Carnivorous Plant Responses to Resource Availability

Carnivorous plant responses to resource availability:

environmental interactions, morphology and biochemistry

Christopher R. Hatcher

A doctoral thesis submitted in partial fulfilment of requirements for the award of Doctor of

Philosophy of Loughborough University

November 2019

Abstract Understanding how organisms respond to resources available in the environment is a fundamental goal of ecology. Resource availability controls ecological processes at all levels of organisation, from molecular characteristics of individuals to community and biosphere. Climate change and other anthropogenically driven factors are altering environmental resource availability, and likely affects ecology at all levels of organisation. It is critical, therefore, to understand the ecological impact of environmental variation at a range of spatial and temporal scales. Consequently, I bring physiological, ecological, biochemical and evolutionary research together to determine how plants respond to resource availability.

In this thesis I have measured the effects of resource availability on phenotypic plasticity, intraspecific trait variation and metabolic responses of carnivorous sundew plants. Carnivorous plants are interesting model systems for a range of evolutionary and ecological questions because of their specific adaptations to attaining nutrients. They can, therefore, provide interesting perspectives on existing questions, in this case trait-environment interactions, plant strategies and plant responses to predicted future environmental scenarios.

In a manipulative experiment, I measured the phenotypic plasticity of naturally shaded Drosera rotundifolia in response to disturbance mediated changes in light availability over successive growing seasons. Following selective disturbance, D. rotundifolia became more carnivorous by increasing the number of trichomes and trichome density. These plants derived more N from prey and flowered earlier. This suggests that D. rotundifolia extend their fundamental niche in part due to phenotypic plasticity into environments that are shaded and are, for non- carnivorous plants, high in stress and disturbance—untenable environments according to universal adaptive theory.

In an observational study on peatlands across Europe, I investigated how habitat heterogeneity and climate-driven changes in nutrient availability impact on D.

i rotundifolia trait variability. I found that patterns of hydrology and light are consistent across peatlands. But, climate driven N accumulation alters the expression of carnivory which can reverse expected patterns of trait expression. These findings have implications for predicting the ecological impacts of climate change, as patterns of precipitation and evaporation may have different consequences depending on specific local topology and climate norms and can have significant impacts at small spatial scales, to the extent that entire habitats may be lost.

I conducted a research review to determine the role of secondary plant metabolites in carnivory. This identified similar patterns of metabolite production among closely- and distantly-related carnivorous taxa, which suggests that secondary metabolites are important for plants to facilitate carnivory. The review supports the hypothesis that secondary plant metabolites are drivers of plant diversification and evolution of plants into new environments.

In a laboratory experiment using Drosera capensis, I measured the metabolic response of trap tissue to simulated prey capture, using an omics approach. There was a substantial response to simulated prey capture. More compounds were up-regulated following prey capture than expected (over 30 times more than currently identified across all carnivorous species). In addition, changes in several metabolites suggested that their function is different to that previously reported.

The findings of this thesis highlight substantial phenotypic plasticity and intraspecific trait variation as a means for plants to cope with environmental resource availability and variability. Furthermore, it is clear that climate interacts with processes of nutrient availability, hydrology, and subsequent vegetation patterns which means that accurate predictions of species responses to climate change must be highly contextualised to local environments. The findings of this thesis also suggest that secondary metabolites may hold a significant role in the response and survival of plants to changing environments as they have the propensity for plants to adapt at a rate faster than mutation.

ii Acknowledgements I would like to extend my thanks to my primary supervisor, Dr Jon Millett, I thank you for your continued professional support and advice, and your belief in me when

I first approached you as a PhD candidate. I thank you for your support in this thesis and also in my development as a scientist. To my second supervisor, Dr Dave Ryves,

I thank you for your feedback. To Prof Paul Wood, I thank you for your support throughout my PhD–not just as my internal assessor, but for the day-to-day support you provide to myself, other postgraduates and undergraduates studying at

Loughborough University.

To Prof Adam Hart, I thank you for directing my passion for the natural world into research and your perpetual source of inspiration and aspiration as an academic.

I thank the technical support at Loughborough University, in particular Richard

Harland. You are always willing and in good spirit to help and without you, aspects of this project would have been considerably more difficult to complete.

I give my gratitude to Ulf Sommer, Dr Liam Heaney & Prof Duncan Cameron for advice on metabolomic analyses. Thanks to Angus Davidson for assistance with the drawings for the evolutionary tree. I thank Dr Oliver Preedy for access to the pXRF for tissue analysis and I thank the South West Carnivorous Plants nursery for supplying Drosera capensis for Chapter 5.

To my wife, Mrs Sarah Hatcher, I thank you for your unyielding encouragement and tolerance of me. To my parents, Mr Leslie Hatcher and Mrs Jacqueline Hatcher, thank you for your continued support over the years which have made me the person I am today.

iii Funding This PhD was funded by a Loughborough University PhD studentship. Stable isotope analysis in Chapters 2 & 3 was funded by the UK Natural Environment

Research Council (NERC), through funding to the Life Sciences Mass Spectrometry

Facility (LSMSF). Fieldwork was supported in Chapter 3 by Transnational Access funding provided through the European Union H2020 programme as part of the

Long-Term Ecosystem Research in Europe (eLTER) project, and by a separate grant from the International Peatland Society. Metabolomic analysis, in Chapter 5, was funded by a NERC grant at the NERC Biomolecular Analysis Facility (NBAF) in

Birmingham.

Publication At the time of thesis submission, a version of Chapter 4 has been accepted for publication in Annals of Botany.

C.R. Hatcher, D.B. Ryves, J. Millett (in press) The function of secondary plant metabolites in plant carnivory. Annals of Botany

iv Table of Contents

Abstract ...... i

Acknowledgements ...... iii

Funding ...... iv

Publication ...... iv

List of tables ...... x

List of figures ...... xiii

Preface...... 1

Chapter 1 Introduction ...... 3

1.1 Carnivorous plants ...... 9

1.2 Responses of carnivorous plants to resource availability ...... 16

1.2.1 Mineral nutrition and responses to nutrient availability ...... 16

1.2.2 Responses to light ...... 20

1.2.3 Responses to water availability ...... 22

1.3 Secondary metabolites and carnivorous plants ...... 25

1.4 Habitats ...... 28

1.5 Plant responses to climate...... 31

1.6 Research questions ...... 33

1.7 Thesis structure ...... 34

Chapter 2 Phenotypic plasticity in the carnivorous habit of the round-leaved sundew, Drosera rotundifolia, to disturbance ...... 36

v 2.1 Introduction ...... 37

2.2 Methods ...... 40

2.2.1 Site information...... 40

2.2.2 Experimental set-up...... 40

2.2.3 Measurements ...... 41

2.2.4 Trait measurement protocol using Adobe Photoshop...... 43

2.2.5 Tissue nutrient concentration and nitrogen isotope analysis ...... 44

2.2.6 Data analysis ...... 46

2.3 Results ...... 46

2.3.1 Morphological traits ...... 52

2.3.2 Drosera rotundifolia nutrient stoichiometry and proportion of prey derived

N ...... 53

2.4 Discussion ...... 56

Chapter 3 Climate-driven changes in small-scale habitat characteristics affects phenotypic variation of the carnivorous herb Drosera rotundifolia in patterned peatlands...... 62

3.1 Introduction ...... 63

3.2 Methods ...... 66

3.2.1 Site information and plot selection...... 66

3.2.2 Measurements ...... 67

3.2.3 Trait measurement using Adobe Photoshop ...... 69

3.2.4 Nitrogen isotope and tissue nutrient concentration ...... 69

3.2.5 Data analysis ...... 71

vi 3.3 Results ...... 72

3.3.1 Patterns of light and nutrient availability in microforms across sites ...... 72

3.3.2 Influence of habitat heterogeneity and nutrient availability driven by

climate factors across sites on trait expression and nutrient composition of

Drosera rotundifolia plant traits...... 76

3.3.3 Drosera rotundifolia nutrient stoichiometry and N derived from prey ...... 81

3.4 Discussion ...... 83

3.4.1 Habitat heterogeneity ...... 85

3.4.2 Prey capture and tissue N concentrations between microforms among sites

...... 86

3.5 Appendix C3 ...... 89

Chapter 4 The function of secondary metabolites in plant carnivory...... 92

4.1 Introduction ...... 92

4.2 Secondary metabolites involved in the attraction of prey ...... 95

4.2.1 Olfactory cues...... 95

4.2.2 Visual cues ...... 99

4.2.3 Extra-floral nectaries ...... 101

4.2.4 Presence of metabolites across independent lineages of plant carnivory for

prey attraction ...... 102

4.3 Secondary metabolites involved in the capture of prey ...... 103

4.3.1 Physical trapping mechanisms ...... 104

4.3.2 Anti-adhesive wax production used in prey capture ...... 104

4.3.3 Mucilage and glue production for prey capture ...... 105

4.3.4 Prey capture signalling ...... 106

vii 4.4 Secondary metabolites involved in the digestion of prey ...... 108

4.4.1 Presence of compounds involved in digestion within the same carnivorous

lineage ...... 110

4.5 Secondary metabolites involved in nutrient assimilation and whole plant

responses ...... 111

4.6 Occurrence of secondary plant metabolites involved in carnivory across plant

taxa ...... 112

4.7 Concluding remarks and future perspectives ...... 117

4.8 Supplementary Material ...... 119

Chapter 5 Metabolomic analysis reveals a reliance on secondary plant metabolites to facilitate carnivory in the Cape sundew, Drosera capensis ...... 162

5.1 Introduction ...... 163

5.2 Methods ...... 164

5.2.1 Plant rearing and preparation ...... 164

5.2.2 Experimental design ...... 165

5.2.3 Prey addition ...... 166

5.2.4 Plant harvests ...... 166

5.2.5 Sample preparation for metabolite profile analysis ...... 166

5.2.6 UHPLC-MS analysis (positive ion mode) ...... 167

5.2.7 Data processing ...... 167

5.2.8 Statistical analysis and identification of metabolites ...... 168

5.2.9 Multivariate analysis ...... 169

5.2.10 Annotation summary and pathway analysis ...... 170

5.2.11 Metabolite comparisons to other carnivorous plants ...... 171

viii 5.3 Results ...... 173

5.3.1 Multivariate analyses ...... 173

5.3.2 Volcano plot ...... 180

5.3.3 Pathway analysis of biologically important metabolites...... 187

5.3.4 Convergence of compounds for plant carnivory ...... 188

5.4 Discussion ...... 195

5.5 Conclusion ...... 200

5.6 Appendix C5 ...... 202

Chapter 6 Synthesis ...... 203

6.1 Thesis overview ...... 203

6.2 Synthesis of findings and new avenues for future work ...... 205

6.3 Climate ...... 206

6.4 Light ...... 209

6.5 Root-derived and prey-derived N in carnivorous plants ...... 212

6.6 Biochemical responses (secondary plant metabolites) ...... 214

6.7 The omics ...... 216

6.8 Conclusions ...... 218

References ...... 220

ix List of tables Table 1.1: Tissue nutrient concentration of nitrogen, phosphorous and potassium in leaves and shoots (N, P, K respectively, median dry weight %, simplified from

Ellison & Adamec 2018)...... 17

Table 1.2: Variability in the proportion of nitrogen (N) derived from prey, compared to root uptake within and across select carnivorous plant species (adapted and updated from Ellison & Gotelli, 2001 and Ellison & Adamec, 2018)...... 19

Table 2.1: Results of univariate statistical analysis of differences between Drosera rotundifolia plants growing in different vegetation manipulation treatments.

Presented are baseline statistical comparisons (ANOVA) to confirm homogeneity across plots prior to treatment. Also included are RCB repeated measures ANOVA of treatment effects. Effect size is indicated by mean values for the two growing seasons after treatment application (Jul ’16 to Sept ’17) with SEM below. Arrows indicate direction of effect where treatment effect is significant. Statistically significant differences between the Light treatment and other treatments are emboldened with the exception of baseline measurements of CIE Lab in which Light and Natural are different to Artificial and Control. Figures rounded to 2 d.p. Plant traits are per leaf. CIE lab is International Commission on Illumination L*a*b* space.

...... 47

Table 2.2: Results from univariate statistical analysis of differences in nutrient composition of D. rotundifolia growing in different vegetation manipulation treatments. Effect size is indicated by mean values and SEM from natural abundance isotope ratio and pXRF analysis. Data in this table derived from plants collected at the end of the experiment. Emboldened results are comparisons of Light to other treatments (unless stipulated as otherwise in direction of treatment effect)...... 48

Table 3.1: Site information for three peatlands across Europe. Predicted nutrient accumulation based on definitions of Eppinga et al. (2010) ...... 66

x Table 3.2: Summary for hierarchical ANOVA analysis of environmental variables in hummocks and hollows across three sites (Sw = Sweden, F = Finland, Sc = Scotland).

‘/’ denotes a significant difference ordered largest to smallest alpha 0.05. Duplicate letters appearing either side of ‘/’ signifies a non-significant difference to adjacent letters, but the other two sites are significantly different. Microform within site comparisons are represented using hollows as a base level and denoting whether hummocks have significantly higher [+] or lower [-] values or not significantly different [O], P value significance threshold (alpha) 0.05. Where 0.05 < P < 0.1 arises, differences are in brackets. PC signifies planned contrast analysis of microform within each site...... 74

Table 3.3: Summary for hierarchical ANOVA analysis of Drosera rotundifolia morphological measurements and nutrient status in hummocks and hollows across three sites (Sw = Sweden, F = Finland, Sc = Inverewe, Scotland). ‘/’ denotes a significant difference ordered largest to smallest. Duplicate letters appearing either side of ‘/’ signifies a non-significant difference to adjacent letters, but the other two sites are significantly different. Microform (M) within site comparisons are represented using hollows as a base level and denoting whether hummocks have significantly higher [+] or lower [-] values or not significantly different [O], P value threshold of alpha 0.05. Where 0.05 < P < 0.1 arises, differences are in brackets. PC signifies planned contrast analysis of microform within each site...... 78

Table 4.1: Secondary plant metabolites involved in carnivory present in more than one genera. Role indicates role in the carnivorous habit, A: Attraction, C: Capture, D:

Digestion. Ticks represent the 26 compounds present across independent lineages of carnivory. Ticks within boxes represent presence of compound found within different genera of the same carnivorous ancestor. Purple are of Nepenthales, blue are of Ericales lineage of carnivory. Note that some compound isomers are identified, but in some studies, this is not included. References are included in Table

S 4.2...... 115

xi Table S 4.2: Secondary plant metabolites involved in plant carnivory. Light grey bars signify the compound is found in more than one genus of the same carnivorous lineage. Dark grey bars signify the compound is found in more than one independent lineage of carnivory. No compound has been found in more than 2 separate lineages of carnivory. Compounds reported from Hotti et al., 2017 and

Kreuzweiser et al., 2014 have 70% confidence minimum. Function of metabolite is known or highly likely. Note that some compound isomers are identified, but in some studies, this is not included...... 119

Table 5.1: Cross comparison of univariate and OPLS-DA analysis. T-tests confirm grouping from OPLS-DA. Only annotated compounds included in this table. This table was used for metabolic pathway analysis...... 185

Holm p is used to control multiple testing. FDR = False Discovery Rate. Pathway

Impact is a combination of centrality and pathway enrichment results calculated by the sum of the importance of each metabolite divided by the sum of all metabolites in each pathway...... 187

Table 5.3: Secondary metabolites annotated in this experiment and have been found to have a role in carnivory in a previous study, highlighting occurrence of this compound in other carnivorous plants and whether this experiment shows agreement of previously proposed function. Light grey bars are metabolites newly found present for Drosera. Dark grey bars are newly found in Nepenthales. White bars are compounds found previously for D. capensis. Ticks indicate that findings of this experiment are congruent with the previously proposed function. Crosses do not agree with previously proposed function and brackets indicate partial agreement or

xii disagreement and these compounds require further investigation, “ns” = no significant fold-change across any treatment, arrows indicate concentration increase or decrease by the end of the experiment. No. lineages = number of lineages in which the metabolite has been found to have a carnivorous function...... 189

List of figures Figure 1.1 Grime's CSR theory (1979) for plant strategies, C: competitors; S: stress tolerators; R: ruderals. Blue triangle represents environments plants can occupy. Red triangle represents theoretically uninhabitable environments of high stress and high disturbance (the untenable triangle). Wilson’s r/K model is overlaid, indicating how

Grime’s model builds upon the previous r/K model (MacArthur & Wilson, 1967). .... 8

Figure 1.2: Phylogenetic tree of carnivorous plants (non-carnivorous phylogenies removed). Classification follows APG (2016) and NCBI except for Caryophyllales which are divided from Nepenthales. Letters beside plants illustrate trap type: a: flypaper, b: snap-trap, c: pitfall, d: lobster pot, e: bladder, f: pigeon. Number of species (non-carnivorous species in parenthesis if any) and distribution derived from

Ellison & Adamec, 2018. Coloured blocks signify a shared carnivorous ancestral lineage. Aldrovanda & Utricularia are aquatic. Illustrations by Angus Davidson...... 11

(B1) compared to more light or water limited sites (B2) as carbon is the most limiting resource. Dashed lines indicate net difference in photosynthetic cost (C, yellow solid lines) and benefit of carnivory. Carnivory is evolutionarily favourable where there are marginal benefits as a result of allocation to carnivory i.e. a peak in dotted blue lines (B-C) where C is near 0. The optimal investment in carnivory occurs where the difference between the benefit and cost curves is highest. Where there is no peak in

xiii dotted line, there is no allocation to carnivory that is better than less allocation. (b) In sunny, moist environments, photosynthetic rate is greater on fertile sites compared to infertile sites. In fertile sites, carnivory is, therefore, unlikely to be favoured. (c)

Relative cost of carnivory is likely to be reduced with the presence of pre-existing traits or precursors to carnivory such as glandular hairs, specific enzymes or metabolites already present. (d) Carnivory is unlikely to evolve on extremely infertile soils as the initial cost to allocation in carnivory shows negative carbon balance, even though further investment may provide net gains in carbon. Adapted from Ellison & Adamec, 2018...... 13

Drosera (b) share a common carnivorous ancestor but have divergent trap morphology; c: Sarracenia purpurea filled with rain water filled digestive fluid; d: close up of Sarracenia purpurea with rain filled digestive fluid and drowned prey; e:

Pinguicula sp. is a flypaper trap of independent origin to Drosera (b); f: Cephalotus sp. is a pitcher plant of independent origin to Sarracenia which fills and controls pitcher fluid without rain. Photos by Christopher R. Hatcher...... 24

Figure 1.5: Secondary plant metabolites involved in the attraction of prey. Left:

Nepenthes khasiana under normal light. Right: Nepenthes khasiana under UV light shows pigment accumulation around the peristome of the trap to attract prey.

Photographs courtesy of Dr Sabulal Baby (Kurup et al., 2013)...... 27

Figure 2.1: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted line indicates end of growing season 1. Presented are mean ± SEM of top: number of trichomes per leaf; bottom: trichome density of trap.

xiv Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001)...... 49

Figure 2.2: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted line indicates end of growing season 1. Presented are mean ± SEM of top: leaf lamina colour (high values indicate red colour; low values indicate green colour); bottom: lamina area of longest leaf. Tukey HSD for comparison within a time point between vegetation removed (Light) and intact

(Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, (*) P = 0.0549...... 50

Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001...... 51

Figure 2.4: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted line indicates end of growing season 1. Presented are mean ± SEM. Proportion of D. rotundifolia with flowers. Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001...... 52

Figure 2.5: Drosera rotundifolia nutrient information at the end of the second growing season after vegetation and light manipulation treatments. Presented are mean ±

SEM of top: Drosera rotundifolia δ15N; bottom: Drosera rotundifolia tissue N concentration (%) * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001 ...... 54

Figure 2.6: Ternary plots of nutrient stoichiometry for Top: Drosera rotundifolia, bottom: Sphagnum spp. Nutrient limitation boundaries in the ternary plot are based on the criteria of Olde Venterink et al., 2003...... 55

Figure 3.1: Photo protocol for measurement of Drosera rotundifolia morphological traits...... 68

xv Figure 3.2: Differences between microforms (hummocks and hollows), for three sites in A: vegetation coverage, and B: Sphagnum spp. tissue %N. Error bars represent

SEM. *** P < 0.001 (*) P <0.10, ns = P > 0.10. C: Ternary diagram of N, P, K nutrient stoichiometry of Sphagnum spp. within microform at each site (%). Nutrient limitation guides from Olde Venterink et al., (2003)...... 73

Figure 3.3: Differences between D. rotundifolia leaf traits growing on different bog microforms (hummock vs hollow) in three different locations. Presented are A: trichome density, B: average (mean) number of trichomes per leaf, C: lamina area and D: length of the longest leaf (petiole plus lamina). Annotations above bars show the results of a priori planned comparisons tests for differences in leaf traits between microform within each bog. * P < 0.05, ** P < 0.005, *** P < 0.0001, (*) P < 0.10...... 77

Figure 3.4: Differences between D. rotundifolia growing on different microforms at three sites A: Proportion of prey derived N, B: Tissue %N, C: Ternary diagram of nutrient stoichiometry of D. rotundifolia within microform at each site. Nutrient limitation guides from Olde Venterink et al. (2003)...... 82

Figure 4.1: Phylogenetic tree of carnivorous plants and research and metabolites identified with known function in carnivory. Branches are to emphasise the location of carnivorous taxa and do not signify time since divergence or a metric of relatedness. Trap morphology is grouped into: (a) flypaper; (b) snap-trap; (c) pitfall;

(d) lobster pot; (e) suction bladder; and (f) pigeon-trap. Colour bands signify a common carnivorous ancestor. Tree generated using NCBI taxonomy browser, iTOL and APG (APG, 2016; Letunic & Bork, 2016). Nepenthales sensu stricto sister to

Caryophyllales. 24 common compounds are found in the Nepenthales lineage and

Sarraceniaceae lineage. Note that Aldrovanda and Utricularia are genera of aquatic carnivorous plants...... 94

xvi confirmed to be related to carnivory; (b) upper figure: Nepenthes species show redder locations, possibly to contrast trap peristome to background for prey attraction.

Lower figure: Nepenthes spp. also use UV inflorescence for prey attraction; (c)

Sarracenia purpurea with high diversity of red and green pigmentation from secondary compounds with high background contrast in situ...... 100

Figure 4.3: Evolution and production zones of metabolites for attraction of prey via use of extrafloral nectaries. Highlighted areas are highest density of extrafloral nectaries for attraction of prey. Areas for secretion of secondary metabolites and nectar are important for prey attraction and capture. Length of phylogenetic branches do not signify relatedness or any other metric of evolution...... 102

Figure 4.4: Drosera capensis leaf bending caused by jasmonate accumulation after prey capture. The regulation of this is not yet known. Credit: Christopher R. Hatcher

...... 107

Figure 5.1: Data workflow for metabolomics analysis used in this experiment (Di

Guida et al., 2016; Grace & Hudson, 2016; Considine et al., 2018)...... 172

Figure 5.2: Ordination of the first two principle components from a PCA of metabolite profiles of Drosera capensis following insect substrate addition, with unfed controls. Treatments indicated are hours after substrate addition (c = unfed controls).

Presented are the PC1 and PC2 scores for each plant. Different symbols represent each time point. The ellipsoids are normal contour lines with probability 68% for each group. Presented PCA is without inclusion of the anomalous result (one control sample)...... 174

Figure 5.3: Orthogonal partial least squares-discriminant analysis model 2- dimensional score plot for unfed plants (black circles) and plants up to 6 hrs after feeding (green triangles). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data...... 176

xvii Figure 5.4: S-plot for modelled variables using an orthogonal partial least squares- discriminant analysis model with unit variance scaling for unfed versus ≤ 6 hrs after feeding samples. Important compounds are highlighted with red ellipses...... 176

Figure 5.5: Orthogonal partial least squares-discriminant analysis model 2- dimensional score plot for unfed plants (black circles) and plants ≥ 24 hrs after feeding (blue squares). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data...... 177

Figure 5.6: S-plot for modelled variables using an orthogonal partial least squares- discriminant analysis model with unit variance scaling for unfed versus ≥ 24 hrs after feeding samples. Important compounds are highlighted with red ellipses. .... 177

Figure 5.7: Orthogonal partial least squares-discriminant analysis model 2- dimensional score plot for plants ≤ 6 hrs after feeding (green triangles) and plants ≥

24 hrs after feeding (blue squares). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data. ... 178

Figure 5.8: S-plot for modelled variables using an orthogonal partial least squares- discriminant analysis model with unit variance scaling for plants ≤ 6 hrs after feeding versus plants ≥ 24 hrs after feeding. Important compounds are highlighted with red ellipses...... 178

Hotelling’s t2 as a quality check to see all the data are within normal boundaries.

Ellipsoid is 95% confidence interval for the data...... 179

Figure 5.10: Loading plot of metabolites (X) with relatedness to (Y) plants ≥ 24 hrs after feeding, ≤6 hrs after feeding and unfed controls (left to right respectively). Red dots are metabolites most associated with plants ≥ 24 hrs after feeding compared to unfed and ≤ 6 hrs after feeding...... 179

xviii Figure 5.11: Number and proportion of discriminatory features in treatment groups compared to the control group. Shown here are log fold changes (x-axis), in comparison to unfed control plants, and the y-axis displays the negative log of the p- value from a two-sample t-test. Data points that are far from the origin (near the top or far left and right) are considered important variables with potentially high biological relevance. Blue and red features (down-regulated and up-regulated respectively) are those chosen based on the presented criteria of thresholds of 2 and -

2 for the log2 fold change and 2 for the -log FDR adjusted p value (p < 0.01, indicated with red margins) to highlight those features that showed the largest differences

(Grace & Hudson, 2016)...... 181

Figure 5.12: Metabolic patterns over time after prey substrate addition. Presented are the 100 most important features, based on statistical significance and fold change.

Colour and colour intensity represent Z score for the change in peak intensity from the mean of the compound (one compound per row). Each column represents a single plant, these are grouped according to treatment and time point (randomly assigned within these). Each row represents a single metabolite grouped as a dendrogram by similarity in change of intensity...... 182

Figure 5.13: Normalised intensities of a selection of metabolites which were highlighted as important. A-H are compounds highlighted as biologically important by all analyses (Volcano plot group ≤ 6 hrs and ≥ 24 hrs and from OPLS-DA of fed groups compared to unfed controls). These compounds identified by OPLS-DA are therefore confirmed by FDR adjusted t-tests that they are likely to be of high biological importance. Green, yellow and orange are treatments ≤ 6 hrs, blue are ≥ 24 hrs, black are unfed controls. Note that the x-axis is not scaled and therefore distances are not proportional between treatment boxes. If metabolite is not possible to be annotated, its peak is ascribed...... 183

Figure 5.14: (Figure 5.13 cont.). Normalised intensities of metabolites highlighted as important from volcano plot or OPLS-DA. I: Compound highlighted as biologically

xix important by all analyses (Volcano plot group <6 hrs and ≥ 24 hrs and from OPLS-

DA of treatment groups compared to controls). C, E, J, K, L: Compounds previously known to be involved in carnivory and confirmed to be significant by volcano plot in this experiment. M, N, O, P: Compounds identified by three-way OPLS-DA (Figure

5.9, Figure 5.10) as highly associated with a separation of ≥ 24hrs from unfed controls and plants ≤ 6 hrs after feeding...... 184

Figure 6.1: Framework for considering the impact of resource availability on carnivorous plant traits from climate to plant physiology. Yellow arrows indicate result from an increase in investment in carnivory. Numbers signify which empirical chapter examines the specific process...... 206

Figure 6.2: Colour response likely due to accumulation of anthocyanins in response to increased light availability during the first growing season from Chapter 2. Left:

D. rotundifolia naturally shaded; right; the same D. rotundifolia individual 2 months after vegetation removal...... 211

Figure 6.3: Pathway for omics technologies to identify gene expression through to phenotypic expression linking genes to the environment...... 217

xx Preface “At this present moment I care more about Drosera than the origin of all the species in the

world”—Charles Darwin

Over a century and a half since Darwin penned those words, researchers and non- scientists alike have continued to be fascinated by carnivorous plants. Carnivorous plants are remarkable examples of co-evolution and evolutionary convergence

(Thorogood et al., 2017). Enhancing our understanding of carnivorous plants has led to a better understanding of symbioses in plants, animals, fungi and microbes, in some cases comprising over 150 commensals within a single carnivorous plant

(Anderson & Midgley, 2002; Lam et al., 2017; Schöner et al., 2017). Genetic analysis of carnivorous plants is providing breakthroughs in design of synthetic plants—

Genlisea and Utricularia have the smallest known plant genomes, half the size of model species Arabidopsis thaliana (Leushkin et al., 2013; Lan et al., 2017).

Understanding morphological and physiological processes of the carnivorous syndrome has advanced our understanding of physical and chemical defences across the plant kingdom inspiring new biotechnologies (Wong et al., 2011), while defining and modelling the environments carnivorous plants are restricted to is developing hypotheses for adapting to and persisting in environments of varying resource availability in space and time (Brewer, 2008; Karagatzides & Ellison, 2009;

Abbott & Brewer, 2016).

This thesis takes advantage of the utility of carnivorous plants for a variety of disciplines within biology at varying scales. Plant trade-offs are reliably measured in carnivorous plants due to the separation of investment in carnivory for nutrients versus photosynthesis within the same leaf tissue and are ideal species for understanding allocation and investment under different levels of resource availability. They are therefore useful systems to understand the fine balance between carbon and nutrient acquisition and availability, and how plants can alter investment in strategies in response to demand and supply of resources. Owing to

1 carnivorous plants being highly sensitive to changes in their environment for both nutrients and light and ease in which to measure trait responses, they are becoming effective systems to test hypotheses of climate-driven nutrient cycling and forecasting how species and communities may respond to patterns and processes of climate change on a global scale (Ellison & Adamec, 2018). Carnivorous plants are used in this thesis to understand plant biochemical, morphological and physiological responses to resource availability and as potential indicators of the impacts of climate change to environments that are of global significance (Gotelli & Ellison,

2002; Cook et al., 2018).

2 Chapter 1 Introduction

A fundamental goal in ecology is to understand how organisms respond to resource availability in order to determine ecosystem patterns and processes globally

(Sutherland et al., 2013). Resource uptake from the environment and resource allocation to survival, growth and reproduction controls ecological processes at all levels of organisation, from molecular and physiological characteristics, to the morphological structure of individuals, to communities and the biosphere (Stuart

Chapin et al., 2012; Fronhofer et al., 2018; Tamminen et al., 2018). The composition of the environment, in terms of resource availability and limitation, strongly constrains how and where these resources are allocated within the organism (Putman &

Wratten, 1983; Yang & Midmore, 2005). Understanding the dynamics of trade-offs and how plants alter allocation of resources in response to resource availability at different temporal and spatial scales is therefore critical for understanding the distribution of plants, plant life-history strategies, ecosystem function and enables the prediction of ecological responses to environmental change.

In ecology, ‘resource’ and ‘condition’ have been described in subtly different ways.

Tilman described resources as “all things consumed by an organism” (Tilman, 1982), and Begon et al. (1986) defined it as that which “may be consumed by an organism and as a result becomes unavailable to another”. Examples of resources according to

Grime are concurrent and include “quanta of light, ion of a mineral nutrient, molecule of water and volume of space” (Grime, 1973). Conditions are abiotic environmental factors which vary in space and time, but cannot be consumed or made less available e.g. temperature, humidity, diurnal light cycles, wind, disturbance (Townsend et al., 2003). Conditions can, however, change the availability of resources by interacting with other environmental processes. For example, temperature and rainfall can influence nutrient cycles within the environment and therefore impact on nutrient availability (Eppinga et al., 2008). Resource availability

3 can, therefore, be affected by biotic and abiotic factors across a range of spatial and temporal scales.

The resource and condition limits within which individuals of a species can survive and reproduce is defined as its ecological niche (Chase & Leibold, 2003). These factors form axes of an n-dimensional hypervolume representing the environmental states in which a species can exist indefinitely (Hutchinson, 1957). The fundamental niche defines the upper and lower limits of environmental factors which determine the potential ecological space in which a species can exist (Peterson et al., 2015). The realised niche represents the niche dimensions of a species that is observed, which is distinct from the fundamental niche as it also includes ecological processes of movement and interactions with other organisms such as competition, facilitation and predation (Whittaker et al., 1973).

Species traits facilitate adaptation to the prevailing niche conditions (Cheplick, 2015).

A trait at the individual level is defined as any morphological, physiological or phenological feature measurable from cell to whole organism level (Violle et al.,

2007). These traits are quantified without reference to the environment or any other level of organisation and as such is simply defined and unambiguous. Traits, expressed as phenotypes in individuals, are related to the selection pressures within the environment and may provide a competitive advantage over other organisms within its geographical range (Frommlet et al., 2014). In addition to direct competitive ability, traits may also instigate facilitation with other organisms, enhancing at least one of the participants’ fitness without reducing the fitness of the other (Molles, 2015). Traits can therefore be important measures of growth, survival and reproduction (the three levels of individual performance) and can be a key determinant of species composition within different environments.

The expression of traits requires investment of energy or nutrients, of which are inherently limited or in competition for within the environment (Craine et al., 2005).

Allocation of resources are distributed among the three aspects of performance

4 which maximise fitness in the ecological niche in which they occupy (Roff &

Fairbairn, 2007). This optimisation of trade-offs between growth, survival and reproduction over evolutionary time define the life-history strategy of a species

(Charnov, 1989). The ability of a species to exist in an environment is therefore dependent on the availability of resources in that area as well as the life-history strategy of the plant.

There have been numerous models to explain the emergence of plant strategies

(Jenkins & Pierce, 2017). The unifying concept of these theories is that plants must make trade-offs within the environment to optimise fitness. Liebig’s law states that plant growth is restricted by the most limiting resources, rather than the total amount of resources available (Liebig, 1840, 1855). In accordance of Liebig’s law of the minimum, plant adaptation acts to allocate resources to compensate for this limitation (Allaby, 2010).

One of the first models to attempt to explain life history strategies within the environment is r/K selection theory (MacArthur & Wilson, 1967). This relates to the trade-off in quality and quantity of offspring, with increased quantity of offspring at one end of the spectrum (r-strategists), and reduced quantity of offspring at the other corresponding with increased parental investment and quality of offspring (K- strategists). Though r/K selection is still applied in biology, it is now considered ineffective at explaining organism’s life history traits because of its focus on reproductive strategies, when other factors may provide more causative links between the environment and optimal life history strategies (Reznick et al., 2008).

Key questions, therefore, are what are the key plant traits for certain conditions and how do these traits trade off in relation to the environment?

An issue with these questions is that there is no single optimal strategy that dominates an environment. Instead, there are co-existing species occupying the same habitats (Kunstler et al., 2016). There have been several attempts at explaining the coexistence of competing plants and communities (Kraft et al., 2015). Tilman devised a

5 mechanistic explanation of competitive interactions in plants based on resource utilisation, the resource ratio hypothesis (Tilman, 1985). The hypothesis is centred around how there can be co-existence of competing species by their differing abilities to compete for pairs of shared resources where the ratio of abundance of the resource varies. Each species differs in their use of the resource and in their impact on the shared resource.

Tilman’s model is useful in explaining how two species interact to compete for resources in an ecosystem (Thompson, 1987). Essentially plants will search for pockets of resources e.g. by investing in roots to explore areas of soil. As the plant takes up the resource, its availability in the environment decreases. When the plant dies, the resource is released into the environment and the cycle is then repeated.

When new plant growth and resource use is equal to the mortality of plants, the population is stable over time (Connell, 2014). Assuming other factors do not inhibit growth, when two species compete for the same resource, either the species that can continue growing with the least resources will eventually take over the space, or alternatively the species that can maximise resource uptake e.g. by proliferating roots in zones of high resource availability, will dominate the resource and outcompete conspecifics (Hodge, 2004). Competitive ability is, therefore, not only a function of uptake of a resource, but also on the efficiency of resource utilisation.

This means that there are multiple strategies that can potentially exist in similar environments. Tilman’s model is based on pairs of resources, however, and it has been highlighted that there are not enough different resources for plants to differentiate and specialise into niches following this model (Miller et al., 2005).

Plants can also co-exist in an environment by using different ratios of resources

(Miller et al., 2005). Niche differentiation explains the process in which co-existing species may be stable over time by utilisation of different resources (e.g. nutrients) within their environment or by occupying different areas within the environment

(Zuppinger-Dingley et al., 2014). For example, plant species may be distributed along

6 a water table gradient. Plant species preferring wetter areas are separated from congeneric species that preferentially grows in drier areas such as the carnivorous sundew plants Drosera anglica and Drosera rotundifolia resulting in a reduction in competition between the two species within their shared environment (Nordbakken et al., 2004). Where a species strongly overlaps in resource use in the same environment, it is likely that one species outcompetes the other, termed the competitive exclusion principle (Kraft et al., 2015).

Grime’s CSR theory is one attempt to define the key resources, conditions and life- history strategies, and how they are linked. This model considers that primary strategies are controlled by just two key factors: the amount of stress an organism experiences, and the amount of disturbance present in a habitat (Grime, 2006). Stress is defined as the external constraints which limit the rate of dry matter production of all or part of the vegetation e.g. nutrient limitation, low water availability, low light availability (Grime, 1979). Disturbance is defined as the mechanisms or events which limit the plant biomass by causing its partial or total destruction, the state from which tissue cannot recover such as herbivory, fire or vegetation removal (Grime,

2009). This creates four possible extremes. Under these environmental constraints, three primary strategies have evolved: C (competitive) strategists that use traits that maximize resource acquisition and control in consistently productive niches; S

(stress tolerant) strategists survive via maintenance of metabolic performance and thrive in variable and unpredictable niches; or R (ruderal) strategists that have rapid gene propagation via fast completion of lifecycle and regeneration in niches that are frequently lethal to the individual (high disturbance environments, e.g. areas of frequent grazing) (Grime & Pierce, 2012). This theory also predicts that there are environments that are not possible for plants to occupy, environments of high stress and high disturbance, termed the untenable triangle (Figure 1.1).

7 The theories of Grime and Tilman attempt to explain plant strategies, community

composition, responses and interaction with the environment (Tilman, 1985; Grime,

2006). Recently, areas have been highlighted where Grime’s theory does not

adequately incorporate the importance of non-heterogeneous supplies of nutrients

and how these supplies are allocated long term (Craine, 2005). It has also been

highlighted that the model needs to address the importance of disturbance in

nutrient limited habitats and the carbon balance of shade tolerant or temporarily

shaded plants. Grime’s model predicts that disturbance is a mechanism that

eliminates species at low nutrient supply (high stress) and that plants in low-

nutrient environments are “severely restricted in their ability to replace even quite

small losses of tissues or individuals” (Grime, 2007). Some studies have suggested,

however, the necessity of disturbance for some carnivorous plant species in high

stress environments for survival (Brewer, 1998, 2008). Tilman’s model, in addition, is

bi-variate and now falls short of being used to predict mechanistic patterns or

processes in natural populations in terms of diffusion of nutrients in the soil.

Stress K

C Untenable

Disturbance r R

Figure 1.1 Grime's CSR theory (1979) for plant strategies, C: competitors; S: stress tolerators; R: ruderals. Blue triangle represents environments plants can occupy. Red triangle represents theoretically uninhabitable environments of high stress and high disturbance (the untenable triangle). Wilson’s r/K model is overlaid, indicating how Grime’s model builds upon the previous r/K model (MacArthur & Wilson, 1967).

8 Furthermore, one aspect of Tilman’s model asserts that species in low-N habitats will respond with high resource allocation to roots. However, again in contrast, carnivorous plants are low-N specialists and have highly modified leaves with a typically reduced root system in biomass and complexity and in some cases the root system has been completely lost through evolution (Adlassnig et al., 2005a).

Attempts have been made to reconcile the approaches of Grime and Tilman, along with other models of plant ecology, to understand plant strategies, community composition and competition (Craine, 2005, 2007; Brooker & Kikvidze, 2008; Jabot &

Pottier, 2012), yet there are still important basic questions regarding the adaptation of specialist plants, and nutrient uptake and allocation in plant ecology. Particularly in environments highlighted above, of low nutrient or resource availability, more research is needed to understand how plants respond to resource availability across different spatial and temporal scales (Sutherland et al., 2013). Given carnivorous plants are potential exceptions to these general plant ecological models, carnivorous plants are particularly interesting for addressing questions about the effects of resource availability on the physiological and ecological responses of plants.

1.1 Carnivorous plants

Carnivorous plants capture, kill, assimilate and utilise nutrients from animal prey and are found on every continent except Antarctica (Adamec, 2013; Ellison &

Adamec, 2018). They are a diverse, polyphyletic group of flowering plants, generally restricted to nutrient-poor sites such as peatlands (Ellison & Adamec, 2018). Plant carnivory has evolved independently in at least ten lineages spanning five separate orders (Figure 1.2). Over 800 species of carnivorous plant have been recognised to- date (Givnish, 2015; APG, 2016). The distribution of these genera is highly variable.

For example, Drosera spp. are found globally, yet some genera are restricted to very local areas, e.g. The Venus flytrap, Dionaea muscipula (the sole species of the genus), is endemic only to peatlands of North and South Carolina.

9 Six distinct trap morphologies have been described: flypaper, pit fall, lobster pot, snap-trap, pigeon-trap and bladder traps (Figure 1.2 letter annotations). Some trap morphologies exhibit striking evolutionary convergence across carnivorous lineages

(e.g. pitcher plants, trap annotation “c”, Figure 1.2), but also divergence within lineages (Thorogood et al., 2017). The Nepenthales, for example, share a common carnivorous ancestor but have evolved species of flypaper, snap-trap and pitcher plants and consist of both terrestrial and aquatic species within the clade (Figure 1.2 purple block) (Ellison & Adamec, 2018). Others have only evolved on a single occasion within a genus of non-carnivorous plants, for example the recently discovered plants to be carnivorous: Catopsis berteroniana (Frank & O’Meara, 1984) and Paepalanthus bromeliodes (Nishi et al., 2013). Carnivorous plants, therefore, are an interesting group for studying a variety of evolutionary and ecological questions.

The benefits of acquiring additional nutrients from prey via modified leaves can be numerous. Increased growth rate, allocation to photosynthetic tissue and increased photosynthetic output has been shown in Dionaea muscipula (Gao et al., 2015),

Sarracenia spp. (Farnsworth & Ellison, 2007) and Drosera capensis (Krausko et al.,

2014). Nutrient uptake via roots after prey capture increases for Drosera spp. and

Dionaea muscipula (Adamec, 2002). Reproductive potential i.e. seed set, number of flowers and earlier flowering, is also shown to increase following prey capture

(Hanslin & Karlsson, 1996; Schulze et al., 2001; Adamec, 2002; Pavlovič et al., 2009), as well as increased seedling growth in Dionaea muscipula (Hatcher & Hart, 2014).

Carnivory, therefore, can enhance the fitness of carnivorous plants by increasing nutrient uptake stimulating increased carbon sequestration, higher growth rates and increased allocation to reproduction. Though these benefits, in general, only occur when plants are growing in low nutrient conditions.

10 Figure 1.2: Phylogenetic tree of carnivorous plants (non-carnivorous phylogenies removed). Classification follows APG (2016) and NCBI except for Caryophyllales which are divided from Nepenthales. Letters beside plants illustrate trap type: a: flypaper, b: snap-trap, c: pitfall, d: lobster pot, e: bladder, f: pigeon. Number of species (non-carnivorous species in parenthesis if any) and distribution derived from Ellison & Adamec, 2018. Coloured blocks signify a shared carnivorous ancestral lineage. Aldrovanda & Utricularia are aquatic. Illustrations by Angus Davidson.

11 Adaptations involved in the carnivorous habit, however, also have associated energetic costs, reducing the competitive ability of these plants in other areas

(Ellison & Gotelli, 2002; Pavlovič & Saganová, 2015). These costs include the two-fold carbon expense of leaf trap formation. These costs relate to the resource requirements for building the traps (though traps are generally quite ‘cheap’ to construct compared to non-carnivorous leaves), and the reduction in photosynthetic efficiency of the leaf as function shifts towards capturing and digesting prey

(Karagatzides & Ellison, 2009; Pavlovič & Saganová, 2015). Many carnivorous plants also synthesise polysaccharides to make sticky mucilage (Erni et al., 2008), nectar from extrafloral nectaries around traps (Płachno, 2007), volatile organic compounds

(Jürgens et al., 2009; Kreuzwieser et al., 2014) and digestive enzymes (Krolicka et al.,

2008; Kováčik et al., 2012a; Huang et al., 2015; Ravee et al., 2018).

The costs of carnivory in plants can be substantial. Ellison (2006) found lower rates of photosynthesis per unit leaf mass and whole plant growth in carnivorous plants than in any other non-carnivorous plants from a global database. Some species of

Drosera in southwestern Australia use 3-6% of their total photosynthetic output solely on the production of mucilage for prey capture (Pate, 1986). Uptake of nutrients also requires active transport across the leaf membrane to be incorporated into the plant (Renner & Specht, 2013). The result of these costs is that payback times for carnivorous leaves and photosynthetic net output tend to be much lower than for non-carnivorous plants (Karagatzides & Ellison, 2009). Consequently, carnivory is a favourable strategy only under specific environmental conditions, and as previously mentioned, is not the only strategy for plants to occupy nutrient poor environments.

The evolutionary adaptation of carnivory was first addressed in detail, as a cost/benefit model for the conditions under which marginal carbon gains can be made from increased investment in strategies to acquire nutrients from prey, using

Brocchinia reducta as a model species (Givnish et al., 1984). Given the costs of carnivory, benefits from botanical carnivory are predicted to be restricted to specific

12 environmental conditions, i.e. under environments of increasing light and water and

decreasing nutrient availability (Figure 1.3). Givnish et al.’s model for the

adaptiveness of carnivory has remained largely unchanged despite its application

across carnivorous genera distributed globally, and has been the conceptual

a b B3: fertile

B1: sunny, moist B1: infertile C: Cost C: Cost

B2: shady or dry B3 - C

B1 - C B1 - C Carbon gain or loss Carbon gain or loss

B2 - C Allocation to carnivory Allocation to carnivory c d C: Cost B1: sunny, moist C: Cost

B1 – C1 C1: Cost with pre- existing traits B5: sunny, moist, extremely infertile B – C

Carbon gain or loss 1 Carbon gain or loss

B4 - C Allocation to carnivory Allocation to carnivory

Figure 1.3: Cost/benefit model for the evolution of carnivory in plants, with photosynthetic output against investment in carnivory under different environmental or evolutionary scenarios: (a) in environments of low nutrient availability, the rate of photosynthesis (B, green solid lines) increases with investment in carnivory and shows less tendency to plateau in well-lit, moist sites (B1) compared to more light or water limited sites (B2) as carbon is the most limiting resource. Dashed lines indicate net difference in photosynthetic cost (C, yellow solid lines) and benefit of carnivory. Carnivory is evolutionarily favourable where there are marginal benefits as a result of allocation to carnivory i.e. a peak in dotted blue lines (B-C) where C is near 0. The optimal investment in carnivory occurs where the difference between the benefit and cost curves is highest. Where there is no peak in dotted line, there is no allocation to carnivory that is better than less allocation. (b) In sunny, moist environments, photosynthetic rate is greater on fertile sites compared to infertile sites. In fertile sites, carnivory is, therefore, unlikely to be favoured. (c) Relative cost of carnivory is likely to be reduced with the presence of pre-existing traits or precursors to carnivory such as glandular hairs, specific enzymes or metabolites already present. (d) Carnivory is unlikely to evolve on extremely infertile soils as the initial cost to allocation in carnivory shows negative carbon balance, even though further investment may provide net gains in carbon. Adapted from Ellison & Adamec, 2018. 13 framework for considerable research through which amendments have been proposed rather than any shifts in paradigm (Zamora et al., 1998; Ellison & Gotelli,

2002, 2009; Alcalá & Domínguez, 2005; Gibson & Waller, 2009; Pavlovič & Saganová,

2015; Gaume et al., 2016).

Givnish et al. (1984) predicted three main aspects of benefit from plants being carnivorous under these conditions: (i) an increase to the overall photosynthetic rate or leaf mass able to be supported, (ii) excess nutrients from prey can be invested in reproductive processes, (iii) carnivory may be an alternative or additional source of chemical energy such as carbon from polysaccharides, thus reducing the required output from photosynthesis. Givnish’s model has received much attention over the last 30 years with studies generally supporting the first two benefits, and less evidence supporting the importance of energetic gain from carnivory.

Though ultimately inconclusive on the effect of prey-derived nutrients on photosynthetic output of the plant (Méndez & Karlsson, 1999; Wakefield et al., 2005;

Adamec, 2008), there is an abundance of indirect evidence from increases in plant growth when plants are able to acquire nutrients from prey (Farnsworth & Ellison,

2007; Adamec, 2008; Hatcher & Hart, 2014; Krausko et al., 2014). The second measure of benefit has been observed in several carnivorous species. Plants that acquire prey- derived nutrients have been shown to flower earlier, have a higher seed set, and have higher success rates of seedling establishment (Thorén & Karlsson, 1998).

Increased investment in sexual and vegetative reproduction was also observed in

Drosera (Stewart & Nilsen, 1992), Sarracenia (Ne’eman, G., Ne’eman, R. and Ellison,

2006) and Pinguicula (Eckstein & Karlsson, 2001). There are only a few studies supporting the third criterion of benefit in heterotrophic uptake of carbon. Michalko et al. (2013) confirmed the production of glucanases from Drosera rotundifolia and the active absorption of saccharide-based compounds into leaf tissue. However, there is little evidence to support this hypothesis in situ, and the majority of carnivorous plants obtain mostly nutrients, not carbon, from carnivory (Adamec, 1997).

