AEM Accepted Manuscript Posted Online 17 July 2015

Appl. Environ. Microbiol. doi:10.1128/AEM.01098-15

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

1 The role of the microbiome of truffles in aroma

2 formation: a meta-analysis approach
3 Maryam Vahdatzadeh1,2, Aurélie Deveau3,4, Richard Splivallo1,2,*
1
4 Goethe University Frankfurt, Institute for Molecular Biosciences, 60438 Frankfurt, Germany
2
5 Integrative Fungal Research Cluster (IPF), Georg-Voigt-Str. 14-16, 60325 Frankfurt, Germany
3
6 INRA, Interactions Arbres-Microorganismes, UMR 1136, Champenoux, F-54280, France
4
7 Université de Lorraine, Interactions Arbres-Microorganismes, UMR 1136, Vandœuvre-lès-
8 Nancy, F-54506, France
9

10 * Corresponding author: richard.splivallo@a3.epfl.ch

11

12 Format: Invited mini-review AEM

13

14

15

1
16 Summary

17 Truffles (Tuber spp.) are ascomycete subterraneous fungi that form ectomycorrhizas in a

18 symbiotic relationship with plant roots. Their fruiting bodies are appreciated for their

19 distinctive aromas which might be partially derived from microbes. Indeed truffle fruiting

20 bodies are colonized by a diverse microbial community made of bacteria, yeasts, guest

21 filamentous fungi and viruses. The aim of this mini-review is double. First the current

22 knowledge on the microbial community composition of truffles has been synthesized to

23 highlight similarities and differences among four truffle species (T. magnatum, T.

24 melanosporum, T. aestivum and T. borchii) at various stages of their life cycle. Second

25 the potential role of the microbiome in truffle aroma formation has been addressed for the

26 same four species. Our results suggest that on one side odorants which are common to

27 many truffle species might be of mixed truffle and microbial origin while on the other

28 side less common odorants might be derived from microbes only. They also highlight that

29 bacteria, the dominant group in the truffle’s microbiome, might also be the most

30 important contributors to truffle aromas not only in T. borchii as already demonstrated

31 but also in T. magnatum, T. aestivum and T. melanosporum.

32

33

34

35 Keywords: Truffle, Tuber spp., microbiome, aroma, volatile, yeast, bacteria

36

37

2
38 Introduction

39 Microbes can be found almost everywhere on our planet. They colonize many different

40 types of habitats, among them living organisms such as plant roots or insect and human

41 guts. Classical microbiological methods have for long offered a spotlight view on

42 microbial diversity. Recent high throughput molecular techniques have revolutionized the

43 field of microbial ecology by unravelling an enormous microbial diversity in numerous

44 organisms and highlighting the deep impact of microbiomes of their host physiology and

45 behaviour (1, 2). Truffles fungi are no exception since they are colonized by a complex

46 microbial community made up of bacteria, yeasts, guest filamentous fungi and viruses

47 (3–14).

48 Truffles are ascomycete subterraneous fungi that form ectomycorrhizas in symbiotic

49 relationship with plant roots (15). Their fruiting bodies are appreciated for their

50 distinctive aromas which is partially derived from microbes (6, 14, 16). The aim of this

51 mini-review is to synthesize the current knowledge on the microbial community

52 composition of truffles and discuss their potential role in truffle aroma formation,

53 specifically focusing on volatiles which are responsible for human-perceived truffle

54 aroma (defined as odorants).

55

56

57 Truffle microbiomes

58 Truffle are colonized by microbes at all stages of their life cycle, which include a

59 symbiotic stage in association with a host plant (ectomycorrhiza), a sexual stage (fruiting

3
60 bodies) and a “free living mycelial stage” which might serve for exploratory purpose in

61 the soil. To date, microbes and microbial communities have been characterized in truffles

62 with culture dependant and independent techniques in more than 15 papers (3–14, 17–

63 21). Various life cycle stages of four commercially relevant species have been

64 investigated: the white truffles T. magnatum, T. borchii and the black species T.

65 melanosporum and T. aestivum. Similarities and differences in microbial community

66 compositions among truffle species are highlighted here for bacteria, fungi (yeast and

67 filamentous) and viruses.

68

69 Bacterial communities

70 Most studies investigating microbes in truffles have been performed on bacteria. Those

71 can heavily colonize inner and outer parts of truffle fruiting bodies as their density range

72 from a million to a billion cells per gram (dry weight) of fruiting bodies (4, 5, 19–22).

73 The aim of these studies varied from the characterization of taxonomic and/or functional

74 community composition to the influence of specific variables. Indeed bacterial

75 community composition has been investigated in relation to fruiting body maturation,

76 aging, season or life cycle (i.e. mycorrhizas vs. fruiting body) and tissue specificity

77 (gleba, the inner part of the fruiting bodies versus the peridium, the outer protective

78 layer).

79 Combinations of culture dependent and independent methods have demonstrated that all

80 truffle species analysed so far are colonized by complex bacterial communities made

81 mostly of Proteobacteria, Bacteroidetes, Firmicutes and Actinobacteria (Figure 1) (4, 5,

82 12–14, 19). Similarities among fruiting bodies of the three truffle species investigated to

4
 83 date include a dominance of α-Proteobacteria and a relative paucity of Firmicutes and

84 Actinobacteria. In the contrary, differences among truffle species might exist for β- and γ

85 -Proteobacteria and Bacteroidetes which might be more abundant in T. borchii compared

86 to T. melanosporum and T. magnatum (Figure 1). As a matter of fact a Bacteroidetes

87 strain might even co-exist inside T. borchii mycelium grown under axenic laboratory

88 conditions (18), suggesting a tight possible association between bacteria and truffles. The

89 occurrence of endosymbiont has not been described so far in other truffle species.

