The semen microbiome of miniature pony stallions

Subjects

Four crossbred miniature pony stallions, aged between 2 and 15 years and stabled at an equestrian centre in New South Wales, Australia, were recruited for this study in October 2019. They comprised a cohort of horses that routinely provide semen specimens for scientific purposes and in accordance with the ethical requirements of this study.

The ponies grazed on native grasses in adjacent paddocks, and lucerne hay was added to their diet throughout the year. No exercise regime was undertaken other than free paddock roaming, and the ponies were stabled overnight. The body condition score of the ponies was 6–7 on the US body condition score chart (Henneke et al. 1983).

Study design

Semen samples were collected on the same morning from each pony with a minimum of two days rest between each semen collection. Seminal specimens were prepared and subjected to genomic profiling analysis of bacterial microbial composition, and classified by phylum, order, class, family, genus, and alpha and beta diversity for a descriptive analysis.

Collection of semen samples

After careful washing of each stallion’s prepuce and penis with warm water to remove smegma and other residues, semen collections were undertaken. Semen was collected from the miniature pony stallions using a Missouri artificial vagina (Minitube Australia, Ballarat, Vic.) with an in-line gel filter, and a phantom mount. Following semen collection, 250 μL was immediately removed and diluted with 20 μL of DNA Microshield, after which it was stored at −30°C until analyses.

Microbial DNA profiling of semen microbiota

ZymoBiomics DNA Miniprep kit (Zymo Research) was used to extract DNA from 150 μL specimens of seminal fluid, having been firstly digested with Proteinase K at 55°C for 30 min, following the manufacturer’s instructions. Full-length 16S rRNA gene sequencing (Hypervariable regions V1-9) was performed in preparation of the PacBio libraries, with SMRTBell® Library Preparation and Sequencing Procedure and Checklist with 15 ng extracted DNA input for the semen samples library preparation. Initially, PCR was conducted with KAPA HiFi HotStart ReadyMix (Roche), with 20 cycles in 25 μL reactions following protocol cycling conditions. Amplified DNA was indexed using the barcoded F/R Universal Primers Plate 96 v2 – part number 100-466-100 (Pacific Biosciences), and PCR products were purified by using AMPure PB (Agencourt Bioscience), in accordance with PacBio’s protocol. Two primers were employed in this study. Primer *27F, 5’ block /5AmMC6/ with universal sequence gcagtcgaacatgtagctgactcaggtcac 16S rRNA gene Primer AGRGTTYGATYMTGGCTCAG. The second primer used was **1492R 5’Block /5AmMC6/ with universal sequence tggatcacttgtgcaagcatcacatcgtag, 16S Primer RGYTACCTTGTTACGACTT.

Verification of amplicon sizes was carried out with a Fragment Analyser 5200 (Agilent Technologies) and Qubit 4.0 Fluorometer (ThermoFisher Scientific) using the High Sensitivity dsDNA assay. Pooling of barcoded amplicons was performed before the library construction process with the SMRTBell® Express Template Prep Kit 2.0 (Pacific Biosciences). Semen specimens were pooled in groups of 20, then sequenced on separate chips. A PacBio sequel (Pacific Biosciences) sequencing process was used with v3 sequencing chemistry in CCS collection mode for 20 h on a 1M v3 LR SMRT Cell. 16S rRNA gene sequencing used in this protocol is 1600 base pairs. Following this, CCS analysis was run prior to demultiplexing and fastq conversion. The CCS (Circular Consensus Sequence) analysis algorithm is applied to processes which require distinguishing of closely related molecules of DNA.

Taxonomical classifications of microbiota

The semen microbiota of the subjects were categorised according to taxonomical classifications of phylum, order, class, family and genus.

Bioinformatics for microbiome profiling

Prior to data analysis, the quality of the raw reads was checked using the package FastQC (Ver. 0.11.9). An average 80–95% of the raw reads were retained across the samples. This was followed by inspection of the length distribution of the reads. Most of the sequences were found to be in the expected range of approximately 1400 to 1550 bp. A filter was applied to retain the sequences between 1000 and 1600 bp length. The next-generation microbiome bioinformatics platform QIIME2 (version q2cli ver. 2020.8.0) was used for analysing the amplicon data and further generation of Amplicon Sequence Variants (ASVs) was conducted from the raw amplicon data, using the algorithm DADA2 (ver. 1.19.2). DADA2-formatted training fasta files were derived from the Silva Project’s ver. 128 release and used for taxonomy assignment. Three output files were generated using DADA2. These included the sequence feature count tables, the taxonomy identification file per ASV and a fasta file containing sequences of all ASVs. These files were exported in a phyloseq format for further analysis using QIIME2. These files along with the sample metadata files were used as input for different modules in the QIIME2 workflow.

