Taxonomy and phylogeny of the Leptographium olivaceum complex (Ophiostomatales, Ascomycota), including descriptions of six new species from China and Europe

Abstract The Leptographium olivacea complex encompasses species in the broadly defined genus Leptographium (Ophiostomatales, Ascomycota) that are generally characterized by synnematous conidiophores. Most species of the complex are associates of conifer-infesting bark beetles in Europe and North America. The aims of this study were to reconsider the delineation of known species, and to confirm the identity of several additional isolates resembling L. olivacea that have emerged from recent surveys in China, Finland, Poland, Russia, and Spain. Phylogenetic analyses of sequence data for five loci (ACT, TUB, CAL, ITS2-LSU, and TEF-1α) distinguished 14 species within the complex. These included eight known species (L. cucullatum, L. davidsonii, L. erubescens, L. francke-grosmanniae, L. olivaceum, L. olivaceapini, L. sagmatosporum, and L. vescum) and six new species (herein described as L. breviuscapum, L. conplurium, L. pseudoalbum, L. rhizoidum, L. sylvestris, and L. xiningense). New combinations are provided for L. cucullatum, L. davidsonii, L. erubescens, L. olivaceum, L. olivaceapini, L. sagmatosporum and L. vescum. New Typifications: Lectotypes are designated for L. olivaceum, L. erubescens and L. sagmatosporum. Epitypes were designated for L. olivaceapini and L. sagmatosporum. In addition to phylogenetic separation, the synnematous asexual states and ascomata with almost cylindrical necks and prominent ostiolar hyphae, distinguish the L. olivaceum complex from others in Leptographium.


Isolates
All isolates included in this study are listed in Table 1. Reference isolates were obtained from the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa. Ex-type isolates of newly described species were deposited in the Westerdijk Fungal Biodiversity Institute (CBS), Utrecht, in the Netherlands. Type specimens of new species were deposited in the National Collection of Fungi (PREM), Pretoria, South Africa. Taxonomic novelties and new typification events for known taxa were registered in MycoBank (Robert et al. 2013).
PCR products were sequenced with the same primers used for PCR, together with the Big Dye Terminator 3.1 cycle sequencing premix kit (Applied Biosystems, Foster City, California, USA). BigDye PCRs were conducted in 12 μL: sequencing Buffer 4.0 μL, Big Dye 1.0 μL, PCR Grade Water 4.0 μL, primer 1.0 μL, PCR product 2.0 μL; PCR conditions were: 1 min at 96 °C; 25 cycles of 10 sec at 96 °C, 5 sec at 50 °C, and 1min at 60 °C; and finally held at 12 °C. BigDye PCR products were also cleaned up with Sephadex. Sequence analyses were done on the ABI PRISM 3100 Genetic Analyzer (Applied Biosystems, Foster City, California, USA). Consensus sequences were generated from forward and reverse sequences in the CLC Main Workbench 6.0 (CLC Bio, Aarhus, Denmark).

