Whole genome sequencing of methicillin-resistant and methicillin-sensitive Staphylococcus aureus isolated from 4 horses in a veterinary teaching hospital and its ambulatory service

Abstract

Genomic characterization was conducted on 2 methicillin-resistant Staphylococcus aureus (MRSA) strains isolated from 2 horses hospitalized during an overlapping period of time and 2 methicillin-sensitive S. aureus (MSSA) strains isolated from 2 distinct horses. Phylogenetic proximity was traced and the genotypic and phenotypic characteristics of the antimicrobial resistance of the strains were compared.

Whole genome sequencing of MRSA strains for this report was similar but differed from whole genome sequencing of MSSA strains. The MRSA strains were closely related, belonging to sequence type (ST) 612, spa type t1257, and SCCmec type IVd2B. The MSSA strains were also closely related, belonging to ST1660, spa type t3043, and having no detectable staphylococcal cassette chromosome mec elements. All MSRA and MSSA strains were Panton-Valentine leukocidin negative. There were discrepancies in the genotypic analysis and the antimicrobial susceptibility testing (phenotypic analysis) of MRSA strains for rifampin, trimethoprim-sulfamethoxazole, gentamicin, amikacin, and enrofloxacin.

Infections caused by methicillin-resistant Staphylococcus aureus (MRSA) are a serious concern in equine medicine. Infections involving MRSA in horses have a clinical impact, represent a risk for nosocomial diseases and have the potential to be transmitted to humans (1–5). Unfortunately, there is a limited number of antimicrobials approved for use in horses in Canada to address such important infectious diseases.

To the authors’ knowledge, whole genome sequencing (WGS) of Canadian equine MRSA has not been described. The WGS of 2 MRSA strains isolated from 2 horses hospitalized at the veterinary teaching hospital (VTH) of the Faculty of Veterinary Medicine of the Université de Montréal, during an overlapping period is described. Two equine methicillin-sensitive S. aureus (MSSA) strains were described, to compare MRSA and MSSA genomic characteristics. The genotypic and phenotypic characteristics for antimicrobial resistance of these S. aureus strains were compared and their phylogenetic relatedness was assessed.

Two frozen MRSA strains collected from 2 horses hospitalized at the VTH over the same period in 2017 (71 d overlapping period) were selected. The horses shared the same facilities and were kept in separate stalls with the same hospital staff handling them. The first MRSA sample (MRSA 1) was taken from the left hind limb fetlock wound of an 8-year-old Quarter Horse mare hospitalized at the VTH for recurrent lymphangitis. Prior to admission, this horse was treated with intravenous (IV) penicillin (day −35 to day −27), IV gentamicin (day −35 to day −32), IV enrofloxacin (day −32 to day −19), and oral chloramphenicol (day −28 to day −12) based on bacterial culture results (MRSA and Enterobacter cloacae) and antimicrobial susceptibility testing (AST). Upon admission (day 0), a sample was taken for a bacterial culture and AST. While the results of the bacterial culture and AST were pending, the horse received 4 d of IV enrofloxacin and oral chloramphenicol (day 0 to day 3). Methicillin-resistant S. aureus, Proteus vulgaris, Pseudomonas aeruginosa, and Streptococcus equi spp. zooepidemicus were identified on the bacterial culture. Based on the results of the AST, the clinical findings, and a second bacterial culture performed 26 d following admission (day 25), IV ceftiofur and oral chloramphenicol were administered for 30 d (day 4 to day 33) and 69 d (day 0 to day 68), respectively.

The second MRSA sample (MRSA 2) was taken from a chronic wound between the right tuber coxae and ischium of a 25-year-old Hanoverian mare. Before her admission, trimethoprim-sulfamethoxazole (TMS) (unknown route of administration) was administered for 7 d (day −18 to day −12), followed by intramuscular penicillin and enrofloxacin (unknown route of administration) for 7 d (day −11 to day −5), and trimethoprim-sulfadiazine (unknown route of administration) and enrofloxacin (unknown route of administration) for 5 d (day −4 to day 0), due to treatment failures (i.e., antimicrobial treatments were not based on a bacterial culture and AST results). Upon admission, a sample was submitted to the bacteriology laboratory for bacterial culture, but no growth was observed. A second sample was obtained for bacterial culture 14 d after being admitted (day 13), which revealed a pure culture of MRSA. During the hospitalization, the horse was treated with oral enrofloxacin and metronidazole for 20 d (day 0 to day 19), followed by oral chloramphenicol for 10 d (day 20 to day 30).

