Ultrastructure, transport and the state of phosphorus in Pisolithus tinctorius

David Allan Orlovich Doctor of Philosophy

The University of New South Wales

1994 UNIVERSITY Or N.S.VV.

1 6 J'.'-j \TA LI BR/' TIES Contents

Abstract...... i Acknowledgements...... ii Certificate of originality...... iii

1 Phosphorus and mycorrhizas Introduction...... 1 Phosphorus nutrition and mycorrhizas...... 1 Phosphorus uptake...... 2 Phosphorus storage...... 2 Phosphorus transport...... 4 Aim...... 5

2 Polyphosphate granules are an artefact of specimen preparation in Pisolithus tinctorius Introduction...... 6 Materials and methods...... 7 Treatment of hyphae to precipitate metachromatic granules...... 7 Structure, cytochemistry and X-ray microanalysis of granules in chemically fixed hyphae...... 7 Freeze-substitution, ion localisation and X-ray microanalysis...... 8 Results...... 9 Precipitation of metachromatic granules during conventional specimen preparation...... 9 Evidence that metachromatic granules contain polyphosphate...... 10 Comparison of content after conventional procedures and freeze-substitution...... 10 Discussion...... 11 Polyphosphate granules and specimen preparation...... 11 Cation exchange with polyphosphate in vivo...... 12 Significance in transport...... 13

3 Ultrastructure of the vacuole system in Pisolithus tinctorius Introduction...... 23 Materials and Methods...... 24 Rapid freezing and freeze-substitution...... 24 Measurements and statistics...... 24 Results...... 25 General ultrastructure of the tip region and and tubules in cells close to the hyphal tip...... 25 Ultrastructure of cells behind the growing tip...... 27 Discussion...... 28 Clusters and tubules...... 28 Fine filaments...... 30 Golgi bodies and smooth membrane cistemae...... 31 Mature hyphae...... 34

4 Structure and development of the dolipore septum in Pisolithus tinctorius Introduction...... 45 Materials and Methods...... 46 Results...... 47 Clamp connections and dolipore septa...... 47 Early development of the dolipore septum...... 47 The pore complex of mature septa...... 48 Discussion...... 50 General considerations...... 50 Membrane furrowing and the filamentous ring...... 51 Structure of the septal pore complex at maturity...... 53 Transport and communication through the dolipore...... 54

5 Observations on ion redistribution within a specimen during freeze-substitution, embedding and X-ray microanalysis Introduction...... 62 The application of freeze-substitution...... 62 General procedure for freeze-substitution...... 62 Theoretical aspects of freeze-substitution for ion localisation...... 63 Limitations of X-ray microanalysis...... 64 Experimental rationale...... 65 Materials and Methods...... 65 Sandwiched droplets...... 65 Solutions frozen in contact with each other...... 66 Freeze-dried droplets...... 66 Freeze-substitution...... 66 X-ray microanalysis...... 67 Results...... 68 Ultrastructure of freeze-substituted solutions...... 68 Elemental distribution in solutions sandwiched together after freezing...... 68 Elemental distribution in solutions frozen in contact with each other...... 69 Elemental distribution in freeze-dried droplets...... 70 Discussion...... 70 Retention of water soluble ions by freeze-substitution...... 70 Interpretation of X-ray microanalysis results...... 71 Beam damage...... 71 Contamination of sections with atmospheric water...... 72

6 Symplastic transport in Pisolithus tinctorius Introduction...... 88 Symplastic translocation in fungi...... 88 Compartmentation of P, into vacuoles in hyphal tips...... 89 Storage and translocation in large vacuoles...... 89 Translocation in mycorrhizas...... 90 Mechanism of translocation to the Hartig net...... 91

7 References...... 93

8 Appendix...... 105 Abstract

The ultrastructure of hyphae was studied in the ectomycorrhizal Pisolithus tinctorius (Pers.) Coker & Couch. Polyphosphate granules were found to be artefacts of conventional specimen preparation. Granules were not produced by glutaraldehyde fixation but appeared at early stages of ethanol dehydration. In contrast, vacuoles of freeze-substituted hyphae contained evenly-dispersed material; phosphorus and potassium were distributed more or less uniformly throughout. Vacuoles formed clusters in the cortex of the and were located at intervals along the length of the tip . Elongate vacuoles spanned those regions of the cytoplasm which did not contain clusters. identified as Golgi bodies had a single, occasionally fenestrated, confluent with a peripheral tubular reticulum. Smooth membrane cisternae with electron-lucent content occasionally formed ring-like structures which have been described as Golgi bodies in other fungi. In some mature cells, the cytoplasm contained homogeneously dispersed fibrillar material. The structure and development of dolipore septa and associated clamp connections are described. Septa were produced by a furrowing of the plasma membrane and concurrent wall deposition. At maturity, the pore is surrounded by a dome-shaped, perforate parenthesome on either side and is filled with filamentous electron-opaque material. Continuities between the lumen of this rough (ER) and the parenthesome were not found. The evidence indicates that the rim of the parenthesome is anchored to the plasma membrane rather than adjacent ER. Ion redistribution during freeze-substitution and X-ray microanalysis was tested using water-soluble salts solutions containing dextran. Phosphorus, potassium, calcium and sodium were all successfully localised although beam damage and contamination of sections by atmospheric water occasionally caused redistribution of sodium. The implications of the results for nutrient transfer in mycorrhizal associations are discussed. It is suggested that nutrients including phosphorus and potassium are supplied to the plant at the fungus-host interface by transfer across the sheath through interconnected vacuoles from the sites of uptake in the soil mycelium. Acknowledgements

Thankyou very much to my supervisor Anne Ashford, co-supervisor Guy Cox, Suzanne Bullock for preparing the plates, Bill Allaway, Larry Peterson, Margaret McCully, Mum, Dad, Susan, Matthew, Nicola Young, Anika Mostaert, Bettye Rees, Eleni Taylor-Wood, Lydia Kupsky, Virginia Shepherd, Charles Morris, Melissa Fitzgerald, Greg Howell, Marita Sydes, Patrick Driver, Kim Plummer, Jamie Solo, Martin Canny, Robyn Murphy, Fred Rost, Mel Dickson, Patrick Marks, Chris Martinic, Ian Kaplin, Dennis Dwarte, Thor Bostrom, Andrew Moore and Lewis Melville. Thankyou also to Andrew Drinnan, Pauline Ladiges, Anthony Vadala, Andrew Rozefelds, Kirsten Parris, Peter Neish, Marco Duretto, Michael Bayly, Frank Udovicic, Nicole Middleton, Indira Narayan, Arlene McDowell, Ros Gleadow, Hugh Campbell and Jocelyn Carpenter for support and encouragement during 1993. I am grateful for the award of an Australian Postgraduate Research Award. Certificate of originality

I hereby declare that this submission is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another person nor material which to a substantial extent has been accepted for the award of any other degree or diploma of a university or other institute of higher learning, except where due acknowledgement is made in the text. Phosphorus and mycorrhizas

Introduction The level of total phosphorus is low in many Australian soils, being about 0.03 - 0.06% P compared with up to 0.1% in some American soils (Wild, 1958; Groves and Keraitis, 1976; Taylor, 1983). Around plant roots actively absorbing phosphorus, there is a zone of phosphorus depletion (Bieleski, 1973) governed by the low mobility of phosphorus in the soil. In these circumstances, uptake of phosphorus by the plant is likely to be limited by the rate of diffusion of phosphorus through the soil rather than the actual uptake properties of the root surface (Bowen and Theodorou, 1967). The root must either outgrow the zone of depletion or more efficiently exploit the same volume of soil. By increasing the available surface area of the root/soil interface and extending the uptake surface beyond the zone of depletion, fungal hyphae may be able to obtain phosphorus which may otherwise have been unavailable to the non-mycorrhizal plant root (Harley and Smith, 1983). Therefore, mycorrhizas are thought to play a particular role in phosphorus nutrition of plants.

Phosphorus nutrition and mycorrhizas Many studies have shown that inorganic phosphate (P,-) is taken up from the soil in greater amounts by mycorrhizal plants than non-mycorrhizal plants. Malajczuk et al. (1975) found that mycorrhizal Eucalyptus calophylla R. Br. absorbed more phosphorus from a soil low in P,• than did non-mycorrhizal E. calophylla. Mycorrhizal Pinus radiata D. Don. seedlings were found by Bowen (1969) to take up 2.5 times as much P/ as non- infected seedlings. Harley and McCready (1950) observed a higher rate of P, uptake in mycorrhizal Fagus sylvatica L. roots than in non-mycorrhizal roots. Mitchell and Read (1981) compared the growth (dry weight increase) of mycorrhizal and non-mycorrhizal cranberry (Vaccinium macrocarpon Ait.) and concluded that the mycorrhizal association was beneficial to the plant. Alnus viridis (Chaix) D. C. mycorrhizas were shown by Mejstrik and Benecke (1969) to absorb phosphorus up to five times faster than non- mycorrhizal roots. This and other evidence, convincingly shows that mycorrhizas are very important for P nutrition in a great variety of plants. It is not therefore surprising that mycorrhizas are found on almost all vascular plants. The distribution of mycorrhizas among flowering plants and gymnosperms has been compiled from various sources by Harley and Smith (1983) and Harley and Harley (1987). Vesicular-arbuscular mycorrhizas, ie: those in which hyphae penetrate the root cells and form structures known as vesicles and arbuscules, account for most of the mycorrhizal associations,

1 Phosphorus and mycorrhizas 2 however ectomycorrhizas are found in members of at least 43 families including 140 genera. Ectomycorrhizal plants of economic importance include members of the genera Vicia, Eucalyptus, Pinus, Populus and Prunus (Harley and Smith, 1983).

Phosphorus uptake Pisolithus tinctorius (Pers.) Coker & Couch is a cosmopolitan ectomycorrhizal fungus which is known to form an association with a wide variety of plants, including members of the genus Eucalyptus (Ashford et al., 1975, 1986). Ectomycorrhizas are characterised by a sheath of hyphae which completely encircles some lateral roots of the host. From this sheath, hyphae may grow out into the soil and also inwards, between the epidermal and cortical root cells to form what is called the Hartig net (Harley and Smith, 1983). The distribution of P; between the fungal sheath and host was studied by Harley and McCready (1952a,b). They found that in both excised and attached mycorrhizal root tips of beech exposed to labelled inorganic phosphate (32P/), about 90% of the accumulated phosphate was found in the fungal sheath which surrounds the plant root, regardless of the time of exposure to 32P,. Labelled inorganic phosphate was found in the fungal and host tissue by Harley and Loughman (1963) after only thirty seconds exposure to 32P/. After a few minutes, labelled organic phosphate was found in both the fungus and the host. Apparently the phosphorus is passed to the plant as P/ and not as organic P. The cytoplasmic Pz is buffered by the vacuolar P, and is thus kept at a relatively constant level. At the interface, the host cytoplasm and the fungal vacuole compete for the fungal cytoplasmic P, pool. Photosynthate is transferred from the plant to the fungus (Cairney et al., 1989) and is converted to mannitol and trehalose by the fungus (see Bieleski, 1973). Since the plant cannot use these compounds, it is hypothesised that a sink is established.

Phosphorus storage Some of the phosphate that enters fungal cells is converted to polyphosphate. The state of polyphosphate in P. tinctorius is investigated in Chapter 2. Polyphosphate is one of a group of phosphorus compounds known as condensed phosphates (Thilo, 1962). They can be formed by the elimination of water from a solution of phosphate ions, causing them to condense and become linked by oxygen bridges. The condensed phosphates thus formed may be cyclic, linear or cross-linked. The chemistry of these is reviewed by Thilo (1962), Harold (1966) and Dawes and Senior (1973). Polyphosphates are linear condensed phosphates. Polyphosphate production in microorganisms is thought to be catalysed by the enzymes: polyphosphate kinase, polyphosphate synthetase and polyphosphate glucokinase (Harold, 1966; Durr et al., 1979). Phosphorus and mycorrhizas 3

Polyphosphates are unbranched and conform to the general formula Mn+2Pn03n+i (Harold, 1966) where M is a metal atom. The chain length can vary from two P/ units (pyrophosphate), to approximately 104 units which is the average chain length of mixtures known as Kurrol's and Maddrell's salts (Harold, 1966). Polyphosphate is stable in solution at neutral pH and at room temperature. As temperature rises and pH falls, the rate of hydrolysis increases (Thilo, 1962). Polyphosphate is stable to alkali but the presence of divalent cations in alkaline solutions catalyzes degradation (Harold, 1966). Upon heating with organic solvents, polyphosphate is degraded to trimetaphosphate. This reaction is also catalysed by the addition of cations, especially Mg2+ (Harold, 1966). Evidence supporting the existence of linear polyphosphate in cells has been gained from three lines of research: (i) extraction of polyphosphate, (ii) nuclear magnetic resonance (NMR) and (iii) in situ cytochemistry. Callow et al. (1978) extracted polyphosphates from onion roots infected with Glomus mosseae (Nicol. & Gerd.) Gerdemann and Trappe using phenol-detergent extraction and separated them from nucleic acids by polyacrylamide gel electrophoresis (PAGE). No polyphosphate was detected in extracts from non-mycorrhizal roots. Infected roots which were given nutrient solution contained more polyphosphate than infected roots not given nutrients. From comparison with polyphosphate standards, they determined that the chain length of the extracted polyphosphate was no greater than 200 P, residues. Polyphosphate with an average chain length of 15 P, subunits has been extracted from P. tinctorius (Ashford et al., 1994). 31P NMR has been used to investigate the state and metabolism of polyphosphate in ectomycorrhizal fungi (Martin et al., 1985; Grellier et al., 1989; Ashford et al., 1994). The average polyphosphate chain length in Cenococcum graniforme Ferd. & Wing, and Hebeloma crustuliniforme (Bull.) Quel, was 11 phosphorus residues (molecular mass of 1 x 103) (Martin et al., 1985). They found that in both species, the level of cellular polyphosphate was lowest in rapidly growing mycelia and highest in cultures that were old or starved of nitrogen. When cultures were starved of phosphorus, polyphosphate was rapidly degraded in H. crustuliniforme but degradation was slower in C. graniforme. These authors suggest that the polyphosphate may be in a different physical state in each species. Martin et al. (1985) list three possible states of polyphosphate: (a) as freely mobile polymers, (b) as macromolecular aggregates and (c) as completely insoluble precipitates. They point out that NMR would not detect insoluble precipitates because the P would not resonate. It is thought that this form of polyphosphate corresponds to an ‘acid-insoluble’ polyphosphate pool. When basic polypeptides like polylysine and polyarginine were added to a solution of polyphosphate, a precipitate formed which could not be detected by NMR. These polymers have a molecular mass of about 20 kDa. Martin et al. (1985) also added Mn2+ ions to solutions of polyphosphate and found that Phosphorus and mycorrhizas 4 as concentration of Mn2+ increased, the relaxation time of the polyphosphate solution decreased. A short relaxation time means that the polyphosphate is more closely associated with other ions and therefore does not resonate to the same degree. A long relaxation time means that the polyphosphate is more freely mobile in water. NMR of intact mycelia showed that the relaxation time for the inner P of polyphosphate was quite low. It is unlikely that there will be a high concentration of Mn2+ at the site of polyphosphate accumulation in fungal vacuoles to account for the reduced relaxation time and that this (or other paramagnetic ions: Cu2+, Fe2+, Co2+) is the sole ion associated with polyphosphate. Rather, Martin et al. (1985) suggest that the polyphosphate may be interacting with other macromolecules and/or other polyphosphate molecules (as well as some metal ion) as macromolecular aggregates. The state of polyphosphate in microorganisms is poorly understood because of difficulty correlating often conflicting data from various experiments. The presence of polyphosphate granules in most microorganisms (Keck and Stich, 1957) has led to many studies on the size, composition and dynamics of these (usually) vacuolar bodies (see Chapter 2) and to hypotheses about their significance in phosphorus nutrition in mycorrhizas. Data presented in this thesis indicate that the granules are an artefact of specimen preparation and are not present in living hyphae.

Phosphorus transport Phosphorus translocation in mycorrhizal fungi is discussed in detail by Harley and Smith (1983) and Bolan (1991). If the hypothesis is accepted that mycorrhizas enhance the phosphorus nutrition of the host plant by improving phosphorus nutrition, it is obvious that phosphorus stored by the fungal component must be translocated to the host via the fungus. This is particularly clear in ectomycorrhizas where the fungal sheath completely surrounds the region of uptake of the host root: thus there is no direct pathway from the soil to the root tip surface. Transport of nutrients through soil mycelium is likely to have both a symplastic and an apoplastic component (Cairney, 1992). Symplastic transport is thought to occur along hyphae by diffusion, cytoplasmic streaming or osmotically generated mass flow (Harley and Smith, 1983; Jennings, 1987, 1989; Thomson et al., 1987). The dolipore septum of basidiomycete fungi would present a barrier to such intercellular transport in these fungi for some molecules. It has recently been shown by V. A. Shepherd (Shepherd et al., 1993a,b) that young hyphae of P. tinctorius contain a pleiomorphic system of vacuoles and tubules which accumulates the vital fluorochrome 6-carboxyfluorescein. The system is capable of intra- and intercellular transport of its contents in tip cells. This is direct evidence from living cells of an alternative mechanism for symplastic transport through fungal hyphae. A similar system also occurs in other fungi and oomycetes (Rees et al., 1994). Phosphorus and mycorrhizas 5

Aim The aim of the experiments described in this thesis was to investigate the vacuole system and state of polyphosphate in P. tinctorius. It is shown in Chapter 2 that vacuoles are the major site of phosphorus accumulation and that this is in the form of soluble polyphosphate stored with potassium, not as calcium polyphosphate granules as previously thought (Ashford et al., 1986). A paper resulting from this work (Orlovich and Ashford, 1993) and two papers containing preliminary results not described in this thesis (Orlovich et al., 1989, 1990) are in the Appendix. The ultrastructure of the vacuole system of P. tinctorius is shown in Chapter 3. The ultrastructure correlates with data from living hyphae shown in Chapter 2 and by V. A. Shepherd showing that the vacuoles and tubules form an interconnected continuum in hyphal tips. Papers from this work (Shepherd et al., 1993a,b) are included in the Appendix. The ultrastructure of Golgi bodies and more mature hyphae is also investigated. The dolipore septum presents a substantial barrier to intercellular transport in basidiomycete fungi. However, it is shown using light microscopy in Chapter 2 and using fluorochrome tracers (Shepherd et al., 1993b) that tubules can cross the septum. The structure and development of the dolipore septum is shown in Chapter 4. A paper arising from this work (Orlovich and Ashford, 1994) is in the Appendix. The technique of using freeze-substitution to localise water soluble ions was tested in experiments described in Chapter 5. This technique was used to localise phosphorus and potassium in fungal cells in Chapter 2 and has revealed different results from those obtained after chemical fixation. A model system of salt solutions containing salts and a macromolecule was used as test specimens. The implications of the results of experiments on P. tinctorius on ideas concerning the translocation of phosphorus in mycorrhizal fungi are discussed in Chapter 6. Polyphosphate granules are an artefact of specimen preparation in Pisolithus tinctorius

Introduction Polyphosphate, a linear condensed polymer of inorganic phosphate (P/), occurs in most microorganisms (for example see review Kulaev and Vagabov, 1983) and some higher plants (see Bieleski, 1973). It has been found in all major groups of fungi (Chilvers et al., 1985). Mycorrhizal fungi are thought to enhance the phosphorus content of the host plant and polyphosphate accumulation in the fungus is likely to play an important part in this process (see Harley and Smith, 1983). Most biochemical work on fungi has been done with yeasts, where it is known that polyphosphate is a vacuolar store of P, involved in amino acid and cation retention and osmoregulation (reviewed by Klionsky et al., 1990). Polyphosphate has been identified in extracts and intact fungi, using 31P nuclear magnetic resonance (NMR) spectroscopy and a number of other biochemical techniques (Callow et al, 1978; Martin et al., 1983; Straker and Mitchell, 1985). At least some of the polyphosphate was thought to exist as vacuolar granules that exhibit y-metachromasy when stained with dyes such as toluidine blue O at low pH and are extracted by cold acid (Keck and Stich, 1957). Such granules have been localised in vacuoles of mycorrhizal fungi using cytochemical techniques following aldehyde or ethanol fixation. They are identified also as electron- opaque deposits that volatilise under the electron beam and show high phosphorus peaks with X-ray microanalysis (see Ashford et al., 1986). However, there is some evidence from the literature that polyphosphate granules may be artefacts of specimen preparation (Delaporte, 1939; Nakai, 1976). This has been largely ignored and estimates of numbers of granules have been used as a measure of the amount of polyphosphate in many systems (Cox et al., 1980; Lapeyrie et al., 1984). Polyphosphate granules have been shown in the vacuoles of Pisolithus tinctorius (Ashford et al., 1986) and polyphosphate with a chain length of about 15 P, subunits has been characterised by 31P NMR and other biochemical methods (Ashford et al. , 1994). Vacuoles in the terminal and subapical cells of growing hyphal tips have been shown to be continuous with a motile tubule system that can transport material by peristalsis along and between cells (Shepherd et al., 1993a, b). In more mature regions, this develops into a system of larger, less mobile vacuoles that are also interconnected. Here it is shown that these vacuoles contain polyphosphate, but not as metachromatic granules. Granules were formed during various treatments, including conventional specimen preparation involving aldehyde

6 Polyphosphate in Pisolithus tinctorius 7

fixation and ethanol dehydration. Polyphosphate occured in soluble form, dispersed throughout the vacuole in association with potassium in living hyphae. This offers the possibility that phosphorus may be transferred along hyphae as polyphosphate in soluble form within the vacuole system.

Materials and methods

Treatment of hyphae to precipitate metachromatic granules P. tinctorius (isolate DI-15, see Grenville et al., 1986) was grown at 21°C in the dark on modified Melin Norkrans (MMN) agar medium (Marx, 1969) with the following variations: 1.0% agar, 1.0% D-glucose instead of sucrose, light dried malt extract instead of paste and 3.0 mL.L'1 of 1.0% ferric citrate instead of FeCl3. The composition of the liquid medium was the same as MMN agar medium except that agar and malt extract were omitted and trace elements were added. Pieces of agar with mycelium (2 mm x 2 mm) were cut from the growing edge of 3-4 week old cultures and mounted for light microscopy in 2 drops of MMN liquid medium, pH 6.2, on a slide using a coverslip as a spacer. After about 20 min, when the fungus had resumed cytoplasmic streaming, the hyphae were subjected to a number of treatments, while individual hyphae were kept under continuous observation, using Nomarski differential interference contrast (DIC) optics on a Zeiss Universal light microscope. Photomicrographs were taken with Kodak Technical Pan film rated at 50 ASA. Each solution was added drop by drop to the slide and then drawn under the coverslip with a piece of absorbent tissue paper for a minimum of 10 min. Each treatment was duplicated. The treatments were as follows: (i) glutaraldehyde (2.5%) in MMN, followed by successive dehydration in ethanol (5, 10, 15, 20, 30, 40, 50, 60, 70, 80, 95, 100%), and then staining in 0.05% toluidine blue O in 0.05 mol.L'1 KC1/HC1 buffer, pH 1.0. (ii) glutaraldehyde (2.5%) in MMN followed directly by staining with 0.05% toluidine blue O at pH 1.0, in 0.05 mol.L'1 KC1/HC1 buffer, or (iii) fixation as in (ii) followed by staining with 0.05% toluidine blue O in MMN, pH 6.2. (iv) treatment of fresh hyphae with 0.05% toluidine blue O, pH 1.0. (v) treatment of fresh hyphae with 30%, 50% or 100% ethanol followed by staining with 0.05% toluidine blue O, pH 1.0. Staining with toluidine blue O at pH 1.0 is more specific, since low pH depresses the ionisation of many groups with the potential to stain (see Ashford et al., 1975).

Structure, cytochemistry and X-ray microanalysis of granules in chemically fixed hyphae Mycelium was fixed with 3% glutaraldehyde in 0.3 mol.L'1 NaOH-PIPES buffer (Salema and Brandao, 1973), pH 7.5 for 4 h at 4°C, rinsed for 10 min in the same Polyphosphate in Pisolithus tinctorius 8

buffer, dehydrated to 100% ethanol @ 20 min per 10% step at 4°C and embedded in glycol methacrylate (GMA). Sections (1 pm) were mounted on poly-L-lysine-coated slides (van Noorden and Polak, 1983) and stained with 0.05% toluidine blue O pH 1.0 (0.05 mol.L'1 KC1/HC1 buffer) and photographed. The mountant was removed, the sections were incubated with 10% trichloroacetic acid at 4°C for 6 h (Keck and Stich, 1957), restained with toluidine blue O, pH 1.0 and then rephotographed (n=2). Some unstained sections were cut at 1 pm, collected on formvar and carbon coated nickel slot grids, and analysed by X-ray microanalysis as for freeze-substituted material. For ultrastructure, material was fixed and dehydrated as for light microscopy, but embedded in Spurr's resin (Spurr, 1969). Ultrathin sections were triple stained (Daddow, 1983) with undiluted lead citrate (Reynolds, 1963) for 5 min, 4% uranyl acetate in water for 20 min and lead citrate for 40 min. Sections were examined with an Hitachi H-7000 transmission electron microscope at 100 kV.

Freeze-substitution, ion localisation and X-ray microanalysis Mycelium was grown over discs of Nuclepore brand ‘Membra-fil’ gridded membrane filter (8.0 pm pore size) placed on the agar surface ahead of the growing hyphal front and frozen on a liquid nitrogen-cooled copper block. For ultrastructural work the samples were freeze-substituted in 2% osmium tetroxide in acetone at -70°C for 6 d, flat embedded in Spurr's resin, sectioned and stained for 10 min in 2% uranyl acetate in methanol and 20 min in undiluted lead citrate. For ion localisation cells were frozen in the same way but then substituted in 10% acrolein in diethyl ether with 3 A molecular sieve for 32 d at -90°C. After warming to room temperature in stages (20 h at -20°C, 4 h at 4°C and 4 h at room temperature), specimens were transferred to a sealed plastic chamber, kept dry (5-10% relative humidity) with molecular sieve and a stream of dry nitrogen (Marshall, 1980; Canny and McCully, 1986) and infiltrated for one week in chloride-free resin (Pallaghy, 1973), also stored over molecular sieve. Fresh sieve was added to the specimens at each resin change. The resin was polymerised overnight at 70°C and the blocks were stored in a desiccator for one week to allow complete polymerisation. Sections 1 pm thick were cut dry (ie. without a trough of fluid) at 30% relative humidity or less, maintained using an Ebco Oasis 2700L dehumidifier (see Fitzgerald and Allaway, 1991), and were placed on formvar and carbon coated nickel slot grids. Unless extreme care was taken to keep all solvents and conditions very dry, redistribution of ions occurred. Single spots were analysed at ambient temperature (17°C) by energy-dispersive X-ray microanalysis using a Philips CM-12 scanning transmission electron microscope (STEM) and an Edax PV9900 analyser (120 kV). Analysis time was 100 s (live detector time) for chemically fixed samples and 200 s for freeze-substituted samples. Probe current was not altered between Polyphosphate in Pisolithus tinctorius 9

the analyses of granules or vacuoles and cytoplasm in each cell. X-ray mapping was done on an Hitachi H-7000 STEM and a Kevex analyser (100 kV) at a resolution of 128x64 pixels and 1500 ms dwell time.

Results

Precipitation of metachromatic granules during conventional specimen preparation The precipitation of metachromatic granules in vacuoles during ethanol dehydration is shown in Fig. 2.1. Figures 2.1A-D first show characteristic changes in the vacuole in living hyphae. The vacuoles were large, elongate and often occupied most of the diameter of the (Fig. 2.1 A). They were mobile but generally much less so than those at the hyphal tip. Some large vacuoles were seen to fuse with adjacent ones and others to divide (Figs 2.1A-D). Occasionally vacuoles extended into thin tubules (Fig. 2. ID) which fused with other vacuoles for short periods and then separated. In their living state these large vacuoles had no visible refractive contents under DIC optics (Figs 2.1A-D). Fixation with glutaraldehyde caused rapid cessation of the movement of vacuoles and other cytoplasmic components but caused no visible change in the appearance of the vacuole contents; no refractive granular material appeared in the vacuoles (Fig. 2.IE) for at least 1 h after cytoplasmic streaming had stopped. On occasions when fine tubular extensions of vacuoles were present, these broke up into a string of vesicles during fixation (Fig. 2. IE). Addition of 5% ethanol to the fixed cells also caused no visible change to the vacuole content (Fig. 2. IF), but the subsequent addition of 10% ethanol caused the precipitation of small granules (Fig. 2.1G). These showed Brownian motion at first but this ceased with time. Most granules remained in the same position as the cells were dehydrated to 100% ethanol in a graded series (Figs. 2.1H-J); in fact the number and distribution of granules did not change after the cessation of Brownian motion. Addition of toluidine blue O, pH 1.0 increased the contrast of some granules under DIC optics (Fig. 2.IK) and, under bright field optics, these granules were shown to be stained red (Fig. 2.1L). Another sequence (Figs 2.2A-D) shows that the addition of toluidine blue O, pH 1.0, directly to glutaraldehyde-fixed cells (Fig. 2.2A,B) without dehydration, similarly caused the precipitation of vacuolar granules (Fig. 2.2C), and that they were also stained red (Fig. 2.2D). Similar results were obtained when glutaraldehyde-fixed cells were treated with 0.05% toluidine blue O at higher pH (for example pH 6.2). Direct addition of toluidine blue O (pH 1.0) to living hyphae also caused the cessation of cytoplasmic streaming and granulation of the vacuole contents (compare Fig. 2.2E and F). Again, many of the precipitated granules stained red. Treatment of living hyphae (Fig. 2.2G) directly with 30% (Fig. 2.2H), 50%, or 100% ethanol also caused precipitation of granules, although there was massive Polyphosphate in Pisolithus tinctorius 10

disruption of the cells such that the distribution of cell contents including granules could not be related to the structure of the cell in the living state. Granules in the ethanol-treated cells increased in contrast when stained with toluidine blue O at pH 1.0 (Fig. 2.21) and were shown to be red under bright field optics (Fig. 2.2J). These experiments collectively show that refractive granules are precipitated by the addition of toluidine blue O, or ethanol, to either living or glutaraldehyde-fixed cells, and that these granules are y- metachromatic when stained with toluidine blue O at low pH.

Evidence that metachromatic granules contain polyphosphate Chemically-fixed, GMA-embedded hyphae also contained vacuolar granules which stained red with toluidine blue O, pH 1.0 (Fig. 2.3A). These were extracted either partially or completely with cold trichloroacetic acid (Fig. 2.3B) and showed peaks for phosphorus and calcium (Fig. 2.4A) when 1 pm thick sections were subjected to X-ray microanalysis (n=6). Regions of the vacuole not containing granules did not show peaks above background (Fig. 2.4B). This indicates that the vacuolar granules contained polyphosphate (see Ashford et al., 1975, 1986).

