Discovery of novel treponemes associated with pododermatitis in elk (Cervus canadensis)

ABSTRACT Pododermatitis, also known as treponeme-associated hoof disease (TAHD), presents a significant challenge to elk (Cervus canadensis) populations in the northwestern USA, with Treponema spp. consistently implicated in the lesion development. However, identifying species-specific Treponema strains from these lesions is hindered by its culture recalcitrance and limited genomic information. This study utilized shotgun sequencing, in silico genome reconstruction, and comparative genomics as a culture-independent approach to identify metagenome-assembled Treponema genomes (MATGs) from skin scraping samples collected from captive elk experimentally challenged with TAHD. The genomic analysis revealed 10 new MATGs, with 6 representing novel genomospecies associated with pododermatitis in elk and 4 corresponding to previously identified species—Treponema pedis and Treponema phagedenis. Importantly, genomic signatures of novel genomospecies identified in this study were consistently detected in biopsy samples of free-ranging elk diagnosed with TAHD, indicating a potential etiologic association. Comparative metabolic profiling of the MATGs against other Treponema genomes showed a distinct metabolic profile, suggesting potential host adaptation or geographic uniqueness of these newly identified genomospecies. The discovery of novel Treponema genomospecies enhances our understanding of the pathogenesis of pododermatitis and lays the foundation for the development of improved molecular surveillance tools to monitor and manage the disease in free-ranging elk. IMPORTANCE Treponema spp. play an important role in the development of pododermatitis in free-ranging elk; however, the species-specific detection of Treponema from pododermatitis lesions is challenging due to culture recalcitrance and limited genomic information. The study utilized shotgun sequencing and in silico genome reconstruction to identify novel Treponema genomospecies from elk with pododermatitis. The discovery of the novel Treponema species opens new avenues to develop molecular diagnostic and epidemiologic tools for the surveillance of pododermatitis in elk. These findings significantly enhance our understanding of the genomic landscape of the Treponemataceae consortium while offering valuable insights into the etiology and pathogenesis of emerging pododermatitis in elk populations.

the consistent presence of invasive spiral-shaped bacteria typical of spirochetes (1).Characteristic histologic changes within the elk pododermatitis along with the presence of Treponema spp.exhibit similarities to bovine digital dermatitis (BDD) and contagious ovine digital dermatitis (CODD) in domestic livestock (1,4).The consistent identification of lesion-associated Treponema spp., a member of the family Treponemata ceae (formerly Spirochaetaceae), led to the designation of treponeme-associated hoof disease (TAHD) of elk (1,4,5).Although the presence of Treponema spp. is characteristic of BDD, CODD, and TAHD, and is likely the primary agent associated with these diseases, there is speculation that each of these diseases may involve multiple bacterial species (4,6,7).
In one study, Treponema spp.representing the three classical BDD phylotypes [Treponema medium/Treponema vincentii-like, Treponema phagedenis-like, and Treponema denticola/Treponema putidum-like (8)] were identified by PCR on isolates cultured using specialized techniques from a small number of samples from elk hooves (4).However, the isolation of fastidious Treponema spp. in pure culture is challenging because hoof samples are frequently tainted with soil, feces, and various other environmental microorganisms.Moreover, the investigation of diseases in free-ranging wildlife such as elk is further complicated by difficulties in obtaining fresh tissues that maintain the viability of Treponema spp.from a sizeable sample group.Therefore, most epidemiologic investigations of hoof disease in elk have relied on culture-independent techniques, such as 16S rRNA amplicon sequencing, to detect Treponema spp.(2,3,9).In our recent report, we discovered that in addition to the well-known treponemal phylotypes, partial sequencing of the 16S rRNA amplicons identified distinct signatures of previously uncultured Treponemataceae, including Treponema sp.clone 19 (PT19), from hooves of free-ranging elk diagnosed with TAHD (2).These novel signatures of uncultured spirochetes were also detected in early lesions, and before the development of graded TAHD lesions, in experimentally challenged captive elk (3).Signatures of uncultured PT19 were previously detected from BDD lesions in Norwegian cattle (10).In the last decade, other novel uncultured Treponema spp.have also been detected within BDD lesions from domestic cattle (6,11,12).Collectively, these findings have led to the hypothesis that PT19, and perhaps other novel spirochetes, are likely associated with TAHD in free-ranging elk (3).A few complete genome sequences of Treponema spp.cultured from BDD lesions from cattle are now publicly available (13,14); however, the genome sequences of Treponema spp.detected within TAHD lesions in elk are not yet available.Therefore, the identity of these uncultured spirochetes remains obscure.This lack of genomic information poses significant limitations to etiological and epidemiological investigations of pododermatitis in elk.
Genome reconstruction approaches hold promise for investigating wildlife disea ses where the etiology of disease is poorly understood, and sample quality is often compromised for routine microbiological processing especially when infectious agents, such as Treponema, are difficult to culture.Construction and comprehensive characteriza tion of genomes of microbiologically fastidious Treponema spp.detected in TAHD lesions (2,3) is fundamental to epidemiologic investigations, the development of improved tests for reliable diagnosis, and routine monitoring and surveillance of TAHD.Shotgun sequencing coupled with in silico genome reconstruction offers a viable alternative for novel pathogen discovery to improve our understanding of the etiology and pathogene sis of pododermatitis in elk.
The objective of this study was to construct and characterize the metagenomeassembled genomes (MAGs) of fastidious and culture-recalcitrant Treponema spp.from elk with TAHD lesions that were previously suspected to carry novel Treponema spp.signatures based on 16S rRNA amplicon sequencing (2,3).To accomplish this, we developed a complete bioinformatics analysis pipeline to construct MAGs of culturerecalcitrant Treponema from shotgun metagenome data obtained from hooves of elk with TAHD.We then performed a comprehensive comparative genomics analysis of MAGs to identify novel genomospecies of Treponemataceae involved in pododermatitis in elk.We also demonstrate that multiple Treponema spp., including the novel genomo species identified in this study, are indeed associated with pododermatitis in free-rang ing elk.

