Addictive plasmids drive hospital transmission of mupirocin-resistant Staphylococcus aureus
aDepartment of Medicine, Division of Infectious Diseases, NYU Grossman School of Medicine, New York, NY, USA
bDepartment of Microbiology, NYU Grossman School of Medicine, New York, NY, USA
cAntimicrobial-Resistant Pathogens Program, NYU Grossman School of Medicine, New York, NY, USA
dDepartment of Population Health, NYU Grossman School of Medicine, New York, NY, USA
eClinical Microbiology and Diagnostic Immunology, NYU Grossman School of Medicine, New York, NY, USA
fPublic Health Research Institute, New Jersey Medical School, Rutgers University, Newark, NJ, USA
gDepartment of Microbiology, Biochemistry & Molecular Genetics, New Jersey Medical School, Rutgers University, Newark, NJ, USA
*Correspondence: Bo Shopsin MD, PhD; 430 E 29th St, New York, NY 10016; bo.shopsin@nyulangone.org, Sarah E. Hochman MD; 545 1st Ave, SC1-173, New York, NY 10016; sarah.hochman@nyulangone.orgSUMMARY
Background
Mupirocin, a widely used topical agent for decolonization of Staphylococcus aureus, is increasingly compromised by resistance. Although plasmid-mediated mupirocin resistance is a recognized cause of decolonization failure, its role in facilitating hospital-wide transmission is unknown.
Methods
We conducted genomic surveillance of S. aureus at two interconnected urban hospitals where mupirocin decolonization is routine. Genome sequencing of >10,000 isolates was integrated with patient data to identify transmission and resistance determinants. Bacterial phenotypes and fitness were evaluated in vitro and in murine colonization models.
Findings
Genome sequencing identified 475 hospital transmission events; none were detected by conventional surveillance. The mupA (ileS2) resistance determinant, carried on conjugative plasmids, was enriched eightfold in methicillin-resistant S. aureus (MRSA) relative to methicillin-susceptible strains. mupA was associated with nearly a threefold greater chance of hospital transmission, especially within endemic healthcare-associated MRSA lineages, and was enriched twofold in hospital-onset infections compared with admission colonizing isolates. Multiple independently evolved inactivating mutations in the essential chromosomal gene ileS1 co-occurred with mupA, creating plasmid addiction in which mupA became indispensable for bacterial survival. Addiction arose most frequently within the dominant community-acquired MRSA lineage, where plasmid carriage reduced colonization fitness in mice. Plasmid-containing strains exhibited stringent-response activation, explaining the fitness costs and collateral tolerance to disinfectants, such as ethanol and peroxide. Although addiction reduced S. aureus fitness, it increased plasmid transfer, and addicted variants spread across hosts, demonstrating adaptation that mitigates these costs. Unexpectedly, we identified a mupirocin–dependent vulnerability to isoleucine limitation, revealing a potential strategy to target mupA-mediated resistance.
Interpretation
Plasmids promote hospital transmission of mupirocin-resistant S. aureus and create an evolutionary trap in which antibiotic use selects for bacterial dependence on otherwise costly resistance elements. This dependence revealed a collateral bacterial vulnerability that could be exploited to target resistant strains and preserve the effectiveness of mupirocin.
Funding
National Institutes of Health.
Research in context
Evidence before this study
We searched PubMed for articles published in any language from database inception to July 2025 using the terms “Staphylococcus aureus,” “MRSA,” “mupirocin,” “chlorhexidine,” “resistance,” “plasmid,” “addiction,” and “transmission.” We also reviewed the reference lists of relevant studies. Previous work showed that mupirocin resistance, mediated either by non-inactivating chromosomal ileS1 mutations or plasmid-encoded mupA genes, decreases the success of S. aureus decolonization efforts. However, no study had systematically examined how plasmid-mediated mupirocin resistance affects S. aureus transmission within hospitals. Existing literature describes fitness costs of mupirocin plasmids, but not mechanisms that enforce plasmid maintenance through gene essentiality. Additionally, the relationship between mupirocin resistance, stringent response activation, cross-tolerance to other disinfectants, and collateral vulnerabilities has not been reported.