90 Bacterial community composition of truffle fruiting bodies might evolve over time and in

91 relation to the physiology of the truffle host. Indeed truffle fruiting bodies mature as their

92 inner part (gleba) undergoes melanisation due to the spore forming process taking place

93 inside the fungal ascii. This maturation / melanisation process generally lasts a few

94 months and occurs in late autumn/winter for T. borchii, T. melanosporum and T.

95 magnatum harvested in Europe. Using FISH in the latter species, a slight but significant

96 decrease in total bacterial count was observed with increasing maturity, nevertheless no

97 difference in the relative community composition was detectable for α-, β-, γ-

98 Proteobacteria, Bacteroidetes, Firmicutes or Actinobacteria (5). A different pattern was

99 observed in T. melanosporum using high throughput sequencing methods (13). The

100 composition of the bacterial community present inside the gleba and in the peridium

101 significantly changed along the maturation of the ascocarps. The community composition

102 in the peridium was very close to the soil community in young ascocarps but strongly

103 diverged from the soil community in mature ascocarps. The differences were, mainly in

104 the peridium, due to a progressive increase in the abundance of Bacteroidetes and α-

105 Protebacteria members while the abundance of β-Proteobacteria members decreased. In

5
106 contrast, the gleba bacterial community was very early dominated by α-Protebacteria

107 members. Moreover this dominance kept increasing with the maturity level just as it did

108 in the peridium. Altogether, these data prompted Anthony Babu et al. (2014) to propose

109 the following model: soil bacteria would colonize truffle primordia before the

110 differentiation of ascocarpic tissues would occur. Then, the bacteria would be trapped in

111 the gleba and partly protected from soil exchanges by the warted peridium. Because of

112 this compartmentalization, bacterial community composition would mainly evolve in

113 response to changes in physiology of the maturing ascocarp. In contrast, the peridium

114 would remain in contact with the soil all along the development process of the ascocarp

115 due to crack openings during growth of the ascocarp (13).

116 In addition to natural variations, the harvest of truffle fruiting bodies is likely to induce

117 changes in the composition of the associated bacterial community. This could be due to

118 modifications of physico-chemical parameters such as temperature and CO2 (23). For

119 example, Splivallo et al., (2014) observed the appearance of colonies belonging to

120 Firmicutes and Actinobacteria while the abundance of members of α- and β-

121 Proteobacteria decreased in fruiting bodies of T. borchii after six days of post-harvest

122 storage at room temperature (14).

123

124 The composition of bacterial communities associated to truffles is not only influenced by

125 the stage of maturity of the fruiting bodies but also by the stage of the life cycle of the

126 fungus. Comparative analysis of the bacterial communities associated to fruiting bodies

127 and ectomycorrhizae (EcM) of T. melanosporum showed striking differences suggesting

128 that the fungus could provide two different habitats to bacteria. For example,

6
129 Actinobacteria are dominant in EcM but rare in fruiting bodies of T. melanosporum (13).

130 Interestingly, an enrichment in several genera of Actinobacteria has also been

131 demonstrated for specific zones within orchards of T. melanosporum - referred to as

132 “brûlé”- (24) which are especially rich in truffle mycelium (25).

133

134 Overall, these observations demonstrate that truffles provide several habitats to complex

135 bacterial communities. α-Protebacteria, among which mainly members of the

136 Bradyrhiziaceae and Rhizobiaceae families, seem to form the core component of these

137 communities, whatever the truffle species considered. The parameters that control the

138 selection of this very specific community are still to be discovered. A tempting

139 hypothesis is that truffle fruiting bodies would be more than a habitat for bacteria and that

140 mutualistic interactions could occur between the fungi and their microbiota. Some

141 members of the Rhizobiales order are well known for their ability to fix atmospheric

142 nitrogen either as free living organisms or in symbiosis with plants (26). Barbieri et al.,

143 (2010) demonstrated that nitrogen fixation occurs inside fruiting bodies of the white

144 truffle T. magnatum (27). Nif genes encoding for the enzymes responsible for nitrogen

145 fixation were also detected in bacteria associated to T. melanosporum (13). Thus it is

146 tempting to speculate that part of the nitrogen captured by bacteria in fruiting bodies

147 could benefit the host fungus. However, it remains to be demonstrated that the nitrogen

148 fixed by bacteria inside truffle fruiting bodies is indeed transferred to the fungus.

149

150

151

7
152 Yeast communities

153 Besides bacteria, yeasts are ubiquitous organisms which occupy most terrestrial

154 ecological niches. Yeasts community composition has been investigated in fruiting

155 bodies (T. aestivum, T. melanosporum and T. magnatum), in ectomyocrrhizas and truffle

156 orchards soil (T. aestivum) (3, 6, 21). These studies were based on culture-dependant

157 methods and might hence miss the real diversity; nevertheless they also do provide useful

158 insights. In a study comparing yeasts distribution within an orchard of T. aestivum,

159 Zacchi et al., (2003) demonstrated that yeasts were enriched on truffle ectomycorrhizas

160 and fruiting bodies reaching up to 3x107 CFU/g dry weight compared to bulk soil (1x102

161 CFU/g dry weight). The total yeast diversity was made of five species, namely

162 Cryptococcus albidus, Cryptococcus humicolus, Rhodotorula mucilaginosa,

163 Debaryomyces hansenii and Saccharomyces paradoxus (3). Interestingly Cryptococcus

164 sp., Rhodotorula mucilaginosa, Debaryomyces hansenii and Saccharomyces spp were

165 also isolated from T. melanosporum, T. magnatum or T. aestivum by others (6, 21) and

166 might therefore be common to distinct truffle species. Yeast density might also vary

167 between peridium and gleba. Indeed, based on culture-dependant methods, yeasts could

168 only be isolated from the peridium of T. aestivum and T. melanosporum (103-104 CFU / g

169 (fresh weight)) but not from the gleba of intact truffles (21).

