Evolution and Ecophysiology of the Industrial Producer Hypocrea jecorina (Anamorph Trichoderma reesei) and a New Sympatric Agamospecies Related to It

Background Trichoderma reesei, a mitosporic green mould, was recognized during the WW II based on a single isolate from the Solomon Islands and since then used in industry for production of cellulases. It is believed to be an anamorph (asexual stage) of the common pantropical ascomycete Hypocrea jecorina. Methodology/Principal Findings We combined molecular evolutionary analysis and multiple methods of phenotype profiling in order to reveal the genetic relationship of T. reesei to H. jecorina. The resulting data show that the isolates which were previously identified as H. jecorina by means of morphophysiology and ITS1 and 2 (rRNA gene cluster) barcode in fact comprise several species: i) H. jecorina/T. reesei sensu stricto which contains most of the teleomorphs (sexual stages) found on dead wood and the wild-type strain of T. reesei QM 6a; ii) T. parareesei nom. prov., which contains all strains isolated as anamorphs from soil; iii) and two other hypothetical new species for which only one or two isolates are available. In silico tests for recombination and in vitro mating experiments revealed a history of sexual reproduction for H. jecorina and confirmed clonality for T. parareesei nom. prov. Isolates of both species were consistently found worldwide in pantropical climatic zone. Ecophysiological comparison of H. jecorina and T. parareesei nom. prov. revealed striking differences in carbon source utilization, conidiation intensity, photosensitivity and mycoparasitism, thus suggesting adaptation to different ecological niches with the high opportunistic potential for T. parareesei nom. prov. Conclusions Our data prove that T. reesei belongs to a holomorph H. jecorina and displays a history of worldwide gene flow. We also show that its nearest genetic neighbour - T. parareesei nom. prov., is a cryptic phylogenetic agamospecies which inhabits the same biogeographic zone. These two species thus provide a so far rare example of sympatric speciation within saprotrophic fungi, with divergent ecophysiological adaptations and reproductive strategies.


Introduction
The holomorphic fungal genus Hypocrea/Trichoderma (Hypocreales, Ascomycota) contains several hundred species comprising both sexually propagating (Hypocrea) as well as apparently mitosporic (Trichoderma) taxa without observed ability to sexual reproduction. One species of Trichoderma -T. reesei -is particularly well known because it is used in the biotechnological industry for the production of cellulolytic and hemicellulolytic enzymes and recombinant proteins [1,2]. This property, and strategies for its further improvement, have recently regained strong interest because of the attempts to produce second generation biofuels to combat carbon dioxide emission and dependence on fossil oil [3,4].
T. reesei was originally isolated as in the Solomon Islands during WW II, where it destroyed canvas and other cellulose-containing material of the US army [5]. This species is unique among industrial fungi, as it was known only from this single wild-type isolate ( = QM 6a) for last 50 years, and all mutant strains used in biotechnology today have thus been derived from it. Kuhls et al. [6], however, used molecular characters (internal transcribed spacers 1 sequences of the rRNA gene cluster and RAPD fingerprinting) to conclude that T. reesei was indistinguishable from the pantropical ascomycete Hypocrea jecorina, and suggested that the latter species is its teleomorph. Yet, although H. jecorina is heterothallic [7], the authors were unable to mate T. reesei with wild type isolates of H. jecorina in vitro. Further, Kuhls et al. [6] noted small phenotypic differences between T. reesei QM 6a and wild-type strains of H. jecorina. Consequently, the authors suggested that T. reesei is a clonal derivate from H. jecorina but this argument was not considered in the taxonomic status of the first fungus. Seidl et al. [8] recently described that T. reesei can indeed be crossed with wild-type isolates of H. jecorina, thus rejecting at least one of the above arguments for the clonality of T. reesei.
The progress in molecular evolutionary tools and theoretical concepts for species recognition has recently led to the detection of a continuously increasing number of cryptic species within individual morphological or biological fungal taxa [9][10][11][12][13] including Hypocrea/Trichoderma [14][15][16]. Recent world-wide sampling has detected a putative anamorph of H. jecorina as being common in soils of South East Asia and particularly in South America [17][18][19]. The frequent occurence of these strains in soil led us to hypothesize that they may eventually be co-specific with T. reesei, and the latter in fact is not being an anamorph of but a cryptic species to H. jecorina. The attribution of T. reesei to holomorph of H. jecorina has so far been only claimed from identical ITS 1 and 2 sequences, but recent studies have shown that sequences from this locus are unable to distinguish between several closely related Hypocrea/Trichoderma species [15][16][17].
Here we will show that T. reesei sensu stricto ( = QM 6a) indeed belongs to the holomorphic species H. jecorina, and exhibits a history of recombination and world-wide gene flow, thus rejecting the hypothesis of being a genetically separated agamospecies derived from H. jecorina. In addition, we will also show that strains recently isolated from soils as putative anamorphs of H. jecorina in fact form two sibling species and that at least one of them (T. parareesei nom. prov.) reproduces asexually, and is the result of sympatric speciation in parallel to H. jecorina. Ecophysiological analyses reveal that this sympatric speciation is due to the adaptation to two different habitats.

