Testing spore amyloidity in Agaricales under light microscope: the case study of Tricholoma

Although species of the genus Tricholoma are currently considered to produce inamyloid spores, a novel standardized method to test sporal amyloidity (which involves heating the sample in Melzer’s reagent) showed evidence that in the tested species of this genus, which belong in all 10 sections currently recognized from Europe, the spores are amyloid. In two species, T. josserandii and T. terreum, the spores are also partly dextrinoid. This result provides strong indication that a positive reaction of the spores in Melzer’s reagent could be a character shared by all genera in Tricholomataceae s. str.


Introduction
It has been known for about 150 years that some ascomycete and basidiomycete sporomata may contain elements which stain grey to blue-black with iodinecontaining solutions. Such a staining was termed amyloid reaction, sometimes written as I+ or J+ (the term "amyloid" being derived from the Latin amyloideus, i.e. starch-like), because plant starch gives a similar reaction with iodine (starch-reaction) (Bailey and Whelan 1961;Locquin and Langeron 1978;Immel and Lichtenthaler 2000). The blue colour of the stain is due to the amylose component of plant starch (Takahashi and Ono 1972;Bluhm and Zugenmaier 1981;Moulik and Gupta 1984;Murdoch 1992;Immel and Lichtenthaler 2000). The amylose chain forms a helix shape, and iodine (as triiodide anion I 3 − ) can be bound inside the helix channel. The other component, amylopectin, gives a red-purple colour which is much less intense than the amylose stain (Bailey and Whelan 1961;Shen et al. 2013). The nature of the starch-iodine interaction is extremely complex and still remains imperfectly known (Bluhm and Zugenmaier 1981;Immel and Lichtenthaler 2000;Shen et al. 2013;Du et al. 2014;Okuda et al. 2020).
An overview of the historical use of Melzer's was provided by Leonard (2006). Iodine was used in Mycology in the mid-1800s (as alcoholic solutions) mainly for studying lichens and asci (entire ascus wall or apical apparatus). The earliest reference to the use of such a bluing reaction in fungi as a character having a systematic significance is a report of the bluing of a cleistothecial marine ascomycete, Amylocarpus encephaloides by Currey (1859). Then the Tulasne brothers (1865), Nylander (1865) and Rolland (1887) noted and described iodine bluing in lichens and Ascomycetes at ascus and ascospore level. Boudier (1885,(1905)(1906)(1907)(1908)(1909)(1910) was the first to describe and illustrate the use of iodine to classify apothecial Ascomycetes. The first reports of iodine solutions used in Basidiomycetes are those of Patouillard (1887), who described a violet coloration with iodine of the spores of Cyphella vitellina (now Aleurodiscus vitellinus), and of Rolland (1887), who signaled a blue coloration in the stipe trama of Mycena tenerrima. At the time, however, such colour changes observed in microanatomical elements of Basidiomycetes were not considered to be systematically relevant by the above mentioned authors. As highlighted by Leonard (2006), it is worth noting that the following year Patouillard (1888) published a paper in which he stressed the importance of spore colour in the classification of fungi, even if he did not mention the use of iodine.
The turning point came when Melzer (1924), after studying Russula species for many years, developed a method based on the use of an iodine reagent mixed with chloral hydrate to clear and stain microscopical preparations, which made it possible to study the fine details of the sporal ornamentation of Russula species. Baral (1987a) believed that Melzer was "probably inspired by Arthur Meyer", a German botanist, renowned for his studies on chloroplasts, who introduced the use of chloral hydrate/iodine in Botany under the name "Chloraljod" (Meyer 1883). The formula for what became later known as Melzer's reagent -and which quickly came into standard use in mycology -is as follows: KI -1.5 g, iodine -0.5 g, distilled water -20 g, chloral hydrate -22 g). The resulting brownish solution stained dark black-blue the Russula sporal ornamentations.
With this reagent, Melzer supplied mycologists with an extraordinarily valuable tool, which has now become so widely used that the term Melzer's solution (or reagent or simply "Melzer's") appears in almost all taxonomic works on the higher fungi. At first, with the exception of some papers by Boursier (1925) and Kühner (1926a, b), Melzer's reagent was only used by Russula specialists because it was invaluable in allowing for an indepth study of the ornamentation of Russula spores (Crawshay 1930;Moreau 1930;Malençon 1931;Singer 1932).