Added value of this study
This study provides the first comprehensive genomic evidence that plasmid-mediated mupirocin resistance directly contributes to S. aureus transmission in hospitals. By sequencing thousands of isolates from two interconnected hospitals, we show that plasmids encoding mupA (ileS2) are strongly associated (11·3% increase in nosocomial transmission, 95% CI 5.8–16·9) with nosocomial spread. We further identify a previously undescribed form of plasmid addiction caused by inactivation of the essential chromosomal gene ileS1, rendering plasmid-encoded ileS2 indispensable for survival. Addiction did not itself enhance strain fitness but stabilizes otherwise costly resistance elements, enabling their continued transmission and dissemination. By revealing an unexpected dependence of MRSA on a resistance plasmid, our findings identified a collateral vulnerability to isoleucine limitation that could be leveraged to sustain the effectiveness of mupirocin.
Implications of all available evidence
Our findings highlight a crucial paradox: mupirocin decolonzation works—susceptible strains transmit less—but its use selects for a previously unappreciated form of plasmid-addicted strain having cross-tolerance to multiple disinfectants. Addiction helps explain the maintenance of resistance plasmids that drive MRSA spread within hospitals. At the same time, resistance-fitness interactions that create genetic dependencies also expose collateral vulnerabilities, providing a rationale for resistance-breaking adjuvant strategies aimed at preserving the effectiveness of mupirocin. By showing how antimicrobial use can create irreversible genomic dependencies, this study also reframes infection-control strategies toward proactive genomic surveillance to identify and mitigate the unintended consequences of mupirocin use. More broadly, the work challenges the assumption that reducing antibiotic exposure alone will reverse resistance once genetic dependence has evolved.
Article notes
Competing Interest Statement
Declaration of interests: B.S. has served on a scientific advisory board for Innoviva Specialty Therapeutics and has received research funding from Analog Devices Inc.
INTRODUCTION
Nasal mupirocin, often combined with chlorhexidine bathing, is widely used for decolonization and control of Staphylococcus aureus, especially methicillin-resistant strains (MRSA).1,2 These interventions help sustain recent reductions in healthcare-associated infections.3 Although mupirocin use selects for resistance,4 its effects on hospital-wide transmission and clonal dynamics are unknown. As the principal decolonizing agent worldwide, the emergence and spread of resistance threaten infection prevention in vulnerable patients.
Mupirocin resistance arises through two mechanisms. Low-level resistance results from point mutations in the essential chromosomal isoleucyl-tRNA synthetase gene ileS1, whereas high-level resistance is mediated by mupA, encoding an alternative isoleucyl-tRNA synthetase (ileS2),5 carried on plasmids.6 Plasmid-mediated resistance is enriched in settings having extensive topical antimicrobial use. We recently reported the emergence of a dual mupirocin-resistant community-acquired (CA-)MRSA lineage undergoing community transmission in New York City.4 However, whether such plasmids promote hospital transmission and what evolutionary forces maintain7 them is unknown.
Here, we analyzed >10,000 S. aureus genomes collected between January 2022 and January 2025 to define the distribution, spread, and evolutionary consequences of mupA-carrying plasmids. Strong selection for mupA was evident in preferential nosocomial transmission of mupA-carrying strains, and more strikingly, in the repeated emergence (>100 independent times) of inactivating ileS1 mutations that created plasmid addiction. In these strains, plasmid- encoded ileS2 becomes essential for survival. The association of addiction with increased conjugative transfer of mupA and broad biocide tolerance suggests additional selective pressures reinforcing plasmid maintenance and spread. Conversely, mupirocin unexpectedly synergized with exogenous isoleucine to re-sensitize mupA-positive strains.These findings underscore the need for genomic surveillance and resistance testing. They also suggest that risk-targeted decolonization strategies and exploitation of plasmid-dependent vulnerabilities may help control mupirocin-resistant S. aureus.
METHODS
Methods are summarized here; additional details are provided in the Supplementary Materials.
Study isolates and procedures
S. aureus isolates were obtained from an institutional biobank at two tertiary-care hospitals within NYU Langone Health (Brooklyn and Manhattan; 444 and 813 beds). The biobank and decolonization procedures have been described previously.8,9 For the present study, we analyzed an expanded surveillance cohort comprising 10,321 sequenced isolates collected between January 2022 and January 2025, including 4,129 MRSA and 6,192 MSSA isolates. MRSA isolates were obtained from various clinical and admission screening cultures, whereas MSSA isolates were obtained from bloodstream infection and admission screening cultures only. Adult patients admitted to medicine, oncology, transplant, and intensive care units underwent routine admission screening for nasal colonization. Patients colonized with MRSA or MSSA received twice-daily intranasal mupirocin and daily bathing with 2% chlorhexidine gluconate for 5 days. During the study period, admission screening compliance was 77.8%. Among colonized patients, 49% received at least one day of decolonization therapy and 10.8% received three or more days.
