lunularia cruciata, a close relative, growth is reported to slow under long day lengths and stops in continuous light, a condition that is reversible, with growth resuming in short-days (Nachmony-Bascomb and Schwabe, 1963). There is an effect due to photoperiod on liverwort growth and with significantly greater thallus length (Carter and Romine, 1969; Hedger et al., 1972; Voth and Hamner, 1940), dry weight and fresh weight of Marchantia polymorpha gemmalings (young, recently germinated gemmae as opposed to mature thallus) produced in long days than short-days (Voth and Hamner, 1940).

33 Introduction and literature review Chapter 1

Mache and Loiseaux (1973) found an inhibiting effect of light intensities above 6000 lux on liverwort growth. Morphological changes occurred in the thallus, which became thicker, brittle and curved downwards into the medium, and in the structure of chloroplasts where small grana (stacked thylakoid membranes, the site of photosynthesis) connected by fret membranes were replaced by continuous grana

Effect of environmental conditions on gemma cup production In experiments using Lunularia cruciata (L.) Dum. it was found that thalli produce more gemma cups in high light intensities, but only if accompanied by low temperatures (Nachmony-Bascomb and Schwabe, 1961). They also found little effect due to light intensity (1350 lux and 247 lux) on thallus growth or branching, or gemma cup production, with an average of 0.025 cups mm-2 of new growth.

More gemmae cups were produced per unit area on plants grown in short days in several experiments using Marchantia po/ymorpha (Carter and Romine, 1969; Voth, 1943; Voth and Hamner, 1940). Similarly, for L. cruciata more gemma cups were produced at short-day lengths, at 18 and 24°C (Schwabe and Nachmony-Bascomb, 1963). Conversely, Terui (1981) found that for M. polymorpha gemma cup production was promoted under long day lengths.

An effect due to temperature is reported, with more gemma cups produced in lower temperatures 0.43 cups per new thallus branch at 12°C compared with 0.11 cups at 18°C for Lunu/aria cruciata. Temperatures above 24°C affected growth detrimentally, resulting in fewer new branches that are also shorter and smaller overall (Nachmony- Bascomb and Schwabe, 1961; Nachmony-Bascomb and Schwabe, 1963).

Experiments reported by Dickson (1932) indicate dominance by the apical cells which decreases as distance from the apex increases, dominance of one area of tissue over another depending on their relative age, with younger cells dominating older cells. This dominance may be affected by environmental conditions such as light and heat. The inhibiting influence of younger cells is not transmitted across dead cells.

34 Introduction and literature review Chapter 1

Nutrition and gemma cups Voth (1941) conducted experiments using separate male and female M. polymorpha thalli grown under comparable conditions. More gemma cups were produced on male than female plants, >6:1 across all nutrient treatments. Female thalli had fewer gemma cups, a broader thallus tip, and a smoother surface than male thalli when grown in the conditions of this experiment. The margins of the male plant were more undulating and curved downwards, particularly under high pH conditions (Voth, 1941).

The effect of environment on gametophore production M. polymorpha responds to long day lengths by producing sexual structures, gametophores. When exposed to long day length (day lengths were extended using tungsten lighting) during winter mature antheridiophores are formed in 3-4 weeks, mature archegoniophores within 6-8 weeks and mature sporophytes within 10-12 weeks. These processes are quicker in relatively high humidity, while low humidity appears to retard or inhibit sexual development, especially in archegoniophores (Wann, 1925).

However, when growing liverwort in 6, 8, 14, 16, 18 and 24 hr day length, Benson- Evans (1964) found that gametophores were produced in all treatments, with 16 hrs the optimum, producing 90 (18 and 24 hr days produced 84.7 & 87 respectively). The response time to produce these was 18-20 days (male) and 21-28 (female) in 16 hr days (Benson-Evans, 1964).

Temperature also has an influence, with gametophores produced at 21 °C but not 10 °C with an 18 hr photoperiod, and with no gametophores produced at either temperature with a 6 hr photoperiod (Benson-Evans, 1964).

Effects of light on spore germination Early research gave an indication that there is a minimum light requirement for spore germination with no germination in complete darkness. At low light levels germination was retarded and caused abnormal morphology, as a filament of cells with little chlorophyll was produced that did not develop into the common gametophyte form (de Forest Heald, 1898). Low light intensity retards cell division; high light intensity and long photoperiod increases the frequency of cell division, particularly when the protonema filament is three or four cells long. A marginal row of meristematic cells is

35 Introduction and literature review Chapter 1 soon established (O'Hanlon, 1926). More recent research confirms that a minimum of 10 hrs of light is necessary for germination, with the entire spectrum (near UV to red) effective, although red light was most effective (Nakazato et al., 1999). Intermittent 15 min pulses of red light every one or two hrs for 24 hrs was also found to induce germination.

1.5 Lunularic acid Lunularic acid (LNA) is an endogenous plant growth regulator involved in growth inhibition (Pryce, 19714 It is a natural dihydrostilbene (Pryce, 1972) identified as 3, 4'-dihydroxybibenzyl-2-carboxylic acid (Figure 1-7) (Gorham, 1977). LNA was first extracted from the liverwort L. cruciata and characterised (Table 1-2) by Valio (Valio, 1969; Valio et al., 1969) and the presence of lunularic acid in M. polymorpha was established by Pryce (1971b)

OH OH

Figure 1-7 Structure of lunularic acid.

Lunulana cruciata becomes dormant in high temperature, long day conditions, effectively drying and preserving the liverwort until conditions improve (Valio et al, 1969). Temperatures of 24 °C and continuous light induce dormancy at around 6 days; growth recommences within 3-4 days of applying short-day conditions.

Table 1-2 Description of lunularic acid by Valio (1969)

Characteristic Description Appearance Pale yellow needles (2mg) Melting point 192 °C Molecular weight 258 Suggested molecular formula Ci5H1.404

36 Introduction and literature review Chapter 1

A leachable inhibitory factor was postulated that accumulates within the liverwort, and becomes effective above a critical level. Light interruption (of night) also induces growth inhibition, suggesting a photoperiodic response; every part of the thallus is sensitive to photoperiod. With continuous light, either the production of the inhibitor could be increased or the rate of loss reduced. L. cruciata was found to enter dormancy in response to photoperiod only if the temperature was raised to 30°C (Schwabe and Nachmony-Bascomb, 1963). LNA has been detected in all parts of 111. polymorpha, including the rhizoids, with the highest concentrations found at the thallus apex (Gorham, 1977). Growth inhibition by LNA is directly related to concentration over the range of 0.1 to 10 ppm; very high concentrations are lethal (Valio and Schwabe, 1970).

Intact thalli (Gorham, 1975) and cell suspensions of M. polymorpha (Abe and Ohta, 1983) exhibit the same relationship between growth and lunularic acid content in response to day length as that demonstrated in Lunu/aria. There is also an effect of light intensity on LNA accumulation with higher concentrations of LNA found in liverwort species grown in light intensities of 5600 lux than 560 lux (Gorham, 1975).

Taren (1958) had suggested that a substance present in gemma cups and undifferentiated liverwort thallus inhibited germination of gemmae within the cup. Dormancy can occur in short days if leaching of LNA is prevented, allowing it to accumulate and effect growth reduction even when liverwort is actively growing. This may help to explain the arrest of gemma growth within gemma cups, where LNA may accumulate and be unable to leach away (Valio and Schwabe, 1970).

Regulation of LNA and growth inhibition is not fully understood. Both Valio (1969) and Pryce (1972) have suggested the phytochrome-mediated photoperiodic regulation of lunularic acid production as a growth control in liverwort.

A change of LNA content dependent on the cell growth cycle has also been recorded. Low LNA content during the initial exponential cell growth phase increased rapidly, coincidental with the reduced cell growth rate as the stationary growth phase began. Increased LNA content was also correlated with the reduction of media phosphate content over time. This 'antagonistic regulation between primary and secondary metabolites' is suggested as a means of LNA control (Abe and Ohta, 1983).

37 Introduction and literature review Chapter 1

Exogenous application of LNA at the same concentrations as is present endogenously (up to 600 pg g-1 fresh weight) would inhibit photosynthesis, but this was not confirmed as the mechanism of growth control (Gorham and Coughlan, 1980). This high LNA content was clarified by the discovery of pre-lunularic acid (pre-LNA), a direct precursor of LNA that readily converts to LNA in acid or basic conditions, inflating the level of LNA present in liverwort (Ohta et aL, 1983; Ohta et al., 1984). It is also suggested that pre-LNA acts as a regulator of LNA levels: a reservoir of pre-LNA is retained that can be rapidly metabolised to LNA as required (Abe and Ohta, 1984). The control mechanism for LNA has not yet been fully established.

The proposed biosynthetic pathway of marchantins incorporates the production of lunularic acid, lunularin and prelunularic acid from L-phenylalanine (Friederich et al., 1999a; Friederich et at., 1999b; Pryce, 1971a). Marchantins are macrocyclic bis- bibenzyls, secondary products of liverworts that exhibit antifungal, antimicrobial, muscle relaxing and cytotoxic activities (Asakawa, 2001). Several research groups have successfully synthesised LNA (Arai et al, 1972; Arai et al, 1973; Bracher et al, 2000; Eicher et al, 1988; Furstner and Nikolakis, 1996; Hashimoto et al., 1988; Huneck and Schreiber, 1977)

Lunularic acid, however, does not inhibit growth of Lunularia or Marchantia gemmalings (young, recently germinated gemmae) or cress roots any more than other similar compounds and is less effective than lunularin, pinosylvin and several other naturally occurring stilbenoids and analogues of LNA (Gorham, 1975; Nakayama et al, 1996). Gorham's investigations into the effects of analogues of LNA on plant growth identified 3-hydroxy-4'-methoxybibenzyl and 4-hydroxybibenzyl as the most active of those tested (Gorham, 1978).

LNA has been reported to inhibit growth in higher plants: water cress (Nasturtium officinale), timothy grass (Phleum pratense) (Nakayama et at., 1996), lettuce hypocotyls (at concentrations normally toxic to Lunularia gemmalings), and rice coleoptiles. LNA also showed 100% inhibition of germination of rice seeds at concentrations above 300ppm (Hashimoto et at., 1988). Germination tests using Lactuca sativa and Lepidium sativum indicated that seeds that would not germinate in a solution of LNA did germinate when it was washed off, thus the inhibition caused

38 Introduction and literature review Chapter 1 dormancy rather than death (Yoshikawa et al., 2002). Oat coleoptile sections showed no inhibition or promotion of growth due to the inhibitor (Arai et al., 1973).

An allelopathic effect on other plants could be suggested by the inhibitory effect of LNA on lettuce hypocotyl growth and cress seed germination. However, this would require accumulation of greater concentrations of LNA in the growing medium than is normally produced (Valio and Schwabe, 1970). Gorham was unable to find any trace of lunularic acid in either agar or compost media, concluding that any lunularic acid would quickly be taken up and inactivated by the liverwort (Gorham, 1975).

1.6 Liverwort control Research in the UK and America provides important information regarding chemical and cultural liverwort control (Altland, 2003; Atwood, 2005; Scott and Hutchinson, 2001).

Liverwort is becoming a major weed problem in some areas of the US, spreading to cover many states in recent years. The studies referred to in this thesis provide a comprehensive overview; however it should be recognised that success with herbicides is variable across the US due to differing climatic and growing conditions and results may differ under UK growing conditions (Atwood, 2006). Comments relating to US herbicides have been limited to those in current use that are most successful against liverwort. Few reports on liverwort control have been found from Europe, including the UK.

As for weeds generally, the most successful management system would prevent liverwort germination, establishment and dissemination, using a combination of nursery hygiene, cultural practices and correct application technology necessary for adequate control.

39 Introduction and literature review Chapter 1

1.6.1 Cultural and environmental controls

Nursery Hygiene Herbicide programmes are more effective where weed inoculum is reduced; a high level of general nursery hygiene helps to reduce liverwort infestation (Altland, 2003).

Empty glasshouses, polytunnels and outside growing areas should be cleaned and kept weed-free to reduce liverwort infestations that may spread to other areas. Ground areas between pots and non-cropping areas also need to be maintained in a clean and weed-free state; these areas often become covered with plant and compost debris, on which liverwort spores and gemmae can settle and germinate. Covering these areas with a fabric such as Mypex makes them easier to clean whilst suppressing most weeds. Plant handling areas, such as propagation and potting benches, should also be cleaned thoroughly regularly to remove any liverwort debris (Altland, 2003).

Fresh media should be covered to prevent contamination by spores or gemmae. Purchase of plug plants or liners that are already infested should be avoided where possible, and if there is no alternative then plants should be inspected and any liverwort removed before potting (Svenson et al., 1997). To help reduce liverwort contamination of new plants clean pots should be used for propagation and potting as weed seeds, moss and liverwort tissues adhere to the sides of containers (Altland, 2003).

Irrigation Liverworts are primitive non-vascular plants preferring damp conditions, therefore allowing the surface of compost to dry before applying irrigation provides unsuitable growing conditions. Allowing young plants in 9cm pots (liners) to start wilting before irrigation is applied, significantly reduces liverwort establishment (Scott and Hutchinson, 2001). Svenson and Deuel (2000) found increased liverwort pot coverage with high frequency (daily) irrigation compared to low (every 3 days) across a range of surface mulch treatments (hazelnut shells, oyster shells, and copper-treated geotextile discs). Although details of the type of irrigation used were not described, the use of sub-irrigation to reduce liverwort establishment was recommended. Clemens et al (1991) compared various irrigation methods (capillary, ebb-and-flow and overhead, and commented that greater liverwort presence on the compost surface was a particular problem with the capillary systems.

40 Introduction and literature review Chapter 1

In this study liverwort dispersal by overhead irrigation under different combinations of water pressure, nozzle height and nozzle size was characterised followed by a closer look at the replenishment of gemmae in gemma cups. The results were used to design a long term experiment comparing liverwort spread by different irrigation systems (drip, capillary matting and overhead) popular with growers, using two different application timings (2-daily and every 2 days).

Media Atwood (2005) investigated the use of compost amendments and found that SylvafibreTM (woodfibre) incorporated at 30% v/v effectively reduced liverwort infestation from 40% to 5%. When combined with the industry standard herbicide programme (Ronstar G, Flexidor 125 + Panacide M, Ronstar G) infestation was further reduced to 0.7% pot cover. Control was attributed to reduced media water holding capacity combined with the biological activity of microbes breaking down the wood fibre (Atwood, 2005).

Mulches that dry rapidly such as Miscanthus and pine bark help media surfaces to dry out (Richardson, 2003; Svenson and Deuel, 2000; Svenson, 1998). Hazelnut shells, oyster shells, filter-fabric weed barriers and copper-treated geotextile discs have also been found to suppress liverwort growth (Svenson et al., 1997). Any mulch used should completely cover the media surface, not blow away easily in the wind and dry rapidly between irrigation cycles (Richardson, 2003; Svenson and Deuel, 2000; Svenson, 1998). Surface mulches such as chopped Miscanthus can also be unsightly and time consuming to apply (Atwood, 2005).

In compost with loam incorporated, antagonistic micro-organisms are present that are thought to help prevent liverwort establishment (Galloway, 2004). In this study some fungi were isolated from dying liverworts and other selected potential antagonists were obtained and used to inoculate liverwort colonies in laboratory bioassays and glasshouse experiments and their effects measured.

Fertilisation Recommendations are to apply the minimum amount of nitrogen or phosphorus fertilisers necessary for optimum crop development, as increased nutrients encourage liverwort growth (Altland, 2003; Clemens et al., 1991). The method of fertiliser

41 Introduction and literature review Chapter 1 application is also important; growers usually either add fertiliser to the top of the pot, incorporating it into the compost or dibbling (placing into a hole made in the compost and then covering up) it into the compost. Dibbling is the preferred method as this prevents liverwort from deriving any benefit from the nutrients provided (Altland, 2003).

Shading Crop canopies can be utilised to provide shade to the growing medium, helping to reduce liverwort growth (Svenson et al., 2001). Experiments carried out in Oregon, US found greater liverwort coverage with 50% shading between 11 am and 4 pm daily than continual 30% or no shading; least liverwort was present with no shade, or where light levels were consequently highest (Scott and Hutchinson, 2001). Altland (2003) also observed greatest liverwort colonisation in plugs where foliage provided less shade.

In this study growth and development of liverwort, initially in controlled conditions at specific light and temperature levels and subsequently with three different levels of shade, established how the environment can affect liverwort establishment, growth and development, and how this relates to plant nursery conditions.

Experimental work focused on the effect of environmental conditions (light levels and temperature) on liverwort growth and development, specifically in relation to using the results to provide practical advice to UK growers as there was little detailed information found in the literature.

1.6.2 Chemical controls

Numerous chemical controls are used to control algae, mosses, liverworts and higher plants, with some chemicals more successful than others against each weed species. It is therefore essential to make the correct choice of herbicide, formulation (granular or sprayable), and application method for effective weed control. Sufficient and uniform herbicide should be applied at the correct rate: too little and control would be incomplete, too much and crop injury may occur. Equipment should be properly calibrated and function correctly (Altland, 2003).

42 Introduction and literature review Chapter 1

Pre-emergent herbicides form a chemical barrier over the compost surface providing control when seeds emerge through the barrier; if the barrier is incomplete weeds will germinate in the gaps and grow, therefore good nursery practice should involve not damaging the chemical barrier whilst moving containers and not dropping or allowing containers to blow over (Altland, 2003).

Timing of herbicide application is also important. Nursery growers tend to use three to four applications each year, but this could be reduced to two with proper hygiene regimes and minor hand weeding. Early spring application prior to weed germination, or soon after potting, is essential for the use of pre-emergence herbicides (Scott and Hutchinson, 2001).

The UK industry standard herbicide programme for liverworts is Ronstar G (active ingredient (a.i.) oxadiazon) followed by Panacide M (a.i. dichlorophen) + Flexidor 125 (a.i. isoxaben), and then Ronstar G. However, Panacide M will be withdrawn in the UK after 31 December 2007 (Atwood, 2005). Scott and Hutchinson (2001) found the chemicals Mogeton (a.i. quinoclamine) and Panacide M to be the most effective of the treatments trialled at both plug and liner stage, although some phytotoxicity was observed with both compounds. Lenacil 80W provided control for a range of weeds, although exhibiting more phytotoxic effects when applied to soft new growth of some species. Few chemical herbicides are approved for use under protection (Atwood, 2004; Certis, 2005)

Subsequent research (Atwood, 2005) obtained best control using Ronstar G followed by Flexidor 125 + Panacide M, and then Lenacil 80W, applied as a winter treatment to avoid phytotoxicity. Application of Alpha Simazine 50 SC (a.i. simazine), Helmsman granules (a.i. oxadiazon + diflufenican + carbetamide) and Butisan S (a.i. metazachlor) as the winter treatment also produced good control with no phytotoxicity (Atwood, 2005).

In the UK, Mogeton is currently listed as a biocide and surface cleaner, providing contact and residual activity but is not approved for application on pots containing plants. A SOLA (specific off-label approval) is currently being sought by the HDC (Brough, 2007) using mutual label recognition rules, as the product is approved in Germany. The Crompton Uniroyal Company is exploring the possibility of gaining

43 Introduction and literature review Chapter 1 approval in the US and Canada, providing data on the performance and phytotoxicity of Mogeton under their growing conditions (Newby et al., 2004; Richardson, 2005). Mogeton appears to provide better control of minor and mature liverwort infestations than Terracyte or Broadstar (flumioxazin) when applied pre- and post-emergence (Altland, 2003).

Quinoclamine is a naphthoquinone derivative that acts as a photosynthesis inhibitor, first developed by Uni-royal Co. as an algacide for industrial use in the 1960s, and subsequently developed as an herbicide (Vea and Palmer, 2006). Its orange colouration is used to confirm even application and gradually disappears over 10 to 12 weeks, dependent on the rate used. It can be tank-mixed with other herbicides such as Ronstar Liquid to cover a wider range of weeds. Certis (UK suppliers) claim that quinoclamine does not affect higher plants except ferns (Certis, 2005). However other research using Mogeton 25W in the US indicates phytotoxicity towards a number of herbaceous perennials, (particularly Aquilegia spp., Athyrium nipponicum var. pictum, Liriope muscari and Verbena spp.), shrubs (Berberis spp. and Hydrangea spp.) and grasses (Miscanthus sinensis) (Scott and Hutchinson, 2001). Altland (2003) also observed some phytotoxicity to plug plants treated with Mogeton.

In the US mixed success has been achieved using flumioxazin, applied as a granular post-emergence herbicide, (sold as Broadstar and SureGuard) with some researchers reporting good control, although with some phytotoxicity to Hosta spp., (Richardson, 2005) and others observing acceptable control when applied post-emergence, with declining effectiveness through the year (Chase, 2000; Svenson et al., 1997; Svenson, 1997). Flumioxazin is not currently approved for use in the UK; however samples have been obtained for research purposes, and trials are planned for 2007 (Atwood, 2007)

In the US, encouraging results have also been obtained using CinnecureTM', a cinnamon oil extract (active ingredient cinnamic aldehyde) although some phytotoxicity was seen on soft, new growth of Rhododendron 'Jean Marie Montague' (Svenson, 1997). Agrigerm 2000 (100 g dimethyl didecyl ammonium chloride, 40 g glutaraldehyde, 32 g gluoxal and 31.5 g formaldehyde per litre) has proved as efficient as Mogeton 25 WP (1.5 g m2) 4 days after application at 2% in Poland (Mompert and Orlikoski, 2000).

44 Introduction and literature review Chapter 1

Terracyte is a granular formulation of hydrogen peroxyhydrate that, on contact with water, degrades to sodium carbonate and hydrogen peroxide, which causes damage to cell membranes. It is used by the US greenhouse industry for sanitation. It has been found to be very effective early in the year, although showing little effect later in the year on colonies of comparable size and maturity (Altland, 2004)

Older colonies of liverwort have different generations in one pot, growing over each other to create a thick layer which even the most effective herbicides find difficult to eradicate, so that manual removal is the only solution. Herbicide applications can be reduced by using cultural control methods, such as good nursery hygiene and keeping hard surfaces liverwort-free. Dead liverwort is equally undesirable from a cosmetic point of view as live.

1.6.3 Non-herbicidal controls

In the research described in this thesis, two non-herbicidal methods are investigated for their activity against liverworts: GSL hydrolysis products and fungal antagonists.

Why use non herbicidal products for weed control? There are a number of benefits in using natural plant products as pesticides: they have been naturally selected for very specific biological activities, some with novel molecular sites of action not used by synthetic herbicides (Duke et al., 2000). Due to their chemical composition and structural characteristics natural compounds tend to degrade rapidly, for example ITC concentrations in soil can decrease by 90% in 24 hrs (Brown et al., 1991) and are thus more environmentally friendly (Dayan et al., 1999). However, they target specific plant species and have correspondingly small markets, historically making the cost of registration prohibitive. Biological and chemical controls are registered through the Pesticide Safety Directorate in the UK, who have launched a biopesticide scheme, covering pheromone and plant extract products, and products containing biological organisms, and through which three have already gained approval: a pheromone to control codling moth in apple and pear orchards, a virus for protection from damage caused by Zuchini Yellow Mosaic Virus of cucurbits and a fungal agent for the control of Scierotinia scierotiorum and S. minoron in susceptible crops. Within the scheme maximum approval fees are £13,000 (pheromones and other semiochemicals) and £22,500 (products containing microorganisms or based on plant

45 Introduction and literature review Chapter 1 extracts), compared to in excess of £105,000 for approval of chemical products containing new active substances (Pesticide Safety Directorate, 2006).

Many natural compounds produced by plants are as yet undescribed, providing a potentially huge chemical palette for scientists to explore and exploit in direct contrast with a lack of new synthetic herbicides in development. Many natural products have been patented as herbicides, but a limited number have been successfully commercialised. Often the active agent in a natural product is identified and then reproduced synthetically (Dayan et al., 1999). One example is glufosinate-ammonium, a non-selective, non-residual herbicide used for the control of broad leaf and grass weeds) and marketed in the UK in a number of formulations (e.g. Challenge, Finale, Harvest) (Whitehead, 2007). The previously unknown amino acid phosphinothricin was isolated from the soil bacteria Streptomyces, subsequently found to have herbicidal bioactivity and is produced synthetically as glufosinate (Duke et al., 2002)

Callisto, an effective herbicide against broadleaved weeds in corn and crabgrass in lawns, contains the active ingredient mesotrione, an analogue of leptospermone, itself a weak herbicide and an allelochemical produced by lemon bottlebrush (Callistemon citrinus) (Encore Technologies, 1999a).

A number of biological controls using fungal antagonists have also been successfully commercialised, including Collego, the first biological herbicide to be registered in the US. It is a post-emergence biological control agent of northern joint vetch in rice and soybeans (Sandrin et al., 2003). Spores of the fungus Colletotrichum gloeosporioidesf. sp. aeschynomene, formulated as a wettable powder, are rehydrated and applied as a spray (Encore Technologies, 1999b).

DeVine is a mycoherbicide used to control milkweed vine in citrus trees (Morrenia odorata) in Florida, US (Encore Technologies, 1999b). Chlamydospores of Phytophthora pa/mivora MWV are applied as a spray (Hallett, 2005).

1.6.3.1 Natural products Glucosinolates (GSLs) and their hydrolysis products are responsible for the distinctive pungent smell and hot taste of cabbages and other brassicas and are known to have fungicidal (Gamliel and Stapleton, 1993; Hooker et al., 1943; Manici et al., 1997)

46 Introduction and literature review Chapter 1 phytotoxic, bacteriocidal, nematocidal, allelopathic and cancer-protective properties (Bialy et at., 1990; Fahey et al., 2001; Matthiessen and Kirkegaard, 2006). Found in dicotyledonous angiosperms, GSLs are predominately produced by members of the order Capparales, particularly Brassicaceae, and all members that have been tested are capable of GSL synthesis. Some 500 plus non-cruciferous dicotyledonous species are reported to contain GSLs, including members of Limnanthaceae (Fahey et al., 2001).

The delivery mode and distribution of GSLs within the soil includes leaching from the roots of living plants (allelopathy) and diffusion from cover crops ploughed into the soil (biofumigation), each resulting in the breakdown of GSLs (Lovett, 2005).

Allelopathy The effects of allelopathy have been observed for thousands of years, reported for 2000 years, and the term finally coined in 1937 by Hans Molisch (Mattner, 2001). Allelopathy describes the beneficial and detrimental chemical interactions that occur between plants, microorganisms, and fungi; allelochemicals, usually secondary metabolites, are produced in all parts of plants as by-products of the acetate and shikimic acid biosynthetic pathways. They act on neighbouring plants or microorganisms, either stimulating or suppressing growth (Weston, 2005) and are released into the soil rhizosphere by decomposition of residues, volatilisation and root exudation (Rice, 1979). Such chemicals, including GSLs, are thought to be secondary products produced as part of the plants' defence strategies against competitors, pathogens and insects.

Members of Brassicaceae are considered poor companion plants, and this is attributed to their allelopathic effects on other plants: GSLs infuse into the soil, leaching from roots of living plants or from whole plants ploughed into the soil.

There are a number of reasons why plants have been selected for these mechanisms in natural ecosystems, for example: many seeds contain or produce microbial inhibitors to prevent their decomposition before germination; others release allelochemicals to enable adequate spatial distribution, an extreme example is Imperata cylindrica, a pernicious weed in the US that produces phytotoxins that attack neighbouring plants (Rice, 1979).

47 Introduction and literature review Chapter 1

Allelopathic compounds have been collected and analysed by various means from root exudates of Hemarthria altissima (Tang and Young, 1982), Desmodium uncinatum (Khan et al., 2002), Rorippa indica Hiern (Yamane et al., 1992) and Pope et at (1985) collected root exudates from nine different plant species. For this project, four target plants (Diplotaxis tenuifolia, Sisymbrium oriental, Brassica juncea and Limnanthes a/ba) were selected to provide a range of bioactive products via their roots grown within a hydroponic system, as a soil-free plant growth technique, and their root exudates collected and analysed.

Biofumigation One aspect of allelopathy is biofumigation, a term originally coined by Kirkegaard et al (1993) to describe the beneficial use of Brassica species to eliminate soil borne pests (e.g. weeds, fungi, microorganisms, insects, nematodes) through the release of toxic GSL hydrolysis products, isothiocyanates (ITC's), into the soil. The term has since broadened to encompass any beneficial effects of green manure, rotation crops and composts (Matthiessen and Kirkegaard, 2006). This effect of Brassicas has also been observed and used for thousands of years.

Brassica cover crops are planted in the autumn and ploughed into the soil in spring prior to sowing the subsequent crop. Phytotoxic allelochemicals are released into the soil as the damaged plant material is incorporated into the soil; these preventing establishment of other plant species within the soil ecosystem, by inhibiting seed germination and suppressing growth (Mattner, 2001).

In vitro investigations indicate that Brassica tissue can inhibit growth of soil borne pathogens (e.g. Rhizoctonia, Fusarium, Phytophthora, Gaeumannomyces) (Kirkegaard et al., 1996), although according to Brown and Morra (2005) control in the field is rarely achieved. Brassica napus has been shown to reduce Gaeumannomyces graminis var. tritici (a pathogen causing take-all in cereals) inoculum in soil, and this was attributed to a biofumigation effect (Kirkegaard et al., 2000).

Additional benefits of biofumigation to growers include: reduced cost of herbicides, or at least a delay in the need for their application; improved soil structure by addition of plant material and soil excavation by cover crop roots; reduced soil erosion and nitrogen leaching over winter compared to unplanted soil; and reduced fertiliser

48 Introduction and literature review Chapter 1 requirements as cover crops break down to provide nutrients (Matthiessen and Kirkegaard, 2006).

Biofumigation is potentially a biological replacement for the use of the soil fumigant methyl bromide, which was phased out in 2005 following agreements contained within the Montreal Protocol (Batchelor, 1998); it is currently approved for use in critical use exemptions only (Pesticide Safety Directorate, 2006). Methyl bromide (bromomethane, CH3Br) is a toxic organic halogen compound occurring naturally in plants, predominately members of Brassicaceae. It is produced synthetically for use as a soil sterilant and fumigant against pests (rodents, insects, fungi, nematodes, viruses, bacteria, fungi, and mites) and is phytotoxic (Matthiessen and Kirkegaard, 2006). It is a gaseous fumigant, applied to plastic tarpaulin-covered soil, which retains the gas as it rapidly diffuses through the soil (Batchelor, 1998). It is toxic to humans, with attendant health and safety issues during application, and has a detrimental effect on soil biodiversity and pollutes surface and ground water.

Confirmed as a 'significant ozone depleting compound' by the United Nations Environment Program (UNEP) in 1994, methyl bromide was brought into the 1989 Montreal Protocol on Substances that Deplete the Ozone Layer, which aimed to eliminate ozone depleting substances, with all imports and manufacture to cease by January 2005 unless the use was agreed as critical (The Food Commission, 2006).The International Partnership for Phasing out Methyl Bromide planned to complete the phasing out process by September 2007 (United Nations Environment Programme, 2006).

Metham sodium (sodium N-methyldithiocarbamate) (Figure 1-8), often considered a potential replacement for methyl bromide, is less potent and degrades to soluble methyllTC upon contact with moist soil, through which it then diffuses, enabling its application via irrigation systems. MethyllTC is effective against nematodes (Santo, 1999), pathogens and weed seeds (Matthiessen and Kirkegaard, 2006; Matthiessen and Shackleton, 2002; Teasdale and Taylorson, 1986). Natural ITCs are similar in structure and some are known to have a toxicity more than fifty times that of synthetic methyllTC (Matthiessen and Kirkegaard, 2006).

Introduction and literature review Chapter 1

H3C S

\ / N C + / \ _ Na H s

Figure 1-8 Molecular diagram of metham sodium

A potential problem with metham sodium first observed in the 1980's is its enhanced biodegradation; the loss of pesticidal effectiveness due to its unusually rapid breakdown in the soil, and consequent failure to manage the target pest. Soil microbes have the ability to use the pesticide as a food source, and enhanced biodegradation is induced by repeated pesticide use on the same area of soil. Many potential methyl bromide replacements are limited in use by this characteristic (Matthiessen and Kirkegaard, 2006).

Glucosinolate chemistry Glucosinolates (GSLs) are non-toxic thioglucosides, having a common core comprised of a 8-D-thioglucose group with a sulphonated oxime, and a variable side chain ('R' group) derived from an amino acid that largely determines the biological activities of the degradation products (Figure 1-9), and which are biosynthesised from amino acids and classified into three groups depending on their origin: aliphatic (methionine, alanine, valine, leucine, isoleucine), aromatic (tyrosine, phenylalanine) and indole (tryptophan) (Fahey et al., 2001; Mithen, 2001; Wittstock and Halkier, 2002).

s f3 Glucose / R C 11 N OS03-

Figure 1-9 Structure of glucosinolate. R represents the variable side chain (Mithen, 2001)

Introduction and literature review Chapter 1

GSL hydrolysis within the soil (Figure 1-10), catalysed by a myrosinase enzyme (thioglucoside glucohydrolase, EC 3.2.3.1) released following mechanical damage in the presence of water, involves cleavage of the thioglucoside linkage, yielding D- glucose and an unstable aglycone intermediate (thiohydroximate-Gsulphonate). This spontaneously rearranges, producing sulphate and one of a range of degradation products, primarily ITCs, thiocyanates, nitriles, or epithionitriles, dependent on the 'R' group present on the GSL substrate, pH and availability of ferrous ions (Fe2±) (Bones and Rossiter, 1996). Nitriles are produced at lower pHs, ITCs in neutral and high pH conditions (Gil and MacLeod, 1980). Aliphatic and aromatic GSLs produce ITCs, the most bioactive of the hydrolysis products (Vaughn et al., 1996; Vaughn et a/., 2006), with ITCs more phytotoxic than the corresponding nitriles (Borek et al., 1995).

H2O Glucosinolate ► Intermediate -► + Glucose Thioglucosidase i + HSO4 Isothiocyanate Nitrile Thiocyanate

Figure 1-10 General structure of GSLs and their major hydrolysis products. Adapted from (Vaughn etal., 1996).

GSLs and myrosinase exist throughout the plant, separated either physically in different areas of the same cell or in different cells; or chemically with the myrosinase in inactive form, being brought together after mechanical cell damage (Bones and Rossiter, 1996; Wittstock and Halkier, 2002).

There are a number of forms of myrosinase, and more than one can be present within a single plant, for example James and Rossiter (1991) found that Brassica napus has two myrosinases, one restricted to the cotyledons with possible involvement in nutrient mobilisation and the other sited throughout the plant and linked to the plant defence system. Myrosinase is localised and synthesised within scattered myrosin cells specifically associated with the tonoplast-like membrane surrounding myrosin grains (Bones and Rossiter, 1996; Rosa et al., 1997).

51 Introduction and literature review Chapter 1

GSL abundance due to environmental conditions Researchers investigating the effect of soil conditions on 2-propenyllTC and allylnitrile production, and methyllTC production from metham sodium (Gerstl et a4, 1971; Smelt and Leistra, 1974) found that ITCs were generated more rapidly in soils with less moisture, higher temperature and in loamy soil with greater organic carbon and clay content. Allyl nitrile production increased in wetter conditions with lower temperatures and higher inorganic carbon content (Gerstl et al., 1971).

Metham sodium, the active ingredient in a number of herbicides including Vapam, is highly soluble in water and is applied via irrigation systems, so its rate of degradation is crucial to its success. It must be rapid enough to maintain the methyllTC concentration in soil water above a critical level (50 ppm) for long enough to be effective as a pesticide before the methyllTC breaks down to non-toxic substances. MethylITC breaks down more slowly in coarse textured soil with less than 20% clay, than fine textured soil with a higher clay content and this helps to maintain higher levels (Gerstl et a4, 1971). Half-lives vary between 20 to 60 hrs (2-propenyllTC), and 80 to 100 hrs (allylnitrile) (Borek et al., 1995). Metham sodium degrades to methyllTC within mins to hrs depending on soil conditions; methyllTC was found to degrade to non-toxic substances within 4 to 13 days (Teasdale and Taylorson, 1986)

The GSL profile and level of ITC toxicity varies between plant species and within individual plants. Sang et a/(1984) found that the major GSL present in radish seeds is absent from the leaves and roots, whereas the major GSL of Brassica juncea is found in all tissue types, demonstrating the variability in pesticidal potential of GSL products in individual plants. Plants species may produce single or multiple GSL hydrolysis products (Vaughn and Berhow, 2005).

Glucosinolate (GSL) content of plants is affected by plant growth stage and cultural conditions, such as boron (Ju et a4, 1982), sulphur and nitrogen fertilisation (Josefsson, 1970), and water stress (Louda et al., 1987). Increased GSLs occur in high sulphur fertilisation conditions, particularly on sandy soils where lack of sulphur may normally be a limiting factor; decreased GSLs in high nitrogen conditions in soil-free- culture (Josefsson, 1970); and more GSLs are produced in water-restricted conditions (Louda et a4, 1987). There is a trade-off, however, by growing brassicas in low nitrogen, low water conditions, where the concomitant reduction in plant biomass

52 Introduction and literature review Chapter 1 results in less GSLs available to be returned to the soil (AI-Khatib and Boydston, 1999). Similarly, a reduction of GSL concentration in the autumn is counterbalanced with increased biomass, resulting in an overall increase in GSL production (Sarwar and Kirkegaard, 1998).

GSL concentration within Brassica napus plant tissue varies throughout its life cycle: decreasing during germination and early growth, stabilising after 15 days, and increasing between 35-100 days before decreasing again during flowering (Matthiessen et al., 2001). The greatest biofumigation potential occurs during budding and early flowering (Clossais-Besnard and Larher, 1991). This suggests metabolism of GSLs during germination and early development and flowering when components (nitrogen, carbon and sulphur) may be required to sustain growth; GSLs accumulate, mainly in shoots and roots once seedlings start to photosynthesise and biomass increases (Kirkegaard, 1998; Sarwar and Kirkegaard, 1998). GSL concentration tends to be greater in spring- than autumn-sown plants (Sarwar et al., 1998). Kirkegaard (1998) found that roots produced 23.6% of total plant GSLs, limited only by low biomass. GSLs were predominately aliphatic in shoots and aromatic in roots, with low concentrations of indole GSLs present (below 1 p mol g-1) in all tissues.

Phytotoxic effects of ITCs There have been many studies providing evidence of phytotoxic effects of ITCs:

Experiments using liquid ITCs on Panicum texanum, Digitaria sanguinalis and Senna obtusifolia, (Norsworthy and Meehan, 2005a) and subsequently Cyperus rotundus and Cyperus esculentus showed a reduction in shoot density and biomasss at the various concentrations used (Norsworthy et al., 2006). Vaughn et al (2006) evaluated the phytotoxic effect of 15 GSL-containing seed meals on wheat and sicklepod seedling emergence. The greatest effect was produced by 2-propenyllTC (brown mustard seed meal), allylthiocyanate and allyllTC and 4-methylthiobutyllTC (erucin) (arugula seed meal), 3-butenyllTC and 6-methylthiohexyllTC (lesquerellin) (sweet alyssum seed meal). Brown and Morra (1995) investigated the effects of GSL-containing seed meal of Brassica napus L. 'Dwarf Essex' on germination of Lactuca sativa seeds; the results suggested an inhibitory effect of ITCs, nitriles and water soluble components. Norsworthy and Meehan (2005b) compared the phytotoxic effects of eight soil- incorporated ITCs on Amaranthus palmeri, Ipomoea lacunosa and Cyperus esculentus

53 Introduction and literature review Chapter 1 seedling emergence, finding 3-methylthiopropyllTC and phenyllTC the most effective. Bialy et al (1990) investigated the inhibition of wheat (Triticum aestivum 'Bop') seedling germination and growth by various GSLs and ITCs. He found that ITCs showed the greatest activity, particularly 2-phenethyllTC, completely inhibiting germination at 500 ppm concentration, and 2-propenyllTC.

Liverwort control properties of Limnanthes alba Hartweg ex. Benth The general principle of biofumigation has been adopted in the use of L. a/ba seed meal, usually discarded as a waste product following oil extraction, as mulch, harnessing its pesticidal and growth promotion properties.

Figure 1-11 L, a/ba growing in a hydroponic system

L. a/ba (meadowfoam), a member of the family Limnanthaceae (Table 1-3) is an herbaceous winter annual native to northern California, Southern Oregon, and Vancouver Island, British Colombia. It is rosette-forming with bright green pinnately dissected leaves and buds produced in terminal clusters on stems arising directly from the rosette to a height of 20-30 cm (Figure 1-11). The lightly scented white flowers have 5 petals and are followed by seeds, usually in groups of 3 nutlets that turn mid- brown, striped or mottled. Flowering is related to the onset of late spring, and summer drought and temperature. The fibrous root system penetrates to 150 mm (IENICA, 2006; Oelke et al., 1990).

Table 1-3 Taxonomy of L, alba

Taxonomic level Taxonomic group Class Dicotyledoneae Order Geranialea Family Limnanthaceae Genus & species L. alba

54 Introduction and literature review Chapter 1

Seeds are sown in October, below 15.5 °C in darkness to prevent secondary dormancy. Low nitrogen is applied (below 50 kg hal to improve seed yield and quality and to prevent delayed flowering and decreased seed oil content, and soil pH is adjusted to 5.5-6. Yield is variable, with an average of 0.8 t ha-1. Meadowfoam is adapted to poorly drained soils with good water retention and has a low tolerance to water stress, as it is well adapted to cool wet climates. Meadowfoam is an entomophilous species, primarily pollinated by honey bees, so pollination is poor in cool damp conditions at flowering time due to a dearth of insects flying. To ensure cross pollination, pollen is released before the stigma is receptive (Burden, 2005).

Meadowfoam is grown for the high volume industrial oilseed market in direct competition with rapeseed, and is used in cosmetic and hair care products (Burden, 2005; Oelke et al., 1990). Seeds are high yielding, producing 20-30% oil with outstanding oxidative stability and are unique in having over 90% long chain fatty acids (20-22 carbon) with high levels of mono-unsaturation and low levels of poly- unsaturation, making it useful as lubricants, inks, detergents and platicisers and an ideal substitute for sperm whale and jojoba oils (Vaughn etal., 1996).

Preliminary field experiments into alternative methods of pest control, coupled with the search for a use for waste meadowfoam seedmeal suggested a possible reduction in weed numbers following seedmeal application as a soil amendment. Vaughn et al (1996) compared the phytotoxicity of meadowfoam GSL hydrolysis products and found that 3-methoxybenzyllTC (Limnanthin) was more toxic than (3-methoxyphenyl acetonitrile (3-MPAN) against both velvetleaf (Abutllon theophrasti) and wheat (Triticum aestivum 'Cardinal') seedling radicle elongation (Vaughn etal., 1996).

Svenson and Deuel (2000) recognised the potential of L. a/ba seed meal as a natural liverwort control and plant growth promoter. Trials were carried out using different quantities of seed meal applied as a top dressing on pots containing Rhododendron 'Cannon's Double' and which had been inoculated with a slurry of liverwort gemmae and thallus tissue blended with buttermilk and water 40 days previously. The seed meal provided excellent control after 30 days, although efficacy had reduced after 60 days; no phytotoxicity towards the Rhododendron was observed. Disadvantages of the seed meal were: difficulty in application, an unpleasant odour, grass seed infestation and fungal growth was observed on treated pots (Svenson and Deuel, 2001). During

55 Introduction and literature review Chapter 1 subsequent investigations the seed meal was substituted with AlbaGroTm, a pelleted product derived from L.alba seed meal (with the addition of granular sulphate, produced by Natural Plant Products, Salem, Oregon, US) and the previous experiments were repeated applying AlbaGroTM to the surface of liverwort-infested pots. Results were improved using this product, with reduced control only in the 1/4 cup treatment after 60 days; the pellets were more easily applied, and the odour and grass seed problems removed (Svenson and Deuel, 2001). UK research incorporating meadowfoam seed meal into compost at 1% an 2% levels provided only short term liverwort control (Atwood, 2005).

Meadowfoam seed meal has been marketed in the US by Natural Plant Products in various forms: AlbaAideTM, AlbaGroTm and Limnaxm (Novak, 1999), with LimnaxTM as a high grade mulch and growth promoter for farm and nursery soils, and for weed control. At present only AlbaAide is marketed as a fertiliser, in the state of Oregon, and it does not require a herbicide licence (Martinez, 2006).

There are reports of L. a/ba GSL products' toxicity to fall armyworm larvae and European corn borer (Bartelt and Mikolajczak, 1989). Soil-incorporated L. a/ba seed meal is toxic towards the root-knot nematode Meloidogyne chitwoodi, affecting potatoes (Santo, 1999). Deuel and Svensen (1999) incorporated L.a/ba seed meal into infested soil and demonstrated its potential in controlling the plant pathogen Plasmodiophora brassicae, the causal agent of clubroot in cauliflower and Chinese mustard, possibly due to the effects of thionin and aglycone allelochemicals.

L. a/ba seedmeal contains glucolimnanthin which is known to form primarily 3- methoxybenzyllTC following hydrolysis. ITCs have been found in other members of the Limnanthaceae (Table 1-4) in varying quantities, some of which may grow more successfully in the UK (Miller et al., 1964).

Table 1-4 ITC content of members of Limnanthaceae (Miller et al., 1964)

L. species Isothiocyanate content (mg/g) L. alba 15.6 L. alba var. versicolor 19.3 L. douglasii 18.0

56 Introduction and literature review Chapter 1

In this study investigations were carried out into L. a/ba, establishing its GSL profile and optimising the extraction of glucolimnanthin and 3-methoxybenzyllTC. In vitro bioassays were used to characterise the inhibitory effects of selected ITCs (2- phenylethyl, benzyl, 2-propenyl and 3-methoxybenzyl) and two synthetic herbicides (metazachlor and lenacil) approved for use in the UK.

1.7.3.2 Fungal antagonists Many fungal species are exploited for use as biocontrol agents. By 2004, 26 fungal species had been used as classical biocontrol agents, since the introduction in 1971 of Puccinia chondrillina into Australia to control skeleton weed (Chondrilla juncea), with no reports of unpredicted damage to non-target plants (van Driesche and Bellows, 1996).

According to van Driesche and Bellows (1996) members of the Basidiomycotina (including rusts and smuts), Deuteromycotina and Ascomycotina are of importance to biological control. Many of the fungal species used for classical biocontrol are rusts, members of the Uredinales, their characteristic high virulence, efficient dispersal mechanisms, and high host specificity contributing to their success (Barton (nee Frohlich), 2004).

Strategies: classical and inundative Three main strategies can be used for weed control with fungal antagonists: classical, inundative and use of phytotoxic fungal metabolites.

In the classical approach the pathogen, is released, disseminates and then perpetuates so that the weed population is regulated at an acceptable level. The antagonist, a natural pathogen of the target plant, is usually an exotic species imported from an area where it co-exists with the host (Watson, 1991).

For the inundative method a large population of a pathogen is applied as a mycoherbicide, and can be defined as 'a living product that controls weeds in agriculture as effectively as chemicals' (Templeton et a/., 1986) and is applied in a similar manner to chemical herbicides. These can be used against native and exotic weeds and are most successful with weeds that grow in dense patches over a wide area, making them difficult to control chemically for economic or environmental

57 Introduction and literature review Chapter 1 reasons; also with weeds in annual crops where specificity is paramount. Weeds such as northern joint-vetch (Aeschynomene virginica) that grow densely in damp areas provide an ideal microclimate for pathogens to establish (Charudattan, 1991). The pathogen needs to survive only until the target weed is under control, thus preventing inoculum build up to unacceptable levels (Templeton et al., 1979). Pathogens used for this method must be easy to produce in culture; therefore the rusts often used for the classical approach are unsuitable.

Another strategy that is used is identification and development of phytotoxic metabolites from fungi, or necrogenic micro-organisms, into chemical herbicides or for use as additives to bioherbicides, however they may also be toxic to humans (Greaves, 1996). An example of a phytotoxic fungal compound is maculosin, produced by Alternaria alternata, which has been found effective against Centaurea maculosa (Bobylev et al., 1996).

Mycoherbicide formulation This project concentrates on using an inundative strategy, to provide maximum control of liverwort whilst the crop plant is containerised. Formulation of such mycoherbicides includes the fungal pathogen as the active ingredient, a carrier and a surfactant (e.g. tween 80 or tween 20, 0.02 — 0.05%) to help fungal propagules to disperse and to reduce loss of the pathogen in run-off; addition of an adjuvant (e.g. sucrose) can stimulate fungi to germinate and infect target weeds more effectively, or increase formulation viscosity (e.g. gelatine), reducing runoff (Daigle and Connick, 1990). Water, inexpensive, easy to handle and necessary to the pathogen, is often used as a carrier of weed pathogens. Invert oil emulsion carriers, a water-in-oil mixture which slows water evaporation aids fungal establishment and can also be successful, although application is more complicated (Daigle and Connick, 1990).

Soil-borne fungi can also be applied with a solid carrier, such as vermiculite, rice, wheat kernels, alginate-kaolin granules or peat compost. Alginate is a water-soluble polysaccharide gum extracted from kelp and seaweed that forms biodegradable gel beads that entrap organisms such as fungal inoculum (Connick, 1988). Vestberg et al (2004) combined a conidial suspension of Trichoderma harzianum with peat, while Blok and Bollen (1997) used a soil meal culture combining potting compost with 5% oatmeal. 'Pesta' (pasta-like) granules have been widely used, encapsulating conidial

58 Introduction and literature review Chapter 1 inoculum in a wheat gluten matrix consisting of wheat flour, fungal inoculum and water, air dried, ground into granules and incorporated into soil (Connick et al., 1991). Choice of carrier depends on compatibility with the selected fungal pathogen and has to be tested thoroughly to provide optimum conditions for the species under consideration.

Pathogen viability needs to be conserved throughout processing and storage so the pathogen will be in a condition to establish rapidly once applied to the target weed and this is one of the functions of solid carriers (Daigle and Connick, 1990). When investigating the use of Trichoderma harzianum to control Botrytis cinerea infection of apples Batta (2004) formulated conidia in an invert oil emulsion based on coconut and soybean oils, which then remained viable for 36 months, with a 50% reduction in viability after 5.3 months at 20 °C; dry, non-formulated conidia remained viable for 2.7 months, reduced by 50% after 0.7 months. Viability of Fusarium oxysporum mycelium, macroconidia, and chlamydospores formulated in pesta granules was tested by Elzein et al (2004) who achieved 85-100% viability of Fusarium oxysporum chlamydospore inoculum for at least 1 year, stored at 4 °C; mycelia! and macroconidial were not viable after 1 year. Commercially produced DeVine uses live chlamydospores of Phytophthora pa/mivora, capable of remaining viable when refrigerated for six weeks (Greaves, 1996).

Environmental requirements of mycoherbicides A number of fungi need sufficient time in free water, the dew period, for efficient spore germination and pre-infection growth, with 12-36 hrs of free water commonly required. Foliar applied inoculum in particular is prone to desiccation prior to germination, severely reducing efficacy; soil provides a buffer against moisture and temperature fluctuations (Greaves, 1996). Use of a humectant type adjuvant, promoting retention of moisture, such as sucrose as in Collego, helps with this problem; sucrose also promotes pathogen growth (Daigle and Connick, 1990).

Older target plants require a greater spore density and sometimes longer dew periods for infection, possibly due to host resistance variation with age; younger plants are more vulnerable (Daigle and Connick, 1990; Morin et al., 1998). Symptom development can be slow, e.g. an incubation period of 4-7 days before symptoms become apparent and then up to 5 weeks for the fungus to kill the weed; efficacy can

59 Introduction and literature review Chapter 1 be improved by using higher application rates and humidity (Daigle and Connick, 1990). Sometimes a combined microbial and sub-lethal rate of chemical herbicide are used, the stress produced by the herbicide making some microbial agents more effective (Altman et at., 1990). Herbicides can predispose crop plants to increased disease, which can then be exploited by the increasing efficacy of the mycoherbicides. Herbicides can also stimulate hyphal growth and spore production, e.g. triflualin increased growth and spore production of Fusarium solani f. sp. Phaseoli (Charudattan, 1991).

Advantages of mycoherbicides compared to chemical controls No unexpected non-target effects of mycoherbicides have been reported using the classical approach. Rigorous host-range testing identifies potential susceptible non- target plants (Barton (nee Friihlich), 2004), and mycoherbicides need to be specific to the target weed, reducing risk of non-target infection (Greaves, 1996).