Comparison of vacuole content after conventional procedures and freeze- substitution Glutaraldehyde-fixed, ethanol-dehydrated hyphae invariably showed discrete electron- opaque deposits in the vacuoles. These were more or less spherical and tended to fall out of the material during sectioning or staining, leaving an electron-lucent space (Fig. 2.5). Vacuoles containing these granules were otherwise largely devoid of electron-opaque material. This contrasts with the appearance of vacuoles after freeze-substitution (Fig. 2.6). These contained evenly dispersed, electron-opaque material which varied in amount between cells. Vacuoles without any contents were not seen, nor were vacuoles ever found with large electron-opaque deposits like those in chemically fixed hyphae. Vacuoles were identified in unstained sections (cut dry at 1 pm) of hyphae freeze- substituted in acrolein-ether, as electron-opaque regions that could be readily distinguished from the cytoplasm, nuclei and wall. These were of comparable size, shape and distribution to those seen in ultrathin sections of osmium-acetone substituted material. Some were damaged by ice-crystal growth during freezing and only those that were recognisable and well preserved were chosen for X-ray microanalysis (n>50). All of these vacuoles contained relatively high levels of potassium and phosphorus (Fig. 2.7A), whereas these ions were barely detectable in the cytoplasm (Fig. 2.7B). X-ray mapping of thick freeze-substituted sections (Fig. 2.8A) showed that most potassium and phosphorus were located in the vacuole and were evenly distributed throughout. These elements were also detected in the walls and septum (Figs. 2.8B,C). Relatively high levels of potassium were detected in the vicinity of the septal pore and in the extracellular Polyphosphate in Pisolithus tinctorius 11

crystals (Fig. 2.8C). Occasionally other elements, notably sodium and magnesium, were detected with potassium and phosphorus in vacuoles, but calcium levels were not above background.

Discussion

Polyphosphate granules and specimen preparation Chemically fixed, dehydrated hyphae invariably show vacuoles containing discrete granules that only give signals for P and Ca with X-ray microanalysis (see Ashford et al., 1986 for references). These granules are surrounded by electron-lucent areas that give no signal above background. In contrast, in cells that have not been conventionally fixed (for example in air-dried, freeze-dried or freeze-substituted cells) potassium, not calcium, is the element most frequently associated with phosphorus (Doonan et al., 1979; Walker and Powell, 1979; Crang, 1980; Orlovich, 1987; Orlovich et al., 1989, 1990). This agrees with biochemical data from non-aqueous extracts of freeze-dried yeast, where large amounts of potassium and small amounts of calcium are associated with polyphosphate (Jacobson et al, 1982). Polyphosphate balanced by potassium is reported to be soluble. Miller (1984) for example showed that potassium ions alone did not precipitate polyphosphate in vitro. Potassium was the only element consistently detected together with phosphorus in most vacuoles. This is circumstantial evidence that polyphosphate is likely to be in a soluble form in the vacuoles of living P. tinctorius hyphae. A polyphosphate with an average chain length of 15 Pt subunits has been identified in extracts and whole hyphae of P. tinctorius by various methods including 31P NMR spectroscopy (Ashford et al., 1994). No polyphosphate of longer chain length was detected. 31P NMR spectroscopy is known to detect soluble polyphosphate but not polyphosphate precipitates since these do not spin. In earlier work (Appendix: Orlovich et al., 1989, 1990) potassium was found to be associated with phosphorus in what were initially thought to be granules. It is now believed that these were whole vacuoles seen in the thick, unstained sections, rather than granules in vacuoles. The finding that ethanol, but not glutaraldehyde, precipitates the polyphosphate as granules is in accord with known properties of these fixatives. Glutaraldehyde does not eliminate osmotic responsiveness in many tissues (see Brunk and Ericsson, 1973; Davey, 1973) and this is usually taken to indicate that membrane permeability is not very much altered by aldehyde fixation, at least in the short term. Ethanol, in contrast, is well known to change membrane permeability substantially. As the tonoplast becomes leaky during ethanol dehydration, ions will enter the vacuole and may precipitate polyphosphate to form granules. In chemically fixed and embedded material, the ion accompanying phosphorus in granules invariably is calcium (see for example Ashford et al., 1986; Polyphosphate in Pisolithus tinctorius 12

Grellier et al., 1989), so the precipitate would appear to be calcium polyphosphate. Commercial glutaraldehyde contains calcium as a contaminant (Morgan et al., 1978), and this may be the source of the calcium detected in the granules of chemically fixed, ethanol-dehydrated material. Another possibility is that precipitation of polyphosphate is caused by the removal of water by ethanol, and other ions may not exchange until a later stage. Precipitation of polyphosphate by toluidine blue O is not surprising, given that the dye is cationic and has been shown to precipitate polyphosphate in vitro (Miller, 1984). There is evidence from the literature that polyphosphate granules may be an artefact of specimen preparation in other fungi. Martin et al., (1983) found that most of the polyphosphate accumulated in three ectomycorrhizal fungi could be detected by 31P NMR spectroscopy, indicating that it was present in solution or as a macromolecular aggregate (Martin et al, 1985) rather than as solid granules. Delaporte (1939) reported that metachromatin exists in colloidal solution in vacuoles of lower organisms and is precipitated by the action of vital dyes and certain fixatives to form metachromatic red granules. Nakai (1976) found that vacuolar inclusions were induced by glutaraldehyde fixation in Lentinus edodes (Berk.) Sing., while Wiemken et al., (1979), using rapid freezing and freeze-substitution, showed that isolated yeast vacuoles and vacuoles in intact cells contained material dispersed through the entire vacuole rather than discrete granules. In spite of this, in almost every attempt to localise and quantify polyphosphate in microorganisms, conventional specimen preparation techniques and toluidine blue staining or precipitation with lead have been used (Cox et al., 1980; Lapeyrie et al., 1984; Straker and Mitchell, 1985; Tijssen et al., 1985; Ashford et al., 1986; Strullu et al., 1986). Under these circumstances it is very unlikely that the number and size of granules will accurately represent the amount of polyphosphate initially present, because precipitation must depend on the nature of the counter-ion, experimental conditions and the chain length of the polyphosphate.

Cation exchange with polyphosphate in vivo The cation associated with polyphosphate at any time in vivo will depend on the access of the ion to the vacuole and its capacity to displace less electronegative cations from the polyphosphate. If precipitating cations such as Mg2+ are sequestered in the vacuole, formation of granules may, under some circumstances, occur in vivo and this possibility is not ruled out by our results. Similarly, many of the toxic ions such as Al3+, Cd2+, Ni2+, Cu2+ and Zn2+ will bind to and precipitate polyphosphate if they have access to the vacuole, and it has been suggested that vacuolar polyphosphate may in this way play a role in detoxification of heavy metals, enabling mycorrhizal plants to grow better on contaminated soils (see Vare, 1990). However, when analysis of the cations associated with vacuolar polyphosphate following treatments with heavy metals has been attempted, methods have been used where leakage of heavy metals into the cells and secondary Polyphosphate in Pisolithus tinctorius 13

binding to polyphosphate have not been ruled out. Granules could be precipitated following an influx of heavy metals from the surrounding solution, or extracellular spaces, into the cells during tissue manipulation. In this case there will be a rapid binding to the polyphosphate and the cation composition will then simply reflect the most efficient ion exchanger, and will not represent an in vivo detoxifying mechanism. This is likely to happen if conventional tissue preparation involving chemical fixation and ethanol dehydration are used. In one case where freeze-substitution was used the metal (in this case Al) was confined to the cell walls and and was not found inside the cells (Denny and Wilkins, 1987), suggesting that exclusion at the plasma membrane is the primary mechanism for protoplast protection. The phenomenon of polyphosphate precipitation when appropriate cations are present may also explain the many conflicting reports of polyphosphate in various cellular locations other than the vacuole, especially the cell surface (see Kulaev and Vagabov, 1983; Tijssen et al, 1985; Vare, 1990). If low molecular weight polyphosphates are not retained in the vacuole during procedures that damage the cells, they may be precipitated at various other locations either by incoming counter ions or cellular cationic substances (such as basic proteins) that are immobile at other cellular locations (for example in nuclei). In freeze substituted P. tinctorius hyphae no evidence of polyphosphate in locations other than the vacuole could be found. Nicola Young (in Young et al, 1993: Appendix) showed that phosphorus compounds, identified as polyphosphate, are redistributed to new cellular sites when freeze-substituted fungal sclerotia, cut dry, are stained with toluidine blue in aqueous solution. Vesiculation of tubular vacuoles occurred as a result of glutaraldehyde fixation. The ultrastructure of tubular vacuoles as preserved by freeze-substitution is shown in Chapter 3. Vesiculation has been reported in numerous other systems, for example of canalicular vacuoles in Heracleum mantegazzianum (O'Brien et al., 1973), elongated vacuoles in basidiospores of Gymnosporangium juniperi-virginianae (Mims et al., 1988b) and membranous cleavage sheets in Phytophthora (Hyde et al., 1991). Tubular of isolated murine macrophages are particularly sensitive to aldehyde fixatives and improved preservation was achieved using a lower concentration of fixative (0.2% glutaraldehyde-0.5% formaldehyde) than that normally used (Luo and Robinson, 1992) or by rapid freezing and freeze-substitution (Robinson and Karnovsky, 1991). The occurence of elongate vacuoles has important implications for our understanding of potential transport pathways in fungal hyphae.

Significance in transport The role of mycorrhizal fungi in enhancing the phosphorus status of their higher plant associate is well documented (see Harley and Smith, 1983). This is thought to be achieved by absorption of phosphorus from the soil by extramatrical hyphae and its Polyphosphate in Pisolithus tinctorius 14

translocation through the mycelium to the root surface, where it can be exchanged with the higher plant. Fungal hyphae grow towards sources of soil P and zones of P depletion around individual roots may thereby be avoided (Jennings, 1989). Transport of phosphorus along hyphae is clearly of great importance to this whole process. It is known to occur at high rates and is metabolically dependent (see Gianinazzi-Pearson and Gianinazzi, 1989). The current view is that the transport of P along hyphae occurs largely by transport of orthophosphate, either by diffusion, osmotically generated mass flow, or cytoplasmic streaming (Jennings, 1987, 1989; Thompson et al., 1987), and possibly also via movement of small polyphosphate granules, presumably in vacuoles, also carried by cytoplasmic streaming (Yanagita, 1964; Cox et al, 1975, 1980; Callow et al., 1978). It is shown here, however, that polyphosphate is present in solution in the vacuoles. The vacuole system in P. tinctorius has been shown to be motile and interconnected via tubular elements. These transport a fluorochrome by peristaltic movements, along hyphae and from cell to cell, at least in growing hyphal tips (Shepherd et al., 1993a,b). Movement is independent of the cytoplasmic streams. Since phosphorus and potassium are also present in this vacuole system, this is a potential pathway for the transport of these elements along the hyphae by the same mechanism, in a compartment separate from the cytoplasm. Maintaining polyphosphate in soluble form would facilitate this transfer. It would also avoid some of the problems inherent in other mechanisms, namely: the feasibility of transfer of large amounts of orthophosphate via the cytoplasm when the P, concentration in the cytoplasm should be controlled because of its effects on glycolysis, via hexokinase and phosphofructokinase; the problems of establishing appropriate gradients in living cells; and the incompatibility of bidirectional movement with pressure driven bulk flow of solution (Jennings, 1987, 1989). 15 Polyphosphate in Pisolithus tinctorius

Fig. 2.1. Nomarski DIC (A - K) and bright field (L) light micrographs of whole- mounted hyphae. Bar = 10 pm. (A - D) Living hypha photographed at 5 minute intervals. (A) The hypha is highly vacuolate. The two vacuoles (v) at the left fused to form one larger vacuole in (B). (B - D) Two vacuoles at the right of (B) have formed from a single vacuole in (A). A tubule (t) has extended from the large vacuoles and traverses the region containing a nucleus. In (D), a tubule almost connecting two larger vacuoles is present. This was the situation just prior to fixation. There are no refractive granules in the vacuoles. (E) The same hypha after fixation with 2.5% glutaraldehyde in MMN. Movement of cytoplasmic contents and vacuoles ceased. The thin tubule seen in (D) has vesiculated, forming a chain of small vesicles Q>). No granular refractive contents are seen in the vacuoles. (F) Same hypha after treatment with 5% ethanol. There are still no observable granular contents in the vacuoles. (G - H) The same hypha after addition of 10% (G) and 15% (H) ethanol. Refractive granules are seen in the vacuoles (>). Some granules moved within the vacuoles with Brownian motion but this ceased after 15% ethanol was added. (I - J) Hypha after 95% (I) and 100% (J) ethanol. Refractive granules are present in vacuoles (>). (K) The hypha after staining with toluidine blue O, pH 1.0. Some of the vacuolar granules have increased in contrast. (L) Bright field image showing the sub-vacuolar granules which exhibited y- metachromasy. V V 16 Polyphosphate in Pisolithus tinctorius

Fig. 2.2. Nomarski DIC (A - C, E, G -1) and bright field (D, F, J) light micrographs of whole-mounted hyphae. Bar = 10 pm. (A) The living hypha contains vacuoles (v) which have no visible refractive contents. Many vacuoles are closely associated with the lower edge of the hypha. Movement of cytoplasmic contents and vacuoles was observed prior to fixation. (B) After fixation with 2.5% glutaraldehyde in MMN, streaming of the cytoplasm and vacuolar movement ceased. There is a distinct change in the shape of the vacuoles after fixation. They do not appear to be as closely associated with the edge of the cell as in the living hypha. The vacuoles still have no granular contents. (C) After subsequent staining with toluidine blue O, pH 1.0, small refractive granules (>) are seen in the vacuoles. The granules were not present in the same vacuoles when the hypha was alive. (D) The granules are clearly seen under bright field optics and they exhibit y- metachromasy at pH 1.0. (E) The vacuoles (v) in this living hypha again contained no refractive granules. After staining this cell with toluidine blue O, pH 1.0, without prior fixation, the protoplasm was severely disrupted. Numerous small granules formed and movement of vacuoles and cytoplasmic components stopped. (F) Under bright field optics the small granules (►) were y-metachromatic. (G) This living hypha is again highly vacuolate and the vacuoles (v) are associated with one edge of the hypha. A cluster of small vacuoles is present in the young upper branch. There are no refractive contents in the vacuoles. (H) The contents of the hyphae are disrupted after the addition of 30% ethanol. The protoplasm has become highly granular and granules have appeared in the vacuoles (►). (I) Staining with toluidine blue O, pH 1.0, increases the contrast of some granules within the hyphae (►). (J) Granules in both hyphae stain red under bright field optics. Light green staining of some parts of the protoplasm occurred in this case but this was easily distinguished from the red granules.

17 Polyphosphate in Pisolithus tinctorius

Fig. 2.3. Light micrographs of a 1 pm thick section of glutaraldehyde-fixed, GMA- embedded hyphae. Bar = 10 pm. (A) Section stained with toluidine blue O, pH 1.0. Numerous y-metachromatic (red staining) granules are seen in the cells (>). (B) The same section after incubation with cold 10% trichloroacetic acid and restaining with toluidine blue O, pH 1.0. The granules are smaller and some have been removed completely.

18 Polyphosphate in Pisolithus tinctorius

Fig. 2.4. Typical energy-dispersive X-ray spectra from analyses of an unstained section (cut at 1 |Lim) of glutaraldehyde-fixed, ethanol dehydrated and GMA-embedded hyphae. The number at the top right of each spectrum is the full vertical scale measured in X-ray counts. (A) Spectrum from a spherical electron-opaque showing peaks for P and Ca. (B) Spectrum from an electron-lucent region in the same hypha and analysed under the same conditions as (A). There are no peaks significantly above the background. There were no peaks above background in spectra from the resin.

19 Polyphosphate in Pisolithus tinctorius

Fig. 2.5. TEM of part of a glutaraldehyde-fixed hypha embedded in Spurr’s resin and triple stained. Bar = 1 pm. There are spherical electron-opaque granules (★) in the vacuoles (v). The holes in the section are presumably from where granules have been lost during sectioning or staining.

20 Polyphosphate in Pisolithus tinctorius

Fig. 2.6. Transmission electron micrograph of a stained ultrathin section of part of a freeze-substituted hypha near a clamp connection. Bar = 1 |im. The vacuoles (v) contain electron-opaque material which is evenly dispersed throughout. They do not contain discrete sub-vacuolar granules. The content of some vacuoles was more electron-opaque than others but this did not vary within a single cell.

21 Polyphosphate in Pisolithus tinctorius

Fig. 2.7. Energy-dispersive X-ray spectra from the analysis of an unstained freeze- substituted hypha cut dry at 1 pm. The vertical full scale, measured in X-ray counts, is at the top right of each spectrum. (A) A typical spectrum from the analysis of a vacuole. There are large peaks for phosphorus and potassium. (B) A typical spectrum from the analysis of the cytoplasm of the same hypha and analysed under the same conditions as (A). The peaks for phosphorus and potassium are very small. There were no peaks above background in spectra from the analyses of resin outside the cell profiles.

22 Polyphosphate in Pisolithus tinctorius

Fig. 2.8. (A) Transmission electron micrograph of an unstained 1 |nm thick dry cut section of a freeze-substituted hypha near a clamp connection. Bar = 2 (im. A large vacuole (v) is recognisable, as is the septum. The region in the centre of the septum is interpreted to be either the annular septal pore swelling or the parenthosome (p). (B-C) X-ray maps of the same hypha in (A) showing the distribution of phosphorus (B) and potassium (C). The positions of the vacuole (v) and septal pore (p) are indicated. Phosphorus and potassium are localised together in the vacuole. A Ultrastructure of the vacuole 3 system in Pisolithus tinctorius

Introduction Vacuoles in fungi are important storage sites for inorganic phosphate (Pi), polyphosphates, other inorganic ions and basic amino acids (Chapter 2, Durr et al., 1979, Beever and Burns, 1980). An electrochemical gradient generated by proton pumps in the tonoplast membrane drives ion porters which in turn influence both the vacuolar and cytoplasmic pH and ion composition (Klionsky et al., 1990). Fungal vacuoles contain many hydrolytic enzymes (Klionsky et al., 1990) and are similar to lysosomes of animal cells in this regard. Thus the vacuole is involved in the control of osmoregulation and metabolism in addition to its role in storage and lytic activities. Vacuoles in actively growing hyphal tip cells of Pisolithus tinctorius accumulate the fluorochrome 6-carboxyfluorescein (6-CF) (Shepherd et al., 1993a) which is compartmented into an interconnected, motile system of pleiomorphic vacuoles and tubules. The tubules move independently of the cytoplasmic stream and transport 6-CF between clusters of small vacuoles situated at intervals along the apical cell. 6-CF is transferred from one vacuole cluster to another, apparently by dilation and contraction along the length of the tubules. Tubules also transfer 6-CF across the dolipore septum between the tip and penultimate cells just after septum formation (Shepherd et al., 1993b). The system of vacuoles and tubules thus has the capacity to act as an inter- and intracellular transport pathway which is separate from the cytoplasm. This is particularly significant in filamentous fungi where food sources and sinks may be separated by large distances. The mechanisms driving the movements of the vacuole system of young hyphae are relevant to the role of vacuoles in phosphorus translocation. Although some components of the ultrastructure of fungi have been described from freeze-substituted material (Howard and Aist, 1979; Hoch and Howard, 1980, 1981; Howard and O’Donnell, 1987; Hoch, 1991; Lii and McLaughlin, 1991; Hippe- Sanwald et al., 1992), little attention has been focused on the vacuole system in young cells. Observations of living hyphae using fluorescence or differential interference contrast (DIC) microscopy reveal details about the movements of subcellular organelles but it is difficult to identify organelles definitively and classify various compartments without reference to complimentary electron micrographs. The relationships between vacuoles, tubules, other organelles and elements of the can be investigated in detail using transmission electron microscopy. Membrane continuity between organelles indicates a structural connection between them and lumen continuity can indicate a potential pathway for the movement of material from one to the other. When correlated

23 Ultrastructure o/Pisolithus tinctorius 24

with data from living cells, either untreated or loaded with tracer fluorochromes, conclusions can be drawn concerning what organelles are involved in the transport processes. It is important that the results from light and electron microscopy can be correlated with confidence. The disrupting effects of conventional fixation and dehydration on the vacuole morphology and composition of P. tinctorius hyphae (Chapter 2) illustrates the difficulty of correlating electron micrographs of chemically-fixed cells with light micrographs of living hyphae. Organelles move and change shape during fixation and dehydration and their relationship to other organelles is altered. Rapid freezing followed by substitution in organic solvents at low temperature is used to avoid many artefacts known to occur using chemical fixation (Hoch, 1991). Disorganisation of the ultrastructure is much less likely to occur during freezing and substitution because of the very rapid freezing rate and low thermal energy of the frozen sample during substitution. This chapter describes the ultrastructure of tip and older cells in freeze- substituted hyphae, emphasising the distribution and identity of the vacuole system and its relation with other organelles. It complements and clarifies observations of the motile system obtained from fluorescence microscopy, giving evidence of a potential intravacuolar pathway for the transfer of vacuolar content along hyphae and showing that tubular cisternae are able to cross completed dolipore septa.

Materials and Methods

Rapid freezing and freeze-substitution Hyphae of P. tinctorius were grown on 5 mm diameter discs of cellulose nitrate/acetate filters (see Chapter 2). Five samples of mycelium were taken from the growing edge of each of five culture plates and freeze-substituted in acetone/Os04 as described in Chapter 2. Each of these 25 pieces of flat-embedded mycelium was subdivided into lxl mm pieces from which young or older cells which did not show evidence of ice crystal damage were selected by examination using light microscopy. Block faces were trimmed to approximately 150 x 150 pm to facilitate serial sectioning of individual hyphae. Sections were cut and stained for 10 min in 2% uranyl acetate in methanol and 20 min in undiluted lead citrate (Reynolds, 1963).

Measurements and statistics Measurements of dimensions were made from prints enlarged by approximately 2.5 times using a 6 x eyepiece magnifier with a 0.1 mm graticule (Graticules Ltd, England). Averages are cited as means ± 95% confidence limits with the sample size in brackets. Confidence limits were calculated for sample size n using the formula: Ultrastructure o/Pisolithus tinctorius 25

X(Jn - 1 X t[a = 0.05; v = n -1] 95% limits yfn where xOn - \ is the sample standard deviation, t[a = o.osjv = n - n is the critical value for Student’s /-distribution at a probability with v degrees of freedom.

Results

General ultrastructure of the tip region and vacuoles and tubules in cells close to the hyphal tip Vacuoles were the most obvious organelles in electron micrographs of freeze-substituted cells from hyphal tips. They were situated in interconnected clusters located at intervals along the terminal cell and contained a characteristic evenly dispersed electron-opaque material (Figs 3.1, 3.2 and 3.3). Their profiles ranged from approximately circular to ovate, pyriform, irregular or elongate; narrow connections between adjacent vacuoles were frequent. Vacuole width (measured in a plane perpendicular to the long axis of the hypha) averaged 0.52 ± 0.06 pm (n=36). The clusters were a complex system in which vacuoles were connected, often to more than one adjacent vacuole, via narrow membrane- enclosed bridges (Fig. 3.3). Elongate vacuoles commonly spanned the region of the cytoplasm between adjacent vacuole clusters (Figs 3.1, 3.2, 3.4B). Invariably they were oriented parallel to the longitudinal axis of the hypha and were occasionally situated close and parallel to elements of the cytoskeleton (Figs 3.1, 3.2). Elongate vacuoles up to 14 pm long were observed. They were characterised by alternate constricted and dilated regions (Fig. 3.4B) with measurements of diameter ranging from 0.3 - 0.8 pm at the dilations (mean = 0.41 ±0.11 pm, in 3 vacuoles at n=15 points) to 40 - 150 nm at constricted regions (mean = 83 ± 72 nm, n=4). The narrowest regions of these vacuoles approached the diameter of the lumen of the rough endoplasmic reticulum (mean = 39 ± 3 nm, n=21) which also lay approximately parallel to the long axis of the hypha (Fig. 3.2). Rough ER was present in regions that contained vacuole clusters and in those that did not, but was more obvious in the latter (Figs 3.1, 3.2). Profiles of rough ER were continuous over long distances in serial sections with occasional breaks in individual sections, indicating that the ER occurred in the form of perforated sheets. Sheets of rough ER also lay parallel to the septum (see Chapter 4). Tubules of smooth membrane cisternae were widespread in both peripheral and central regions of the terminal cells in zones with and without vacuole clusters (Figs 3.1, 3.2), in cytoplasm around the nucleus (Fig. 3.4C) and associated with the septal pore complex at the base of the cell (Fig. 3.5a,b). Smooth membrane cisternae often formed characteristic ring-like structures (Fig. 3.4A,C), also extending into the extreme hyphal Ultrastructure o/Pisolithus tinctorius 26

tips which did not usually contain vacuoles (Fig. 3.4A). The diameter of the smooth membrane cisternae varied from 60 ± 8 nm (n=14) at dilated regions to totally constricted at some points. Small vacuoles and smooth membrane cisternae were associated with the dolipore septum. Smooth membrane cisternae passed through the parenthesome perforations as far as the threshold of the dolipore, at least partially occluding the pore entrance (Fig. 3.5a). There was enough space for two cisternae to pass through the pore at any one time, although more than two passed through the parenthesome (Fig. 3.5b). Smooth membrane cisternae and small vacuoles were the only organelles found between the parenthesome and the dolipore entrance in the hyphae. Electron-opaque material was present around the pore swellings and in the pore itself in dolipores with smooth membrane cisternae associations. This electron-opaque material is treated in further detail in Chapter 4. Diameters of the parenthesome perforations and of the pore neck are compatible only with the passage of elements of the dimensions of the smooth membrane cisternae or constricted vacuoles. Further measurements of components of the dolipore septum are given in Chapter 4. Profiles of smooth membrane cisternae often also lay parallel to elements of the cytoskeleton, many of which were oriented approximately parallel to the long axis of the hypha. had a diameter of 26.4 ± 0.4 nm (n=18) and were easily identified in longitudinal (Fig. 3.3B) and transverse (Fig. 3.4C,D,E) profile. Structures tentatively identified as in longitudinal profile from Figures 3.1 and 3.2 had a mean diameter of 14 ± 2 nm (n=10). Most did not have any internal substructure discernable in this plane of section. Groups of microfilaments (Fig. 3.4E,F) that may constitute small bundles are seen in cross section in transverse sections of the hyphae. Fine filaments had varied diameters and a number of different types of filament are likely to be present. Microfilaments within bundles were 9.5 ±1.0 nm (n=5) wide in transverse section. Some filaments (Fig. 3.4F) had an electron-lucent core 6 ± 1 nm (n=5) in diameter. Other distinct organelles, identified as Golgi bodies, consisted of clusters of vesicle-like profiles and short branched tubules with dilated regions and electron-opaque contents which formed a reticulum (Figs 3.1, 3.2, 3.4A, 3.6). The tubules were occasionally connected to a single flat cisterna (Fig. 3.7). Many of the circular profiles were tubules oriented in various planes, rather than vesicles (Fig. 3.3, 3.6) when examined in serial sections. The outer diameter of the dilated regions of these tubules was 85 ± 6 nm (n=17). The tubules frequently had a core of electron-opaque material (Fig. 3.6D, 3.7A). Fine electron-opaque strands connected the core to the membrane. The core of electron-opaque material was obvious in some transverse profiles of dilated tubules (Fig. 3.7B) and was surrounded by an electron-lucent layer. Where the tubules were dilated, they contained more electron-opaque material. The electron-lucent layer was of more or less constant width in profile (15 ± 3 nm, n=8). Strands of electron- Ultrastructure o/Pisolithus tinctorius 27

opaque material traversed the electron-lucent region. Occasionally profiles of two membranes lying parallel and at a constant distance (33 ± 2 nm, n=6) from each other were observed (Fig. 3.7C). These often had dilated ends of similar appearance to the tubular profiles in the Golgi bodies. Tubular profiles were frequently connected to flat sheets (Fig. 3.7D-G) from which they appeared to radiate. The sheets were either perforated (Fig. 3.7F) or not (Fig. 3.7D, E, G) and were more electron-opaque than the surrounding cytoplasm. It is possible that the parallel membranes shown in Fig. 3.7C are the flat sheets sectioned transversely. The vesicles at the hyphal tip (Fig. 3.4A) had a similar electron-opaque content to the vesicle-like profiles of the Golgi bodies but lacked an electron-lucent region inside the membrane and were narrower in diameter (69 ± 4 nm, n=22). Electron-opaque deposits of unknown origin or composition occurred adjacent to the plasma membrane along the sides of the hyphal tip, but not at the tip itself, appeared to get wider away from the apex and were very irregular (Fig. 3.4A).

Ultrastructure of cells behind the growing tip Cells became increasingly vacuolate with distance from the tip. Observation of living hyphae using DIC optics (Chapter 2: Fig. 2.2A) showed that even the tip cell may be highly vacuolate towards the base by the time a septum forms to delimit a new penultimate cell. Cells further back from the hyphal tip contained vacuoles that were commonly located at the cell periphery. This was observed in living (Chapter 2: Fig. 2.2A, 2.2G) and freeze-substituted hyphae (Fig. 3.8A,B, 3.9). Ultrastructure of hyphae chosen randomly from regions not at the growing front of the mycelium (but still within 5 mm of it) was highly variable and distinctly different to that in the tip cells. A mosaic of hyphal profiles frequently occurred in single sections (Fig. 3.8A). Vacuoles tended to be larger in older cells and contained more electron-opaque material than those in younger cells although quantitative measurements were not made. Clusters of small interconnected vacuoles like those shown in Chapter 2: Fig. 2.4 were not observed in more mature hyphae although the vacuoles were still often pleiomorphic (Fig. 3.8A, 3.9C). A notable difference in the cytoplasm of many older cells was an increase in the number of glycogen rosettes per unit area of cell profile. While no such deposits were observed in tip cells (Figs 3.1-3.4), glycogen was common in more mature cells (Fig. 3.8A). Conversely, profiles of Golgi bodies were prominent in tip cells (Fig. 3.4A) but not observed as frequently in older cells. Other organelles were present, notably mitochondria, rough endoplasmic reticulum, vacuoles, tubular cisternae and multivesicular bodies (Fig. 3.8A). Interspersed amongst mature hyphae were vacuolate cells where the cytoplasm consisted of evenly dispersed material with only occasional, often peripheral, organelles (Figs 3.8A,B, 3.9). The cytoplasm contained microtubules (Fig. 3.8B) but not free Ultrastructure o/Pisolithus tinctorius 28

or glycogen. The content of the vacuoles varied: it was highly electron-opaque in some cells (Fig. 3.8B, 3.9A) and less so in others (Fig. 3.9B) but always contained evenly dispersed material. Vacuolar content in a single cell was usually of a similar electron-opacity for all vacuoles but this varied between cells. Vacuoles were either approximately hemispherical (Fig. 3.9B) or irregular (Fig. 3.9C) and often with one edge parallel and closely appressed to the plasma membrane (Figs 3.8, 3.9A-C). The extent of contact between the tonoplast and the plasma membrane varied according to vacuole size. The tonoplast was closely appressed to the plasma membrane over the full length of smaller vacuoles, forming long straight profiles (Fig. 3.9B) but only attached at certain regions along the profile of larger vacuoles (Fig. 3.9C,D) with regions of cytoplasm between. Electron-opaque material in these regions spanned the gap between the two membranes (Fig. 3.9D). The electron-opaque material consisted of short strands of fibrillar material attached to each membrane (Fig. 3.9E). The mean distance between the plasma membrane and the tonoplast membrane was 19 ± 4 nm (n=l 1). The tonoplast membrane was coated with material resembling the contents of the surrounding cytoplasm (Fig. 3.9F), which occasionally radiated some distance from the tonoplast membrane into the cytoplasm (Fig. 3.9G). Additional evidence of a close association between the plasma membrane and the tonoplast membrane was gained during the exposure of sections to the electron beam. Sections tended to tear during such irradiation. Tearing occurred between the wall and the plasma membrane, presumably due to incomplete polymerisation of the resin in this region. The plasma membrane remained attached to the only in those regions where it was already closely associated with vacuoles (Fig. 3.9G), indicating a strong connection between the two membranes and the wall in these regions.