170 These observations suggest that, similarly to bacteria, yeast community composition

171 might vary with tissues and that a “core yeast community” might exist among truffle

172 species. Culture-independent techniques will nevertheless be necessary to confirm these

173 hypotheses and get a better view of the variability in space and time of the yeast

174 communities associated with truffles.

8
175 Truffle’s guest filamentous fungi and viruses

176 Besides yeasts and bacteria, truffles might also be colonized by filamentous fungi

177 (hereafter “guest filamentous fungi” as opposed to the “host” truffle mycelium) and

178 viruses. As in the case of yeasts, only a few reports exist on the occurrence of the latter

179 organisms in truffles. Guest filamentous fungi, mostly Ascomycetes, have been isolated

180 from T. rufum, T. brumale, T. magnatum, T. melanosporum, T. nitidum, T. excavatum,

181 T.aestivum, T. borchii and T. puberulum (7). However their occurrence in fruiting bodies

182 might be seldom since guest filamentous fungi could be isolated from only 26% of all

183 truffles (n=30), suggesting a loose association. The density of guest filamentous fungi

184 might vary between gleba and peridium. In T. melanosporum and T. aestivum,

185 correspondingly to what has been observed for yeasts, guest filamentous fungi

186 (Ascomycete molds) predominantly colonized the peridium with a density of 102 CFU/ g

187 (fresh weight) but seem absent from the gleba (21). Similarly a recent report described

188 the occurrence of viruses (Totivirus, Mitovirus and Endornavirus from T. aestivum and

189 Mitovirus from T. excavatum) without however addressing their occurrence frequency

190 within fruiting bodies or in orchards (8–11). Some authors have also suggested viral gene

191 integration in the genome of T. melanosporum (28). Surely guest filamentous fungi and

192 viruses might interact with truffles in nature; however additional ecological data is

193 needed at this stage to understand how frequent they might be and to assess how relevant

194 they are in the microbiome of truffles.

195

196

197

9
198 The involvement of microbes in human-sensed truffle aroma

199 Unique and delightful aromas are partially responsible for the high demand of truffles in

200 the world market. The particular aromas of truffles are made up of a mixture of various

201 volatiles namely alcohols, esters, ketones aldehydes, aromatic and sulfur compounds. To

202 date, the number of identified volatiles from various truffle species exceeds 200, however

203 only a small fraction of these, the so-called odorants, are responsible for what humans

204 perceive as truffle smell (16, 29, 30).

205 Historically the aroma of the white truffle T. magnatum was the first one characterized

206 and ascribed to a single sulfur compound (2,4-dithiapentane) (31). In the 80’s, a mixture

207 of two constituents, 2-methylbutanal and dimethyl sulfide, were patented to reproduce the

208 smell of the Périgord truffle T. melanosporum (32). Essentially due to increasingly

209 sensitive techniques in sensory science, the number of key odorants in T. melanosporum

210 was recently revised to more than 15 volatiles (29). A comparable number of odorants

211 (about 10 to 20) have also been described in four black truffle species (T. aestivum, T.

212 himalayense, T. indicum, T. sinense (29, 30) and in the white truffle T. borchii (16).

213 Interestingly most of these odorants are common to almost all truffle species (i.e.

214 methylthiomethane, 3-methyl-1-butanol, oct-1-en-3-ol) and only a few are species

215 specific or occur in a rather small number of species (i.e. thiophene derivatives, 2,4-

216 dithiapentane) (Figure 2). The exact origin of truffle volatiles and specifically of most

217 odorants reported in Figure 2 is unclear. It has been speculated that truffle aroma might

218 result from the intimate interaction of truffles and their microbiomes (6, 33, 34). Indeed

219 some volatiles might be produced by both truffles and microbes while others might be

220 derived from a single player (i.e. yeasts, bacteria, or truffle). Only recently has the role of

10
221 bacteria in the formation of thiophene derivatives, odorants unique to T. borchii, been

222 demonstrated. In the latter species, only bacteria but not truffle metabolize a precursor of

223 unknown origin into volatile thiophene derivatives (14). As the matter of fact, the

224 biosynthetic pathway leading to thiophene derivatives remains elusive (14) and this is

225 also the case for 2,4-dithiapentane, the major odorant of T. magnatum. By contrast, based

226 on the genome of T. melanosporum, pathways leading to odorants commonly produced

227 by yeast and bacteria most likely exist in truffles as well (34, 35). This is the case for

228 example for the Ehrlich pathway which consists in the catabolism of specific amino acids

229 and which results in dimethyl sulfide, 2-phenylethanol and 2- and 3-methylbutanol and

230 numerous other volatiles common to microbes and truffles (35). The Ehrlich pathway

231 consists of a three step process involving the initial transamination of an amino acid,

232 followed by a decarboxylation and a reduction step (36). Indeed enzymes fulfilling these

233 steps most likely exist in T. melanosporum (34, 35), their functions have nevertheless not

234 yet been demonstrated. At this stage, however, genomes provide limited insights on the

235 possible identity of the producer of specific odorants because either the pathways leading

236 to those odorants are highly conserved among yeasts, bacteria and truffles (i.e. the

237 Ehrlich pathway) or the biosynthetic pathways are not known (i.e. thiophene derivatives,

238 2,4-dithiapentane).