Constraints and Commercial disadvantages Difficulties posed by using mycoherbicides within an inundative strategy are the need for the pathogen to be effective in a range of environments pre- and post- application. The pathogen must grow well in culture, producing adequately infective inoculum, and once applied the inoculum must establish under ambient conditions (Templeton et al., 1979). Failure to be effective outside a restricted geographical area, season or environmental could make a product economically untenable (Templeton, 1990).

Mycoherbicides have to compare well to chemical herbicides to make them attractive to growers. The long incubation periods of some fungi prior to germination (Greaves, 1996) and subsequent control, and the requirement for a dew period for infection are both negative indicators. Growers may be resistant to using microbials closely related to serious crop plant pathogens and will need to become accustomed to a number of issues: speed of action, with mycoherbicides acting more slowly than synthetic pesticides; the need to apply them with greater consideration given to timing; and the more likely successful control of younger target plants than old (van Driesche and Bellows, 1996).

There are also serious economic constraints to the development and use of mycoherbicides. The host-specific nature of pathogens, with each usually controlling

60 Introduction and literature review Chapter 1 one weed species, the weed would need to be a large problem to many growers to make product development commercially viable (Templeton, 1990), therefore such products are rarely economically attractive to the private sector (Charudattan, 1991). Economic viability is improved if there is a lack of other suitable controls, if a patent is obtained to protect product ownership, and where there is a technical ease in production (Daigle and Connick, 1990). Mycoherbicides need to be compatible with chemical herbicides (Weidemann and TeBeest, 1990) and fungicides, as a program of treatments may be used by growers throughout the season to control a variety of pests and diseases.

Fungal genetic variability. There are two aspects of genetic variability of importance to mycoherbicides. The species used require sufficient genetic stability to prevent loss of virulence towards the target weed or, potentially more damaging, change their host range to include non- weeds. Conversely a heterogenous gene pool can be beneficially utilised to select for desirable characteristics, such as increasing virulence to the target weed or improving tolerance to a broader range of environmental conditions. Genetic variation may be obtained by selecting different strains or races of one fungal species to which the target weed may be susceptible in varying degrees (Weidemann and TeBeest, 1990).

Associations between bryophytes and fungi The main bryophyte groups (Metzgeriales, Sphaerocarpales, Jungermanniales, Marchantiales) probably arose in the first quarter of the Mesozoic era (654 to 248 million years ago). The oldest bryophytes appeared along with early vascular plants in the late Silurian and early Devonian (443 to 354 million years ago) periods. Subclasses and orders of hepatics were probably established as early as the Devonian period (Paleos, 2005). Fungal fossils have been found probably dating from the early Devonian, with the first Ascomycetes appearing during the Carboniferous era, 359 to 299 million years ago (Boullard, 1988). A temporal link therefore exists whereby bryophytes and fungal groups evolved together, providing the opportunity for the ancient fungal relationships to form that we see today in members of Bryophyta and Hepatophyta.

During the course of time members of Marchantiales have adapted to grow in intense light and very wet conditions; Marchantiidae, which arose by the early Mesozoic,

61 Introduction and literature review Chapter 1 resisted desiccation, as they do today, by having xeromorphic thalli (structurally modified to withstand drought), conversely members of Monocleales and Sphaerocarpales are drought-intolerant (Felix, 1988).

Bryophytes are often found in damp areas, providing a microenvironment beneficial to fungal growth. Bryophytes can also live on hot, dry surfaces, or other nutrient deficient conditions, where fungi may solubilise otherwise unobtainable organic matter, that may be taken up by hepatics, so that growing on bryophytes allows fungi to proliferate in otherwise inhospitable environments. There is therefore an association between fungi and bryophyte that is generally beneficial to fungi without bringing serious harm to the bryophyte (Dobbeler, 1997).

A number of fungal species, around 300 Ascomycetes are known to grow exclusively on mosses and hepatics, including the obligate bryophilous genus Bryoscyphus, and other genera which include bryophilous species, such as Dacty/ospora heimerlk Muellera frullaniae, and members of Hymenoscyphus and Nectria (Kirk and Spooner, 1984). Of these only a few, including Btyoscyphus marchantiae and B. conocephali, are known to parasitise thalloid hepatics. B. marchantiae was initially reported, incorrectly, on Marchantia polymorpha, and has since been corrected to the liverwort Reboulia hemisphaerica; it is uncertain if B. marchantiae grows on M. polymorpha, (Dobbeler, 1997). Phaeodothis winteri is one of a few facultative cosmopolitan species, able to grow on higher plants, fungi and bryophytes including M. polymorpha (Dobbeler, 1997). Hyphae of fungi that destroy bryophytes grow intracellularly. However, most bryophilous species are biotrophic parasites, requiring living host cells and so only slightly damaging their hosts, as their hyphae grow superficially on the host or within cell walls (Webster, 1980).

Ascomycetes A number of potential fungal antagonists of Marchantia polymorpha were investigated in this research (Chapter 6), all of which were Ascomycetes. Ascomycete fungi include species that are saprophytes and parasites of plants and animals, and occupy a variety of habitats. Ascomycetes characteristically reproduce by sexual spores (ascospores), but also by asexual conidia; some reproduce by ascospores only and others have more than one conidial form, for example Fusarium equiseti produces both microconidia and macroconidia. Ascomycetes tend to produce eight sexual spores in an ascus from

62 Introduction and literature review Chapter 1 which they are often explosively ejected. Their hyphae are septate, each compartment containing multiple nuclei, sometimes of more than one genotype, with septal pores allowing movement of cytoplasm, mitochondria and nuclei throughout the mycelium. Cell walls have a chitinous microfibrillar skeleton along with other compounds, including protein, mannose and glucose (Webster, 1980).

Ascospores are produced within an ascus, a structure that develops from an ascogenous hypha, itself developed from a (female) ascogonium. In mycelial Ascomycetes the asci are surrounded by hyphae, forming an ascocarp, of which there are four forms (Table 1-5, Figure 1-12), and by which they are grouped (non- taxonomically). Ascocarps may occur singly or in groups. There are two ascus types, unitunicate with rigid, inelastic inner and outer walls, and bitunicate with distinct outer and inner walls (Webster, 1980).

Asci and paraphyses

Asci and paraphyses (hymenium)

Figure 1-12 Ascocarps produced by members of Ascomycetes. (a) Perithecium (b) Apothecium and (c) Cleistothecium. Adapted from (Alexopoulos, 1979)

In temperate regions, almost all pathogenic Ascomycetes infections are caused by the asexual stage (conidia) during the growing season; at the end of the season or when the food supply is diminishing the sexual (perfect) stage is produced on or in infected leaves, fruits or stems. The perfect stage is the over wintering stage, however, usually the fungus can also over winter as mycelium and sometimes conidia. In spring primary

63 Introduction and literature review Chapter 1 infections are usually caused by ascospores, which then produce the conidia that cause infections throughout the growing season (Agrios, 1997).

In this study liverworts were inoculated with fungal antagonists under laboratory conditions and then in growth cabinets under environmental conditions selected to favour the fungal growth. For successful infection to take place the fungi would have to invade the liverwort and feed and proliferate under conditions that favour the host. For these experiments to be a success both the liverwort and pathogen would have to be at an appropriate stage: the liverwort susceptible to the pathogen, the pathogen in an active stage, and with temperature and moisture levels favouring pathogen growth

(Agrios, 1997).

Table 1-5 Ascocarp grouping of Ascomycetes (Agrios, 1997)

Group Ascocarp Description Plectomycetes Cleistothecium Globose structure with no special opening, (unitunicate) disintegrates to release a mass of sticky spores that ooze out in a drop e.g. Penicillium velutinum. Pyrenomycetes Perithecium Flask-shaped, open to the exterior via a narrow pore (unitunicate) (ostiole). Asci elongate and are discharged singly e.g. Fusarium equiseti, Trichoderma harzianum. Discomycetes Apothecium Open saucer-shape with exposed asci tips at (unitunicate) maturity. Increased pressure causes the spores to shoot out in a puff or cloud. Some have a pore in the ascus tip (operculate), others don't (inoperculate), e.g. Bryoscyphus atromarginatus. Loculoascomycetes Pseudothecium Similar to perithecia, may be embedded in a bed of (bitunicate) tissue (stroma). 'Jack in the box' spore discharge mechanism - the rigid outer wall ruptures as the ascus expands, the inner wall stretches, the ascus explodes, expelling the spores simultaneously e.g. Phaeodothis winter/.

64 Introduction and literature review Chapter 1

Higher plants have developed mechanisms for resisting pathogen attacks: Non-host resistance by definition is manifested by taxonomic groups outside the pathogen's host range. Constitutive resistance, involving mechanical resistance in the form of cell wall reinforcement by callose deposition, a thick cuticle, and chemical resistance. Inducible resistance is generated by possessing resistance genes which interact with specific avirulence genes of a pathogen (gene-for-gene) resistance. Tolerance of infection by the pathogen (apparent resistance) can also occur (Agrios, 1997).

There is little information in the literature concerning liverwort defence against pathogen attack, but there are numerous reports of close associations between fungal species and liverworts. The ascomycete endophyte Hymenoscyphus ericae infects both ericaceous plants and leafy liverworts (Duckett and Read, 1995); other examples are Xylarla spp., found infecting liverworts (Davis et al., 2003); and Mniaecia nivea, a biotrophic parasite of leafy liverworts (Sloover, 2001); and a mycorrhiza-like infection in the gametophyte of Conocephalum conicum, a member of Marchantiales (Ligrone and Lopes, 1989). Hyphae of mycorrhizal fungi can can sometimes be found in patches of cells with purple walls either side of the middle line of the ventral surface of the thallus of the liverwort Pressia quadrata (Watson, 1971).

However, there are reports of high rates of extracellular superoxide production in bryophytes in response to rehydration following desiccation (Minibayeva and Beckett, 2001). Hydrogen peroxide is formed in cell walls of cultured M. polymorpha, formed by NAD(P)H and mediated by cell wall associated peroxidase (Ishida et al., 1987), although not linked with defence by these authors. The origin of the reactive oxygen species (ROS) could be generation of peroxide (H202) by cell wall peroxidise (Wojtaszek, 1997). Oxidative bursts, the rapid and transient production of reactive oxygen species (ROS) (e.g. superoxide, hydrogen peroxide and nitric oxide) by plants when stressed are one of the indications of a plant defence system that could protect against invading pathogens. The release of ROS occurs mainly on the cell surface and could be a signal activating further defence reaction, including hypersensitive response (HR) of infected cells (Wojtaszek, 1997).

There is evidence of chemical resistance to infection by liverworts, which produce oil bodies composed of lipophilic terpenoids and aromatic compounds. Marchantin A, is

65 Introduction and literature review Chapter 1 one such product found in many species of Marchantiales, including M. polymorpha, which shows antifungal, antimicrobial and cytotoxic properties (Asakawa, 2001).

To determine whether a pathogen is responsible for a plant disease, a procedure, known as Koch's postulates is followed: the pathogen must be found in association with the infected plant, and isolated from it, producing a pure culture on nutrient media or susceptible host and its characteristics described; it must then be reintroduced to healthy plants of the original diseased species where it must produce the same disease symptoms; and finally it must again be reisolated and reproduce the characteristics previously recorded .

For this study Bryoscyphus atromarginatus and Phaeodothis winteri were selected as potential antagonists of liverwort, the former due to its reported reputation as a parasite of M. po/ymorpha (Verkley et al., 1997); the latter as a cosmopolitan species found world wide including fungi, liverworts (M. po/ymorpha) and phanerogams (Aptroot, 1995). Strains were obtained from Centraalbureau voor Schimmelcultures (CBS) in the Netherlands. Isolates originated in both temperate and tropical areas (Chapter 6), with different growth condition requirements leading to the possibility of variation in vigour and parasitising potential. Ascomycete fungal species were also isolated from dying liverwort and identified as Fusarium equisek Trichoderma harzianum and Penicillium velutinum.

Summary of experimental chapters Experimental work carried out in this study was designed to utilise information found in the literature whilst also making the work relevant to liverwort control in plant nursery conditions, and eventual incorporation into integrated pest management systems. Details of general procedures used that are common to more than one experimental chapter or are of routine method development are contained within Chapter 2. Otherwise the specific methods used are described in the relevant experimental chapters.

Preliminary investigations into liverwort growth and development under controlled conditions (Chapter 3), with varying temperature and light levels, were developed firstly to extend the temperature range considered and then in a final nursery scale

66 Introduction and literature review Chapter 1 experiment liverwort growth and development were compared under different light regimes over a longer time span to confirm the results of the earlier experiments.

Asexual reproduction and dispersal by gemmae was investigated (Chapter 4) in an initial experiment using overhead sprinkler irrigation with different nozzles, nozzle heights and water pressures. Aspects that became apparent during this experiment were investigated further: the dispersal of gemma in tightly packed clumps and the replenishment of gemma cups which provided insight into the effect of the timing of application of irrigation. Again, the results of the preliminary experiments were used to design a longer term nursery scale experiment to confirm the results of the earlier experiments, and also to compare the results with other irrigation systems (drip and capillary matting).

The potential of using natural control methods was then investigated (Chapter 5), developing methods to optimise production of GSL hydrolysis products from seed meal and plant material, and carrying out bioassays to identify which products would be most effective as a control. Isothiocyanate levels in root exudates were also considered, by collecting and analysing root exudates of plants grown hydroponically.

Finally potential fungal antagonists were obtained (Chapter 6), and others were isolated from infected, dying liverworts. These were used in bioassays to eliminate those that showed little effect against liverwort; those that showed most promise were used in a long term glasshouse experiment applied as pre- and post-emergent treatments.

67 General procedures and method development Chapter 2

Chapter 2 General procedures and method development

General procedures and methods developed for the research are described in this chapter, whilst methods specific to individual experimental topics are described in Chapters 3-6.

2.1 General procedures 2.1.1 Identification of liverwort species infesting nurseries Marchantia polymorpha is the most commonly reported liverwort in nurseries (Svenson et at., 1997). Members of the horticulture industry were invited via articles in the horticultural press to contact me with details of infestations of liverworts that appear to be different to M. polymorpha. Only one other report was received. A liverwort occupying the same nursery niche, Lunularia cruciata, was reported growing on a wall at Hadlow College, Kent (Figure 2-1)(Carey, 2005). Lunu/aria cruciata is closely related to M. po/ymorpha and easily distinguished by the crescent shaped gemma cups, found on the thallus surface; M. po/ymorpha has round gemma cups. Although growing within the nursery L. cruciata was not found invading plant pots and was not considered to be a nursery weed. All plant material was assumed to be M. po/ymorpha subsp. ruderalls.

68 General procedures and method development Chapter 2

Figure 2-1. Lunularia cruciata found at Hadlow College, Kent (a) growing on a wall (b) detail of crescent-shaped gemma cups (Carey, 2005)

2.1.2 Glasshouse liverwort culture Liverwort cultures were maintained on compost media in half seed trays (220 x 175 mm) in the glasshouse for experimental use. Benches, covered with plastic and capillary matting to maintain humidity and reduce compost drying, were inspected twice daily, irrigated with rainwater as necessary, and fed weekly with Solufeed 'Super' water soluble fertiliser, N: P: K ratio: 18:10:10 at the recommended rates (Supplier: MonroSouth).

2.1.2.1 Compost media 'Imperial College' compost media was used except when stated, consisting of a peat and grit substrate with base fertilisers (Table 2-1). After six weeks liquid feed was applied weekly. Fargro P Base is a compound fertiliser, N: P: K ratio: 6.8:2.5:6.1 (Supplier: Fargro).

Table 2-1. 'Imperial College' compost media components

Manufacturer Item Quantity

Shamrock Irish moss peat 300 L

Horticultural grade grit 75 L

Fargro Fargro P Base 2.72 kg

SHL Professional Potting Compost (Table 2-2), manufactured by William Sinclair Horticulture Ltd. (Supplier: MonroSouth) was used for the liverwort dispersal experiment (Chapter 4.34). N:P:K ratio:1:1.2:2.

69 General procedures and method development Chapter 2

Table 2-2. Peat content of SHL Professional Potting Compost

Peat grade (mm) Proportion (%) 0-10 70 3-15 30

2.1.2.2 Glasshouse environmental conditions Liverwort is reported to tolerate pH levels between 3.5 (Taren, 1958) and 10 (Voth, 1943). Wye is located in a hard water area (pH 7.3) and liverwort growth improved when irrigated with rainwater. Schneider (1967a) suggested optimum growth conditions within a temperature range 18-20 °C, with a 12-18 hour photoperiod. Environmental conditions for all glasshouse housed experiments were set within these limits (Table 2-3).

Table 2-3 Glasshouse environmental conditions

Parameter Setting Day length 16 hrs Supplementary lighting 400 watt MBF/U Temperature, 24 hrs 18 °C Vent temperature 21 it Irrigation Rainwater

2.1.2.3 Inoculation procedure A simple inoculation procedure using liverwort gemmae was devised for stock production of new colonies. Pots of liverwort were flooded with water which was decanted into a jug and poured over the surface of pre-irrigated compost, avoiding compost disturbance, preventing the gemmae from being covered over.

2.1.3 In vitro liverwort culture In order to measure the effect of various factors (e.g. light levels, temperature, glucosinolate hydrolysis products, fungal antagonists) on the growth and development of liverworts a reliable in vitro culture method was developed.

70 General procedures and method development Chapter 2

2.1.3.1 Media selection method development 2.1.3.1.1 Aim To identify a media suitable for in vitro cultivation of M. polymorpha from gemmae.

2.1.3.1.2 Introduction The effects of three media on liverwort gemma growth and development were examined: M51C, Murashige and Skoog with vitamins (MS) (Supplier: Duchefa), and Murashige and Skoog Shoot Multiplication Medium C (MSMC) (Supplier: Sigma Aldrich)

A literature review identified a nutrient medium M51C (Table 2-5) as suitable for M. polymorpha cell suspension culture (Ono et al., 1979), consisting of a base medium and Gamborg's 'B5' nutrients (Gamborg, 1975), pH 5.8. This was modified with the addition of agar and omission of 2,4-D, and was used to cultivate M. polymorpha from gemmae (Fujisawa et al., 2003; Ishizaki et al., 2002; Okada et al., 2000; Takenaka et a/., 2000).

MSMC media, used for Boston fern cultivation, has fewer vitamins but more plant hormones than M51C. MS media is often used at half the recommended concentration for higher plants, and has a higher magnesium sulphate (MgSO4.7H20), potassium hydrogen phosphate (KH2PO4), boric acid (H3B04) and hydrated zinc sulphate (ZnSO4.7H20) content than M51C media. MS and MSMC media are produced commercially; M51C has to be mixed from the individual components.

2.1.3.1.3 Method Media preparation An initial assessment of liverwort growth was made using the three different media. 12 petri dishes were filled with each (a total of 36) and a single gemma that had previously been rinsed 10 times in sterile water placed in the centre of each. Phytagel (Supplier: Sigma Aldrich) was used as the gelling agent in place of agar; agar has impurities toxic to some bryophytes, and does not have the transparency of phytagel (Duckett et al., 2004).

71 General procedures and method development Chapter 2

MS and MSMC media Pre-prepared granules were dissolved in distilled water according to the manufactures instructions (Table 2-4), the pH adjusted (5.8) and autoclaved (Table 2-8). MS media was used at half the recommended concentration

Table 2-4 MS and MSMC media preparation

Media Media (g L-1) Sucrose (g 1-1) Phytagel (g L-1) MS with Vitamins 2.2 20.0 3.0 MSMC 34.7 3.0

M51C media Stock solutions of the macronutrients, micronutrients, vitamins and potassium iodide were prepared in quantities to produce the final concentrations indicated (Table 2-5). The remaining ingredients (except phytagel and the stock solutions) were combined, dissolved in distilled water (500 ml), stirred, made up to 1 L with water the pH adjusted (5.8) using sodium hydroxide or hydrochloric acid, phytagel added, heated until dissolved and autoclaved (Table 2-8).

A number of replicates of each media became contaminated and were discarded. The remaining liverworts were assessed morphologically to identify which of the media provided the correct nutrient balance for normal liverwort growth, i.e. that most comparable to that exhibited by liverworts grown in the glasshouse on peat-based growing media. The only liverwort exhibiting 'normal' growth and morphology was on the M51C media, growing to 10 mm diameter after 19 days and producing gemma cups by 26 days.

For the liverworts grown on MSMC media, the surviving thalli were predominately unstructured, blob-like and fragmented into more than one part; four produced some rhizoids. Sizes ranged from less than 1 mm to 3 mm after 19 days. Four liverworts grown on MS media survived. They were predominately light green with no rhizoids, and one had fragmented into 3 unstructured portions. Sizes ranged between 2 mm and 11 mm diameter after 19 days. The liverwort gemmae grown on M51C were morphologically most normal and appeared healthy.

72 General procedures and method development Chapter 2

Table 2-5 In vitro liverwort cultivation media components

Stock Ingredients M51C MS with vitamins MSMC Supplier solutions (mg L'1) (mg I:1) (mg I:" NH4NO3 400 1650 1650 KNO3 2,000 1900 1900 Macro- MgS0.4.7H20 370 180.54 180.7 nutrients KH2PO4 275 170 170 FeSO4.7H20 27.8 MnSO4.H20 10 16.9 16.9 ZnSO4.7H20 2 8.6 8.6 H3B03 3 6.2 6.2 Micro- NaMo0.4.2H20 0.25 0.25 0.25 nutrients CuSO4.5H20 0.025 0.025 0.025 CoC12.6H20 0.025 0.025 0.025 NaH2PO4.H20 148 Nicotinic Acid 1 0.5 Sigma Aldrich Thiamine.HCI 10 0.1 0.4 Sigma Aldrich Vitamins Pyridoxine.Hcl 1 0.5 Sigma Aldrich My0-Inositol 100 100 100 Sigma Aldrich Glycine 2 KI 0.75 0.83 0.83 CaCl2.2H20 300 332.2 332.2 EDTA-FE 40 36.7 37.26 BDH Other N-Z Amine (A) 200 Sigma Aldrich Ingredients L-Glutamine 292 - Sigma Aldrich Sucrose 20000 20000 20000 Phytagel 3000 3000 3000 Sigma Aldrich Adenine hemi-sulphate 80 Hormones Kinetin 1 NAA 0.1

73 General procedures and method development Chapter 2

2.1.3.2 Liverwort sterilisation treatment 2.1.3.2.1 Aim To identify a sterilisation process for aseptic in vitro cultivation of M. po/ymorpha from gemmae.

2.1.3.2.2 Introduction For aseptic liverwort culture a sterilisation method harmless to fragile liverwort gemmae was required. Some researchers consider it impossible to surface sterilise gemmae (Taren, 1958); others use a variety of methods, four of which were selected for assessment.

2.1.3.2.3 Method

Six gemma cups were removed from male and female liverwort thalli, placed into an Eppendorf, washed in unsterilised water, shaken gently and the gemma cup tissue removed, leaving the gemmae in the water. 40 Petri dishes were prepared with M51C media and a quarter of the gemmae sterilised using each of the sterilisation treatments listed below; a single gemma was placed in the centre of each Petri dish, 10 dishes per sterilisation treatment, sealed with Parafilin PM-922 sealing film (Supplier: American National Can) and placed in the growth cabinet (Table 2-7).

Sterilisation treatments: 1. Washed 10 times with sterile water, adding Tween-20 and centrifuged (1 min, 2000 rpm) between each wash (Schwabe, 1951). 2. Washed with 3% sodium hypochlorite bleach solution for 10 min, then rinsed 3 times in sterile water (Schneider et al., 1967a). 3. Soaked in 1.4% potassium permanganate (KmnO4) (Supplier: Sigma Aldrich) for 10 min, filtered and washed three times with sterile water (Gorham, 1978). 4. Washed with 70% alcohol (10-60 seconds) and rinsed 3 times in sterile water (Kandasamy et al., 1996).

74 General procedures and method development Chapter 2

2.1.3.2.4 Results Few of the liverworts survived the 70% alcohol, potassium permanganate and sterile water treatments; of the survivors, those treated with sterile water and potassium permanganate were comparable in size to those treated with bleach (Table 2-6). The thalli of those treated with bleach were pale green and curled at the edges suggesting the bleach was too concentrated. All gemmae survived without contamination in a subsequent assessment using 2% and 1.5% bleach; therefore 1.5% bleach dilution was used.

Table 2-6. The effect of sterilization treatments on gemma size and survival.

Treatment Average gemma width No. of gemmae surviving (mm) (out of 10 per treatment) 70% alcohol 5.5 2 3% bleach 13.4 10 10 sterile water rinses 13.5 2 Potassium permanganate 10.0 1

2.1.3.3 Growth cabinet environmental conditions Liverwort cultures on M51C media were maintained in growth cabinets with short day lengths to maintain the liverworts in a non-sexual state (Wann, 1925) (refer to

Table 2-7) with a slow growth rate.

Table 2-7 Standard growth cabinet conditions

Parameters Setting Temperature 18 oc

Light levels 49 p mol nn 2 st

Photoperiod 8 hrs

2.1.3.4 Inoculation procedure Liverworts were subcultured in vitro by removing a number of gemmae from gemma cups of aseptically-grown subjects (keeping male or female gemmae separate), placed in an Eppendorf containing sterile water and shaken to separate gemmae. A small

General procedures and method development Chapter 2

number were poured onto sterile tissue and an individual gemma placed onto the centre of the media in each container using a sterile teasing needle. The containers used were clear, plastic, lidded, Bellaplast pots, 10 cm diameter, 4.5 cm deep providing room for growth. The lids were not airtight (Supplier: W K Thomas & Co. Ltd.).

2.1.4 Autoclave procedures Media was autoclaved to ensure sterility using standard procedures (Table 2-8). The compost used for the fungal experiment (Chapter 6) was autoclaved to eliminate fungi other than those applied intentionally, in 1 L quantities in open autoclave bags to allow steam penetration and sealed immediately on completion of the process.

Table 2-8. Autoclave procedures (Haines, 2006; Kandasamy etal., 1996). Media Temp. (°C) Time (mins) Agar & phytogel 121 20

Compost 126 30

2.1.5 Data analysis 2.1.5.1 Statistical analysis Statistical analysis was carried out using: • Genstat for Windows, 8th Edition, Release 8.2 (PC/Windows XP). Copyright 2005, Lawes Agricultural Trust (Rothamsted Experimental Station). • SPSS Release 11.0.1 for windows. Copyright SPSS Inc 1989-2001 (Chapter 3.2.2) • SPSS Release 12.0 Copyright SPSS (Chapter 3.2.1)

2.1.5.2 Image.] data analysis Image) is a free, downloadable digital analysis software package, developed at the National Institute of Health (National Institute of Mental Health, Bethesda, Maryland, USA), and available at http://rsb.info.nih.gov/ij/.

This software allows measurement and count of subjects contained within digital photographs individually or in groups and was used in this project for measurement of areas: water droplet area and liverwort gemmaling surface area (Chapters 3 and 5) also liverwort colonisation and fungal infection (Chapter 6).

76 General procedures and method development Chapter 2

2.2 Analytical methods 2.2.1 Thin layer chromatography Thin layer chromatography (TLC) was used to purify lunularic acid (Chapter 3) and 3- methoxybenzylisothiocyanate (Chapter 5) using Whatman Partisil K6 TLC chromatography plates. The choice of solvent system used to separate sample components depended on the substance being analysed. Table 2-9 indicates the solvents used in this study and their polarities.

Table 2-9. Polarity of solvents (Murov, 2007)

Solvent Relative Polarity Acetic acid 0.648 Chloroform 0.259 Ethyl acetate 0.228 Toluene 0.099

A pencil line was drawn on the TLC plate approximately 2.5 cm from the bottom, marking the spotting line (Figure 2-2). 20 pl of each sample was applied in even streaks along the start line, producing a narrow line, building up layers of each. Sufficient solvent was poured into a chromatography tank to cover the bottom of the TLC plate which was placed vertically into the solvent with the start line at the bottom. When the solvent front had almost reached the top of the plate its position was marked with pencil, the plate removed and placed vertically in a fume cupboard to dry. It was then viewed under ultraviolet light and the positions of compound fronts marked. Retention factors (Rf) (Figure 2-3) were calculated to help identify separated compounds as those with identical Rf values on the same TLC plate may be the same compound. The area between the spotting line and solvent front line was divided into 16 sections, 2cm apart, marked out with pencil (Figure 2-2). The silica gel containing the separated compounds between each pencil line was scraped off the plate, transferred to small conical flasks, the eluate extracted with methanol (20m1), refrigerated overnight and filtered to remove the silica gel. The methanol was removed by evaporation at reduced pressure (Buchi RE120 rotoevaporator), the dried filtrate dissolved in a small amount of methanol, transferred to small, pre-weighed vials and the remaining methanol remove with a Techne Dri-Block DB-3A sample concentrator.

77 General procedures and method development Chapter 2

Solvent 8 front 7 7 2 cm

6 6

5 5

4 4

3 3

2 2

1 1 Sample 1 Sample 2 — 4 Spotting line 2.5 cm

Figure 2-2 Chromatography plate layout

distance travelled by the compound Rf = distance travelled by the solvent front

Figure 2-3 Calculation of the retention factor

2.2.2 Gas chromatography (GC-MS) Gas chromatography-mass spectography (GC-MS) was used to separate compounds and identify their components. In this study a capillary column was used in 'split' mode due to its limited capacity, so that only a small proportion of each sample was injected onto the column. Gas chromatography (GC) equipment and operating conditions (Table 2-10 and Table 2-11) were standard for isothiocyanate analyses, however, samples were injected with either a 3:1, 5:1 or 50:1 split depending on isothiocyanate abundance, to prevent overloading the column. GC data was analysed using MSD ChemStation Build 75, 26 August 2-3, Version G1701DA D.01.00., Agilent Technologies. Spectra were compared with known standards or by comparison with the Wiley 275 mass spectra library.

78 General procedures and method development Chapter 2

Table 2-10. Gas chromatography equipment

GC System HP 6890 Detector HP 5973 mass selective detector Column Agilent HP-5ms GC 30m x 250 pm internal diameter Column dimensions 0.25 pm film thickness Carrier gas Helium

2.2.3 High performance liquid chromatography (HPLC) system Reversed phase HPLC techniques were used for lunularic acid (Chapter 3) and glucosinolate (Chapter 5) sample analysis. The solvent system, delivery programs and detectors were varied to separate and detect the compounds being analysed. The Waters Chromatography System comprised an Automated Gradient Controller 680 and two Chromatography pumps: pump A (Waters 510 HPLC pump) and pump B (Waters chromatography pump) (solvent flow rate 1 ml min-'). A Phenomenex SphereClone 5p ODS (2), Size 250 mm x 4.6 mm x 5 micron, reverse phase column was used at 35 °C, containing classic C18 reverse phase material. The detectors used were a Waters 486 tuneable absorbance detector (UV) and a Merck/Hitachi F-1050 fluorescence spectrophotometer (FL) detector, for lunularic acid analysis. As most molecules do not exhibit fluorescence this can be a sensitive tool; the structure of lunularic acid (Chapter 1), incorporating phenyl, carboxyl and hydroxide groups suggested that fluorescence detection would elicit a good response. Jones Chromatography JCL6000 for Windows (version 2.0) detector and data manipulation software was used. The area under each peak on the HPLC trace produced was estimated using integration software and the concentration estimated from the pure lunularic acid sample.

79 General procedures and method development Chapter 2

Table 2-11. Gas chromatography methods 40 °C, equilibration time 0.5 min Initial tempterature 40 °C, then held for 5 min Temperature 5 °C min-1 to 180 °C parameters Temperature increases Method 10 °C min-1 to 280 °C

3:1 split Final oven temp 280 °C, held for 10 min MSD transfer line heater 290 °C Injection volume 1.0 pl. Run time 53 min 125 °C, equilibration time 0.5 min Initial temperature Temperature 125 °C, then held for 5 min parameters Temperature increases 10 °C min-1 to 280 °C Method Final oven temp 280 °C, held for 5 min 50:1 split MSD transfer line heater 290 °C Injection volume 1.0 pl Run time 25.50 min 50 °C, equilibration time 0.5 min Initial temperature 50 °C, then held for 5 min Temperature 5 °C min-1 to 180 °C parameters Temperature increases Method 10 °C min-1 to 280 °C

5:1 split Final oven temp 280 °C held for 10 min MSD transfer line heater 290 °C Run time 51 min Injection volume 1.0 pl

80 The effect of environment on liverwort establishment, growth and development Chapter 3

Chapter 3 The effects of environment on liverwort establishment, growth and development

3.1 Introduction

Vegetative growth There is limited information in the literature concerning the effect of environment on Marchantia po/ymorpha growth and development. Maximum vegetative growth has been achieved under medium light intensity (3500 lux), whilst high light (5000 lux) suppresses growth, suggesting that photosynthesis and growth were saturated at 3500 lux, and inhibited by an excess of light at 5000 lux (Terui, 1981). Mache and Loiseaux (1973) found light levels above 6000 lux limited growth, with morphological changes occurring in the thallus and the structure of the chloroplasts (Chapter 1).

Number of gemma cups In experiments using Lunularia cruciata it was found that thalli produce more gemma cups in high light intensities, but only if accompanied by low temperatures. They also found little effect due to light intensity 1345 lux and 247 lux on thallus growth or branching, or gemma cup production, with an average of 0.025 cups mm-2 of new growth (Nachmony-Bascomb and Schwabe, 1963).

An effect due to temperature was also reported, with more gemma cups produced at lower temperatures, 0.43 cups per new thallus branch at 12 °C compared with 0.11

81 The effect of environment on liverwort establishment, growth and development Chapter 3 cups at 18 °C for L. cruciata. Temperatures above 24 °C affected growth detrimentally, resulting in fewer new branches that were also shorter and smaller overall (Nachmony- Bascomb and Schwabe, 1963).

Voth (1941) reported on experiments using separate male and female M. polymorpha thalli grown under comparable conditions. More gemma cups were produced on male than female plants in a ratio of 6.6:1 across all nutrient treatments. Female thalli had fewer gemma cups, a broader thallus tip and smoother surfaces than male thalli when grown in the conditions of this experiment. The margins of the male plant were more undulating and curved downwards, particularly in basic conditions (Voth, 1941).

An effect due to photoperiod has also been found on liverwort growth and gemma cup production, with significantly greater thallus length, dry weight and fresh weight of M. po/ymorpha gemmalings (young, germinated gemmae) produced under long days than short observed by a number of researchers: (Carter and Romine, 1969; Hedger et at., 1972; Miller et al, 1962; Voth, 1943; Voth and Hamner, 1940). With L. cruciata growth slows under long days and stops under continuous light, a condition that is reversible, with growth resuming under short-days (Nachmony-Bascomb and Schwabe, 1963).

More gemmae cups were produced per unit area on plants grown under short days in several experiments using M. po/ymorpha (Voth, 1943). Similarly, for L. cruciata more gemma cups were produced under short-days at 18 °C and 24 °C (Schwabe and Nachmony-Bascomb, 1963). Conversely, Terui (1981) found that for M. po/ymorpha gemma cup production was promoted under long day conditions.

Lunularic acid Lunularic acid is thought to be an endogenous growth regulator, with light and temperature levels and photoperiod (Chapter 1) affecting its formation, and consequently the growth rate of liverwort. Characterisation of the lunularic acid content of M. po/ymorpha throughout its life cycle in relation to ambient nursery conditions of light, temperature and photoperiod could then be related to liverwort development. For example, increased gemmae production during short days (Voth, 1943) and increased gametophore production during long days (Wann, 1925) could potentially contribute to liverwort control strategies involving the manipulation of the growth environment. For

82 The effect of environment on liverwort establishment, growth and development Chapter 3 example, knowing that LNA content is high during long, hot days with high light conditions, thereby reducing liverwort growth, liverwort removal could be timed to coincide with these conditions in the knowledge that growth is slower, therefore delaying and reducing re-establishment.

3.2 Experimental work

3.2.1 The effect of light level and temperature on the growth and development of Marchantia polymorpha (25 °C and 15 °C) 3.2.1.1 Introduction An experiment using four Fitotron growth cabinets compared the effect of two different temperatures (25 °C and 15 °C) and two light levels (800 pmol m-2 st and 400 pmol m- 2 S-1) on the growth (radial expansion), dry weight accumulation and development of male and female liverworts over 6 weeks.

3.2.1.2 Method Temperature was replicated at cabinet level, with two cabinets at 25 °C and two at 15

°C. Within each cabinet the shelf heights were adjusted to provide two light levels, the upper shelf: 800 pmol m-2 s-1 and the lower shelf: 400 pmol m-2 s-1; this arrangement did not allow for random allocation of light treatments. For each temperature, equal numbers of male and female gemmae were placed in separate groups. Humidity was set at 65% with a 12 hour photoperiod.

288 Al. po/ymorpha gemmae (72 per cabinet) were placed individually onto M51C media (Section 2.3.1.3) in lidded 10 cm diameter Bellaplast pots (Supplier: W K Thomas & Company Ltd.), one gemma per pot. Placement of male and female groups was randomly allocated, an example of the within cabinet layout is shown in Figure 3-1.

83 The effect of environment on liverwort establishment, growth and development Chapter 3

MMMMMM F F F F F F High Light MMMMMM F F F F F F Treatment MMMMMM F F F F F F

F F F F F F MMMMMM Low Light F F F F F F MMMMMM Treatment F F F F F F MMMMMM Figure 3-1. Example of the layout within each Fitotron cabinet. Shelves containing high light treatment (800 pmol m-2 s-1) were positioned higher (nearer to the fluorescent tubes) than those containing the low light treatment levels (400 pmol m-2 s-1). M and F refer to male and female liverwort gemmae respectively.

Data collection Three randomly selected pots were removed from each treatment on 6 occasions, at 7- day intervals (48 pots); 3 male and 3 female gemmalings from each light treatment (a gemmaling is young liverwort developing from a gemma). Each gemmaling was placed on a light box, photographed using a Nikon Coolpix 995 digital camera and its area calculated by digital image analysis using Image] software (Chapter 2.7.2).

The number of gemma cups present on each gemmaling was counted and once three had developed the number of gemmae they contained was estimated by transferring three gemma cups containing gemmae into an Eppendorf containing 500 pL water and a drop of Tween 20 to reduce surface tension. The gemma cup tissue was removed, 200 pL water containing gemmae transferred to a piece of filter paper and the gemmae counted. The fresh and dry weight of the final set of replicates was also recorded.

84 The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.1.3 Results Growth For statistical analysis of gemmaling growth (radial expansion) a Univariate General Linear Model (SPSS 12.0) was used, with data transformed to natural logarithms. Results indicated that the effect on growth due to temperature, light gender and week were significant at the 0.1% level as were interactions between temperature and light and temperature and gender (Table 3-1).

Graphs (Figure 3-2) show an interaction between temperature and week as the trend lines are not parallel, intersecting at week 1 at 15 °C and week 2 in the 25 °C treatment. Comparison of data points on the two graphs corresponding to low light at 15 °C, and then 25 °C, the interaction between light levels and temperatures, can be seen.

Table 3-1. Statistical analysis of gemmaling growth by applying the Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interections; data transformed to natural logs. Significant results only are presented. R2 = 0.865 (Adjusted R2 = 0.854) df v.r. p-value Temperature 1 173.474*** <0.001 Light 1 147.863*** <0.001 Gender 1 11.721*** <0.001 Week 5 227.058*** <0.001 Temp x light 1 46.665*** <0.001 Temp x gender 1 13.973*** <0.001 Temp x week 5 5.961*** <0.001 Light x gender 1 6.665*** <0.001 Light x week 5 16.347** <0.01

The growing conditions had a different effect on male than female gemmaling growth (Figure 3-3); in the high temperature treatments female gemmalings grew larger than male in low light, however under high light conditions the reverse is seen. Although there is no intersection of trend lines in the graph for 15 °C the lines are not parallel, indicating an interaction. The interactions between gender, temperature and light were significant at the 0.1% level.

85 111 Figure 3-2.Gemmaling growth (In)at(a)15 natural logs. The effectofenvironmentonliverwortestablishment,growthanddevelopment -2 S -1 , low

Average growth (In) Average growth (In) light =400limo! Week 123456 Week 12 .

, M -2 S Week 34 -1 . .

Growth (mm)datawas transformedusing Week 86 Week ° C and(b)25 i ° C. Highlight=800

Chapter 3 IJMOI (mm) datawastransformedusingnaturallogs. Figure 3-3.Growth(In)ofmaleandfemalegemmalings attemperaturesof(a)15 and (b)25 The effectofenvironmentonliverwortestablishment,growthanddevelopment

Average growth (In) ° C. Highlight=800pmolm Low light i

-2 Light level Light level 87 s -1 , low light=400pmolm High light

-2 s -1 . Growth Chapter 3 ° C The effect of environment on liverwort establishment, growth and development Chapter 3

2 3 4 5 6

Week no.

2 3 4 5 6

Week no.

Figure 3-4. Relative growth (radial expansion) rate of gemmalings in (a) 15 °C (b) 25 °C. High light = 800 pmol M-2 S-1, low light = 400 pmol m-2 s-1

88 The effect of environment on liverwort establishment, growth and development Chapter 3

Relative growth At the higher temperature, gemmalings tended to have a high relative growth rate up to week 2 when compared to other treatments (Figure 3-4); levels fell sharply during week 2 until they were similar to other treatments. In low light conditions relative growth rate was more constant over six weeks, although in high light there were two occasions of increased relative growth rate during weeks 3 and 5. The gemmalings with the more constant, slower growth in low light conditions were those that appeared to be more morphologically 'normal' (Figure 3-5).

10 °C, female, 15 °C, male, IR high light low light phillifililli1111111111111111i11111111111111 [ 11[1 CM 1 3 4 7 a

25 °C, male, 25 °C, male, high light low light

IIII11111111111111111111111111111 1 I

12 13 14 15 I, tz u 11 0 0 17 0 10

Figure 3-5 Comparison of liverwort gemmalings after 6 weeks

Fresh weight Fresh weight at week 6 was analysed statistically by applying a Univariate Weighted General Linear Model using SPSS 12.0 (Table 3-2 and Table 3-3). The effect of light on fresh weight was significant at the 0.1% level; the interactions between temperature and light, and temperature and gender were significant at the 1% and 5% levels, respectively.

Table 3-2. Statistical analysis of gemmaling fresh weight after 6 weeks analysed by applying a Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interactions. Significant results only are presented. R2 = 0.746 (Adjusted R2 = 0.711). df v.r. p-value Light 1 88.629*** <0.001 Temp x light 1 8.564** <0.01 Temp x gender 1 4.725* <0.05

89 The effect of environment on liverwort establishment, growth and development Chapter 3

Table 3-3. Statistical analysis of gemmaling dry weight after 6 weeks by applying a Univariate General Linear Model, using SPSS 12.0, showing effects of treatments and interactions. Significant results only are presented. R2 = 0.642 (Adjusted R2 = 0.623).

df v.r. p-value Temp 1 19.187*** <0.001 Light 1 29.187*** <0.001

At both 15 °C and 25 °C fresh weights after six weeks were greater when grown under low light levels compared to high light (Figure 3-6). In low light, growth was greater at low temperature than high; and conversely, under high light, growth was greater in high temperatures than low. Both male and female gemmalings had greater fresh weight in low light levels at both temperature treatments; however there was an interaction between temperature and gender on growth at 15 °C, whereby male gemmalings were larger than female.

Dry weight At both 15 °C and 25 °C dry weights after six weeks were greater when grown under low light levels compared to high light, and greater when grown at 25°C than 15 °C. Statistically there was a significant effect of temperature and light on dry weight at the 0.1% level; however there was no interaction between light and temperature, as can be seen in the parallel lines of the graph (Figure 3-7).

Development Liverwort development was also monitored, with the presence and number of gemma cups on each gemmaling used as a measurement. Counts of gemma cups was statistically analysed using a Negative Binomial Regression Model using STATA 8.2. Temperature was not found to be significant; however interactions of temperature with light and week were significant (Table 3-4). Gemmaling development was faster at 25 °C, with the first gemma cup produced in week 2 at both light levels (Figure 3-8). More gemma cups were produced at the low light treatments than high light. During week six there was a large increase in the number of gemma cups present in the 15 °C treatment, where more gemma cups were produced overall.

90 The effectofenvironmentonliverwortestablishment,growthanddevelopment High light=800 Figure 3-6.Freshweight of(a)femaleand(b)malegemmalingsat15

Fresh weight (g) Average fresh weight (g) 1.0 - 1.5 - 0.0 - 0.5 - 2.0 - 2.5 1.1M01 M —0— Lowlight —•— Highlight -2 5 1 , low 15 15 light =400pmolm

Temperature ( Temperature ( 91 ° ° C) C) 25 25 -2 s -1

(b)

° C and25 Chapter 3 ° C. The effect of environment on liverwort establishment, growth and development Chapter 3

15 25

Temperature (C)

Figure 3-7. Dry weight of gemmalings after 6 weeks of growth. HL=high light (800 pmol m-2 S-1), LL=low light (400 pmol M-2 S-1) Temperature=15 °C & 10 °C.

Table 3-4. Statistical analysis of no. of gemma cups by applying the Negative Binomial Regression Model, using STATA 8.2, showing effects of treatments and interactions. Significant results only are presented. coef. and std. errors p-value Light -2.54 ±0.28 <0.001 Gender 0.74 ±0.15 <0.001 week 3 -2.64 ±0.40 <0.001 week 4 -1.27 ±0.31 <0.001 week 5 -0.62 ±0.30 <0.05 Temp x light 2.07 ±0.34 <0.001 Temp x week 3 1.45 ±0.49 <0.01 Temp x week 4 1.07 ±0.41 <0.01

92 The effectofenvironmentonliverwortestablishment,growthanddevelopment High light=800pmolM Figure 3-8.No.ofgemma cupsproducedattemperaturesof(a)15

Average no. of gemma cups Average no. of gemma cups Week 123456 Week 123 Week 456 -2 S t , lowlight=400m 93 Week Week ij o i m -2 s -1

° C (b)25 Chapter 3 ° C. The effect of environment on liverwort establishment, growth and development Chapter 3

There was an effect on the number of gemma cups produced depending on liverwort gender, with the results significant at the 0.1% level. More gemma cups were produced on male plants than female at both light levels. Some gemmae produced within each gemma cup were found to clump together, making it difficult to estimate numbers. This aspect of liverwort development was investigated further (Chapter 4).

Radial expansion of the gemmalings was used as the measure of growth as the majority of growth is two dimensional; fresh and dry weight measurements were found to be consistent with the radial expansion measurements.

Observations were made concerning the morphology and colour of the gemmalings throughout the experiment. By week 5 those grown under high light and in high temperature conditions were a dull green colour; by week 6, 8 of the 12 replicates were developing a dark brown colouration with a reduced relative growth rate (Figure 3-5). The thallus had a dome-shaped appearance, rather than growing flat over the surface of the media.

Gemmalings grown under lower temperature and light conditions were flatter and had a lighter green colour. Although relative growth rates were variable they were sustained and, on occasions, increased to produce liverworts with more normal morphology and fresh weights similar to the high temperature, low light treatment after six weeks. These replicates may have had the potential to reach greater sizes than those grown in higher light and temperature treatments, given a longer growth period.

The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.2 The effect of light level and temperature on the growth and development of male and female Marchantia polymorpha at 15 °C and 10 °C 3.2.2.1 Introduction A second growth and development experiment using four Fitotron growth cabinets investigated growth and development at 15 °C and 10 °C; the temperature range was thus extended, with 15 °C used as a common reference point. Light levels remained unaltered at 800 pmol m-2 st (high light) and 400 pmol m-2 s-1 (low light).

3.2.2.2 Method This experiment compared the effect of two different temperatures (15 °C and 10 °C) and two light levels (800 pmol m-2 S-1 and 400 pmol m-2 s-1) on the growth (radial expansion), fresh and dry weight accumulation and development of male and female liverworts. The experimental design was as previously used, with cabinet, temperature, and liverwort gender randomly allocated.

Data was recorded after 4 and 6 weeks growth; nine randomly selected pots were removed from each treatment (144 in total) after 4 weeks for data collection and data was collected from the remaining replicates after 6 weeks. Area, fresh and dry weights and the number of gemma cups present on each gemmaling were recorded. Liverworts were photographed using a Nikon Coolpix 995 digital camera and their areas calculated using Image.] software.

3.2.2.3 Results Dry weight, fresh weight, growth (radial expansion), and no. of gemma cups were all greater in high temperature than low; greater in low light than high light after six weeks (Figure 3-9 to Figure 3-12). Statistical analysis was carried out using a Univariate General Linear Model, using SPSS 11.0; significant results only are presented (Table 3-5).

Growth There was a very highly significant effect of light and temperature on growth (radial expansion) after six weeks (Figure 3-9) with greater growth at 15 °C than 10 °C under both light levels; and greater growth in low light under both temperatures. There was also a very highly significant interaction between light and temperature; although there was greater growth under low light than high at both 15 °C and 10 °C, it was greatest

95 The effect of environment on liverwort establishment, growth and development Chapter 3 in the 15 °C, low light treatment. There was also a highly significant effect of gender, with greater growth by female than male gemmalings.

Table 3-5. Statistical analysis of results after 6 weeks analysed by applying a Univariate General Linear Model, using SPSS 11.0. Only significant results are presented. Parameter df v.r p-value

Growth Light 1 114.081 <0.001 Temperature 1 103.590 <0.001

(R2 = 73.1 Temp x light 1 44.709 <0.001 Adjusted R2 = 69.3) Gender 1 7.093 0.009

Fresh weight Light 1 84.787 <0.001 Temperature 1 118.584 <0.001

(R2 = 72.1 Temp x light 1 36.440 <0.001 Adjusted R2 = 68.2) Temp x gender 1 4.521 0.036 Gender 1 18.822 <0.001 Dry weight Light 1 168.785 <0.001 Temperature 1 42.777 <0.001 (R2 = 71.2 Gender x temp 1 15.255 <0.001 Adjusted R2 = 67.1) Temp x light 1 7.963 0.006 No. gemma cups Light 1 60.105 <0.001

(R2 = 61.5 Temperature 1 86.877 <0.001 Adjusted R2 = 55.7)

Dry weight At week six the effect of temperature, light and gender were all very highly significant, as were the interactions between temperature and light, and temperature and gender. Dry weight was greater in low light than high at both temperatures, and greater in high temperature than low. There was a significant interaction between light and temperature, with markedly greater dry weight observed in high temperature, low light than other conditions.

There was also a significant effect of gender, with a significant interaction between temperature and gender. Greater growth of gemmalings was seen at 15 °C than 10 °C, with little growth of either male or female gemmalings at 10 °C; At 15 °C greatest growth was observed in female than male gemmalings.

