Front line defenders of the ecological niche! Screening the structural diversity of peptaibiotics from saprotrophic and fungicolous Trichoderma/Hypocrea species

Approximately 950 individual sequences of non-ribosomally biosynthesised peptides are produced by the genus Trichoderma/Hypocrea that belong to a perpetually growing class of mostly linear antibiotic oligopeptides, which are rich in the non-proteinogenic α-aminoisobutyric acid (Aib). Thus, they are comprehensively named peptaibiotics. Notably, peptaibiotics represent ca. 80 % of the total inventory of secondary metabolites currently known from Trichoderma/Hypocrea. Their unique membrane-modifying bioactivity results from amphipathicity and helicity, thus making them ideal candidates in assisting both colonisation and defence of the natural habitats by their fungal producers. Despite this, reports on the in vivo-detection of peptaibiotics have scarcely been published in the past. In order to evaluate the significance of peptaibiotic production for a broader range of potential producers, we screened nine specimens belonging to seven hitherto uninvestigated fungicolous or saprotrophic Trichoderma/Hypocrea species by liquid chromatography coupled to electrospray high resolution mass spectrometry. Sequences of peptaibiotics found were independently confirmed by analysing the peptaibiome of pure agar cultures obtained by single-ascospore isolation from the specimens. Of the nine species examined, five were screened positive for peptaibiotics. A total of 78 peptaibiotics were sequenced, 56 (=72 %) of which are new. Notably, dihydroxyphenylalaninol and O-prenylated tyrosinol, two C-terminal residues, which have not been reported for peptaibiotics before, were found as well as new and recurrent sequences carrying the recently described tyrosinol residue at their C-terminus. The majority of peptaibiotics sequenced are 18- or 19-residue peptaibols. Structural homologies with ‘classical representatives’ of subfamily 1 (SF1)-peptaibiotics argue for the formation of transmembrane ion channels, which are prone to facilitate the producer capture and defence of its substratum.


Introduction
Currently, the fungal genus Trichoderma/Hypocrea 1 comprises more than 200 validly described species, which have been recognised by molecular phylogenetic analysis . This high taxonomic diversity in Trichoderma/Hypocrea is not only reflected in a permanently increasing number of species (Jaklitsch 2009(Jaklitsch , 2011 Both the taxonomic and metabolic diversity of Trichoderma/Hypocrea are hypothesised to originate from mycoparasitism or hyperparasitism, which may represent the ancestral life style of this genus (Kubicek et al. 2011). The unique bioactivities of peptaibiotics, resulting from their amphipathicity and helicity, make them ideal candidates to support the parasitic life style of their fungal producers: Under in vitro-conditions, the parallel formation of peptaibiotics such as the 19-residue trichorzianins 2 and of hydrolytic enzymes, above all chitinases and β-1,3-glucanases (Schirmböck et al. 1994), could be demonstrated. This observation led to a widely accepted model describing the synergistic interaction of peptaibiotics and hydrolases in the course of mycoparasitism of Trichoderma atroviride towards Botrytis cinerea (Lorito et al. 1996). Despite this, reports on in vivo-detection of peptaibiotics have scarcely been published in the past. Examples include the isolation of hypelcins A and B obtained from ca. 2 kg of dried, crushed stromata of the mycoparasite Hypocrea peltata (Fujita et al. 1984;Matsuura et al. 1993Matsuura et al. , 1994 3 as well as the detection of antiamoebins in herbivore dung, which have been produced by the coprophilous Stilbella fimetaria (syn. S. erythrocephala) (Lehr et al. 2006).
