Botryandromyces, a morphology-based genus concept scrutinized by molecular data

Laboulbeniales (Ascomycota) are an order of understudied, biotrophic microfungi uniquely associated with arthropods. More than 2300 species are described but only a fraction of those have been sequenced. Molecular studies have shown that cryptic diversity and phenotypic plasticity are present within the Laboulbeniales. Thus far, all of the 146 genera described in Laboulbeniales have been based on morphological characteristics; features commonly used to delineate genera are the organization of receptacle cells and the number of perithecial outer wall cells. The genus Botryandromyces was erected to accommodate two species, B. heteroceri and B. ornatus (type), which share similar morphological characteristics and are different from other genera in their number of perithecial outer wall cells. Here, we generated sequences of multiple loci (18S, ITS, and 28S) of B. heteroceri and several Laboulbenia species. Our phylogenetic analyses retrieved Botryandromyces within Laboulbenia with high support. The two Botryandromyces species are similar to related Laboulbenia species in their upper receptacle (i.e., cells IV and V). We propose to transfer Botryandromyces ornatus and B. heteroceri to Laboulbenia as L. heteroceri and L. mairei nom. nov., respectively, due to a complicated taxonomic history. These results advocate the use of molecular data and the necessity of an integrative taxonomy approach in the study of Laboulbeniales not only to delineate species, but also to investigate relationships among species, genera, and higher taxa as well as to understand the evolution of morphology in this group of fungi.


Introduction
Macroscopic and microscopic morphology of sporocarps has traditionally been of great importance in identifying and describing species of fungi and classifying them into higher taxa (Bridge et al. 2005;Cao et al. 2021;Maharachchikumbura et al. 2021).The use of molecular data has challenged several proposed morphogroups.A well-studied example within Russulales (Agaricomycetes) is the convergent evolution of sequestrate fruiting bodies in the genera Lactarius and Russula.Multiple genera were erected to accommodate these sequestrate forms, but early molecular work has indicated that these genera are polyphyletic and that these sequestrate forms independently evolved multiple times within Lactarius and Russula (Miller et al. 2001;Nuytinck et al. 2003;Eberhardt and Verbeken 2004).Similar cases can be found in other genera including Agaricus, Amanita, Cortinarius, and Entoloma (Peintner et al. 2001;Co-David et al. 2009;Justo et al. 2010;Sánchez-García et al. 2020).Similarly, morphologically defined higher taxa of Ascomycota have also been challenged by molecular data (Arzanlou et al. 2007;Crous et al. 2007Crous et al. , 2009Crous et al. , 2021;;Wynns 2015).
A severely understudied group of Ascomycota is the order Laboulbeniales (Laboulbeniomycetes).These fungi obligately live on the exoskeleton of arthropod hosts.They do not form a typical hyphal system but a multicellular, 3-dimensional structure called a thallus (Haelewaters et al. 2021a).More than 2300 species in 146 genera are currently recognized in Laboulbeniales.The vast majority of this diversity is described based on morphology while only a fraction has been sequenced (Haelewaters et al. 2021b).Molecular studies have presented evidence that both cryptic diversity and phenotypic plasticity are present within Laboulbeniales (Goldmann and Weir 2012;Goldmann et al. 2013;Haelewaters et al. 2018;Haelewaters and Pfister 2019;Van Caenegem et al. 2023a).This makes delineating the taxa of Laboulbeniales solely based on morphology difficult.The first molecular phylogenies of Laboulbeniales showed that several morphologically defined higher taxa are non-monophyletic.Antheridial characteristics appear to have a low systematic value, while features of the perithecium seem to be phylogenetically informative (Goldmann and Weir 2018;Haelewaters 2018).
In the protologue of Laboulbenia heteroceratis, Thaxter (1912) wrote that the production of sessile antheridia from proliferous cells had not yet been reported within the genus Laboulbenia.However, he stated that "the basal cells of its appendages may assume an appearance very similar to that of some of the aquatic forms on Gyrinidae."The insertion cell of L. heteroceratis is concolorous with the surrounding cells.It also tends to divide into several smaller cells (Thaxter 1912).In contrast, the vast majority of species in Laboulbenia have a simple, blackened insertion cell.Thaxter (1912) did not include figures in his description, but L. heteroceratis is illustrated in Tavares and Majewski (1976: Fig. 2) and Goldmann and Weir (2018: Fig. 4 I).