14 Patterns of the distribution of carnivorous plants generally support the predictions of evolutionary models (Figure 1.3, Ellison & Gotelli, 2001). The majority of carnivorous plants are found in sites that are nutrient-poor, sunny and moist, at least during the growing season (Juniper et al., 1989; McPherson et al., 2010; Pereira et al.,

2012). The model also rules out the direct evolution of carnivory to extremely infertile soils, explaining its rarity (though not exclusion) from these environments

(Figure 1.3 d). This is because on bare rock or extremely infertile soil, non- carnivorous ancestors are predicted to have negative carbon balance under any initial allocation of resources to carnivory. Instead, adaptation to these environments is likely to arise in a stepwise fashion, with initial investments on low nutrient sites, leading to plants being able to survive and adapt to extremely infertile sites after the initial allocation to carnivory under more favourable environments (Ellison &

Adamec, 2018). Seasonal growth or expression of carnivory explains other apparent exceptions. Some Drosera spp. occupy nutrient poor but semi-arid sites, but are only active carnivores during the moist winter and spring (Erickson, 1978) and butterworts, Pinguicula, stop producing mucilage during the dry season (Alcalá &

Domínguez, 2005).

Although carnivorous plants are generally unable to inhabit environments outside those stipulated in Givnish et al.’s (1984) model, there are some unexplained exceptions. For example the carnivorous Drosophyllum lusitanicum grows on arid

Mediterranean heathland and has an extensive, deep root system that may act as a taproot (Adlassnig et al., 2005a; Paniw et al., 2017). Some Drosera spp. inhabit the shaded rainforest floors of Queensland (Givnish et al., 1984; McPherson, 2008) and some occur on calcareous sands (Adlassnig et al., 2005a). Of the aquatic genus

Utricularia, some species grow in hard cation-rich waters or tolerate shade (Adamec,

1997; McCormick et al., 2011). There are also carnivorous plant species present in both open and shaded areas within the same local environment (Thorén et al., 2003;

Suárez-Piña et al., 2016).

15 An update to the cost/benefit model addresses some of these potential contradictions with one subtle modification that allows carnivory to occur in a variety of environmental conditions (Benzing, 2000; Ellison & Adamec, 2018). A phenotypically plastic or adaptive investment in carnivory would be evolutionarily favourable to occupy habitats of reduced light, water or increasing nutrient availability, to an extent, beyond the conditions originally considered as viable for carnivory to be a successful strategy. Though, originally, the model was designed to explain the evolution of plant carnivory, considering phenotypic plasticity under the same conditions enables the application of this model in ecological context.

1.2 Responses of carnivorous plants to resource availability

1.2.1 Mineral nutrition and responses to nutrient availability

An element or nutrient is essential to plant growth if: the plant is unable to complete a normal life cycle in the absence of the element; or the element is part of an essential metabolite or tissue (Epstein & Bloom, 2016). There are at least 17 elements considered essential nutrients for plants (Mengel et al., 2001). Plants take up essential nutrients usually from the soil through their roots, but can also take up nutrients through their leaves (Adamec, 1997; Harrison et al., 2000). The rate or efficiency of this nutrient uptake can be altered by the structure, form and physiology of the roots and leaves as well as symbioses of some plants such as mycorrhizae (Allen, 1992;

Gilroy & Jones, 2000; Brundrett, 2002; López-Bucio et al., 2003). Once sequestered, nutrients can be allocated within a plant to specific tissue, diverting resources to maintenance or production of new root or leaf tissue or diverting resources to reproduction at the cost of normal growth, though this varies based on the life history of the plant (Kurup et al., 2013). Nutrients can also be stored and recycled, usually occurring when certain resources are scarce or costly to acquire (Masclaux-

Daubresse et al., 2010).

16 Carnivorous plants mostly are stress tolerators sensu Grime’s CSR theory (Grime,

2006; Ellison & Adamec, 2018). They reutilise nutrients for new plant growth

(Adamec, 1997, 2002; Butler & Ellison, 2007), both during a growing season and in some species during senescence over winter. The efficiency of this reutilisation is high (N: 33-99%, P: 51-98%, K: 41-99%, Adamec, 1997, 2002) and on average (mean)

20-25% greater than co-occurring non-carnivorous plants (for N and P, Aerts et al.,

1999). Carnivorous plants also typically have very low growth rates (median 0.020 –

0.025 g g-1 d-1, Adamec, 2011; Ellison & Adamec, 2011; Kruse et al., 2014). Generally, foliar N and K in carnivorous plants (terrestrial) are two times lower than those for non-carnivorous plants, while P values are similar (Table 1.1). The low N and K contents are associated with the low growth and photosynthetic rates, which may lower the demands for N and P for normal growth in comparison to non- carnivorous plants (Ellison, 2006; Ellison & Adamec, 2011).

Table 1.1: Tissue nutrient concentration of nitrogen, phosphorous and potassium in leaves and shoots (N, P, K respectively, median dry weight %, simplified from Ellison & Adamec 2018).

Plant category N P K All plants 1.81 0.10 1.86 (n = 397-1973) Terrestrial carnivorous plants 0.88 0.11 0.91 (n = 51-94) Aquatic carnivorous plants 2.4 0.22 1.55 (n = 43-54) Carnivorous plants have the same mineral nutrient requirements as non-carnivorous plants, but they attain a range of nutrients from prey in addition to root uptake to compensate for the nutrient deficient environments they inhabit (Adamec, 2002).

Nutrient stoichiometry of carnivorous plants indicates co-limitation of N & P or N, P

& K (Ellison, 2006). In addition to N, P and K, carnivorous plants have been found to derive from prey: S, Na, Ca, Mg, Fe, Mn, Zn and Co (Adamec, 1997; Adlassnig et al.,

17 2009). The ecological and physiological significance of the nutrients derived from prey depends on the uptake efficiency and there are limited studies on the relative importance of most of these nutrients—the mineral nutrition of carnivorous plants is based on studies of only ca. 70 species (i.e. less than one tenth of all species) and most are focused on N.

Natural abundance isotope analysis, particularly of N, can reveal important information on the proportion of nutrients derived from prey compared to roots and can even identify carnivorous plants that are prey generalists to prey specialists

(Moran et al., 2001; Brearley, 2011). This is because δ15N accumulates under increasing trophic levels (Post, 2002). Insects are thus likely to have enriched δ15N within their tissues compared to the vegetation they feed upon. Sampling the δ15N abundance in non-carnivorous plants, carnivorous plants and their prey, therefore, enables the calculation of the proportion of prey derived N compared to roots

(Brearley, 2011). The proportion of nitrogen uptake derived from prey in carnivorous plants can be highly variable across and within species (Table 1.2), which is likely a consequence of their evolutionary history and local environmental resource availability (Adamec, 1997).

Carnivory has independently evolved under similar, but not identical circumstances which can explain the disparity in reliance on prey for N across carnivorous species

(Ellison & Adamec, 2018). Nutrient limitation in plants may be limited by absolute concentrations of available nutrients or by their relative concentrations (Aerts &

Chapin, 2000). Limitation of nutrients is therefore not only by nitrogen, and therefore prey derived N uptake may become less beneficial as other nutrients become limiting to performance. Reliance on prey for specific nutrients is therefore also likely to be variable across independent taxa as nutrient limitation varies across environments (Ellison, 2006).

18 Table 1.2: Variability in the proportion of nitrogen (N) derived from prey, compared to root uptake within and across select carnivorous plant species (adapted and updated from Ellison & Gotelli, 2001 and Ellison & Adamec, 2018). %N Trap Type Species derived Ref from prey Karlsson, 1988; Karlsson et al., 1994; Flypaper Pinguicula vulgaris 26 - 40 Hanslin & Karlsson, 1996 Karlsson, 1988; Karlsson et al., 1994; Pinguicula alpine 8 - 14 Hanslin & Karlsson, 1996 Karlsson, 1988; Karlsson et al., 1994; Pinguicula villosa 7 - 15 Hanslin & Karlsson, 1996 Thum, 1989; Schulze & Schulze, 1990; Drosera rotundifolia 22 - 63 Millett et al., 2003, 2012, 2015 Dixon et al., 1980; Watson et al., 1982; Drosera erythrorhiza 11 - 100 Schulze et al., 1991 Snap-trap Dionaea muscipula 45 - 75 Schulze et al., 2001 Pitcher Sarracenia purpurea 10 - 100 Chapin & Pastor, 1995 Most carnivorous plants maintain the capacity for nitrogen uptake from the soil via their roots (Adamec, 2002; Adlassnig et al., 2005a; Gao et al., 2015). There is a wide range of root types in carnivorous plants. Some species have roots that only function for part of the year (Slack, 2000), some have deep (likely tap-roots) roots (Paniw et al., 2017) and others have poorly developed roots or have completed lost their roots

(e.g. aquatic carnivorous plants), which in some cases has led to the stem and leaves replacing their function entirely (Adlassnig et al., 2005a). The majority of species, however, have retained some form of root system which takes up nutrients to varying extents (Adamec, 1997).

Root N uptake can be an important driver of phenotypic variability and investment in carnivory, which can explain intraspecific variability in reliance on prey for N along environmental gradients of N deposition (Table 1.2, Ellison & Gotelli, 2002;

Alcalá & Domínguez, 2005). For example, leaves of D. rotundifolia produce mucilage that is 45% less sticky after soil was fertilised (Thorén et al., 2003). Reliance on prey for N decreases linearly with increasing N deposition in D. rotundifolia (Millett et al.,

2012). Sarracenia spp. similarly become less carnivorous under increasing N levels by

19 producing more phyllodia (non-carnivorous leaves) compared to traps (Ellison &

Gotelli, 2002). Reduction in investment in carnivorous plants can also present as lower trichome or digestive gland density, changes in pitcher trap morphology i.e. narrower pitchers with wider keels (photosynthetic portions of leaves) or changes in the number of bladder traps (Knight & Frost, 1991; Ellison & Gotelli, 2002; Thorén et al., 2003; Alcalá & Domínguez, 2005). These changes occur because high levels of root N uptake results in less demand for whole-plant N and therefore a reduction in expression of carnivorous traits as the energetic benefits of carnivory are diminished under increasing N availability (Figure 1.3 b). Measurements of root and prey derived N acquisition is therefore a useful indicator of the effects of plant trait responses to resource availability.

1.2.2 Responses to light

Light, or specifically, photosynthetically active radiation, is an essential resource to facilitate plant nutrition and growth (Craine & Dybzinski, 2013). To deal with light limitation, plants can adopt strategies of shade avoidance or tolerance, the latter arising when avoidance is not possible e.g. species from forest understories that cannot outgrow competitors (Gommers et al., 2013). Even within an individual, expression of leaf traits in response to varying light availability can occur, such as sun and shade leaves (Gommers et al., 2013). The shade avoidance syndrome is characterised by longer stems, wider leaves, lower specific leaf area i.e. the mass per unit leaf area as a strategy to intercept more light in shaded environments, earlier flowering and increased apical dominance (Casal, 2012). Shade tolerant plants typically invest in molecular adaptations, such as utilising far-red light (a consequence of competing vegetation intercepting more red light), have higher photosynthetic and nutrient use efficiency and grow broader leaves to intercept more light with lower cost for producing the leaf (Gommers et al., 2013). Leaves are normally for the sole purpose of carbon sequestration and is rarely or is negligible in the role of nutrient uptake from the environment (Gulman & Chu, 1981; Buchanan et

20 al., 2015). For carnivorous plants, however, light availability is predicted to have an impact on both photosynthetic and carnivorous traits. When light availability is limited, allocation to carnivory is predicted to be low, when considering Givnish’s model in an ecological context (Figure 1.3 a, Givnish et al., 1984). This is because carbon assimilation decreases as a consequence of low photosynthetic output

(Karagatzides & Ellison, 2009), and the extent to which carnivory can enhance overall carbon gain is diminished (Ellison & Gotelli, 2009). Resource limitation shifts from nutrition to carbon and allocation of investment is directed to carbon uptake and away from the primarily carbon-based carnivorous adaptations (Méndez &

Karlsson, 1999). In these conditions, a reduction in allocation to carnivory results in a marginal gain in net carbon uptake. The optimal strategy is, therefore, one of reduced investment in carnivory under increased shade (Figure 1.3 a).

Relatively few studies focus on how light availability affects investment in traits specifically linked to carnivory (Zamora et al., 1998; Thorén et al., 2003; Alcalá &

Domínguez, 2005). The butterwort, Pinguicula vallisneriifolia, increased mucilage secretion (a sticky glue used to capture prey) along a gradient of increasing light availability (Zamora et al., 1998). Pinguicula moranensis also increased mucilage and digestive gland density in response to increasing light along an environmental gradient (Alcalá & Domínguez, 2005). Similarly, the round-leaved sundew, Drosera rotundifolia, under controlled botanical garden conditions reduced the stickiness of its mucilage by 14% due to shading, though this was significantly lower than the effect of addition of soil N (45%), and the shading effect varied with soil N availability (Thorén et al., 2003). Mucilage production by the rainbow plant Byblis guehoi, did not, however, vary along a gradient of light intensity in a greenhouse experiment (Lam et al., 2018). Byblis guehoi, in this experiment, showed patterns of anthocyanin accumulation under high light levels and etiolation responses to low light, confirming that light levels were sufficient to impact on photosynthetic output and non-carnivorous trait responses. These findings suggest impacts of light on

21 resource allocation to carnivory varies across taxa, potentially due to environmental conditions specific to each species. There is missing information on the impact of temporal variability in light availability, particularly in situ, which might be important because if carnivorous plants are able to tolerate periods of shade by altering investment in carnivorous traits, they are able to extend the range of habitats in which they can persist.

Some carnivorous plants have been shown to persist in shaded environments

(Zamora, 1995; Zamora et al., 1998; Thorén et al., 2003), and in some cases may prefer shade (McPherson, 2008), which is at odds with predictions of favourable environments for carnivory to exist according to the cost-benefit model. Though prey capture is in some cases reduced in shaded habitats, the carnivorous habit is maintained, prey are still captured and plants persist in these areas (Nordbakken et al., 2004). This may be due to limits on the plasticity of carnivorous traits (Auld et al.,

2010). What is not clear, however, is how carnivorous plants are able to persist in shade, because they are predicted to be outcompeted in these environments by non- carnivorous plants (Ellison & Gotelli, 2009).

1.2.3 Responses to water availability

Water is integral in plants for almost all metabolic processes, nutrient transport and plant cell structure (Tanaka & Makino, 2009). Plants typically adapt to drier environments with enhanced root systems in addition to specific leaf characteristics

(i.e. curled leaves, waxy cuticle etc) (Fahn & Cutler, 1992). Carnivorous plants attain nutrients from leaves and generally have a reduced root system (Adlassnig et al.,

2005a), hypothesised to be a consequence of allocation of investment from root nutrient uptake to leaf nutrient uptake due to anoxic soils which increase the cost of efficient roots (Juniper et al., 1989; Adlassnig et al., 2005a). Consequently, most carnivorous plants are poorly equipped to deal with arid environments. If investment in carnivory is indeed inversely related to root allocation for nutrient acquisition and not water, the presence of carnivory will decrease under

22 environments of increasing water limitation because of the increased cost-to-benefit ratio of investment in carnivory as water availability decreases, resulting in carnivorous plants being at a competitive disadvantage in drier habitats (Givnish et al., 1984).

Studies testing the hypothesis that carnivorous plants are restricted to wet environments is limited. Pinguicula vallisneriifolia had higher fitness gains from increasing prey and light availability in environments where water was not limiting

(Zamora et al., 1998). Some species of carnivorous plant persist through dry periods through seed or bud dormancy or water storage (Dixon et al., 1980; Juniper et al.,

1989; Luken, 2007). Brewer (2011) found that, compared to non-carnivorous plants,

Sarracenia and Drosera spp. responded more negatively when soil moisture was reduced, primarily due to intolerance of drier conditions (regardless of presence of competitors which also negatively impacted carnivorous plants) due to low root depth and porosity. Though water availability has not been directly manipulated to identify changes in the investment of carnivory, carnivorous plants are restricted to wet environments, perhaps more so than non-carnivorous plants due to the nature of the carnivorous syndrome being based in leaves at the potential expense of roots.

Furthermore, aspects of carnivory require the use of water to function, such as the snap-trap of Venus flytrap following stimulation of mechanosensory hairs through turgor pressure (Figure 1.4 a), mucilage of Drosera spp. and Pinguicula spp. (Figure

1.4 b & e) or the rain filled pitcher plants of Sarracenia spp (Figure 1.4 c & d), which is distinct from the self-filling pitchers of Cephalotus (Figure 1.4 f) and Nepenthes.

23 a b

c d

e f

Figure 1.4: Trap morphology, secondary metabolites in carnivory, prey capture and evolutionary convergence and divergence of carnivorous plants; a: Venus flytrap snap-trap requires a change in turgor to flip the lobes from convex to concave for rapid closure. Trigger hairs are visible on the lobe; b: Drosera capensis leaf bending caused by turgor pressure due to accumulation of jasmonates; Dionaea (a) and Drosera (b) share a common carnivorous ancestor but have divergent trap morphology; c: Sarracenia purpurea filled with rain water filled digestive fluid; d: close up of Sarracenia purpurea with rain filled digestive fluid and drowned prey; e: Pinguicula sp. is a flypaper trap of independent origin to Drosera (b); f: Cephalotus sp. is a pitcher plant of independent origin to Sarracenia which fills and controls pitcher fluid without rain. Photos by Christopher R. Hatcher.

24 1.3 Secondary metabolites and carnivorous plants

There are multiple systems involved in carnivory to enable a response to the environment to facilitate carnivory, i.e. the need to attract prey, respond to prey capture, instigate leaf movement, digest prey, assimilate nutrients and utilise nutrients derived from prey (Ellison & Adamec, 2018). These processes require a combination of morphological and physiological adaptations that have evolved and in some cases are exapted from previous adaptations (Nakamura et al., 2013;

Pavlovič & Saganová, 2015). Digestive enzymes and trichomes (Gaume et al., 2016;

Ellison & Adamec, 2018; Ravee et al., 2018), for example, and their utilisation in breakdown of prey and protection of trap tissue during prey decomposition, originates from pathogenic and herbivore defence (Rischer et al., 2002; Renner &

Specht, 2013). There remains aspects of plant carnivory at the molecular level, however, of which we know little.

Secondary plant metabolites provide an explanation for how carnivorous plants interact with the environment to facilitate prey capture. Primary metabolism is involved in all processes essential for growth and development (Rees, 1994).

Secondary metabolism, however, is not vital for these processes but is integral to the survival of an individual within its environment (Hartmann, 2007). It is because growth can continue in the absence of secondary metabolism that these compounds can be continuously modified and adapted under environmental selective pressure

(Moore et al., 2014). The use of secondary metabolites by plants to take on a diverse range of functions has led to the evolution of ca. 200,000 distinct secondary plant metabolites on record and an estimated 1 million unique compounds in the plant kingdom (Hartmann, 2007). Secondary metabolites are involved in regulating growth, responding to abiotic stress, defence and competition, attraction and stimulation (Hartmann, 2007; Kessler & Baldwin, 2007; Agrawal, 2011).

It is likely, therefore, that secondary plant metabolites are key aspects of facilitating plant carnivory by responding to the environment. The biochemical components of

25 plant carnivory, however, have not been extensively studied. This represents a critical knowledge gap because growing understanding of the importance of the role of metabolites in rapid plant diversification and adaptation means that important insight into the evolution of plant carnivory and how carnivorous plant species can respond to changes in the environment is missing (Theis & Lerdau, 2003; Wink,

2003). Chapter 4 provides a comprehensive review of secondary metabolites involved in carnivory.

Briefly, studies have shown that carnivorous plants utilise a number of secondary metabolites to facilitate the carnivorous habit (Krolicka et al., 2008; Banasiuk et al.,

2012; Król et al., 2012). Of those documented, these consist mostly of jasmonates (for production of digestive enzymes or leaf bending in response to prey capture, Figure

1.4 a & b respectively), napthalenes and naphthoquinones, juglone and plumbagin, phenolic acids, flavonoids and volatile organic compounds. The origin of secondary metabolites in carnivory is largely associated with defence, similar to enzyme production, but many aspects of prey attraction are linked to mimicry of fruit and floral scents or pigmentation and represent the majority of known secondary plant compounds involved in carnivory (Jürgens et al., 2009; Renner & Specht, 2013)

(Figure 1.5). These studies, however, have mostly been targeted analyses of a few known metabolites found in other plants or only sample the volatile emissions in the air, rather than directly analysing trap tissue, and likely represent only a fraction of the metabolic diversity associated with carnivory within this unique group of plants

(Hotti et al., 2017).

26 Figure 1.5: Secondary plant metabolites involved in the attraction of prey. Left: Nepenthes khasiana under normal light. Right: Nepenthes khasiana under UV light shows pigment accumulation around the peristome of the trap to attract prey. Photographs courtesy of Dr Sabulal Baby (Kurup et al., 2013).

While volatiles emitted from plant traps have been investigated from sampling the

air (Jaffé et al., 1995; Jürgens et al., 2009; Kreuzwieser et al., 2014), no exploratory or

untargeted experiment has been conducted on the metabolites within leaf trap tissue

of carnivorous plants, nor of the metabolic response to prey capture to facilitate prey

retention, digestion and assimilation. Environmental metabolomics provides a

potentially powerful approach for understanding the biochemical basis of nutrient

uptake in these plants. Metabolomics can be deployed in a targeted or non-targeted

approach that generates a comprehensive biochemical profile of the plant (Di Guida

et al., 2016; Viant et al., 2017). Metabolic profiling of trap tissue before and after prey

capture may provide insight into the biochemical basis of carnivory and progress

our understanding of the evolution of plant carnivory and the role of secondary

plant metabolites in adaptation to the environment.

There are therefore several biotic and abiotic factors at different spatial and temporal

scales associated with investment and plastic responses of carnivory. The presence

and intensity of carnivory is a combination of these resources’ availability within the

environment and the plant’s evolutionary history.

27 1.4 Habitats

Habitats vary over a range of spatial and temporal scales (Cramer & Willig, 2005;

Van Dyke, 2008). Even over very small spatial (< 1 m) and temporal scales, the biotic and abiotic environment can vary dramatically (Cramer & Willig, 2005; de Smedt et al., 2018). This variability can be a consequence of a variety of mechanisms such as seasonality, climate, management or instances of disturbance (Hegazy et al., 2005;

Eppinga et al., 2010; Palacio-López et al., 2015). Habitat heterogeneity is realised at the organismal level by differences in resource availability or limitation, which control patterns of species distribution, and so biodiversity by altering niche diversity (Cramer & Willig, 2005). To be able to exist across a range of conditions found at a site, a species must adapt through genotypic adaptation or phenotypic plasticity or variation (Hutchings et al., 2000; Crowley et al., 2013). Understanding plant responses across a range of spatial and temporal scales is therefore important for determining community composition and predict impacts of environmental variability.

Accurate prediction of species’ responses to habitat variability at different scales is now a critical issue as there is a need to understand how key factors associated with climate change, such as temperature and rainfall are predicted to impact on plant traits directly and indirectly through light availability, nutrient cycling processes and interaction of these processes at different scales (Seddon et al., 2016; Cohen et al.,

2018; Madani et al., 2018). This research is imperative for understanding globally significant ecosystems and ecosystem health of environments that either mitigate global carbon emissions or provide habitat for endemic or unique species. Peatlands are environments that meet both criteria (Leifeld & Menichetti, 2018), in that they store twice as much carbon than all the forests in the world, an estimated 30% of all soil carbon, only cover 3% of the Earth (Josten & Clarke, 2002; Frolking et al., 2011), and often are inhabited by the polyphyletic, and often highly endemic, group of carnivorous plants (Jennings & Rohr, 2011).

28 Peatlands are the habitat for many carnivorous plant species, due to their low nutrient, waterlogged conditions. They are found globally and have high heterogeneity across a range of scales (Rydin et al., 2013). This thesis is mainly concerned with temperate ombrotrophic (rain fed) peatlands (peatlands hereafter).

Generally, peatlands are categorised as (i) raised bogs that are domed or plateau, (ii) non-raised bogs that are flat and riparian, and (iii) blanket bogs where there is high rainfall and low evapo-transpiration over large expanses of land (Rydin & Jeglum,

2006).

Within this general characterisation, however, peatlands have high habitat heterogeneity. The emergent microforms can be grouped as raised drier hummocks and lower, wetter hollows and lawns (Van Der Molen et al., 1994; Rietkerk et al.,

2004; Rydin & Jeglum, 2006). Hummocks also generally have higher vascular plant abundance as a consequence of hydrology which in turn impacts on the light environment below the canopy (Rydin et al., 2013). Nutrient availability also varies across peatland microforms, but the way it varies differs between different peatlands

(Eppinga et al., 2010). Therefore, it is clear that there is habitat heterogeneity over a variety of niche components, but the way these components interact may change fundamentally across environmental gradients and have very localised impacts on plant traits (Madani et al., 2018).

Peatlands can be characterised under Grime’s CSR model as stressful environments.

They are generally reliant on wet and dry atmospheric deposition for most of the nitrogen supply (Lamers et al., 2000). This makes these environments nutrient deficient (Rydin et al., 2013). They are also predominantly low pH environments due to the acidifying nature of Sphagnum moss (though in some cases can be alkaline,

Givnish et al., 1984; Clarke & Moran, 2016), a dominant component of the flora in these ecosystems, and anaerobic due to the obstruction of oxygen flow from the atmosphere to the organic-rich, decomposing substrate because of its consistently wet conditions (Clymo & Hayward, 1982; Rydin et al., 2013; Novak et al., 2015). Due

29 to the frequent or permanent waterlogging of peatlands, nutrient deficiency, high acidity and anoxia (Berendse et al., 2001), only a relatively small number of plant species are able to survive, but are often endemic only to these environments (Rydin et al., 2013).

Peatland vegetation is typified by an abundance of Sphagnum moss in terms of both biomass and species, as well as other species adapted to nutrient deficient conditions

(Rydin et al., 2013). Bog communities also typically contain shrubs, graminoids, and forbs or herbs, including: bryophytes, lichens and liverworts (Pastor et al., 2002;

Rydin et al., 2013). Species typical to Europe include common/ling heather, Calluna vulgaris, and cottonsedge, Eriophorum vaginatum (Rodwell, 1991). These species have evolved a variety of morphological and physiological adaptations to the nutrient- deficient conditions (Rydin & Jeglum, 2006). Adaptations include mycorrhizal associations (e.g. arbuscular mycorrhizal fungi associations with Calluna Vulgaris,

Yesmin et al., 1996), high internal nutrient recycling efficiency (e.g. Eriophorum vaginatum, Rydin & Jeglum, 2006), evergreen leaves for nutrient retention (e.g. Erica tetralix Aerts et al., 1999), and carnivory. There are, therefore, multiple strategies for plants to exist in these environments, and understanding plant species presence and responses across spatial and temporal resource availability is important for determining ecosystem processes and community composition.

Models for the presence of carnivorous plants predict that they are restricted to open, sunny areas of nutrient poor peatlands (Givnish et al., 1984; Ellison & Adamec,

2018). However, as previously mentioned, there are exceptions where carnivorous plants persist in shade. Carnivorous plants may only be able to temporarily tolerate shade, and it has been proposed that, to persist in the environment without being outcompeted, they require disturbance (Brewer, 1999). This scenario provides an interesting setting in which to assess the role of disturbance. This is because they are stressful environments, and according to Grime’s CSR model, areas that experience high disturbance and stress are predicted to be untenable environments for plants to

30 exist (Figure 1.1, Grime, 2006). Disturbance, trampling, grazing or land management is common in peatlands, yet few studies have investigated the impact of these disturbances on the fitness and responses of carnivorous plants (Turetsky & St.

Louis, 2006; Robroek et al., 2010).

There is some support for the reliance of carnivorous plants on disturbance.

Peatlands are typically permanently waterlogged environments, but paradoxically experience frequent fire regimes which some carnivorous species benefit from by increased seedling establishment and reduced competition through absent niche complementarity and comparatively higher growth rates of carnivorous plants following fire mediated disturbance (Brewer, 1999, 2001; Garrido et al., 2003; Kesler et al., 2008; Paniw et al., 2015). Seedling establishment of the Pink Sundew, Drosera capillaris, is also found to be higher in habitats where vegetation is removed (Brewer,

1998, 2008), but currently the effects of disturbance on mature individuals that are persisting in shade is not well known. An experiment to measure how disturbance mediated increases in the light availability of a naturally shaded population of carnivorous plants in situ impact on the expression of carnivorous traits would address whether existing models for botanical carnivory can successfully predict resource allocation and carnivorous plant responses to varying light availability under ecological scenarios of natural populations. This question is addressed in

Chapter 2.

1.5 Plant responses to climate

Research over the past few decades has studied plant responses to changes in soil nutrients, hydrology and light availability (Falster & Westoby, 2005; Ordoñez et al.,

2009; Douma et al., 2012), there is a major lack of information, however, on how plant responses to these abiotic factors are indirectly influenced by local climate patterns.

Climate regimes indirectly influence nutrient cycling, which may vary across site microforms and among biomes (Eppinga et al., 2008). Temperate ombrotrophic

31 peatlands typically receive all of their nitrogen from atmospheric deposition (Lovett et al., 2009), which subsequently moves with water flow (Pastor et al., 2002). This pattern of water and nutrient movement to hollows or hummocks depends on the primary driver of water loss from the site—drainage correlated to precipitation or water loss by evapotranspiration (ET:P) (Eppinga et al., 2010). This relationship between nutrient transport and ET:P functions along a gradient, such that intermediate sites will have equal nutrient distribution (accumulation) between hummocks and hollows (Rietkerk et al., 2004; Larsen et al., 2007). Therefore, climate change may have a critical impact on plant traits within peatlands, the impact of which may change fundamentally across environmental gradients due to the complex interactions of climate and habitat heterogeneity of water, nutrients and light availability across spatial and temporal scales (Madani et al., 2018). What impact do these climatic factors (evapotranspiration and precipitation) have on processes within peatlands, and to what extent does climate driven nutrient cycling impact on plant traits along an environmental gradient on a continental scale? If habitat heterogeneity is important in determining plant trait expression and species distribution, and that this varies between bogs and in response to climate change, then management to establish heterogeneity might help to mitigate the negative impacts of climate change. This forms the basis of Chapter 3.

The resilience of plants to climate change may be strongly dependent on the production of secondary plant metabolites (Theis & Lerdau, 2003; Iason et al., 2012;

Reen et al., 2015). Secondary plant metabolites are increasingly being considered as important drivers of plant diversification and evolution (Stevenson et al., 2017). This is in part due to their involvement in a large number of fundamental processes associated with plant fitness: regulating growth and interacting with the environment, conspecifics, competitors and symbionts and as a defence against abiotic and biotic stress (Hartmann, 2007; Kessler & Baldwin, 2007; Agrawal, 2011).

Plants retain a high metabolic diversity and have the genetic capacity to quickly

32 generate novel compounds due to gene duplication and divergence (Pichersky &

Gang, 2000). Plants are therefore able to potentially adapt to new environments at higher rates than explained by mutation (Moore et al., 2014).

If secondary plant metabolites play an important role in plant evolutionary adaptation, they should be inherently involved in processes specific to the characteristics of the plant’s ecology. The role of metabolites in adaptation may, therefore, be greater than previously considered, which would imply that we are missing important information on plant evolutionary ecology (Nishida, 2014; Moore et al., 2014). Consequently, similar patterns of metabolite production may emerge among close and distantly related taxa that have evolved under similar environmental conditions. Chapter 4 seeks to determine the patterns of metabolite production to facilitate carnivory across taxa, and Chapter 5 aims to determine the extent (i.e. a metabolic profile) to which secondary metabolites are important in the response of plants to prey capture.

1.6 Research questions

This chapter has highlighted carnivorous plants as model study systems, established the current understanding of carnivorous plant energetics, the role of resource availability on plant functional traits and the global importance of peatlands and study of these environments. This synthesis identifies key research areas that need to be addressed and how certain empirical studies will contribute to key discussions in the field. In this thesis, I have measured the effects of resource availability on carnivorous plant responses, influenced by disturbance, direct and indirect impacts of climate and foliar nutrient addition. I have done this experimentally in situ and ex situ and through a continental study to obtain a picture of the patterns present and the driving mechanisms behind plant trait responses in terms of phenotypic plasticity and intraspecific trait variation. I have used Drosera rotundifolia (the round- leaved sundew) in situ and Drosera capensis (the Cape sundew) ex situ. Drosera

33 rotundifolia is common throughout Europe (and globally) and is often present on ombrotrophic peatlands (Stewart & Nilsen, 1992; Nordbakken et al., 2004; Baranyai &

Joosten, 2016). This species is therefore a model species for determining plant responses to resource availability and predicting patterns and processes at a continental scale. Drosera capensis was utilised in the laboratory because it is an easier plant to grow, is adapted to warmer environments more closely matched to greenhouse conditions, provides more leaf tissue for chemical analysis per leaf and has been used for targeted analysis of secondary plant metabolites previously

(Adlassnig et al., 2005b; Bruzzese et al., 2010; Mithöfer et al., 2014). It is therefore a useful model system for studying carnivory ex situ. The overall aim of this thesis is to determine plant responses to resource availability at a range of temporal and spatial scales. This was achieved using carnivorous plants as a model system to determine potential impacts of climate change on heterogeneous environments and the importance secondary metabolites may have in the resilience of plants to climate change.

The thesis addresses the following overarching research questions:

(i) To what extent is the carnivorous habit of Drosera rotundifolia

phenotypically plastic in response to disturbance?

(ii) Do climate and microtopography interact at small spatial scales to impact

on intraspecific trait variation of Drosera rotundifolia?

(iii) What is the role of secondary metabolites in plant carnivory?

(iv) What is the metabolic response to prey capture of Drosera capensis?

1.7 Thesis structure

This thesis comprises six chapters. To address the research questions, the four research chapters are presented as three standalone experiments and a review. This introductory section (Chapter 1) has outlined the importance of understanding plant responses to varying resource availability at different spatial and temporal scales

34 and identified areas that require further research. Chapter 2 presents the results of an in-situ manipulative study investigating the impact of vegetation disturbance on carnivorous trait expression and fitness in Drosera rotundifolia. Chapter 3 presents the results of a multi-site, in-situ study of the impact of habitat heterogeneity within patterned ombrotrophic peatlands on D. rotundifolia nutrition and carnivorous trait expression. Variation in within-site habitat heterogeneity was investigated along a climate gradient across Europe. Chapter 4 presents a systematic review of existing studies investigating the role of secondary plant metabolites in botanical carnivory.

Chapter 5 presents a metabolomic investigation of the response of D. capensis to prey capture. The aim of this is to understand the biochemical responses to prey capture.

The final chapter (Chapter 6) presents a synthesis of the four research chapters, in the context of existing understanding of plant carnivory, general plant responses and plant ecology and understanding the potential impact of climate change.

35 Chapter 2 Phenotypic plasticity in the carnivorous habit of the round-leaved sundew, Drosera rotundifolia, to disturbance

36 2.1 Introduction

Disturbance and stress are the two fundamental processes predicted to control the evolution of plant strategies (Grime, 2006). Disturbance is the mechanism which limits plant biomass by causing its partial or total destruction (Grime, 1979, 2009;

Palacio-López et al., 2015). Stress is defined as the external constraints which limit the rate of dry matter production of all or part of the vegetation e.g. nutrient limitation, low water availability, low light availability (Grime, 1979). These two processes can be spatially and temporally variable, which creates habitat heterogeneity (Cramer & Willig, 2005; Van Dyke, 2008). This heterogeneity controls the distribution of species within a habitat, and influences biodiversity by altering niche diversity (Bergholz et al., 2017). Understanding plant responses to variability in stress and habitat disturbance is therefore of critical importance for understanding plant strategies in heterogeneous environments and predicting species distributions.

Real species distributions, however, often do not match predictions, because species are able to adapt to their environment to some extent, expanding their niche space

(Benito Garzón et al., 2019). Disturbance, for example, can be avoided by some species within a habitat, resulting in a selective disturbance (Wilson & Lee, 2000).

Selective disturbances impact a habitat and limit biomass, but not the biomass of all species within the area of disturbance. Instances of selective disturbance do, however, change the environmental conditions in which the surviving species exist.

Species that remain in this disturbed area, therefore, need to be able to cope in both the pre- and post-disturbance environments.

One approach for a species to be able to exist across a range of conditions, which may occur in instances of selective disturbance, is phenotypic plasticity, the capacity for a genotype to produce multiple phenotypes in different environments

(Schlichting, 2002; Gratani, 2014). Plasticity, therefore, provides a mechanism for surviving in habitats which change, such as through selective disturbance (Reed et al., 2011). Such plastic responses may be a fundamental component of habitat

37 disturbance adaptation. CSR theory predicts that adaptation to forms of disturbance is not possible in habitats of high stress (Craine, 2005), so measuring disturbance responses of plants in stressful environments may provide new information explaining how plants may exist in Grimes’ untenable environment (Grime, 2006).

Carnivorous plants are ideal systems to study the extent to which phenotypic plasticity is key to survival in high stress environments that undergo instances of selective disturbance within the lifetime of each generation. Carnivorous plants supplement low root nutrient availability with the capture and digestion of animal prey (Ellison & Adamec, 2018). The trade-off between the costs and benefits of carnivory mean carnivory in plants is predicted to only evolve in high stress, open, wet habitats (Givnish et al., 1984). It does appear, however, that some carnivorous plant species are able to tolerate a degree of shade (e.g. Millett et al., 2018), thereby existing outside their predicted niche. Brewer (1999) predicted that the response to habitat variability, in this case disturbance, was an important factor in establishing the success and distribution of carnivorous plants through seedling establishment and growth in recently disturbed microhabitats. Seedling proliferation in recently disturbed habitats, however, does not explain how carnivorous species persist in shaded environments. Little work has been undertaken to understand how carnivorous species respond to selective disturbance (Ellison et al., 2003).

Understanding the interaction between response to shade and selective disturbance will provide important information on the evolution of plant strategies and more accurate predictions of the distribution and restriction of species.

For carnivorous plant species, we might expect phenotypic plasticity to be expressed in the context of the existing cost-benefit model (Givnish et al., 1984). While this was originally conceived as an evolutionary model, it should also be relevant ecologically; these same trade-offs are relevant during the life of a plant, where environmental conditions change, such as through disturbance. When applied in an ecological context this model predicts that investment in carnivory will follow the

38 relative availability of root nutrients and light. Increased light availability will result in increased investment in carnivory. Such responses have been observed in ex-situ studies (Thorén et al., 2003), and phenotypic variability has been observed (Ellison &

Gotelli, 2002; Bott et al., 2008). Phenotypic plasticity has not yet, however, been demonstrated in-situ in response to selective disturbance.

In this study, I investigated the response of heavily shaded individuals of the small, carnivorous herb Drosera rotundifolia, to simulated selective disturbance. Drosera rotundifolia is predominantly found in open habitats on ombrotrophic (rain-fed) peatlands (Crowder et al., 1990), but is also known to tolerate and persist in shade, where it is subordinate to the shrub canopy in these habitats (Baranyai & Joosten,

2016; Millett et al., 2018). Disturbance from biotic (e.g. grazing, trampling, herbivory, climate) (Alonso et al., 2001; Robroek et al., 2010) and anthropogenic factors (e.g. trampling and management) (Arnesen, 1999; Turetsky & St. Louis, 2006) is a ubiquitous characteristic of peatlands. These selective disturbances typically damage peatland shrubs (i.e. Calluna vulgaris), but prostrate plants, such as D. rotundifolia, remain intact. Drosera rotundifolia trap insect prey using sticky mucilage on the end of leaf trichomes (Huang et al., 2015); the extent of the allocation of resource to this carnivorous trait is easily measured, along with the impact on prey N uptake (Millett et al., 2003; Erni et al., 2008; Cook et al., 2018) making D. rotundifolia an ideal study species for measuring phenotypic response to a variety of abiotic factors. My aim was to determine the extent to which the responses found ex-situ are realised in-situ, and in response to ecological manipulation. I hypothesised that traits involved in carnivory will be phenotypically plastic in response to selective disturbance

(vegetation removal). I predict that in response to removal of the surrounding plant vegetation, D. rotundifolia plants will (i) increase investment in carnivory through an increase in the number and density of leaf trichomes, due to increased light availability, (ii) increased investment in carnivory will result in increased N uptake from prey, (iii) 15N proportion will be higher, (iv) flower more and grow larger.

39 2.2 Methods

2.2.1 Site information

This experiment was carried out on a rain-fed (ombrotrophic) peat bog,

Humberhead Peatland National Nature Reserve, Thorne Moor, South Yorkshire, UK

(53.6359° N, 0.9071° W). Mean annual precipitation is 574 mm, mean annual temperature is 10.0 °C (1981-2010, Met Office, 2018), mean annual N deposition is ca.

22.54 kg N ha-1 yr-1 (2013-2015, APIS, 2016a). This peatland has been partially excavated for peat in the past, but peat extraction ceased in 2004 and site management is now focussed on restoration of previous species composition. The study site is an area of uncut peatland surrounded by previous hand cutting. The area has ditches which have been recolonised by Sphagnum spp., separated by uncut raised baulk which consist of a community of Sphagnum species (predominantly

Sphagnum cuspidatum, S. imbricatum, and S. campactum) under a dense shrub canopy

(mostly Calluna vulgaris). Populations of D. rotundifolia are present under the shrub canopy.

2.2.2 Experimental set-up

I selected twenty 50 cm ´ 50 cm plots quasi-randomly on one baulk in a 30 m ´ 80 m area. Locations of the plots were randomly selected within areas that were not on the edge of the baulk, were naturally covered by vegetation, not characterised by a rare or non-native community to the environment and contained at least 100 individual

D. rotundifolia. Plots were arranged in five groups of four by proximity. Within each group of four plots, each plot was randomly allocated to one of four treatments, giving a randomised complete block design.

The plots were subjected to one of four treatments: vegetation intact (referred to as

Natural [shading]), all aboveground vegetation (with the exception of D. rotundifolia plants) removed in an 80-cm2 area around the plot to simulate vegetation disturbance (referred to as Light), to separate the impact of light from other factors

40 arising from disturbance, vegetation removed and an 80-cm2 shade net (blocking ca.

75% of solar radiation) level with above-ground canopy was added to the plot on a wooden frame (referred to as Artificial [shading]), and an observer control treatment to account for observer effects which were monitored less frequently than other plots

(referred to as [observer effect] Control). The relative impact of the treatments on the light environment experienced by the D. rotundifolia plants was quantified by measuring photosynthetically active radiation (PAR = 400-700 nm wavelengths) above vegetation height, and at ground level using a quantum sensor (SKP 200 PAR

Quantum Sensor, Skye Instruments Ltd., Wales UK). Light measurements were made between the hours of 10:00 and 12:00. Five measurements per plot were made on a clear, sunny day in June 2016 immediately after treatment addition. While measurement made at a single time point does not accurately describe the light environment, I consider this adequate for comparing relative amounts of shade between treatments. Due to the fragility of Calluna vulgaris and the requirement for frequent measurements of D. rotundifolia, there was a potential to inadvertently increase the light environment in naturally shaded plots—necessitating the inclusion of observer effect control plots. Plants in these plots were only measured at the start, middle and end point of each growing season and prey capture rates were not monitored.

2.2.3 Measurements

Five healthy, mature D. rotundifolia plants were selected at random from each plot.

These plants were labelled and followed throughout the entire experiment from June of the first growing season (2016) to the end of the second growing season

(September 2017). Morphological measurements (see below) were made in June immediately prior to treatment addition, July, August, September 2016 and June,

August and September 2017.

Because the survivability of individual plants under these treatments was at this point unknown, five additional plants were selected at random from each plot,

41 measured and removed at three time points during each growing season (June,

August, September). Therefore, the data comprises two separate sets of plant measurements under the same treatments; one which follows individual plants throughout the experiment, and one which measures and subsequently removes plants from the same plots.

All morphological measurements of D. rotundifolia were taken prior to treatment addition to determine baseline measurements and confirm there were no significant differences before vegetation removal.

Prey availability and capture rates were measured (details below) before manipulation of the plots and in subsequent months to account for the potential for seasonal and inter-annual variability in prey availability which may be affected by treatment addition (Farnsworth & Ellison, 2007).

Prey capture rate was calculated at each measuring interval. Five randomly-selected plants per plot were observed for five days and the number of prey captured was recorded each day and identified at least to order. Prey were removed each morning, after being recorded. These plants were marked and not used for measurements for the rest of the experiment. Potential prey availability was calculated during the same week by placing five 4-cm2 odourless, transparent flypaper squares made of laminate and a coating of the non-drip insect trapping glue Oecotak A5 (Oecos) on the plots and monitoring in the same protocol as plants for prey capture, insects captured outside of the spectra captured by D. rotundifolia were not included in the analysis (i.e. insects too large to be captured by D. rotundifolia but were captured by the artificial trap).

The aim was to determine the response of D. rotundifolia to disturbance. As such, it was necessary to measure a large number of traits in a short amount of time whilst minimising disturbance to the plots. Therefore, a protocol to take non-destructive measurements using digital photographs was devised. Drosera rotundifolia grows in a basal rosette in approximately 2d space and, as such, leaf area and lengths were

42 calculated from 2d imagery of in-situ plants. I recognise the potential for slight differences in leaf angle, or leaf surface morphology to affect these results, but I consider these small relative to the changes I consider to be ecologically significant, and this is a necessary trade-off against the need to minimise disturbance.

Photographs were taken using a Sony SLR SLT-a37 fitted with a SAL-100M28 Sony

100mm F/2.8 macro lens and saved as RAW format to enable colour calibration during processing (Yao, 2012). At each plot, a photo was taken of a QP203card for colour calibration. Photographs of all plants were completed within a week at each measuring interval. All photographed subjects were taken at least twice and the clearest used for measurement recordings. Plant photographs included a tape measure placed adjacent to the plant for scale in analysis. Photos were taken parallel to the face of the subject and as close as possible (approximately 20 cm from the plant). Photographs were taken with a narrow depth of field (F = 2.4) to confirm that the scale (tape measure) and the subject (plant or plant trap) were level and that the angle of the camera was parallel to the subject. If all parts of the plant and scale are in focus, it means the camera angle is parallel to the plant, the scale is level, and the plant is growing in approximately 2d space. Where there is blurring of limited parts of the plant only, leaves were manually flattened until at least two photographs of the plant were in complete focus along with the scale. For each plant, photos were taken to capture the entire rosette with scale, and a closer photograph capturing the entire trap (lamina & trichomes) of a healthy leaf with scale. This process also allowed internal validation from lamina area calculation between whole plant and trap photos.

2.2.4 Trait measurement protocol using Adobe Photoshop

Photographs were opened in ADOBE PHOTOSHOP CC V2015.1 (2015) and a camera profile derived from QPCALIBRATION V1.99C was applied to the photo to standardise colour across photos. A scale was calculated by selecting Image, Analysis, Set measurement scale, Custom and selecting a length of 20 mm running along the tape

43 measure as a number of pixels e.g. Pixel length 331, Logical length 20, Logical units mm. Plant measurements were taken through pixel counts converted to metrics in

Adobe Photoshop and saved to a CSV file.

Using the whole-plant photo I measured total leaf-length of the four longest leaves, petiole and lamina length of two largest leaves using the ruler tool, lamina area of two largest leaves were calculated using the magnetic lasso tool. Number of leaves were counted using the count tool and presence/absence of flower(s) recorded.

Using the trap photo, I measured lamina area using the magnetic lasso tool (to cross validate photos) and trichome number was counted using the count tool. Total number of trichomes is correlated with the number of trichomes that terminate within the circumference of the leaf lamina of each image, therefore I only used the number of trichomes terminating within lamina area for statistical analysis

(Pearson’s Correlation coefficient = 0.94, P < 0.001, N = 750). Leaf colour was measured by averaging the magnetic lasso selection of leaf lamina by selecting Filter,

Blur, Average and then sampling the colour using the Colour Sampling tool to record the CIE L*a*b which expresses colour as three numerical values, “a” provides the green-to-red colour components. CIE L*a*b colour space is a device independent model that can be used as a reference and as such can also be used to quantitatively compare colours (Westland et al., 2012).

2.2.5 Tissue nutrient concentration and nitrogen isotope analysis

At the end of the final measuring interval, ten randomly selected plants per plot were harvested for nutrient analysis. All samples were dried to a stable dry mass by placing in a forced-air oven at 70 °C for 72 hrs on the day of harvesting. Plant samples were ground to a fine homogenised powder using a ball mill. Samples were analysed for δ15N and N proportion at the NERC Life Science Mass Spectrometry

Facility, UK. Nitrogen isotope ratios were analysed by continuous flow using a

Thermo Scientific DELTA V Plus isotope ratio mass spectrometer (Thermo Scientific

Germany) interfaced with a Costech ECS 4010 elemental analyser (Costech

44 Instruments, Italy). Three in-house standards (alanine, gelatine and glycine) were run every ten samples for quality assurance. All data were reported with respect to

15 the international standard of atmospheric N2 (AIR) for δ N. Results were reported in the δ notation as the deviation from standards in parts per mille (‰) where

(Equation 2.1):

'( )+,-./0 "# '*) δ $ = &'( − 17 × 1000 (Equation 2.1) )102010340 '*)

Comparing δ15N in D. rotundifolia across treatments provides evidence for differences in patterns of reliance on prey for total N uptake. This assumption takes advantage of the inherent difference in δ15N between trophic levels to determine the proportional contributions of N sources (prey- or root-derived) into a single sink (D. rotundifolia). If a higher proportion of nitrogen is derived from prey (15N enriched compared to root N), then δ15N of D. rotundifolia plant tissue will be enriched relative to plants which obtain less of their N from prey. For this study is was not possible to collect appropriate plants to act as a non-carnivorous plant end-point, because the site is relatively dry and therefore Sphagnum spp. were rooted in peat and may be taking up different sources of N to D. rotundifolia. I, therefore, used raw measures of

D. rotundifolia δ15N, to give an understanding of qualitative differences in N sources between treatments. This assumes no difference in prey δ15N signature between treatments, or differences in δ15N of root derived N between treatments. These are reasonable assumptions, due to the blocked design—i.e. the treatments were closely located within blocks, and randomly assigned.