Elk hoof samples for shotgun sequencing
Nine elk hoof samples including seven biopsies from free-ranging elk and two skin scrapings from experimentally challenged captive elk collected and archived from previous studies were utilized in this study (2,3).Biopsies were collected, post-mortem, from elk that were harvested or culled for management purposes by state wildlife agencies and, thus, were exempt from Institutional Animal Care and Use Commit tee review.These samples were selected for shotgun sequencing because 16S rRNA sequencing revealed that each had tested positive for the presence of unclassified Treponemataceae signatures, particularly PT19 (2, 3), raising the possibility that unique Treponema spp.associated with early lesions were likely present.The biopsy samples originated from the IDS of hooves of free-ranging elk previously submitted to the Washington Animal Disease Diagnostic Laboratory (Pullman, WA) for TAHD surveillance.Samples were from elk with abnormal hooves in areas where TAHD occurs in Washing ton, Idaho, and California (2).Multiple feet from each elk were examined histologically and TAHD was confirmed in every case; however, biopsies did not originate from a TAHD-confirmed foot in every case (Table S1).Biopsies were placed in Allprotect Tissue Reagent (Qiagen, Germantown, Maryland, USA) and frozen at -80 C until processed for genomic DNA (gDNA) extraction as previously described (2).The two skin scrapings selected for the current study originated from two captive experimental research elk that were sampled 98 days following the initial challenge with inoculum prepared from the hooves of TAHD-affected free-ranging elk from Washington as previously reported (3).Scrapings were collected from the IDS of feet with mild gross changes that did not meet the criteria for a graded lesion; however, each experimental elk exhibited an apparent grade I lesion on another foot at the time of sampling.The lesions on the experimental elk progressed and TAHD was confirmed by histologic examination of lesion biopsies (3) (Table S1).Using 16S rRNA amplicon sequencing, the unclassified Treponemataceae signatures, particularly PT19, were detected in both skin scrapings, which corroborated with the detection of these signatures in biopsies from free-ranging elk (2).However, the unclassified Treponemataceae signatures were not detected in these elk before the experimental challenge or in the control elk used in that study (3).Skin scrapings from experimental elk were placed in sterile vials and frozen at −80°C until processed for gDNA extraction as previously described (3).

Shotgun metagenome sequencing of elk hoof DNA
An aliquot of gDNA was submitted to Novogene (Novogene Corporation Inc., Sacra mento, California, USA) for shotgun sequencing using the Illumina platform with the objective of obtaining microbial reads for reconstructing potentially novel MAGs of Treponema spp.(hereafter referred to as MATGs) likely associated with early TAHD lesions.The sequencing libraries were constructed by random shearing of gDNA using NEBNext Ultra DNA Library Prep Kit (Illumina, USA), resulting in the generation of fragments with an approximate length of 350 base pairs (bp) with dA-tailing and NEBNext adapter ligation.Base calling was performed using the Illumina real-time analysis (RTA), thereby generating paired-end sequence reads with 150 bp length at each end.The resulting reads were trimmed, filtered, and deduplicated using the fastp tool with the following parameters "-l 100 --dedup --dup_calc_accuracy 6" (15).Reads with Phred score <25 and a length <150 bp were discarded from downstream processing.Host reads were removed by mapping to the elk reference genome (NCBI assembly accession GCF_019320065.1) using a Bowtie 2 aligner (16,17).Host-subtracted reads were assembled into metagenomes using the Megahit program (v.1.2.9) (18).Follow ing assembly, reads were individually mapped using Burrows-Wheeler Aligner to the metagenomic contigs for each sample (19), and mapping files were generated with SAM tools (20).Contigs ≥1,500 bp were selected for downstream processing using four different binning algorithms, including BinSanity v.0.2.6.1 (21), CONCOCT (22), MaxBin 2.0 v.2.2.4 (23), and MetaBAT 2 v.2.12.1 (24).To ensure that genome bins acquired from the binning method contained non-redundant MATGs, they were individually dereplica ted using the DAS_Tool to obtain quality-filtered consensus MATGs (25).The bin quality and completeness of MATGs were evaluated based on the number of single-copy marker genes using CheckM v.1.0.7 program (26).MATGs with ≥90% completeness and ≤3% contamination were selected to forecast the taxonomic assignment using the taxonomic classification procedure of GTDB-Tk v.0.2.1 (27).MATGs of Treponema species curated from elk samples were annotated using the NCBI PGAP pipeline (28).To estimate the relative abundance of spirochetes, shotgun read sequences for each sample were mapped against the standard Kraken database.NCBI accession numbers were assigned to each mapped sequence followed by the assignment of taxonomy using Kraken 2 and KrakenTools software suite (29)(30)(31) with a confidence score of 0.5.To determine the relative abundance of novel MATG-specific reads in each sample, we performed back mapping of shotgun sequencing reads against each MATG using the script map-bowtie2markduplicates.sh (32).