To identify putative progenitors of the ileS1 mutant lineage E499, we screened a historical collection of 1,649 consecutive clinical MRSA isolates collected between 2017 and 2020 (pre- COVID-19 pandemic). These 604 isolates were obtained from bloodstream, respiratory, and wound infections in adult patients. The 31 mupA-positive CC8 isolates with closed genomes analyzed for chromosomal integration were also derived from this historical collection.
The study was approved by the NYU Langone Health Institutional Review Board (s24-01872). Additional details regarding strains, plasmids, primers, growth conditions, RNA sequencing, and phenotypic assays are provided in the Supplementary Methods.
Genome sequencing and transmission detection
Whole-genome sequencing, assembly, quality control, bioinformatic analyses, and genomic surveillance–based transmission inference were performed as previously described for the institutional genomic surveillance program.8,9 Additional details, including transmission cluster definitions and linkage criteria, are provided in the Supplementary Methods.
Role of the funding source
The study sponsors had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all study data and had final responsibility for the decision to submit for publication.
RESULTS
Genomic Epidemiology of mupA-positive S. aureus
Susceptibility testing for mupirocin resistance is not routinely performed by most clinical laboratories because resistance is not included on automated testing platforms. We therefore used whole-genome sequencing to systematically identify mupA across 10,321 S. aureus isolates collected through institutional genomic surveillance. Overall, mupA was detected in 11·8% of isolates, indicating a substantial reservoir of mupirocin resistance within the healthcare system.
Mupirocin resistance was strongly enriched in MRSA. Although MSSA isolates were more numerous overall (6,192 versus 4,129), 85·8% (1,041) of mupA-positive isolates occurred in MRSA. Phylogenetic analysis revealed, as previously reported,10 that mupA within MRSA was concentrated in CC8 and, to a lesser extent, CC5. These lineages represent the predominant community- and healthcare-associated MRSA lineages, respectively, in New York City and the United States.9 Together, these lineages accounted for 96·7% of mupA-positive MRSA, whereas the remainder of mupA-positive isolates occurred in sporadic isolates and small clusters. We therefore focused subsequent analyses on CC8 and CC5.
mupA was distributed across multiple independent branches of both the CC5 and CC8 phylogenies (Fig. 1A), consistent with repeated horizontal acquisition, not single ancestral expansion. Nevertheless, several large clonal expansions were evident, particularly within CC8.
Assembly of a subset of complete plasmids revealed several structurally distinct plasmid groups (representative plasmid in Fig. 1B; group comparisons in Supplementary Fig. 1). These complete plasmids informed plasmid classification across the surveillance cohort (Fig. 1). Three major plasmid groups (Groups 1–3) accounted for >91% of mupA-positive isolates, with Group 1 representing 60%. Thus, mupA dissemination was dominated by a small number of plasmid groups. Mobility predictions of the reference plasmids corresponded closely with the major structural groups and were consistent with a recent regional survey (Supplementary Fig. 1);11 representative Group 1 plasmids were conjugative, Group 2 included mobilizable and non- mobilizable plasmids, and Group 3 plasmids were mobilizable. Conjugative plasmids contained a tra region homologous to those of the S. aureus resistance plasmids pUSA03 and pSK41.6,12 Plasmids exhibited extensive mosaicism, consistent with recombination and exchange of accessory genetic elements (Supplementary Fig. 1). In addition to mupA, these plasmids encoded multiple accessory features, including antimicrobial resistance determinants, transposases, and insertion sequences. Notably, 53·7% of mupA-positive strains also carried chlorhexidine resistance genes, establishing dual resistance to both hospital decolonization agents, consistent with previous reports.4 Some plasmids also carried a pentose phosphate pathway locus arranged in a non-native, operon-like configuration (Fig. 1B).