Tissue elemental composition was obtained using pXRF Bruker S1 Titan 600 with bench-top assembly in oxide three phase method in GeoExplorer mode for 90 seconds (30 seconds for each phase), max power 50kV at 39uA. Homogenised powdered samples were housed in glass vials covered with a prolene film and placed directly onto the sensor to determine P and K tissue concentration (Towett et

45 al., 2016). As a control for effect of nutrient perturbation from removal of vegetation, sphagnum tissue nutrient composition was also tested.

2.2.6 Data analysis

Data were analysed using R within RSTUDIO V1.1.456 (R Core Team, 2016; RStudio

Team, 2016). To confirm baseline measurements and confirm no difference between treatments, baseline measurements were analysed using a randomised complete block ANOVA (n= 5).

Data after treatments were applied were analysed using a randomised complete block ANOVA with repeated measures (lme), using the mean values of the five harvested plants per plot and a first-order autoregressive structure (AR1). Pairwise post-hoc analysis was performed with general linear hypotheses and multiple comparisons for parametric models using the glht function in the “multcomp” package. Statistical significance was restricted to an alpha of 0.05.

2.3 Results

Prior to treatment, canopy vegetation intercepted on average 74.2% ± 3.63% (mean ± standard error of the mean, SEM) of PAR reaching the bog surface, there were no statistically significant differences between plots allocated to different treatments in the extent of this PAR interception to treatment addition (ANOVA: F2,12 = 0.5112, P =

0.6123). After vegetation disturbance treatments were applied, there were statistically significant differences between treatments in the amount of PAR reaching the bogs surface (ANOVA: F2,12 = 1463.7, P < 0.0001). Specifically, PAR interception by the vegetation canopy in naturally shaded plots remained broadly similar to pre-treatment conditions (PAR interception = 77.3% ± 1.58%) the removal of canopy vegetation (Light) led to a decrease in PAR interception to 10.1% ± 1.65%

(mean ± SEM). Where artificial shade was added to vegetation removal plots, the amount of light interception returned to very similar levels to control plots (79.8% ±

0.61).

46 Table 2.1: Results of univariate statistical analysis of differences between Drosera rotundifolia plants growing in different vegetation manipulation treatments. Presented are baseline statistical comparisons (ANOVA) to confirm homogeneity across plots prior to treatment. Also included are RCB repeated measures ANOVA of treatment effects. Effect size is indicated by mean values for the two growing seasons after treatment application (Jul ’16 to Sept ’17) with SEM below. Arrows indicate direction of effect where treatment effect is significant. Statistically significant differences between the Light treatment and other treatments are emboldened with the exception of baseline measurements of CIE Lab in which Light and Natural are different to Artificial and Control. Figures rounded to 2 d.p. Plant traits are per leaf. CIE lab is International Commission on Illumination L*a*b* space. Baseline After treatment application Direction of Trait Pre-treatment Treatment Time Treatment x time Mean value after treatment treatment d.f. F P d.f. F P d.f. F P d.f. F P Light Artificial Control Natural effect 134.18 105.22 109.68 104.91 Trichome number 3,12 0.30 0.823 3,92 24.76 < 0.0001 5, 92 33.45 < 0.0001 15,92 1.87 0.0361 ↑ ± 4.3 ± 3.2 ± 3.4 ± 2.8 Trichome density 4.92 3.35 3.04 3.04 3,12 1.23 0.342 3,92 47.11 < 0.0001 5, 92 4.03 0.0024 15,92 4.57 < 0.0001 ↑ (trichomes/mm2) ± 0.17 ± 0.12 ± 0.08 ± 0.08 Lamina area 29.04 34.34 37.58 36.65 3,12 1.84 0.193 3,92 9.19 < 0.0001 5, 92 26.29 < 0.0001 15,92 2.04 0.0201 ↓ (mm2) ± 1.3 ± 1.4 ± 2.9 ± 2.78 0.66 0.74 0.76 0.76 Petiole:Leaf length 3,12 2.73 0.091 3,92 71.76 < 0.0001 5, 92 18.76 < 0.0001 15,92 4.33 < 0.0001 ↓ ± 0.01 ± 0.08 ± 0.01 ± 0.01 Longest leaf 17.30 24.29 26.59 26.77 3,12 0.45 0.720 3,92 59.89 < 0.0001 5, 92 28.87 < 0.0001 15,92 4.20 < 0.0001 ↓ length (mm) ± 0.68 ± 0.91 ± 1.49 ± 0.79 22.05 -12.40 -14.07 -14.51 CIE Lab a* 3,12 4.53 0.024 3,92 590.61 < 0.0001 5, 92 41.83 < 0.0001 15,92 19.51 < 0.0001 ↑ ± 1.12 ± 0.95 ± 0.47 ± 0.424 Flowering 0.42 0.30 0.45 0.42 3,12 0.77 0.533 3,92 1.85 0.1431 5, 92 11.52 < 0.0001 15,92 2.01 0.0227 ns (↑ Jun ’17) proportion ± 0.07 ± 0.04 ± 0.06 ± 0.07 Artificial trap 32.36 21.72 21.08 2,8 1.40 0.301 2,68 2.71 0.0594 5, 68 39.34 < 0.0001 10,68 0.85 0.4321 na ns capture (N in plot) ± 4.22 ± 2.15 ± 2.32 Prey capture (N 6.44 4.68 5.52 2,8 0.41 0.679 2,68 2.36 0.1025 5, 68 12.85 < 0.0001 10,68 1.36 0.2196 na ns per plant) ± 1.02 ± 0.67 ± 0.54

47 Table 2.2: Results from univariate statistical analysis of differences in nutrient composition of D. rotundifolia growing in different vegetation manipulation treatments. Effect size is indicated by mean values and SEM from natural abundance isotope ratio and pXRF analysis. Data in this table derived from plants collected at the end of the experiment. Emboldened results are comparisons of Light to other treatments (unless stipulated as otherwise in direction of treatment effect). End of experiment Direction Treatment Mean value after treatment of Trait treatment d.f. F P Light Artificial Control Natural effect Drosera δ15N -5.20 -7.34 -7.71 -7.31 3,12 27.64 < 0.0001 ↑ (‰) ± 0.27 ± 0.26 ± 0.10 ± 0.23 Drosera tissue 1.56 1.89 2.01 2.05 3,12 17.50 < 0.0001 ↓ %N ± 0.04 ± 0.06 ± 0.03 ± 0.08 0.30 0.32 0.32 0.33 N/N(10P)K 3,12 0.69 0.5778 ns ± 0.01 ± 0.01 ± 0.01 ± 0.02 0.38 0.41 0.38 0.38 10P/N(10P)K 3,12 2.09 0.1553 ns ± 0.01 ± 0.01 ± 0.01 ± 0.01 0.31 0.27 0.30 0.30 * Artificial K/N(10P)K 3,12 4.48 0.0249 ± 0.01 ± 0.01 ± 0.01 ± 0.02 lower Sphagnum 1.18 0.94 0.93 1.14 3,12 1.71 0.2184 ns tissue %N ± 0.10 ± 0.07 ± 0.05 ± 0.13 No statistically significant differences in plant traits between plots were present

before the application of treatments, with the exception of a small but statistically

significant difference in colour (Table 2.1). Plants in plots allocated to vegetation

removed [Light] and vegetation intact [Natural] treatments were slightly less green to

plots allocated to artificially shaded [Artificial] and observer control [Control]

treatments but were not different to each other (Table 2.1, Figure 2.2 top). Lamina

trichome density (and number of trichomes per lamina) for a single plant pre-

treatment was six times higher than all other plants. Findings with or without this

anomalous plant included in the analysis were qualitatively similar so this plant was

therefore considered an outliner and removed from further analyses.

Treatment Light Artificial Control Natural

ns ** ** *** ** *

150 number

100 Trichome

Pre-treatment 0 Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17

6 ns *** *** *** *** *** ) 2 mm

4 trichomes (

density

2 Trichome

Pre-treatment 0 Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17 Figure 2.1: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted line indicates end of growing season 1. Presented are mean ± SEM of top: number of trichomes per leaf; bottom: trichome density of trap. Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001). 49 Treatment Light Artificial Control Natural

30 ns *** *** *** *** ***

0 CIE Lab a* Lab CIE

-10

-20 Pre-treatment

Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17 50 ns (*) ns ns *** ***

40 ) 2 30 mm (

area

20 Lamina

Pre-treatment 0 Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17

30 ) mm (

20 length

leaf

Longest 10

Pre-treatment 0 Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17 0.9

ns *** *** *** *** ***

0.8 ratio

length 0.7 leaf : Petiole 0.6

Pre-treatment 0.5 Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17

Figure 2.3: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted line indicates end of growing season 1. Presented are mean ± SEM of Top: length of longest leaf; Bottom: petiole length: leaf length ratio. Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤

0.05, ** P ≤ 0.01, *** P ≤ 0.001. 51 Treatment Light Artificial Control Natural

1.00 ns ns ns * ns ns

0.75

0.50 Proportion flowering Proportion 0.25

0.00 Pre-treatment

Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17 Month across growing seasons Figure 2.4: Drosera rotundifolia response to vegetation and light manipulation treatments over time. Dotted nsline indicates end of ***growing season *** 1. Presented *** are mean ± SEM***. Proportion *** of D. rotundifolia6 with flowers. Tukey HSD for comparison within a time point between vegetation removed (Light) and intact (Natural) treatments: * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001. ) 2 mm

2.3.1 Morphological traits 4 For plants where there was no manipulation (Natural & Control) the following traits trichomes (

generally increased, then decreased during the growing season: leaf length, lamina density

area and trichome number. Trichome density generally decreased during the season, 2 followed by a small increase at the end of each growing season. Leaf colour and

petiole:leafTrichome length ratio remained consistent throughout each growing season.

Where the vegetation canopy was removed to simulate disturbance (Light) D. Pre-treatment rotundifolia0 leaf traits diverged from those under the intact canopy. The leaves on Jun '16 Jul '16 Aug '16 Sept '16 Jun 17' Jul '17 Aug '17 Sept '17 these plants had more trichomes (Figure 2.1 top), a higher density of trichomes

(Figure 2.1 bottom), were redder (Figure 2.2 top), had smaller leaf laminas (Figure

52 2.2 bottom), shorter leaves (Figure 2.3 top) and lower petiole:leaf length ratio (Figure

2.3 bottom, See Table 2.1). This divergence occurred quickly for colour, trichome number and trichome density, all being different from shaded plots from the first measurements after treatment applications. Decreases in leaf lamina area were not present until the middle of the second year after treatment additions. For trichome number the difference between the vegetation removed (Light) and shaded (Natural) treatments increased at first, but the size of this difference decreased towards the end of the second growing season (Figure 2.1).

There were no statistically significant differences in D. rotundifolia leaf traits between plants in plots where vegetation was removed but shade was added (Artificial) when compared to shaded treatments (Natural). This indicates that the differences between the treatments with vegetation removed compared to intact-vegetation treatments are due to changes in the light environment created by the vegetation canopy.

Plants in the vegetation removed (Light) environment flowered at similar frequencies at the start and end of the experiment, but did flower earlier and in higher probability in the first month of the second growing season (Table 2.1, Figure 2.4).

Plants caught approximately two prey items per plant over four days. Plants in the vegetation removed plots (Light) did capture slightly more prey but there were no statistically significant differences in prey capture between treatments. Artificial traps also did not capture a significantly different amount of prey across treatments

(Table 2.1). The amount of prey captured by both D. rotundifolia plants and artificial traps did, however, vary over time such that more prey were captured in the middle and end points of the growing season compared to the start of each growing season.

2.3.2 Drosera rotundifolia nutrient stoichiometry and proportion of prey derived N

Plants that were in treatments where vegetation was removed (Light) had a higher proportion of 15N at the end of the experiment compared to the other treatments.

This resulted in significantly less negative tissue δ15N values (Table 2.1, Figure 2.5

53 top). These plants also had lower tissue N concentration (Table 2.1, Figure 2.5 *** bottom).

N:P:K-5 ratios were not significantly different among treatments at the end of the

experiment, with the exception of the artificially shaded treatment [Artificial] having ) ‰

a slightly( lower proportion of K in total N:P:K (Table 2.1, Figure 2.6 top). Sphagnum

N 15

spp.δ N:P:K ratios did not differ among treatments but did differ from Drosera

-6 rotundifolia in that Sphagnum spp. were more P limited (Table 2.1, Figure 2.6).

General patterns, however, consistently indicate marginally higher concentrations of rotundifolia

all rotundifolia D. nutrients of Sphagnum spp. in the light environment, but this did not affect D. -7 Drosera rotundifolia nutrient stoichiometry in the same way and was not statistically

significant.

***

(%) 2.0

conc.

tissue

1.8 D. rotundifolia D. D. rotundifolia 1.6

Light Artificial Control Natural

Figure 2.5: Drosera rotundifolia nutrient information at the end of the second growing season after vegetation and light manipulation treatments. Presented are mean ± SEM of top: Drosera rotundifolia δ15N; bottom: Drosera rotundifolia tissue N concentration (%) * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001

54 Treatment 100 Light

Artificial 20 Control 80 Natural 100 40 N/(N(10P)K P or P + N Limitation 60

P/(N(10P)K)60 40 80 K or K + N Limitation

40 20 N/(N(10P)K

100 N Limitation P or P + N Limitation60

20 40 60 80 100

P/(N(10P)K)60 K/N(10P)K

100 40 K or K + N Limitation

80 80

40 N/(N(10P)K P or P + N Limitation 20 60

P/(N(10P)K)60 100 N Limitation40

K or K + N Limitation

20 40 20 60 80 100

100 N Limitation K/N(10P)K

20 40 60 80 100 K/N(10P)K

55 2.4 Discussion

My data demonstrate phenotypic plasticity in the response of Drosera rotundifolia to selective disturbance in situ. I show that changes in the light environment, caused by the removal of vascular vegetation, result in changes in leaf traits and N nutrition of sub-canopy D. rotundifolia plants. These changes occur at different rates, for different traits. When vegetation was removed, D. rotundifolia leaves responded with a typical light acclimation response—decreasing relative petiole length, and increased redness within the first two months (Figure 2.2, Figure 2.3). Lamina area was only significantly different from the middle of the second growing season (Figure 2.2).

Importantly, carnivorous traits reflected the shift in the balance of nutrient and carbon limitation such that strategies to invest in nutrient acquisition became dominant following canopy vegetation removal, resulting in increased investment in carnivory (higher trichome number and densities, Figure 2.1), and a resulting increase in the contribution of prey N (indicated by higher δ15N, Figure 2.5). For the first time, carnivorous trait differences and contribution of prey N within individual plants over the course of two growing seasons has been shown following selective disturbance. Phenotypic plasticity of D. rotundifolia in response to selective disturbance, as evidenced in this experiment, may facilitate persistence in dynamic environments where these particular disturbances are likely to occur within the lifetime of the plant. My data confirm the predictions that D. rotundifolia will increase investment in carnivory (i.e. increased trichome number and density) which will increase the proportion of prey derived N and increase the frequency of earlier flowering. Confirmation of these predictions derived from current evolutionary models being applied in an ecological context (Givnish et al., 1984; Pavlovič &

Saganová, 2015; Givnish, 2015) support the hypothesis that traits involved in plant carnivory are phenotypically plastic and respond to selective disturbance in situ.

Habitat disturbances provide a temporary microsite that favours either the recruitment and establishment of seedlings, or a phenotypic response of organisms

56 that survive and remain in the disturbed microsite (Briske, 1996; Pausas, 2015;

Herben et al., 2018). Selective disturbances such as trampling and grazing in temperate peatlands remove canopy vegetation leaving non-vascular and prostrate plants, such as D. rotundifolia, that survive the perturbation, in an exposed, lighter environment (Smith et al., 2003; Groome & Shaw, 2015). Previous studies have shown phenotypic plasticity ex-situ in D. rotundifolia leaf traits (Thorén et al., 2003) and phenotypic variability, but not plasticity, in situ (Millett et al., 2012). In this experiment I show that D. rotundifolia can persist in shade (Stewart & Nilsen, 1992), and demonstrate phenotypic plasticity to changes in the environment from selective disturbance, and are able to acquire a greater proportion of nutrients from prey

(Figure 2.2 & Figure 2.5 top, Karagatzides & Ellison, 2009). Previous studies have confirmed that Drosera spp. also benefit from disturbance due to high seedling establishment (Brewer, 1998, 1999). Not only is D. rotundifolia seedling proliferation high in disturbed areas (Brewer, 1998, 1999), but mature individuals are able invest more and earlier in reproduction as a consequence of increased nutrients derived from prey resulting from increased investment in carnivory, facilitating a two-sided exploitation of transient disturbances, whether they remove all species from an area or only canopy biomass as demonstrated in this experiment. I suggest that these life history traits make D. rotundifolia particularly suited to habitats that experience intermittent disturbance or selective disturbance.

Disturbance and stress are the two fundamental processes predicted to control the evolution of plant life history strategies (Grime, 2006). Grime’s universal adaptive strategy theory identifies three primary strategies which emerge from high and low combinations of disturbance and stress: competitors, stress-tolerators and ruderals

(CSR). According to this theory, there is no viable strategy for plants in environments of high stress and disturbance—the ‘untenable triangle’. It has been suggested that carnivory evolved to facilitate persistence in these conditions

(Brewer, 1998, 1999). It may, however, be the case that carnivorous plant species

57 escape the untenable triangle by accessing sources of nutrients unavailable to other plants, thus reducing the stress they experience. I simulated selective disturbance in a high stress (low nutrient) environment, and demonstrated that D. rotundifolia invest more in carnivory, acquire more N from prey and flower earlier. This shows that D. rotundifolia respond positively to selective disturbance and have the capacity to allocate resources quickly to take advantage of the new environmental conditions as demonstrated by the quick and sustained trait changes (Figure 2.2 & Figure 2.3) .

Plants experiencing stress in these environments should be heavily nutrient limited and should be severely restricted in their ability to invest even small amounts into new growth or in this case, carnivorous traits (Grime, 2007). Plants in these environments should not be able to respond positively to any form of disturbance. A more accurate description of carnivorous plant strategies is therefore, to occupy other plant’s untenable environments. Peatlands are considered high stress environments due to their low root nutrient availability, competition for which is overcome in carnivorous plants by accessing a source of nitrogen that is not possible for their non-carnivorous counterparts (Karagatzides & Ellison, 2009). And disturbances common in these peatlands generally remove canopy vegetation, which carnivorous plants then take advantage of by increased seedling establishment and, as shown in this experiment, a phenotypic response to take advantage of the increase in light availability. While I think it is not accurate to confirm that D. rotundifolia are evolved to exist in high stress, high disturbance environments, this experiment indicates that they gain an advantage in instances of other plant’s untenable triangles through accessing otherwise inaccessible sources of nutrition and responding to selective disturbance.

The represented lack, or low presence, of carnivorous plants under shade conditions in the literature may be a consequence of the difficulty to identify these plants under dense vegetation, rather than an accurate prediction of their restriction to consistently light environments. The premise of this experiment relies on pre-

58 existing populations of densely shaded carnivorous plants, and a caveat of this exception to models restricting carnivorous plants to well-lit environments may be that carnivorous plants can persist in shaded environments temporarily if the rate of selective disturbance is sufficient to maintain populations (Brewer, 1999). Persistence in shade is only possible, however, through phenotypic plasticity to tolerate these different environments and increased expression of carnivory and resulting uptake of nutrients derived from prey which enhance reproduction and seedling establishment during these disturbance events.

I found that some aspects of nutrition changed but others did not: N derived from prey was higher, tissue %N was lower but prey capture and N:P:K stoichiometry did not differ (Figure 2.5, Figure 2.6). These different results can be explained biologically. Prey capture may not change due to very low densities of prey available (two prey items per plot during recording times), but investment in resources to digest and assimilate nutrients from prey may increase, similar to other measured carnivorous traits. To my knowledge, phenotypic plasticity of digestion and assimilation has not been investigated in carnivorous plants. If digestion and assimilation are phenotypically plastic components of carnivory, evidence of a higher proportion of N derived from prey without an increase in prey capture rate

(as found in this experiment) would support this theory. I was unable to quantify the proportion of N from prey because it was only possible to sample these plants at the end of the growing season. Therefore, important information on the amount of nutrients invested in reproduction during the growing season is not measured, meaning that any quantitative proportion of N from prey would be inaccurate.

Reproductive investment may also explain lower tissue N concentration of plants in the Light treatments. The benefits of increased nutrient uptake from prey are known to increase earlier flowering and seed set (Givnish, 2015), and the flowering of plants inherently reduces the tissue nutrient content of the rest of the plant. There is evidence of earlier and more frequent flowering in the vegetation removed

59 treatments which support this explanation (Figure 2.4). However, our measures of

δ15N are a robust qualitative comparison because of the natural abundances of nitrogen isotopes across increasing trophic levels. I can conclude with reasonable confidence that the plants in the vegetation removal treatments obtained more N from prey, indicated by a higher proportion of δ15N. The allocation of resources to reproduction is an important fitness parameter in these species, particularly if, as I predict, they benefit from disturbance events for seedling establishment and selective disturbances through increased reproductive potential. This aspect is often overlooked and, based on evidence presented here, may be a significant area of investment of nutrients derived from carnivory, and our nutrient analysis at the end of the growing season would not show this investment of nutrients in reproduction.

I conclude that D. rotundifolia can survive in shade and respond to selective disturbance through the phenotypic plasticity of traits, carnivorous and non- carnivorous, which may explain how this species is able to exist partially outside of current models predicting their environmental restriction to well-lit environments.

Phenotypic plasticity is poorly considered in species distribution models, which may be a substantial oversight as it has the potential to alter the constraints of trade-off relationships which contribute to being able to survive in environments of varying stress and disturbance (Shipley et al., 2006; Benito Garzón et al., 2019). I hypothesise that phenotypic plasticity of D. rotundifolia contributes to their ability to exist in environments of low N and intermittent habitat disturbance because they can take advantage of disturbance mediated increases in light availability in two distinct ways–seedling establishment and, as shown in this experiment, increase in prey derived N which can be invested in reproduction. Adaptation to disturbance and selective disturbance may provide carnivorous plants a key advantage over non- carnivorous competitors as they cannot take advantage of disturbed areas in the same way, due to the constraints imposed by the adaptations required for the restrictions in N availability they experience. In this case, it is unclear to what extent

60 other aspects of the carnivorous habit are plastic, such as digestion and assimilation, and furthering our understanding of this aspect of the carnivorous habit would give important information on the unique costs of carnivorous plants to acquire nutrients via leaves. Furthermore, if substantial proportions of nutrients derived from prey are directed towards earlier and increased flowering, future studies may provide a more holistic understanding of plant carnivory if investment in carnivory is associated with aspects of reproductive investment and seedling establishment throughout the growing seasons in addition to measures of fitness correlates used in this study.

61 Chapter 3 Climate-driven changes in small-scale habitat characteristics affects phenotypic variation of the carnivorous herb Drosera rotundifolia in patterned peatlands.

62 3.1 Introduction

Species distributions are controlled by the biotic and abiotic conditions they experience (Van der Putten et al., 2010). Ecological conditions vary at a range of scales, from inter-continental to micro-habitat (Statzner & Moss, 2004; Bergholz et al.,

2017; Frainer et al., 2018). It is also likely that impacts at different scales will interact, resulting in complex biological responses to environmental change. Phenotypic variability in morphology and physiology in response to limiting resources in the environment such as nutrients, light and water, is a key determinant of species to persist across a range of ecological conditions (de Smedt et al., 2018). Understanding how phenotypes vary in situ and at the same scales as, and in connection with, ecological and environmental variability is, therefore, essential to understand the impacts of different drivers of environmental change.

Patterned peatlands mitigate aspects of climate change and contribute to biodiversity because they are characterised by specific biogeological processes that result in the sequestration of carbon and provision as a habitat for a number of rare or highly endemic species (Josten & Clarke, 2002; Leifeld & Menichetti, 2018).

Patterned peatlands are inherently heterogeneous environments, characterised by raised drier hummocks with high vascular plant cover which impacts on the sub- canopy light environment (Korrensalo et al., 2018), and lower, wetter hollows and lawns typified by non-vascular vegetation (Van Der Molen et al., 1994; Rietkerk et al.,

2004; Rydin & Jeglum, 2006). Hummocks are, consequently, more shaded than hollows. Nutrient availability can also vary between peatland microforms, but the pattern of this variation differs between peatlands, due to differences in climate.

Mixed mires or patterned peatlands typically receive most of their nitrogen from atmospheric deposition (ombrotrophic), which moves with water flow, or by surface/ground water (minerotrophic) (Lovett et al., 2009). The distribution of nitrogen within these peatlands, therefore, depends on the direction of water flow.

Where flow is from hollows to hummocks (when evapotranspiration >

63 precipitation), nitrogen accumulates in hummocks; where flow is from hummocks to hollows (when evapotranspiration < precipitation), nitrogen accumulates in hollows

(Eppinga et al., 2010). This pattern is predicted to determine the primary pattern- forming processes, but clearly nutrient distribution will also influence plants growing on the peatland. This potential habitat variability across microforms might impact on patterns of phenotypic variability—as this is how plants can persist in these variable environments (Eppinga et al., 2008; de Smedt et al., 2018). This impact has, however, never been tested. Though peatland formation has been demonstrated to be dependent on these processes, the biological impact of these patterns of nutrient transfer is not yet known. It is clear that there is habitat heterogeneity over a variety of niche components, but the way these components interact may change fundamentally along a climate gradient, which may have a critical impact on plant trait variation because these habitats are sensitive to changes in the environment

(Koller & Phoenix, 2017). Understanding the impact of climate-driven nutrient accumulation on intraspecific plant trait variation within sites along a climatic gradient will enable general predictions to be made of the impact climate change will have on peatland ecosystems globally (Madani et al., 2018).

Carnivorous plants are commonly found in peatlands and are sensitive to changes in light and nutrient availability (Ellison & Gotelli, 2002; Luken, 2007; Bruzzese et al.,

2010; Millett et al., 2012), due to their specialised adaptations for acquiring nutrients from prey via modified leaves to supplement low root nutrient uptake (Ellison &

Adamec, 2018). Carnivorous plants are therefore useful indicators of broader ecosystem processes as they are indicative of patterns of peatland nutrient availability and their presence is characteristic of functioning peatland ecosystems

(Ellison & Gotelli, 2002; Ellison & Adamec, 2018). Some carnivorous species occupy both hummocks and hollows, and so experience a range of habitat conditions. They are therefore ideal study systems to understand the impact of habitat heterogeneity on intraspecific trait variability.

64 Intraspecific variation in carnivorous traits has been demonstrated across large spatial scales in response to differences in resource availability (e.g. Ellison & Gotelli,

2002; Millett et al., 2012, 2015), suggesting that such phenotypic variability is important as a trait to occupy different environments. Carnivorous plants increase allocation to carnivory in high light, N deficient conditions (Thorén et al., 2003;

Alcalá & Domínguez, 2005). What has not been demonstrated to date is the extent that this variation might manifest within a habitat (Thorén et al., 2003). Furthermore, no study to date has linked the impact of climate-driven nutrient accumulation within microforms in patterned peatlands to intraspecific trait variation.

In this study I measured phenotypic variability in allocation to carnivory and nutrition of the carnivorous herb, Drosera rotundifolia, in response to habitat heterogeneity at small (< 1 m) scales. I measured leaf traits and used natural abundance stable isotopes (15N) to estimate the contribution of root vs prey N to total plant N across microforms at three sites with varying levels of ET:P. I hypothesise that intraspecific trait variation of D. rotundifolia is influenced by general patterns of shade in hummocks and hollows and also by local patterns of ET:P driven nutrient accumulation. I predict that increased shade and nutrients will result in decreased investment in carnivory (trichome density), and reduced contribution of prey N to total N content. Correspondingly, I predict that, in bogs with high ET:P, nutrient accumulation and shade on hummocks will result in reduced investment in carnivory (trichome density) and reduced contribution of prey N to N content, compared to plants on hollows; in bogs with lower ET:P, decreased, or zero nutrient accumulation, but continued increased shade, on hummocks will result in smaller or no differences in investment in carnivory (trichome density) or the contribution of prey N to plant tissue N content.

65 3.2 Methods

3.2.1 Site information and plot selection

Fieldwork was conducted from July to August 2017. Measurements were made at three patterned ombrotrophic peatlands in Europe, across a gradient of ET:P (Table

3.1). Bogs were characterised by a hummock-hollow microtopography with D. rotundifolia present on both microforms. At all sites, vascular vegetation on hummocks was dominated by Calluna vulgaris; vascular vegetation on hollows was predominantly Carex spp.

Table 3.1: Site information for three peatlands across Europe. Predicted nutrient accumulation based on definitions of Eppinga et al. (2010) 1.

Expected Latitude, Precipitation Evapotranspiration microform of Study Site ET:P Longitude P (mm year-1) ET (mm year-1) nutrient accumulation Inverewe, 57.460120, 1700 315 0.19 Hollow Scotland (Sc) 5.331009 Hallebomossen, 59.729101, 741 282 0.38 Intermediate Sweden (Sw) 14.977923 Intermediate/ Siikaneva, 61.843523, 707 380 0.54 weakly Finland (F) 24.288918 hummock

The three sites are along a latitudinal gradient and comprise three climate classifications (Köppen Geiger climate classification: Scotland – Cfb, Sweden – Dfb and Finalnd – Dfc, Kottek et al., 2006). While these climes are different and the three sites have different levels of nitrogen deposition (Scotland = 7, Sweden = 4 & Finland

= 7 kg N ha-1 year-1, Korhonen et al., 2013; APIS, 2016; SUAS, 2017), these factors act at a site level. The ET:P at each site affects the accumulation of nutrients in hummocks or hollows. Therefore, though absolute values for different

1 Climate data derived from; Scotland: Eppinga et al., 2010, Sweden: Kindla II weather monitoring site mean for 1996–2016, Finland: Korhonen et al., 2013 and Siikaneva and Hyytiala weather monitoring stations 1986–2016.

66 environmental factors (i.e. light availability and N availability) will differ between sites, the overall influence of ET:P on intraspecific trait variation within each site is predicted to reflect the local pattern of ET:P. As a consequence, I can quantitatively compare microforms within individual sites and qualitatively compare the different site patterns.

Fifty 50-cm2 plots containing D. rotundifolia plants growing in Sphagnum spp. were selected using a stratified random design at each site. These were distributed evenly among hummocks and hollows (25 in hummocks and 25 in hollows), with a total of five hummocks and five hollows selected per site (five plots per hummock or hollow).

3.2.2 Measurements

The aim was to determine the differences in D. rotundifolia plant traits between microforms. This necessitated the measurement of many traits in a very limited time available in the field at each site. It was therefore necessary to devise a protocol to minimise time in the field whilst enabling the accurate and precise measurement of multiple traits. This study utilised the method of photographic measurements described in detail in Chapter 2. Briefly, D. rotundifolia grow in a basal rosette in approximately 2D space, making macro-photography a viable method for measuring leaf length and area. Photographs were taken using a Sony SLR SLT-a37 fitted with a SAL-100M28 Sony 100 mm F/2.8 macro lens and saved as RAW format to enable colour calibration during processing (Yao, 2012). At each plot a photo was taken of a

QP203 card for colour calibration. Photographs of all plants in each site were completed within five days. All subjects were photographed at least twice and the clearest used for measurement recordings ensuring the scale and the entire plant were in focus using a narrow depth of field (F = 2.4, Figure 3.1). I recognise the potential for slight differences in leaf angle, or leaf surface morphology to affect these results, but I consider these small relative to the changes I consider to be

67 ecologically significant, and this is a necessary trade-off against the need to minimise disturbance, time and cost in the field.

Figure 3.1: Photo protocol for measurement of Drosera rotundifolia morphological traits.

Five D. rotundifolia plants per plot were selected at random for measurement of leaf traits using digital photography. These, and a further five plants were collected giving a total of ten plants collected from each plot. From these ten plants, five were selected at random and a fully-formed, healthy leaf was removed, and stored separately for measurement of leaf area and mass.

The interception of light by the vascular plant canopy was measured as the percentage difference between measurements of light intensity above and below the vascular plant canopy. Measurements were made at five locations for each plot, between 10:00 and 14:00 of photosynthetically active radiation (PAR = 400-700nm wavelengths) above vegetation height, and at ground level using a quantum sensor

(SKP 200 PAR Quantum Sensor, Skye Instruments Ltd., Wales UK). Vascular vegetation cover was estimated by eye. Vegetation height was measured at three points per plot using a tape measure. I recognise that this does not fully characterise the light environment experienced, but this was a pragmatic way of obtaining meaures that can be used to comapare for large differences between microforms.

For natural abundance isotope analysis and tissue nutrient concentration, one 5 ´ 5 ´

2 cm (width, length, depth) sample of Sphagnum spp. capitula was collected from each plot. Potential prey was sampled by placing a 10 ´ 5-cm sheet of yellow

68 flypaper at each plot; captured insects were collected daily for four days. Insects likely to be too large to be captured by D. rotundifolia (i.e. > 10-mm length) were not included in this sample. Samples were stored with desiccant before being dried. As soon as possible and within 48 hrs, plant and insect samples were dried to a stable mass in a forced-air oven at 70 °C for 72 hrs. Mass of the five separated leaves was measured, and these leaves were returned to their respective plot samples. Plant samples were ground to a fine homogenised powder using a ball mill, insects were ground using a pestle and mortar, to ensure sample homogeneity.

3.2.3 Trait measurement using Adobe Photoshop

Photographs were opened in ADOBE PHOTOSHOP CC V2015.1 (2015) and a camera profile derived from QPCALIBRATION V1.99C was applied to the photo to standardise colour across photos. Plant measurements were taken through pixel counts converted to metrics in Adobe Photoshop and saved to a CSV file.

Using the whole-plant photo I measured total leaf-length of the four longest leaves, petiole and lamina length of the two largest leaves, lamina area of two largest leaves, number of leaves on rosette and presence or absence of flowers. Using the trap- photo, I measured lamina area and trichome number. Red or green colour of D. rotundifolia was measured by averaging the lamina colour in CIE L*a*b which expresses colour as three numerical values, “*a” provides the green-to-red colour components. CIE L*a*b colour space is a device independent model that can be used as a reference and as such can also be used to quantitatively compare colours

(Westland et al., 2012).

3.2.4 Nitrogen isotope and tissue nutrient concentration

Carnivorous plant prey tends to be 15N enriched compared to N taken up through the roots. The 15N of a carnivorous plant can therefore be used (with estimates of prey 15N and 15N of root N uptake) to estimate the contribution of root and prey N to the total N content of the plant (Millett et al., 2012; Cook et al., 2018). Drosera

69 rotundifolia, Sphagnum spp. and prey samples were analysed for δ15N and N concentration at the NERC Life Science Mass Spectrometry Facility, UK. Nitrogen isotope ratios were analysed using a Thermo Scientific DELTA V Plus isotope ratio mass spectrometer (Thermo Scientific Germany) interfaced with a Costech ECS 4010 elemental analyser (Costech Instruments, Italy). Three in-house standards (alanine, gelatine and glycine) were run every ten samples for quality assurance. All data were reported with respect to the international standard of atmospheric N2 (AIR) for

δ15N. Results were reported in the δ notation as the deviation from standards in parts per mille (‰) where (Equation 3.1):

'( )+,-./0 "# '*) Equation 3.1: δ $ = &'( − 17 × 1000 )102010340 '*)

The proportion of prey derived N and corresponding root derived N of D. rotundifolia was calculated by using the δ15N of D. rotundifolia, δ15N of Sphagnum and

δ15N of prey into a single isotope ratio, two end-point linear mixing model (Equation

3.2) (Shearer & Kohl, 1989). The assumption of the model is that non-carnivorous plants can be used as a proxy for δ15N of carnivorous plants that have obtained none of their N from prey and prey δ15N can be used as a proxy for carnivorous plants that rely on prey for all their N.

'( '( A BCDEFGDH DEIJKLMNEOMH P A BQRSHTKJU Equation 3.2: %$<01=>0< 21?- .10@ = '( '( A BVDGW P A BQRSHTKJU

This model takes advantage of the inherent difference in δ15N between trophic levels to discern the proportional contributions of two isotope ratio sources (prey or root derived) into a single sink (D. rotundifolia δ15N).

Tissue elemental composition was determined using X-ray, with a pXRF Bruker S1

Titan 600 with bench-top assembly. This is increasingly used as a viable alternative to acid digestion and ICP analysis (McGladdery et al., 2018; Etienne et al., 2018). The homogenised powder of the plant tissue samples housed in glass vials covered with a prolene film were placed directly on the sensor following guidelines set by Towett

70 et al. (2016). The pXRF machine was set up for the Oxide three phase method for 90 seconds (30 seconds for each phase), max power 50 kV at 39 uA.

3.2.5 Data analysis

Measurements of plants within a plot were averaged (arithmetic mean). All data analyses were conducted using R within RSTUDIO V1.1.456 (R Core Team, 2016;

RStudio Team, 2016). Trichome density was calculated using Equation 3.3, trichomes that terminated within the circumference of the lamina within the photo were counted, rather than counting all trichomes on the plant. This is a more efficient measure with little difference in outcome. Phytovolume was calculated using

Equation 3.4 and Specific Leaf Area (SLA) was calculated using Equation 3.5:

e?f,/ 3g-h01 ?2 f1=4i?-0+ ?3 /0,2 /,-=3, Equation 3.3: XYZ[ℎ]^_ `_abZcd = j,-=3, ,10, (--l)

Equation 3.4: nℎdc]o]pq^_ (^r^Ps) = t_ua o_v_cucZ]a ℎ_Zvℎc ]w xp]c (^) × ^_ua o_v_cucZ]a []o_Yuv_ ]w xp]c(^s^Ps)

|0,3 ,10, ?2 210+i /0,2 .01 ./?f (--l) Equation 3.5: yz{ = |0,3 <1@ -,++ (-})

Due to the high degree of covariance between the measured parameters, I used principal component analysis (PCA) to reduce the number of response variables in the analyses. I performed two separate PCAs: plant morphological traits and plant tissue nutrient concentration and nutrient ratios. Bivariate matrices were then used to determine correlations between PCA scores from each of these separate analyses

(See Appendix C3.1).

The reduced D. rotundifolia data set, and other biotic and abiotic measurements (light interception, canopy characteristics, Sphagnum nutrient concentrations) were analysed using nested ANOVA, with bog microform nested within bog. Initially, hierarchical models for light interception, vegetation coverage, phytovolume and nutrient availability were used to confirm differences in microforms, the light environment and nutrient availability within each site. A priori contrasts (planned

71 comparisons) were used to determine pairwise comparisons of differences between these measurements between microforms within each site (and not microforms across sites). Assumptions of normality and homoscedasticity were tested using Q-Q plots and Shapiro-Wilk tests. Trichome density was log-transformed for analysis.

3.3 Results

3.3.1 Patterns of light and nutrient availability in microforms across sites

Across all sites, vascular vegetation cover, phytovolume and light interception were consistently higher on hummocks than on hollows. These differences were statistically significant (Figure 3.2 A, Table 3.2). There was a significant difference in

Sphagnum sp. N content between sites and microforms (Figure 3.2 B, Table 3.2). At the Finland site there was no difference in N content between microforms; at the

Sweden site there was a weak evidence of nutrient accumulation in hollows; at the

Scotland site N content in hollows was higher than in hummocks (Figure 3.2 B, Table

3.2). Nutrient limitation of N:P:K did not vary significantly between microforms, but did vary slightly across sites (Figure 3.2 C).

72 A A Hollow Hummock 0.8 ) 2

− *** *** *** m

2 0.6 m (

0.4 coverage 0.2 Veg. 0.0 BB ns (*) ***

0.9

0.6 Moss N (%) N Moss 0.3

0.0 Finland Sweden Scotland Site CC 100

Site 20 Finland 80 Sweden

40 Scotland N/(N(10P)K) P or P + N Limitation 100 60 Microform

P/(N(10P)K) 60 Hollow 20 40 Hummock K or K + N Limitation

80 80

40 100 N Limitation N/(N(10P)K) P or P + N Limitation 60 20 40 60 80 100 K/(N(10P)K) Figure 3.2: Differences between microforms (hummocks and hollows), for three sites in A: P/(N(10P)K)60 vegetation coverage, and B: Sphagnum spp. tissue %N. Error bars represent SEM. *** P < 0.001 (*) P <0.10, ns = P > 0.10. C: Ternary diagram of N, P, K nutrient stoichiometry of Sphagnum spp. within microform at each site (%). Nutrient limitation guides40 from Olde Venterink et al., (2003). K or K + N Limitation 73

100 N Limitation

20 40 60 80 100 K/(N(10P)K) Table 3.2: Summary for hierarchical ANOVA analysis of environmental variables in hummocks and hollows across three sites (Sw = Sweden, F = Finland, Sc = Scotland). ‘/’ denotes a significant difference ordered largest to smallest alpha 0.05. Duplicate letters appearing either side of ‘/’ signifies a non-significant difference to adjacent letters, but the other two sites are significantly different. Microform within site comparisons are represented using hollows as a base level and denoting whether hummocks have significantly higher [+] or lower [-] values or not significantly different [O], P value significance threshold (alpha) 0.05. Where 0.05 < P < 0.1 arises, differences are in brackets. PC signifies planned contrast analysis of microform within each site. Site Site M x Site Finland Sweden Scotland differences Hollow Hummock Hollow Hummock Hollow Hummock Variable d.f. F P d.f. F P High first PC P Mean ± Mean ± PC P Mean ± Mean ± PC P Mean ± Mean ± SEM SEM SEM SEM SEM SEM Veg. cover + 0.16 0.42 + 0.34 0.61 + 0.14 0.39 2,24 20.21 < 0.001 3,24 27.09 < 0.001 Sw / F Sc (m2 m-2) 0.003 ± 4.5 ± 1.0 0.002 ± 4.5 ± 5.0 0.004 ± 1.6 ± 2.2

- 0.83 0.65 - 0.88 0.52 - 0.72 0.47 % light trans. 2,24 7.99 0.002 3,24 28.06 < 0.001 Sw F / Sc 0.009 ± 0.01 ± 0.04 < 0.001 ± 0.03 ± 0.05 < 0.001 ± 0.04 ± 0.03

Phytovol. (m3 + 2.70 5.64 + 5.04 9.34 + 2.43 7.21 2,24 8.41 0.002 3,24 13.81 < 0.001 Sw / Sc F m-2) 0.042 ± 0.26 ± 0.95 0.004 ± 0.82 ± 1.13 0.002 ± 5.28 ± 6.34

Sphag. spp. O 0.64 0.62 O (-) 0.81 0.69 - 0.98 0.67 2,24 19.38 < 0.001 3,24 17.14 < 0.001 Sw Sc / F %N 0.820 ± 0.08 ± 0.02 0.091 ± 0.02 ± 0.01 < 0.001 ± 0.08 ± 0.06

+ 0.03 0.04 + 0.02 0.043 - 0.05 0.04 Sphag. spp. P 2,24 7.19 0.004 3,24 12.86 < 0.001 Sc / F Sw 0.032 ± < 0.01 ± < 0.01 0.003 ± < 0.01 ± < 0.01 0.004 ± < 0.01 ± < 0.01

O 0.94 1.03 + 0.79 1.05 - 1.08 0.83 Sphag. spp. K 2,24 0.46 0.639 3,24 6.13 0.003 F Sw Sc 0.289 ± 0.08 ± 0.05 0.005 ± 0.06 ± 0.07 0.007 ± 0.04 ± 0.04

74 Site Site M x Site Finland Sweden Scotland differences Hollow Hummock Hollow Hummock Hollow Hummock Variable d.f. F P d.f. F P High first PC P Mean ± Mean ± PC P Mean ± Mean ± PC P Mean ± Mean ± SEM SEM SEM SEM SEM SEM Sphag. spp. - 24.21 16.25 - 34.03 16.91 O 21.41 20.96 2,24 4.96 0.016 3,24 18.91 < 0.001 Sw Sc / F Sc N:P 0.010 ± 1.85 ± 0.94 < 0.001 ± 2.98 ± 0.75 0.878 ± 1.00 ± 2.03

Sphag. spp. O 0.726 0.622 - 1.122 0.69 O 0.95 0.835 2,24 7.93 < 0.001 3,24 8.21 < 0.001 Sw Sc / F N:K 0.370 ± 0.031 ± 0.039 < 0.001 ± 0.12 ± 0.06 0.340 ± 0.06 ± 0.043

Sphag. spp. - 34.01 26.27 - 33.04 25.49 O 24.02 25.76 2,24 3.48 0.047 3,24 4.41 0.013 Sw F / Sw Sc K:P 0.026 ± 3.75 ± 0.79 0.030 ± 2.54 ± 1.65 0.601 ± 1.32 ± 1.28

Sphag. spp. - 34.99 30.55 - 43.82 32.21 O 38.35 36.08 2,24 9.98 < 0.001 3,24 16.61 < 0.001 Sw Sc / F N/N(10P)K 0.067 ± 0.81 ± 1.09 < 0.001 ± 1.70 ± 1.57 0.337 ± 0.84 ± 1.30

Sphag. spp. + 15.79 19.44 + 14.34 19.60 O 19.13 19.22 2,24 4.60 0.020 3,24 12.21 < 0.001 Sc F / Sw F 10P/N(10P)K 0.005 ± 1.21 ± 0.46 0.002 ± 0.71 ± 0.52 0.944 ± 0.66 ± 0.68

Sphag. spp. O 49.22 50.01 + 41.84 48.20 O 42.51 44.71 2,24 9.46 < 0.001 3,24 3.66 0.027 F / Sw Sc K/N(10P)K 0.766 ± 1.58 ± 0.98 0.023 ± 1.84 ± 1.90 0.411 ± 1.27 ± 0.69

75 3.3.2 Influence of habitat heterogeneity and nutrient availability driven by climate factors across sites on trait expression and nutrient composition of Drosera rotundifolia plant traits

Patterns of differences in trait values between microform were generally similar in the Finland and Sweden bogs, but these patterns tended to be different in the

Scotland bog. In the Finland and Sweden bogs, D. rotundifolia growing on hummocks had more leaf trichomes, larger lamina area and longer leaves, but lower leaf-trichome density and were greener compared to hollows (Figure 3.3, Table 3.3).

These same traits on the Scottish bog, in contrast, did not differ between hummocks and hollows (Figure 3.3, Table 3.3). Specific leaf area and proportion of plants flowering did not change significantly across microforms in any sites. Comparisons across sites indicate that plants in Scotland have significantly lower trichome number and trichome density, whilst lamina area, and thus area for trichomes to form, is consistent across sites. Scotland had a significantly lower proportion, however, of flowering plants overall (Table 3.3).