Two frozen MSSA strains (MSSA 1 and 2) sampled in 2016 were also sequenced. Methicillin-sensitive S. aureus 1 was from a skin sample of a horse hospitalized at the VTH for colic and presenting papules. Methicillin-sensitive S. aureus 2 was isolated from a nasal swab sampled by the ambulatory practice of the VTH.

All specimens were obtained using sterile cotton swabs (BBL CultureSwab; Becton Dickinson Canada, Mississauga, Ontario) with Stuart Transport Medium (Becton Dickinson Canada).

The samples were cultured aerobically. A single colony subculture was frozen at −80°C in tryptic soy broth (CM0129) (Oxoid Ltd, Basingstoke, England) with 15% glycerol. The S. aureus clone was inoculated on Difco™ Columbia blood agar base (279240) with 5% sheep blood (Becton Dickinson and Company, Sparks, Maryland, USA) and incubated at 35°C ± 2°C with 5% carbon dioxide. Bacteria were identified with conventional biochemical tests for MSSA 1 and the matrix-assisted laser desorption ionization time-of-flight mass spectrometry method was used for MSSA 2 and both MRSA. The same clone was used for sequencing.

All AST was repeated in September 2019 using BBL™ Mueller Hinton II agar (Becton Dickinson and Company, Sparks, Maryland, USA) for the Kirby-Bauer disk diffusion method, according to the most recent guidelines published by the Clinical Laboratory Standards Institute (CLSI). If CLSI criteria were not available for horses or for bacteria isolated from other animals, human criteria were used for the interpretation. Clinical Laboratory Standards Institute criteria specific to horses were only available for ceftiofur and enrofloxacin. All samples were tested for the following antimicrobials: amikacin, ampicillin, cefoxitin, ceftiofur, chloramphenicol, enrofloxacin, erythromycin, gentamicin, penicillin G, rifampin, tetracycline, and TMS. Strains phenotypically resistant to cefoxitin were considered MRSA.

Bacterial DNA was extracted using the DNeasy Blood & Tissue Kit (Qiagen, Germantown, Maryland, USA), as described in the company’s protocol. DNA was quantified using the Qubit dsDNA High Sensitivity Assay Kit (ThermoFisher Scientific, Waltham, Maine, USA) and a Qubit Fluorometer (ThermoFisher Scientific). The next-generation sequencing libraries were synthesized with the Nextera XT DNA Library Preparation Kit (Illumina, San Diego, California, USA), as described in the Nextera XT protocol. The quality of the library was assessed using the Agilent High Sensitivity DNA Kit (Agilent, Santa Clara, California, USA) in a Bioanalyzer (Agilent). Libraries were sequenced in a v3 600-cycle cartridge using a MiSeq instrument and PhiX Control at around 1% (Illumina).

Reads were trimmed by the MiSeq system together with the CLC Genomics Workbench software (CLC-GW) (version 12.0.3; Qiagen, Redwood City, California, USA). De novo analysis was performed using the De Novo Assemble Metagenome application in the CLC-GW and confirmed by SPAdes Genome Assembler in BaseSpace (Illumina). Contigs were used to prepare the phylogenetic trees and the Bacterial Analysis Pipeline GoSeqIt ApS. All genome contig sequences for the 4 cases described in this paper are available in GenBank with the BioSample accession number SAMN15351203-6 for strains FMV17-17660 (MRSA 1), FMV17-17156 (MRSA 2), FMV16-5587 (MSSA 1), and FMV16-7294 (MSSA 2).

