Phylogenetic analyses allow species-level recognition of Leptographium wageneri varieties that cause black stain root disease of conifers in western North America

Leptographium wageneri is a native fungal pathogen in western North America that causes black stain root disease (BSRD) of conifers. Three host-specialized varieties of this pathogen were previously described: L. wageneri var. wageneri on pinyon pines (Pinus monophylla and P. edulis); L. wageneri var. ponderosum, primarily on hard pines (e.g., P. ponderosa, P. jeffreyi); and L. wageneri var. pseudotsugae on Douglas-fir (Pseudotsuga menziesii). Morphological, physiological, and ecological differences among the three pathogen varieties have been previously determined; however, DNA-based characterization and analyses are needed to determine the genetic relationships among these varieties. The objective of this study was to use DNA sequences of 10 gene regions to assess phylogenetic relationships among L. wageneri isolates collected from different hosts. The multigene phylogenetic analyses, based on maximum likelihood and Bayesian inference, strongly supported species-level separation of the three L. wageneri varieties. These results, in conjunction with previously established phenotypic differences, support the elevation of L. wageneri var. ponderosum and L. wageneri var. pseudotsugae to the species level as L. ponderosum comb. nov. and L. pseudotsugae comb. nov., respectively, while maintaining L. wageneri var. wageneri as Leptographium wageneri. Characterization of the three Leptographium species, each with distinct host ranges, provides a baseline to further understand the ecological interactions and evolutionary relationships of these forest pathogens, which informs management of black stain root disease.


Introduction
Black stain root disease (BSRD) is a vascular wilt disease of conifers that is damaging to forests in western North America, especially in the western United States (Washington, Oregon, and California), and southwestern Canada (British Columbia) (Lockman and Kearns, 2016).BSRD was first found producing a dark sapwood stain in hard pines [Jeffrey pine (Pinus jeffreyi) and ponderosa pine (P.ponderosa)] and single-leaf pinyon (P.monophylla) in 1938 and 1941, respectively, in California (Wagener and Mielke, 1961).Although BSRD was not well recognized in the Pacific Northwest of United States before 1969, it has emerged as one of the five most-damaging root diseases in western forests of the United States (Hadfield et al., 1986;Cobb, 1988).The disease is caused by a native, insect-vectored ascomycete fungal pathogen, originally named Verticladiella wageneri W.B Kendr.(Kendrick, 1962) and later transferred to Leptographium wageneri (W.B Kendr.)M.J. Wingf.(Wingfield, 1985).
Leptographium wageneri can be introduced to hosts in two ways: 1) via root contacts or grafts between healthy and infected trees, or 2) via insect vectors.Wounds may be required for non-root graft infection because the pathogen's hyphae cannot penetrate bark or break down cellulose (Hessburg, 1984), though small roots may be infected directly (Cobb, 1988).Once the BSRD fungal pathogen enters the host, it colonizes tracheid cells in sapwood xylem, and dark-brown to purple-black staining starts to appear when the xylem is colonized by hyphae (Figure 1) (Wagener and Mielke, 