76 Hollow Hummock A 15

) *** *** ns 2 − mm ( 10 density

5 Trichome 0 B 200 *** *** ns

150

100

50 Number of trichomes of Number

0 Hollow Hummock C 30 *** *** ns ) 2

mm 20 (

Area

10 Lamina

0 20 D ** *** ns

15 (mm)

10 length

5 Leaf

0 Finland Sweden Scotland Site Figure 3.3: Differences between D. rotundifolia leaf traits growing on different bog microforms (hummock vs hollow) in three different locations. Presented are A: trichome density, B: average (mean) number of trichomes per leaf, C: lamina area and D: length of the longest leaf (petiole plus lamina). Annotations above bars show the results of a priori planned comparisons tests for differences in leaf traits between microform within each bog. * P < 0.05, ** P < 0.005, *** P < 0.0001, (*) P < 0.10. 77 Table 3.3: Summary for hierarchical ANOVA analysis of Drosera rotundifolia morphological measurements and nutrient status in hummocks and hollows across three sites (Sw = Sweden, F = Finland, Sc = Inverewe, Scotland). ‘/’ denotes a significant difference ordered largest to smallest. Duplicate letters appearing either side of ‘/’ signifies a non-significant difference to adjacent letters, but the other two sites are significantly different. Microform (M) within site comparisons are represented using hollows as a base level and denoting whether hummocks have significantly higher [+] or lower [-] values or not significantly different [O], P value threshold of alpha 0.05. Where 0.05 < P < 0.1 arises, differences are in brackets. PC signifies planned contrast analysis of microform within each site. Site Site M x Site Finland Sweden Scotland differences Hummoc Hollow Hummock Hollow Hollow Hummock k Variable d.f. F P d.f. F P High first PC P Mean Mean PC P Mean PC P Mean Mean Mean ± SEM ± SEM ± SEM ± SEM ± SEM ± SEM Trichome + 158.38 197.58 + 154.80 186.18 O 153.64 150.880 2,24 27.29 < 0.001 3,24 32.86 < 0.001 F Sw / Sc number 0.001 ± 3.56 ±3.95 0.001 ± 2.897 ± 2.43 0.756 ± 5.08 ± 2.932

Trichome - 13.02 8.46 - 12.38 8.04 O 8.39 9.210 2,24 13.34 < 0.001 3,24 44.48 < 0.001 F Sw / Sc density < 0.001 ± 0.59 ± 0.60 < 0.001 ± 0.50 ± 0.30 0.294 ± 0.33 ± 0.563

+ 12.62 24.26 + 13.01 24.65 O 19.07 17.590 Lamina area 2,24 0.15 0.864 3,24 48.33 < 0.001 F Sw Sc < 0.001 ± 0.24 ± 0.81 < 0.001 ± 0.56 ± 1.17 0.274 ± 1.25 ± 1.283

O 0.67 0.67 + 0.66 0.70 - 0.70 0.67 Petiole:Leaf 2,24 2.15 0.138 3,24 4.57 0.011 F Sw Sc 0.956 ± 0.01 ± 0.01 0.008 ± 0.01 ± 0.01 0.047 ± 0.01 ± 0.01

Longest leaf + 11.41 15.59 + 11.21 17.33 - 16.42 14.143 2,24 4.10 0.030 3,24 25.58 < 0.001 Sc Sw / F Sw length 0.003 ± 0.26 ± 0.69 < 0.001 ± 0.28 ± 0.70 0.029 ± 0.94 ± 0.6030

O 5473.26 5041.95 (-) 6238.87 5548.04 - 6231.600 5458.918 SLA 2,24 5.02 0.015 3,24 4.21 0.016 Sw Sc / F 0.244 ± 150.46 ± 56.24 0.067 ± 391.77 ± 284.48 0.042 ± 86.844 ± 178.670

78 Site Site M x Site Finland Sweden Scotland differences Hummoc Hollow Hummock Hollow Hollow Hummock k Variable d.f. F P d.f. F P High first PC P Mean Mean PC P Mean PC P Mean Mean Mean ± SEM ± SEM ± SEM ± SEM ± SEM ± SEM CIE *a (High = O 18.89 11.92 - 30.14 13.70 O 3.07 7.90 2,24 35.44 < 0.001 3,24 14.76 < 0.001 Sw / F / Sc green, low = red) 0.200 ± 1.37 ± 1.06 0.005 ± 1.14 ± 2.60 0.371 ± 1.95 ± 2.89

O (+) 0.17 0.34 O 0.20 0.35 O 0.072 0.088 Flowering 2,24 7.92 0.002 3,24 3.08 0.0465 Sw F / Sc 0.075 ± 0.046 ± 0.04 0.121 ± 0.03 ± 0.11 0.867 ± 0.027 0.034

N derived from - 0.46 0.30 - 0.48 0.34 O 0.73 0.72 2,24 113.90 < 0.001 3,24 12.40 < 0.001 Sc / Sw F prey < 0.001 ± 0.012 ± 0.02 0.014 ± 0.04 ± 0.02 0.811 ± 0.03 ± 0.02

O 1.15 1.16 O 1.22 1.29 - 1.60 1.38 Total % N 2,24 150.50 < 0.001 3,24 23.10 < 0.001 Sc / Sw / F 0.909 ± 0.0103 ± 0.01 0.479 ± 0.03 ± 0.02 0.034 ± 0.02 ± 0.02

O 2.10 2.05 O 2.18 2.22 - 2.37 1.91 K 2,24 0.92 0.413 3,24 4.71 0.010 F Sw Sc 0.623 ± 0.1010 ± 0.22 0.729 ± 0.40 ± 0.05 0.001 ± 0.16 ± 0.11

O 0.2356 0.24 O 0.22 0.23 - 0.28 0.24 P 2,24 11.64 < 0.001 3,24 6.86 0.002 Sc / Sw F 0.869 ± 0.0059 ± 0.01 0.249 ± 0.01 ± 0.01 0.004 ± < 0.00 ± 0.01

O 4.8891 4.92 O 5.70 5.61 O 5.80 5.87 N:P 2,24 20.83 < 0.001 3,24 0.09 0.965 Sc / Sw / F 0.994 ± 0.2994 ± 0.53 0.809 ± 0.69 ± 0.45 0.843 ± 0.16 ± 0.09

O 0.5491 0.57 O 0.57 0.58 O 0.68 0.72 N:K 2,24 18.01 < 0.001 3,24 0.65 0.591 Sc / Sw / F 0.697 ± 0.0237 ± 0.03 0.816 ± 0.04 ± 0.02 0.433 ± 0.01 ± 0.01

79 Site Site M x Site Finland Sweden Scotland differences Hummoc Hollow Hummock Hollow Hollow Hummock k Variable d.f. F P d.f. F P High first PC P Mean Mean PC P Mean PC P Mean Mean Mean ± SEM ± SEM ± SEM ± SEM ± SEM ± SEM O 8.9443 8.72 O 10.08 9.66 O 8.56 8.13 K:P 2,24 9.08 0.001 3,24 0.508 0.680 Sw / F Sc 0.740 ± 0.2463 ± 0.63 0.527 ± 0.36 ± 0.40 0.525 ± 0.17 ± 0.16

O 20.502 20.78 O 22.09 22.18 O 23.80 24.44 10P/N(10P)K 2,24 23.349 < 0.001 3,24 0.317 0.813 Sc Sw F 0.814 ± 0.524 ± 0.30 0.941 ± 1.0 ± 0.40 0.941 ± 0.19 ± 0.07

O 41.980 42.51 O 38.82 39.65 O 41.08 41.69 10P/N(10P)K 2,24 8.904 0.001 3,24 0.414 0.745 Sc F / Sw 0.693 ± 0.330 ± 1.47 0.533 ± 0.36 ± 0.79 0.648 ± 0.33 ± 0.33

O 37.519 36.71 O 39.09 38.17 O 35.12 33.88 10P/N(10P)K 2,24 10.088 < 0.001 3,24 0.589 0.628 Sw F / Sc 0.641 ± 0.766 ± 1.49 0.596 ± 1.15 ± 0.86 0.474 ± 0.42 ± 0.39

+ 45.442 68.70 + 47.62 94.08 (-) 93.15 72.75 Plant size 2,24 11.47 < 0.001 3,24 17.75 < 0.001 Sc Sw / F 0.032 ± 1.897 ± 5.00 0.001 ± 1.16 ± 6.29 0.058 ± 9.13 ± 4.76

80 3.3.3 Drosera rotundifolia nutrient stoichiometry and N derived from prey

The proportion of D. rotundifolia N that was derived from prey did not directly follow patterns of shade or nutrient availability (as measured by Sphagnum spp. N content), but they did follow patterns of differences in trichome density (Figure 3.3

A, Figure 3.4 A). On the Scottish bog there was no difference in %N derived from prey between plants on hummocks and hollows. Plants on hollows in the Swedish bog obtained more N from prey than those on hummocks; the same pattern was present, but to a greater extent, on the Finnish bog. Drosera rotundifolia tissue N concentrations followed the patterns of N availability to some extent (Figure 3.2 B,

Figure 3.4 B). In Finland and Sweden, tissue N does not differ between microforms, but in Scotland there was significantly higher tissue N concentration in hollows than in hummocks (Table 3.3). This trend is the same for K and P tissue nutrient concentrations (Table 3.3). Nutrient stoichiometry of D. rotundifolia within microforms at all three sites remains consistent regardless of the proportion of nutrients derived from prey or root nutrient availability across microforms (Figure

3.4 C).

81 Hollow Hummock A A 1.00 AA 1.0015 A 1.00

) *** *** ns 2 − ** * ns** * ns** * ns 0.75 0.75mm 0.75 ( (%) (%) (%)

10 N N N

0.50 0.50 0.50 density derived derived derived

5 0.25 0.25 Prey 0.25 Prey Prey Trichome 0 0.00 B 0.00 0.00 200 B 2.0 BB 2.0 *** *** B 2.0 ns

** * ns** * ns** * ns 1.5 1501.5 1.5 N (%) N (%) N N (%) N N (%) N

1.0 1001.0 1.0

0.5 0.550 0.5 D. rotundifolia Number of trichomes of Number D. rotundifolia D. D. rotundifolia D. D. rotundifolia D.

0.0 0.0 0.0 Finland Sweden 0 ScotlandFinland Sweden ScotlandFinland Sweden Scotland Site Site Site CC 100

20 Site 80 Finland Sweden

40 N/(N(10P)K) Scotland P or P + N Limitation 100 60 Microform P/(N(10P)K)60 Hollow 20 40 Hummock K or K + N Limitation

80 80 20

40 100 N Limitation N/(N(10P)K) P or P + N Limitation 60 20 40 60 80 100 K/(N(10P)K)

P/(N(10P)K)60 Figure 3.4: Differences between D. rotundifolia growing on different microforms at three sites A: Proportion of prey derived N, B: Tissue %N, C: Ternary diagram of nutrient stoichiometry of D. rotundifolia within microform at each site. Nutrient40 limitation guides from Olde Venterink et al. (2003). K or K + N Limitation 82

100 N Limitation

20 40 60 80 100 K/(N(10P)K) 3.4 Discussion

I found clear evidence that microhabitat variability between hummocks and hollows influences within-habitat phenotypic variability in leaf traits and nutrition of the carnivorous plant D. rotundifolia. How these patterns manifested varied between peatlands due to differences in climate. On the site where ET:P was lowest

(Scotland), Sphagnum N content was highest in hollows, consistent with nutrient accumulation in hollows. The expected reduced allocation to carnivory was not present, presumably due to reduced shade in hollows compared to hummocks

(Figure 3.2 A). The impact of heterogeneity in shade and nutrients appears to be balanced in terms of the trade-off of inherent carbon costs associated with carnivory and the nutritional gain from prey capture (Ellison & Gotelli, 2009; Karagatzides &

Ellison, 2009). The equal investment in carnivory is due to entirely different resource limitations—high light reducing the cost of carnivory, and high N reducing the demand for N from prey. On the site with highest ET:P (Finland), root N availability appeared to be evenly distributed between hummocks and hollows. On this site, however, allocation to carnivory, and N from prey did vary between microform because of the increased shade on hummocks (Figure 3.3, Figure 3.4). The Sweden site was intermediate in ET:P and phenotypic variability. These results demonstrate that the impacts of climate variability on nutrient distribution in patterned peatlands

(Eppinga et al., 2010) alter phenotypic expression of D. rotundifolia in addition to variability in the distribution of vascular vegetation and subsequent shade.

Phenotypic variability of D. rotundifolia is consistent with the cost-benefit model

(Givnish, 2015), by responding to the combination of light and N. The controls (i.e. local climate regimes, nutrient flow and vegetation coverage) are predicted to be consistent on patterned peatlands across their range (Pastor et al., 2002; Eppinga et al., 2008, 2010; Rydin et al., 2013; Korrensalo et al., 2018), so it is likely that the impacts of these controls on within-site phenotypic variability in D. rotundifolia can be similarly generalised. This is the first evidence of within site variability in

83 allocation to carnivory and consequent prey nutrition across microforms in relation to interaction with vegetation, N and light availability. Moreover, this is the first evidence of the impact of differences in nutrient accumulation patterns within peatlands on phenotypic variability in plants driven by the local climate regime. This demonstrates the potential for climate change to indirectly impact on phenotypic variability and may ultimately determine habitat change, microhabitat loss and species composition.

Phenotypic trait variation, plasticity and local adaptation is historically not incorporated into species distribution models, and only very recently is this approach being considered for predicting the impact of climate change (Benito

Garzón et al., 2019). I demonstrate that this can be important and can itself vary in response to climate. Here, I show that even within a site there is systematic variability in the phenotype, and that variability changes between sites, in a systematic way related to local climate regimes and general patterns of vegetation, light and water availability. Not only does this indicate that local adaptation and trait variability are essential considerations for predicting future species distributions, but within site variation is also a vital aspect to be incorporated in these models (Van Der Molen et al., 1994; Randin et al., 2009; Benito Garzón et al.,

2019). My findings of variation in leaf traits and carnivory highlight the intraspecific trait variation of plants within a site, but also shows that microhabitat presence is also influenced by the local climate regime (Figure 3.3, Figure 3.4). Over time, specific habitats may no longer exist, and/or novel habitats may emerge. In this study, high N, low vegetation cover habitats only existed in the Scottish sites where precipitation was sufficiently high. It is only at the resolution of within-site impacts, that it is possible to accurately predict global changes in species distributions as even small-scale environmental changes are influencing trait variation.

84 3.4.1 Habitat heterogeneity

I found clear within-bog variability in resources, over small spatial scales (<1 m). The presence of vascular vegetation on hummocks and resultant shade was consistent across sites, where heather and evergreen trees occupy the drier areas of the bogs, and lower wetter areas are typified by a lawn of sphagnum with sedge grass (Van

Der Molen et al., 1994; Rydin et al., 2013). The extent of this variation is also different between bogs, which may be a function of topography, geology, overall N availability and the nutrient accumulation mechanism driven by ET:P. In this respect, the sites in this study are highly heterogeneous at small spatial scales. At this resolution, it was possible, however, to identify patterns of nutrient accumulation as a function of local ET:P (Figure 3.2 B). Therefore, not only are there general patterns of habitat heterogeneity across patterned peatlands, but localised climate regimes can also impact on habitat heterogeneity and trait variability within these environments. I show that carnivorous traits (e.g. trichome density) and resultant proportion of prey derived N significantly vary across microforms due to local ET:P (Figure 3.3, Figure 3.4 B). These findings suggest that understanding the impacts of habitat heterogeneity in these environments at regional resolutions may not be sufficient and future research on the impact of habitat heterogeneity must be highly contextualised.

Drosera rotundifolia have been shown to have phenotypic trait variation in response to patterns of light and nutrients (Zamora et al., 1998; Ellison & Gotelli, 2002).

Expression of carnivory decreases in association with increases in nitrogen availability along an environmental gradient (Alcalá & Domínguez, 2005; Millett et al., 2012). Additionally, decreases to the light environment have been shown ex situ to result in decreased expression in carnivory due to the increased cost of carnivory as carbon assimilation decreases (Thorén et al., 2003; Karagatzides & Ellison, 2009).

These instances of intraspecific trait variation had, until now, not been investigated in relation to habitat and microhabitat heterogeneity. Here, I show that intraspecific

85 trait variation arises from habitat heterogeneity. Furthermore, this intraspecific variation can be impacted indirectly by levels of ET:P which can lead to differing intensities of trait variation across habitat heterogeneity, and in some cases remove intraspecific variation across distinct microforms.

3.4.2 Prey capture and tissue N concentrations between microforms among sites

Total tissue N in Drosera rotundifolia is likely due to a combination of light and root N availability and a function of prey availability or plant-prey instances which was not measured in this study. The site in Scotland has much higher prey derived N in D. rotundifolia tissue compared to the other sites (Figure 3.4 A). Though not measured in this study, the frequency of midges, a major component of D. rotundifolia diet (Cook et al., 2018), was extremely high at the Scotland site compared to the other three sites.

Given Scotland hollows have more root N availability than hummocks, and high light, and almost certainly a higher prey availability, this results in hollows having significantly higher tissue %N than hummocks and Scotland having higher D. rotundifolia tissue N overall than the other two sites. Comparisons between the three sites in terms of ET:P driven patterns between microforms within each site remain valid. Prey abundance has no impact on the nutrient accumulation mechanism and therefore N accumulation was still present within this site. High prey abundance in

Scotland may, however, explain why trichome number and density were lower here than in Finland and Sweden at the site level as carnivory can be invested to a lesser extent whilst still having a high probability of prey capture. Understanding the impact of habitat heterogeneity, in the case of carnivorous plants, needs to also consider the abundance of potential prey as an additional variable of nutrient availability (Farnsworth & Ellison, 2007).

This study uses Sphagnum moss tissue nutrient composition to determine nutrient availability within site microforms. The patterns of Sphagnum N concentration confirm the expected patterns from modelling studies (Table 3.1 & Figure 3.2), and as found by Eppinga et al. (2010). Sphagnum spp. absorb inorganic N from

86 atmospheric deposition (Bragazza et al., 2005), and have been widely used as indicators of nutrient availability in peatlands as they are a more accurate indicator of environmental nutrient availability than phylogenetic differences (Limpens et al.,

2017). Sphagnum tissue N concentrations have been used as indicators of differences in nutrient availability between ombrotrophic bogs across a range of levels of atmospheric N deposition and are a good indicator up to a threshold, where uptake of nutrients is likely limited by some other limiting factor (Bragazza et al., 2005). The field sites for this experiment have sufficiently low N deposition (F = 7; Sw = 4; Sc = 7 kg N ha-1 year-1, Korhonen et al., 2013; APIS, 2016; SUAS, 2017) that I am confident of using Sphagnum tissue nutrient concentrations as an indicator of within site variability in N availability. Therefore, general patterns, but not absolute values, can be used to compare nutrient status of D. rotundifolia across sites in relation to nutrient availability within microforms.

In conclusion, the findings of this study suggest that climate variability—ET:P ratio—impacts on habitat heterogeneity in patterned peatlands, but not in the way predicted by existing models. Climate change will result in increased temperatures, evaporation and altered patterns of precipitation, which will alter local ET:P

(Breeuwer et al., 2009; Molau, 2010; Bragazza et al., 2013; EEA, 2016).

Evapotranspiration-precipitation ratio has been shown previously to impact on nutrient distribution within peatlands, but I demonstrate, for the first time, that these changes will also likely impact on plant intraspecific trait variation. Drosera rotundifolia is a useful indicator species, due to its sensitivity to environmental conditions such as water table, light availability and nutrient availability

(Nordbakken et al., 2004; Bruzzese et al., 2010; Cook et al., 2018). There is high potential for other important plant species and processes to also be impacted due to the fundamental importance of nutrient availability and shade. My results suggest important impacts on within site habitat composition, and microhabitat presence.

Such changes are important because existing species distribution models assume

87 that the movement of species will follow changes in temperature and precipitation

(Randin et al., 2009). But if they result in a fundamental change in the organisation and characteristics of habitats within peatlands, then these assumptions, and so the models based on them will not be valid. The implication of an increasing annual precipitation on sites that have a nutrient accumulation in hummocks (not in this study) or are intermediate (Finland & Sweden) is that there may be a fundamental regime shift in the nutrient cycling within the peatlands and therefore plant trait expression (Blanco et al., 2014). Though D. rotundifolia are present across the hummock-hollow microform, other species that are restricted to certain conditions within microforms because of N availability may lose their fundamental niche and will result in loss of biodiversity within these sites. The importance of understanding niche components within microforms within these sites and how these differ across sites along an environmental gradient of ET:P is paramount. The likelihood is that there will be environmental thresholds which, if they are breached, will result in a shift in nutrient accumulation to hummocks or hollows and as a result, plants may be stretched beyond their phenotypic variability. And at a rate far higher than can be adjusted by evolutionary adaptation.

88 3.5 Appendix C3

Principal Components analysis of Drosera rotundifolia morphological traits

Figure Appendix C3.1: PCA biplot of Drosera rotundifolia morphological traits (and

PC weightings and variable contributions.

The first three principal components explain a cumulative 88% of the total variance and each contribute at least 5% to the overall variance. From the biplot, five sections are well defined: Trichome Density, Colour (RedGreen), specific leaf area, flowering

+ leaf mass + trichome number, and a cluster of variables related to plant size (right side). PC1 is most associated with plant size, the strongest weightings being longest leaf length (and other leaf lengths), trichome density (negatively) and lamina area

(which also slightly separate to two clusters: Area and leaf length above and number below). PC2 is a combination of trichome number, specific leaf area (negatively) and flowering proportion. PC3 is predominantly flowering, specific leaf area and colour.

Biologically, flowering and trichome number are very different traits and should be considered separately in the analysis. Traits to remain in the analysis: Lamina area

(Right cluster top & PC1), Longest leaf (Right cluster centre & PC1), Petiole to leaf ratio (Right cluster bottom & PC1), Flowering (PC2, PC3 & upper cluster), Total trichome number (PC2 & Upper cluster), Trichome density (PC1 and left side),

89 RedGreen [colour] (Upper left side & PC3), Specific leaf area (PC2, PC3 & lower side)

Principal Components analysis of Drosera rotundifolia tissue nutrient percent concentration, N:P:K ratios and proportion of prey derived N

Figure Appendix C3.2: PCA of tissue element ratio and concentrations of Drosera rotundifolia.

The first five principal components explain 93% of the variance and each contribute at least 5% to the total variance. PC1 is predominantly prey derived N, total N and S

% concentration. PC2 is most associated with Zinc (negative), Mn and K:P ratio

(negative). PC3 is mostly K and P (and N:K). PC4 is mostly N:P ratio and K:P ratio.

PC5 is mostly Mg tissue concentration. The biplot suggests a general separation of tissue elemental composition with Ca, Co, Fe being accounted for by other variables more strongly associated with the predominant principal components. I am not interested in Mn or Zn in this study. Tissue element concentrations to remain in analysis: Prey derived N (PC1), % total N (PC1), P (PC2 + PC3), K (PC2 + PC3), N:P:K ratios (PC4 + PC4).

I then created bivariate matrices of the correlation of Drosera rotundifolia plant traits.

These data show a high but not exclusive correlation of lamina area with trichome density.

Table Appendix C3.1: Bivariate matrix of Drosera rotundifolia trait correlations. Correlations in red indicate a non-significant

Pearson’s correlation (P > 0.05).

Trichome Trichome Lamina Petiole:l Prey Drosera Petiole Specific Red Plant size Flowering number density area eaf derived N %N length leaf area Green Trap outer area 0.6095204 -0.71202 0.830626 0.178454 -0.18301 0.042436 0.576812 0.564558 -0.24582 -0.30969 0.188427 Trichome number 1 -0.31786 0.687787 0.089621 -0.43931 -0.25716 0.389061 0.341509 -0.2399 0.031231 0.279348 Trichome density -0.31786048 1 -0.83437 -0.29698 -0.01535 -0.28734 -0.66595 -0.64189 0.115687 0.441092 -0.14085 Lamina area 0.68778748 -0.83437 1 0.269735 -0.22831 0.052118 0.71028 0.675405 -0.21288 -0.29159 0.259986 Petiole:leaf 0.08962123 -0.29698 0.269735 1 0.029779 0.147503 0.778268 0.46949 0.091854 -0.22893 0.118674 Prey derived N -0.43931059 -0.01535 -0.22831 0.029779 1 0.632428 -0.0135 0.085724 0.131932 -0.24292 -0.21941 Drosera % N -0.25715862 -0.28734 05118 0.147503 0.632428 1 0.240362 0.368293 0.233252 -0.41682 -0.1639 Petiole length 0.3890614 -0.66595 0.71028 0.778268 0.240362 1 0.780995 -0.03694 -0.38945 0.195014 Plant size 0.34150905 -0.64189 0.675405 0.46949 0.085724 0.368293 0.780995 1 -0.02248 -0.41713 0.085138 Specific leaf area -0.23990005 0.115687 -0.21288 0.091854 0.131932 0.233252 1 0.051264 -0.07933 Red Green 0.03123074 0.441092 -0.29159 -0.22893 -0.24292 -0.41682 -0.38945 -0.41713 1 0.061168

91 Chapter 4 The function of secondary metabolites in plant carnivory

4.1 Introduction

Secondary plant metabolites are increasingly being considered as important drivers of plant diversification and evolution (Stevenson et al., 2017). This is in part due to their involvement in a large number of fundamental processes associated with plant fitness: regulating growth and interacting with the environment, conspecifics, competitors and symbionts and as a defence against abiotic and biotic stress

(Hartmann, 2007; Kessler & Baldwin, 2007; Agrawal, 2011). Plants retain a high metabolic diversity and have the genetic capacity to quickly generate novel compounds due to gene duplication and divergence (Pichersky & Gang, 2000).

Plants are therefore able to potentially adapt to new environments at higher rates than explained by mutation (Moore et al., 2014). If secondary plant metabolites play an important role in plant evolutionary adaptation, they should be inherently involved in processes specific to the characteristics of the plant’s ecology. The role of metabolites in adaptation may, therefore, be greater than previously considered, which would imply that we are missing important information on plant evolutionary ecology (Nishida, 2014; Moore et al., 2014). Consequently, similar patterns of metabolite production may emerge among close and distantly related taxa that have evolved under similar environmental conditions.

Carnivorous plants are a diverse, polyphyletic group of flowering plants, generally restricted to nutrient poor sites (Ellison & Adamec, 2018). Plant carnivory has evolved independently at least ten times in five orders (Fleischmann et al., 2018).

Carnivorous plants represent an example of convergent evolution of a single syndrome (Figure 4.1), to enhance growth and reproduction by attaining nutrients from prey. The benefits of carnivory are numerous: increased growth rate, earlier flowering, increased seed set, increased rate in photosynthetic output, increased nutrient uptake via roots after prey capture and increased seedling growth (Schulze et al., 2001; Adamec, 2002; Pavlovič et al., 2009; Hatcher & Hart, 2014). Within

92 carnivorous plants, there is striking evolutionary convergence (e.g. pitcher plants, see Figure 4.1, Figure 4.2), but also a wide diversity of morphological and physiological approaches to attract, capture, digest and assimilate prey, even within single evolutionary lineages (e.g. Nepenthales). It is now clear that, to some extent, plant carnivory has a biochemical basis (Table 4.1). As such, carnivorous plants are an ideal study system to understand the role of secondary plant metabolites in evolutionary ecology (Thorogood et al., 2017).

Determining the function of secondary metabolites in the carnivorous habit is considerably more challenging and less well understood than determining morphological function (Ellison & Adamec, 2018). This lack of knowledge is critical because growing understanding of the importance of the role of metabolites in rapid plant diversification and adaptation means that we may be missing important insight into the evolution of plant carnivory (Moore et al., 2014). Advanced analytical capabilities and reduced costs have resulted in an increase in the number of studies identifying secondary metabolites in carnivorous plants. The result is a growing body of literature describing secondary metabolites, mostly in relation to their anthropogenic use (Banasiuk et al., 2012; Egan & Van Der Kooy, 2013). We lack, however, a synthesis of current knowledge of specific biochemical function in relation to the tissue in which these compounds are produced; this is required to begin to link gene expression to secondary metabolite synthesis, carnivorous plant ecology and evolution.

93 Figure 4.1: Phylogenetic tree of carnivorous plants and research and metabolites identified with known function in carnivory. Branches are to emphasise the location of carnivorous taxa and do not signify time since divergence or a metric of relatedness. Trap morphology is grouped into: (a) flypaper; (b) snap-trap; (c) pitfall; (d) lobster pot; (e) suction bladder; and (f) pigeon-trap. Colour bands signify a common carnivorous ancestor. Tree generated using NCBI taxonomy browser, iTOL and APG (APG, 2016; Letunic & Bork, 2016). Nepenthales sensu stricto sister to Caryophyllales. 24 common compounds are found in the Nepenthales lineage and Sarraceniaceae lineage. Note that Aldrovanda and Utricularia are genera of aquatic carnivorous plants.

94 In this review I have synthesised current understanding of the secondary plant metabolites that are involved in the carnivorous syndrome. The aim is to use this synthesis to explore the extent to which secondary plant metabolites play a role in all aspects of plant carnivory across independent lineages. It will also highlight comparisons of secondary plant metabolite function across lineages, providing insight on the biochemical evolutionary convergence and divergence of carnivorous plants. Secondary metabolites were identified through a systematic search of

‘Google Scholar’ and ‘Web of Science’ using search terms ‘metabolite’, ‘carnivorous plant’ and their synonyms, carnivorous plant genera and prominent study species. I also searched the reference list of relevant journal articles in case papers were not identified from the previous search terms. I did not include studies where the focus was on extracting secondary metabolites, rather than considering their role in carnivory. Because I am interested in the role of metabolite function in plant carnivory, I have collated metabolites under the four key stages of the carnivorous habit: prey attraction, capture, digestion, and assimilation of nutrients.

4.2 Secondary metabolites involved in the attraction of prey

Darwin (1875) was the first to suggest that carnivorous plants might attract their prey through the production of odour. It has been demonstrated that some (but by no means all) carnivorous plant species actively attract prey by mimicking fruit and flowers using olfactory or visual cues, or through the production of extrafloral nectar

(Bennett & Ellison, 2009; Kreuzwieser et al., 2014). Attraction mechanisms are still, however, relatively poorly understood or inconsistent across species and the evidence is, in some cases, controversial.

4.2.1 Olfactory cues

Volatile organic compounds (VOCs), compounds volatile at ambient temperature, are involved in the attraction of prey; such animal attraction is not uncommon elsewhere in nature (Pichersky & Gershenzon, 2002). VOCs can increase the

95 frequency of prey interaction but have also been demonstrated to enable carnivorous plants to select prey by species and potentially by size (Di Giusto et al., 2010;

Kreuzwieser et al., 2014; Hotti et al., 2017). VOCs are by far the most researched group of metabolites produced by carnivorous plants (over half of the ~170 compounds listed in Table S1). Nepenthes rafflesiana, a vine species of pitcher plant, synthesises at least 54 different VOCs that, as a blend of compounds, attracts their prey (Di Giusto et al., 2010). Kreuzwieser et al. (2014) identified more than 60 different VOCs released by Dionaea muscipula (Venus flytrap). Their comparison of

VOC concentrations before and after prey capture suggested that 20 compounds were directly related to prey attraction, through mimicry of fruit and flowering scents. They also found differences in levels of 1-(1-cyclohexen-1-yl)-ethanone and dodecanoic acid, propyl ester before and after prey capture. These compounds have not previously been attributed as attractants to insects, suggesting that D. muscipula produce a mix of VOCs, some of which are the same as those in flowers and fruit, whilst others are novel. The synthesis of fruit and flower mimicking VOCs alongside novel VOCs may be a strategy for avoiding pollinator-prey conflict (see El-Sayed et al., 2016). Alternatively, it might be the case that novel VOCs were part of the evolution of snap-traps in D. muscipula as an adaptation for the capture of larger prey than co-occurring carnivorous plants (Hutchens & Luken, 2009).

Prey spectrum specialisation through the use of modified secondary plant metabolites is evident in other carnivorous plant genera. Nonanal is a floral scent compound, the synthesis of which is widespread in plants. This VOC is an attractant for Nematocera spp., which are abundant in the natural habitats of carnivorous plant species (i.e. bogs) and a common prey taxa (Ellison & Adamec, 2018). This VOC is synthesised by the closely related carnivorous plants; Sarracenia spp. and

Darlingtonia californica and confirmed to be a prey attractant through Y-tube olfactory choice assays ex situ (Di Giusto et al., 2010; Hotti et al., 2017). In the pitcher plant, Nepenthes rafflesiana, synthesis of VOCs differs between upper and lower

96 pitchers; this attracts a different spectra of prey—aerial insects to upper pitchers and ants to lower pitchers (Di Giusto et al., 2010). In addition, juvenile or freshly opened

N. rafflesiana pitchers that have captured few or no prey emit flower-like scents

(Bauer et al., 2009); putrefaction odours are produced by older pitchers that have accumulated prey. It is not clear whether this is through compounds emitted by the leaf tissue, or is a function of actual prey decomposition (Jürgens et al., 2009). This demonstrates the potential for prey spectrum specialisation within different tissues of a single individual, controlled by secondary metabolite synthesis. These results also highlight how the main biochemical drivers of attraction may change throughout the life of a plant or leaf.

There are several instances in which a VOC linked to the attraction of organisms may also serve an additional function in plant carnivory. Coniine is a similar metabolite to floral scent compounds found in non-carnivorous plants (Roberts &

Wink, 1998). This compound has been found to be produced in the lids and pitchers of Sarracenia spp., suggesting a prey attraction function (Hotti et al., 2017). In addition, coniine may play a role in retaining prey if concentrations are sufficient as it can act as an insect stunning agent, though actual concentrations required for this to be effective have not been found in situ. Pulegone, a floral scent compound also found in Sarracenia spp., functions as an attractant but may also function as an insecticide (Franzios et al., 1997), demonstrating the possibility that it plays a role in the killing of prey as well as attraction. Tetradecanoic acid and hexadecanoic acid methyl ester are both floral scent compounds and have also been found in the traps of D. muscipula and Sarracenia spp. (Kreuzwieser et al., 2014). Hexadecanoic acid production is increased by D. muscipula after feeding, suggesting some carnivorous function. Hexadecanoic acid is a major component of oils in plants and may act as a surfactant or be involved in enzymatic activity during prey digestion, though this has not been tested.

97 Production of insect reproductive pheromones for attraction of pollinators is well established in nature (Nishida, 2014). Some VOCs produced by carnivorous plant species are also thought to attract prey by mimicking insect reproductive pheromones. One individual of Sarracenia flava was found to contain 14-b-Pregna, a

VOC known to be a sex pheromone of the insect Eurygaster maura, though whether the plant produced this cannot be confirmed (and hence this is not included in Table

S1, see Hotti et al., 2017). In addition, Sarracenia spp. also produce actinidine and trans-jasmone, pheromones used by species of Hymenoptera. The direct role of these

VOCs for prey attraction has not been tested under a single compound regime, though a blend of VOCs is likely to enhance prey attraction efficiency similar to floral blends (Kreuzwieser et al., 2014).

This co-option of existing VOCs—from fruit and floral compounds for use as a prey attractant—demonstrates the versatility of secondary metabolites to provide multiple functions, differentiated by the location or tissue from which the compound is emitted. The potential independent evolution of VOCs similar to those used by insects to attract other insects is an example of convergent evolution driven by evolutionary pressure to attract insects, though for entirely different purposes.

Plants’ deception of insects is well documented, but carnivorous plants have adopted a multitude of strategies to lure and deceive animals for the ultimate goal of nutrient acquisition, in addition to reproduction (Ellison & Adamec, 2018).

Further work is needed to identify the specific roles of these novel and pre-existing compounds and to determine whether their effectiveness is dependent on the compounds acting as a blend, or if each compound acts as an effective attractant to prey. To date, no studies have investigated how VOC blends may change over time or growth stages i.e. longer than immediately before and after prey capture. Dionaea muscipula, for example, captures an entirely different spectrum of prey as seedlings compared to mature stages (Hatcher & Hart, 2014); it seems possible that their volatile blend may change over the lifecycle of the plant to reflect this specialisation.

98 A research gap exists in understanding how plastic the metabolic profile of carnivorous plants is in response to growth stage, nutrient status and environmental stimuli.

4.2.2 Visual cues

Secondary metabolites constitute the pigments that plants use to create the visual cues which play a key role in plant-animal communication. Anthocyanins, for example, produce a red colour that advertises fruit ripeness. Anthocyanins, flavonoids, carotenoids and betalains are responsible for flower colour and pattern which attracts pollen vectors for reproduction and dispersal (Tanaka et al., 2008).

These pigments have been assumed to attract prey to carnivorous plants, due to the distinctive red trap colouration of many carnivorous plants (Figure 4.1). The anthocyanins and co-pigments which constitute this colouration in trap leaves include kaempferol, quercetin and cyanidin (Sheridan & Griesbach, 2001). A variety of carnivorous plant traps also display UV reflectance patterns. These patterns are also suggested to have a prey-attraction function, and are attributed to the accumulation of phenolics, quinones and sugars (Kurup et al., 2013) (Figure 4.2 b).

There is, however, as yet no robust evidence that pigments in traps serve a prey attraction function ubiquitously (Bennett & Ellison, 2009). Red pigmentation is not likely to be an attractant to many invertebrate species as they cannot perceive red wavelengths of light, and in some carnivorous plants red pigmentation has been shown not to impact prey capture rate (Foot et al., 2014). The potential for pigmentation as a prey attractant may be plant species-prey specific, or colouration may indirectly attract prey due to a contrast to background colouration for insect species that do not perceive light in the red frequencies (Figure 4.2 c). Other functions for trap pigments have been suggested, such as photoprotection or light stress as seen in Drosera rotundifolia (Figure 4.2 a) (Millett et al., 2018). Analysis of pigments in carnivorous plants has, however, been useful in determining phylogeny. The Venus flytrap (Dionaea muscipula), formerly grouped under

99 Caryophyllales, synthesise anthocyanins rather than betalains (characteristic of the

majority of Caryophyllales) and provided further support for grouping these plants

under ‘Nepenthales’ using molecular phylogenetic data (Ellison & Adamec, 2018).

Interestingly, pigment concentrations were shown to be significantly different in

Figure 4.2: Function of secondary metabolites in trap colouration requires further investigation; (a) Drosera rotundifolia change in colour from accumulation of anthocyanin (red pigmentation) in response to environmental stimuli but this is not confirmed to be related to carnivory; (b) upper figure: Nepenthes species show redder locations, possibly to contrast trap peristome to background for prey attraction. Lower figure: Nepenthes spp. also use UV inflorescence for prey attraction; (c) Sarracenia purpurea with high diversity of red and green pigmentation from secondary compounds with high background contrast in situ.

100 petioles compared to traps of Dionaea muscipula. In particular, cyanidin was present in the digestive glands, but absent from the non-carnivorous leaf tissue (Henarejos-

Escudero et al., 2018). Given the limitations of current approaches, further work is required to determine the adaptive carnivorous function, if any, of the synthesis of these pigment metabolites.

4.2.3 Extra-floral nectaries

Nectar is a saccharide-rich liquid produced by plants in glands or nectaries and is composed of primary metabolites including sucrose, fructose and glucose, but is also often a combination of many secondary compounds (Stevenson et al., 2017).

Flowering plants possess nectaries that function as a reward or bribe to attract pollinating insects (Roy et al., 2017). Many plant species also possess extra-floral nectaries that perform functions other than pollinator attraction, such as recruiting animals that deter herbivores, and the nectar often contains additional metabolites

(Marazzi et al., 2013). Extra-floral nectaries are present on the peristome of carnivorous pitfall traps in the genera Darlingtonia, Heliamphora, Sarracenia,

Cephalotus and Nepenthes, and on the trap lobes of the snap-trap of D. muscipula

(Płachno, 2007) (Figure 4.3). Nepenthes are also known to have coprophagous relationships with birds and small mammals that perch on the pitcher rim to feed on the nectar (Clarke et al., 2009; Chin et al., 2010; Grafe et al., 2011). The chemical constituents of the nectar produced by these extra-floral nectaries have not been closely studied in carnivorous plants, though such nectaries have evolved independently at least three times in carnivorous plants and are likely to be a combination of primary metabolites (sugars) and secondary compounds such as

VOCs (Figure 4.3) (Givnish, 2015). The degree to which this convergent evolution also applies to nectar chemistry would be valuable to study to explain how carnivorous plants may alleviate the overlap between pollinators and prey.

Currently, as far as I am aware, there has been no study comparing the metabolite blend of a species’ floral, extrafloral and trap related extrafloral nectaries.

101 Figure 4.3: Evolution and production zones of metabolites for attraction of prey via use of extrafloral nectaries. Highlighted areas are highest density of extrafloral nectaries for attraction of prey. Areas for secretion of secondary metabolites and nectar are important for prey attraction and capture. Length of phylogenetic branches do not signify relatedness or any other metric of evolution.

4.2.4 Presence of metabolites across independent lineages of plant carnivory for prey attraction

This review has identified 26 secondary plant metabolites involved in the attraction of prey which are present across independent lineages of carnivory. These secondary metabolites are the only known aspect of carnivory where function of the same compound has been demonstrated across independent lineages of plant carnivory, with the exception of benzoic acid, of which the function is not confirmed but is suggested to have a function in attraction and digestion (Table S 4.2). There is also high overlap within different genera of the same carnivorous lineage (boxed ticks of

Table 4.1 and grey bars in Table S 4.2). These secondary plant metabolites are all volatile organic compounds and provide evidence of convergent evolution of carnivory at a biochemical scale.

102 The majority of these volatile organic compounds are common across non- carnivorous plant taxa as fruit or floral scents to attract pollinators. There are only a handful of genera, however, that have been studied for the presence of volatile organic compounds for prey attraction, and the majority are found in a few studies on Nepenthes spp., Dionaea muscipula and the separate carnivorous lineage of

Sarracenia spp. Future work on the role of secondary metabolites in attraction should focus on determining whether other lineages of carnivorous plant attract prey through the use of VOCs, particularly VOCs already known in other species. Given that most of the compounds found to be used in the attraction of prey are also widespread as fruit or floral scents, determining whether these compounds were present before evolutionary divergence or afterwards would show the frequency of exaptation of current plant metabolites or generation of novel compounds for the carnivorous syndrome.

4.3 Secondary metabolites involved in the capture of prey

Prey capture mechanisms are the most studied aspect of carnivory in plants (Ellison

& Adamec, 2018). The different approaches to prey capture result in novel and unusual morphologies and plant movements. In carnivorous plants, six distinct trapping mechanisms have been identified: (a) flypaper—leaves that secrete a sticky mucilage; (b) snap-traps—some of the most rapid plant movements in the kingdom;

(c) pitchers/pitfall traps; (d) lobster pot traps; (e) suction traps/ bladder traps; and (f) pigeon traps (Figure 4.1). These trapping mechanisms rely on distinct morphological adaptation, but in many cases, they also require the use of secondary metabolites to enhance capture efficiency and prey retention. Secondary metabolites can be split into two functional groups for prey capture: those which play a physical role, and those which signal prey capture.

103 4.3.1 Physical trapping mechanisms

The physical role of metabolites for trapping prey includes adhesive properties as found, for example, in mucilage of Drosera spp., glue of Roridula species and the fluids of pitcher plants, some of which have viscoelastic properties, or wax crystals and metabolites in nectar within the pitcher plants relying on slippery surfaces for capture (Riedel et al., 2003; Płachno, 2007; Bauer et al., 2008). One example of how secondary metabolites are involved in these functions is the use of extra-floral nectar to decrease friction on the peristome of Nepenthes spp. pitchers (Figure 4.3; Bauer et al., 2008). Reduced friction on the peristome increases prey capture efficiency by reducing the ability of prey (primarily ants, Formicidae) to adhere to the pitcher lip, increasing the likelihood of falling into the pitcher trap. Secondary metabolites specifically implicated in this process are as yet unknown, but given the novel, mechanical function of this nectar, chemical composition may differ from nectar produced specifically for attraction. A similar functional role is performed by erucamide in the pitcher plants of the unrelated genus Heliamphora, where it acts as a lubricating agent in the nectar produced around the peristome of the traps to increase trapping efficiency (Płachno, 2007). Plumbagin may also play a key role in prey capture for Nepenthes khasiana. Plumbagin is found on the rim of species of

Nepenthes pitcher plants and is suggested to work in tandem with the slippery surface by inflicting an anaesthetic or toxic influence on prey inhibiting their escape responses (Raj et al., 2011). Plumbagin is a secondary metabolite found in all genera of the Nepenthales and rarely outside this group of plants, though its specific function in carnivory in these species is yet to be fully determined (Schlauer et al.,

2005).

4.3.2 Anti-adhesive wax production used in prey capture

Epicuticular wax crystals are widespread in the plant kingdom and fulfil multiple functions. Within carnivorous plants, formation of wax is present in pitcher forming plants; Catopsis berteroniana, Brocchinia reducta, Cephalotus, Sarracenia, Darlingtonia,

104 Heliamphora and Nepenthes spp. (Riedel et al., 2003; Gaume et al., 2004; Gorb et al.,

2005; Poppinga et al., 2010). However, the specific structure and composition of the wax has only been documented in Nepenthes spp.

In pitcher plants of Nepenthes spp., the inner surface of the trap is densely covered with epicuticular wax crystals (Riedel et al., 2003). This wax impedes insect adhesion reducing their ability to escape from the pitcher. The wax constituents are secondary metabolites from primary alcohols, fatty acids, alkyl esters and n-alkanes, with the exact composition determining the specific formation pattern of the crystals.

Triacontanal is the crucial compound in the formation of the wax crystals. This aldehyde homologue contributes 43% of the entire wax crystal mass (Riedel et al.,

2003). Triacontanal is common in a variety of non-carnivorous flora cuticular waxes and epicuticular wax platelets (Naeem et al., 2012). The other components of

Nepenthes spp. pitcher wax are currently unknown. It is, however, likely that novel wax-forming metabolites are distributed throughout the crystal network or segregate into individual platelets and also form wax crystals to enhance prey capture (Gorb et al., 2005).

4.3.3 Mucilage and glue production for prey capture

Two variants of the sticky substance used in flypaper traps are produced by carnivorous plants. Within the Nepenthales (Drosera, Drosophyllum, Triphyophyllum), and within the Lamiales (Pinguicula and Byblis), they produce a polysaccharide- based mucilage. The mucilage of Drosera is composed mostly of primary metabolites and hence not included in Table S 4.2. Myo-Inositol represents a significant proportion of the non-polysaccharide organic compounds, though is also a primary metabolite, derived from glucose (Loewus & Murthy, 2000; Kokubun, 2017). Though no detailed analyses are available, the other mucilage-producing carnivorous plants likely produce a similar product to the polysaccharide-based mucilage described.

The mucilage is similar in appearance and characteristics, though confirmation of its chemical constituents would provide additional insight into the evolution of

105 strategies to capture prey as the flypaper strategy spans at least five separate lineages (Figure 4.1 a, Pereira et al., 2012).

The second type of flypaper glue is not mucilage, nor an aqueous solution of acidic polysaccharides as described above; it is a lipophilic resin with high viscosity and is produced by species of Roridula. The resinous glue is composed of triterpenoids and acylglycerides (Simoneit et al., 2008). No saccharides or proteins have been previously detected within this glue, but traces of flavonoids have been identified in two species (Wollenweber, 2007). This glue is found to be considerably more viscous than the mucilaginous secretions of other carnivorous plants. In addition, this resin- based glue is much more resilient to water submergence and does not wash away as does Drosera spp. mucilage. Here, two morphologically similar strategies to capture prey (flypaper) have evolved with distinctly contrasting chemical composition to act as a glue. One of a saccharide-based (primary metabolite) substance, one of a resin based (secondary metabolite) glue to achieve effective prey capture.

4.3.4 Prey capture signalling

Carnivorous plants have evolved strategies to detect prey capture and respond by synthesising specific metabolites to advance the process of nutrient uptake from prey (Nakamura et al., 2013; Krausko et al., 2017). These signals may instigate trap movement, the accumulation of digestive enzymes, prey digestion and prevention of microbial decay of prey (Table S 4.2).

Jasmonates are usually a component of plant defence systems. In Droseraceae, however, they have been co-opted to control the signalling of prey capture to instigate trap movement (Krausko et al., 2017). Jasmonates are lipid-based metabolites that typically act as phytohormones known to be involved in a wide range of regulatory processes. In particular, plant responses to stress, including herbivory and other kinds of biotic and abiotic stress, elicit jasmonate production, though they also serve as regulators of growth, photosynthesis and reproductive

106 development (Wasternack & Hause, 2013). Accumulation of these metabolites is usually a fast response, occurring within minutes or hours of wounding or herbivory

(Wasternack & Hause, 2013).

The sundew Drosera capensis increases concentrations of jasmonic acid and its conjugates in response to prey capture (Krausko et al., 2017). These increasing concentrations of jasmonic acid instigate leaf bending (Figure 4.4), which is targeted, due to jasmonate accumulation being controlled at an individual cell level, preventing a whole-plant response to jasmonates (Nakamura et al., 2013). This ensures that the traps curl and envelop around prey, increasing the trap surface area important in prey retention, digestion and assimilation. Drosera capensis has also retained the herbivory defence response instigated by levels of jasmonic acid.

Drosera capensis therefore differentiates between prey capture and herbivory through propagation of electrical signals and consequent jasmonic acid accumulation

(Krausko et al., 2017).

Figure 4.4: Drosera capensis leaf bending caused by jasmonate accumulation after prey capture. The regulation of this is not yet known. Credit: Christopher R. Hatcher 107 4.4 Secondary metabolites involved in the digestion of prey

Once captured, the prey of carnivorous plants are digested. Enzymes are fundamental to the digestion of prey and identification of these specialised proteins is a current and ongoing endeavour (for a review see Ravee et al., 2018). Secondary metabolites can instigate enzyme production, enhance the efficiency of digestion and protect the plant from infection by preserving prey whilst it is being digested.

Three species that share the same carnivorous ancestor—Nepenthes alata and the two closely related species, Drosera capensis and Dionaea muscipula—demonstrate how the same secondary metabolites (jasmonates) can act as a biochemical signal for indicating the same stimuli but instigating a range of processes. Drosera spp. and D. muscipula share a common Drosera-like ancestor (Gibson & Waller, 2009; Givnish,

2015), but they have different prey trapping mechanisms. Drosera spp. use sticky mucilage and slow leaf movement; Dionaea muscipula uses fast snap-traps (Forterre et al., 2005; Erni et al., 2008). Prey capture results in the accumulation of jasmonic acid,

12-oxo Phytodienoic acid and isoleucine conjugate of jasmonic acid in all three species, Nepenthes alata, Drosera capensis and Dionaea muscipula. While jasmonates instigate enzyme secretion in all three species, it also initiates trap movement (to slowly envelop prey) in Drosera capensis (Mithöfer et al., 2014; Krausko et al., 2017), and in Dionaea muscipula the accumulation of these secondary metabolites also instigates sealing of the ‘outer-stomach’ (Libiaková et al., 2014; Pavlovič et al., 2017).

In Dionaea musciupula, jasomonates instigate production of endopeptidase dionain, a compound involved in the chemical breakdown of prey (Libiaková et al., 2014). Trap movement in D. muscipula is, similarly to Drosera spp., instigated by intercellular electrical signals but relies on mechanical snap-buckling instability for the initial prey capture, followed by slow trap closure to form an ‘outer-stomach’ instigated by phytohormones. Reliance on a physical snap-buckling mechanism over phytohormone controlled leaf movement in D. muscipula is presumably because of the faster reaction times required for a snap-trap compared to a sticky flypaper trap

108 for prey capture and retention. It seems likely that the prey-capture signalling function is common to the three species because of their shared carnivorous ancestor, but the function of this signal has somewhat diverged because a different trap strategy evolved in Dionaea muscipula and Nepenthes spp.

Two additional groups of secondary metabolites appear to be part of the process of prey digestion: naphthoquinones and phenolics. In non-carnivorous plant species these secondary metabolites act as antibacterial protection against pathogenic attack on plant tissue (Krolicka et al., 2008). In carnivorous plants they appear to have been co-opted to preserve and aid in the digestion of prey by preventing decay from bacteria as well as functioning to protect the plant tissue during the process of prey breakdown.

Naphthoquinones are a class of organic compounds derived from naphthalene. In non-carnivorous plants, the naphthoquinone plumbagin is used as a deterrent against herbivorous insects, and has fungicidal and microbicidal effects (Paiva et al.,

2003). Similarly, juglone is inhibitory to insects, and so prevents herbivory.