However, this new reagent was soon also used for identification purposes in other groups of pale-spored agarics. Gilbert and Kühner (1928) highlighted that the spore wall is amyloid (blue) in some species of Amanita, but unreactive (or inamyloid, i.e. remaining pale yellow-brown or clear) in others. Positive iodine reactions of the spore wall were observed by Kühner (1931Kühner ( , 1936aKühner ( , 1938 in many species of the genus Mycena and, by Kühner and Maire (1934), in many other white-spored agarics. Further, Metrod (1932), Kühner and Maire (1934) and Kühner (1936b) discovered that Melzer's reagent caused a different, purple-red staining in the spore wall of many species of the genus Lepiota and in the basidiome hyphae (especially those of the stipe in Mycena) of other agarics. Kühner and Maire (1934) and Kühner (1936a, b) used indiscriminately the term "amyloid" for both the blue and the purple-red reaction on elements of the basidiome. The dark red-brown reaction in Melzer's reagent was later named pseudoamyloidity by Singer (1938Singer ( , 1951 and some decades later, dextrinoidity by Orton (1960). Gilbert (1940) used the following modified solution for Amanita spores: distilled water 10 ml, KI 0.5 g, iodine 0.05 g, chloral hydrate 10 g. This solution is almost colourless under the microscope, but the Amanita spore walls turn dark grey. Langeron (1945), in the formulation of Melzers' s reagent included in his book, used 22 g of chloral hydrate instead of Melzer's 20 g, but there was no discussion as to why he changed the proportions. This change in the proportions of Melzer's is now known as "Langeron's modification" (Ainsworth 1961) and was adopted by Singer (1951Singer ( , 1963Singer ( , 1975Singer ( , 1986. The contributions of further authors to the generalized use of the Melzer's reagent led to rapid progress in fungal taxonomy. This was not only especially significant in the taxonomy of the pale-spored agarics (Kühner and Romagnesi 1953;Moser 1967Moser , 1978Kühner 1980Kühner , 1984Singer 1982Singer , 1986Horak 2005;Bas et al. 1988;Ludwig 2012;Knudsen and Vesterholt 2018), but it was also beneficial to other non-agaricoid fungal groups in Agaricomycotina (e.g., Imler 1950;Donk 1964;Miller 1964;Smith and Zeller 1966;Dodd 1972;Jülich and Stalpers 1980;Watling 1980;Jülich 1984;Hjortstam et al. 1988;Stalpers 1996;Ginns 1998;Ryvarden and Melo 2014;Agerer 2018) and Pezizomycotina (Kohn and Korf 1975;Nannfeldt 1976;Rossman 1980;Baral 1987aBaral , b, 2007Baral , 2009Common 1991;Frey 2016).
In some papers (Kohn and Korf 1975;Nannfeldt 1976;Redhead 1977;Baral 1987aBaral , b, 2007Baral , 2009) the importance of using a KOH pretreatment of the sample to enforce amyloid reactions in ascal tips was stressed. Baral (1987aBaral ( , b, 2007Baral ( , 2009) highlighted that Lugol's, a chloral hydrate-free iodine solution developed in 1829 by the French physician J.G.A. Lugol (initially as a cure for tuberculosis, Calissendorff and Falhammar 2017) is more efficient than Melzer's in inducing amyloid reactions on apical apparatus in asci.
The chemical components of Melzer's reagent have different functions. Chloral hydrate, being a clearing agent, clarifies and improves the transparency of various dark-colored microscopic materials. Iodine is hardly soluble in water, therefore, potassium iodide is used to improve its solubility (thanks to the reaction I 2 + I − = I 3 − ); iodine is thought to be the main active staining agent in Melzer's, even if its mechanism of action in the amyloid reaction is not entirely understood. It is thought to react/interact with starch-like polysaccharides present in the walls of spores and hyphae, but data in this regard are dated and incomplete (Smith 1965;Blackwell et al. 1985;Morton 1986;Leonard 2006). Dodd 1971, Dodd andMcCracken (1972) and McCracken et al. (1973) noted that fungal "starch" is different from plant starch in that it is not produced in plastids, is not in granular form, is mainly a cell-wall component (rather than an energy source), and is made up of "only short-chained amylose molecules." They hypothesized that the amylose in the spore cell helped the spore stay viable until conditions were good for germination. Webster and Weber (2007) suggested that the amylose-like substances prevent oxygen from entering the spore, thereby enforcing dormancy by slowing down metabolism, a hypothesis re-proposed by Halbwachs and Bässler (2015) in their review of spore morphology in agarics. Amylose is insoluble in water and also more resistant to digestion compared to other forms of starch (i.e. amylopectin) (Chen 1990;Bertoft 2017), which could then also contribute to a longer spore dormancy.
