Isolates and typing
The isolates characterised as well as strain affiliations, geographic origins and clinical presentations are summarised in Table 1. Autopsy images showing typical aspects of putrid infections in some animals are shown in Fig. 1. The complete microarray hybridisation patterns are provided as Supplemental file 2 and some relevant features will be discussed in the descriptions of the respective strains. While all German isolates yielded hybridisation signals for lukF/S-PV, frequently only weak positive or ambiguous results for the lukS-PV probe were observed. This prompted further investigations, including the detection of PVL by lateral flow assay21 (Table 1) and whole genome sequencing (see below).
Pathological lesions of Eurasian beavers (C. fiber) infected with BVL-positive S. aureus. (A) Severe suppurative necrotizing pneumonia (animal B); (B) severe suppurative pyelonephritis (animal G); (C) caseous lymphadenitis, popliteal lymph node (animal E); (D) urinary bladder with pyuria (animal C).
Phenotypic and genotypic resistance properties of the S. aureus isolates
Antimicrobial susceptibility testing revealed that all beaver isolates from Germany were susceptible to all antimicrobial agents tested. The distribution of minimal inhibitory concentration (MIC) values and test ranges are displayed in Supplemental File 3a. The phenotypic data corresponded well with microarray data, since none of the corresponding resistance genes was identified. In contrast, two of the Austrian isolates showed macrolide resistance with one of them also being lincosamide resistant. One isolate also exhibited tetracycline resistance. These phenotypes corresponded with the detection of genes erm(A), erm(C) and tet(M), respectively (Supplemental file 2 and 3b).
The chromosomal variant of the metallothiol transferase gene fosB was present in all CC1956 isolates. Sequence analysis revealed a frame shift at position 108 creating a stop codon at positions (pos.) 146.0.148 compared to the reference sequence (N315, GenBank BA000018.3 [2,389,328.0.2,389,747]). This resulted in a truncated protein of 48 amino acids (aa) rather than 139 aa as for the original fosB gene product. The mutation was present in all available sequences (i.e., Oxford Nanopore and Illumina of WT19 as well as Illumina of WT63, WT64, WT66, WT67a, WT67b, WT68, WT69, WT70, WT71, WT110 and WT111). While fosB was originally implicated in fosfomycin resistance, it appears to be linked to certain CCs. Indeed, it was also present in the CC8 and CC12 beaver isolates (B2, B3, B4) as well as in the reference sequences of the respective CCs (Supplemental File 2). The fosB gene was absent from the CC49 isolate WT65 and from the CC49 reference sequence of Tager 104, GenBank CP012409.1, as well as from the CC398 isolate B1. Moreover, all sequenced isolates (from animals A to G) harboured a gene designated tet(38), encoding a major facilitator superfamily permease. While this gene was implicated in low-level tetracycline resistance when overexpressed22, its mere presence certainly is not associated with phenotypic tetracycline resistance as it can be found in virtually every S. aureus genome.
Biocide susceptibility testing of the CC49/1956 isolates revealed unimodal MIC distributions (Supplemental File 3b), with ranges encompassing not more than three to four dilution steps for each of the biocides (benzalkonium chloride, 0.00003–0.00025%; polyhexanide, 0.000125–0.0005%; chlorhexidine, 0.00006–0.00025% and octenidine, 0.00006–0.00025% with percentages given as mass per volume). The four remaining isolates showed MIC values of 0.0000125–0.00025% for benzalkonium chloride, 0.0005–0.001% for polyhexanide, 0.00006–0.000125% for chlorhexidine, and 0.000125–0.00025% for octenidine.
The chromosomal heavy metal resistance markers arsB/R and czrB were detected by hybridisation in all four CC1956 isolates tested as well as in the CC49 isolate. This was confirmed by sequencing. There was no evidence for plasmid- or SCC-borne heavy metal resistance markers.
The sequence of the phage-borne leukocidin genes in WT19 and WT65
As mentioned above, CC49/CC1956 beaver isolates yielded occasionally ambiguous hybridisation intensities for lukS-PV probes prompting further investigation assuming that the specifically designed oligonucleotides were not able to bind optimally at the target due to mismatches, i.e., allelic variants. Sequencing revealed the presence of distinct alleles of phage-borne leukocidin genes (Figs. 2a/b and 3a/b). The sequences from the two sequenced beaver isolates were identical to each other despite their origin from different prophages in different CCs. In general, the beaver alleles, hitherto referred to as “Beaver Leukocidin” or BVL, lukF/S-BV, appeared to be closer related to the PVL genes from human strains of S. aureus than to those from ruminants and horses (see Figs. 2a/b and 3a/b and the percentages of homologies as provided in Supplemental File 4). There was no evidence for recombination/chimerism in lukF-BV and lukS-BV as mismatches compared to other sequences were evenly distributed across the entire sequences. Sequences of lukF-BV and lukS-BV were also related but clearly distinct from core genomic lukF/S-int of S. intermedius/pseudintermedius.