239 By combining knowledge about the structure of truffle microbiomes (Figure 1) with

240 literature data on the ability of specific microbes to produce odorants, we speculate here

241 on the origin of these volatiles in truffles. For this purpose, we first established a list of

242 all odorants described in truffles and reported in four publications (16, 29–31). We then

243 used the mVOC database (37) and the data from a review on fungal volatiles (38) to

11
244 understand which organisms had the ability to produce those volatiles, specifically

245 focusing on the phyla and classes which are reported in Figure 1. For the purpose of this

246 review, volatile occurrence is expressed for bacterial and fungal phyla/classes and

247 presented as a heatmap in Figure 2.

248

249 Sulfur containing volatiles

250 Sulfur containing volatiles (hereafter “sulfur volatiles”) represent the most important

251 group of odorants in truffles since they confer the typical garlicky and sulfurous notes

252 which characterize all truffle species (see the aroma descriptors in Figure 2). The most

253 common sulfur containing volatile in truffle fruiting bodies is dimethyl sufide which has

254 been detected from 85% of the species investigated to date (Figure 2). Along with

255 dimethyl disulfide, dimethyl trisulfide and 3-(methylthio)propanal, dimethyl sulfide

256 might be derived from the catabolism of methionine through the Ehrlich pathway (36, 39,

257 40). According to the mVOC database on microbial volatiles (37), the latter four volatiles

258 occur in the fungal phyla of Pezizomycetes (i.e. truffles) and Agaricomycetes as well as in

259 eight bacterial phyla (Figure 2). Since most of these volatiles are also produced by axenic

260 cultures of truffle mycelium (39, 41–43), they might be synthesized in truffle fruiting

261 bodies by both bacteria and truffle mycelium. Of special interest is dimethyl sulfide since

262 it might be produced by some α- and β-Proteobacteria (Figure 2) which are also

263 dominant in truffle fruiting bodies (Figure 1).

264 By contrast to the relatively common sulfur volatiles just described, other sulfur odorants

265 might be more specific (i.e. specific to a single or a limited number of species). Four

266 sulfur volatiles, namely 2-methyl-4,5-dihydrothiophene, 3-methyl-4,5-dihydrothiophene,

12
267 2,4-dithiapentane and 2-methylfuran-3-thiol occur in one or two truffle species only

268 (Figure 2). As for the common sulfur containing volatiles, they might be derived from

269 methionine; however this has not yet been appropriately demonstrated (i.e. through

270 feeding with labelled precursors). Interestingly none of these specific sulfur odorants

271 have been reported in axenic cultures of truffle mycelium, but some microbes have the

272 ability to produce them (Figure 2). Based on this observation, 2,4-dithiapentane might be

273 produced by β-Proteobacteria in T. magnatum fruiting bodies. Interestingly this might be

274 a similar case to the one of thiophene derivatives, which were recently shown to originate

275 from bacteria inhabiting T. borchii (14). 2-Methylfuran-3-thiol has been reported from

276 fruiting bodies of T. melanosporum and T. aestivum (29), but this volatile has been

277 detected from neither axenic mycelial cultures nor microbes (Figure 2). Its origin,

278 therefore, remains elusive; nevertheless it can be speculated that the latter odorant might

279 be specifically produced during the sexual stage of truffles.

280 Overall, this suggests that common sulfur volatiles might be produced inside truffle

281 fruiting bodies by both truffles and microbes (mixed origin) whereas more specific sulfur

282 volatiles might be derived from truffles or microbes only.

283

284 Alcohols, esters, ketones and aldehydes

285 Another important group of truffle odorants is made of alcohols, esters, ketones and

286 aldehydes that are possibly derived from amino acid and fatty acid catabolism (36). As

287 for sulfur volatiles, some commonly occur in numerous truffle species while others are

288 more specific (Figure 2). Axenic cultures of truffle mycelium as well as numerous fungal

289 and bacterial phyla are able to produce the most common volatiles (3-methylbutanal,

13
290 octan-3-one, oct-1-en-3-ol, 3-methyl-1-butanol, hexanal, acetaldehyde) which occur in

291 more than 50% of all species. Interestingly enough, eight carbon-containing volatiles (i.e.

292 octan-3-one, oct-1-en-3-ol) were believed to be strictly of fungal origin but Figure 2

293 suggests that similarly to 3-methylbutanal, 3-methyl-1-butanol, hexanal and acetaldehyde,

294 they might also be produced by specific bacterial classes. Eight carbon-containing

295 volatiles are important contributors to fungal aroma and have a characteristic mushroom

296 flavor (44).

297 The remaining less common alcohol, ketone, ester and aldehyde odorants found in truffle

298 fruiting bodies have not been detected from truffle mycelium and might be potentially

299 produced only by guest filamentous fungi, yeasts and bacteria. This is the case for

300 example of 3-hydroxybutan-2-one, which can potentially be produced by fungi of the

301 Sordariomycetes phyla or by the β-Proteobacteria phylum, which is a dominant group in

302 the truffle’s microbiome (Figure 1). Other rare volatiles might not be produced by

303 microbes or by axenic cultures of truffle mycelium. It can be hypothesized that they

304 might be specific to the sexual stage of truffle fruiting bodies.

305 Similarly to sulfur volatiles, the trend with alcohols, esters, ketones and aldehydes is that

306 common volatiles might be of mixed origins while more specific ones might be produced

307 either by microbes or by truffles.