Mapping plasmid structural groups onto the chromosomal phylogeny revealed both multiple plasmid groups within individual lineages and identical plasmid groups across lineages (Fig. 1A). Thus, mupA plasmids disseminate through a combination of vertical clonal expansion and repeated inter-lineage transfer.
mupA is Associated with Increased Patient-to-patient Transmission
Patients harboring mupA-positive isolates differed in several previously identified risk factors for S. aureus colonization and transmission, including greater hospital exposure (Supplementary Table 1).9 To determine whether mupA was associated with transmission, we analyzed transmission using a previously described genomic-epidemiologic framework (Supplemental Methods).8,9 Compared with mupA-negative strains, mupA-positive strains had a 10·5 percentage point higher transmission rate (95% CI 8·4 – 12·7); this association was robust to propensity score adjustment for lineage, hospitalization history, antibiotic exposure, and demographic factors (adjusted difference 11·3 percentage points, 95% CI 5·8–16·9; Supplementary Table 2). Consistent with transmission contributing to infection risk,13 mupA was also enriched among hospital-onset infections compared with admission colonizing isolates (17·0% vs 10·2%).
The transmission advantage associated with mupA was observed across both MRSA and MSSA but was most pronounced in MRSA lineages (Fig. 2, Supplementary Table 3). MSSA increased from 2·1% to 7·0% with mupA but remained low overall; mupirocin-resistant MSSA transmitted at rates comparable to mupirocin-susceptible MRSA. In CC5 MRSA, transmission increased from 12·9% to 26%, and in CC8 MRSA the increase was from 7·4% to 14·6%. Thus, although mupA was most prevalent in CC8, its association with transmission was greatest in CC5, the predominant healthcare-associated MRSA lineage.
Chromosomal ileS1 point mutations conferring low-level mupirocin resistance (V588F, V631F, G593V)14 were not associated with increased transmission. The consistency of the mupA effect across genetic backgrounds and patient populations, its persistence after statistical adjustment, and the absence of effect from chromosomal resistance mutations support a role for plasmid- mediated mupirocin resistance in promoting hospital transmission.
Recurrent Evolution of Plasmid Addiction Through Chromosomal ileS1 Inactivation
Unexpectedly, analysis of ileS1 revealed widespread inactivation of this essential chromosomal isoleucyl-tRNA synthetase among mupA-positive isolates. We identified 105 distinct loss-of- function mutations, consisting of frameshifts and truncations (Fig. 3A). Most mutations were observed in only one or a few isolates; however, several underwent substantial clonal expansion, including one lineage involving >155 patients. Isolates carrying the same inactivating ileS1 mutation clustered within monophyletic groups, indicating that recurrently observed mutations reflected clonal expansion rather than parallel evolution (Fig. 3B; Supplementary Fig. 2).
Addiction (leS1 inactivation–mediated dependence on mupA) was most common in CC8 and in plasmid groups associated with this lineage, reflecting the high prevalence of mupA and expansion of addicted clones. Thus, plasmid addiction both reflects MRSA population structure and contributes to its evolution.
Because chromosomal integration events are largely invisible in short-read surveillance data, we screened all available closed CC8 genomes from our collection. Among 31 mupA-positive isolates, chromosomal integration was identified in 6 (19·4%). All six occurred among the seven addicted isolates (85·7%), involving three independent addiction variants, whereas none of the 24 non-addicted isolates showed chromosomal integration. Thus, chromosomal acquisition of mupA may be an additional outcome of plasmid addiction.
Fitness Costs of mupA Plasmids Drive Selection for Addiction
To determine whether ileS1 inactivation conferred a fitness advantage, we introduced an ileS1 addiction mutation identified in a clinical isolate into a laboratory strain carrying a mupA plasmid. Addiction increased neither resistance nor growth in the presence of subinhibitory mupirocin concentration (Supplementary Fig. 3). Instead, addicted strains grew more poorly than their plasmid-carrying parental counterparts, indicating that addiction imposes a burden rather than enhancing resistance.