Phylogenetic analyses
Five sequence datasets were analyzed. The ITS2-LSU sequences of the ex-type isolate of every species in the L. olivaceum complex (Table 1) were compared with sequences of other known species in Leptographium obtained from GenBank to show the placement of the complex within the genus. Sequences of Fragosphaeria purpurea and F. reniformis were used to represent the outgroup taxa. Four protein coding gene regions (ACT, TUB, CAL, and TEF-1α) were sequenced (Table 1) for 39 isolates (Table 1) in order to delineate closely related species in the L. olivaceum complex. Sequences for L. procerum and L. profanum from the study of Yin et al. (2015) were selected to represent the outgroup taxa for the four protein-coding gene regions as well as in the combined dataset. Alignments of loci were conducted in MAFFT 7.0 online (Katoh and Standley 2013), then checked manually in MEGA X (Kumar et al. 2018) and compared with the gene maps (Yin et al. 2015) to ensure that introns and exons were aligned appropriately. Three methods were used for phylogenetic analyses including Maximum parsimony (MP), Maximum Likelihood (ML), and Bayesian Inference (BI). A partition homogeneity test was conducted in PAUP* 4.0b10 (Swofford 2002) to consider the congruence of the four protein-coding gene regions before analyses of the combined dataset. The most important parameters used in phylogenetic analyses and statistical values related to all datasets analyzed are presented in Table 2.
MP analyses were executed in PAUP* 4.0b10 (Swofford 2002) with heuristic searches of 1000 replicates and tree bisection and reconnection (TBR) branch swapping options. Gaps were treated as the fifth base. Bootstrap analysis (1000 pseudo replicates) was performed to determine the confidence levels of the branch nodes. Tree length (TL), consistency Index (CI), retention Index (RI), Homoplasy Index (HI), and Rescaled Consistency Index (RC) were recorded after generating the trees.
The best substitution models (Table 2) for the two likelihood methods (ML and BI analyses) were selected congruously in jModelTest 2.1.1 (Pasoda 2008). MEGA X (Kumar et al. 2018) was used for ML analyses with Nearest-Neighbor-Interchange (NNI) branch swapping option. Node support values were determined using analysis of 1000 bootstrap pseudo replicates.
For BI analyses, the Markov Chain Monte Carlo (MCMC) method was used in MrBayes 3.2 (Ronquist et al. 2012). Four MCMC chains were simultaneously run from a random starting tree for five million generations. Trees were sampled every 100 generations. Burn-in values were determined in Tracer v1.7 (Rambaut et al. 2018). Trees sampled in the burn-in phase were discarded and posterior probabilities were calculated from all the remaining trees.

Morphology and growth studies
In order to describe their morphology, isolates of new species were inoculated on to 2% water agar (WA, 20 g Difco agar and 1000 ml deionized water) amended with sterilized pine twigs (Pinus pinaster) and examined microscopically as described by Yin et al. (2015). Culture characteristics were recorded on Oatmeal agar (OA, 30 g oatmeal, 20 g Difco Bacto malt extract, from Becton, Dickinson and Company, and 1000ml deionized water) incubated at 25 °C for 10-14 days. Color descriptions were defined using the charts of Rayner (1970). Growth studies were conducted on 2 % Malt extract agar (MEA) following the procedure described by Yin et al. (2015).

Phylogenetic analyses
The phylogenetic trees arising from the analyses of the ITS2-LSU data for Leptographium s.l. showed the L. olivaceum complex grouping between the L. galeiformis and L. procerum complexes with strong statistical support (Fig. 1). Within the complex, the ITS2-LSU sequences could not distinguish between some of the species, e.g. between L. rhizoidum and L. sagmatosporum; L. davidsonii and L. vescum; L. conplurium, L. pseudoalbum and L. erubescens. Leptographium francke-grosmanniae grouped peripheral to other species in the complex, but remained part of a strongly supported lineage including all the species under consideration.
The ACT data matrix included part of exon 5 (sites 1-678), intron 5 (sites 679-785) and part of exon 6 (sites 786-809). The intron/exon composition of this gene region was congruent with that of the L. procerum complexes (Yin et al. 2015). Analyses of this gene region (Fig. 2) separated all known species and revealed six new taxa in the complex.

ACT+CAL+TEF-1α+TUB
Outgroup The partition homogeneity test conducted on the combined data set for the four protein coding genes (ACT, TUB, CAL and TEF-1α) resulted in a P-value of 0.081, indicating that these regions could be combined. The MP, ML, and BI analyses gener-  ated were consistent with each other. Fourteen species with significant statistical support were defined in the L. olivaceum complex (Fig. 1), including eight known species (L. cucullatum, L. davidsonii, L. vescum, L. olivaceapini, L. erubescens, L. olivaceum, L. sagmatosporum, and L. francke-grosmanniae) and six new species from Europe and China.