Discussion

Clusters and tubules Numerous membrane-enclosed systems were identified in young freeze-substituted hyphae. The largest organelles in the tip cells were vacuoles, recognisable by their size, content and irregular shape. Vacuoles formed clusters in the cortex of the cytoplasm and the clusters were located at intervals along the length of the tip cell. The vacuolar content formed a continuum: the vacuoles in each cluster being interconnected by narrow bridges. Elongate vacuoles with dilated and constricted regions spanned those regions of cytoplasm which did not contain clusters and were also observed within clusters. All of these vacuoles had similar contents within each cell. The ultrastructural identification of a vacuole/tubule system in hyphae of P. tinctorius accords with results of V. A. Shepherd (Shepherd et al., 1993a,b) which show that the hyphal tips contain an interconnected Ultrastructure o/Pisolithus tinctorius 29

system where clusters of vacuoles containing a fluorescent tracer, connected along the hypha by tubules, were observed in living hyphae. Living, untreated hyphae also contained narrow tubules capable of extending and contracting to connect large vacuoles and also spanned the region of the developing septum (Chapter 2). The ultrastructural evidence is consistent with the hypothesis that elongate vacuoles transport their content between the saltatory clusters of small vacuoles situated at intervals along the hypha. It is concluded that these tubules and clustered vacuoles are part of an interconnected system involved in intra- and intercellular transport near the growing apex of hyphae. The ultrastructural identification of this system was made possible by the use of freeze-substitution which, as is discussed in Chapter 2, avoids many artefacts associated with chemical fixation. It is especially significant that the morphology of elongate vacuoles was preserved by freeze-substitution since it is shown in Chapter 2 that elongate vacuoles seen in living cells became vesiculated in response to the addition of glutaraldehyde. Thus they are rarely observed in electron micrographs of chemically fixed hyphae. Microtubules and microfilaments were oriented parallel to the long axis of the hyphae and thus also parallel to the elongated vacuoles. It is likely that the complex movements exhibited by the vacuoles in vivo are controlled by elements of the cytoskeleton, although the mechanism cannot be elucidated from electron microscopy alone. Neither microtubules nor microfilaments were preferentially associated with constricted or dilated parts of the elongated vacuoles and it is possible that the characteristic shape changes are the result of purely physical processes acting on the vacuole. Recently, Matsuo and Tanaka (1992) showed that patterns similar to those seen here in elongate vacuoles could be induced in polyacrylamide gels. In that study, narrow cylindrical hydrated gels were dehydrated in acetone/water mixtures. Dilated and constricted regions formed along the gels during dehydration under certain conditions. Pattern formation in the gels as a result of dehydration was attributed to: the formation of an impermeable outer layer, the buildup of internal pressure and phase separation of the polymer. There are parallels between these three factors in the shrinking polyacrylamide gel and the elongate vacuole system. All vacuoles are likely sinks for molecules such as Pi and basic amino acids which are incorporated into vacuole-specific compounds and cannot escape through the tonoplast membrane. Internal pressures in the vacuole could conceivably be generated by a combination of osmotic pressure and the result of the transfer of vacuole content from a tubule to a cluster without a concomitant transfer of membrane. This vacuole system does not only occur in P. tinctorius. In tetraploid cells of the yeast, Saccharomyces cerevisiae, the vacuole divides just prior to nuclear division and in doing so a tubule extends from the vacuole of the mother cell into the bud (Jones et al., 1993). These authors noted that the tubule was capable of a periodic contraction which Ultrastructure o/Pisolithus tinctorius 30

started at the vacuole and proceded along the tubule into the bud in a similar manner to the dilations and contractions observed in P. tinctorius. An elongate vacuole about 6 pm long with dilated and constricted regions was observed in freeze-substituted cells of the fungal rice blast pathogen Magnaporthe grisea (Herbert) Barr by Bourett et al. (1993). Glycoconjugates of glucose or mannose were localised in these elongate vacuoles distal to the hyphal tip and other cell compartments using concanavalin A-gold binding. No information is given on the movements of the vacuoles in living hyphae. Elongate vacuoles were observed in basidiospores of freeze-substituted cedar-apple gall rust, Gymnosporangium juniperi-virginianae Schw. (Mims et al., 1988/?). These vacuoles also had dilated and constricted regions although they contained unevenly dispersed material unlike that seen in P. tinctorius vacuoles. The elongate vacuoles in P. tinctorius resemble tubular lysosomes in murine macrophages as preserved by freeze-substitution (Robinson and Karnovsky, 1991). This is circumstantial evidence supporting the hypothesis that fungal vacuoles have similar properties to lysosomes. Fungal vacuoles are known to contain phosphatases: phosphatases have been localised in the tonoplast membrane of yeasts (see Klionsky et al., 1990) and ectomycorrhizal fungi are known to produce hydrolytic enzymes (Antibus et al., 1981; Dighton, 1983; Kroehler et al., 1988; Bae and Barton, 1989; Ho, 1989). Sub-cellular localisation of phosphatase enzymes in P. tinctorius would be a valuable contribution to further studies on this organism. Most biochemical work on the composition of vacuoles has been done on yeasts (Klionsky et al., 1990). As unicellular organisms, yeasts are not likely to require a mechanism for the long distance transport of nutrients, metabolites etc. either through the cytoplasm or otherwise. This contrasts with filamentous fungi where the translocation of material over long distances is essential because of the nature of the growth of such organisms. The conventional image of vacuoles as large, static components of the cytoplasm is possibly a result of this difference. This belies the potential role of the vacuole in the translocation of its contents over long distances.

Fine filaments The movement of organelles is mediated by elements of the cytoskeleton (Schroer and Kelly, 1985, Alberts et al., 1989). The immunolocalisation of microfilaments was not investigated in P. tinctorius', as a result it is difficult to identify more than tentatively cytoskeletal elements other than easily distinguished microtubules. Fine filaments were seen in freeze-substituted hyphae but measurements varied from section to section and depended on the profile of the filament and the degree of staining of the section. It is likely that there is more than one type of filament in the sections. Large bundles of filaments were not observed in longitudinal section except at the developing septum (Chapter 4), whilst small groups or single filaments, often associated with smooth membrane cisternae, were. Hoch and Staples (1983) identified nine different Ultrastructure o/Pisolithus tinctorius 31

types in Uromyces phaseoli var typica but none of these was as wide as those found in longitudinal section here (14 nm). Of the nine different morphologies identified by Hoch and Staples (1983), many had overlapping size ranges and were distinguished on the basis of associations with particular organelles. The larger filaments identified here may in fact be collapsed microtubules, which have been shown to occur as a result of ice-crystal damage (Ding et ai, 1992) and have similar morphology to those observed in Fig. 3.1. Hollow filaments 9.5 nm wide with a hollow core 6 nm wide were identified in serial sections. Hollow filaments were not recorded by Hoch and Staples (1983) or Lii and McLaughlin (1991) in Auricularia auricula-judae but the outer diameter is similar to that of cytoplasmic microfilaments seen in both organisms. Moreover, the size of microfilaments recorded in the literature is variable. Microfilament bundles of higher plant cells are reported to contain filaments between 2 and 5 nm wide (Gunning and Steer, 1975). Those from high pressure frozen and freeze-substituted leaf tissue (Ding et al., 1992) were 6.9 ± 0.6 nm wide (n=6, not measured by those authors). A source of this variation is likely to be the method of specimen preparation: different techniques will result in the precipitation of different amounts of osmium and other heavy metal stains on the microfilaments giving them a wider or narrower profile. In any case, the differentiation of microfilaments from other protein components of the cytoplasm on the basis of ultrastructure alone, particularly for short profiles of single filaments 1 or 2 nm wide, is likely to be subject to significant errors of measurement.

Golgi bodies and smooth membrane cisternae Organelles identified as Golgi bodies had a single, occasionally fenestrated, cisterna confluent with a peripheral tubular reticulum. The reticulum had dilated regions and a central core of electron-opaque material. Significantly, the cisternae were not stacked. The tubular reticulum was highly branched near the cisterna and single tubules radiated away from each Golgi body. However, connections between individual Golgi cisternae via the tubular reticulum were not demonstrated here. Further electron microscopy, possibly with the use of thick sections and high voltage EM, is required to elucidate any connections between individual Golgi cisternae. Golgi bodies have remained poorly described in fungi. Reasons for this include: they are difficult to fix for electron microscopy; they have little morphological similarity to those of other eucaryotes; and they are difficult to isolate for molecular or biochemical studies (Graham and Emr, 1991; Ayscough et al., 1993). Golgi bodies have been described in several filamentous fungi following freeze-substitution although considerable confusion surrounds the distinction between Golgi bodies and smooth membrane cisternae. Golgi bodies consisting of a single fenestrated cisterna with an associated tubular reticulum containing electron-opaque material occur in Aspergillus nidulans (Mims et al., 1988a). The tubular reticulum often formed ring-like structures which had a Ultrastructure o/Pisolithus tinctorius 32

similar electron-opaque content to tubules confluent with the cisternae. However, other ring-like smooth membrane cisternae with electron-lucent contents, similar to those seen in P. tinctorius in Fig. 3.4A and 3.4C, also occurred in A. nidulans but were identified as Golgi bodies. Gymnosporangium juniperi-virginianae (Mims et al., 1988b) has Golgi bodies which have dilated vesicles interconnected by narrow bridges. The Golgi bodies appear to be clustered and have electron-opaque content. Some of the tubules in the reticulum have a more electron-opaque core (Mims et al., 1988b: Fig. 18). Again, ring­ like smooth membrane cisternae with electron-lucent content are also labelled as Golgi bodies (eg: Mims et al., 1988/7: Fig. 9) despite the lack of a clear morphological or mechanical connection between the two organelles. In barley powdery mildew (Erysiphe graminis DC. f. sp. hordei Em. Marchal), structures identified as ‘Golgi-like structures’ (Hippe-Sanwald et al., 1992: Fig. 14) are similar to the ring-like smooth membrane cisternae in P. tinctorius, whereas in the same paper (Hippe-Sanwald et al., 1992: Fig. 19) both a ring-like smooth membrane cistema and a single cisterna with electron-opaque contents associated with dilated vesicles are labelled as ‘Golgi equivalents’. Clearly there is a difficulty associated with identifying Golgi bodies in fungal cells, partly because of a variable and poorly described morphology and partly due to a lack of knowledge about the function of Golgi bodies and smooth tubular cisternae in these organisms. The function of Golgi bodies has been studied in more detail in yeasts. While Golgi bodies in Schizosaccharomyces pombe have stacked cisternae (Ayscough et al., 1993), those of Saccharomyces cerevisiae have only a single cisterna (or rarely up to three: Preuss et al., 1992). Despite this variation, there is considerable evidence of a functional similarity between Golgi bodies of yeasts and those of animal cells (Rothman and Orci, 1992). Thus the occurrence of single Golgi cisternae in P. tinctorius or other filamentous fungi need not mean that their function is markedly different from that in other organisms. For example, in Chinese hamster ovary cells, proteins being transported through the Golgi complex need not move from cisterna to cistema within one stack (Rothman et al., 1984). Rather, vesicles moved randomly; they could move from a cisterna in one Golgi complex to a target cisterna in either the same or another complex. Ayscough et al. (1993) recently showed that for up to 2 hours after stacked Golgi cisternae in Schizosaccharomyces pombe had been artificially ‘unstacked’ by the addition of an anti- agent there was little change in cell growth or acid phosphatase secretion (measures of Golgi function). In other words, stacking of cisternae is not always necessary for Golgi function. The composition of the electron-opaque core or the electron-lucent region surrounding it in the Golgi tubular reticulum is unknown although it is likely to contain material being processed by the Golgi complex. The electron-lucent region surrounding the content of some Golgi bodies may contain glycoconjugates attached to the inside of the membrane. Concanavalin A-binding glycoconjugates were localised in smooth Ultrastructure o/Pisolithus tinctorius 33

membrane cisternae in Magnaporthe grisea by Bourett et al. (1993). The cisternae were fenestrated and were confluent with a ring-like-tubular component similar to the Golgi bodies in Aspergillus nidulans (Mims et al., 1988a). The cisternae in M. grisea did not have an electron-lucent region inside the membrane but such a region has been observed in other organisms. Glycoconjugates have been localised in an electron-lucent region in stacked Golgi cisternae by Roth (1983), on lysosomal membranes in rat kidney by Neiss (1984) and in tubular lysosomes of macrophages (Sawnson et al., 1987). Clusters of narrow tubules in HeLa cells, resembling the Golgi tubular reticulum seen in P. tinctorius, were identified as mitotic Golgi fragments that reassemble to form stacked cisternae after mitosis (Lucocq et al., 1989). The Golgi bodies in P. tinctorius bear a striking structural resemblance to the system of tubular reported in cultured animal cells (Tooze and Hollinshead, 1991) in which tubular endosomes and mitotic Golgi clusters were demonstrated to be synonymous (Tooze and Hollinshead, 1992). However, there is considerable controversy over the identification of these endomembrane systems in animal cells (for example see Pypaert et al., 1993). An endocytic marker and acid phosphatase have simultaneously been localised in a similar tubular reticulum in macrophages (Luo and Robinson, 1992). The implication of such findings for fungal hyphae cannot be judged until components of the endosomal and lysosomal systems are distinguished. Smooth membrane cisternae with an electron-lucent content formed localised networks of narrow tubules and occasionally formed ring-like structures which have been described as Golgi bodies in other fungi (see above). They show parallels with structures that load 6-CF, seen by fluorescence microscopy (V. A. Shepherd, unpublished observations). The identity of the ring-like structure, its function and relationship with other organelles remains unknown. Smooth membrane cisternae were also seen in association with the dolipore septum, where tubules passed through the parenthesome pores and occluded the dolipore opening. It is tempting to propose that smooth membrane cisternae passing through the dolipore septum are part of the same system of tubules which transfer fluorescent material across the septum during and after cytokinesis. The evidence is strong that tubules traverse the developing septum at this stage in the cell cycle: this has been shown in untreated living cells by DIC microscopy (Chapter 2) and in living cells loaded with 6-CF (Shepherd et al., 1993b). As these tubules frequently originate from vacuoles and bear a strong resemblance to those that interconnect clusters of vacuoles, it would be parsimonious to hypothesise that the tubules seen crossing the septum in living cells were part of the vacuole system and not smooth membrane cisternae. It is likely that the smooth membrane cisternae are indeed capable of uptake and transport of 6-CF as evidenced by the identification of the ring-like organelles in 6-CF-loaded cells. However, no ultrastructural evidence of a connection between structures identified as smooth membrane cisternae and the vacuole system was Ultrastructure o/Pisolithus tinctorius 34

found. Therefore it is possible that those components of the vacuole system involved in intercellular transport await ultrastructural identification.

Mature hyphae There are a number of ultrastructural differences between cells at the growing tip and those further back in the hypha. Characteristics of older hyphae of P. tinctorius are (i) they contain more glycogen, (ii) the vacuoles contain more electron-opaque material and are generally associated with the plasma membrane at the edge of the cell; and (iii) in some cells the cytoplasm is changed, containing homogeneously dispersed fibrillar material. Aging of the hyphae is likely to be accompanied by an increase in the amount of reserve material laid down, shown here by an increase in glycogen deposits in the cytoplasm and electron-opaque material in the vacuole. Glycogen is a storage carbohydrate in ectomycorrhizal fungi (Foster and Marks, 1966; Ling-Lee et al., 1977). PAS-positive carbohydrates occur in hyphae of P. tinctorius (Orlovich, 1987) and have a similar distribution to that of the glycogen rosettes observed in freeze-substituted cells. A change in the morphology of vacuoles and an increase in the amount of electron-opaque vacuolar contents was observed in older hyphae. Vacuoles in older cells were connected to the plasma membrane by short strands which presumably held the vacuole closely appressed to the edge of the cell. This close association was also observed in living hyphae in Chapter 2. It is significant that the close attachment to the cell periphery is lost once the cells are exposed to glutaraldehyde (Chapter 2: Fig. 2.2A,B). This may explain why the close relationship between vacuoles and the cell periphery has not previously been investigated in detail, although evidence that this is a widespread phenomenon is available. Living hyphae of Pyronema confluens were observed by Buller (1933) who found that they contained large peripheral vacuoles which, while capable of movement, were firmly attached to the periphery of the hypha and also the septum. Buller (1933) hypothesised that the cytoplasm formed a continuum capable of movement from cell to cell, while vacuoles were held in place by peripheral attachment. Vacuoles are associated with the cell periphery in living Rhizoctonia solani hyphae (Hoch, 1991). The vacuoles are disrupted by the addition of phosphate-buffered glutaraldehyde and the distinct association with the edge of the cell is lost within 2 minutes after fixation. In hyphae in which the cytoplasm contains fibrillar material, the ultrastructure of the cytoplasm bears a striking resemblance to that of cytoplasm containing P-protein in sieve-elements of higher plants (eg. Evert and Deshpande, 1969). It is tempting to draw comparisons between the two systems concerning the role of the fungal cytoplasm in translocation of water and dissolved material. Based on measurements of water flux through mycelium of Serpula lacrimans, Thompson et al. (1985) concluded that water flux through the mycelium occurs by turgor pressure-driven bulk flow of solution. This Ultrastructure o/Pisolithus tinctorius 35

is unlikely to occur in the cytoplasm of actively growing hyphae, so it follows that certain hyphae must be specialised for the translocation of water. P. tinctorius forms aggregated vegetative structures similar to rhizomorphs in axenic culture on agar plates (D. Orlovich, personal observations) and it can be inferred from this that under these conditions of growth, there is a net transfer of water and/or nutrients from the growing edge of the mycelium back to the older regions. It follows that as cells mature at the growing edge, some hyphae will become specialised for that function. Hyphae with a modified cytoplasm have vacuoles that are tightly attached to the plasma membrane and there are few organelles free in the cytoplasm. If material flows through the cytoplasm then it is conceivable that organelles need to be firmly held so that they are not moved as well. No biochemical information about the cytoplasmic material in P. tinctorius is available. Developmental sequences of maturing hyphae using freeze-substituted mycelium would be useful for further understanding of these cells. In the hyphae examined here, there was still a septum present between mature cells and this would be a considerable barrier to the transfer of large amounts of cytoplasm from cell to cell. There is evidence that in Armillaria mellea (Cairney et al., 1988), the septa break down during rhizomorph development so perhaps the hyphae seen here represent an intermediate stage where some reserves are accumulated in the vacuoles as the cytoplasm changes and the cell ages. 36 Ultrastructure o/Pisolithus tinctorius

Fig. 3.1. Adjacent areas of the sub-apical zone of a freeze-substituted terminal cell, in longitudinal section. The bottom of (A) is continuous with the top of (B). t, direction of the hyphal tip. Bar = 1.0 jam.

There are two clusters of vacuoles in this section of the hypha. The one with fewer, smaller vacuole profiles (v-|) is nearer the tip than the other (V2). The contents of the vacuoles are evenly dispersed. There is some evidence of ice-crystal damage in the vacuole content, particularly in the vacuoles of V2. Several of the vacuoles in each cluster are connected by narrow bridges. One of the vacuoles in vj is elongate. There are many profiles of rough ER cisternae (rer) oriented more or less parallel to the long axis of the hypha between, and adjacent to, the vacuole clusters. Ribosomes are sparsely distributed on the rough ER profiles. Clusters of vesicle-like profiles and short tubules, identified as Golgi-like organelles (g 1,2,3) by their size, shape and electron- opaque content, occur at intervals along the hypha. Several sectioned cytoskeletal elements are identified as microfilaments (mf) and run more or less parallel to the long axis of the hypha. m, mitochondrial profiles. mx . ; *;^n mm*

. ' mi v SJj A J \ ( j/ V1/ > - a a!

^mmm fog*' twiwi

■ \hrf^Vi' IM insfii £P««*sWSSf.

j « £ ':: fe»i tf'f ^ w.-.-

l--A '^mw\v . TiKSt *li 37 Ultrastructure 6>/Pisolithus tinctorius

Fig. 3.2. Another section from the same series as Fig. 3.1, showing the same vacuole clusters (v-| and V2). T, direction of the hyphal tip. Bar = 1.0 jam.

In this section there are fewer vacuole profiles in \\ than shown in Fig. 3.1. An elongate vacuole (V), showing the typical dilated and narrow regions, is oriented parallel to the long axis and spans the gap between vj and V2. In addition to the many rough ER profiles, there are profiles of smooth tubular cisternae (tc). They are wider and more irregular in diameter than the rough ER (rer) cisternae, and show distinct constrictions. Although often longitudinally oriented, they are not always parallel to the long axis, as are many of the rough ER cisternae. Microfilaments (mf) run for long distances parallel to the hyphal long axis. Individual Golgi-like bodies (g-1,2,3) can be traced through the series. Mitochondrial profiles (m) are continuous through the series of sections, indicating a mitochondrial reticulum. *77* • Nit *»< 0* .> & mm .. if! ft V Mx. %i: IPiSJW)r'.'y'^.' M- ' vW I ■it 4 i; 3 >Ji

- b— :. ■: . = : ■vi„m

■-‘i J8r/‘P$s® m Wk?£Z gfe-v&m J lip Afcifi %; *

4 mm-smS t MBS ! jft < -' ssiila Jopw■\ y w%%m iff sgrslmmmB mmtmmSSslgs™; m> i Us, Sf Mm i ■H IlillfeiiM % 9^, '•T;..:- .93^ ■■ '•, , ', •;• ■ 'i&'iei'i.#.' ■< M 38 Ultrastructure of Pisolithus tinctorius

Fig. 3.3. Four consecutive sections from a series through a vacuole cluster show interconnections between all vacuoles in the cluster. Bar = 1.0 pm.

The position of individual vacuoles (V1.5) can be determined relative to the other organelles such as the adjacent Golgi body (g). All vacuoles are interconnected in some plane or other by narrow bridges, showing that the vacuole cluster is a continuum. Note the heterogeneity in the cytoplasm; there is an area relatively free of rough ER and ribosomes immediately around the vacuole cluster. A cluster of microtubules (mt) running parallel to the long axis of the hypha may be traced in several micrographs. The Golgi-like body (g) is continuous in all sections and forms an interconnected system of small tubules and vesicles.

39 Ultrastructure q/’Pisolithus tinctorius

Fig. 3.4. Transmission electron micrographs of freeze-substituted hyphae. Bar = 1.0 pm (A-C), 0.5 pm (D-F).

(A) Apical zone at the hyphal tip. of the extreme tip is indicated by the cluster of electron-opaque vesicles (ve), characteristic of the region surrounding the Spitzenkorper. Behind this region there are many profiles of smooth tubular cisternae (tc) oriented in various planes including two rings of tubules (arrow) but no vacuoles. Behind this are several Golgi bodies (g), mitochondrial (m) and ER profiles, then a vacuole cluster (v) and some rough ER cisternae.

(B) An elongate vacuole (v), with dilated and constricted regions and characteristic content, runs longitudinally for a distance of at least 8.0 pm. The section also shows short lengths of several microtubules (mt), all oriented parallel to the hyphal long axis.

(C) Transverse section from a series through the nucleus (n) in one hypha and a vacuole cluster (V) in another. Many circular profiles of organelles with smooth membranes and of different dimensions are obvious in both hyphae. The large profiles (v) can be identified by their size and content as dilated regions of the interconnected vacuole system, while the smaller profiles (tc) are of tubules of the dimensions of smooth tubular cisternae. A ring of tubules (>) similar to those in (A) is present in one of the hyphae. Single microtubules and groups of two or more, all in cross section, occur throughout both hyphae.

(D-F) Enlargement of three areas from (C) showing transverse profiles of microtubules (mt) and microfilaments (mf) in more detail. Some of the microfilaments in (F) are sectioned obliquely. They appear to be composed of a narrow electron-opaque cylinder (you have to look very closely!). mt mi 40 Ultrastructure of Pisolithus tinctorius

Fig. 3.5. Sections of freeze-substituted dolipore septa, showing tubules of smooth membrane cistemae in the region of the septal pore complex. Bars = 0.5 pm.

(a) Several smooth membrane cistemae (tc) radiate towards the parenthesome and in this section three pass through the parenthesome perforations. The membrane profiles appear precisely aligned so that it passes through the perforations more or less perpendicular to the parenthesome, towards the mouth of the septal pore. A sheet of endoplasmic reticulum (er) (determined from the examination of serial sections) lies parallel to one side of the septum but there is none on the other side. There is electron- opaque material in the pore and around the pore swelling. The cytoplasm inside the parenthesome is of a different appearance to that outside it. More electron-opaque material is accumulated under the parenthesome on the same side at the smooth membrane cistemae than the other.

(b) Some smooth membrane cistemae can be traced from the cytoplasm, through the parenthesome perforations, to the dolipore entrance where they partially occlude the septal pore. The diameter of the parenthesome perforations (76 nm) is similar to that of the septal pore entrance, but the latter widens to 120 nm at the midpoint of the wall. Two tubules occur at the septal pore (arrows) although more than this number penetrate the parenthesome. A vacuole profile (v) occurs close to the septal pore inside the region bounded by the parenthesome. Vacuoles and the smooth membrane cistemae were the only organelles found in this region. Profiles of endoplasmic reticulum (er) lie adjacent and parallel to both sides of the septum. VV‘> . /' , T* 1 V'**j ^Si «*?#€$ £P1 ^ tTpvH- 4' \ ^er Jr** EBB2sti$ r tp v V *1 *>Ss-i* 41 Ultrastructure o/Pisolithus tinctorius

Fig. 3.6. Serial sections through a cluster of Golgi-like tubular cisternae in a freeze-substituted hypha. Bar = 200 nm.

(A) Circular profiles of dilated membrane cisternae (arrow) are interconnected by narrow branched tubules (>). Some of the dilated cisternae contain electron-opaque material.

(B) Consecutive section showing again the branched nature of the narrow membrane cisternae which interconnect dilated circular profiles. The circular profiles are approximately 80 nm in diameter, similar to the thickness of the sections. As dilated circular profiles do not appear in the same position in consecutive sections, they are likely to be approximately spherical. However, they are frequently connected to the branched tubules (arrow).

(C) One of the tubules is dilated (>) in this consecutive section and has an electron- lucent lumen (compare with Fig. 3.7B). Most cisternae are not sectioned in the median plane, however electron-opaque material indicates their position.

(D) Cisternae cut in this section are mostly circular in profile and contain an electrron- opaque core. The circular profiles underlie tubular cisternae in the previous section indicating a close association with them. £«’* *.** is . a&agfipi if %.y*

* * 42 Ultrastructure o/Pisolithus tinctorius

Fig. 3.7. Transmission electron micrographs of Golgi-like clustered tubular cisternae.

(A) Circular profiles of cisternae associated with the same Golgi-like cluster shown in Fig. 3.6. One of the circular profiles (>) is continuous with an elongated tubule. Electron-opaque ‘spokes’ (arrows) span the region between the electron-opaque core and the limiting membrane. Bar = 200 nm.

(B) Section through Golgi-like tubular cisternae. The cisternae contain a core of electron-opaque material which is surrounded by an electron-lucent layer (►) across which there are irregularly placed strands of material. The electron-opaque core is wide in the dilated regions and constricted at narrow regions. Bar = 200 pm.

(C) An elongate cisterna with dilated ends in close proximity to circular membrane profiles. Elongate cisternae such as this were found on numerous occasions. The limiting membranes lie parallel to one another for some distance. It is possible that the tubules sectioned on the right are part of the tubular reticulum associated with the flat cisterna sectioned in profile on the left. Bar = 200 pm.

(D-E) Serial sections cut parallel to a sheet-like cisterna. The section before that shown in D did not contain the sheet-like cisterna and that following E showed only tubules.

(D) The cisterna extends into dilated processes which contain evenly distributed electron-opaque material and extend into the surrounding cytoplasm. Bar = 200 pm.

(E) The elongated processes are similar to the dilated cisternae associated with Golgi- like clusters in this consecutive section. The cisternae are connected to the flat membranous sheet and extend from this into the surrounding cytoplasm. Bar = 200 pm.

(F) Dilated tubules containing electron-opaque material radiate from an electron- opaque sheet which is perforated with small holes. Bar = 200 pm.

(G) Another section where tubules with an electron-opaque core are connected to a flat sheet. Bar = 200 pm. WF aftrrvgNfcV ■> *» - - t» - *TV*Ct47A* *«CVk '{fe, ?«, >J

•v<*?a* •

*«fjr ~ *fM - V V,~J, ,fi »' *£& V- * <* ’4 6'' 4«#. .—*7? I .:■

-m^S ■ ■ -r-^ V. 'iSA.^-v;- „*m fe.s ■ % & J| w.: >& 5|aJL- ;, <1%* -. 4W.tti .'• ^ra i%. .^.:•■ ~• &k.• ,—*,- ,-1

* \\ ali^y’^SsjS^kEB -**■ 33 :-vjp t> :***i •-Sfr v' **.i*V sii

m* m C 4. /TFV IrP"» ..* j ** v*A;^ *V

* *5. i k‘ ?.>;■' ■* ,a* L; . , flrriir ‘ “"ii t ., *>ia^ ■ h * ^3* 43 Ultrastructure o/Pisolithus tinctorius

Fig. 3.8. (A) The cell containing the two nuclei (n) has glycogen deposits (gl) in the cytoplasm. In addition, there is a similar complement of organelles to younger cells: m, mitochondria; v, vacuoles; rer, rough endoplasmic reticulum; tc, tubular smooth membrane cisternae; g, Golgi bodies; mvb, multivesicular body. Adjacent cells also contain glycogen (gl). The cytoplasm (★) of one cell consists of evenly dispersed flocculent material and contains few organelles and little glycogen. The content of the vacuoles (v) in this cell is more electron-opaque. One of these vacuoles is closely associated with the edge of the cell (V). Bar = 2.0 pm.

(B) Transverse section of a cell containing a peripheral vacuole (v), some smooth tubular cisternae (tc) and transverse microtubule profiles (mt). The cytoplasm has a flocculent appearance. Bar = 500 nm.

44 Ultrastructure o/Pisolithus tinctorius

Fig. 3.9. Transmission electron micrographs of mature freeze-substituted hyphae.

(A) Portion of a hypha with large peripheral vacuoles (V) and containing few other large organelles (m, ). The vacuoles are closely appressed to the plasma membrane and they are filled with an electron-opaque material. The appearance of the cytoplasm is mostly of evenly dispersed fibrillar material similar to that in Fig. 3.8B. The adjacent younger narrow hypha has a dense cytoplasm with small vacuoles not attached to the plasma membrane. Bar = 5.0 pm.

(B) Peripheral vacuoles (v) in this mature hypha contain electron-opaque material that appears to be more dispersed than (A). Again the cytoplasm contains dispersed fibrillar material and few large organelles. There are microtubules (mt) present in the cytoplasm and some smooth tubular cisternae (tc). Bar = 5.0 jam.

(C) Vacuoles (v) contain evenly dispersed electron-opaque material. They are pleiomorphic and are appressed to the plasma membrane (>). The cytoplasm consists of flocculent material. Bar = 2.0 pm.

(D) Part of a hypha containing a large vacuole (v). The vacuole contains electron- opaque material. The tonoplast membrane adjacent to the plasma membrane is joined at certain regions (V) by electron-opaque material. Bar = 1.0 pm.

(E) Ultrastructure of the plasma membrane/tonoplast membrane interface. Short strands span the region between the plasma membrane and tonoplast membrane (>). Bar = 200 nm.