Botryandromyces heteroceri was described as Misgomyces heteroceri by Maire (1920).Species of Misgomyces have perithecia with four outer cell walls of unequal height in each vertical tier and compound antheridia (Tavares 1985).The genus Botryandromyces was erected to accommodate M. heteroceri, as it has a perithecium with different cellular organization and simple sessile antheridia (Tavares and Majewski 1976).Botryandromyces heteroceri differs from B. ornatus by showing considerable variation in the number of cells in the lower receptacle, ranging from two to eight, and even 33 in aberrant, filiform thalli.Botryandromyces ornatus always has a two-celled lower receptacle and shows a blackening on the perithecial apex, which B. heteroceri lacks (Tavares 1985;De Kesel 2009;Santamaria and Pedersen 2021).
Recent molecular work based on the small ribosomal subunit (18S) of the ribosomal RNA (rRNA) gene placed B. ornatus within Laboulbenia, which makes the latter a paraphyletic group (Goldmann and Weir 2018).The authors refrained from making taxonomic changes given their restricted sample size (one isolate of Botryandromyces, three isolates of Laboulbenia).Although morphological differences between the two genera are clear and well-defined, this result was not surprising given Thaxters' (1912) decision to place the taxon currently accepted as B. ornatus in Laboulbenia.Haelewaters (2018) retrieved B. ornatus as a sister to the genus Laboulbenia, which was represented by 13 isolates.Also, this analysis was only based on the 18S region.
Here, based on recently collected material, we present a phylogeny incorporating new sequence data from three loci for B. heteroceri and show the placement of both species of Botryandromyces in relation to Laboulbenia.

Collection and identification of beetles and Laboulbeniales, and morphological study
Specimens of Heterocerus Fabricius, 1792 (Coleoptera, Heteroceridae), were captured alive in 2022 and 2023 using a light trap (160w ML) in a private garden in Herzele, Belgium.Specimens were immediately screened alive for infections with Laboulbeniales using a dissecting microscope.Infected specimens were stored in 99% ethanol, and uninfected specimens were released back into nature.Other hosts included in this study (Coleoptera, Carabidae) were sent by entomologists or collected by W.V.C. and A.D.K. using pitfall traps and by hand, from multiple localities in Belgium, Latvia, the Netherlands, Uganda, and the USA.These specimens were used to broaden the phylogenetic diversity and are also part of an ongoing study about the molecular diversity in the genus Laboulbenia.Thalli of Laboulbeniales were removed from their host at the point of attachment and mounted in permanent slides using the double-coverslip technique as described by Liu et al. (2020), with one modification: thalli were placed in a droplet of 1:1 Hoyer's medium:glycerin mixture instead of pure Hoyer's medium, because our Hoyer's medium dried quickly.Mounted thalli were viewed at 200-1000 × magnification under an Olympus BH-2 microscope (Olympus, Center Valley, PA, USA).Images of thalli were made with a Nikon DS-Fi3 microscope camera mounted on an Eclipse Ni-U compound microscope (Nikon, Nelville, NY, USA), equipped with differential interference contrast optics, and processed using NIS-Elements BR 5.0.03 imaging software (Nikon).Photos were enhanced and the background was removed using cutout.pro (https:// www.cutout.pro/) and figures were assembled in PowerPoint v.2306 (Microsoft, Redmont, WA, USA).
Studied slides are deposited at the Herbarium Universitatis Gandavensis (GENT) and Meise Botanic Garden Herbarium (BR).Hosts are stored in the Taxon Expeditions collection (TXEX) and the personal collection of Oscar Vorst.

