Mammaliicoccus sciuri: The Resistome Behind Bovine Mastitis
All six mastitis-associated Mammaliicoccus sciuri genomes in a recent dairy survey carried mecA, yet only half of the isolates looked resistant on a standard cefoxitin disk test. The resistome spanned nine antibiotic classes, and two of the six isolates met the definition of multidrug resistance. Virulence and biofilm genes were widespread, but none of it proves this organism was the main cause of the mastitis it was found in.
Why Mammaliicoccus sciuri matters in dairy herds
Most of the damage mastitis does to a dairy business never shows up on the udder. Surveillance reports put annual incidence at roughly 20 to 40 percent of dairy animals worldwide, and subclinical cases make up the majority, often more than 60 to 70 percent of the cases in an affected herd. In Indian dairy systems the numbers run higher: subclinical prevalence is commonly reported between 40 and 60 percent, clinical mastitis nearer 10 to 20. Producers pay through lost yield, altered milk composition, discarded milk, veterinary costs and early culling.
The cast of pathogens behind those losses has also changed. Staphylococcus aureus still matters, but a lot of attention has moved to non-aureus staphylococci and mammaliicocci, shortened to NASM, now recognized contributors to both clinical and subclinical mastitis. These bacteria live on teat skin and mucosal surfaces, get into the gland opportunistically, and tend to settle in for persistent infections marked by elevated somatic cell counts.
Mammaliicoccus sciuri, formerly Staphylococcus sciuri, sits right in the middle of that shift. It is a Gram-positive, coagulase-negative member of the Staphylococcaceae, common on the skin and mucosal surfaces of animals and humans alike. Long dismissed as a commensal, it now reads as an opportunistic pathogen in bovine mastitis and, for anyone tracking antibiotic policy, a reservoir of antimicrobial resistance genes. The group carries determinants against beta-lactams, tetracyclines, macrolides, aminoglycosides and sulfonamides, and evolutionary work suggests mecA may have originated in this lineage before it spread into more pathogenic staphylococci such as S. aureus.
That is the part worth sitting with. If commensals and opportunists quietly maintain resistance determinants and pass them around, then surveillance that only watches the major pathogens is watching the wrong half of the picture.
How the study was run
The work screened lactating dairy cows on rural household farms in southern India, then sequenced the genome of every isolate recovered. The setup is below.
| Parameter | Detail |
| Animals screened | 54 lactating dairy cows, all four quarters sampled, 216 quarter milk samples |
| Mastitis screening | California Mastitis Test, reactions graded from negative to strong positive |
| Isolation | Enrichment in brain heart infusion broth, then mannitol salt agar, 37 degrees Celsius, 18 to 24 hours |
| Species confirmation | PCR of the species-specific gap gene, amplicon about 398 bp |
| Susceptibility testing | Kirby-Bauer disk diffusion, CLSI 2023 veterinary guidance, interpreted with coagulase-negative staphylococci breakpoints |
| Antibiotics tested | Nine agents: penicillin, cefoxitin, gentamicin, erythromycin, linezolid, chloramphenicol, enrofloxacin, tetracycline, trimethoprim-sulfamethoxazole |
| Sequencing | Illumina MiSeq, paired-end reads of 2 x 300 bp |
| Assembly | SPAdes via Shovill, contigs under 500 bp removed, quality checked with QUAST and CheckM |
| Resistance genes | CARD via the Resistance Gene Identifier, cross-checked with ResFinder 4.1 |
| Virulence genes | VFDB |
| Typing | MLST against the PubMLST scheme, species identity confirmed by average nucleotide identity |
| Phylogeny | 106 publicly available genomes, Roary core genes at 95 percent BLASTp identity, RAxML with 1,000 bootstrap replicates |
Two caveats travel with the methods. No breakpoints exist specifically for Mammaliicoccus, so disk results were read against coagulase-negative staphylococci criteria, with cefoxitin standing in as the phenotypic marker for methicillin resistance. And the enrichment step was there to maximize recovery of M. sciuri for genomic work, not to name the predominant mastitis pathogen. That distinction shapes how much weight the findings can carry.