Morphology and growth studies
Isolates of the six new species emerging from this study were similar in growth in culture, with colors initially hyaline, later turning pale yellowish or pale olivaceous. Mononematous synnemata were common in the cultures and hyphae were superficial on the agar. The droplets containing conidia were initially hyaline, becoming yellowish with age. Morphological differences among all these new species are discussed in the Notes sections provided with the new species descriptions in the Taxonomy section. A sexual state was induced only in isolates of L. sylvestris after incubation at 25 °C for three weeks.
Other than L. sylvestris that grew fastest at 30 °C, the optimal growth temperature for all isolates of the new species was 25 °C. None of the isolates of the new species grew at 5 °C or 35 °C, only L. rhizoidum was able to grow (2.5 mm/d) at 35 °C.

Taxonomy
Sequence data for 39 isolates included in the present study revealed 14 taxa in the L. olivaceum complex. One of these species, L. erubescens, was previously treated as a synonym of L. cucullatum but our data distinguished clearly between the two species. A new combination is thus provided for L. erubescens. Lectotypes and epitypes are designated here for L. olivaceum, L. sagmatosporum and L. erubescens. The remaining six taxa in the complex represented novel species and descriptions are provided for them.   Description. Sexual state not observed. Conidiophores occasionally observed on wood of WA, macronematous, synnematous, short, wide at the stipe, light brown to yellowish, expanding branches at the apex, 150-230 μm in length including conidiogenous apparatus, 20-25 μm wide at base, 40-45 μm wide at apex, 100-150 μm wide at conidiogenous apparatus. Conidiogenous cells discrete, hyaline, cylindrical, percurrent proliferation, (8-)9-13(-15) × 1.8-2.5 μm. Conidia hyaline, onecelled, smooth, ellipsoidal, (3.7-)4-4.5(-5) × 2.5-3 μm. Culture characteristics: Colonies on OA, hyaline at first, later becoming light yellowish in the center, mycelium superficial on agar. Mostly mycelium observed in culture, synnemata sparse. Optimal temperature for growth 25 °C, growth reduced at 10 °C and 30 °C, no growth at 35 °C.

Leptographium breviuscapum
Host tree. Picea crassifolia. Insect vector. Polygraphus poligraphus. Distribution. Qinghai, China. Note: The asexual state of L. breviuscapum has very short conidiophores making it very easy to distinguish from that of other species in the complex. Etymology. The epithet refers to synnemata produced abundantly in culture.
Host tree. Picea abies. Insect vectors. Dryocoetes autographus, Hylastes brunneus. Distribution. Finland. Notes. All isolates of this species were initially recognized as a cryptic species closely related to L. cucullatum and L. olivaceapini by Linnakoski et al. (2012). Our results confirmed that they represent an undescribed taxon.