In order to close this gap, we initiated a screening project aimed at resolving the question as to whether peptaibiotic production in vivo is a common adaptation strategy of Trichoderma/ Hypocrea species for colonising and defending ecological niches: Several Hypocrea specimens were freshly collected in the natural habitat and analysed for the presence of peptaibiotics. Sequences of peptaibiotics found were independently confirmed by analysing the peptaibiome 4 of pure agar cultures obtained by single-ascospore isolation from the specimens. Using liquid chromatography coupled to electrospray high resolution mass spectrometry we succeeded in detecting 28 peptaibiotics from the polyporicolous Hypocrea pulvinata (Röhrich et al. 2012). Another 49 peptaibiotics were sequenced in Hypocrea phellinicola, a parasite of Phellinus sp., especially Ph. ferruginosus (Röhrich et al. 2013a).
Due to these encouraging results, our screening programme was extended to another nine specimens belonging to seven hitherto uninvestigated mycoparasitic or saprotrophic Trichoderma/Hypocrea species, respectively (Table 2).

Materials and methods
Specimens of Hypocrea teleomorphs were collected from four different locations in Austria (Table 3). Pure agar cultures were obtained by single-ascospore isolations from the respective, freshly collected specimens as previously described by Jaklitsch (2009): Parts of stromata were crushed in sterile distilled water. The resulting suspension was transferred to cornmeal agar plates (Sigma, St. Louis, Missouri) supplemented with 2 % (w/v) D(+)-glucose-monohydrate (CMD), and 1 % (v/v) of an aqueous solution of 0.2 % (w/v) streptomycin sulfate (Sigma) and 0.2 % (w/v) neomycin sulfate (Sigma). Plates were incubated overnight at 25 °C. In order to exclude possible contamination by spores of other fungal species, few germinated ascospores from within an ascus were transferred to fresh plates of CMD using a thin platinum wire. The plates were sealed with Parafilm (Pechiney, Chicago, Illinois) and incubated at 25 °C. As all species listed in Table 2 could unambiguously be identified by their morphological and growth characteristics (Jaklitsch 2009(Jaklitsch , 2011, no molecular phylogenetic analyses needed to be performed.
Detailed descriptions of chemicals, extraction and work-up procedures for specimens and agar plate cultures, cultivation methods, as well as comprehensive protocols for HPLC/ QTOF-ESI-HRMS were given by Röhrich et al. (2012Röhrich et al. ( , 2013a. For routine screening, a high-resolution micrOTOF Q-II mass spectrometer with orthogonal ESI source (Bruker Daltonic, Bremen, Germany), coupled to an UltiMate 3000 HPLC (Dionex, Idstein, Germany), was used. Samples, which have been screened negative with the above HPLC/MS system, were re-examined using a maXis 3G QTOF mass spectrometer with orthogonal ESI source (Bruker Daltonic, Bremen, Germany), coupled to an UltiMate 3000 UHPLC (Dionex, Idstein, Germany) as previously described (Röhrich et al. 2012(Röhrich et al. , 2013a.

General considerations
All strains investigated in this study represent phylogenetically well-defined species (Tables  2 and 3). This is in contrast to most of the reports published until the end of the 1990s, when peptaibiotic production by the genus Trichoderma/Hypocrea was -according to Rifai's classification (1969) -mostly attributed to one of the four common species T. viride, T. koningii, T. harzianum, T. longibrachiatum, and sometimes to T. pseudokoningii and T. aureoviride. Careful inspection of the literature published prior to the turn of the millennium revealed that only three of the Trichoderma strains, reported as sources of 'classical' peptaibiotics have correctly been identified and appropriately been deposited, viz. the paracelsin-producing T. reesei QM 9414 (Brückner and Graf 1983;, the trichosporin/trichopolyn producer T. polysporum TMI 60146 (Iida et al. 1990(Iida et al. , 1999, and the paracelsin E-producing T. saturnisporum CBS 330.70 (Ritieni et al. 1995). Furthermore, none of the numerous peptaibiotic-producing strains, reported to belong to those six Trichoderma species mentioned above, has subsequently been verified by phylogenetic analyses. Statements on the identity of the producers must therefore be regarded with great caution, unless it is being described how isolates were identified (Degenkolb et al. 2008). Unfortunately, most of the peptaibiotic-producing Trichoderma/ Hypocrea strains investigated prior to 2000 have never been appropriately deposited either i) in a publicly accessible culture collection or ii) in an International Depositary Authority (IDA) under the conditions of the Budapest Treaty; thus, they are not available to independent academic research. As misidentifications persist to be a continuous problem, not only in the older literature (Neuhof et al. 2007), the authors prefer to introduce new names for the peptaibiotics sequenced in this study. Those new names refer to the epithets of the producing species.