Six isolates at a glance
Culture of the 58 mastitis-positive samples produced 58 bacterial isolates, and six of them, 10.34 percent, were confirmed as M. sciuri by the gap gene PCR and genome comparison. Their drug profiles span the full range in the study.
| Isolate | Mastitis status (CMT) | Resistant antibiotics | Classes resistant | Multidrug-resistant |
| A | Subclinical (++) | Penicillin, cefoxitin, gentamicin, erythromycin | 4 of 9 (44.44%) | Yes |
| B | Subclinical (++) | Penicillin, cefoxitin, gentamicin | 3 of 9 (33.33%) | Yes |
| C | Clinical (+++) | Penicillin, cefoxitin | 2 of 9 (22.22%) | No |
| D | Clinical (+++) | Penicillin, linezolid | 2 of 9 (22.22%) | No |
| E | Clinical (+++) | Penicillin | 1 of 9 (11.11%) | No |
| F | Clinical (++++) | Penicillin | 1 of 9 (11.11%) | No |
Three things catch the eye. Penicillin resistance was universal at 6 of 6, and the fuzzy edges around those disks hint at borderline beta-lactam resistance, the signature of beta-lactamase activity. Everything else was patchy, from a single drug in one isolate to four in another, an overall span of 11.11 to 44.44 percent of the nine agents tested. And the only linezolid-resistant isolate came from a clinical case, a useful reminder that last-line drugs do not get a free pass.
Both multidrug-resistant isolates, A and B, came from subclinical cases, while the three clinical isolates carried narrower profiles. Six isolates make that an anecdote, not a statistic, but it leans the same way as the rest of the survey: the quiet infections are hauling the heavier resistance loads.
What the Mammaliicoccus sciuri resistome contains
The genomes carried resistance determinants against nine antibiotic classes. Every gene, its class and its frequency is listed below.
| Gene | Resistance conferred | Isolates | Share |
| mecA | Methicillin and most beta-lactams | 6 of 6 | 100% |
| sal(A) | Lincosamides, MLS group | 6 of 6 | 100% |
| mprF | Peptide antibiotics | 6 of 6 | 100% |
| gidB | Aminoglycosides | 6 of 6 | 100% |
| folP | Sulfonamides | 6 of 6 | 100% |
| blaZ | Penicillin, via beta-lactamase | 4 of 6 | 66.7% |
| mph(C) | Macrolides | 5 of 6 | 83.3% |
| folA/dfr | Trimethoprim | 5 of 6 | 83.3% |
| sdrM | Multidrug efflux | 3 of 6 | 50% |
| sepA | Multidrug efflux | 3 of 6 | 50% |
| salC | MLS group | 3 of 6 | 50% |
| norA | Efflux, associated with fluoroquinolone tolerance | 2 of 6 | 33.3% |
| fosD/fosB and FosBx1 | Fosfomycin | 1 of 6 | 16.7% |
Two readings matter. The universal core, mecA, sal(A), mprF, gidB and folP, was present in every genome, so the baseline resistome of this organism is broad before anything acquired gets added. The variable tail, blaZ, mph(C), the efflux genes and the fosfomycin genes, came and went between isolates, which makes resistance breadth a per-isolate property here rather than a species constant.
Worth being clear about what a resistome is. The table is a gene inventory. It says nothing about expression levels, and it does not predict what a disk test will show, a point the beta-lactam results drive home hard.
Beta-lactams: mecA in every genome and blaZ in two thirds
The mecA gene was in every isolate, confirmed by PCR of the expected 162 bp fragment and by whole-genome screening, and the alternative methicillin resistance gene mecC was nowhere. Four of six genomes carried the blaZ gene, the beta-lactamase behind most penicillin resistance in staphylococci.
The beta-lactams also delivered the study’s cleanest lesson about method. Conventional PCR found blaZ in three isolates, A, B and E. Genome sequencing found it in four, adding isolate F. The likely culprit is sequence variation in a primer binding site that blocks amplification even with the gene sitting right there. Sequencing compares whole sequences and does not need a primer to anneal perfectly, so it catches what PCR misses. Neither method is wrong. They fail differently.
Macrolides, efflux and the rest of the resistome
Past the beta-lactams, the macrolide resistance gene mph(C) turned up in five of six genomes, the lincosamide gene salC in three, and sal(A) in all six. The trimethoprim determinants folA/dfr were in five of six and the sulfonamide gene folP in all six, so the folate pathway was covered panel-wide.