Sample Design and Genetic Markers
Our sample consisted of 34 strains from teleomorphs and anamorphs (Table 1, already sorted in relation to the results from phylogenetic analysis given below), which were originally identified as H. jecorina or T. reesei both by morphological analysis as well as ITS1 and 2 based oligonucleotide barcode (TrichOKey [17]; online at www.ISTH.info). They covered the whole geographic variability known for these species (South and Central America, Caribbean archipelago, Africa, South Pacific, South East Asia, Macro-and Micronesia and the Indian subcontinent).
A preliminary screening for phylogenetic markers, which were previously used with success in other studies on Hypocrea/ Trichoderma (such as tef1, cal1, chi18-5, rpb2 or act1 [15][16][17]20]), showed that only the 4th intron of tef1 and the 2 nd and 3 rd introns of cal1 provided sufficient phylogenetic information. Coding regions such as chi18-5 or act1 provided insufficient polymorphism (data not shown). We therefore searched for additional genes with long introns and tested their ability to differentiate within H. jecorina. One locus fulfilling this requirement is the las1 gene (fgenesh5_pg.C_scaffold_1000016), which encodes the orthologue of an essential nuclear protein regulating bud formation and morphogenesis in Saccharomyces cerevisiae [21]. It is interrupted by four introns, of which the second (307 nts) was selected as the phylogenetic marker (Fig. 1).
Sequencing of tef1, cal1 and las1 for the whole strain sample provided 1241 nts, of which 1098 were constant sites and 109 polymorphic characters were parsimony-informative (49 in tef1, 38 in las1 and 22 in cal1, respectively). Nucleotide characteristics of the three genes are shown in Table 2.

Molecular Phylogeny of H. jecorina Sensu Lato
We used Bayesian phylogenetic analysis of both the concatenated as well as individual gene data sets to test for a phylogenetic structure within H. jecorina sensu lato ( Fig. 2 and Fig. S1). One single teleomorph isolate from North Sulawesi, Indonesia (C.P.K. 1281 = G.J.S. 85-238 -CBS 638.92), which was described as H. jecorina by Kuhls et al. [6] and Samuels et al. [22], consistently formed a single branch separated from all other clades in the tef1, las1 and cal1 trees (data not shown) indicating that it represents a still undescribed species outside the H. jecorina sensu lato clade. Thus, we named it as H. sp. nov. G.J.S. 85-238 and excluded it from further analyses. The remaining 33 isolates formed three significantly supported clades in all single gene trees as well as in the combined phylogram, thus fulfilling the criteria of the genealogical concordance phylogenetic species recognition concept [23,24] which recognizes a clade as an evolutionary lineage if its separation is supported by at least two gene trees and not contradicted by the others. One of the obtained clades contained all the teleomorph isolates and also the original isolate of T. reesei. As the type strain of H. jecorina (specimen No. 989 of the Berkeley and Broom collection from Ceylon at Kew, UK) is not available for molecular analysis, we rely on the attribution of all these strains to the above taxon by Samuels et al. [22] and recognize this clade as a holomorph H. jecorina/T. reesei sensu stricto.
The next large clade contained all strains that were isolated from soil as anamorphs (Fig. 2). Since T. reesei was not clustered in this clade we call it T. sp. nov. 'parareesei nom. prov.' (for simplicity T. parareesei nom. prov.) throughout the manuscript, assuming that its formal taxonomic description will soon be published elsewhere. Two isolates (C.P.K. 523 and C.P.K. 524), collected from tree bark in Taiwan formed a separate branch in all of these trees and are therefore recognized as another phylogenetic species T. sp. nov. C.P.K. 524, whose formal description will be possible when additional isolates become available.
Geographic Distribution of H. jecorina and T. parareesei Nom. Prov.
Our sample shows that both H. jecorina and T. parareesei nom. prov. are cosmopolitan species (Fig. 2), yet restricted to a narrow latitudinal belt around the equator (620u). They must therefore be considered to be sympatric species, particularly in Central and South America. In order to test, whether the gene sequences would reveal some intraspecific geographic separation within H. jecorina, we determined the F ST values for pairwise combinations of strains from different locations (e.g. Caribbean vs. Indopacific, South American vs. Africa etc.). However, the F ST values were within 0.011-0.024 for all combinations (data not shown), thus documenting a high rate of exchange of genetic material over these wide geographic distances, and no evidence for geographic segregation.
In vitro Mating between and within H. jecorina and T. parareesei Nom. Prov.