Data produced by QIIME 2 exists as QIIME 2 artifacts, which contains data and metadata. The feature tables imported from DADA2 were filtered to retain the features, which were observed in at least two samples. A rooted phylogenetic tree was generated using the filtered feature table for downstream diversity analysis.

Data analysis

This cross-sectional data sample was prepared in Microsoft Excel v16.67 (2022). ASV count data was aggregated first at the classification of phylum and sorted from highest to lowest frequency. This method was applied again for each of the remaining taxonomical ranks, namely order, class, family and genus. The top 10 organisms identified in each taxon for each pony, with the addition of an overall microbial profile were reported. Principal coordinates analysis (PCoA) was applied to analyse and visualise biodiversity between the semen microbiota of the four subjects using Bray–Curtis distance matrix. The alpha diversity measures evenness (abundance), richness and Shannon’s diversity index summarise the cohort’s semen microbiota. Summary statistics report the distribution of those three alpha diversity measures.

Ethic statement

All procedures undertaken in this study were approved by the University of Newcastle Animal Ethics Committee and adhere to the Australian code for the care and use of animals for scientific purposes (approval number A-2011-122). The Australian code for the care and use of animals for scientific purposes adheres to the EU standards for the protection of animals used for scientific purposes.

Phylum

The eight phyla predominating the semen samples of four miniature stallions were Bacteroidetes, Firmicutes, Proteobacteria, Fusobacteria, Spirochaete, Actinobacteria, Tenericutes and Synergistetes; those less frequently represented were categorised as other (Fig. 2). Overall, the phyla Bacteroidetes and Firmicutes were most represented in this small cohort, comprising 76%, and Proteobacteria was the third most abundant phylum, comprising another 15% of the phyla identified. Less than 1% of those detected were classified as taxon undescribed (NA).

Order

Within the taxonomical classification of order, Bacteroidales, Clostridiales, Cardiobacteriales, Fusobacteriales, Bradymonadales, Corynebacteriales, Propionibacteriales, Spirochaetales and Mycoplasmatales were most abundant (Fig. 3). The three predominating orders were Bacteroidales, Clostridiales and Cardiobacteriales, accounting for 86% of the abundance distribution.

Class

Within the taxonomic classification of microbial class, Bacteroidia, Clostridia, Gammaproteobacteria, Fusobacteriia, Actinobacteria, Deltaproteobacteria, Spirochaetes, Mollicutes, Betaproteobacteria and Alphaproteobacteria were the most abundant, with 87% of the microbiota composed of Bacteroidia, Clostridia and Gammaproteobacteria (Fig. 4).

Family

The most abundant microbial families were Porphyromonadaceae, Family_XI, Cardiobacteriaceae, Bacteroidales_BS11_gut_group, Ruminococcaceae, Fusobacteriaceae, Peptostreptococcaceae, Rikenellaceae and Corynebacteriaceae. Of these, Porphyromonadaceae, Family_XI (Family XI Incertae sedis) and Cardiobacteriaceae predominated, comprising 62% (Fig. 5).

Genus

Within the taxonomical classification of microbial genus, Porphyromonas, Suttonella, Peptoniphilus, Fastidiosipila, Ezakiella, Petrimonas, Fusobacterium, Gallicola and Finegoldia were the most abundant genera identified in the semen. Of these, Porphyromonas, Suttonella, Peptoniphilus, Fastidiosipila, Ezakiella, Petrimonas and several undescribed taxa (NA) predominated (Fig. 6). In terms of the microbiota represented in individual ponies, some differences were observed in the abundance of the family and genus level classifications. For example, Pony B had greater abundance of Fastidiosipila than the other ponies, Pony C had greater abundance of Fusobacterium, Pony P had greater abundance of Suttonella and Pony Z had greater abundance of Ezakiella genus.

Beta diversity

Fig. 7 illustrates the beta diversity measure (microbial diversity between subjects). The microbial composition of Ponies P and Z was similar compared to Ponies B and C.

Fig. 7.

Beta diversity of four miniature stallions’ semen microbiota using the Bray–Curtis distance matrix. The microbial composition of Ponies P and Z was similar, compared to Ponies B and C.