DNA extraction, PCR amplification, and sequencing
DNA extractions were done using the REPLI-g Single Cell Kit (Qiagen, Stanford, CA, USA).All steps were performed wearing disposable latex gloves.Thalli of Laboulbeniales were removed from their host using a hypodermic needle, which was inserted into a glass syringe for holdfast, under a dissecting microscope.Removed thalli were placed in a droplet of glycerin on a microscope slide.The thalli were cut into multiple smaller pieces with the sharp tip of the needle.These pieces were placed in 0.2-ml PCR tubes with 4 µl of phosphate-buffered saline (PBS).Next, we followed the instructions as indicated in the manufacturer's manual (Qiagen).

Phylogenetic analyses
We used a broad selection of newly generated Laboulbenia sequences, supplemented with 18S and 28S sequences of Laboulbenia spp.and 18S sequences of other genera downloaded from NCBI GenBank.Accession numbers of sequences and additional information about the isolates can be found in Table 1.As outgroup, we used taxa from family Dimorphomycetaceae (Dimeromyces, Nycteromyces, and Polyandromyces) (Goldmann and Weir 2018).
We aligned 18S and 28S sequences by locus with the G-INS-i strategy and ITS with the E-INS-i strategy using the online version 7 of MAFFT (Katoh et al. 2005(Katoh et al. , 2019;;Kuraku et al. 2013).Sequences were manually trimmed using BioEdit Sequence Alignment Editor version 7.2.6 (Hall 1999) and combined in SequenceMatrix 1.9 (Vaidya et al. 2011) to construct one concatenated dataset (18S-ITS-28S).The final dataset included five partitions: 18S, the ITS1 and ITS2 spacer regions, the 5.8S gene, and 28S.Models for nucleotide substitution were selected for each partition with ModelFinder (Kalyaanamoorthy et al. 2017) according to the corrected Akaike information criterion (AICc).A maximum likelihood (ML) reconstruction was inferred using IQ-TREE (Nguyen et al. 2015) under partitioned models (Chernomor et al. 2016).Ultrafast bootstrapping was performed with 1000 replicates (Hoang et al. 2017).
Bayesian inference was done using MrBayes (Ronquist et al. 2012), available on the CIPRES Science Gateway web portal (Miller et al. 2010).Four Markov chains were run for 80 million generations, sampling every 8000 generations.Our concatenated dataset (18S-ITS-28S) was not partitioned.The analysis was performed using the GTR substitution model, with some sites being invariable and gamma-distributed rate variation across the remaining sites (GTR+I+G) (Abadi et al. 2019).A burn-in of 8000 trees was selected.
The 18S sequence of B. heteroceri (D.Haelew.4197b) shares 98.57% identity with B. ornatus (AW821) and 95.13-98.57%identity with other species of Laboulbenia, with L. collae, L. notiophili, L. pedicellata, and L. thaxteri as the highest ranked ones.To compare these results with the divergence in the 18S region among species of Laboulbenia, we blasted an 18S sequence of L. slackensis, which shares between 100% (L.slackensis) and 95.50% (L.cf.dorstii) identity.The ITS sequence of B. heteroceri is highly divergent compared to those of Laboulbenia species, with a query cover of only 21-37% (which roughly corresponds to the conserved 5.8S region and the beginning of ITS2).It shares between 91.43 and 96.89% with other species of Laboulbenia, with L. clivinalis, L. littoralis, L. pedicellata, and L. slackensis as the highest ranked.The 28S sequence of B. heteroceri shares between 81.74 and 87.44% identity with other species of Laboulbenia, with L. benjaminii, L. slackensis, and L. pedicellata as the highest ranked.To compare these results with the divergence in the 28S region among species of Laboulbenia, we blasted a 28S sequence of L. slackensis, which shares between 100% (L.slackensis) and 83.92% (L.oioveliicola) identity.If not flattened and blackened, then not distinguishable from the surrounding cells and surrounded by proliferating cells.Appendages variable: simple to highly branched; short to long; with or without blackened septa; hyaline, colored, or blackened; typically consisting of an inner, usually fertile, appendage and an outer, sterile, appendage.Antheridia terminal or lateral simple phialides and then usually born on the inner appendage, rarely sessile.Solitary perithecium at least free at the ventral side, with four tiers, typically with four outer wall cells of usually unequal to rarely equal size in each vertical tier or with two tiers with four unequal wall cells and two tiers with three unequal wall cells.Edited from descriptions by Tavares (1985), Majewski (1994), and Santamaria and Pedersen (2021).