The mating type loci of the heterothallic species H. jecorina have recently been identified and conditions for successful mating have been established [8]. We thus tested whether the two newly recognized species would contain one of the two mating types of H. jecorina. Using primers within conserved regions of the genes mat1-1-1, mat1-1-2, mat1-1-3 (indicative for a MAT1-1 mating type locus, Fig. 3A) and mat1-2-1 (indicative for a MAT1-2 locus, Fig. 3A), we in fact identified that all isolates of T. parareesei nom. prov. have a MAT1-1 locus except C.P.K. 661 from northern Argentina, which has a MAT1-2 locus (Table 1). Both isolates of T. sp. C.P.K. 524 also possess a MAT1-1 locus. RFLP of the complete mating type loci in comparison with H. jecorina confirmed the segregation of all three species in three and two different haplotypes of MAT1-1 and MAT1-2 respectively (Fig. 3B). Although most of the SNPs of the mating type loci of T. parareesei nom. prov. and T. sp. C.P.K. 524 were silent mutations, only the amino acid sequence of MAT1-1-2 of T. parareesei nom. prov. was identical to that from H. jecorina, whereas it was polymorphic in three positions in T. sp. C.P.K. 524. The amino acid sequence of the MAT1-1-3 protein of T. parareesei nom. prov. differed in two residues from that of H. jecorina including one case where the helix breaker P was exchanged by a polar S residue. The most significant difference was observed in MAT1-2-1, which contained a PS insertion (Fig. S2). Thus these data demonstrate that the mating type proteins from T. parareesei nom. prov. have undergone alterations which may have affected their functionality. The availability of two mating types for T. parareesei nom. prov. provided us with the possibility to test whether these isolates already developed an infertility barrier to H. jecorina, and if they would still be able to mate with each other. Therefore, selected isolates of H. jecorina and T. parareesei nom. prov. carrying the MAT1-1 and MAT1-2 locus respectively, were subjected to pairwise crossing experiments on plates under daylight conditions (indicated by arrows on Fig. 4). Only the heterothallic strain pairs belonging to H. jecorina indeed produced either mature fruiting bodies or primordia within 6-12 weeks of incubation, whereas all other combinations did not mate, even after further prolonged incubation time. These data are consistent with the conclusion that -at least in vitro -T. parareesei nom. prov. is unable to cross with H. jecorina, and is also not capable of sexual reproduction with the opposite mating partner of itself.
In silico Analysis of Reproduction Strategies of H. jecorina and T. parareesei Nom. Prov.
The results of mating experiments, the visual inspection of tree topologies ( Fig. 2 and Fig. S1) and eventually the origin of strains (as teleomorph or anamorph, respectively) lead to the assumption that only H. jecorina performs sexual reproduction, whereas T. parareesei nom. prov. would be a clonal taxon ( = agamospecies). In order to investigate this by means of sequence analysis we first used the split decomposition method [25,26] to test for the presence of network relationships in H. jecorina and T. parareesei nom. prov., using a concatenated dataset of tef1, cal1 and las1. This method presents conflicting phylogenetic data, presumably arising from recombination, as an interconnected network of lineages. As shown in Fig. 4, such a network was evident within H. jecorina ( = the strains isolated as teleomorphs and T. reesei) whereas it was absent within T. parareesei nom. prov. We especially note that the ex-type strain of T. reesei was tightly linked to the H. jecorina network, thus arguing against its origin as a clonally separated isolate.  We then used the partition homogeneity test (PHT; [27,28]) to examine the congruence between individual gene trees. This test produces artificial datasets by multiple (10 000) re-sampling and random swapping of observed datasets and subsequent construction of maximum-parsimony trees for every newly sampled 'gene' sequence. For clonally reproducing populations ( = no sexual recombination), the sums of the lengths of the gene trees for the observed and re-sampled data should be similar. However, under recombination the sums of the tree lengths should be longer than those for the actual data because of introduction of homoplasy into unlinked sites. This test confirmed our analysis of topologies of single locus trees -the clades containing T. parareesei nom. prov. showed congruence of all three loci suggesting the clonality of this species (Fig. 4). The topologies of H. jecorina subclades appeared to be not concordant between individual trees, providing the evidence for recombination (Fig. 4).
As another means, we used the index of association (IA) test on a subset of 'clone corrected' data (i.e. individuals with identical alleles of the three loci were excluded so that each haplotype was represented only once; cf. [29]). In this test, complete panmixia (sexual compatibility resulting in recombination) would be indicated by a value of 0 ( = the null hypothesis). This value was neither obtained with the complete dataset nor with any of the individual clades (data not shown).