A completely different cell wall composition should be the basis of dextrinoidity. Blackwell et al. (2001) suggested that the red-brown reaction with Melzer's reagent does not involve starch or amylose, but is a reaction with glycine betaine, an osmolyte (an organic osmotic solute) which they found in high concentrations in the Basidiomycetes they studied (Antrodia carbonica, Calvatia sp., Chlorophyllum molybdites, Ganoderma lucidum and Laetiporus sulphureus).
Focusing on the spores of the pale-spored agarics, the amyloid positive reactions obtained with this medium range from pale grey to blue to violet-black (Meixner 1975;Locquin and Langeron 1978;Erb and Matheis 1983;Brunori et al. 1985;Singer 1986;Charbonnel 2004;Basso 2005). The solution can be stored for a long time (Ammirati et al. 1985;Clémençon 2000Clémençon , 2012Lecomte 2017). Usually a positive reaction occurs almost immediately; when the reaction is doubtful, leaving the material in solution for 30 min is recommended (Clémençon 2012). The amyloid reaction may not be of the same intensity on the entire surface of a spore (e.g. Mycena, Gloiothele, H. Clémençon, pers. comm.). Fresh material sometimes gives a slower reaction than dried material. Melzer's cannot be used in combination with alkali because iodine reacts with hydroxide ions and a cloudy precipitate develops. Consequently, when potassium hydroxide or ammonium are used as a pretreatment, the alkalinity must be first neutralized before adding Melzer's (Largent et al. 1977;Ammirati et al. 1985).
At present, it is also known that the amyloid reaction of spore ornamentations is a character that may have evolved several times. Compelling evidence for this is offered by Leucopaxillus and Melanoleuca, genera traditionally included in the tribe Leucopaxilleae or subfamily Leucopaxilloideae of the Tricholomataceae (Singer 1986;Bon 1991) but which recent molecular studies (Matheny et al. 2006;Dentinger et al. 2016) now place in the tricholomatoid clade (suborder Tricholomatineae) and in the Pluteoid clade (suborder Pluteineae), respectively.
The amyloid reaction of the spores can sometimes be so weak as to be difficult to assess properly [e.g. in Pseudoclitocybe (Bigelow 1982); some species of Mycena sect. Calodontes (Harder et al. 2012); Pseudolaccaria (Lavorato et al. 2015); Musumecia ] and, consequently, there might be some subjectivity on the part of the operator. To avoid the possible distorting consequences of subjectivity in the assessment of this character, a novel standardized method to test sporal amyloidy has recently been used by the present authors (Alvarado et al. 2018). It allowed us to obtain unambiguous results in cases (e.g. in Pseudoclitocybaceae, Alvarado et al. 2018) in which the traditional testing method led to results of dubious interpretation. Based on these outcomes, a project was set up to test with this new method amyloid reactions in the spores of Agaricales when observed under light microscopy. It was decided to start, in the present paper, with the species of the large genus Tricholoma, typified with Tricholoma flavovirens (= T. equestre), a widely distributed genus of ectomycorrhizal agarics that was traditionally considered to produce inamyloid spores (Kühner and Maire 1934;Bon 1984Bon , 1991Singer 1986;Riva 1988Riva , 2003Christensen and Heilmann-Clausen 2013). However, it has recently been shown to belong to a clade (Tricholomataceae s. stricto) containing mostly taxa with positive reaction of the spores in Melzer's reagent (Sánchez-García et al. 2014). Vizzini et al. IMA Fungus (2020) 11:24 Page 3 of 20

Morphology
The eighteen Tricholoma collections (corresponding to seventeen species) used for the microscopic analyses ( Fig. 1, Table 1) were identified using the monographic work by Christensen and Heilmann-Clausen (2013) and were selected to represent all the ten sections of Tricholoma recognized in Christensen and Heilmann-Clausen  2017), were also studied. Seventeen of these collections have been deposited at AMB and sequenced in the present work, while for one, T. lascivum C-F-96230 (shown in Fig. 1 with an asterisk), the nrITS sequence was already present in the public databases GenBank ( Tulloss and Possiel 2020), were tested with all the methods (under all the different conditions) above mentioned. These species were chosen as an additional control for the new proposed method. A spore print on glass (microscope slides) obtained from a basidiome of T. josserandii (TO AV260920) was tested for amyloidity adding directly a drop of Melzer's and observed after complete absorption/evaporation of the reagent (2 h) (Fig. 8).