(F) Peripheral vacuole (v) surrounded by dispersed material in the flocculent cytoplasm. Strands of electron-opaque material of similar appearance to that in the cytoplasm extend radially from the tonoplast (►). Again there is electron-opaque material between the pm and tm. Bar = 0.5 pm.

(G) Hypha with peripheral vacuoles (v). The protoplasm has been torn from the cell wall by irradiation with the electron beam. The section has split between the hyphal wall (w) and the plasma membrane except at the sites adjacent to the peripheral vacuoles (►). At these sites, the tonoplast membrane and plasma membrane remain attached to the wall. The vacuoles are coated with material which radiates into the surrounding cytoplasm (»). Bar = 1.0 pm.

Structure and development of the dolipore septum in Pisolithus tinctorius

Introduction The dolipore septum with its associated clamp connection is thought to be part of a complex mechanism evolved in dikaryotic mycelia of many basidiomycetes ensuring an ordered migration of nuclei so that new cells each receive a copy of both nuclei following mitosis (Alexopoulos and Mims, 1979). The classical view of the dolipore is that, whilst it prevents random migration of nuclei and mixing of organelles, it maintains cytoplasmic continuity between adjacent cells so that transport of molecules can occur along the hyphae in the symplast. The extent to which the septum is a barrier to this depends on the size of the pore, viscosity of its content and whether any occlusions are present, as for any symplastic connection (Gunning and Robards, 1976; Gunning and Overall, 1983; Robards and Lucas, 1990). The structure of the dolipore septum in basidiomycetes has been extensively studied, but mostly from a taxonomic viewpoint and using conventional methods of fixation and embedding (Moore, 1984; Lii and McLaughlin, 1991). Unfortunately, methods which use chemical fixatives in aqueous solutions produce a number of artefacts. These include characteristic swelling of the septal wall at the rim around the pore, changes in dimensions of the pore, loss of filamentous structures and changes in the structure and relationships of membranes (eg. Hoch and Howard, 1980, 1981; Howard and O’Donnell, 1987). These artefacts make it difficult to draw useful conclusions about the structure and transport capacity of the dolipore. Freeze- substitution avoids many of these artefacts and preserves structures which are not usually found after chemical fixation (Howard and O’Donnell, 1987; Lingle, 1989) and a large amount of work illustrates the value of this technique in preserving fungal ultrastructure. This is especially important for relationships of membranous structures, the size and shape of compartments or pores, and elements of the cytoskeleton. Hoch and Howard (1980, 1981) were the first to examine septal structure in freeze-substituted hyphae of basidiomycetes. Their work with the holobasidiomycete Laetisaria arvalis Burds. showed that septal swelling after freeze-substitution was only about one third of that after chemical fixation. A recent paper on septal structure in the heterobasidiomycete Auricularia auricula-judae (Bull.:Fr.) Schroet., also using freeze- substitution, revealed new cytoplasmic features of relevance to the maintenance of the dolipore septum and its role in transport (Lii and McLaughlin, 1991).

45 The dolipore septum in Pisolithus tinctorius 46

In this chapter, the development of the dolipore septum between the tip and penultimate cell is examined, and also the structure of mature, apparently functional, septa in Pisolithus tinctorius. The structure of septal pores is relevant to longitudinal transport of nutrients along hyphae from the soil to the tree root system. The hyphal tips of P. tinctorius contain a highly motile pleiomorphic vacuole and tubule system which accumulates the fluorochrome 6-carboxyfluorescein. The ultrastructure of components of this system is described in Chapter 3. Tubules of this system transport fluorochrome between vacuole clusters situated at intervals along the apical and penultimate cells (Shepherd et al., 1993a) and across the dolipore septum (Shepherd et al., 1993b). This transfer is very active during and just shortly after the completion of the dolipore complex between the apical and penultimate cells. Here, the dolipore structure and development in these septa is examined precisely at this stage in freeze- substituted hyphae. The observations have been correlated with those from differential interference contrast microscopy of living hyphae.

Materials and Methods P. tinctorius, isolate DI-15 (Grenville et al., 1986), was cultured as described in Chapter 2. Five samples of mycelium (grown on Nuclepore filters: Chapter 2) from the growing edge of each of 5 culture plates were cut from the agar and frozen on a copper block which was cooled in a bath of liquid nitrogen. They were then freeze-substituted in 2% OsC>4 in acetone at -70°C for 6 d, and embedded in Spurr's resin (Spurr, 1969) as described in Chapter 2. Each of these 25 pieces (5x5 replicates) of flat-embedded mycelium were subdivided into 1 x 1 mm pieces and examined by light microscopy to select either young or older cells which did not show evidence of ice crystal damage. Sections were cut and stained for 10 min in 2% uranyl acetate in methanol and 20 min in undiluted lead citrate (Reynolds, 1963). Some hyphae were freeze-substituted in 20% acrolein in diethyl ether for 32 d, embedded in Spurr's resin, and ultrathin sections were triple stained (Daddow, 1983). Electron micrographs were taken with an Hitachi H-7000 transmission electron microscope at 100 kV. Averages of measurements made from electron micrographs are cited as means ±95% confidence limits (sample size given in brackets). At least ten mature septa and three developing septa were examined. Many were serially sectioned. For scanning electron microscopy, samples were fixed in 2.5% glutaraldehyde in 0.025 mol.L-1 potassium phosphate buffer, dehydrated in an ethanol series, transferred to 100% acetone and critical point dried. They were sputter coated with gold/palladium and examined in a Cambridge S-360 scanning electron microscope with an accelerating voltage of 20 kV. The dolipore septum in Pisolithus tinctorius 47

Results

Clamp connections and dolipore septa The formation of the dolipore septum in P. tinctorius follows the pattern typical for basidiomycete fungi. As the hypha grows it periodically forms septa which separate the apical cell from the penultimate cell. The septa form after the two nuclei have undergone mitosis. One nucleus divides in the plane parallel to the longitudinal axis of the hypha and a septum forms perpendicular to this between the two daughter nuclei. Concurrently the second nucleus divides in an oblique plane so that one of its progeny moves into the clamp connection, a short anabranch that is initiated anterior to the site of the main septum, and grows back to fuse with the main hypha basal to it. Formation of the clamp connection temporarily isolates this nucleus in the clamp. The clamp cell then fuses with the subterminal segment re-establishing the dikaryon. Therefore, there are two septa formed: one in the clamp and the other across the main hypha. If branching occurs, the new penultimate cell produces the branch on the basal side of the main septum, and a second clamp forms when this branched penultimate cell divides to cut off a tip cell (Figs 4.1 and 4.2).

Early development of the dolipore septum Development of the septa of the clamp and main hypha is synchronised (Fig. 4.3). The septa are formed by invagination of plasma membrane perpendicular to the long axis of the cell, with concurrent wall deposition (Figs 4.3-4.7). The septal wall material was similar in appearance to the lateral hyphal wall and was confluent with it (Figs 4.3 and 4.4) . At this stage the septum was centripitally tapered. The plasma membrane was closely appressed to the new wall and appeared more electron-opaque than elsewhere. At the inner edge of the ingrowing septum, the membrane was less clearly defined and its trilaminar structure was not apparent. The cytoplasm in this region appeared more electron-opaque than elsewhere (Figs 4.3 and 4.4) and contained fine filaments (Figs 4.3-4.7). A group of filaments, situated just internal to the centripetally expanding membrane, was in transverse profile in median sections of the ingrowing septum (Fig. 4.4) while filaments in non-median sections passing through the edge of the pore were in longitudinal profile. (Figs 4.3, 4.5 and 4.7). The latter extended across the pore and some continued as far as the lateral walls of the parent hypha (Fig. 4.5). Serial sections showed that many of these filaments formed a ring just inside the invaginating membrane. Clusters of filments in transverse section often occur at the junction of the septal and lateral walls (Figs 4.3 and 4.6). Small vesicles were frequent in the cytoplasm along the sides of the newly developing septum, interspersed amongst the filaments (Figs 4.3 and 4.7). These The dolipore septum in Pisolithus tinctorius 48

vesicles were surrounded by a unit membrane (Fig. 4.7) and had a moderately electron- opaque content. They were circular in profile and were shown in serial sections to be small spherical organelles rather than tubules. Multivesicular bodies and mitochondrial profiles were also frequently seen in the vicinity of the septum. The mitochondria, which were long and cylindrical, tended to span the incipient pore region (Fig. 4.3). Smooth membrane cisternae also frequently spanned this region at this stage. The multivesicular bodies were circular or elongate in profile and contained small vesicles which were usually more or less spherical (Fig. 4.8). These internal vesicles were complex in structure with a centre of low electron-opacity surrounded by a more electron-opaque region with a fuzzy coat. Most were similar in size (34 ± 2 nm, n=30) but some larger ones were occasionally seen (Fig. 4.8). Rough ER was seen in the cytoplasm in the vicinity of the developing septum and occasionally was found parallel and adjacent to the longitudinal wall of the main hypha. Broad flat profiles of a membrane enclosed compartment with finger-like projections and sometimes seen to be connected with smooth tubular cisternae were commonly found surrounding the invaginated membrane apex (Figs 4.9 and 4.10). An elongate vacuole, of a similar width to those seen interconnecting groups of vacuoles in other hyphae, was observed in the vicinity of the developing septum in one instance. A later stage of development is shown in Figures 4.11-4.16, where the pore was much narrower. The clamp tip had not yet fused with, or dissolved, the lateral wall of the main hypha and the nucleus was still within the clamp (Fig. 4.11). The septum in the main hypha was at a similar stage of development to the clamp septum. Serial sections through the clamp septum show that the septal pore apparatus has not yet developed (Figs 4.12-4.16). Electron-opaque material is seen on either side of the septum at the mid-point but no structural detail may be discerned (Fig. 4.12). The orifice of the pore was more electron-opaque than the surrounding cytoplasm (Figs 4.13 and 4.14). A mitochondrion was continuous through the pore (Fig. 4.14). Tubular cisternae occurred in the vicinity of the septum (Fig. 4.12) and may be traced along adjacent to the septum (Figs 4.14 and 4.15), but there was no parenthesome or rough ER in an arrangement characteristic of mature septa. In Figure 4.15 the septum is again continuous, indicating that only two sections in the series cut through the pore. There was electron-opaque material in the cytoplasm at the mid-point of the septum (Fig. 4.15) , as there was on the other side of the pore. In the final section of the series (Fig. 4.16) this electron-opaque material was not present.

The pore complex of mature septa The pore was barrel-shaped with a diameter of about 76 nm (n=10) at the orifices and 120 nm (n=10) at the midpoint of the wall (Figs 4.17-4.20). The pore was surrounded by the dolipore which was about 265 nm (n=3) thick in median sections. Adjacent to The dolipore septum in Pisolithus tinctorius 49

the orifices of the pore on either side was a perforated parenthesome characteristic of holobasidiomycetes. The parenthesome appeared as a cisterna-like structure apparently bounded by a unit membrane and with a content of alternating electron-opaque and -lucent layers in a very complex pattern (Figs 4.17-4.19). The thickness of the parenthesome, ie. the distance between the two outer surfaces (43 ± 2 nm, n=14), was slightly greater than the diameter of the rough ER. The rim of the parenthesome was located very close to the plasma membrane with a more or less perpendicular orientation, often with electron-opaque material bridging the gap. On the plasma membrane and more or less in line with the edge of the rim of each parenthesome there was a series of small electron-opaque deposits (Figs 4.17-4.19). In the best preserved material, these were approximately aligned with the inner and outer membrane and central region of the parenthesome (Figs 4.18 and 4.20). In the ether/acrolein freeze- substituted material, where sections were extracted, there were small filaments radiating from these deposits for a short distance towards the parenthesome. In one area, where the fine layer of electron-opaque material is pulled away from the plasma membrane, a fine filament connected an electron-opaque deposit within this layer to the parenthesome (Fig. 4.20). A single rough ER cisternal profile ocurred on each side of the parenthesome on either side of the septum and lay parallel to the septal wall. It was obvious in acetone/osmium freeze-substituted material (Figs 4.17-4.19), but not ether/acrolein freeze-substituted material (Fig. 4.20), although very close scrutiny of Figure 20 shows electron-transparent areas where it may have been. Lumen continuity could not be demonstrated between the parenthesome and any of these rough ER profiles, although the two often lay very close together and electron-opaque material appeared in some instances to bridge the gap between them (eg. Fig. 4.18). Ribosomes were sparse and were only seen on the cytoplasmic side of the rough ER (Fig. 4.19). Small filaments connected the septal side of the rough ER to the plasma membrane (Figs 4.18 and 4.19). Electron-opaque material was usually found in the lumen of the septal pore often filling it and extending to spread across the pore entrance and the surrounding septal pore swelling (Figs 4.17-4.19). This material varied in appearance according to the freeze-substitution technique used (compare Fig. 4.18 with 4.20). It often contained fine filaments and showed distinct electron-opaque bands with an orientation perpendicular to the pore channel. Filaments parallel to the pore channel connected with the transverse bands. The material which spilled out of the pore entrance and over the adjacent septal pore swelling also contained filaments (Figs 4.17 and 4.18) and in several places fine filaments radiated from it to the parenthesome (Figs 4.17, 4.18 and 4.20). In Figure 4.20 at the pore orifice on one side the material has contracted away from the wall indicating that it has cohesiveness. At the other side it is connected with electron-opaque layer from which a filament radiates to the rim of the parenthesome. The dolipore septum in Pisolithus tinctorius 50

Serial sections through the dolipore septum (Figs 4.21-4.26) showed rough ER cisternae ocurred as a perforate sheet on either side and parallel to the septum at a distance of 18.3 ± 0.3 nm from the plasma membrane (n = 10). The parenthesome formed a dome over each side of the dolipore and was perforated with regularly arranged hexagonal pores with an average diameter of 75 nm (n=27) (Figs 4.22 and 4.23). Regularly arranged filaments 10 ± 2 nm wide (n=7) were seen to radiate from the electron-opaque material around the pore swelling to the parenthesome with which they appeared to be confluent (Figs 4.24 and 4.25). Again the electron-opaque material was continuous through the septal pore. Some of the filaments were narrow and straight-sided (Fig. 4.25) while others appeared coated with electron-opaque material (Fig. 4.24). The parenthesome rim was very close to the plasma membrane in several places. It appeared confluent with it in tangential view (Fig. 4.22) and appeared to be attached to it via electron-opaque deposits in median section (Fig. 4.26). There was also a very close spatial relationship between the parenthesome and rough ER (notably in Figures 4.22, 4.23 and 4.26), but in no instance could the membranes or lumen be demonstrated as continuous. In sections cut approximately parallel to the septum (Figs 4.27-4.32), the hexagonal pores of the parenthesome are seen to be arranged in a ‘honeycomb’ configuration. There were three concentric layers containing 1, 6 and 12 pores respectively. Again, filaments were seen in some sections, radiating from what appears to be a ring of electron-opaque material surrounding the septal pore, to contact the parenthesome (Fig. 4.31). Some of the filaments appeared as a continuous series of arrowhead-like structures (Fig. 4.31). These filaments were present in most sections (Figs 4.28-4.31), but are sectioned at various angles and not always obvious. The electron-opaque material around the pore contained a complex network of interconnected filaments (Figs 4.30 and 4.31) and filamentous structures are shown in the entrance to the pore itself (Fig. 4.31). Presumably these are the longitudinal filaments sectioned obliquely as they radiate out from the pore entrance. A very complex pattern of filaments is therefore indicated in the sub-parenthesome area.

Discussion

General considerations The two septa in the main hypha and clamp are laid down synchronously over a matter of minutes. This was demonstrated in Chapter 2 using differential interference contrast microscopy of live cells. In Chapter 2: Fig. 2.3 there is no septum across the main hypha, while in Chapter 2: Fig. 2.4, taken 5 minutes later, the septum is complete. The fixed image of this septum (Chapter 2: Fig. 2.6) shows a dilated structure associated The dolipore septum in Pisolithus tinctorius 51

with the septum which is interpreted as a dolipore. A similar very rapid rate of septum formation (ie. completed within 4 minutes) is reported in Auricularia auricula-judae by Lti and McLaughlin (1991). Septum formation involves invagination of the plasma membrane and appears basically similar to cleavage in animal cells (Beams and Kessel, 1976), except that the furrowing membrane concurrently secretes cell wall material and this presumably further stabilises the structure. This has also been reported for other higher fungi and contrasts with cytokinesis in higher plant cells, where the cell wall is deposited as a cell plate which extends out centrifugally to fuse with the parent cell wall in a predetermined region of the cell cortex (Gunning, 1982). In animal cells the cleavage furrow forms precisely in the plane of the metaphase plate at right angles to the long axis of the mitotic spindle. Its position is controlled midway between the two asters originating from the two and it frequently traps longitudinal arrays of microtubules as cytokinesis progresses (Beams and Kessel, 1976). A spatial relationship between nuclear division and cytokinesis has been reported in fungi (see Girbardt, 1979) and may also be the case in P. tinctorius, but there are very few microtubules left in the region as the septum forms. Several general features of septal development, such as the occurrence of fine filaments adjacent to the developing septum, membrane invagination and microvesicles in the vicinity, are reported from chemically-fixed as well as freeze-substituted septa in a number of other fungi. However, freeze-substitution also reveals new features in P. tinctorius, some of which have been shown only recently in freeze-substituted septa of the heterobasidiomycetes Tremella globospora Reid and Auricularia auricula-judae (Berbee and Wells, 1988; Lti and McLaughlin, 1992).

Membrane furrowing and the filamentous ring Septa form by invagination of the plasma membrane. Although controversial (see Wessels, 1986), in growing hyphal tips it seems most likely that the protoplast exerts a positive pressure on the wall to create the necessary expansion at the tip, that is the apex is under turgor pressure. Under these conditions formation of a cleavage furrow is only possible if the plasma membrane is physically pulled inwards by some cytoskeletal element. The plasma membrane of the developing septum is in close association with filaments some of which are oriented in a ring around the circumference of the narrowing pore. A similar structure, sometimes called a filamentous septal belt, has been shown to be associated with septum formation in a number of other fungi (eg. Patton and Marchant, 1978; Girbardt, 1979; Hoch and Howard, 1980; Roberson, 1992). The ring of filaments in P. tinctorius may be a similar structure. This septal belt has been compared with the contractile ring in animal cells, found beneath the furrowing membrane and known to be composed of circumferentially oriented microfilaments associated with other cytoskeletal elements (Rappaport, 1986). Actin has been The dolipore septum in Pisolithus tinctorius 52

demonstrated at the site of developing septa in the ascomycete Sclerotium rolfsii Sacc. and other fungi (see Roberson, 1992), and cytochalasin E inhibits septum and septal ring formation in bean rust, Uromyces appendiculatus (Pers.) Unger, another basidiomycete (Tucker et al., 1986). However, confirmation that the septal belt consists of F-actin awaits labelling at the ultrastructural level. Any structure which causes membrane furrowing must have good lateral contact with the membrane. This applies to the filamentous electron-opaque material at the apex of the furrowed membrane around the edge of the pore. There appear to be three sets of filaments all perpendicular to each other. Two of these ran parallel to the membrane while the third set was perpendicular to it. Those filaments perpendicular to the membrane appeared to anchor it and are probably responsible for pulling it inwards, while the set represented by the electron-opaque circular profiles that these pass through could be part of the contractile ring. The third set are parallel to the membrane and interconnect the other two. Their arrangement, and the appearance of the membrane all indicate an interaction with the filaments, with the membrane being pulled in at this point. The difference between the appearance of the membrane around the rim of the pore and along the flanks of the septa may simply result from the membrane being more fragile at the innermost region, perhaps resulting from a slight delay between membrane assembly and wall deposition. In animal cells net membrane biosynthesis is reported to occur just before cell division and it is believed that extra membrane required during cytokinesis is stored at the cell surface (Alberts et al., 1989). In P. tinctorius, finger-like processes commonly seen extending from flattened cisternae around the apex of the furrowing membrane could perhaps supply membrane very rapidly to the developing septum. The electron microscopy does not shed any light on how the parenthesome forms, but it does show quite clearly that it appears late in septal development when the pore has become quite narrow. This indicates that structures such as the larger organelles can move across the septum until very late in septal development and, as shown here, organelles can be trapped and constricted by the rapidly narrowing pore. The septal pore swelling, a feature that characterises holobasidiomycetes and advanced heterobasidiomycetes (Hoch and Howard, 1981; Lii and McLaughlin, 1991), is not present at this stage either. The septal pore swelling at maturity was about 260 nm at its widest point. This agrees with results from other basidiomycetes. For example in Laetisaria arvalis the swelling ranged from 220 to 300 nm in freeze-substituted and 760 to 800 nm in chemically fixed material indicating that it was enlarged about threefold after chemical fixation. In Auricularia auricula-judae it was a somewhat narrower 100 to 184 nm but was still expanded by about a factor of three by chemical fixation. The dolipore septum in Pisolithus tinctorius 53

Structure of the septal pore complex at maturity The septum with its dolipore is thought to provide symplastic continuity while excluding organelles. It has been shown here, as in other basidiomycetes (see Lti and McLaughlin, 1991), that the sub-parenthesome space is largely free of organelles. The only structures shown in this area in P. tinctorius are the tubular cisternae which pass through the parenthesome pores to the septal pore entrance (Chapter 3 and Shepherd et al., 1993b) and occasional circular membrane profiles which may be these tubules in transverse section or small vacuole profiles. This indicates that the parenthesome, even though perforate, is very effective in excluding organelles from this zone, so that the septal pore entrance is not occluded by any of the larger organelles such as mitochondria. It is noteworthy that the diameter of the parenthesome pores is very similar to the diameter of the septal pore entrance. Therefore in this fungus, structures allowed through the parenthesome should also be able to enter and pass through the septal pore. However, it might be envisaged that for efficient movement through the pore, any structure passing through the parenthesome pores would need to be oriented towards the septal pore, so that it does not strike the septal wall rather than the pore entrance. The domed shape of the parenthesome, the position of the parenthesome pores and any guiding elements radiating towards these pores and between these and the septal pore orifice, may all play a part in such guidance. In the absence of microtubules in the sub-parenthesome region it would seem likely that the radiating filaments play some role in direction of movement. No attempt was made to identify the filaments seen here in P. tinctorius because no chemical information is known about them. Little significance can be ascribed to differences in size and appearance of the various filaments in the sub-parenthesome region since this is influenced by staining, which was quite variable after freeze-substitution in osmium/acetone. The occurrence of rough ER parallel to the septum on either side of the pore has been reported in a range of basidiomycetes (see Moore, 1984). This occurs as perforated sheets, apparently connected to the plasma membrane by somewhat broader filaments (Berbee and Wells, 1988; Lii and McLaughlin, 1991) at a fixed distance from it. The situation was similar in P. tinctorius. Continuity between this rough ER adjacent to the septum and the parenthesome has also been reported from both chemically-fixed and freeze-substituted hyphae (see Moore, 1984; Lii and McLaughlin, 1991). Lumen continuity between the rough ER and the parenthesome of mature septa could not be demonstrated here in P. tinctorius despite searching extensively in serial sections. The ER was always very close to the parenthesome, but the edges of the parenthesome were frequently rounded off and invariably pointed towards the plasma membrane to which they appeared anchored via fine filaments. Our conclusion is that in the mature, apparently fully functional septa of non-senescent cells of P. tinctorius, lumen continuities are either very rare or absent altogether. This does not, however, The dolipore septum in Pisolithus tinctorius 54

exclude a mechanical contact and electron-opaque material frequently bridged the zone between the parenthesome and ER. The parenthesome contained electron-opaque material which showed a characteristic banding pattern. The structure, origin and function of this material is not clear but it is tempting to speculate that it may play a role in anchoring the parenthesome to the filaments and may be important in maintaining its shape. The orientation of the rough ER sheets parallel and at a fixed distance from the septal plasma membrane indicate that the filaments bridging the two are probably acting as anchor points between ER and membrane.

Transport and communication through the dolipore The dolipore frequently appears to be filled by electron-opaque material that is continuous through the pore and spreads across the entrance on both sides. This material is heterogeneous, shows dark transverse bands and appears to contain filaments. It also has some measure of cohesion. Similar ‘occlusion’ of the pores with electron-opaque material with darker transverse bands is described from chemically- fixed (see Moore, 1984) and freeze-substituted septa (Lingle, 1989; Bourrett and McLaughlin, 1986; Lii and McLaughlin, 1991). It varies in appearance and position according to species and fixation conditions, but in general it appears much better preserved after freeze-substitution when the fine filaments and more cross bands are apparent. It is difficult to determine a role for this material in transport since neither its viscosity or permeability are known. No conclusions can be drawn about its role in blocking or controlling the permeability of the channel. It appears to be pushed aside when tubular cisternae pass across the parenthesome pores and enter the mouth of the pore channel (Chapter 3 and Shepherd et al, 1993b). This provides circumstantial evidence that the smooth membrane cisternae do pass through the channel and that the material offers some resistance, at least to larger structures. It has been shown that tubules connected with the vacuole system transfer a fluorochrome across the dolipore septum between adjacent cells (Shepherd et al, 1993b). This transfer appears to be most active during and shortly after completion of the septum. It is possible that this transfer might occur at a stage when the septum was incomplete and the dolipore had not yet formed. However, the data here provide evidence that this is not so. Commonly, movement of tubules occurred for periods of at least 45 minutes (V. A. Shepherd, unpublished observations), while the septum took only about 5 minutes to complete. The dolipore septa in the main hypha and clamp formed synchronously and it is shown here that both are completed by the time the clamp fuses with the main hypha, prior to migration of the nucleus back into the main hypha. Tubule movements continued back and forth across these septa until well after this stage, when the heterokaryon has been restored (Shepherd et al., 1993/7, Shepherd and Ashford, unpublished observations). The dolipore septum in Pisolithus tinctorius 55

Apart from the tubules, organelles are generally excluded from the region between the parenthesome and the pore and so the parenthesome clearly plays a role in keeping the pore free from structures that have the potential to occlude it. At the same time it would appear that the parenthesome is instrumental in controlling the approach of the smooth tubular cisternae towards the septal pore entrance. Under these circumstances the parenthesome should be firmly anchored and its position precisely maintained in relation to the septal pore entrance. This requirement for precise orientation also applies to the spatial relationship between the septal pore entrance and the parenthesome pores. The filaments that occur between the edge of the parenthesome and plasma membrane would be in an ideal position to anchor the parenthesome around the septal pore, while the many filaments that radiate between the pore entrance and the parenthesome membrane at various points could fulfil the function of keeping the parenthesome in its convex configuration, like the spokes of an umbrella. This would maintain both the shape and orientation of the parenthesome and thus maximise control over movement of tubular cisternae towards the pore entrance, from a wide radius across the hypha. The radiating filaments may also be involved in guiding the movement of tubules or cytoplasmic streams towards the pore entrance. In conclusion, the parenthesome may be seen not only as an occluding, filtering device as envisaged by Lti and McLaughlin (1991), but also as a structure that positively guides elements towards the septal pore. For efficient function it is imperative that the parenthesome maintains its shape and position in relation to the pore. It is proposed that the various filaments radiating from it towards the pore and those attached to the plasma membrane all play a role in this. 56 The dolipore septum in Pisolithus tinctorius

Figs 4.1-4.2. Branched hyphae showing two clamp connections. The direction of the hyphal tip is indicated by the arrow labelled T in each case. Bars = 2.0 pm. (1) Scanning electron micrograph of a chemically fixed hypha showing a branch point with two clamp connections (C). Arrows indicate the direction of growth of the clamps. The predicted positions of the septa in the main axis (Si) and its clamp connection (S2), and of the branch (S3) and its clamp (S4) are indicated. (2) Transmission electron micrograph of a freeze-substituted hypha sectioned in a similar orientation to that in (1) to show the actual positions of the septa which are labelled Si, S2 etc. Three (Si, S3 and S4) are section more or less in the median plane showing mature dolipores and their parenthesomes (P). Vacuoles (V) are prominent in all three cells. Cistemae of rough ER lie alongside and parallel to each septum and extend for some distance along the longitudinal wall in some cases. The surface of the hypha has an irregular coating of fibrillar material (EC) radiating from its surface; this is common in freeze-substituted material but is mostly removed during conventional specimen preparation. ' wtSSt

7 HU m v

.* ^ i vw fcf 1v, 57 The dolipore septum in Pisolithus tinctorius

Figs 4.3-4.10. Transmission electron micrographs from two sets of serial sections of freeze-substituted hyphae cut longitudinally through the clamp connection to show early stages of development of the two septa. (3) The two septa are laid down more or less synchronously after the clamp has formed and the nucleus (n) has migrated into it, but before its tip has fused with the penultimate cell. No septal pore apparatus is present at this stage and there are no rough ER cisternae lying alongside the septum. The septum across the main hypha (Si) is in tangential section at the edge of the pore, while that across the base of the clamp (S2) is more or less median. Profiles of mitochondria (m) and multivesicular bodies (mvb) are abundant and there are short ER cisternal profiles throughout the area. Microtubules (mt) near the septum (S1) are in various orientations. Filaments (f) in longitudinal profile occur in a band across the pore of S1, while clusters of filaments (f) in transverse section occur adjacent to the innermost part of the invaginating membrane in S2. Small circular profiles of microvesicles (mv) with electron-opaque contents are common. Bar = 1 pm. (4) Another septum sectioned close to the median plane shows the transversely sectioned filament (f) clusters in greater detail. A membranous reticulum (r) is present at the apex of the invaginated septal membrane and microvesicles (mv) also occur in the vicinity. Profiles of mitochondria (m) and multivesicular bodies (mvb) can also be seen. Bar = 0.5 pm. (5) Tangential view of a septum, more or less at the edge of the pore. The filaments (f) are now mostly in longitudinal profile. They extend across the pore and along the flanks of the newly developed septum showing a more complex pattern around the apex of the infurrowing membrane. The membrane becomes indistinct at its innermost point and in one area (arrowheads) there appears to be a semicircle of fine filaments radiating from it at its tip. Bar = 0.25 pm. (6) Section showing the complexity of arrangement of fine filaments around the edge of the septum (arrowheads). Two filaments, partly beaded structures, follow the contour of the membrane in a concentric pattern, while radiating spokes perpendicular to the membrane pass across these. One of thesj spokes passes through a circular profile of electron opaque material, and there are other circular profiles following the same concentric pattern. Bar = 0.25 pm. (7) Section showing electron-opaque microvesicles on both sides of the developing septum. One of these (mv) is delimited by a distinct unit membrane. Bar = 0.25 pm. (8) Multivesicular bodies (mvb) in the vicinity of the septum are delimited by a characteristic unit membrane with a central electron-lucent region flanked on either side by a layer of electron-opaque material. The internal ‘vesicles’ do not show this, but are bounded by a single electron-opaque layer which commonly has short electron-opaque ‘spokes’ radiating from it. Bar = 0.25 pm. (9) Dilated sheet-like membranous cisternae extend finger-like processes around the apex of extending membrane of the septum (arrowheads). Bar = 0.25 pm (10) Similar membrane-enclosed sheets (arrowheads) occur around the membrane apex and link up wit) smooth tubular cisternae (t) which extend away from the septum. Bar = 0.25 pm. #* v'liilaiM