96 1 The effectofenvironmentonliverwortestablishment,growthanddevelopment at 10 light (800pmolm Figure 3-10.Dryweight ofgemmalingsafter4and6weeksgrowth. HL=high Figure 3-9.Growth(radialexpansion)ofgemmalingsafter4and6weeksgrowth )

° C and15

Dry weight (mg) ° C. HL=highlight(800pmolM -2 S -1 ), LL=lowlight(400pmol m 10-HL 10-HL

10-LL 10-LL 97 Treatment

Treatment

-2 15-HL S 15-HL -2 -1 ), LL=lowlight(400pmolm S

-1

) Temperature=15 15-LL 15-LL

° C &10 Chapter 3 -2 ° s C - Oc Figure 3-11.Freshweightofgemmalingsafter 4and6weeksofgrowth.HL=high light (800pmolnI ° temperature andgender,withgreatergrowthoffemalethanmalegemmalingsat15 gemmaling growthwasalsogreatestunderlowlightthanhigh. levels andgreaterunderlowlightthanhighinbothtemperatures.However,the There werealsohighlysignificantinteractionsbetweenlightandgender, The effectofenvironmentonliverwortestablishment,growthanddevelopment high light,15 increase ingemmalingareawasgreaterunderlowlight,15 Fresh weight of lightandtemperature. At weeksixtherewasaveryhighlysignificanteffectofbothlightandtemperature, 10 with greatergrowthunderlowlightthanhighinbothtemperatures,andat15 C than10 ° C forbothlightlevels.Freshweightwasgreaterat15

Fresh weight (mg) ° ° C, withgreatestgrowthbyfemalegemmalingsat15 C treatments.Femalegemmalingswerelargerthanmaleinbothlevels 2 s -1 ), LL=lowlight(400pmolrn 10-HL

10-LL Treatment

98 15-HL

-2 S -1 ) Temperature=15 ° C than10 15-LL ° C thaneither10 ° C forbothlight

° C. Female ° ° Chapter 3 C &10 C than ° C or The effectofenvironmentonliverwortestablishment,growthanddevelopment temperatures thanlow.However,at10 Development The effectsoftemperaturewerethesamefornumbergemmacupsmm After sixweekstherewasaveryhighlysignificanteffectofbothlightandtemperature thallus. Again,withmoregemmacupspresentinlowlightthanhigh,andhigh cups perunitareawasless. gemma cupswereonlyproducedinthefemale,lowlighttreatment. on liverwortdevelopmentintermsofthenumbergemmacupspresent Temperature =15 growth. HL=highlight(800pmolm Figure 3-12.No.ofgemmacupspresentongemmalingsafter4and6weeks produced inhighlightthanlowafterfourandsixweeks(Figure3-13).Inthe were producedthanunderhighlightconditions (Figure3-12),thenumberofgemma temperature, lowlighttreatmentalthoughgrowthwasgreaterandmoregemmacups the exceptionofhightemperaturetreatmentwheremoregemmacupsmm

No. gemma cups ° C &10 ° C ° -2 Treatment C thegemmalingswereparticularlysmalland s 99 4 ), LL=lowlight(400pmolm

Chapter 3 -2 -2 -2 , with were s -1 ) The effect of environment on liverwort establishment, growth and development Chapter 3

ci E E to 0_ m 0 0.10 Co E E a) CD

1 1 10-LL 15-HL

Treatment

Figure 3-13. No. of gemma cups mm-2 after 4 and 6 weeks of growth. HL=high light (800 pmol m-2 s-1), LL = low light (400 pmol m-2 s-1) Temperature = 15 °C & 10 °C

3.2.3 The effect of lunularic acid on the growth and development of M. polymorpha - H PLC optimisation. 3.2.3.1 Introduction Lunularic acid is thought to be an endogenous plant growth regulator, predominately found in liverworts (Chapter 1). Higher concentrations are known to be produced in high light, high temperature conditions; however these environmental effects on liverwort growth have not been characterised for liverworts growing in plant nursery conditions, therefore it was considered that this aspect of liverwort biology merited further investigation in relation to experiments concerning the effect of growth and development in this study. As lunularic acid (LNA) is not produced commercially samples were initially extracted and purified using a known method (PhD thesis, (Valio, 1969) for use as an HPLC standard. A sample of lunularic acid subsequently provided by Professor Asakawa of Tokushima Bunri University, Japan as a gift was authenticated by mass spectroscopy (MWt 258). A simple extraction and quantification method was then developed and used to measure lunularic acid levels in different liverwort tissues when grown under nursery conditions.

HPLC techniques were developed to identify the optimum conditions for detecting LNA, initially using fluorescence (by producing fluorescence and excitation and emission

100 The effect of environment on liverwort establishment, growth and development Chapter 3 curves) and UV detectors; various solvent systems were tested to determine which produced a peak corresponding with LNA at an optimal retention time. The appropriate UV detector setting was known (285 nm) (Abe and Ohta, 1983). As most molecules do not exhibit fluorescence this can be an extremely sensitive tool; the structure of lunularic acid (Chapter 1), incorporating phenyl, carboxyl and hydroxide groups suggested that fluorescence would elicit a good response (Chapter 2).

3.2.3.2 Method Solvent system A stock solution of LNA (0.19 mg m1 1 ethanol) was diluted with solvent (H20 + 0.1% acetic acid + 95% acetonitrile) to give 19 pg m14 LNA. Four solution ratios of solvent A (H20 + 0.1% acetic acid) and solvent B (95% acetonitrile) (30:70, 40:60, 50:50, 60:40) were compared.

Fluorescence excitation/emission curves: A fluorescence excitation/emission curve was produced, taking energy level readings with a static emission wavelength (400 nm) and varying excitation wavelengths between. 250 nm and 365 nm (20 nm increments). This was repeated with a static excitation wavelength (200 nm) and varying emission wavelengths between 355 nm and 475 nm (20 nm increments). The results were plotted graphically (Figure 3-14) to identify the optimum detector settings.

3.2.3.3 Results A 60:40 solvent ratio produced a peak after 8 min (300 nm excitation, 400 nm emission wavelength fluorescence detector settings). The fluorescence excitation and emission curves identified optimum detector settings as 300 nm (excitation X) and 405 nm (emission X) (Figure 3-14). Although a single peak was produced by both detectors the response was 35 fold greater using the fluorescence detector (Figure 3-15) than ultra violet (Figure 3-16) confirming the greater sensitivity of this detection method with LNA.

101

The effect of environment on liverwort establishment, growth and development Chapter 3

250

200 -

150

E 100 - C9

50 -

o - Excitation spectrum Emission spectrum

200 250 300 350 400 450 500

Wavelength ? (nm)

Figure 3-14. HPLC analysis of lunularic acid using fluorescence detection. Excitation spectrum produced at 400 nm wavelength emission. Emission spectrum produced at 200 nm wavelength excitation.

mV 885.8 7.9

800.0 750.0 700.0 650.0 600.0 550.0 500.0 450.0

350.07 300.01 250.0 200.0, 150.0- 100.0: 50.01

-94.7 ll, 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0

Figure 3-15 HPLC chromatogram produced by lunularic acid, detected by fluorescence

102

The effect of environment on liverwort establishment, growth and development Chapter 3

25.0- 7.99 my

20.0

15.0

10.0

5.0-

0.0

3.4 I 00 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0

Figure 3-16. HPLC chromatograms produced by lunularic acid and detected by ultra violet.

3.2.4 Extraction and purification of lunularic acid 3.2.4.1 Introduction A method for extraction and purification of lunularic acid from freeze dried liverwort was developed with samples submitted to thin layer chromatography to separate and collect extracted LNA using a known method (Valio, 1969).

3.2.4.2 Method Freeze dried liverwort tissue (13 g) was ground in a Moulinex coffee grinder, placed into a conical flask with 80% methanol (500m1), stirred overnight (2 °C) and filtered (Hartley funnel) with Hyflo Supercel 'Filter Aid' (Supplier: BDH) (Figure 3-17). The filtrate was removed by rotary film evaporation at 35-40 °C under reduced pressure (Biichi Rotavapor RE120). The residue was made up to 15 ml with distilled water, acidified (pH3) using 5 M HCI acid and extracted 3 times with an equal volume of ethyl acetate. From this the aqueous fraction was used to obtain the basic fraction, and the organic fraction was used to obtain the neutral and acidic fractions.

Organic fraction The organic fraction was reduced to a small volume (50 ml) (rotary film evaporation at 35-40 °C under reduced pressure, Buchi Rotavapor RE120) and extracted 3 times with equal volumes of 5% sodium bicarbonate (NaHCO3), producing a neutral organic fraction and an aqueous fraction. The aqueous fraction was acidified (pH3, 5M HCI)

103 The effect of environment on liverwort establishment, growth and development Chapter 3 and extracted 3 times with equal volumes of EtOAc to produce an acidic organic fraction.

Aqueous fraction The aqueous fraction from the first extraction was basified (pH 8, NaHCO3) and extracted 3 times with equal volumes of EtOAc to produce a basic organic fraction.

Drying of organic fractions The three organic fractions (acidic, basic and neutral) were separately dried over magnesium sulphate (MgSO4), filtered and evaporated to dryness.

Purification of lunularic acid Quantities of the neutral (6 mg) and basic (5 mg) fractions were dissolved in EtOAc (150 pl), streaked onto a TLC plate separately and developed (Section 2.2.1); the acid fraction was discarded as the quantity was small and Valio (1969) found it contained little LNA. The solvents used initially (chloroform, ethyl acetate, acetic acid, 60:40:5) for TLC did not adequately separate the compounds, and so were replaced with a second solvent solution (toluene, ethyl acetate, acetic acid, 50:5:2) and the TLC repeated. After approx. 1 hr the solvent front had almost reached the top of the plate. Its position was marked, the plate dried, viewed under UV light, marked into sections which were scraped from the plate (Figure 2-2) and from which the separated compounds were extracted to produce 16 samples.

The process was repeated with a second TLC plate to further separate the compounds. The spotting line was marked in 16 places, spotted twice with each sample (2.5 pl x 2), the plate developed (solvent solution: toluene, ethyl acetate, acetic acid, 50:5:2), dried and viewed under UV light. Two spots were visible; the samples used to produce these

104

The effect of environment on liverwort establishment, growth and development Chapter 3

(Liverwort freeze dried, ground and mixed with 80% methanol.

• Refrigerated (2°C) overnight.

• Filtered using Hartley filter

Removed methanol under reduced pressure.

• Adjusted pH to 3 using 5M HCI

• Extracted 3 times using equal volumes of EtOAc

ORGANIC FRACTION AQUEOUS FRACTION Reduced to small volume under reduced pressure pH adjusted to 8 using solid 5% NaHCO3

• Extracted 3 times with equal volumes of 5% Extracted 3 times with equal volumes of .\ NaHCO3 EtOAc

t ORGANIC FRACTION AQUEOUS FRACTION H2O removed using MgSO4. Acidified to pH 3 using 5M MCI , ORGANIC FRACTION _of • H2O removed using MgSO4. Evaporated to dryness to give Extracted 3 times with EtOAc NEUTRAL FRACTION V 'Evaporated to dryness to give • BASIC FRACTION • ORGANIC FRACTION AQUEOUS FRACTION AQUEOUS FRACTION H2O removed using MgSO4. Discarded Discarded • Evaporated to dryness to give -‘ ACIDIC FRACTION

Figure 3-17. Summary of lunularic acid extraction method

105

The effect of environment on liverwort establishment, growth and development Chapter 3

spots were used to produce a third TLC plate, separating a larger quantity of the substance. Each was streaked separately along the spotting line, (20 pl quantity) and developed three times (solvent solution: toluene, ethyl acetate, acetic acid, 50:5:2).

3.2.4.3 Results An authentic sample of lunularic acid (LNA) sent as a gift (Professor Asakawa of Tokushima Bunri University, Japan) was used to confirm the identification of lunularic acid extracted.

3.2.5 Investigation into compounds suitable for use as an internal standard 3.2.5.1 Introduction A liverwort tissue sample preparation method was developed to produce clean samples containing lunularic acid for HPLC analysis. Salicylic acid (SA) was initially selected as an internal standard due to its similar structure to LNA (phenyl, carboxyl and hydroxide groups) to detect losses during sample preparation and for LNA quantification.

SA was found inappropriate as an internal standard therefore an alternative compound, 3-hydroxy-2-nahthoic acid (HNA) (CAS: 92-70-6), was investigated (Figure 3-18), the absorbance values are given in Table 3-6. HNA was selected due to its hydroxyl, carboxyl and two phenyl groups, which are arranged differently than in LNA and SA, suggesting it should have a different retention time whilst still being detectable by fluorescence.

OH Figure 3-18. Structure of 3-hydroxy-2-naphthoic acid (HNA)

106 The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.5.2 Method Salicylic acid (SA) as an internal standard Liverwort thallus (1 g) was ground using a pestle and mortar in a mixture of Me0H (10 ml) and glass fragments, transferred to a 100 ml conical flask with anti-bumping granules and a further 10 ml solvent. The extract was heated (steam bath, 15 min), filtered, the residue returned to a conical flask, solvent added (MeOH, 10 ml), re- heated (5 min), re-filtered, the two filtrates bulked and the process repeated to produce two samples, each of which was evaporated to a watery residue under reduced pressure (Buchi Rotavapor RE120, 35 °C). A salicylic acid (SA) standard (0.01 mg m1-1 ether, 2 ml) was added to one sample, made up to 10 ml with distilled water, extracted three times (diethyl ether, 10 ml), the upper (chlorophyll) layer removed each time. The extracts were combined and dried through a magnesium sulphate (MgSO4) column. The second sample was prepared without SA. For each sample an Amino-propyl solid phase extraction (SPE) column (Supplier: Alltech) was conditioned with ether (20 ml), filled with six volumes of sample and the liquid collected discarded. The column was washed with three volumes of dichloromethane (DCM): 2-propanol (2:1), the LNA eluted with MeOH:AcOH (95:5) (10m1) and the eluate collected. This was repeated with an SA standard (2 ml, 0.01 mg m1-1 ether) producing a third sample. Each sample was evaporated under reduced pressure (Bach' Rotavapor RE120, 35 °C) azeotroped twice with butan-1-ol (2 ml, 50 °C) to remove the acetic acid, and the residue dissolved in acetonitrile (1 ml) for analysis (HPLC). An LNA standard (19 pg ml-

1 solvent, Section 3.2.3.1) was also analysed (HPLC).

3-hydroxy-2-nahthoic acid (HNA) as an internal standard A standard solution of HNA (0.01 mg m1-1) in ether was not detected when subjected to HPLC analysis. To optimise the analysis the UV absorbance spectrum for HNA was produced (spectrophotometer: Shimazu UV-210-PC UV-VIS) to identify the UV wavelengths that produced the largest peaks, indicating maximum absorbances (?max) (Table 3-6) to optimise the UV detector calibration.

To test this, HNA standard (0.1 mg m1-1) was prepared in HPLC solvent (solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40) and analysed with the UV detector at 360, 284 and 310 nm, from which 284 nm was adopted for the UV setting. The analysis was repeated firstly with a higher concentration of HNA (1.0 mg m1-1) then purified (99.99% pure) HNA (1 mg m1-1 and 0.1 mg m1-1) in acetonitrile (UV

107 The effect of environment on liverwort establishment, growth and development Chapter 3 detector, 284 nm; fluorescence 300 nm excitation/405 nm emission). Finally the proportion of acetic acid and acetonitrile in the solvent was increased and the column replaced in an attempt to produce symmetrical peaks.

3.2.5.3 Results The salicylic acid (SA) standard produced a peak after 4.47 min, with a greater response detected by fluorescence than UV. Analysis of the plant sample produced a second peak after 8.68 min, confirmed as LNA by comparison with the LNA standard. However, a peak was also produced after 4.47 min by the plant sample without SA, suggesting the liverwort tissue contained SA rendering it inappropriate as an internal standard.

The UV absorbance spectrum for HNA produced three peaks (273, 284 and 295) close to that of LNA (285 nm) and one at 360 nm (Table 3-6). The HPLC analysis of HNA produced small peaks (Table 3-7) that were difficult to identify using both the UV and fluorescence detectors.

Table 3-6. UV absorbance spectra for 3-hydroxy-2-naphthoic acid (HNA) Wavelength (nm) Absorbance 360 0.172 295 0.227 284 0.391 273 0.339

The peak area produced using 0.1 mg m1-1 purified HNA (64) was still very small; the more concentrated standard (1 mg m1-1 HNA) produced a broad, tailing peak (peak area 187), and with increased retention times (8.0 and 7.75 min respectively), similar to that of LNA; if this were the normal retention time for HNA it would be difficult to identify if peaks were produced by LNA or HNA. The peaks produced using the acidified solvent and after changing the column remained small, broad and tailing, and not detectable by fluorescence at the optimum settings for LNA, therefore it was unsuitable to use as a standard.

108

The effect of environment on liverwort establishment, growth and development Chapter 3

Table 3-7. HPLC using 3-hydroxy-2-naphthoic acid (HNA) UV setting HNA Conc Retention time Peak area mg m1-1 (min) 360 0.1 5 14.75 310 0.1 5 5.0 284 0.1 5 25.0 284 1.0 5 292.7

3.2.6 Preliminary investigation into simplifying LNA extraction 3.2.6.1 Introduction A simplified extraction and detection process was developed, minimising the opportunity for LNA losses and removing the absolute need for an internal standard. Liverwort tissue was crushed with a pestle and mortar and the LNA extracted with acetonitrile. SA was again tested as an internal standard.

3.2.6.2 Method 25 mg ground fresh liverwort gemmae were extracted three times with 95% acetonitrile (150 pl), centrifuged, the supernatants collected, bulked and analysed by HPLC, along with SA (0.001 mg m1-1) and LNA (0.019 mg m1-1) standards. The solvents (solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40) were used for the HPLC system and SA and LNA standards preparation, to regain the original retention times for LNA. The guard column was replaced and the samples reanalysed. To confirm the identity of the peak produced by the liverwort sample it was 'spiked' by mixing aliquots of sample and LNA standard, diluted 1:20 as the peak size of the standard was approx. 20 times larger than that produced by the gemmae extract, and analysed.

3.2.6.3 Results As the peaks produced were initially unacceptable the guard column was replaced and the run repeated, when symmetrical peaks were produced for both SA and LNA standards (Table 3-8).

The peak produced in the plant sample had a retention time of 8.08 min that could have been LNA; the peak for the LNA standard had a retention time of 8.58 min. The identity of the peak produced by the gemma extract was confirmed as LNA as the total peak area for the LNA standard and the gemmae extract was 887726; very similar to

109 The effect of environment on liverwort establishment, growth and development Chapter 3 the single peak size produced using the combined gemmae extract and LNA (955117) (Table 3-9). This suggested the two peaks co-migrated to form one peak and confirmed the peak produced by the gemmae extract as LNA.

Table 3-8. HPLC analysis of liverwort gemmae, LNA and SA standards detected using fluorescence Sample Retention Peak area time (min) SA std 0.001 mg m1-1 4.88 5,217,797 LNA std 0.019 mg m1-1 8.58 9,244,885 25 mg gemmae extract 8.08 487,064

Table 3-9. HPLC analysis of liverwort gemmae detected using fluorescence Sample Retention time Peak area (min) LNA standard 1:20 dilution 8.68 400,662 25 mg gemmae extract 8.08 487,064 LNA standard 1:20 dilution + 25 mg gemmae extract 8.25 955,117

110

The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.7 LNA recovery rates using the simplified extraction method 3.2.7.1 Introduction To calculate how many extractions were necessary to extract the maximum amount of LNA from tissue samples recovery rates were calculated using serial extractions of LNA from liverwort tissue without bulking the supernatant after centrifuging, and analysing them individually.

3.2.7.2 Method Samples of liverwort thallus (3 x 25mg) were weighed into separate Eppendorfs and extracted with 95% acetonitrile + 0.1% acetic acid (50 pL). They were ground up with a pestle, and centrifuged for 5 min (13000 rpm). The supernatant was removed into a vial, and the extraction repeated, keeping the supernatants separate. Five extractions were completed from each sample and analysed by HPLC. Areas beneath the peaks produced by fluorescence were recorded and percentage recovery calculated for each (Table 3-10).

3.2.7.3 Results The average peak area for each extraction across the three samples and the percentage of the total amount recovered was calculated (Table 3-10). The peak areas show that the majority of the LNA was recovered during the first four extractions. The presence of LNA in extractions indicates that further extractions would be necessary for 100% recovery. The results can be used to determine the number of extractions necessary to provide an acceptable recovery rate, which can then be incorporated into the extraction protocol.

Table 3-10. LNA quantities obtained with five extractions from 25 mg liverwort thallus. LNA was still detected after 5 extractions using this method. Figures relate to peak areas produced by LNA during HPLC analysis, detected by fluorescence. Extraction Average of 3 samples % of total LNA extracted 1 20434469 67.5% 2 7819372 25.8% 3 1320731 4.4% 4 509006 1.7% 5 185155 0.6%

111

The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.8 Method development for extraction techniques using TissueLyser 3.2.8.1 Introduction Grinding of plant material with a pestle and mortar is slow, particularly when preparing large numbers of samples therefore using a TissueLyser (Manufacturer: Qiagen) was investigated. The TissueLyser pulverised plant tissue by shaking samples in Eppendorfs with a steel ball, with each movement referred to as a motion. Several different volumes of extraction solvent were compared to find which extracted LNA most efficiently. The extraction solvent was optimised, as using high concentrations of acetonitrile produced dirty samples which adversely affected the HPLC chromatography, producing variable retention times and unacceptable peak shapes. Extractions using various concentrations of acetonitrile (ACN) and methanol (MeOH) were compared, aiming to extract LNA but not chlorophyll from liverwort tissue. LNA extraction rates were investigated using ACN (60%, 95%) to identify how many extractions were required to obtain maximum LNA from tissue. 95% ACN was used as it was known to extract efficiently (although too green to use regularly), to compare overall extraction rates. A control was also used, comprising 60% ACN but no plant extract, to establish if any peaks were produced by plasticisers from the Eppendorfs during extraction. All solvents were acidified with 0.1% acetic acid.

3.2.8.2 Method Samples of liverwort thallus (3 x 25 mg) were extracted six times in 50, 100, 150 and 200 pL acetonitrile + 0.1°/0 acetic acid. For each extraction liverwort tissue was placed in Eppendorfs with solvent, TissueLysed (2 min, 30 motions min-1), centrifuged (5 min, 12651 rpm) (Eppendorf Centrifuge 5810R with FA45-30-11 Eppendorf adaptor) and the supernatant removed to a vial; supernatants were not combined.

To produce samples free of particles of plant tissue the method was adjusted, reducing the acetonitrile concentration and acidifying the solvent; samples were centrifuged for longer. Samples of liverwort thallus (25 mg) were extracted six times in 100, 150 and 200 pL quantities of solvent (solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40), TissueLysed (2 min, 30 motions min-1), centrifuged (10 min, 12651 rpm) and the supernatant removed to individual vials; supernatants were not combined. Samples were analysed by HPLC with 10 pL injection (solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40).

112 The effect of environment on liverwort establishment, growth and development Chapter 3

Extractions using different solvents (ACN and Me0H) at various concentrations were compared. Samples of thallus tissue (25 mg), taken from a single liverwort, were extracted three times with various solvents (80%, 60%, 40% acetonitrile (ACN); 80%, 60%, 40% Me0H), TissueLysed (2 min, 30 motions min-1) and centrifuged (10 min, 12651 rpm). Supernatant colour after each extraction was recorded; supernatants were then bulked and selected samples (60% and 40% ACN) analysed (HPLC, solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40).

To establish LNA extraction rates six samples of liverwort thallus tissue (25 mg) were extracted five times, with 60% and 95% ACN, TissueLysed (2 min, 30 motions min-1), and centrifuged (10 min, 12651 rpm). Supernatants of those extracted in 95% ACN were bulked; samples extracted in 60% ACN were kept separate and analysed (HPLC, solvent: solvent A, H2O + 0.1% acetic acid, and solvent B, 95% acetonitrile, 60:40). The control sample, comprised of 60% ACN only, was TissueLysed and centrifuged along with the plant samples but contained no plant tissue.

Following the procedure in earlier experiments, 95% acetonitrile was used; however this also extracted too much chlorophyll, adversely affecting the column, therefore the process was repeated with more samples, extracted with the ACN content of the solvent reduced to 65%. The first extraction from each sample of the 65% extractions and one 95% extraction were then run with standards to establish the retention times and peak areas were consistent.

3.2.8.3 Results The samples produced were very green and dirty, suggesting the solvent should be adjusted to remove more chlorophyll and centrifuged for longer to improve the separation; the TissueLyser pulverised plant tissue more thoroughly than a pestle and mortar, producing smaller particles of tissue that required longer centrifuge time. 50 pL solvent quantities proved too small to provide usable samples, so only the larger quantities were used. The extended centrifuge time resulted in improved separation, producing cleaner samples. However, peaks produced were small with inconsistent retention times (standards 11.53 and 11.71 min) when compared with those of previous analyses (standards 8.97 and 9.93 min). Some variation in peak size was expected as the plant material was not grown under controlled conditions.

113 The effect of environment on liverwort establishment, growth and development Chapter 3

The comparison of different solvents and solvent concentrations showed that Me0H removed less chlorophyll from tissue than ACN (Table 3-11). For both solvents 40% conc. extracted least chlorophyll; the greatest amount was extracted by 80% ACN. HPLC analysis for samples extracted in methanol and 40% ACN produced small, unmeasurable peaks (Table 3-12), suggesting that little LNA was extracted. A sizeable peak was produced extracting with 60% ACN, and a smaller peak for 40% ACN. The retention times for both standards and samples was extended beyond 11 min.

Table 3-11. Appearance of extracts of liverwort thallus, comparing solvent and solvent concentration. G = green, LG = light green, Y = yellow, C = clear Acetonitrile Methanol Extraction 80% 60% 40% 80% 60% 40% 1 G Y/G CL/Y LG LG C 2 G L/G C LG LG C 3 C C/Y C YG LG C Bulked LG Y C Y Y C

Table 3-12. HPLC analysis of LNA extracted from liverwort thallus in methanol and acetonitrile (ACN) Fluorescence Sample Rt Peak area 0.019 mg/mI LNA std 11.89 9935181 40% methanol unmeasurable 60% methanol unmeasurable 40% ACN - unmeasurable 60% ACN 11.98 422842 0.019 mg/ml LNA std 12.45 12567227

100% LNA was recovered from samples after 3 extractions and 98% after 2 extractions (Table 3-13). The control produced a single peak after 2.48 min, which did

not compromise the LNA results.

114 The effect of environment on liverwort establishment, growth and development Chapter 3

Table 3-13. Recoveries for LNA tissue extracted five times with 60% ACN, 0.1% acetic acid Extraction Average of 3 samples % of total LNA extracted 1 3,956,031 85.7

2 573,320 12.4 3 89,248 1.9 4 5

The retention times of the samples were changeable, but moved in tune with the LNA standards. Possible reasons for variations were an adjustment to room temperature, which had become markedly warmer (23 °C). The first extraction from each sample of the 60% extractions and one 95% extraction and three LNA standards were reanalysed and produced peaks with similar retention times and symmetrical peaks (Table 3-14), indicating an acceptable level of consistency.

Table 3-14. Comparison of HPLC analysis results obtained from selected samples analysed on 16th and 17th November. 16/11/2005 17/11/2005 Fluorescence Rt Area Rt Area 60%-A-1 11.04 4,342,611 8.76 6,056,885 60%-B-1 7.53 4,575,219 8.85 5,979,054 60%-C-1 7.61 2,950,263 8.78 3,961,495 95%-A 7.71 1,732,178 9.08 1,607,969

115

The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.9 The effect of day length and environment on lunularic acid content of liverwort grown in nursery conditions. 3.2.9.1 Introduction The extraction and sample analysis methods developed were used to investigate the lunularic acid content of liverwort tissue grown in nursery conditions. Two extractions of lunularic acid were carried out during short days (November and December, 2005) and two during long days (May and June, 2006); liverwort samples were taken from Madrona Nursery, Pluckley, and Imperial College Nursery at Wye from inside and outside environments.

Liverwort tissue used for these experiments were thallus edge (TE), thallus centre (TC), archegonia (AR), antheridia (AN), and gemmae & cups (G) and rhizoids (R). The thallus edge was defined as tissue within 2 mm of the outside edge of thallus, and the thallus centre was thallus tissue only with no other structures, such as gemma cups nearby. Archegonia and antheridia tissue included the stalks. Gemmae and gemma cups were analysed together. For the November extractions there was not enough clean rhizoid material collected for analysis. For the December experiment liverwort tissue was collected from different areas of the nursery as the plants used previously were not in good condition and the area was being heated.

3.2.9.2 Method Tissue samples (25 mg) were extracted three times in 150 pL 60% ACN with 0.1% acetic acid; each tissue type was replicated three times. Samples were TissueLysed for two minutes at 30 motions per minute (the TissueLyser shakes samples back and forth, each movement is referred to as a motion), centrifuged for 10 minutes (12651 rpm) and supernatants for each sample bulked. Extracts were subjected to HPLC analysis with 20 minute runs, 10 pL injections and detected by fluorescence. The system was washed through with 95% ACN prior to the analysis.

As no internal standard was used in these experiments a calibration chart (Figure 3-19) was constructed to quantify LNA content following HPLC analysis. Approximately 0.1 mg LNA was weighed and using the molecular extinction coefficient the actual LNA concentration was found to be 0.0613 mg m1-1.

116 The effect of environment on liverwort establishment, growth and development Chapter 3

The molecular extinction coefficient (E) of LNA at a UV absorption wavelength of 308 nm = 4200 (Table 3-15); molecular weight of LNA = 258. The equation used was A, =

EcL where Ax = absorbance wavelength (nm), E = molar extinction coefficient, c = concentration, L = light path length (1 cm). LNA absorbencies were determined using a

Shimazu UV-210-PC UV-VIS scanning spectrophotometer, Table 3-16)

Serial dilutions of this standard were analysed by HPLC and the results used to construct the calibration curve, with LNA concentrations adjusted to the actual LNA concentration previously calculated. Using linear regression a curve was fitted to the data (Figure 3-19), and the regression line used to quantify LNA concentration of peak areas produced by HPLC analysis of liverwort tissue samples.

Table 3-15. Molecular extinction coefficients of LNA in neutral ethanol at given UV wavelengths (Valio and Schwabe, 1969). Absorbance wavelength Extinction coefficient

Amax (nm) 308 4,200 287 3,600 280 3,300

Table 3-16. Absorbances of 0.1 mg m1-1 LNA in ethanol Absorbance wavelength (nm) Absorbance (A) 308 0.998 287 0.766 280 0.720

117 The effect of environment on liverwort establishment, growth and development Chapter 3

3.5e+6 y=2 x 1010x-1.651 + 38990 R2= 0.9997 3.0e+6 -

2.5e+6 -

"a 2 Oe+6 - E' ' ID _v as w 0- 1.5e+6 -

1.0e+6 -

5.0e+5 -

0.0 0.00000 0.00002 0.00004 0.00006 0.00008 0.00010 0.00012 0.00014

Lunularic acid concentration (mg) Figure 3-19. Calibration curve produced for lunularic acid

3.2.9.3 Results The analysis of samples proved difficult initially because of the problem finding a suitable internal standard, and although a simpler extraction process was established it was difficult to produce clean samples. Maceration of liverwort tissue using a pestle and mortar was successful, but time consuming for a large number of samples; using the TissueLyser pulverised the tissue into smaller fragments that could not be completely removed and caused difficulties with the HPLC. Thorough cleaning of the system with 95% acetonitrile before putting samples through and changing the guard column alleviated this problem.

There was a trend where greater LNA levels were found in the long day experiments than in short days (Figure 3-20), in agreement with Gorham (1975) who found that liverwort growth decreases during long days with high light intensities above 5,600 lux due to high levels of lunularic acid (Gorham, 1975). In the outside environment and long day treatments, higher LNA levels were found in May than July; however, for liverworts grown inside higher levels were found in July than May, other than in rhizoids and gemma cups.

118 The effectofenvironmentonliverwortestablishment,growthanddevelopment R= rhizoids Figure 3-20.Analysisoflunularicacidcontent ofliverworttissue.Resultsarean AN=antheridiophores, AR=archegoniophores, TE=thallusedge,TC=thalluscentre, average ofthreereplicates.O=outside,I=inside. G=gemmaeandgemmacups,

Lunularic acid concentration (mg) Lunularic acid concentration (mg) 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 MN O-G 0-AN0-AR0-TEO-TCO-RI-GI-ANI-ARI-TEI-TCI-R November2005 December 2005

Liverwort tissuetype 119 T 7,7 (a)

Chapter 3 The effect of environment on liverwort establishment, growth and development Chapter 3

With the short day length, lunularic acid levels were more variable than with long day length. Highest levels were found in archegoniophores, particularly in the November, inside treatment; the LNA content of the archegoniophores was markedly greater than the other tissues.

Results for both short and long day experiments were analysed, separately, using ANOVA, and found not to be significant (vr 3.12, p<0.05).

Rhizoids were not included in the November 2005 experiments, in the December 2005 LNA was detected in only one of the six samples analysed, but in May and July all rhizoid samples contained lunularic acid.

3.2.10 The effect of shading on liverwort establishment and growth 3.2.10.1 Introduction The previous growth and development experiments under controlled environments (3.3.1 and 3.3.2) indicated that radial growth of liverwort was greater under high light levels (800 p mol ms-2 s-1) than low (400 p mol ms-2 s-1) with morphological changes occurring to the gemmalings under high light as they became brown and started to die, indicating that high light levels for extended periods were damaging liverwort; there was least growth at low light levels.

A further experiment, located at Palmstead Nurseries Ltd., Wye, was designed to investigate the effect of light levels on liverwort establishment and growth under nursery conditions, from May to September, 2006. 'Seed' pots of liverworts and pots of liverwort-free compost were placed in shaded tunnels designed to provide different levels of light. Liverwort gemmae were dispersed by the overhead irrigation system and their establishment and growth measured.

The aim of this experiment was to investigate the effect of light levels on liverwort establishment and growth and to ascertain whether shading could be utilised as an a cultural practice to control liverwort. It was hypothesised that under high light levels liverwort would initially grow and establish, but in the longer term would be damaged, resulting in a reduced liverwort presence; liverwort would establish but grow less in very low light levels; and liverwort would proliferate more in medium light levels.

120 The effect of environment on liverwort establishment, growth and development Chapter 3

3.2.10.2 Methods Treatments Liverworts were grown within shade tunnels providing three light treatments: 44% and 73% shading, and no shading, with one 2 L 'seed' pot of healthy liverwort and 20 x 2 L pots of compost placed in each shade tunnel. The 'seed' pot of liverwort was placed in the centre of the pots of compost. The shading fabric used was selected from a limited range available, and the level of shade afforded was measured by the manufacturer. All treatments were replicated twice inside a polytunnel and twice outside, as in the experiment layout (Figure 3-21); the placement of blocks and treatments were fully randomised.

Within polytunnel Outside

Block I Block II Block I Block II

73% 44% 73% 0%

44% 0 44% 44%

0 73% 0 73

Figure 3-21 Experiment layout. Each treatment is replicated twice within the polytunnel and outside, and provides 730/0, 44% and 0 shading. Actual light levels are presented in Table 3-17)

121 The effect of environment on liverwort establishment, growth and development Chapter 3

Shade tunnels The shade structures were constructed from galvanised steel hoops, the ends of which were forced into the ground, two per structure placed 1 metre apart with 300 mm between tunnels. They were clad in woven polypropylene shading fabric (Supplier: Growing Technologies) tied to the hoops to provide stability (Figure 3-22, Figure 3-23).

2800 mm Hoop r___,----- '----,..

i_ 1700 mm —o. 9001mm Ground level—o-

Figure 3-22. Construction of shading structures. Galvanised steel hoops were forced into the ground and clad with shading fabric.

Hoops prior to cladding

Figure 3-23. Shade tunnels (a) outside (b) inside the polytunnel

122 The effect of environment on liverwort establishment, growth and development Chapter 3

Compost Compost was 100% peat (Maxi Bale Natural AOGBF 2.25 m, Kekkila Oyj, Tuusula, Finland), with added Osmocote Pro (3500 g m-3) and lime as Dolodust magnesium limestone (2000 g m-3), provided in 2 L pots by Palmstead Nurseries Ltd.

Irrigation The overhead irrigation system at Palmstead Nursery is based on vapour pressure deficit, with water applied in two short 2 minute bursts, with a few minutes between each burst, and minimal water application, equating to daily in winter and twice daily in spring and summer applications.

Data collection Percentage coverage of each pot by liverwort was recorded after 119 days. A 10 mm grid printed on a transparency was placed over each pot and the area of liverwort coverage recorded.

Supplementary environmental data was recorded at pot level within each shade tunnel: light levels using a Basic Quantum Meter model QMSW-SS (Apogee Instruments Inc. Logan, UT, USA), and relative humidity and temperature using an Al Hygromer (Rotronic Instruments (UK) Ltd., Crawley, W. Sussex).

Statistical analysis For statistical analysis the inside and outside treatments were treated as independent experiments as the results in the outside treatments would be confounded by uncontrollable environmental conditions, especially rain, which affected gemma dispersal.

3.2.10.3 Results Temperature, relative humidity and light levels were recorded weekly for each treatment, and are presented in Figure 3-25. Liverwort growth was measured as the percentage area of pot covered.

A range of growth patterns were observed throughout the treatments, a sample of which is shown in Figure 3-24. The polytunnel itself reduced the light level by approximately 50% (Table 3-17, treatments 0-0 and I-0) so an effect of the lower light

123 The effect of environment on liverwort establishment, growth and development Chapter 3 levels may be evident in liverwort establishment and growth in treatments with no additional shade applied. The shade tunnels afforded 73% and 44% shade to the treatments (Figure 3-21), with no shading provided to the controls.

Table 3-17. Comparison of light levels, temperature and humidity inside the polytunnel and outside for each treatment. Figures are averages of 12 readings with block I and II figures amalgamated. 0 = outside, I = inside, 0, 44 and 73 = 0/0 shade. Light level Temperature Relative Humidity Treatment (pmol M-2 S-1) (C) (%) 0-0 714.13 23.25 41.22 I-0 338.92 24.49 42.60 0-44 316.08 22.75 46.12 1-44 171.29 24.41 45.77 0-73 127.33 22.72 47.43 1-73 83.25 24.63 46.71

Figure 3-24. Examples of liverwort growth showing (a) very dry compost (I-1-73) (b) an area of liverwort die back with new growth (I-0-0) (c) many small, congested plants (1-0-73) (d) sparse liverwort establishment (I-0-0) (e) vigorous growth, (I-1-44) (f) large gemma cups with many gemmae (I-0-0). Scale bars = 10 cm.

124

The effect of environment on liverwort establishment, growth and development Chapter 3

C) ( e tur era Temp

1 2 3 4 5 6 7 8 9 10 11 12

100

g:,) 80 -

:2 60 -

20 -

0 1 2 3 4 5 6 7 8 9 10 11 12

1600 1400 - ' Lc) 1200 - rsi • 1000- • 800- • 600 - • 400 - 200- a' 71 0 -

1 2 3 4 5 6 7 8 9 10 11 12

Week No.

—II— Outside 0 —v— Outside 44% —m— Outside 73% —0— Inside 0 —v— Inside 44% —0— Inside 73%

Figure 3-25. Temperature, relative humidity and light level readings taken from each shade tunnel during the experiment. Legend refers to treatments, with 0 (none), 44 and 73 % shading.

125 The effect of environment on liverwort establishment, growth and development Chapter 3

120

100

%) ( e ag er t cov Po

t liverwor d he blis ta h es it w ts o p f o No

hores top ame g ing ar be

ts r wo liver h it w ts o f p o No.

Treatment

Figure 3-26. The effect of light level on (a) liverwort growth, expressed as pot coverage, (b) establishment. (c) The number of pots in each treatment containing gametophore-bearing liverwort. I = inside, 0 = outside. 0, 44 and 73 = % shade.

126 The effect of environment on liverwort establishment, growth and development Chapter 3

Light levels The effect of light levels on pot coverage is shown in (Figure 3-27 and Figure 3-26a). There was less growth at the highest and lowest light levels, and maximum pot coverage for inside and outside treatments where liverworts experienced mid-range light levels between 300 to 400 pmol ms-2 s-1 (I-0 and 0-44).

60 —ii— Inside light levels —A— Outside light levels 50 -

40 -

rag 30 - ve t co o

p 20 - °A)

10 -

o 100 200 300 400 500 600 700 800

Light level gmol ms 2 s 1 Figure 3-27. % pot coverage compared to light levels for inside and outside treatments. Light levels are means of weekly readings for each treatment, taken over the full term of the experiment. I = inside, 0 = outside. 0, 44 and 73 = 0/0 shade.

Mean liverwort area was less inside than out; with average pot coverages of 17.91% and 39.84% respectively. Inside there was less liverwort growth in the 73% than 44% shading treatment; greatest growth was observed with no shading applied. Fewer liverworts established per pot, but these grew larger and more vigorously, with larger thalli than outside treatments, and gemma cups tended to be large containing many gemmae (Figure 3-24f). In the 73% shade treatments, however, fewer liverworts established and were smaller than other treatments. The compost of the inside, no shade treatment was very dry, typical for Palmstead Nursery; previously established liverwort had died and new colonies were subsequently developing.

127 The effect of environment on liverwort establishment, growth and development Chapter 3

Outside, there was less liverwort growth with no shade than with 73%; greatest growth occurred with 44% shade. The liverwort that grew under these conditions tended to be small and congested, and present in large numbers, possibly the result of the dispersal of a greater number of gemmae than inside due to rainfall. There were fewer gemma cups than inside, containing fewer gemmae, particularly where liverwort establishment was sparse.

The results for the outside treatments could have been affected by rain, which would have increased gemma dispersal. No significant differences were found, accepting the null hypothesis of no difference in pot coverage as a result of shade level.

For inside treatments data was transformed using a logit transformation. Analysis of variance showed that there was a significant effect of light levels (0 and 73% shade) on pot coverage by liverwort (F2,2 = 39.93, P < 0.05) (Table 3-18).

Table 3-18. Analysis of variance for inside treatments using logit transformation. Source of variation d.f. s.s. m.s. v.r. F pr. BLOCK stratum 1 0.8282 0.8282 2.37 BLOCK.*Units* stratum SHADE 2 27.9487 13.9743 39.93 0.024 Residual 2 0.7000 0.3500 Total 5 29.4768

Establishment Establishment was calculated as the number of pots containing established liverwort (Figure 3-26b). The results suggest an effect of light level on liverwort establishment that reflects the effect of light on liverwort growth, with less liverwort established successfully inside than out in all treatments.

Statistical analysis (calculated separately for inside and outside treatments) using a Chi-squared contingency test confirmed a significant effect of shade on liverwort establishment for both the inside and outside experiments. The critical value at p=0.05 with two degrees of freedom is 5.99; chi-squared values for outside were 0.113, and for inside 6.46; shading therefore did have a statistically significant effect on liverwort

128 The effect of environment on liverwort establishment, growth and development Chapter 3 establishment for the inside environment, rejecting the null hypothesis, but not for the outside environment.

Further inspection of each set of calculations revealed that the inside, 73% shade treatment had a greater effect than either 44% shade or no shade; fewer pots contained established liverwort in the 73% shade treatment. Outside, less liverwort was established in the treatment with no shade, corresponding to the maximum light level, than 44% shade. Inside there was little difference between the 44% and '0' shade treatments, and again fewer pots contained established liverwort in the 73% shade treatment.

Maturity There appeared to be an effect of light level on the sexual maturity of liverworts, with more pots containing liverworts bearing gametophores where no shade was provided, even though one treatment was very dry, had died back, was re-growing and bore no gametophores. No gametophores were produced within the 73% shade treatments, either inside or outside.

For the inside treatments the growth and establishment trends are reflected in the number of pots of liverwort bearing gametophores (Figure 3-26c), with fewer gametophores (none), less establishment and less growth in treatments with 73% shade provided, suggesting that light levels could affect liverwort maturity; most gametophores and greater growth occurred in the treatments without shade. Where there was greatest growth and more gametophores the liverworts were more mature, and more of them had entered the sexual phase. However, there was greater pot coverage in the outside 73% treatment than inside, but still no gametophores were produced and pots contained greater numbers of smaller liverworts.

Statistical analysis using a Chi-squared contingency test confirmed a significant effect of light on gametophore production. Using the critical value of chi-square at p<0.05 with two degrees of freedom (5.99), chi-squared values for outside were 10.49, and for inside 30.17, therefore the null hypothesis of shade having no effect on gametophore production was rejected for both inside and outside treatments. Further inspection of the calculations for the inside treatments revealed that the '0' shade

129 The effect of environment on liverwort establishment, growth and development Chapter 3 treatment had a greater effect than other treatments, with more gametophores present than either 44% or 73% shade treatments.

For the outside treatments the greatest effect was also due to the '0' shade treatment, although the difference in the number of gametophores present was smaller than for the inside treatments.

Liverwort in the seed pots for the inside treatments died back early in the experiment; some were re-establishing themselves either from an influx of gemmae from other pots or regeneration of surviving tissue. Gemmae from the seed liverworts had dispersed prior to this and newly established colonies developed their own gemma cups and gemmae which were subsequently dispersed. Outside, some seed pots had died, but others were growing vigorously.

3.3 Discussion

Growth and development The two growth cabinet experiments compared liverwort growth and development in varying temperature conditions (25 °C, 15 °C and 10 °C), while light levels, humidity and day length remained constant.

Growth Gemmalings grown in high light, high temperature (25 °C) conditions initially appeared to benefit from this regime; however in the longer term these conditions proved detrimental as evidenced by the abnormal morphology and death of some replicates. When grown at the lower temperature and lower light levels although relative growth rates were variable they were generally sustained and, on occasions increased producing liverworts with more normal morphology and fresh weights similar to the high temperature, low light treatment after six weeks. These replicates may have had the potential to reach similar levels of growth as those in the high temperature, low light treatments, given a longer growth period.

Both experiments showed that generally growth, dry weight, fresh weight and number of gemma cups are all greater in high temperature than low temperature; and in low light than high light. These results are broadly in line with the literature, where Terui

130 The effect of environment on liverwort establishment, growth and development Chapter 3

(1981) and Mache and Loiseaux (1973) found that higher light levels (5000 and 6000 lux respectively) inhibited liverwort growth.

Development The results of the first experiment do not appear to offer any indication that, by altering light or temperature within the parameters used, development of gemma cups could be slowed down significantly, in fact gemma cups appeared sooner in high light, high temperature treatments. Overall development of gemma cups followed the same pattern as growth, with more gemma cups produced under high temperature than low, and in low light than high light conditions. It is difficult to relate the results of this experiment to others reported in the literature, as the purpose of the research, and experimental conditions used differed. Voth (1941) found more gemma cups on male L. cruciata than female, as was found in this experiment at both light levels. They also found that temperatures above 24 °C were detrimental to growth, as in this experiment.

During each of these experiments, it was noticed that gemmae clump together. This affected results by making it difficult to count gemmae in (Section 3.2). Possibly, as no water was applied over the gemmae they were not dispersed, but replacements were still being produced, resulting in gemmae being pushed out of the cup and away from the influence of lunularic acid (Chapter 1), a mechanism thought to prevent germination of gemmae within the cup (Taren, 1958); as the gemmae start to grow they become intertwined and locked together. A further experiment was carried out to investigate this 'clumping' more closely (Chapter 4).

Gender In the second growth cabinet experiment (Section 3.2.2) dry weights, fresh weights and growth (radial expansion) were all greater for female gemmalings than male, except for the high light, low temperature treatment where male gemmalings were larger. These gemmalings were the smallest of all the treatments, showing very little growth.

In the first experiment (Section 3.2) there was a significant effect at the 0.1% level depending on gender, with more gemma cups produced on male plants than female at both light levels. In year 2, after four weeks there were more gemma cups on male

131 The effect of environment on liverwort establishment, growth and development Chapter 3 than female gemmalings in high light than low light treatments. However, at 10 °C the gemmalings were particularly small and gemma cups were only produced in the female, low light treatment, with no discernable trend relating to gemmaling gender overall.

Morphology In the first experiment (Section 3.2) growing liverworts in high light at 25 °C had an effect on their colour and morphology. By week 5 they were a dull green colour, and by week 6, 8 of the 12 replicates were developing dark brown colouration and reduced relative growth rate, showing signs of tissue damage due to adverse growing conditions. The thallus had a thicker, dome-shaped appearance. Liverworts grown in lower temperature and light conditions were flatter and were a brighter green colour.

In the second experiment (Section 3.3), at the lower temperatures none of the replicates developed the dark brown colouration although some replicates grown at 15 °C in high light levels again became domed shaped, with sparse radial growth. Liverworts grown under low light levels were, again, flatter and brighter green in colour.

The results of the growth and development experiments suggested that liverwort would be more prevalent in areas protected from full sunlight and high temperature, and during cooler weather. Light shading of propagation areas, along with their high humidity levels, for example, would promote liverwort growth. The implications are that a reduction in shading and humidity could reduce liverwort infestations.

The results of these experiments indicate that liverworts grown in high temperature and light conditions suffer damage that causes them to degenerate more quickly than when grown in lower light and temperature conditions. Mache and Loiseaux (1973) found that the thallus became thicker and brittle in high light conditions (6000 lux) and changes occurred in the structure of chloroplasts, with small grana and fret membranes replaced by continuous grana.

Lunularic acid concentrations were variable as liverworts were taken directly from nursery conditions; had they been grown in constant environmental conditions in a growth room, less variability may have been present and more definitive results

132 The effect of environment on liverwort establishment, growth and development Chapter 3 obtained. However, the experiment was designed to characterise LNA levels in nursery conditions, to ascertain whether such information could provide clues to the timing of nursery liverwort control practices.

The inhibiting effect of extreme high and low light levels on liverwort growth reflects the results of the growth cabinet experiments where overall growth of liverwort was greater in low light (400 pmol s-1) than high light (800 pmol m-2 s-1).

Conclusions Growers of protected crops could take advantage of these observations by manipulating light and shade to provide light levels unfavourable to maximum liverwort growth. This could be tailored to crop plant requirements, grouping sun loving plants in areas of high sunlight and shade loving plants in areas of low light, with plants being grown at the extremes of their tolerance where possible. However, this may not be practical for those growing many varieties in small numbers.

Alternatively, shade could be provided to the compost surface by use of filter-fabric pot covers (widely available) (Svenson et al., 1997) or free-draining mulch such as bark, or shells (Svenson, 1998). Miscanthus used as mulch can also be beneficial, but is often unsightly. For the grower, any material used as mulch has to withstand wind without being blown away and dry rapidly between irrigation cycles; it also has to be inexpensive and easy to apply.

133 Liverwort epidemiology Chapter 4

Chapter 4 Liverwort epidemiology

4.1 Introduction This section of work investigates the dispersal and establishment of Marchantia polymorpha by vegetative propagules (gemmae). Mature liverwort gemmae are produced in gemma cups and are connected to the parent liverwort thallus by a single- celled stalk, their dispersal is described in Chapter 1. Glands that grow up from the base of the cup among the gemmae (Cavers, 1903a) are thought to imbibe water and swell, breaking the gemmae from their stalks and forcing them out of the cup (Round, 1969).

Dispersal of liverwort gemmae within a plant nursery setting has not been characterised and therefore it is unknown how irrigation systems, particularly the overhead sprinkler systems commonly used, affect dispersal. However, there has been research into the dispersal of spores of fungal pathogens by splash droplets under simulated rain and in field conditions with natural rain. This has broad relevance for liverwort gemmae dispersal.

Where large incident water drops hit the inoculum source, they break up into smaller splash droplets that pick up and carry spores within them. Smaller incident drops, less than 50 pm, do not normally bounce off surfaces (Chamberlain, 1975). The

134 Liverwort epidemiology Chapter 4 fragmentation of the incident drop is dependent on its release height, as splash droplets are not produced by incident drops of 1 to 2 mm diameter when released from heights of 10 and 20 cm, respectively, above the source (Reynolds et al., 1989). The number of splash droplets produced, number of spores dispersed and the travel distance all increase with increased size of incident water drop (Madden, 1997; Madden et al., 1996) and with increased drop height (Geagea et al., 1999; Reynolds et a/., 1989). Travel distances are usually short (< 15 cm) so that repeated bouts of splashing may be required to spread a disease throughout a crop canopy. Higher rain intensity is also associated with increased droplet diameter. The related increase in drop kinetic energy results in a higher number of spores being dispersed, as does increased drop release height (Madden, 1997).

Spores may be dry-dispersed, dislodged by water drops striking the leaf boundary layer, or wet-dispersed within water droplets (Geagea et al., 1999). For spores transported within droplets, dispersal is related to incident drop size; most spores are carried in large droplets, which travel shorter distances (Geagea et al., 1999; Madden, 1997), with the resulting dispersal gradients tending to be steeper. Geagea et at (1999) found that as water drops are applied to a source, each subsequent drop disperses fewer spores, exhausting lesions of spores after 18 drops for all drop diameter and fall height combinations tested. For brown rust (Puccinia recondita f. ap. &No) 50-80% of both rain-splashed and dry-dispersed spores were released after the first 3 drops (Geagea et a/., 1999).. It is unknown what duration of water application would exhaust supplies of liverwort gemmae within gemma cups, or if there is an effect of irrigation duration or frequency on dispersal. A preliminary experiment was designed to investigate the replenishment of gemma cups when gemmae were removed at different time intervals, and the results were integrated into a final dispersal experiment where irrigation was applied via various irrigation systems with different frequencies.

The mucilage found in gemma cups may be a mixture of water and soluble carbohydrates that holds the gemmae together in clumps (Equihua, 1987). 'Clumping' together of gemmae in large numbers was observed during both growth and development (Chapter 3) and gemma dispersal experiments (Section 4.2.1). This phenomenon raised the question of whether there was any biological advantage in clumping of asexual propagules for the liverwort life cycle, perhaps as a strategy to aid

135 Liverwort epidemiology Chapter 4 establishment and growth of new colonies; it has been suggested that this is a strategy to ensure gemmae remain close together after dispersal in an environment where they thrive until conditions are adequate to produce the sexual spores which facilitate long distance dispersal (Equihua, 2005). An experiment was designed to compare the growth and establishment of liverwort colonies initiated with clumps of liverwort gemmae or with freely dispersed groups of individual gemmae.