308

309

310 Aromatic compounds

311 Aromatic odorants produced by truffles include for example the volatile 2-phenylethanol

312 with a characteristic rose smell and benzaldehyde, an odorant with a characteristic bitter

14
313 almond flavour (Figure 2). Aromatic odorants might be derived from the catabolism of

314 phenylalanine (36). Interestingly, with the exception of benzaldehyde and 2-

315 phenylethanol, none of these volatiles have been detected from truffle mycelium (Figure

316 2). These common aromatic odorants (which also include phenylacetaldehyde) are

317 potentially also synthesized by numerous fungal and or bacterial phyla and might

318 therefore be of mixed origins (Figure 2). The less common aromatic odorant 2-methoxy-

319 4-methylphenol is potentially produced by the two bacterial phyla Bacillii and γ-

320 Proteobacteria whereas other rare odorants might be derived from the sexual stage of

321 truffles.

322 Overall common aromatic odorants might either be of mixed fungal (truffle) or only

323 microbial origins. The absence or rare occurrence in microbes of specific aromatic

324 odorants suggests that they might be synthesized only by truffles and possibly only

325 during their sexual stage.

326

327

328 Other volatiles

329 Butanoic acid and 4-hydroxy-2,5-dimethylfuran-3-one are odorants specific to T.

330 melanosporum and T. aestivum (29). Based on Figure 2, they have not been detected

331 from truffle mycelia, and in the case of 4-hydroxy-2,5-dimethylfuran-3-one neither from

332 microbes, which suggests that the latter volatile might be synthesized only during the

333 sexual stage of truffles. Numerous microbes have the ability to produce butanoic acid

334 suggesting that it might be of microbial origin in truffle fruiting bodies (Figure 2).

335

15
336

337 Overall we are well aware that the absence of evidence is no evidence of absence. In

338 other words not having detected a volatile from a given organism does not demonstrate

339 that the organism in question is not able to produce it under specific circumstances. For

340 example this might be the case with truffles which might produce specific odorants only

341 during their sexual stage (fruiting body) but not as free living mycelium (axenic cultures).

342 Our approach, nevertheless, allows constructing hypothesis on the identity of the possible

343 producers of specific odorants. Demonstrating who produces what will not only require

344 fully characterizing pathways leading to specific odorants in truffles and microbes but

345 also obtaining microbe-free truffles. This is a specially challenging task considering that

346 to date all truffle fruiting bodies harvested from the wild contain microbes and that

347 microbe-free fruiting bodies cannot be obtained under axenic conditions.

348

349 Do truffles or actually microbes attract animals?

350 Truffles are hypogeous fungi, meaning that they form their fruiting bodies below the soil

351 surface. Since their belowground habitat prevents them from dispersing spores through

352 the air/wind, truffles have developed intense aromas to attract small rodents and larger

353 mammals. These animals eat fruiting bodies and subsequently disperse truffle spores

354 through their faeces. Mammals are not the only animals which are able to locate fruiting

355 bodies belowground; a beetle (Leiodes cinnamomea Panzer) and a fly (Suillia pallida)

356 can achieve the same. However it remains unclear whether these insects participate in

357 spore dispersal or whether they just feed on truffles (45, 46).

16
358 Mammals are able to locate truffles belowground due to the dimethyl sulfide emitted by

359 fruiting bodies (47). Dimethyl sulfide is obviously not the only volatile that animals can

360 smell, since for example dogs, similarly to humans, are able to distinguish among truffle

361 species. Nevertheless, besides dimethyl sulfide, species-specific attractants have not been

362 identified in truffles. Neither are the structures of the compounds that attract flies and

363 beetles to truffles known (45). The question of who actually produces these attractants

364 raises interesting hypotheses about multitrophic interactions. Indeed dimethyl sulfide

365 might be of mixed fungal (truffle) and bacterial origin since truffle mycelia as well as α-

366 Proteobacteria, which are dominant in fruiting bodies, are able to produce it. Assuming

367 that dimethyl sulfide is partially derived from bacteria would imply that bacteria

368 participate in attracting mammals and small rodents to truffles. A similar case has

369 actually been demonstrated for the fruit fly Drosophila melanogaster which is not

370 attracted by fruit volatiles but rather by microbial volatiles emitted by the yeasts which

371 colonize the fruit’s surface (48). Finding the answer to who produces truffle attractants

372 will require obtaining microbe free truffles, and this has not yet been achieved.

373

374

375 Conclusion

376 Understanding to which extent the microbiomes of truffles participate in truffle aroma

377 formation promises to be a complex and challenging task. Literature data on the ability of

378 organisms to produce volatiles suggests that truffle and microbes might be able to

379 produce common truffle odorants whereas more specific compounds might be of

380 microbial origin only. Disentangling who produces what within truffle fruiting bodies

17
381 will require elucidating the biosynthetic pathways for specific odorants and using

382 innovative techniques to follow the fate of aroma precursors in situ. Overall truffles offer

383 a unique opportunity to better understand the ecological function of microbes associated

384 with fungi and their involvement in aroma formation.

385

386 Acknowledgments

387 The authors are thankful to Dr. Mello (CNR Torino, Italy) and Dr. Gryndler (Academy of

388 Sciences of the Czech Republic) for making accessible the raw data of their manuscripts

389 published on truffle microbiomes. Prof. Piechulla and Mrs. Lemfack (University of

390 Rostock, Germany) are kindly acknowledged for providing raw data from the mVOC

391 database.