Finally we applied the Phi-test, which uses the pairwise homoplasy index (PHI, W) to detect refined incompatibility [30]. This method assumes the infinite sites model of evolution in which the detection of incompatibility for a pair of sites indicates recombination. Application of this test to H. jecorina did find statistically significant evidence for recombination (P%0.001) while it was not the case for T. parareesei nom. prov. (P = 1). Since the Phi-test is a very robust means which can detect recombination even in the presence of recurrent mutation, we decided to use this method to define the borders of recombining populations. To this end, we first detected a Bayesian circular phylogram inferred from the concatenated dataset of tef1, cal1 and las1 phylogenetic markers. Symbols at nodes correspond to posterior probabilities (PP) .95%. Filled circles correspond to PP in the concatenated tree, open stars, squares and polygons to PP in las1, cal1 and tef1 gene trees, respectively. The corresponding phylograms are given in Figure S1. The color code indicates the geographic region from which the isolates were obtained, as explained in the right top inset. doi:10.1371/journal.pone.0009191.g002 non-recombining subsample consisting of six most terminal strains of H. jecorina (C.P.K. 160 and 1282 from Indonesia, C.P.K. 155 and 938 from Brazil, C.P.K. 3419 from Cameroon and C.P.K. 1273 from Pacific; P = 0.34), and then gradually added strains from two other phylogenetic species until evidence for recombination was detected (P,0.05). No recombination was detected between T. parareesei nom. prov. and H. jecorina. In contrast, a positive recombination signal (P = 0.012) was obtained when both strains of T. sp. C.P.K. 524 were tested together with the above listed subset of H. jecorina. It suggests that T. sp. C.P.K. 524 may also reproduce sexually but the current sample is too small to reveal it.
Thus the four alternative tests suggest that T. parareesei nom. prov. is an agamospecies which did not undergo sexual recombination in its recent evolutionary history.
The fact that H. jecorina and T. parareesei nom. prov. evolved in sympatry raises the question about the differences in their ecological niches. Since both fungi are saprotrophs, we tested whether they would differ in their carbon metabolism, response to environmental stimuli or antagonistic ability against other fungi.  [8]. Primers used to amplify the complete MAT-loci are indicated by gray arrows and primers for fragments of the mating type genes (Table 4) by black arrows. Numbers correspond to the respective proteins IDs in the T. reeesei genome database. (B) Restriction fragment patterns of the mating type loci, amplified with primers aF and aR (Table 4)  Carbon utilization. With respect to carbon metabolism, H. jecorina and the two new species exhibited qualitatively very similar carbon source utilization profiles. The full scale profile will be published elsewhere alone with the formal description of the new species; the list of carbon sources is in Table S1 and in Druzhinina et al. [31]. On a quantitative basis, T. parareesei nom. prov. and T. sp. C.P.K. 524 generally exhibited faster growth rates on so-called ''Cluster I'' carbon sources, i.e. the ones which provide fastest growth (such the chitin monomer n-acetyl-b-d-glucosamine or the hemicellulose monomers l-arabinose, d-xylose, d-galactose and corresponding polyols; cf. [31]), and also displayed a broader intraspecific variation (Fig. 5A). In contrast, the majority of H. jecorina strains showed a much more conserved quantitative pattern of growth on these carbon sources, yet with a significantly lower growth rate (except C.P.K. 938 and C.P.K. 160). Thus, we conclude that T. parareesei nom. prov. has become more versatile and efficient in the utilization of its preferred carbon sources.
Consistent results were obtained when the experiment was repeated with carbon sources typical for soil (EcoPlate TM ; Biolog Inc., Hayward, CA, USA, see Table S1 for individual carbon sources). Although the growth of both species on EcoPlates was slower than in the experiment above, the results confirm that in general T. parareesei nom. prov. is more competent in utilizing these carbon sources (ANOVA, F(1, 133) = 38.32, P = 0.000).
Since the only known anamorphic strain of H. jecorina -T. reesei QM 6a is a model fungus for cellulase formation, we tested whether the superiority of T. parareesei nom. prov. in carbon assimilation would also be reflected in this trait. We therefore tested 10 randomly chosen strains of H. jecorina and five strains of T. parareesei nom. prov., and included the cellulase twofold-  overproducing and the cellulase negative mutants QM 9414 and QM 9978 respectively (Fig. 5B). The results show that both species contain extremely efficient cellulase producing strains which gave values close to that of QM 9414. For H. jecorina these were two strains isolated from Brazil (C.P.K. 938 and C.P.K. 155); for T. parareesei nom. prov. these are also South American C.P.K. 661 and C.P.K. 665 isolated from northern Argentina. Interestingly, three strains of H. jecorina showed no cellulase activity, while all strains of T. parareesei nom. prov. were efficient producers. Analysis of variance, however, did not detect any statistically significant difference between two species (ANOVA, P.0.05). This suggests that the ability to produce cellulases for the degradation of cellulose in the environment has been maintained in both species.
Photosensitivity. Light sensing is an important mechanism in the ecophysiological adaptation of fungi as it is strongly involved in regulation of their reproduction. Seidl et al. [6] reported that light is important for formation of perithecia of H. jecorina. Friedl et al. [32] demonstrated that light plays the role in the conidiation of H. atroviridis. Our recent studies have shown that light influences the mycelial growth of Hypocrea but the effect varies depending on the species: temperate H. atroviridis is strongly stimulated by illumination [33] while mutant strains derived from tropical T. reesei are frequently photoinhibited [34]. Thus, photosensitivity may reflect the ecological niche of the fungus.