Alpha diversity

Table 1 presents the summary statistics for alpha diversity measures of the semen microbiota of the four subjects. Variations in the species richness (number of unique species within the sample) had the highest variation as reflected by the large standard deviation (s.d.). The mean and median were the same for evenness and Shannon diversity indicating an approximately even distribution. A skewed distribution was observed for microbial richness of the semen samples as reflected by the 11.5 points difference in mean and median.

Semen microbial phyla and families

Quiñones-Pérez et al. (2022) evaluated the phyla and families present in the semen of 12 healthy Arab, Anglo-Arab and Andalusian stallions (aged 7–24 years) located at the Spanish Army’s breeding centre in Seville, Spain. Like Quiñones-Pérez et al. (2021), we also found that Bacteroidetes and Firmicutes predominated equine semen. However, Actinobacteria was the third most abundant phylum present in the semen of the Spanish stallions, whereas Proteobacteria was the third most abundant phylum in the miniature ponies’ semen. These Gram-negative bacteria play an important metabolic, energy-generating role in the equine intestinal microbiome (Li et al. 2022). Proteobacteria have also been reported as being more represented in the gastrointestinal tract of horses with equine grass sickness (Garber et al. 2020) and colitis (Ayoub et al. 2022). Somewhat contrary to this, in a separate study, Proteobacteria colonies were reported to be less represented in the faeces of horses affected by diarrhoea (Li et al. 2022). The presence of Proteobacteria in the semen was not associated with these conditions in the miniature ponies studied, which were of good health prior to and throughout the study period, with no reports of diarrhoea or other illness. Interestingly, many of the less abundant phyla, including Fusobacteria, Spirochaetes, Synergestes, Tenericutes and Chloroflexi, were present in similar amounts in both studies. Such similarities across demographic locations signal there may be a characteristic semen microbiome of stallions that is independent of environmental factors.

At the taxonomic classification of family, Quiñones-Pérez et al. (2021) identified 69 families of microbes in their stallions’ semen, of which only 22 families were present in proportions greater than 1% and only nine were common to all samples (Quiñones-Pérez et al. 2021). The microbial family represented in the Spanish stallion’s semen were predominantly Porphyromonadaceae, which was similar in preponderance to our miniature ponies. Of interest to fertility research, Porphyromonadaceae is reported as a predominant family present in fertile stallion semen (Quiñones-Pérez et al. 2021), where it is found to be more dominant than in human semen (Hou et al. 2013). In contrast, the Spanish stallions had a much greater abundance of Corynebacteriaceae in the semen compared to the miniature ponies in this study, who had more Family XI present. The Corynebacteriaceae family, which has also been associated with fertility in stallions, was not highly represented in our cohort of miniature ponies.

Differences in phylum and family identification between the studies are not readily attributed to the chosen methods. In many regards, the methods in both studies were similar, involving similar collection processes during the spring season, to obtaining one semen specimen and using the 16S rRNA gene sequencing. One distinction which may account for the variations between the two studies was that the Spanish study used the 16S rRNA gene hypervariable V3 region for microbial detection. However, in our study full length (hypervariable regions 1–9) 16S rRNA gene amplicon sequencing was used to detect microbial taxa. In addition, using different DNA extraction protocols could influence sequencing outcomes and the bio-informatic pipelines employed by both studies were different, which may have influenced the respective findings. Differences between the Spanish study and the findings reported here are more likely to relate to breed, diet, geographical location and associated environmental factors (Theelen et al. 2021). The stallions studied by Quiñones-Pérez et al. (2022) in Spain were fed alfalfa hay and oats, rather than lucerne hay, which may have produced a different semen microbial profile to that of our stallions. Although, it could be argued that dietary variations are more likely to be associated with differences observed in the gastrointestinal rather than extraintestinal microbial communities. The faecal microbiome of ponies and horses in good health is known to be impacted by environmental factors, such as air exposure, farm location (region and country), season and access to pasture (Theelen et al. 2021), but these influences are yet to be demonstrated in relation to extraintestinal microbial communities such as semen.