Plumbagin and juglone accumulate in the trap leaves of Drosera spp., in response to prey capture, suggesting that they play a role in the process of carnivory (Egan &

Kooy, 2012). While their specific function has not yet been tested in carnivorous plants in situ, it seems likely that they enhance prey digestion by protecting the plant from pathogens that may otherwise infect the plant as a result of decaying prey.

Nepenthes khasiana utilises naphthoquinones as a chitin-induced antifungal agent found in its specialised pitcher fluid (Eilenberg et al., 2010). It appears to be used to preserve caught prey for digestion and assimilation of nutrients. Droserone and 5-O- methyldroserone have also been identified specifically for antifungal abilities in

Drosera spp., D. muscipula and Nepenthes spp. (Table 4.1, Acton, 2012).

109 4.4.1 Presence of compounds involved in digestion within the same carnivorous lineage

Plumbagin is present in the trap tissue of every genus of the Nepenthales—

Aldrovanda, Drosophyllum, Triphyophyllum, Dionaea, Drosera and Nepenthes—which share a common carnivorous ancestor (Table 4.1). Plumbagin is considered to function as an antimicrobial agent during digestion (Aung et al., 2002). Other possible roles, however, have been suggested. Plumbagin may function as an allelopathic, insecticidal, molluscidal, antifeedant, while it also potentially may be important in attraction as a method to differentiate prey with pollinators (Krolicka et al., 2008; Raj et al., 2011).

It is suggested that, as the potential benefits of carnivory increase, the morphology of carnivorous plant traps become more complex (Ellison & Adamec, 2018). Secondary metabolites involved in digestion seem, however, to be fairly consistent across several different trap types. If plumbagin and its isomers are critical in protecting the plant during digestion or enhancing prey digestion, it would explain why these compounds have persisted regardless of trap morphology. Their presence across other taxa has, in this case, been investigated—7-methyl-juglone and plumbagin are not present in five other lineages of carnivory (Cephalotus, Byblis, Roridula, Sarracenia and Heliamphora (of Ericales lineage) and Utricularia) (Zenk et al., 1969; Culham &

Gornall, 1994). These findings contribute important extensions to confirmation of compounds found in the Nepenthales. If preservation of prey or protection of trap tissue during prey digestion are such important adaptations, to the extent that they have persisted in all Nepenthales, why are these compounds not present across all carnivorous lineages? Other carnivorous lineages may have evolved novel compounds for the same function as the 1,4-naphthoquinones such as plumbagin in

Nepenthales, but to date this has not been investigated.

110 4.5 Secondary metabolites involved in nutrient assimilation and whole plant responses

Currently, the metabolic role of, and responses to, assimilation of nutrients from prey is unexplored. Carnivorous plants balance nutrient uptake and rate of photosynthesis within the same tissue—leaves, though root nutrient uptake is maintained in some species (Gao et al., 2015). Retained root nutrient uptake capability suggests, for some species, an interesting leaf–root signalling mechanism for nutrient status, distinct from typical nutrient dynamics in plants. Uptake of nutrients from prey via leaves has been shown either to maintain or increase root activity, the reverse response of leaf nitrogen uptake via stomata in non-carnivorous flora (Adamec, 2002; Gao et al., 2015). A novel plant response to leaf nutrient uptake suggests there is a new plant signalling process in this complex nutrient regulation system.

Other aspects of the assimilation of nutrients from prey are known. For example, breakdown of nonabsorbable compounds by secretion of enzymes in specialised glands and transmembrane transport of nutrients can occur through use of membrane carriers and endocytosis (Schulze et al., 1999; Plachno et al., 2006;

Adlassnig et al., 2012). Pulse–chase experiments and research into metabolites that are taken up by the plant via prey digestion have confirmed the uptake of carbon and nitrogenous compounds; but identification of subsequent plant response has varied (Adamec, 2002; Rischer et al., 2002; Karagatzides et al., 2009), Carbon from prey-derived protein in Nepenthes insignia was found to be utilised for production of plumbagin (Rischer et al., 2002). Conversely, Drosera capensis was found to have similar levels of quercetin and kaempferol before and after prey capture (Kováčik et al., 2012a). Plant response and signalling using plant synthesised metabolites in response to the assimilation of animal metabolites remains largely unexplored.

Metabolic profiling of trap, non-carnivorous leaf and root tissue before, during and after nutrient uptake from prey may provide insight into any biochemical signalling

111 within the plant to control assimilation or regulate nutrient balance within the whole plant. Research in this area may provide insight into exaptation of current plant regulation processes or a novel method for plants to control nutrient uptake via leaves.

4.6 Occurrence of secondary plant metabolites involved in carnivory across plant taxa

This review has identified ~170 secondary plant metabolites that have, or are highly likely to have, a role in some aspect of the carnivorous habit, some of which are isomers or derivatives or select groups of compounds. Of these, 44 secondary metabolites have been found to play a role in carnivory in more than one genus

(boxed ticks of Table 4.1, light grey bars in Table S 4.2). Twenty-six compounds are shared across independent evolutionary lineages of plant carnivory (Nepenthales and Sarraceniaceae, unboxed ticks in Table 4.1). However, only Nepenthales and

Ericales have been investigated for secondary plant metabolites from an ecological function perspective (Figure 4.1), and of the 33 studies used for the generation of

Table S 4.2, only one examines the metabolic role in carnivory across multiple evolutionary lineages (Jürgens et al., 2009). There are, however, many studies identifying these and other compounds across other taxa for some kind of anthropogenic use (pharmaceuticals or biotechnology), or as taxonomic markers, but their plant function is not specified (Zenk et al., 1969; Culham & Gornall, 1994;

Budzianowski, 1996; Muhammad et al., 2013); I excluded these studies from this review because these artificially derived compounds may have no function in the carnivorous habit.

The origins of the evolution of carnivory is beginning to be determined, but there remain key components of the evolution of the carnivorous syndrome that are yet to be identified. Aspects of the evolution of carnivory have recently been linked to plant defence (Pavlovič & Saganová, 2015). This link is appropriate given the similarity in function of response to and digestion of prey, compared to responses to

112 herbivory using anti-microbial and insecticidal compounds and phytohormones.

Similarly, most metabolites used for attraction are related to flower or fruiting scents

(Jaffé et al., 1995; Kreuzwieser et al., 2014). Given carnivorous plants, so far, are all angiosperms, it would be interesting to investigate the genetic origins of these compounds, as it is currently unclear whether compounds for attraction are similarly co-opted from pollinating and dispersal mechanisms, or whether novel co-evolution has taken place to mimic olfactory cues.

The majority of metabolites, and metabolites identified so far across different taxa, are involved in the attraction and or capture phases of plant carnivory, this may however, be falsely disproportionate (Table S 4.2). Over representation of metabolites involved in attraction for plant carnivory is understandable, given the relative simplicity in inferring their functional role. Presence of volatiles found on traps are logically assumed to have some olfactory attraction for animals and up- or down-regulation of these compounds following prey capture reinforces this assumption. Similarly, composition of sticky mucilage exuded on traps leads to a conclusion of a function in capture or retention (though potentially also attraction).

Only a handful of studies identifying compounds in attraction or capture specifically test these compounds with prey capture assays (e.g. Jürgens et al., 2009).

Furthermore, classification of many VOCs is from a technical chemical standpoint and may not have a significant impact on olfactory attraction in situ as their volatility is not realised under biologically relevant temperatures. More stringent restrictions of inclusion of these compounds would reduce the overall number of compounds significantly, though perhaps unnecessarily. Further investigation of these compounds is required to determine their function, rather than their outright exclusion as important compounds. This, coupled with the challenge of identifying a carnivorous role of a compound in digestion and assimilation, may have led to an overrepresentation of certain compounds in the literature and may erroneously suggest that only these aspects are strongly reliant on secondary plant metabolites.

113 There are currently many secondary metabolites extracted from specifically carnivorous plants for pharmaceutical, pseudo-medicinal or biotechnological use

(see Ellison & Adamec, 2018 and references therein). These could be compared with chemicals where a functional role is known in order to establish metabolic similarities across taxa. Experiments testing the function of these known compounds would be an effective way to rapidly grow the library and taxonomic presence of compounds strictly related to plant carnivory that function in processes other than attraction or capture.

114 Table 4.1: Secondary plant metabolites involved in carnivory present in more than one genera. Role indicates role in the carnivorous habit, A: Attraction, C: Capture, D: Digestion. Ticks represent the 26 compounds present across independent lineages of carnivory. Ticks within boxes represent presence of compound found within different genera of the same carnivorous ancestor. Purple are of Nepenthales, blue are of Ericales lineage of carnivory. Note that some compound isomers are identified, but in some studies, this is not included. References are included in Table S 4.2.

Role Metabolite

Aldrovanda Drosophyllum Triphyophyllum Dionaea Drosera Nepenthes Darlingtonia Heliamphora Sarracenia A Benzyl acetate ü ü A Benzyl benzoate ü ü A Ethyl benzoate ü ü A (Z,E)-α-farnesene ü ü A Heptadecane ü ü Hexadecanoic acid ethyl A ü ü ester A Isoamyl acetate ü ü A Linalool ü ü A Methyl benzoate ü ü A Myrcene ü ü A p-Cymene ü ü A Palmitic acid methyl ester ü ü A Phenylethyl acetate ü ü A Tetradecanoic acid ü ü A Tridecane ü ü A α-Copaene ü ü A α-Terpinene ü ü A 2-Phenylethanol üþ üþ ü A 6-Methyl-5-hepten-2-one üþ üþ üþ ü A Benzaldehyde üþ üþ ü

115 Role Metabolite

Aldrovanda Drosophyllum Triphyophyllum Dionaea Drosera Nepenthes Darlingtonia Heliamphora Sarracenia A or D Benzoic acid üþ üþ üþ ü A Benzyl alcohol üþ üþ ü A Limonene üþ üþ üþ ü A Methyl salicylate üþ üþ ü A Nonanal üþ üþ ü A α-Pinene üþ üþ ü A Coniine þ þ þ A & C Erucamide þ þ þ A Sarracenin þ þ A Acetophenone þ þ A Caryophyllene oxide þ þ A Decanal þ þ A Hexadecane þ þ A Humulene þ þ A Pentadecane þ þ A Tetradecane þ þ C & D Jasmonates þ þ þ C & D Plumbagin þ þ þ þ þ þ D 7-methyl-juglone þ þ þ þ Droserone (plumbagin D þ þ derivative) D Gallic acid þ þ þ D Isoshinanolone þ þ D Myricetin þ þ D PlumbasideA þ þ D Protocatechuic acid þ þ þ

116 4.7 Concluding remarks and future perspectives

Secondary metabolites are known to be important to carnivorous plants in the attraction, capture and digestion of prey for nutrient assimilation. There is a high degree of co-option as well as generation of novel metabolites to aid the carnivorous syndrome. This is likely due to inherently high chemical diversity and the utility of these compounds for multiple functions (Moore et al., 2014). A key question is the importance of secondary metabolites in carnivorous plant evolution. Have similar metabolic approaches to carnivory evolved independently in different carnivorous plant lineages? There is evidence of metabolites involved in carnivory present in more than one genus and, in some cases, these are found across independently evolved lineages (Table 4.1, also highlighted dark grey in Table S 4.2). There may be great potential in examining the role of novel secondary metabolites for carnivory in different species, which differ in their morphological approach to carnivory, but which share a common ancestor (e.g. Nepenthales, purple block of Figure 4.1 &

Table 4.1).

There is growing interest in the role of secondary metabolites in the rapid diversification and evolution of plants to new environments. This is due to high chemical diversity, with rapid generation of novel compounds being attributed to adaptation rather than mutation (Moore et al., 2014). Carnivorous plants have been found to have huge chemical diversity, of which we know most about through their use in biotechnology (see Ellison & Adamec, 2018 and references therein). I hypothesise that metabolite diversity provided a mechanism for the evolution of carnivory in plants, and that this continued diversity facilitates rapid evolution into new environments. Due to the multiple times carnivory has independently evolved, and the general restriction of carnivorous plants to high-stress environments, these plants are an ideal system to investigate whether metabolic diversity may have been or is a way for novel traits to evolve and be maintained. Metabolites are clearly integral to the survival of carnivorous plants in nutrient-deficient environments and

117 further investigation of the carnivorous habit will continue to provide new insights into novel pathways of plant evolution.

118 4.8 Supplementary Material

Table S 4.2: Secondary plant metabolites involved in plant carnivory. Light grey bars signify the compound is found in more than one genus of the same carnivorous lineage. Dark grey bars signify the compound is found in more than one independent lineage of carnivory. No compound has been found in more than 2 separate lineages of carnivory. Compounds reported from Hotti et al., 2017 and Kreuzweiser et al., 2014 have 70% confidence minimum. Function of metabolite is known or highly likely. Note that some compound isomers are identified, but in some studies, this is not included. Role in plant Metabolite Group Likely function Plant genus/ species studied Family Order Source carnivory

Floral or fruit Attraction -Eudesmol Sesquiterpenoid Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 scent

(E,E)-a- Floral or fruit Nepenthales Attraction Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 farnesene scent (Caryophyllales)

119 Floral or fruit Attraction (E)-Coniferol Monolignols Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 scent

Fruit or floral Nepenthales Attraction (E)-dendrolasine Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Nepenthales Attraction (E)-nerolidol Sesquiterpenoid Floral scent Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 (Caryophyllales)

Floral or fruit Nepenthales Attraction (E)-ocimene Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

120 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction (E)-!-Ocimene Monoterpenoid individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

Nepenthales (Z,E)-α- Scent (wood, Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Sesquiterpenoid (Caryophyllales) farnesene sweet) Sarracenia flava, S. minor Sarraceniaceae Giusto et al., 2010 , Ericales

Scent (green (Z)-3-hexen-1- grass/ attractant Attraction Alkenoic alcohol Sarracenia flava Sarraceniaceae Ericales Jürgens et al., 2009 ol for predatory insects)

Floral or fruit Nepenthales Attraction (Z)-dendrolasine Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

121 Nepenthales Attraction (Z)-nerolidol Sesquiterpenoid Waxy scent Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 (Caryophyllales)

Floral or fruit Nepenthales Attraction (Z)-ocimene Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Scent (citrus, Sarracenia flava, S. leucophylla, Attraction (Z)-!-ocimene Monoterpenoid Sarraceniaceae Ericales Jürgens et al., 2009 herb, flower) S. minor, S. purpurea

VOC (emitted 1-(1- from trap but not Nepenthales Kreuzwieser et al., Attraction Cyclohexen-1- Alkenoic ketone individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 yl)-ethanone confirmed to attract prey)

122 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction 1-Dodecanol Alkanoic alcohol individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not individually Nepenthales Kreuzwieser et al., Attraction 1-Octadecanol Alkanoic alcohol Dionaea muscipula Droseraceae confirmed to (Caryophyllales) 2014 attract prey) waxy odour

Floral or fruit scent Nepenthales Kreuzwieser et al., Attraction 1-Octanol Alkanoic alcohol (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture) waxy

1-Phenylbutan- Floral or fruit Nepenthales Attraction Aromatic alcohol Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 2,3-diol scent (Caryophyllales)

123 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction 1-Tetradecanol Alkanoic alcohol individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

Fruit or floral Attraction 2-Decen-1-ol Alkenoic alcohol Sarracenia purpurea Sarraceniaceae Ericales Jürgens et al., 2009 scent

Floral or fruit Nepenthales Attraction 2-Heptanone Alkanoic ketone Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Nepenthales Attraction 2-Nonanone Alkanoic ketone Fruit scent Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 (Caryophyllales)

124 Fruit or floral Attraction 2-Nonen-1-ol Alkenoic alcohol Sarracenia purpurea Sarraceniaceae Ericales Jürgens et al., 2009 scent

Scent (honey, spice, rose, lilac). Attracts a wide Jürgens et al., 2009; Di 2-Phenylethanol Dionaea muscipula, Nepenthes Droseraceae, Nepenthales range of taxa Giusto et al., 2010; Attraction (Phenylethyl Aromatic alcohol rafflesiana, Sarracenia flava, S. Nepenthaceae, (Caryophyllales) (downregulated Kreuzwieser et al., alcohol) leucophylla, S. minor Sarraceniaceae , Ericales after prey capture 2014 in Dionaea muscipula)

2-Phenylethyl Nepenthales acetate Aromatic acid Scent (fruity, Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction (Caryophyllales) (Phenethyl ester sweet) Sarracenia flava, S. leucophylla Sarraceniaceae Giusto et al., 2010 , Ericales acetate)

2,5-di-tert- Floral or fruit Attraction Aromatic amine Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 butylaniline scent

125 VOC Nepenthales Kreuzwieser et al., Attraction 3-Carene Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

VOC instigates 4- response from Sarracenia leucophylla, S. Attraction Ketone Sarraceniaceae Ericales Jürgens et al., 2009 Oxoisophorone butterflies, moths purpurea and bees

4,8,12,16- Alkanoic acid Floral or fruit Attraction Tetramethylhept Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 ester scent adecan-4-olide

Scent (unknown 5-Hydroxy- Sarracenia flava, S. leucophylla, Attraction Aromatic alcohol specific Sarraceniaceae Ericales Jürgens et al., 2009 methylfurfural S. minor attraction)

126 Dionaea muscipula, Drosera Jürgens et al., 2009; Di Droseraceae, Nepenthales 6-Methyl-5- binata, Nepenthes rafflesiana, Giusto et al., 2010; Attraction Monoterpenoid Scent Nepenthaceae, (Caryophyllales) hepten-2-one Sarracenia flava, S. leucophylla, Kreuzwieser et al., Sarraceniaceae , Ericales Sarracenia minor, S. purpurea 2014

Nepenthales Attraction α -Selinene Sesquiterpenoid Scent (amber) Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 (Caryophyllales)

VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Acetic acid Alkanoic acid individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not Acetic acid Alkanoic acid Nepenthales Kreuzwieser et al., Attraction individually Dionaea muscipula Droseraceae methyl ester ester (Caryophyllales) 2014 confirmed to attract prey)

127 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Acetone Ketone individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from Dionaea muscipula trap Di Giusto et al., 2010; Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Acetophenone Aromatic ketone but not Kreuzwieser et al., rafflesiana Nepenthaceae (Caryophyllales) individually 2014 confirmed to attract prey)

Pyridine Attraction (insect Attraction Actinidine Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 derivative pheromone)

Floral or fruit Nepenthales Attraction b-Selinene Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

128 Scent (almond, burnt sugar) attracts Lepidoptera, Jürgens et al., 2009; Di Pieridae). Emitted Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Aromatic Giusto et al., 2010; Attraction Benzaldehyde from Dionaea rafflesiana, Sarracenia flava, S. Nepenthaceae, (Caryophyllales) aldehyde Kreuzwieser et al., muscipula trap leucophylla Sarraceniaceae , Ericales 2014 but not individually confirmed to attract prey.

Attraction Floral or fruit (also see Benzoic acid Aromatic acid Sarracenia purpurea Sarraceniaceae Ericales Jürgens et al., 2009 scent. digestion)

Aromatic Attraction Benzothiazole Floral scent Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 heterocycle

Scent (floral - Nepenthales Aromatic acid attracts Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Benzyl acetate (Caryophyllales) ester Lepidoptera, Sarracenia flava, S. leucophylla Sarraceniaceae Giusto et al., 2010 , Ericales Noctuidae)

129 Sweet, flower scent (emitted from Dionaea Jürgens et al., 2009; Di Dionaea muscipula, Nepenthes Droseraceae, Nepenthales muscipula trap Giusto et al., 2010; Attraction Benzyl alcohol Aromatic alcohol rafflesiana, Sarracenia flava, S. Nepenthaceae, (Caryophyllales) but not Kreuzwieser et al., leucophylla Sarraceniaceae , Ericales individually 2014; Hotti et al., 2017 confirmed to attract prey)

Scent (balsamic, Nepenthales Aromatic acid oil, herb, attracts Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Benzyl benzoate (Caryophyllales) ester Lepidoptera, Sarracenia leucophylla Sarraceniaceae Giusto et al., 2010 , Ericales Noctuidae)

Scent (vegetative, Aromatic acid Attraction Benzyl tiglate earthy, Sarracenia flava, S. leucophylla Sarraceniaceae Ericales Jürgens et al., 2009 ester mushroom)

VOC (emitted from trap but not Butanal, 3- Alkanoic Nepenthales Kreuzwieser et al., Attraction individually Dionaea muscipula Droseraceae methyl- aldehyde (Caryophyllales) 2014 confirmed to attract prey)

130 Scent Nepenthales Kreuzwieser et al., Attraction Camphene Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

Floral or fruit scent Di Giusto et al., 2010; Caryophyllene (downregulated Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Sesquiterpene Kreuzwieser et al., oxide after prey capture rafflesiana Nepenthaceae (Caryophyllales) 2014 in Dionaea muscipula)

Attraction Cinerone Alkenoic ketone Scent Heliamphora spp. Sarraceniaceae Ericales Jaffé et al., 1995

cis-linalool Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide (furan) scent (Caryophyllales)

131 cis-linalool Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide (pyran) scent (Caryophyllales)

cis-ocimene Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide scent (Caryophyllales)

Volatile organic cis-para-2- compound Nepenthales Kreuzwieser et al., Attraction Monoterpenoid Dionaea muscipula Droseraceae Menthen-1-ol (downregulated (Caryophyllales) 2014 after prey capture)

cis-α- Fruit or floral Sarracenia flava, S. leucophylla, Attraction Sesquiterpenoid Sarraceniaceae Ericales Jürgens et al., 2009 Bergamotene scent S. minor

132 Floral or fruit scent (emitted from Dionaea Di Giusto et al., 2010; Alkanoic muscipula trap Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Decanal Kreuzwieser et al., aldehyde but not rafflesiana Nepenthaceae (Caryophyllales) 2014 individually confirmed to attract prey)

Rotten egg scent. Sulphur Dimethyl Attracts Diptera, Attraction containing Sarracenia purpurea Sarraceniaceae Ericales Jürgens et al., 2009 disulphide Calliphoridae, compound Muscidae

Dimethyl Aromatic acid Fruit or floral Attraction Sarracenia leucophylla Sarraceniaceae Ericales Jürgens et al., 2009 salicylate ester scent

VOC (emitted from trap but not Alkanoic Nepenthales Kreuzwieser et al., Attraction Dodecanal individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

133 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Dodecane Alkane individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC Attraction Dodecanoic acid Alkanoic acid of prey Nepenthales Kreuzwieser et al., Attraction Dionaea muscipula Droseraceae propyl ester ester (downregulated (Caryophyllales) 2014 after prey capture)

Unknown Nepenthales Kreuzwieser et al., Attraction Eicosane Alkane (upregulated after Dionaea muscipula Droseraceae (Caryophyllales) 2014 prey capture)

Ethyl 2- Alkanoic acid Nepenthales Attraction Fruit scent Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 methylbutanoate ester (Caryophyllales)

134 Pineapple scent, attracts Diptera, Muscidae, Alkanoic acid Sarcophagidae, Attraction Ethyl acetate Sarracenia flava Sarraceniaceae Ericales Jürgens et al., 2009 ester Drosophilidae, Coleoptera, Scarabaedidae, Nitidulidae

Nepenthales Aromatic acid Scent (chamomile Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Ethyl benzoate (Caryophyllales) ester flower, fruit) Sarracenia flava, S. leucophylla Sarraceniaceae Giusto et al., 2010 , Ericales

Floral or fruit Nepenthales Attraction Germacrene A Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Floral scent (emitted from Dionaea Nepenthales muscipula trap Dionaea muscipula, Sarracenia Droseraceae, Kreuzwieser et al., Attraction Heptadecane Alkane (Caryophyllales) but not spp. Sarraceniaceae 2014; Hotti et al., 2017 , Ericales individually confirmed to attract prey)

135 VOC (emitted from trap but not Alkanoic Nepenthales Kreuzwieser et al., Attraction Heptanal individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

Scent Di Giusto et al., 2010; Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Hexadecane Alkane (downregulated Kreuzwieser et al., rafflesiana Nepenthaceae (Caryophyllales) after prey capture) 2014

Floral scent Hexadecenoic Attraction Alkenoic acid compound for Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 acid attraction

VOC (emitted from trap but not individually Dionaea muscipula, Nepenthales Jaffé et al., 1995; Hexadecanoic Alkanoic acid confirmed to Droseraceae, Attraction Heliamphora heterodoxa, H. (Caryophyllales) Kreuzwieser et al., acid ethyl ester ester attract prey), Sarraceniaceae tatei , Ericales 2014 Found on the spoon of Heliamphora spp.

136 VOC (emitted from trap but not Alkanoic Nepenthales Kreuzwieser et al., Attraction Hexanal individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted Humulene (α - from trap but not Di Giusto et al., 2010; Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction humulene, α - Sesquiterpene individually Kreuzwieser et al., rafflesiana Nepenthaceae (Caryophyllales) caryophyllene) confirmed to 2014 attract prey)

Nepenthales Alkanoic acid Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Isoamyl acetate Scent (banana) (Caryophyllales) ester Sarracenia flava Sarraceniaceae Giusto et al., 2010 , Ericales

Floral or fruit Attraction Isochiapin B Sesquiterpene Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 scent

137 VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Isomenthol Monoterpenoid individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not Isopinocampheo Nepenthales Kreuzwieser et al., Attraction Monoterpenoid individually Dionaea muscipula Droseraceae l (Caryophyllales) 2014 confirmed to attract prey)

Aromatic amino Floral or fruit Attraction Lagumicine Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 acid ester scent

Lemon, orange scent. Attracts Lepidoptera, Dionaea muscipula, Drosera Jürgens et al., 2009; Di Droseraceae, Nepenthales Noctuidae binata, Nepenthes rafflesiana, Giusto et al., 2010; Attraction Limonene Monoterpenoid Nepenthaceae, (Caryophyllales) (downregulated Sarracenia flava, S. leucophylla, Kreuzwieser et al., Sarraceniaceae , Ericales after prey capture S. minor, S. purpurea 2014 in Dionaea muscipula)

138 Floral, lavender scent. Attracts Nepenthales Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Linalool Monoterpenoid Lepidoptera, (Caryophyllales) Sarracenia minor Sarraceniaceae Giusto et al., 2010 Noctuidae, , Ericales Hymenoptera

VOC (emitted from trap but not Alkenoic Nepenthales Kreuzwieser et al., Attraction Methacrolein individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

Methyl 2- hydroxyy-3- Aromatic acid Attraction Scent Sarracenia flava Sarraceniaceae Ericales Jürgens et al., 2009 phenylpropionat ester e

Methyl 5- Aromatic acid Floral or fruit Attraction Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 ethylnicotinate ester scent

139 Methyl Aromatic amino Attracts Diptera Attraction Sarracenia purpurea Sarraceniaceae Ericales Jürgens et al., 2009 anthranilate acid ester and Chloropidae

Scent (prune, Nepenthes rafflesiana, Nepenthales Aromatic acid lettuce, herb, Nepenthaceae, Jürgens et al., 2009; Di Attraction Methyl benzoate Sarracenia spp., Sarracenia (Caryophyllales) ester sweet attracts Sarraceniaceae Giusto et al., 2010 flava , Ericales Lepidoptera)

Found on the spoon of the Methyl Alkanoic acid Heliamphora heterodoxa, H. Attraction pitcher assumed Sarraceniaceae Ericales Jaffé et al., 1995 linolelaidate ester tatei to be for attraction

VOC (emitted from trap but not Methyl Alkanoic acid Nepenthales Kreuzwieser et al., Attraction individually Dionaea muscipula Droseraceae myristate ester (Caryophyllales) 2014 confirmed to attract prey)

140 Methyl Aromatic acid Floral or fruit Attraction Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 nicotinate ester scent

Scent (Peppermint, attracts Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Methyl Aromatic acid Jürgens et al., 2009; Di Attraction Lepidoptera, rafflesiana, Sarracenia flava, S. Nepenthaceae, (Caryophyllales) salicylate ester Giusto et al., 2010 Noctuidae, leucophylla Sarraceniaceae , Ericales usually acts as a stress response)

Scent (balsamic, must, spice, Nepenthales Nepenthes rafflesiana, Nepenthaceae, Jürgens et al., 2009; Di Attraction Myrcene Monoterpenoid attracts (Caryophyllales) Sarracenia flava, S. leucophylla Sarraceniaceae Giusto et al., 2010 Lepidoptera, , Ericales Noctuidae)

Scent n-Butyl Alkanoic acid Nepenthales Kreuzwieser et al., Attraction (downregulated Dionaea muscipula Droseraceae myristate ester (Caryophyllales) 2014 after prey capture)

141 VOC (emitted from trap but not n-Hexadecanoic Nepenthales Kreuzwieser et al., Attraction Alkanoic acid individually Dionaea muscipula Droseraceae acid (Caryophyllales) 2014 confirmed to attract prey)

Alkanoic acid Attraction n-Hexyl acetate Fruit or herb scent Sarracenia flava Sarraceniaceae Ericales Jürgens et al., 2009 ester

VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Nonadecane Alkane individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

Droseraceae, Nepenthales Di Giusto et al., 2010; Alkanoic Dionaea muscipula, Nepenthes Attraction Nonanal Floral scent Nepenthaceae, (Caryophyllales) Kreuzwieser et al., aldehyde rafflesiana, Sarracenia spp. Sarraceniaceae , Ericales 2014; Hotti et al., 2017

142 Scent (unknown Nepenthales Attraction Nonanoic acid Alkanoic acid specific Drosera binata Droseraceae Jürgens et al., 2009 (Caryophyllales) attraction)

VOC (emitted from trap but not Alkanoic Nepenthales Kreuzwieser et al., Attraction Octanal individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not Oleic acid Alkanoic acid Nepenthales Kreuzwieser et al., Attraction individually Dionaea muscipula Droseraceae methyl ester ester (Caryophyllales) 2014 confirmed to attract prey)

Floral or fruit scent Monoterpenoid Nepenthales (downregulated Dionaea muscipula, Sarracenia Droseraceae, Kreuzwieser et al., Attraction p-Cymene (in this instance (Caryophyllales) after prey capture spp. Sarraceniaceae 2014; Hotti et al., 2017 is aromatic) , Ericales in Dionaea muscipula)

143 VOC (emitted from trap but not individually Dionaea muscipula, Nepenthales Jaffé et al., 1995; Palmitic acid Alkanoic acid confirmed to Droseraceae, Attraction Heliamphora heterodoxa, H. (Caryophyllales) Kreuzwieser et al., methyl ester ester attract prey), Sarraceniaceae tatei , Ericales 2014 found on the spoon of Heliamphora spp. Floral or fruit scent (emitted from Dionaea Di Giusto et al., 2010; muscipula trap Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Pentadecane Alkane Kreuzwieser et al., but not rafflesiana Nepenthaceae (Caryophyllales) 2014 individually confirmed to attract prey)

Floral or fruit Nepenthales Attraction Perillene Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Attraction Phenol Aromatic alcohol Scent Heliamphora spp. Sarraceniaceae Ericales Jaffé et al., 1995

144 Phenylacetaldeh Aromatic Attraction Scent Heliamphora spp. Sarraceniaceae Ericales Jaffé et al., 1995 yde aldehyde

Attraction Pulegone Monoterpene Floral scent Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017

VOC (emitted from trap but not Pyrazine, 2,6- Aromatic Nepenthales Kreuzwieser et al., Attraction individually Dionaea muscipula Droseraceae diethyl- heterocycle (Caryophyllales) 2014 confirmed to attract prey)

Scent Nepenthales Kreuzwieser et al., Attraction Sabinene Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

145 Jaffé et al., 1995; Monoterpene Found to be Heliamphora heterodoxa, H. Attraction Sarracenin Sarraceniaceae Ericales Newman et al., 2000; (iridoid) attractants of prey tatei, Sarracenia spp. Hotti et al., 2017

Fruit or floral Attraction Terpinolene Monoterpenoid Sarracenia leucophylla Sarraceniaceae Ericales Jürgens et al., 2009 scent

Floral or fruit scent (emitted from Dionaea Di Giusto et al., 2010; muscipula trap Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Attraction Tetradecane Alkane Kreuzwieser et al., but not rafflesiana Nepenthaceae (Caryophyllales) 2014 individually confirmed to attract prey) Floral scent (emitted from Dionaea Nepenthales Tetradecanoic muscipula trap Dionaea muscipula, Sarracenia Droseraceae, Kreuzwieser et al., Attraction Alkanoic acid (Caryophyllales) acid but not spp. Sarraceniaceae 2014; Hotti et al., 2017 , Ericales individually confirmed to attract prey)

146 Scent trans- Nepenthales Kreuzwieser et al., Attraction Sesquiterpene (downregulated Dionaea muscipula Droseraceae Geranylacetone (Caryophyllales) 2014 after prey capture)

Floral or fruit Attraction trans-Jasmone Sesquiterpenoid Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 scent

trans-Linalool Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide (furan) scent (Caryophyllales)

trans-Linalool Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide (pyran) scent (Caryophyllales)

147 trans-Ocimene Floral or fruit Nepenthales Attraction Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 oxide scent (Caryophyllales)

Scent trans- Nepenthales Kreuzwieser et al., Attraction Monoterpenoid (downregulated Dionaea muscipula Droseraceae Pinocarveol (Caryophyllales) 2014 after prey capture)

Scent Nepenthales Kreuzwieser et al., Attraction trans-Verbenone Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

Scent Nepenthales Kreuzwieser et al., Attraction Tricyclene Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

148 Floral scent Nepenthales Dionaea muscipula, Sarracenia Droseraceae, Kreuzwieser et al., Attraction Tridecane Alkane (downregulated (Caryophyllales) spp. Sarraceniaceae 2014; Hotti et al., 2017 after prey capture) , Ericales

VOC (emitted from trap but not Alkanoic Nepenthales Kreuzwieser et al., Attraction Undecanal individually Dionaea muscipula Droseraceae aldehyde (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction Undecane Alkane individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

Aromatic Floral or fruit Attraction Vanillin Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 aldehyde scent

149 Found on the spoon of the Xylene (o-, m-, Heliamphora heterodoxa, H. Attraction Alkyl Aromatic pitcher assumed Sarraceniaceae Ericales Jaffé et al., 1995 p-) tatei to be for attraction

Nepenthes rafflesiana, Nepenthales Scent (wood, Nepenthaceae, Jürgens et al., 2009; Di Attraction α-Copaene Sesquiterpenoid Sarracenia flava, S. leucophylla, (Caryophyllales) spice) Sarraceniaceae Giusto et al., 2010 S. minor , Ericales

Scent Nepenthales Kreuzwieser et al., Attraction α-Phellandrene Monoterpene (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

Scent (pine, turpentine, emitted from Jürgens et al., 2009; Di Dionaea muscipula, Nepenthes Droseraceae, Nepenthales Dionaea Giusto et al., 2010; Attraction α-Pinene Monoterpenoid rafflesiana, Sarracenia Nepenthaceae, (Caryophyllales) muscipula trap Kreuzwieser et al., purpurea Sarraceniaceae , Ericales but not 2014 individually tested)

150 Lemon scent (downregulated Nepenthales Jürgens et al., 2009; Dionaea muscipula, Sarracenia Droseraceae, Attraction α-Terpinene Monoterpenoid after Dionaea (Caryophyllales) Kreuzwieser et al., leucophylla Sarraceniaceae muscipula prey , Ericales 2014 capture)

Fruit or floral Sarracenia leucophylla, S. Attraction α-Ylangene Sesquiterpenoid Sarraceniaceae Ericales Jürgens et al., 2009 scent minor

Sarracenia flava, S. leucophylla, Attraction !-Bourbonene Sesquiterpenoid Scent (herb) Sarraceniaceae Ericales Jürgens et al., 2009 S. minor

!- Di Giusto et al., 2010; Floral or fruit Nepenthales Attraction Caryophyllene Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Kreuzwieser et al., scent (Caryophyllales) (Caryophyllene) 2014

151 Floral or fruit Nepenthales Attraction !-Elemene Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction !-Guaiene Sesquiterpene individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

VOC (emitted from trap but not Nepenthales Kreuzwieser et al., Attraction !-Linalool Monoterpenoid individually Dionaea muscipula Droseraceae (Caryophyllales) 2014 confirmed to attract prey)

Floral or fruit Nepenthales Attraction !-Pinene Monoterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

152 VOC Nepenthales Kreuzwieser et al., Attraction γ-Terpinene Monoterpenoid (downregulated Dionaea muscipula Droseraceae (Caryophyllales) 2014 after prey capture)

Floral or fruit Nepenthales Attraction δ-Cadinene Sesquiterpenoid Nepenthes rafflesiana Nepenthaceae Di Giusto et al., 2010 scent (Caryophyllales)

Apigenin Flavonoid Not tested Attraction (derivatives (aglycones - (present in sticky Roridula gorgonias Roridulaceae Ericales Wollenweber, 2007 or capture acacetin & 7,4'- sugar free resin) dimethyl ether) molecules)

Flavonoid Not tested Attraction (aglycones - Kaempferol (present in sticky Roridula dentata Roridulaceae Ericales Wollenweber, 2007 or capture sugar free resin) molecules)

153 Luteolin (and Flavonoid Not tested Attraction derivatives (aglycones - (present in sticky Roridula gorgonias Roridulaceae Ericales Wollenweber, 2007 or capture velutin and sugar free resin) apometzgerin) molecules)

Flavonoid Not tested Attraction (aglycones - Quercetin (present in sticky Roridula dentata Roridulaceae Ericales Wollenweber, 2007 or capture sugar free resin) molecules)

Flavonoid Tricetin Not tested Attraction (aglycones - (derivative of (present in sticky Roridula gorgonias Roridulaceae Ericales Wollenweber, 2007 or capture sugar free corymbosin) resin) molecules)

Attraction, Darlingtonia californica, Capture Nectar lubricant Jaffé et al., 1995; Hotti Erucamide Alkenoic amide Heliamphora spp., Sarracenia Sarraceniaceae Ericales and for Heliamphora et al., 2017 spp. Retention

154 Minor compound Capture 3α-lupeol Triterpenoid in sticky capture Roridula gorgonias Roridulaceae Ericales Simoneit et al., 2008 fluid

Darlingtonia californica, Alkanoic Insect paralysing Capture Coniine Heliamphora spp., Sarracenia Sarraceniaceae Ericales Hotti et al., 2017 secondary amine agent spp.

Major compound Dihydroxyolean Capture Triterpenoid in sticky capture Roridula dentata, R. gorgonias Roridulaceae Ericales Simoneit et al., 2008 -12-ene fluid

Major compound Dihydroxyurs- Capture Triterpenoid in sticky capture Roridula dentata, R. gorgonias Roridulaceae Ericales Simoneit et al., 2008 12-ene fluid

155 Minor compound Capture Germanicol Triterpenoid in sticky capture Roridula gorgonias Roridulaceae Ericales Simoneit et al., 2008 fluid

Major compound Capture Taraxeradiol Triterpenoid in sticky capture Roridula gorgonias Roridulaceae Ericales Simoneit et al., 2008 fluid

Toxic or Capture anaesthetic Nepenthales [also see Plumbagin Naphthoquinone influence on prey Nepenthes khasiana Nepenthaceae Raj et al., 2011 (Caryophyllales) digestion] on the rim of the pitcher

Trap seal and secretion of (Escalante-Pérez et al., Jasmonates enzymes in 2011; Nakamura et al., (Jasmonic acid, Dionaea 2013; Libiaková et al., 12-Oxo- Capture/ muscipula. Leaf 2014; Mithöfer et al., Phytodienoic Dionaea muscipula, Drosera Droseraceae, Nepenthales digestion/ Oxylipins bending in 2014; Yilamujiang et Acidic & capensis, Nepenthes alata Nepenthaceae (Caryophyllales) retention Drosera spp. after al., 2016; Krausko et Isoleucine prey capture. al., 2017; Pavlovič et conjugate of Digestive enzyme al., 2017; Pavlovič & jasmonic acid) secretion in Mithöfer, 2019) Drosera spp.

156 Dionaea muscipula and Nepenthes alata

Capture Forms wax Alkanoic Nepenthales and Triacontanal crystals slippery Nepenthes spp. Nepenthaceae Riedel et al., 2003 aldehyde (Caryophyllales) retention to prey

Trap closure Capture, sensitivity in Nepenthales Escalante-Pérez et al., retention & Abscisic acid Isoprenoid Dionaea muscipula, Droseraceae Dionaea (Caryophyllales) 2011 digestion muscipula

5-O- Eilenberg et al., 2010; methyldroserone Nepenthales Digestion Naphthoquinone Anti-fungal Nepenthes khasiana Nepenthaceae Raj et al., 2011; Buch (plumbagin (Caryophyllales) et al., 2013 derivative)

157 Dionaea muscipula, Drosera aliciae, D. burkeana, D. 7-methyl- Zenk et al., 1969; capensis, D. cistiflora, D. Droseraceae, juglone Culham & Gornall, Antimicrobial/ cuneifolia, D. hamiltonii, D. Nepenthaceae, Nepenthales Digestion (Ramentaceone) Naphthoquinone 1994; Krolicka et al., antibacterial madagascariensis, D. Drosophyllace (Caryophyllales) (Plumbagin 2008, 2009; Buch et spathulata, D. tracii, D. a isomer) al., 2013 trinervia, Drosophyllum lusticanum, Nepenthes khasiana

Not individually Digestion * tested but Dionaea muscipula, Drosera Droseraceae, Nepenthales also see Benzoic acid Aromatic acid suggested Kováčik et al., 2012a capensis, Nepenthes anamensis Nepenthaceae (Caryophyllales) attraction antioxidant during digestion

Cannon et al., 1980; Droserone Nepenthes khasiana, Nepenthes Nepenthaceae, Bringmann et al., Anti-fungal/ Nepenthales Digestion (plumbagin Naphthoquinone rafflesiana, Triphyophyllum Dioncophyllac 2000; Eilenberg et al., microbial (Caryophyllales) derivative) peltatum eae 2010; Raj et al., 2011; Buch et al., 2013

Dionaea muscipula, Drosera Droseraceae, Nepenthales Digestion Gallic acid Phenolic acid Anti-fungal Kováčik et al., 2012a,b capensis, Nepenthes anamensis Nepenthaceae (Caryophyllales)

158 Hydroxydrosero Nepenthales Digestion ne (plumbagin Naphthoquinone Anti-microbial Nepenthes rafflesiana Nepenthaceae Cannon et al., 1980 (Caryophyllales) derivative)

Anti-bacterial & Nepenthales Digestion Hyperoside Flavonoid Drosera rotundifolia Droseraceae Egan & Kooy, 2012 antioxidant (Caryophyllales)

Drosera adelae, D. aliciae, D. Not specifically Nepenthales Digestion Isorhamnetin Flavonoid capensis, D. cuneifolia, D. Droseraceae Marczak et al., 2005 tested (Caryophyllales) ramentacea

Nepenthaceae, Bringmann et al., Nepenthes gracilis, Nepenthales Digestion Isoshinanolone Naphthoquinone Anti-microbial Dioncophyllac 2000; Aung et al., Triphyophyllum peltatum (Caryophyllales) eae 2002

159 Dionaea muscipula; Drosera Nepenthales Marczak et al., 2005; Digestion Myricetin Flavonoid Antibacterial adelae, D. aliciae, D. capensis, Droseraceae (Caryophyllales) Krolicka et al., 2008 D. cuneifolia, D. ramentacea

Dioncophyllac Nepenthes insignis, Nepenthales Digestion PlumbasideA Naphthoquinone Antimicrobial eae, Rischer et al. 2002 Triphyophyllum peltatum (Caryophyllales) Nepenthaceae

Protocatechuic Not specifically Dionaea muscipula, Drosera Droseraceae, Nepenthales Digestion Phenolic acid Kováčik et al., 2012a acid tested capensis, Nepenthes anamensis Nepenthaceae (Caryophyllales)

Not specifically Nepenthales Digestion Rossoliside Naphthoquinone Nepenthes insignis Nepenthaceae Rischer et al., 2002 tested (Caryophyllales)

160 Not specifically Nepenthales Digestion Shinanolone Naphthoquinone Nepenthes gracilis Nepenthaceae Aung et al., 2002 tested (Caryophyllales)

Aldrovanda vesiculosa, Dionaea Zenk et al., 1969; muscipula, Drosera adelae, D. Cannon et al., 1980; aliciae, D. auriculata, D. binata, Culham & Gornall, Droseraceae, D. capensis, D. cistiflora, D. 1994; Bringmann et Drosophyllace Digestion dichotomata, D. indica, D. al., 2000; Rischer et Antibacterial, ae, Nepenthales [also see Plumbagin Naphthoquinone longifolia, D. lunata, D. al., 2002; Aung et al., insecticide Nepenthaceae, (Caryophyllales) capture] ramentacea, D. whitakeri, 2002; Marczak et al., Dioncophyllac Drosophyllum lusitanicum, 2005; Krolicka et al., eae Nepenthes gracilis, N. 2008; Buch et al., rafflesiana, Triphyophyllum 2013; Gonçalves et al., peltatum 2015

161 Chapter 5 Metabolomic analysis reveals a reliance on secondary plant metabolites to facilitate carnivory in the Cape sundew, Drosera capensis

162 5.1 Introduction

Carnivorous plants are a diverse, polyphyletic group of flowering plants adapted specifically to nutrient poor environments (Darwin, 1875; Albert et al., 1992; Ellison

& Gotelli, 2001). They are increasingly being considered as model systems for addressing a diverse range of biological questions (Adlassnig et al., 2005b; Schöner et al., 2017; Lan et al., 2017; Gergely et al., 2018). The morphological basis for carnivory is well known and details of different leaf modifications to capture prey are well documented (Knight & Frost, 1991; Gibson & Waller, 2009; Gaume et al., 2016;

Hedrich & Neher, 2018). But, there are multiple systems involved in carnivory, i.e. the need to attract prey, respond to prey capture, instigate leaf movement, digest prey, assimilate nutrients and utilise nutrients derived from prey (Ellison & Adamec,

2018). Recent advances in molecular and genetic technologies are highlighting a molecular basis in the function of carnivory in plants (Leushkin et al., 2013;

Libiaková et al., 2014; Fukushima et al., 2017). Recent studies have shown a large diversity of digestive enzymes synthesised by carnivorous plants to break down prey including chitinase, nuclease, protease and lipase (See Ravee et al., 2018). There remain aspects of plant carnivory at the molecular level, however, of which we know little.

Secondary metabolites are integral for many aspects of plant fitness: regulating growth, attracting pollinators (Hartmann, 2007; Kessler & Baldwin, 2007), facilitating interactions between conspecifics, competitors and symbionts, and as defence against abiotic and biotic stress (Agrawal, 2011). To date, however, relatively few examples have been found of secondary metabolites involved in the carnivorous habit (Chapter 4). The majority are associated with prey attraction and are classified as volatile organic compounds (VOCs) (Kováčik et al., 2012a; Nakamura et al., 2013;

Kreuzwieser et al., 2014; Krausko et al., 2017; Hotti et al., 2017), but are also important in leaf bending (Mithöfer et al., 2014), production of digestive enzymes (Nakamura et al., 2013; Yilamujiang et al., 2016) and protection of leaf tissue from decay as prey are

163 digested (Raj et al., 2011). Given the apparent importance of secondary metabolites in evolutionary adaptation of plants to the environment and recent findings of a metabolic basis in carnivory, I hypothesise that there is a substantial biochemical role, and plant regulatory response to prey capture in other aspects of carnivory i.e. retention, digestion and assimilation.

To determine the extent of the biochemical role in carnivory, I measured the metabolic profile of the carnivorous herb Cape sundew, Drosera capensis, before and after the addition of insect substrate to the traps. Up- or down-regulation of compounds following substrate addition indicates the involvement of that compound in carnivory (as per Kreuzwieser et al., 2014). I predict that there is a whole leaf metabolic response to substrate addition incorporating multiple secondary metabolites to facilitate or enhance carnivory. Specifically, I predict that

(i) compounds involved in leaf bending i.e. jasmonates, will increase in concentration within 45 minutes and will be sustained for the remainder of the experiment; (ii) compounds involved in attraction will decrease in concentration following nutrient addition as the benefit of these compounds is reduced, (iii) there will be a latency of certain compounds before upregulation which are suggested to be compounds involved in the digestion or assimilation of nutrients from prey.

5.2 Methods

5.2.1 Plant rearing and preparation

Drosera capensis plants were purchased from an established nursery. Prior to this, seeds were grown in an open polytunnel on a standard medium of peat:sand:perlite

(6:1:1 respectively) in individual pots. Plants were able to catch available prey during growth to maturity. On the 1st July 2016, four weeks prior to the experiment,

500 mature individuals were translocated to a greenhouse at Loughborough

University. Exposure to prey and prey items on leaves were restricted as much as possible, the few prey that were captured were removed daily. The greenhouse was

164 maintained at ambient temperature, with supplementary heating to prevent the temperature dropping below 18 °C, additional fluorescent lighting on a light/dark cycle of 12:12 h was used and plants were watered regularly with de-ionised water when required. To prevent allocation to carnivory being confounded by reproductive investment, flower stems were removed when visible during this time up to the final week before the experiment, when removal of stems may have mitigated a stress response. Plants were selected at random from the 500 individuals in the greenhouse for use in this experiment.

5.2.2 Experimental design

The metabolic response of Drosera capensis to prey capture was determined by adding insect substrate to one leaf trap of an individual plant; the leaf (to which substrate was added) was subsequently harvested for analysis (protocol below). Leaf response to prey (e.g. leaf bending instigated by accumulation of jasmonates) is known to occur within the first 6 hours of prey capture but digestion, assimilation and induced root activity following prey nutrient uptake can take place over multiple days or weeks to completion (Adamec, 2002; Nakamura et al., 2013;

Krausko et al., 2017). To identify these short- and longer-term plant responses, I created a time-series of harvests from prey addition. Plants were harvested at 0.75,

1.5, 3, 6, 24, 48 and 72 hrs after prey addition. Circadian variability in the metabolome of plants is common (Kim et al., 2017). I therefore staggered prey addition throughout the day (prey additions from 5.15 am to 12:50 pm) to keep harvest times in the shortest possible time window. Plant harvests were all carried out between 10:30-13:30. Each time interval had six replicates. These replicates were spread evenly throughout the harvest period and at least one unfed control was harvested within 10 minutes of a fed plant being harvested to account for natural circadian changes and comparison of the metabolite profile to unfed plants (total number of controls = 32). There was a total of 74 plant samples. Due to the amount of work involved, the experiment was carried out in two sections. The first started on

165 day 1 (2nd August 2016) and finished on day 4, the second started on day 2 and finished on day 5 (6th August 2016).