Screening of Hypocrea voglmayrii
The most notable species screened is by far H. voglmayrii (Fig. 3), the specimen of which produced two 18-residue deletion sequences, compounds 35 and 36, which lack the Cterminal amino alcohol, as well as 15 19-residue peptaibols, compounds 37-51 (Tables 8  and 9, Table S3a and S3b). As all of them are new, the names voglmayrins 1-17 are introduced. They partly resemble the building schemes of trichokonin V (Huang et al. 1995) and of trichorzianins B (Rebuffat et al. 1989). Six of the major compounds (40-45) carry a C-terminal phenylalaninol (Pheol) residue, whereas three minor compounds (37-39) terminate in tyrosinol (Tyrol) -a residue that has not been described for peptaibiotics until only recently (Röhrich et al. 2013a). Another six major compounds (46-51) display an additional fragment ion 68.0628 ± 2.3 mDa at their C-terminus (Fig. 4). Thus, the p-OH group of their Tyrol residue is hypothesised to be substituted by a prenyl or isoprenyl residue (C 5 H 8 , for details see paragraph below). In contrast to this, major 19-residue peptaibols produced by the plate culture, compounds 40, 41, 43, 44, and two additional compounds, 52 and 53, voglmayrins-18 and -19, terminate in Pheol. HR-MS data clearly confirm the presence of additional minor components carrying a C-terminal Tyrol or prenylated Tyrol residue, respectively. Unfortunately, the intensities were too low for MS/MS sequencing of the respective y 6 ions. Two 11-residue lipopeptaibols, compound 54 and 55, resembling lipostrigocin B-04/B-05 (Degenkolb et al. 2006a) and trichogin A IV (Auvin-Guette et al. 1992), have also been sequenced.

Screening of Hypocrea sulphurea
All three specimens of H. sulphurea were negatively screened for peptaibiotics. From two of them, plate cultures could be obtained; however, those were also screened negatively (data not shown).

Screening of Hypocrea parmastoi
Neither specimen, nor plate culture of H. parmastoi displayed the presence of peptaibiotics (data not shown).
Screening of specimens collected in the natural habitat(s) corroborated the distinguished importance of the genus Trichoderma/Hypocrea as the currently richest source of peptaibiotics. Five of the nine specimens were screened positively, and the results of this screening confirmed by the sequences obtained from screening of the plate cultures. Notably, 56 of the 78 peptaibiotics (72 %) detected represent new sequences.
Screening of H. voglmayrii and H. citrina revealed five peptaibols (compounds 37-39, 70, and 73) carrying a C-terminal Tyrol, a residue quite recently described for H. phellinicola (Röhrich et al. 2013a), which is considered comparatively rare. The additional substituent of the C-terminal Tyrol of voglmayrins 12-17 (compounds 46-51), which has tentatively been assigned as a prenyl or isoprenyl (C 5 H 8 ) residue, is hypothesised to be located at the phydroxy group. A regiospecific O-prenylation at the 4-position of the aromatic ring has recently been demonstrated for SirD (Zou et al. 2011), a tyrosine O-prenyltranferase (Kremer and Li 2010) catalysing the first pathway-specific step in the biosynthesis of the phytotoxin sirodesmin PL. The latter is produced by Leptosphaeria maculans (anamorph: Phoma lingam), the causal agent of blackleg of canola (Brassica napus). Recently, Oprenyltyrosine diketopiperazines have been described from Fusarium sp. and Penicillium crustosum (Guimarães et al. 2010).