Numerous different irrigation systems are used by growers: overhead sprinkler systems, ebb-and-flow, capillary matting, sand bed, drip, overhead gantry, trough track and hand watering. The type of irrigation used is based on a number of factors such as installation costs, ease of installation, crop type. Briercliffe (2000) calculated the set up cost for each system and found variations, with hand watering the least expensive (£0.46 m2) then overhead, drip and capillary matting (£1.71 - £2.92 m2), and ebb and flood, overhead gantry and trough track the most expensive (U6.62 - £33.29 m2 respectively). Running costs ranged between £0.39 to £1.38 m2 (drip and hand watering respectively); although the cheapest to install, hand watering, had the highest running costs due to the high labour input. The greatest amount of water was used under overhead irrigation (30 L m2 per week) and ebb and flood (34.6 L m2 per week) compared to drip systems (3.7 L m2 per week), hand watering (8.4 L m2 per week) and capillary matting (12 L m2 per week) (Briercliffe, 2000). Conversely, other investigations into nutrient application systems and growing media within different irrigation systems (capillary matting, ebb and flood and overhead irrigation) have shown that ebb and flood systems used least water, whilst overhead systems were most wasteful (Maher et aL, 1996), so water usage may depend on the way the system is used (e.g. whether the water is re-circulated) as well as the design. Many nurseries irrigate to excess to ensure all plants receive enough water, draining the surplus to waste, a system that is known to increase liverwort infestation. Plants having different irrigation requirements are often grouped within the same irrigation regime rather than accurately scheduling water application to meet the crop's needs (Burgess, 2003a).

Water drops produced in nursery overhead irrigation systems are affected by the water pressure forcing the water through the nozzle, altering the range of droplets sizes produced and affecting the irrigation rate. Increased water pressure produces smaller water droplets. A glasshouse experiment was designed to investigate the effect of

136 Liverwort epidemiology Chapter 4

overhead irrigation systems on gemma dispersal, specifically to investigate different nozzle types, water pressure and nozzle height, each combination of which produced a characteristic range of incident droplet sizes and number of drops. Overhead irrigation was selected as it is the most common method used commercially. A further large- scale experiment under nursery conditions compared the effect of different irrigation systems (drip, overhead and capillary beds) and frequency of water application on gemma dispersal.

4.2 Experimental section 4.2.1 The effect of nozzle, water pressure and nozzle height on gemmae dispersal using an overhead sprinkler system 4.2.1.1 Introduction This experiment investigated the dispersal of liverwort gemmae by a glasshouse overhead sprinkler system, using three different sprinkler nozzle sizes, four different water pressures (1.5, 2, 2.5 and 3 bar) and two different nozzle heights (1 and 2 metres); the irrigation system was designed so that nozzle, nozzle height and water pressure could be varied, therefore any effects on liverwort dispersal were due to the adjustments made. The nozzle heights used reflect commercial practice where placement is variable depending on crop height and whether the growing environment is protected (by glass or plastic) or outside; generally they are supported on risers of various heights (predominately outside) or suspended from above (predominately inside). For each treatment water sensitive paper was used to catch droplets, their area and the number of droplets produced was calculated from the stains that remained once dry.

4.2.1.2 Method The nozzles used were Agridor 700 Dynamic Sprayers, manufactured by Ein Dor (Supplier: Access Irrigation Ltd), colour coded by the manufacturer to denote the water flow rate they produce (Brown = 160 L hr-', Blue = 105 L hr-1 and Grey = 60 L re). A system was set up in the glasshouse with a hose attached to a water tap and an adjustable in-line water pressure adapter fitted (Flamco pressure-reducing valve type '3/4" PR'. Flamco UK Ltd, www.flamco.co.uk). The hose was fitted around the glasshouse cubicle to a central position along one wall with an anti-drip device and sprinkler nozzle attached to the end and held in position by a length of steel projecting

137 Liverwort epidemiology Chapter 4 towards the centre of the area, such that the nozzle could be operated at 1 and 2 m heights. An Ein Dor 530 Non-Drip Valve was inserted above the nozzle to prevent the heavy water drops normally experienced as the water was turned off (Figure 4-3 and Figure 4-4). A half tray (of similar height as the collection pots) with approximately a third of the surface covered with mature liverwort thalli bearing gemma cups (Figure 4-4) was placed on the ground beneath the nozzle.

Collection pots were arranged so they were touching, (Figure 4-1), with three lines of 16 placed at right angles to each other. The number of collection pots required was determined by test trials. For each treatment, water was applied to the liverwort via the sprinkler system for 15 mins, and then the number of gemmae that had fallen into each collection pot was counted. This arrangement was designed to reflect how liverwort could be dispersed within a nursery situation by water passing through an overhead sprinkler system above seed trays or pots containing liverworts.

Each nozzle is characterised by the flow rate and range of droplet sizes produced at each water pressure. Measuring the amount of water collected in a measuring cylinder in one min and converting the result to litres per hour established the actual flow rates obtained in this system. Flow rates given by the manufacturers relate specifically to the conditions under which they are tested, and will therefore differ from those obtained in this experiment, although still providing a good relative estimate of flow. A measure of relative droplet size for each nozzle was obtained using water sensitive paper (Supplier: Syngenta Crop Protection AG) which is card (76 x 52 mm) with a coating that stains dark blue when in contact with water. Three cards per treatment were individually exposed to droplets produced by each treatment, mid-flow, and then placed in a box with silica gel to prevent humid air from affecting the droplet stains.

138 Liverwort epidemiology Chapter 4

Liverwort source 16 i 00000 0. 16 Collection pots •

Figure 4-1. Arrangement of collection pots, with 16 pots were arranged in each line

2m height position

Nozzle 1m height position

Source of liverwort gem mae Collection or dye pots

Figure 4-2. Equipment layout

139 Liverwort epidemiology Chapter 4

In-line water 2m height position pressure adapter

1 m height position p

4 Anti-drip device

Water from mains supply

Flow of incident 4-- water droplets

Collection pots

Source of liverwort gemmae or dye

1.• Water flow

Figure 4-3. Experiment layout

A sequence of tests using the blue nozzle and red dye in place of plant material, with water pressures of 2, 3 and 4 bar, indicated how gemmae are dispersed. If the dye droplets travelled similar distances as the gemmae, this would imply the gemmae are splashed within the splash droplets after they hit the gemma cups. If the gemmae travelled further than the dye the implication would be that at least some are propelled into the air by the incident water droplets.

140 Liverwort epidemiology Chapter 4

Figure 4-4. Example of liverwort used as gemmae source.

4.2.1.3 Results In the preliminary experiment using red dye (Table 4-1) red splash droplets travelled a maximum of 45 cm, with the nozzle 2 metres high. However, when using liverworts as a source, gemmae were recorded at a maximum distance of 160 cm, with the nozzle at 1 metre high, suggesting that whilst some gemmae may be transported within splash droplets, they were also propelled out of the gemma cups by the incident drops.

Table 4-1. Average distance travelled (cm) by splash droplets containing red dye using the blue nozzle at two different heights Water pressure 1m 2m 4 Bar 33 38 3 Bar 43 37 2 Bar 28 45

Dispersal gradients were constructed for each treatment, representing the quantity of gemmae and the distance from the source that they that fell (Figure 4-5 to Figure 4-8), summarised in Table 4-2. Using a log-linear model, regression curves fitted to gemma dispersal data were highly significant, (P <0.001) for all treatments (Table 4-2). Regression coefficients indicated steeper dispersal gradients at 2 and 2.5 bar than 3 and 1.5 bar; steeper dispersal gradients suggest more gemmae were deposited nearer to the source, travelling on average shorter distances.

141 Liverwort epidemiology Chapter 4

Table 4-2. Results obtained from gemma dispersal experiment analysed using log linear regression, providing intercepts and regression coefficients. Nozzles: grey = 60 L hr1 , blue = 105 L hr-1, brown = 160 L hr1 . R2 values relate to transformed data (In). Smaller coefficients indicate steeper gradients and more gemmae nearer to the source. Regression Intercept Std Std Nozzle coefficient F prob R2 (constant) error error (slope)

Blue-3 bar-2 m 7.72 0.243 0.945 0.008 <0.001 0.18

Blue-2.5 bar-2 m 279.30 0.129 0.924 0.006 <0.001 0.77 Blue-2 bar-2 m 256.30 0.096 0.916 0.003 <0.001 0.82 Blue-1.5 bar-2 m 140.60 0.075 0.942 0.003 <0.001 0.83 Blue-3 bar-1 m 6.44 0.216 0.959 0.006 <0.001 0.26 Blue-2.5 bar-1 m 397.00 0.121 0.912 0.006 <0.001 0.75 Blue-2 bar-1 m 263.40 0.096 0.937 0.004 <0.001 0.81 Blue-1.5 bar-1 m 93.00 0.122 0.945 0.004 <0.001 0.70

Brown-3 bar-2 m 82.62 0.206 0.957 0.006 <0.001 0.76

Brown-2.5 bar-2 m 103.70 0.158 0.943 0.006 <0.001 0.66 Brown-2 bar-2 m 482.20 0.207 0.913 0.010 <0.001 0.56 Brown-3 bar-1 m 40.05 0.196 0.946 0.007 <0.001 0.75 Brown-2.5 bar-1 m 82.02 0.178 0.942 0.006 <0.001 0.63 Brown-2 bar-1 m 128.00 0.185 0.924 0.008 <0.001 0.48

Grey-3 bar-2 m 457.90 0.104 0.944 0.004 <0.001 0.76 Grey-2.5 bar-2 m 680.90 0.090 0.936 0.004 <0.001 0.84 Grey-2 bar-2 m 169.70 0.099 0.950 0.003 <0.001 0.82 Grey-1.5 m-2 m 80.91 0.114 0.955 0.003 <0.001 0.72

Grey-3 bar-1 m 8.16 0.234 0.963 0.006 <0.001 0.16 Grey-2.5 bar-1 m 240.10 0.116 0.903 0.006 <0.001 0.71 Grey-2 bar-1 m 522.80 0.100 0.912 0.005 <0.001 0.80

Grey-1.5 bar-1 m 107.40 0.138 0.943 0.005 <0.001 0.69

Whilst the coefficients were similar, steeper dispersal gradients were obtained at 2 and 2.5 bar than 3 and 1.5 bar; the exception being the 3 bar, 2m height treatment using the grey nozzle. Using the grey nozzle, the distance travelled by gemmae clearly increased with increased nozzle height. However, for the blue and brown nozzles a number of gemmae travelled further when the nozzle was at the lower position; although the bulk of the gemmae did travel further when nozzles were higher.

142 Liverwort epidemiology Chapter 4

For each nozzle and nozzle height, gemmae tended to travel further at the extreme water pressures (1.5 and 3 bar), except for the brown nozzle at 1 m and blue nozzle at

2 m height treatments. Using the brown (160 L hr-1) and blue (105 L hr-1) nozzles, the maximum distance travelled by gemmae was greater with the nozzle at 1 m, although the bulk of the gemmae travelled further when the nozzle was at 2 m. Using the grey

(60 L hr-1) nozzle gemmae travelled further with the nozzle at 2 m (Table 4-3).

Maximum dispersal distances obtained were 160 cm; distances previously recorded were 60 cm and 121.4 cm by Brodie (1951) and Equihua (1987) respectively.

The experiment was initially designed to be carried out at 1, 2, 3 and 4 bar. However, the nozzles did not operate below 1.5 bar, and at 4 bar very few gemmae were dispersed, thus only 1.5, 2, 2.5 and 3 bar water pressures were used. The brown nozzle (160 L hr-1) did not operate at 1.5 bar, therefore no results were obtained at this water pressure.

Table 4-3. Maximum distances travelled by gemmae (cm). The brown nozzle (160 L hr') did not operate at 1.5 bar. Nozzle height Pressure Grey Blue Brown

(m) (bar) (60 L hr1 ) (105 L hrl) (160 L hr 1) 1.5 110 130 - 2 90 120 100 1 2.5 80 150 160 3 110 150 90 1.5 150 110 - 2 130 140 110 2 2.5 90 140 90 3 140 140 110

143

Liverwort epidemiology Chapter 4

Blue-3 bar-1 m Bkie-3 bar-2 m

Grey-3 bar-1 m Grey-3 bar-2 m 430

350

11 E E Ear

150

100

Brown-3 bar-1 m Brown-3 bar-2 m

1 'Z''

1 !il 8o z z ,,,,

X X X

to 80 130 ,e 1E0 Distance (cm) Distance (cm)

Figure 4-5. Dispersal gradients obtained operating nozzles at 3 bar.

144

Liverwort epidemiology Chapter 4

Blue-2.5 bar-1 m Blue-2.5 bar-2 m

X 250

E E E E E E x X a

103

40 60 BO 103 . 140 160 Distance (cm) Distance (cm)

Grey-2.5 bar-1 m Grey-2.5 bar-2 m

250

of -

ICO

X II X x K x M M x x x X x

CO 103 133 140 WO 03 BO

Distance (cm) Distance (cm)

Brown-2.5 bar-1 m Brown-2.5 bar-2 m

/50

200 E E O z

x so X X

k N NNNNWNNN X

BO BO 100 123 140 103 Distance (cm)

Figure 4-6. Dispersal gradients obtained operating nozzles at 2.5 bar.

145 Liverwort epidemiology Chapter 4

Blue-2 bar-1 m Blue-2 bar-2 m

E E

Distance (cm) Distance (cm)

Grey-2 bar-1 m Grey-2 bar-2 m

9E5

222

250

E E E

0 X 150 X

102

N m x x 20. H. KV 120 140 100 Distance (cm)

Brown-2 bar-1 m Brown-2 bar-2 m

14 x XNX X X X X X ea 80 103 120 140 1f.) Distance (cm)

Figure 4-7. Dispersal gradients obtained operating nozzles at 2 bar.

146

Liverwort epidemiology Chapter 4

Blue-1.5 bar-i m Blue-I.5 bar-2 m

E p

Distance (cm)

Grey-1.5 bar-1 m Grey-I.5 bar - 2 m

Distance (cm) Distance (cm)

Figure 4-8. Dispersal gradients obtained operating nozzles at 1.5 bar.

147 Liverwort epidemiology Chapter 4

Generally more gemmae were dispersed at 2 and 2.5 bar water pressure (Table 4-4) (exceptions were the grey nozzle, 2 bar, 2m and brown nozzle, 2.5 bar, 2 m). Overall more gemmae were dispersed at the 2m height for the brown and grey (60 L hr-1) nozzles, and at 1 m height for the blue nozzle (105 L hr-1). However, different gemma sources were used for each irrigation application, which would not have contained the same number of gemmae, and this could account for some of the variation in gemma numbers.

Table 4-4. Total no. of liverwort gemmae dispersed Grey Nozzle Blue Nozzle Brown Nozzle Water Pressure (60 L hr-1 ) (105 L hr-1) (160 L hr-1) 3 bar 159 110 481 2.5 bar 1229 2364 906 1m 2 bar 3086 2574 963 1.5 bar 1219 1087 3 bar 5260 91 1347 2.5 bar 6502 2061 1166 2m 2 bar 2287 3611 2921 1.5 bar 1253 1531 -

Figure 4-9. Water sensitive paper showing droplet stains for the blue nozzle (105 L hr-1), 2 m height and 2 bar water pressure treatment.

A measure of relative droplet size and number of droplets produced by each treatment was obtained using water sensitive paper (Figure 4-9); the droplet pattern was photographed digitally and analysed using Image J software. Distribution charts constructed using this data identified the range of droplets sizes produced during each treatment (Figure 4-10 to Figure 4-14). In general the higher the water pressure the smaller the size of the majority of the droplets produced (exceptions were blue nozzle,

148 Liverwort epidemiology Chapter 4

1.5 bar, 2m and grey nozzle, 2 bar, 1 m treatments). Treatments with the least number of gemmae dispersed (Blue nozzle, 3 bar, 1 m, Blue nozzle, 3 bar, 2 m and

Grey nozzle, 3 bar, 1 m) had mean droplet diameters between 171 and 176 pm (Figure 4-11, Table 4-5), however these treatments did not produce the fewest droplets. However, there were exceptions to each trend.

Table 4-5. Comparison of number of gemmae dispersed with no. of droplets and droplet size.

Treatment No. gemmae dispersed No. droplets Mean droplet diameter (pm) Blue nozzle, 3 bar, 2 m 91 3719 172 Blue nozzle, 3 bar, 1 m 110 2766 176 Grey nozzle. 3 bar 1 m 159 5078 171 Brown nozzle, 2 bar, 1 m 963 5489 177 Blue nozzle, 1.5 bar, 2 m 1531 1065 216 Blue nozzle, 2.5 bar, 1 m 2364 3116 172 Grey nozzle, 2.5 bar, 2 m 6502 1451 271

Nozzle flow rates obtained (Table 4-6) were lower than those quoted by the manufacturer which refer to tests carried out under laboratory conditions, full details of which are not provided with the nozzles; differences are expected when nozzles are used in field or glasshouse situations (Agridor, Undated).

Table 4-6. Nozzle flow rates (L hr1 ) at different water pressures Water Pressure Grey Nozzle Blue Nozzle Brown Nozzle 1.5 Bar 43.20 60.30 0.00 2.0 Bar 51.60 75.30 90.80 2.5 Bar 56.04 88.60 102.20 3.0 Bar 68.60 97.74 110.60

149 Liverwort epidemiology Chapter 4

2500 2500 (a) (b)

2000 - 2000 -

1500 - 1500- S P

'4 woo - g 1000-

500 - 500 -

0 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Slain diameter of droplets on water-sensItive paper (pm) Slain diameter of droplets on water-sensitive paper (pm)

2500 2500 (c) (d)

2000 - 2000 -

1500- 1500 - ; 1 g 1000- g 1000-

500 - 500

0 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (µ111)

Figure 4-10. Droplet size distribution graphs for the blue nozzle (105 L hr'), 1 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar

150 Liverwort epidemiology Chapter 4

2500 2500 (a) (6)

2000 2000 -

1500 " 1500 -

g 1000 - Iwo -

500 - 500-

0 t, 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1900 Stain diameter of droplets on water-sensItIve paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

2500 2500 (d)

2000 2000 -

1500 - 1500 2

g 1000 - Iwo

500 - 500

0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (p.m) Stain diameter of droplets on water-sensitive paper (pm)

Figure 4-11. Droplet size distribution graphs for the blue nozzle (105 L hr 1), 2 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar

151

Liverwort epidemiology Chapter 4

2500 2500 (a) (b)

2000 2000

1500 1500- e g 1000 4 l000

500 500

0 O 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

2500 2500 (c) (d)

2000 2000 -

1500 2

•g 1000

500 500 -

O 200 400 600 800 1000 1200 1400 200 400 600 BOO 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

Figure 4-12. Droplet size distribution graphs for the grey nozzle (60 L hr-1), 1 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar

152 Liverwort epidemiology Chapter 4

2500 2500 (a) (b)

2000 - 2000 -

1500 - I noo

I moo - g woo -

500 - 500 -

0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplet on water-sensItive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

2500 2500 (C) (d)

2000 - 2000 -

g 1500 -

noo -

500 - SOO -

0 0 200 400 600 800 1000 1200 1900 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

Figure 4-13. Droplet size distribution graphs for grey nozzle (60 L hr-1), 2 m (a) 1.5 bar (b) 2 bar (c) 2.5 bar (d) 3 bar

153

Liverwort epidemiology Chapter 4

2500 2500 (a) (b)

2000 2000

1500 1500 - S

g 1000 g 1000 -

500 500 -

0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

2500 2500 (t) (d)

2000 2000 -

1500 t 1500 - 2

•g 1000 g l000 -

500 500 -

O 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensitive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

2500 2500 (e) (0

2000 2000 -

1500 1500 -

▪ 1000

500 500

O 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400

Stain diameter of droplets on water-sensltive paper (pm) Stain diameter of droplets on water-sensitive paper (pm)

Figure 4-14. Droplet size distribution graphs for the brown nozzle (160 L hr1 ) (a) 1 m 2 bar (b) 1 m 2.5 bar (c) 1 m 3 bar (d) 2 m 2 bar (e) 2 m 2.5 bar (f) 2 m 3 bar

154

Liverwort epidemiology Chapter 4

4.2.2 Gemma cup replenishment 4.2.2.1 Introduction An experiment was designed to investigate how quickly gemma cups are replenished and if the rate of replenishment was a response to gemma dispersal or removal from the cup. If so the rate of gemmae replenishment could alter as a consequence of continual emptying, and the time interval between water applications could affect the amount of gemmae available for dispersal. Gemma cups were manually emptied in

different time frames and the number of gemmae removed was recorded.

4.2.2.2 Method All gemmae were removed from pre-identified cups using a 100-1000 pl pipette and counted. Three treatments were applied whereby gemmae were removed from cups and counted either every 3 days, weekly, or at the end of the experiment (4-weekly).

The gemma retrieval method used was to apply water to each gemma cup with a pipette, then to remove the water with gemmae, continuing to add water and remove

gemmae until no more could be removed.

Three healthy, active gemma cups per pot of liverwort were identified using coloured pins (red, blue, green); 10 pots of liverwort per treatment (30 pots in total). Identification of gemma cups allowed the tracking of gemma numbers produced by specific cups. Treatments were identified using coloured labels: 3-day, red (R); weekly, blue (B); 4-weekly, green (C). The pots of liverwort were positioned in a completely randomised design on capillary matting on a shaded glasshouse bench, providing damp, shady conditions (Table 4-7).

Table 4-7. Position of replicates. R = red, B = blue, C = 4-weekly. Numbers refer to replicates of each treatment. R1 C1 R2 R3 R4 R5 C2 B1 R6 R7 B2 B3 B4 R8 B5 C3 C4 C5 C6 B6 B7 R8 C7 B8 R10 C8 C9 C10 B9 B10

155 Liverwort epidemiology Chapter 4

4.2.2.3 Results The number of gemmae collected during the initial removal of gemmae was fairly constant across all treatments (Table 4-8). For subsequent collections, however, markedly more gemmae were collected during the 3-day treatment than either the weekly or 4-weekly (Figure 4-15)

Table 4-8. Initial gemma collection (12th May)

No. Gemmae Average for Standard

Collected treatment error 3-day 3131 104 21.28 Weekly 3180 106 21.02 4-weekly 3646 122 25.74

7000

6000

5000

a) E 4000 E0, 6 3000

2000

1000

3-day Weekly 4-weekly

Treatment

Figure 4-15. Total number of gemmae collected, excluding initial gemmae count

156 Liverwort epidemiology Chapter 4

250- 3-daily 200 —

150-

100-

50 —

0 iii 12 May 15iissi May 18 May 21 May 24 May 27 May 30 May 250 —

Weekly 200 —

0) gc 150 — E 0) cn ci 100 — Z

50-

0 I 1 i i 12 May 19 May 26 May 2 June 250- 4-weekly 200-

150-

100-

50-

0 I i 12 May 2 June

Treatment

Figure 4-16. Gemma cup replenishment; 3-day and weekly treatments and 4- weekly. 12th May was the initial collection.

157 Liverwort epidemiology Chapter 4

For the 3-day and weekly treatments the average number of gemmae collected each week declined overall during the course of the experiment. This effect was more pronounced for the first three data collections of the 3-day treatment (Figure 4-16), becoming more or less constant for the final four collections. For the 4-weekly treatment, more gemmae were counted in the final than the initial collection.

Statistical analysis (Table 4-9) using analysis of variance showed that the number of gemmae collected in each treatment differed significantly (F2,11 = 7.0, P<0.05). Further investigation using analysis of variance (Table 4-10) indicated that there was a highly significant difference between the 3-day and 4-weekly treatments (F1,8 = 15.39, P<0.01)., with more gemmae collected during the 3-day treatment.

Table 4-9. Analysis of variance comparing gemma cup replenishment of 3-day, weekly and 4-weekly treatments.

Source of Variation d.f. s.s m.s v.r. F.pr. Treatment 2 11982234 5991117.0 7.0 0.011 Residual 11 9414726 855884.0

Total 13 21396959

Table 4-10. Analysis of variance comparing gemma cup replenishment of 3-day and 4-weekly treatments.

Source of Variation d.f. s.s m.s v.r. F.pr. Treatment 1 11783103 11783103 15.39 <0.01 Residual 8 6122740 765342.4

Total 9 17905842

During the experiment some cups began to degenerate, becoming brown, and producing few, if any, gemmae. Some previously healthy gemma cups became discoloured and were surrounded by areas of dieback, possibly due to scorching during a period of hot, sunny weather; other gemma cups may have been adversely affected by gemma removal. This was particularly evident for the 3-day treatment where only

158

Liverwort epidemiology Chapter 4

30% of the designated gemma cups were still active, compared to 36% for the weekly treatment and 60% for the 4-weekly treatment, at the final data collection. The affected cups were still used for the remainder of the experiment. The high growth rate of thallus caused some gemma cups to be overgrown by new thallus. These were uncovered and gemmae counted as intended.

4.2.3 Clumping of gemmae 4.2.3.1 Introduction The aim of this experiment was to investigate if the biological rationale for gemmae to disperse in intact tightly packed clumps, rather than individually is to aid establishment and subsequent growth of liverwort thallus. The weight of liverworts established by clumps and freely dispersed groups of three different weight classes of gemmae were compared after four weeks growth. 'Clumps' were numbers of gemmae found naturally in tightly packed masses and dispersed intact. 'Groups' were clumps of gemmae that had first been separated in water and then poured onto the compost surface in a group.

4.2.3.2 Method A preliminary bioassay was used to devise a means of estimating the number of gemmae in a clump: clumps of gemmae were weighed, separated in water, and the number of gemmae counted. A regression equation was then used to estimate the number of gemmae within a clump of known weight.

Pots of mature liverwort were grown in shaded conditions and watered from below only, to promote clumping of gemmae by discouraging dispersal. Clumps were removed from gemma cups, weighed on a Sartorius 1712 balance and placed into Eppendorfs containing water, to separate the gemmae. They were then tipped onto filter paper and counted.

The linear regression constructed using this data gave the relationship between number of gemmae vs. clump weight with an adjusted R2 value of 76.8 and F prob. <0.001 (Figure 4-17). The balance used was not sensitive enough to weigh individual gemmae, and clumps of less than 30 were not found in gemmae cups, providing no reliable weights for small numbers of gemmae. To obtain such data, groups of 10 gemmae were formed from individually separated gemma, and weighed on a more

159 Liverwort epidemiology . Chapter 4 sensitive balance (Sartorius 4503 micro balance). The regression equation obtained with these additional data had an adjusted R2 of 84.6 and F prob. <0.001.

95% confidence intervals were plotted on the regression curve, indicating a 95% confidence that the population mean fell within the interval limits (Figure 4-17). The relationship between the number of gemmae and clump weight was close enough to use the data to predict the number of gemma within a clump of a specific weight, and this formed the basis of size class definitions. From these bioassays the value for individual gemmae was calculated using the equation: y = 0.1717 + 0.007124x (y = a + 13x, where x= no. of gemmae, y = weight of x no. of gemmae a 7 intercept and 1 slope).

The bioassay regression curve data was used to predict the weight of different sized groups of gemmae. From these, three discrete size classes were used as the basis of the experiment: small (26-50 gemmae), medium (101-125 gemmae) and large (176- 200 gemmae) (Table 4-11).

Table 4-11. Weights and predicted number of gemmae in small, medium and large size classes Size class No. gemmae Weight prediction (mg) 26 0.2569 Small (S) 50 0.4278 101 0.7911 Medium (M) 125 0.9621 176 1.3254 Large (L) 200 1.4964

160 Liverwort epidemiology Chapter 4

3.0 • No. gemmae vs weight Regression 2.5 - — — 95% confidence interval ▪ Additional data

2.0 - • • rn E • 1.5 - • co • //• 1.0 - • •

0.5 -

0.0 A'

0 50 100 150 200 250 300 350 No. gemmae

Figure 4-17. Linear regression of number of gemmae vs gemmae weight. 95% confidence interval indicates a 95% certainty that the population mean will fall between the limits. Adjusted R2 = 76.8. F probability = <0.001. 'Additional data' refers to groups of 10 gemma weighed on a more sensitive balance and not included in regression calculations.

Liverworts were grown under glasshouse conditions, watering from below only to encourage gemmae clumping. Naturally formed clumps (C) and groups (G) of separated gemmae of each size class (Table 4-12) were placed on damp growing medium in 10 cm pots (five pots per treatment). 'Groups' of gemmae were obtained by weighing clumps, placing them into Eppendorfs with water and shaking vigorously to separate the gemmae. The Eppendorfs were then emptied onto the growing medium, distributing the gemmae randomly, then rinsed and emptied twice more.

The experiment was replicated three times, giving a total of 90 pots. Pots were arranged in a completely randomised design on capillary matting on a shaded glasshouse bench, providing damp, shady conditions. Fresh and dry weights of liverworts were recorded after four weeks and an establishment rate calculated.

161

Liverwort epidemiology Chapter 4

Table 4-12. Experimental design showing size classes and treatment structure. CI = clumps, G = groups. Size classes: small (S) = 26-50 gemmae, medium (M) = 101- 125 gemmae, large (L) = 176-200 gemmae Small (S) Medium (M) Large (L)

Replications 1 2 3 1 2 3 1 2 3 Treatment C G C G C G C G C G C G C G C G C G No. pots 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5

Table 4-13. Completely randomised design of pots on glasshouse bench. CI = clumps, G = groups. Size classes: small (S) = 26-50mg, medium (M) = 101-125mg, large (L) = 176-200mg

M-3-C-1 L-1-G-1 M-3-G-1 S-1-C-1 L-1-C-1 S-1-G-1 L-2-G-1 L-2-G-2 S-1-G-2

S-2-G-1 L-1-C-2 L-2-G-3 L-3-C-1 M-3-C-2 S-2-G-2 L-2-C-1 L-2-G-4 S-2-G-3

M-1-C-1 M-2-G-1 L-3-G-1 M-2-G-2 M-3-C-3 S-3-C-1 M-2-C-1 M-1-G-1 S-3-G-1

L-3-G-2 L-1-G-2 S-1-G-3 S-3-C-2 L-2-C-2 M-2-C-2 S-1-C-2 L-1-G-3 M-3-C-4

S-2-C-1 M-2-G-3 L-1-C-3 M-2-C-3 S-3-G-2 L-1-C-4 S-1-C-3 L-2-G-5 M-3-C-5

M-2-C-4 M-2-C-5 M-1-G-2 L-2-C-3 M-1-G-3 L-1-G-4 M-1-C-2 M-1-G-4 L-3-C-3

M-1-G-5 S-1-G-4 L-3-G-3 S-2-G-4 L-1-C-5 L-3-C-2 M-1-C-3 M-3-G-3 S-2-G-5

S-2-C-2 M-3-G-2 S-3-C-3 L-3-C-4 M-1-C-4 S-3-G-3 M-1-C-5 M-3-G-4 S-2-C-3

S-1-C-4 M-3-G-5 M-2-G-4 L-3-G-5 S-3-C-4 L-1-G-5 S-1-G-5 L-3-C-5 S-2-C-4

S-1-C-5 L-3-G-4 S-3-C-5 L-2-C-5 M-2-G-5 L-2-C-4 S-3-G-4 S-2-C-5 S-3-G-5

162 Liverwort epidemiology Chapter 4

4.2.3.3 Results The results of this experiment were that both the fresh and dry weights were greater for groups of individual gemmae than clumps in all three size classes (Table 4-16, Figure 4-19, Figure 4-18). Considering the fresh weight of the gemmae, analysis of variance showed there was a significant effect due to the form of application Cclumps' or 'groups') (F1,10 = 17.28, P<0.002), with greater fresh weights obtained when gemmae were applied as 'groups' than 'clumps' (Table 4-14). For groups the effect was increased with size class, with the greatest dry weights obtained by the large size class and the least in the small. For clumps, however, there did not appear to be an overall trend, with very similar results obtained for large and small size classes.

Table 4-14. Analysis of variance of fresh weights of gemmae distributed by different application methods (`clumps' and 'groups') in different size classes.

Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 2 0.02830 0.01415 0.61 Block.*Units* stratum

Application method 1 0.39817 0.39817 17.28 0.002 Size class 2 0.01865 0.00932 0.40 0.678 Application.Size 2 0.02987 0.01494 0.65 0.544 Residual 10 0.23037 0.02304 Total 17 0.70536

When analysis of variance was applied to the dry weight results, a similar outcome resulted, with a significant effect of application form (F1,10 = 18.34, P<0.002), with greater dry weights obtained when gemmae were applied as 'groups' than 'clumps' (Table 4-15).

163 Total Table 4-15.Analysisofvariancedryweightsgemmaedistributedbydifferent gemmae, large(L)=176-200gemmae.'Clumps'groupsofgemmaefoundina Block.*Units* stratum Block stratum application methods('clumps'and'groups')indifferentsizeclasses. been separatedandthendispersed. weights wereobtainedbythelargesizeclassandleastinsmall.Forclumps, Size class Source ofvariation tightly packedmassanddispersedintact.'Groups'=clumpsofgemmaethathave again theredidnotappeartobeanoveralltrend,withverysimilarresultsobtainedfor Figure 4-18.Freshweightsofsmall(S)=26-50gemmae,medium(M)101-125 Liverwort epidemiology Residual Application.Size Application method As withfreshweights,theeffectincreasedsizeclass;forgroups,greatestdry large andsmallsizeclasses.

Fresh weight (g)

2 2 2 17 10 1 d.f. s

0.0043777 0.0013408 0.0002267 0.0002382 0.0024587 0.0001134 s.s. TrpAtmpnt 164 M

0.0001341 0.0001133 0.0001191 0.0024587 0.0000567 m.s. L 18.34 0.85 0.89 0.42 v.r. 0.458 0.442 0.002 F pr. Chapter 4 `Groups' =clumpsofgemmaethathavebeenseparatedandthendispersed. `Clumps' =groupsofgemmaefoundinatightlypackedmassanddispersedintact. Groups ofgemmaeestablishedmoresuccessfullythanclumps,with100% been separatedandthendispersed. 50 gemmae,medium(M)=101-125large(L)176-200gemmae. establishment betweenanyofthetreatments.Theredidnotappeartobetrend establishment inthesmallandmediumsizeclasses(Figure4-20).However,statistical analysis usingaChi-squaredcontingencytestdidnotshowsignificantdifferencein Table 4-16.Freshanddryweightspercentageestablishmentofsmall(S)=26- gemmae, large(L)=176-200gemmae.'Clumps' =groupsofgemmaefoundina Liverwort epidemiology within either'clump'or'group'treatmentsofgemmaeduetosizeclass. Figure 4-19.Dryweightsofsmall(S)=26-50 gemmae,medium(M)=101-125 tightly packedmassanddispersedintact.'Groups' =clumpsofgemmaethathave Establishment (%) Dry weight(mg) Fresh weight(mg)

Dry weight (g)

151.5 67 S-C 8.8 s

Clumps 60 96.1 M-C 6.5 165 Treatment M 111.0

73 9.1 L-C 287.9 100 22.5 S-C L 382.3 100 31.1 M-C Groups 533.3 41.8 87 Chapter 4 L-C 'groups' ofliverwortcomparedtothe'clumps'usedinthisexperiment,butitwouldbe These resultssuggestthatgemmaedispersinginaggregatedclumpsdonotaid Liverwort epidemiology that innaturegemmaclumpsareformedthesamemannerandtodegree gemmae werenotdispersedandclumpsformedartificially.Innaturetheymay as thoseusedinthisexperiment,whereoverheadwateringwasstoppedsothat establishment andsubsequentgrowthofliverwortcolonies.However,thisassumes would notsurvive.Thistreatmentdoesappeartosignificantlyenhancegrowthof form asamatterofcourseduringperiodswithoutrainoriftheatmosphereishumid; they mayaggregateasarenotdispersed.Thereforethe'groups'usedinthis environment. Itisalsopossiblethatclumpsofgemmaearemoreresistanttodrying experiment maybeabetterrepresentationofgemmadispersalinnatural out inthenaturalenvironment,witheffectonsurvivalofliverwortcoloniesonly provide amorecompletepicture. becoming apparentinextremedryconditionswhenfreelydispersedgroupsofgemmae useful totesttheseresultsagainstindividuallyplacedgemmaeonpotsofcompost gemmae, andlarge(L)=176-200liverwortclumps andgroups.'Clumps'=groups of gemmaethathavebeen separatedandthendispersed. of gemmaefoundinatightly packedmassanddispersedintact.'Groups' =clumps Figure 4-20.Establishmentofsmall(S)=26-50 gemmae,medium(M)=101-125

Average establishment (counts) 1 5 4 3 6 0 2

Treatment 166 M

L Chapter 4

Liverwort epidemiology Chapter 4

4.2.4 Characterisation of gemma dispersal in nursery irrigation systems 4.2.4.1 Introduction Previous experiments have suggested that liverwort gemma dispersal is affected by the irrigation system (nozzle type, water pressure) and by watering regime; the gemma cup regeneration experiment suggested that by reducing irrigation events fewer gemmae are produced and therefore available for dispersal.

A larger scale experiment, using water beds at East Mailing Research was designed to test these possibilities, combining different irrigation systems: two different nozzles, drip, capillary matting and no irrigation. Irrigation was provided either every two days or twice daily to the nozzle and drip irrigation systems. Capillary matting required markedly less water to maintain a moist environment. A 'seeder' pot of liverwort was placed in each treatment as a gemma source.

As gemmae are dispersed by water droplets, it should be the case that no dispersal would be observed in the capillary bed and drip treatments; no gemmae should establish and the 'seed' liverwort would be unlikely to survive summer temperature in the control treatment with no irrigation. Results of previous experiments suggest that more liverwort gemmae should be dispersed by the irrigation systems, and of these more should be dispersed in treatments irrigated twice daily than every two days.

4.2.4.2 Method For each treatment twenty 2 L pots were prepared with SHL Professional Potting Compost manufactured by William Sinclair Horticulture Ltd (Chapter 2.2.1); along with a 2 IL pot containing healthy liverwort as a gemma source, placed centrally. The position of each treatment, except the capillary bed, was randomly selected; all treatments were replicated twice, arranged in two blocks (Table 4-17). In the experimental design, all treatments were replicated once in each block, with drip and overhead irrigation applied using two time schedules, every two days or twice daily.

For the nozzle and drip irrigation applications, water was applied either twice daily or every two days. For the 2-day treatment, water was applied for 60 mins at 22.00 hrs, and for the twice daily treatment it was applied for 15 mins at 06.00 and 22.00 hrs. Therefore pots in each treatment received a total of 60 mins irrigation during each two day time period. For the capillary bed it was essential to prevent the capillary matting

167 Liverwort epidemiology Chapter 4 from drying out, thereby maintaining good capillary action between the growing medium in the pot and the matting. In preliminary tests, irrigation was applied for 1 min on 6 occasions in each 24 hr period. However, this resulted in the growing medium being too wet and application time was reduced to 1 min at 06.00 and 18.00. The overhead irrigation systems used two nozzles, MP Rotator model 1000 (nozzle 1) and Dan modular 180° spread (nozzle 2) (Supplier: Revaho Ltd) arranged as in Figure 4-24, supported on 60 cm risers.

Table 4-17. Experimental design: all treatments were replicated once in each block, with drip and overhead irrigation applied using two time schedules, every two days or twice daily.

Nozzle 1 Nozzle 2 Capillary MP Rotator model Dan modular 180° Drip Control bed 1000 spread Block 2-day Twice daily 2-day Twice daily 2-day Twice daily Twice daily Nil I (2-D) (TWD) (2-D) (TWD) (2-D) (TWD) (CM) Block 2-day Twice daily 2-day Twice daily 2-day Twice daily Twice daily Nil II (2-D) (TWD) (2-D) (TWD) (2-D) (TWD) (CM)

The MP Rotator® sprinkler, manufactured by the Walla Walls Sprinkler Company, USA, delivers adjustable arcs and radii of droplets in individual rotating streams, with flow changes proportional to the area being covered following changes to match the original precipitation; it is marketed as having a low application rate and high uniformity.

The capillary bed construction consisted of a layer of plywood with a wooden frame covered with a waterproof black plastic membrane overlapping the frame to prevent water run-off and to provide a smooth, level base. Within the frame was layered

168 Liverwort epidemiology Chapter 4

Pots

Irrigation feed pipe

Microtubing

Figure 4-21. Example drip irrigation arrangement with microtubes connecting the feed pipe with a dripper held in position in each pot with a stake.

Florimat 2 capillary matting (Supplier: Flowering Plants Ltd) to hold water and an upper layer of micro-perforated polythene to keep the matting clean whilst still allowing free water movement, (Figure 4-22(a)). Water was supplied to the matting via NetafimTM Streamline 80 trickle tape with 20 cm hole spacing (Figure 4-22(b)). The beds provided were on a slope, so the most level area of each block was selected for the capillary beds, rather than being randomly allocated, therefore this treatment was used for observation only, and was not included in statistical analysis.

Drip irrigation consists of a feed pipe with micro-tubes with a dripper held in place with a small stake; this system provides water direct to the growing medium, with no wastage due to water falling outside the pot. Although efficient, this system takes longer to set up than overhead or capillary matting systems.

Irrigation timing was controlled with two Orbit Sunmate automatic garden watering systems, one each for the 2-day and twice daily treatments; and a Kompernass Watering Computer KH 4038 (Supplier: City Irrigation Ltd) for the capillary beds.

169

Liverwort epidemiology Chapter 4

(a)

• 4 .!....!&!:—.....NINIeNtNt&NA 3 2

(b)

Wooden frame

Netafim tape

41— —Water supply

Figure 4-22. Capillary bed construction (a) Layers comprising the base: 1-plywood, 2- black plastic waterproof membrane, 3-capillary matting, 4-perforated plastic (b) Layout of Netafim tape, positioned between layers 2 and 3, and water supply.

Two beds, (block I and block II), were constructed from Mypex laid on grass and edged with concrete blocks. A Spanish tunnel construction protected the beds from rain whilst allowing free air movement. Screens were constructed from clear plastic supported by nylon rope connected to tunnel support and stakes protected treatments from contamination from neighbouring sprinkler irrigation, down the side of the tunnel and at each end, providing protection from the prevailing weather conditions (Figure

4-23, Figure 4-25).

170 Liverwort epidemiology Chapter 4

Figure 4-23. Spanish tunnel construction: (a) side view (b) treatment area with pots protected by plastic screens (c) screen attachments using elasticated tarpaulin ties and balls (d) view of general layout.

Data collection: The number of gemmae dispersed was measured by positioning 10 collection pots per treatment for the duration of a full irrigation cycle, 2 days, and the number of gemmae collected recorded.

Final results taken were the surface area of the pot covered by liverwort and the number of liverwort colonies present. Water droplet size was measured using water sensitive paper using the method described in section 4.2.1.2.

171 Liverwort epidemiology Chapter 4

4.2.4.3 Results Results for the capillary matting treatment were not included in the statistical analysis as their position within the experiment could not be randomly allocated due to the slope of the ground, and the requirement for capillary beds to be level.

Seed pots The liverwort in the seed pots for the control treatments, with no water applied predictably died within eight days (when the first results were collected). The seed liverwort in the Block II, nozzle 1, two daily treatment had died by 56 days. Some dieback was observed on the Block II, Drip, twice daily treatment and Block H, drip, two daily treatment. After 56 days the seed pots were still healthy in the remaining treatments, with the majority bearing gametophores, which were removed manually to prevent spore dispersal.

Gemma distribution within pot and relative to source Observations were made relating to the distribution pattern of gemmae; in the Block II-N2-2D and Block I-N1-2D treatments, more gemmae were established in pots nearest to the seed pot, whilst for the remainder of the treatments they appeared randomly distributed, showing no clear trend. Some gemmae had established in the drip treatment, and although some of these were near to the dripper, others were randomly distributed, suggesting the drippers were not involved in gemmae dispersal.

Where treatments had many gemmae established in a pot they were small and congested, however where a small number of liverworts had established the liverwort thalli tended to be of larger size. This pattern of growth was also observed in the shading experiment (Chapter 3).

172 Liverwort epidemiology Chapter 4

Block I Block II

• III • , 1 • 111 a ENE • •

N N2- N1- 2D 2D f 4

IIEMI MM. Il NM 11 J _ _ _ _ ,I

D- C 2D t .:

N1- N2- TWD TWD + I+ . _ . _ .1_

D- iii.. TWD CM •

1 I 1 ...... 1 L MIME • •0 T

Water supply in

Figure 4-24. Layout of treatments showing irrigation design. T=timer, N=nozzle, C=control, CM=capillary bed, D=drip, TWD=twice daily, 2D=2-daily. Dotted lines indicate pipes connecting mains water supply, timers, sprinklers and drippers and capillary beds. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread).

173

Liverwort epidemiology Chapter 4

Block I Block II + r r "T" T 1 I I I 1 I I I N1- N2- N2- N1- 1 I I I TWD TWD 2D 2D 1 I I I I I I I I- 11 .1- -I- -1. 1 I I

1 D- I N2- I D- 1 TWD I 2D I C 2D 1 I I 1 I I r r lr" T 1 I I I 1 N1- I I N1- I N2- 1 2D I C I TWD I TWD 1 I I I 1 1 I I 1 I 1 I 1 D- I D- 1 2D CM I TWD CM 1 I 1 I I

Figure 4-25. Layout of treatments showing irrigation design. T=timer, N=nozzle, C=control, CM=capillary bed, D=drip, TWD=twice daily, 2D=2-daily. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 1800 spread). Dotted lines indicate position of clear plastic screens. The solid line to the right indicates the position of the neighbouring polytunnel. Screens were omitted between treatments where overhead irrigation was not used and was therefore unlikely to affect adjacent plots.

174 Liverwort epidemiology Chapter 4

Number of gemmae dispersed A measure of the number of gemmae dispersed was made by counting the number of gemmae that fell into collection pots on five occasions (Figure 4-26 b). A trend soon emerged, with no gemmae collected in any of the drip, control or capillary bed treatments. With the nozzles, gemmae were collected in all treatments except the two replications of the N2-TWD treatment. More gemmae were collected in the N1-TWD than the N1-2D or N2-2D treatments.

Liverwort establishment The number of pots with any liverwort established was recorded (Figure 4-26a). It had been expected that no liverworts would be dispersed or established in the capillary bed, drip and control treatments; however, this was not the case. These could have come from a number of sources: spores produced either by liverworts used in the experiment, or growing elsewhere on the site present in the air, contamination by gemmae from other treatments, or gemmae from seed pots when they were placed in their positions. Care was taken to prevent such contamination by erecting screens between treatments, removing gametophores from seed liverworts and by careful handling of pots. A greater number of pots contained established liverwort in the nozzle and drip treatments irrigated twice daily than every 2-days, with more established in the N1 than the N2 treatments.

Statistical analysis using a Chi-squared contingency test confirmed a significant effect of dispersal treatment on establishment (P = 0.001). Further investigation of the results indicated the greatest difference was found between the control and the other treatments; large differences were also found between the twice daily and 2-day applications for both nozzles and the drip treatments. Differences between application methods were smaller than differences between application timing.

175 Rotator model1000),N2=nozzle2(Danmodular 180 Figure 4-26.Liverwortinfestationafter56 days:(a)Liverwortestablishment, days, TWD=irrigationtwicedaily. gemmae collectedinpotsduringonetwoday irrigationcycleN1=nozzle1(MP gemmae dispersed.Resultsforeachtreatment areanaverageofthenumber each treatmentaretheaverageoffortypotsarranged intwoblocks.(b)Numberof measured asthenumberofpotswithmorethan onegemmapresent.Resultsfor C =control(noirrigation),CMcapillarybed irrigation.2D=irrigationeverytwo Liverwort epidemiology

Average no. gemmae No of pots withliv erwort established 100 - 300 - 200 - 400 0 I i1 N1-2D N2-2DN1-TWDN2-TWDD-2DD-TWDC N1-2D N2-2DN1-TWDN2-TWDD-2DD-TWDC

176 Treatment Treatment ° spread),D=dripirrigation,

CM CM (b) Chapter 4 Liverwort epidemiology Chapter 4

The amount of liverwort in each pot was measured using two methods (Figure 4-27): counting the number of liverwort colonies present and recording their surface area. The same trends were exhibited using either method. More liverwort was present in the twice daily than 2-day treatments of the nozzle 1 and drip treatments. Results for nozzle 2 were almost identical with average liverwort areas of 1.84 cm2 (2-day treatment) and 1.82 cm2 (twice daily treatment). The number of colonies was greater in the twice daily (533) than the 2-day (445) treatment. The area of liverwort present in the capillary bed treatments were less than all other treatments except for the drip,

2-day and the control. However, the number of liverworts established was less than all treatments other than the control.

Analysis of variance (Table 4-18) showed that there was a significant effect of the treatments on liverwort area in pots (F7,8 v.r.= 6.07, P<0.05). Further investigation using t-tests showed that significant differences occurred between the Nozzle 1, twice daily treatment and the drip 2-daily treatment (p value = 0.039155, P<0.05); and between the Nozzle 1, twice daily treatment and the control. However, the number of colonies present in the pots did not differ significantly.

Table 4-18. Analysis of variance results of area of liverwort

Source of Variation d.f. s.s m.s. v.r. P-value Between Groups 7 1330.18 190.0257 6.072897 0.010539 * Within Groups 8 250.3263 31.29078

Total 15 1580.506

177 Liverwort epidemiology Chapter 4

20 (a) 18 -

) 16 - 2

(cm 14 - e

12 - coverag t 10 -

8 - liverwor

e 6 -

Averag 4 -

2 - o I1111 li Nam i 1111 N1-2D N2-2D N1-TWD N2-TWD D-2D D-TWD C CM Treatment

140 (b)

120 -

ies 100 - lon co t r 80 - liverwo 60 - no. e

40 - Averag

20 -

0 1i a , --T"-- N1-2D N2-2D N1-TWD N2-TWD D-2D D-TWD C CM Treatment

Figure 4-27. Liverwort infestation after 56 days. (a) Liverwort area. (b) No. of liverwort colonies. Results for each treatment are the average of forty pots arranged in two blocks. N1 = nozzle 1 (MP Rotator model 1000), N2 = nozzle 2 (Dan modular 180° spread), D = drip irrigation, C = control (no irrigation), CM = capillary bed irrigation. 2D = irrigation every two days, TWD = irrigation twice daily.

These results are consistent with those of the gemma cup replenishment experiment, where more gemmae were produced in gemma cups from which gemmae were moved in the 3-day than either the weekly or 4-weekly treatments.

178 Liverwort epidemiology Chapter 4

Droplet size distribution graphs were constructed for each nozzle treatment (Figure 4-29), and overall graphs for the MP Rotator model 1000 nozzles (nozzle 1) and Dan modular 180° spread nozzles (nozzle 2) (Figure 4-28). The greatest number of droplets was produced by the Dan modular nozzle, with a mean diameter of 341 pm, compared to a mean diameter of 546 pm diameter droplets produced by the MP Rotator nozzles. The greatest mean number of gemmae (total) dispersed by the nozzle treatments was 60.7 (MP rotator model 1000), with a mean of 12.2 by the Dan modular nozzles. This suggests that overall the number of gemmae dispersed is again dependent on droplet size rather than number of droplets, in accordance with the results of the previous gemma dispersal experiment (Table 4-5), where the mean diameter of droplets was 161 to 271 pm, smaller than those produced by either the Dan modular nozzles (341 pm) or the MP Rotator (546 pm). The overall mode of droplet size produced by the Dan modular nozzles was 160 pm, and of the MP Rotator 182 pm, both falling within the size range of the mode droplet diameters produced by nozzles in the previous experiment. Had smaller droplet sizes been produced during this experiment it may have resulted in a greater number of gemmae being dispersed.

Table 4-19. Mean droplet sizes (pm) produced by nozzle 1 (MP Rotator model 1000) and nozzle 2 (Dan modular 180° spread). 2-day Twice daily N1 554 539 N2 249 433

When the data for timing of application and nozzle type are considered separately the trends are less clear (Table 4-19): the greatest number of gemmae were dispersed by the Nozzle 1, twice daily, and the least by the Nozzle 2, 2-day treatments.

179 Liverwort epidemiology Chapter 4

3000

2500 -

2000 -

ts le op 1500 - f dr o No. 1000 -

500 -

0 T I

0 500 1000 1500 2000 2500 3000

Stain diameter of droplets on water-sensitive paper (µm)

3000 Dan modular 1800 spread nozzle

2500 -

2000 -

t le

drop 1500 - f o No. 1000 -

500 -

0 T

0 500 1000 1500 2000 2500 3000

Stain diameter of droplets on water-sensitive paper (µm)

Figure 4-28. Characteristic size distributions of droplets produced by the MP Rotator model 1000 (nozzle 1) and Dan modular 1800 spread (nozzle 2) nozzles.