Host tree. Pinus sylvestris.
Insect vector. unknown. Distribution. Sweden. Notes. This species was first described by Mathiesen-Käärik (1953) from pine timber in Sweden. No specimen numbers and very little detail (e.g. no host locality or collection dates) were provided in the protologue. Furthermore, no specimen number and little detail are listed under this species name in the herbarium of the Museum of Evolution, Uppsala, which incorporated Mathiesen-Käärik's collection. However, in 1954 she deposited an isolate (No. Sk 13-52) in the CBS labeled as L. erubescens. Two dried specimens (CBS H-7193, CBS H-7194) are linked to this isolate and these are labeled as isotypes. It is reasonable to assume that this isolate represents the original material, but there is no conclusive evidence that this is true. We have thus designated the line drawings from the protologue (Mathiesen-Käärik 1953) as the lectotype. Harrington et al., (2001) suggested that Graphium erubescens (as Phialographium erubescens) represented the asexual state of L. cucullatum (as O. cucullatum) based on ITS sequences. However, based on sequences produced in the present study, the extype culture of L. erubescens differs from that of L. cucullatum in 1bp in ITS2-LSU, 17 bp in ACT, 17 bp in BT, 30 bp in CAL, and 48 bp in TEF-1α. We have thus treated these species as distinct and have provided a new combination for L. erubescens.   (Figs 1-3). Morphologically, the ascospores are almost cylindrical and its ascomatal necks correspond with other species in the complex. But L. francke-grosmanniae produces mononematous conidiophores, in contrast to the synne-mata produced by the other species, which also explains why it is the only species in the complex previously treated in Leptographium. The mode of conidiogenesis of L. francke-grosmanniae (Mouton et al. 1992) appears similar to that of other species where the conidiogenous cells that appear phialidic under a light microscope arise from percurrent proliferation (Wingfield et al. 1989, Wingfield et al. 1991, Mouton et al. 1993). However, the apices of the apparent "phialides" are substantially more flared than those of other species in the complex and they could be more different than assumed by Mouton et al. (1993). Leptographium francke-grosmanniae is also unusual in the L. olivaceum complex in having an angiosperm host.
Leptographium francke-grosmanniae was originally described as Ceratocystis franckegrosmanniae from larval galleries of Hylecoetus dermestoides on Quercus sp. in Germany (Davidson 1971).  showed that sequences for this species produced in different studies were inconsistent. Based on comparisons of the ITS2 region, the sequences of ex-holotype generated in the present study are consistent with those produced by Mullineux and Hausner (2009) for ATCC 22061 and Hamelin et al. (unpublished) for CBS 356.77, but differ substantially from sequences produced by Jacobs et al. (2001b). In the LSU gene region, our sequences are identical to those of Hausner et al. (2000), but they differed from that of Jacobs et al. (2001a, b) for CMW 445.In the β-tubulin gene region, the sequence of CMW 445 in the present study was consistent with that provided by Kim et al. (2004) for CMW 445 and Hamelin et al. (unpublished sequence in GenBank) for CBS 356.77. We thus suggest that the two sequences for L. francke-grosmanniae produced by Jacobs et al. (2001a, b) are incorrect. Sequences of another isolate from the USA (CMW 2975), previously identified as L. francke-grosmanniae (Zipfel et al. 2006), differ substantially from the ex-holotype culture. Thus, this isolate (CMW 2975) does not represent L. francke-grosmanniae, and its taxonomic placement needs reconsideration.  Zipfel et al., Stud. Mycol. 55: 91 (2006 Olchowecki and Reid (1974, pp 1699-1700); Upadhyay (1981, pp 52-54, figs 116-121); Mouton et al. (1993, pp 376-377, figs 19-22 Notes. This species was first described invalidly (no Latin diagnosis) from Pinus sylvestris infested by a longhorn beetle Acanthocinus aedilis in Sweden (Mathiesen-Käärik 1950). Mathiesen-Käärik (1951) then validated the name with a more detailed description accompanied by a Latin diagnosis. In the original descriptions of L. olivaceum by Mathiesen-Käärik (1950, 1951, the host tree, beetle and location of the collection was noted, but no mention was made of a specimen. The herbarium specimens of Mathiesen-Käärik were initially curated in the herbarium of the Statens Skogsforsknings institut, Experimentalfältet, Sweden. The collection was later incorporated into the herbarium of the Museum of Evolution, Uppsala. Only one herbarium specimen (UPS:BOT:F-130986) of L. olivaceum, collected from the same host, beetle and location by T. Hedquist, is available from that collection. However, an isolate of L. olivaceum , collected in 1949, also from the original host and location, was deposited in the CBS by Mathiesen-Käärik in 1951. Although we were not able to confirm that this isolate was from the original collection, it was treated as the ex-type culture of the species in previous studies , Linnakoski et al. 2012. In view of the absence of concrete evidence that this isolate represents the original material, we have designated the line drawings from the protologue (Mathiesen-Käärik 1951) as lectotype.
More recently, it was reported from Picea abies and Pinus sylvestris infested by Ips typographus and Dryocoetes autographus in Finland and Russia, in a study where the identities were confirmed using DNA sequence analyses (Linnakoski et al. 2012). Griffin (1968) reduced L. vescum to synonymy with L. olivaceum, but data from the present study confirmed that these two species are phylogenetically distinct.