Microscopic pictures were taken on a Canon Power Shot A640 digital camera connected to a Zeiss Axioscope 40 microscope with both interferential contrast and phase-contrast optics.

DNA extraction, amplification and sequencing
Total DNA was extracted from dry specimens (Table 2) employing a modified protocol based on Murray and Thompson (1980). PCR amplification (Mullis and Faloona 1987) included 35 cycles with an annealing temperature of 54°C. Primers ITS1F and ITS4 (White et al. 1990;Gardes and Bruns 1993) were employed for the region. PCR products were checked in 1% agarose gels, and positive reactions were sequenced with one or both PCR primers. Chromatograms were checked searching for putative reading errors, and these were corrected. The accession numbers of the sequences are reported in Table 2.

Phylogenetic analyses
The nrITS dataset was assembled based on that of Heilmann-Clausen et al. (2017). BLASTn (Altschul et al. 1990) was used to select the most closely related sequences from GenBank and UNITE. Hypsizygus marmoreus (HM561970) was used as outgroup taxon to root the tree.
Sequences first were aligned in MEGA 6.0 software (Tamura et al. 2013) with its MUSCLE application (Edgar 2004) and then corrected manually. The nrITS alignment was not partitioned into ITS1, 5.8S and ITS2 regions. The Bayesian analysis was performed through the CIPRES Science Gateway platform (Miller et al. 2010) by using the MrBayes v. 3.2.7 algorithm with two simultaneous runs, four chains, temperature fixed at 0.2 and sampling every 1000 generations until reaching the convergence parameters (standard deviation less than 0.01). The first 25% trees were discarded as burn-in. Finally, a full search for the best-scoring Maximum likelihood tree was performed in RAxML v.8.2.10 (Stamatakis 2014) using the standard search algorithm (GTRCAT model, 2000 bootstrap replications). Significance threshold was set ≥0.95 for posterior probability (PP) and ≥ 70% for bootstrap proportions (BP).

Results
The correct determination of all the Tricholoma collections used in this work was confirmed molecularly (Fig. 2). The ten sections (major clades) recognized in Heilmann-Clausen et al. (2017) are also recovered in the present analysis. Tricholoma apium, T. arvernense, T. fucatum and T. josserandii are not included in these sections and occupy an isolated position. Vizzini et al. IMA Fungus (2020) 11:24 Page 4 of 20 All the Tricholoma collections showed spores with at least partly dark grey to blue amyloid walls (apiculus included) under the light microscope both by the novel standardized protocol (HM) and the classical method with spores observed after 30 min. (CM30) and 120 m. (CM120), regardless of the soaking medium used (Table 1) (Figs. 3 and 4c-i). The amyloid reaction was also observed on the apical part of basidia (Figs. 3a-d and 4g, i) with both HM and CM30 Spores of Tricholoma josserandii and T. terreum, which were negative with the CMi method (Table 1) (Figs. 5a-b and 6a-b) showed an evident dextrinoid reaction (Figs. 5c-f and 6c-f) coupled or not with an amyloid plaque reaction ( Fig. 3c-d, j) only with the HM and CM30 methods, regardless of the soaking medium used. The spore print of T. josserandii is amyloid (Fig. 8c). Spores of Amanita phalloides are clearly amyloid (Fig. 7ae) and spores of A. argentea are inamyloid (Fig. 7f-h) by whatever method is used. Some apically amyloid basidia are also present in A. phalloides (Fig. 7b-d).

Spore amyloidity
Species identification of the Tricholoma collections analyzed, based on pileus colour, structure of the pileipellis, presence/absence of clamp-connections, size and shape of the basidiospores, was supported by molecular data (Fig. 2). All of the eighteen collections studied (seventeen species), which include taxa representative of all the ten sections recognized in Tricholoma in Europe by Christensen and Heilmann-Clausen (2013) and Heilmann-Clausen et al. (2017), showed spores with a positive reaction in Melzer's reagent under light microscope (Figs. 3, 4c-i, 5c-f and 6c-f) by using the preheating and the CM30 and CM120 methods ( Table 1). The spore print of T. josserandii shows an amyloid reaction when treated with a drop of Melzer's (Fig. 8).