1961;Hessburg et al., 1995).Other than the characteristic staining, infected hosts often show general wilt disease symptoms, such as tufted needles, needle loss, chlorosis, and/or reduced growth (Supplementary Figure 1), typically followed by tree mortality.Basal resin flow may also occur, and a stress cone crop may be produced in Douglas-fir (Pseudotsuga menziesii).Because the crown symptoms are similar to other root diseases, the best way to confirm BSRD in the field is by verifying the presence of the dark stains in sapwood of roots and/or at the base of the stem (Figure 1) (Hadfield et al., 1986).
The sexual stage of the BSRD pathogen was described as Ceratocystis wageneri Goheen & F.W. Cobb (Goheen and Cobb, 1978;Harrington and Cobb, FIGURE 1 Black streaking of sapwood, which is associated with black stain root disease (caused by Leptographium wageneri varieties s.l.), appears in (A) root collars above ground and (B) underground roots from a single-leaf pinyon (Pinus monophylla).Staining appears above ground level in (C) ponderosa pine (P.ponderosa) and (D) Douglas-fir (Pseudotsuga menziesii).Choi et al. 10.3389/fpls.2023.1286157Frontiers in Plant Science frontiersin.org1987; Zipfel et al., 2005).Perithecia and ascospores were found in an insect gallery in a root of a ponderosa pine in a diseased stand in California (Goheen and Cobb, 1978), but no culture of the teleomorph is available, and subsequent searches in this same location failed to find perithecia of L. wageneri (Harrington, 1988).No further reports of a sexual stage have been made.Further, the sexual stage of other heterothallic Leptographium species are readily produced in culture if opposite mating types are paired (Duong et al., 2016), but pairing of isolates from ponderosa pine and other hosts failed to produce perithecia (Harrington and Cobb, 1986).
In culture, the BSRD pathogens produce stalked conidiophores with sticky drops of conidia, which appear suitable for insect dispersal, and such conidia are presumably formed in galleries produced by the insect vectors.When young, adult insects emerge from infected roots, their exoskeleton may be contaminated by conidia.These conidia can be subsequently introduced to roots of healthy trees when these adult insect vectors bore into roots for egg laying or maturation feeding (Hessburg et al., 1995;Ferguson, 2009).Root-feeding insects, such as bark beetles (Hylastes nigrinus and H. macer) and weevils (Pissodes fasciatus and Stereminius carinatus) (Coleoptera: Curculionidae), are the primary vectors involved in spread of the BSRD pathogen (Hessburg et al., 1995;Ferguson, 2009).Among these potential vectors, the conifer seedling weevil, S. carinatus, is considered an important agent for within stand spread because it is flightless (Hadfield et al., 1986;Ferguson, 2009).
We hypothesized that the three varieties of L. wageneri are phylogenetically distinct and genetic differences are sufficient to warrant their recognition at the species level.Thorough phylogenetic analyses of L. wageneri requires ample isolates and accurate sequences from a sufficient number of diverse loci.The overall objective of this study was to use DNA sequences derived from multiple loci to genetically characterize the three L. wageneri varieties collected from western North America and determine their phylogenetic relationships.