180

Liverwort epidemiology Chapter 4

N1-2D N1-1W5

2000 2000 -

1500 t 1500 -

g 1000 •d 1000

500 500 -

500 1000 1500 2000 2500 3000 500 1000 1500 2000 2500 3000

Stain diameter of droplets on water-sensitNe paper (pm) Stain diameter of droplets on water-sensitNe paper (pm)

N2-25 N2-TWD

2000 2000 -

1500 1500 -

g 1000 •g 1000 -

500 500 -

I 0 _11111.._

500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000

Stain diameter of droplets on water-sensitNe paper (µm) Stain diameter of droplets on water-sensitive paper (0m)

Figure 4-29. Characteristic size distributions of droplets produced by nozzles. N1 = MP Rotator model 1000 nozzle, N2 = Dan modular 1800 spread nozzle, I = block 1, II — block 2, 2D = irrigated every two days, TWD = irrigated twice daily.

4.3 Discussion Dispersal In the initial dispersal experiment (4.2.1) generally more gemmae were dispersed using 2 and 2.5 bar water pressures, and these were dispersed closer to the source, coincidentally corresponding with the manufacturer's recommendation that 2 bar is the optimum operating pressure for this particular nozzle range. There were exceptions to each trend shown in the results due to the complex interactions between water pressure, nozzle and nozzle height, all of which impacted on drop size and the number of drops produced; however the number of gemmae dispersed does appear to be related to water drop size, not number of drops.

The dispersal distances of 160 cm obtained in this experiment far exceed the distances previously reported (Brodie, 1951; Equihua, 1987). However the results are not directly comparable as the dispersal equipment and conditions used were different.

181 Liverwort epidemiology Chapter 4

In each treatment the distance travelled by gemmae reaffirmed the ability of liverwort to disperse and colonise new areas in nursery situations. The preliminary experiment using dye indicated that whilst gemmae may be dispersed within splash droplets, they are also propelled through air by the incident water drop.

When using the brown nozzle, more gemmae were dispersed with increased nozzle heights, in accordance with findings reported by Geagea (1999). However, there was no clear effect of nozzle height on the number of gemmae dispersed for the grey (flow rate 60 LW') and blue nozzles (flow rate 105 L hr-1) as the results were more variable. Distances travelled by the gemmae increased with increased nozzle height. In contrast with Geagea's findings, for the blue and brown nozzles some gemmae travelled further when the nozzle was at the lower position, although the bulk of the gemmae did travel further when nozzles were higher. Steeper dispersal gradients were obtained at 2 and 2.5 bar than 3 and 1.5 bar; the exception was the 3 bar, 2m height treatment using the grey nozzle.

At the two extreme water pressures (1.5 and 3 bar), generally fewer gemmae were dispersed for all nozzles types, indicating there may be an optimum combination of water pressure and flow rate for dispersing maximum gemmae numbers. This would appear to coincide with the manufacturers recommendation that 2 bar is the optimum operating pressure for this particular nozzle, as above 2.5 bar and below 1.5 bar water distribution becomes uneven, and damage may be caused to the emitters, so they fall outside the terms of the guarantee (Agridor, Undated).

There may be a relationship between gemma cup size and droplet size. Typical gemma cup diameter ranges between 0.5 and 2.0 mm, however gemma cups were not measured for these experiments, therefore no data is available to specifically relate gemma dispersal, droplet size and gemma cup size.

Gemma cup replenishment The aim of this experiment was to investigate how gemma cups replace dispersed gemmae, and whether this is affected by irrigation regime. The results suggest that when gemmae are removed, more are produced to replenish them; the total number of gemmae produced by a gemma cup could be minimised by reducing or eliminating overhead irrigation events.

182 Liverwort epidemiology Chapter 4

Clumping of gemmae The hypothesis for this experiment was that gemmae clump together as an aid to establishment and growth. However, the results obtained under the conditions of this experiment suggest that dispersed groups of gemmae establish better, with 100% establishment rates in the small and medium size classes.

Fresh and dry weights were also greater for groups of gemmae than clumps, increasing with class size for 'groups' of gemmae, but remaining more constant for clumps.

The clumps of gemmae used in this experiment were encouraged to form by watering the liverworts from below only, so that gemmae were not dispersed. However, the clumps observed in previous experiments occurred naturally, even with overhead watering; there may be other factors that influence their formation and these may affect subsequent establishment and growth.

The results of the final nursery-scale dispersal experiment (Section 4.2.4) reflect those of the gemma cup replenishment experiment where more gemmae were produced in gemma cups from which gemmae were removed in the 3-day than either the weekly or 4-weekly treatments.

Droplet size distribution graphs were constructed for each nozzle treatment (Figure 4-29), and overall graphs for the MP Rotator model 1000 nozzle and Dan modular 180° spread nozzles (Figure 4-28). The greatest number of droplets was produced by the Dan modular nozzle, with a mean diameter of 341 pm diameter, compared to 546 pm diameter droplets produced by the MP Rotator nozzles. The greatest mean number of gemmae dispersed by the nozzle treatments was 60.68 by the MP rotator and 12.25 by the Dan modular nozzles. This suggests that overall the number of gemmae dispersed is dependent on droplet size rather than number of droplets, in accordance with that of the previous gemma dispersal experiment (Section 4.1.1); the mean diameter of those droplets was 161 to 271 pm, smaller than those produced by either the Dan modular nozzles (341 pm) or the MP Rotator (546 pm). The overall mode of droplets produced by the Dan modular nozzles was 160 pm, and of the MP Rotator 182 pm, both falling within the size range of the mode droplet diameters produced by nozzles in the

183 Liverwort epidemiology Chapter 4 previous experiment. Had smaller droplet sizes been produced during this experiment it may have resulted in a greater number of gemmae being dispersed.

These results add further to findings (Burgess, 2003a; Burgess, 2003b) that sub- irrigation should be practiced in nurseries to reduce liverwort infestation. Sub-irrigation can also provide cost savings of 25-35% over well designed overhead irrigation systems, improving water use efficiency, uniformity of water distribution and plant quality.

184 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Chapter 5 The use of glucosinolate hydrolysis products as herbicides

5.1 Introduction In this section of work glucosinolates (GSLs) and their hydrolysis products, isothiocyanates (ITCs) were investigated for their effect on liverwort gemmae. Methods were developed for the extraction and identification of 3-methoxybenzyl GSL (glucolimnanthin) and 3-methoxybenzyllTC from Limnanthes a/ba seed meal, and the GSL profile of the whole plant established. L. a/ba was used as previous research (Svenson and Deuel, 2000) indicated some success in controlling liverwort using L. a/ba seed meal as a mulch (Chapter 1). Although the GSL glucolimnanthin (Figure 5-1) and its hydrolysis products were previously known (Bartelt and Mikolajczak, 1989; Vaughn et a/., 1996), the full GSL profile of L. a/ba had not been elucidated.

OH HO \•••_ HO \/S OH

OC H3

Figure 5-1. Chemical structure of glucolimnanthin.

185 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Various experimental methods have been described for harnessing plant allelochemicals, including ITCs, by collecting and utilising root exudates to control weeds. Examples are plants grown in soil culture and their root exudates collected either in run-off water (Pope etal., 1985), or the plants transferred to a soilless system for exudate collection (Khan et al., 2002). Alternatively, seedlings have been grown in a hydroponic system with bioactive compounds collected on in-line C-18 columns (Tsanuo et al., 2003). Yamane et a! (1992) collected Rorippa indica (Indian cress) plants from the wild and transferred them to a hydroponic system with continuous root exudate trapping based on that designed by Tang and Young (1982), collecting the ITCs on an in-line column containing Amberlite XAD-4 resin (Supplier: Sigma-Aldrich). Amberilite is a hydrophobic polyaromatic compound used to remove small hydrophobic compounds from a solution, for example organic substances from aqueous systems and polar solvents (Sigma-Aldrich, 2006).

Plant species (Figure 5-2) were selected for their high root GSL content (Fahey et al., 2001; Kirkegaard, 1998), to provide a range of bioactive products for use in bioassays; their taxonomic detail, germination conditions and root GSL profiles are summarised in Table 5-1. These plants were grown in a nutrient film technique (NFT) hydroponic plant growth system, and methods were developed to collect and analyse their root exudates from the nutrient solution, based on methods described by Yamane et al (1992). L. alba plants were also grown in this system to provide clean plant tissue for GSL and ITC extraction and bioassays.

In vitro bioassays were designed using the predominant ITCs (benzyl-, 2-phenylethyl-, 2-propenyl- and 3-methoxybenzyl-) (Figure 5-3) found in the root tissues of the plants grown hydroponically (Table 5-1). These were used to investigate the effect of ITCs on the growth of liverwort thalli and on cress (Lepidium sativum) radicle germination and elongation to provide a comparison with higher plants.

As an indicator of the potential usefulness of ITCs in controlling liverwort comparative bioassays were carried out using two herbicides, lenacil and metazachlor (Figure 5-4), found within different herbicide groups and currently recommended for use with liverwort (Atwood, 2005).

186 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Figure 5-2. Plant species used in GSL experiments (a) Limnanthes a/ba (b) Diplotaxis tenuifolia (c) Sisymbrium orientale (d) Brassica juncea.

187 Root GSL profile

Seed Germination Plant Plant family Chemical name Trivial name GSL classification source* conditions 2-phenylethyl Gluconasturtiin, Phenethyl Aromatic 2-propenyl Sinigrin, Allyl Aliphatic - Olefin (straight chain, double bond) 3-butenyl Gluconapin Aliphatic - Olefin (straight chain, double bond) John 4-methoxy-3- 4-Methoxyglucobrassicin Indole Brassica juncea L. Czern. Rossiter, 2 mm deep, indolylmethyl (Brown or Indian mustard) Brassicaceae Glucobrassicin Indole Imperial 15-25 °C 3-indolylmethyl Herbaceous perennial College 1-methoxy-3- Neoglucobrassicin Indole indolylmethyl benzyl Glucotropaeolin Aromatic 4-pentenyl Glucobrassicanapin Aromatic 4-hydoxy-3- 4-Hydroxyglucobrassicin Indole indolylmethyl benzyl Glucotropaeolin Aromatic 4-methylsulphinylbutyl Glucoraphanin Aliphatic - AlkylThioAlkyl - Dip/otaxis tenuifolia D. DC Herbiseeds (Perennial wall rocket) Brassicaceae None p-hydroxybenzyl Glucosinalbin Aromatic UK Herbaceous perennial 2-phenylethyl Gluconasturtiin, Phenethyl Aromatic 3-Indolylmethyl Glucobrassicin Indolyl 2-phenylethyl Gluconasturtiin, Phenethyl Aromatic 3-butenyl Gluconapin Aliphatic - Olefin (straight chain, double bond) Sisymbrium orientale L. p-hydroxybenzyl Glucosinalbin Aromatic Herbiseeds (Indian hedge mustard) Brassicaceae 5mm deep UK 1-methoxy-3- Neoglucobrassicin Indole Herbaceous annual indolyl methyl 4-methoxy-3- 4-Methoxyglucobrassicin Indole i ndolyl methyl 3-indolylmethyl Glucobrassicin Indole Massey Umnanthes alba Hartw University ex Benth. (Meadowfoam). Limnanthaceae New 12 °C, dark Herbaceous winter annual Zealand Table 5-1. Details of selected plant species with root GSLs listed in order of abundance. The root GSL profile for L. a/ba is unknown (Fahey et al., 2001; Kirkegaard, 1998).

188 The use of glucosinolate hydrolysis products as herbicides Chapter 5

CH, ____,NCS --CH2 H2C...,,,,..„ CH2 "--,„,„, ,,,,•---- ..., CH NCS

2-PhenylethyllTC 2-PropenyllTC

. CH2 . - 'C S-CN 9 NCS CH3

BenzylITC 3-MethoxybenzyllTC

Figure 5-3. Chemical structures of ITCs used in bioassays

0 II / cH2_ H CI„cH2 N N 2,C, N C O

H3C CH3

0 NICII

Metazachlor (C14hl16CIN30) Lenacil (C13H18N1202)

Figure 5-4. Chemical structures of herbicides used in bioassays

Lenacil, introduced in 1964, is a residual, selective, systemic herbicide (Tomlin, 2000), one of three belonging to the uracil group of herbicides, which are comprised of a uracil nucleus with various chemical groups substituted. Uracils are carried to the root zone by water, absorbed by the roots and move through plants in the transpiration stream (Ware and Whitacre, 2004). Lenacil's mode of action is to block photosynthetic electron transport at the photosystem H receptor site. It is used to control annual grass, broad-leaved weeds in various field crops, ornamental plants and shrubs, applied pre-planting or pre-emergence (Tomlin, 2000). Residual activity is reduced in soils with high organic matter; moist conditions are required to sustain residual activity, dry conditions reduce efficacy. A number of products are available with lenacil as the

189 The use of glucosinolate hydrolysis products as herbicides Chapter 5 active ingredient in the UK, with Clayton lenacil 80W (MAPP No. 09488) approved for use in ornamental plant production. It is supplied as a wettable powder containing 80% w/w lenacil (Pesticide Safety Directorate, 2006; Tomlin, 2000; Whitehead, 2006).

Metazachlor, introduced in 1982, is a residual, selective chloroacetamide (syn. acetanilide) herbicide absorbed by plant hypocotyls and roots, with activity dependent on root uptake (Tomlin, 2000; Whitehead, 2006). Chloroacetamides are meristematic growth inhibitors, with inhibition of long chain fatty acids the predominant mode of action (Ware and Whitacre, 2004). Metazachlor is used pre- and early post-emergence, either incorporated pre-planting or surface applied; effectiveness is reduced on soils with over 10% organic matter content. It is approved for use in brassicas, nurseries and forestry, where it is used to control winter and annual grasses and broad-leaved weeds in fruit and vegetable crops and ornamental plants and shrubs. A number of products are available for use in ornamental plant production with metazachlor as the active ingredient, e.g. Butisan S (Pesticide Safety Directorate, 2006).

It was considered that the GSL hydrolysis product 3-methoxybenzyllTC, found in L. a/ba seed meal, was the active substance that reduced liverwort presence on the compost surface in experiments carried out by Svenson and Deuel (2000). Vaughn et a/ (1996) had previously found 3-methoxybenzyllTC to be toxic against velvetleaf and wheat seedlings. One aim was, therefore, to extract this ITC from L. a/ba seed meal, identify it and apply it to liverwort gemmae in bioassays to measure the direct effect on their growth. Three other selected ITCs (benzyl, 2-propenyl and 2-phenylethyl) and the herbicides metazachlor and lenacil were similarly applied to liverwort gemmae and their effects measured. Additionally the previously unknown GSL profile of all L. a/ba tissues was established.

ITCs have been collected from plant root exudates using an in-line column containing XAD-4 resin (Tang and Young, 1982). It was hypothesised that these methods could be used to collect ITCs, which could then be tested for herbicidal effect against liverwort gemmae. The objectives were to grow selected plants with known high levels of root GSLs in a hydroponic system, collect and identify the ITCs produced and then apply them to liverwort gemmae in laboratory bioassays.

190 The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.2 General methods

5.2.1 Hydroponic growth of plant species

A nutrient film technique (NFT) hydroponic system was designed using black guttering for the NFT channel to contain plants (Figure 5-5, Figure 5-6). Nutrient solution was pumped around the system from a 20 litre reservoir by a submersible pump via plastic tubing, into one end of the channel, over the plant roots and then returned to the reservoir, keeping the solution aerated. Seedlings were grown in 3 cm rockwool blocks which rested in the nutrient solution. The guttering, rockwool and reservoir were wrapped in co-extruded black on white plastic film, with the white layer to the outside to reduce algal growth.

Feeding regime Genesis Formula, manufactured by Green Air Products a proprietary hydroponic nutrient solution fertiliser concentrate system was used (supplier: Growth Technology), supplied as 2 parts: Microbase providing nitrogen, phosphorus, potassium and micronutrients; and Grow 2 providing nitrogen and potash (Table 5-2). Both Grow 2 and Microbase were applied at 3.78 ml 1_4 to provide optimum growth conditions; the nutrient solution was changed weekly.

As the plants were growing in alkaline rockwool, which raises the pH, the nutrient solution was adjusted to pH 5.8 using 'pH Down' (81% phosphoric acid), manufactured by Essentials Hydroponics (supplier: Growth Technology). Using the recommended nutrient doses an acceptable electric conductivity (E.C.) of 1.2 mS cm"1 was obtained. E.C. is a measure of nutrient concentration which, if too high can result in root dehydration, and if too low in lack of growth.

Seed germination Rockwool cubes (3 cm) were soaked in water, a well made in the centre and two seeds sown in each. They were then placed in a growth chamber with the correct germination conditions for the plant species (Table 5-1). Once germinated the weakest seedling was removed from each rockwool block, the block transferred to the NFT system (Figure 5-6), and fed with nutrient solution at half the required concentration for the first week.

191 The use of glucosinolate hydrolysis products as herbicides Chapter 5

NFT channel Seedlings growing in rockwool blocks

Reservoir

containing

Submersible

Figure 5-5. Diagram of hydroponic system. Nutrient solution is pumped from the reservoir by the submersible pump to the end of the NFT channel, flows down the channel over the plant roots and back into the reservoir in the direction of the blue arrows.

Figure 5-6. Hydroponics system set up: a) two NFT channels with their nutrient reservoirs, b) NFT channel containing a Brassica juncea seedling in its rockwool block, c) a Limnanthes alba seedling with its block and NFT channel covered in light omitting plastic

192 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Table 5-2. Genesis Formula nutrient solution components

Microbase B. N:P:K 3:10:19

Component 0/0 Total nitrogen 3 Available phosphate (P205) 10 Soluble potash (K20) 19 Total magnesium (Mg) 3 Sulphur (S) 4 Boron(B) 0.03 Cobalt (Co) 0.002 Copper (chelated) (Cu) 0.02 Iron (chelated)(Fe) 0.2 Manganese (chelated)(Mn) 0.05 Molybdenum (Mo) 0.003 Chlorine (CI) <.12 Zinc (chelated)(Zn) 0.05 Grow nutrient 2. N:P:K 15:00:08 Component 0/0 Total Nitrogen 15 Ammoniacal nitrogen 3 Nitrate nitrogen 12 Soluble potash (K20) 8 Calcium (C) 8

5.2.2 Buffer preparation

Buffer solutions were prepared as follows:

Tris buffer — pH 7.0, 8.0 and 9.0: 100 ml 0.1 M Tris(hydroxymethyl)methylamine (Tris) (MW 121.14 g moll in milli-Q water, the pH adjusted to 7.0 with 5 M hydrochloric acid.

Acetate buffer — pH 5.0 and pH 4.0 100 ml 0.1 M glacial acetic acid (MW = 60.05 g moll in water, the pH adjusted with 5 M NaOH.

Phosphate buffer - pH 6.0 0.2 M solutions of NaH2PO4 (MW = 119.98 g L-1) and Na2HPO4 (MW = 141.96 g Ll were each made up to 100 ml with water, mixed (6.1 ml: 43.9 ml, NaH2PO4:Na2HPO4)(Dawson et al., 1969) to obtain pH 6.0.

193 The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3 Experimental work

5.3.1 Glucosinolate extraction from L. a/ba seed meal

5.3.1.1 Introduction

L. alba seed meal has been shown to have an effect on liverwort proliferation in nursery containers (Svenson and Deuel, 2000). To examine if GSLs present in the seeds could have caused this inhibition, methods were developed to isolate and identify any GSLs; the GSL hydrolysis products were then used in bioassays to determine any effect on liverwort gemma growth. Seeds of L. a/ba were defatted to remove the oil and volatile components, and the GSL content extracted, using a known method (Thies, 1988), quantified and identified.

5.3.1.2 Method

Defatting seeds

L. a/ba seed (150 g) was ground until fine using a coffee grinder and weighed. It was defatted using Soxhlet apparatus with petroleum ether (40-60 °C bp) and a few anti- bumping granules added. The ground seed was placed in a cellulose extraction thimble, the top plugged with cotton wool and placed into the glass Soxhlet chamber and, with the condenser operating, sufficient heat applied to allow the petroleum ether to gently reflux. When all the oil had been extracted the thimble was removed and left for the seed powder to dry thoroughly.

GSL extraction For the GSL extraction, defatted seed meal (18.2 g) was extracted in methanol (150 ml, 4 hrs) with a Soxhlet extractor, and the methanol evaporated until nearly dry under reduced pressure (Rich' Rotavapor RE120, 35 °C). The extract was twice dissolved in water (10m1) and the concentrates transferred to a conical flask (50 ml). 0.5 M Lead acetate (1 ml) and 0.5 M barium acetate (1 ml) were added to the extract, mixed, allowed to stand at 4 °C and centrifuged for 30 min (3,500 rpm). The supernatant was decanted through a column of 1g dry weight Sephadex DEAE-A-25, formiate form (Supplier: Sigma-Aldrich), in a tube of 12 mm internal diameter, 100 mm long, and washed through twice with 10 ml water (discarding the run off solution each time); twice with 5 ml formic acid:i-propanol:water, 3:2:5; three times with water (10 ml);

194

The use of glucosinolate hydrolysis products as herbicides Chapter 5

and then eluted with 12.5 ml 0.5 M potassium sulphate (K2SO4) in 5% isopropyl alcohol, allowing the eluate to drop into 12.5 ml ethanol. The eluate was filtered in a No. 3 sintered glass funnel to remove the potassium sulphate, agitated and cooled for 10 min at +4 °C. It was centrifuged, the supernatant collected and then evaporated to nearly dry under reduced pressure (13iichi Rotavapor RE120, 35 °C). After drying in vacuo over phosphorus pentoxide a sticky, glassy solid was obtained which did not crystallise.

5.3.1.3 Results

18.2 g seed meal produced 504 mg of residue, and further drying reduced this to 260 mg of solid. This equates to 14.3 mg g-1 (32.5 pmol 4'1) seed meal, providing a yield of 1.4%, greater than the 265 nmol g-1 obtained by Vaughn et a/ (2006). Although Vaughn also extracted the GSL in methanol, he agitated the seedmeal in boiling 70% methanol for 15 min before filtering, a 16-fold reduction in time compared with the method used in this experiment.

5.3.2 Glucosinolate extraction from L. a/ba plant tissue

5.3.2.1 Introduction

GSLs were extracted from different plant tissues harvested from L. a/ba plants (grown hydroponically) using the method described by Heaney (1986), and then analysed and identified.

5.3.2.2 Method

Plant material preparation Plant material of L. a/ba was removed from plants grown in the hydroponics system, different tissues separated, weighed (Table 5-3) and freeze dried (Chapter 2).

Extraction of GSLs from plant tissue Freeze dried plant material was ground until very fine using a coffee grinder. Quantities of stems, leaves, buds and rhizoids (3 x 50 mg) were weighed and placed into Eppendorfs, with an anti-bumping granule in each and a hole made in each lid to release vapour. A quantity of the GSL sinalbin (p-hydroxybenzyl GSL) was used as an

195 The use of glucosinolate hydrolysis products as herbicides Chapter 5 internal standard (2 mg m1-1). GSLs were extracted using 80% (v/v) methanol (1 ml) at 85 °C for 10 min. 20 pl sinalbin standard was then immediately added to each Eppendorf, and the lid closed.

Table 5-3. Fresh and dry weights of L. a/ba tissue.

Plant tissue Fresh weight (g) Stem 16.4 Buds 2.5 Roots 7.6 Leaves 12.2

Eppendorfs were centrifuged for 5 min (13,000 rpm, Eppendorf centrifuge 54141) and the supernatant transferred to clean Eppendorfs. The extraction was repeated with 80% (v/v) methanol (1 ml) at 85 °C (5 min) and centrifuged (5 min). The supernatant was collected and combined with the first extraction. The methanol was evaporated to almost dry using the Techne Dri-Block DB-3A sample concentrator at 60 °C and samples were made up to 1 ml with water.

Removal of proteins Proteins contaminating GSL samples were precipitated with 30 pM Pb(0Ac)2 and 30 pM Ba(0Ac)2 per sample, mixed, left to stand (5 min), centrifuged (5 min, 13,000 rpm) and the supernatant removed.

Preparation of ion exchange columns Ion exchange columns (DEAE Sephadex A25) were prepared, for the plant tissue extracts and external standards, using Sephadex DEAE A-25 (2-(Diethylamino)ethyl- Sephadex) gel, formiate form (Supplier: Sigma-Aldrich), supplied in powder form and stored as a stock solution in 20% ethanol. The stock was shaken well, approx. 15 ml removed, washed twice to remove the ethanol by shaking it in a centrifuge tube with plenty of water, leaving it to settle and the upper, water layer removed. An equal quantity of water was added to the gel (50/50), mixed well, then pipetted (1.2 ml) into each graduated ion exchange column, the water draining away to leave 0.5-0.6 ml gel in each. The gel in each was washed twice with 1 ml water, leaving a smooth, level

196 The use of glucosinolate hydrolysis products as herbicides Chapter 5

surface to the gel. One column was prepared for each sample plus two for external standards.

Load samples onto columns The supernatant was pipetted onto the gel columns; the two external standard columns were loaded by pipetting sinalbin standard (2 mg m1-1, 20 pL), mixed with water (200 pL) for ease of even application, onto each of the two columns.

Each column was washed twice with water (1 ml) and twice with 0.02 M acetate buffer (pH 5.0, 0.5 ml). Sulphatase (75 pl) was added to each column to desulphate the GSLs bound to the column, left overnight, eluted 3 times with water (0.5 ml), collecting the eluate which was transferred to 2 ml Eppendorfs, frozen to -20 °C and freeze dried. Samples were resuspended in water (200 pl) prior to HPLC analysis.

Preparation of glucolimnanthin standard Glucolimnanthin previously extracted from L. a/ba seeds (Section 5.3.1) was used as a standard. Glucolimnanthin (2.38 mg) was dissolved in water (1.5 ml) to give a concentration of 1.59 alga-1. Two ion exchange columns were prepared as before, omitting the barium acetate step as the sample was already protein-free. A sample (0.75 ml, containing 1.19 mg glucolimnanthin) was loaded onto each column, washed twice with water (1 ml), twice with 0.02 M acetate buffer (pH 5.0, 0.5 ml), sulphatase (75 pL) added and left overnight. The next day the sample was eluted 3 times with water (0.5 ml), collecting the eluate in vials for HPLC analysis.

HPLC analysis Samples were analysed for GSL content using HPLC equipment previously described (Chapter 2). Samples were injected (20 pl) and analysed using a gradient solvent system (Table 5-4), with a 42 min run time at 35 °C with 1.5 ml mint flow rate. GSLs were detected using the UV detector set at A 230 nm.

197

The use of glucosinolate hydrolysis products as herbicides Chapter 5

Table 5-4 Solvent gradient conditions used for HPLC analysis. Solvent A = water, Solvent B = 20% acetonitrile Time Flow % A % B Initial 1.50 99 1 1.0 1.50 99 1 21.0 1.50 1 99 34.0 1.50 1 99 36.0 1.50 99 1 42.0 0.00 99 1

5.3.2.3 Results

Recovery rates (Table 5-5), comparing peak sizes of the internal and external sinalbin

standards were above 80% for all tissues except the roots (average 41%), possibly due to losses during the methanol boiling stage.

Glucolimnanthin was the only GSL detected throughout the plant; most plant species have a more complex GSL profile. The GSL peak was confirmed as glucolimnanthin, with retention times averaging at 17.08 min, compared to 17.25 for the glucolimnanthin standard, which was then used to calibrate the GSL peak sizes.

Table 5-5. Glucolimnanthin content of L. a/ba tissue. All values given are an average of three samples. Retention time (Re) of glucolimnanthin standard was 17.25 min. Tissue type Glucolimnanthin content Recovery Rt (nmol me dry weight) % (min) ± SD Stem 86.9 ±1.27 100 17.01 Root 40.9 ±1.04 39 17.14 Leaf 78.8 ±1.15 99 17.06 Bud 79.5 ±2.85 82 17.09

198

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.3 Extraction and identification of glucolimnanthin hydrolysis products from seed meal using a GSL hydrolysis time course assay, preliminary experiment

5.3.3.1 Introduction

A small scale preliminary experiment to extract ITCs from seed meal was carried out, using a method adapted from Vaughn and Berhow (2005). An appropriate time scale for the extraction was established and the identity of the compounds extracted was confirmed.

5.3.3.2 Method

50 mg samples of ground, defatted L. a/ba seed meal was placed into Eppendorfs with buffer (250 pl Tris pH 7.0, 0.1 M) (Section 5.2.2), mixed thoroughly and incubated at 35 °C. The Eppendorfs were removed at predetermined time intervals, dichloromethane (DCM) (1 ml) added, mixed thoroughly for 30 seconds to form an emulsion, and centrifuged (Eppendorf 54141 centrifuge, 13,000 rpm) for 3 min. The DCM was removed with a syringe, transferred to a clean Eppendorf containing a small amount of anhydrous magnesium sulphate (to remove water), mixed and centrifuged again (2 min, 13,000 rpm). Supernatant (500 pL) was removed for analysis. Samples were removed from the water bath at hourly intervals: 0, 1, 2, 3, 4 and 5 hrs. As this was a preliminary experiment samples were not replicated. Samples were analysed by GC-MS using a 5-1 split, and repeated using a 50-1 split due to high abundance of the primary compound detected (Chapter 2.5.2).

5.3.3.3 Results

A peak, detected by GC-MS analysis (Chapter 2) (Figure 5-7a), which appeared in the 0 hr sample, was larger in the 1 hr sample and subsequently reduced, producing a degradation curve (Figure 5-9) suggesting the molecule was unstable at pH 7.0. The compound extracted was identified by mass spectography as 3-methoxybenzyllTC (C9H9NOS) (limnanthin)(Figure 5-8) by comparison with mass spectral data presented by Vaughn and Berhow (2005).

199

The use of glucosinolate hydrolysis products as herbicides Chapter 5

3.5e+6

3.0e+6 -

2.5e+6 -

t 2.0e+6 - en

curr 1.5e+6 - l ion ta

To 1.0e+6 -

5.0e+5

0.0

0 10 20 30 40 50 60

Time (min)

5e+5 (b)

121 4e+5 -

t n 3e+5 - rre cu n io l

ta 2e+5 - To 179

91 le+5 - 78 65 II

0 • "r'''II II P -- 20 40 60 80 100 120 140 160 180 200 220 240

m/z

Figure 5-7. (a) Chromatogram and (b) mass spectrum (EI) of 3-methoxybenzyllTC. Extracted from 50 mg L. a/ba seed meal in 1 ml DCM at pH 7.0.

,N ,C' 'C O S' CH3

Figure 5-8. 3-MethoxybenzyllTC

200

The use of glucosinolate hydrolysis products as herbicides Chapter 5

3e+6

3e+6 -

2e+6 -

t 2e+6 - ren r cu

n le+6 - io l ta

To 5e+5 -

0 1 2 3 4 5 6 Time (hours)

Figure 5-9. Preliminary glucolimnanthin degradation curve produced during a time course assay at hourly intervals, pH 7.0, extracted with DCM (1 mL)

5.3.4 Optimisation of glucolimnanthin hydrolysis products extraction

5.3.4.1 Introduction

The preliminary experiment was repeated with shorter time intervals, providing results at more time points over 1.5 hrs, to characterise more precisely the glucolimnanthin degradation curve.

5.3.4.2 Method

The preliminary experiment (Section 5.3.3) was repeated with eight extractions at 20 min intervals, pH 7.0 with two samples at each time point to further characterise glucolimnanthin hydrolysis. Time points were chosen to ensure that complete GSL degradation could be observed.

5.3.4.3 Results

The maximum amount of GSL product was present after 50 min, sharply declining thereafter (Figure 5-10).

201

The use of glucosinolate hydrolysis products as herbicides Chapter 5

30x106

25x106 -

20x106 - t n rre cu

n 15x106 - l io ta To 10x106 -

5x106 -

0 10 30 50 70 90 110 130 150

Time (min)

Figure 5-10. Glucolimnanthin degradation curve produced during a time course assay, samples taken at 20 min intervals, pH 7.0, extracted with DCM (1 mL)

5.3.5 Optimisation of 3-methoxybenzyllTC extraction from seed meal

5.3.5.1 Introduction

Degradation of GSLs produces different products and in different concentrations dependent on pH, for example producing nitriles at low pHs, and ITCs in neutral conditions (Gil and MacLeod, 1980). This experiment was designed to identify an optimum pH for ITC extraction from glucolimnanthin.

5.3.5.2 Method

This time course experiment used the previous method (Section 5.3.3) with six extractions at hourly intervals (0, 1, 2, 3, 4, 5 hrs), with two samples taken at each time point and at pH 4.0, 5.0 6.0, 8.0 and 9.0, 30 samples in total; pHs were adjusted using 0.1 M acetate, 0.2 M phosphate and 0.1 M Tris buffers (Section 5.2.2).

5.3.5.3 Results

Samples were analysed using GC-MS (50-1 split method) (Chapter 2.5.2), producing the degradation curves shown in Figure 5-11. Additional GSL products were detected

202 The use of glucosinolate hydrolysis products as herbicides Chapter 5 at lower pHs, 4.0 and 5.0; less at pH 7.0, 8.0 & 9.0, with a second peak appeared in pH 7.0, 8.0 and 9.0, identified as 3-methoxybenzylamine, (C8H11N0)(Figure 5-13, Figure 5-14, Figure 5-12).

At lower pHs, 4.0, 5.0, 6.0, 9.0 hydrolysis reactions took place immediately prior to degrading; the reaction was slower at pH's 7.0 & 8.0. Previous analysis (Section 5.3.4) suggested maximum ITC was produced after 50 min at pH7.

1.2e+8

1.0e+8 -

8.0e+7 -

t en 6.0e+7 - curr l ion

ta 4.0e+7 - To

2.0e+7 -

0.0 -

0.15 1.15 2.15 3.15 4.15 5.15 Time (hr)

Figure 5-11. Degradation curves for 3-methoxybenzylITC extracted from L. a/ba seed meal at pH 5.0, 6.0, 7.0, 8.0, 9.0, at hourly time intervals. 50 mg seed meal extracted with DCM (1 mL). Samples were analysed by GC-MS, 50-1 split method (Section 2.5.2)

9 CH3

Figure 5-12. 3-Methoxybenzylamine.

203 The useofglucosinolatehydrolysisproductsasherbicides Figure 5-13.Chromatogramof by GC-MS,50-1splitmethod(Section2.5.2) meal atpH8.0for4hrs.50mgseedwasextractedwithDCM(1mL).Analysed product extractedfrom 50-1 split(Section2.5.2).50mgseedmealwas extractedwithDCM(1mL) Figure 5-14.Massspectrogram(IE)of3-methoxybenzylamine, GSLhydrolysis

Total ion current Total ion current 5e+5 4e+5 - 3e+5 - 2e+5 - 0 le+5 - 10000 15000 20000 5000 0 2 20 3-Methoxybenzylamine 30 39 40 I LI L. a/ba 4 I 51 Ilin, i...I 60 65 L. a/ba seed mealatpH8.0after4hrs,analysedbyGC-MS, 6 77 80 GSL hydrolysisproductsextractedfromseed 94 204 I 8 Time (min)

100 3-MethoxybenzylITC 106 mz 120 10 121 1 36 140

160 14 180 16 200 Chapter 5

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.6 Scale up of ITC extraction and purificaton

5.3.6.1 Introduction

To produce larger quantities of pure 3-methoxybenzyllTC for use in bioassays the extraction process was scaled up, bioactive fractions identified and separated, producing purified ITC. 0.1 M acetate buffer (pH 5.0) was used, as this produced the maximum amount of ITC.

5.3.6.2 Method

L. a/ba seed meal (40g) in 0.1 M acetic acid buffer (200 ml, pH 5.0) (Section 5.2.2) was agitated (100 RPM, 37 °C for 1 hr), transferred to a separating funnel (1 L), extracted with DCM (400 ml) (ratio of 2:1 with the buffer) and centrifuged (MSE Hi- spin 21, 10,000 rpm) for 15 min. The seed meal was extracted three times, collecting the lower DCM layer each time. Extracts were bulked, dried over anhydrous magnesium sulphate, mixed on a stirrer (10 min) and filtered under pressure (Hartley funnel). DCM was removed under reduced pressure (13Lichi Rotavapor RE120, 35 °C) and the extract stored (-20°C).

Thin layer chromatography (TLC) techniques were used to identify the bioactive

fraction of the extract (Chapter 2). Using a sheet of Silica Gel 60 F254 TLC aluminium sheet (Merck catalogue no. 1.05554.0001) cut to 3 cm x 9 cm, a drop of ITC extract was dissolved in a small amount of hexane and a small drop placed in the centre of the sheet approximately 1 cm from the bottom and the position marked. The plate was developed (solvent, hexane, 10 ml) until the solvent front almost reached the top of the TLC sheet. It was removed, dried, viewed under ultra violet light and the extract front marked.

As the extract front moved little the TLC was repeated using hexane:ethyl acetate

solvent mixtures with ratios of 60:40, 80:20, 90:10 and 95:5, aiming for an Rf value of around 0.3 (Chapter 2) until the optimal proportions were found.

A Teledyne Isco CombiFlash® RETRIEVE TM purification system with an Isco RediSeptm 40g normal phase silica column, with an optimum flow rate of 40 ml min-1, volume 48

205 The use of glucosinolate hydrolysis products as herbicides Chapter 5 ml was used to purify the extract. Hexane:ethyl acetate (1 litre, 95:5) solvent was prepared and used to prime and equilibrate the CombiFlash purification system using 3 x column volumes of solvent. The extract was dissolved in solvent (4 ml), injected into the purification system and 35 x 15 ml fractions were collected. A TLC plate was spotted with drops from fractions 1-35 and viewed under ultraviolet light. ITC was found in fractions 6-15, with greater concentrations in the darker spots of fractions 9- 12.

Further TLC was carried out using the crude extract produced previously and fractions 6-15 with the 95:5 hexane:ethyl acetate solvent. Fractions 9-12 again contained the greatest amount of ITC, the position confirmed by the Rf value. Fractions 9-12 were bulked, the solvent removed under reduced pressure (13Lichi Rotavapor RE120, 35 °C), diethyl ether (1 ml) added to the flask to dissolve the ITC and then transferred to a vial. The ether was removed with dry argon.

5.3.6.3 Results

The NCS group is polar and interacts with the silica, moving a shorter distance up the

TLC sheet than a less polar molecule would. The ideal Rf value of around 0.3, was produced with a solvent ratio of 95:5 hexane:ethyl acetate (Table 5-6).

Table 5-6. Movement of extract in various solvent ratios during TLC.

Distance to Distance to Hexane EtOAc extract front solvent front Rf (mm) (mm) 100 0 10 64 0.2 95 5 23 65 0.4 90 10 29 65 0.5 80 20 36 60 0.6 60 40 38 48 0.8

780 mg green residue was extracted from 40 g seed meal, which was then purified to obtain 372 mg pure ITC, equating to 9.3 mg IV (57 pmol g4). Previously 32.492 pmol 0 glucolimnanthin was extracted from 18.2 g seed meal.

206

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.7 Preliminary bioassay to investigate the effect of ITCs on liverwort gemma growth

5.3.7.1 Introduction

To test the effect of ITCs (benzyl-, 2-phenylethyl-, 3-methoxybenzyl- and 2-propenyl- on liverwort growth an in vitro bioassay was developed whereby ITCs of varying concentrations were incorporated into M51C nutrient media (Chapter 2), from which they were taken up by individual gemmae laid on the surface. They were incubated for fourteen days and their radial growth measured using image analysis techniques. The ITCs needed to be dissolved in solvent prior to incorporation into the media, however earlier work indicated some sensitivity of gemmae to alcohol; therefore a preliminary experiment investigated gemma tolerance of ethanol. The methods used were developed by Dornbos (1990) who found that agar bioassays required smaller quantities of the bioactive compound than filter paper bioassays, partly due to the size of the cells used compared to Petri dishes, providing a more sensitive system.

5.3.7.3 Methods

Previous liverwort growth experiments (Table 5-7) indicated that liverwort gemmalings grow to approximately 5-6 mm in 3-4 weeks, therefore 25 ml well plates were an appropriate size to use.

Table 5-7. Gemmaling diameter (mm) when grown at 25 °C, 400mp 01 m-2 s-1.

Week Wk 1 Wk 2 Wk 3 Wk 4 Wk 5 Wk 6 Gemmaling 1.0 2.9 5.1 6.9 7.3 8.3 diameter (mm)

M51C media was autoclaved in 5 x 100 ml quantities, cooled slightly, ethanol added in quantities as shown in Table 5-8 and mixed by pouring back and forth into a sterilised jar four times. 2 ml aliquots were pipetted into each well, 5 wells per treatment, replicated four times. One gemma was placed on the media in the centre of each well and incubated in growth rooms (20 °C, 8 hr day). Plates were digitally photographed after 14 days and gemma areas calculated using Image) software.

207 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Table 5-8. Treatments used in preliminary experiment.

Treatments 0/0 Ethanol Control 0.1% 1% 2.5% 5% Volume ethanol (ml) /100 ml media 0 0.1 1 2.5 5

5.3.7.3 Results

The ethanol reduced gemma growth in all treatments, markedly more so in the 2.5% and 5% treatments; 0.1% ethanol was used in subsequent bioassays (Table 5-9,

Figure 5-15).

0.1 1 2.5 5 Ethanol content of media (%)

Figure 5-15. Preliminary investigation into the effect of ethanol on gemmaling growth (radial expansion). Gemmaling areas shown are a percentage of the control.

Table 5-9. The effect of ethanol on liverwort gemma growth after 14 days.

Treatment 0/0 Ethanol Nil 0.1% 1% 2.5% 50/0 Gemmaling area after 14 days 4.6 2.1 1.8 0.6 0.6 Gemmaling area as a percentage of control 45.5 39.0 13.8 13.5

208

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.8 Bioassay investigating the effect of ITCs on liverwort gemma growth

5.3.8.1 Introduction

The bioassay method (Section 5.3.7) was used to test the effect of different concentrations of ITCs (benzyl-, 2-phenylethyl-, 3-methoxybenzyl- and 2-propenyl-) on liverwort gemmaling growth (radial expansion). Each ITC was dissolved in 0.1% ethanol prior to incorporation into autoclaved M51C media.

5.3.8.2 Method

ITC (100 mg) was dissolved in ethanol (1 ml), and then four 10-fold serial dilutions each of 100 and 50 mg m1-1 ITC, and zero ITC (alcohol only) were prepared. The control contained media only, no ITC or alcohol. Using the previous method (Section 5.3.7), 12 aliquots (50 ml) of media were autoclaved and maintained at 50 °C in a water bath. ITC (50 pl) (to achieve 0.1% ethanol) was added to each, mixed and pipetted (2 ml) into each well. 5 wells were used for each treatment and treatments were replicated 4 times.

5.3.8.3 Results

Whilst low doses of 2-propenyllTC (0 to 0.5 pg m1-1), benzyllTC (0 to 0.05 pg m1-1) and 2-phenylethyllTC (0.005 pg m1-1) ITCs promoted gemmaling growth compared to the control, all ITCs except 2-propenyllTC reduced growth almost to zero at concentrations of 1.0 pg m1-1 and above (Figure 5-16). These results were reflected in the estimated ED50 figures, with greater doses of 2-propenyllTC required to limit growth of 50% of the gemma population to an area of 1.5 mm2 (Table 5-10).

Table 5-10. Estimated ED5os and standard errors of ITCs applied to liverwort gemmae, obtained using probit analysis. ED50 is the effective dose where 50% of the gemma population has an area <1.5 mm2. ED50 (mol (±SE) ITC m1-1) Values and std errors x 10-3 3-Methoxybenzyl 2.6 (±0.098) 2-Phenylethyl 3.1 (±0.106) Benzyl 3.9 (±0.084) 2-Propenyl 30.8 (±0.129)

209

The use of glucosinolate hydrolysis products as herbicides Chapter 5

30 30 (a) 3-MethoxybenzylITC (b) 2-PhenylethyllTC

25 25 -

20 - 20 - E E 15 -

10 -

5 - 5 -

0 - 0 - • • •

0 G d1 el ep Isothiocyanate concentration (mol Isothiocyanate concentration (mol mlr )

30 30 (c) BenzylITC (d)2-PropenylITC

25 - 25 -

20 -

E 15 15 -

10 -

5 5 -

.;• o b cv 4P* 41 45 45 45 4?' 45 Isothiocyanate concentration (mol Isothiocyanate concentration (mol

Figure 5-16. Dose-response curves of ITC concentrations and gemmaling growth (radial expansion). Control: treatment with no isthiocyanate or alcohol

5.3.8 Herbicide bioassays

5.3.8.1 Introduction

Further in vitro bioassays were carried out to investigate the effect of the herbicides lenacil and metazachlor (Figure 5-4) (Supplier: Sigma-Aldrich) on the radial growth of liverwort gemmae as a general comparison with ITCs.

5.3.8.2 Method

Lenacil is not soluble in ethanol or water, therefore a stock solution was prepared in dimethyl sulphoxide (DMSO), dissolving 100 mg lenacil in 18.26 ml DMSO; the solubility of lenacil in DMSO is 6 mg m1-1, DMSO density is 1.1 g m1-1. Treatments used were 5 x 10-fold serial dilutions each of 6 and 3 mg m1-1 lenacil, zero lenacil (DMSO only), and a control containing media only, no lenacil or DMSO.

210 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Metazachlor (100 mg) was dissolved in ethanol (1 ml), then 4 x 10-fold serial dilutions each of 100 and 50 mg m1-1 metazachlor, zero metazachlor (alcohol only), and a control containing no metazachlor or alcohol were prepared. Bioassays were prepared (Section 5.3.7), using 1 pl herbicide m1-1 media.

5.3.8.3 Results

Metazachlor produced a higher estimated ED50, than either lenacil or the ITCs, however this does not fully describe its effect on gemma growth. Doses as low as 0.05 pg m1-1 reduced the average gemma area to 2.78 mm2, and gradual growth reduction continued, reaching an average of 1.13 mm2 with a dose of 100 pg m1-1 (Figure 5-17). Similarly, the estimated ED50 for lenacil was slightly higher than the majority of ITCs, although it produced a sharp growth reduction response at lower doses, between 0.06 and 0.6 pg m1-1 compared with 0.1 and 1.0 pg m1-1. This suggests that these herbicides may produce a great enough effect on liverwort gemma growth to reduce liverwort infestation to an acceptable level at lower doses than ITCs used in these bioassays (Table 5-11). Selecting an ED50 standard that reflects the economic injury level would be more realistic in any future bioassays.

Table 5-11. Estimated EDsos and standard errors of herbicides applied to liverwort gemmae, obtained using probit analysis. ElDso is the effective dose where 50% of the gemma population has an area <1.5 mm2. ED50 (mol m1-1) (±SE) Herbicide Values and std errors x 10-9 4.2 (±0.162) Lenacil 16.0 (±0.124) Metazachlor

211 The use of glucosinolate hydrolysis products as herbicides Chapter 5

35 (a) Lenacil 30 -

25 -

5 -

0 -

ON.N.00<)N •0)00 • • "1(3 • bfr) vsf' I; (de nO (0° -5e. coo 50 (0° of\ N. N: • N.• S.• N.• N.• 1 Lenacil concentration (mol m1 )

30 (b) Metazachlor

25 -

20 -

5 -

0 -

0 0 93 1 'A 10 0 <0 <1 00 (0° 0° 0° 0° 0°

Metazachlor concentration (mol m11)

Figure 5-17. Dose-response curves of herbicides a) lenacil and b) metazachlor and liverwort gemma growth (radial expansion). Control: treatment with no solvent (DMSO or alcohol) or herbicide.

212

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.9 Cress ITC bioassays using Petri dishes

5.3.9.1 Introduction

A further bioassay was developed to investigate the phytotoxicity of ITCs (benzyl-, 2- propenyl-, 2-phenylethyl- and 3-methoxybenzyl-) on a higher plant species, cress (Lepidium sativum), to confirm whether there would be a comparable effect on germination and growth. A method was adapted from that used by Kasama (2003), with seeds incubated on filter paper in Petri dishes and fed with the bioactive compound dissolved in distilled water, recording seed germination and radicle elongation.

5.3.9.2 Method

ITC (100 mg) was dissolved in ethanol (1 ml), then 4 x 10-fold serial dilutions each of 100 and 50 mg m1-1. ITC, zero ITC (alcohol only), and a control containing no ITC or alcohol were prepared. Each of these was then made up into a 10 ml solution prepared using 10 pl ITC in sterile distilled water; these solutions were fed to the cress seeds. This was repeated for each of the ITCs used.

10 seeds were placed on filter paper (Whatman 1, 90 mm) in each Petri dish, fed with ITC solution (2 ml) on the first day and sterile distilled water thereafter for 14 days as required maintaining moist, humid conditions. Petri dishes were arranged in a completely randomised design, with two replications, on laboratory benches out of direct sunlight in ambient conditions. Radicle lengths were measured using digital image analysis (Chapter 2) techniques and the number of seeds that germinated was recorded.

5.3.9.3 Results

Little effect of these ITCs on cress seed was observed, with average radicle lengths and germination rates similar at all concentrations of ITC used (Figure 5-18, Figure 5-19). This contrasts with the results observed when the same ITCs were applied to liverwort gemmae, when thallus radial growth was reduced with greater concentrations of ITC. This suggests these ITCs could potentially be applied as a liverwort herbicide over higher plants without affecting their growth.

213 The use of glucosinolate hydrolysis products as herbicides Chapter 5

' b e° goo •,;:` ^)..s* Isothiocyanate concentration (ma Mr') Isothiocyanate concentratbn (mol

30 30 (c) BenzyIITC (d)2-PropenylITC

25 - 25 -

20 E 20 - 5 15 - 10

10 - 5 -

0 5 1f ° Ae1 Ai 04 e4 4,j; 4 b' b. 'S. b 1. 04' Lsothiocyanate concentration (mol Isothiocyanate concentration (mol ml')

Figure 5-18. Dose-response curves of ITCs and cress (Lepidium sativum) radicle elongation. Control: treatment with no alcohol or ITC.

214 The useofglucosinolatehydrolysisproductsasherbicides method forcollecting and identifyingITCsfromthefourtestplants (Table5-1). 5.3.10.1 5.3.10 to complexextractionprotocols;simplelab-based bioassayscouldthenbedesignedto Figure 5-19.Dose-responsecurvesofITCsandcressseed designed tocollecthydrophobiccompounds.It wasconsideredthatfromthisanon- germination (%).Control:treatmentwithnoalcoholorITC. measure theireffectonliverwortthallusgrowth. resin. A preliminaryexperiment usingITCstandardswasdevelopedadapting Yamane's Yamane (1992)successfullytrappedai-methylsulfonylalkylisothiocyanatefromroot Samples wereelutedfrom thecolumnandanalysedbyGC-MS(Section 2.5.2). destructive methodforcontinuousITCcollection couldbedevelopedwithoutresorting exudates usingAmberliteXAD-4polymericresin adsorbent(Supplier:Sigma-Aldrich), 50 50 110 100 70 60 90 80

o 's

° Preliminary investigationintoITCcollectionusingAmberlite Introduction Isothiocyanate concentration(molml') Isothlocyanate concentration(mol e l b t •;:r i

a J.; 3 d ° t' 43. ." I 215 50 Germlnatbn (%) 100 110 110 100 50 70 - 70 - SO 80 90 - 60 - 60 -

(P c ' c; 4 Isothkryanate concentration(mol Isothlocyanate concentration(toolml

..5":" 4 e .

5; . "

(Lepidium sativum) 4 c; 4 4 <;; ml') (b) 2-PhenylethylITC .1$ (d) 2-PropenyllTC 4 7 Chapter 5 4 e's ;

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.10.2 Method

A mixture was prepared containing 10 mg of each ITC (2-propenyl-, benzyl-, 3- methoxybenzyl- and 2-phenylethyl-) in distilled water (2.5 L) shaken well to dissolve and left to settle.

A trap column was prepared using a Pharmacia XK26 column filled with Amberlite XAD- 4 resin polymeric adsorbent (15 ml) (Sigma-Aldrich). The resin was mixed with enough distilled water to produce a slurry, put into the column and washed with methanol (200 ml), then 200 ml distilled water using an Eyela micro tube pump MP-3 (manufactured by Tokyo Rikakikai Co. Ltd) to circulate the solutions.

The ITC solution was pumped through the resin, collecting approximately 1 litre. The resin was removed from the trap column into a funnel over a side-arm conical flask and vacuum filtered. The resin was washed with distilled water (200 ml), and the ITC's eluted with acetone (100 ml), collecting the filtrates separately. Water was removed from the acetone filtrate with sodium sulphate, stirred (20 min) then filtered. It was then concentrated to 5 ml (Techne Dri-Block DB-3A sample concentrator) and subjected to GC-MS, 3-1 method (Chapter 2.5.2) analysis.