Core gene phylogeny of MATGs from elk hooves
To identify core gene content, draft and complete genome sequences of 74 non-Trepo nema pallidum Treponema species along with one T. pallidum Str.SamoaD (accession number CP002374.1),as an outgroup, were retrieved from NCBI GenBank database (accessed 18 March 2023) (Table S2).Contamination and completeness of genomes retrieved from the NCBI GenBank database were estimated using CheckM v. 1.1.3 (26).Genomes with completeness >90% and contamination <3% were included for the downstream analysis (33).Single-copy core genes with an e-value of 1.0 e − 15 were identified using best-bidirectional hits through genomic comparison (34,35).The core gene hits were selected with ≥95% sequence identity and ≥85% alignment coverage.Non-recombinant core gene families were included for the construction of a maximum likelihood phylogenetic tree using RAxML v.8 (36) under Gamma + General Time Reversible model (37).To identify recombinant genes, the PhiTest feature included in the PhiPack program was employed (38).

Genomic relatedness, functional profiling, and genomic clustering of MATGs from elk hooves
The taxonomic boundaries and genomic similarities of MATGs constructed from elk hooves were determined by calculating and comparing the average nucleotide identity (ANI) values using the Python package pyani.py(39) and average amino acid iden tity (AAI) using the "aai_wf" function implemented in CompareM program (v.0.1.2) with the default parameters (40).Conserved proteins were identified using a BLASTP match (e-value <1e − 5, sequence identity of >40%, and alignable area of the query sequence >50%) (41).The percentage of conserved proteins (POCP) between two genomes was calculated using the formula [(C1 + C2) / (T1 + T2)] * 100%, where C1 and C2 represent the number of conserved proteins in the two genomes under comparison and T1 and T2 represent the total number of proteins in two genomes being compared (41).A POCP value of ≥50% was considered as a criterion to taxonomically classify genomes within the same genus.ANI and AAI values ≥95% were used to taxonomically classify genomes within the same species (42,43).To measure genomic distance among MATGs along with other publicly available Treponema genomes, amino acid usage, codon usage, and k-mer profile were determined and Euclidean distance was calcula ted by "comparem diss" commands implemented in CompareM program (40,44,45).The weighted Bray-Curtis approach was used to generate hierarchical cluster heatmaps based on Euclidean distance metric as a measure of distance between genomes (46).The clusters of orthologous groups (COG) functional annotation of coding sequences was performed using the Perl script CDD2COG.plv.0.2 (47).The Manhattan distance and average clustering method built into the heatmap2 function of the gplots package of the R program were used to construct the heatmap to examine the distribution of genes in different COG functional categories (48,49).The regions of differences (RODs) between reference Treponema genomes and MATGs were visualized by pairwise comparison using BLASTP searches in the CGView Comparison Tool (CCT) (50).MultiGeneBlast (51) was used to identify homologs of multigene clusters or operons between MATGs and publicly available Treponema genomes.

General genomic features of MATGs obtained from the elk hoof microbiome
Read statistics from shotgun metagenomes for the nine hoof samples analyzed in this study are shown in Table 1.Out of nine hoof sample metagenomes, only the two IDS skin scraping samples (20-01RF98 and 20-04LF98) collected from experimentally challenged elk yielded >10% bacterial reads, leading to successful construction of high-quality MATGs.In contrast, the metagenomes sequenced from biopsy samples yielded <5% bacterial reads and, as a result, these samples failed to yield high-quality MATGs for downstream analysis.The two skin scraping samples yielded a total of 10 high-quality MATGs with completeness of >90% and contamination of ˂3% (Table 2).Consequently, for the downstream analysis, we focused on these 10 MATGs constructed from skin scraping samples 20-01RF98 and 20-04LF98.The length of the 10 reconstructed MATGs ranged between 2.27 Mb (Treponemataceae_phy1_A17) and 2.82 Mb (T.phagedenis_A13) with the GC (guanine and cytosine) percentages ranging between 35% (Treponemata ceae_phy1_A13) and 45.3% (Treponema sp._A13) (Table 2).The Treponema genomes available in the NCBI GenBank database (last accessed on 18 March 2023) were retrieved for comparative genomics analysis of the newly constructed MATGs.The lengths of the Treponema genomes downloaded from the NCBI GenBank database ranged from 1.13 (Treponema paraluiscuniculi Cuniculi A) to 4.05 Mb (Treponema primitia ZAS-2) and the GC percentage ranged from 36.8% (Treponema pectinovorum strain Marseille-1-CSURP6641) to 53.2% (Treponema saccharophilum DSM 2985) (Table S2).