Efflux was the other recurring theme. sdrM and sepA each appeared in three of six genomes, norA in two of six. Efflux pumps rarely deliver outright resistance by themselves. They move enough drug out of the cell to buy tolerance, trimming susceptibility across several classes at once and keeping cells alive through fluctuating exposure. In a mammary gland that sees antibiotics come and go, that is a handy trait for a bacterium.
Fosfomycin resistance genes, fosD/fosB plus FosBx1, were confined to isolate A, the most broadly resistant of the six. Single-isolate findings like this are exactly why genome-level surveys exist: a drug that looks usable in five herds may not be usable in the sixth.
Virulence and biofilm genes
Resistance is half the story. The genomes also carried adhesion, capsule and biofilm determinants, listed below.
| Gene | Role | Isolates | Share |
| clpP | Protease, virulence and stress associated | 6 of 6 | 100% |
| mprF, pgsA, gdpD | Biofilm formation and membrane lipid modification | 6 of 6 | 100% |
| clfB | Adhesion to host tissue | 4 of 6 | 66.7% |
| sdrE | Adhesion, SD protein family | 4 of 6 | 66.7% |
| cap8B | Capsule biosynthesis | 3 of 6 | 50% |
| cap8D | Capsule biosynthesis | 3 of 6 | 50% |
| liaF, liaR, liaS | LiaFSR envelope stress response regulation | 5 of 6 | 83.3% |
| bceA, bceB, bceR | Bacitracin tolerance transporters and regulators | 5 to 6 of 6 | 83 to 100% |
The biofilm trio, mprF, pgsA and gdpD, was in every genome, and earlier work ties those genes to biofilm development. Biofilm matters at the udder because it shelters bacteria from antimicrobials and immune clearance, which fits the pattern of persistent intramammary infection and raised somatic cell counts. The LiaFSR stress-response regulators, present in five of six genomes, point the same direction: this organism is built to sense and survive envelope stress, the very kind antibiotic exposure creates.
One boundary needs stating. Biofilm capacity here is read from gene presence alone. No isolate was phenotyped for biofilm formation, so the genes show potential, not demonstrated behavior.
The counter-intuitive finding: mecA-positive yet cefoxitin-susceptible
Here is the finding that deserves the most attention, and it cuts against the usual shorthand. All six isolates carried mecA. Only three of six resisted cefoxitin on disk testing. The gene was everywhere; the phenotype was not.
That is the pattern known as oxacillin-susceptible mecA-positive staphylococci, reported more and more in human and veterinary microbiology. It matters because the mental shortcut, mecA present means methicillin resistant, fails often enough to change decisions. The study listed candidate explanations without testing any of them: altered regulation of mecA expression through the mecI-mecR1 system, incomplete or unusual SCCmec elements that impair stable expression, heteroresistance where only part of the population expresses resistance, and environmental influences such as antibiotic exposure and growth conditions. All of them remain hypotheses.
The consequences run both ways. Phenotypic testing alone will underestimate how much methicillin resistance is circulating, which feeds underreporting in surveillance and can steer therapy toward drugs that look safe on paper. Molecular testing alone overstates the problem if a silent gene gets read as active resistance. The defensible move is to run both and read them together, which is also the right way to see the blaZ discrepancy in isolate F: detection method changes the answer, and knowing the method is part of knowing the result.
Heavy metals, prophages and CRISPR-Cas
The genomes carried more than antibiotic resistance. Arsenic resistance genes, including the ars operon for arsenate reduction and arsenite efflux, were in most isolates, the cadmium resistance gene cadD in three, and sh-fabI in five. Metal and antibiotic resistance genes turning up together fits the idea of co-selection, where metal exposure helps keep antibiotic resistance alive because the determinants travel as a package or share regulatory ground. In livestock systems, feed, water and farm waste have all been proposed as the metal sources holding that pressure on.