(a) Alignment of the lukF-BV sequences, of other phage-borne leukocidin F component sequences from S. aureus and of lukF-int from S. intermedius/pseudintermedius. (b) Alignment of the amino acid sequences of the corresponding lukF gene products.
(a) Alignment of the lukS-BV sequences, of other phage-borne leukocidin S component sequences from S. aureus and of lukS-int from S. intermedius/pseudintermedius. (b) Alignment of the amino acid sequences of the corresponding lukS gene products.
lukF/S-BV and the agr locus
Two isolates from one animal, WT110 and WT111 (Table 1), differed in hemolysis on Columbia blood agar and were thus handled separately although array analysis eventually revealed the same strain affiliations. They also differed in BVL production as shown by lateral flow tests. Sequencing using both, Illumina and Oxford nanopore technologies, revealed a substitution from A to T in position 706 of the agrA gene that results in a premature stop codon at position 236 of the agrA gene product (Supplemental File 5) suggesting that agr played a role in the observed phenotype and the regulation of BVL.
Core genome and genomic islands of the CC1956 isolate WT19
As revealed by array experiments (Supplemental File 1) and confirmed by genome sequencing of WT19, CC1956 isolates presented with agr IV alleles and capsule type 5. They were positive for cna, but they lacked seh and egc enterotoxin genes, ORF CM14 as well as sasG. Leukocidin genes lukX/Y, lukD/E and lukF/S-hlg were present. This is also in accordance with previously sequenced BVL-negative CC1959 isolates (SAMEA3251370, SAMEA3251372, SAMEA3251377, SAMEA3251376, SAMEA3251380; Supplemental File 2).
The WT19 genome (Supplemental Files 6a and 6b) harboured two uncharacterised enterotoxin genes (pos. 1,940,148..1,940,900 and pos. 1,939,378..1,940,121). Both were also found in DAR4145 (CC772) where they also formed a genomic island at approximately the same position within the genome (GenBank CP010526.1: RU53_RS09775, pos. 1,968,336..1,969,061 and RU53_RS09780, pos. 1,969,088..1,969,840). One of these two genes (“seu2” = RU53_RS09780) was covered by the second array-based assay23 and it was found in all four isolates tested with this array.
Mobile genetic elements in the CC1956 isolate WT19
The lukF/S-BV prophage was integrated into the lipase 2 gene (lip2, “geh”, “sal3”, “salip35”, GenBank CP000253.1 [314,326..316,398]), and spanned pos. 322,629 to 365,636. Besides leukocidin genes, it also included genes associated with the different modules of a typical Siphoviridae genome (lysogeny, DNA metabolism, packaging and capsid morphogenesis, tail morphogenesis, host cell lysis24,25; see Supplemental File 7/Fig. 4).
Schematic representation of the aligned sequences of the lukF/S-BV prophages from WT19 and WT65.
Furthermore, there was a small pathogenicity island at pos. 869,706 to 884,748 that included pif encoding a phage interference protein, a gene for a small terminase subunit, genes for “putative proteins” as well as a gene (scn2) coding for a paralog of a complement inhibitor SCIN family protein and a gene for a variant of the von Willebrand factor binding protein Vwb (vwb3). Thus, it is considered a staphylococcal pathogenicity island (SaPI) related to the one in S0385, GenBank AM990992.1.
Another prophage integrated between rpmF and isdB, pos. 1,107,447 to 1,146,132. A third prophage was located between a truncated nikB and Q5HG37, pos. 1,425,279 to 1,481,870. Finally, there was a forth prophage between Q5HDU4 and sarV (actually interrupting an MFS transporter between those genes), pos. 2,340,832 to 2,386,591. This prophage sequence corresponded to the phage that was detected by nanopore sequencing after induction by Mitomycin C (see below and Supplemental File 8).