The genotypic characterizations of isolates [sequence type (ST), spa typing schemes, staphylococcal cassette chromosome mec (SCCmec) type, Panton-Valentine leukocidin (PVL), leucocidin-F and -S subunits (lukF, lukS), staphylococcal complement inhibitor (scn), staphylokinase (sak), clumping factor A and B (cflA, cflB), fibronectin binding protein (fnBP), collagen adhesin (cna), elastin-binding protein EbpS (ebpS), extracellular adherence protein (eap), intercellular adhesion protein icaA (icaA), intercellular adhesion protein icaB (icaB), intercellular adhesion protein icaC (icaC), intercellular adhesion protein icaD (icaD), and phospholipase C (hlb)] were acquired by performing genomic comparisons with other complete S. aureus genomes in the Center for Genomic Epidemiology public database (https://cge.cbs.dtu.dk/services/) (6,7).

The online version of the ResFinder 3.2 tool (https://cge.cbs.dtu.dk/services/ResFinder/database.php) was used to identify the antimicrobial resistance genes, the single-end and paired-end reads, and the FastQ file, as previously described (8). The classes of antibiotics of interest reported were phenicol, β-lactam antibiotics, quinolone, tetracycline, trimethoprim, macrolide, aminoglycoside, sulfonamide, and rifampin (9–11).

For the phylogenetic analysis, the whole genome sequenced strains were compared to others available on GenBank (https://www.ncbi.nlm.nih.gov/genbank/) (Table I). A neighbor-joining phylogenetic tree with whole genome sequences was created using Comparison (beta) 1.0 based on the average nucleotide identity calculation from the CLC Genomics Workbench (Qiagen) Whole Genome Alignment 20.0 plugin. Phylogenetic analysis using the core genome multilocus sequence typing (cgMLST) approach based on the 1865 cgMLST loci in the scheme for S. aureus was performed using BacWGSTdb 2.0 (http://bacdb.cn/BacWGSTdb/).

Table I.

Reference strains.

Host specificity	Isolation site	Collection date	Sample type	Isolation source
Equus caballus	Germany	2016	Pure culture	Ocular surface swab
Equus caballus	Germany	2016	Pure culture	Ocular surface swab
Equus caballus	Germany	2014	Nasal swab	N/A
Equus caballus	Germany	2014	Fecal sample	N/A
Equus caballus	Germany	2014	Nasal swab	N/A
Equus caballus	Germany	2016	Pure culture	Ocular surface swab
Equus caballus	Germany	2015	Nasal swab	N/A
Sus scrofa domesticus	Germany	2012	Cell culture	N/A
Homo sapiens	United Kingdom	2003	N/A	Blood
Homo sapiens	Australia	2009	N/A	Vet Hospital
Not Available (N/A)	USA	N/A	N/A	N/A
Homo sapiens	Saudi Arabia	2017	N/A	N/A
Equus caballus	Japan	2018	N/A	N/A
Homo sapiens	Jordan	2007	N/A	Wound

The genotypic characteristics of the S. aureus strains are summarized in Table II. The ST obtained from MRSA 1 and 2 was similar (ST612). Both MRSA strains harbored the SCCmec type IVd2B. Methicillin-sensitive S. aureus 1 and 2 displayed the same ST (ST1660) and did not harbor the SCCmec gene. The spa types identified for both MRSA strains and for MSSA 1 were type t1257 and t3043, respectively; the spa type for MSSA 2 was unidentified.

Table II.

Equine methicillin-resistant S. aureus (MRSA) and methicillin-sensitive S. aureus (MSSA) genomic characteristics identified in 4 strains.