392 MV and RS are supported by the LOEWE research funding programme of the

393 government of Hessen, in the framework of the Integrative Fungal Research Cluster (IPF)

394 and by the German Research Foundation/ Deutsche Forschungsgemeinschaft (DFG) –

395 grant number 1191/4-1. AD is partially supported by a grant overseen by the French

396 National Research Agency (ANR) as part of the "Investissements d'Avenir" program

397 (ANR-11-LABX-0002-01, Lab of Excellence ARBRE).

398

399

400 Figures legends

401 Figure 1. Bacterial communities in fruiting bodies, ectomycorrhizas and

402 soil. The most abundant bacterial communities associated to four truffle species based

403 on culture independent methods. Bars represent the min. and max. values reported in

18
404 literature, whereas points display a single literature value (T. aestivum: (12), T. magnatum:

405 (5), T. borchii: (4, 14) and T. melanosporum: (24) and period from December to January

406 for (13). Cells were left empty where no literature data was available.

407

408 Figure 2. Ability of microbes to produce typical odorants of truffle fruiting

409 bodies. List of odorants and aroma descriptors from T. melanosporum and T. aestivum

410 (29), T. indicum, T. himalayense and T. sinense (30), T. borchii (16) and T.magnatum

411 (31). Occurrence in fungal and bacterial phyla/ classes is derived from the mVOC

412 database (37) and the data from a review on fungal volatiles (38). It is shown as a

413 heatmap representing percentage occurrence in each classes with n being the total number

414 of organisms in each classes (Ascomycetes: 1=Dothideomycetes (n=4),

415 2=Eurotiomycetes (n=29), 3=Pezizomycetes (n=26), 4=Saccharomycetes (n=4),

416 5=Sordariomycetes (n=47); Basidiomycetes: 6=Agaricomycetes (n=135),

417 7=Exobasidiomycetes (n=3), 8=Pucciniomycetes (n=4); Actinomycetes: 9=Actinobacteria

418 (n=62), Bacteroidetes: 10=Bacteroidetes (n=17), 11=Bacteroidia (n=24),

419 12=Flavobacteria and Sphingobacteriia (n=3); Firmicutes: 13=Bacilli (n=55),

420 14=Clostridia (n=10); Proteobacteria: 15=α-Proteobacteria (n=25), 16=β-

421 Proteobacteria (n=43), 17= δ-Proteobacteria (n=16), 18=γ-Proteobacteria (n=61).

422 Occurrence in axenic cultures of truffle (Tr. mycel.) is shown as presence/absence for T.

423 borchii (33, 39, 41), T. melanosporum (42), T. formosanum (43). Origin refers to the

424 speculative origin of the odorants in truffle fruiting bodies, where some odorants could be

425 produced by microbes only (Microbes), by truffle only at its sexual stage (Tr. sex. st.) or

426 by both microbes and truffle (Mixed). Frequency represents the percentage occurrence of

19
427 each odorants in fruiting bodies of 13 truffle species (T. aestivum, T. brumale, T.

428 himalayense, T. indicum, T. sinense, T. melanosporum, T. mesentericum, T. borchii, T.

429 excavatum, T. magnatum, T. oligospermum, T. panniferum, T. rufum (16, 29–31, 37))

430

431 References

432 1. Mayer EA, Knight R, Mazmanian SK, Cryan JF, Tillisch K. 2014. Gut microbes
433 and the brain: paradigm shift in neuroscience. J Neurosci 34:15490–15496.

434 2. Van der Heijden MGA, Schlaeppi K. 2015. Root surface as a frontier for plant
435 microbiome research. PNAS 112:2299–2300.

436 3. Zacchi L, Vaughan-Martini A, Angelini P. 2003. Yeast distribution in a truffle-

437 field ecosystem. Annals of Microbiology 53:275–282.

438 4. Barbieri E, Bertini L, Rossi I, Ceccaroli P, Saltarelli R, Guidi C, Zambonelli A,

439 Stocchi V. 2005. New evidence for bacterial diversity in the ascoma of the
440 ectomycorrhizal fungus Tuber borchii Vittad. FEMS Microbiology Letters 247:23–
441 35.

442 5. Barbieri E, Guidi C, Bertaux J, Frey-Klett P, Garbaye J, Ceccaroli P, Saltarelli

443 R, Zambonelli A, Stocchi V. 2007. Occurrence and diversity of bacterial
444 communities in Tuber magnatum during truffle maturation. Environmental
445 Microbiology 9:2234–2246.

446 6. Buzzini P, Gasparetti C, Turchetti B, Cramarossa MR, Vaughan-Martini A,

447 Martini A, Pagnoni UM, Forti L. 2005. Production of volatile organic compounds
448 (VOCs) by yeasts isolated from the ascocarps of black (Tuber melanosporum Vitt.)
449 and white (Tuber magnatum Pico) truffles. Archives of Microbiology 184:187–193.

450 7. Pacioni G, Leonardi M, Aimola P, Ragnelli AM, Rubini A, Paolocci F. 2007.

451 Isolation and characterization of some mycelia inhabiting Tuber ascomata.
452 Mycological Research 111:1450–1460.

453 8. Stielow B, Menzel W. 2010. Complete nucleotide sequence of TaV1, a novel

454 totivirus isolated from a black truffle ascocarp (Tuber aestivum Vittad.). Archives of
455 Virology 155:2075–2078.

456 9. Stielow B, Klenk H-P, Menzel W. 2011. Complete genome sequence of the first
457 endornavirus from the ascocarp of the ectomycorrhizal fungus Tuber aestivum
458 Vittad. Archives of Virology 156:343–345.

20
459 10. Stielow B, Klenk H-P, Winter S, Menzel W. 2011. A novel Tuber aestivum
460 (Vittad.) mitovirus. Archives of Virology 156:1107–1110.