Genomic characterization of MATGs constructed from the elk hoof micro biome
To examine the genetic relatedness between the 10 MATGs constructed in this study and publicly available Treponema genomes, the ANI, AAI, and POCP values were compu ted for all 10 MATGs.Based on the ANI and AAI analyses, the MATGs were grouped into five genogroups each with two MATGs representing the same species (Fig. 1a  and b).When compared with the publicly available reference Treponema genomes, group I MATGs (T.pedis_A13 and T. pedis_A17) and group II MATGs (T.phagedenis_A13    S3) and AAI of <96% (Table S4).Taxonomic validation (41) could be achieved at the genus level for group III MATGs (POCP >50%); however, taxonomical validation could not be achieved at the genus level for group IV or group V MATGs (POCP <50%).The POCP analysis suggests that group III MATGs hold distinct taxonomic positions within the genus Treponema, and group IV and group V MATGs hold distinct taxonomic positions within the family Treponemataceae (Table S5) (41).

Core genome phylogeny
To determine the genomic relatedness of MATGs obtained from the elk hoof microbiome, we compared the core gene content of each MATG with 74 publicly available Treponema genomes, and one T. pallidum reference genome, all with quality criteria of >90% completeness and <5% contamination (Table S6) (33).Comparative genomics analysis identified 88 core genes that are conserved across all compared genomes.These 88 core genes were used to construct a non-recombinant single-copy core gene phylogenomic tree.The results show the existence of eight well-defined monophyletic clades, L1 to L8 (Fig. 2).Two T. phagedenis MATGs constructed from elk hoof metagenomes clustered within clade L1, which is composed of multiple T. phagedenis strains isolated and

Profiling metabolic functions of MATGs derived from elk hooves
Metabolic profiling of bacterial genomes serves as an important adjunct in resolv ing taxonomic ambiguities of the newly discovered genomes (52)(53)(54).For metabolic profiling of MATGs constructed in this study, we determined the COG metabolic functions of all gene content of each MATG and compared them with 75 publicly available Treponema genomes.Hierarchical clustering showed that hoof-derived MATGs constructed in this study were clustered into COG functional groups (Fig. 3) that aligned with the genogroup clusters (Fig. 2) based on the core genome phylogeny.MATGs possessed more than 1,300 functionally active genes (Fig. 3).As expected, T. pallidum (794 functional genes) and T. paraluiscuniculi (797 functional genes) harbored fewer functional genes, indicating overall diminished metabolic capability of these organisms when compared with the other Treponema species.Among all the elk hoof-derived MATGs, Treponemataceae_phy1, Treponemataceae_phy2, and Treponema sp.showed significantly lower COG functional enrichment relative to other non-Treponema pallidum genomes (Fig. 3), indicating metabolic distinction of these MATGs.