These findings suggest addiction is selected not to improve bacterial growth but to stabilize plasmids that would otherwise be lost. Consistent with this interpretation, plasmid carriage produced no measurable defect in vitro across a range of growth conditions, but it imposed substantial costs in vivo. In a murine nasal colonization and transmission model (Fig. 4A), introduction of the plasmid into a CC8 background, where addiction is common, significantly reduced both colonization and transmission (Fig 4B, Supplementary Fig. 4A). By contrast, in CC5, where addiction is rare, plasmid carriage had minimal or slightly beneficial effects (Fig. 4C, Supplementary Fig. 4B). Introduction of an addicted chromosomal background further reduced transmission (Supplementary Fig. 5), demonstrating that addiction can persist despite exacerbating plasmid-associated fitness costs. These in vitro and in vivo observations suggest that mupA plasmids impose lineage-specific fitness costs and suggest that such costs help drive both selection for addiction and its uneven distribution across MRSA populations.
mupA Plasmids/Addiction Impose Dependence on the Stringent Response
Because aminoacyl-tRNA synthetases operate near saturation, even small activity reductions impose steep fitness costs.15 Reduced charging efficiency of the mupA-encoded isoleucyl-tRNA synthetase may therefore elevate uncharged tRNA levels and activate the stringent response via production of (p)ppGpp, a nutrient-stress pathway triggered by accumulation of uncharged tRNA that suppresses growth and biosynthesis.16 Experimental disruption of the stringent- response is complicated in S. aureus because the RelA/SpoT homolog (RSH), which mediates (p)ppGpp synthesis and hydrolysis, is essential for viability.17 Therefore, we evaluated growth in the presence of relacin, an inhibitor of (p)ppGpp synthesis, to test whether mupA-carrying strains exhibit increased dependence on stringent-response signaling, as predicted for aminoacyl-tRNA synthetase limitation.15 Relacin impaired growth of plasmid-carrying and addicted strains, but it had little effect on plasmid-free, parental strains (Fig. 5). Although relacin exhibits low affinity for Rel and may have off-target effects,18 these findings support the idea that mupA plasmids increase dependence on the stringent response. This dependence provides a mechanistic explanation for both the fitness cost of plasmid carriage and selection of addicted ileS1 mutants.
Because stringent-response activation promotes biofilm formation,19 we examined this phenotype. Addicted strains, but not their plasmid-carrying parental counterparts, exhibited increased biofilm formation compared with plasmid-free controls (Fig. 6A). Because conjugative transfer is enhanced within biofilms,20 we next tested whether addiction influences plasmid transfer. Conjugation frequencies were higher in addicted strains than in their plasmid-carrying parental strains (8·88e-7 vs 2·31e-7 (mean values); Fig. 6B). Thus, addiction not only ensures vertical inheritance but it also promotes horizontal dissemination, thereby expanding the reservoir of transmissible resistance.21
Stringent-Response Dependence Creates a Therapeutic Vulnerability
Expression analysis of selected stringent-response genes in biofilm populations was consistent with activation of stringent-response pathways in both mupA plasmid-carrying and addicted strains (Supplementary Fig. 6), as evidenced by increased expression of oligopeptide transport genes (oppA, oppC), branched-chain amino acid (BCAA) biosynthesis genes (ilvB, leuA, leuB), purine salvage (xpt), and reduced expression of ribosomal genes (rplR, rplV), purine biosynthesis (purH), and metabolism (sucA, atpA, gltB).22 The addicted strain generally showed larger effect sizes across several markers, potentially explaining its association with increased biofilm formation. No change in rsh expression was observed, consistent with prior work showing that stringent-response transcriptional programs can persist after adaptation to sublethal stress without increased rsh transcription.23 Reduced translational capacity and impaired energy metabolism provide a potential mechanistic basis for the fitness cost of plasmid carriage.
The stringent-response dependence imposed by mupA, and more generally by mupirocin, suggested a vulnerability arising from the interaction between stringent-response signaling and CodY-mediated regulation. In S. aureus, intracellular pools of BCAA are integrated with stringent-response signaling through the global regulator CodY.24,25 When BCAA are abundant, CodY represses genes involved in amino acid biosynthesis and stress adaptation; depletion of these metabolites relieves repression and activates these pathways. Thus, exogenous isoleucine can uncouple this regulatory relationship by reactivating CodY even when stringent-response signaling persists.25 Among BCAA, isoleucine is the most potent mediator of this effect, restoring CodY-dependent repression despite continued metabolic stress.25 This creates a regulatory mismatch in which stringent-response signaling remains active while CodY suppresses the biosynthetic programs needed for growth.