Because of the differences in carbon assimilation, we reasoned that the response to light may also be different in H. jecorina and T. parareesei nom. prov. We thus incubated Biolog FF MicroPlates either in complete darkness or under the conditions of day light (May-June, N 48u) for 72 hours and compared the data (Fig. 6). Indeed, all strains of H. jecorina show their best growth rates in darkness, which is basically not changed by exposure to light. We have detected about 10 cases of photoinhibition (on individual carbon sources, see Fig. 6), which were strain-but not speciesspecific. Growth of only one strain, C.P.K. 160, was considerably stimulated by light on several carbon sources (for example, dcellobiose, d-mannitol, gentiobiose, d-trehalose, maltotriose and lfucose). In contrast, the best growth of the majority of T. parareesei nom. prov. (all possessing the MAT1-1) was under illumination indicating that this species is not capable of normal growth in darkness (Fig. 6). No cases of photoinhibition were detected for T. parareesei nom. prov. These data show a striking difference in physiological adaptation of two species and confirm that although they are sympatric they have different ecological niches. The detection of photostimulation in T. parareesei nom. prov. is particularly interesting as all its strains were isolated from soil.
Conidiation. H. jecorina has a complete holomorphic life cycle and forms both conidia and ascospores, while propagation of T. parareesei nom. prov. is dependent on the distribution of its mitospores. A visual inspection of cultures indicated that T. parareesei nom. prov. conidiates essentially more intensively than the mycelial forms of H. jecorina which leads to the striking difference in culture morphology (data not shown). Since conidiation is carbon source dependent [33], we tested it on 95 carbon sources. The results show that in total H. jecorina conidiates only on 7 out of 95 carbon sources (6 of those shown on Fig. 6, see the corresponding legend) while T. parareesei nom. prov. forms mitospores on 62 carbon sources (36 of those shown on Fig. 6). A quantification of conidial density per cm 2 of a MEA plate showed that on an average H. jecorina formed 3.6 (61.78)610 6 per 10 tested strains, while the six available strains of T. parareesei nom. prov. produced an average of 16.05 (64.8)610 6 conidia under the same conditions.
Antagonistic potential against soil competent and epigeal plant pathogens. Species of Hypocrea/Trichoderma are renowned for their mycoparasitic behaviour [1]. In order to test whether the two species may also differ in their antagonistic abilities, we have selected five plant pathogenic fungi which differ in their primary habitat: Sclerotinia sclerotiorum and Fusarium oxysporum (FOX, F. oxysporum species complex) representing soil competent mycobionts; and F. xylarioides, Alternaria alternata and Botrytis cinerea representing pathogens of green plant tissue and thus predominantly epigeal fungi. Fig. 7 shows the summary of dual confrontation tests assessed after 10 days of co-cultivation on MEA medium. H. jecorina is able to considerably inhibit growth of S. sclerotiorum, A. alternata and B. cinerea. The latter plant pathogen was so completely combated by H. jecorina C.P.K. 160 and C.P.K. 1127 that its mycelium was overgrown and killed. No antagonism to both Fusarium species by H. jecorina was detected. Strains of T. parareesei nom. prov. showed a conserved pattern of mycoparasitic activity with almost no variation among strains, and they also exhibited a superior antagonistic potential against S. sclerotiorum, B. cinerea and particularly A. alternata. T. sp. C.P.K. 524 behaved similar to T. parareesei nom. prov. except that it was the most resistant strain against the isolate of FOX used. We should like to stress that the in vitro dual confrontation tests are only an indirect means to assess antagonistic potential of the fungus. However, the significant difference shown between H. jecorina and T. parareesei nom. prov. suggests that the former is only a moderate mycoparasite with significant variation among isolates, whereas the latter species may exhibit a stronger antagonistic potential against all epigeal fungi tested (Fig. 7).

Discussion
T. reesei QM 6a has been a taxonomic riddle. Originally isolated in 1942, it was a victim of then almost undeveloped taxonomy for Trichoderma. Therefore it was first determined to be T. viride (because the genus was in that time believed to consist only of this single species [35]), and later on recognized as a unique species and named in honor of its detector Elwyn T. Reese T. reesei [36]. Bissett [37] then revised it as being co-specific with T. longibrachiatum. Finally, it was recognized to be identical to the pantropical ascomycete H. jecorina [6], which was itself just distinguished as a separate tropical species closely related to H. schweinitzii [22]. Yet small morphological differences and the inability of these authors to cross it with other H. jecorina isolates in the lab led Kuhls et al. [6] to assume that it is actually a clonally derived asexual form of H. jecorina.
The present data have clearly rejected this hypothesis: our analysis shows that the original isolate T. reesei QM 6a reveals a history of recombination similar to that of the teleomorphic isolates of H. jecorina. These findings are also supported by our recent discovery that T. reesei QM6a is a MAT1-2 idiotype and can indeed be crossed with MAT1-1 partners of H. jecorina [8]. Previous failures to obtain crossing is therefore likely due to the use of the incorrect mating type partner or inappropriate experimental conditions. T. reesei clearly is an isolate identical to H. jecorina, and the minor differences in anamorph morphology [6] and nutrient assimilation [38] to other (more recent) isolates of H. jecorina are likely the result of its maintenance in the laboratory for the last 65 years.