	Virulence factors
	mecA	Toxins	Immune evasion			Adherence factors
		lukF, lukS	scn	sak	hlb	clfA, clfB	fnBP	ebpS	eap	icaA, icaB, icaC, icaD	cna
MSSA 1
MSSA 2
MRSA 1
MRSA 2

luk — leucocidin; scn — staphylococcal complement inhibitor; sak — staphylokinase; hlb — beta-hemolysin; clf — clumping factor; fnBP — fibronectin-binding protein; ebpS — elastin-binding protein; eap — extracellular adherence protein; ica — intercellular adhesion protein; cna — collagen adhesin.

gene absent

gene identified

non-conclusive result

The same genetic patterns for virulence factors and antimicrobial resistance were identified on both MRSA strains. Methicillin-resistant S. aureus 1 and 2 were positive for mecA, scn, sak, clfA, clfB, ebpS, eap, icaA, icaB, icaC, icaD, and hlb genes. It was inconclusive for the fnBP gene. Both MSSA strains were positive for clfA, clfB, fnBP, cna, ebpS, eap, icaA, icaB, icaC, icaD, and hlb genes. All S. aureus were PVL negative.

The identified antimicrobial resistance genes were in accordance with the AST results, except for amikacin, rifampin, TMS, and enrofloxacin for MRSA 1 and 2. The results are presented in Table III.

Table III.

Genotypic and phenotypic characteristics of antimicrobial resistance in equine methicillin-resistant S. aureus (MRSA) (n = 2) and methicillin-sensitive S. aureus (MSSA) (n = 2). Italicized words represent discrepancies between the genotypic and phenotypic patterns of antimicrobial resistance.

	MSSA 1	MSSA 2	MRSA 1	MRSA 2
Genotypic analyses	Absence of resistance genes	Absence of resistance genes	P: β-lactam antibiotics	P: β-lactam antibiotics
			N: phenicols	N: phenicols
			N: quinolones	N: quinolones
			P: tetracyclines	P: tetracyclines
			P: macrolides	P: macrolides
			P: aminoglycosides	P: aminoglycosides
			N: rifampin	N: rifampin
			N: trimethoprim	N: trimethoprim
			N: sulphonamid	N: sulphonamid
Phenotypic analyses (AST)	Absence of antibiotic resistance	Absence of antibiotic resistance	R: β-lactam antibiotics	R: β-lactam antibiotics
			S: chloramphenicol	S: chloramphenicol
			I: enrofloxacin	I: enrofloxacin
			R: tetracycline	R: tetracycline
			R: erythromycin	R: erythromycin
			S: amikacin	S: amikacin
			R: gentamicin	I: gentamicin
			R: rifampin	R: rifampin
			R: TMS	R: TMS

AST — antimicrobial susceptibility testing; P — positive for the gene; N — negative for the gene; S — susceptible; I — intermediate; R — resistant.

The genes identified for each S. aureus and their pattern of branching in the phylogenetic tree suggests that the MRSA and MSSA strains of this report are distinct. The phylogenetic trees based on the whole genome (Figure 1) and based on the core genome indicate that both MRSA strains are closely related. These 2 different ways of comparing the strains are 2 alignment methods currently used in the literature (12–14). Whole genome alignment compares full length genomes and plasmid sequences, while core genome alignment compares the set of genes shared by all strains in a species (i.e., 1865 cgMLST loci for S. aureus) that mostly encode functions related to basic cellular biology (12–14). The MRSA strains are not exactly on the same branch, but they belong to a common sub-cluster and harbor similar genes. Indeed, they harbored the same ST, spa type, and SCCmec, and were negative for PVL. Moreover, the genes associated with virulence factors and antimicrobial resistance were similar.

Figure 1.

a — Neighbor joining phylogenetic tree of FMV17-17660 (MRSA 1), FMV17-17156 (MRSA 2), FMV16-5587 (MSSA 1), FMV16-7294 (MSSA 2), and whole sequenced S. aureus genomes of different species. The more genetic characteristics shared between strains, the closer they are in the phylogenetic tree. Each isolate is listed as strain name, species, country of origin, and sequence type (ST). The scale bar indicates the percentage of difference in the nucleotide sequence based on the average nucleotide identity. * Represents the nearest ST. N/A — Not available. b — Phylogenetic analysis was performed by using the core genome multilocus sequence typing (cgMLST) approach based on 1865 cgMLST loci. BacWGSTdb 2.0 was used to construct and visualize the minimal spanning tree generated based on the cgMLST allelic profiles, which supports manipulations of both tree layout and the user specified metadata attributes. The lines connecting the circles indicate the clonal relationship between different isolates and the numbers on the connecting lines illustrate the numbers of allelic differences. Numbers below 70 are not shown on the Figure. The scale bar represents a pairwise allelic difference of 400 cgMLST loci (13,14).