Notes.
No living culture associated with the holotype (BPI 595910) or isotype (BPI 595914) of L. olivaceapini exists. However, T. Hinds, a collaborator of R.W. Davidson and later curator of the RWD culture collection, provided an isolate (COLO 479) labeled as C. olivaceapini to M.J. Wingfield, who later deposited this in the CBS (CBS 504.86). The species name and origin provided by Hinds with the isolate corresponds to a second specimen mentioned by Davidson (1971, p. 10) in the protologue (RWD 581-D = BPI 596223). In our opinion, the isolate (COLO 479) most probably originated from the specimen (RWD 581-D). We could not confirm with certainty that BPI 296223 originated from RWD 581-D and thus designated a dried culture of COLO 479 as the epitype for L. olivaceapini.

Leptographium rhizoidum
Host tree. Pinus radiata. Insect vectors. Hylastes ater, H. attenuatus, Hylurgops palliatus, Ips sexdentatus. Distribution. Spain. Note: Isolates of L. rhizoidum from pine-infesting bark beetles in Spain were initially identified as L. olivaceum based on ITS sequences by Romon et al. (2007). Our data showed them to be distinct from that species. This species produced more abundant and longer rhizoids than others in the complex.
Host trees. Pinus strobus, Picea mariana. Insect vectors. unknown bark beetle species. Distribution. Canada. Notes. This species was originally described from bark beetle galleries and freshly cut surfaces of Picea mariana, Pinus resinosa and Pinus strobus in Canada (Wright and Cain 1961). The Royal Ontario Museum Fungarium (TRTC), Canada, informed the authors of this study that the holotype (TRTC 36427) of L. sagmatosporum was permanently lost. There is also no living culture available from the holotype. We have thus designated the line drawings in the protologue as the lectotype. An isolate (CMW 34135), also from pine in Ontario, identified as L. sagmatosporum based on morphology (K. Jacobs, unpublished) and used in previous studies to represent the species , Linnakoski et al. 2012, its dry specimen is designated here as the epitype.
Host trees. Pinus sylvestris, Picea abies. Insect vector. Ips typographus. Distributions. Poland, Finland. Notes. The Finnish isolate (CMW 23300) was considered by Linnakoski et al. (2012) to be the same undescribed species as the isolates described above as L. conplurium. The addition of a newly obtained isolate from Poland in the present study, confirmed that the two isolates represented a distinct taxon, clearly separated from all other species in the complex. This is the only new species for which ascomata were obtained in culture. Single ascospore isolates of this species produced ascomata in culture, suggesting that the species is homothallic. The common characters of sexual states of species in this complex are having ascomata with sheath and ostiolar hyphae on the top of neck. This species differs from others by its fusiform to orange section shaped ascospores and slightly curved neck.  Descriptions. Davidson (1958, p. 666);De Hoog and Scheffer (1984, p. 295, fig. 2); Samuels (1993, p. 16, fig. 1C-F).

Distribution. China.
Note. This species groups closely with L. conplurium and L. erubescens, but can be distinguished by its dark conidial droplets. In addition, the synnematous conidiophores of this species were shorter, and its conidia were bigger than that of L. erubescens. Additional