Comparative genomics and cluster analysis of MATGs
To further explore genomic distances between Treponema genomes and newly constructed MATGs, we compared codon usage, amino acid usage, and k-mer frequen cies (44,52).The clustering pattern among MATGs based on codon usage (Fig. 4), amino acid usage (Fig. S1), and k-mer frequency analysis (Fig. S2) corroborated with the ANI and AAI analyses results.To identify major genomic differences between elk hoof-derived MATGs and other closely related Treponema genomes, we compared publicly available Treponema genomes sharing the same clade with the elk hoof-derived MATGs using the CGView Comparison Tool (Fig. 2).This analysis identified a few unique genomic regions, referred to as RODs, in elk hoof-derived MATGs.For instance, comparison of T. phagedenis MATGs with the T. phagedenis genomes within clade L2 identified two RODs ranging in size from 14 kbp to 68 kbp (T.phagedenis_A13-ROD 1) and 200 kbp to 240 kbp (T.phagedenis_A13-ROD 2).These RODs are present in T. phagedenis_A13 MATG, but not in T. phagedenis_A17 MATG (Fig. 5).Noteworthy is that these two MATGs (T.phagede nis_A13 and T. phagedenis_A17) were constructed from the same sample and showed only 2% to 6% dissimilarity at the nucleotide level.T. phagedenis_A13-ROD 1 contains approximately 53 genes, whereas T. phagedenis_A13-ROD 2 contains approximately 18 genes that are exclusive to the elk hoof-derived T. phagedenis_A13 MATG.The majority of the genes within these RODs encode proteins associated with flagellar structures, transport functions, and glycosyltransferase activity (Table S7).These RODs are distinct from recently reported bovine-specific gene clusters associated within the T. phagedenis genome ( 14), one encoding machinery for outer membrane protein (OMP) synthesis and another constituting two polycistronic operons for phosphate utilization systems (Fig. 6).The analysis of the two bovine-specific gene clusters revealed that the elk hoof-derived T. phagedenis_A13 MATG possessed the OMP gene cluster while lacking a complete phosphate utilization gene cluster (14).Conversely, both gene clusters are absent in the T. phagedenis_A17 MATG.It is worth noting that both gene clusters were also missing or incomplete in T. phagedenis 4A, a bovine isolate from the USA (13).Therefore, although T. phagedenis_A13 MATG and T. phagedenis_A17 MATG cluster with T. phagedenis isolated from cattle (Fig. 2), carriage of two distinct RODs within T. phagedenis_A13 MATG shows that this MATG is genetically distinct from T. phagedenis_A17 MATG and all other T. phagedenis isolated to date from domestic cattle.The comparison of two elk hoof-derived T. pedis MATGs with T. pedis genomes isolated from cattle also identified two RODs ranging from 264 to 269 kbp (T.pedis-ROD 1) and 1,940 to 1,960 kbp (T.pedis-ROD 2) that are exclusive to the elk hoof-derived T. pedis MATGs (Fig. 7).T. pedis_A13-ROD 1 contains seven genes, while T. pedis_A13-ROD 2 contains 19 genes that are specific to the T. pedis MATGs.The unique genes identified in these elk hoof-derived MATGs encode diverse metabolic functions, without an over representation of genes associated with any specific functional group (Table S7).T. pedis MATGs (T.pedis_A13 and T. pedis_A17) exhibited 100% genome-wide similarity despite being reconstructed from two different samples (Fig. 7).
Contrary to the above observations on T. phagedenis and T. pedis MATGs, the elk hoofderived MATGs within the other three taxonomic genogroups (Treponema sp., Trepone mataceae_phy1, and Treponemataceae_phy2) showed no significant similarity to any of the genomes with their closest relatives within clades L7 and L4 (Fig. 8).These MATGs showed genome-wide dissimilarity of ≥80% at the nucleotide level when compared with publicly available Treponema genomes within clade L4, suggesting their taxonomic novelty within the Treponemataceae family.

Taxonomic analysis for identification of MATG-specific signatures in biopsy samples
Due to the relatively low numbers of microbial reads obtained from biopsy samples when compared with the skin scraping samples (Table 1), none of the binning approaches could successfully reconstruct MATGs from the biopsy samples.Conse quently, to determine if biopsy samples carried any known or novel Treponema signa tures, we employed an alternative approach including taxonomic analysis using the Kraken 2 pipeline (30), and back mapping reads to the MATGs constructed in this study (Fig. 9).With the Kraken 2 pipeline, each read was assigned to NCBI taxonomic ID, using the lowest common ancestor-based method, whereas with back mapping, each read was assigned a MATG-specific taxonomic ID.For comparison, skin scraping samples were included in this analysis.The results show that the percent abundance of microbial reads in biopsy samples mapping to known Treponema signatures in the current NCBI GenBank database (0.063%) was 10 times lower when compared with the microbial reads from the biopsy samples mapping to the MATGs (0.62%) constructed in this study (Fig. 9a).Similarly, a lower proportion of microbial reads from skin scrapings mapped to known Treponema signatures in the NCBI GenBank database (22%) when compared with reads mapped to MATGs (49%) (Fig. 9b).More specifically, results of analysis using Kraken 2 pipeline revealed that six biopsy samples carried a low abundance of reads mapping to Treponemataceae (0.002% to 0.05%) and T. phagedenis (0.003% to 0.04), five samples showed a low abundance of reads mapping to Treponema (0.02% to 0.06%), three samples showed a low abundance of reads mapping to T. pedis (0.01% to 0.03%), and one sample (15074RH) failed to yield any taxonomic ID (Fig. 9c).In contrast, a higher abun dance of reads (1.76% to 5.11%) from each skin scraping mapped to these four taxa (Fig. 9d).Back mapping of microbial reads from biopsy samples revealed varying proportions of reads from all seven (100%) biopsy samples mapped to five representative MATGs identified in this study including Treponemataceae_phy1_A13 (0.01% to 0.14%), Treponemataceae_phy2_A13 (0.01% to 0.16%), Treponema sp._A13 (0.001% to 0.14%), T. pedis_A17 (0.003% to 0.15%), and T. phagedenis_A13 (0.003% to 0.73%) (Fig. 9e).Similar results were obtained when microbial reads from skin scrapings were back mapped to MATGs (Fig. 9f).Given that the current NCBI GenBank database does not include the genomes of novel MATGs constructed in this study, taxonomic analysis is not expected to detect any MATG-specific signatures.These results show that the biopsy samples contained both the reads specific to known Treponema signatures in the current database and reads that are unique to MATGs.Furthermore, the detection of novel MATG-specific signatures in biopsy samples collected from free-ranging elk with naturally occurring TAHD and from the skin scrapings collected from captive elk early during the experimentally induced disease provides robust evidence that the novel MATGs identified in this study are likely associated with pododermatitis in free-ranging elk.