^Sip^-: -Sll

■^fhvb . : .-^j.,, 4’4*^PPJ •• »> ?>*« W*fcwm. ip Gift -,: JM

fe^c

*/ +-mv 58 The dolipore septum in Pisolithus tinctorius

Figs 4.11-4.16. Serial sections through a septum of a freeze-substituted hypha late in development when the pore is almost complete. Bars = 0.5 pm except Fig. 4.11. (11) Although the septum is almost complete, the clamp, which contains a nucleus (n), has not yet fused with the main hypha. Mitochondrial profiles occur on both sides of the narrowing pore and are connected by a narrow thread (arrowhead) across the pore. The small size of the pore is indicated in Figs 4.12-4.16, which are consecutive sections in a series, only two of which pass through the pore. None of these sections show the parenthesome typical of mature dolipore septa. A tubular cisterna (tc) is detected lying alongside the septum in two of the sections. Bar =1.0 pm. (12) Section adjacent to the pore. Note the electron-opaque material (arrowheads) on either side of the septum more or less at its midpoint and the tubular cisternae (tc) near the septum. (13) Section through the pore shows electron opaque material (arrowhead) in the pore and spreading on either side. There is no evidence of microtubules or microfilamentous bundles running longitudinally through the pore. There is no parenthesome and no septal pore swelling. (14) An enlargement of Fig. 4.11 showing the bridge through the pore (arrowhead) connecting the mitochondrial profiles (m) on either side of the septum. The pore is 140 nm wide. A tubular cisterna (tc) is present at the junction of the septum with the longitudinal wall. (15) The next section again shows the septum as continuous wall, indicating that the outer edge of the pore has been reached. Electron-opaque material (arrowheads) spreads across the wall on either side of the septum close to its mid-point. The tubular cisterna (tc) may again be detected and appears continuous for some distance along the septum. (16) The next section also shows continuous wall but with little electron-opaque material adjacent to it. V '

» ' . > 59 The dolipore septum in Pisolithus tinctorius

Figs 4.17-4.20. Ultrastructure of fully-developed dolipore septa freeze-substituted in either 2% OSO4 in acetone (Figs 4.17-4.19) or 20% acrolein in diethyl ether (Fig. 4.20). Bars = 0.25 p.m. (17) Median section through a pore. The septum (s) is swollen around the septal pore. The plasma membrane (pm) is continuous through the septal pore, which contains electron-opaque material. This material spreads out on either side over the septal pore swelling. The diameter of the pore at its mid­ point is 106 nm while at the entrance the pore diameter is only 44 nm. Rough ER (er) with occasional ribosomes lies parallel to the septum and is not continuous with the parenthesome (p) in this section. The parenthesome shows concentric fine bands of electron-opaque material alternate with areas of lower electron-opacity and it is assumed that this parenthesome is sectioned just at the edge of several pores. At one of its ends, the parenthesome is rounded (arrow) and appears to have fine threads radiating from it. There are small electron-opaque deposits on the plasma membrane opposite to the end of the parenthesome on each side of the pore; these are most clear in the regions labelled with arrowheads. In some parts, the septal wall swelling has a ‘bubbled’ appearance at its edge (double arrow), which may be an artefact where the wall has retracted away from the outer leaflet of the plasma membrane. (18) The parenthesome contains alternating electron-opaque and -lucent bands, the outermost group of which is presumed to be the limiting unit membrane. These bands are occasionally spanned by transverse filaments (tf). The parenthesome surface carries an irregular deposit of electron-opaque material; this is most obvious on the septal side, where occasionally it extends into fine radiating strands of electron-opaque material (St). At one end (double arrow) the parenthesome is rounded. Small electron-opaque deposits (arrowheads) on the plasma membrane are again found opposite to the ends of the parenthesome, and fine filaments appear to attach the parenthesome ends to these deposits (eg. arrow). Other filaments (f) radiate from the parenthesome at the edge of one of its pores. Several transverse electron-opaque bands (b) are seen in the septal pore. Fine filaments run across these longitudinally through the pore. (19) This section passes obliquely through the septal pore. Ribosomes (r) on the rough ER parallel to the septum are present only on the cytoplasmic side. Small filaments connecting the plasma membrane to the rough ER can just be detected (arrowheads). There are also fine, sometimes beaded filaments (arrows) radiating from the region adjacent to the pore mouth towards the parenthesome. (20) In this section of ether/acrolein freeze-substituted material, five distinct electron opaque bands (b) are seen in the septal pore. Some of these are beaded. Again the pore filling material contains longitudinal filaments and these terminate at cross bands. Electron-opaque deposits occur on the plasma membrane adjacent to the ends of the parenthesome and some of these are extended into short filaments (arrowheads). At one point where filamentous electron-opaque pore-filling material (om) is spread over the pore swelling, but appears to have been displaced from the plasma membrane, this material is connected to the end of the parenthesome by a single fine strand (f). The electron-opaque material (om) is continuous across the pore and extends over the plasma membrane on the other side along the septum. There is an electron-lucent layer around the edge of the septum across which there are also very fine transverse filaments. The septal pore swelling is mostly electron-opaque, with an electron-lucent longitudinal band in the mid-point of the wall. t 4 60 The dolipore septum in Pisolithus tinctorius

Figs 4.21-4.26. Serial sections through a freeze-substituted dolipore septum cut approximately perpendicular to the septum. Bars = 0.5 jam. (21) The first section misses the dolipore and shows a septal wall with no gap, bounded by a plasma membrane. Lying on both sides of the septum and parallel to it is rough ER (er). This is discontinuous and is seen in subsequent sections (Figs 4.22-4.26), indicating that it occurs as perforated sheets. (22) The edge of the parenthesome (p) is in this section. It extends to contact the plasma membrane on both sides of the septum. Rough ER is very close to the parenthesome at several points (arrowheads) but no lumen continuity is apparent. On the upper side of the septum, the beginnings of the septal pore swelling are apparent. Several perforations, one of which is hexagonal in profile, can be identified in the parenthesome of the lower cell. (23) The septal pore swelling is now apparent on both sides of the septum. In the upper cell the electron-opaque material can be seen between the parenthesome and the membrane-wall interface. Perforations are seen in the parenthesome (p) of the upper cell. (24) The septal pore swelling now appears larger and electron-opaque material lies in a layer on either side of it beneath the parenthesome; it is spread out to cover the entire area below the parenthesome dome. Fine filaments (arrows) radiate from this electron- opaque material to the parenthesome. (25) At least two filaments (arrows) radiate between the parenthesome and the septum in the lower cell. They are embedded in electron-opaque material and one radiates from the edge of the pore entrance. (26) Median section shows that the electron-opaque material is continuous through the central pore.

61 The dolipore septum in Pisolithus tinctorius

Figs 4.27-4.32. Serial sections through another freeze-substituted dolipore septum, cut approximately parallel to the septum. Bars = 0.5 jam. (27) The first section shows the top of a parenthesome (p) surrounded by cytoplasm. There is a single perforation. (28) The adjacent section shows that the single perforation is surrounded by a second layer of six perforations. Electron-opaque dots in the sub-parenthesome cytoplasm (arrowheads) are interpreted to be filaments sectioned at various angles. (29) In the next section a third layer of perforations is seen. Electron-opaque dots (arrowheads) are again present in the sub-parenthesome cytoplasm. (30) This section just grazes into the sub-parenthesome cytoplasm situated just above the pore, showing electron-opaque filaments radiating from the parenthesome membrane to a mass of electron-opaque material (arrowed region) containing a network of fine filaments. (31) The section shows several filaments (arrowheads). All radiate from what appears to be an inner electron-opaque mass towards the parenthesome. Most contact the parenthesome and in one case the direct contact is seen to be with the electron-opaque material overlying the membrane. (32) The final section passes through the septal pore which is partially occluded by electron-opaque material. A number of electron-opaque circular profiles (arrows) are seen embedded in this material, overlying the septal pore swelling.

Observations on ion redistribution 5 within a specimen during freeze- substitution, embedding and X-ray microanalysis

Introduction

The application of freeze-substitution Freeze-substitution has been used to prevent the redistribution and loss of water-soluble ions in biological samples prepared for energy-dispersive X-ray microanalysis (Marshall, 1980; Harvey, 1982). It is assumed that by rapidly freezing the specimen, the distribution in vivo of diffusable elements is preserved. The existence of elemental gradients in sections of freeze-substituted specimens has been taken as circumstantial evidence that redistribution does not occur during ice removal (substitution), warming or infiltration (Marshall, 1980). However, without knowing the original distribution of elements in the sample, it cannot be known with certainty that preservation of the distribution in vivo has been achieved. The aim of the work described in this chapter is to evaluate the suitability of freeze-substitution and resin embedding for X-ray microanalysis using test samples where the original distribution of ions is known. Particular attention is paid to ensuring that the test samples are prepared and analysed under similar conditions to those commonly used for biological specimens. Some of the principles of freeze-substitution are presented as a background to the rationale of the experimental design and as a basis for the interpretation of the results.

General procedure for freeze-substitution The specimen is rapidly frozen, dehydrated (substituted) at a temperature well below the freezing point of water and embedded in an hydrophobic resin. Cellular water dissolves in, and is replaced by, the substitution solvent during substitution. Since macromolecules and (depending on the choice of substitution solvent) low molecular weight ions are insoluble in the solvent, they co-precipitate and are retained in the specimen (Kellenberger, 1987) which is then infiltrated with resin. The choice of solvent and resin depends on the subsequent use of the specimen. For ultrastructural analyses where the retention of macromolecules is important but the retention of very low molecular weight ions is not considered so, a relatively polar solvent such as acetone or methanol is used. Substitution times of between 48 h and 7 d were sufficient to remove the water from test specimens using acetone (Harvey et al., 1976; 62 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 63

Kellenberger, 1987; Steinbrecht and Muller, 1987) and methanol substituted tritiated ice in less than 8 h (Steinbrecht and Muller, 1987). On the other hand, when the retention of soluble ions is required, a less polar solvent (diethyl ether) is generally used. This requires a longer substitution time (at least 3-4 weeks) because water is less soluble in the apolar solvent. In all cases, the temperature at which substitution occurs, the amount of water initially present in the substitution solvent, the amount of water in the specimen and the duration of warming to room temperature are of particular importance. A fixative (acrolein) is sometimes included (Van Zyl et al., 1976). When a relatively apolar solvent such as diethyl ether is used, molecular sieve is required to drive the substitution process, removing water as it dissolves in the solvent.

Theoretical aspects of freeze-substitution for ion localisation Ions are surrounded by water molecules when in aqueous solution. The electrostatic charge attractions between dissolved ions (including macromolecules) are dampened by the electrostatic charge of water. The ability of a solvent to dampen the charge attractions of its solutes is indicated by its dielectric constant (Table 1), which is a measure of capacitance relative to the capacitance of a vacuum. To retain dissolved ions in a specimen, its thermal energy is first reduced by rapid freezing (to reduce diffusion) and water is replaced by a solvent with a lower dielectric constant. The attraction between ions is greater during substitution and subsequent warming to room temperature than when previously dissolved in water and the ions therefore co­ precipitate. This occurs because the new solvent does not interact with (and thus dampen) the electrostatic charge of the solutes to the same degree as water. The distance over which ions move during precipitation is likely to be small provided there is a matrix of macromolecules with which the ions can precipitate. This precipitation should not limit the resolution of detection if the content of macromolecules is >10% (Zierold and Steinbrecht, 1987). A salt solution containing no macromolecules could not be freeze-substituted without a loss of the structure of the original frozen droplet. Also, a solvent with the lowest possible dielectric constant would be preferable. In this

Table 1. Dielectric constants for various substitution solvents. From Zierold and Steinbrecht (1987).

Solvent Dielectric constant water 80 methanol 33.5 ethanol 25.1 acetone 21.4 diethyl ether 4.3 vacuum 1 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 64

way, the charge attraction between the solutes, and thus ion retention, is optimised. Ions should be retained during freeze-substitution in a solvent such as diethyl ether, owing to its very low dielectric constant. There is a need to test freeze-substitution using a model system in which the original element distribution in the sample is known. If the distribution of elements after freeze-substitution and embedding in epoxy resin is the same as that in the original sample, then it could be assumed that freeze-substitution and subsequent embedding and analytical procedures did not cause detectable redistribution of elements. This is not possible using a biological specimen in which the original distribution of elements is not known. Absolute loss of water soluble ions during freeze-substitution of test specimens has been investigated (Harvey, 1980, 1982). Retention of Na, Cl and K was >97% using diethyl ether as the substitution solvent. An ideal model system may have a number of compartments which contain various water-soluble elements that could be detected using X-ray microanalysis. The preparation of the sample must be reproducible and should not cause the elements in the various compartments to mix prior to freezing. This was a major argument against the use of erythrocyte ghosts or liposomes, as it seemed likely that leakage across the limiting membrane during preparation could not reliably be prevented. To test for ion redistribution in freeze-dried, resin embedded samples, Elder et al. (1988) used frozen droplets of solutions containing dextran and various ions. The frozen droplets were sandwiched together, freeze-dried and resin embedded. Sections were cut at the interface of the two droplets and analysed by X-ray microanalysis. The ions present in one droplet could not be detected in the other droplet and vice versa. Rather than measuring absolute losses from the droplets, this system allowed any detectable redistribution within the specimen to be examined.

Limitations of X-ray microanalysis X-ray microanalysis of thick sections requires a high accelerating voltage (80 - 100 kV for TEM/STEM compared with 10-20 kV for SEM) to avoid specimen damage by heating from trapped electrons. The signal/noise ratio of low molecular weight elements (notably Na) decreases as accelerating voltage increases above approximately 20 kV (Russ, 1978) with a necessary compromise between reducing specimen damage on one hand and optimising the signal/noise ratio of the detectable elements on the other. Under specified operating conditions, the minimum detectable limits for elements of biological interest may be defined for any particular microscope. The minimum dectable limit of an element will be higher for TEM/STEM than for SEM because of the higher accelerating voltage required in the former. For the same reason, X-ray microanalysis using TEM/STEM is unlikely to detect very low levels of redistributed elements if the amount is lower than the minimum detectable limit for that Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 65

element. However, being able to analyse sections of biological material has obvious advantages (most importantly, improved resolution) over the analysis of bulk specimens using SEM.

Experimental rationale The experimental system Elder et al. (1988) used with freeze-dried samples has been adapted here to test whether freeze-substitution and subsequent embedding can cause detectable redistribution of ions within a sample. Droplets of salt/dextran solutions were frozen and sandwiched together, freeze-substituted, embedded and analysed using X-ray microanalysis. Other solutions were frozen in contact with each other so that there was no gap between the different compartments of ions in the specimen. The frozen droplets had a similar concentration of macromolecules to that which may be expected in living cells (Alberts et al., 1989). A relatively high salt concentration was used in order to maximise concentration gradients between the two frozen solutions.

Materials and Methods

Sandwiched droplets. Solutions of 20% (w/v) dextran (Sigma, average molecular weight: 39 100) in distilled water were made containing 500 mmol.L-1 of either NaCl, KH2PO4, CaCl2 or Na2H2P2C>7. Each solution was loaded into a new plastic 5.0 mL syringe fitted with a gauge 20 hypodermic needle and suspended one metre above the polished surface of a liquid nitrogen-cooled copper block (Fig. 5.1 A). The surface of the block was washed with liquid nitrogen-soaked non-absorbent cotton wool and allowed to dry prior to freezing. The surface of the block was approximately 10.0 cm below the rim of the dewar thus preventing contamination of the surface by atmospheric water. For each pair of salts, a single droplet was allowed to fall from the syringe, freezing on contact with the copper block and forming a disc approximately 1 mm thick and 5 mm in diameter. Using fine teflon-coated forceps, the droplet was placed in a small container of liquid nitrogen. Two droplets, each containing different salts, were placed in a small aluminium foil envelope lined with nylon gauze. Droplets were randomly sandwiched together so that the best- or worst-frozen edges were not necessarily adjacent. The droplets were in constant physical contact whilst in the aluminium envelopes, although the surfaces were not perfectly flat because of distortion during freezing, so some droplets were in contact for only part of their surface area. Windows had been cut in the aluminium to facilitate movement of the substitution solvent between the external solution and the frozen Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 66

droplets. The envelopes containing the droplets were transferred to the frozen substitution mixture.

Solutions frozen in contact with each other A solution was frozen on the copper block as previously described. A second solution, either at room temperature, 4°C or 80°C and containing a different salt, was allowed to fall next to the first and overlapping it (Fig. 5.IB). Solutions were also frozen in the reverse order. The combined frozen droplets were transferred to the frozen substitution mixture.

Freeze-dried droplets. Solutions at room temperature were frozen on the copper block in contact with each other (Fig. 5.IB) and freeze-dried overnight at -80°C. The frozen droplets were transferred directly to 100% chloride-free resin and infiltrated in a dry box as described later.

Freeze-substitution The samples were substituted in vials containing either 4 mL of 20% acrolein in diethyl ether or 4 mL of 100% diethyl ether, and approximately 30 pellets of 3A molecular sieve, for 21 days at -85°C. The vials were subsequently stored at -20°C for 24 hours, 4°C for 2 hours and room temperature (approximately 20°C) for 2 hours, before being placed in a dry box. The relative humidity of the dry box was maintained at less than 10% with 3A molecular sieve and by flushing the box with dry nitrogen gas. The droplets were infiltrated for 3-5 days with chloride-free resin (Pallaghy, 1973) which had been stored over molecular sieve for a minimum of 24 hours prior to use. All resin components were stored over molecular sieve prior to mixing. Between six and ten pellets of molecular sieve were placed in each vial during infiltration and were replaced during each daily change of resin. The vials were removed from the dry box following infiltration, the samples placed in fresh resin in flat embedding moulds and immediately transferred to an oven at 70°C for 24 hours. To ensure complete polymerisation, the blocks were stored for up to seven days at room temperature in a desiccator with molecular sieve. The blocks were trimmed and sectioned across the interface of the droplets in a small room where the relative humidity was maintained at approximately 30% with an Ebco Oasis 2700L dehumidifier (Fitzgerald and Allaway, 1991). Blocks sectioned under conditions of uncontrolled humidity were occasionally contaminated with atmospheric water. Sections approximately 1.0 pm thick were cut dry and transferred to formvar-coated nickel slot grids using a mounted eyelash. Most wrinkles in the section were eliminated by pulling the free edge of the section while it was still Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 67

attached to the knife edge. The sections were flattened on the formvar with gentle pressure. All sections were stored in sealed plastic containers with molecular sieve. Sections were carbon-coated using high-voltage evaporation of either carbon rods or carbon thread. This was essential to improve the stability of the sections during exposure to the electron beam, as uncoated sections invariably became dislodged from the formvar support film when the objective aperture was removed for analysis. The thickness of the carbon layer was not measured. Sections were usually analysed within two hours of sectioning.

X-ray microanalysis Analyses were conducted using a Philips CM 12 scanning transmission electron microscope (STEM) and an Edax PV9900 energy-dispersive X-ray analyser. A low background 50 jam condenser aperture was used, grids were placed in a beryllium- tipped low background holder and the objective aperture was removed prior to analysis. An accelerating voltage of 120 kV was used with a static condensor lens current. Analyses were done at either ambient temperature (17°C) or -160°C using a Gatan cold stage. Sections were analyzed at the interface of the two droplets. Where a gap existed between the two discs, sections were analyzed at the most narrow region of this gap. X-ray maps were produced at a resolution of 64 x 50 pixels and 500 ms per pixel (dwell time). Spectra were collected in transects for 50 or 100 seconds (live detector time) per spectrum with the aid of beam control software. The large number of spectra collected necessitated automatic processing using the THIN program (Edax Co.) which calculated peak/background ratios (rather than the preferable (peak- background)/continuum ratios) for the elements of interest. These ratios normalise the data for variations in section thickness and probe current and were plotted against distance to produce graphs of element distribution along the transect. Some sections were examined using an Hitachi H-7000 electron microscope in STEM mode equipped with a Kevex Quantum energy-dispersive X-ray detector. Analyses were conducted at ambient temperature under the conditions of: lOOkV accelerating voltage, static probe current, analysis time 100 seconds (live detector time) for spot analyses and 500-1500 ms dwell time for X-ray maps. The resolution of the X- ray maps was 128 x 64 pixels. Rows of spot analyses were collected with the aid of beam control software and the ratio: (peak-background)/continuum was calculated for elements of interest using the Quantex program (Kevex). The continuum region used in calculations was that from 4.5 to 6 keV, where there were no peaks above background. Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 68

Results

Ultrastructure of freeze-substituted solutions The ultrastructure of the freeze-substituted solutions was variable and this depended on the degree of ice crystal damage and elemental composition of the solutions (Fig. 5.2A). The edge of the solution that froze against the copper block was the best frozen edge and this usually had an homogeneous appearance in the sections after freeze- substitution (Fig. 5.2A). The degree of ICD increased with distance from the frozen edge and resembled a dendritic or herringbone pattern at the poorly frozen edge (Fig. 5.2A). Variation in levels of detectable elements in sections was always greater due to ice crystal damage than any likely variation due to whether or not acrolein was used in the substitution solvent. No difference in ultrastructure or elemental content of the droplets could be attributed to the use of acrolein. However, the qualitative and semi- quantitative nature of the X-ray microanalysis techniques used here did not permit the accurate comparison of this treatment.

Elemental distribution in solutions sandwiched together after freezing The general distribution of elements in sections was determined by X-ray mapping. In a frozen solution containing KH2PO4 sandwiched against a frozen solution containing NaCl (Fig. 5.2A), the distribution of K and P was confined to one droplet (Fig. 5.2D,E) and the distribution of Na and Cl was confined to the other droplet (Fig. 5.2B,C). At the interface, the NaCl droplet had considerable ice crystal damage (Fig. 5.2A) and the distribution of Na and Cl followed the pattern of ice crystal damage, owing to the separation of water and solutes caused by sub-optimal freezing rates. The reflection of this pattern in the X-ray maps indicates no detectable redistribution of elements back into the electron-lucent regions after the initial freezing and substitution. Electron- opaque regions contained high levels of the elements and electron-lucent regions contained low levels (Fig. 5.5). Under prolonged exposure to the electron beam, the electron-opaque regions of the NaCl droplets volatilised leaving small holes in the sections (Figs 5.2A, 5.3A, 5.5A). This was not observed in sections of droplets containing KH2PO4. Droplets containing NaCl frequently contained small electron- opaque deposits which were not formed during freezing and which were present on sections with ice crystal damage (Fig. 5.3) and without (Fig. 5.4). Such deposits were exclusive to NaCl droplets and were enriched with Na but not Cl (compare Fig. 5.3B with Fig. 5.3C and Fig. 5.4B with Fig. 5.4C). X-ray maps of sandwiched droplets containing CaCl2 and Na-pyrophosphate (Fig. 5.7) showed that the original distribution of Na, P, Cl and Ca was preserved during freeze-substitution and embedding. In Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 69

sandwiched droplets containing NaCl and KH2PO4 where the two droplets were touching (Fig. 5.9), the original elemental distribution was also retained. Spectra were collected in a transect across the interface of droplets and processed to remove background radiation. The remaining X-ray counts from the elemental peaks were converted to ratios of either peak/background or (peak- background)/continuum and the values graphed. A transect across the interface of two droplets containing NaCl and KH2PO4 (Fig. 5.6) showed that Na and Cl were retained in one droplet and were not present in the narrow gap between the two or in the other droplet. K and P were retained in the other droplet and were not present in the gap or in the NaCl droplet. Thus the original ion distribution was retained. A similar result was obtained when one of the droplets contained CaCl2 instead of NaCl (Fig. 5.8). Again, there was no cross-contamination of elements from one side to the other.

Elemental distribution in solutions frozen in contact with each other In order to determine whether or not elements could be redistributed within a specimen where there was no gap between the droplets, the second solution was dropped on to a droplet already frozen on the copper block. In this way, two different salt solutions could be freeze-substituted in contact with each other and the interface between them analysed. It was to be expected that some melting would occur when the second solution contacted the first. The behaviour of the solutions was dependent on the temperature of the second solution. For example, when a solution at room temperature was dropped on to a frozen one, the two droplets fused and did not separate during substitution. If the second solution was cooled to 2-4°C prior to freezing, the two droplets did not fuse and usually fell apart during transfer from the copper block to the substitution vial. However, if the second solution was heated to approximately 80°C prior to freezing, the droplets were fused and occasionally the first droplet was distorted by the second. Figure 5.10A depicts a section of a freeze-substituted CaCl2 solution surrounded by a KH2PO4 solution. The CaCl2 was frozen first and the KH2PO4 solution (originally at room temperature) flowed around it during freezing. Ice crystal damage is substantial. The X-ray maps (Fig. 5.10B-E) show that Ca and Cl are localised in one droplet surrounded by K and P belonging to the other droplet. The solutions have not mixed indicating that redistribution did not occur during substitution. In most samples in which two solutions were frozen in contact with each other, a boundary between the two could be easily distinguished (Fig. 5.11 A). Whether there was a sharp line or a diffuse boundary depended on the degree of melting between the solutions, which varied from one part of the boundary to the next within each sample (compare Figs 5.11A and 5.12A). There was less melting (and thus a sharper Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 70

boundary) near the edge of the two droplets rather than where overlap between them was greatest. This was equally true in samples where the temperature of the second droplet was increased or decreased prior to freezing. Analysis of droplets containing KH2PO4 and CaCl2 in a transect across the contact zone (Fig. 5.1 IB) showed that P and K were localised in one droplet and Ca and Cl in the other. All four elements were detected at the boundary indicating that melting occurred between the two solutions during freezing. However, the separation of elements in the two solutions and the rapid fall-off in the element levels from one side of the boundary to the other (Fig 5.1 IB) is evidence that redistribution did not occur during substitution. Distribution of elements was maintained even when the second droplet was heated to 80°C prior to freezing (Fig. 5.12). Analysis across the interface of a NaCl solution frozen in contact with a KH2PO4 solution produced equivocal results (Fig. 5.13). Whilst a distinct boundary may be seen at the interface of the droplets (Fig. 5.13A), Na was present on both sides (Fig. 5.13A and B). However, beyond the limits of the transect, Na was not detected in the KH2PO4 side and was present in higher levels in the NaCl side away from the interface. The loss of Na was also apparent when sections were analysed at -160°C, where specimen damage due to beam exposure should be reduced, indicating that the altered distribution of Na was not likely to be caused by the analysis conditions.

Elemental distribution in freeze-dried droplets Droplets were incompletely dried following freeze-drying overnight. Ice crystal damage with similar herringbone patterns to freeze-substituted droplets was observed in the sections. In some electron-lucent regions of the sections there was no pattern of ice crystal damage and no elements were detected in these regions. These were likely to be areas of the droplets which had not dried and therefore the structure of the droplets was lost. Distribution of elements was otherwise similar to that observed in freeze- substituted droplets. Despite a narrow overlap of elements at the interface, P and K were confined to one side and Ca and Cl to the other (Fig. 5.14).

Discussion

Retention of water soluble ions by freeze-substitution The use of solutions of salt and dextran as a model system is in many ways a ‘worst case’ situation which would be unlikely to exist in a biological specimen given that the salt concentrations were very high and there was no barrier between the pairs of solutions. Mixing would have occurred if the solutions had been in the aqueous phase. However, phosphorus, potassium, sodium, calcium and chlorine were all retained in Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 71

their original position in the droplets after freeze-substitution and embedding and were localised by X-ray microanalysis. The localisation of sodium was more problematic than other elements, however it was successfully localised in a number of instances. Since the original distribution of these water soluble ions was known, the preservation of the original location is evidence that, provided careful precautions are taken, the techniques used are suitable for the qualitative localisation of these elements in biological material. This is the first direct evidence that such localisation is possible after freeze-substitution and embedding.

Interpretation of X-ray microanalysis results One disadvantage of the X-ray maps presented here is the presence of background X- rays. Any point in the section containing high levels of a particular element is mapped as a bright pixel. However, points lacking any particular element will still emit a low level of background radiation, shown as a dim pixel on the map. When only low levels of an element are present, the peak/background ratio (ie. signal/noise ratio) is low and the map looks 'noisy'. Sodium had the lowest peak/background ratio of any of the elements detected (because of its low atomic weight and the high accelerating voltage used). Consequently X-rays emitted by sodium may not be appreciably higher than the background. Background radiation is often more prominent in maps of sodium distribution as a result. To further confound the interpretation of X-ray maps, the level of background is higher in any region of the sample compared to regions of pure resin because of the higher density of the sample. This accounts for the faint image in the map of sodium in droplets not containing sodium when there are only very low levels detectable in the adjacent droplet. A more advanced system for elemental mapping is used by Ingram et al. (1989) to collect ‘continuum maps’ in addition to elemental maps of sections through the use of specially written computer software. The continuum maps are subtracted from the element maps and the resultant image is calibrated against standards to give quantitative elemental images. Occasionally sodium could not be detected at the interface of droplets which were known to contain it despite being present in original solutions at a very high concentration (500 mM). In these cases, it was detectable in more ice crystal damaged regions further away from the interface. This is an anomalous and puzzling result as sodium could be detected with great accuracy in some sections.

Beam damage Sections irradiated during analysis often became ‘burnt’, leaving small holes in the sections. This was almost always observed in droplets containing NaCl but rarely in droplets containing any other salt. The effect of this exposure was to reduce the Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 72

contrast of the NaCl droplet until it was barely visible. This effect was more pronounced in sections that were not ice crystal damaged. The electron-opaque herringbone pattern characteristic of ice crystal damaged droplets was more resistant to the effect of the electron beam. Additionally, small circular electron-opaque deposits sometimes formed during irradiation of sections which contained sodium. These deposits did not contain chlorine and occurred on droplets with and without ice crystal damage. Specimen damage as a result of irradiation by the electron beam causes damage by heating and damage by ionising radiation. Both are likely to affect the elemental composition of the sections significantly. As outlined in the Introduction, the optimal accelerating voltage for the detection of elements of biological interest is around 20 kV (Russ, 1978). However, the need to have good penetration of the section to reduce heating damage and permit visualisation of the specimen requires a much higher accelerating voltage of 80-100 kV: 4-5 times in excess of that required for optimum detection. Consequently, minimising heating damage to the specimen also decreases the ability to detect the component elements. This is especially the case for sodium which is at the lower limit of element detection by X-ray microanalysis. The excess energy causes radiation damage which contributes to the loss of particular elements, notably sulphur and chlorine (Echlin, 1992). Specimen damage can be avoided, or at least minimised, by analysis at a very low temperature (von Zglinicki and Uhrik, 1988). While the analysis at -160°C of droplets containing NaCl did significantly reduce the amount of ‘etching’ or visible loss of material from sections, sodium could still not be reliably detected in regions which were not ice crystal damaged, indicating that some movement must have occurred prior to analysis.