Because we reinstated L. heteroceri Thaxt., B. heteroceri based on Misgomyces heteroceri needed a replacement name in Laboulbenia: Laboulbenia mairei.Laboulbenia mairei is positioned on a long branch in our phylogenetic tree (Fig. 2).This is mainly due to the divergence in sequences of the ITS and 28S regions between L. mairei and other species in the genus.For L. heteroceri, however, only one sequence is available: that of the conserved 18S region (Goldmann and Weir 2018).This explains the large evolutionary distance between L. heteroceri and L. mairei on the one hand and the shorter distances between L. heteroceri and closely related species of Laboulbenia on the other hand.Also, a few other species are found on relatively long branches in our phylogenetic reconstruction: Laboulbenia bicornis, L. bruchii, and L. fasciculata.This can, in part, be attributed to taxon sampling error.Indeed, only 14 of the 667 currently accepted species of Laboulbenia (Haelewaters et al. 2023) are included in our phylogenetic analysis.A revision of this genus based on molecular phylogenetic data, with increased sampling, both taxonomically (more taxa) and geographically (from a wide geographic coverage), is desirable and may result in the disintegration of Laboulbenia in meaningful taxonomic groups (sections, subgenera, or different genera).We conclude that the proposed transfer of Botryandromyces species to Laboulbenia is on par with our current morphological and molecular knowledge of the genus.
A few considerations arise after including these species in the genus Laboulbenia.The difference in perithecial outer wall cells is striking.The number of outer wall cells in each tier is a commonly used and reliable character to delineate and identify genera of Laboulbeniales (Tavares 1985;Majewski 1994;De Kesel et al. 2020;Santamaria and Pedersen 2021).The difference in the number of these cells was one of the main reasons why Tavares and Majewski (1976) erected Botryandromyces.Tavares (1985) proposed that it "was undoubtedly derived from a more typical arrangement of four cells in each row."In addition, Tavares (1985) erected Dixomyces and Scalenomyces to accommodate a few other species, based on the number of outer wall cells in each tier and Note: Accession numbers of sequences generated during this study are in boldface characteristics of the appendages and the receptacle.Eventually, Rossi and Santamaria (2008) synonymized Scalenomyces with Laboulbenia, as the morphology of their newly described L. magrinii was similar to S. endogaea.Both species are known from endogean ground beetles (Coleoptera, Carabidae).Whether their specific morphology is an adaptation to their host, their host's ecology, the environment, due to random genetic drift, or other factors, is unknown.Similar thoughts can be made regarding morphological changes of L. heteroceri and L. mairei compared to phylogenetically related species (Fig. 2).Both species are found on Heteroceridae, while L. clivinalis and L. slackensis are found on Carabidae, like most species of Laboulbenia are.Our phylogeny provides evidence for a host shift, which might have driven the observed changes in morphology.
Remarkably, the lower receptacle of L. mairei often shows secondary divisions, while L. heteroceri consistently has a two-celled lower receptacle (Fig. 1) (Thaxter 1912;Maire 1920;Tavares and Majewski 1976).The receptacle of L. heteroceri resembles that of a typical species of Laboulbenia, which was already acknowledged by Thaxter (1912).Laboulbenia mairei is not the only species in the genus that has more than two cells in the lower receptacle.Laboulbenia dohrni and L. partita also have this peculiar organization of the lower receptacle but differ in other characteristics, e.g., they have a typical blackened insertion cell (Thaxter 1914;Spegazzini 1915;Tavares 1985).In L. mairei, the number of cells in the lower receptacle is variable and may depend on the position of thalli on the host integument or thallus age (Majewski 1994;De Kesel 2009;Santamaria and Pedersen 2021).Thalli of L. mairei with a typical Laboulbenia receptacle are illustrated by Majewski (1994).