5.3.10.3 Results

All four ITCs were detected in the samples (Figure 5-20), their identity confirmed by comparison of their mass spectra with spectral data published by Spencer and Daxenbichler (1980) and Vaughn and Berhow (2005), confirming the methods were appropriate for the collection and elution of these ITCs from solution.

216

The useofglucosinolatehydrolysisproductsasherbicides Figure 5-20.(a)Chromatogramofthefour standard ITCsusedandtheirmass phenylethyllTC spectra (EI):(b)benzyllTC (c)2-propenyllTC(d)3-methoxybenzyllTC (e)2- a 50000 10000 - current 40000 30000 20000 - Totalion 6e+4 - 8e+4 - 4e+4 1e+5 2e+4 0 2e+6 1e+6 - 2e+6 2e+6 - 6e+5 - le+6 - 8e+5 - 2e+5 4e+5 - 0 le+6 - 0 20

0 40 6080100120140

60 1 65 1 65

78 11

10 91 1.1 91 100 m/z

mlz 121 120 140 16018020( 20 149 160 180200 Time (min) 179 (d) 30 (b)

P, 20000- 4 a 217

50000 40000 30000 5e+5 3e+5 10000 - 6e+5 4e+5 2e+5 0 1e+5 0 20 20 406080 40 40 608010012014016018020( 41

58 ,11

72 77

50 9

I )1 99 00 120140160180 105 miz DIA 60 163 I. (C) Chapter 5 (e) 200 The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.11 Collection of root exudates from the glasshouse hydroponics system

Introduction The method developed (Section 5.3.10) proved successful for ITC trapping and elution and was adapted for the collection of ITCs exuded by plant roots direct from the hydroponic growing system (Section 5.2.1) onto the trap column. D. tenuifolia was used for this section of work as the L. a/ba plants had deteriorated and been replaced with fresh plants which had young, less developed root systems at this time.

5.3.11.2 Method An in-line filter was constructed using a glass tube filled with glass wool at either end and with acid washed sand (BDH Lab supplies, 0.1-0.3 mm grain size, product no. 330945E) between (Figure 5-21) to prevent plant debris from blocking the plastic tubing.

Glass tube Sand 1

Glass wool

Figure 5-21. In-line filter construction

Equipment used in the preliminary experiment was prepared (Figure 5-22) with the in- line filter attached, and with the open end positioned in the irrigation stream within the root zone of D. tenuifolia. Nutrient solution containing root exudates was pumped through the trap column (27 hrs). ITCs were eluted from the Amberlite resin, concentrated to 5 ml (Techne Dri-Block DB-3A sample concentrator) and subjected to GC-MS analysis, 3-1 method (Chapter 2.5.2).

Samples were purified by diluting with distilled water (15 ml), extracting twice with an equal quantity of DCM and the lower layer removed. Extracts were bulked, dried with anhydrous magnesium sulphate, filtered, evaporated to a small volume under reduced

218 The use of glucosinolate hydrolysis products as herbicides Chapter 5 pressure (13Lichi Rotavapor RE120, 35 °C), concentrated to approx. 200 pL, and subjected to GC-MS analysis using the 3-1 split method (Chapter 2.5.2).

Pharmacia

Peristaltic pump

In-line filter —► Flow direction

Figure 5-22. ITC extraction equipment

To quantify the amount of ITCs in the sample a calibration curve was produced. The sample was dried down to remove the remaining DCM and re-dissolved in DCM (300 pL). A standard was prepared by dissolving quantities of benzylITC, 3- methoxybenzyITTC and 2-phenylethyllTC (Table 5-12) separately in acetone (1 ml). 100 pL of each was then dissolved in acetone (700 pL) and mixed with acetone to produce: 1:250, 1:500, 1:750, 1:1,000 and 1:2,000, 1:4,000, 1:6,000, 1:8,000 and 1:10,000 dilutions. The standards were analysed by GC-MS, 3-1 split method (Chapter 2.5.2). Graphs were constructed using peak areas, subjected to linear regression and used to quantify the ITCs present.

Table 5-12. ITC weights used for standards

ITC Weight (Ng) Benzyl 16.0 2-Phenylethyl 12.5 3-Methoxybenzyl 13.3

219

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.11.3 Results

The data produced from the GC-MS analysis of the ITC standard solutions was used to construct a linear regression, showing the relationship between peak area and ITC quantity, providing values for a and 3. The fitted equation was y= ax + p, where x = quantity of ITC, y= peak area, a = intercept and p = slope; calculated weights of the ITCs are shown (Table 5-13).

Table 5-13. Quantities of ITCs extracted from root exudates o D, tenuifolia. ITC Weight (pg) 3-Methoxybenzyl 220 Benzyl 5 2-Phenylethyl 94

3-Methmbenzy ITC 28.817 2e+6 -

2e+6 -

n 2-PhenylethyllTC 25.58 urre c

n le+6 - l io ta To 5e+5 - BenzyIITC 23.01

0 5 10 15 20 25 30 35 Time (min)

Figure 5-23. Diplotaxis tenuifolia root exudates chromatograph, analysed by GC-MS, 3-1 split method (Chapter 2.5.2).

220 The use of glucosinolate hydrolysis products as herbicides Chapter 5

GC-MS analysis initially contained numerous peaks probably produced by plasticisers from the plastic equipment used in the hydroponics system, and no ITCs were identified. Following purification benzyllTC, 2-phenylethyllTC and 3-methoxybenzylITC were found in the samples (Figure 5-23) and their identities confirmed by comparison with spectral data published by Spencer and Daxenbichler (1980) and Vaughn and Berhow (2005). However, as glucolimnanthin had not previously been identified in D. tenuifolia roots (Table 5-1) its hydrolysis product 3-methoxybenzyllTC was not expected to be found in this sample and was a suspected contaminant.

5.3.12 Collection of root exudates from plants grown in glass containers

5.3.12.1 Introduction

As plasticisers were detected in the root exudate samples collected direct from the hydroponics system (Section 5.3.11), mature plants of D. tenuifolia were removed to glass vessels and the collection process repeated, followed by GC-MS analysis.

5.3.12.2 Method

3 plants of D. tenuifolia were suspended in fresh water in a glass chromatography tank. Collection equipment set up (Figure 5-21 and Figure 5-22) with the filter in the plant root zone and root exudates collected over 24 hrs. The roots of the same plants were then damaged by squeezing between forceps and root exudates eluted and concentrated as previously.

Samples were diluted with distilled water (15 ml), and extracted twice with an equal quantity of dichloromethane (DCM) by shaking, allowed to settle and the lower (DCM) layer removed. Extracts were bulked, anhydrous magnesium sulphate added to remove any water, filtered and DCM removed under reduced pressure (Buchi Rotavapor RE120, 35 °C), concentrated to approx. 200 pL and subjected to GC-MS analysis, 3-1 method (Chapter 2.5.2).

5.3.12.3 Results

None of the 1TCs found in the earlier samples were detected in these samples, suggesting the compounds were contaminants as suspected.

221

The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.3.13 Glucosinolate extraction from 0 tenuifolia root tissue

5.3.13.1 Introduction

The previous experiment seemed to confirm the ITCs previously detected in D. tenuifolia root exudates were contaminants. To verify this GSLs were extracted from root tissues and identified on the premise that without GSLs in the root tissue, no ITCs could have been produced in the root exudates.

5.3.13.2 Method

Roots from D. tenuifolia used in the root exudate extraction were washed with water, freeze dried, ground in a coffee grinder and the GSLs were then extracted from a 50 mg sample as described in Section 5.3.2 along with 2-propenyl GSL and benzyl GSL standards. The sample and standards were analysed by HPLC (Chapter 5.3.2), however the peaks produced by the sample did not correspond to those produced by the standards. The sample was then subjected to LC-APCI MS/MS and UV (230 nm) analysis for identification.

5.3.13.3 Results

Chromatographs and spectra (Figure 5-24) were analysed and compounds identified by comparison with known standards or by computer comparison with the Wiley 275 mass spectra library. Of these compounds only 4-methylsulphinylbutyl and p-hydroxybenzyl have previously been identified in D. tenuifolia roots by others (Table 5-1). The GSLs relating to the ITCs identified in root exudates (Section 5.3.1.1.3) were not found. This could suggest that either they were contaminants or as the plants' condition had deteriorated by this time, the GSLs were no longer present.

222

The use of glucosinolate hydrolysis products as herbicides Chapter 5

1e+9 600 (b) 9e+8 (u) 500 80+8

7e+8 400 3 6e+8 300 5e+8 4 200 4e+8

3e+8 - 100 (I) (I) 2e+8

10+13

100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 El 9 10 11 12 13

Retention tine (min) Retention time (min)

8e+6 (I)

6e+6

4e+6

2e+6 -

0 100 200 300 400 500 100 200 300 400 500

nu mz

1e+8 3.5e+8 (iv)

3.0e+8 - 8e+7 -

2.5e+8 -

6e+7 - 2.0e+8

40+7 1.5e+8 -

1.00+8 - 2e+7 -

5.0e+7 -

0 - 0.0 -

0 100 200 300 400 500 100 200 300 400 500

mz mz

Figure 5-24. Analysis of GSL content of Diplotaxls tenuifolia root tissue: chromatographs analysed by (a) total ion current (90-400 atmospheric mass units) and (b) detected by UV absorption at 230 nm wavelength. Chromatogram peaks and mass spectra (EI) refer to GSLs: (i) 4-methylsulphinylbutyl (ii) p-hydroxybenzyl (iii) 4-methylthiobutyl (iv) 4-methoxy-3-indolemethyl.

5.4 Discussion The GSL work in this study was undertaken primarily to identify whether selected GSL hydrolysis products have a herbicidal effect on liverwort. Three main approaches were taken: bioassays investigating the effect of four ITCs, including that found in L. a/ba, on liverwort gemma growth and higher plants; an investigation into the GSLs present in L. a/ba seed meal, and the direct harvesting of ITCs from the root exudates of plants grown hydroponically.

223 The use of glucosinolate hydrolysis products as herbicides Chapter 5

5.4.1 Phytotoxicity

The results of the bioassays using both ITCs and herbicides against liverwort gemmae did show a phytotoxic effect, with estimated ED50 values showing that 3- methoxybenzyl ITC (Limnanthin) was the most potent of those tested and it was effective at lower concentrations. ED50s for metazachlor and lenacil were higher than for the ITCs, suggesting that more herbicide than ITC would be required to be effective against liverwort gemmae. The cress seeds were tolerant of the ITCs at the concentrations used; compounds toxic to target weeds without harming crop plants could suggest a potential use as a selective herbicide that could be applied over crop plants.

ED50 values obtained suggest that of those tested 2-propenyllTC was the least toxic towards liverwort and is also the most volatile. According to Sarwar et at (1998) application method should be geared to ITC characteristics, applying volatile ITCs in a headspace and water soluble ITCs in agar media to gain greater inhibitory effects, and this could provide an explanation for the marked difference in toxicity of 2- propenyllTC. Losses could also have occurred during incorporation with warm media. The aliphatic ITCs tend to be more volatile than aromatic as volatility decreases with increased molecular weight (Norsworthy and Meehan, 2005a).

Norsworthy and Meehan (2005b) found aromatic ITCs more phytotoxic towards annual dicotyledonous weeds and suggested this is due to the greater stability of their structures provided by their conjugated ring, and therefore greater contact with the weed before degrading. It is the variable side chain found on GSLs that is thought to determine the biological activity of ITCs (Brown and Morra, 1999). Vaughn et (3/(2006) found that aliphatic ITCs with short chain R-groups, such as 2-propenyllTC, were amongst the most inhibitory; according to Sarwar et a/ (1998) the shorter the chain length the greater the toxicity of ITCs, when used against soil-borne fungal pathogens, and suggests such a trend occurs within each ITC class.

This hypothesis was not borne out by this research, where 2-phenylethyllTC (ED50 = 0.513 ug ml-') was marginally more toxic than benzyllTC (ED50 = 0.583 ug however the values obtained were similar; had there been greater disparity in their structure a more definitive difference may have been observed.

224 The use of glucosinolate hydrolysis products as herbicides Chapter 5

Low ITC concentrations can have a stimulatory effect on plants, as found in these bioassays. Norsworthy and Meehan (2005a) found that phenyl and 2-phenylethyl]TC had a stimulatory effect on sicklepod emergence at lower concentrations, becoming toxic at higher concentrations, and this effect was observed with benzyllTC and 2- propenylITC, and also the two herbicides metazachlor and lenacil.

5.4.2 Limnanthes a/ba

The GSL profile of L. alba was established, and identified a single GSL, glucolimnanthin, present throughout the plant. In other plants the profile can be widely variable, with most species producing a range of GSLs in each tissue, e.g. nine have been identified in the roots of B. juncea (Table 5-1).

The GSL glucolimnanthin was extracted from L. a/ba seed meal, with yields of 32.492 pmol g-1 seed meal (1.43%), and the quantity of limnanthin (3-methoxybenzylITC) extracted was 56.992 pmol g-1 seed meal (0.93%). These values compare favourably to quantities of 265 nm glucolimnanthin $9-1 L. a/ba seed meal, and 65 nm limnanthin extracted from a mixture of 0.5 g seed meal and soil by Vaughn et a/ (2006), the variance possibly due to differences in the extraction method. GSL levels obtained from the remainder of the plant ranged between 79 and 87 nmoles mg-1 except for the roots, where the recovery dropped to 39%, compared with over 82% for other tissues.

Although ITCs are generally produced in pH 7.0 conditions (Chew, 1988), in the extraction optimisation experiment more limnanthin was produced in acidic conditions (pH 4.0 and 5.0) and least in neutral to alkaline (pH 7.0, 8.0 and 9.0) conditions; Vaughn and Berhow (2005) found a 0.1 M HCL crude extract from L. alba produced the greatest amount of 3-methoxybenzyllTC, although it was also present in extractions in pH 7 and 10 buffers.

5.4.3 Hydroponics

The method for collecting and analysing ITCs direct from a solution was successful in principle, however when used to collect ITCs direct from the root exudates of D. tenuifolia growing in the hydroponics system the samples were contaminated. Further GSL analysis of the root tissue did not identify any of the expected ITCs. Whilst four

225 The use of glucosinolate hydrolysis products as herbicides Chapter 5 plant species were grown hydroponically for root exudate collection only D. tenuifolia had great enough root mass for ITC exudate collection as the others had succumbed to powdery mildew infection (B. juncea, Sorlentale, and L. a/ba), outgrown the equipment (B. juncea) or their general condition deteriorated. Replacement plants were not sufficiently developed to be used.

The intention of this section of work was to develop a system whereby ITCs could be harvested simply from root exudates, thereby providing a continual supply and avoiding destructive methods requiring complex ITC extraction techniques. Time constraints did not allow for further development of this work within this project. Improvements could easily be made to the design of the hydroponics system to suit the root mass produced by plants; the narrow width of the plastic guttering used resulted in the roots blocking the water flow on occasions, and free flow is essential to the supply of optimum levels of nutrients and oxygen to the roots. The plastic construction materials caused complications with GC-MS analysis as samples were contaminated, with detectable plasticisers appearing in the chromatographs and requiring analysis before discounting them as compounds of interest.

Proprietary nutrient solutions were used in this hydroponics system. However, by designing a bespoke nutrient solution based on analysis of the local water and plant species requirements, plant growth could be optimised and individual components altered as required. It is suggested that GSL production varies under different fertilisation (Al-Khatib and Boydston, 1999; Josefsson, 1970; Ju et al, 1982) and pH conditions, and according to plant age and life cycle stage (Clossais-Besnard and Larher, 1991); manipulation of the hydroponic growing environment would allow ITC and other bioactive compound levels in root exudates to be characterised in optimal and sub-optimal conditions, as would the application of other stresses such as mechanical root damage.

226 Fungal antagonists Chapter 6

Chapter 6 Fungal antagonists

6.1 Introduction

This section of work investigates the use of fungal pathogens to control liverwort establishment and growth. A number of fungal antagonists have been shown to be effective biocontrol agents of weeds (van Driesche and Bellows, 1996), such as Co//etotrichum gloeosporioides f. sp. aeschynomene, marketed as Collego which is used to control northern joint vetch in rice and soybeans (Sandrin et al., 2003), and Phytophothora pa/mivora MWV, marketed as DeVine and used to control milkweed vine in citrus trees (Morrenia odorata) in Florida, US (Encore Technologies, 1999). A mycoherbicide effective in liverwort control would provide an alternative to chemical controls.

Fungal species identified as potential parasites on liverworts were sourced from the Centraalbureau voor Schimmelcultures (CBS), Baarn, The Netherlands: Bryoscyphus atromarginatus and six strains of Phaeodothis winter". Four further fungal species were isolated from dying liverworts provided by John Atwood (ADAS). They were obtained from Darby Nursery Stock, Methwold, Norfolk where they were growing in unsterilised loam compost; these were established on potato dextrose agar (PDA) and identified by CABI Bioscience, Egham as Fusarium equisek velutinum and Trichoderma harzianum (2 samples). Full details of the isolation and subsequent culture of these

227 Fungal antagonists Chapter 6 species is described (Section 6.3.1). These five species, all Ascomycetes (Table 6-1) are described below; descriptions are not intended to be comprehensive, but highlight relevant aspects of fungal morphology.

6.1.1 Bryoscyphus atromarginatus Verkley Aa & G.W. De Cock

Bryoscyphus is an obligate bryophilous genera, unknown beyond mosses and hepatics and one of the few Ascomycetes reported to parasitise thalloid hepatics (Dobbeler, 1997). B. atromarginatus (Table 6-1) is a rare ascomycete first collected in the Netherlands in 1995 that appears to parasitise Marchantia polymorpha, killing the liverwort rapidly, turning it brown and bearing numerous apothecia (Verkley et al., 1997). However, reinfection of M. polymorpha with this species has not been attempted, hence Koch's postulates have not been tested (Verkley, 2004).

Figure 6-1. Bryoscyphus atromarginatus, holotype (a) asci with ascospores (b) ascospores in water (c) detail of ascus apical apparatus (scale bars — 10 pm). Adapted from Verkley, 1997.

The genus Bryoscyphus is identified by the structure of the excipulum, truncate apices of the asci and fusoid, and frequently rhomboidal ascospores' (Figure 6-1) although there is a detailed description of the apothecia and associated structures, there is no description of asexual structures. Colonies are slow growing, producing both aerial and submerged mycelium which becomes ochre, maturing to shades of rose and purple; 228 Fungal antagonists Chapter 6 the underside has similar colouration with furrows radiating from the centre. Sterile apothecia are produced from aggregations of white mycelium on oatmeal agar, therefore no ascospores are produced in culture. Mycelium is hyaline, thin walled, septate with limited branching, filled with large green gutules containing oil droplets (Verkley et al., 1997).

6.1.2 Phaeodothis winter! (Niessl) Aptroot syn. Didymosphaeria marchantiae

P. winter! (Table 6-1) is a cosmopolitan, saprophytic species occurring on a wide range of hosts in many countries (Aptroot, 1995a). It has been found on numerous substrates, including grasses, herbs, bamboos, fungi, wood, cones, nuts and hepatics; it has been found on M. po/ymorpha in various locations, including Austria, Poland, Germany, and Sweden (Aptroot, 1995b).

Figure 6-2. Ascospores and ascus tips (a) Didymosphaeria marchantiae (b) Phaeosphaerella marchantiae, neotype B (c) Didymosphaeria schroeteria (d) D. thalictri (e) D. petrakiana . b-f have been reassigned as P. winter!. Adapted from Aptroot, 1995.

229 Fungal antagonists Chapter 6

P. winteri is an endophyte, with ascomata found immersed within the thallus, on archegoniophores and antheridiophores. Black spots on the thallus surface indicate the presence of fruiting bodies (Dobbeler, 2002). The black ascomata are pseudothecioid, immersed or superficial. Ascospores are characteristically euseptate with a broad, short, conical upper cell, and thinner longer lower cell (Figure 6-2)(Aptroot, 1995b).

6.1.3 Penicillium velutinum3.F.H. Beyma

There are over a hundred species of Penicillium, a common soil saprophyte mainly of temperate countries, also found on damp leather, decaying vegetation, and grain with spores invariably present in the air. Many species are economically beneficial, in their uses as antibiotics (P. notatum and P. chrysogenum) and in cheese ripening (P. camembertii and P. roqueforti), and detrimental, causing post harvest diseases in citrus fruit (P, digitatum, and P. italicum) (Ingold, 1973) and brown-rot of apples (P. expansum) (Webster, 1980).

Figure 6-3 Eupenicillium stolkiae (CSIR 1041) (A) Penicilli, (B) conidia, (C) asci producing singly on ascogenous hyphae, (D) asci containing different numbers of ascospores, (E) Ascospores (de Scott, 1975).

230 Fungal antagonists Chapter 6

Pen/c////um spp. are phialidic; variously branched conidiophores, depending on species, bear whorls of branches with terminal clusters of phialides (Figure 6-3) within which chains of dry, often green, conidia are produced and then dispersed by air movement (Webster, 1980). Conidiophores of P. velutinum are septate with smooth walls mostly monoverticillate penicilli, that are sometimes branched, each having terminal clusters of 6 to 10 phialides, bearing long, narrow tubes producing chains up to 100 pm in length of sage-green unicellular conidia, 3.0 — 4.0 pm diameter. When growing on Czapek media the colony is almost white and radially furrowed, producing spores after two weeks. The texture is velvety, ranging in colour through shades of blue- to grey- green and slate olive to deep slate olive with maturity; the underside begins dark yellow or greenish brown, darkening to almost black with age (de Scott, 1975).

The teleomorphic state of P. velutinum, Eupenicillium stolkiae (Figure 6-3) produces orange-brown cleistothecia from branched aerial hyphae, after 2-3 weeks producing hyphae bearing asci of various shapes and sizes, small or elongate, singly, either terminally or laterally. Ascospores are lenticular and approximately 4.0 x 4.0 pm (de Scott, 1975).

6.1.4 Fusarium equiseti (Corda) Saccardo

F. equiseti is a cosmopolitan species found in soils worldwide and isolated from many different plant materials. Although not an aggressive pathogen, it does infect numerous crops and causes root and post-harvest fruit rots, and is pathogenic to cereal seedlings (Booth, 1971). It has been shown to be an effective biocontrol agent against water hyacinth (Eichhornia crassipes)(Naseema et al., 2001). Phytotoxicity is reported and strains of F. equiseti were found to inhibit seed germination and seedling growth of various plants including brassicas and cucurbits, producing the toxin equisetin as they infected plant tissue (Wheeler et al., 1999). One strain of F. equiseti produces nematode-antagonistic trichothecene compounds that inhibit root knot nematode (Meloidogyne incognita) (Nitao et al., 2001).

Young cultures of F. equiseti have white floccose mycelium, tinged peach, aging to beige then deep olive buff above; beneath maturing from peach to dark brown. It produces septate (4-7) macroconidia, approximately 50 x 4.5 pm, from lateral phialides, producing Penicillium-like branched conidiophores. Conidia are falcate, with a

231 Fungal antagonists Chapter 6 foot cell and attached apical cell, bent inwards (Figure 6-4)(Booth, 1978). Intercalary chlamydospores, 6-10 pm are produced singly, in chains or knots, although perithecia are rare. When grown on media, cultures can produce mycelium only, which remains peach or buff, or pionnote, producing spores over the mycelium, the surface covered in spores from many sporodochia. The teleomorph is Gibberella intricans Wollenw (Booth, 1971).

There is a concern that Fusarium spp. can cause mycotoxicosis or invasive disease in humans by the production of the toxins fumonisin (Kroschel etal., 1996) and equisetin. Strains of F. equiseti tested by Thiel (1991) did not produce the cancer promoting mycotoxins Fumonisin B1 or B2, however equisetin has repeatedly been found in environments where several genetically unrelated individuals each developed leukaemia (Phillips etal., 1989).

Figure 6-4. F. equiseti (a) conidia (b) conidiophores and (c) chlamydospores (Booth, 1971)

232 Table 6-1. Taxonomic details of fungal specimens (Index Fungorum, 2004) Bryoscyphus T harzianum F. equiseti (Corda) P. winteri (Niessl) atromarginatus P. velutinum J.F.H. Genus & species Rifai (1969) Saccardo Aptroot (1995) Verkley, Aa Beyma (1935) G.W. De Cock (1997) Family Hypocreaceae Nectriaceae Phaeosphaeriaceae Helotiaceae Trichocomaceae Order Hypocreales Hypocreales Pleospora les Helotiales Eurotiales Subclass Sordariomycetidae Sordariomycetidae Dothideomycetidae Leotiomyceldae Eurotiomycetidae Class Ascomycetes Ascomycetes Ascomycetes Ascomycetes Ascomycetes Phylum Ascomycota Ascomycota Ascomycota Ascomycota Ascomycota Kingdom Fungi Fungi Fungi Fungi Fungi Ascocarp group Pyrenomycetes Pyrenomycetes Loculoascomycetes Discomycetes Plectomycetes (Webster, 1980)

233 Fungal antagonists Chapter 6

6.1.5 Trichoderma harzianum Rifai

T harzianum is a saprophytic fungus antagonistic to plant pathogens, widely distributed in soil, plant material, decaying vegetation and wood. The antagonist nature of T harzianum towards plant pathogenic fungi, its antibiotic production and nutrient competitiveness, is widely harnessed as a biocontrol agent against Rhizoctonla solam; Sclerotinia sclerotiorum and Botlytis cinerea (Batta, 2004). A number of commercial products containing T harzianum are available in several formulations in the UK: wettable powders (e.g. Bio-fungus®, Rootshield®, Trichodex®), granular formulations (e.g. Bio-Fungus®, T-22 Planter Box®), pelleted (Binab r) and as a conidial suspension (e.g. Promote®) (Copping, 2001).

Figure 6-5 Trichoderma conidiophore bearing conidia. Adapted from Agrios (1988)

Trichoderma produces repeatedly branched conidiophores (Figure 6-5) which rise above the mycelium forming a pyramidal structure, bearing hyaline, flask-shaped conidiogenous phialides attached to the conidiophores at right angles (Kubicek and Harman, 1998). The unicellular conidia are mostly green, less than 3 pm diameter having smooth or rough walls which are slimy, causing masses of sticky conidial heads (glioconidia) to form at the apex of the phialide, dispersed by a number of means including insects and rain splash (Webster, 1980). Resting, intercalary unicellular chlamydospores capable of surviving adverse conditions are also formed. When grown on potato dextrose agar (PDA) colonies are woolly and white, becoming blue- or

234 Fungal antagonists Chapter 6 yellow-green as conidia form, with the reverse pale tan or yellow and have an odour of coconut or camphor (Kubicek and Harman, 1998).

These fungal strains were cultured on agar media and methods developed to test their pathogenicity towards healthy liverwort, initially in the laboratory and subsequently in glasshouse experiments.

6.2 General methods

6.2.1 Fungal culture on agar media

Media formulation used for fungal culture was supplied in powder form and prepared according to the manufacturers instructions (Table 6-2), mixed with distilled water, autoclaved (Chapter 2) and poured into sterile Petri dishes within a sterile tissue flow cabinet and stored in ambient conditions. Once inoculated, cultures were incubated (23 °C) in a Gallenkamp Incubator Model 1H-150.

Table 6-2 Components of media used for fungal culture. Supplier: Sigma. Malt extract Potato dextrose Potato dextrose Component agar (MEA) agar (PDA) broth (PDB) M 6907 P 2182 P 6685 Agar 15.00 15.0 Dextrin 2.75 Glucose (dextrose) 20.0 20.0 Glycerol 2.35 Maltose 12.75 Peptone 0.78 Infusion from potatoes 4.0 4.0 Grams of powder required to prepare 1L 33.60 39.0 24.0

235 Fungal antagonists Chapter 6

6.2.2 Spore suspension preparation

A small amount of sterile distilled water was poured over a Petri dish of sporulating fungi. A flame sterilised inoculating loop was used to scrape the surface of the fungi, releasing spores, and the fluid filtered via a double layer of muslin to remove mycelium into a centrifuge tube, centrifuged (2.5 mins, 3000 rpm), the water poured off and the spores resuspended in sterile distilled water. The spore concentration was calculated using a haemocytometer, aiming for around 1 x 105 spores cm-2.

6.2.3 Incubation chamber preparation

Incubation chambers used for bioassays were transparent, lidded, non-airtight plastic boxes with the lid and base lined with tissue moistened with sterile water, creating a humid atmosphere. Liverwort specimens rested on sterilised squares of green plastic mesh, approximately 3 cm2, raising them above the wet tissue (Figure 6-6).

Figure 6-6. Incubation chamber containing inoculated liverworts.

6.2.4 Liverwort tissue preparation for microscopic examination

Liverwort tissue was soaked in 95% methanol in a water bath (65 °C, approx. 30 mins) to remove chlorophyll, dipped in chloralhydrate to clear the tissue, and then stained with lactophenol tryphan blue before mounting onto microscope slides. To make slides semi-permanent tissue was mounted in 50% glycerol (Mansfield, 2005).

236 Fungal antagonists Chapter 6

6.3 Experimental work

6.3.1 Preparation of fungal pathogens from dying liverwort samples

6.3.1.1 Introduction

Each of the dying liverworts was colonised by several different fungal species which needed to be isolated onto agar media and individual species isolated to separate Petri dishes for identification and future experimental work.

Species of P. winteri and B. atromarginatus were supplied on agar slants (Table 6-3) and were subcultured onto PDA for experimental work and long term storage. All fungal species used in this work were regularly subcultured every three to four weeks to maintain viable cultures on fresh media.

6.3.1.2 Method

Pathogens were removed from dying liverwort samples (Figure 6-7) in sterile conditions using a heat sterilised wire loop and spread onto potato dextrose agar (PDA) media in Petri dishes, sealed with Parafilm PM-922 sealing film (Manufacturer: American National Can, USA) and incubated (20 °C) (Section 6.2.1). The liverwort tissue was not surface sterilised, as the fungal colonies to be isolated were growing on the surface. Once colonies had developed enough to distinguish between different species, each was reisolated onto new media, using a 'streaking' technique to separate colonies, and repeating this procedure until Petri dishes containing individual species were produced.

Figure 6-7. Samples of dying liverwort provided by John Atwood (ADAS)

237

Fungal antagonists Chapter 6

Specimens sourced from CBS (Figure 6-9 and Table 6-1) were subcultured, in sterile conditions, onto PDA and malt extract agar (MEA) in Petri dishes, sealed with parafilm and incubated according to the information provided. Each of the fungal species was then established on sealed slants and refrigerated as a long-term maintenance procedure, and in Petri dishes for pathogenicity testing.

Table 6-3. Fungal strains sourced from Centraalbureau voor Schimmelcultures Strain Name Country of origin Pathogen growing conditions

CBS 162.31 P. winter! Indonesia 30 °C /PDA

CBS 182.58 P. winter! France *PDA CBS 551.63 P. winter/ Liberia 30 °C / PDA CBS 429.96 P. winter! Papua New Guinea 21 °C / PDA CBS 102466 P. winter! El Salvador 24 °C / PDA CBS 102483 P. winter! El Salvador 24 °C / PDA CBS 211.96 B. atromarginatus Netherlands 21 °C / MEA *Temperature unknown.

6.3.1.3 Results

Species isolated from liverworts were identified by CABI Biosciences (Egham, UK) as T harzianum, F. equiseti and P. ve/utinum (Figure 6-8); P. winter! and B. atromarginatus were successfully subcultured (Figure 6-9) and were found tolerant of the general incubation temperature (20 °C). There was great variation in growth rate of the different species, with F. equiseti the most vigorous; B. atromarginatus and P. winter!

CBS 182.58 were slow to establish and proved difficult to maintain.

(a)

Figure 6-8. Fungal cultures isolated from dying liverwort samples. (a) T. harzianum (b) F. Equiseti (c) P. velutinum.

238

Fungal antagonists Chapter 6

Figure 6-9. Fungal specimens sourced from CBS: (a) P. winteri CBS 102466 (b) P. winteri CBS 102483 (c) P. winteri CBS 162.31 (d) P. winteri CBS 182.58 (e) P. winteri CBS 429.96 (f) P. winteri CBS 551.63 (g) B. atromarginatus.

6.3.2 Preliminary spore germination bioassay

6.3.2.1 Introduction

Preliminary spore germination bioassays using droplets of spore suspensions on microscope slides were designed to test the viability of spores, so they could easily be viewed microscopically.

6.3.2.2 Method

T harzianum was tested initially to confirm spore viability. Spores were prepared (Section 6.2.2) and three 20 pl drops of spore suspension in sterile water placed individually on a microscope slide, incubated (Section 6.2.3) and examined microscopically for germination after 24 hrs.

The T harzianum spores did not germinate in sterile water only. It is known that nutrients can encourage spore germination (Chapter 1) therefore the process was

239 Fungal antagonists Chapter 6

repeated, suspending spores in sterile water (i.e. no added nutrients), and 50%, 25%, 12.5% and 6.25% nutrient solution (potato dextrose broth, PDB). P. velutinum, P. winter' CBS 551.63, and F. equiseti were included in this bioassay; B. atromarginatus and other strains of P. winteri were excluded as they had not produced spores in culture by this time.

6.3.2.3 Results

T harzianum and P. velutinum did not germinate in sterile water only (Table 6-4); however spores of all species germinated within 24 hrs when suspended in the above 25% PDB, confirming the viability of spores and that nutrients improve spore germination for these species.

Table 6-4. Germination bioassay of spores suspended in water and PD broth. ✓ = spores germinated; x = germination failed. Fungal species PDB concentration 50% 25% 12.50% 6.25% Water

T. harzianum ✓ ✓ x x x

P. winter' CBS 551.63 ✓ ✓ ✓ ✓ ✓

F. equiseti ✓ ✓ ✓ ✓ ✓

P. velutinum ✓ ✓ ✓ ✓ x

6.3.3 Preliminary liverwort inoculation with fungal spores

6.3.3.1 Introduction

A bioassay was designed placing liverworts in incubation chambers, inoculating them with fungal spores suspended in different concentrations of PDB and recording the effects using a scale incorporating the proliferation of fungal mycelium and the response of the liverwort. Spore germination bioassays (Section 6.3.2.1) were carried out simultaneously confirming spore viability, using only sporulating species.

240 Fungal antagonists Chapter 6

6.3.3.2 Method

Spores of P. winteri CBS 551.61, F. equiseti, P. velutinum and T harzianum were prepared (Section 6.2.2) to give spore densities indicated in Table 6-5, suspended in either sterile water, or 50%, 25%, 12.5% or 6.25% PDB. Liverworts of approximately 2 cm diameter were used for inoculations; they were removed from their growing media, M51C (Chapter 2) rinsed and placed on plastic mesh in incubation boxes (Section 6.2.3), four per box. The surface of the liverwort was allowed to dry then inoculated with three 20 pl drops of spore suspension; two liverworts were used for each broth dilution treatment. Spore germination bioassays were carried out as before.

Fungal growth was reisolated from the liverwort thallus surface by scraping fungal mycelium from the liverwort with a heat sterilised inoculation loop and streaking it onto PDA or MEA in Petri dishes and incubating.

Table 6-5. Concentration of fungal spores used in liverwort inoculation Fungal species Spore density

T hatztanum 2.4 x 107 cm-3

P. winteri CBS 551.63 9.05 x 105 cm-3

F. equiseti 9.0 x 105 cm-3

P. velutinum 6.48 x 106 cm-3

6.3.3.3 Results

Semi-permanent slides were made of inoculated liverwort tissue; those inoculated with P. velutinum showed mycelium attached to the liverwort, although the hyphae appear to be growing over the surface of thallus and rhizoid tissue rather than penetrating (Figure 6-10). The liverworts inoculated with 50% and 25% PDB collapsed; one treatment using 50% PDB had died completely and rotted after 2 months. Others that had almost died away showed signs of new growth at the outer edges.

241

Fungal antagonists Chapter 6

Figure 6-10. Liverwort infected with P, velutinum (a) thallus (b) rhizoid. Magnification X400

T harzianum did not appear to have infected the liverwort, however microscopic examination indicated that spores had germinated and begun to invade the tissue, but then stopped. The majority of liverworts inoculated with P. winteri CBS 551.63 were growing strongly after 2 months with no real trends with regard to PDB dilution level. F. equiseti was the most successful of the species tested, causing severe collapse with all dilutions of PDB and water, although three samples did exhibit signs of new growth at the edges. Round orange sporodochia developed on the surface of one liverwort which released many macroconidia when crushed (Figure 6-11).

Sporodochium

Figure 6-11(a) Liverwort infected with F. equiseti with orange sporodochia. (b) Sporodochium releasing macroconidia, found on liverwort inoculated with F. equiseti. (c) F. equiseti macroconidia. Magnification x 400.

Inoculated liverworts showed signs of infection, with tissue collapsing (Figure 6-12). Fungal growth was observed on the surface of liverwort tissue inoculated with F. equiseti, T harzianum and P. velutinum; attempts made to reisolate the fungi were successful with F. equiseti and P. velutinum (Figure 6-12). The spore germination bioassays confirmed the viability of the spores used (Table 6-6), with the addition of nutrients unnecessary for F. equiseti:

242 Fungal antagonists Chapter 6

Figure 6-12(a) Liverwort infected with F, equiseti (b) fungus reisolated from liverwort inoculated with F. equiseti.

Table 6-6. Results of spore germination bioassay. ✓ = spores germinated; x = ermination failed. Nutrient Solution Fungal species Water 50% 25% 12.5% 6.25%

T harzianum ✓ ✓ x x x

P. velutinum ✓ ✓ ✓ ✓ x

F. equiseti ✓ ✓ ✓ ✓ ✓

P. winteri CBS 551.63 ✓ ✓ ✓ ✓ x

6.3.4 Liverwort inoculation with fungal spores

6.3.4.1 Introduction

A further set of inoculations were carried out to include the other P. winter/ strains that had subsequently started to produce spores; B. atromarginatus had not produced any spores as yet and was therefore applied as a mycelial suspension. The F. equiseti was again included to confirm its vigorous effect on liverwort could be replicated.

6.3.4.2 Method

Liverwort sections (2 cm2) were inoculated using spores in 50% PDB, water, others had no treatment applied (control); spores suspensions were prepared as previously

(Section 6.2.2), applied at densities shown in Table 6-7. Two liverwort sections were used per treatment, each inoculated with five 20 pl drops of inoculant. Spore germination bioassays were carried out as before, using mycelium suspension for B. atromarginatus.

243 Fungal antagonists Chapter 6

Table 6-7. Spore densities applied to liverworts Fungal species Spore Density

P. winteri CBS 551.63 1.12 x 106 cm -3 P. winteri CBS 102466 7.0 x 106 cm-3 P. winteri CBS 429.96 5.32 x 106 cm-3 B. atromarginatus CBS 211.96 mycelium F. equiseti 5.2 x 106 cm-3 Control

6.3.4.3 Results

The spores and mycelium in 50% PDB all germinated within 24 hrs. P. winteri CBS 102466 spores did germinate in water only, but more sparsely than when suspended in nutrient solution (Table 6-8).

Table 6-8. Spore germination bioassay inoculation. ✓ = spores germinated; x = germination failed.

Fungal species Water 50% PDB

P. winteri CBS 551.63 ✓ ✓

P. winteri CBS 102466 ✓ ✓

P. winteri CBS 429.96 x ✓

B. atromarginatus CBS 211.96 ✓ ✓

F. equiseti ✓ ✓

Table 6-9. Levels of collapse of liverwort 21 days post inoculation. X = no response. Spores + Water Spores + 50% PDB Fungal species Sample 1 Sample 2 Sample 1 Sample 2 P. winteri CBS 551.63 x x 25% slight P. winteri CBS 102466 x x x x P. winteri CBS 429.96 x 50% slight x B. atromarginatus CBS 211.96 x x slight slight F. equiseti 50% 25% 25% 25%

Control slight slight .. -

244 Fungal antagonists Chapter 6

The results of this inoculation were varied (Table 6-9), with only P. winteriCBS 429.96 sample 2 adversely affecting the liverwort when applied without nutrients. The F. equiseti again produced the greatest response from the liverwort, with 95% collapse of one sample after 34 days. Subsequently, some weak liverwort re-growth was observed around the edge of the dying liverwort, but this did not develop further.

6.3.5 Preliminary liverwort inoculation with fungal agar plugs

6.3.5.1 Introduction

As some fungal strains had still not produced spores, an alternative inoculation method was developed to inoculate liverworts using agar plugs colonised with fungi.

6.3.5.2 Method

Plugs of established mycelium on agar media were taken using a No. 2 cork borer and inverted onto the liverwort thallus so that contact was made between the mycelium and liverwort tissue. For each fungal strain, four liverwort samples were inoculated with five plugs each; of these, half were wounded with a needle to investigate whether the response would be greater in wounded or unwounded liverworts. The control treatment was liverwort without agar or fungal inoculum applied.

Observations were made on the amount of fungal growth using the characteristics: ■ Growth visible on plug • Growth extending beyond plug by less than 1 plug diameter ■ Growth extending beyond plug, greater than 1 plug diameter ■ The effect on the condition of the liverwort was graded as 'slight', or 'collapsed'

6.3.5.3 Results

All fungal species were well established on the agar plugs, covering the upper plug surface 8 days post inoculation (Table 6-10); P. winter' CBS 102483 and CBS 162.31 had not extended beyond the plug by this time. F. equi:seti again proved vigorous, with liverwort tissue collapsing around all plugs. The greatest effect on the liverwort was produced by F. equiseti, P. velutinum, P. winteri CBS 551.63 and CBS 182.58, where collapse of liverwort tissue was observed associated with the greatest number of 245 Fungal antagonists Chapter 6 inoculated plugs. The control liverworts were all healthy on the day when data was collected, except for one sample that was soft and beginning to collapse in the centre; after 15 days they were all pale green and two were collapsing.

Wounding the liverworts did seem to produce an effect (Table 6-10), with more samples collapsing with P. winteri CBS 551.63, CBS 182.58, B. atromarginatus and P. ve/ut/num; less effect was observed with P. winteri CBS 102466 and CBS 429.96; an equal number with F. equiseti, and no effect with or without wounding for P. winter' CBS 102483 and CBS 162.31. Small, black structures were found growing on the surface of liverworts inoculated with P. winteri CBS 551.63. These were examined microscopically and found to be pycnidia which when crushed released thousands of spores (Figure 6-13);

Figure 6-13. P. winteri CBS 551.63 (a) Mycelium infecting liverwort thallus. (b) Pycnidia releasing thousands of conidia, (c) Pycnidia on the thallus surface. (a) and (b) Stain = lactophenol tryphan blue. Magnification x 400

It was decided not to continue with pathogenicity testing on fungal species that were slow to grow and elicited little response from the liverwort: P. winter/ CBS 182.58, P. winter/ CBS 429.96, P. velutinum and T harz/anum. Although P. ve/ut/num did affect liverwort tissue and cause some collapse, the liverwort was not overcome and continued to develop new growth.

246 Table 6-10. The effect of fungal plug inoculation of liverwort sections, 8 days post inoculation. Figures refer to the number of plugs affected

Growth extended Growth extended

Growth visible on plug beyond plug within 1 beyond plug Liverwort condition

Fungal species plug diameter > 1 plug diameter Wounded Unwounded Wounded Unwounded Wounded Unwounded Wounded Unwounded Slight Collapsed Slight Collapsed

P. winteriCBS 551.63 10 10 10 10 6 4 - 10 2 4

P. winteriCBS 182.58 10 10 2 2 2 8 3 7 P. winteri CBS 102466 10 10 10 9 3 5 4 6 P. winteriCBS 102483 10 10 10 10 1 5 - 6

P. wfriteriCBS 162.31 5 3 - 5 3 P winteri CBS 429.96 10 10 3 - 4 1 3 2 B atromarginatus 9 10 10 8 - 1 5 4 3 CBS 211.96

F equiseti 10 10 10 10 9 7 10 10

P velutinum 10 10 10 10 7 6 Control (no plugs applied)

247 Fungal antagonists Chapter 6

6.3.6 Liverwort inoculation with fungal plugs

6.3.6.1 Introduction

A second experiment using inoculation of liverworts with fungal agar plugs was made to confirm previous results.

6.3.6.2 Method

Plugs of mycelium were again used for this set of inoculations. For each fungal strain used, six liverwort samples were inoculated with 2 plugs each; of which three were wounded and 3 remained unwounded. The control was uninoculated plugs.

6.3.6.3 Results

The only two fungal strains that did not establish well on the agar plugs were P. winteri CBS 162.31 and B. atromarginatus. F. equiseti had the most vigorous effect on the liverwort, with collapse surrounding each inoculated plug applied (Table 6-11). After 11 days all replicates inoculated with P. winteri CBS 102483 were covered with fungal mycelium; wounded liverworts inoculated with P.winteri CBS 551.63 had collapsed, becoming black and slimy. The controls and those liverworts inoculated with P. winteri CBS 102466 and 162.31, were dull green and collapsing but without evidence of fungal infection, suggesting a cause other than the inoculant. The most promising strains were F. equiseti, P. winteri CBS 551.63 and P. winteri CBS 102466. Reisolation of fungal strains of liverwort continued, being successful with F. equiseti, B. atromarginatus and P. winter! CBS 162.31 and 102466. B. atromarginatus was slow to establish and had little effect on the liverwort. As it had been reported to attack M. po/ymorpha specifically, efforts were made to reinvigorate it by repeatedly passaging it through liverworts. However, in this process the cultures became contaminated with Stachybotrys chartarum, a species commonly found on plant debris, soil and in damp indoor environments, responsible for stachybotryotoxicosis in animals and humans (Nelson, 2001). No further studies with B. atromarginatus were carried out.

248 Table 6-11 The effect of fungal plug inoculation of liverwort sections, 11 days post inoculation. Figures refer to the number of plugs affected. ** denotes bacterial contamination Growth extended Growth extended

Fungal species Growth visible on plug beyond plug within 1 beyond plug > 1 plug Liverwort condition plug diameter diameter

Wounded Unwounded Wounded Unwounded Wounded Unwounded Wounded Unwounded

Slight Collapsed Slight Collapsed

P. winteri CBS 551.63 6 6 5 6 6 2 4

P, winter' CBS 102466 6 6 6 6 - - 2 2 3 3

P. winteri CBS 102483 6 6 6 6 2 2 2 3 3

P. winteri CBS 162.31 1 B. atromarginatus CBS 211.96

F. equiseti 6 6 6 6 6 6 - 6 - 6 Control (no plugs applied) ** ** - -

249 Fungal antagonists Chapter 6

6.3.7 Fungal glasshouse experiment

6.3.7.1 Introduction

A glasshouse experiment was designed to investigate the effect of F. equiseti and P. winted on liverwort when applied pre- and post- emergence to liverwort. These two species were those that had the greatest effect on liverwort in earlier laboratory tests (Sections 6.3.5 and 6.3.6). A successful pre-emergence mycoherbicide would combat liverwort prior emergence of liverwort thalli, minimising spread and colonisation of new areas; post-emergence treatments would act on existing infestations to prevent further spread and kill the liverwort.

Sterilised compost was used for the fungal treatments to eliminate interference by other fungal species. Three different controls were used (Table 6-12), sterilised compost; unsterilised compost to compare any effects of the sterilisation process; and sterilised compost with formulation but no fungal inoculation, to compare any effect of the formulation. Table 6-12 Controls and fun al treatments Control Treatment Control 1 (C1) Unsterilised compost only Control 2 (C2) Sterilised compost only Control 3 (C3) Sterilised compost with formulation only Fungal application (F) Formulation with fungal inoculum

For the pre-emergence treatment the formulation was prepared with and without fungal inoculation, incorporated into sterilised compost and liverwort gemmae dispersed over the surface. For the post-emergence treatment liverwort gemmae were established on sterilised and unsterilised compost to produce healthy, vigorous colonies ready for inoculation with a combination of spores and mycelium.

An appropriate inoculation carrier was required for the pre-emergence treatment. A method was adapted from Blok and Bollen (1997), who used a soil meal culture containing potting compost amended with 5% oatmeal, autoclaved (30 mins) on two consecutive days, inoculated with fungus and then incubated (23 °C, 2-3 weeks).

250 Fungal antagonists Chapter 6

6.3.7.2 Method

Compost sterilisation 700 g quantities of Imperial College compost media, comprised of a peat and grit substrate with base fertilisers (Chapter 2) were autoclaved (126 °C, 30 mins) in unsealed bags, six at a time to allow maximum steam penetration.

Fungal species Fungal species used for this experiment were F. equiseti and a mixture of four isolates of P. winter;, all of which showed vigour and had exhibited parasitic potential against liverwort in laboratory bioassays (Table 6-13).

Table 6-13. Fungal species used for pre and post-emergence treatment; inoculum quanti Fungal species No. fungal plugs P. winteri, 102466, 102483, 581.63, 162.31 3 F. equiseti 6

Post-emergence treatment Liverwort preparation Seed trays (19 x 23 cm) containing 700g autoclaved and unsterilised compost were each inoculated with liverwort gemmae: 5 gemma cups were placed with 10 ml sterile water in a centrifuge tube, agitated to separate and suspend the gemmae, and dispersed evenly over the compost; any remaining gemmae were rinsed out onto the compost with a further 10 ml water. The gemmae were then allowed to establish for eight weeks in normal glasshouse liverwort culture conditions (Chapter 2.2.2).

Spore inoculum preparation Spore suspensions were prepared as previously (Section 6.2.2) with the exception that the liquid and fungal tissue scraped from the cultures was transferred to a centrifuge tube without filtering, retaining both mycelium fragments and spores. Once resuspended in sterile water the number of colony forming units (spores and mycelium fragments) was estimated using a haemocytometer, the samples recentrifuged, the supernatant removed and fungal material resuspended in the carrier, a solution of 0.05% (0.5 mL) Tween 20 and 0.01% (0.1 g) glucose in 1 L sterile distilled water to achieve the required concentrations (Table 6-14); the glucose provided nutrients and 251

Fungal antagonists Chapter 6 the Tween 20 helped to separate the fungal material establishment, aiding even dispersal (Daigle and Connick, 1990).

Table 6-14 Post-emergence treatment fungal inoculum concentration Fungal species Colony forming unit concentration P. winteri 551.63, 163.91, 102466, 102483 3.86 x 106 F. equiseti 1.0 x 106

A second application of spores was carried out after 3 weeks (Table 6-15) with inoculum applied to one randomly selected tray per cabinet of the post-emergence treatments: I-post-Fe-fl, II-post-Fe-f3, I-post-Pw-f2, II-post-Pw-f3 as the liverwort appeared to be outgrowing the fungal infection initially observed.

Table 6-15. Post-emergence treatment fungal inoculum concentration — rd inoculation Fungal species Colony forming unit concentration P. winteri 551.63, 163.91, 102466, 102483 1.93 x 106 F. equiseti 1.64 x 106

Post-emergence treatment application For the post-emergence fungal treatments, 20 ml quantities of fungal inoculum of either F. equiseti or P. winteri were applied as a spray onto each tray of established liverwort. For control 3 (Table 6-12), sterilised compost with formulation, 20 ml quantities of carrier (glucose and Tween 20 in distilled water) without fungal inoculum was applied. Controls 1 and 2 did not require any additions to the compost.

Pre-emergence treatment Inoculum preparation Two pre-emergence formulations were prepared using oatmeal and bran as a nutrient supply. Sixteen conical flasks, eight containing oatmeal (70 g) finely ground in a food processor, and eight containing flaked bran (35 g) (bran was bulkier than ground oatmeal), with water (300 ml), were autoclaved (121°C, 20 mins) on two consecutive days to ensure sterility. Fungal plugs, taken using a No. 3 cork borer, were added to 252 Fungal antagonists Chapter 6 half the flasks of formulation as detailed in Table 6-13, two flasks each of oatmeal and bran per fungal group, and incubated on a shaker (InnovaTh 4330 refrigerated incubator shaker, New Brunswick Scientific) (120 rpm, 23 °C, for 7 days). Sterile water (75 ml) was added to each flask and incubated for 14 days, at which time 100 g sterilised compost was added to each flask to bulk up the inoculum and incubated for a further 12 days.

It was clear that both P. winteri and F. equiseti mycelia had proliferated more in the oatmeal formulation, and this was used for the experiment. 70 g quantities of formulation, with and without inoculum (fungal mycelium broken into small pieces by hand) was evenly incorporated into sterilised compost (700 g) to form six trays each of control 3 and fungal treatment for each fungal species.

Both pre- and post-emergence fungal treatments were placed in a humid environment prior to transference to growth chambers for 24 hrs to encourage the fungi to establish.