5.2.3 Prey addition

To simulate an accurate and standardised response of Drosera capensis to prey capture, fruit fly powder (Drosophila melanogaster) was used as the prey addition

(fed) treatment using the protocol outlined by Gao et al. (2015). Fruit flies were reared on a standard medium, freeze dried and ground into a homogenised powder using a ball mill. Prey addition consisted of 25 mg fruit fly powder in 100 µl of de- ionised water spread evenly and carefully across the trap using a pipette.

5.2.4 Plant harvests

At each harvest, leaf and root tissue was cut from the plant and washed immediately in de-ionised water to remove prey residue and soil. The leaf was then flash-frozen in liquid nitrogen (−196°C). The time from cutting to freezing was always < 30 seconds. Tissue samples were then weighed before being placed in a Precellys tube in a −80°C freezer until analysis.

5.2.5 Sample preparation for metabolite profile analysis

For a simplified workflow from sample analysis through to statistical analysis see

Figure 5.1. Monophasic extractions used CK-14 Precellys tubes and acetonitrile:methanol:water 2:2:1 as the final solvent. Differences in weight of tissue was accounted for by using a fixed amount of the original homogenate, set by the smallest sample (14.9 mg). After centrifugation, equal volumes of the supernatant were transferred into new glass vials and dried in a SpeedVac before storage at -

80°C. LC-MS samples were taken up in 50 µl, 20% aqueous methanol each, and 10 µl each were pooled into one Quality Control QC sample per sample type, 16 QC samples in total.

166 5.2.6 UHPLC-MS analysis (positive ion mode)

After centrifugation, (15000 rpm for 10 min at 4°C, Biofuge), 20 µl per sample were pipetted into a 96-well plate, starting with a blank, QC samples, samples with intermittent further QC samples, and ending with a blank after two QC samples. The samples were run in controlled randomised order, with QC samples equidistant between them. They were analysed by Ultra High Precision Mass Spectrometry

(UHPLC-MS) on a Thermo Scientific Q Exactive mass spectrometer interfaced with a

Thermo Dionex Ultimate 3000 RS system, equipped with a Thermo Hypersil Gold column (100 x 2.1 mm, 1.9 µm particles, C18 material). Solvents: (A) 0.1% formic acid in water and (B) 0.1% formic acid in methanol. Method: 14 min LC from (A) to (B),

400 µl / min, with 18 µl injections per sample and MS start at 0.1 min, with the flow up to 0.45 min directed towards waste. Quality control sample (QC) 02 contains data-dependent tandem mass spectrometry (MS/MS) data and is used with the retention time index to annotate metabolites (details below) and was run at 140k resolution. Data were collected in positive ion and profile mode, m/z 100-1000, at 70k resolution (See supplementary material for total ion chromatogram for a blank and

QC run).

5.2.7 Data processing

Data processing and quality control methods followed guidelines outlined by Guida et al. (2016) and Kirwan et al. (2014) using NBAF-B in house scripts in MATLAB (v8.1;

The MathWorks, Natick, MA, USA), the SIMStitch pipeline. Briefly, Thermo.raw data files in profile mode were converted into mzML files in centroid mode using

MSCONVERT (Proteowizard 3.0.7665). An R (3.2.0) based XCMS / CAMERA script was used for alignment and resulted in an intensity matrix in a csv file (9309 features). The matrix was imported into MatLab and inserted into a SIMSTITCH pipeline. During replicate filtering, peaks were retained if they were found in two out of three technical replicates. A blank filter of > 2× sample over blank signal was applied and a sample filter of peak presence of at least 75% of all samples. Samples

167 were filtered with a 2-ppm mass error and a 75% filter. Data were normalised using probabilistic quotient normalisation (PQN) and missing values being filled using the k-nearest neighbours (KNN) algorithm (k=5). This matrix was used for univariate statistics including fold-changes. Generalised logarithm (g-log) transformation was optimised on QC samples only and then applied to all samples. This matrix was used for multivariate statistics (over Pareto scaling as there is less bias towards more intense signals). Normalisation and g-log transformation are applied to make the samples more comparable and to reduce heteroscedasticity.

The XCMS matrix and therefore the dataset object for the samples contained 5423 peaks, the blank filtered dataset object had 3259 peaks, and the sample filtered dataset object had 3257 peaks. The median relative standard deviation (RSD) for the

QC samples in the normalised dataset was 7.48 % (without QC01-QC04). This confirms the very high technical quality of this dataset (for biomarker studies, the

FDA guidelines specify an RSD of < 20% as an acceptable level of precision and presenting as a median RSD is the best practice, (Parsons et al., 2009; Kirwan et al.,

2014)).

5.2.8 Statistical analysis and identification of metabolites

The analysis utilises separate supervised and unsupervised methods to extract the most important compounds likely associated with carnivory. Furthermore, univariate analyses investigate metabolites independently and do not consider interactions or correlations with other metabolites. The multivariate analysis in this experiment investigates metabolites and patterns among metabolites and therefore a dynamic approach to analysis reveals different information about the data (Grace &

Hudson, 2016). These data are then assessed in a cross-comparative approach to determine the extent of the role of secondary plant metabolites in the carnivorous plant response to feeding.

168 5.2.9 Multivariate analysis

Principal component analysis (PCA) was conducted initially to assess the overall metabolic differences among the control and time-after-feeding treatments in an unbiased manner using the prcomp function in R which uses singular value decomposition to calculate eigenvalues. Principal component analysis confirms unbiased separation of metabolic profiles across the treatment groups to determine the appropriateness of subsequent analysis. Following this, there was obvious separation of the treatments into three groups: Controls, ≤ 6 hrs and ≥ 24hrs. Further analyses were, consequently, conducted for these three pooled groups. PCA of the samples highlighted a clear single anomalous control sample and was removed.

Supervised multivariate analyses were subsequently performed using orthogonal partial least-squares discriminant analysis (OPLS-DA) using SIMCA-P+ (v 14,

Umetrics Umeå, Sweden) to find the direction of maximum covariance between controls and grouped treatments ≤ 6 hrs or ≥ 24 hrs or between grouped treatments

(≤ 6 hrs and ≥ 24 hrs). Partial least squares discriminant analysis (OPLS-DA) was also used to separate differences across all three treatment clusters. OPLS-DA separates better the difference between groups of observations by rotating PCA components such that maximum separation among groups is obtained and identifies variables that are most contributing to this separation. SIMCA-P+ validates the OPLS-DA model by removing 1/7th of data and producing a model for the 6/7th of remaining data. The new data are then predicted using the new model, continuing in this process until all the data have been predicted. The contribution of each metabolite to the separation of treatments is evaluated using S-plots and deriving metabolites of biological relevance indicated from these plots. S-plots are analysed for metabolites that have the greatest selectivity and sensitivity in discriminating between the data, presented as points at the upper right and lower left corners of the plot isolated within red ovals.

169 Univariate statistical analyses were used to confirm the statistical significance of changes in individual mass-spectral signals between unfed (controls) and grouped fed treatments ≤ 6 hrs or ≥ 24 hrs. Specifically, analysis of variance (ANOVA) was conducted with a false discovery rate (FDR) of 5% to determine if any compounds across all fed treatments changed significantly from the unfed controls. These identified compounds were compared to previously identified metabolites in other studies highlighted in Chapter 4. Then multiple Welch’s t-tests were conducted between controls and the ≤ 6hrs or the ≥ 24 hrs treatment groups. Welch’s t-tests were conducted with a false discovery rate (FDR) of 5% to correct for multiple hypothesis testing of all treatment peak intensities compared to control to identify any changes in compound intensities across all treatments (Benjamini & Hochberg,

1995). These were then combined with log-fold changes to identify the significant and intense signal changes between control samples and the ≤ 6 hrs or ≥ 24 hrs treatments presented as volcano plots. Conservative thresholds of ± 2 for the log2 fold change and 2 for the -log FDR adjusted p value (p < 0.01) to highlight those features that showed the largest differences (Grace & Hudson, 2016). Of the significant compounds (high statistical significance and fold change), the 100 most significant compounds are presented in a heat map to illustrate whole plant metabolome response.

Metabolites that are highlighted by both OPLS-DA and univariate analysis were presented in box plots to examine individual metabolite patterns.

5.2.10 Annotation summary and pathway analysis

The in-house MIPACK software matched 419 signals of the total of 3257 signals to the

BIOCYC/A. thaliana database (5 ppm m/z error), and 1290 signals up to m/z 600 to the KEGG database with 2 ppm m/z error and including molecular formula search

(as per Weber & Viant, 2010). Choline was annotated manually. Compounds annotated and found to be biologically important from the analyses were inserted into the METABOANALYST pathway generator. METABOANALYST was implemented

170 using PRIMEFACES library (v6.1) based on the JavaServer Faces Technology. The communication between Java and R is established through TCP/IP using the RSERVE program (Li et al., 2018). Pathway library selected was ‘Arabidopsis thaliana’. Over representation analysis method was ‘Hypergeometric test’. Node importance measure for topological analysis is ‘relative betweenness centrality’.

5.2.11 Metabolite comparisons to other carnivorous plants

Compounds were compiled to produce a library of metabolites that have previously been associated with a role in carnivory (See Chapter 4). These were cross-checked for matches with compounds identified from the present study. Exact compounds or isomers were identified and their regulation before and after prey capture and the statistical significance of these changes in the experiment were produced in a table and compared to previously annotated compounds. Metabolites derived from carnivorous plants purely for pharmaceutical use are not discussed in this paper as their derivation is stimulated under conditions not analogous to the plant’s ecology and therefore do not assist in the identification of compounds likely to function in plant carnivory.

171 Data acquisition

• Sample preparation

• UHPLC-MS

Raw file data processing • File conversion

• XCMS to produce a .csv file

Data pre-processing • Filtering

• Missing values (knn algorithm n = 5)

Data pre-treatment

• Normalisation (PQN) (Used for univariate)

• Transformation and scaling (glog as less bias towards intense signals over Pareto)

• Centring (Used for multivariate)

• Univariate analyses metabolomic features independently and does not take interactions and correlations between features into account. Multivariate does and therefore both types of analysis reveal different information about the data.

Univariate Multivariate • Fold changes • PCA to identify patterns in data, • T-test to identify significant remove outliers, identify group metabolites between groups separation (with and without QC for

• Identify compounds of high quality assurance and quality checks) significance and/or high fold change • OPLS-DA forces group separation, results require further investigation

Compound identification and cross validation of analyses • Annotation of metabolites

• Collating important metabolites from univariate and multivariate analysis

Figure 5.1: Data workflow for metabolomics analysis used in this experiment (Di Guida et al., 2016; Grace & Hudson, 2016; Considine et al., 2018).

172 5.3 Results

In response to simulated prey capture, Drosera capensis upregulate and downregulate a large number of secondary plant metabolites over the course of 72 hours. Of the

3257 peaks present, statistically significant changes in intensity were found in 2383 peaks in at least one treatment compared to controls. Over 1700 compounds were annotated putatively from the 3257 peaks identified in the leaf tissue, of which notable compounds were collated. Of the 2383 compounds where change was detected, multiple patterns were identified. Generally, compounds that were upregulated increased in concentration within 0.75 hrs following substrate addition.

After this point, in general, concentrations stopped increasing and the elevated concentration was maintained for the duration of the experiment. In some cases, metabolites rapidly increase at 0.75 hrs, then slowly increase for the duration of the experiment, other metabolites had a more consistent rate of increase over time which lasted the duration of the experiment. For a small proportion of compounds, increases in concentration were very large. Over twenty compounds have tenfold increases and are significant compared to controls, and 12 of these have over 30-fold increases compared to controls. Decreases in concentration of downregulated compounds were mostly small in the first 6 hours, followed by much larger decreases in concentration for the remaining 66 hours. In very few cases, there are compounds that decrease in concentration initially, and then have an overall increase in concentration by the end of the experiment.

5.3.1 Multivariate analyses

Principal component analysis (PCA) of the metabolome of plant samples show a clear separation between time points along PC1, increasing in time from left to right.

PCA axis 1 and 2 (44.4% and 12.7% respectively) explain over 50% of the variance.

Control groups are clustered at the left side, shorter times (0, 75, 1.5, 3, 6) cluster in the middle of the ordination, and longer treatment times (24, 48, 72) cluster at the right side of the ordination (Figure 5.2). Further analysis, therefore, focused on

173 compounds within these three clustered time range groups. These results show it is appropriate to perform supervised multivariate analysis.

OPLS-DA was performed with these fed treatment time groups (≤ 6 hrs and ≥ 24 hrs) compared to the unfed controls. S-plots of metabolites between controls and ≤ 6 hrs highlighted 28 upregulated and 8 downregulated compounds (Figure 5.3, Figure

5.4). S-plots of metabolites between controls and ≥ 24 hrs highlighted 10 upregulated and four downregulated compounds (Figure 5.5, Figure 5.6). OPLS-DA comparing the ≤ 6 hrs to ≥ 24 hrs highlighted 13 upregulated and three downregulated

Treatment c 0.75 0 1.5 3 6 24

PC 2 (12.7%) 2 PC 48 72

-20

-20 0 20 40 PC 1 (44.4%) Figure 5.2: Ordination of the first two principle components from a PCA of metabolite profiles of Drosera capensis following insect substrate addition, with unfed controls. Treatments indicated are hours after substrate addition (c = unfed controls). Presented are the PC1 and PC2 scores for each plant. Different symbols represent each time point. The ellipsoids are normal contour lines with probability 68% for each group. Presented PCA is without inclusion of the anomalous result (one control sample).

174 compounds in the ≥ 24 hrs compared to ≤ 6 hrs (Figure 5.7, Figure 5.8). A final OPLS-

DA comparing all three groups (unfed, ≤ 6hrs after feeding and ≥ 24 hrs after feeding) highlighted four compounds that are highly associated specifically to the ≥

24 hrs after feeding group (Figure 5.9, Figure 5.10 annotated metabolites from these analyses are compiled in Table 5.1).

175

Figure 5.3: Orthogonal partial least squares-discriminant analysis model 2-dimensional score plot for unfed plants (black circles) and plants up to 6 hrs after feeding (green triangles). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data.

Figure 5.4: S-plot for modelled variables using an orthogonal partial least squares- discriminant analysis model with unit variance scaling for unfed versus ≤ 6 hrs after feeding samples. Important compounds are highlighted with red ellipses.

176

Figure 5.5: Orthogonal partial least squares-discriminant analysis model 2-dimensional score plot for unfed plants (black circles) and plants ≥ 24 hrs after feeding (blue squares). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data.

177

Figure 5.7: Orthogonal partial least squares-discriminant analysis model 2- dimensional score plot for plants ≤ 6 hrs after feeding (green triangles) and plants ≥ 24 hrs after feeding (blue squares). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data.

178

Figure 5.9: Orthogonal partial least squares-discriminant analysis model 2-dimensional score plot for unfed plants (black circles), plants up to 6 hrs after feeding (green triangles) and plants 24 hrs or later after feeding (blue squares). Hotelling’s t2 as a quality check to see all the data are within normal boundaries. Ellipsoid is 95% confidence interval for the data.

179 5.3.2 Volcano plot

In the ≤ 6 hrs group compared to unfed controls, 62 compounds were upregulated and five downregulated with a statistically significant fold-change greater than four.

In the ≥ 24 hrs group compared to unfed controls, 287 compounds were upregulated and 14 were downregulated with a statistically significant fold-change greater than

4. All of the 62 upregulated compounds in the ≤ 6 hrs group were also identified present among the 287 upregulated compounds in the ≥ 24 hrs group. None of the six downregulated compounds in the < 6 hrs group were present in the downregulated compounds in the ≥ 24 hrs group (Figure 5.11, Table 5.1).

These compounds were investigated further to illustrate the general metabolic change in D. capensis following prey capture. A heatmap with a compound dendrogram shows a clear treatment grouping of controls, followed by the treatments up to 6 hours, and then the treatment times over 24 hours (Figure 5.12).

There are four clear separations of metabolite patterns. There are compounds that are downregulated 24 hrs after prey capture, compounds that have a gradual change in concentration over the course of the experiment, and some compounds appear to have a high increase in concentration initially, followed by a steady increase in concentration for the remainder of the experiment. The results of the volcano plot also confirmed the significance of compounds highlighted in OPLS-DA, a sample of which are included in Figure 5.13 A-H and Figure 5.14 I, M-P.

180 ≤ 6 hrs ≥≤ 246 hrs hrs ≥ 24 hrs

20 20

15 15

10 10 Significance (-logP) Significance Significance (-logP) Significance

5 5

0 0

-2.5 0.0 2.5 5.0 7.5 -2.5-2.5 0.0 2.5 5.0 7.5 -2.5 0.0 2.5 5.0 7.5 Log2 Fold Change Log2 Fold Change Figure 5.11: Number and proportion of discriminatory features in treatment groups compared to the control group. Shown here are log fold changes (x-axis), in comparison to unfed control plants, and the y-axis displays the negative log of the p-value from a two-sample t-test. Data points that are far from the origin (near the top or far left and right) are considered important variables with potentially high biological relevance. Blue and red features (down-regulated and up-regulated respectively) are those chosen based on the presented criteria of thresholds of 2 and -2 for the log2 fold change and 2 for the -log FDR adjusted p value (p < 0.01, indicated with red margins) to highlight those features that showed the largest differences (Grace & Hudson, 2016). 181 Color Key Heatmap of significant features from PCA loadings

-4 -2 0 2 4 Row Z-Score Color Key c 0.75 1.5 3 6 24 48 72 Heatmap of significant features from PCA loadings

306 303 -4 -2 0 2 4 302 304 Row Z-Score 278 305 294 290 287 299 238 306 291 303 302 231 304 278 261 305 294 232 290 287 242 299 260 238 291 267 231 261 220 232 242 259 260 267 218 220 248 259 218 239 248 239 251 251 229 229 296 288 296 212 288 292 253 212 284 298 292 214 264 253 282 284 236 243 298 230 283 214 216 217 264 221 271 282 208 236 211 241 243 213 222 230 252 240 283 255 263 216 224 217 295 237 221 234 228 271 249 225 208 227 226 211 254 241 250 233 213 257 275 222 277 210 252 247 273 240 270 255 276 280 263 258 281 224 293 c 223 295 279 235 237 268 0.75 234 301 269 228 286 289 1.5 249 265 209 225 300 227 246 285 3 226 307 297 254 219 245 6 250 244 266 233 215 257 256 24 262 275 274 277 272 48 210 247 c.0 c.1 c.2 c.4 c.5 c.6 c.7 c.8 c.9 273

72 c.10 c.11 c.12 c.13 c.14 c.15 c.16 c.17 c.18 c.19 c.20 c.21 c.22 c.23 c.24 c.25 c.26 c.27 c.28 c.29 c.30 c.31 X3.0 X3.1 X3.2 X3.3 X3.4 X3.5 X6.0 X6.1 X6.2 X6.3 X6.4 X6.5 270 X24.0 X24.1 X24.2 X24.3 X24.4 X24.5 X48.0 X48.1 X48.2 X48.3 X48.4 X48.5 X72.0 X72.1 X72.2 X72.3 X72.4 X72.5 X1.5.0 X1.5.1 X1.5.2 X1.5.3 X1.5.4 X1.5.5 276 X0.75.0 X0.75.1 X0.75.2 X0.75.3 X0.75.4 X0.75.5 280 258 281 293 223 279 235 268 301 269 286 289 265 209 300 246 285 307 297 219 245 244 266 215 256 262 274 272 Figure 5.12: Metabolic patterns over time after prey substrate addition. Presented are the 100 most

important features, based on statisticalc.0 c.1 c.2 c.4 c.5 c.6 c.7 c.8 c.9 significance and fold change. Colour and colour intensity c.10 c.11 c.12 c.13 c.14 c.15 c.16 c.17 c.18 c.19 c.20 c.21 c.22 c.23 c.24 c.25 c.26 c.27 c.28 c.29 c.30 c.31 X3.0 X3.1 X3.2 X3.3 X3.4 X3.5 X6.0 X6.1 X6.2 X6.3 X6.4 X6.5 X24.0 X24.1 X24.2 X24.3 X24.4 X24.5 X48.0 X48.1 X48.2 X48.3 X48.4 X48.5 X72.0 X72.1 X72.2 X72.3 X72.4 X72.5 represent Z score for the change in peak intensity from the mean of theX1.5.0 X1.5.1 X1.5.2 compoundX1.5.3 X1.5.4 X1.5.5 (one compound per X0.75.0 X0.75.1 X0.75.2 X0.75.3 X0.75.4 X0.75.5 row). Each column represents a single plant, these are grouped according to treatment and time point (randomly assigned within these). Each row represents a single metabolite grouped as a dendrogram by similarity in change of intensity.

182 280.1390431 294.154507 isoleucine tyramine Group 4 A B C D c 0.75 1.5 2 3 6 24 48 72 0

-2

phenyacetaldehyde 133.1048579 231.1702626 401.2152413 4 E F G H Normalised intensity Normalised

-2

c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 Time after substrate addition

-2.5

427.231257 527.2315188 258.13362116 563.2684374 M N O P

Normalised intensity Normalised 2.5

0.0

-2.5

c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 c 0.751.5 3 6 24 48 72 Time after substrate addition Figure 5.14: (Figure 5.13 cont.). Normalised intensities of metabolites highlighted as important from volcano plot or OPLS-DA. I: Compound highlighted as biologically important by all analyses (Volcano plot group <6 hrs and ≥ 24 hrs and from OPLS-DA of treatment groups compared to controls). C, E, J, K, L: Compounds previously known to be involved in carnivory and confirmed to be significant by volcano plot in this experiment. M, N, O, P: Compounds identified by three-way OPLS-DA (Figure 5.9, Figure 5.10) as highly associated with a separation of ≥ 24hrs from unfed controls and plants ≤ 6 hrs after feeding.

184 Table 5.1: Cross comparison of univariate and OPLS-DA analysis. T-tests confirm grouping from OPLS-DA. Only annotated compounds included in this table. This table was used for metabolic pathway analysis. 24 to C 6 to C OPLS-DA OPLS-DA OPLS-DA OPLS-DA Putative annotation Regulation t-test t-test 6 TO C 24 TO C 24 to 6 24 to 6 & C Isoleucine ✓ ✓ ✓ ✓ ↑ Phenylacetaldehyde ✓ ✓ ✓ ✓ ↑ Tyramine ✓ ✓ ✓ ✓ ↑ 1-(3-aminopropyl)-pyrrolinium ✓ ✓ ✓ ↑ 8-Oxodeoxycoformycin ✓ ✓ ✓ ↓ Carnitine isomer ✓ ✓ ✓ ↑ Choline sulphate ✓ ✓ ✓ ↑ D-Methionine ✓ ✓ ✓ ↑ Hypoxanthine ✓ ✓ ✓ ↑ Idanpramine ✓ ✓ ✓ ↑ Methionine oxide ✓ ✓ ✓ ↑ Napthylamine isomer ✓ ✓ ✓ ↑ SR 12813 ✓ ✓ ✓ ↑ Tryptophan ✓ ✓ ✓ ↑ Xanthine ✓ ✓ ✓ ↑ 1-(3-Aminopropyl)-4- ✓ ✓ ↑ aminobutanal 11-Aminoundecanoic acid ✓ ✓ ↓ Butabarbital ✓ ✓ ↑ Dethiobiotin ✓ ✓ ↑ Glycyl-leucine ✓ ✓ ↑ Guanosine isomer ✓ ✓ ↑ Ipratropium ✓ ✓ ↑ Leucyl-leucine ✓ ✓ ↑ Linmarin ✓ ✓ ↑ Rhizocticin isomer ✓ ✓ ↓ Tryptophan isomer ✓ ✓ ↑ Tyrosine ✓ ✓ ↑ (2S,3S)-2-Hydroxytridecane-1,2,3- ✓ ↑ tricarboxylate 11-Aminoundecanoic acid ✓ ↑ 2-Phenylacetamide ✓ ↑ 3-Oxododecanoic acid ✓ ↑ 5-Methylthioadenosine ✓ ↑ Acetylagmatine ✓ ↑ Adenine ✓ ↑ Benzoylagmatine ✓ ↑ Buchananine ✓ ↑ Caryophyllene oxide (epoxide) ✓ ↑

185 cis-2-Carboxycyclohexyl-acetic ✓ ↑ acid CPX ✓ ↑ Crotono-betaine ✓ ↑ Cycloate ✓ ↑ D-lyposine ✓ ↑ Danorubicin ✓ ↓ Dibutyl adipate ✓ ↑ Eudesmol / nerolidol ✓ ↑ Fagomine ✓ ↑ Fortimicin isomer ✓ ↑ Gibberellin A41 ✓ ↑ Guanine ✓ ↑ HC-toxin ✓ ↑ Hexahomomethionine ✓ ↑ Isoleucine isomer ✓ ↑ Kynurenine ✓ ↑ Leu-Gly-Pro ✓ ↑ Leucine ✓ ↑ Lycodine ✓ ↑ Muramic acid ✓ ↑ N1-Acetylspermine ✓ ↑ Naphthoquinone ✓ ↓ Nicotinate ✓ ↑ Nifuradene ✓ ↓ Octanal ✓ ↑ Octhilinone ✓ ↑ Pantetheine ✓ ↑ Phenazocine ✓ ↑ Phenylethylamine ✓ ↑ Pheophorbide a ✓ ↑ Pirbuterol ✓ ↑ Proacacipetalin ✓ ↑ Quinoline isomer ✓ ↑ Rhododendrin ✓ ↑ Schizonepetoside E ✓ ↑ Spironolactone ✓ ↑ Styrene ✓ ↑ Succinic anhydride ✓ ↑ UCL 1608 ✓ ↑ Xanthopterin-B2 ✓ ↑

186 5.3.3 Pathway analysis of biologically important metabolites

The compounds involved in carnivory are derived from multiple metabolic

pathways (Table 5.2). Four pathways were found to be significantly involved in

Drosera capensis response to prey addition: flavone and flavonol biosynthesis,

phenylalanine metabolism, isoquinoline alkaloid biosynthesis and flavonoid

biosynthesis. Sixteen other metabolic pathways were associated with compounds

found to significantly change in response to simulated prey capture in this

experiment, but these pathways were not statistically significant. They are included

here for future work.

Table 5.2: Results of metabolic pathway analysis of Drosera capensis after prey substrate addition. Presented are the number of metabolites in certain metabolic pathways. Statistically significant pathways are emboldened, non-significant pathways are included for reference in future studies. Match status is number of metabolites in this experiment out of number of compounds involved in a pathway. Holm p is used to control multiple testing. FDR = False Discovery Rate. Pathway Impact is a combination of centrality and pathway enrichment results calculated by the sum of the importance of each metabolite divided by the sum of all metabolites in each pathway. Match P- Pathway Name -log(P) Holm P FDR Impact Status value Flavone and flavonol biosynthesis 4/9 0.0005 7.7043 0.039226 0.039226 0.8 Phenylalanine metabolism 3/8 0.0046 5.3911 0.39189 0.19822 0.16667 Isoquinoline alkaloid biosynthesis 2/6 0.0284 3.5625 1 0.82268 0.5 Flavonoid biosynthesis 5/43 0.0456 3.0868 1 0.99287 0.00566 Purine metabolism 5/61 0.1497 1.8992 1 1 0.04869 Tyrosine metabolism 2/18 0.2031 1.5942 1 1 0.45455 Glucosinolate biosynthesis 4/54 0.2407 1.4244 1 1 0.00952 Phenylalanine, tyrosine and tryptophan 2/21 0.2558 1.3635 1 1 0 biosynthesis Indole alkaloid biosynthesis 1/7 0.2845 1.2571 1 1 0 Valine, leucine and isoleucine 2/26 0.3438 1.0677 1 1 0.01865 biosynthesis Aminoacyl-tRNA biosynthesis 4/67 0.3809 0.96519 1 1 0 Nicotinate and nicotinamide metabolism 1/12 0.4373 0.82719 1 1 0 Valine, leucine and isoleucine 2/34 0.4768 0.74074 1 1 0 degradation Zeatin biosynthesis 1/19 0.5987 0.51308 1 1 0 Ubiquinone and other terpenoid-quinone 1/23 0.6694 0.40135 1 1 0 biosynthesis

187 alpha-Linolenic acid metabolism 1/23 0.6694 0.40135 1 1 0 Diterpenoid biosynthesis 1/26 0.7143 0.33646 1 1 0.01368 Tryptophan metabolism 1/27 0.7279 0.31762 1 1 0.17059 Porphyrin and chlorophyll metabolism 1/29 0.7532 0.28347 1 1 0.00806 Glycine, serine and threonine 1/30 0.7649 0.26797 1 1 0 metabolism

5.3.4 Convergence of compounds for plant carnivory

Thirty-four secondary plant metabolites already known to function in plant

carnivory are also present in Drosera capensis in this experiment from putative

annotation. Of these, only nonanal (pelargonaldehyde) did not have significant fold

changes in concentration after prey capture when compared to the results of the

multiple ANOVA. From the volcano plot highlighting high significance and fold

changes, five compounds were identified that are also important in carnivory for

other carnivorous plants (Figure 5.13 C, E; Figure 5.14 J, K, L). Of the 34 compounds,

21 compounds have been found in Drosera capensis that are also produced and linked

to carnivory in plants that do not share a common carnivorous ancestor (light grey

bars of Table 5.3). Eleven compounds have been identified for the first time in

Nepenthales (a carnivorous clade of five genera with multiple trap types) that are

also present in an unrelated carnivorous lineage (dark grey bars of Table 5.3).

188 Table 5.3: Secondary metabolites annotated in this experiment and have been found to have a role in carnivory in a previous study, highlighting occurrence of this compound in other carnivorous plants and whether this experiment shows agreement of previously proposed function. Light grey bars are metabolites newly found present for Drosera. Dark grey bars are newly found in Nepenthales. White bars are compounds found previously for D. capensis. Ticks indicate that findings of this experiment are congruent with the previously proposed function. Crosses do not agree with previously proposed function and brackets indicate partial agreement or disagreement and these compounds require further investigation, “ns” = no significant fold-change across any treatment, arrows indicate concentration increase or decrease by the end of the experiment. No. lineages = number of lineages in which the metabolite has been found to have a carnivorous function.

Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

Leaf bending in Yilamujiang et al., Drosera rotundifolia, Isoleucine (conjugate Drosera spp., Droseraceae, Nepenthales 2016; Mithofer et al., Digestion Drosera capensis & ü 1 ↑ of jasmonic acid) Nepenthes sp. Nepenthaceae (Caryophyllales) 2014; Nakamura et digestion in Nepenthes al., 2013 sp.

Attraction Phenylacetaldehyde Scent û Heliamphora spp. Sarraceniaceae Ericales Jaffe et al., 1995 2 ↑

Attraction -Eudesmol Floral or fruit scent û Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 2 ↑

Seal outer stomach Methyl Jasmonate and secretion of Nepenthales Escalante-Perez et al., Digestion (Jasmonic acid ü Dionaea sp. Droseraceae 1 ↑ digestive enzymes in (Caryophyllales) 2011 methyl ester) Dionaea muscipula

189 Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

Floral or fruit scent Kreuzwieser et al., (Downregulated after Dionaea sp., Droseraceae, Nepenthales Attraction Caryophyllene oxide û 2014; Di Giusto et al., 1 ↑ prey capture in Nepenthes sp. Nepenthaceae (Caryophyllales) 2010 Dionaea muscipula) Nepenthales Nepenthes sp., Nepenthaceae, Di Giusto et al., 2010; Attraction Linalool Flower, lavender scent (û) (Caryophyllales) 2 ↑ Sarracenia sp. Sarraceniaceae Jurgens et al., 2009 , Ericales VOC (Emitted from Nepenthales Kreuzweiser et Attraction Octanal trap but not confirmed (û) Dionaea sp. Droseraceae 1 ↑ (Caryophyllales) al.,2014 to attract prey) Floral or fruit scent Nepenthales Hotti et al., 2017, (Downregulated after Dionaea sp., Droseraceae, Attraction p-Cymene û (Caryophyllales) Kreuzweiser et al., 2 ↑ prey capture in Sarracenia spp. Sarraceniaceae , Ericales 2014 Dionaea muscipula) Scent (Floral - attracts Nepenthales Nepenthes sp., Nepenthaceae, Di Giusto et al., 2010; Attraction Benzyl acetate Lepidoptera, û (Caryophyllales) 2 ↑ Sarracenia spp. Sarraceniaceae Jurgens et al., 2009 Noctuidae) , Ericales Scent (Emitted from Dionaea muscipula Kreuzweiser et al., Dionaea sp., Droseraceae, Nepenthales trap but not 2014; Di Giusto et al., Attraction Benzaldehyde û Nepenthes sp., Nepenthaceae, (Caryophyllales) 2 ↑ individually 2010, Jurgens et al., Sarracenia spp. Sarraceniaceae , Ericales confirmed to attract 2009 prey) Capture, Ion transport in Dionaea sp., Droseraceae, Nepenthales Escalante-Perez et al., retention & Abscisic acid Nepenthes, trap closure ü 1 ↑ Nepenthes spp. Nepenthaceae (Caryophyllales) 2011 digestion sensitivity in Venus

190 Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

flytrap

Found on the spoon of Attraction xylene (o-, m-, p-) the pitcher assumed to û Heliamphora spp. Sarraceniaceae Ericales Jaffe et al., 1995 2 ↑ be for attraction Seal outer stomach and secretion of Yilamujiang et al., digestive enzymes in Capture Dionaea muscipula, 2016; Mithofer et al., Dionaea muscipula. Leaf Droseraceae, Nepenthales and Jasmonic acid ü Drosera spp., 2014; Nakamura et 1 ↑ bending in Drosera spp. Nepenthaceae (Caryophyllales) digestion Nepenthes sp. al., 2013; Escalante- Digestive enzyme Perez et al., 2011 secretion in Nepenthes spp. Floral or fruit scent (Emitted from Dionaea Kreuzweiser et Dionaea sp., Droseraceae, Nepenthales Attraction Decanal muscipula trap but not û al.,2014; Di Giusto et 1 ↑ Nepenthes sp. Nepenthaceae (Caryophyllales) confirmed to attract al., 2010 prey) VOC (Emitted from Nepenthales Kreuzweiser et Attraction Isomenthol trap but not confirmed û Dionaea sp. Droseraceae 1 ↑ (Caryophyllales) al.,2014 to attract prey) Kreuzweiser et al., Scent (Downregulated Dionaea sp., Droseraceae, Nepenthales 2-Phenylethanol 2014; Di Giusto et al., Attraction after prey capture in û Nepenthes sp., Nepenthaceae, (Caryophyllales) 2 ↑ (Phenylethyl alcohol) 2010; Jurgens et al., Dionaea muscipula) Sarracenia spp. Sarraceniaceae , Ericales 2009

191 Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

Nepenthales Kreuzweiser et al., 6-Methyl-5-hepten-2- Dionaea sp., Drosera Droseraceae, Attraction Scent (û) (Caryophyllales) 2014; Jurgens et al., 2 ↑ one sp., Sarracenia spp. Sarraceniaceae , Ericales 2009

Attraction Trans-jasmone Scent (û) Sarracenia spp. Sarraceniaceae Ericales Hotti et al., 2017 2 ↑

Hotti et al., 2017; Dionaea sp., Droseraceae, Nepenthales Kreuzweiser et al., Attraction Nonanal Floral Scent ns Nepenthes sp., Nepenthaceae, (Caryophyllales) 2 ns 2014; Di Giusto et al., Sarracenia spp. Sarraceniaceae , Ericales 2010

Nepenthales 2-Phenylethyl acetate Nepenthes sp., Nepenthaceae, Di Giusto et al., 2010; Attraction Fruit, sweet, scent (û) (Caryophyllales) 2 ↑ (Phenethyl acetate) Sarracenia spp. Sarraceniaceae Jurgens et al., 2009 , Ericales

Attraction Not tested (Present in Kaempferol (û) Roridula sp. Roridulaceae Ericales Wollenweber 2007 2 ↓ or capture sticky resin)

Luteolin (and Attraction Not tested (Present in derivatives velutin (û) Roridula sp. Roridulaceae Ericales Wollenweber 2007 2 ↓ or capture sticky resin) and apometzgerin) Sweet, flower scent Dionaea sp., Droseraceae, Nepenthales Hotti et al., 2017; Attraction Benzyl alcohol (Emitted from Dionaea (û) Nepenthes sp., Nepenthaceae, (Caryophyllales) Kreuzweiser et al., 2 ↑ muscipula trap but not Sarracenia sp. Sarraceniaceae , Ericales 2014; Di Giusto et al.,

192 Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

individually 2010; Jurgens et al., confirmed to attract 2009 prey) Induces secretion of Nepenthales Yilamujiang et al., Digestion Salicylic acid digestive enzymes in (ü) Nepenthes sp. Nepenthaceae 1 ↓ (Caryophyllales) 2016 Nepenthes sp. Scent (Peppermint. Attracts Lepidoptera, Dionaea sp., Droseraceae, Nepenthales Di Giusto et al., 2010; Attraction Methyl salicylate Noctuidae. (Though ü Nepenthes sp., Nepenthaceae, (Caryophyllales) 2 ↓ Jurgens et al., 2009 usually acts as a stress Sarracenia spp. Sarraceniaceae , Ericales response))

Dionaea sp., Drosera Droseraceae, Nepenthales Digestion Gallic acid Anti-fungal ü Kovácik et al., 2012 1 ↓ sp., Nepenthes sp. Nepenthaceae (Caryophyllales)

5-Hydroxy- scent (unknown Attraction ü Sarracenia spp. Sarraceniaceae Ericales Jurgens et al., 2009 2 ↓ methylfurfural specific attraction) Apigenin (and Attraction derivatives acacetin Not tested (Present in (ü) Roridula sp. Roridulaceae Ericales Wollenweber 2007 2 ↓ or Capture and 7,4'-dimethyl sticky resin) ether) Dionaea sp., Drosera Nepenthales Krolicka et al., 2008; Digestion Myricetin Antibacterial (û) Droseraceae 1 ↓ spp. (Caryophyllales) Marczak et al., 2005

193 Support for Regulation Plant genus/ Proposed suggested No. after prey Metabolite Proposed function species previously Family Order Source role function in lineages capture in studied this study this study

Nepenthales Digestion Isorhamnetin Not specifically tested û Drosera spp. Droseraceae Marczak et al., 2005 1 ↓ (Caryophyllales)

Attraction Not tested (Present in Quercetin (ü) Roridula sp. Roridulaceae Ericales Wollenweber 2007 2 ↓ or capture sticky resin)

Attraction Tricetin (derivative Not tested (Present in (ü) Roridula sp. Roridulaceae Ericales Wollenweber 2007 2 ↓ or capture corymbosin) sticky resin) Darlingtonia sp., Insect paralyzing Capture Coniine ü Heliamphora sp., Sarraceniaceae Ericales Hotti et al., 2017 2 ↑ agent Sarracenia sp. Raj et al., 2011 (Digestion - Buch et al., 2013; Gonçalves Nepenthes sp. Toxic or anaesthetic et al., 2008; Krolicka (Digestion: Droseraceae, influence on prey on et al., 2008; Marczak Aldrovanda sp., Drosophyllaceae Capture the rim of the pitcher Nepenthales et al., 2005; Aung et Plumbagin ü Dionaea sp., Drosera , Nepenthaceae, 1 ↓ *[Digestion] (*also digestion (Caryophyllales) al., 2002; Rischer et spp., Drosophyllum Dioncophyllacea Antibacterial, al., 2002; Bringman sp., Nepenthes sp., e insecticide) 2000; Culham & Triphyophyllum sp. Gornall 1994; Cannon et al., 1980; Zenk et al., 1969)

194 5.4 Discussion

I have quantified, for the first time, the whole-leaf metabolome response to the addition of animal substrate to a carnivorous plant trap. Though some metabolites have been previously shown to be involved in plant carnivory (Chapter 4), I show that there is a substantial metabolic response, which involves a large proportion of known biochemical systems as well as at least 164 unidentified secondary metabolites. The largest response (fold changes and number of metabolites) was of defence related processes, and compounds that are associated with the attraction of pollinators in other plant systems. This finding is important because it indicates that carnivory, at the metabolic level, is derived from at least two distinct plant functions–defence and attraction of organisms. Many of the secondary metabolites that demonstrated a response had fold increases over 30 times the concentration of unfed plants. This demonstrates that there is a clear and substantial metabolic response to prey capture.

Drosera capensis have co-opted or exapted the use of existing secondary plant metabolites known to function in non-carnivorous plant processes for responding to prey capture (Nakamura et al., 2013; Pavlovič & Saganová, 2015). Some of these compounds increased to considerably higher concentrations and/or are sustained for much longer in response to prey capture than is typical of the known response in non-carnivorous plant functions. For example, defence related jasmonates such as isoleucine increased by 3000% within 24 hrs of simulated prey capture and was sustained above this level for the full 72-hr experiment. Wounding does not instigate a significant change in concentration of this compound, but response of this compound to herbivory peaks similarly to prey capture after 1.5 hrs but returns to the original concentration by 3 hrs post herbivory (Mithöfer et al., 2014). In addition, phenylacetaldehyde is a floral scent and flavour. In tomato (Solanum lycopersicum), over expression of genes controlling phenylacetaldehyde increased concentration by up to ten times (Schwab et al., 2008). In the present study, however, a 36-fold

195 increase in concentration in response to prey capture was recorded. Metabolites can have multiple functions within the plant based on biological thresholds. Here,

Drosera capensis appears to have metabolic thresholds at much higher levels to instigate and maintain a carnivorous response to prey capture. Drosera capensis, in response to substrate via the leaves instigates a highly complex biochemical response utilising many compounds across multiple metabolic pathways, differentiating from non-carnivorous functions through either higher concentrations of the compound, or by sustaining high levels for longer.

In this experiment, I identified previously known compounds involved in carnivory and confirmed their predicted regulation following prey capture in D. capensis (Table

5.3), demonstrating that this is a robust experiment and that identification of other compounds involved in carnivory is valid. Compounds that are up- or downregulated following prey capture are expected to be involved in some capacity in carnivorous processes in plants (Kreuzwieser et al., 2014; Grace & Hudson, 2016).

Accumulation of jasmonates are known to instigate leaf bending in D. capensis and is therefore expected to feature as an important compound in this experiment

(Mithöfer et al., 2014; Krausko et al., 2017). As expected, isoleucine conjugate of jasmonic acid (Libiaková et al., 2014), is one of the highest fold-increasing compounds in this study. Along with isoleucine, there are also numerous other conjugates of similar chemical composition. For example, leucyl-leucine, a dipeptide composed of two leucine residues. Whether persisting levels of this compound are caused by synthesis of these compounds to instigate carnivory or are a product of breakdown of nitrogenous substances within prey is not determined but are likely important in the carnivorous habit. Given isoleucine, along with other compounds known to be important in carnivory for D. capensis, is present in this study and highlighted as important in the analysis (Figure 5.13 & Figure 5.14), it is reasonable to assume that other compounds highlighted from this experiment also hold importance in carnivory, particularly compounds that are upregulated following

196 substrate addition. Compounds that increase in concentration are likely to be involved in the retention, digestion or assimilation of nutrients from prey.

The function of compounds that decrease in concentration is not entirely clear. The downregulation of compounds may have multiple biological explanations and, though these compounds are included here as important, should be interpreted with caution, for example isorhamnetin, tricetin, gallic acid and kaempferol (Table 5.3 ↓).

Kreuzwieser et al. (2014) hypothesise that VOCs are down regulated following prey capture. They speculate that this is because if the compound functions as an attractant to prey it is not necessary to produce following prey capture. Reduced synthesis of attractants following prey capture is logical to assume as a decrease in the production of this compound minimises the net cost of production (Grace &

Hudson, 2016). I found that some, but not all, VOCs were downregulated which partially supports this hypothesis, though further work is necessary to determine the role of some VOCs in carnivory. Furthermore, it is not only VOCs that are downregulated following prey capture. The present study directly analysed trap tissue, as opposed to only volatiles emitted from the leaves. Many metabolites are stored as an inactive compound until some form of stimulus instigates the activation of these compounds at specific times (Gachon et al., 2005). Important compounds can, as a result, rapidly become active rather than the plant having to synthesise the whole compound on demand. The presentation of these processes would be one metabolite increasing in concentration, whilst another decreases. While it remains reasonable to state that compounds up- and down-regulated following prey capture are likely to be involved in some aspect of carnivory, their specific function, particularly for fold decreases, cannot be explicitly determined. It can be stated, however, that due to the high number of compounds involved that change in intensity following prey capture (~2383 metabolites), there is a reliance on secondary plant metabolites to facilitate plant carnivory, and that some VOCs previously

197 associated with attraction decrease following prey capture which may reduce the cost of carnivory (Ellison & Gotelli, 2009).

The identification of compounds involved in attraction due to their grouping as a volatile organic compound requires further investigation. Kreuzweiser et al. (2014) hypothesised that decreases in compounds are likely to be related to attraction of prey and are therefore downregulated once prey are captured. Of the ~171 secondary plant metabolites known to be involved in plant carnivory (Chapter 4),

142 have been proposed to play a role in the attraction of prey, mostly due to their classification as volatile organic compounds (Jürgens et al., 2009; Kreuzwieser et al.,

2014; Hotti et al., 2017). This classification appears to be misleading as, rather than a description of biological relevance, this classification is a chemical grouping and these compounds usually have very low volatility under ambient conditions (under

European classification have boiling points under 250°C (US EPA, 2019)). In this experiment, 24 of the 34 compounds also found in other carnivorous plants have previously been suggested to function in attraction (Table 5.3) (Jürgens et al., 2009;

Kreuzwieser et al., 2014; Hotti et al., 2017). If compounds are involved in attraction, it is expected that these compounds will decrease in concentration after prey capture, or at the very least not change in concentration. Of the 24 compounds, 18 increase in concentration after prey capture, in some cases by over 30 times the concentration of unfed plants (Table 5.3). This is opposite to expected responses if the compound is solely involved in the attraction of prey. Though these compounds’ involvement in carnivory is not disputed here, I propose that these compounds provide an alternative function, at least for D. capensis but probably other carnivorous plants, that aids in either capture, retention, digestion or assimilation and requires further investigation to determine their specific function (Table 5.3).

This experiment identified a large number of metabolites which showed a concentration response to substrate addition. The approach used in this study, however, was not exhaustive. Here, I only used monophasic extraction and HPLC-

198 MS in positive ion mode. While it can be stated that there are many aspects of the biochemical response that is not assessed this study, it remains that the metabolites identified as biologically important in the carnivorous syndrome of D. Capensis is valid. The aim of this experiment was to determine the extent to which there is a biochemical response to prey capture using metabolomics and identify directions for future work. In this regard, the extraction and sampling protocol used in this experiment is sufficient and fulfils the requirements to address the questions appropriately (Zukunft et al., 2017). Even within more limited exploratory analyses, I identify hundreds of compounds that are significantly up- and down-regulated in response to prey capture. Increasing the spectrum of metabolites able to be detected by using bi-phasic extractions and negative and positive ion identification will only enhance our findings. Future work should be directed at both identification of metabolites and further exploratory analysis of specific compounds, particularly of other carnivorous species in response to prey capture.

One possibility with this type of study, particularly with carnivorous plants in which many of their adaptations are exapted from defence related responses (Eilenberg et al., 2010; Renner & Specht, 2013; Van de Weyer et al., 2016; Lan et al., 2017), is that I may be detecting a stress response rather than a carnivorous response. I do not consider this to be the case in this experiment. Leaf bending occurred in response to substrate addition within hours and persistently high levels of metabolites (e.g. isoleucine had a thirty-fold increase, Figure 5.13) are above thresholds seen in wounding or stress responses (Escalante-Pérez et al., 2011; Nakamura et al., 2013).

Furthermore, if substrate addition was affecting some of the plants, the response would have occurred more variably across the treatments and individual plants.

That the plants responded in a highly uniform behaviour in which only the leaf with substrate added responded indicates that the results in this experiment are indeed carnivorous responses to simulated prey capture (Krausko et al., 2017).

199 Secondary plant metabolites are hypothesised to play an important role in plant diversification and evolution into novel environments (Theis & Lerdau, 2003; Wink,

2003; Lewinsohn & Gijzen, 2009). If secondary metabolites are a conduit for plant evolutionary adaptation, co-occurring metabolites are expected to be present among unrelated taxa that are evolved to occupy similar habitats or possess similar syndromes (Fang et al., 2018). The extent to which convergent evolution of secondary metabolites is true for carnivorous plants, a polyphyletic group of plants adapted to gain nutrients from prey (Ellison & Gotelli, 2009), is unclear. I putatively identified

~34 compounds that are also present in other carnivorous plants, 11 of which are new to the Nepenthales lineage but have been previously identified in other carnivorous lineages (Table 5.3). That two thirds of the matched compounds are found across independent lineages of carnivory support the hypothesis that secondary plant metabolites are important for evolutionary adaptation to specific environments (Theis & Lerdau, 2003; Wink, 2003). This experiment highlights that even across independent lineages of carnivorous plants, many of the same compounds are used to some extent for the purpose of acquiring nutrients from prey i.e. phenylacetaldehyde, nonanal and quercetin. These findings partly support the hypothesis that secondary metabolites are important for evolutionary adaptation to the environment (Pichersky & Gang, 2000; Hartmann, 2007; Lewinsohn & Gijzen,

2009). It should be noted, however, that these compounds are annotated from metabolite databases, and therefore require specific targeted analysis to comprehensively confirm their presence and function.

