Targeted next-generation sequencing assay to detect 3 Moraxella spp. directly from bovine ocular swabs

Abstract

Infectious bovine keratoconjunctivitis (IBK) is associated with 2 species of Moraxella: M. bovis and M. bovoculi. A third novel Moraxella spp., designated tentatively as M. oculi, has been identified from the eyes of cattle with and without pinkeye. These 3 Moraxella spp. can be found in various combinations within the same clinical sample, making speciation of this genus directly from a sample impossible with Sanger sequencing. Assessing Moraxella diversity found in IBK- and non-IBK–affected cattle eyes, independent of culture, may provide additional information about IBK by avoiding the selectivity bias of culturing. We developed a targeted NGS panel to detect and speciate these 3 Moraxella spp. directly from bovine ocular swabs. Our targeted panel amplifies bacterial essential genes and the 16S-23S ribosomal RNA intergenic spacer region (ITS) of the 3 Moraxella spp. and speciates based on these sequences. Our panel was able to differentiate the 3 species directly from DNA extracted from 13 swabs (6 from healthy animals, 7 from animals with IBK), and every swab except one (clinically healthy eye) had the 3 Moraxella spp. Targeted NGS with sequencing of Moraxella spp. housekeeping genes appears to be a suitable method for speciation of Moraxella directly from ocular swabs.

Infectious bovine keratoconjunctivitis (IBK), or pinkeye, is an economically important disease of cattle, resulting in significant production losses and losses associated with treatment costs.^1,14 IBK has been documented experimentally to be caused by Moraxella bovis. ⁹ M. bovoculi has been associated with this disease and has been cultured from pinkeye cases in the absence of M. bovis,^2,10 but to date, adequate evidence to support a causal role for this organism is lacking.^1,8,11 Although both of these Moraxella spp. can also be detected from clinically healthy eyes, ⁴ distinct genotypes of both species have been identified that may associate differently with eyes from healthy cattle compared with eyes from cattle with IBK.^11,14

We identified another Moraxella species (suggested name M. oculi) associated with both IBK and clinically healthy eyes. ⁵ In addition, more than one Moraxella spp. can be detected in a single pinkeye case. ¹⁰ Speciation of Moraxella directly from a clinical sample (no culture/isolate) is impossible with PCR and direct Sanger sequencing, although these techniques have been used for speciating isolated colonies. ¹⁰ As well, the regions that are used routinely for speciation, such as 16S rRNA gene variable regions, ¹³ may fail to provide accurate results.^5,7 A multiplex PCR assay designed to detect and speciate both M. bovis and M. bovoculi using the cytotoxin A gene has been described. ¹⁶ However, that PCR assay has not been evaluated with M. ovis, an organism that is detected occasionally from bovine eyes. ¹⁰ This is a significant issue given that there is high sequence identity between the M. bovoculi and M. ovis cytotoxin A genes. Thus, the complications with accurate Moraxella speciation limit our ability to adequately classify the organisms involved in the disease or prevent IBK.

Detection of the Moraxella spp. associated with pinkeye cases directly from a clinical sample in the absence of culturing, which may select for growth of one organism over another,^6,10 will aid in assessing how these Moraxella spp. contribute to pinkeye and potentially lead to better prevention or treatment strategies. We designed a targeted next-generation sequencing (NGS) panel (AmpliSeq Designer, White Glove Team-Ion Torrent; Thermo Fisher) with primers to speciate M. bovis, M. bovoculi, and M. oculi directly from ocular swabs that amplify 3 essential genes (ATP synthase F1 epsilon subunit gene [ATP], RNA polymerase subunit B [rpoB], and phospho-N-acetylmuramoyl-pentapeptide transferase), as well as the 16S-23S ribosomal RNA intergenic spacer region (ITS) of these 3 species. There are 100 intellectual property–protected primers available as 2 primer pools from Thermo Fisher; FASTA and Bed files are available from the corresponding author upon request.

We evaluated the specificity of the primers in silico using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Specificity was also evaluated by testing previously verified field isolates of M. bovis, M. bovoculi, M. oculi, and M. ovis (Suppl. Table 1), individually with our assay. Each of these field isolates had been verified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS; Biotyper Sirius, Bruker) and 16S-23S ITS Sanger sequencing using primers described previously. ³ The Biotyper MBT Compass library (Bruker) was used for species identification, with identifications accepted for MALDI-TOF MS scores of ≥2. The MALDI-TOF MS could not identify M. oculi using the commercial database (no reliable identification). For ITS sequencing, species identification was determined with BLAST and based on ≥99% nucleotide identity. The ITS sequence for M. oculi has been deposited in GenBank (OR428531). Additionally, dilutions of DNA from the verified field isolates were prepared to contain ~10,000, 5,000, 1,000, and 100 genome equivalent copies of the organisms to evaluate assay sensitivity. Then, a preliminary evaluation was performed with ocular swabs obtained for a separate study from 13 cattle that were age- and breed-matched (7 with pinkeye, 6 with clinically healthy eyes) from 2 separate farms known to have M. oculi based on culture and ITS Sanger sequencing. ⁵ The nucleic acid was extracted from the ocular swab collected from each animal (DNeasy blood and tissue kit; Qiagen) and stored at −80°C until tested. These samples had been found to contain Moraxella spp. based on a targeted 16S rRNA gene metagenomics approach using primers to amplify and sequence the V2, V4, V8, V3–6, and V7–9 hypervariable regions (Ion 16S metagenomics kit; Thermo Fisher). ⁵ Despite obtaining a genus-level identification for these samples and the potential for speciation based on sequencing multiple hypervariable regions, the 16S rRNA gene is a poor target for speciating Moraxella. ⁵ A previous study showed interspecies recombination within the 16S rRNA gene of Moraxella, which necessitates evaluation of other genetic segments of these organisms to accurately resolve to the species level. ⁷