DISCUSSION
Identifying and characterizing putative pathogens is crucial to support the development of sensitive and specific assays and for etiological and epidemiological investigations for emerging diseases such as pododermatitis in elk.Treponema is identified as a consistent and dominant bacterial group in digital dermatitis in livestock and TAHD in elk through immunohistochemistry, PCR, and V3-V4 amplicon sequencing (2,55).Although research on digital dermatitis in domestic livestock has yielded a limited number of complete genomes of a few culturable Treponema spp., the lack of publicly available Treponema genomes derived from pododermatitis lesions in elk leaves a significant gap in our knowledge of TAHD's etiology and epidemiology.
The comprehensive approach employed in this study, including the reconstruction of novel MATGs from the shotgun metagenome of clinical samples and comparative genomics analysis, broadens our knowledge by capturing a wider range of yet uncul tured Treponema genomes linked to pododermatitis in elk.Among the nine elk hoof samples examined in this study, the bioinformatics pipeline effectively constructed novel MATGs from two skin scraping samples (20-01RF98 and 20-04LF98) from live elk.Given that the tissue biopsies used in this study originated from a convenience sampling of elk mortalities from the field, inconsistencies in the sample freshness, collection, storage, and transportation may have resulted in poor gDNA quality leading to low bacterial read outputs for downstream shotgun metagenomic assemblies.Furthermore, biopsies are naturally highly enriched with host reads as sampling includes deeper tissues with higher host DNA biomass, relative to skin scraping, potentially masking the microbial read outputs and leading to decreased sensitivity for microbial detection.Despite these challenges hindering our ability to reconstruct MATGs from biopsy samples effectively, our alternative approach to utilize a combination of taxonomic analysis of reads from each biopsy sample using the Kraken 2 pipeline and back mapping of biopsy reads against recovered MATGs allowed successful detection and determination of the proportion of novel MATG-specific signatures in all biopsy samples tested.These results suggest that the novel MATGs identified in this study are associated with pododermatitis.The availability of this new genomic information in the publicly available genomic database will facilitate follow-up investigations into the etiology and epidemiology of pododermatitis in elk.While MATGs constructed in this study were from samples collected from captive experimental elk, the study design and results of the previously reported captive elk study provide confidence that the source of MATGs was from lesions induced by challenge with inoculum prepared from TAHD-infected feet of free-ranging elk (3).MATGs were associated with the development of lesions in experimentally challenged elk that were indistinguishable from those in free-ranging elk (3).This finding com bined with the detection of MATGs in biopsies from free-ranging elk by back mapping conducted in this study confirms that these MATGs are present not just in captive experimental elk, but in free-ranging elk.Thus, the successful reconstruction of MATGs from skin scraping samples, which may contain relatively less host biomass and serve as a rich source of microbial signatures, suggests that skin scrapings may potentially serve as more representative samples than biopsies for investigating pododermatitis in elk.
The bioinformatics pipeline employed in this study effectively constructed MATGs with robust evidence of their taxonomic rank through core genome phylogeny, ANI, AAI, and POCP analysis.Notably, four MATGs (group I and group II) provided species-level resolution, and six MATGs (group III, group IV, and group V) provided identification of novel genomospecies within the Treponemataceae family (Fig. 1).The three unresolved taxa, namely Treponema sp., Treponemataceae_phy1, and Treponemataceae_phy2, can be regarded as new additions to the treponemal group associated with TAHD.Hierarchical clustering based on amino acid usage, codon usage, and k-mer frequencies reinforced the genomic differences between currently known Treponema species and novel MATGs and supports distinct species groups identified in this study.The discovery of novel Treponema MATGs suggests the continued expansion of members within the family Treponemataceae and increased genetic diversity within the phylum Spirochaetota (56,57).Our bioinformatics pipeline does not yield full-length 16S rDNA sequences.Thus, it remains unknown if any of the MATGs constructed in this study indeed represent PT19 as reported previously using 16S amplicon sequencing (2).Follow-up studies including reconstruction of novel MATGs from additional samples from free-ranging elk will further strengthen the association and distribution of these MATGs with pododermatitis in wild populations.