Contamination of sections with atmospheric water Contamination of sections by atmospheric water was an intermittent problem. It is likely that water from the atmosphere will be attracted preferentially to the sections given their high concentrations of salts. In some cases, water was actually visible on the surface of sections in the microscope, despite keeping the sections in a dry environment until just prior to insertion into the microscope. In these cases, there was redistribution of elements and loss of ultrastructure of the droplets. While every attempt was made to reduce exposure of the sections to water, this may have occurred during transfer from the storage container to the microscope or even during carbon coating. Coating sections with carbon before any exposure to the atmosphere is apparently the best preventative measure against water contamination (von Zglinicki, 1989). However, during the evacuation of the chamber prior to carbon coating, a cloud of water vapour often condensed due to the sudden drop in pressure. This cloud was a likely source of water contamination and was not produced when a less powerful pump Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 73

was used. This risk should be assessed before carbon-coating is used. Contamination of sections by atmospheric water during specimen transfer was identified by Hagler and Buja (1986) as the cause of an artificial increase in the sodium content of freeze-dried cryosections of isolated cardiac myocytes. This is a likely explanation for the anomalous distribution of sodium in some sections. The use of a microtome which can be flushed with dry air or nitrogen gas (cryomicrotomes often have this facility) and specimen transfer devices which prevent the contamination of sections with atmospheric water (cryotransfer devices would be suitable) may be the only solution to the problem of redistribution of ions in sections prior to analysis. Freeze-substitution, embedding and X-ray microanalysis was successful in preserving ions at their original site in sections and gave similar results to those of freeze-dried specimens. Freeze-substitution works well for the localisation of phosphorus, potassium and calcium which are of particular interest in the present study of Pisolithus tinctorius. Sodium was localised successfully in some cases despite occasional problems encountered with water contamination during section transfer and beam damage during analysis. These led to the occasional redistribution of sodium. The fact that sodium could be localised in the majority of cases indicates that it is most likely that freeze-substitution and embedding did not cause redistribution per se, as this would have resulted in all sections having similar results with respect to sodium. A paper describing an application of this freeze-substitution technique (Fitzgerald et al., 1992) is in the Appendix. There, it is shown that sodium chloride localisation was successful in grey mangrove salt glands following freeze-substitution. It is concluded that the freeze-substitution technique described here is suitable for ion localisation in biological specimens provided stringent anhydrous specimen preparation and analysis conditions are maintained. 74 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.1. Diagram of the experimental setup for preparing frozen salt/dextran solutions (not to scale). (A) Two droplets containing different salt solutions were frozen separately, sandwiched together in an aluminium envelope and placed in a vial containing the substitution mixture. (B) Alternatively, after one solution was frozen, the second was dropped adjacent to the first so that they overlapped. The combined frozen droplets were placed in a vial containing the substitution mixture. I

i—i (A)

^AvvvVvVvW; #LN2«i •VvV/^\vVv\v':\-v mmmm.mrnrn

(B) 75 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.2. (A) The interface of a KH2PO4 (left) and a NaCl (right) droplet sandwiched together and freeze-substituted in 20% acrolein in diethyl ether (digitised STEM image). Bar = 20 pm. The flat edge (ie. the edge that hit the copper block) of the KH2PO4 droplet at this interface was sandwiched against the curved (upper) edge of the NaCl droplet. The KH2PO4 side shows no evidence of ice crystal damage and has an homogeneous appearance. The NaCl droplet has considerable ice crystal damage indicated by the characteristic herringbone pattern formed by the separation of the solutes from the water during ice crystal growth. There is a narrow gap between the droplets.

(B-E) Energy dispersive X-ray maps showing the distribution of Na, Cl, K and P in the section shown in (A). Na and Cl are co-localised in one side and K and P are co­ localised in the other side. The distribution of Na and Cl follows the pattern of ice crystal damage seen in the right side of (A). The distribution of K and P is more homogeneous. There is a sharp edge in the distribution of each element corresponding to the edge of each droplet.

76 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.3. (A) Interface of a KH2PO4 (left) and NaCl (right) droplet sandwiched together and freeze-substituted in 20% acrolein in diethyl ether (digitised STEM image). Bar = 20 |im. Again the best frozen edge of the KH2PO4 droplet is sandwiched against the poorly frozen edge of the NaCl droplet. There are small electron opaque deposits (>►) on the NaCl side. The KH2PO4 side is homogeneous.

(B-E) X-ray maps corresponding to (A) showing the distribution of Na, Cl, K and P. Na and Cl are localised in one side and K and P in the other. There is a sharp interface between the two droplets. The distribution is similar to the pattern of ice crystal damage. Deposits of Na (>) correspond with electron opaque deposits shown in (A).

77 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.4. (A) Interface of a KH2PO4 (top) and NaCl (bottom) droplet sandwiched together and freeze-substituted (digitised STEM image). Bar = 20 |Lim. In this case, the best frozen edges of each droplet are facing each other. The KH2PO4 droplet is electron opaque and has a more opaque edge. The NaCl side has numerous small electron opaque deposits (►).

(B-E) X-ray maps for Na, Cl, K and P corresponding to the area shown in (A). There is very little Na in the lower droplet. Deposits of Na (>) correspond in position to the electron opaque deposits seen in (A). Chlorine is not similarly localised in these sites; rather, it is homogeneously distributed throughout the lower droplet. K and P are localised in the upper droplet with higher levels at the edge corresponding in position to the electron opaque edge in (A).

78 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.5. (A) The interface of a NaCl (right) and KH2PO4 (left) droplet at higher magnification (TEM). Bar = 5.0 pm. The NaCl droplet has been affected by the electron beam. The material excluded during ice crystal growth aggregated into small electron opaque deposits. The KH2PO4 droplet has not been noticeably affected in this way.

(B-E) X-ray maps of Na, Cl, K and P for the region shown in (A). Na and Cl are localised in the right side and K and P are localised in the left side. There is a sharp edge at the interface of each droplet and there are no elements in the gap.

79 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.6. (A) The interface of a KH2PO4 (left) and NaCl (right) droplet sandwiched together and freeze-substituted in 20% acrolein in diethyl ether (TEM). Bar = 10 |im, The row of spots are contamination deposits from the analysis of a transect of points from one droplet to the other across the interface. There was a gap between the two droplets and there were two analysis points in this gap.

B

CQ I ^ 15

0 w iwMHa w ■ m 1 0 5 11 16 22 27 33 38 44 49 55 60 66

Distance (pm)

------■------Na ------n------P ------♦------Cl ------O------K

(B) A graph of (peak-background)/continuum ratios ((P-B)/C) calculated for Na, Cl, K and P at each point analysed. The arrows indicate the two analysis points in the gap between the droplets. Levels of K and P are variable due to the ice crystal damage. These elements are present in the left side but not in the gap and not in the right side. Levels of Na and Cl are not as variable. These elements are present in the right hand droplet but not in the gap or the left hand side.

80 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.7. (A) The interface of CaCl2 (left) and Na2H2?207 (right) freeze-substituted droplets sandwiched together after freezing. Both sides have considerable ice crystal damage. Bar = 10 jam.

(B-E) X-ray maps showing the distribution of Na, P, Cl and Ca in (A). Na and P are localised in the right side and Ca and Cl are localised in the left side. The distribution of each element follows the pattern of ice crystal growth.

81 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.8. (A) Two droplets containing KH2PO4 (left) and CaCl2 (right) freeze- substituted in 20% acrolein in ether (TEM). Bar = 10 pm. The well frozen edges of each droplet are sandwiched together. The row of spots are the analysis points from spectra collected in a transect across the interface. Each point is 1.3 pm apart and there is one analysis point in the gap between the droplets.

B

Distance (pm)

------□------P ------♦------Cl ------0------K ------•------Ca

(B) A graph of (peak-background)/continuum ratios ((P-B)/C) showing the levels of Ca, Cl, K and P at each point in the transect. The arrow indicates the analysis point in the gap between the droplets. There are high levels of K and P at the left side of the transect but not in the gap and not in the right side. Similarly Ca and Cl are present in the right hand side but not in the gap and not in the left hand side.

82 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.9. (A) Interface of two droplets containing NaCl (right) and KH2PO4 (left). The ice crystal damaged edges of each droplet are in contact with each other in this section. Bar = 3.0 pm

(B-E) X-ray maps for Na, Cl, K and P from the section in (A). The Na and Cl are localised in the right droplet. There is some signal in the left side of the Na map but this is attributed to background and is visible as a result of the low signal/noise ratio of Na (see text). K and P are localised in the left droplet. The distribution of each element follows the pattern of ice crystal damage.

83 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.10. (A) STEM image of CaCl2 and KH2PO4 solutions frozen in contact with each other. Both solutions were at room temperature prior to freezing. Bar = 20 pm. The edge of the CaC^ droplet is shown O) and is surrounded by KH2PO4.

(B-E) X-ray maps showing the distribution of K, P, Ca and Cl. The distributions of K and P and Ca and Cl are exclusive of each other indicating that the original distribution has been retained.

84 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.11. (A) TEM of CaCl2 and KH2PO4 solutions frozen in contact with each other. Both solutions were at room temperature prior to freezing. There is a diffuse boundary between the two solutions indicative of a small degree of melting. The dark lines on the section are a result of contamination during raster scanning of the section prior to analysis. The accompanying transect is from the upper row of analysis points in this section. Bar = 2.0 jam.

B

8 9 10 11 12 13 14 15 16 17 18 19 20

Distance (pm)

------O------p ------♦------Cl ------O------K ------•------Ca

(B) A transect of peak/background ratios (P/B) for Ca, Cl, K and P. At the boundary point (arrow) all four elements are present but they fall off rapidly away from the interface.

85 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.12. (A) STEM image of CaCl2 (left) and KH2PO4 (right) solutions frozen in contact with each other. Bar = 10 pm The CaCl2 solution was frozen first, then the KH2PO4 solution which was heated to 80°C before being frozen. The transect shown is from the upper row of analysis points seen in the micrograph. There is one analysis point (arrow) on the electron-opaque interface of the two droplets. The large particle in the centre contained A1 and Si and is interpreted to be either a piece of molecular sieve or a small grain of sand (a common component of dust).

B

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Distance (pm)

------□------p ------♦—Cl ------0------K ------•------Ca

(B) A transect of peak/background ratios (P/B) for Ca, Cl, K and P across the interface of the two droplets. The arrow indicates the analysis point at the interface of the two droplets. All four elements are present at the interface. Ca and Cl are present in the left hand side of the transect but fall off rapidly across the interface and are at very low levels in the right hand side. There is considerable variation in the levels of Ca and Cl due to ice crystal damage. K and P were barely detectable in the left hand side but were present in the right side. I 86 Ion redistribution during freeze-substitution, embedding and X-ray microanalysis

Fig. 5.13. (A) TEM of a NaCl solution (left) frozen in contact with a KH2PO4 solution (right). Both solutions were at room temperature prior to freezing. There is a distinct boundary between the solutions. The transect shown (B) is from the lower row of analysis points seen in the micrograph. Bar = 2.0 pm.

B

Distance (pm)

------■------Na ------°------P ------♦------Cl ------0------K

(B) Transect across the interface showing peak/background ratios (P/B) of Na, Cl, K and P. The arrow indicates the position of the interface of the two droplets. The elements are not as well separated as the previous figure. P is present in the right droplet, falls off rapidly over the boundary and is not detectable in the left side. K decreses over the boundary too but it is still present in very low levels in the left hand side (note that the vertical scale of this graph is much more exaggerated than that of the previous figure). The level of Cl in the left side is variable due to ice crystal damage. It falls off gradually over the interface and is present in low levels in the right side. Very low levels of Na were detected throughout the entire transect. Na was not preferentially associated with the side also containing Cl.

Ion redistribution during freeze-substitution, embedding and X-ray microanalysis 87

6 7 9 10 11 12 13 14 15 16 17 18 19

Distance (pm)

------□------P ------♦------Cl ------0------K ------•------Ca

Fig. 5.14. Transect across the interface of a solution of KH2PO4 frozen in contact with a CaCl2 solution and freeze-dried. Both solutions were at room temperature prior to freezing. The section analysed here was dislodged from the grid into the column of the microscope before a photograph was taken. The arrow indicates the position of the interface between the two droplets. The peak/background ratios (P/B) for K, P, Ca and Cl are plotted. Ca and Cl are present in the left side and in the analysis point at the interface. They are barely detectable in the other droplet. K and P are present in the right side and at the interface but not in the left side. The variation in levels of K and P are due to ice crystal damage. Symplastic transport in 6 Pisolithus tinctorius

Introduction The results described in this thesis have important implications for the understanding of intra- and intercellular transport in mycorrhizas. Transport of nutrients through the fungal sheath to the host plant is likely to involve at least some transport through the fungal symplast. The uptake of water-soluble fluorochromes by mycorrhizal roots has been shown to be blocked in Eucalyptus mycorrhizas at the surface of the sheath (Ashford et al., 1989) indicating that the walls and interhyphal spaces have lower permeability than non-mycorrhizal fungal or plant cell walls. It is likely that dissolved nutrients enter the fungal symplast at the sheath surface, if they have not done so by entering hyphae in the soil mycelium. An understanding of mechanisms for translocation through the fungal symplast is essential to fully understand the process of nutrient uptake by mycorrhizas, which may involve transport from growing hyphae in the soil some distance away from the root/mycorrhiza surface, through the soil mycelium and sheath to the fungus/plant interface.

Symplastic translocation in fungi Mechanisms for symplastic translocation in fungi are discussed by Harley and Smith (1983) and Jennings (1987, 1989). Translocation of phosphorus through the hyphae of mycorrhizal fungi is not likely to occur by diffusion alone unless the diffusion pathway is less than 40 jam long (Harley and Smith, 1983). While diffusion may be sufficient to account for transport in some situations, such as at cell junctions, it is generally considered that this is not a satisfactory hypothesis to account for observed phosphorus flux through hyphae for distances greater than 40 pm. An alternative mechanism to simple diffusion suggested by Cox et al. (1975, 1980) and Callow et al. (1978) involved the quantised transport of P in VA mycorrhizas via vacuoles containing polyphosphate granules propelled by cytoplasmic streaming. They suggested that vacuolar granules are unloaded near the fungus-host interface. This was criticised by Harley and Smith (1983) and they argued that if it is assumed that individual polyphosphate-containing vacuoles are moved by cytoplasmic streaming, they will be distributed randomly and there would be no longitudinal progression. Thus the vacuoles do not represent a separate, ‘tight’ compartment of P. The mechanism of translocation of P in a separate vacuolar compartment can be supported either if (i) the polyphosphate is not in equilibrium with cytoplasmic P; and (ii) polyphosphate-containing vacuoles are not distributed at random

88 Symplastic transport in Pisolithus tinctorius 89

(Harley and Smith, 1983). Data presented here indicates that vacuoles do not move randomly in P. tinctorius hyphae: vacuoles are capable of complex bidirectional intra- and intercellular movement with concomitant transfer of vacuolar content, apparently moving independently of cytoplasmic streaming. Thus it is possible that translocation of material in vacuoles including P can be controlled by directed movements of the vacuole system. The system has not been demonstrated in VA mycorrhizal fungi but it is present in closely related zygomycetes (Rees et al., 1994). It is of particular importance to hypotheses of P translocation that the vacuole movements are independent of cytoplasmic streaming. It was previously thought that translocation of Pt- occurs by the establishment of a source-sink gradient through the cytoplasm, the vacuoles having only a storage role. Based on the results in this thesis, an alternative hypothesis is suggested: P, is collected near the tip region of hyphae by clusters of small vacuoles. Some of the P, is converted into polyphosphate and the entire vacuolar content is translocated along hyphae by means of bulk transfer from one vacuole to the next. It is likely that both P; and polyphosphate are transported. As cells age, the vacuoles coalesce and develop attachments to the plasma membrane. However, they are still capable of movement and interconnection. Movements of the vacuole system may be under the control of the cytoskeleton but this conclusion awaits confirmation using immunolabelling at the LM or EM level. It is likely that maintaining cytoplasmic P, concentration and pH are important functions in cells close to the hyphal tip. As cells age, the function of the vacuole as a storage site would become more important. This is especially the case in axenic cultures where there is not likely to be a large sink for nutrients away from the growing front of the mycelium.

Compartmentation of P, into vacuoles in hyphal tips Cells at the hyphal tip contain vacuole clusters and tubules. The vacuole clusters have a higher surface area/volume ratio than a single vacuole and therefore would be well suited to collect excess P; from the cytoplasm. A possible explanation for the saltatory movements of the vacuole clusters is that they may be moving in response to localised concentration gradients generated by the uptake of elements into the vacuole. The vacuolar content is transferred along the hypha via elongate vacuoles between clusters and from cell to cell. The tubule movements between cells are bidirectional (Shepherd et al., 1993b) indicating that transfer of vacuolar content in the cells near the hyphal tip is also likely to be bidirectional.

Storage and translocation in large vacuoles Cells behind the growing tip often contained larger vacuoles which were still capable of movement although many were frequently attached to the cell periphery. They contained Symplastic transport in Pisolithus tinctorius 90

polyphosphate and potassium. Phosphorus-containing vacuoles extending from tip cells back along the hypha would seem to be interconnected to form a continuum. The need to regulate cytoplasmic P; concentration makes the uptake of P, into vacuole clusters of paramount importance. In axenic cultures used here, the concentration of P,• in the modified Melin Norkrans growth medium is likely to be far in excess of that required for cell metabolism. Once the vacuole clusters have collected Pj, it is in a compartment shown here to be capable of cell to cell transport by what is apparently a non-random mechanism. This is evidence of a new symplastic pathway in fungal hyphae. By conversion into polyphosphate, the osmotic effect of transporting P/ is reduced (Dawes and Senior, 1973) and this may increase the capacity of the vacuole for P/ uptake from the cytoplasm. However, the amount of polyphosphate produced in hyphal tips of different fungi varies according to conditions of growth. In Hebeloma cylindrosporum grown in pure culture, only 4-9% of the total P was converted into polyphosphate (Rolin et al, 1984), whereas up to 30% of total P was converted in experiments on beech ectomycorrhizas (Harley and Me Cready, 1981) and 40% in V-A mycorrhizas (Callow et al., 1978).

Translocation in mycorrhizas The fungal cytoplasm is in direct contact with the apoplast in the sheath and at the fungus- host interface. Therefore it is assumed that the cytoplasm is the most likely compartment through which nutrients are transferred in mycorrhizas (eg. Ashford et al., 1989). However from X-ray analysis data, the major site of P and K accumulation was the vacuole, not the cytoplasm, which is as would be expected (Loughman and Ratcliffe, 1984). There are two sites where these nutrients must pass through the cytoplasm: (i) at the site of uptake - likely to be at hyphal tips and (ii) at the plant-fungus interface in the Hartig net, where nutrients are released into the sealed apoplastic compartment. It is difficult to relate directly data obtained from studying hyphae grown in axenic culture to a mycorrhizal association. The state of fungal cells in different areas of the mycorrhiza (for example the Hartig net compared with the inner or outer sheath) is likely to vary and direct comparison with either growing hyphal tips containing vacuole clusters or older hyphae with peripheral vacuoles cannot be made without reference to freeze- substituted mycorrhizal roots. It is thought that the hyphae forming the Hartig net are likely to be less metabolically active as they are not growing rapidly (Rolin et al., 1984; Cairney and Smith, 1993). Hyphal tips were not observed amongst Hartig net or sheath cells of pine ectomycorrhizas (Timonen et al., 1993). Preliminary results from the freeze-substitution of a Eucalyptus pilularis-Pisolithus tinctorius ectomycorrhiza (Fig.

Symplastic transport in Pisolithus tinctorius 91

6.1) indicate that cells in the sheath have vacuoles similar to those in more mature cells behind the growing tip of axenic cultures. It must be pointed out that this is a preliminary result and observations were few. These vacuoles had content similar to those in axenic mycelium preserved by freeze-substitution and did not contain polyphosphate granules. Cells with tubular vacuoles resembling those of hyphal tip cells were not found. Many cells contained large vacuoles with less electron-opaque material than those from axenic cultures. This could indicate a larger amount of vacuolar content in axenic cultures (eg. phenolics, polyphosphate etc) or be due to the difference in the freeze-substitution protocols used. The fibrillar nature of the cytoplasm of some sheath cells was similar to that of older cells in axenic cultures.

Mechanism of translocation to the Hartig net The pathway for unloading at the interface is unknown. At the interface it is thought that the release of P, is stimulated by high concentrations of monovalent cations (Caimey and Smith, 1993). This release is likely to come from the cytoplasm as it is in contact, via the plasma membrane, with the interfacial space. Current models of nutrient translocation (eg. Ashford et al, 1989) suggest that nutrients exchanged at the interface are replenished by transfer across the sheath through the cytoplasm of the adjacent fungal cell via the septum. However, the highest concentration and closest source of P, in the Hartig net cells is the vacuole in hyphae of the interface (eg. the Hartig net cells), not the cytoplasm of adjacent cells. Polyphosphate is known to accumulate in vacuoles of Hartig net cells in Eucalyptus ectomycorrhizas (Ashford et al., 1975) and beech ectomycorrhizas (Chilvers and Harley, 1980). It is discussed in Chapter 2 that a large flux of P, through the cytoplasm would interfere with fundamental metabolic processes (eg. glycolysis). It is paradoxical that large amounts of polyphosphate accumulate in the Hartig net cells

Fig. 6.1. Transmission electron micrograph of part of the sheath from a freeze- substituted Eucalyptus pilularis-Pisolithus tinctorius ectomycorrhiza. Some cells at the inner sheath (left) are embedded in a layer of homogeneous electron-opaque material (T) similar in appearance to the tannin layer in E. fastigata mycorrhizas (Ashford et al., 1989). Cells in the sheath frequently contained large vacuoles (v) which occasionally were associated with the cell periphery (v*). The vacuoles contained a fine matrix of evenly dispersed material and did not contain polyphosphate granules. Some cell profiles did not contain large vacuoles. The cytoplasm (C) of these cells consisted of dispersed electron-opaque material and had few organelles (notably mitochondria (m)) and deposits of glycogen (gl). Ectomycorrhiza synthesised in a growth pouch (see Grenville et al., 1986) by Dr W. G. Allaway. Frozen in super critical nitrogen using an Oxford Instruments freezer. Freeze-substituted in 20% acrolein in diethyl ether for 21 d at -70° C. Embedded in resin (Spurr, 1969) and triple stained (Daddow, 1983). Bar = 2.0 pm. Symplastic transport in Pisolithus tinctorius 92

when it would be expected that these cells have a lower cytoplasmic P, concentration as a result of being adjacent to the sink at the interfacial apoplast. This could be explained if polyphosphate is supplied to Hartig net cells via the vacuole system rather than the cytoplasm of adjacent cells. As P/ is lost from the Hartig net cytoplasm to the interface, it is readily replaced by the hydrolysis of vacuolar polyphosphate. If similar processes of vacuole movement are present in the sheath as have been seen in hyphal tips, they provide evidence for a vacuole transfer mechanism in mycorrhizas: nutrients including phosphorus (as Pt and polyphosphate) and potassium are supplied to the plant at the fungus-host interface by transfer across the sheath through interconnected vacuoles from the sites of uptake of potassium and P;. In addition to the accepted roles of storage and osmoregulation, the complex movements of the vacuole system in P. tinctorius and ability for intra- and intercellular transport indicate that vacuoles have a crucial role in nutrient transport in ectomycorrhizal fungi. References

Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. (1989). Molecular Biology of the Cell, Second Edition. Garland Publishing, Inc., New York, London. Alexopoulos, C. J. and Mims C. W. (1979). Introductory Mycology, Third Edition. Wiley, New York. Antibus, R. K., Croxdale, J. G, Miller, O. K. and Linkins, A. E. (1981). Ectomycorrhizal fungi of Salix rotundifolia. III. Resynthesized mycorrhizal complexes and their surface phosphatase activities. Canadian Journal of Botany 59, 2458-2465. Ashford, A. E., Allaway, W. G., Peterson, C. A. and Cairney, J. W. G. (1989). Nutrient transfer and the fungus root interface. Australian Journal of Plant Physiology 16, 85-97. Ashford, A. E., Ling-Lee, M. and Chilvers, G. A. (1975). Polyphosphate in eucalypt mycorrhizas: a cytochemical demonstration. New Phytologist 74, 447-453. Ashford, A. E., Peterson, R. L., Dwarte, D. and Chilvers, G. A. (1986). Polyphosphate granules in eucalypt mycorrhizas: determination by energy dispersive X-ray microanalysis. Canadian Journal of Botany 64, 677-687. Ashford, A. E., Ryde, S. and Barrow, K. D. (1994). Demonstration of a short chain polyphosphate in Pisolithus tinctorius and the implications for phosphorus transport. New Phytologist 126, in press. Ayscough, K., Hajibagheri, N. M. A., Watson, R. and Warren, G. (1993). Stacking of Golgi cisternae in Schizo sac char omyces pombe requires intact microtubules. Journal of Cell Science 106, 1227-1237. Bae, K. -S. and Barton, L, L. (1989). Alkaline phosphatase and other hydrolases produced by Cenococcum graniforme, and ectomycorrhizal fungus. Applied and Environmental Microbiology 55, 2511-2516. Beams, H. W. and Kessel, R. G. (1976). Cytokinesis: a comparative study of cytoplasmic division in animal cells. American Scientist 64, 279-290. Beever, R. E. and Burns, D. J. W. (1980). Phosphorus uptake, storage and utilization by fungi. Advances in Botanical Research 8, 127-219. Berbee, M. L. and Wells, K. (1988). Ultrastructural studies of mitosis and the septal pore apparatus in Tremella globospora. Mycologia 80, 479-492. Bielesky, R. L. (1973). Phosphate pools, phosphate transport, and phosphate availability. Annual Review of Plant Physiology 24, 225-252.

93 References 94

Bolan, N. S. (1991). A critical review on the role of mycorrhizal fungi in the uptake of phosphorus by plants. Plant and Soil 134, 189-207. Bourett, T. M., Picollelli, M. A. and Howard, R. J. (1993). Postembedment labeling of intracellular concanavalin A binding sites in freeze-substituted fungal cells. Experimental Mycology 17, 223-235. Bourett, T. M. and McLaughlan, D. J. (1986). Mitosis and septum formation in the basidiomycete Helicobasidium mompa. Canadian Journal of Botany 64, 130- 145. Bowen, G. D. (1969). The uptake of orthophosphate and its incorporation into organic phosphates along roots of Pinus radiata. Australian Journal of Biological Sciences 22, 1125-1135. Bowen, G. D. and Theodorou, C. (1967). Studies on phosphate uptake by mycorrhizas. International Union of Forestry Research Organizations 14th Congress, Munich. Proceedings 5, 116-136. Brunk, U. T. and Ericsson, J. L. E. (1973). The demonstration of acid phosphatase activity in in vitro cultured tissue cells. Studies on the significance of fixation, tonicity and permeability. In Fixation in Histochemistry (ed. P. J. Stoward), pp. 121 - 135. Chapman and Hall, London. Buller, A. H. R. (1933). Translocation of protoplasm through the septate mycelium of certain Pyrenomycetes, Discomycetes and Hymenomycetes. In Researches in Fungi Vol. V, pp. 75-167. Longmans, London. Cairney, J. W. G. (1992). Translocation of solutes in ectomycorrhizal and saprotrophic rhizomorphs. Mycological Research 96, 135-141. Cairney, J. W. G., Ashford, A. E. and Allaway, W. G. (1989). Distribution of photosynthetically fixed carbon within root systems of Eucalyptus pilularis plants ectomycorrhizal with Pisolithus tinctorius. New Phytologist 112, 495-500. Cairney, J. W. G., Jennings, D. H. and Veltkamp, C. J. (1988). Structural differentiation in maturing rhizomorphs of Armillaria mellea (Tricholomatales). Nova Hedwigia 46, 1-25. Cairney, J. W. G. and Smith, S. E. (1993). Eflux of phosphate from the ectomycorrhizal basidiomycete Pisolithus tinctorius: general characteristics and the influence of intracellular phosphorus concentration. Mycological Research 97, 1261- 1266. Callow, J. A., Cappacio, L. C. M., Parish, G. and Tinker, P. B. (1978). Detection and estimation of polyphosphate in vesicular-arbuscular mycorrhizas. New Phytologist 80, 125-134. Canny, M. J. and McCully, M. E., (1986). Locating water-soluble vital stains in plant tissues by freeze-substitution and resin-embedding. Journal of Microscopy 142, 63-70. References 95

Chilvers, G. A. and Harley, J. L. (1980). Visualization of phosphate accumulation in beech mycorrhizas. New Phytologist 84, 319-326. Chilvers, G. A., Lapeyrie, F. F. and Douglass, P. A. (1985). A contrast between oomycetes and other taxa of mycelial fungi in regard to metachromatic granule formation. New Phytologist 99, 203-210. Cox, G., Moran, K. J., Sanders, F., Nockolds, C. and Tinker, P. B. (1980). Translocation and transfer of nutrients in vesicular-arbuscular mycorrhizas. III. Polyphosphate granules and phosphorus translocation. New Phytologist 84, 649-659. Cox, G., Sanders, F. E., Tinker, P. B. and Wild, J. A. (1975). Ultrastructural evidence relating to host-endophyte transfer in a vesicular-arbuscular mycorrhiza. In Endomycorrhizas (eds F. E. Sanders, B. Mosse and P. B. Tinker), pp. 297-312. Academic Press, London. Crang, R. E. (1980). Polyphosphate bodies in Aureobasidium pullulans. Micron 11, 3-4. Daddow, L. Y. M. (1983). A double lead stain method for enhancing contrast of ultrathin sections in electron microscopy: A modified multiple staining technique. Journal of Microscopy 129, 147-153. Davey, D. F. (1973). The effect of fixative tonicity on the myosin filament lattice volume of frog muscle fixed following exposure to normal or hypertonic Ringer. In Fixation in Histochemistry (ed. P. J. Stoward), pp. 103 - 120. Chapman and Hall, London. Dawes, E. A. and Senior, P. J. (1973). The role and regulation of energy-reserve polymers in microorganisms. In Advances in Microbial Physiology. Vol. 10 (eds A. J. Rose and D. W. Tempest), pp. 135-266. Academic Press, London, New York. Delaporte, B. (1939). Sur les acides nucleiques des levures et leur localisation. Revue General de Botanique 51, 449-482. Denny, H. J. and Wilkins, D. A. (1987). Zinc tolerance in Betula spp. IV. The mechanism of ectomycorrhizal amelioration of zinc toxicity. New Phytologist 106, 545-553. Dighton, J. (1983). Phosphatase production by mycorrhizal fungi. Plant and Soil 71, 455-462. Ding, B., Turgeon, R. and Parthasarathy, M. V. (1992). Effect of high- pressure freezing on plant microfilament bundles. Journal of Microscopy 165, 367- 376. Doonan, B. B., Crang, R. E., Jensen, T. E. and Baxter, M. (1979). In situ X-ray energy dispersive microanalysis of polyphosphate bodies in Aureobasidium pullulans. Journal of Ultrastructure Research 69, 232-238. References 96

Durr, M., Urech, K., Boiler, Th., Wiemken, A., Schwencke, J. and Nagy, M. (1979). Sequestration of arginine by polyphosphate in vacuoles of yeast (,Saccharomyces cerevisiae). Archives of Microbiology 121, 169-175. Echlin, P. (1992). Low-temperature Microscopy and Analysis. Plenum Press, New York, London. Elder, H. Y., Bovell, D. L., Pediani, J. D., Wilson, S. M., McWilliams, S. A. and McEwan Jenkinson, D. (1988). On the validity of block freeze­ drying and vacuum resin embedding of cryoquenched tissues for quantitative intracellular X-ray microanalysis. In Institute of Physics Conference Series (EUREM 88) No. 93, 3 (eds H. G. Dickinson and P. G. Goodhew), pp. 575-576. IOP Publishing Ltd, Bristol. Evert, R. F. and Deshpande, B. P. (1969). Electrron microscope investigation of sieve-element ontogeny and structure in Ulmus americana. Protoplasma 68, 403- 432. Fitzgerald, M. A. and Allaway, W. G. (1991). Apoplastic and symplastic pathways in the leaf of the grey mangrove Avicennia marina (Forsk.) Vierh. New Phytologist 119, 217-226. Fitzgerald, M. A., Orlovich, D. A. and Allaway, W. G. (1992). Evidence that abaxial leaf glands are the sites of salt secretion in leaves of the mangrove Avicennia marina (Forsk.) Vierh. New Phytologist 120, 1-7. Foster, R. C. and Marks, G. C. (1966). The fine structure of the mycorrhizas of Pinus radiata D. Don. Australian Journal of Biological Sciences 19, 1027-1038. Gianinazzi-Pearson, V. and Gianinazzi, S. (1989). Phosphorus metabolism by mycorrhizas. In Nitrogen, Phosphorus and Sulphur Utilization by Fungi (eds L. Boddy, R. Marchant and D. J. Read), pp. 226-241. Cambridge University Press, Cambridge. Girbardt, M. (1979). A microfilamentous septal belt (FSB) during induction of cytokinesis in Trametes versicolor (L. ex Fr.). Experimental Mycology 3, 215-228. Graham, T. R. and Emr, S. D. (1991). Compartmental organization of Golgi- specific protein modifications and vacuolar protein sorting events defined in a yeast sec 18 (NSF) mutant. Journal of 114, 207-218. Grellier, B., Strullu, D. G., Martin, F. and Renaudin, S. (1989). Synthesis in vitro, microanalysis and 31P-NMR study of metachromatic granules in birch mycorrhizas. New Phytologist 112, 49-54. Grenville, D. J., Peterson, R. L. and Ashford, A. E. (1986).Synthesis in growth pouches of mycorrhizae between Eucalyptus pilularis and several strains of Pisolithus tinctorius. Australian Journal of Botany 34, 95-102. References 97