A large number of studies have shown that fungal taxa which were defined on the basis of consistent invariant morphological features in fact contain multiple, well differentiated phylogenetic species [9,[12][13][14][15]. Here we provide a further example for this growing list, the pantropical ascomycete H. jecorina, and show that the 33 strains that were available for this study actually consist of not less than four different species, i.e. H. jecorina, T. parareesei nom. prov., T. sp. nov. C.P.K. 524 and H. sp. nov. G.J.S. 85-238. The formal taxonomic description of T. parareesei nom. prov. will be published elsewhere (L. Atanasova, W.J. Jaklitsch, C.P. Kubicek, and I.S. Druzhinina, manuscript in preparation) but we will refrain from describing the other two phylogenetic taxa based on the small number of strains. Taxonomies with arbitrarily named fungal species (T. parareesei nom. prov. in the present case) have frequently been published [11][12][13][14] and enable researchers to communicate effectively before the formal species nomenclature has been established [9].
H. jecorina yielded evidence for sexual recombination even via a large distance and geographic barriers, thus indicating the presence of a very efficient system for transfer of spores, and/or of the fungus itself (e.g. via wood logs, insects etc.). On the other hand, the same tests clearly rejected this possibility for T. parareesei nom. prov. Thus, these two species represent a diverged species pair, similar to H. orientalis and T. longibrachiatum [15].
Our findings show that those strains of H. jecorina, which were recently isolated as anamorphs from soil and litter and identified as H. jecorina by morphological analysis and molecular barcodes [17] are in fact a cryptic agamospecies, T. parareesei nom. prov. It is very intriguing in this regard that this sibling anamorphic sister species of H. jecorina is similarly restricted to the same narrow belt of620u altitude around the equator, and otherwise similarly pantropical. It consequently represents a rare example of sympatric speciation (i.e. the evolution of reproductive isolation between co-distributed populations) of a saprotrophic fungus [39].
One of the most obvious reasons for such a divergent speciation resulting in formation of an agamospecies would be a mutation in a gene required for sexual reproduction. In such a scenario T. parareesei nom. prov. would have arisen from its ancestor because of a loss of a subpopulation ability to mate. However, our data reject this hypothesis as in such a case all strains of the anamorphic population should have the same mating type. The possession of both MAT1-1 and MAT1-2 loci, although unevenly distributed in our (small) sample, suggest the operation of other speciation mechanisms. The other indirect argument against altered mating mechanism as the driving force for speciation comes from the analysis of topologies of phylogenetic trees. If T. parareesei nom. prov. would have arisen as an agamospecies by a sudden loss of its ability to sexually recombine, its gene sequences should display lower evolutionary rates compared to H. jecorina. In other words the length of the branch leading to T. parareesei nom. prov. from the hypothetical ancestor of both species should be then shorter than the one leading to H. jecorina. However, the individual trees presented here as well as the analysis of an exon of the gene coding for RNA polymerase subunit B II (C.P. Kubicek, I.S. Druzhinina, unpublished data) revealed similar genetic distances between H. jecorina and T. parareesei nom. prov. from their hypothetical ancestor respectively.
The currently most favored explanation for reproductive isolation postulates that hybrid inferiority is caused by antagonistic epistasis between incompatible alleles at interacting loci [40][41][42]. Theoretical models have shown that sympatric speciation may occur when the same genes control both mating and habitat preference or fitness [43]. Our data would be compatible with this hypothesis: growth and conidiation of MAT1-1 strains of T. parareesei nom. prov.-in contrast to H. jecorina -are stimulated by intensive illumination, whereas the MAT1-2 strain behaves similarly to H. jecorina. Evidence for regulation of both mating and carbon source utilization by the blue light has previously been obtained [32][33][34] demonstrating that sexual reproduction and carbon assimilation may indeed share the same regulatory circuits. The availability of further MAT1-2 strains of T. parareesei nom. prov. will help to differentiate between the effect of mating type loci and ecophysiological divergence of two species.
Sympatric speciation is well known from the evolution of plant and human pathogenic fungi [39], but has not been reported for non-pathogenic saprotrophs like Hypocrea/Trichoderma. In this work we applied a bioinformatic sequence analyses and ecophysiological characterization to obtain some insights into the forces thriving speciation in H. jecorina and T. parareesei nom. prov. We have analyzed the global versatility of their carbon metabolism, response to light, conidiation intensity and mycoparasitic potential (for a summary see Table 3). The differences detected are indeed striking and reveal that T. parareesei nom. prov. and H. jecorina occupy different ecological niches in the vertical profile of the tropical forest. T. parareesei nom. prov. displays all the properties of an environmental opportunist: it shows faster growth on a wider spectrum of carbon sources than H. jecorina, and produces a much higher number of propagules on a greater variety of carbon sources. The species is also able to strongly compete with other mycobionts and is mycoparasitic on the epigeal plant pathogenic fungi tested. The latter fact, combined with its profound photostimulation of conidiation, allows us to speculate that T. parareesei nom. prov. might occupy an ecological niche connected with photosynthesizing parts of higher plants, i.e. the canopy of the tropical forest.