Discussion
Among the five loci used in the phylogenetic analyses, ACT, CAL, and TEF-1α were able to distinguish among all species in the L. olivaceum complex. In contrast, TUB sequences could not distinguish between L. davidsonii and L. vescum. Although ITS2-LSU sequences provided reasonable resolution for species complexes at the genus level, this region could not be used to distinguish among closely related species. Of the five gene regions, TEF-1α had the most variable sites and this is consistent with the results of Yin et al. (2015) for the L. procerum complex. This also supports their suggestion that TEF-1α is suitable for use as a barcoding gene for accurate species identification in Leptographium.
In this study, we have clarified the previous confusion related to the ex-type isolate of L. francke-grosmanniae, and although our phylogenetic data placed it close to the complex, it grouped separated from all other species. This is consistent with its mon- onematous morphology that distinguishes it from all other species in the complex that produce synnematous asexual states. Furthermore, it is unique in that it does not come from the galleries of a conifer-infesting scolytine bark beetle like the other species, but from the large timberworm beetle, Hylecoetus dermestoides (Coleoptera: Lymexylidae), infesting a Quercus sp. (Davidson 1971). Some beetles in the latter genus are known to vector ambrosial yeasts (Batra and Francke-Grosmann 1961), but the role and biology of L. francke-grosmanniae in these galleries on oak remains unknown. If these beetle ecosystems in hardwoods are explored further, it seems reasonable to expect that additional species related to L. francke-grosmanniae could be discovered. These would most likely emerge as a species complex distinct from the L. olivaceum complex.
All species in the L. olivaceum complex, with the exception of L. francke-grosmanniae, share various characteristics. Apart from similar sexual and asexual morphology (as discussed in the introduction), these species are all associated with scolytine bark beetles infesting primarily species of pine (Pinus) and spruce (Picea). Only L. davidsonii has been reported from another conifer genus, namely Pseudotsuga (Douglas-fir). However, there is no evidence for strong host or beetle specificity among these fungi. The European spruce bark beetle, Ips typographus, for example, infests various species of spruce and pine, and L. cucullata, L. olivacea, and L. poloniae, have been isolated from this beetle or its galleries. Nothing is known regarding the pathogenicity of any of the species in the complex, but Griffin (1968) and Davidson (1958) showed that some species were responsible for the blue-stain of the timber.
In terms of the distribution of species in the L. olivaceum complex, our data suggest that most of these taxa are geographically restricted to the continents from which they have been recorded. Four species have been reported only from North America, namely L. davidsonii, L. olivaceapini, L. sagmatosporum, and L. vescum, while L. olivaceum, L. erubescens and four of the new species have been found only in Europe and western Russia. Two of the new species originate from China. Only L. cucullatum has been found in Europe and East Asia, specifically Japan.
The results of this study incorporating data for morphology, ecology, and phylogenetic inference based on DNA sequences for five loci have confirmed that the L. olivaceum complex is a well-defined species complex in Leptographium. Moreover, this integrative approach has been recently employed to resolve lower-level taxonomy in several other groups of fungi such as the Ophiocordycipitaceae (Araújo et al. 2015), Pyronemataceae (Sochorová et al. 2019), Laboulbeniaceae (Haelewaters et al. 2018), Geastraceae (Sousa et al. 2017), and Helvellaceae (Skrede et al. 2017). The combination of multiple properties as independent lines of evidence (e.g., morphology, DNA, substratum, and/or geography) is the way to move forward in fungal taxonomy in general.

Conclusions
In the present study, DNA sequences for five loci were amplified and used to reconstruct phylogenies for species in the L. olivaceum complex. Multilocus phylogenies distinguished clearly among the eight previously described species and also revealed six species: L. breviuscapum, L. conplurium, L. pseudoalbum, L. rhizoidum, L. sylvestris, and L. xiningense that are newly described. TEF-1α was recognized as the best candidate gene to distinguish all species in the complex. For several of the previously known species, problems relating to type specimens were identified, and to resolve these, seven new combinations, two epitypes and three lectotypes have been designated. Following the "one fungus one name" principles, this study provided a model solution to resolving interspecific relationships within the species complexes in the Ophiostomatales. More work should be done on other unresolved species complexes of Leptographium and other lineages in the Ophiostomatoid fungi in the future.