Another notable structural element, dihydroxy-Pheol was found at the C-terminus of hypocitrin-1 (compound 69). While the presence of either Pheol or Tyrol may be assumed to originate from the relaxed substrate specificity in the terminal adenylate domain of the respective peptaibol synthetase, the direct incorporation of dihydroxy-Phe, presumably 3,4dihydroxy-L-Phe (DOPA), is one possible biosynthetic route. Fungal tyrosinases are known to oxidise not only Tyr and various other monophenols, e.g. in the route to melanins, but also act on tyrosyl residues within peptides and proteins, leading to the formation of interand intra-molecular crosslinks (Selinheimo et al. 2007). Thus, Tyrol-containing peptaibols could be further oxidised by tyrosinases, and even become attached to components of the fungal cell wall (Mattinen et al. 2008).
The results of our screening programme further extend the list of peptaibiotic-producing species of Trichoderma/Hypocrea compiled in Table 14. Most notably, the sequences of peptaibiotics produced by the freshly collected specimens are either identical to those found in the plate cultures, or represent -at least -closely related homologues and positional isomers of the latter. Thus, our LC-MS/MS screening approach confirmed that all peptaibiotic-producing specimens and plate cultures obtained thereof represent one and the same species. Consequently, the same type (= subfamily) of peptaibiotics is produced both in the natural habitat and under artificial (= laboratory) conditions -a fact, which is important for the application of Trichoderma formulations in biocontrol and integrated pest management schemes. ATrichoderma/Hypocrea species capable of producing peptaibiotics under the conditions of its natural habitat may defend its ecological niche more effectively compared to a non-producing species, as will be outlined below. At present, ca. 15 % of the phylogenetically verified Trichoderma/Hypocrea species have been positively screened for peptaibiotics; however, it appears that the inventory of peptaibiotics of the remaining 85 % is still waiting to be scrutinised by state-of-the-art bioanalytical -particularly mass spectrometric -methods. Of approximately 130 Trichoderma/Hypocrea species prescreened by LC/HRMS (Nielsen et al. 2011), ca. 60 were found to produce peptaibiotics 8 . Thus, the production of peptaibiotics in the natural habitat seems to be independent of the habitat preference, i.e. mycoparasitism vs. saprotrophy (Chaverri and Samuels 2013), but neither predictable per se nor universal.
Given that peptaibiotics are readily biosynthesised in the natural habitat of the producers, they could significantly contribute to the complex interactions of phytoprotective Trichoderma species, which are used in commercial or semi-commercial biocontrol agents (BCAs) against plant pathogenic fungi (Harman et al. 2004;Viterbo et al. 2007;Vinale et al. 2008a, b). Examples of successful biocontrol approaches using Trichoderma strains include 'Tricovab', a Brazilian formulation recently approved (Anonymous 2012) for integrated management of Crinipellis (syn. Moniliophthora) perniciosa, the causal agent of Witches' broom of cacao (Pomella et al. 2007;Loguercio et al. 2009;Medeiros et al. 2010). Notably, 'Tricovab' contains a peptaibiotic-producing strain (Degenkolb et al. 2006a) of the hyperparasitic endophyte Trichoderma stromaticum. Moreover, the in vivo-detection of peptaibiotics corroborates the recently demonstrated pro-apoptotic in vitro-activities of the 19-residue peptaibols trichokonin VI 9 (Huang et al. 1995) from Trichoderma pseudokoningii SMF2 towards plant fungal pathogens such as Fusarium oxysporum (Shi et al. 2012).