H. jecorina in turn seems to be a specialized on a narrow habitat, where its surviving strategy mainly relies on the advantages of sexual reproduction. The species has definitely reduced conidiation efficiency (compared for the general mean for the genus, I.S. Druzhinina, personal observations) and is less aggressive against potentially competing fungi. It is remarkable that the quantitative pattern of carbon metabolism of H. jecorina is highly conserved (with exception of two fast growing strains, Fig. 5A) which may indicate its nutritional specialization. The fact that its anamorph has only very rarely been found in its natural environment [38] further supports the hypothesis that this species is strongly specialized. In our own work on the assessment of the general Hypocrea/Trichoderma biodiversity, we failed to encounter H. jecorina in an asexual form in more than 1000 samples, collected worldwide from soil and litter. It is probable that asexual reproduction in H. jecorina -like it apparently occurred in the case of the original strain of T. reesei and the three strains described in [38] -can be observed only under certain conditions or in certain habitats. Yet the origin of the anamorphic strains G.J.S. 97-177 (CBS 102271) and G.J.S. 97-178 (CBS 102270) both found on dead cacao brooms in Brazil, and G.J.S. 97-38 (CBS 999.97), from soil at a storage lake in French Guiana [38] does not provide a hint towards ecological adaptation of the species. Moreover details of the environment at the US Army camp in Guadalcanal during WW II, where and when T. reesei QM 6a was originally isolated, are also not available. Further samplings in tropical regions may eventually disclose the habitat of one of the most prominent producer in biotechnology.

Material Studied
The strains, their origin and the sequence accession numbers used in this work are listed in Table 1. The isolates are stored at 280uC in 50% glycerol in the laboratory of Vienna University of Technology (TUW). Strains are grouped according to their identification in the present work. For convenience, TUW-lab codes (C.P.K.) are used for the strains throughout, but other collection numbers are also listed in Table 1.

Molecular Genetic Analysis
DNA extraction, PCR amplification and sequencing.
Mycelia were harvested after 2-4 days of growth on MEA at 25uC and genomic DNA was isolated using QIAGEN DNeasyH Plant Mini Kit following the manufacturer's protocol. Amplification of nuclear rRNA gene cluster, containing the ITS1 and 2 and the 5.8S rRNA gene, and of fragments of tef1 (translation elongation factor 1-a) and cal1 (calmodulin) was performed as described previously [15]. The las1 gene (fgenesh5_ pg.C_scaffold_1000016), which encodes the orthologue of an essential nuclear protein regulating bud formation and morphogenesis in S. cerevisiae [21] was amplified using primers given in Table 4. PCR amplification was carried out in an i-cycler (BIO-RAD, USA) for 30 cycles of 94uC for 1 min denaturing, 58uC for 1 min annealing, and 74uC for 50 sec extension. Initial denaturing was at 94uC for 1 min.and the final extension was at 74uC for 7 min. PCR amplification of the mating type loci and the mating type genes was carried out as described [8]. The primers used for MAT1 loci are listed in Table 4. PCR fragments were purified (PCR purification kit, Qiagen, Hilden, Germany), and sequenced at MWG (Ebersberg, Germany).
Phylogenetic analysis. For the main phylogenetic analysis DNA sequences were aligned with Clustal X 1.81 [44] and then edited using GeneDoc 2.6 [45]. The possibility of intragenic recombination, which would prohibit the use of the respective loci for phylogenetic analysis, was tested by linkage disequilibrium based statistics as implemented in DnaSP 4.50.3 [46]. The neutral evolution of coding fragments (cal1 and las1) was tested by Tajima's test implemented in the same software. The interleaved NEXUS file was formatted using PAUP*4.0b10 [47]. The best nucleotide substitution model for the each locus was determined using jMODELTEST [48]. As Akaike and Bayesian Information criteria (AIC [49] and BIC [50] respectively) selected different nucleotide substitution models for every locus and due to the relatively small size of individual datasets (1242 characters per 34 sequences for the biggest) the unconstrained GTR + I + G substitution model was applied to all sequence fragments (Table 2). Metropolis-coupled Markov chain Monte Carlo (MCMC) sampling was performed using MrBayes v. 3.0B4 with two simultaneous runs of four incrementally heated chains that performed 5 million generations. The length of run (number of generations) for each dataset was determined using AWTY graphical system [51] to check the convergence of MCMC. Bayesian posterior probabilities (PP) were obtained from the 50% majority rule consensus of trees sampled every 100 generations after removing the first trees using the ''burnin'' command. Number of discarded generations was determined for each run based on visual analysis of the plot showing generation versus the log probability of observing the data. PP values lower than 0.95 were not considered significant while values below 0.9 are not shown on the resulting phylograms. Model parameters summaries after MCMC run and burning first samplings as well as nucleotide characteristics of used loci are collected in Table 2.