Study sites and sample collection
During 2019-2021, a total of 24 BSRD samples were collected from various National Forests where BSRD was previously reported in Idaho, Oregon, and California (Supplementary Table 1).Samples were collected from three single-leaf pinyon pines, six ponderosa pines, two Jeffrey pines, and nine Douglas-fir trees.The number of samples collected from each host species varied depending on site accessibility, disease incidence, and previous records.Trees with BSRD crown symptoms were examined for black or dark-brown streaks in the sapwood near the roots and/or basal stems.For each tree, three to five wood samples (ca. 2 x 7 x 1 cm) were collected directly from the stained tissue and stored in a sealed plastic bag assigned to each tree with paper towels dampened in sterile water.Samples were kept cool by storage on ice or refrigeration before fungal isolation was performed.Precise locations were collected using a GPSMAP 64s GPS receiver (Garmin, Olathe, KS, USA).Diameter at breast height (DBH), health status (dead/live/declining and crown status), and other noticeable field conditions (e.g., elevation, vegetation, stand conditions) were also recorded.Twothirds of the collected samples were sent to USDA Forest Service -Pacific Northwest Research Station, Forest Heath Laboratory (Corvallis, OR) for isolation within 48 hours of collection, and one-third of samples were isolated on-site.

Sample isolation
For surface disinfestation, the wood samples were cut into smaller pieces containing mostly stained areas (Supplementary Figure 2).The wood pieces were submerged and soaked in 95% ethanol for 30 seconds, and transferred to 10% commercial bleach (0.05% sodium hypochlorite) for a 3-minute soak.The samples were then rinsed four times in a separate beaker with sterile, deionized H 2 O. Damp surfaces were dried with sterile tissues, and the samples were cut into pieces (ca. 10 x 2.5 x 2.5 mm) for culturing.Approximately ten pieces were embedded in a Petri dish containing a selective medium with malt extract agar (MEA) amended with cycloheximide and streptomycin sulfate (CSMA) (one-fourth strength MEA media: 0.75% malt, 0.75% dextrose, 0.25% peptone, and 1.5% agar with 200 ppm cycloheximide and 200 ppm streptomycin sulfate) (Harrington, 1992).Three to five Petri dishes were established for samples from each tree.During incubation, samples were stored at 15°C, checked daily, and discarded if contamination occurred.After 2-3 weeks, when L. wageneri conidiophores were observed on the wood piece, drops of conidia were transferred from the conidiophores onto MEA medium (3% malt, 3% dextrose, 1% peptone, 1.5% agar).After 1-2 weeks of incubation, when fungal hyphae started to grow from the transferred conidia, hyphal tips were transferred onto fresh MEA medium with the use of a flame-sterilized dissecting pin.In addition, 15 isolates were studied from collections made in the 1980s (Harrington and Cobb, 1986;Harrington and Cobb, 1987).These previously collected isolates were used in descriptions of the varieties, including ex-types (CAP-19 and CAD-18), and they are now archived at the USDA Forest Service -Pacific Northwest Research Station, Forest Heath Laboratory (Corvallis, OR) (Supplementary Table 1).

Phylogenetic analyses
The resulting sequences of each L. wageneri s.l.isolate consisted of forward and reverse ABI sequencer data files (.ab1).Geneious Prime version 2022.2.2 (https://www.geneious.com/)was used to pair and edit the sequences following IUPAC (International Union of Pure and Applied Chemistry) codes.The 39 paired sequences for each of the 10 loci were generated as consensus sequences, and they were aligned using Clustal Omega 1.2.2.(Sievers et al., 2011) with default parameters and trimmed to the same length for comparison.
The sequences of the loci were analyzed in four parts: Analysis 1) individual loci; Analysis 2) concatenated sequences of two (CAL + ACT*) and three (CAL + ACT* + TEF-1a) loci; Analysis 3) concatenated sequences of 10 loci; and Analysis 4) concatenated sequences of two loci (CAL + ACT*) analyzed with multiple outgroups (Supplementary Table 4).Leptographium douglasii was used as an outgroup for single outgroup analyses (Analysis 1, 2, and 3).The sequences of ACT**, GPD, and CHS of L. douglasii were not available in GenBank, therefore, these sequences were generated using the same primer sets and protocols as for L. wageneri s.l.For Analysis 2, the two-and three-loci combinations were selected because each of the loci have relatively large numbers of parsimony-informative sites.Parsimony-informative sites were identified using IQ-TREE (Nguyen et al., 2014).For Analysis 4, six outgroup taxa (L.douglasii, L. rhodanense, L. gracile, L. castellanum, G. alacris, and G. serpens) were selected based on the phylogenetic tree from de Beer et al. ( 2022) and availability of sequences in GenBank.
A partition homogeneity test was performed using PAUP* version 4.0a169 (Swofford and Sullivan, 2003) to determine whether the sequence data for the 10 loci were congruent for concatenation.A p-value over 0.05 indicates that partitions can be combined, and the sequence data of the 10 loci resulted in a p-value = 0.08.The best-fit substitution model for each locus was selected by Bayesian Information Criterion (BIC) calculated in ModelFinder in IQ-TREE (Nguyen et al., 2014;Kalyaanamoorthy et al., 2017) (Supplementary Table 5).Phylogenetic analyses were implemented using maximum likelihood and Bayesian inference.Maximum likelihood trees were constructed using IQ-TREE (Nguyen et al., 2014;Hoang et al., 2018) with 1,000 bootstrap pseudo-replications.Maximum likelihood trees were imported and visualized using Geneious Prime, with bootstrap support values (BS) over 50%.MrBayes (Huelsenbeck and Ronquist, 2001) plugin in Geneious Prime was used for the Bayesian inference.Because Tamura-Nei (TN) models could not be implemented using MrBayes in Geneious Prime, the next best models were selected.Bayesian trees were generated using 1,100,000 generations with the four heated-chains setting in Markov Chain Monte Carlo (MCMC).The first 100,000 chains were discarded as burn-in, posterior probability (PP) for each dataset was calculated, and a consensus tree was produced with PP over 0.80 illustrated on the branches.