Liverwort gemmae were applied to the surface of the inoculated compost as for the post-emergence treatment, using 5 gemma cups in 10 ml sterile water in a centrifuge tube, dispersing them evenly over the compost and then rinsing any remaining gemmae out onto the compost with water (10 ml).

Experimental design Eight growth cabinets were prepared with a base of perforated polythene covered with capillary matting. They were set at 23 °C, using natural daylight and day length.

The growth cabinets were fitted with fans to circulate air, providing an even temperature, therefore each tray was fitted with a clear lid (supplier: William Sinclair Horticulture) to prevent contamination (Figure 6-15); trays therefore required careful handling during data collection, with lids removed from one tray at a time once removed from the growth cabinet.

253

Fungal antagonists Chapter 6

Block 1 Block 2 Post-emergence, F. equiseti Pre-emergence, P. winteri C2 F F C1 Cl F F C1 C1 C3 C3 C3 Cl F C2 C3 C2 F C1 C2 C2 C3 C3 C2

Pre-emergence, F. equiseti Pre-emergence, F. equiseti C1 F F F C1 C3 C3 C2 Cl C3 C3 C3 F Cl C1 F C1 C2 C2 C2 C2 C3 F C2

Post-emergence, P. winteri Post-emergence, F. equiseti Cl C2 Cl C1 Cl C3 F C2 F C3 C3 C3 F F C3 C3 F F C2 C2 C1 C2 C2 C1

Pre-emergence, P. winter! Post-emergence, P. winteri Cl Cl Cl F F C3 C3 C3 F F C3 C3 C1 C1 F Cl C2 C2 C2 C3 F C2 C2 C2

Figure 6-14. Experimental layout of randomly assigned controls and treatments, each 12 tray set contained within one growth chamber. Block 1=glasshouse 1; block 2=glasshouse 2; C1=control 1, unsterilised compost; C2=control 2, sterilised compost only; C3=control 3, sterilised compost with formulation only; F=fungal treatment applied in formulation.

%L- ,.417,;5‘mtle-

Figure 6-15. Growth cabinet arrangement

254 Fungal antagonists Chapter 6

Data collection An A4 size grid of 1 cm squares on clear plastic was placed over trays of liverwort and used as a template for recording growth. An outline of liverwort and dieback areas were drawn, scanned into a computer and the area calculated using Image) digital image analysis.

Areas of liverwort coverage and dieback were recorded weekly for 5 weeks, and then a final assessment was made after 11 weeks. Possible contaminants present on the compost surface were re-isolated from trays onto PDA.

At four weeks post-inoculation, re-isolations from fungal growth in the P. winteriand F. equiseti treatments were made onto potato dextrose agar (PDA), incubated and identified. At week 5 tissue samples with lesions were taken as previously (Section 6.2.4) and examined microscopically.

6.3.7.3 Results

Although the compost was sterilised and efforts were made to produce uncontaminated inoculum, it was not practical to keep the experiment completely sterile. However, the aim was to test the effect of fungal treatments in conditions reflecting those found in nurseries, albeit with temperature control and minimising cross-contamination of treatments and controls. Post-emergence treatments were prepared using sterilised compost to mitigate the effects of other fungal species in the compost, so that any infections observed should be due to those applied in the treatments; however they were subject to contamination from air-borne fungal spores prior to inoculation and from run-off water via the capillary matting throughout the experiment.

Five days post-inoculation, fungal mycelium was observed growing over the compost surface in the pre-emergence trays. Liverwort gemmae had germinated and liverwort thalli were growing vigorously (Figure 6-16), and by 21 days F. equiseti mycelium could be seen colonising the liverwort thallus (Figure 6-17). Fungi were successfully reisolated from inoculated treatments and identified (Table 6-16), except from the post -emergence F. equiseti treatments (II-post fe-f).

255 Fungal antagonists Chapter 6

Figure 6-16 Pre-emergence treatments (a) II-pre pw—f (b) II-pre-fe-f, 5 days post- inoculation. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P. winteri, f = fungal treatment.

Figure 6-17 Liverwort thallus colonised with F. equiseti mycelium. Pre-emergence treatment (II-pre-fe-f) 14 days post-inoculation. II = block 2, pre = pre-emergence, fe = F. equiseti, f = fungal treatment.

Table 6-16 Record of successful reisolation of fungi from fungal treatments. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, Fe = F. equiseti, Pw — P. winteri. Treatment Date I-Pre Fe 01/06/06 30/07/06 II-Pre Fe 01/06/06 30/07/06 I-Pre Pw 01/06/06 30/07/06 II-Pre Pw 01/06/06 30/07/06 I-Post Fe 01/06/06 30/07/06 II-Post Fe I-Post Pw 01/06/06 - II-Post Pw 01/06/06 30/07/06

256 Fungal antagonists Chapter 6

Figure 6-18 Microscopic analysis of infected liverwort tissue 5 weeks post inoculation. F. equiseti (a) macroconidia x400 (b) mycelium, x200 and P. winteri (c) and (d) mycelium wrapped around rhizoids x400 (e) spores x400 (f) an unidentified nematode among liverwort rhizoids x400

Microscopic analysis (Section 6.2.4) of infected liverwort tissue indicated presence of F. equiseti and P. winteri in their respective treatments. The F. equiseti was identifiable from its macroconidia (Figure 6-18a) although the mycelium appears to be growing around liverwort rhizoids rather than penetrating them (Figure 6-18b). Similarly, spores and mycelium of P. winteri were identified around liverwort rhizoids (Figure 6-18c-d). An unidentified nematode was also observed (Figure 6-18f) which may have been introduced with the unsterilised control treatment.

After 11 weeks liverwort growth was extremely vigorous in most treatments (Figure 6-19a) and outgrew the trays so the lids no longer fitted closely. However, in a minority of treatments the liverwort did not cover the tray (Figure 6-19b). Although efforts had been made to use sterile compost in controls 2 and 3, because the trays had been exposed in the glasshouse whilst the liverwort established they were unlikely to be sterile at the start of the experiment, and fungal mycelium was present in some (Figure 6-19c). Fungal mycelium was present within the gemma cups of a number of treatments, across both fungal species and all controls, in pre- and post-emergence treatments, indicating some cross contamination had occurred (Figure 6-20a-c).

257

Fungal antagonists Chapter 6

Figure 6-19. Liverwort growth after 11 weeks (a) II-pre fe-C3 (b) II-post pw-C3 (c) area of contamination, I-post pw-C2. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P, winteri, f = fungal treatment, C2=control 2 sterilised compost only, C3=sterilised compost with formulation.

Figure 6-20 Liverworts with gemma cups infested with fungal mycelium (a) II-post pw-C3 (b) I-pre fe-f (c) II-post pw-f. I= block 1, II = block 2, pre = pre-emergence, post = post-emergence, fe = F. equiseti, pw — P. winteri, f = fungal treatment, C3 = control 3, sterilised compost with formulation,

The data for this experiment is presented and analysed separately for total, healthy (total area less dieback area), and dieback areas of liverwort. Analysis was by analysis of variance (ANOVA) and mean separations where appropriate, using Genstat for Windows, 8th Edition (Chapter 2). The results of the analyses are summarised within the text and ANOVA tables are presented in Appendix 4.

Pre-emergence P. winteri treatment

For the pre-emergence P. winteri treatment, at week 3 (Figure 6-21) the total (v.r.3,3= 14.97, P<0.05) and healthy areas (v.r.3,3= 21.98, P<0.05) of liverwort were significantly less when grown in controls 1 and 2 (unsterilised and sterilised compost) than when treated with the fungus formulation; there was no dieback by this stage. By week 5, least liverwort dieback was observed with control 2 (sterilised compost) and

258 Fungal antagonists Chapter 6 then the fungal treatment. Greatest dieback occurred using control 3 (sterilised compost with formulation). By weeks 5 and 11 (Figure 6-22) there was no significant difference in liverwort area between the P. winter/ treatment and any of the controls; at week 11 the greatest dieback occurred in the fungal treatments, but again this was not significant.

Pre-emergence F. equiseti treatment For the pre-emergence F. equiseti treatment the total area of liverwort was significantly less at week 3 (v.r.3,3= 17.27, P<0.05) and significantly less at week 5 (v.r.3,3= 46.74, P<0.05)(Figure 6-23 ), and healthy liverwort area was significantly less at weeks 3 and 5, when grown in control 2 (sterilised compost) than when treated with the fungal formulation or controls 1 (unsterilised compost) and 3 (sterilised compost with formulation). However, two of the six liverworts grown in control 1 were markedly smaller than the remainder, at 7822 and 6614 mm2 compared to the mean (of six liverworts) of 17,043 mm2 at week 5, and this may be distorting the results. Figure 6-23 represents growth with the outliers removed. Dieback at week 5 was least with control 1. As with the pre-emergence P. winteri treatment, greatest dieback was observed with control 3, suggesting the formulation had the greatest effect on liverwort dieback. However, closer inspection of the data revealed that for F. equiseti; within control 3 only one liverwort exhibited any dieback (10,892 mm2) in week 5 and again this seems to be distorting the figures, and the small areas of dieback present in control 3 in weeks 3 and 4 were no longer apparent. By week 11 total and healthy areas of liverwort were not significantly different among the treatments (Figure 6-22), although dieback area of the F.equiSeti-treated liverwort was greater than the controls (v.r.3,3= 5.36, P<0.05).

Post-emergence F. equiseti treatments For the post-emergence F. equiseti treatment, at week 3 the total (v.r.3,3= 18.86, P<0.05) (Figure 6-24) and healthy (v.r.3,3= 15.79, P<0.05) (Figure 6-25) liverwort areas were significantly less when grown in unsterilised compost than the F. equiseti and sterilised

259 Fungal antagonists Chapter 6

35000

30000 -

25000 - E E Iv 20000 -

15000 -

O r 10000 -

5000 -

Week 1 Week 2 Week 3 Week 4 Week 5 Time

40000 -

"E 30000 - E

20000 - g 10000 - m

0 -

Week 1 Week 2 Week 3 Week 4 Week 5 Time

600 Dieback rs, 500

3 400

300 a

200

g 100 - j 0- 0

Week 1 Week 2 Week 3 Week 4 Week 5 Time

Control 1 - unsterilised compost only 0. Control 2 - sterilised compost only - - -v- -- Control 3 - sterilised compost with formulation — •-v — • Fungal treatment with formulation

Figure 6-21. Areas of liverwort treated pre-emergence with P. winteri

260 Fungal antagonists Chapter 6

Total 60000 -

E 40000 -

20000 -

Control 1 Control 2 Control 3 Pre Fe Pre Pw Treatment

Healthy 60000 -

E 2 40000 - Io

20000 -

Control 1 Control 2 Control 3 Pre Fe Pre Pw Treatment

Dieback 60000

40000

g 20000

0 I I r

Control 1 Control 2 Control 3 Pre Fe Pre Pw Treatment Figure 6-22. Areas of pre-emergence treated liverwort after 11 weeks. Control 1 = unsterilised compost only. Control 2 = sterilised compost only. Control 3 = sterilised compost with formulation. Pre Fe = pre-emergence F. equisetitreatment. Pre Pw = pre-emergence P. winteritreatment.

261 Fungal antagonists Chapter 6

35000

30000 -

25000 - NE 20000 -

15000 -

10000 -

5000 -

Week 1 Week 2 Week 3 Week 4 Week 5 Time

35000

30000 -

N 25000 - E E 20000 -

15000 - a 10000 -

5000 -

Week 1 Week 2 Week 3 Week 4 Week 5

Time

2000 Dieback 1800 - 1600 - E E 1400 - 1200 - 1000 - 5th 800 - 2 0 e 600 - 400 - .a 200 - 0- e • •

Week 1 Week 2 Week 3 Week 4 Week 5 Time • Control 1- unsterilised compost only o Control 2 - sterilised compost only Control 3- sterilised compost with formulation Fungal treatment applied with formulation

Figure 6-23. Areas of liverwort treated pre-emergence with F. equiseti. Total and healthy liverwort areas represented with two outliers removed.

262 Fungal antagonists Chapter 6 treatments. There was no significant difference between dieback areas of liverwort grown in any treatments, including the unsterilised compost (v.r.3,3= 2.60, P<0.05)(Figure 6-26).

At week 3 one randomly selected fungal-treated tray from each block was inoculated for a second time; therefore for statistical analysis and graphs, figures are averages of one tray per block for '2 application' treatments, and two trays per block for '1 application' treatments.

By week 5, for the post-emergence F. equiseti treatment both the total (v.r.3,3= 63.89, P<0.05) (Figure 6-24) and healthy (v.r.3,3= 83.08, P<0.05) (Figure 6-25) liverwort area of control 1 (unsterilised compost) was significantly less than when treated with 1 application of fungal inoculum. For those trays treated with two fungal applications, although liverwort growth was less than with a single fungal application, total and healthy liverwort area was still greater than with control 1 (unsterilised compost). There was no significant difference between control 1 when compared with 2 fungal application treatments for either total (v.r.3,3= 3.39, P<0.05) or healthy (v.r.3,3= 1.21, P<0.05) liverwort area. There were no significant differences in dieback area for any of the treatments (Figure 6-26). Although liverwort dieback was greatest for the fungal treatment with the second fungal application (Figure 6-26b), this was not found to be significant (v.r3,3 = 0.92, P<0.005) and was due to a single tray of liverwort with a dieback area over 1000 mm2 at 8375 mm2.

By week 11, there was no significant difference in total (Figure 6-27), healthy (Figure 6-28) or dieback (Figure 6-29) liverwort areas with one or two post-emergence F. equiseti applications. With one fungal application total liverwort area was least in control 1 (unsterilised compost) and greatest with the fungal application; and with a second fungal application least in control 1 (unsterilised compost) and greatest in control 3 (sterilised compost with formulation), indicating an effect of the second fungal application, albeit not significant. This trend was reflected with healthy liverwort area (Figure 6-28) where liverwort area was least in control 1 (unsterilised compost) with both one and two fungal applications. When comparing fungal treatments, liverwort area was less with two fungal applications than one, again indicating an effect of the second fungal application.

263 Fungal antagonists Chapter 6

45000

40000 -

(NI E E 35000

▪8•3 30000 -

25000 - > =1•

20000 -

15000 Week 0 Week1 Week 2 Week 3 Week 4 Week 5

Time

45000

40000 -

(N E E 35000

ca2 8 30000 - mca

25000 - =1•

20000 -

15000 Week 0 Weekl Week 2 Week 3 Week 4 Week 5

lime

—It— Control 1 - unsterilised compost only o Control 2 - sterilised compost only Control 3 - sterilised compost with formulation —••-v•—•• Fungal treatment applied in formulation

Figure 6-24. Growth curve of total liverwort area of liverwort subjected to (a) one and (b) two applications of F. equiseti

264

Fungal antagonists Chapter 6

cN E E co

t 0

Week 0 Week1 Week 2 Week 3 Week 4 Week 5

Time

45000

40000 -

E E 35000 - m

fl) 30000 - 0 t 25000 - r,>

20000 -

15000 Week 0 Weekl Week 2 Week 3 Week 4 Week 5

Time

Control 1 - unsterilised compost only 0 Control 2 - sterilised compost only Control 3 - sterilised compost with formulation Fungal treatment applied in formulation

Figure 6-25. Growth curve of healthy liverwort subjected to (a) one and (b) two applications of F. equiseti

265

Fungal antagonists Chapter 6

(a) 4000 -

(N E 3000 - a9 co"

2000 -

N 'Es 1000 - 0o •. ••••••...... • ,„ . „. ... 0 -

Week 0 Weekl Week 2 Week 3 Week 4 Week 5

Time

5000 (b) V 4000 - rs1 E E

3000 -

_6 2000 - (0 a) t, 1000 - /

g) •••••...... • . . 0...... 0-

Week 0 Weekl Week 2 Week 3 Week 4 Week 5

Time

• Control 1 - unsterilised compost only

0 Control 2 - sterilised compost only Control 3 - sterilised compost with formulation Fungal treatment applied with formulation

Figure 6-26. Growth curve of liverwort dieback, subjected to (a) one and (b) two applications of F. equiseti

266 Fungal antagonists Figure 6-27.Totalarea ofliverwortsubjectedto(a)oneand(b)two applicationsof fungal inoculum,after elevenweeks. control 3sterilisedcompost withformulation. winteri.

Control 1—unsterilised compostonly,control2-sterilised only, 2 Liverwort area mm2 Liverwort area mm 70000 30000 30000 60000 50000 50000 10000 10000 20000 20000 40000 40000 70000 60000 0 0

Control 1 Control 1

Control 2 Control 2

267 Treatment Treatment Control 3 Control 3 Fe=Fusarium equiseti, Pw=Phaeodothis

Post Fe Post Fe

Post Pw Post Pw (b) (a) Chapter 6 Figure 6-28.Areaofhealthyliverwortsubjected to(a)oneand(b)twoapplications Fungal antagonists control 3sterilisedcompost withformulation. of fungalinoculum,after elevenweeks. winteri,

Control 1—unsterilised compostonly,control2-sterilised only, 2 Liverwort area mm2 Liverwort area mm 10000 50000 20000 40000 70000 70000 30000 30000 60000 50000 10000 20000 40000 60000 0 0

Control 1 Control 1

Control 2 Control 2

268 Treatment Treatment Control 3 Control 3 Fe=Fusarium equiseti, Pw=Phaeodothis

Post Fe Post Fe

Post Pw Post Pw (a) (b) Chapter 6

Fungal antagonists Chapter 6

(a)

60000 E E

40000

▪gri N t 20000

0 Control 1 Control 2 Control 3 Post Fe Post Pw

Treatment

(b)

60000 N E E to

40000

.o CJ

0 20000

Control 1 Control 2 Control 3 Post Fe Post Pw

Treatment Figure 6-29. Area of dieback of liverwort subjected to (a) one and (b) two applications of fungal inoculum, after eleven weeks. Fe= Fusarium equiseti, Pw= Phaeodothis winter!, Control 1— unsterilised compost only, control 2 - sterilised compost only, control 3 sterilised compost with formulation.

269 Fungal antagonists Chapter 6

Post-emergence P. winteritreatments After 3 weeks, for the post emergence treatment of P. winteri there was a significant difference between the total (v.r.3,3= 11.96, P<0.05) (Figure 6-30) and healthy (v.r.3,3= 11.92, P<0.05) (Figure 6-31) liverwort areas compared to control 1 (unsterilised compost), with the liverworts in the control smaller than those treated with fungal inoculum. This effect was also seen with the post-emergence F. equiseti treatments, strongly suggesting that unsterilised compost is more effective than the fungal treatments at suppressing liverwort growth. There were no significant differences between dieback areas of fungal treatments and controls (v.r.3,3= 0.73, P<0.05) (Figure 6-32). As with the F. equiseti, a second inoculation of one tray from each block was carried out in week 3, and the effect of this was observed in week 5. Total liverwort area was significantly less in control 1 (unsterilised compost) than the fungal treatment with either one (v.r.3,3= 11.82, P<0.05) or two (v.r.3,3= 9.80, P<0.05) fungal applications (Figure 6-30) However, there was no significant difference in healthy liverwort area (Figure 6-31) or dieback area (Figure 6-32) between the fungal treatment and controls with one or two fungal applications. With two fungal applications (Figure 6-32b) the dieback area peaked during week 3, but then decreased in size suggesting that the liverwort had overgrown the area of dieback. This trend was also seen between weeks 3 and 4, but the dieback area was increasing again by week 5 (Figure 6-32a). In the final assessments, at week 11, the total liverwort area of control 1 (unsterilised compost media) was significantly less than the fungal treatment with two fungal applications when data was transformed using natural logs (v.r.3,3= 16.64, P<0.05) (Figure 6-27); there were no significant differences observed with one fungal application. No significant difference was observed between healthy (Figure 6-28) or dieback (Figure 6-29) liverwort areas with either one or two fungal applications; however, the healthy liverwort area was less with one fungal application than two, indicating no benefit in the second application. Dieback area was greater with one and two fungal applications than control 3 (sterilised compost with formulation) suggesting a greater effect with fungal treatment than with formulation alone.

270

Fungal antagonists Chapter 6

45000

40000 -

N E E 35000 - m cu 4 30000 -

25000 -

20000 -

15000 Week 0 Weekl Week 2 Week 3 Week 4 Week 5

Time

45000

40000 -

E E 35000 -

• 30000 - Li O 25000 - :71

20000 -

15000 Week 0 Weekl Week 2 Week 3 Week 4 Week 5

Time

Control 1 - unsterilised compost only 0 Control 2 - sterilised compost only ---v--- Control 3 - sterilised compost with formulation —••-v•—•• Fungal treatment with formulation

Figure 6-30. Growth curve of total liverwort area of liverwort subjected to (a) one and (b) two applications of P. &Wilted.

271 Fungal antagonists Chapter 6

45000

40000 -

N E E 35000 ra

a) ,E3 30000

25000 -

20000 -

15000 Week 0 Week 1 Week 2 Week 3 Week 4 Week 5

Time

45000

40000 -

N E E 35000 - o3a)

a) 30000 -

25000 -

20000 -

15000 Week 0 Week 1 Week 2 Week 3 Week 4 Week 5

Time

Control 1 - unsterilised compost only Control 2 - sterilised compost only Control 3 - sterilised compost with formulation Fungal treatment applied in formulation Figure 6-31. Growth curve of healthy liverwort subjected to (a) one and (b) two applications of P. winteri

272

Fungal antagonists Chapter 6

E 3000 - mI' (13 8 2000 - Ln

113 _10a) 1000 -

Week 0 Week 1 Week 2 Week 3 Week 4

Time

4000

E 3000 - E ra° "fa 8 2000

1000

0 -

Week 0 Week 1 Week 2 Week 3 Week 4 Week 5

Time

Control 1 - unsterilised soil only 0 Control 2 - sterilised soil only Control 3 - sterilised soil with formulation Fungal treatment applied with formulation

Figure 6-32. Growth curve of liverwort dieback subjected to (a) one application (b) two applications of P. winteri

273 Fungal antagonists Chapter 6

6.4 Discussion The laboratory experiments gave an initial indication of the strong effect of P. winteri species, and F. equiseti in particular, on liverwort and provided information that enabled the selection of the fungal species most likely to be successfully in subsequent experiments.

Pre-emergence treatments At week 5 in the pre-emergence treatments for F. equiseti control 2 (sterilised compost) appeared to limit liverwort growth (total and healthy liverwort area) more than either the fungal treatments or other controls, however the removal of apparent outliers reduced the extent of the growth limitation. There was marginally greater liverwort growth in treatments incorporating formulation (control 3 and the fungal treatment). Control 1 (unsterilised compost) also appeared to have a greater growth- limiting effect than the fungal treatment. For P. winteri, however, both control 1 (unsterilised compost) and control 2 (sterilised compost), neither of which contained formulation, had a similar limiting effect on liverwort growth that was greater than the fungal treatment or control 3 (sterilised compost with formulation), both of which did contain formulation. The total and healthy liverwort area of control 2 in the pre- emergence F. equiseti growth cabinets was less than in the corresponding controls in the P. winteri cabinets, although their treatment was the same. Examination of the liverwort areas indicated two replicates of control 2 (sterilised compost) were small at 7,822 and 6,614 mm2 compared with the overall average of 17,043 mm2, suggesting that this effect was not due to sterilisation of the compost. By week 11, however, the fungal treatments (F. equiseti and P. winters) both limited the healthy liverwort area more than any of the controls, and the dieback area was greater in the fungal treatments than any of the controls, suggesting the fungal treatments had an increasing effect on the liverwort, although this effect was not significant.

The effect of the formulation in pre-emergence treatments At week 5 liverwort area was greater in treatments containing formulation (with and without fungal inoculum) than without (controls 1 and 2) for both F. equiseti and P. winter,, however these differences were not significant. Liverwort growth was not limited by the fungal inoculum, with minimal difference between formulation only and formulation with fungal inoculum. Dieback area was greater in the sterilised compost applied with formulation (control 3), for both fungal species, however this effect was

274 Fungal antagonists Chapter 6 not evident in the total liverwort areas. For F. equiiseti only one tray of liverwort exhibited any dieback, distorting the results. By week 11 healthy liverwort area was less in all treatments containing formulation (including the fungal treatments) than controls 1 and 2, although the difference was not significant. When considering both fungal treatments, this could suggest that the formulation either slightly mitigated the liverwort growth-limiting effect of the sterilised and unsterilised compost or promoted liverwort growth.

The effect of unsterilised compost (control 1) in post-emergence treatments In the glasshouse experiment the most significant effects were found in the post- emergence treatments for both fungal species, where the unsterilised compost (control 1) consistently reduced liverwort growth without necessarily causing any dieback. Autoclaving of compost in the post-emergence treatments of this experiment resulted in increased liverwort area, suggesting that sterilisation removed a growth-restricting factor. Production of peat-based compost used by the horticulture industry involves the harvesting, milling and drying of peat, and during this process the aeration is improved and consequently the number and activity of microorganisms is increased (Kuster, 1972). It is unclear from this experiment whether microorganisms in the compost caused the reduction in liverwort growth, whether they provided conditions particularly conducive to fungal health and vigour, or if some other factor contributed to the liverwort's ability to tolerate fungal pathogens was involved. A similar effect was observed in an experiment looking at the effect of mulches and growing media amendments on liverwort and moss infestation, where incorporating 10% (v/v) unsterilised loam in the growing media reduced liverwort infestation, although there was considerable infestation by other weeds (Atwood, 2005). The reisolated fungal species, including F. equiseti, were found on dying liverwort growing in unsterilised loam and its progress had been observed for a number of weeks in the nursery, with the fungal species growing over and killing the liverwort (Atwood, 2004).

The effect of the second fungal application in post-emergence treatments In the post-emergence treatment with two fungal inoculations, for F. equiseti total and healthy liverwort area was reduced by the second fungal application, and had a greater limiting effect on liverwort growth than either control 2 (sterilised compost) or control 3 (sterilised compost with formulation). After 5 weeks the difference in total liverwort area due to control 1 was highly significant with one fungal application but not 275 Fungal antagonists Chapter 6 significant with two as liverwort growth was reduced due to the second inoculation. An increased amount of dieback was observed due to the second fungal application.

For P. winter' there was little observable difference due to the number of fungal inoculum applications suggesting that either the inoculum did not survive or the liverwort vigour did not allow it to proliferate. For P. winter'', the dieback area observed in treatments with two fungal applications peaked in week 3 and reduced over the next two weeks, suggesting overgrowth by liverwort.

By week 11, when comparing fungal species, the total and healthy liverwort area was greater when treated with one fungal application of F. equiseti than P. winter The second fungal application increased the effect of F. equiseti, so that total and healthy liverwort area were both reduced, however for P. winter/the total liverwort area was similar whether treated with one or two applications of P. winter': Control 1 (unsterilised compost) had a greater limiting effect on total and healthy liverwort than any of the other treatments. Dieback area at week 11 was greater with the two fungal applications for F. equiseti, and with one fungal application for P. winter', with the greatest amount of dieback found in control 1.

In the post-emergence treatments the liverwort was already growing vigorously when inoculated, therefore the fungal inoculants would have required greater growth and vigour to produce any observable effect on the liverwort. Different management of the fungal inoculation could have yielded better results, as greater spore densities and longer dew periods are reported to increase biocontrol efficacy with older plants. Naseema et a/ (2001) found that maximum infection of F. equiseti, when used against water hyacinth (Eichhornia crassipes (Mart) Solms), increased from 62.7% to 89.7% by increasing the spore concentration from 109 to 1011 spores m1-1.

Concluding comments In summary, there was some, but variable, evidence for fungal antagonism having some effect on liverwort growth. In pre-emergence tests total and healthy liverwort area were greatest in treatments containing formulation (control 3 or the fungal treatment), which could suggest a growth promoting effect of the formulation. However, by week 11 this effect had disappeared and healthy liverwort area was least,

276 Fungal antagonists Chapter 6 and dieback area greatest in the fungal treatments (both species), showing evidence of some effect of fungal antagonism, although not significant.

In the post-emergence tests total and healthy liverwort areas were least with control 1 (unsterilised compost), sustained throughout the experiment, and which may suggest a growth-inhibiting effect of microflora that were removed by sterilisation of the compost. There is some evidence for a growth-limiting effect of two applications of F. equiseti after 5 weeks and which was still observed in the healthy and dieback areas after 11 weeks.

Ideal growth conditions are similar for liverwort and fungi (Chapter 1), hence there are many close associations reported. Conditions provided for this experiment had to include high humidity to encourage the fungi to persist long enough to establish either in the compost or on the liverwort; these conditions also allowed the liverwort to establish and grow vigorously. Other limitations of this experiment were the size of the growth cabinets which meant that each could accommodate a limited number of replicates. A further treatment where inoculum was applied to unsterilised compost would have identified any effect of the sterilisation procedure on fungal establishment.

Young plants are more vulnerable than older plants (Hallett, 2005; Morin et a/., 1998), therefore it was expected that the pre-emergence treatments would be more successful in controlling liverwort growth, preventing the gemmae from developing into large colonies. Once a dense mat of liverwort has built up it is far more difficult to eradicate, with chemical herbicides often requiring repeated applications.

The appearance of established liverwort successfully treated by fungal antagonists would be an unsightly brown, decomposing mass of liverwort and fungi which would still require removal prior to marketing. A pre-emergence treatment would be easier and more cost-effective to apply as it could be incorporated in the compost rather than applied as a separate task, and as it would prevent liverwort from establishing it would remove the need for attention prior to sale

277 Discussion Chapter 7

Chapter 7 Discussion

The overall aims of this study were twofold: to provide biological knowledge on the liverwort Marchantla polymorpha and to investigate novel control methods using natural products. As part of this, areas were to be identified where current control systems could be improved and where future research could be targeted. In particular the following was achieved:

• Current knowledge of the biology, life cycle and epidemiology of liverwort was reviewed, and aspects where knowledge was lacking were investigated. • The effects of light and temperature on liverwort growth and development were characterised, initially using growth cabinets and finally in a nursery scale experiment using different grades of shading fabric. • With regard to epidemiology, liverwort dispersal via gemmae within various nursery irrigation systems was investigated in depth. • Natural control methods using glucosinolates (GSL) and their hydrolysis products, isothiocyanates (ITCs) were investigated, with bioassays used to compare the herbicidal effects of selected ITCs and herbicides. Additionally, a simplified method

278 Discussion Chapter 7

of collecting ITCs via root exudates of plants grown hydroponically rather than time consuming extraction processes was developed. • Fungal antagonists were both isolated from dying liverworts and purchased from fungal collections and evaluated via laboratory bioassays, with those most effective against liverwort compared in a scaled up glasshouse experiment.

The results of preliminary investigations were applied to nursery scale experiments to explore their use in nursery situation, and their incorporation in an integrated liverwort control program for future research.

7.1 Environmental conditions

The environment experiments show there is a clear effect of light and temperature on liverwort growth (radial expansion), with greater growth under high temperature than low, and under low light than high. High light conditions combined with high temperatures were particularly injurious to liverwort, as evidenced by the abnormal morphology observed. The environment experiments indicated that development is also affected, with gemma cups containing gemmae appearing after 2 weeks in the high light, high temperature (25 °C) treatment, although this would not be a sustainable advantage as these conditions had proven injurious to growth. Gemma cups appeared after 3 weeks in the low light, high temperature (25 °C) and low temperature (15 °C) low light treatments; generally gemma cup development followed the same trend as growth, dry weight and fresh weight. These results were broadly in line with those reported in the literature, where Terui (1981) found that maximum growth occurred at light intensities of 3500 lux (70 pmol m-2 s-1), and growth was inhibited at 5000 lux (120 pmol m-2 s-1); Mache and Loiseaux (1973) also found that higher light levels (5000 and 6000 lux respectively, converting to 100 and 120 pmol M-2 s-1) inhibited liverwort growth. Nachmony-Bascomb and Schwabe (1963) found no effect of light level on thallus growth or gemma cup production, however, the light levels quoted are lower than those used in this study. They did observe an effect of temperature on gemma cup production, with more cups produced in lower temperatures (12 °C than 24 °C).

279 Discussion Chapter 7

The results of the growth cabinet experiments were reflected in the final shading experiment, with greater liverwort growth in low light (400 pmol m-2 s-1), suggesting that liverwort would be more prevalent when protected from full sunlight and with lower temperatures; light shading would raise humidity levels whilst reducing light levels to provide conditions most conducive to liverwort growth.

Methods were developed for the extraction, HPLC analysis and quantification of lunularic acid from liverwort using tissue produced under nursery conditions in short days (November and December) and long days (May and June). Although the results did not prove significant, the overall trend was that that more lunularic acid was present in liverworts grown in long days than short, in agreement with the findings of Gorham (1975). Had the liverworts been grown in strictly controlled conditions the results are likely to have been less variable. However any relevance to growers may not have become apparent as lunularic acid levels in liverwort would not have been measured under nursery conditions.

Growers could use these results to help to manage liverwort infestation. Without compromising plant growth, shade and light levels could be manipulated to provide environmental conditions unfavourable to maximum liverwort growth whilst maintaining plant quality. Grouping of plants according to their environmental growing requirements, would enable light levels to be reduced to provide sub-optimal conditions for liverwort, but tolerable to shade-loving plants (e.g. Hosta spp. and ferns). Conversely high light levels, with no shade could be provided to groups of plants that proliferate under these conditions (e.g. Etyngium spp. and Lavender spp.) again reducing liverwort growth. These extreme environmental conditions would also ensure that liverwort does not proliferate as widely by limiting gametophore development. Light levels could also be manipulated at the compost surface using products such as 'pot toppers' to cover the compost surface, eliminating light and preventing liverwort from establishing directly on the compost. However, there is a labour cost involved in fitting 'pot toppers' to individual pots, and then removing them prior to sale; this would have to be compared with the cost of removing liverwort during the plant growth cycle and/or prior to sale and replacing it with compost, depending on the production system.

280 Discussion Chapter 7

The significance of the lunularic acid work for growers is that by removing liverwort during the hot summer period, with long days and high light intensity, the higher lunularic acid levels would result in lower liverwort growth rates, therefore it would take longer for the liverwort to re-establish; higher summer air temperatures would also promote drying of the compost surface between irrigation cycles, and higher light levels would also be detrimental to liverwort growth.

7.2 Epidemiology

Epidemiology of liverwort was investigated, with specific regard to the dispersal of vegetative propagules (gemmae). An initial experiment comparing dispersal by overhead irrigation (using three different nozzles, two nozzle heights, and four water pressures) found that gemmae can be dispersed further than previously thought (165 cm) by water drops (Equihua, 1987), with the greatest number of gemmae dispersed at 2 and 2.5 bar. The number of gemmae dispersed appeared to be related to water drop size, not number of drops. Mean drop sizes ranged between 161 to 271 pm diameter; maximum number of gemmae were dispersed by water drops 63 to 236 pm mode diameter (172 to 271 pm mean). Practical implications of this would be the use of nozzles which would produce water droplets outside the optimum range for liverwort dispersal. However, whilst manufacturers provide details of droplet sizes for nozzles used for pesticide application this information has not previously been required for irrigation nozzles.

The gemma cup replenishment experiment showed that the more often gemmae are removed, then more are produced to replace them. This suggests that by irrigating plants less often gemma dispersal could be reduced as fewer gemmae would then be produced to replace those dispersed.

These two experiments and the initial environment experiment identified a phenomenon where large numbers of gemmae clump together, making them difficult to count. It was suggested that this could be a life-cycle strategy that ensures gemmae fall close to the parent plant, in environmental conditions known to be conducive to liverwort survival in anticipation of hot, dry weather when sexual reproduction would predominate, enabling long distance spore dispersal (Equihua, 2005). To identify any biological significance to the

281 Discussion Chapter 7 liverwort, the establishment and growth of 'clumps' and 'groups' of liverwort were compared ('Clumps' were numbers of gemmae found naturally in tightly packed masses, and which were dispersed intact; 'Groups' were freely dispersed gemmae obtained from clumps that had been separated by agitating in water and then applied to the compost). The results suggested that dispersed clumps established better and showed greater growth than clumps. However, to ensure enough clumps were available the irrigation regime had been altered from overhead- to sub-irrigation. It was concluded that the 'dispersed clumps' used were more comparable to those found in nature, and if so this could be an advantage to liverwort proliferation. A further experiment comparing these dispersed clumps with individual gemmae placed onto compost would add further insight to this topic of interest.

The final experiment in this section of work brought together the conclusions of the earlier dispersal experiments by investigating dispersal using different irrigation systems (overhead, capillary bed, drip and no irrigation) and cycles (twice daily and 2-daily) under more realistic conditions. The results supported those of the previous experiments, with greater pot coverage of liverwort found in the overhead irrigation than other treatments, and also in the twice daily than 2-daily treatments. Again, overall the number of gemmae dispersed was related to water drop size, not number of drops.

Those who use overhead systems should consolidate irrigation events from many short to fewer longer applications, ensuring the compost surface dries out between each water application, providing conditions unfavourable to liverwort. This group of experiments gives a clear indication that by revising current irrigation practices growers could achieve reduced liverwort spread, with financial benefits and improved water use efficiency. Use of sub or drip irrigation systems, particularly when renewing or upgrading existing equipment, factoring water and reduced labour cost savings (for liverwort removal) into installation estimates, would provide financial benefit to growers. These results complement the conclusions of previous research which found that sub-irrigation can provide substantial cost savings (25-35%) over well designed overhead irrigation, improve water use efficiency, uniformity of water distribution and plant quality (Burgess, 2003b).

282 Discussion Chapter 7

Overhead irrigation was found to use 30 L m2 per week (Briercliffe, 2000). A reduction in the number of irrigation events would provide growers with water cost savings, as would replacing overhead irrigation with capillary bed (12 L m2 per week) or drip (3.7 L m2 per week) irrigation systems. Reduced liverwort infestation and the consequent reduced need for its removal before crop marketing contains implicit labour and media (used to replace removed liverwort) savings.

7.3 Glucosinolates

A literature review suggested that the seed meal of Limnanthes a/ba had a herbicidal effect on liverwort when applied as a mulch (Svenson and Deuel, 2000). This was investigated initially by developing methods for the extraction and quantification of the GSL hydrolysis product 3-methoxybenzyllTC (limnanthin) from L. a/ba seeds, optimising the extraction. The optimum pH for extraction of limnanthin was 5.0, therefore this was used during subsequent experiments. The optimum incubation time for maximum limnanthin extraction was measured under pH 7.0 conditions and found to be 40 mins. The GSL profile of the whole plant was established by extracting and analysing GSLs from the leaves, stems, flowers and roots. Unusually, it was found that only one GSL was produced throughout the whole plant.

Growth bioassays were carried out on liverwort gemmae grown in culture on phytagel media treated with 2-phenylethyl-, benzyl-, 2-propenyl- or 3-methoxybenzyllTC; the herbicides Metazachlor and Lenacil were also used in bioassays to provide a reference point to gauge the effectiveness of the ITCs used. Dose response curves indicated that less 3-methoxybenzyllTC was required than any of the other ITCs tested to obtain an ED50 where 50% of the population was less than 1.5 mm2. Less ITCs than herbicides were required to obtain the ED50, with the exception of 2-propenyllTC. Finally, bioassays carried out using the same ITCs and cress (Lepidium sativum) seeds showed no real herbicidal effect at the concentrations used.

There is no literature on the use of ITCs to control liverwort for comparison; however, some of the ITCs used in this study have been used against higher plants. ITCs produced by soil-incorporated seed meals at different concentrations including L. a/ba, Indian

283 Discussion Chapter 7 mustard (producing 2-propenylITC) and garden cress (producing benzylITC) all reduced wheat seedling emergence, with the greatest effect shown by 2-propenyllTC, although the concentration of ITC produced was not quantified (Vaughn et al, 2006). Bialy (1990) found, using purified ITCs applied at 500 ppm, that 2-phenylethyllTC completely inhibited wheat seedling germination, and 2-propenyllTC showed high activity. Norsworthy and Meehan (2005b) found that concentrations of 10,000 nmol g-1 soil inhibited emergence of Palmer amaranth by over 95%, with LC50 (concentration required to reduce emergence by 50%) values ranging between 32 nmol g-1 soil for phenylethyllTC, 83 nmol g-1 soil for benzylITC and 269 nmol g-1 soil for 2-propenyl ITC. Although the LC50 values used cannot be directly compared to the EC50 values used in this study, both pieces of work found 2- propenyllTC, the most volatile of those used, the least phytotoxic, and phenylethyllTC more phytotoxic than benzylITC.

These results suggest the use of GSLs against liverwort would benefit from further research to confirm the ITC to be used and also to develop an application system that could be tested in glasshouse experiments. Difficulties may occur in further product development. Products could be developed for application as a pre-emergent or soil conditioning treatment similar to those already in use, such as Basamid (a.i. dazomet), which degrades to methyllTC and is applied to soil at least 28 days prior to crop establishment (Whitehead, 2007), however a more appropriate treatment for containers may be pre-emergent application with seed meal incorporated into compost or applied as a mulch, formulated to provide a longer term or slow release herbicidal effects. The results of the bioassays using liverworts and cress seeds suggest that it may be possible to use ITCs at concentrations injurious to liverworts, but non-phytotoxic to higher plants.

The ITC collection system did prove effective when used with ITC standards, however the key objective was to grow plants within a hydroponic system and harvest ITCs from root exudates when they are at peak conditions and this needs to be improved. The hydroponics apparatus would need to be constructed from non-plastic materials to eliminate problems with plasticisers during GC analysis. The apparatus used by Tang and Young (1982) consisted of an inverted glass solvent bottle with the base removed, however this could only support a limited number of plants, and a larger system was

284 Discussion Chapter 7 developed for this study, to increase the number of plants being used and consequently the amount of ITC collected. Further work would establish this as a non-destructive, practical system for ITC collection without the need for complex extraction protocols, which could then be used as a model for other GSL work in addition to liverwort control.

7.4 Fungal antagonists

Potential fungal antagonists were purchased or isolated from dying liverwort specimens, established in culture and used to inoculate liverwort plants in laboratory bioassays. Penicillium velutinum, Trichoderma harzianurn, and Phaeodothis winteri strains CBS 182.58 and 429.96 were weak antagonists of liverwort. However, Fusarium equisetiand P. winter/ strains CBS 551.63, 102466, 163.91 and 102483 successfully infected liverwort to varying degrees in laboratory experiments.

The outcome of these bioassays was used to design glasshouse experiments which compared the growth of liverwort on sterilised compost and treated either pre- or post- emergence with F. equiseti or a mixture of four strains of P. winteri with three controls (sterilised compost, unsterilised compost, sterilised compost with formulation only).

The most significant results were the consistent reduction of liverwort growth in control 1 (unsterilised compost) of the post-emergence treatment, which suggested that sterilisation removed growth-restricting microflora. However, this effect was not fully reflected in the pre-emergence treatments where control 2 (sterilised compost) and control 1 had a similar effect in limiting growth in the P. winter/treatments and control 2 had the greatest limiting effect on liverwort growth in the F. equiseti treatments (at week 5). By week 11, healthy liverwort area (pre-emergence) was less and dieback area greater for both fungal treatments than any of the controls suggesting the fungal antagonists were beginning to have a growth-limiting effect that may potentially have increased had the experiment continued for longer. These results suggest the formulation used for the pre-emergence treatments may have had a growth promoting effect on liverwort; as the formulation included nutrients to support fungal growth. The rates may have been excessive for the purpose and utilised by the liverwort, with a reduced effect as the nutrients depleted.

285 Discussion Chapter 7

Another notable effect was seen in the post-emergence treatments using F. equiseti where an increased amount of dieback was observed, due to the second fungal inoculation, although this did not reduce total or healthy liverwort surface area.

Difficulties with this experiment were posed by using fungal species and liverwort which have common growth environment preferences, such as humidity and temperature levels. Conditions provided for this experiment had to include high humidity to encourage the fungi to persist long enough to establish either in the compost or on the liverwort; these conditions also allowed the liverwort to establish and grow vigorously. In these conditions the liverwort essentially out-competed the fungal antagonists in both the pre-emergence and post emergence treatments.

Established liverwort successfully treated with fungal antagonists is unattractive; it would initially become covered with fungal mycelium, and then become dark and slimy as the mass of liverwort starts to decompose. This would require removal prior to marketing and would therefore be of little benefit to growers, albeit removing the need for chemical application. The most successful control would be applied pre-emergence, preventing liverwort from building up to epidemic proportions. There would be no need to apply herbicides or remove any liverwort remains. Such a product would be easier to manage, as it could be incorporated into the compost with other amendments be maintained in moist conditions conducive to fungal colonisation prior to potting up. No specialist equipment would be required for its application and minimal additional labour costs would be incurred. However, we have no suggestions to make based on the results of this study.

The reisolated fungal species, including F. equiseti, were found on dying liverwort growing in unsterilised loam and had been observed for a number of weeks in the nursery, with the fungal species growing over and killing the liverwort (Atwood, 2004). This species may therefore establish well in less humid conditions; further investigation could provide a method for successfully infecting liverwort with this antagonist. To be commercially viable any mycoherbicide would need to be viable under a range of growing conditions and against liverworts of varying maturities.

286

Discussion Chapter 7

7.5 Implications for nursery management

Grouping of plants by their light and water requirements would help to incorporate these findings into general nursery practice. This would allow less irrigation to be applied to those species requiring less water (e.g. Choisya, Euphorbia characias, Stipa spp.), reducing liverwort dispersal (less overhead irrigation) and establishment (drier compost surface). Light levels could also be applied to groups of plants with similar light requirements (e.g. low light to Hosta spp. and ferns, higher light levels to Eryngium spp., Senecio spp. and Lavender spp)., rather than managing for the most sensitive plant from a group with a range of different light level requirements for healthy growth. Nursery practice depends on issues such as glasshouse size, flexibility of environmental control systems, the number species grown, the number of plants of each species grown and their space requirements. Some, but not all, nurseries do tend to group plants with like requirements where possible. Sub irrigation systems would reduce liverwort dispersal, however installation is expensive and if there is no alternative to overhead irrigation a minimum number of irrigation cycles should be used. Timing of liverwort removal to summer periods with high temperatures and high light intensity, when liverwort growth is reduced due to high levels of lunularic acid should delay liverwort re-establishment.

7.6 Further work

Several aspects of this study would benefit from further study:

• Study of the importance of the sexual stage in both liverwort growth and dispersal • The herbicidal effect of GSL hydrolysis products on liverwort could be investigated under glasshouse conditions, and a viable application system developed. • The collection and identification of isothiocyanates in root exudates requires further work establish protocols to optimise GSL production and collection. • Conditions where fungal antagonists, particularly Fusarium equiseti, would establish on liverwort require further investigation. Sub-optimal environmental conditions for the liverwort would increase the probability of successful control.

287 References Chapter 8

Chapter 8 Reference list

Abe, S., and Ohta, Y. (1983): Lunularic acid in cell-suspension cultures of Marchantia polymorpha. Phytochemistty 22, 1917-1920.

Abe, S., and Ohta, Y. (1984): The concentrations of lunularic acid and prelunularic acid in liverworts. Phytochemistty 23, 1379-1381.

Adams, C. R., Bamford, K. M., and Early, M. P. (1998): Principles of horticulture. Butterworth-Heinemann. Oxford.

Agridor, E. D. (Undated): State of the art irrigation. Current Catalogue issued 2004. Ein Dor Agridor.

Agrios, G. N. (1997): Plant Pathology. Academic Press. London.

Alexopoulos, C. J. (1979): Introductory mycology. Wiley. Chichester.

Al-Khatib, K., and Boydston, R. (1999): Weed control with Brassica green manure crops. In S. S. Narwak (Ed.): Allelopathy Update. Volume 2. Basic and Applied Aspects., Science Publishers Inc., Enfield, New Hampshire, US.

Altland, J. (2003): Weed control in container crops. A guide to effective weed management through preventative measures.: Weed Management in Nursery Crops, Oregon State University North Willamette Research and Extension Centre.

Altland, J. (2004): Personal communication, Oregon State University, Oregon, US.

288 References Chapter 8

Altman, 3., Neate, S., and Rovira, A. D. (1990): Herbicide-pathogen interactions and mycoherbicides as alternative strategies for weed control. In R. E. Hoagland (Ed.): Microbes and Microbial Products as Herbicides, American Chemical Society, Washington, DC, USA.

Anthony, R. E. (1962): Greenhouse culture of Marchantia polymorpha and induction of sexual reproductive structures. Turtox News 40, 2-5.

Aptroot, A. (1995a): A monograph of Didymosphaeria. Studies in Mycology, 4-160.

Aptroot, A. (1995b): Redisposition of some species excluded from Didymosphaeria (Ascomycotina). Nova Hedwigia 60, 325-379.

Arai, Y., Kamikawa, T., and Kubota, T. (1972): A convenient synthesis of lunularic acid. Tetrahedron Letters 13, 1615-1617.

Arai, Y., Kamikawa, T., Kubota, T., Masuda, Y., and Yamamoto, R. (1973): Synthesis and properties of lunularic acid. Phytochemistrj/12, 2279-2282.

Asakawa, Y. (2001): Recent advances in phytochemistry of bryophytes-acetogenins, terpenoids and bis(bibenzyl)s from selected Japanese, Taiwanese, New Zealand, Argentinean and European liverworts. Phytochemistry56, 297-312.

Atwood, 3. G. (2004): Senior horticultural consultant. ADAS. Personal communication.

Atwood, 3. G. (2005): HNS 93C Protected container-grown nursery stock: Chemical and non-chemical screening for moss and liverwort control in liners., Horticultural Development Council, East Mailing.

Atwood, 3. G. (2006): Senior horticultural consultant. ADAS. Personal communication.

Atwood, 3. G. (2007): Senior horticultural consultant. ADAS. Personal communication.

Bartelt, R. 3., and Mikolajczak, K. L. (1989): Toxicity of compounds derived from Limnanthes a/ba seed to fall armyworm (Lepidoptera, Noctuidae) and European corn-borer (Lepidoptera, Pyralidae) larvae. Journal of Economic Entomology 82, 1054-1060.

Barton (nee Frohlich), J. (2004): How good are we at predicting the field host-range of fungal pathogens used for classical biological control of weeds? Biological Control3l, 99-122.

Batchelor, T. (1998): Assessment of alternatives to methyl bromide. Montreal protocol on substances that deplete the ozone layer., United Nations Environment Programme, Nairobi, Kenya.

Batta, Y. A. (2004): Postharvest biological control of apple gray mold by Trichoderma harzianum Rifai formulated in an invert emulsion. Crop Protection 23, 19-26.

Bell, P. R. (1992): Green plants. Their origin and diversity. Cambridge University Press. Cambridge.

Benson-Evans, K. (1964): Physiology of the reproduction of bryophytes. Bryologist67.

289 References Chapter 8

Bialy, Z., Oleszek, W., Lewis, J., and Fenwick, G. R. (1990): Allelopathic potential of glucosinolates (mustard oil glycosides) and their degradation products against wheat. Plant and Soi1129, 227-281.

Blok, W. J., and Bollen, G. J. (1997): Host specificity and vegetative compatibility of Dutch isolates of Fusarium oxysporum f.sp. asparagi. Canadian Journal of Botany75, 383-393.

Bobylev, M. M., Bobyleva, L. I., and Strobel, G. A. (1996): Synthesis and bioactivity of analogs of maculosin, a host-specific phytotoxin produced by Alternaria alternata on spotted knapweed (Centaurea maculosa). J. Agric. Food Chem. 44, 3960-3964.

Bold, H. C. (1938): The nutrition of the sporophyte in the Hepaticae. American Journal of Botany25, 551-557.

Bones, A. M., and Rossiter, J. T. (1996): The myrosinase-glucosinolate system, its organisation and biochemistry. Physiologia Plantarum 97, 194-208.

Booth, C. (1971): The genus Fusarium. Commonwealth Mycological Institute. Kew.

Booth, C. (1978): Fusarium equiseti. Descriptions of pathogenic fungi and bacteria 571, 1-2 20723.

Borek, V., Morra, M. J., Brown, P. D., and McCaffrey, J. P. (1995): Transformation of the glucosinolate-derived allelochemcials allyl isothiocyanate and allylnitrile in soil. J. Agric. Food Chem. 43, 1935-1940.

Boullard, B. (1988): Observations on the coevolution of fungi with hepatics. In K. A. Pirozynski, and D. L. Hawksworth (Eds): Coevolution of fungi with plants and animals, Academic Press, London.

Bracher, F., Krauss, J., and Bornatsch, A. (2000): The total synthesis of lunularic acid and lunularine, a new approach to natural dihydrostilbenes. Natural Product Letters 14, 305-310.

Briercliffe, T. (2000): PC 166. Protected ornamentals: the efficiency of water use in different production systems.