The upper receptacle of L. heteroceri and L. mairei is reminiscent of the ones from species in the Laboulbenia luxurians group as defined by Tavares (1985).Similar to the species in this group, the height of their cells IV and V is equal; the vertical septum between these cells reaches cell III.Laboulbenia clivinalis and L. slackensis also belong to this group and form a well-supported clade with L. heteroceri and L. mairei in our phylogeny (Fig. 2).In addition, most species of this group (e.g., L. clivinalis and L. slackensis) have a blackened septum between the basal and suprabasal cells of their outer appendage.Laboulbenia heteroceri also has a blackened septum in this position (Fig. 1).Species in this group are commonly found on hosts that live in humid environments like sandy or muddy river banks, seashores, and wet grasslands.Carabidae (hosts for L. clivinalis, L. pedicellata, and L. slackensis) and Heteroceridae (hosts for L. heteroceri and L. mairei) are often found together in these environments (Holeski and Graves 1978;A. De Kesel, pers. obs.).This shared habitat preference makes host shifts of Laboulbeniales between those two families likely (Rossi 2011;De Kesel and Haelewaters 2014).
The morphology of Dixomyces clivinae and D. pallescens is similar to that of L. heteroceri and L. mairei.They were transferred from Laboulbenia in which they were originally described by Thaxter (1896Thaxter ( , 1908) ) to Dixomyces by Tavares (1985).Both Dixomyces species were described from carabid beetles, suggesting that the adjusted morphology was already present on carabid hosts.No sequence data of these species are available, but we hypothesize that D. clivinae and D. pallescens may also need to be transferred back to Laboulbenia.Thaxter (1912) reported morphological differences between the holotype of L. heteroceri from Argentina and thalli found on beetles collected in KS, USA.Several studies also reported differences in length between the holotype of L. mairei from Algeria and specimens from Europe.Reasons behind these differences are unknown but they have been attributed to either inaccurate measurements or environmental differences (Scheloske 1969;Tavares and Majewski 1976;Majewski 1994;Weir 1994).Given that cryptic diversity in Laboulbeniales is proven using molecular data (Haelewaters et al. 2018(Haelewaters et al. , 2019) ) and that both L. heteroceri and L. mairei are reported from different genera of Heteroceridae, it is only a matter of time and effort to confirm or reject whether there are multiple cryptic species hidden under these two names.Host specimens should be freshly collected to sequence species of Dixomyces and Scalenomyces, L. heteroceri, and L. mairei and resolve these outstanding taxonomic issues.

Conclusions
Based on molecular phylogenetic data, we synonymized Botryandromyces with Laboulbenia and emended the description of Laboulbenia to include that (1) the perithecial outer wall can have either four cells in each of the four vertical tiers, or two tiers with four cells and two tiers with three cells, and (2) the lower receptacle can be either two-celled or multi-celled.The species formerly placed in Botryandromyces (now known as Laboulbenia heteroceri and L. mairei) are morphologically similar to species of the Laboulbenia luxurians species group.Their hosts also occupy the same habitats, which increases the chance of a host shift.Although there is a major difference in the morphology of the perithecium, the equal size of cells IV and V and the presence of a blackened septum in the outer appendage correspond to their phylogenetic position inside this species group.Future research should focus on adding sequence data for morphologically described genera related to Laboulbenia, host shifts within the L. luxurians species group, and cryptic diversity in L. heteroceri and L. mairei.

Fig. 2
Fig. 2 Partial phylogeny of Laboulbeniales based on a concatenated 18S-ITS-28S dataset, with the genus Laboulbenia indicated in the red box.Laboulbenia heteroceri and L. mairei (in bold) are retrieved

Table 1
Tavares and Majewski (1976)d in this studyDiscussionHere, we show that the genus Laboulbenia is paraphyletic if B. heteroceri and B. ornatus are retained in a separate genus.Therefore, we propose to synonymize Botryandromyces with Laboulbenia and to transfer B. ornatus and B. heteroceri to Laboulbenia as L. heteroceri and L. mairei, respectively.When Botryandromyces was erected,Tavares and Majewski (1976)combined two species in the genus, as B. heteroceratis (Thaxt.)I.I.Tav.& T. Majewski and B. heteroceri (Maire) I.I.Tav.& T. Majewski.