Groves, R. H. and Keraitis, K. (1976). Survival and growth of seedlings of three sclerophyll species at high levels of phosphorus and nitrogen. Australian Journal of Botany 24, 681-690. Gunning, B. E. S. (1982). The cytokinetic apparatus: its development and spatial regulation. In The Cytoskeleton and Plant Growth and Development, (ed. C. W. Lloyd), pp 229-292. Academic Press, London. Gunning, B. E. S. and Overall, R. (1983). Plasmodesmata and cell-to-cell transport in plants. Bioscience 33, 260-265. Gunning, B. E. S. and Robards, A. W. (1976). Plasmodesmata and symplastic transport. In Transport and transfer processes in plants (eds I. F. Wardlaw and J. B. Passioura), pp 15-41. Academic Press, New York. Gunning, B. E. S. and Steer, M. W. (1975). Ultrastructure and the Biology of Plant Cells. Edward Arnold, London. Hagler, H. K. and Buja, L. M. (1986). Effect of specimen preparation and section transfer techniques on the preservation of ultrastructure, lipids and elements in croysections. Journal of Microscopy 141, 311-317. Harley, J. L. and Harley, E. L. (1987). A check-list of mycorrhiza in the British flora. New Phytologist (Supplement) 105, 1-102. Harley, J. L. and Loughman, B. C. (1963). The uptake of phosphate by excised mycorrhizal roots of the beech. IX. The nature of the phosphate compounds passing into the host. New Phytologist 62, 350-359. Harley, J. L. and McCready, C. C. (1950). The uptake of phosphate by excised mycorrhizal roots of the beech. New Phytologist 49, 388-397. Harley, J. L. and McCready, C. C. (1952a). The uptake of phosphate by excised mycorrhizal roots of the beech. II. Distribution of phosphorus between host and fungus. New Phytologist 51, 56-64. Harley, J. L. and McCready, C. C. (1952b). The uptake of phosphate by excised mycorrhizal roots of the beech. III. The effect of the fungal sheath on the availability of phosphate to the core. New Phytologist 51, 342-348. Harley, J. L. and McCready, C. C. (1981). The relationship between phosphate absorption and oxygen uptake in excised beech mycorrhizas. New Phytologist 88, 675-681. Harley, J. L. and Smith, S. E. (1983). Mychorrhizal Symbiosis. Academic Press, London. Harold, F. M. (1966). Inorganic polyphosphates in biology: structure, metabolism, and function. Bacteriological Reviews 30, 772-794. Harvey, D. M. R. (1980). The preparation of botanical samples for ion localisation studies at the subcellular level. Scanning Electron Microscopy, 1980/II, 409-419. References 98

Harvey, D. M. R. (1982). Freeze-substitution. Journal of Microscopy 127, 209-

221. Harvey, D. M. R., Hall, J. L. and Flowers, T. J. (1976). The use of freeze- substitution in the preparation of plant tissue for ion localization studies. Journal of Microscopy 107, 189-198. Hippe-Sanwald, S., Hermanns, M. and Somerville, S. C. (1992). Ultrastructural comparison of incompatible and compatible interactions in the barley powdery mildew. Protoplasma 168, 27-40. Ho, I. (1989). Acid phosphatase, alkaline phosphatase, and nitrate reductase activity of selected ectomycorrhizal fungi. Canadian Journal of Botany 67, 750-753. Hoch, H. C. (1991). Preservation of cell ultrastructure by freeze-substitution. In Electron Microscopy of Plant Pathogens (eds K. Mendgen and D. -E. Lesemann), pp. 1-16. Springer-Verlag, Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong, Barcelona. Hoch, H. C. and Howard, R. J. (1980). Ultrastructure of freeze-substituted hyphae of the basidiomycete Laetisaria arvalis. Protoplasma 103, 281-297. Hoch, H.C. and Howard, R. J. (1981). Conventional chemical fixations induce artifactual swelling of dolipore septa. Experimental Mycology 5, 167-172. Hoch, H. C. and Staples, R. C. (1983). Visualization of actin in situ by rhodamine-conjugated phalloin in the fungus Uromyces phaseoli. European Journal of Cell Biology 32, 52-58. Howard, R. J. and Aist, J. R. (1979). Hyphal tip cell ultrastructure of the fungus Fusarium: improved preservation by freeze-substitution. Journal of Ultrastructure Research 66, 224-234. Howard, R. J. and O’Donnell, K. L. (1987). Freeze substitution of fungi for cytological analysis. Experimental Mycology 11, 250-269. Hyde, G. J., Lancelle, S., Hepler, P. K. and Hardham, A. R. (1991). Freeze substitution reveals a new model for sporangial cleavage in Phytophthora, a result with implications for cytokinesis in other eukaryotes. Journal of Cell Science 100, 735-746. Ingram, P., Nassar, R., Le Furgey, A., Davilla, S. and Sommer, J. (1989). Quantitative X-ray elemental mapping of dynamic physiologic events in skeletal muscle. In Electron Probe Microanalysis. Applications in Biology and Medicine (eds K. Zierold and H. K. Hagler), pp. 251-264. Springer Verlag, Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong. Jacobson, L., Halmann, M. and Yariv, J. (1982). The molecular composition of the volutin granule of yeast. Biochemistry Journal 201, 473-479. Jennings, D. H. (1987). Translocation of solutes in fungi. Biological Reviews 62, 215-243. References 99

Jennings, D. H. (1989). Some perspectives on nitrogen and phosphorus metabolism in fungi. In Nitrogen, Phosphorus and Sulphur Utilization by Fungi (eds L. Boddy, R. Marchant and D. J. Read), pp. 1-31. Cambridge University Press, Cambridge. Jones, H. D., Schliwa, M. and Drubin, D. G. (1993). Video microscopy of organelle inheritance and motility in budding yeast. Cell motility and the Cytoskeleton 25, 129-142. Keck, K. and Stich, H. (1957). The widespread occurrence of polyphosphate in lower plants. Annals of Botany 21, 611-619. Kellenberger, E. (1987). The response of biological macromolecules and supramolecular structures to the physics of specimen cryopreparation. In Cryotechniques in Biological Electron Microscopy, (eds R. A. Steinbrecht and K. Zeirold), pp. 35-63. Springer-Verlag, Berlin, Heidelberg. Klionsky, D. J., Herman, P. K. and Emr, S. D. (1990). The fungal vacuole: Composition, function, and biogenesis. Microbiology Reviews 54, 266-292. Kroehler, C. J., Antibus, R. K. and Linkins, A. E. (1988). The effects of organic and inorganic phosphorus concentration on the acid phosphatase activity of ectomycorrhizal fungi. Canadian Journal of Botany 66, 750-756. Kulaev, I. S. and Vagabov, M. (1983). Polyphosphate metabolism in microorganisms. Advances in Microbial Physiology 24, 83-158. Lapeyrie, F. F., Chilvers, G. A. and Douglass, P. A. (1984). Formation of metachromatic granules following phosphate uptake by mycelial hyphae of an ectomycorrhizal fungus. New Phytologist 98, 345-630. Ling-Lee, M., Chilvers, G. A. and Ashford, A. E. (1975). Polyphosphate granules in three different kinds of tree mycorrhiza. New Phytologist 75, 551-554. Lingle, W. L. (1989). Enhanced staining of the basidiomycete Panellus stypticus prepared for transmission microscopy by freeze-substitution. Cryptogamic Botany 1, 236-242. Loughman, B. C. and Ratcliffe, R. G. (1984). Nuclear magnetic resonance and the study of plants. In Advances in Plant Nutrition Vol. 1. (eds P. B. Tinker and A. Lauchli), pp 241-283. Praeger Publishers, New York. Lii, H. and McLaughlin D. J. (1991). Ultrastructure of the septal pore apparatus and early septum initiation in Auricularia auricula-judae. Mycologia 83, 322-334. Lucocq, J. M., Berger, E. and Warren, G. (1989). Mitotic Golgi fragments in HeLa cells and their role in the reassembly pathway. Journal of Cell Biology 109, 463-474. Luo, Z. and Robinson, J. M. (1992). Co-localozation of an endocytic marker and acid phosphatase in a tubular/reticular compartment in macrophages. Journal of Histochemistry and Cytochemistry 40, 93-103. References 100

Malajczuk, N., McComb, A. J. and Loneragan, J. F. (1975). Phosphorus uptake and growth of mycorrhizal and uninfected seedlings of Eucalyptus calophylla R. Br. Australian Journal of Botany 23, 231-238. Marshall, A. T. (1980). Freeze-substitution as a prepaation technique for biological X-ray microanalysis. Scanning Electron Microscopy, 1980/11, 395-408. Martin, F., Canet, D., Rolin, D., Marchal, J. -P. and Larher, F. (1983). Phosphorus-31 nuclear magnetic resonance study of polyphosphate metabolism in intact ectomycorrhizal fungi. Plant and Soil 71, 469-476. Martin, F., Marchal, J. -P., Timinski, A. and Canet, D. (1985). The metabolism and physical state of polyphosphates in ectomycorrhizal fungi. A 31P nuclear magnetic resonance study. New Phytologist 101, 275-290. Marx, D. H. (1969). The influence of ectotrophic mycorrhizal fungi on the resistance of pine roots to pathogenic infections. 1. Antagonism of mycorrhizal fungi to root pathogenic fungi and soil bacteria. Phytopathology 59, 153-163. Matsuo, E. S. and Tanaka, T. (1992). Patterns in shrinking gels. Nature 358, 482-485. Mejstrik, V. and Benecke, U. (1969). The ectotrophic mycorrhizas of Alnus viridis (Chaix) D. C. and their significance in respect to phosphorus uptake. New Phytologist 68, 141-149. Miller, J. J. (1984). In vitro experiments concerning the state of polyphosphate in the yeast vacuole. Canadian Journal of Microbiology 30, 236-246. Mims, C. W., Richardson, E. A. and Timberlake, W. E. (1988a). Ultrastructural analysis of conidiophore development in the fungus Aspergillus nidulans using freeze-substitution. Protoplasma 144, 132-141. Mims, C. W., Roberson, R. W. and Richardson, E. A. (1988b). Ultrastructure of freeze-substituted and chemically-fixed basidiospores of Gymnosporangium juniperi-virginianae. Mycologia 80, 356-364. Mitchell, D. T. and Read, D. J. (1981). Utilization of inorganic and organic phosphates by the mycorrhizal endophytes of Vaccinium macrocarpon and Rhododendron ponticum. Transactions of the British Mycological Society 76, 255- 260. Moore, R. T. (1984). The challenge of the dolipore/parenthesome. In Developmental biology of higher fungi, (eds D. Moore, L. A. Casselton, D. A. Wood and J. C. Frankland), pp 175-212. Cambridge University Press, Cambridge. Morgan, A. J., Davies, T. W. and Erasmus, D. A. (1978). Specimen preparation. In Electron Probe Microanalysis in Biology (ed. D. A. Erasmus), pp. 94-147. Chapman and Hall, London. References 101

Nakai, Y. (1976). Fine structure of shiitake, Lentinus edodes (Berk.) Sing. V. Intercellular inclusions in germinating basidiospores induced by glutaraldehyde fixation. Report of the Tottori Mycological Institute (Japan) 14, 91-94. Neiss, W. F. (1984). A coat of glycoconjugates on the inner surface of the lysosomal membrane in the rat kindey. Histochemistry 80, 603-608. O’Brien, T. P., Kuo, J., McCully, M. E. and Zee, S. -Y. (1973). Coagulant and non-coagulant fixation of plant cells. Australian Journal of Biological Sciences 26, 1231-1250. Orlovich, D. A. (1987). An histochemical and X-ray microanalytical study of polyphosphate granules in the ectomycorrhizal fungus Pisolithus tinctorius. B.Sc. (Hons) thesis. School of Botany, The University of New South Wales. Orlovich, D. A. and Ashford, A. E. (1993). Polyphosphate granules are an artefact of specimen preparation in the ectomycorrhizal fungus Pisolithus tinctorius. Protoplasma 173, 91-102. Orlovich, D. A. and Ashford, A. E. (1994). Structure and development of the dolipore septum in Pisolithus tinctorius. Protoplasma in press. Orlovich, D. A., Ashford, A. E. and Cox, G. C. (1989). A reassessment of polyphosphate granule composition in the ectomycorrhizal fungus Pisolithus tinctorius. Australian Journal of Plant Physiology 16, 107-115. Orlovich, D. A., Ashford, A. E., Cox, G. C. and Moore, A. E. P. (1990). Freeze-substitution and X-ray microanalysis of polyphosphate granules in the mycorrhizal fungus Pisolithus tinctorius (Pers.) Coker & Couch. Endocytobiology IV, 139-143. Pallaghy, C. K. (1973). Electron probe microanalysis of potassium and chloride in freeze-substituted leaf sections of Zea mays. Australian Journal of Biological Sciences 26, 1015-1034. Patton, A. M. and Marchant, R. (1978). A mathematical analysis of dolipore/parenthesome structure in basidiomycetes. Journal of General Microbiology 109, 335-349. Preuss, D., Mulholland, J., Franzusoff, A., Segev, N. and Botstein, D. (1992). Characterization of the Saccharomyces Golgi complex through the cell cycle by immunoelectron microscopy. Molecular Biology of the Cell 3, 789-803. Pypaert, M., Nilsson, T., Berger, E. G. and Warren, G. (1993). Mitotic Golgi clusters are not tubular endosomes. Journal of Cell Biology 104, 811-818. Rappaport, R. (1986). Establishment of a mechanism of cytokinesis in animal cells. International Review of Cytology 105, 245-281. Rees, B., Shepherd, V. A. and Ashford, A. E. (1994). Presence of a motile tubular vacuole system in different phyla of fungi. Mycological Research submitted. References 102

Reynolds, E. S. (1963). The use of lead citrate at high pH as an electron-opaque stain in electron microscopy. Journal of Cell Biology 17, 208-210. Robards, A. W. and Lucas, W. J. (1990). Plasmodesmata. Annual Review of Plant Physiology and Plant Molecular Biology 41, 369-419. Roberson, R. W. (1992). The actin cytoskeleton in hyphal cells of Sclerotium rolfsii. Mycologia 84, 41-51. Robinson, J. M. and Karnovsky, M. J. (1991). Rapid-freezing cytochemistry: preservation of tubular lysosomes and enzyme activity. Journal of Histochemistry and Cytochemistry 39, 787-792. Rolin, D., Le Tacon, F. and Larher, F. (1984). Characterization of the different forms of phosphorus in the mycelium of the ectomycorrhizal fungus, Hebeloma cylindrosporum, cultivated in pure culture. New Phytologist 98, 335-343. Roth, J. (1983). Application of lectin-gold complexes for electron microscopic localization of glycoconjugates on thin sections. Journal of Histochemistry and Cytochemistry 31, 987-999. Rothman, J. E. and Orci, L. (1992). Molecular dissection of the secretory pathway. Nature 355, 409-415. Rothman, J. E., Urbani, L. J. and Brands, R. (1984). Transport of protein between cytoplasmic membranes of fused cells: correspondence to processes reconstituted in a cell-free system. Journal of Cell Biology 99, 248-259. Russ, J. C. (1978). Electron probe X-ray microanalysis - principles. In Electron Probe Microanalysis in Biology, (ed. D. A. Erasmus), pp. 5-36. Chapman and Hall Ltd, London. Salema, R. and Brandao, I. (1973). The use of PIPES buffer in the fixation of plant cells for electron microscopy. Journal of Submicroscopic Cytology 5, 79-96. Schroer, T. A. and Kelly, R. B. (1985). In vitro translocation of organelles along microtubules. Cell 40, 729-730. Shepherd, V. A., Orlovich, D. A. and Ashford, A. E. (1993a). A dynamic continuum of pleiomorphic tubules and vacuoles in growing hyphae of a fungus. Journal of Cell Science 104, 495-507. Shepherd, V. A., Orlovich, D. A. and Ashford, A. E. (1993b). Cell to cell transport via motile tubules in growing hyphae of a fungus. Journal of Cell Science 105, 1173-1178. Spurr, A. R. (1969). A low-viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructure Research 26, 31-43. Steinbrecht, R.A., & Muller, M. (1987). Freeze-substitution and freeze-drying. In Cryotechniques in Biological Electron Microscopy, (eds R. A. Steinbrecht and K. Zeirold), pp. 149-172. Springer-Verlag, Berlin, Heidelberg. References 103

Straker, C. J. and Mitchell, D. T. (1985). The characterisation and estimation of polyphosphates in endomycorrhizas of the Ericaceae. New Phytologist 99, 431-440. Strullu, D. G., Grellier, B., Garrec, J. P., McCready, C. C. and Harley, J. L. (1986). Effects of monovalent and divalent cations on phosphate absorption by beech mycorrhizas. New Phytologist 103, 403-416. Swanson, J., Bushnell, A. and Silverstein, S. C. (1987). Tubular morphology and distribution within macrophages depend on the integrity of cytoplasmic microtubules. Proceedings of the National Academy of Science USA 84, 1921-1925. Taylor, G. H. (1983). Preface. In Phosphorus in Australia, (eds A. B. Costin and C. H. Williams), pp. vii-x. Australian National University, Canberra. Thilo, E. (1962). Condensed phosphates and arsenates. Advances in Inorganic Chemistry and Radiochemistry 4, 1-77. Thompson, W., Brownlee, C., Jennings, D. H. and Mortimer, A. M. (1987). Localized, cold-induced inhibition of translocation in mycelia and strands of Serpula lacrimans. Journal of Experimental Botany 38, 889-899. Thompson, W., Eamus, D. and Jennings, D. H. (1985). Water flux through mycelium of Serpula lacrimans. Transactions of the British Mycological Society 84, 601-608. Tijssen, J. P. F., van Steveninck, J. and de Bruijn, W. C. (1985). Cytochemical staining of a yeast polyphosphate fraction, localized outside the plasma membrane. Protoplasma 125, 124-128. Timonen, S., Finlay, R. D., Soderstrom, B. and Raudaskoski, M. (1993). Identification of cytoskeletal components in pine ectomycorrhizas. New Phytologist 124, 83-92. Tooze, J. and Hollinshead, M. (1991). Tubular early endosomal networks in AtT20 and other cells. Journal of Cell Biology 115, 635-653. Tooze, J. and Hollinshead, M. (1992). Evidence that globular Golgi clusters in mitotic HeLa cells are clustered tubular endosomes. European Journal of Cell Biology 58, 228-242. Tucker, B. E., Hoch, H. C. and Staples, R. C. (1986). The involvement of F- actin in Uromyces cell differentiation: the effects of cytochalasin E and phalloidin. Protoplasma 135, 88-101. van Noorden, S. and Polak, J. M. (1983). Immunocytochemistry today. Techniques and practice. In Immunocytochemistry: Practical Applications in Pathology and Biology (eds J. M. Polak and S. van Noorden), pp. 11-42. John Wright and Sons Ltd, Bristol, London, Boston. References 104

van Zyl, J., Forrest, Q.G., Hocking, C. and Pallaghy, C.K. (1976) Freeze-substitution of plant and animal tissue for the localization of water-soluble compounds by electron probe microanalysis. Micron 7, 213-224. Vare, H. (1990). Aluminium polyphosphate in the ectomycorrhizal fungus Suillus variegatus (Fr.) O. Kunze as revealed by energy-dispersive spectroscopy. New Phytologist 116, 663-668. von Zglinicki, T. (1989). Ensuring the validity of results in biological X-ray microanalysis. In Electron Probe Microanalysis. Applications in Biology and Medicine (eds K. Zierold and H. K. Hagler), pp. 47-58. Springer Verlag, Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong. von Zglinicki, T. and Uhrfk, V. (1988). X-ray microanalysis with continuous specimen cooling: is it necessary? Journal of Microscopy 151, 43-47. Walker, G. D. and Powell, C. LI. (1979). Vesicular-arbuscular mycorrhizas in white clover: a scanning electron microscope and X-ray microanalytical study. New Zealand Journal of Botany 17, 55-59. Wessels, J. G. H. (1986). Cell wall synthesis in apical hyphal growth. International Review of Cytology 104, 37-79. Wiemken, A., Schellenberg, M. and Urech, K. (1979). Vacuoles: The sole compartments of digestive enzymes in yeast (Saccharomyces cerevisiae)? Archives of Microbiology 123, 23-35. Wild, A. (1958). The phosphate content of Australian soils. Australian Journal of Agricultural Research 9, 193-204. Yanagita, T. (1964). Germinating conidiospores of Aspergillus niger. In Synchrony in Cell Division and Growth (ed. E. Zeuthen), pp. 391-420. Interscience Publishers, John Wiley and Sons Inc., New York, London, Sydney. Young, N., Bullock, S., Orlovich, D. A. and Ashford, A. E. (1993). Association of polyphosphate with protein in freeze-substituted sclerotia of Sclerotinia minor. Protoplasma 174, 134-141. Zierold, K. and Steinbrecht, R. A. (1987). Cryofixation of diffusible elements in cells and tissues for electron probe microanalysis. In Cryotechniques in Biological Electron Microscopy, (eds R. A. Steinbrecht and K. Zeirold), pp. 272-282. Springer-Verlag, Berlin, Heidelberg. Appendix

105 Aust. J. Plant Physiol., 1989, 16, 107-15

A Reassessment of Polyphosphate Granule Composition in the Ectomycorrhizal Fungus Pisolithus tinctorius

D. A. Orlovich A, A. E. AshfordA and G. C. CoxB 'School of Biological Science, University of New South Wales, P.O. Box 1, Kensington, N.S.W. 2033, Australia. “Electron Microscope Unit, University of Sydney, N.S.W. 2006, Australia.

A bstract Comparison of the elemental composition of freeze-substituted and conventionally fixed phosphorus- containing granules of the ectomycorrhizal fungus Pisolithus tinctorius (Pers.) Coker & Couch, using energy dispersive X-ray microanalysis, shows that the cation composition is altered appreciably by the method of specimen preparation. Following conventional chemical fixation, the major cation detected in the granules together with phosphorus was calcium, while freeze-substituted granules most frequently contained phosphorus, potassium and sodium. We conclude that chemical fixation causes serious loss or redistribution of ions. The occurrence of monovalent cations in polyphosphate granules has not previously been demonstrated for mvcorrhizal fungi and their presence should be considered in determining the role of the granules in mycorrhizas.

Introduction Polyphosphate (polvP) granules are known to be present in the fungal component of a wide range of mycorrhizal associations (Ashford el al. 1975, 1986; Cox el al. 1975, 1980; Ling-Lee el al. 1975; White and Brown 1979; Chilvers and Harley 1980; Strullu el al. 1981, 1982). PolyP is a strongly negative polyanion and there is a requirement for cations to balance the net negative charge (Harold 1966). There has been consider­ able interest in the nature of the balancing cation. There is no evidence from any histochemical reaction that a macromolecule is involved (Ashford et al. 1975). In all X-ray microanalytical studies of mycorrhizas conventionally prepared for electron microscopy, the most abundant*cation found in polyP granules is calcium and various hypotheses have been presented regarding possible links between phosphorus and cal­ cium metabolism in fungi (see Ashford et al. 1986). It is, however, known that polyP is a very effective cation exchanger (Thilo 1962) and during conventional electron microscopy preparation procedures, as membranes are rendered more permeable and aqueous solutions are used, there is a strong possibility that any monovalent cations originally present in vivo may be displaced by more highly charged divalent ions from other areas of the specimen or surrounding medium. Baxter and Jensen (1980) found that the method of specimen preparation altered the ion composition of polyP bodies in the cyanobacterium Plecionema boryanum Gomont. To evaluate the effect of speci­ men preparation on granule composition in mycorrhizal fungi, hyphae of a strain of the fungus Pisolithus tinctorius (Pers.) Coker & Couch (previously shown by Ashford et al. (1986) to accumulate polyP granules), were grown in a medium where they produced metachromatic granules, and then freeze-substituted and analysed by energy dispersive 0310-7841 89 010107503.00 108 I). A. Orlov ich el at.

X-ray (EDX) analysis. Procedures likely to minimise redistribution or loss of ions from the specimen were employed and granule ion composition was compared with that obtained following conventional preparation for electron microscopy.

Materials and Methods Fungal Material The strain of P. linciorius, DI-15, used here was isolated by D. J. Grenville in Australia (Grenville el at. 1986), and was used by Ashford el at. (1986) in an evaluation of polyP granule composition in Eucalyptus pilularis Sm. mycorrhizas following conventional fixation. The fungus was grown in axenic agar culture on modified Melin Norkrans medium (Marx and Bryan 1975) at 25°C in the dark. Using a sharp razor blade, pieces of mycelium 2 mm by 2 mm were cut from the growing edge of cultures 3-4 weeks old, removing excess agar from the sample. The material was prepared for electron microscopy and EDX analysis by one of the following methods.

Conventional Specimen Preparation Pieces of mycelium were fixed in 3°'o glutaraldehvde in 0-3M NaOH-PIPES (piperazine-.Y-.X'-bis- (2-ethanolsulfonic acid)) butler (Salema and Brandao 1973), pH 7-5 at 0°C for 4 h. Those to be used for EDX analysis were not osmicated since there is a considerable overlap between osmium (M) and phosphorus (K) peaks in EDX spectra. After a rinse in fresh buffer, the material was dehydrated through a graded ethanol series to 1000;o at 0 C and infiltrated with glycol methacrylate (GMA) as described in Feder and O’Brien (1968). The monomer mix was polymerised at 25 C by long-wave ultraviolet light in a nitrogen atmosphere (O’Brien and McCully 1981). Sections were cut on to distilled water, expanded with xylene, and collected on copper grids without a support film. The sections were not stained or coated.

Crvofixation and Freeze-substitution Material was plunge-frozen in Freon 22 which was cooled by liquid nitrogen (LN;), according to the method described by Fineran (19^8), and stored in LN: for about 5 min (until required). Two freeze- substitution schedules were used. All acetone used in both schedules was stored over activated molecular sieve for at least 12 h before use and care was taken to avoid contamination with water during the procedures.

Schedule one (CMA-embedding) Frozen mycelium was quickly transferred from LN? to a wire basket suspended in a polypropylene beaker containing dry ice. molecular sieve (Union Carbide-Linde, Type 4A) and anhydrous acetone. The polypropylene beaker was placed in a wide-mouth vacuum flask, containing additional dry ice. The flask was sealed and it was stored at 2°C for 48 h. At the end of this period, there was still some dry ice remaining in the flask. The material was transferred to vials containing fresh anhydrous acetone and molecular sieve at - 20 C for 48 h, then warmed to 4°C over 24 h and finally brought to room tempera­ ture. The mycelium was infiltrated with GMA and polymerised under UV light as described for con­ ventional specimen preparation.

Schedule two (Spurr's resin-embedding) The frozen mycelium was transferred rapidly from the LN: to small glass bottles containing a 1 : 2 mixture of activated molecular sieve and either 100ro anhydrous acetone or 1 ro distilled glutaraldehyde in anhydrous acetone. The vials were sealed and the material left to substitute at - 80°C for 48 h, then warmed to - 20°C for 24 h, 4°C for 4 h and finally brought to room temperature, where the substitution mix was replaced by fresh anhydrous acetone. The material was infiltrated with resin (Spurr 1969) and polymerised at 60°C.

Sectioning Following Freeze-substitution Since the introduction of water at this stage might result in the redistribution of ions in the material, most sections (c. 0-35 /^m-l ^m) were cut dry, placed directly on 100 200 mesh copper oyster (foldover) grids and were not stained. Polyphosphate Granules in Pisolulws linctorius 109

To evaluate the effect of liquid at this stage on the elemental composition of granules, some sections were cut using a trough of distilled water, as described for conventional specimen preparation, and then analysed. To determine the elemental composition of the Spurr’s resin used in the analyses, resin from the same batch as that used in schedule two was polymerised at 60 C without specimens, sectioned dry at I ^m and placed on copper foldover grids for analysis.

Transmission Electron Microscopy and EDX Analysis Sections were examined and analysed with a Philips CM 12 electron microscope equipped with an EDAX PV9900 analyser. EDX analysis was carried out at an accelerating voltage of 120 kV and a probe diameter of c. 10 nm. The objective aperture was removed prior to each analysis. Spectra were collected for 300 s (live detector time).

Fig. 1. Transmission electron micrograph of a freeze-substituted (schedule one), GMA-embedded cell. The large electron-opaque granule (P) shows the characteristic mottled appearance of a polyP granule. EDX analysis of this granule gave the spectrum shown in Fig. 3tz. Bar = 1 /jm.

Results Appearance and Identification of PolyP Grannies Since the sections were not osmicated or stained in the material prepared for EDX analysis, the images were of very low contrast. Membranes could not be resolved and few subcellular organelles could be distinguished. However, polyP granules could be recognised (Fig. 1) because of their inherent electron-opacity and tendency to sublimate in the electron beam, resulting in a mottled appearance (see discussion in White and Brown 1979). Ice crystal damage was apparent in some hyphae of freeze-substituted material regardless of the schedule used, indicating that freezing rates were not ideal. Many granules appeared to have been torn out of the conventionally prepared sec­ tions, as in Ashford et at. (1986), leaving holes containing remnants of electron-opaque 110 D. A. Oricnich et at.

Fig. 2. EDX spectra from glutaraldehyde- fixed, ethanol-dehydrated and GMA- embedded material. The vertical full scale (VFS) is given at the top right-hand corner of each spectrum, (a) An electron-opaque granule showing large peaks for P and Ca, with very small Na and K peaks. (b) and (c) Electron-lucent areas in the protoplast of the same cell, both showing peaks for Ca and Na and a smaller peak for Cl. Small peaks for P and S are present in (b) but were not significantly above background in (c).

Fig. 3. Spectra from material freeze-substituted and GMA-embedded by schedule one. Note that the VFS is much higher in these spectra than in Fig. 2. (a) Granule in Fig. 1 showing a large peak for P and very small peaks for Ca and Na. (b) Another granule with a large P peak, small peaks for Na and Mg, but no Ca peak. material. This loss may have been caused in the wet sectioning step as it was not apparent in freeze-substituted hyphae, sectioned dry.