Brodie, H. J. (1951): The splash-cup dispersal mechanism in plants. Canadian Journal of Botany 29, 224-234.

Brough, W. (2007): Personal communication. Horticulture consultant, ADAS.

Brown, J., and Morra, M. J. (2005): Glucosinolate-containing seed meal as a soil amendment to control plant pests, 2000-2002., pp. 1-96, National Renewable Energy Laboratory, Golden, Colorado.

Brown, P. D., and Morra, M. J. (1995): Glucosinolate-containing plant tissues as bioherbicides. Journal of Agricultural Food Chemistry43, 3070-3074.

Brown, P. D., and Morra, M. J. (1999): Weed control with Brassica green manure crops. In S. S. Narwak (Ed.): Allelopathy Update. Volume 2. Basic and Applied Aspects., Science Publishers Inc., Enfield, New Hampshire, US.

290 References Chapter 8

Brown, P. D., Morra, M. J., McCaffrey, J. P., Auld, D. L., and Williams, L. (1991): Allelochemical produced during glucosinolate degradation in soil. Journal of Chemical Ecology 17, 2021-2034.

Burden, D. (2005): Meadowfoam. Agricultural marketing resource centre. Available at: http://www.agmrc.org/agmrc/commodity/biomass/meadowfoam/. Accessed 7/11/2005.

Burgess, C. (2003a): HNS 97. Improving the control and efficiency of water use in container-grown hardy ornamental nursery stock., Horticultural Development Council., East Mailing.

Burgess, C. M. (2003b): HNS 107. Container HNS irrigation: use of capillary matting under protection, pp. 41, Horticultural Development Council, East Mailing.

Carey, D. (2005): Personal Communication, Hadlow College, Hadlow, Tonbridge, Kent. TN11 OAL.

Carter, J. L., and Romine, K. G. (1969): The effects of long and short photoperiods on the rate of growth and gemma cup production in Marchantia polymorpha L. Transactions of the Kansas Academy of Science 72, 98-107.

Cavers, F. (1903a): On asexual reproduction and regeneration in Hepaticae. New Phytologist 2, 121-133.

Cavers, F. (1903b): On asexual reproduction and regeneration in Hepaticae (Continued). New Phytologist 2, 155-164.

Certis (2005): Mogeton: biocidal activity on liverwort, moss and algae. Found at: www.certiseurope.co.uk. Accessed:16/11/05.

Chamberlain, A. C. (1975): 5. The movement of particles in plant communities, pp. 155-203. In J. L. Monteith (Ed.): Vegetation and the atmosphere, Academic Press, London.

Charudattan, R. (1991): The mycoherbicide approach with plant pathogens, pp. 24-57. In D. 0. TeBeest (Ed.): Microbial Control of Weeds, Chapman and Hall, London.

Chase, A. R. (2000): The slippery plant pests: algae, moss and liverwort. Garden Product News.

Chew, F. S. (1988): Biological effects of glucosinolates. In H. G. Cutler (Ed.): Biologically active natural products. Potential use in agriculture, American Chemical Society, Washington.

Clemens, J., Christie, B. C., and Barnaby, C. J. (1991): Watering container plants five different ways, pp. 98-102: Combined Proceedings International Plant Propagators' Society.

Clossais-Besnard, N., and Larher, F. (1991): Physiological role of glucosinolates in Brassica napus. Concentrations and distribution pattern of glucosinolates among plant organs during a complete lifecycle. Journal of the Science of Food and Agriculture 56, 25-38.

291 References Chapter 8

Connick, W. J. (1988): Formulation of living biocontrol agents with alginate, pp. 241- 250. In B. Cross, and H. B. Scher (Eds): Pesticide Formulations: Innovations and Developments. ACS Symposium Series No. 371, American Chemical Society, Washington D.C.

Connick, W. J., Boyette, D. C., and McAlpine, J. R. (1991): Formulation of mycoherbicides using a pasta-like process. Biological Control1, 281-287.

Copping, L. G. (2001): The Biopesticide manual., British Crop Protection Council, Farnham, Surrey.

Daigle, D. J., and Connick, W. J. (1990): Formulation and application technology for microbial weed control, pp. 288-304. In R. E. Hoagland (Ed.): Microbes and microbial products as herbicides, American Chemistry Society, Washington.

Davis, C. E., Franklin, J. B., Shaw, J. A., and Vilgalys, R. (2003): Endophytic Xylaria (Xylariaceae) among liverworts and angiosperms: phylogenetics, distribution, and symbiosis. American Journal of Botany90, 1661-1667.

Dawson, R. M. C., Elliot, D. C., Elliot, W. H., and Johns, K. M. (1969): Data for biochemical research, Oxford University Press, Oxford.

Dayan, F., Romagni, J., Tellez, M., Rimando, A., and Duke, S. (1999): Managing weeds with natural products. Pesticide Outlook10, 185-188. de Forest Heald, F. (1898): Conditions for the germination of the spores of Bryophytes and Pteridophytes. Botanical Gazette 26, 25-45. de Scott, B. (1975): Studies on the genus Eupenicillium Ludwig IV. New species from soil. Mycopathologia 36, 1-27.

Deuel, W. A., and Svenson, S. (1999): Control of clubroot on Chinese mustard and cauliflower using meadowfoam (Limanthes alba) seed meal or screenings. HortScience 34, 473.

Dickson, H. (1932): Polarity and the production of adventitios growing points in Marchantia polymorpha. Annals of Botany 46, 683-701.

Dobbeler, P. (1997): Biodiversity of bryophilous ascomycetes. Biodiversity and Conservation 6, 721-738.

Dobbeler, P. (2002): Microniches occupied by bryophilous ascomycetes. Nova Hedwigia 75, 275-306.

Dornbos, D., D. L. (1990): Natural products phytotoxicity. A bioassay suitable for small quantities of slightly water-soluble compounds. Journal of chemical ecology16, 339-352.

Doyle, W. T. (1973): Non seed plantsL form and functions. MacMillan. London.

Duckett, J. G., Burch, J., Fletcher, P. W., Matcham, H. W., Read, D. J., Russell, A. J., and Pressel, S. (2004): In vitro cultivation of bryophytes: a review of practicalities, problems, progress and promise. Journal of Btyology26, 3-20.

292 References Chapter 8

Duckett, J. G., Ligrone, R., Andrews, N., and Renzaglia, K. S. (2000): The enigma of pegged and smooth rhizoids in marchantialean hepatics; a functional explanation. American Journal of Botany87 , 6.

Duckett, J. G., and Read, D. J. (1995): Ericoid mycorrhizas and rhizoid-ascomycete associations in liverworts share the same mycobiont: isolation of the partners and resynthesis of the associations in vitro. New Phytologist129, 439-447.

Duke, S. 0., Dayan, F. E., Rimando, A. M., Schrader, K. K., Aliotta, G., Oliva, A., and Romagni, J. G. (2002): Chemicals from nature for weed management. Weed Science 50, 138-151.

Duke, S. 0., Romagni, J. G., and Dayan, F. E. (2000): Natural products as sources for new mechanisms of herbicidal action. Crop Protection 19, 583-589.

Durand, E. J. (1908): The development of the sexual organs and sporogonium of Marchantia polymorpha. Bulletin of the Torrey Botanical Club 35, 321-335.

Eicher, T., Tiefensee, K., and Pick, R. (1988): Synthesis of Bryophyte components .1. New syntheses of lunularic acid and some of its derivatives. Synthesis-Stuttgart, 525-529.

Elzein, A., Kroschel, J., and Muller-Stover, D. (2004): Effects of inoculum type and propagule concentration on shelf life of pesta formulations containing Fusarium oxysporum Foxy 2, a potential mycoherbicide agent for Striga spp. Biological Control30, 203-211.

Encore Technologies (1999a): Collego. Post emergence herbicide biological weed control agent. Encore Technologies. Available at: www.encoretechllc.com. Accessed 16/11/06.

Encore Technologies (1999b): deVine. Biological herbicide. Encore technologies. Available at: www.encoretechllc.com. Accessed 16/11/06.

Equihua, C. (1987): Splash cup dispersal of gemmae in the liverwort Marchantia polymorpha. Oyptogamie: Bryologie et lichenologie 8, 199-217.

Equihua, C. (2005): Personal contact, Instituto de Ecologia, UNAM, Mexico.

Fahey, J. W., Zalcmann, A. T., and Talalay, P. (2001): The chemical diversity and distribution of glucosinolates and isothiocyanates among plants. Phytochemistry 56, 5-51.

Felix, H. (1988): Fungi on Bryophytes, a Review, pp. 239-269: Botanica Helvetica.

Friederich, S., Maier, U. H., Deus-Neumann, B., Asakawa, Y., and Zenk, M. H. (1999a): Biosynthesis of cyclic bis(bibenzyls) in Marchantia polymorpha. Phytochemistry 50, 589-598.

Friederich, S., Rueffer, M., Asakawa, Y., and Zenk, M. H. (1999b): Cytochromes P-450 catalyze the formation of marchantins A and C in Marchantia polymorpha. Phytochemistty52, 1195-1202.

293 References Chapter 8

Fujisawa, M., Nakayama, S., Nishio, T., Fujishita, M., Hayashi, K., Ishizaki, K., Kajikawa, M., Yamato Katsuyuki, T., and Fukuzawa, H. (2003): Evolution of ribosomal DNA unit on the X chromosome independent of autosomal units in the liverwort Marchantia polymorpha. Chromosome Research 11, 695-703.

Fiirstner, A., and Nikolakis, K. (1996): Ethynylation of aryl halides by a modified suzuki reaction: Application to the syntheses of combretastatin A-4, A-5 and lunularic acid. Liebigs Annalen, 2107-2113.

Galloway, A. (2004): Stay dry to deter the green invaders, pp. 21-22: Horticulture Week.

Gamborg, 0. L. (1975): Callus and cell culture. In 0. L. Gamborg, and L. R. Wetter (Eds): Plant tissue culture methods, National Research Council of Canada, Saskatoon.

Gamliel, A., and Stapleton, J. J. (1993): Characterisation of antifungal volatile compounds evolved from solarised soil amended with cabbage residues. Phytopathology 83, 899-905.

Geagea, L., Huber, L., and Sache, I. (1999): Dry-dispersal and rain-splash of brown (Puccinia recondita f.sp. tritici and yellow P. stdiformis) rust spores from infected wheat leaves exposed to simulated raindrops. Plant Pathology48, 472- 482.

Gerstl, Z., Mingelgrin, U., and Yaron, B. (1971): Behaviour of Vapam and methyl isothiocyanate in soils. Soil Science Society of America .7ournal 41, 545-548.

Gil, V., and MacLeod, A. J. (1980): The effects of pH on glucosinolate degradation by a thioclucoside glucohydrolase preparation. Phytochemist0/19, 2547-2551.

Gorham, J. (1975): Some aspects of the distribution, metabolism and physiological role of lunularic acid in liverworts, PhD Thesis, Wye College, University of London.

Gorham, J. (1977): Lunularic acid and related compounds in liverworts, algae and Hydrangea. Phytochemisby16, 249-253.

Gorham, J. (1978): Effect of lunularic acid analogues on liverwort growth and IAA oxidation. Phytochemisby17, 99-105.

Gorham, 1, and Coughlan, S. J. (1980): Inhibition of photosynthesis by stilbenoids. Phytochemistty19, 2059-2064.

Greaves, M. P. (1996): Chapter 13. Microbial herbicides - factors in development, pp. 444-467. In L. G. Copping (Ed.): Crop Protection Agents from Nature: Natural Products and Analogues, The Royal Society of Chemistry.

Haines, J. (2006): Personal communication. Teaching technician, Imperial College London, Wye.

Hallett, S. G. (2005): Where are the bioherbicides? Weed Science 53, 404-415.

294 References Chapter 8

Hashimoto, T., Tori, M., and Asakawa, Y. (1988): A highly efficient preparation of lunularic acid and some biological activities of stilbene and dihydrostilbene derivatives. Phytochemistry27, 109-113.

Heaney, R. K., Spinks, E. A., Hanlet, B., and Fenwick, G. R. (1986): Technical bulletin: analysis of glucosinolates in rapeseed. Norwich, UL:AFRC, Food Research Institute.

Hedger, D. K., Taylor, J., Montague, M. J., and Schindler, B. (1972): The relationship between some environmental parameters and the growth of Marchantia gemmalings. Bryologist75.

Holliday, P. (1998): A dictionary of plant pathology. Cambridge University Press. Cambridge.

Hooker, W. J., Walker, J. C., and Smith, F. G. (1943): Toxicity of 13 -phenetyl isothiocyanate to certain fungi. American Journal of Botany30, 632-637.

Hopkins, W., G. (1999): Introduction to plant physiology. John Wiley & Sons Inc. Chichester.

Huneck, S., and Schreiber, K. (1977): Synthese der lunularsaure. Phytochemistry16, 1013-1016.

IENICA (2006): Meadowfoam. IENICA. Available at :http://wvvw.ienica.net/crops/meadowfoam.htm. Accessed: 37/09/206.

Index Fungorum (2004): Index Fungorum. Custodians: Cabi Bioscience, CBS and Landcare Research. Found at: www.indexfungorum.org. Accessed 2004.

Ingold, C. T. (1973): The biology of fungi. Hutchinson Educational Ltd. London.

Ingold, C. T. (1974): Spore liberation in ayptogams. Oxford University Press. Oxford.

Ishida, A., Ookubo, K., and Ono, K. (1987): Formation of hydrogen peroxide by NAD(P)H oxidation with isolated cell wall-associated peroxidase from cultured liverwort cells, Marchantia po/ymorpha L. Plant Cell Physiology 28, 723-726.

Ishizaki, K., Shimizu-Ueda, Y., Okada, S., Yamamoto, M., Fujisawa, M., Yamato Katsuyuki, T., Fukuzawa, H., and Ohyama, K. (2002): Multicopy genes uniquely amplified in the Y chromosome-specific repeats of the liverwort Marchantia polymorpha. Nucleic Acids Research 30, 4675-4681.

James, D., and Rossiter, J. T. (1991): Development and characteristics of myrosinase in Brassica napus during early seedling growth. Physiol Plant82.

Josefsson, E. (1970): Glucosinolate content and amino acid composition of rapeseed (Brassica napus) meal as affected by sulphur and nitrogen nutrition. Journal of the Science of Food and Agriculture 21, 98-103.

Ju, H., Hak -Yoon, Chong, C., and Bible, B. B. (1982): Influence of boron nutrition on glucosinolates and reducing sugars of turnip. Canadian Journal of Plant Science 62.

295 References Chapter 8

Kandasamy, K., Mantell, S., Blake, 3., Newton, D., Harris, R., and Farris, S. (1996): Plant Tissue Culture laboratory Practices. The UAPS Manual . Unit for Advanced Propagation Systems. Wye College Press. Wye.

Kasama, Y. (2003): Uptake and metabolism of 3,4 dinitrotoluene in cress (Lepidium sativum). pp. 1-90, Imperial College London.

Khan, Z. R., Hassanali, A., Overholt, W., Khamis, T. M., Hooper, A. M., Pickett, 3. A., Wadhams, L. 3., and Woodcock, C. M. (2002): Control of witchweed Striga hermonthica by intercropping with Desmodium spp., and the mechanism defined as allelopathic. Journal of Chemical Ecology 28, 1871-1885.

Kirk, P. M., and Spooner, B. M. (1984): Ascomycetes in :an account of the fungi of Aran, Gigha and Kintyre. Kew Bulletin 38, 503-597.

Kirkegaard, 3. A. (1998): Biofumigation potential of brassicas. Plant and soil 201, 71- 89.

Kirkegaard, 3. A., Gardner, P. A., Desmarchelier, 3. M., and Angus, 3. F. (1993): Biofumigation - using Brassica species to control pests and diseases in horticulture and agriculture. 9th Australian Research Assembly on Brassicas, pp. 77-82.

Kirkegaard, 3. A., Sarwar, M., Wong, P. T. W., Mead, A., Howe, G., and Newell, M. (2000): Field studies on the biofumigation of take-all by Brassica break crops. Australian journal of agricultural research 51, 445-456.

Kirkegaard, 3. A., Wong, P. T. W., Desmarchelier, 3. M., and Sarwar, M. (1996): Suppression of soil-borne cereal pathogens and inhibition of wheat germination by mustard seed meal. Proceedings of the 8th Australian Agronomy Conference. University of Southern Queensland, Toowoomba. Available at: http://www.regional.org.au/au/gcirc/2/334.htm. Accessed: 5 May 2007.

Kroschel, 3., Hundt, A., and Sauerborn, 3. (1996): Pathogenicity of fungi collected in northern Ghana to Stn:ga hermonthica. Weed Research 36, 515-520.

Kubicek, C. P., and Harman, G. E. (1998): Trichoderma and Gliocladium. Volume 1. Basic biology, taxonomy and genetics. Taylor and Francis. London.

Kuster, E. (1972): ISHS Acta Horticulturae 26: III Symposium on Peat in Horticulture. Microbiology of peat. Acta Horticulturae (ISHS) 26, 23-28.

Lawrence, E. (2000): Henderson's dictionary of biological terms. Prentice Hall. London.

Ligrone, R., and Lopes, C. (1989): Cytology and development of a mycorrhiza-like infection in the gametophyte of Conocephalum conicum (L.) Dum. (Marchantiales, Hepatophyta). New Phyto1111, 423-433.

Louda, S. M., Farris, M. A., and Blua, M. J. (1987): Variation in methylglucosinolate and insect damage to Cleome serrulata (Capparaceae) along a natural soil moisture gradient. Journal of Chemical Ecology13, 569-591.

Lovett, j. (2005): Hans Molisch' legacy. 4th World Congress on Allelopathy.

296 References Chapter 8

Mache, R., and Loiseaux, S. (1973): Light saturation of growth and photosynthesis of the shade plant Marchantia polymorpha. J. Cell Sci. 12, 391-401.

Madden, L. V. (1997): Effects of rain on splash dispersal of fungal pathogens. Canadian Journal of Plant Pathology-Revue Canadienne De Phytopathologie 19, 225-230.

Madden, L. V., Yang, X. S., and Wilson, L. L. (1996): Effects of rain intensity on splash dispersal of Colletotrichum acutatum. Phytopathology86, 864-874.

Maher, M. J., Kirkland, C., Hunter, A., and M, P. (1996): The effect of irrigation system and peat type on the production of Hebe 'Mrs Winder'. International symposium on growing media and plant nutrition. Acta Hort. 450., pp. 199-204.

Manici, L. M., Lazzeri, L., and Palmieri, S. (1997): In vitro fungitoxic activity of some glucosinolates and their enzyme-derived products toward plant pathogenic fungi. J. Agric. Food Chem. 45, 2768-2773.

Mansfield, J. (2005): Personal communication. Imperial College London at Wye.

Martin, J. T., and Juniper, B. E. (1970): The cuticle of plants R & R Clark. Edinburgh.

Martinez, M. (2006): Personal communication., Natural Plant Products Inc., Salem, Oregon, US.

Matthiessen, J., Kirkegaard, J., and Morra, M. (2001): Biofumigtion for soil-borne pest and disease suppression - current status and future directions. South Australian Research and Development Institute. Available at:http://www.sardi.sa.gov.au/pages/horticulture/pathology/hort pn biosoilstat us.htm:sectID=336&tempID=111. Accessed 17.2.2004.

Matthiessen, J., and Kirkegaard, J. A. (2006): Biofumigation and enhanced biodegradation: opportunity and challenge in soilborne pest and disease management. Critical reviews in plant sciences 25, 235-265.

Matthiessen, J., and Shackleton, M. (2002): Biofumigation, CSIRO Entomology.

Mattner, S. W. (2001): The role of allelopathy in the control of soilborne diseases. Second Australasian Soilborne Diseases Symposium.

McConaha, M. (1941): Ventral structures effecting capillarity in the Marchantiales. American Journal of Botany 28, 301-306.

Miller, M. W., Garber, E. D., and Voth, P. D. (1962): Nutritionally deficient mutants of Marchantia polymorpha induced by x-rays. Botanical Gazette 124, 94-102.

Miller, R. W., Daxenbichler, M. E., and Earle, F. R. (1964): Search for new industrial oils. VIII. The genus Limnanthes. The Journal of the American Oil Chemists Society41, 167-168.

Minibayeva, F., and Beckett, R. P. (2001): High rates of extracellular superoxide production in bryophytes and lichens, and an oxidative burst in response to rehydration following desiccation. New Phytologist 152, 333-341.

297 References Chapter 8

Mithen, R. F. (2001): Glucosinolates and their degradation products, pp. 213-262: Advances in Botanical Research, Vol 35.

Mompert, A., and Orlikoski, L. B. (2000): Control of algae, mosses and liverworts by Agrigerm 2000 in ornamental plants under covering and in field. 52nd International Symposium on Crop Protection, pp. 151-156.

Morin, L., Gianotti, A., Barker, R., and Johnston, P. (1998): Favourable conditions for the bioherbicide candidate Fusarium turnidllum to infect and cause severe disease on gorse (Illex europaeus) in a controlled environment. Biocontrol Science and Technology 8, 301-311.

Murov, S. L. (2007): Properties and toxicities of organic solvents. Modesto Junior College, Modesto, Ca, USA. Available at: http://virtual.yosemite.cc.ca.us/smurov/orgsoltab.htm. Accessed: 24/2/07.

Nachmony-Bascomb, S., and Schwabe, W. W. (1961): Growth and dormancy in Lunularia Cruciata (L.) Dum.:1. resumption of growth and its continuation. Journal of Experimental Botany 14, 153-71.

Nachmony-Bascomb, S., and Schwabe, W. W. (1963): Growth and dormancy in Lunularia Cruciata (L.) Dum.:1. resumption of growth and its continuation. Journal of Experimental Botany 14, 153-71.

Nakayama, T., Fukushi, Y., Mizutani, J., and Tahara, S. (1996): Inhibiting effects of lunularic acid analogs on the growth of liverwort, watercress, and timothy grass. Bioscience Biotechnology and Biochemistry 60, 862-865.

Nakazato, T., Kadota, A., and Wada, M. (1999): Photoinduction of spore germination in Marchantia po/ymorpha L. is mediated by photosynthesis. Plant and Cell Physiology 40, 1014-1020.

Naseema, A., Praveena, R., Balakrishnan, S., and Peethambaran, C. K. (2001): Management of water hyacinth [Eichhornia crassipes (Mart) Solms] with fungal pathogens. BCPC International Conference on Weeds, pp. 263-268.

Nelson, B. D. (2001): Stachybotrys chartarum: the toxic indoor mold. APSnet. American Phytopathological Society. Available at: http://www.apsnet.org/online/feature/stachybotrvs/. Accessed 25 August 2006.

Newby, A., Altland, J., Gilliam, C., Fare, D., and Wehtje, G. (2004): Liverwort control in container grown nursery grown crops. Southern Nursery Association Research Conference, pp. 396-400.

Nitao, J. K., Meyer, S. L. F., Schmidt, W., F., Fettinger, J., C., and Chitwood, D., J. (2001): Nematode-antagonistic trichothecenes from Fusarium equiseti. Journal of Chemical Ecology 27, 859-869.

Norsworthy, J. K., Malik, M. S., Jha, P., and Oliveira, M. J. (2006): Effects of isothiocyanates on purple (Cyperus rotundus L.) and yellow nutsedge (Cyperus esculentus L.). Weed Biology and Management 6, 131-138.

298 References Chapter 8

Norsworthy, J. K., and Meehan, J. T. I. (2005a): Herbicidal activity of eight isothiocyanates on Texas panicum (Panicum texanum), large crabgrass (Digitaria sanguinalis), and sicklepod (Senna obtusifolia). Weed Science 53, 515-520.

Norsworthy, J. K., and Meehan, J. T. I. (2005b): Use of isothiocyanates for suppression of Palmer amaranth (Amarantus palmed), pitted morning glory (Ipomoea lacunosa), and yellow nutsedge (Cyperus esculentus). Weed Science 53, 884- 890.

Novak, T. (1999): OSU develops new product using meadowfoam. OSU News. Oregon State University. Found at: http://oregonstate.edu/dept/ncs/newsarch/1999/Jun99/morefoam .htm Accessed: 27/09/2006.

Oelke, E. A., Oplinger, E. S., Hanson, C. V., and Kelling, K. A. (1990): Meadowfoam. Wisconsin Corn Agronomy. Available at: http://corn.agronomy.wisc.edu/AlternativeCrops/Meadowfoam.htm. Accessed 7/11/2005.

O'Hanlon, M. E. (1926): Germination of Spores and Early Stages in Development of Gametophyte of Marchantia polymorpha. Botanical Gazette 82, 215-222.

Ohta, Y., Abe, S., Komura, H., and Kobayashi, M. (1983): Prelunularic acid, a probable immediate precursor of lunularic acid - 1st example of a prearomatic intermediate in the phenylpropanoid polymalonate pathway. Journal of the American Chemical Society105, 4480-4481.

Ohta, Y., Abe, S., Komura, H., and Kobayashi, M. (1984): Prelunularic acid in liverworts. Phytochemistty23, 1607-1609.

Okada, S., Fujisawa, M., Sone, T., Nakayama, S., Nishiyama, R., Takenaka, M., Yamaoka, S., Sakaida, M., Kono, K., Masayoshi, T., Yamato Katsuyuki, T., Fukuzawa, H., Brennicke, A., and Ohyama, K. (2000): Construction of male and female PAC genomic libraries suitable for identification of y-chromosome- specific clones from the liverwort Marchantia polymorpha. The Plant Journal 24, 421-428.

Ono, K., Ohyama, K., and Gamborg, 0. L. (1979): Regeneration of the liverwort Marchantia polymorpha L. from protoplasts isolated from cell suspension culture. Plant Science Letters 14, 225-229.

Paleos (2005): The carboniferous. Paleos: The trace of life on earth. Available at:http://www.palaeos.com/Paleozoic/Carboniferous/Carboniferous.htm. Accessed: 14/02/05.

Pesticide Safety Directorate (2006): Pesticide Safety Directorate. DEFRA. Crown Copyright. Available at: http://www.pesticides.gov.uk/aa registration.asp?id=50 Accessed: 26/9/06.

Phillips, N. J., Goodwin, J. T., Fraiman, A., Cole, R. J., and DJ, L. (1989): Characterization of the Fusarium toxin equisetin: The use of phenyl boronates

299 References Chapter 8

in strucure assignment. Journal of the American Chemical Society111, 8223- 8231.

Pope, D. F., Thompson, A. C., and Cole, A. W. (1985): Phytotoxicity of root exudates and leaf extacts of nine plant species. In A. C. Thompson (Ed.): The chemistry of allelopathy. Biochemical interactions among plants, ACS Symposium series 268., American Chemical Society, Washington D.C.

Proctor, M. C. F. (1984): Structure and ecological adaptation. In A. F. Dyer, and J. G. Duckett (Eds): The Experimental Biology of Bryophytes, Academic Press, London.

Pryce, R. J. (1971a): Biosynthesis of lunularic acid--a dihydrostilbene endogenous growth inhibitor of liverworts. Phytochemistry10, 2679-2685.

Pryce, R. J. (1971b): Lunularic acid, a common endogenous growth inhibitor of liverworts. Planta 97, 354-357.

Pryce, R. J. (1972): Metabolism of lunularic acid to a new plant stilbene by lunularia cruciata. Phytochemistryll, 1355-1364.

Raven, P. H., Evert, R. F., and Eichorn, S. E. (1999): Biology of plants Freeman & Co./ Worth Publishers. New York.

Reynolds, K. M., Madden, L. V., Reichard, D. L., and Ellis, M. A. (1989): Splash dispersal of Phytophthora cactorum from Infected strawberry fruit by simulated canopy drip. Phytopathology79, 425-432.

Rice, E. L. (1979): Allelopathy - An update. Bot. Rev. 45, 15-109.

Richardson, R. (2003): Greenhouse alert, liverwort (Marchantia polymorpha). Michigan State University.

Richardson, R. J. (2005): Liverwort control in containerized hosta with selected herbicide treatments. Proceedings of the 59th Annual Meeting of the North Eastern Weed Science Society.

Rook, E. J. S. (1999): Marchantia polymorpha. Liverwort. Available at: http://www.rook.org/earl/bwca/nature/moss/marchantia.html. Accessed: 21/07/03.

Rosa, E. A. S., Heaney, R. K., Fenwick, G. R., and Portas, C. A. M. (1997): Glucosinolates in crop plants, pp. 99-217. In J. Janick (Ed.): Horticultural Reviews. Volume .19., John Wiley & Sons, Inc., New York.

Round, F. E. (1969): Introduction to the lower plants. Butterworth. London.

Sandrin, T. R., TeBeest, D. 0., and Weidemann, G. J. (2003): Soybean and sunflower oils increase the infectivity of Colletotrichum gloeosporioidesf. sp. aeschynomene to Northern Jointvetch. Biological Control26, 244-252.

Sang, J. P., Minchinton, I. R., Johnstone, P. K., and Truscott, R. J. W. (1984): Glucosinolate profiles in the seed, root and leaf tissue of cabbage, mustard, rapeseed, radish and swede. Canadian Journal of Plant Science 64, 77-93.

300 References Chapter 8

Santo, G. (1999): Organic amendment for nematode management on potatoes. Agrichemical and Environmental News.

Sarwar, M., and Kirkegaard, 3. A. (1998): Biofumigation potential of brassicas. II. Effect of environment and ontogeny on glucosinolate production and implications for screening. Plant and Soil 201, 91-101.

Sarwar, M., Kirkegaard, 3. A., Wong, P. T. W., and Desmarchelier, J. M. (1998): Biofumigation potential of brassicas. III. In vitro toxicity of isothiocyanates to soil-borne fungal pathogens. Plant and SoiI201, 103-112.

Schneider, M. 3., Voth, P. D., and Troxler, R. F. (1967a): Methods of propagating Bryophyte plants, tissues, and propagules. Botanical Gazette 128, 169-174.

Schonherr, 3., and Ziegler, H. (1975): Hydrophobic cuticular ledges prevent water entering the air pores of liverwort thalli. Planta 124, 51-60.

Schwabe, W. W. (1951): Physiological studies in plant nutrition. XVI. The mineral nutrition of bracken. Part 1. Prothallial culture and the effects of phosphorus and potassium supply on leaf production in the sporophyte. Annals of Botany 15, 417-447.

Schwabe, W. W., and Nachmony-Bascomb, S. (1963): Growth and dormancy in Lunularia cruciata (L.) Dum. ii. the response to daylength and temperature. Journal of Experimental Botany14, 353-378.

Scott, M., and Hutchinson, D. (2001): HNS 93. Nursery stock propagation: moss, liverwort and slime control, pp. 1-82, Horticultural Development Council.

Sigma-Aldrich (2006): Amberlite XAD polymeric resins. Product information., Sigma- Aldrich, Saint Louis, USA.

Sloover, J. d. (2001): Mniaecia nivea and M. jungermanniae (Ascomycota: Leotiales) reported from Belgium. Lejeunia, 13.

Smelt, 3., H., and Leistra, M. (1974): Conversion of metham-sodium to methyl isothiocyanate and basic data on the behaviour of methyl isothiocyanate in soil. Pesticide Science 5, 401-407.

Smith, G. (1955): Cryptogamic botany. Bryophytes and Pteridophytes. MacGraw-Hill Book Company Ltd. London.

Spencer, G. F., and Daxenbichler, M. E. (1980): Gas chromatography - mass spectrometry of nitriles, isothiocyanates and oxazolidiinethiones. Journal of the Science of Food and Agriculture 31, 359-367.

Svenson, S., and Deuel, W. (2000): Using quinoclamine and meadowfoam seed meal to control liverworts in containers. Southern Nursery Association Research Conference, pp. 391-393.

Svenson, S., and Deuel, W. (2001): Using AlbaGro for Marchantia control. SNA Research Conference, pp. 443-444.

301 References Chapter 8

Svenson, S., Paxson, J., and Sanford, K. (2001): Composts and shading influence Marchantia infestations in container grown nursery crops. SNA research conference, pp. 445-447.

Svenson, S., Smith, B., and Briggs, B. (1997): Controlling liverworts and moss in nursery production. Combined Proceedings International Plant Propagators' Sociely47, 414-422.

Svenson, S. E. (1997): Supression of liverwort growth using cinnamic aldehyde. Southern Nursery Association Research Conference, pp. 494-496.

Svenson, S. E. (1998): Suppression of liverwort growth in containers using irrigation, mulches, fertilizers and herbicides. Southern Nursery Association Research Conference, pp. 396-402.

Takenaka, M., Yamaoka, S., Hanajiri, T., Shimizu-Ueda, Y., Yamato Katsuyuki, T., Fukuzawa, H., and Ohyama, K. (2000): Direct transformation and plant regeneration of the haploid liverwort Marchantia polymorpha L. Transgenic Research 9, 179-185.

Tang, C.-S., and Young, C.-C. (1982): Collection and identification of allelopathic compounds from the undisturbed root system of bigalta limpograss (Hemarthria altissima). Plant Physiol. 69, 155-160.

Taren, N. (1958): Factors regulating the initial development of gemmae in Marchantia polymorpha. The Biyologist 61, 191-204.

Teasdale, 3. R., and Taylorson, R. B. (1986): Weed seed response to methyl isothiocyanate and metham. Weed Science 34, 520-524.

Templeton, G. E. (1990): Weed control with pathogens. Future needs and directions. In R. E. Hoagland (Ed.): Microbes and Microbial Products as Herbicides, American Chemical Society, Washington, DC, USA.

Templeton, G. E., Smith, R. J. J., and TeBeest, D. 0. (1986): Progress and potential of weed control with mycoherbicides. Review of Weed Science 2, 1-14.

Templeton, G. E., TeBeest, D. 0., and Smith, R. 3. 3. (1979): Biological weed control with mycoherbicides. Annual Review of Phytopathology17, 301-310.

Terui, K. (1981): Growth and gemma cup formation in relation to archegoniophore protrusion in Marchantia polymorpha. Annual Report of the Faculty of Education, Iwate University40, 19-28.

The Food Commission (2006): Methyl bromide passes its sell-by date. The Food Commission. London. Found at: http://www.foodcomm.org.uk/latest methyl may 06.htm Accessed: 15/09/2006.

Thiel, P. G., Marasas, W. F. 0., Sydenham, E. W., Shephard, G. S., Gelderblom, W. C. A., and Nieuwenhuis, J. J. (1991): Survey of fumonisin production by Fusarium species. Applied and Environmental Microbiology 57, 1089-1093.

302 References Chapter 8

Thies, W. (1988): Isolation of sinigrin and glucotropaeolin from cruciferous seeds. Felt Wissenschaft Technologie-Fat Science Technology 90, 311-314.

Tomlin, C. D. S. (2000): The Pesticide Manual, British Crop Protection Council, Farnham, Surrey.

Tsanuo, M. K., Hassanali, A., Hooper, A. M., Khan, Z., Kaberia, F., Pickett, 3. A., and Wadhams, L. 3. (2003): Isoflavanones from the allelopathic aqueous root exudate of Desmodium uncinatum. Phytochemistry64, 265-273.

United Nations Environment Programme (2006): Ozone-friendly agricultural products - goal of new global initiative. UNEP news release 2006/23. Available at httoi/www.unep.org/Documents.Multilingual/Defaultasp?DocumentID=475&A rticleID=5247&l=en. Accessed: 25/01/07.

Valio, I. F. M. (1969): Promotion and inhibition of growth in Lunularia cruciata (L) Dum, PhD Thesis. Wye College, University of London.

Valio, I. F. M., Burdon, R. S., and Schwabe, W. W. (1969): New natural growth inhibitor in the liverwort Lunularia cruciata (L.) Dum. Nature 223, 1176-1178.

Valio, I. F. M., and Schwabe, W. W. (1969): Growth and dormancy in Lunularia cruciata (L.) Dum. iv. light and temperature control of rhizoid formation in gemmae. Journal of Experimental Botany 20, 615-628.

Valio, I. F. M., and Schwabe, W. W. (1970): Growth and dormancy in Lunularia cruciata (L.) Dum. vii. the isolation and bioassay of lunularic acid. Journal of Experimental Botany 21, 138-150. van Driesche, R. G., and Bellows, T. S. (1996): Biological control, Chapman & Hall, London.

Vaughn, S. F., and Berhow, M. A. (2005): Glucosinolate hydrolysis products from various plant sources: pH effects, isolation, and purification. Industrial crops and products 21, 193-202.

Vaughn, S. F., Boydston, R. A., and Mallory-Smith, C. A. (1996): Isolation and identification of (3-methoxyphenyl) acetonitrile as a phytotoxin from meadowfoam (Limnanthes alba) seedmeal. Journal of Chemical Ecology 22, 1939-1949.

Vaughn, S. F., Palmquist, D. E., Duval, S. M., and Berhow, M. A. (2006): Herbicidal activity of glucosinolate-containing seedmeals. Weed Science 54, 743-748.

Vea, E., and Palmer, C. (2006): Quinoclamine crop safety. IR-4 Ornamental Horticulture Program. Summary report for 2005. Available at:htto://www.ir4.rutgers.edu/ornamental/SummarvReports/QuinoclamineData Summary2005.pdf. Accessed 18/10/06.

Verkley, G. 3. M. (2004): Personal communication, Mycologist, Fungal Biodiversity Centre, Centraalbureau voor Schimmelcultures.

303 References Chapter 8

Verkley, G. J. M., VanderAa, H. A., and DeCock, G. W. (1997a): Bryoscyphus atromarginatus spec nov (Leotiaceae), a new ascomycete parasitizing the thallus of Marchantia polymorpha. Persoonia 16, 383-387.

Verkley, G. J. M., VanderAa, H. A., and DeCock, G. W. (1997b): Bryoscyphus atromarginatus spec. nov. (Leotiaceae), a new ascomycete parasitizing the thallus of Marchantia polymorpha. Persoonia 16, 383-387.

Vestberg, M., Kukkonen, S., Saari, K., Parikka, P., Huttunen, 3., Tainio, L., Devos, N., Weekers, F., Kevers, C., and Thonart, P. (2004): Microbial inoculation for improving the growth and health of micropropagated strawberry. Applied Soil Ecology 27, 243-258.

Voth, P. D. (1941): Gemmae cup production in Marchantia polymorpha and its response to calcium deficiency and supply of other nutrients. Botanical Gazette 103, 310-325.

Voth, P. D. (1943): Effects of nutrient solution concentration on the growth of Marchantia polymorpha. Botanical Gazette 104, 591-601.

Voth, P. D., and Hamner, K. C. (1940): Responses of Marchantia polymorpha to nutrient supply and photoperiod. Botanical Gazette 102, 169-205.

Wann, F. B. (1925): Some of the Factors Involved in the Sexual Reproduction of Marchantia polymorpha. American Journal of Botany12, 307-318.

Ware, G. W., and Whitacre, D. M. (2004): The Pesticide Book. Meister Media Worldwide. Willoughby, Ohio.

Watson, A. K. (1991): The classical approach with plant pathogens. In D. 0. TeBeest (Ed.): Microbial control of weeds, Chapman and Hall, London.

Watson, E. V. (1971): The structure and life of bryophytes. Hutchinson University Library. London.

Webster, J. (1980): Introduction to fungi: Cambridge University Press. Cambridge.

Weidemann, G. J., and TeBeest, D. 0. (1990): Genetic variability of fungal pathogens and their weed hosts. In R. E. Hoagland (Ed.): Microbes and Microbial Products as Herbicides. ACS Symposium Series 439, American Chemical Society, Washington, DC.

Weston, L. A. (2005): History and current trends in the use of allelopathy for weed management. 4th World Congress on Allelopathy.

Wheeler, M. H., Stipanovic, R. D., and Puckhaber, L. S. (1999): Phytotoxicity of equisetin and epi-equisetin isolated from Fusarium equiseti and F- pallidoroseum. Mycological Research 103, 967-973.

Whitehead, R. (2006): The UK pesticide guide 2006, British Crop Protection Council, Farnham, Surrey.

Whitehead, R. (2007): The UK pesticide guide 2007, CABI and BCPC, Wallingford and Alton.

304 References Chapter 8

Wittstock, U., and Halkier, B. A. (2002): Glucosinolate research in the Arabidopsis era. Trends in Plant Science 7, 263-270.

Wojtaszek, P. (1997): Oxidative burst: an early plant response to pathogen infection. Biochem. J322, 681-692.

Wyatt, R., and Anderson, L. E. (1984): Breeding systems in bryophytes, pp. 39-64. In A. F. Dyer, and J. G. Duckett (Eds): The Experimental Biology of Bryophytes, Academic Press, London.

Yamane, A., Fujikura, J., Ogawa, H., and Mizutani, J. (1992): Isothiocyanates as alleopathic compounds from Rorippa indica Hiern. (Cruciferae) roots. Journal of Chemical Ecology18, 1941-1954.

Yoshikawa, H., Ichiki, Y., Sakakibara, K., Tamura, H., and Suiko, M. (2002): The biological and structural similarity between lunularic acid and abscisic acid. Bioscience Biotechnology and Biochemistry 66, 840-846.

305 Appendix 1

Appendix 1 Glossary

Subject Description Reference Anamorph Imperfect or asexual state (Holliday, 1998) Antheridiophore Structure bearing antheridia (Lawrence, 2000) Antheridium(a) Organ in which male gametes are produced (Lawrence, 2000) Archegoniophore Structure bearing archegonia (Lawrence, 2000) Archegonium(a) Organ in which female gametes are produced (Lawrence, 2000) Ascus (pl. asci) Typical cell of Ascomycotina, bag or hollow club- (Holliday, 1998) shaped, usually containing 8 asco. Azeotrope A mixture of two substances that boil at the same (Lawrence, 2000) specific temperature so they may be distilled off. Butan-1-ol can therefore be used to remove acetic acid

Chlamydopore(s) Asexual spore, primarily a resting stage (Holliday, 1998) Cleitothecium Closed fruiting body (ascocarp) with spores (Lawrence, 2000) produced internally and released by breakdown of the wall Conidioma Any fungal structure bearing conidia (Lawrence, 2000) Conidiophore Hypha or cell bearing conidiogenous cells (Holliday, 1998) Conidium (pl. conidia) Asexual fungal spore (Holliday, 1998)

Dioecious Plant with male and female structures on (Lawrence, 2000) separate plants Endophyte An organism that completes its life cycle in a (Holliday, 1998) plant with no external signs of infection Euseptate Cells separated by multilayered walls, (Holliday, 1998) Excipulum Tissue containing the hymenium, in an (Holliday, 1998) apothecium or forming the walls of a perithecium of an ascoma Falcate Sickel-shaped or hooked (Lawrence, 2000) Fusiform Spindle-shaped, tapering gradually at both ends (Lawrence, 2000) Fusoid Somewhat fusiform (Lawrence, 2000) Gametophore Structure bearing the spore on which the (Lawrence, 2000) archegoniophore and antheridophore Gametophyte Haploid, gamete-forming phase (Lawrence, 2000) Gemma cup Cup shaped structure bearing gemmae (Lawrence, 2000) Gemma(e) Lenticular vegetative propagule (haploid) that (Lawrence, 2000) develops into a thallus Gemmaling Young, recently germinated gemma (Lawrence, 2000) Hyaline Transparent or translucent (Lawrence, 2000) Hymenium Spore-bearing layer in a basidioma of a (Holliday, 1998) macrofungus Macroconidium (pl, Larger conidium of a fungus that may also have (Holliday, 1998) macroconidia) microconidia e.g. Fusarium spp. Mycotoxicosis Poisoning by toxins produced by fungi (Lawrence, 2000)

306 Appendix 1

Subject Description Reference Obligate parasite Organism wholly dependent on another living (Holliday, 1998) organism for its nutrition Penicillus Branched brush-like strucure at the end of a (Holliday, 1998) conidiophore with one or more whores of branches Perithecium Flask-shaped ascoma (Holliday, 1998) Phialide A cell that develops open ended conidiogenous (Holliday, 1998) loci from which basipetal succession of conidiospores are produced Pionnote, Producing spores over the mycelium, with the (Booth, 1971). surface covered in spores from many sporodochia Pseudothecium Fruiting body resembling a perithecium (Holliday, 1998) Saprophyte Organism that gains nutrition from dead or (Holliday, 1998) decaying organic matter Sporodochium Conidioma with the spore mass supported by a (Lawrence, 2000) superficial, pulvinate mass of short conidiophores

Sporophyte Diploid phase, in which meiosis occurs forming (Lawrence, 2000) the haploid gametophyte Teleomorph Perfect or sexual state (Holliday, 1998) Thallus (pl. thalli) Liverwort body, not differentiated into leaf or (Lawrence, 2000) stems, which in Marchantia polymorpha bears circular gemma cups

Verticillate Arranged in whorls or verticils (Lawrence, 2000)

307 Appendix 2 — Abbreviations

Appendix 2 Abbreviations

Abbreviations a.i. Active ingredient ADAS ADAS UK Ltd DCM Dichloromethane DMSO Dimethyl sulphoxide GC Gas chromatography GSL Glucosinolate HDC Horticultural Development Agency HPLC High pressure liquid chromatography ITC Isothiocyanate M51C Liverwort media (Ono, 1979) MEA Malt extract agar WIC Methyl isothiocyanate MS Murashige and Skoog with vitamins (M0222, Duchefa) MSMC Murashige and Skoog Shoot Multiplication Medium C (M0525, Sigma Aldrich) PDA Potato dextrose agar PDB Potato dextrose broth TLC Thin layer chromatography

308 Appendix 3 — Suppliers

Appendix 3 Suppliers

Supplier Address ABCR GmbH Im Schlehert 10, D-76187. Karlsruhe American National Can American National Can, USA CABI Bioscience Bakeham Lane, Egham, Surrey, -nN20 9TY Centraalbureau voor Baarn, The Netherlands Schimmelcultures (CBS) Duchefa Biochemie BV. A. Hofmanweg 71, 2031 BH Haarlem, The Netherlands Growing technologies Ltd Number One Mill, The Wharf, Shardlow, Derby, DE72 2GH Growth technology www.greenair.com/genesis.htm Herbiseed New Farm, Mire Lane, West End, Twyford, RG10 ON) Massey University Private Bag 11-222, Palmerston North, New Zealand MonroSouth Unit 1, Quarrywood Industrial Estate, Burntash Road, Aylesford, Maidstone, Kent, ME20 7AD Oxoid Ltd. Basingstoke, Hants Sigma Aldrich Co. Ltd The Old Brickyard, New Road, Gillingham, Dorset, SP8 4XT W K Thomas & Company Mount Road, Chessington, Surrey, KT9 1HY Ltd. William Sinclair Horticulture Firth Road, Lincoln, LN6 7AH

309 Appendix 4 - ANOVA tables

Appendix 4 Anova tables relating to Section 6.3.7.3

Pre-emergence P. winteri ANOVA table for total liverwort area, pre-emergence P. winteri treatment. Week 3. L.s.d = 4720.5 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 614314 614314 0.28 Block.*Units* stratum Treatment 3 98801047 32933682 14.97 0.026* Residual 3 6600355 2200118 Total 7 106015716

ANOVA table for healthy liverwort area, pre-emergence P. winteri treatment. Week 3. L.s.d = 3762 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 1293586 1293586 0.93 Block.*Units* stratum Treatment 3 92149104 30716368 21.98 0.015* Residual 3 4192251 1397417 Total 7 97634941

Pre-emergence F. equiseti

ANOVA table for total liverwort area, pre-emergence F. equiseti treatment. Week 3. L.s.d = 1692.5 m.s. v.r. F pr. Source of variation d.f. s.s. Block stratum 1 15156294 15156294 5.29 Block.*Units* stratum Treatment 3 148407538 49469179 17.27 0.021* Residual 3 8593308 2864436 Total 7 172157140

ANOVA table for total liverwort area, pre-emergence F. equiseti treatment. Week 5. L.s.d = 4094.2. Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 22826383 22826383 13.79 Block.*Units* stratum Treatment 3 232087238 77362413 46.74 0.005** Residual 3 4965315 1655105 Total 7 259878935

ANOVA table for dieback liverwort area, pre-emergence F. equiseti treatment. Week 11. L.s.d = 7847.8 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 1090854 1090854 0.18 Block.*Units* stratum Treatment 3 97831851 32610617 5.36 0.101 Residual 3 18242992 6080997 Total 7 117165697

310 Appendix 4 - ANOVA tables

Post-emergence F. equiseti

ANOVA table for total liverwort area, post-emergence F. equiseti treatment. Week 3. L.s.d = 5446.2 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 4997410 4997410 1.71 Block.*Units* stratum Treatment 3 165704636 55234879 18.86 0.019* Residual 3 8785833 2928611 Total 7 179487879

ANOVA table for healthy liverwort area, post-emergence F. equiseti treatment. Week 3. L.s.d = 5955.9 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 5112175 5112175 1.46 Block.*Units* stratum Treatment 3 165915661 55305220 15.79 0.024* Residual 3 10507399 3502466 Total 7 181535234

ANOVA table for dieback liverwort area, post-emergence F. equiseti treatment. Week 3. L.s.d = 881.8 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 651 651 0.01 Block.*Units* stratum Treatment 3 597960 199320 2.60 0.227 Residual 3 230340 76780 Total 7 828952

ANOVA table for total liverwort area, post-emergence F. equiseti treatment with 1 fungal application. Week 5. L.s.d = 3074.3 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 996501 996501 1.07 Block.*Units* stratum Treatment 3 178867963 59622654 63.89 0.003** Residual 3 2799539 933180 Total 7 182664003

ANOVA table for healthy liverwort area, post-emergence F. equiseti treatment with 1 fungal application. Week 5. L.s.d = 3737.0 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 1129817 1129817 1.54 Block.*Units* stratum Treatment 3 183000818 61000273 83.08 0.002** Residual 3 2202831 734277 Total 7 186333466

311 Appendix 4 - ANOVA tables

ANOVA table for total liverwort area, post-emergence F. equiseti treatment with 2 fungal applications. Week 5. L.s.d = 12197.5 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 4665898 4665898 0.32 Block.*Units* stratum Treatment 3 149536629 49845543 3.39 0.171 Residual 3 44069810 14689937 Total 7 198272337

ANOVA table for healthy liverwort area, post-emergence F. equiseti treatment with 2 fungal applications. Week 5. L.s.d = 21529.9 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 25548292 25548292 0.56 Block.*Units* stratum Treatment 3 166591536 55530512 1.21 0.439 Residual 3 137304227 45768076 Total 7 329444055

ANOVA table for dieback liverwort area, post-emergence F. equiseti treatment with 2 fungal applications. Week 5. L.s.d = 9507.4 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 8377915 8377915 0.94 Block.*Units* stratum Treatment 3 24576027 8192009 0.92 0.527 Residual 3 26774753 8924918 Total 7 59728696

Post-emergence P. winteri

ANOVA table for total liverwort area, post-emergence P. winter/treatment. Week 3. L.s.d = 6142.9 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 1182046 1182046 0.32 Block.*Units* stratum Treatment 3 133672239 44557413 11.96 0.036* Residual 3 11177598 3725866 Total 7 146031883

ANOVA table for healthy liverwort area, post-emergence P. winteri treatment. Week 3. L.s.d = 5495.5 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 7422117 7422117 2.49 Block.*Units* stratum Treatment 3 106615347 35538449 11.92 0.036* Residual 3 8945580 2981860 Total 7 122983043

312 Appendix 4 - ANOVA tables

ANOVA table for dieback liverwort area, post-emergence P. winteritreatment treatment. Week 3. L.s.d = 6719.7 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 2680218 2680218 0.60 Block.*Units* stratum Treatment 3 9711328 3237109 0.73 0.601 Residual 3 13375317 4458439 Total 7 25766863

ANOVA table for total liverwort area, post-emergence P. winteritreatment treatment with 1 fungal application. Week 5. L.s.d = 7230.0 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 109613 109613 0.02 Block.*Units* stratum Treatment 3 182984896 60994965 11.82 0.036* Residual 3 15483790 5161263 Total 7 198578299

ANOVA table for total liverwort area, post-emergence P. winteritreatment with 2 fungal applications. Week 5. L.s.d = 7953.6 Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 598798 598798 0.10 Block.*Units* stratum Treatment 3 183682764 61227588 9.80 0.046* Residual 3 18738072 6246024 Total 7 203019634

ANOVA table for total liverwort area, post-emergence P. winteritreatment with 2 fungal applications. Week 11. L.s.d = 0.1137. Data transformed using natural logs. Source of variation d.f. s.s. m.s. v.r. F pr. Block stratum 1 0.007385 0.007385 5.79 Block.*Units* stratum Treatment 3 0.063685 0.021228 16.64 0.023* Residual 3 0.003826 0.001275 Total 7 0.074897

313