This interaction has limited consequences under nutrient-replete laboratory conditions, where amino acids are abundant. However, the in vivo environment differs substantially. Nasal secretions, the primary ecological niche for S. aureus colonization, are markedly deficient in free amino acids, especially isoleucine.26 Under these conditions, mupirocin stringent-response activation occurs in the setting of amino-acid limitation, where relief of CodY repression would be expected to facilitate biosynthesis and growth. Exogenous isoleucine disrupts this adaptive response by restoring CodY despite ongoing stringent-response signaling, with consequences for growth.25
Consistent with these observations, moderate concentrations of isoleucine alone modestly impaired growth with both plasmid-free and plasmid-carrying strains grown in defined medium lacking BCAA (Fig. 7A). However, when combined with subinhibitory mupirocin, which further impairs isoleucyl-tRNA charging and amplifies stringent-response signaling, even lower doses of exogenous isoleucine produced near-complete growth collapse (Fig. 7B). Under these conditions, either agent alone had no impact. This synergistic interaction demonstrates a vulnerability in mupA-containing strains during mupirocin exposure, allowing resistance to be sensitized without directly inhibiting the resistance determinant.
Identifying vulnerabilities is important because inhibition of tRNA synthetases is known to induce broad tolerance to disinfectants.27 The resulting stringent response suppresses the reactive oxygen species (ROS) surge induced by bactericidal agents, thereby reducing killing and promoting survival under biocidal stress. We therefore tested whether mupA plasmids altered tolerance to common biocidal agents after growth in BCAA-deficient medium, conditions that approximate the amino-acid limitation of nasal secretions. Plasmid-carrying strains exhibited a 1–2 log reduction in killing following exposure to multiple disinfectants, including ethanol, 2- propanol, and hydrogen peroxide, versus a plasmid-free parental strain (Fig. 8). In contrast, addicted strains did not show further increases in tolerance beyond that conferred by the plasmid alone. Thus, mupA plasmids confer broad biocide tolerance.
Discussion
The present work shows that mupA can lock S. aureus into plasmid carriage, stabilizing mupirocin resistance. In these cases, mutational inactivation of the essential chromosomal ileS1 gene renders bacterial survival dependent on plasmid-encoded ileS2, creating plasmid-host interdependence that ensures retention even in the absence of mupirocin selection. This mechanism follows the logic of toxin-antitoxin systems—plasmid loss is more costly than retention—but loss achieves dependence through essential-gene replacement rather than post- segregational killing.28 Addiction arose repeatedly, underscoring the likelihood of strong selection to stabilize resistance plasmids. Dissemination of addicted lineages beyond hospitals would entrench resistance and pose a public-health threat.
Chromosomal integration of mupA among addicted strains is consistent with the expectation that beneficial accessory genes may ultimately be captured by the chromosome.29 However, unlike classical models, addiction creates dependence on a plasmid-borne function through loss of the chromosomal homolog. Thus, addiction may represent another route by which plasmid- borne genes become assimilated into the host genome.
Addiction increased biofilm formation and conjugative transfer of mupA. Thus, addiction not only seems to ensure vertical inheritance but also promotes horizontal dissemination, generating a plasmid reservoir that spreads resistance within and across MRSA lineages. Consistent with these experimental findings, phylogenetic analyses revealed expansion of addicted clones and repeated acquisition of mupA-containing plasmids (Figure 2A, Supplementary Fig. 2). Together, these findings indicate addiction promotes both persistence and spread of mupA.
Notably, mupA itself was associated with an increased likelihood of transmission, consistent with an advantage that was strongest in MRSA and greatest within CC5, the predominant healthcare-associated MRSA lineage, whereas the smallest effect was observed in MSSA. Thus, mupA preferentially enhanced transmission among strains that already exhibit the highest baseline transmission potential in healthcare settings (Figure 2, ref.9). Rather than creating successful lineages, mupA appears to amplify the fitness of strains already adapted to transmission. The molecular basis of this effect is unknown.
The observation that the absolute transmission advantage associated with mupA was greatest in healthcare-associated CC5 MRSA, despite the substantially higher prevalence of mupA in CC8, suggests that healthcare transmission is only one component of the selective environment shaping mupA dissemination. Consistent with this idea, our prior work identified the emergence of a mupirocin-resistant CC8 lineage undergoing community transmission.4 Additionally, the present study found that CC8, but not CC5, incurred a substantial in vivo fitness cost associated with mupA carriage and exhibited a markedly higher burden of addiction, suggesting stronger selection for plasmid stabilization in this lineage. Collectively, these findings, together with evidence that that transmission risk reflects interactions between host epidemiology and bacterial lineage,9 suggest that distinct combinations of clonal spread, horizontal plasmid transfer, and plasmid addiction sustain mupA persistence in CC5 and CC8, thereby facilitating dissemination in both hospitals and communities even when selective pressure is reduced. Thus, strategies based solely on restricting mupirocin use may be insufficient.