In addition to the discovery of novel MATGs, our findings corroborate the significant genetic diversity reported in T. phagedenis-like treponemes linked to BDD (58,59).The data show distant phylogenetic placement of the two T. phagedenis MATGs based on core genome content (Fig. 2).For instance, T. phagedenis-ROD 1 and T. phagedenis-ROD 2 (Fig. 5) represent T. phagedenis_A13 MATG-specific features containing genes encoding a flagellar protein, transporter protein, glycosyltransferase, and several other transferase enzymes.These ROD 1-and ROD 2-specific genes might play a crucial role in confer ring motility (60), environmental fitness (61), and lipopolysaccharide diversity (62) to the T. phagedenis_A13 strain.These findings suggest that T. phagedenis_A13 MATG is likely genetically distinct from T. phagedenis_A17 MATG and other previously reported T. phagedenis strains (Table S7).A recent study reported that bovine strains of T. phagedenis carry two bovine digital dermatitis-specific gene clusters (14).Of the two gene clusters, the OMP gene cluster was absent in T. phagedenis_A17 MATG and present in T. phage denis_A13 MATG.Interestingly, T. phagedenis 4A, the only bovine isolate from the USA available for comparison, also lacks both gene clusters.It is important to note that the majority of T. phagedenis genomes analyzed in this study originated from sources outside the USA.As a result, it is unclear whether T. phagedenis_A17 and T. phagedenis_A13 MATGs constructed in this study are unique to elk or also occur in livestock in the USA.Variability in genetic clusters across T. phagedenis clones from different geographic regions emphasizes the need for further investigation into elk-and livestock-derived Treponema genomes to effectively capture the genetic diversity and discern the host and geographic distribution of these potential pathogens.
The genome-wide BLAST analysis revealed the two elk hoof-derived T. pedis MATGs constructed in this study share a 100% sequence identity across the genome (Fig. 7).However, the presence of two elk hoof-derived T. pedis MATG-specific genomic RODs (T.pedis-ROD 1 and T. pedis-ROD 2) raises a possibility that elk hoof-derived T. pedis is genetically distinct from other T. pedis strains.The unique regions in T. pedis and T. phagedenis MATGs have likely been acquired through horizontal gene transfer occurring throughout the lineage evolution of Treponema genomes (63).The clustering of T. medium and T. vincentii clade L5 (Fig. 2) also corroborates with previously reported close clustering of these two species (8,64).However, sub-clustering of Treponema sp.MATGs in clade L6 alongside T. medium and T. vincentii clade L5 implies that the Treponema sp.MATG potentially represents a novel addition to the existing cluster 4 treponeme phylogenetic group as reported by Yano et al. (64).
Microbial minimalism, a process of genome reduction in bacterial pathogens, limits the metabolic capabilities of these pathogens to better adapt to the host environment (65).Within the Treponema genus, the differentiation of T. pallidum and T. paraluiscuniculi is marked by their limited genome sizes and reduced metabolic activities (Fig. 3).In T. pallidum and T. paraluiscuniculi, highly reduced metabolic capacity is associated with the ongoing genome reduction process and loss of metabolic and biosynthetic pathways (66,67).Interestingly, in our analysis, strains from the same Treponema species were aggregated in a single group, indicating that each member of the Treponema genus has a distinctive metabolic functional profile at the species level (Fig. 3).Among the elk hoof-derived MATGs, Treponemataceae_phy1, Treponemataceae_phy2, and Treponema sp.MATGs showed significantly lower COG functional enrichment relative to other non-Treponema pallidum genomes (Fig. 3).For instance, Treponema sp.MATGs carry an average of 58 and 88 functional genes associated with nucleotide transport metabolism and cell motility, respectively.In contrast, Treponemataceae MATGs carry 39 and 57 genes for these two respective functions.Overall, the novel MATGs derived from elk hooves show a considerably reduced metabolic capacity (Fig. 3).The observed reduced metabolic capacity in elk hoof-derived Treponemataceae MATGs aligns with metabolically distinct strains within the Treponemataceae family associated with vertebrates (68).The diminished metabolic capability of these elk hoof-derived MATGs is not surprising, given the relatively small genome size, host dependency, and fastidious nature (66,67), and likely contributes to challenges in successful culturing and increased host dependence and adaptation (69).Thus, the culture-independent in silico approach for the identification and characterization of MAGs from clinical samples employed in this study could serve as a valuable tool for continued investigations into the etiology and epidemiology of pododermatitis in elk with potential translational applications in investigations of other diseases where etiological agents remain obscure.