Phage morphology and sequencing of phages from the CC1956 isolate WT19
In three separate preparations, large numbers of phages were observed that were well contrasted with uranyl acetate and with phosphotungstic acid. Phages had elongated capsids. The non-contractile thin tails were straight or slightly curved and ended in a bulb-shaped base plate. Based on these characteristics, they were assigned to the order Caudovirales, family Siphoviridae.
Capsids were measured in 40 phages, tails in 34 and base plates in 33 phages. Based on these measurements, two distinct populations could be differentiated (Fig. 5). In one (Fig. 5A), the prolate, distinctly pentagonal capsids averaged 39 ± 5 nm (range 32–46 nm) in diameter and 92 ± 8 nm (range 80–104 nm) in length. Tails were 276 ± 20 nm (range 243–310 nm) long, had a diameter of 11 ± 1 nm (range 10–12 nm) and had a stacked discs appearance. Their baseplates were 16 nm (range 16–31 nm) by 27 nm (range 19–33 nm). The other population (Fig. 5B) had elongated oval capsids with a maximal diameter of 55 ± 2 nm (range 51–60 nm) diameter and 93 ± 5 nm (range 85–100 nm) length. Their tails measured 287 ± 12 nm (range 275–313 nm) in length and 9 ± 1 nm (8–10) in diameter and had a rail-road-track morphology. Dimensions of baseplates were 25 nm (range 21–30 nm) by 29 nm (range 23–39 nm).
Transmission electron micrograph of two distinct prolate phages resulting from Mitomycin C treatment of S. aureus CC1956 isolate WT19. A, Phage particle with pentagonal 38 nm in diameter capsid and a 12 nm thick tail with stacked disc appearance; B, Two phage particles (1, 2) with oval capsids of 55 nm in diameter and 9 nm thick tails with rail-road-track morphology. The base plate is separated from the tail by a transversal disc (arrow). Negative contrast preparation with uranyl acetate. Bars = 100 nm.
Oxford Nanopore sequencing of one of these phage preparations (Supplemental File 8) yielded just one circular contig with a coverage of 724. Its sequence was identical to that of the forth prophage, between Q5HDU4 and sarV, except for a loss of a single triplet out of a total length of 46,387 nt.
Core genome and genomic islands of the CC49 isolate WT65
The CC49 isolate carried agr group II alleles and capsule type 5. It was positive for sasG, but lacked seh and egc enterotoxin genes, ORF CM14 and the collagen adhesion gene cna. A truncated copy of the enterotoxin S gene (GenBank CP000046, pos. 2,203,972.0.2,204,196) was found as well as leukocidin genes lukG/H = lukX/Y, lukD/E and lukF/S-hlg. With regard to presence and alleles of chromosomal markers such as MSCRAMM or ssl genes, the genome of WT65 (Supplemental Files 7a and 7b) is closely related to the CC49 reference sequences such as Tager 104, GenBank CP012409.1 (Supplemental File 2).
Mobile genetic elements in the CC49 isolate WT65
One prophage was integrated into the lip2 gene spanning pos. 311,401 to 354,724. The prophage included the lukF/S-BV genes as well as genes associated with the different modules of a typical Siphoviridae genome (Supplemental File 7/Fig. 4). Sequences corresponding to the lysogeny and replication modules were clearly different compared to the lukF/S-BV-prophage in the CC1956 isolate WT19 while approximately the second half of the two respective prophage sequences (the lower part of the alignment in Fig. 4) were virtually identical in gene content, order and orientation.
Other mobile genetic elements (Supplemental File 9a/b) included a small pathogenicity island, pos. 402,133 to 416,237 (between rpsR encoding 30S ribosomal protein S18 and its terminator), that included hypothetical proteins, a gene of a terminase small subunit, vwb3 (encoding a “von Willebrand factor” binding protein) and the scn2 gene (putative paralog of complement inhibitor). Between the genes ktrB and groL, pos. 2,029,208 to 2,042,866, another SaPI was identified that contained additional, slightly different copies of vwb3 and scn2 genes as well as terminase small subunit, integrase and excisionase (xis-AIO21657) genes. Finally, five genes between pos.1,334,169 and 1,339,503 were annotated as phage capsid genes although no other phage-related genes were found in this region.
Phage morphology and sequencing of phages from the CC49 isolate WT65
Four separate phage preparations were examined. In one of them, few phage-like structures were detected. These findings could not be confirmed in the following preparations. Thus, they were interpreted as artefacts, also given that it was not possible to induce a sufficient amount of phages for Oxford Nanopore sequencing.
Source: Ecology - nature.com