Individual loci
No parsimony-informative sites were found among the three L. wageneri varieties at the LSU, TUB, and RPB2 loci.The sequences of LSU and RPB2 were identical among the 39 isolates, and only a single isolate of L. wageneri var.ponderosum showed a single base substitution for TUB.

Concatenation of two and three loci
Analyses using the combined data of the most polymorphic loci clearly separated the L. wageneri varieties.The three varieties were well separated in the tree generated with the two loci, CAL and ACT* (Supplementary Figure 9).The clade containing the L. wageneri var.ponderosum isolates was well-supported (0.97 PP and 88 BS), and a L. wageneri var.pseudotsugae clade was also supported (0.93 PP and 56 BS).The three varieties were also clearly separated based on concatenated sequences of CAL, ACT*, and TEF-1a (Supplementary Figure 10).In the phylogenetic tree based on the concatenated sequences of these three loci, each variety was well-supported: L. wageneri var.wageneri at 1 PP and 80 BS, var.ponderosum at 0.96 PP and 83 BS, and var.pseudotsugae at 0.96 PP and 63 BS.

Concatenation of two loci with multiple outgroup taxa
For two loci, sequences were publicly available for six related Leptographium and Grosmannia species.A phylogenetic tree was generated based on concatenated sequences of CAL and ACT* of the L. wageneri s.l.isolates with those of L. douglasii, L. rhodanense, L. gracile, L. castellanum, G. alacris, and G. serpens (Figure 4).In this tree, each L. wageneri variety was well-supported: L. wageneri var.wageneri with 0.84 PP and 73 BS, L. wageneri var.ponderosum with 1 PP and 92 BS, and L. wageneri var.pseudotsugae with 0.96 PP and 61 BS.