For our Moraxella targeted NGS assay, we used the custom-made primers to amplify M. bovis, M. bovoculi, and M. oculi target genes, followed by the preparation of libraries (Ion Chef, AmpliSeq library kit for Chef DL8 with incorporated barcodes; Thermo Fisher) according to the manufacturer’s protocol. Libraries were templated and loaded onto an Ion 520 chip and sequenced (Ion Chef instrument, Ion Gene Studio S5 system, 510 & 520 & 530 sequencing kit; Thermo Fisher) according to the manufacturer’s protocol. The sequences were assembled using SPAdes (v.3.15.5; https://github.com/ablab/spades/blob/spades_3.15.5/README.md#sec5) and mapped to a reference FASTA file containing the sequences for the reference genes and the ITS region for each of the 3 Moraxella spp. (M. bovis, M. bovoculi, M. oculi). Geneious software (www.genieous.com) was used for sequence analysis, and products ≥150 bp (based on approximate expected size based on the design) were evaluated by BLAST to confirm the results (query coverage and identity >99%, E values <0.0001). The results of the targeted NGS (defined as total percentage of reads of each of the 3 species) from the 2 tested groups (IBK cases vs. clinically normal controls) were compared by Mann–Whitney U tests.

Based on analysis of the assay with validated field isolates, our targeted NGS assay was able to differentiate M. bovis, M. bovoculi, and M. oculi. The primers designed were specific for each species, except that M. bovoculi primers also detected M. ovis, but the sequences obtained allowed differentiation of these 2 organisms. Of the targets included in our assay, the primers for ITS, rpoB, and the transferase genes consistently produced products for all 3 species (M. bovis, M. bovoculi, M. oculi). The primers for the ATP were less sensitive for all 3 species, evidenced by fewer reads detected with these primers compared to the other primer sets. Our assay was also able to detect every DNA dilution of each organism, with thousands of reads still detectable per species when tested with ~100 genome equivalents.

We found that Moraxella speciation could be performed directly from DNA extracted from ocular swabs, without the need for culture to separate multiple Moraxella spp. that may be present in the swab sample. All swab samples, except one (a sample from a clinically normal bovine eye), contained all 3 Moraxella spp. There was a statistical difference (p = 0.035) between the relative abundance of reads (percentage of total reads) of M. bovis detected from cattle with clinical signs of IBK versus cattle with clinically healthy eyes (Fig. 1A). There was no statistical difference between the 2 groups for M. bovoculi (p = 0.295) or M. oculi (p = 0.628; Fig. 1B, 1C).

Figure 1.

Histograms comparing the distribution of percentage of total reads for: A. Moraxella bovis, B. M. bovoculi, and C. M. oculi between clinically normal (Norm; n = 6) and pinkeye (PE; n = 7) samples. There is a statistically significant difference between Norm and PE for M. bovis (A; p = 0.035) but not for M. bovoculi (B; p = 0.295) or M. oculi (C; p = 0.628).

Our small sample size may have precluded our ability to detect a significant difference between these 2 species (M. bovoculi and M. oculi) in the IBK group versus the non-IBK group. Based on the side-by-side histograms, the IBK group tended to have a lower number of reads for M. bovoculi than for the non-IBK group, and the 2 distributions mostly overlap for M. oculi. Based on our previous findings, ⁵ M. oculi is likely a nonpathogenic Moraxella sp. that is part of the normal ocular microbiota of cattle. This presumption aligns with our current study given that we would expect an increase in M. oculi with disease, and thus a greater abundance of the organism in the pinkeye group, compared to the control group, if it were a causative agent. However, evaluation of individual animals over time (before pinkeye to resolution) may provide a better understanding of causality.

Moraxella spp. were not significantly different between pinkeye cases and controls in a previous ocular microbiota study, although Moraxella was consistently observed in the top 10 most abundant genera detected. ⁶ Moraxella could not be speciated in that study; hence, different species of Moraxella may have been statistically different between samples, or a difference could possibly even be seen at the level of genotype.^7,12,15 Including primers in our proposed assay to genotype M. bovis and M. bovoculi would allow for a more accurate evaluation of the data. Further testing is warranted, in addition to challenge studies, to determine the role, if any, of M oculi in pinkeye.