Genome plasticity showed up in the mobile elements too. Three of six genomes carried prophage regions, including intact regions of 55.7 kb, 20.2 kb and 14.1 kb, while the other three carried none. These genomes are gaining and losing material in real time. All six carried CRISPR arrays, four carried cas3, and two carried both cas3 and csx16, corresponding to Type I and Type III-U systems. No plasmids turned up in any genome.
| Feature | Finding |
| Prophages | 3 of 6 genomes; intact regions of 55.7 kb, 20.2 kb and 14.1 kb, plus one incomplete region of 17.5 kb |
| CRISPR arrays | All 6 genomes, 2 to 6 arrays each, repeats of 23 to 55 bp |
| Cas genes | cas3 in 4 genomes (Type I); cas3 plus csx16 in 2 genomes (Type I and Type III-U) |
| Plasmids | None detected |
| Heavy metal genes | Arsenic resistance genes in most genomes, cadD in 3, sh-fabI in 5 |
Phylogeny and MLST: six isolates, several lineages
The six isolates were not one clone. Core-genome phylogeny dropped them into several distinct branches, and sequence typing agreed. Two got defined types, ST49 and ST335. Four could not be assigned because short-read assembly recovered only three to five of the seven housekeeping loci, and their closest known types were ST48, ST183, ST9 and ST197.
| Isolate | Assigned sequence type | Closest known type |
| B | ST49 | Not applicable |
| F | ST335 | Not applicable |
| A | Not assigned | ST48 |
| C | Not assigned | ST183 |
| D | Not assigned | ST9 |
| E | Not assigned | ST197 |
Line those six up against 106 public genomes from cattle, poultry, goat, human, pig, sheep and zebu cattle across thirteen countries and the study isolates scatter through the global phylogeny instead of forming a local cluster. Some sit with bovine-associated lineages, consistent with their dairy origin, but they mix freely with genomes from human, environmental and other animal sources. That pattern supports genetic diversity and broad host association. It does not demonstrate cross-host transmission, and a cross-sectional design with no epidemiological linkage or SNP-based transmission analysis could not have. Surveys like this one are the raw material for One Health surveillance of antimicrobial resistance across the animal, environment and human interface.
What this means for control and surveillance
For herds and clinicians:
- Treat penicillin as an unreliable choice against this organism. Every isolate resisted it, and the fuzzy zone edges suggest beta-lactamase activity on top of mecA.
- Base therapy on culture and susceptibility results, especially for subclinical cases, which in this survey carried the heaviest resistance loads.
- Do not assume last-line drugs are untouchable. One isolate resisted linezolid.
- Screen for subclinical mastitis with the California Mastitis Test and somatic cell counts, since visible signs miss most cases.
For surveillance programs:
- Pair phenotypic testing with mecA detection for non-aureus staphylococci and mammaliicocci. Disk testing alone will undercount methicillin resistance whenever phenotype-genotype discordance is in play.
- Record the detection method alongside the result. The same isolate can be blaZ-negative by PCR and blaZ-positive by sequencing.
- Read metal resistance genes as a co-selection signal, not a curiosity. They point to selection pressure that antibiotics alone do not explain.
- Keep region-specific genomic surveys running. Six isolates from one district already held two sequence types, thirteen resistance genes and several lineages; national aggregates would have smoothed all of that away.
Limitations
The honest limits matter more than usual here, because the dataset is small and the methods shape the answers.
- Six isolates cannot describe a population. Every percentage in this article describes those six genomes, not dairy herds at large.
- The enrichment culture on mannitol salt agar was chosen to recover M. sciuri, and unlike standard National Mastitis Council protocols it may preferentially enhance non-aureus staphylococci. Finding the organism in mastitis-positive milk does not establish it as the primary cause of infection.
- Susceptibility results were interpreted with coagulase-negative staphylococci breakpoints because no Mammaliicoccus-specific criteria exist, and oxacillin MIC testing and PBP2a detection were not performed.
- Short-read assembly could not resolve SCCmec elements, plasmids or the mobile context of resistance genes, and it left four of six MLST profiles incomplete.
- No transcriptomic or proteomic work was done, so gene presence says nothing about expression. The discordance between mecA and cefoxitin resistance is documented here, not explained.
- The cross-sectional design supports no conclusions about transmission, and the phylogenetic mixing with human and environmental genomes reflects shared ancestry as much as anything else.
Bottom line
A region-specific genome survey of mastitis-associated Mammaliicoccus sciuri found a broad, nine-class resistome anchored by universal mecA, a virulence and biofilm repertoire built for persistence, and a phylogenetically diverse, non-clonal set of isolates. The most useful finding is the discordance: the resistance gene was universal while the resistant phenotype was not, so any surveillance or treatment decision built on one test alone rests on half the evidence.