Conclusion
This study addresses the critical need for comprehensive identification of Trepo nema species associated with pododermatitis in elk.By employing a robust bioinfor matic pipeline for metagenome reconstruction and taxonomic classification of novel Treponema from clinical samples, this study enhances our knowledge of genomic diversity within the Treponemataceae family.The study successfully uncovered novel treponemal groups within the Treponemataceae family, emphasizing the need to expand the data set of Treponema genomes linked to elk pododermatitis.The application of this bioinformatic pipeline to hoof samples from cattle and sheep holds promise for yielding similar insights into Treponema associated with digital dermatitis in livestock.Observations on the reduced metabolic capacity in elk hoof-derived Treponemataceae MATGs offer valuable insights into potential fastidiousness, which presents challenges in successful culturing, and host dependency, consistent with patterns observed in other Treponema species associated with vertebrates.In summary, this study identifies novel genomospecies of Treponema associated with pododermatitis in elk, thereby enhancing our understanding of the etiology and epidemiology of this disease.It underscores the ongoing need for research to expand the data set of Treponema genomes, which is essential for capturing genetic diversity and unraveling host and geographic distribution of these pathogens.These efforts are crucial for combating emerging diseases such as pododermatitis in both wildlife and livestock populations.

a
GC, guanine (G) and cytosine (C); CDS, coding sequence.and T. phagedenis_A17) displayed ANI values of >95% and AAI values of >96%, and thus were confidently assigned to the well-defined Treponema species group, i.e., T. pedis and T. phagedenis, respectively.However, group III MATGs (Treponema sp._A13 and Treponema sp._A17), group IV MATGs (Treponemataceae_phy1_A13 and Treponemata ceae_phy1_A17), and group V MATGs (Treponemataceae_phy2_A13 and Treponemata ceae_phy2_A17) displayed ANI values of <95% (Table Collectively, out of 10 MATGs identified in this study, two MATGs are identified as T. pedis (T.pedis_A13 and T. pedis_A17) and two are identified as T. phagede nis (T.phagedenis_A13 and T. phagedenis_A17).Out of the remaining six MATGs, two MATGs have been taxonomically classified as Treponema sp.(Treponema sp._A13 and Treponema sp._A17), exhibiting a POCP greater than 50%, indicating they belong to the same genus with an unknown species.On the contrary, the four MATGs recognized as Treponemataceae (Treponemataceae_phy1_A13, Treponemataceae_phy1_A17, Trepone mataceae_phy2_A13, and Treponemataceae_phy2_A17) exhibit a POCP of less than 50% against reference Treponema genomes, leading to their taxonomic classification as a distinct genus within the Treponemataceae family.

FIG 1 FIG 2
FIG 1 (a) ANI and (b) AAI matrix of 10 MATGs constructed from elk hoof samples.MATGs align into five groups based on ANI.The color gradient represents percent ANI (a) and AAI (b).

FIG 3
FIG 3 Heatmap depicting COG function profiles across Treponema genomes.Differentially enriched metabolic functions are presented for reference genomes (colored in black) and metagenome-assembled Treponema genomes from elk hooves (colored in red, blue, and pink).Raw z-score (scale range from −4 to +4) indicates abundance of genes associated with each functional category illustrated with brown (low abundance) and blue (high abundance) colors.The hierarchical clustering of genomes was performed using a weighted Bray-Curtis approach.

FIG 4
FIG 4 Heatmap representing codon usage distance among Treponema genomes and MATGs from elk hooves.The hierarchical clustering of strains was conducted using the average linkage method and the Bray-Curtis dissimilarity metric.

FIG 5
FIG 5 Genome-wide comparison of Treponema phagedenis strains.Circular map representing T. phagedenis_A13 genome-wide comparison with other publicly available T. phagedenis strains.The percentage similarity is represented by different color codes from outer circle to inner as follows: genome size, genes on the forward strand, genes on the reverse strand, tRNA, and rRNA.The outermost rings (blue) show two strands of T. phagedenis_A13 MATG.The inner 16 rings (red) show the BLAST sequence similarity comparisons between the T. phagedenis_A13 and 16 T. phagedenis genomes, including T. phagedenis_A17 MATG (ring 11), sequentially from inner to outer circle as referenced in the bottom left of the figure.

FIG 6
FIG 6 Comparison of T. phagedenis T320A genome with other T. phagedenis strains for the presence of two bovine-specific gene clusters.(a) Phosphate utilization gene cluster.(b) Gene cluster associated with outer membrane production.

FIG 7
FIG 7 Circular map representing Treponema pedis_A13 genome-wide comparison with other reference T. pedis strains.The percentage similarity is represented by different color codes.The information is read from outer circle to inner as follows: genome size, genes on the forward strand, genes on the reverse strand, tRNA, and rRNA.The next three rings show the sequence similarity by BLAST comparisons between the T. pedis_A13 and three other T. pedis genomes, sequentially from inner to outer circle as referenced in the bottom left of figure.

FIG 8
FIG 8 Circular plot of Treponema sp._A13 (a), Treponemataceae_phy1_A13 (b), and Treponemataceae_phy2_A13 (c) MATGs compared with other Treponema genomes via CCT.The details of each concentric ring are described in File S1.

FIG 9
FIG 9 Relative abundance of Treponemataceae members in biopsy (left panel) and skin scraping (right panel) samples obtained via taxonomic analysis of reads using Kraken 2 pipeline (c and d) and by back mapping of reads to MATGs constructed in this study (e and f).The top panels show the total percent microbial reads detected in biopsy (a) and skin scraping (b) samples using Kraken 2 pipeline and back mapping to MATGs.

TABLE 2
General genomic features and quality metrics of 10 MATGs constructed from two elk hoof metagenomes