5.5 Conclusion

I conclude that secondary plant metabolites can play a significant role in the multifaceted plant adaptation of carnivory to attract, capture, digest and assimilate nutrients from prey. Secondary metabolites may therefore hold an important role in plant evolution and diversification to specific environments (Hartmann, 2007;

Wagner, 2017). Not only does it appear that there is a strong biochemical basis in all

200 aspects of carnivory, there also appears to be a large diversity of compounds that are important for the process of capturing and digesting prey for nutrients, as indicated by the large number of up-regulated compounds following prey capture. This experiment highlights consistent up- and down-regulation of compounds over time, but also more complex metabolic responses which may indicate late stage metabolism as different aspects of digestion and assimilation take effect. The experiment also shows that certain pre-existing secondary metabolites are co-opted for novel processes either by having higher concentration threshold or requirement for sustained concentrations over time to differentiate and instigate biological activity.

Future work should be directed at accurate identification of biologically relevant metabolites and their functions proposed in this paper. This experiment also highlights the co-occurrence of secondary metabolites in related and unrelated lineages of carnivory that are linked to the carnivorous habit by fold changes in the concentration of these compounds after simulated prey capture. The finding that independently evolved carnivorous plants utilise many of the same compounds for carnivory further supports the hypothesis that secondary plant metabolites are important drivers of plant evolution and diversification into new environments. Due to the extremely large number of secondary metabolites found to respond to prey capture, it appears that the importance of secondary plant metabolites in complex adaptations may be even greater than originally thought.

201 5.6 Appendix C5

RT: 0.00 - 12.01 9.89 NL: 100 1.04E10 TIC MS 80 Blank02

40 9.64 9.95 8.68 10.78 20 7.42 7.84 Relative Abundance Relative 7.25 0.43 0.64 1.43 2.70 3.29 3.79 4.36 5.11 5.48 0 9.85 NL: 100 1.11E10 TIC MS 5.37 80 4.72 QC10 4.69 4.75 60 4.47 5.46 3.29 40 0.46 0.47 5.74 7.02 9.62 4.02 8.68 9.93 10.74 20 0.87 1.73 7.85 11.47 1.94 3.75 6.76 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Time (min)

Figure Appendix C5.1: Total ion current (TIC) chromatogram for a blank and a QC

run.

202 Chapter 6 Synthesis

6.1 Thesis overview

Variability in the environment impacts on resource availability to organisms and influences ecological processes at all scales from molecular to ecosystem, from small- scale to inter-continental, from minutes to millennia (Chase & Leibold, 2003; Cramer

& Willig, 2005; Gratani, 2014). It is therefore important to understand how organisms respond to changes in resource availability at different spatial and temporal scales.

This understanding can give insight into physiological, ecological and evolutionary processes and can help to predict the ecological impact of climate change (Gratani,

2014; Bozinovic & Pörtner, 2015). This thesis therefore investigated plant responses at a range of spatial and temporal scales.

This thesis has addressed the following research questions to further our general understanding of plant responses to resource availability; (i) to what extent is the carnivorous habit phenotypically plastic in response to disturbance? (ii) do climate and microtopography interact at small spatial scales to impact on intraspecific trait variation? (iii) what is the role of secondary metabolites in plant carnivory? (iv) what is the metabolic response to prey capture?

To answer research question (i) I measured the phenotypic response of naturally shaded Drosera rotundifolia to disturbance (via the removal of canopy vegetation in

Chapter 2) over successive growing seasons. In response to disturbance mediated increases in light, D. rotundifolia increased allocation to carnivory and derived a higher proportion of N from prey. Such phenotypic plasticity likely enables the plants to survive in environments with contrasting resource availability. The outcome of this experiment suggests that D. rotundifolia are potentially specialists of environments that experience disturbance and can cope with environmental variability. Not only is seedling proliferation likely higher in disturbed areas, similar to D. capillaris (Brewer, 1998), but findings from this experiment highlight a

203 phenotypic response enabling shade tolerance and utility of disturbance which may contribute to D. rotundifolia extending its fundamental niche outside current model predictions of carnivorous plant habitat restrictions (Givnish et al., 1984; Givnish,

2015). Furthermore, these findings suggest that these plants are occupying environments that are, for non-carnivorous plants, untenable (i.e. high stress and disturbance) according to universal adaptive theory (Grime & Pierce, 2012).

To answer research question (ii), I investigated how habitat heterogeneity and climate-driven changes in nutrient availability impact on D. rotundifolia trait variability across peatlands in Europe. I confirmed that patterns of hydrology, vegetation cover and consequent light availability are consistent across microforms in peatlands (Rydin et al., 2013). I also found, however, that climate variability indirectly influences nutrient availability between microforms (Eppinga et al., 2010) which reversed expected patterns of trait expression. These findings mean that climate will likely alter within-habitat patterns of heterogeneity. Species will, therefore, not only need to adapt to changes in temperature and precipitation as expected from future climate change, but also within-habitat patterns of resource availability will be affected. These findings demonstrate that climate impacts are complex, and that climate interacts with local ecology in complex ways at very small spatial scales. My findings show that even at these small spatial scales, intraspecific trait variation in response to these differences is well-defined.

To answer research question (iii), I conducted a research review to determine the role of secondary plant metabolites in carnivory. There is a clear biochemical basis in plant carnivory, and there are patterns of metabolite production for carnivory across close- and distantly-related taxa. To answer research question (iv), I compared the metabolic profiles of trap tissue of unfed and fed plants of the Cape sundew, D.

Capensis. Identification of the up- and down-regulated compounds in response to prey capture provided substantial evidence that plant carnivory is facilitated by many secondary plant metabolites. The magnitude of the biochemical response was

204 much higher than expected (i.e. more than 30 times more compounds were up- or down-regulated in response to simulated prey capture than are currently identified across all carnivorous plants). These findings of the review and experiment on the role of secondary plant metabolites in carnivory support the hypothesis that secondary plant metabolites are important in plant responses to the environment and partially support that secondary plant metabolites are important in plant diversification and evolution.

The research presented here establishes general responses of two congeneric species of carnivorous plant, in terms of physiology and morphology and resultant nutrient stoichiometry. The results of these studies highlight carnivorous plants as ideal study species to determine the impact of climate change on the condition and responses of plants within temperate peatlands, a globally significant environment to mitigate anthropogenically-driven climate change (Leifeld & Menichetti, 2018).

6.2 Synthesis of findings and new avenues for future work

The findings of this thesis highlight substantial phenotypic plasticity and intraspecific trait variation as a means for plants to cope with environmental resource availability and variability. Furthermore, climate interacts in complex ways with many environmental resources and conditions such as nutrient availability, hydrology, and subsequent vegetation patterns which can impact on the light environment. The ecosystem impact of these processes therefore must be understood in terms of local norms in addition to how these processes may be altered by climate changes. Findings in this thesis are thus synthesised following a theoretical framework for the influence of abiotic and biotic processes and carnivorous plant responses outlined in Figure 6.1. The findings from the secondary metabolite review and experiment highlight a significant biochemical component of plant carnivory and indicate that secondary plant metabolites may have a critical role in the resilience of plants to environmental change. To determine the responses of plants to

205 Figure 6.1: Framework for considering the impact of resource availability on carnivorous plant traits from climate to plant physiology. Yellow arrows indicate result from an increase in investment in carnivory. Numbers signify which empirical chapter examines the specific process.

resource availability and enable accurate predictions of the impacts of future

environmental change, a multi-faceted approach to identifying these responses is

required, as in this thesis.

6.3 Climate

Understanding the impact of anthropogenically driven (or otherwise) climate

change on the status of ecosystems is a global concern (Walther et al., 2002; Bellard et

al., 2012). Chapter 3 illustrates a simple method to determine the impact of climate

gradients on abiotic processes within an environment and the subsequent effect on

plant trait expression which can occur at community level and at the molecular level

Figure 6.1. The study found that climate has a substantial impact on plant traits

206 indirectly, by interacting with nutrient cycling processes that are specific to the local environment and across very small (< 1m) distances. These findings are important, because it means that climate is a significant factor in plant trait expression in direct and indirect ways, influencing ecosystem processes globally in more discrete ways than simply temperature or rainfall alone (Bellard et al., 2012; Madani et al., 2018).

The three sites in Chapter three were selected along a gradient of ET:P that ranges from hollow to intermediate, or neutral, accumulation of nutrients. To fully understand the dynamics of climate driven impacts on nutrient accumulation and resultant plant traits, it is necessary to examine sites that are situated along more points of the ET:P gradient, specifically environments such as Siberian peatlands that are predicted to have nutrient accumulation in hummocks (in addition to the hollow and neutral accumulation sites in this thesis). If findings from these sites are congruent with results in this thesis, the implications this has for predicting climate change impacts on peatlands are substantial.

Evaporation:precipitation patterns arise from several different contributing factors, including latitude, distance to water masses and geographical topography or elevation (Hegazy & Kabiel, 2007; Randin et al., 2009; Trenberth, 2011).

Consequently, climate change has different ecological implications for different peatlands as climate clearly interacts with environmental processes at very small spatial scales. Understanding and predicting the impacts of climate change must, therefore, be highly contextualised; there is a need to have much finer spatial detail than only at a regional level.

This thesis focuses on the impact of precipitation and evapotranspiration, the patterns of which are expected to change due to anthropogenically driven climate change. There are other anthropogenic factors, however, that are predicted to impact on ecosystem health (Molau, 2010; Bellard et al., 2012). The burning of fossil fuels and N from fertilisers becoming airborne is exacerbating issues of nitrogen deposition in addition to emission of greenhouse gases and climate change (Chang et

207 al., 2016; Payne et al., 2017). The tripartite impact of anthropogenically driven changes to rainfall, temperature and nutrient deposition on ecosystems, in particular nutrient sensitive environments, such as temperate peatlands used in this thesis, potentially will result in regime shifts in the general patterns of vegetation composition of these environments, as can be followed in Figure 6.1. Over time this may drastically change species distributions with catastrophic effects in these environments as habitats are lost (Phoenix et al., 2006; Pastor et al., 2013; Leifeld &

Menichetti, 2018).

Secondary plant metabolites are identified in Chapter 5 to be important for rapid plant responses to resources and the environment and may be a key component of plant tolerance to climate change (Theis & Lerdau, 2003). They are increasingly being considered important as conduits for evolution at rates higher than mutation, a mechanism in which plants can rapidly diversify into new environments (Wink,

2003; Lewinsohn & Gijzen, 2009). Chapter 3 highlights the predicted impact climate change is likely to have on environments and microhabitats, and this research alluded to the fact that the rate of change in the environment may be at a rate faster than can be responded to, in terms of adaptation or migration. Current species distribution models often overlook trait variability and phenotypic plasticity

(Randin et al., 2009; Van der Putten et al., 2010; Benito Garzón et al., 2019), and perhaps more importantly, do not consider how the propensity for plants to produce novel metabolites may provide a rapid mechanism for plants to cope with the predicted changes in the environment. Chapter 4 and 5 highlight the likelihood that secondary plant metabolites have played a significant role in the evolution of carnivorous plants to enable them to survive in nutrient deficient sites. If secondary plant metabolites are a common source of rapid evolutionary adaptation to cope with the environment, then they may be the key for future resilience to climate change (Holopainen et al., 2018).

208 6.4 Light

There is scope for further study of the impact of light on carnivorous plants specifically. Carnivorous plants are generally restricted to high light environments, as predicted by energetics models for the evolution of carnivory (Ellison & Gotelli,

2009). Carnivorous plants persist, however, in well shaded habitats, as comprehensively evidenced in Chapter 2. If light is such a fundamental factor of carnivory, why is shade tolerance so common, at least for D. rotundifolia? Do they have other methods to reduce substantially the costs of carnivory? Is carbon uptake from prey more frequent or substantial per prey item than we currently understand?

Light is, undoubtedly, an important resource for investment in carnivory for most carnivorous plants and they respond to increasing light by investing in carnivorous traits–some even incorporate the reflection of light into their primary capture strategy (Moran et al., 2012). There remain, however, unexplored aspects of the importance of light. The impact of light may be less significant for individuals that are at mature growth stages. Drosera capillaris showed an increased emergence from seed following fire (Brewer, 1999), which was confirmed to be a result of clearing of above-ground vegetation, increasing the availability of light, as opposed to changes to the soil.

Potentially the importance of light for carnivorous plants is most apparent during seedling establishment and growth, though mature plants do benefit from the increased light (Zamora et al., 1998; Alcalá & Domínguez, 2005). It is already known that seedling growth is significantly enhanced by prey capture, and some species even capture a different spectra of prey to enable this habit during seedling developmental growth stages (Hatcher & Hart, 2014), so light availability may also have additional importance for carnivorous plant seedling proliferation.

It has been proposed that carnivorous plants may require disturbance to survive, to prevent being outcompeted (Brewer, 1999). At this stage, whether disturbance is a requirement, or whether D. rotundifolia are simply better able to take advantage of

209 these perturbations because of its ability to capture a nutrient resource otherwise unavailable to plants, cannot be exclusively confirmed, and therefore the extent of the importance of disturbance in their life history cannot be concluded yet.

Assessing the importance of disturbance on the success of carnivorous plants will likely require long term studies in which plants experiencing a gradient of disturbance frequency are monitored.

For these studies to be ecologically relevant, more information on the frequency and extent of disturbances in peatlands is required. Disturbances are known to occur in peatlands and can remove vegetation cover for prostrate plants (Turetsky & St.

Louis, 2006; Robroek et al., 2010). However, the rate in which these happen is not clear. Communities are often decades old in peatlands, which requires either infrequent disturbance or low disturbance intensity, but disturbance does occur at small spatial scales through herbivory and trampling. Clearly, there is a strong plastic response by D. rotundifolia to disturbance (Chapter 2), but the extent to which this is crucial to their success requires further investigation. The findings of which, has important implications for understanding the importance of disturbance and competition for plants in stressful environments—a point of contention between

Grime’s CSR theory and Tilman’s R* theory (Tilman, 1985; Craine, 2005; Grime,

2006).

The findings from Chapter 2, in which a population of heavily shaded D. rotundifolia subsequently had vegetation removed to increase the light environment, raise intriguing areas for future work in general plant science. Increasing the light transmission from ca. 35% to 100% had significant effects on aspects of the plant other than carnivorous trait expression. Alongside the expected decrease in etiolation related traits, the most notable was the accumulation of red pigmentation soon after the removal of vegetation (Figure 6.2). Drosera rotundifolia leaf colouration spans the red-green spectrum (i.e. shaded plants are obviously green, and plants in open well-lit areas are a deep red). The colouration of carnivorous plants of Drosera

210 Figure 6.2: Colour response likely due to accumulation of anthocyanins in response to increased light availability during the first growing season from Chapter 2. Left: D. rotundifolia naturally shaded; right; the same D. rotundifolia individual 2 months after vegetation removal. and other Nepenthales become redder through the accumulation of anthocyanins in the leaf tissue (Culham & Gornall, 1994; Schaefer & Ruxton, 2008). The role of this pigmentation, however, remains poorly understood.

Historically, it had been assumed that the function of red colouration of carnivorous plants was to attract prey (Juniper et al., 1989). However, evidence is controversial.

Red colouration is an important adaptation to capture prey in the pitcher plant

Nepenthes ventricosa, but this is likely due to contrast with background vegetation, rather than the red colour being specifically attractive (Schaefer & Ruxton, 2008;

Bennett & Ellison, 2009). Prey common to carnivorous species often are not able to perceive the red wavelengths specifically. Clearly, visual cues are important for attraction of prey in some species (Schaefer & Ruxton, 2008), but colouration has not been found to influence prey attraction or capture in D. rotundifolia or Pinguicula spp.

(Foot et al., 2014; Annis et al., 2018). UV reflectance has, however, been shown to be important in the attraction of prey (Kurup et al., 2013), but these patterns are not ubiquitous across carnivorous plants and was not measured in this thesis. It may also be that the red colouration is a deterrent to herbivores, as is common in other organisms (Menzies et al., 2016). Herbivory of carnivorous plants is not well

211 researched currently, especially regarding the impact of colour (Suárez-Piña et al.,

2016). Insight into the extent of herbivory on carnivorous plants will contribute to understanding of the extent of the cost of producing traps and whether there are adaptations to deter herbivores as well as the well documented carnivorous and photosynthetic leaf traits.

Anthocyanin function in non-carnivorous plants has been associated with photoprotection, with evidence of accumulation most commonly in the vacuoles of photosynthetic cells (Gould, 2004; Tanaka et al., 2008). It is hypothesised that an important function of anthocyanins is to protect shade-adapted chloroplasts from brief exposure to high intensity sunlight (Ramakrishna & Ravishankar, 2011).

Evidence in Chapter 2 partially supports this hypothesis, but pigmentation persists over successive growing seasons in high light treatments. These findings are congruent with plants growing in the hollows being consistently redder than shaded hummocks as hollows are consistently open microforms (Chapter 3). Drosera rotundifolia, although predicted to favour high light environments, likely accumulate anthocyanins in response to light stress to protect leaf tissue. Findings of Chapter 3 indicate, in addition, that nutrient stress may also have an impact on pigmentation of leaf tissue, which is a known response in Arabidopsis thaliana (Diaz et al., 2006). It is evident that anthocyanins have the potential for a broad array of functions, not just related to photoprotection (Gould, 2004; Egan & Van Der Kooy, 2013; Tušek et al.,

2016). Research as to the importance of anthocyanins is ongoing and carnivorous plants potentially provide a useful system to disentangle anthocyanin function and utility.

6.5 Root-derived and prey-derived N in carnivorous plants

The factors involved in the expression of carnivory can be complex (Figure 6.1) and understanding the outcome of investment in carnivory can be problematic under certain environmental conditions. For example, investment in carnivory, as shown in

212 Chapter 3, appears similar in high light, high N environments compared to low light, low N environments, though for entirely different resource limitations.

Carnivorous plants supplement their low root nutrient uptake by assimilating nutrients from animal prey (Ellison & Gotelli, 2009). Root N uptake is, however, often retained and responses to N uptake can be contrasting based on the location of uptake within carnivorous plants i.e. roots or leaves. Increases in N uptake via roots, as expected, reduces the expression of carnivory (Millett et al., 2012). Decreases in carnivory following root N uptake is logical—why waste resources on being carnivorous if there is enough N via roots?

Responses to N addition via leaves, however, is somewhat surprising. Prey nutrient uptake can maintain or even increase root activity (Adamec, 2002; Gao et al., 2015;

Paniw et al., 2017). Explanations for this include that a main benefit of carnivory is to enhance the nutrient uptake efficiency from the roots, or that root uptake remains important for nutrients other than N and P (Ellison, 2006). Drosera spp. and

Pinguicula spp. that were fed insects via leaves grew faster and accumulated higher tissue concentrations of N, P, K, Ca and Mg relative to unfed plants than they could have acquired from prey alone (Adamec, 2002). Carnivorous plants therefore respond to prey or root nutrient uptake in a multitude of ways.

In addition to my main findings, results from the light manipulation experiment

(Chapter 2) and the study (Chapter 3) suggest that carnivorous plants regulate nutrient stoichiometry via uptake of nutrients via roots (see ternary diagrams from

Chapter 2 and 3, Figure 2.6, Figure 3.4 C). The ratio of available nutrients from prey compared to soil suggests that the balance of nutrients in carnivorous plants should be different depending on the proportion of nutrients derived from prey compared to roots (Ellison & Adamec, 2018). However, my findings indicate that nutrient stoichiometry is stable within D. rotundifolia (whole plant nutrient analysis) across varying levels of root nutrient, light availability, and prey capture. Further research investigating nutrient balance within carnivorous plants would extend current

213 knowledge of plant nutrient regulation. Potentially, as shown in Chapter 5, there is a biochemical basis in the regulation of root:shoot signalling that is separate for prey derived N to root derived N. Labelling nutrients via foliar and root uptake and analysing chemical signalling within the plant would elucidate whether novel root- shoot signalling systems have evolved or whether regulation is simply a feedback mechanism of tissue nutrient composition that is controlled only by root nutrient uptake.

6.6 Biochemical responses (secondary plant metabolites)

Prey death is an important, and mostly overlooked, aspect of plant carnivory.

Mucilage and other sticky traps typically suffocate their victims. Pitcher plants normally drown their prey, but the rate of prey death in pitcher fluid occurs much faster than in pure water (Adlassnig et al., 2011). Increase rate of prey death may be due to the production of secondary plant metabolites such as naphthoquinones or compounds that impact on neurosensory pathways (Eilenberg 2010). The experiment in Chapter 5 putatively identified compounds produced by Drosera capensis, that are linked to sensory impairment of prey (e.g. coniine which is also present in Sarracenia spp.). The acceleration of prey death is an important aspect of prey retention and further study of these specialised compounds and their importance to prey retention could provide insight into plant evolution and advances in other areas such as agricultural pesticide research.

It is now clear that secondary plant metabolites serve specific functions in all aspects plant carnivory and are intrinsically linked to multiple environmental processes

(Chapter 4 & 5, Figure 6.1). Secondary plant metabolites are increasingly being considered as important drivers of plant diversification and adaptation (Iason et al.,

2012; Reen et al., 2015). The fact that information on their role in plant carnivory is scarce is likely due to a lack research, rather than a reflection of importance or extent of their function in the carnivorous syndrome. The laboratory experiment in Chapter

214 5 clearly supports the suggestion that secondary plant metabolites are integral to plant carnivory and that a vast and complex process of metabolite production is required to facilitate carnivory. The confirmation of a highly complex metabolic function in carnivory is important, because it means that we are likely missing key information on the origins of plant carnivory which has occurred across at least 10 separate plant lineages (Ellison & Adamec, 2018).

The findings of Chapter 5 raise numerous areas for further study, as is intended from untargeted metabolomics experiments, and refines our current understanding of the biochemical basis of carnivory in plants. Jasmonates play a key role in the leaf bending of Drosera capensis (Nakamura et al., 2013), but currently there is no information on how this accumulation is regulated and what limits or reverses this process. Jasmonates are important compounds in non-carnivorous plants, functioning as a plant response to herbivory, and therefore progressing our understanding of how these compounds are regulated in carnivorous plants may also provide insight into non-carnivorous plant function which may be key for developing pest-resistant crops (Ramakrishna & Ravishankar, 2011; Wasternack &

Hause, 2013; Dar et al., 2015).

The requirement for rapid response of compounds in plant carnivory provides scope for further research in biochemical regulation. Biologically inactive compounds may be converted into active compounds to instigate a rapid chemical response to prey capture. Non-carnivorous plants perform this through storage of inactive compounds through glycosylation (Vaistij et al., 2009). This aspect of plant metabolism may explain in part the proportion of glycosylated compounds in the metabolic profiles in Chapter 5 and requires further investigation.

If these compounds can then be converted back into their biologically inactive state, then the cost of producing such compounds would be drastically reduced, which could then be applied to current models for the evolutionary adaptation of plant carnivory (Givnish et al., 1984). This aspect of secondary plant metabolism may be

215 understood through longer metabolite studies over the course of prey capture and digestion. Presumably at some point the plant returns to the original metabolic state of an unfed plant. The rate that this occurs and use of biomarkers may provide a method of determining this process and the potential recycling of metabolites for the carnivorous syndrome.

6.7 The omics

The omics era is in its infancy, particularly regarding species such as carnivorous plants that remain outside of the traditional model species, i.e. Arabidopsis thaliana

(Viant et al., 2017). Chapter 4 and 5 address the significance of metabolites in plant carnivory, highlighting evolutionary convergence of certain compounds. Discoveries made in this thesis (Chapter 5) indicate that, as is common in non-carnivorous plants, carnivorous plants produce a high diversity of metabolites (Hartmann, 2007).

Interestingly, it also seems that there are many co-occurring compounds in related and unrelated carnivorous taxa (Chapter 4 & 5). Evolutionary convergence of compounds involved in plant carnivory may be a combination of inherent biochemical pathways shared across the plant kingdom, but also may indicate that certain paths of evolution or co-option of metabolites are preferable (likely) in nutrient deficient environments that can lead to the evolution of carnivory

(Buchanan et al., 2015). Only now are we uncovering some of the patterns of biochemical pathways that facilitate plant carnivory, and research using metabolomics in addition to other ‘omics’ technologies provides a cutting-edge method to understand plant evolutionary biology (Bylesjö et al., 2007; Beale et al.,

2016; Brunetti et al., 2018).

Combining metabolomics with genomic, transcriptomic and proteomic methods provides a link for understanding the relationships between gene expression, plant traits and responses to the environment (Figure 6.3, Beale et al., 2016). Carnivorous plants are taxonomically spread across 10 independent lineages (Ellison & Adamec,

216 2018). Carnivorous plant phylogeny, therefore, provides a well-defined syndrome in which to follow evolutionary adaptation from genetic coding, through to trait expression, linking bioinformatics directly to the ecology of these species. Recent molecular discoveries are being used to define and confirm taxonomic groupings, and are an effective tool for determining angiosperm phylogeny that is now defined by genetics over morphology (Cameron et al., 2002). Transcriptomics provides the

Genomics Transcriptomics Proteomics Metabolomics

DNA RNA Proteins Metabolites

Figure 6.3: Pathway for omics technologies to identify gene expression through to phenotypic expression linking genes to the environment. link between genes, gene expression and synthesis of compounds or expression of adaptations. Metabolomics provides a powerful approach to profile these plants to understand better the complexities of the less well understood chemical nature of this syndrome.

To my knowledge, the omics have not been completely combined to study carnivorous plants. There are only a few instances of partial overlap in these technologies (Fukushima et al., 2017), which are widely considered separate from ecophysiological studies, but are rapidly growing fields of research in themselves

(Brunetti et al., 2018; Flexas & Gago, 2018). Utricularia and Genlisea species to date contain the smallest of all plant genomes and for this are model systems for understanding structure and organisation of plant genomes across the kingdom

(Leushkin et al., 2013; Lan et al., 2017). Proteomics are being deployed to determine the enzymatic activity of carnivorous plants to digest prey, which has strongly supported the hypothesis that many carnivorous processes have evolved from defence related strategies (Eilenberg et al., 2010; Pavlovič & Saganová, 2015).

217 Metabolomics is similarly associating the biochemical basis for carnivory as heavily linked to defence-related processes and provides the mechanism for these plants to rapidly respond to the environment (Nakamura et al., 2013; Mithöfer et al., 2014).

The biochemical aspect of carnivory is more represented in the literature, however, in other aspects of the carnivorous syndrome (Chapter 4). Attraction of prey is an important method for increasing potential nutrient uptake and carnivorous plants perform this through deception mostly, it would appear, through olfactory cues

(Jürgens et al., 2009; Kreuzwieser et al., 2014; Hotti et al., 2017). Volatile organic compounds identified through metabolomic techniques have shown that several carnivorous plants emit a blend of fruit and floral scents to lure prey into their traps

(Kreuzwieser et al., 2014; Bertol et al., 2015; Hotti et al., 2017). Research combining these bioinformatic approaches will not only improve our understanding of plant carnivory but will also refine and advance our understanding of plant evolution and adaptation.

6.8 Conclusions

This thesis highlights the importance of phenotypic plasticity for plants to respond to short-term changes in the environment, the impacts of habitat heterogeneity and climate on phenotypic variability within and across sites, and the metabolic response of plants to environmental variability. Climate interacts with the environment in direct and discrete ways at a range of scales, and results in this thesis highlight the impact changes climate may have on ecosystems across spatial scales. Processes of nutrient availability, hydrology, and subsequent vegetation patterns all impact on plant resource allocation which determines the realised strategies of plants within their environment. The concern, however, is that predicted changes to the environment caused by anthropogenically-driven climate change may be at a higher rate than plants are able to adapt. The findings of this thesis suggest that secondary metabolites may hold a significant role in the response and survival of plants to

218 changing environments as they have the propensity for plants to adapt at a rate faster than mutation. The findings of this thesis are therefore applicable to a range of disciplines and research using carnivorous plants as model systems may be more relevant for predicting ecological scenarios than contemporary model systems.

219 References Abbott M, Brewer SJ. 2016. Competition does not explain the absence of a carnivorous pitcher plant from a nutrient-rich marsh. Plant Soil 409: 495–504.

Acton A. 2012. Naphthalenes—Advances in Research and Application. Scholarly

Editions.

Adamec L. 1997. Mineral nutrition of carnivorous plants: A review. The Botanical

Review 63: 273–299.

Adamec L. 2002. Leaf absorption of mineral nutrients in carnivorous plants stimulates root nutrient uptake. New Phytologist 155: 89–100.

Adamec L. 2008. Mineral nutrient relations in the aquatic carnivorous plant

Utricularia australis and its investment in carnivory. Fundamental and Applied

Limnology / Archiv für Hydrobiologie 171: 175–183.

Adamec L. 2011. Ecophysiological look at plant carnivory: Why are plants carnivorous? Cellular Origin, Life in Extreme Habitats and Astrobiology 16: 455–489.

Adamec L. 2013. Foliar mineral nutrient uptake in carnivorous plants: what do we know and what should we know? Frontiers in plant science 4: 10.

Adlassnig W, Koller-Peroutka M, Bauer S, Koshkin E, Lendl T, Lichtscheidl IK.

2012. Endocytotic uptake of nutrients in carnivorous plants. The Plant Journal 71:

303–313.

Adlassnig W, Peroutka M, Lambers H, Lichtscheidl IK. 2005a. The roots of carnivorous plants. Plant and Soil 274: 127–140.

Adlassnig W, Peroutka M, Lang I, Lichtscheidl IK. 2005b. Glands of carnivorous plants as a model system in cell biological research. Acta Botanica Gallica 152: 111–

124.

Adlassnig W, Peroutka M, Lendl T. 2011. Traps of carnivorous pitcher plants as a habitat: composition of the fluid, biodiversity and mutualistic activities. Annals of

220 Botany 107: 181–194.

Adlassnig W, Steinhauser G, Peroutka M, Musilek A, Sterba JH, Lichtscheidl IK,

Bichler M. 2009. Expanding the menu for carnivorous plants: Uptake of potassium, iron and manganese by carnivorous pitcher plants. Applied Radiation and Isotopes 67:

2117–2122.

Aerts R, Chapin FS. I. 2000. Advances in ecological research. Academic Press.

Aerts R, Verhoeven JTA, Whigham DF. 1999. Plant-Mediated Controls on Nutrient

Cycling in Temperate Fens and Bogs. Ecology 80: 2170.

Agrawal AA. 2011. Current trends in the evolutionary ecology of plant defence.

Functional Ecology 25: 420–432.

Albert V, Williams S, Chase M. 1992. Carnivorous plants: phylogeny and structural evolution. Science 257: 1491–1495.

Alcalá RE, Domínguez CA. 2005. Differential selection for carnivory traits along an environmental gradient in Pinguicula moranensis. Ecology 86: 2652–2660.

Allaby M. 2010. A dictionary of Ecology. Oxford: Oxford University Press.

Allen MF. 1992. Mycorrhizal functioning : an integrative plant-fungal process. Chapman

& Hall.

Alonso I, Hartley SE, Thurlow M. 2001. Competition between heather and grasses on Scottish moorlands: Interacting effects of nutrient enrichment and grazing regime. Journal of Vegetation Science 12: 249–260.

Anderson B, Midgley J. 2002. It takes two to tango but three is a tangle: mutualists and cheaters on the carnivorous plant Roridula. Oecologia 132: 369–373.

Annis J, Coons J, Helm C, Molano-Flores B. 2018. The Role of Red Leaf Coloration in Prey Capture for Pinguicula planifolia. Southeastern Naturalist 17: 433–437.

APG. 2016. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG IV. Botanical Journal of the Linnean

221 Society 181: 1–20.

APIS. 2016a. Simple site-based assessment result for total N deposition (wet and dry) from 2013-2015 (3 year average) at Humberhead Peatlands.

APIS. 2016b. Simple site-based assessment result for total N deposition (wet and dry) from 2013-2015 (3 year average) at Inverewe Peatlands.

Arnesen T. 1999. Vegetation dynamics following trampling in rich fen at Sølendet,

Central Norway; a 15 year study of recovery. Nordic Journal of Botany 19: 313–327.

Auld JR, Agrawal AA, Relyea RA. 2010. Re-evaluating the costs and limits of adaptive phenotypic plasticity. Proceedings of the Royal Society B: Biological Sciences

277: 503–511.

Aung H, Chia L, Goh N, Chia T, Ahmed A, Pare P, Mabry T. 2002. Phenolic constituents from the leaves of the carnivorous plant Nepenthes gracilis. Fitoterapia 73:

445–447.

Banasiuk R, Kawiak A, Krölicka A. 2012. In vitro cultures of carnivorous plants from the Drosera and Dionaea genus for the production of biologically active secondary metabolites. Biotechnologia 93: 87–96.

Baranyai B, Joosten H. 2016. Biology, ecology, use, conservation and cultivation of round-leaved sundew (Drosera rotundifolia L.): a review. Mires and Peat 18: 1–28.

Bauer U, Bohn HF, Federle W. 2008. Harmless nectar source or deadly trap:

Nepenthes pitchers are activated by rain, condensation and nectar. Proceedings of the

Royal Society B: Biological Sciences 275: 259–265.

Bauer U, Willmes C, Federle W. 2009. Effect of pitcher age on trapping efficiency and natural prey capture in carnivorous Nepenthes rafflesiana plants. Annals of Botany

103: 1219–1226.

Beale DJ, Karpe A V., Ahmed W. 2016. Beyond metabolomics: A review of multi- omics-based approaches. Microbial Metabolomics: Applications in Clinical,

222 Environmental, and Industrial Microbiology. Cham: Springer International

Publishing, 289–312.

Begon M, Harper JL, Townsend CR. 1986. Ecology. Individuals, populations and communities. Ecology. Individuals, populations and communities.

Bellard C, Bertelsmeier C, Leadley P, Thuiller W, Courchamp F. 2012. Impacts of climate change on the future of biodiversity. Ecology Letters 15: 365–377.

Benito Garzón M, Robson TM, Hampe A. 2019. ΔTraitSDM: species distribution models that account for local adaptation and phenotypic plasticity. New Phytologist.

Benjamini Y, Hochberg Y. 1995. Controlling The False Discovery Rate: A Practical and Powerful Approach to Multiple Testing,. Journal of the Royal Statistical Society 57:

289.

Bennett KF, Ellison AM. 2009. Nectar, not colour, may lure insects to their death.

Biology letters 5: 469–72.

Benzing DH. 2000. Bromeliaceae: profile of an adaptive radiation. Cambridge, UK ; New

York, NY, USA : Cambridge University Press.

Berendse F, Van Breemen N, Rydin H, Buttler A, Heijmans M, Hoosbeek MR, Lee

JA, Mitchell E, Saarinen T, Vasander H, et al. 2001. Raised atmospheric CO2 levels and increased N deposition cause shifts in plant species composition and production in Sphagnum bogs. Global Change Biology 7: 591–598.

Bergholz K, May F, Giladi I, Ristow M, Ziv Y, Jeltsch F. 2017. Environmental heterogeneity drives fine-scale species assembly and functional diversity of annual plants in a semi-arid environment. Perspectives in Plant Ecology, Evolution and

Systematics 24: 138–146.

Bertol N, Paniw M, Ojeda F. 2015. Effective prey attraction in the rare Drosophyllum lusitanicum, a flypaper-trap carnivorous plant. American Journal of Botany 102: 1–6.

Blanco G, Gerlagh R, Suh S, Barrett J, de Coninck HC, Diaz Morejon CF, Mathur

223 R, Nakicenovic N, Ofosu Ahenkora A, Pan J, et al. 2014. Drivers, Trends and

Mitigation. Climate Change 2014: Mitigation of Climate Change. Contribution of Working

Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change:

351–412.

Bott T, Meyer GA, Young EB. 2008. Nutrient limitation and morphological plasticity of the carnivorous pitcher plant Sarracenia purpurea in contrasting wetland environments. New Phytologist 180: 631–41.

Bozinovic F, Pörtner HO. 2015. Physiological ecology meets climate change. Ecology and Evolution 5: 1025–1030.

Bragazza L, Limpens J, Gerdol R, Grosvernier P, Hájek M, Hájek T, Hajkova P,

Hansen I, Iacumin P, Kutnar L, et al. 2005. Nitrogen concentration and δ15N signature of ombrotrophic Sphagnum mosses at different N deposition levels in

Europe. Global Change Biology 11: 106–114.

Bragazza L, Parisod J, Buttler A, Bardgett RD. 2013. Biogeochemical plant–soil microbe feedback in response to climate warming in peatlands. Nature Climate

Change 3: 273–277.

Brearley FQ. 2011. Natural abundance of stable isotopes reveals the diversity of carnivorous plant diets. Carnivorous Plant Newsletter: Journal of the International

Carnivorous Plant Society 40: 84–87.

Breeuwer A, Robroek BJM, Limpens J, Heijmans MMPD, Schouten MGC,

Berendse F. 2009. Decreased summer water table depth affects peatland vegetation.

Basic and Applied Ecology 10: 330–339.

Brewer SJ. 1998. Effects of competition and litter on a carnivorous plant, Drosera capillaris (Droseraceae). American Journal of Botany 85: 1592–1596.

Brewer JS. 1999. Effects of fire, competition, and soil disturbances on regeneration of a carnivorous plant (Drosera capillaris). American Midland Naturalist 141: 28–42.

224 Brewer JS. 2001. A demographic analysis of fire-stimulated seedling establishment of Sarracenia alata (Sarraceniaceae). American Journal of Botany 88: 1250–1257.

Brewer JS. 2008. Why Don’t Carnivorous Pitcher Plants Compete with Non-

Carnivorous Plants for Nutrients? Ecological Society of America 84: 451–462.

Brewer JS, Baker DJ, Nero AS, Patterson AL, Roberts RS, Turner LM. 2011.

Carnivory in plants as a beneficial trait in wetlands. Aquatic Botany 94: 62–70.

Bringmann G, Rischer H, Wohlfarth M, Schlauer J, Assi LA. 2000. Droserone from cell cultures of Triphyophyllum peltatum (Dioncophyllaceae) and its biosynthetic origin. Phytochemistry 53: 339–43.

Briske DD. 1996. Strategies of Plant Survival in Grazed Systems : A Functional

Interpretation. CAB International: The Ecology and Management of Grazing Systems: 37–

67.

Brooker RW, Kikvidze Z. 2008. Importance: an overlooked concept in plant interaction research. Journal of Ecology 96: 703–708.

Brundrett MC. 2002. Coevolution of roots and mycorrhizas of land plants. New

Phytologist 154: 275–304.

Brunetti AE, Carnevale Neto F, Vera MC, Taboada C, Pavarini DP, Bauermeister

A, Lopes NP. 2018. An integrative omics perspective for the analysis of chemical signals in ecological interactions. Chemical Society Reviews 47: 1574–1591.

Bruzzese BM, Bowler R, Massicotte HB, Fredeen AL. 2010. Photosynthetic light response in three carnivorous plant species: Drosera rotundifolia, D. capensis and

Sarracenia leucophylla. Photosynthetica 48: 103–109.

Buch F, Rott M, Rottloff S, Paetz C, Hilke I, Raessler M, Mithöfer A. 2013. Secreted pitfall-trap fluid of carnivorous Nepenthes plants is unsuitable for microbial growth.

Annals of Botany 111: 375–383.

Buchanan BB, Gruissem W, Jones RL. 2015. Biochemistry and molecular biology of

225 plants. New Jersey: Wiley Blackwell.

Budzianowski J. 1996. Naphthohydroquinone glucosides of Drosera rotundifolia and

D. intermedia from in vitro cultures. Phytochemistry 42: 1145–1147.

Butler JL, Ellison AM. 2007. Nitrogen cycling dynamics in the carnivorous northern pitcher plant, Sarracenia purpurea. Functional Ecology 21: 835–843.

Bylesjö M, Eriksson D, Kusano M, Moritz T, Trygg J. 2007. Data integration in plant biology: The O2PLS method for combined modeling of transcript and metabolite data. Plant Journal 52: 1181–1191.

Cameron KM, Wurdack KJ, Jobson RW. 2002. Molecular evidence for the common origin of snap-traps among carnivorous plants. American Journal of Botany 89: 1503–

1509.

Cannon J, Lojanapiwatna V, Raston C, Sinchai W, White A. 1980. The Quinones of

Nepenthes rafflesiana. The Crystal Structure of 2,5-Dihydroxy-3,8-dimethoxy-7- methylnaphtho-1,4-quinone (Nepenthone-E) and a Synthesis of 2,5-Dihydroxy-3-

Methoxy-7-methylnaphtho-1,4-quinone (Nepenthone-C). Australian Journal of

Chemistry 33: 1073.

Casal J. 2012. Shade avoidance. American Society of Plant Biologists.

Chang J, Ciais P, Viovy N, Vuichard N, Herrero M, Havlík P, Wang X, Sultan B,

Soussana J-F. 2016. Effect of climate change, CO2 trends, nitrogen addition, and land-cover and management intensity changes on the carbon balance of European grasslands. Global change biology 22: 338–50.

Chapin CT, Pastor J. 1995. Nutrient limitation in the northern pitcher plant

Sarracenia purpurea. Canadian Journal of Botany 73: 728–734.

Charnov EL. 1989. Phenotypic evolution under Fisher’s Fundamental Theorem of

Natural Selection. Heredity 62: 113–116.

Chase JM, Leibold MA. 2003. Ecological niches : linking classical and contemporary

226 approaches. University of Chicago Press.

Cheplick GP. 2015. Approaches to plant evolutionary ecology. Oxford University Press.

Chin L, Moran JA, Clarke C. 2010. Trap geometry in three giant montane pitcher plant species from Borneo is a function of tree shrew body size. New Phytologist 186:

461–470.

Clarke CM, Bauer U, Lee CC, Tuen AA, Rembold K, Moran JA. 2009. Tree shrew lavatories: a novel nitrogen sequestration strategy in a tropical pitcher plant. Biology

Letters 5: 632–635.

Clarke C, Moran JA. 2016. Climate, soils and vicariance - their roles in shaping the diversity and distribution of Nepenthes in Southeast Asia. Plant and Soil 403: 37–51.

Clymo RS, Hayward PM. 1982. The Ecology of Sphagnum. Bryophyte Ecology.

Dordrecht: Springer Netherlands, 229–289.

Cohen JM, Lajeunesse MJ, Rohr JR. 2018. A global synthesis of animal phenological responses to climate change. Nature Climate Change 8: 224–228.

Connell JH. 2014. Apparent versus “Real” Competition in Plants. Perspectives on

Plant Competition. Academic Press, Inc., 9–26.

Considine EC, Thomas G, Boulesteix AL, Khashan AS, Kenny LC. 2018. Critical review of reporting of the data analysis step in metabolomics. Metabolomics 14: 7.

Cook JL, Newton J, Millett J. 2018. Environmental differences between sites control the diet and nutrition of the carnivorous plant Drosera rotundifolia. Plant and Soil 423:

41–58.

Craine JM. 2005. Reconciling plant strategy theories of Grime and Tilman. Journal of

Ecology 93: 1041–1052.

Craine JM. 2007. Plant strategy theories: Replies to Grime and Tilman: Forum.

Journal of Ecology 95: 235–240.

Craine JM, Dybzinski R. 2013. Mechanisms of plant competition for nutrients,

227 water and light (D Robinson, Ed.). Functional Ecology 27: 833–840.

Craine JM, Lee WG, Bond WJ, Williams RJ, Johnson LC. 2005. Environmental constraints on a global relationship among leaf and root traits of grasses. Ecology 86:

12–19.

Cramer MJ, Willig MR. 2005. Habitat heterogeneity, species diversity and null models. Oikos 108: 209–218.

Crowder A, Pearson M, Grubs P, Hanglois P. 1990. Biological flora of the British

Isles: Drosera. Journal of Ecology 78: 233–267.

Crowley PH, Hopper KR, Krupa JJ. 2013. An insect-feeding guild of carnivorous plants and spiders: does optimal foraging lead to competition or facilitation? The

American naturalist 182: 801–19.

Culham A, Gornall RJ. 1994. The taxonomic significance of naphthoquinones in the

Droseraceae. Biochemical Systematics and Ecology 22: 507–515.

Dar TA, Uddin M, Masroor M, Khan A, Hakeem KR, Jaleel H. 2015. Jasmonates counter plant stress: A Review. Environmental and Experimental Botany 115: 49–57.

Darwin C. 1875. Insectivorous Plants. London - John Murray.

Diaz C, Saliba-Colombani V, Loudet O, Belluomo P, Moreau L, Daniel-Vedele F,

Morot-Gaudry J-F, Masclaux-Daubresse C. 2006. Leaf Yellowing and Anthocyanin

Accumulation are Two Genetically Independent Strategies in Response to Nitrogen

Limitation in Arabidopsis thaliana. Plant and Cell Physiology 47: 74–83.

Dixon K, Pate J, Bailey W. 1980. Nitrogen Nutrition of the Tuberous Sundew

Drosera erythrorhiza Lindl. With Special Reference to Catch of Arthropod Fauna by Its

Glandular Leaves. Australian Journal of Botany 28: 283.

Douma JC, Bardin V, Bartholomeus RP, van Bodegom PM. 2012. Quantifying the functional responses of vegetation to drought and oxygen stress in temperate ecosystems (J Watling, Ed.). Functional Ecology 26: 1355–1365.

228 Van Dyke F. 2008. Conservation biology : foundations, concepts, applications. Springer.

Eckstein RL, Karlsson PS. 2001. The effect of reproduction on nitrogen use- efficiency of three species of the carnivorous genus Pinguicula. Journal of Ecology 89:

798–806.

EEA. 2016. Climate change, impacts and vulnerability in Europe 2016 An indicator-based report.

Egan PA, Kooy F Van Der. 2012. Coproduction and Ecological Significance of

Naphthoquinones in Carnivorous Sundews (Drosera). Chemistry & Biodiversity 9:

1033–1044.

Egan PA, Van Der Kooy F. 2013. Review: Phytochemistry of the Carnivorous

Sundew Genus Drosera (Droseraceae) – Future Perspectives and

Ethnopharmacological Relevance. Chemistry and Biodiversity 10: 1774–1790.

Eilenberg H, Pnini-Cohen S, Rahamim Y, Sionov E, Segal E, Carmeli S, Zilberstein

A. 2010. Induced production of antifungal naphthoquinones in the pitchers of the carnivorous plant Nepenthes khasiana. Journal of Experimental Botany 61: 911–922.

El-Sayed AM, Byers JA, Suckling DM. 2016. Pollinator-prey conflicts in carnivorous plants: When flower and trap properties mean life or death. Scientific Reports 6:

21065.

Ellison AM. 2006. Nutrient Limitation and Stoichiometry of Carnivorous Plants.

Plant Biology 8: 740–747.

Ellison AM, Adamec L. 2011. Ecophysiological traits of terrestrial and aquatic carnivorous plants: Are the costs and benefits the same? Oikos 120: 1721–1731.

Ellison AM, Adamec L. 2018. Carnivorous Plants: Physiology, Ecology, and Evolution.

Oxford University Press.

Ellison AM, Gotelli NJ. 2001. Evolutionary ecology of carnivorous plants. Trends in

Ecology & Evolution 16: 623–629.

229 Ellison AM, Gotelli NJ. 2002. Nitrogen availability alters the expression of carnivory in the northern pitcher plant, Sarracenia purpurea. Proceedings of the National

Academy of Sciences of the United States of America 99: 4409–4412.

Ellison AM, Gotelli NJ. 2009. Energetics and the evolution of carnivorous plants -

Darwin’s ‘most wonderful plants in the world’. Journal of Experimental Botany 60: 19–

42.

Ellison AM, Gotelli NJ, Brewer JS, Cochran-Stafira DL, Kneitel JM, Miller TE,

Worley AC, Zamora R. 2003. The evolutionary ecology of carnivorous plants.

Advances in Ecological Research Volume 33. Elsevier, 1–74.

Eppinga MB, Rietkerk M, Belyea LR, Nilsson MB, De Ruiter PC, Wassen MJ. 2010.

Resource contrast in patterned peatlands increases along a climatic gradient. Ecology

91: 2344–2355.

Eppinga MB, Rietkerk M, Borren W, Lapshina ED, Bleuten W, Wassen MJ. 2008.

Regular surface patterning of peatlands: Confronting theory with field data.

Ecosystems 11: 520–536.

Epstein E, Bloom A. 2016. Mineral Nutrition of Plants Principles and Perspectives, 2nd

Edition. John Wiley & Sons.

Erickson R. 1978. Plants of prey in Australia. Nedlands: University of Western

Australia Press.

Erni P, Varagnat M, McKinley GH, Co A, Leal GL, Colby RH, Giacomin AJ. 2008.

Little Shop of Horrors: Rheology of the Mucilage of Drosera sp., a Carnivorous Plant.

AIP Conference Proceedings. AIP, 579–581.

Escalante-Pérez M, Krol E, Stange A, Geiger D, Al-Rasheid KAS, Hause B, Neher

E, Hedrich R. 2011. A special pair of phytohormones controls excitability, slow closure, and external stomach formation in the Venus flytrap. Proceedings of the

National Academy of Sciences of the United States of America 108: 15492–15497.

230 Etienne P, Sorin E, Maillard A, Gallardo K, Arkoun M, Guerrand J, Cruz F, Yvin J-

C, Ourry A, Etienne P, et al. 2018. Assessment of Sulfur Deficiency under Field

Conditions by Single Measurements of Sulfur, Chloride and Phosphorus in Mature

Leaves. Plants 7: 37.

Fahn A, Cutler DF. 1992. Xerophytes. Stuttgart: Gebrüder Borntraeger.

Falster DS, Westoby M. 2005. Tradeoffs between height growth rate, stem persistence and maximum height among plant species in a post-fire succession.

Oikos 111: 57–66.