The genotypic characteristics of Canadian equine MRSA (Canadian epidemic MRSA-5) are partially reported (Table IV) (15,16). Canadian epidemic MRSA-5 was reported as ST8 and spa type t064, which differs from the reported MRSA strains (ST612 and spa type t1257). Canadian epidemic MRSA-5 and both reported MRSA strains were SCCmec type IV and PVL negative. These results add new information regarding the epidemiological data relating to the circulating equine MRSA strains in Canada.

Table IV.

Genetic comparisons between the reported S. aureus strains and the Canadian epidemic methicillin-resistant S. aureus-5 (CMRSA-5) (15).

	MSSA 1	MSSA 2	MRSA 1	MRSA 2	CMRSA-5
ST	1660	1660	612	612	8
Spa type	t3043	unknown	t1257	t1257	t064
SCCmec	Negative	Negative	IVd2B	IVd2B	IV
PVL	Negative	Negative	Negative	Negative	Negative

MSSA — methicillin-sensitive S. aureus; ST — sequence type; SCCmec — staphylococcal cassette chromosome mec; PVL — Panton-Valentine leukocidin.

Interestingly, the MRSA strains in this report share some common genotypic characteristics with the ST612 Australian strain [SVH7513(horse)AUS] (17). Worldwide, the ST612 strain has only been isolated in horses in Australia (17). As for SVH7513(horse)AUS, the MRSA strains reported here also share some similar genotypic characteristics with the human MRSA USA500 strains previously reported in North America. However, a whole genome comparison between the reported MRSA strains and the ST612 Australian and human MRSA USA 500 strains would be required to clearly identify the homology between these MRSA strains. The MRSA USA500 strain (ST8; spa types t064, t008, or t211; SCCmec IV; PVL negative) is responsible for both community- and hospital-associated infections, and is considered to be a significant infectious agent (4,18,19).

We report the staphylococcal genes identified, but their expression was not studied. Therefore, the clinical relevance of the virulence factors identified is unknown. All S. aureus strains in this report harbor genes responsible for adhesion, colonization, and invasion, as well as the ability to form a biofilm. They have the potential to evade the immune system by interfering with the complement cascade, neutrophil function, and diapedesis. They also harbor genes responsible for cytolytic toxins (4,20). The reported MRSA strains harbor genes responsible for the inhibition of the complement system and the degradation of immunoglobulin opsonins (sak, scn) (4). These characteristics may influence the severity of the disease if they are expressed.

The results of the genotypic and phenotypic antibiotic resistance profiles for each MRSA strain show discrepancies. This could be explained by a consensus sequence for the genotypic antibiotic resistance analyses, but this was not the methodology used in this report. Another explanation could be antibiotic resistances carried by mechanisms such as genetic mutations (rifampin, TMS, and enrofloxacin) or unexpressed genes of resistance (amikacin), which were not studied in this report (21). A limitation to the AST results is the absence of horse-specific cutoff values for the interpretation of the antimicrobial resistance of all antimicrobials tested. Although it is unlikely that the discrepancies between the genotypic and phenotypic antibiotic resistance profiles are a consequence of the absence of horse-specific cutoff values for all tested antimicrobials, this might have influenced the interpretation of the AST.

The use of multiple classes of antibiotics for a prolonged period in these horses is likely to be one of the major contributing factors to the pattern of resistance of the MRSA strains described in this report (22).

Considering the phylogenetic relatedness of MRSA 1 and 2, an indirect transmission from one horse to the other cannot be excluded (common shared areas, common hospital handlers), despite the application of standard infection control measures.

In conclusion, complete genome sequencing tools are useful in hospital settings to provide epidemiological information and may assist veterinary hospitals in improving biosecurity and antimicrobial stewardship (23,24).