Detection of recombination. he congruence or incongruence of the three gene genealogies was used to infer recombination between isolates. To this end, three different tests were employed: the incongruence length difference/partition homogeneity test (ILD/PHT) [27,28] using a score of P,0.05 to reject the null hypothesis of congruence between loci; the Index of Association (IA) test [29], in which the data were compared to the IAs of artificially recombined datasets; and the Phi-test implemented in SplitsTree [26], which uses the pairwise homoplasy index, PHI ( = W) statistic, to detect refined incompatibility indicating recombination [30].
In addition we applied split decomposition implemented in the SplitsTree program, version 4.0 [25,26], using pairwise distances under the Kimura 3ST model [52].

Ecophysiological Characterization
Phenotype profiling. The carbon assimilation patterns were investigated using Biolog FF MicroPlate TM and EcoPlate TM (Biolog Inc., Hayward, CA, USA) according to the protocol published recently [31]. The complete lists of carbon sources implemented in both plates are given in Table S1. Briefly, strains were grown on 3% malt extract agar (MEA), and 90 ml of a conidial suspension from them (7562% transmission at 590 nm) was dispensed into each of the wells of a corresponding Biolog microplate. Inoculated microplates were incubated at 28uC, and optical density (O.D.) determined after 12,18,24,36,42,48,66 and 72 h at 750 nm. Analyses were repeated at least three times for each strain.
In order to estimate the effect of illumination on mycelial growth and conidiation the protocols of Friedl et al. [32,33] were used respectively. Biolog FF MicroPlates were incubated either at natural day light (May-June, N 48u), while plates for darkness experiment were not exposed to any light source in between the measurements. Biolog EcoPlates were incubated in darkness. Data exploratory statistical analyses were performed using Statistica 6.1 (StatSoft, Inc., Tulsa, OK, USA) data analysis software system. Antagonistic potential. To assess the antagonistic potential of anamorphic cultures of H. jecorina five potential prey fungi have been selected: Sclerotinia sclerotiorum C.P.K. 3593 and FOX (Fusarium oxysporum species complex, strain C.P.K. 1842) to represent soil and rhizosphere competent pathogens and F. xylarioides C.P.K. 3453, Alternaria alternata C.P.K. 3594 and Botrytis cinerea C.P.K. 3592 to represent epigeal plant pathogens. Potential prey fungi were inoculated as agar blocks of the standard size always 1 cm from the edge of the Petri plate and pre-cultivated on 3% PDA in darkness at 25 C. Then similar agar blocks with Trichoderma cultures were introduced on the opposite side of the plate and cultivated for 10 days. Antagonistic potential was semiquantified based on both ability to inhibit the growth of a pathogen and ability to overgrow the mycelium of the pathogenic fungus. One of 5 phases for each confrontation was recorded: 0no inhibition; I -started to inhibit; II -clear signs of inhibition; IIImostly or strongly inhibited; IV -totally inhibited. The ability to overgrow was based on the same scale but recorded using Arabic numbers.
Quantitative and qualitative conidiation assessments.
Quantitative assessment of conidiation was done by measuring conidia density per cm2 produced on Petri dishes on 3% MEA after 10 days of cultivation under natural light/darkness cycle. For this purpose the 6.2 cm2 fragment of an agar plate was cut and rinsed in 15 ml of water containing 0.1% of Tween-80 until visually all conidia were washed out. The concentration of conidia was estimated based on optical density at 540 nm (Biolog Turbidimeter) and transferred into density values based on the calibration curve inferred from the serial dilutions of the standard suspension. In addition the qualitative conidiation assessment when the ability to produce conidia was estimated in respect to carbon metabolism was done using Biolog FF Microplates. In this case conidiation intensity was estimated according to the nominative scale [32] after 72 and 168 hours of incubation. Values 0 and 1 were assigned to the cases when no aerial mycelium and no conidia were detected respectively. Values above 1 corresponded to different intensities of conidiation from single spores (2) to the full coverage of the microplate well by a thick conidial mat (4).
Conidia (final concentration, 10 8 per liter) were used as the inoculum. Cellulase activity in the extracellular culture supernatant was measured using 4-nitrophenyl-b-d-lactopyranoside as a substrate dissolved in a 50 mM sodium citrate buffer, pH 5.0. Other conditions for the assay were the same as used for bglucosidase [54]. One unit (1 U) of enzyme activity is given as the amount of enzyme needed to liberate 1 mmole 4-nitrophenole from the substrate per min. under the conditions of the assay.

Mating Experiments
This was done as described by Seidl et al. [8]. Briefly, the two putative mating partners were placed onto 3% MEA plates 5 cm apart from each other, and incubated at 25uC for 7-10 days in the presence of a natural illumination cycle. In compatibility reactions fruiting bodies were formed at the interaction zone between the two cultures. Figure S1 Single loci phylograms. Bayesian circular phylogram inferred from the concatenated dataset of las1 (A), tef1 (B), cal1 (C) phylogenetic markers. Symbols at nodes correspond to posterior probabilities (PP) .95%.