Semen microbial genera

The aforementioned factors may also explain the differences in genera identified in equine semen by Al-Kass et al. (2020a), who examined the microbial composition of semen specimens from seven healthy stallions (aged 7–17 years) on a commercial stud in Stockholm, Sweden. Characterising the microbiota from one ejaculate per animal to genus level (Al-Kass et al. 2020a), they identified almost twice as many genera in the semen in comparison to those discovered in our cohort, with only Porphyromonas and Peptoniphilus being common to both cohorts. The four most common genera found in the Swedish study were Porphyromonas, Corynebacterium, Finegoldia and Peptoniphilus, followed by Mobiluncus, Chondromyces, Suttonella, Treponema, Actinetobacter, Campylobacter and a group of ‘other’. In comparison, the three most abundant genera in the semen of our miniature ponies were Porphyromonas, Suttonella and Peptoniphilus, followed by Fastidiosipila, Ezakiella, Petrimonas, Fusobacterium, Gallicola, Finegoldia and ‘other’. No Ureaplasma, Spirochaeta, Prevotellaceae, Fibrobacter, Acinetobacter, Chondromyces or Clostridium were identified in our miniature ponies. Like the Al-Kass study, our miniature ponies displayed predominantly Porphyromonas genus, an organism also common to healthy human semen (Hou et al. 2013; Liu et al. 2014). Consistent with other equine studies, we did not find any Lactobacillus genus in our samples, a bacterium which has been associated with good sperm quality in humans (Baud et al. 2019).

The methodology and technology employed in the aforementioned Swedish study were similar in many regards to those employed in our study i.e. semen specimens were collected in spring from one ejaculate, and in the same regular collection routine and manner (using a Missouri AV without mare contact). Although the stallions of the Swedish study were engaged in an artificial insemination program, the procedure for collecting and storing each aliquot of semen before its transport to the laboratory for genomic sequencing was identical to that employed by our Australian stallion handlers, hence one would not expect to find an effect of stallion management on semen microbiome composition to influence the results of these two studies.

Soil bacteria may contact the stallion’s urogenital tract from its environment. A miniature pony stallion’s erect penis can more easily contact the ground than that of a full-sized stallion – thus giving rise to extrinsic contamination and differences in microbes detected. Petrimonas genus, found in the semen of our cohort of miniature pony stallions, resides in soil, where it is able to effectively degrade hazardous compounds (Zhao and Zhang 2021) to innocuous acetic acid and carbon dioxide (Grabowski et al. 2005). Fastidiosipila, a Gram-positive anaerobe previously unrecognised in equine semen, which was identified in the miniature stallions, has also been identified in stored waste of landfill (Wang et al. 2017), hence may have been acquired from the soil by these ponies. These soil-dwelling organisms were not isolated by Al-Kass et al. (2020a) from their full-size breeds. Gallicola genus was also unique in the semen of the miniature stallions. Very little exists in the literature about these microbes, other than that they are Gram-positive, strictly anaerobic cocci associated with dark fermentation of food waste (Shi et al. 2021) and may be a faecal contaminant of ground water sourced from chicken manure (Naphtali et al. 2019).

The seminovaginal microbiome

The influence of the vaginal microbiome on the composition of the semen microbiome is important to consider. The stallion urethral and possibly semen microbiomes may be contaminated by the mare’s vaginal microbiota following mating, and although there are no detailed descriptions of a shared complementary equine seminovaginal microbiome, one study has found microbiome network links between semi-feral Welsh Mountain pony stallions and mares (Antwis et al. 2018). Porphyromonas is the most abundant genus in Arabian mares’ vaginal microbiomes (approximately 14–17%) and Bacteroidales is the highest abundance by order (approximately 31%) (Barba et al. 2020). By comparison, our stallions’ semen contained 25% Porphyromonas and 42% Bacteroidales, although they had not had genital contact with mares, so perhaps the colonisation of the mare’s vaginal microbiome arises from the stallion’s penis, which harbours larger abundances of these genera. At phylum level, the core species of the vaginal microbiome of the mare are comparative in ranking to those phyla found in miniature pony stallion semen (Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria) (Barba et al. 2020). Microbes from the Corynebacteriaceae family are found in low abundance on equine skin (Costa et al. 2016; Kaiser-Thom et al. 2021) as well as in semen, which might suggest a degree of cross-contamination of this bacterium that occurs during specimen collection.

Limitations of this study

Only one breed of stallion was recruited for this study, with single semen samples collected and profiled for each miniature pony stallion, thus reducing the generalisability of the findings. The ASVs display advantages over the more traditional approach of Operational Taxanomical Units (OTUs) as they represent the exact amplicon sequences without coarse graining into OTUs and can resolve even single nucleotide base differences. However, it should be considered that ASVs can sometimes over-split sequences, creating separate clusters for very similar sequences. This ‘over-splitting’ can lead to the separation of sequences that belong to the same bacterial genome, creating artificial diversity. (Schloss 2021).

Due to the sample size, detailed statistical analysis and comparison with various measures of sperm quality were not possible, thus associations between the composition of the semen microbiome and sperm quality could not be made. The detailed characteristics of the semen microbiome in miniature pony stallions may not be generalisable but provides preliminary data to inform such research.