The value of peptaibiotics for chemotaxonomy of Trichoderma/Hypocrea has scarcely been scrutinised in the past (Neuhof et al. 2007;Degenkolb et al. 2008). To exhaustively answer this question, a larger number of strains, belonging to recently described species, are required to be included in an LC-MS/MS-based study aimed at analysing the peptaibiome of strains and species within different clades of Trichoderma/Hypocrea. However, statements on peptaibiotic production by a particular Trichoderma/Hypocrea species must always be treated with great caution as they are highly habitat-, isolate-, and/or cultivation-dependent. Furthermore, 'peptaibol subfamilies' were introduced at a time when the total number of peptaibiotics described did not exceed 200 (Chugh and Wallace 2001) -less than a sixth of the currently known sequences. Notably, the additional 1,000-1,100 individual peptaibiotics published since then exhibit both new building schemes and constituents. This issue becomes even more complex as 'peptaibol subfamilies' were published when phylogenetic methods have not yet been recognised as an indispensable tool in fungal taxonomy. Thus, a considerable number of peptaibiotics, the sequences of which have been elucidated correctly, cannot be linked to an unambiguously identified producer that is deposited in a publicly accessible culture collection. These facts illustrate the urgent need to reconsider the classification into the nine subfamilies -a task that has to be completed before the aforementioned study can be performed.
Currently, any approach for a peptaibiotics-based chemotaxonomy of Trichoderma/ Hypocrea must be regarded as extremely complicated -even within a defined clade -, because i) peptaibiotics only represent one single class of secondary metabolites produced by Trichoderma/Hypocrea, ii) most of the producers reported in literature have never been deposited appropriately, and iii) the persistently high degree of misidentification makes any comparison between members of different clades problematic and challenging. This is illustrated by the following examples (references are compiled in Table 14): i. The 20-residue alamethicins (ALMs) have hitherto been found in four species belonging to the Brevicompactum clade of Trichoderma; however, it is not yet possible to estimate if the Pro 2 residue of the ALMs could be regarded as a structurally highly conserved position, comparable to the Pro 14 residue. Chemotaxonomy of the Brevicompactum clade encompassed the comparison of hydrophobins, peptaibiotics, and low-molecular weight secondary metabolites, including simple trichothecene-type mycotoxins.
iii. Trichoderma citrinoviride strains S 25 and IMI 91968 are rich sources of 20residue peptaibols of the paracelsin/saturnisporin/trichocellin/suzukacillin/ trichoaureocin-type. These are the only two strains of T. citrinoviride that have been investigated for peptaibiotics. Hypocrea schweinitzii ICMP 5421, which has also been verified phylogenetically (Réblová and Seifert 2004), had only been screened positive for Aib by GC/MS; but -to the best of the authors' knowledgespecimens of that species have never been investigated for its inventory of peptaibiotics. Parcelsins, which have been isolated from T. reesei QM 9414, are also produced by a member of the Longibrachiatum clade. However, the producer of saturnisporin (T. saturnisporum MNHN 903578: Rebuffat et al. 1993) has never iv. T. flavofuscum CBS 248.59 is the only species of Trichoderma/Hypocrea, which produces 13-residue sequences -notably trichofumins C and D are the only two peptaibols of that chain length reported to date. They display the rare Gln-Gln motif in positions 5 and 6. Looking at the sequences, their biosynthesis seems to be distantly related to that one of trichofumins A and B (and positional isomers thereof). The latter are 11-residue SF4-peptaibols and widespread amongst Trichoderma/Hypocrea species.
v. T. virens strain Tv29-8 produces common 11-and 14-residue peptaibols, and it is the only phylogenetically verified source of 18-residue peptaibols of the trichorzintype.
However, the results of our LC-MS/MS screening are also of interest for analysis of environmental samples as well as extraterrestrial materials such as carbonaceous meteorites as their contamination by propagules of soil-or airborne peptaibiotic-producing fungi has to be taken into account (Brückner et al. 2009;Elsila et al. 2011).
To sum up, production of peptaibiotics may generally be regarded as a sophisticated ecological adaptation for the producing fungus providing it with an obvious advantage over non-producing fungal and other competitors. This group of 'chemical weapons' in their 'armoury' may effectively assist a remarkable number of strains currently identified as belonging to ca. 30 Trichoderma/Hypocrea species in colonising and defending their ecological niches.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.