EDX Analysis Conventionally fixed GMA-embedded cells The most abundant elements detected in electron-opaque granules of glutaraldehyde- fixed, GMA-embedded cells were phosphorus and calcium (Fig. 2a). There were small peaks for other cations, notably potassium and sodium. Spectra of electron-lucent Polyphosphate Granules in Pisolilhus tinciorius 111

regions in the same sections showed either only a small peak for phosphorus (Fig. 2b), or no phosphorus detectable above background (Fig. 2c). These spectra (Figs 2b and 2c) also showed peaks of sodium, chlorine and calcium. A small peak for sulfur was also detected in Fig. 2b. No elements were detected above the background in areas of resin adjacent to the specimen.

Fig. 4. Analyses of freeze-substituted, Spurr’s resin embedded material (schedule two) sectioned dry. Note the large differences in VFS in these spectra. Counts in spectra from different cells can not be directly compared since differences in section thickness and beam conditions can alter the total number of counts detected. Spectra a. b and c are from material freeze-substituted in lff'o glutaraldehyde in anhydrous acetone (spectra b and c are from the same cell). Spectra d, e and /are from material freeze- substituted in anhydrous acetone only (spectra d and e are from the same cell). Analysis of electron- opaque granules (a. b and d) each showed a relatively large peak for P and peaks of different heights for Cl, K and Na. Note the large relative variation in Cl peak height. The large peak at the extreme left-hand-side of b is due to noise. Analysis of electron-lucent regions adjacent to granules (c and e) showed much less P (compare b with c and d with e). Peaks for Na and K were not obviously lower in c but were much lower in e. Analysis of extra-cellular material (/) showed a peak for Cl.

Freeze-substituted cells embedded in GMA (schedule one) Analysis of the electron-opaque granule shown in Fig. 1 indicated a large peak for phosphorus and small peaks for sodium and calcium (Fig. 3a). Other electron-opaque regions of cells prepared in the same way showed spectra with a large phosphorus peak and small cation peaks (e. g. Fig. 3b), but the amount and type of cations varied. All of the granules contained low levels of sodium. Magnesium and calcium peaks were 112 D. A. Orlov ich et at.

present in only a few granules. In all cases much lower amounts of calcium were present than in conventionally fixed granules. The very low relative content of metal ions in this freeze-substituted material prompted us to consider the possibility that this freeze- substitution schedule was in need of improvement.

Freeze-substituted cells embedded in Spurr's resin (schedule two) The most abundant cations detected together with phosphorus in the electron-opaque granules analysed after Spurr’s resin embedding by schedule two were sodium and pot­ assium. Representative spectra for freeze-substituted, Spurr’s resin-embedded granules are shown in Figs 4a, 4b and 4^, with spectra for electron-lucent areas for comparison (Figs 4c and 4c). Small peaks of phosphorus were detected in electron-lucent areas, which also showed large peaks for chlorine. Potassium was not always present in

Fig. 5. Spectrum from dry-cut Spurr’s resin not containing a biological specimen. VFS is given in top right-hand corner. Only a small Cl peak was detected above the background. [The peak on the extreme left is attributed to noise and the adjacent small peak is the Cu (La) peak.]

Fig. 6. Spectra from the analysis of wet-cut freeze-substituted (schedule 2, l°'o glutaraldehyde in acetone, Spurr’s resin-embedded) material. Note the difference in VFS between spectra a and b. A spectrum from an electron-opaque granule (a) showed peaks for P, Cl, Ca and Na with smaller K and S peaks. The Na peak was enhanced and broadened by the presence of Cu and may therefore be an overestimate. Analysis of an electron-lucent region (b) showed a peak for Cl. Si was detected just above background but was considered to be a contaminant since Si-rich particles were observed on the surface of these sections. electron-lucent areas (compare Figs 4c and 4c). It is important to remember that soluble ions are retained in the sections by freeze-substitution and that they will be detected in the electron-lucent areas analysed. The presence or absence of glutaraldehyde in the substituting fluid did not apparently influence the cation composition of the granules (e. g. compare Figs 4a and 4d). Chlorine peaks were present in all spectra of material prepared according to schedule two. There was variability in chlorine levels within hyphal cells (compare Figs 4a and Ad) and between cells and extracellular regions of the section. A small peak of chlorine was present in areas of the resin outside the specimen (Fig. 4f). This was expected since Spurr’s resin is known to contain chlorine, especially the flexibiliser D.E.R. 736 (Pallaghy 1973). Analysis of Spurr’s resin which did not contain biological tissue (Fig. 5) also showed very low levels of chlorine. Polyphosphate Granules in Pisutilhus tinctorius 113

In view of the differences in granule composition between conventionally prepared material which had been wet-sectioned and freeze-substituted material sectioned dry, it was thought worthwhile to test the effect of wet-sectioning on freeze-substituted material. In this case, electron-opaque areas with high phosphorus and likely to be granules contained chlorine, calcium, sodium and low potassium (Fig. 6a). Analysis of electron-lucent regions showed low levels of chlorine (Fig. 6b).

Discussion The results have shown that the composition of polyP granules in P. tinctorius varies depending on the method of specimen preparation. Calcium was the major cation detected in granules from chemically fixed, conventionally prepared material, while pot­ assium and sodium were the most abundant cations in areas identified as granules of freeze-substituted material. Freeze-substitution has been shown to reduce ion loss and redistribution during specimen preparation for X-ray microanalysis (see Marshall 1980; Humbel and Muller 1986) and it is reasonable to assume that the cation composition of freeze-substituted granules is much more likely to reflect that found in vivo. The change in ion composition when freeze-substituted material is wet-sectioned, compared with dry sectioning, demonstrates firstly, that ions are not strongly held to polyP in the granules and are potentially mobile during tissue preparation and, secondly, that cal­ cium can replace monovalent cations at this stage. Similar mobility of ions during the wet sectioning step was shown by Roomans (1980), in a study of phosphorus-rich cyto­ plasmic granules in yeast. In this case, sectioning onto a trough of liquid resulted in a large loss of radioactive strontium from the sections. Since all conventional pro­ cedures have a wet-sectioning step, if ion exchange had not occurred earlier, it will certainly occur at this stage. It is therefore very likely that the universal occurrence of calcium in granules is an artefact which reflects specimen preparation. The freeze- substitution data indicate that the accompanying cations are much more likely to be monovalent cations and that cation composition of granules may vary more in kind and amount than the previous observations would have indicated. The finding that conventional preparation produces artefacts does not necessarily imply that the freeze-substitution techniques used here are perfect and have given an accurate indication of the precise cation composition of the granules. The sections were relatively thick, granules were hard to locale and many sections showed freezing dam­ age. Thus, it is very likely that the electron-opaque areas, determined to be granules by their high P content, will also have included some vacuole or cytoplasm. This may well account for differences in level and composition of cations as indicated by relative peak heights in the various spectra. In view of the composition of Spucr’s resin, it was not surprising to find chlorine in all of the spectra of Spurr's resin-embedded material. The apparently higher chlorine levels associated with embedded specimens (rather than resin alone) presumably result from chlorine in the tissue and associated agar medium and from increased electron scattering due to the higher density of the specimen. All previous studies of polyP granule composition in mycorrhizal fungi have detected calcium as the major cation and all have used conventional specimen preparation methods (White and Brown 1979; Strullu et al. 1981, 1982; Ashford et al. 1986). Our findings are much more in accord with the findings of Lapevrie et al. (1984) that presence or absence of calcium or magnesium in the medium has no effect on metachromatic granule formation in either calcicole or calcifuge isolates of Paxillus involutus (Batsch. ex Fr.) Fr. Lapeyrie et al. (1984) also found that mycelium starved of potassium or sodium formed granules when salts of any one of a number of cations were added to the medium, including ammonium. A result of this kind is most consistent with the hypothesis that cation composition of polyP granules varies according to the availability of ions in the medium. 1 14 D. A. Orlov ich et al.

There are in fact three alternative ways the cell may balance the polvP charge. The cell may accumulate a cation from the medium, it may make up the deficit with a metabolically produced organic cation, or it may mobilise particular cations from other cellular sites and transfer them into the granule-containing vacuoles. Presumably once in the vacuole, the cation with the greatest exchanging capacity will be the one to associate with the granule. Accumulation of organic cations in association with polvP would not be detected either by histochemical staining or EDX analysis. It has been shown by Durr et al. (1979) that the amino acid arginine, which has a net positive charge, is taken up with polvP into vacuoles of Saecharomyces cerevisiae L. Further perfection of freeze-substitution techniques and their application to polvP granule analysis will allow the investigation of the relationship between cation compo­ sition in the granules and in the medium. This applies not only to metabolically import­ ant ions, but also to toxic ions. Since granules are ion exchangers, there has been some speculation that they may play a role in sequestering any toxic ions that may have entered the cells. This could in part explain why mvcorrhizal infection can reduce heavy metal toxicity in plants (Bradley et al. 1981; Brown and Wilkins 1985), although there are of course several other explanations (e. g. Denny and Wilkins 1987). Jones and Hutchinson (1988) reported that in Betula papyrifera Marshall seedlings mvcorrhizal with Scleroderma flavidum Ell. & Ev., the calcium concentration in the roots decreased as they accumulated nickel and suggested that this may result from replacement of calcium by nickel in polyP granules. A rigorous study of polyP composition involving analysis of granules unmodified from the natural state is required to substantiate such claims. Freeze-substitution methods currently available provide the opportunity to do this.

Acknowledgments The authors thank Dr Maret Vesk for advice on freeze-substitution and Dr Thor Bostrom for assistance in processing the EDX spectra. Financial assistance from the Australian Research Grants Scheme to A.E.A. is gratefully acknowledged. Part of this work was completed while D.A.O. was in receipt of a Commonwealth Post-graduate Research Award.

References Ashford, A. E., Ling-Lee, \L, and Chilvers, G. A. (1975). Polyphosphate in eucalypt mycorrhizas: a cytochemical demonstration. New Phytologist 74, 447-53. Ashford, A. E., Peterson, R. L., Dwarte, D., and Chilvers, G. A. (1986). Polyphosphate granules in eucalypt mycorrhizas: determination by energy dispersive X-ray microanalysis. Canadian Journal of Botany 64, 677-87. Baxter, M., and Jensen, T. (1980). A study of methods for in situ X-ray energy dispersive analysis of polyphosphate bodies in Plectonema boryanum. Archives of Microbiology 126, 213-15. Bradley, R., Burt, A. J., and Read, D. J. (1981). Mycorrhizal infection and resistance to heavy metal toxicity in Calluna vulgaris. Nature (London) 292, 335-7. Brown, M. T., and Wilkins, D. A. (1985). Zinc tolerance of Amanita and Paxillus. Transactions of the British Mvcological Society 84, 367-9. Chilvers, G. A., and Harley, J. L. (1980). Visualization of phosphate accumulation in beech mycorrhizas. New Phytologist 84, 319-26. Cox, G., Moran, K. J., Sanders, F., Nockolds, C., and Tinker, P. B. (1980). Translocation and transfer of nutrients in vesicular-arbuscular mycorrhizas. 111. Polyphosphate granules and phosphorus trans­ location. New Phytologist 84, 649-59. Cox, G., Sanders, F. E., Tinker, P. B., and Wild, J. A. (1975). Ultrastructural evidence relating to host- endophyte transfer in a vesicular-arbuscular mycorrhiza. In ‘Endomycorrhizas’. (Eds F. E. Sanders, B. Mosse and P. B. Tinker.) pp. 297-312. (Academic Press: London.) Polyphosphate Granules in Pisolilhus tinctorius 115

Denny, H. J., and Wilkins, D. A. (1987). Zinc tolerance in Belula spp. IV. The mechanism of ectomycorrhizal amelioration of zinc toxicity. New Phytologist 106, 545-53. Durr, M., Urech, K., Boiler, Th., Wiemken, A., Schwencke, J., and Nagy, M. (1979). Sequestration of arginine by polyphosphate in vacuoles of yeast (Saccharomyces cerevisiae). Archives of Microbiology 121, 169-75. Feder, N., and O’Brien, T. P. (1968). Plant microtechnique: some principles and new methods. American Journal of Botany 55, 123-42. Fineran, B. A. (1978). Freeze-etching. In ‘Electron Microscopy and Cytochemistry of Plant Cells'. (Ed. J. L. Hall.) pp. 279-341. (Elsevier North Holland Biomedical Press: Amsterdam, Oxford and New York.) Grenville, D. J., Peterson, R. L., and Ashford, A. E. (1986). Synthesis in growth pouches of mycorrhizae between Eucalyptus pilularis and several strains of Pisolilhus tinctorius. Australian Journal of Botany 34, 95-102. Harold, F. M. (1966). Inorganic polyphosphates in biology: structure, metabolism and function. Bac­ teriological Reviews 30, 772-94. Humbel, B., and Muller, M. (1986). Freeze substitution and low temperature embedding. In ‘The Science of Biological Specimen Preparation 1985’. (Eds M. Muller, R. P. Becker, A. Bovde and J. J. Wolosewick.) pp. 175-83. (SEM Inc., A.V1F O'Hare: Chicago.) Jones, M. D., and Hutchinson, T. C. (1988). Nickel toxicity in mycorrhizal birch seedlings infected with Lactarius rufus or Scleroderma flavidum. II. Uptake of nickel, calcium, magnesium, phosphorus and iron. New Phytologist 108, 461-70. Lapeyrie, F. F., Chilvers, G. A., and Douglass, P. A. (1984). Formation of metachromatic granules following phosphate uptake by mycelial hyphae of an ectomycorrhizal fungus. New Phytologist 98, 345-60. Ling-Lee, M., Chilvers, G. A., and Ashford, A. E. (1975). Polyphosphate granules in three different kinds of tree mycorrhiza. New Phytologist 75, 551-4. Marshall, A. T. (1980). Freeze-substitution as a preparation technique for biological X-ray micro- analysis. Scanning Electron Microscopy/1980/11, 395-408. Marx, D. H., and Bryan, W. C. (1975). Growth and ectomycorrhizal development of loblolly pine seedlings in fumigated soil infested with the fungal symbiont Pisolilhus tinctorius. Forest Science 21, 245-54. O'Brien, T. P., and McCully, M. E. (1981). ‘The Study of Plant Structure: Principles and Selected Methods.’ (Termarcarphi: Melbourne.) Pallaghy, C. K. (1973). Electron probe microanalysis of potassium and chloride in freeze-substituted leaf sections of Zea mays. Australian Journal of Biological Sciences 26, 1015-34. Roomans, G. M. (1980). Localization of divalent cations in phosphate-rich cytoplasmic granules in yeast. Phvsiologia Plantarum 48, 47-50. Salema, R., and Brandao, 1. (1973). The use of PIPES buffer in the fixation of plant cells for electron microscopy. Journal of Submicroscopic Cytology 5, 79-96. Spurr, A. R. (1969). A low-viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructure Research 26, 31-43. Strullu, D. G., Gourret, J. P., Garrec, J. P., and Fourcy, A. (1981). Ultrastructure and electron-probe microanalysis of the metachromatic vacuolar granules occurring in Ta.xus mvcorrhizas. New Phytologist 87, 537-45. Strullu, D. G., Harley, J. L., Gourret, J. P., and Garrec, J. P. (1982). Ultrastructure and microanalysis of polyphosphate granules of the ectomycorrhizas of Fagus svlvatica. New Phytologist 92, 417-23. Thilo, E. (1962). Condensed phosphates and arsenates. Advances in Inorganic Chemistry and Radiochemistry 4, 1-77. White, J. A., and Brown, M. F. (1979). Ultrastructure and X-ray analysis'of phosphorus granules in a vesicular-arbuscular mycorrhizal fungus. Canadian Journal of Botany 57, 2812-8.

Manuscript accepted 29 November 1988 Endocytobiology IV Lyon. July 4-8. 1989

Freeze-substitution and X-ray microanalysis of polyphosphate granules in the mycorrhizal fungus Pisolithus tinctorius (Pers.) Coker & Couch

D. A. ORLOVICH*, A. E. ASHFORD*, G. C. COX** and A. E. P. MOORE*** * University of New South Wales, School of Biological Science, P. O. Box 1, Kensington, NSW 2033, Australia. **University of Sydney, Electron Microscope Unit, NSW 2006, Australia. *** University of Guelph, Department of Botany, Guelph, Ont. NIG 2W1, Canada.

Introduction Polyphosphate (polyP) granules occur in most microorganisms and are thought to represent a store of inorganic phosphorus in insoluble form (KECK and STICH, 1957; HAROLD, 1966). In mycorrhizas, the ability of the fungal symbiont to store inorganic phosphorus in polyP granules may be of great benefit to the host where phosphorus is transiently available. It would be expected that cations would associate with polyP to balance the negative charge it carries. Histochemical evidence from fixed material indicates that in Pisolithus tinctorius (Pers.) Coker & Couch, proteins, complex polysaccharides, lipids and RNA are not associated with polyP granules in vivo (D. A. ORLOVICH, unpublished data). In fixed material, calcium has invariably been found to be the cation associated with polyP in mycorrhizal fungi (WHITE and BROWN, 1979; ASHFORD et al., 1986), whereas in freeze- substituted hyphae phosphorus, potassium and sodium but not calcium were found in the granules (ORLOVICH et al., 1989). Calcium could readily be introduced in the wet- sectioning step. This indicated that the cation composition of granules in fixed material was not likely to be a reliable indication of their composition in vivo. However we were not satisfied that the freeze-substitution schedule used in these early experiments was retaining the in vivo composition. Acetone was used as a substitution solvent (see HARVEY, 1980) and special care was not taken to ensure that water was excluded from the material at all stages of preparation (see PALLAGHY, 1973 and MARSHALL, 1980). Using more rigorously controlled experimental conditions (i.e. use of anhydrous ether/acrolein as substitution solvent, a dry-box for infiltration, storage of all solvents and resin over molecular sieve prior to use, and inclusion of filter paper soaked in a water soluble dye to test for redistribution of mobile substances (see CANNY and McCULLY, 1986)), we have attempted to determine the in vivo elemental composition of polyP granules in the ectomycorrhizal fungus P. tinctorius. To gain some idea of the relative proportions of the elements detected in the granules, the X-ray spectra were compared with spectra from analyses of resin-embedded crystal standards for which the atomic ratios of the component elements were known. Materials and Methods The fungus Pisolithus tinctorius (Pers.) Coker & Couch (isolate DI-15, GRENVILLE et al., 1986) was grown at 25°C in the dark on Modified Melin Norkrans agar medium (MARX and BRYAN, 1975). Small pieces of mycelium (2mm x 2mm) were cut 140 from the growing edge of 4-week-old cultures and freeze-substituted as follows by the method of CANNY and MCCULLY (1986). Excess agar was removed and the material was frozen in a 1:1 mixture of isopentane and methylcyclohexane cooled by liquid nitrogen and then transferred to a pool of melted liquid overlying frozen 10% distilled acrolein (Polysciences Inc.) in anhydrous diethyl ether. The substitution mixture was refrozen in liquid nitrogen and the vials transferred to a melting acetone bath which was then allowed to warm up to -80°C in a deep freeze. They were substituted for seven days at -80°C, then warmed to room temperature overnight. The vials were transferred to a dry box and infiltrated for seven days with Spurr's resin (SPURR, 1969) dried over molecular sieve. Molecular sieve was added to the vials during infiltration and fresh sieve was used for each resin change. The material was polymerised in fresh dried resin at 70°C overnight. Polymerised blocks were stored over silica gel until used. The blocks were sectioned within 5 min of EDX analysis. Sections (1 to 1.5 pm thick) were cut dry using glass knives and mounted dry on carbon coated nylon mesh grids. It was necessary to sandwich the section between two grids and apply pressure to cause sections to stick adequately to either the top or bottom grid. Sections were not stained. As an internal control, small pieces of filter paper (Whatman No. 4) were soaked in 10'4 mol L1 sulforhodamine G (Aldrich). Half were frozen as for the fungal material and added to one of the substitution vials containing mycelial pieces. The other half were oven dried overnight (60°C) and embedded directly in SPURR's resin. For analytical standards, crystals of analytical grade KH2P04 were finely ground with a mortar and pestle and a small amount of the powder was placed in the bottom of BEEM capsules. Spurr's resin was added, air bubbles were gently tapped out, and the blocks were polymerised at 70°C overnight. Sections (1 to 1.5 pm thick) were treated as for the freeze- substituted material. The grids were mounted in a graphite holder and examined using a JEOL JEM 100CX scanning transmission electron microscope (STEM) at lOOkV accelerating voltage. Sections were analysed using a Tracor Northern TN 5500 energy dispersive X-ray analyser. Analysis time was 100 s (live detector time). Spectra were obtained from 3 replicate blocks freeze-substituted in separate vials. Results and Discussion The capacity of the freeze-substitution method to retain small mobile molecules is indicated by the retention of the sulforhodamine G on the filter paper throughout the procedure. There was no detectable difference in the dye colour of the filter paper embedded following freeze-substitution compared with sulforhodamine-soaked filter paper dried directly in an oven and embedded. Two types of granule were found. Both were electron opaque. Most granules were within the range 0.4-1.4 pm and were heterogeneous, as shown in Fig. 1. However a few cells contained smaller (0.1-0.4 pm) granules which were more uniformly electron-opaque (Fig. 2), although these often also contained areas of mottled appearance characteristic of polyP granules. The two granule types were found in different cells. Both contained phosphorus but had a different elemental composition. EDX analysis of the commonest granules, similar in appearance to those shown in Fig. 1, indicated that they contained P and K (Figs 3 and 4). The peak for Cl in the spectra was attributed to the Spurr's resin: analysis of resin outside the specimen showed only a Cl peak and this is characteristic (PALLAGHY, 1973). Electron-lucent regions of the section adjacent to granules usually also showed spectra with only Cl peaks but occasionally there were also small P and K peaks (Fig. 5). Spectra of the smaller more homogeneous granules, as shown in Fig. 2, showed a major Ca peak in addition to a P peak (Fig. 6). Occasionally small Mg and K peaks were found but neither element was detected in large quantities in these granules. No cell analysed contained both K- and Ca-rich granules. The Ca-rich granules were much less frequently found than the K-rich granules but we have not yet attempted to quantify the abundance or distribution of each type. Since the K-rich granules were the most abundant, we attempted to determine the relative proportions of P and K in the granules to see whether K was present in sufficient 141

FIG. I. Thick unstained freeze-substituted cell containing large electron-dense, mottled granules (*). Analysis of the labelled granule in this cell produced the spectrum shown in Figure 3. Bar = lpm. Viewed in STEM mode. FIG. 2. Thick unstained freeze-substituted cell containing smaller electron-dense granules (*). The arrowheads show the position of the cell wall. Granules in this cell gave spectra very similar to that shown in Fig. 6. Bar = l pm. Viewed in STEM mode. amounts to act as the only cation balancing the negative charge of the polyP. To balance the charge P and K would need to be present in equal amounts in the granules. Although equal peak heights for different elements do not indicate equivalent amounts of those elements, comparison of relative peak heights does give an indication of relative amounts of particular elements. To assess ratios of P and K in granules we compared spectra of granules with those of embedded crystals of known elemental composition. Analysis of anhydrous potassium dihydrogen orthophosphate crystals (which have equivalent amounts of K and P) embedded in Spurn's resin and sectioned at c. 1.5 pm gave the spectra shown in Figs 7 and 8. Mass loss was not seen during analysis of very large (greater than 5 pm in profile) crystals (Fig. 7) but did occur during analysis of small (less than 1 pm) crystals (Fig. 8). The effect was progressively to reduce the height of the K peak relative to the P peak as analysis time increased. With the crystal used to produce the spectrum shown in Fig. 8 (analysis time 100 s), initially the K peak was higher than the P peak. Then after 40 s of analysis, the K and P peaks were of equal height and at 100 s, as shown in Fig. 8, the K peak was lower. The crystal became more electron-lucent in the area of analysis indicating that mass loss had occurred. This was also observed during the analysis of the polyP granules resulting in a reduction of the height of the K peak relative to the P peak in each spectrum, as seen with the crystal: Taking this into account, comparison of spectra of the crystals and K-rich granules indicates that the relative amounts of K in the granules are sufficient to balance the negative charge on the polyP. The finding here that K is the major element accompanying P in polvP granules agrees with results obtained from earlier analyses of similar fungal material freeze- substituted by less rigorous procedures (ORLOVICFI et al., 1989) and differs fundamentally from results obtained with chemically fixed material, where Ca was invariably the major element apart 142

LI Cl

1024 1024

2048

i l Cl

LEGENDS OF THE FIGURES 3-8: FIG. 3. EDX spectrum from the analysis of the granule shown in Fig. I. The vertical full scale is shown at the top left hand side of each spectrum. Note that this is not the same in all spectra. The spectrum shows large peaks for P and K and a small Cl peak. The small peak to the right of the K peak is the K (Kp peak, not a Ca peak. FIG. 4. Another EDX spectrum from the analysis of freeze-substituted granules with a similar appearance to those shown in Fig. 1. FIG. 5. EDX spectrum from an electron-lucent region adjacent to the granule which gave the spectrum shown in Fig. 4. There are small P, K and Cl peaks. FIG. 6. EDX spectrum from a granule similar to those shown in Fig. 2. The spectrum has large peaks for P and Ca and a small Cl peak. FIG. 7. EDX spectrum from a large potassium dihydrogen orthophosphate crystal embedded in Spurr's resin. There is a large K peak and a smaller P peak. The peak to the right of the K peak is the K (Kp) peak. FIG. 8. EDX spectrum from a small potassium dihydrogen orthophosphate crystal embedded in Spurr's resin for comparison with Fig. 7 (note the large difference in vertical scale). There is a high P peak, a smaller K peak and a small Cl peak. 143 from P (WHITE and BROWN, 1979; ASHFORD et al., 1986). The levels of K detected in the K-rich granules were consistently much higher and sodium was not found when more care was taken to exclude water at all stages. The use of resin stored over molecular sieve, inclusion of molecular sieve in the vials during infiltration and exclusion of water vapour by use of the dry box should have ensured mininal contact with water during the present schedule. The retention of the sulforhodamine G on the filter paper freeze-substituted together with the specimens is evidence that water was excluded successfully, since the dye leaks out of specimens very rapidly into water or incompletely dried resin (M.E. McCully, personal communication) and will diffuse from oven-dried, sulforhodamine-impregnated filter paper into resin that is known to contain water (unpublished observations). The finding that there is a small number of Ca-rich granules and that these are found in different cells from the K-rich granules requires further confirmation. Since Ca would be expected to replace other cations frorti granules in procedures where water is not excluded, we were concerned whether Ca may have been introduced into these granules as a result of ion redistribution in cut cells during removal of mycelium samples prior to freezing. At present we have insufficient data to eliminate this possibility. However it seems unlikely in view of the consistent and different appearance of the Ca-containing granules. Although we have not shown here that the granules analysed are in vacuoles, this has commonly been observed in fungal tissues (see ASHFORD et al., 1986) and, since polyphosphate is a very efficient cation exchanger, the composition of granules might be expected simply to reflect what is available in the vacuole. This would be determined by the specific properties of the vacuolar membrane of that cell. In this case it is surprising that there is so little variability in granule composition. However there are parallels in some seeds (DWARTE and ASHFORD, 1982) where protein- rich vacuoles contain either calcium oxalate crystals or phytin (inositol hexaphosphate) globoids low in Ca, and these are found in different cells of the same tissue. This implies an ion sorting mechanism operating to store specific ions in specific cells. A similar mechanism may well be operating in cells of the fungal mycelium. The role of polyP as a cation exchanger in the cell and the implications for metabolism and cation release when the phosphorus is remobilised remain to be determined.

ACKNOWLEDGMENTS: The work was undertaken while D.A.O. and A.E.A. were visiting the laboratories of Dr. M.E. McCully, Carleton University and Dr. R.L. Peterson, University of Guelph and we thank them for provision of facilities. We also thank Drs M.J. Canny and M.E. McCully for advice and assistance with the freeze-substitution procedure, Dr W.G. Allaway for comments and help with the manuscript, Dr C. A. Peterson for loan of a dry box at Guelph and Mr. L.H. Melville for assistance with electron microscopy. Financial support from the Australian Research Grants Scheme and the National Science and Engineering Research Council of Canada to A.E.A., from the Australian Commonwealth Post Graduate Scheme as a scholarship to D.A.O., and from the Universities of Guelph and New South Wales, are gratefully acknowledged. References ASHFORD A.E., PETERSON R.L., DWARTE D. and CHILVERS G.A., 1986. Can. J. Bot., 64, 677-687. CANNY MJ. and MCCULLY M.E., 1986. J. Microsc., 142, 63-70. DWARTE D. and ASHFORD A.E., 1982. Bot. Gaz., 143, 164-175. GRENVILLE D.J., PETERSON R.L. and ASHFORD A.E., 1986. Aust. J. Bot., 34 , 95-102. HAROLD F.M., 1966. Bacteriol. Rev., 30, 772-794. HARVEY D.M.R., 1980. Scanning Electron Microscopy, 1980/11, 409-419. KECK K. and STICH H., 1957. Ann. Bot., 21, 611-619. MARSHALL A.T., 1980. Scanning Electron Microscopy, 1980/11, 395-408. MARX D.H. and BRYAN W.C., 1975. For. Sci., 21, 245-254. ORLOVICH D.A., ASHFORD A.E. and COX G.C., 1989. Aust. J. Plant Physiol., 16, 107-115. PALLAGHY C.K., 1973. Aust. J. biol. Sci., 26, 1015-1034. SPURR A.R., 1969.7. Ultrastruct. Res., 26, 31-43. WHITE J.A. and BROWN M.F., 1979. Can. J. Bot., 57, 2812-1218. ENDOCYTOBIOLOGY IV

4th International Colloquium on Endocytobiology and Symbiosis

4eme Congres International sur l’Endocytobiologie et la Symbiose

INSA - Villeurbanne (France) July 4-8, 1989

Editors : P. NARDON, V. GIANINAZZI-PEARSON, A.M. GRENIER, L. MARGULIS and D.C. SMITH

ENSTTTUT NATIONAL DE LA RECHERCHE AGRONOMIQUE 147, rue de l’Universite - 75341 Paris Cedex 07 The following articles have been removed from the digital copy of this thesis. Please see the print copy of the thesis for a complete manuscript.

Title: Evidence that abaxial leaf glands are the sites of salt secretion in leaves of the mangrove Avicennia marina (Forsk.) Vierh. Authors: MELISSA A. FITZGERALD, DAVID A. ORLOVICH and WILLIAM G. ALLAWAY Journal: New Phytologist

Title: Polyphosphate granules are an artefact of specimen preparation in the ectomycorrhizal fungus Pisolithus tinctorius Authors: D. A. Orlovich and A. E. Ashford Journal: Protoplasma (1993) 173: 91-102

Title: Association of polyphosphate with protein in freeze-substituted sclerotia of Sclerotinia minor Authors: Nicola Young, Suzanne Bullock, D. A. Orlovich, and Anne E. Ashford Journal: Protoplasma (1993) 174: 134-141

Title: Cessation of cytoplasmic streaming follows an increase of cytoplasmic Ca2 + during action potential in Nitella Authors: M. Kikuyama, K. Shimada, and Y. Hiramoto Journal: Protoplasma

Title: A dynamic continuum of pleiomorphic tubules and vacuoles in growing hyphae of a fungus Authors: V. A. Shepherd*, D. A. Orlovich and A. E. Ashford Journal: Journal of Cell Science

Title: Cell-to-cell transport via motile tubules in growing hyphae of a fungus Authors: V. A. Shepherd, D. A. Orlovich and A. E. Ashford Journal: Journal of Cell Science

Title: Structural and development of the dolipore septum in Pisolithus tinctoriu Authors: R. A. Orlovich* and Anne E. Ashford