461 11. Stielow JB, Bratek Z, Klenk H-P, Winter S, Menzel W. 2012. A novel mitovirus
462 from the hypogeous ectomycorrhizal fungus Tuber excavatum. Archives of
463 Virology 157:787–790.

464 12. Gryndler M, Soukupová L, Hršelová H, Gryndlerová H, Borovička J,

465 Streiblová E, Jansa J. 2013. A quest for indigenous truffle helper prokaryotes:
466 Tuber aestivum-associative prokaryotes. Environmental Microbiology Reports
467 5:346–352.

468 13. Antony-Babu S, Deveau A, Van Nostrand JD, Zhou J, Le Tacon F, Robin C,
469 Frey-Klett P, Uroz S. 2014. Black truffle - associated bacterial communities during
470 the development and maturation of Tuber melanosporum ascocarps and putative
471 functional roles. Environmental Microbiology 16:2831–2847.

472 14. Splivallo R, Deveau A, Valdez N, Kirchhoff N, Frey-Klett P, Karlovsky P. 2014.

473 Bacteria associated with truffle-fruiting bodies contribute to truffle aroma.
474 Environmental Microbiology DOI: 10.1111/1462–2920.12521.

475 15. Le Tacon F, Rubini A, Murat C, Riccioni C, Robin C, Belfiori B, Zeller B, De

476 la Varga H, Akroume E, Deveau A, Martin F, Paolocci F. 2015. Certainties and
477 uncertainties about the life cycle of the Périgord black truffle (Tuber melanosporum
478 Vittad.). Annals of Forest Science DOI: 10.1007/s13595–015–0461–1.

479 16. Splivallo R, Ebeler SE. 2015. Sulfur volatiles of microbial origin are key
480 contributors to human-sensed truffle aroma. Applied Microbiology and
481 Biotechnology 99:2583–2592.

482 17. Citterio B, Cardoni P, Potenza L, Amicucci A, Stocchi V, Gola G, Nuti M.

483 1995. Isolation of Bacteria from Sporocarps of Tuber Magnatum Pico, Tuber
484 Borchii Vitt. and Tuber Maculatum Vitt., p. 241–248. In Biotechnology of
485 ectomycorrhizae. Springer.

486 18. Barbieri E, Potenza L, Rossi I, Sisti D, Giomaro G, Rossetti S, Beimfohr C,

487 Stocchi V. 2000. Phylogenetic characterization and in situ detection of a
488 Cytophaga-Flexibacter-Bacteroides phylogroup bacterium in Tuber borchii Vittad.
489 ectomycorrhizal mycelium. Applied and Environmental Microbiology 66:5035–
490 5042.

491 19. Sbrana C, Agnolucci M, Bedini S, Lepera A, Toffanin A, Giovannetti M, Nuti

492 MP. 2002. Diversity of culturable bacterial populations associated to Tuber borchii
493 ectomycorrhizas and their activity on T. borchii mycelial growth. FEMS
494 Microbiology Letters 211:195–201.

21
495 20. Nazzaro F, Fratianni F, Picariello G, Coppola R, Reale A, Luccia AD. 2007.
496 Evaluation of gamma rays influence on some biochemical and microbiological
497 aspects in black truffles. Food Chemistry 103:344–354.

498 21. Rivera CS, Blanco D, Oria R, Venturini ME. 2010. Diversity of culturable
499 microorganisms and occurrence of Listeria monocytogenes and Salmonella spp. in
500 Tuber aestivum and Tuber melanosporum ascocarps. Food Microbiology 27:286–
501 293.

502 22. Saltarelli R, Ceccaroli P, Cesari P, Barbieri E, Stocchi V. 2008. Effect of storage
503 on biochemical and microbiological parameters of edible truffle species. Food
504 Chemistry 109:8–16.

505 23. Rivera CS, Blanco D, Salvador ML, Venturini ME. 2010. Shelf-life extension of
506 fresh Tuber aestivum and Tuber melanosporum truffles by modified atmosphere
507 packaging with microperforated films. Journal of Food Science 75:E225–E233.

508 24. Mello A, Ding G-C, Piceno YM, Napoli C, Tom LM, DeSantis TZ, Andersen
509 GL, Smalla K, Bonfante P. 2013. Truffle brûlés have an impact on the diversity of
510 soil bacterial communities. PLoS ONE 8:e61945.

511 25. Suz LM, MartÃ-n MP, Oliach D, Fischer CR, Colinas C. 2008. Mycelial
512 abundance and other factors related to truffle productivity in Tuber melanosporum –
513 Quercus ilex orchards. FEMS Microbiology Letters 285:72–78.

514 26. Carvalho FM, Souza RC, Barcellos FG, Hungria M, Vasconcelos AT. 2010.
515 Genomic and evolutionary comparisons of diazotrophic and pathogenic bacteria of
516 the order Rhizobiales. BMC Microbiology 10:37.

517 27. Barbieri E, Ceccaroli P, Saltarelli R, Guidi C, Potenza L, Basaglia M, Fontana

518 F, Baldan E, Casella S, Ryahi O, Zambonelli A, Stocchi V. 2010. New evidence
519 for nitrogen fixation within the Italian white truffle Tuber magnatum. Fungal Biol
520 114:936–942.

521 28. Liu H, Fu Y, Li B, Yu X, Xie J, Cheng J, Ghabrial SA, Li G, Yi X, Jiang D.

522 2011. Widespread horizontal gene transfer from circular single-stranded DNA
523 viruses to eukaryotic genomes. BMC Evolutionary Biology 11:276.