mupA plasmids conferred increased survival under biocide pressure, including ethanol, isopropanol, and chlorhexidine. Increased tolerance to these agents could enhance environmental persistence and opportunities for transmission, potentially contributing to the increased transmission of mupA-carrying strains observed in our surveillance data. Together with additional metabolic and resistance determinants encoded by mupA plasmids, these findings suggest that the success of mupA plasmids may reflect broader adaptive benefits beyond resistance to mupirocin alone.
Exogenous isoleucine synergized with mupirocin to inhibit growth of mupA-containing strains under BCAA-limited, nasal-like conditions. A likely explanation (Supplementary Fig. 7) is that by impairing isoleucyl-tRNA charging, mupirocin reduces utilization of intracellular isoleucine, thereby increasing its effective availability. As a result, even low concentrations of exogenous isoleucine may be sufficient to maintain CodY-mediated repression of amino acid biosynthetic pathways. The resulting regulatory mismatch suppresses biosynthetic programs required for BCAA synthesis, creating a conditional vulnerability that culminates in growth collapse.
Further genomic and experimental studies are needed to confirm the predicted low charging efficiency of mupA-encoded IleS2, define how stringent-response activation generates plasmid- associated fitness costs, determine why these costs are particularly pronounced in CC8, and identify compensatory mutations that mitigate conflicts between plasmid carriage and the host chromosome in successful addicted lineages. Additional in vivo studies, including colonization and transmission models, are underway to determine whether the isoleucine-mediated sensitization observed in vitro can be translated into an effective decolonization strategy. Finally, our epidemiologic analyses were limited to two hospitals within a single urban healthcare system and did not comprehensively sample community reservoirs. Because community- associated selection may contribute to mupA dissemination, studies spanning healthcare and community settings will be needed to define the prevalence and spread of addicted lineages.
Overall, our results suggest that reducing mupirocin use alone will be insufficient to reverse resistance, especially if addiction stabilizes resistance within anchor lineages: adaptive optimization may convert a costly plasmid into a self-sustaining component of the MRSA population. As with penicillin resistance, mupA could become endemic in hospitals and communities. This possibility underscores the need for new decolonization strategies that are less susceptible to resistance, including bacteriophage-derived lysins,30 exploitation of collateral susceptibilities (e.g., isoleucine), and colonization-specific immunotherapy. However, development of new interventions alone may be insufficient. Preserving effectiveness of agents will require more judicious use, guided by identification of and targeted intervention in patients most likely to transmit or acquire MRSA rather than broad empiric treatment. Such risk stratification depends on genomic surveillance to identify transmission events, including those involving asymptomatic colonization, thereby defining high-risk spreaders and enabling development of personalized models that predict transmission. Because resistance may arise despite these efforts, surveillance will also be essential to detect emerging resistant clones before they adapt, spread, and become entrenched.4 Prior work demonstrates that this is possible, but only real-time implementation will enable intervention before resistant lineages become endemic.9
Supporting information
Data Availability
Data produced in the present study are available upon request to the authors. Majority of the data produced in the present work are contained in the manuscript.
Declaration of interests
B.S. has served on a scientific advisory board for Innoviva Specialty Therapeutics and has received research funding from Analog Devices Inc.
Data availability
Sequencing data are available through the NCBI repository using the accession number PRJNA1442056.
Acknowledgments
We thank Drs. Emily Grasso, Paul Zappile, Gael Wesby and Peter Meyn for assistance with library construction, and sequencing, Marc Lipsitch, Lorna Thorpe, and Jeffrey Weiser and for critical comments on the manuscript, and Daria Frolova and John A. Lees for helpful discussions. This work was supported in part by National Institutes of Health grants AI137336 (to B.S. and A.R.); AI140754 (to B.S.); CDC U01CK000590 (B.S.), K08AI163457 (to R.J.U.), and funds from the NYULH Antimicrobial-Resistant Pathogens Program (B.S., S.H.). Genome sequencing was partially supported by Cancer Center Support Grant P30CA016087 from the NYULH Laura and Isaac Perlmutter Cancer Center.