The type of soaking medium used does not appear to have any influence on the iodine reactions of the spore wall, but the use of KOH is not recommended because it can sometimes induce distortions in the shape of some spores; furthermore, if the excess KOH is not well removed, hydroxide ions can react with iodine producing cloudy troublesome precipitates. Instead, very important factors are the preheating or not of the sample and the immersion time of the spores in Melzer's solution.
With the new method (HM), the amyloid reaction is also evident at the apex of the basidia and sterigmata (Fig. 3a-d). This would seem to indicate structural continuity and an identity of chemical composition between the basidial apex wall and the spore wall. By contrast, the amyloidity test of the same above-mentioned seventeen species, using the classical method (without heating) with spores observed immediately after being mounted in Melzer's (CMi), produced no clear results (Table 1) .
As indicated by Clémençon (2000Clémençon ( , 2012, pitfalls in assessment of spore amyloidity may be due to several causes, among which first and foremost allowing insufficient time for the reaction to complete. In Mycena, he observed that, leaving the sample in Melzer's for at least 20 min, the spores became black even in species which show very weak reactions (barely discernible under the microscope) when examined in Melzer's after a short time. Consequently, we tested the eighteen collections also following Clémençon's advice. First, the samples were soaked for 2 min in 5% ammonia ( Fig. 4c-g) or for 2 h in tap water ( Fig. 4h-i), then they were transferred to, and squashed in Melzer's. No amyloid spores could be observed immediately after squashing, but 30 min later some amyloid spores were clearly visible ( Fig. 4c-g). Re-observed after an additional 30 min, the number of amyloid spores and the intensity of the amyloid reaction did not change and the same happened after 120 min. Even with the CM30 method some basidia are apically amyloid (Fig. 4g-i).
Interestingly, the spores of T. josserandii and T. terreum demonstrated a dextrinoid reaction in  and 6c-f), sometimes with small amyloid portions j). To our knowledge this situation is absolutely unprecedented in the Agaricales and in all Basidiomycota.
It is not easy to explain the discrepancy between our results and those of all the other authors who have studied the genus Tricholoma. It might be possible that preheating the sample in Melzer's solution somehow allows the starch-like wall material to better interact with the chemical agent and so make the colour reaction much more visible than when it is obtained in Melzer's at room temperature. The same may happen in unheated samples if the observation is made after waiting the right amount of time (at least 30 min), that is giving iodine the necessary time to get in touch with the amyloid material in the sporal wall. Whenever it is necessary to highlight the amyloidity of spores of species long known as clearly amyloid (see for example A. phalloides, Fig. 7a-e), any method proves effective. But whenever it is intended to analyze the spores of species with known dubious or weak amyloidity (e.g., Tricholoma in the present study; Musumecia, Pseudolaccaria, Alvarado et al. 2018;Vizzini et al. 2020) or of species not yet described or not tested, a negative reaction with the quick classical method (without pre-heating and observation of spores in Melzer's immediately after squashing, CMi) must be regarded as non-definitive and a stepwise process should be carried out.
In the description of tricholomatoid fungi and in preparing dichotomous keys we would like to suggest to indicate the spores which are amyloid after the CMi method as "immediately amyloid"; after CM30 as "tardily amyloid"; and after HM as "amyloid after pre-heating".
In conclusion, the authors recommend that the heating method described here, which is as effective as the CM30 but much faster, becomes of common use, and hope that in the future, taxonomic workers will routinely turn to "hot Melzer's" as a check when determining whether a species which appears to be iodine-negative is truly iodine-negative.
In the light of the data reported in the present paper, and considering that in the monospecific genus Albomagister has recently included a second species with faintly amyloid spores by Moreau et al. (2015), it seems that a positive reaction of the spores in Melzer's reagent could be a character shared by all genera in Tricholomataceae s. str. and consequently be considered, in this case, a systematic marker at the family level.

Conclusions
At variance with what has been hitherto believed, the spores of the Tricholoma collections sampled here have proved to be amyloid when treated with a Melzer's solution (original formulation) both with the classical method (without pre-heating), which involves an immersion in Melzer's for at least 30 min, and a new method proposed in the present work, in which the sample is pre-heated. The soaking media appear to have no influence in inducing the reaction with the Melzer's solution. Before being declared inamyloid, the spores of whatever species should be tested by all the methods suggested in this paper. It is hoped that the new method with preheating will be generally adopted by mycologists and that it will become widely used. A positive reaction of the spores with Melzer's reagent appears to be a character present in all genera of Tricholomataceae s. str.