Fang C, Fernie AR, Luo J. 2018. Exploring the Diversity of Plant Metabolism. Trends in Plant Science.

Farnsworth EJ, Ellison AM. 2007. Prey availability directly affects physiology, growth, nutrient allocation and scaling relationships among leaf traits in 10 carnivorous plant species. Journal of Ecology 0: 213–221.

Fleischmann A, Schlauer J, Smith SA, Givnish TJ. 2018. Evolution of carnivory in angiosperms. In: Ellison AM,, In: Adamec L, eds. Carnivorous Plants: Physiology, ecology, and evolution. Oxford University Press, 22–42.

Flexas J, Gago J. 2018. A role for ecophysiology in the ’omics’ era. The Plant Journal

96: 251–259.

Foot G, Rice SP, Millett J. 2014. Red trap colour of the carnivorous plant Drosera rotundifolia does not serve a prey attraction or camouflage function. Biology letters 10:

20131024–20131024.

Forterre Y, Skotheim J, Dumais J, Mahadevan L. 2005. How the Venus flytrap snaps. Nature 433: 421–425.

Frainer A, Polvi LE, Jansson R, McKie BG. 2018. Enhanced ecosystem functioning following stream restoration: The roles of habitat heterogeneity and invertebrate species traits (Y Cao, Ed.). Journal of Applied Ecology 55: 377–385.

231 Frank JH, O’Meara GF. 1984. The Bromeliad Catopsis berteroniana Traps Terrestrial

Arthropods but Harbors Wyeomyia larvae (Diptera: Culicidae). The Florida

Entomologist 67: 418.

Franzios G, Mirotsou M, Hatziapostolou E, Kral J, Scouras ZG, Mavragani-

Tsipidou P. 1997. Insecticidal and Genotoxic Activities of Mint Essential Oils. Journal of Agricultural and Food Chemistry 45: 2690–2694.

Frolking S, Talbot J, Jones MC, Treat CC, Kauffman JB, Tuittila E-S, Roulet N.

2011. Peatlands in the Earth’s 21st century climate system. Environmental Reviews 19:

371–396.

Frommlet F, Bogdan M, Ramsey D. 2014. Phenotypes and Genotypes. The Search for

Influential Genes.

Fronhofer EA, Legrand D, Altermatt F, Ansart A, Blanchet S, Bonte D, Chaine A,

Dahirel M, De Laender F, De Raedt J, et al. 2018. Bottom-up and top-down control of dispersal across major organismal groups. Nature Ecology & Evolution 2: 1859–

1863.

Fukushima K, Fang X, Alvarez-Ponce D, Cai H, Carretero-Paulet L, Chen C, Chang

T-H, Farr KM, Fujita T, Hiwatashi Y, et al. 2017. Genome of the pitcher plant

Cephalotus reveals genetic changes associated with carnivory. Nature Ecology &

Evolution 1: 0059.

Gachon CMMM, Langlois-Meurinne M, Saindrenan P. 2005. Plant secondary metabolism glycosyltransferases: the emerging functional analysis. Trends in Plant

Science 10: 542–549.

Gao P, Loeffler TS, Honsel A, Krüse J, Krol E, Scherzer S, Kreuzer I, Bemm F,

Buegger F, Burzlaff T, et al. 2015. Integration of trap- and root-derived nitrogen nutrition of carnivorous Dionaea muscipula. New Phytologist 205: 1320–1329.

Garrido B, Hampe A, Marañón T, Arroyo J. 2003. Regional differences in land use affect population performance of the threatened insectivorous plant Drosophyllum

232 lusitanicum (Droseraceae). Diversity and Distributions 9: 335–350.

Gaume L, Bazile V, Huguin M, Bonhomme V. 2016. Different pitcher shapes and trapping syndromes explain resource partitioning in Nepenthes species. Ecology and

Evolution 6: 1378–1392.

Gaume L, Perret P, Gorb E, Gorb S, Labat J-J, Rowe N. 2004. How do plant waxes cause flies to slide? Experimental tests of wax-based trapping mechanisms in three pitfall carnivorous plants. Arthropod Structure & Development 33: 103–111.

Gergely ZR, Martinez DE, Donohoe BS, Mogelsvang S, Herder R, Staehelin LA.

2018. 3D electron tomographic and biochemical analysis of ER, Golgi and trans Golgi network membrane systems in stimulated Venus flytrap (Dionaea muscipula) glandular cells. Journal of Biological Research-Thessaloniki 25: 15.

Gibson TC, Waller DM. 2009. Evolving Darwin’s ‘most wonderful’ plant: Ecological steps to a snap-trap. New Phytologist 183: 575–587.

Gilroy S, Jones DL. 2000. Through form to function: root hair development and nutrient uptake. Trends in Plant Science 5: 56–60.

Di Giusto B, Bessière J-M, Guéroult M, Lim LBL, Marshall DJ, Hossaert-McKey

M, Gaume L. 2010. Flower-scent mimicry masks a deadly trap in the carnivorous plant Nepenthes rafflesiana. Journal of Ecology 98: 845–856.

Givnish TJ. 2015. New evidence on the origin of carnivorous plants. Proceedings of the National Academy of Sciences 112: 10–11.

Givnish TJ, Burkhardt EL, Happel RE, Weintraub JD. 1984. Carnivory in the

Bromeliad Brocchinia reducta, with a Cost/Benefit model for the general restriction of carnivorous plants to sunny, moist, nutrient-poor habitats. The American Naturalist

124: 479–497.

Gommers CMM, Visser EJW, St Onge KR, Voesenek LACJ, Pierik R. 2013. Shade tolerance: when growing tall is not an option. Trends in Plant Science 18: 65–71.

233 Gonçalves S, Gonçalves MA, Ameixa O, Nogueira JMF, Romano A. 2008.

Insecticidal activity of leaf extracts from Drosophyllum lusitanicum against Liriomyza trifolii (Burgess) (Diptera: Agromyzidae). Journal of Horticultural Science and

Biotechnology 83: 653–657.

Gorb E, Haas K, Henrich A, Enders S, Barbakadze N, Gorb S. 2005. Composite structure of the crystalline epicuticular wax layer of the slippery zone in the pitchers of the carnivorous plant Nepenthes alata and its effect on insect attachment. The

Journal of experimental biology 208: 4651–62.

Gotelli NJ, Ellison AM. 2002. Nitrogen Deposition and Extinction Risk in the

Northern Pitcher Plant, Sarracenia purpurea. Ecology 83: 2758–2765.

Gould KS. 2004. Nature’s Swiss Army Knife: The Diverse Protective Roles of

Anthocyanins in Leaves. Journal of biomedicine & biotechnology 2004: 314–320.

Grace SC, Hudson DA. 2016. Processing and Visualization of Metabolomics Data

Using R. Metabolomics - Fundamentals and Applications. InTech, .

Grafe TU, Schöner CR, Kerth G, Junaidi A, Schöner MG. 2011. A novel resource– service mutualism between bats and pitcher plants. Biology Letters 7: 436–439.

Gratani L. 2014. Plant Phenotypic Plasticity in Response to Environmental Factors.

Advances in Botany 2014: 1–17.

Grime JP. 1973. Competition and diversity in herbaceous vegetation (reply). Nature

244: 311.

Grime JP. 1979. Primary strategies in plants. Transactions of the Botanical Society of

Edinburgh 43: 151–160.

Grime JP. 2006. Plant Strategies, Vegetation Processes, and Ecosystem Properties. John

Wiley & Sons.

Grime JP. 2007. Plant strategy theories: A comment on Craine (2005). Journal of

Ecology 95: 227–230.

234 Grime JP. 2009. Primary strategies in plants. Transactions of the Botanical Society of

Edinburgh 43: 151–160.

Grime JP, Pierce S. 2012. The Evolutionary Strategies that Shape Ecosystems. Wiley-

Blackwell.

Groome GM, Shaw P. 2015. Vegetation response to the reintroduction of cattle grazing on an English lowland valley mire and wet heath. Conservation Evidence 12.

Di Guida R, Engel J, Allwood JW, Weber RJM, Jones MR, Sommer U, Viant MR,

Dunn WB. 2016. Non-targeted UHPLC-MS metabolomic data processing methods: a comparative investigation of normalisation, missing value imputation, transformation and scaling. Metabolomics 12: 93.

Gulman S, Chu C. 1981. The effects of light and nitrogen on photosynthesis, leaf characteristics, and dry matter allocation in the chaparral shrub Diplacus aurantiacus. Oecologia 49: 207–21.

Hanslin HM, Karlsson PS. 1996. Nitrogen uptake from prey and substrate as affected by prey capture level and plant reproductive status in four carnivorous plant species. Oecologia 106: 370–375.

Harrison AF, Schulze E-D, Gebauer G, Bruckner G. 2000. Canopy Uptake and

Utilization of Atmospheric Pollutant Nitrogen. Springer, Berlin, Heidelberg, 171–

188.

Hartmann T. 2007. From waste products to ecochemicals: fifty years research of plant secondary metabolism. Phytochemistry 68: 2831–46.

Hatcher CR, Hart AG. 2014. Venus flytrap seedlings show growth-related prey size specificity. International Journal of Ecology 2014: 631–636.

Hedrich R, Neher E. 2018. Venus Flytrap: How an Excitable, Carnivorous Plant

Works. Trends in Plant Science 23: 220–234.

Hegazy AK, Fahmy GM, Ali MI, Gomaa NH. 2005. Growth and phenology of eight

235 common weed species. Journal of Arid Environments 61: 171–183.

Hegazy AK, Kabiel HF. 2007. Significance of microhabitat heterogeneity in the spatial pattern and size-class structure of Anastatica hierochuntica L. Acta Oecologica

31: 332–342.

Henarejos-Escudero P, Guadarrama-Flores B, García-Carmona F, Gandía-Herrero

F. 2018. Digestive glands extraction and precise pigment analysis support the exclusion of the carnivorous plant Dionaea muscipula Ellis from the Caryophyllales order. Plant Science 274: 342–348.

Herben T, Klimešová J, Chytrý M. 2018. Effects of disturbance frequency and severity on plant traits: An assessment across a temperate flora (CE Timothy Paine,

Ed.). Functional Ecology 32: 799–808.

Hodge A. 2004. The plastic plant: root responses to heterogeneous supplies of nutrients. New Phytologist 162: 9–24.

Holopainen JK, Virjamo V, Ghimire RP, Blande JD, Julkunen-Tiitto R,

Kivimäenpää M. 2018. Climate Change Effects on Secondary Compounds of Forest

Trees in the Northern Hemisphere. Frontiers in Plant Science 9: 1445.

Hotti H, Gopalacharyulu P, Seppänen-Laakso T, Rischer H. 2017. Metabolite profiling of the carnivorous pitcher plants Darlingtonia and Sarracenia (V Gupta, Ed.).

PLOS ONE 12: 1–21.

Huang Y, Wang Y, Sun L, Agrawal R, Zhang M. 2015. Sundew adhesive: a naturally occurring hydrogel. Journal of the Royal Society, Interface 12: 20150226.

Hutchens JJ, Luken JO. 2009. Prey capture in the Venus flytrap: collection or selection? Botany 87: 1007–1010.

Hutchings MJ, John LA, Stewart AJA. 2000. The Ecological Consequences of

Environmental Heterogeneity: The 40th Symposium of the British Ecological Society.

Blackwell Science.

236 Hutchinson GE. 1957. Concluding Remarks. Cold Spring Harbor Symposia on

Quantitative Biology 22: 415–427.

Iason G, Dicke M, Hartley S. 2012. The Ecology of Plant Secondary Metabolites: From

Genes to Global Processes. Cambridge University Press.

Jabot F, Pottier J. 2012. A general modelling framework for resource-ratio and CSR theories of plant community dynamics (J Fridley, Ed.). Journal of Ecology 100: 1296–

1302.

Jaffé K, Blum MS, Fales HM, Mason RT, Cabrera A. 1995. On insect attractants from pitcher plants of the genus Heliamphora (Sarraceniaceae). Journal of Chemical

Ecology 21: 379–384.

Jenkins DG, Pierce S. 2017. General allometric scaling of net primary production agrees with plant adaptive strategy theory and has tipping points (H Cornelissen,

Ed.). Journal of Ecology 105: 1094–1104.

Jennings DE, Rohr JR. 2011. A review of the conservation threats to carnivorous plants. Biological Conservation 144: 1356–1363.

Josten H, Clarke D. 2002. Wise use of mires and peatlands: Background and principles including a framework for decision making. International Mire

Conservation Group. International Peat Society, Jyväskylä 304: 304.

Juniper BE, Robins RJ, Joel DM. 1989. The Carnivorous Plants. Academic Press,

London.

Jürgens A, El-Sayed AM, Suckling DM. 2009. Do carnivorous plants use volatiles for attracting prey insects? Functional Ecology 23: 875–887.

Karagatzides JD, Butler JL, Ellison AM. 2009. The Pitcher Plant Sarracenia purpurea

Can Directly Acquire Organic Nitrogen and Short-Circuit the Inorganic Nitrogen

Cycle (J Chave, Ed.). PLoS ONE 4: e6164.

Karagatzides JD, Ellison AM. 2009. Construction costs, payback times, and the leaf

237 economics of carnivorous plants. American Journal of Botany 96: 1612–1619.

Karlsson PS. 1988. Seasonal Patterns of Nitrogen, Phosphorus and Potassium

Utilization by Three Pinguicula Species. Functional Ecology 2: 203.

Karlsson PS, Thoren LM, Hanslin HM. 1994. Prey capture by three Pinguicula species in a subarctic environment. Oecologia 99: 188–193.

Kesler HC, Trusty JL, Hermann SM, Guyer C. 2008. Demographic responses of

Pinguicula ionantha to prescribed fire: a regression-design LTRE approach. Oecologia

156: 545–557.

Kessler D, Baldwin IT. 2007. Making sense of nectar scents: the effects of nectar secondary metabolites on floral visitors of Nicotiana attenuata. The Plant Journal 49:

840–54.

Kim JA, Kim HS, Choi SH, Jang JY, Jeong MJ, Lee SI. 2017. The importance of the circadian clock in regulating plant metabolism. International Journal of Molecular

Sciences 18.

Kirwan JA, Weber RJM, Broadhurst DI, Viant MR. 2014. Direct infusion mass spectrometry metabolomics dataset: A benchmark for data processing and quality control. Scientific Data 1: 1–13.

Knight SE, Frost TM. 1991. Bladder Control in Utricularia Macrorhiza: Lake-Specific

Variation in Plant Investment in Carnivory. Ecology 72: 728–734.

Kokubun T. 2017. Occurrence of myo-inositol and alkyl-substituted polysaccharide in the prey-trapping mucilage of Drosera capensis. Die Naturwissenschaften 104: 83.

Koller EK, Phoenix GK. 2017. Seasonal dynamics of soil and plant nutrients at three environmentally contrasting sites along a sub-Arctic catchment sequence. Polar

Biology 40: 1821–1834.

Korhonen JFJ, Pihlatie M, Pumpanen J, Aaltonen H, Hari P, Levula J, Kieloaho A-

J, Nikinmaa E, Vesala T, Ilvesniemi H. 2013. Nitrogen balance of a boreal Scots pine

238 forest. Biogeosciences 10: 1083–1095.

Korrensalo A, Kettunen L, Laiho R, Alekseychik P, Vesala T, Mammarella I,

Tuittila E-S. 2018. Boreal bog plant communities along a water table gradient differ in their standing biomass but not their biomass production (S Roxburgh, Ed.).

Journal of Vegetation Science 29: 136–146.

Kottek M, Grieser J, Beck C, Rudolf B, Rubel F. 2006. World Map of the Köppen-

Geiger climate classification updated. Meteorologische Zeitschrift 15: 259–263.

Kováčik J, Klejdus B, Repčáková K. 2012a. Phenolic metabolites in carnivorous plants: Inter-specific comparison and physiological studies. Plant physiology and biochemistry : PPB / Société française de physiologie végétale 52: 21–7.

Kováčik J, Klejdus B, Štork F, Hedbavny J. 2012b. Prey-induced changes in the accumulation of amino acids and phenolic metabolites in the leaves of Drosera capensis L. Amino Acids 42: 1277–1285.

Kraft NJB, Godoy O, Levine JM. 2015. Plant functional traits and the multidimensional nature of species coexistence. Proceedings of the National Academy of

Sciences of the United States of America 112: 797–802.

Krausko M, Libiakova M, Pavlovič A. 2014. Feeding on prey increases photosynthetic efficiency in the carnivorous sundew Drosera capensis. Annals of

Botany 113: 69–78.

Krausko M, Perutka Z, Šebela M, Šamajová O, Šamaj J, Novák O, Pavlovič A.

2017. The role of electrical and jasmonate signalling in the recognition of captured prey in the carnivorous sundew plant Drosera capensis. New Phytologist 213: 1818–

1835.

Kreuzwieser J, Scheerer U, Kruse J, Burzlaff T, Honsel A, Alfarraj S, Georgiev P,

Schnitzler J-PJP, Ghirardo A, Kreuzer I, et al. 2014. The Venus flytrap attracts insects by the release of volatile organic compounds. Journal of Experimental Botany

65: 755–766.

239 Król EE, Płachno BJ, Adamec L, Stolarz M, Dziubińska H, Trebacz K, Krol E,

Plachno BJ, Adamec L, Stolarz M, et al. 2012. Quite a few reasons for calling carnivores ‘the most wonderful plants in the world’. Annals of Botany 109: 47–64.

Krolicka A, Szpitter A, Gilgenast E, Romanik G, Kaminski M, Lojkowska E. 2008.

Stimulation of antibacterial naphthoquinones and flavonoids accumulation in carnivorous plants grown in vitro by addition of elicitors. Enzyme and Microbial

Technology 42: 216–221.

Krolicka A, Szpitter A, Maciag M, Biskup E, Gilgenast E, Wegrzyn G, Lojkowska

E. 2009. Antibacterial and antioxidant activity of the secondary metabolites from in vitro cultures of the Alice sundew (Drosera aliciae). Biotechnology and applied biochemistry 53: 175–84.

Kruse J, Gao P, Honsel A, Kreuzwieser J, Burzlaff T, Alfarraj S, Hedrich R,

Rennenberg H. 2014. Strategy of nitrogen acquisition and utilization by carnivorous

Dionaea muscipula. Oecologia 174: 839–851.

Kunstler G, Falster D, Coomes DA, Hui F, Kooyman RM, Laughlin DC, Poorter L,

Vanderwel M, Vieilledent G, Wright SJ, et al. 2016. Plant functional traits have globally consistent effects on competition. Nature 529: 204–207.

Kurup R, Johnson AJ, Sankar S, Hussain AA, Kumar CS, Sabulal B. 2013.

Fluorescent prey traps in carnivorous plants (H Rennenberg, Ed.). Plant Biology 15:

611–615.

Lam WN, Chong KY, Anand GS, Wah Tan HT. 2017. Dipteran larvae and microbes facilitate nutrient sequestration in the Nepenthes gracilis pitcher plant host. Biology

Letters 13: 20160928.

Lam WN, Goh AWH, Kee NJH, Ng GH, Sia CH, Tan HTW. 2018. Within- individual physiology constrains carnivorous investment in the rainbow plant Byblis guehoi more than does environmental light intensity. Botany Letters 165: 274–279.

Lamers LPM, Bobbink R, Roelofs JGM. 2000. Natural nitrogen filter fails in

240 polluted raised bogs. Global Change Biology 6: 583–586.

Lan T, Renner T, Ibarra-Laclette E, Farr KM, Chang T-H, Cervantes-Pérez SA,

Zheng C, Sankoff D, Tang H, Purbojati RW, et al. 2017. Long-read sequencing uncovers the adaptive topography of a carnivorous plant genome. Proceedings of the

National Academy of Sciences 114: E4435–E4441.

Larsen LG, Harvey JW, Crimaldi JP. 2007. A delicate balance: ecohydrological feedbacks governing landscape morphology in a lotic peatland. Ecological

Monographs 77: 591–614.

Leifeld J, Menichetti L. 2018. The underappreciated potential of peatlands in global climate change mitigation strategies. Nature Communications 9: 1071.

Letunic I, Bork P. 2016. Interactive tree of life (iTOL) v3: an online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Research 44:

W242–W245.

Leushkin E V, Sutormin RA, Nabieva ER, Penin AA, Kondrashov AS, Logacheva

MD. 2013. The miniature genome of a carnivorous plant Genlisea aurea contains a low number of genes and short non-coding sequences. BMC Genomics 14: 476.

Lewinsohn E, Gijzen M. 2009. Phytochemical diversity: The sounds of silent metabolism. Plant Science 176: 161–169.

Li C, Soufan O, Chong J, Xia J, Bourque G, Li S, Caraus I, Wishart DS. 2018.

MetaboAnalyst 4.0: towards more transparent and integrative metabolomics analysis. Nucleic Acids Research 46: W486–W494.

Libiaková M, Floková K, Novák O, Slováková L, Pavlovič A. 2014. Abundance of cysteine endopeptidase dionain in digestive fluid of Venus flytrap (Dionaea muscipula

Ellis) is regulated by different stimuli from prey through jasmonates. PloS one 9: e104424.

Liebig J. 1840. Die organische Chemie in ihrer Anwendung auf Agrikultur und

241 Physiologie. Braunschweig, Germany: Friedrich Vieweg und Sohn Publ. Co.

Liebig J. 1855. Die Grundsatze der Agrikultur-Chemie mit Rucksichẗ auf die in England angestellten Untersuchungen,1st and 2nd edn. Braunschweig, Germany: Friedrich

Vieweg und Sohn Publ Co.

Limpens J, Bohlin E, Nilsson MB. 2017. Phylogenetic or environmental control on the elemental and organo-chemical composition of Sphagnum mosses? Plant and Soil

417: 69–85.

Loewus FA, Murthy PPN. 2000. myo-Inositol metabolism in plants. Plant Science 150:

1–19.

López-Bucio J, Cruz-Ramı́rez A, Herrera-Estrella L. 2003. The role of nutrient availability in regulating root architecture. Current Opinion in Plant Biology 6: 280–

287.

Lovett GM, Tear TH, Evers DC, Findlay SEG, Cosby BJ, Dunscomb JK, Driscoll

CT, Weathers KC. 2009. Effects of Air Pollution on Ecosystems and Biological

Diversity in the Eastern United States. Annals of the New York Academy of Sciences

1162: 99–135.

Luken J. 2007. Performance of Dionaea muscipula as Influenced by Developing

Vegetation. Economic Botany 134: 45–52.

MacArthur RH, Wilson EO. 1967. The theory of island biogeography. Princeton, N.J.:

Princeton University Press.

Madani N, Kimball JS, Ballantyne AP, Affleck DLR, van Bodegom PM, Reich PB,

Kattge J, Sala A, Nazeri M, Jones MO, et al. 2018. Future global productivity will be affected by plant trait response to climate. Scientific Reports 8: 2870.

Marazzi B, Bronstein JL, Koptur S. 2013. The diversity, ecology and evolution of extrafloral nectaries: current perspectives and future challenges. Annals of botany 111:

1243–50.

242 Marczak L, Stobiecki M, Łojkowska E. 2005. Secondary metabolites in in vitro cultured plants of the genus Drosera. Phytochemical Analysis 16: 143–149.

Masclaux-Daubresse C, Daniel-Vedele F, Dechorgnat J, Chardon F, Gaufichon L,

Suzuki A. 2010. Nitrogen uptake, assimilation and remobilization in plants:

Challenges for sustainable and productive agriculture. Annals of Botany 105: 1141–

1157.

McCormick P V., Harvey JW, Crawford ES. 2011. Influence of changing water sources and mineral chemistry on the everglades ecosystem. Critical Reviews in

Environmental Science and Technology 41: 28–63.

McGladdery C, Weindorf DC, Chakraborty S, Li B, Paulette L, Podar D, Pearson

D, Kusi NYO, Duda B. 2018. Elemental assessment of vegetation via portable X-ray fluorescence (PXRF) spectrometry. Journal of Environmental Management 210: 210–225.

McPherson S. 2008. Glistening carnivores : the sticky-leaved insect-eating plants. Redfern

Natural History Productions.

McPherson S, Fleischmann A, Robinson A. 2010. Carnivorous plants and their habitats. Redfern Natural History Productions Ltd.

Méndez M, Karlsson P. 1999. Costs and Benefits of Carnivory in Plants: Insights from the Photosynthetic Performance of Four Carnivorous Plants in a Subarctic

Environment. Oikos 86: 105–112.

Mengel K, Kirkby EA, Kosegarten H, Appel T. 2001. Plant Nutrients. Principles of

Plant Nutrition. Dordrecht: Springer Netherlands, 1–13.

Menzies IJ, Youard LW, Lord JM, Carpenter KL, van Klink JW, Perry NB, Schaefer

HM, Gould KS. 2016. Leaf colour polymorphisms: a balance between plant defence and photosynthesis (M Heil, Ed.). Journal of Ecology 104: 104–113.

Met Office. 2018. Humberhead peatlands annual climate data 1981-2010.

Michalko J, Socha P, Mészáros P, Blehová A, Libantová J, Moravčíková J,

243 Matušíková I. 2013. Glucan-rich diet is digested and taken up by the carnivorous sundew (Drosera rotundifolia L.): implication for a novel role of plant β-1,3- glucanases. Planta 238: 715–25.

Miller TE, Burns JH, Munguia P, Walters EL, Kneitel JM, Richards PM, Mouquet

N, Buckley HL. 2005. A critical review of twenty years’ use of the resource-ratio theory. The American naturalist 165: 439–48.

Millett J, Foot GW, Svensson BM. 2015. Nitrogen deposition and prey nitrogen uptake control the nutrition of the carnivorous plant Drosera rotundifolia. Science of the

Total Environment 512–513: 631–636.

Millett J, Foot GW, Thompson JC, Svensson BM. 2018. Geographic variation in

Sundew (Drosera) leaf colour: plant–plant interactions counteract expected effects of abiotic factors. Journal of Biogeography 45: 582–592.

Millett J, Jones RI, Waldron S. 2003. The contribution of insect prey to the total nitrogen content of sundews (Drosera spp.) determined in situ by stable isotope analysis. New Phytologist 158: 527–534.

Millett J, Svensson BM, Newton J, Rydin H. 2012. Reliance on prey-derived nitrogen by the carnivorous plant Drosera rotundifolia decreases with increasing nitrogen deposition. New Phytologist 195: 182–188.

Mithöfer A, Reichelt M, Nakamura Y. 2014. Wound and insect-induced jasmonate accumulation in carnivorous Drosera capensis : two sides of the same coin. Plant

Biology 16: 982–987.

Molau U. 2010. Long-term impacts of observed and induced climate change on tussock tundra near its southern limit in northern Sweden. Plant Ecology & Diversity

3: 29–34.

Van Der Molen PC, Schalkoort M, Smit R. 1994. Vegetation and ecology of hummock-hollow complexes on an Irish raised bog. Biology and Environment 94B:

145–175.

244 Molles MC. 2015. Ecology : concepts and applications. McGraw-Hill Education.

Moore BD, Andrew RL, Külheim C, Foley WJ. 2014. Explaining intraspecific diversity in plant secondary metabolites in an ecological context. New Phytologist 201:

733–750.

Moran JA, Clarke C, Gowen BE. 2012. The use of light in prey capture by the tropical pitcher plant Nepenthes aristolochioides. Plant signaling & behavior 7: 957–60.

Moran JA, Merbach MA, Livingston NJ, Clarke CM, Booth WE. 2001. Termite Prey

Specialization in the Pitcher Plant Nepenthes albomarginata—Evidence from Stable

Isotope Analysis. Annals of Botany 88: 307–311.

Muhammad A, Haddad PS, Durst T, Arnason JT. 2013. Phytochemical constituents of Sarracenia purpurea L. (pitcher plant). Phytochemistry 94: 238–242.

Naeem M, Khan MMA, Moinuddin. 2012. Triacontanol: a potent plant growth regulator in agriculture. Journal of Plant Interactions 7: 129–142.

Nakamura Y, Reichelt M, Mayer VE, Mithöfer A. 2013. Jasmonates trigger prey- induced formation of ‘outer stomach’ in carnivorous sundew plants. Proceedings.

Biological sciences / The Royal Society 280: 1–6.

Ne’eman, G., Ne’eman, R. and Ellison AM. 2006. Limits to reproductive success of

Sarracenia purpurea (Sarraceniaceae). American Journal of Botany 93: 1660–1666.

Newman T, Ibrahim S, Wheeler JW, McLaughlin WB, Petersen RL, Duffield RM.

2000. Identification of sarracenin in four species of Sarracenia (Sarraceniaceae).

Biochemical Systematics and Ecology 28: 193–195.

Nishi AH, Vasconcellos-Neto J, Romero GQ. 2013. The role of multiple partners in a digestive mutualism with a protocarnivorous plant. Annals of Botany 111: 143–150.

Nishida R. 2014. Chemical ecology of insect–plant interactions: ecological significance of plant secondary metabolites. Bioscience, Biotechnology, and Biochemistry

78: 1–13.

245 Nordbakken JF, Rydgren K, Økland RH. 2004. Demography and population dynamics of Drosera anglica and D. rotundifolia. Journal of Ecology 92: 110–121.

Novak M, Veselovsky F, Curik J, Stepanova M, Fottova D, Prechova E, Myska O.

2015. Nitrogen input into Sphagnum bogs via horizontal deposition: an estimate for

N-polluted high-elevation sites. Biogeochemistry 123: 307–312.

Olde Venterink H, Wassen MJ, Verkroost AWM, de Ruiter PC. 2003. Species

Richness – Productivity Patterns Differ Between N-,P-, And K-limited Wetlands.

Ecology 84: 2191–2199.

Ordoñez JC, van Bodegom PM, Witte J-PM, Wright IJ, Reich PB, Aerts R. 2009. A global study of relationships between leaf traits, climate and soil measures of nutrient fertility. Global Ecology and Biogeography 18: 137–149.

Paiva SR de, Figueiredo MR, Aragão TV, Kaplan MAC. 2003. Antimicrobial activity in vitro of plumbagin isolated from Plumbago species. Memórias do Instituto

Oswaldo Cruz 98: 959–961.

Palacio-López K, Beckage B, Scheiner S, Molofsky J. 2015. The ubiquity of phenotypic plasticity in plants: A synthesis. Ecology and Evolution 5: 3389–3400.

Paniw M, Gil-Cabeza E, Ojeda F. 2017. Plant carnivory beyond bogs: reliance on prey feeding in Drosophyllum lusitanicum (Drosophyllaceae) in dry Mediterranean heathland habitats. Annals of Botany 1: 1035–1041.

Paniw M, Salguero-Gómez R, Ojeda F. 2015. Local-scale disturbances can benefit an endangered, fire-adapted plant species in Western Mediterranean heathlands in the absence of fire. Biological Conservation 187: 74–81.

Parsons HM, Ekman DR, Collette TW, Viant MR. 2009. Spectral relative standard deviation: a practical benchmark in metabolomics. The Analyst 134: 478–85.

Pastor J, Peckham B, Bridgham S, Weltzin J, Chen J. 2002. Plant Community

Dynamics, Nutrient Cycling, and Alternative Stable Equilibria in Peatlands. The

246 American Naturalist 160: 553–568.

Pastor J, Solin J, Bridgham SD, Updegraff K, Harth C, Dewey B, Weishampel P.

2013. Global Warming and the Export of Dissolved Organic Carbon from Boreal

Peatlands. Oikos 2: 380–386.

Pate JS. 1986. Economy of symbiotic nitrogen fixation. On the economy of plant form and function.299–335.

Pausas JG. 2015. Bark thickness and fire regime. Functional Ecology 29: 315–327.

Pavlovič A, Jakšová J, Novák O. 2017. Triggering a false alarm: Wounding mimics prey capture in the carnivorous Venus flytrap (Dionaea muscipula). New Phytologist

216: 927–938.

Pavlovič A, Mithöfer A. 2019. Jasmonate signalling in carnivorous plants: copycat of plant defence mechanisms. Journal of Experimental Botany 70: 3379–3389.

Pavlovič A, Saganová M. 2015. A novel insight into the cost-benefit model for the evolution of botanical carnivory. Annals of Botany 115: 1075–1092.

Pavlovič A, Singerová L, Demko V, Hudák J. 2009. Feeding enhances photosynthetic efficiency in the carnivorous pitcher plant Nepenthes talangensis.

Annals of Botany 104: 307–314.

Payne RJ, Dise NB, Field CD, Dore AJ, Caporn SJ, Stevens CJ. 2017. Nitrogen deposition and plant biodiversity: past, present, and future. Frontiers in Ecology and the Environment 15: 431–436.

Pereira CG, Almenara DP, Winter CE, Fritsch PW, Lambers H, Oliveira RS. 2012.

Underground leaves of Philcoxia trap and digest nematodes. Proceedings of the

National Academy of Sciences of the United States of America 109: 1154–1158.

Peterson AT, Papeş M, Soberón J. 2015. Mechanistic and correlative models of ecological niches. European Journal of Ecology 1: 28–38.

Phoenix GK, Hicks WK, Cinderby S, Kuylenstierna JCI, Stock WD, Dentener FJ,

247 Giller KE, Austin AT, Lefroy RDB, Gimeno BS, et al. 2006. Atmospheric nitrogen deposition in world biodiversity hotspots: The need for a greater global perspective in assessing N deposition impacts. Global Change Biology 12: 470–476.

Pichersky E, Gang DR. 2000. Genetics and biochemistry of secondary metabolites in plants: an evolutionary perspective. Trends in Plant Science 5: 439–445.

Pichersky E, Gershenzon J. 2002. The formation and function of plant volatiles: perfumes for pollinator attraction and defense. Current Opinion in Plant Biology 5:

237–243.

Płachno BJ. 2007. Sweet but dangerous: nectaries in carnivorous plants. Acta

Agrobotanica 60: 31–37.

Plachno BJ, Adamec L, Lichtscheidl IK, Peroutka M, Adlassnig W, Vrba J. 2006.

Fluorescence labelling of phosphatase activity in digestive glands of carnivorous plants. Plant Biology 8: 813–820.

Poppinga S, Koch K, Bohn HF, Barthlott W. 2010. Comparative and functional morphology of hierarchically structured anti-adhesive surfaces in carnivorous plants and kettle trap flowers. Functional Plant Biology 37: 952.

Post DM. 2002. Using Stable Isotopes to estimate trophic postition: models, methods and assumptions. Ecology 83: 703–718.

Putman RJ, Wratten SD. 1983. Principles of Ecology. Dordrecht: Springer

Netherlands.

Van der Putten WH, Macel M, Visser ME. 2010. Predicting species distribution and abundance responses to climate change: why it is essential to include biotic interactions across trophic levels. Philosophical Transactions of the Royal Society B:

Biological Sciences 365: 2025–2034.

R Core Team. 2016. R: A language and environment for statistical computing.

Raj G, Kurup R, Hussain AA, Baby S. 2011. Distribution of naphthoquinones,

248 plumbagin, droserone, and 5-O-methyl droserone in chitin-induced and uninduced

Nepenthes khasiana: Molecular events in prey capture. Journal of Experimental Botany

62: 5429–5436.

Ramakrishna A, Ravishankar GA. 2011. Influence of abiotic stress signals on secondary metabolites in plants. Plant Signaling & Behavior 6: 1720–1731.

Randin CF, Engler R, Normand S, Zappa M, Zimmermann NE, Pearman PB, Vittoz

P, Thuiller W, Guisan A. 2009. Climate change and plant distribution: Local models predict high-elevation persistence. Global Change Biology 15: 1557–1569.

Ravee R, Mohd Salleh F ‘Imadi, Goh H-H. 2018. Discovery of digestive enzymes in carnivorous plants with focus on proteases. PeerJ 6: e4914.

Reed TE, Schindler DE, Waples RS. 2011. Interacting effects of phenotypic plasticity and evolution on population persistence in a changing climate. Conservation biology : the journal of the Society for Conservation Biology 25: 56–63.

Reen FJ, Romano S, Dobson ADW, O’Gara F. 2015. The sound of silence: Activating silent biosynthetic gene clusters in marine microorganisms. Marine Drugs 13: 4754–

4783.

Rees T. 1994. Plant physiology. Virtue on both sides. Current biology 4: 557–9.

Renner T, Specht CD. 2013. Inside the trap: gland morphologies, digestive enzymes, and the evolution of plant carnivory in the Caryophyllales. Current opinion in plant biology 16: 436–42.

Reznick D, Bryant MJ, Bashey F. 2008. r- and K-Selection Revisited : The Role of

Population Regulation in Life-History Evolution Published by : Ecological Society of

America Stable URL : http://www.jstor.org/stable/3071970. America 83: 1509–1520.

Riedel M, Eichner A, Jetter R. 2003. Slippery surfaces of carnivorous plants: composition of epicuticular wax crystals in Nepenthes alata Blanco pitchers. Planta 218:

87–97.

249 Rietkerk M, Dekker S, Wassen M, Verkroost A, MF B. 2004. A putative mechanism for bog patterning. The American Naturalist 5: 699–708.

Rischer H, Hamm A, Bringmann G. 2002. Nepenthes insignis uses a C2-portion of the carbon skeleton of L-alanine acquired via its carnivorous organs, to build up the allelochemical plumbagin. Phytochemistry 59: 603–9.

Roberts MF, Wink M. 1998. Alkaloids : Biochemistry, Ecology, and Medicinal

Applications. Springer US.

Robroek BJM, Smart RP, Holden J. 2010. Sensitivity of blanket peat vegetation and hydrochemistry to local disturbances. Science of The Total Environment 408: 5028–

5034.

Rodwell JS. 1991. British plant communities volume 2: mires and heaths. Cambridge

University Press, Cambridge, UK.

Roff DA, Fairbairn DJ. 2007. The evolution of trade-offs: where are we? Journal of

Evolutionary Biology 20: 433–447.

Roy R, Schmitt AJ, Thomas JB, Carter CJ. 2017. Review: Nectar biology: From molecules to ecosystems. Plant Science 262: 148–164.

RStudio Team. 2016. RStudio: Integrated development for R.

Rydin H, Jeglum J. 2006. The biology of peatlands. Oxford: Oxford University Press.

Rydin H, Jeglum JK, Bennett KD. 2013. The biology of peatlands. Oxford University

Press.

Schaefer HM, Ruxton GD. 2008. Fatal attraction: carnivorous plants roll out the red carpet to lure insects. Biology Letters 4: 153–155.

Schlauer J, Nerz J, Rischer H. 2005. Carnivorous plant chemistry. Acta Botanica

Gallica 152: 187–195.

Schlichting CD. 2002. Phenotypic plasticity in plants. Plant Species Biology 17: 85–88.

250 Schöner CR, Schöner MG, Grafe TU, Clarke CM, Dombrowski L, Tan MC, Kerth

G. 2017. Ecological outsourcing: a pitcher plant benefits from transferring pre- digestion of prey to a bat mutualist (S Bonser, Ed.). Journal of Ecology 105: 400–411.

Schulze W, Frommer WB, Ward JM. 1999. Transporters for ammonium, amino acids and peptides are expressed in pitchers of the carnivorous plant Nepenthes. The Plant

Journal 17: 637–646.

Schulze E-D, Gebauer G, Schulze W, Pate JS. 1991. The utilization of nitrogen from insect capture by different growth forms of Drosera from Southwest Australia.

Oecologia 87: 240–246.

Schulze W, Schulze E-D. 1990. Insect capture and growth of the insectivorous

Drosera rotundifolia L. Oecologia 82: 427–429.

Schulze W, Schulze ED, Schulze I, Oren R. 2001. Quantification of insect nitrogen utilization by the Venus flytrap Dionaea muscipula catching prey with highly variable isotope signatures. Journal of experimental botany 52: 1041–1049.

Schwab W, Davidovich-Rikanati R, Lewinsohn E. 2008. Biosynthesis of plant- derived flavor compounds. The Plant Journal 54: 712–732.

Seddon AWR, Macias-Fauria M, Long PR, Benz D, Willis KJ. 2016. Sensitivity of global terrestrial ecosystems to climate variability. Nature 531: 229–232.

Shearer G, Kohl DH. 1989. Estimates of N2 Fixation in Ecosystems: The Need for and Basis of the 15N Natural Abundance Method. Springer, New York, NY, 342–374.

Sheridan PM, Griesbach RJ. 2001. Anthocyanidins of Sarracenia L. flowers and leaves. HortScience 36: 384.

Shipley B, Lechowicz MJ, Wright I, Reich PB. 2006. Fundamental trade-offs generating the worldwide leaf economics spectrum. Ecology 87: 535–41.

Simoneit BRT, Medeiros PM, Wollenweber E. 2008. Triterpenoids as major components of the insect-trapping glue of Roridula species. Zeitschrift fur

251 Naturforschung - Section C Journal of Biosciences 63: 625–630.

Slack A. 2000. Carnivorous Plants. MIT Press. de Smedt P, Ottaviani G, Wardell-Johnson G, Sýkora K V, Mucina L. 2018. Habitat heterogeneity promotes intraspecific trait variability of shrub species in Australian granite inselbergs. Folia Geobotanica: 1–13.

Smith RS, Charman D, Rushton SP, Sanderson RA, Simkin JM, Shiel RS. 2003.

Vegetation change in an ombrotrophic mire in northern England after excluding sheep. Applied Vegetation Science 6: 261–270.

Statzner B, Moss B. 2004. Linking ecological function, biodiversity and habitat: a mini-review focusing on older ecological literature. Basic and Applied Ecology 5: 97–

106.

Stevenson PC, Nicolson SW, Wright GA. 2017. Plant secondary metabolites in nectar: impacts on pollinators and ecological functions (J Manson, Ed.). Functional

Ecology 31: 65–75.

Stewart CN, Nilsen ET. 1992. Drosera rotundifolia growth and nutrition in a natural population with special reference to the significance of insectivory. Canadian Journal of Botany 70: 1409–1416.

Stuart Chapin F, Matson PA, Vitousek PM. 2012. Principles of terrestrial ecosystem ecology. New York, NY: Springer New York.

Suárez-Piña J, Rueda-Almazán JE, Ayestarán LM, Alcalá RE. 2016. Effect of light environment on intra-specific variation in herbivory in the carnivorous plant

Pinguicula moranensis (Lentibulariaceae). Journal of Plant Interactions 11: 146–151.

SUAS. 2017. Climate data from Kindla II weather monitoring station from 1996-2017.

Sutherland WJ, Freckleton RP, Godfray HCJ, Beissinger SR, Benton T, Cameron

DD, Carmel Y, Coomes D a., Coulson T, Emmerson MC, et al. 2013. Identification of 100 fundamental ecological questions. Journal of Ecology 101: 58–67.

252 Tamminen M, Betz A, Pereira AL, Thali M, Matthews B, Suter MJ-F, Narwani A.

2018. Proteome evolution under non-substitutable resource limitation. Nature

Communications 9: 4650.

Tanaka A, Makino A. 2009. Photosynthetic Research in Plant Science. Plant and Cell

Physiology 50: 681–683.

Tanaka Y, Sasaki N, Ohmiya A. 2008. Biosynthesis of plant pigments:

Anthocyanins, betalains and carotenoids. Plant Journal 54: 733–749.

Theis N, Lerdau M. 2003. The Evolution of Function in Plant Secondary Metabolites.

International Journal of Plant Sciences 164: S93–S102.

Thompson K. 1987. The Resource Ratio Hypothesis and the Meaning of

Competition. Functional Ecology 1: 297.

Thorén LM, Karlsson PS. 1998. Effects of supplementary feeding on growth and reproduction of three carnivorous plant species in a subarctic environment. Journal of

Ecology 86: 501–510.

Thorén LM, Tuomi J, Kämäräinen T, Laine K. 2003. Resource availability affects investment in carnivory in Drosera rotundifolia. New Phytologist 159: 507–511.

Thorogood CJ, Bauer U, Hiscock SJ. 2017. Convergent and divergent evolution in carnivorous pitcher plant traps. New Phytologist 217: 1035–1041.

Thum M. 1989. The significance of carnivory for the fitness of Drosera in its natural habitat. Oecologia 81: 401–411.

Tilman D. 1982. Resource competition and community structure. Princeton University

Press.

Tilman D. 1985. The Resource-Ratio Hypothesis of Plant Succession. The American

Naturalist 125: 827–852.

Towett EK, Shepherd KD, Lee Drake B. 2016. Plant elemental composition and portable X-ray fluorescence (pXRF) spectroscopy: Quantification under different

253 analytical parameters. X-Ray Spectrometry 45: 117–124.

Townsend CR, Begon M, Harper JL. 2003. Essentials of ecology. Essentials of ecology.

Trenberth K. 2011. Changes in precipitation with climate change. Climate Research

47: 123–138.

Turetsky MR, St. Louis V. 2006. Disturbance in Boreal Peatlands. Boreal Peatland

Ecosystems. Springer Berlin Heidelberg, 359–379.

Tušek M, Curman M, Babić M, Tkalec M. 2016. Photochemical efficiency, content of photosynthetic pigments and phenolic compounds in different pitcher parts of

Sarracenia hybrids. Acta Botanica Croatica 75: 179–185.

US EPA. 2019. Technical Overview of Volatile Organic Compounds.

Vaistij FE, Lim EK, Edwards R, Bowles DJ. 2009. Glycosylation of secondary metabolites and xenobiotics. Plant-derived Natural Products: Synthesis, Function, and Application. New York, NY: Springer US, 209–228.

Viant MR, Kurland IJ, Jones MR, Dunn WB. 2017. How close are we to complete annotation of metabolomes? Current Opinion in Chemical Biology 36: 64–69.

Violle C, Navas ML, Vile D, Kazakou E, Fortunel C, Hummel I, Garnier E. 2007.

Let the concept of trait be functional! Oikos 116: 882–892.

Wagner A. 2017. The White-Knight Hypothesis, or Does the Environment Limit

Innovations? Trends in ecology & evolution 32: 131–140.

Wakefield AE, Gotelli NJ, Wittman SE, Ellison AM. 2005. Prey addition alters nutrient stoichiometry of the carnivorous plant Sarracenia purpurea. Ecology 86: 1737–

1743.

Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin JM,

Hoegh-Guldberg O, Bairlein F. 2002. Ecological responses to recent climate change.

Nature 416: 389–395.

Wasternack C, Hause B. 2013. Jasmonates: biosynthesis, perception, signal

254 transduction and action in plant stress response, growth and development. An update to the 2007 review in Annals of Botany. Annals of Botany 111: 1021–1058.

Watson AP, Matthiessen JN, Springett BP. 1982. Arthropod associates and macronutrient status of the red-ink sundew (Drosera erythrorhiza Lindl.). Australian

Journal of Ecology 7: 13–22.

Weber RJM, Viant MR. 2010. MI-Pack: Increased confidence of metabolite identification in mass spectra by integrating accurate masses and metabolic pathways. Chemometrics and Intelligent Laboratory Systems 104: 75–82.

Westland S, Ripamonti C, Cheung V. 2012. CIELAB and Colour Difference.

Computational Colour Science using MATLAB®: 49–74.

Van de Weyer A-L, Bemm F, Weber AP, Hedrich R, Larisch C, Schulze WX,

Kreuzer I, Ankenbrand M, Schultz J, Al-Rasheid KA, et al. 2016. Venus flytrap carnivorous lifestyle builds on herbivore defense strategies. Genome Research 26: 812–

825.

Whittaker RH, Levin SA, Root RB. 1973. Niche, Habitat, and Ecotope. The American

Naturalist 107: 321–338.

Wilson JB, Lee WG. 2000. C-S-R triangle theory: Community-level predictions, tests, evaluation of criticisms, and relation to other theories. Oikos 91: 77–96.

Wink M. 2003. Evolution of secondary metabolites from an ecological and molecular phylogenetic perspective. Phytochemistry 64: 3–19.

Wollenweber E. 2007. Flavonoids occurring in the sticky resin on Roridula dentata and Roridula gorgonias (Roridulaceae). Carnivorous plant newsletter. 36: 77–80.

Wong T-S, Kang SH, Tang SKY, Smythe EJ, Hatton BD, Grinthal A, Aizenberg J.

2011. Bioinspired self-repairing slippery surfaces with pressure-stable omniphobicity. Nature 477: 443–7.

Yang Z, Midmore DJ. 2005. Modelling plant resource allocation and growth

255 partitioning in response to environmental heterogeneity. Ecological Modelling 181: 59–

77.

Yao P. 2012. The advantages of RAW format image post-processing.

Communications in Computer and Information Science. Springer, Berlin,

Heidelberg, 625–631.

Yesmin L, Gammack SM, Cresser MS. 1996. Effects of atmospheric nitrogen deposition on ericoid mycorrhizal infection of Calluna vulgaris growing in peat soils.

Applied Soil Ecology 4: 49–60.

Yilamujiang A, Reichelt M, Mithöfer A. 2016. Slow food: Insect prey and chitin induce phytohormone accumulation and gene expression in carnivorous Nepenthes plants. Annals of Botany 118: 369–375.

Zamora R. 1995. The Trapping Success of a Carnivorous Plant, Pinguicula vallisneriifolia - the Cumulative Effects of Availability, Attraction, Retention and

Robbery of Prey. Oikos 73: 309–322.

Zamora R, Gómez JM, Hódar JA. 1998. Fitness responses of a carnivorous plant in contrasting ecological scenarios. Ecology 79: 1630–1644.

Zenk MH, Fürbringer M, Steglich W. 1969. Occurrence and distribution of 7- methyljuglone and plumbagin in the Droseraceae. Phytochemistry 8: 2199–2200.

Zukunft S, Prehn C, Röhring C, Möller G, Hrabě de Angelis M, Adamski J,

Tokarz J. 2017. High-throughput extraction and quantification method for targeted metabolomics in murine tissues. Metabolomics 14: 18.

Zuppinger-Dingley D, Schmid B, Petermann JS, Yadav V, De Deyn GB, Flynn

DFB. 2014. Selection for niche differentiation in plant communities increases biodiversity effects. Nature 515.

256