Taxonomy
Clear phylogenetic distinction as well as several minor morphological and physiological characters (Harrington and Cobb, 1986;Harrington and Cobb, 1987) (Table 2) provide strong justification for separating these three host-specialized, Leptographium pathogens as species.The low level of genetic Leptographium ponderosum (formerly L. wageneri var.ponderosum) indicates that isolates are identical within each species.Underlined isolates (CAP-19 and CAD-18) are ex-type isolate of each species.Choi et al. 10.3389/fpls.2023.1286157Frontiers in Plant Science frontiersin.orgpigmentation of mycelia and conidiophore features (Harrington and Cobb, 1986;Harrington and Cobb, 1987) (Table 2).Subsequently, isozyme variation and RAPD markers supported distinction of the three varieties (Otrosina and Cobb, 1987;Zambino and Harrington, 1989;Zambino and Harrington, 1992;Witthuhn et al., 1997).Phylogenetic analyses based on rDNA (ITS2 and LSU) sequences did not provide clear separation of the three L. wageneri varieties (Jacobs et al., 2001).The present study is the first to demonstrate robust separation of the three L. wageneri varieties using multiple, non-rDNA loci.Although Bennett et al. (2021) demonstrated strong separation between two varieties (L.wageneri var.wageneri and L. wageneri var.pseudotsugae) based on over 100,000 binary SNPs, their analyses included isolates from limited geographic areas (e.g., L. wageneri var.pseudotsugae isolates from Oregon and L. wageneri var.wageneri isolates from California).And, sequences of the three most parsimony informative loci (e.g., CAL, ACT*, and TEF-1a) from L. wageneri var.pseudotsugae isolates were identical to those within the genomic sequences of L. wageneri var.pseudotsugae isolates from Oregon used in the study of Bennett et al. (2021).The clear phylogenetic separation warrants species designation.Phylogenetic separation of L. wageneri, L. ponderosum and L. pseudotsugae was better resolved when multi-locus analyses were used rather than a single locus.In general, these three Leptographium species showed very little genetic variation, and most of the loci showed only minor sequence differences.When analyzing with a single locus, these three species tended to be separated differently depending on the locus being analyzed.For example, for MAT1-1-3 and TEF-1a, L. wageneri was distinct, while for ACT** (ACT512F & ACT783R) and GPD, L. ponderosum was distinct, and for CHS and CAL, L. pseudotsugae was distinct.The very limited genetic variation within each of these three Leptographium species, pathogens generally believed to be native to western North America, strongly suggests that they are predominantly asexual.
In most cases, ACT* (Lepact-F & Lepact-R) could be used to identify the three Leptographium species.ACT*-based separation of three species was generally well-supported, with the exception of one isolate, IDD-1 (L.pseudotsugae), which was placed in the L. wageneri clade.The TUB locus has been used for identification at the species level; however, TUB-based identification of L. ponderosum is distinguished by only a single-base substitution.
Because the majority of parsimony-informative sites were found in the CAL, ACT* (Lepact-F & Lepact-R), and TEF-1a loci, combinations among these loci were found to be most useful in separating the Leptographium species.Of these three loci, CAL had the most parsimony-informative sites.However, with the CALbased tree, only L. pseudotsugae was separated.A phylogeny based on CAL with ACT* (Lepact-F & Lepact-R), which had the next most parsimony-informative sites, resulted in divergence and separation of three species.In the phylogenetic analysis of three combined-loci, CAL, ACT* (Lepact-F & Lepact-R), and TEF-1a, the separation of all three Leptographium species was clear and strongly supported.These results suggest that cumulative phylogenetic information provides more precise divergence with stronger support values for examining evolutionary relationships within and among the species (Taylor et al., 2000).
Distinguishing different species of closely related fungi is a complex process that considers various factors including morphology, biological reproduction, biochemical properties, ecological behaviors, genetic differences, and other factors (Mayr, 1940;Hennig, 1965;Hawksworth et al., 1996;Harrington and Rizzo, 1999;Taylor et al., 2000).In the case of L. wageneri, the three host-specialized pathogens (Harrington and Cobb, 1984) were previously described as varieties instead of species, largely because phylogenetic data were lacking to distinguish these taxa.However, differences among the three L. wageneri varieties were previously noted in terms of minor morphological characteristics, isozyme variations, RAPD profiles, and ecological habitat (host species) (Harrington and Cobb, 1986;Harrington and Cobb, 1987;Otrosina and Cobb, 1987;Zambino and Harrington, 1989;Zambino and Harrington, 1992;Witthuhn et al., 1997).When recognizing new species within a genus, multiple genes have been used in phylogenetic analyses to allow for concordance (Taylor et al., 2000), and several new Leptographium species have been described using multi-locus datasets (Kim et al., 2004;Lee et al., 2005; Jacobs et al., 2006).In this study, the three varieties of L. wageneri were clearly separated phylogenetically on the basis of concatenated sequences of 10 loci and on the two-loci dataset with multiple outgroup taxa.Thus, solid genetic evidence is provided that the three L. wageneri varieties should be elevated to species status.Recognition of the L. wageneri varieties as three separate Leptographium species will help to understand ecological interactions and evolutionary history of these forest pathogens, which can contribute to improved management of the black stain root disease pathosystem.
growth on potato dextrose agar (PDA) at 18°C.Colors are from Ridgway (1912).b On colonized PDA plugs transferred to water agar and incubated for 10 days at 18°C.c Width of stipe apex minus the width of the middle of the uppermost stipe cell.Range of means of 15 isolates, five conidiospores measures per isolate.d After 7 days growth on PDA.