524 29. Culleré L, Ferreira V, Chevret B, Venturini ME, Sánchez-Gimeno AC, Blanco
525 D. 2010. Characterisation of aroma active compounds in black truffles (Tuber
526 melanosporum) and summer truffles (Tuber aestivum) by gas chromatography–
527 olfactometry. Food Chemistry 122:300–306.

528 30. Liu R-S, Li D-C, Li H-M, Tang Y-J. 2012. Evaluation of aroma active compounds
529 in Tuber fruiting bodies by gas chromatography–olfactometry in combination with
530 aroma reconstitution and omission test. Applied Microbiology and Biotechnology
531 94:353–363.

22
532 31. Fiecchi A, Kienle M, Scala A, Cabella P. 1967. Bis-methylthiomethane, an
533 odorous substance from white truffle, Tuber magnatum Pico. Tetrahedron Letters
534 8:1681–1682.

535 32. Talou T, Delmas M, Gaset A. 1989. Black Périgord truffle: from aroma analysis to
536 aromatizer formulation, p 715–728. Flavors and Off- Flavors, Proceedings of the 6th
537 International Flavor Conference Rethymnon, Crete, Greece.

538 33. Splivallo R, Bossi S, Maffei M, Bonfante P. 2007. Discrimination of truffle

539 fruiting body versus mycelial aromas by stir bar sorptive extraction. Phytochemistry
540 68:2584–2598.

541 34. Splivallo R, Ottonello S, Mello A, Karlovsky P. 2011. Truffle volatiles: from
542 chemical ecology to aroma biosynthesis. New Phytologist 189:688–699.

543 35. Martin F, Kohler A, Murat C, Balestrini R, Coutinho PM, Jaillon O,

544 Montanini B, Morin E, Noel B, Percudani R, Porcel B, Rubini A, Amicucci A,
545 Amselem J, Anthouard V, Arcioni S, Artiguenave F, Aury J-M, Ballario P,
546 Bolchi A, Brenna A, Brun A, Buée M, Cantarel B, Chevalier G, Couloux A, Da
547 Silva C, Denoeud F, Duplessis S, Ghignone S, Hilselberger B, Iotti M, Marçais
548 B, Mello A, Miranda M, Pacioni G, Quesneville H, Riccioni C, Ruotolo R,
549 Splivallo R, Stocchi V, Tisserant E, Viscomi AR, Zambonelli A, Zampieri E,
550 Henrissat B, Lebrun M-H, Paolocci F, Bonfante P, Ottonello S, Wincker P.
551 2010. Périgord black truffle genome uncovers evolutionary origins and mechanisms
552 of symbiosis. Nature 464:1033–1038.

553 36. Hazelwood LA, Daran J-M, Maris AJA van, Pronk JT, Dickinson JR. 2008.
554 The Ehrlich Pathway for Fusel Alcohol Production: a Century of Research on
555 Saccharomyces cerevisiae Metabolism. Appl Environ Microbiol 74:2259–2266.

556 37. Lemfack MC, Nickel J, Dunkel M, Preissner R, Piechulla B. 2014. mVOC: a
557 database of microbial volatiles. Nucleic Acids Res 42:D744–748.

558 38. Chiron N, Michelot D. 2005. Odeurs des champignons : chimie et rôle dans les
559 interactions biotiques - une revue. Cryptogamie, Mycologie 26:299–364.

560 39. Splivallo R, Maier C. 2011. Production of natural truffle flavours from truffle
561 mycelium. Patent Publication Number WO/2011/001349.

562 40. Liu R-S, Zhou H, Li H-M, Yuan Z-P, Chen T, Tang Y-J. 2013. Metabolism of L-
563 methionine linked to the biosynthesis of volatile organic sulfur-containing
564 compounds during the submerged fermentation of Tuber melanosporum. Appl
565 Microbiol Biotechnol 97:9981–9992.

566 41. Tirillini B, Verdelli G, Paolocci F, Ciccioli P, Frattoni M. 2000. The volatile
567 organic compounds from the mycelium of Tuber borchii Vitt. Phytochemistry
568 55:983–985.

23
569 42. Li Y-Y, Wang G, Li H-M, Zhong J-J, Tang Y-J. 2012. Volatile organic
570 compounds from a Tuber melanosporum fermentation system. Food Chemistry
571 135:2628–2637.

572 43. Du M, Huang S, Wang J. 2014. The Volatiles from Fermentation Product of Tuber
573 formosanum. Open Journal of Forestry 04:426–429.

574 44. Combet E, Eastwood DC, Burton KS, Combet E, Henderson J, Henderson J,
575 Combet E. 2006. Eight-carbon volatiles in mushrooms and fungi: properties,
576 analysis, and biosynthesis. Mycoscience 47:317–326.

577 45. Hochberg ME, Bertault G, Poitrineau K, Janssen A. 2003. Olfactory orientation
578 of the truffle beetle, Leiodes cinnamomea. Entomologia Experimentalis et Applicata
579 109:147–153.

580 46. Maser C, Claridge AW, Trappe JM. 2008. Trees, truffles, and beasts: how forests
581 function. Rutgers University Press.

582 47. Talou T, Gaset A, Delmas M, Kulifaj M, Montant C. 1990. Dimethyl sulphide:
583 the secret for black truffle hunting by animals? Mycological Research 94:277–278.

584 48. Becher PG, Flick G, Rozpędowska E, Schmidt A, Hagman A, Lebreton S,

585 Larsson MC, Hansson BS, Piškur J, Witzgall P, Bengtsson M. 2012. Yeast, not
586 fruit volatiles mediate Drosophila melanogaster attraction, oviposition and
587 development. Functional Ecology 26:822–828.

588

24