FAQ
What is the Mammaliicoccus sciuri resistome?
The resistome is the complete set of antimicrobial resistance genes a bacterium carries. In these six mastitis-associated genomes it spanned nine antibiotic classes, including mecA, blaZ, sal(A), mph(C) and several efflux genes. A resistome is always broader than the disk test suggests, because gene presence does not guarantee visible resistance.
Why are all mecA-positive isolates not cefoxitin-resistant?
All six isolates carried mecA, but only three of six resisted cefoxitin on disk testing. This phenotype-genotype discordance matches oxacillin-susceptible mecA-positive staphylococci reported in human and veterinary laboratories. The study did not measure mecA expression or PBP2a production, so the mechanism behind it remains unconfirmed.
What is phenotype-genotype discordance in staphylococci?
It means a resistance gene is present while the isolate still tests susceptible by standard methods. Proposed explanations include suppressed mecA regulation, incomplete SCCmec elements and heteroresistance within the bacterial population. Disk and dilution results should be read together with gene detection rather than instead of it.
Is Mammaliicoccus sciuri resistant to penicillin?
In this survey every isolate resisted penicillin on disk testing, making penicillin resistance universal across the panel. Fuzzy zone edges around the disks pointed to borderline beta-lactam resistance, consistent with beta-lactamase activity. Penicillin is therefore a poor empirical choice whenever this organism is a plausible cause of infection.
Which antibiotics were the mastitis isolates still susceptible to?
All six isolates stayed susceptible to chloramphenicol, enrofloxacin, tetracycline and trimethoprim-sulfamethoxazole on disk testing. Resistance clustered in penicillin, cefoxitin, gentamicin, erythromycin and linezolid instead. Susceptibility varied by isolate rather than by herd, so culture results from the individual animal should drive drug choice.
Were any of the isolates multidrug-resistant?
Two of the six isolates met the definition of multidrug resistance, non-susceptibility to at least three antimicrobial categories. They resisted four and three of the nine antibiotics tested, and both came from subclinical cases. The clinical isolates in the same survey carried narrower profiles, so the heaviest resistance loads sat in the quietest infections.
Can disk diffusion testing miss methicillin resistance?
Yes, whenever a mecA-carrying isolate still grows inside the cefoxitin breakpoint. In this survey that applied to three of the six isolates, every one of them mecA-positive by both PCR and genome sequencing. Pairing disk testing with molecular detection of mecA reduces the risk of underreporting methicillin resistance in surveillance data.
What virulence and biofilm genes does Mammaliicoccus sciuri carry?
All six genomes carried clpP plus the biofilm-associated genes mprF, pgsA and gdpD. Adhesion genes clfB and sdrE appeared in four of six, and capsule genes cap8B and cap8D in three of six. Biofilm capacity was inferred from gene presence only and was not measured phenotypically in any isolate, so the genes indicate potential rather than demonstrated behavior.
Why did PCR and whole-genome sequencing disagree on blaZ?
One isolate tested negative by PCR but positive by whole-genome sequencing. Sequence variation in the primer binding site can block amplification even when the target gene is present in the genome. Genome sequencing compares whole sequences rather than relying on a single primer pair, so it is less exposed to that particular failure mode.
How is subclinical mastitis detected in dairy herds?
Subclinical mastitis shows no visible udder changes, so screening relies on tests like the California Mastitis Test and somatic cell counts rather than on clinical signs. In this survey 58 of 216 quarter samples were positive, a quarter-level positivity of 26.85 percent. Trace and weak reactions on the paddle were treated as suggestive of subclinical infection.
Do these isolates show that M. sciuri spreads between cattle and humans?
No transmission conclusion is supported by this work, only relatedness. The isolates clustered with genomes from cattle, humans and other animals, but clustering reflects shared ancestry rather than proven cross-host spread. The cross-sectional design had no epidemiological linkage or transmission analysis to test movement between hosts.
What are the limitations of whole-genome studies with only six isolates?
Six isolates cannot describe a population, so the percentages here are descriptive rather than estimates. Short-read assembly left SCCmec elements, plasmids and some MLST loci unresolved. Enrichment culture also biases recovery toward non-aureus staphylococci, so these isolates may not have been the primary mastitis pathogens in every case.
