Transmission of Antimicrobial-Resistant Staphylococcus Aureus Clonal Complex 9 Between Pigs and Humans, United States

Pranay R. Randad; Jesper Larsen; Hülya Kaya; Nora Pisanic; Carly Ordak; Lance B. Price; Maliha Aziz; Maya L. Nadimpalli; Sarah Rhodes; Jill R. Stewart; Dave C. Love; David Mohr; Meghan F. Davis; Lloyd S. Miller; Devon Hall; Karen C. Carroll; Trish M. Perl; Christopher D. Heaney


Emerging Infectious Diseases. 2021;27(3):740-748. 

In This Article


Sources of S. Aureus Isolates From Humans and From Pigs Raised on IHOs in North Carolina

S. aureus isolates from IHO pigs were collected from a convenience sample of a single IHO in North Carolina (IHO-1), as described previously.[12] We collected additional pig samples by hanging a length of undyed, unbleached cotton rope in pig pens of 20 IHOs in North Carolina (IHO-2–IHO-21) (Appendix). Pig isolates were recovered from IHO-2, IHO-3, IHO-4, IHO-5, and IHO-6 for a total of 6 IHOs (IHO-1–IHO-6). Isolates from IHO-2–IHO-6 have not been published previously. The spa type for all IHO pig isolates was characterized, as previously described,[12] and used to assign each isolate to a putative multilocus sequence type (MLST).

S. aureus isolates from humans were collected from participants who were previously enrolled into 1 of 3 separate epidemiologic studies (study 1, study 2, and study 3) and screened for nasal carriage of S. aureus (Appendix, Sample collection, sample processing, and S. aureus isolation methods were described previously.[1–3] MLST was previously determined for all study 1 isolates.[1] The spa type was previously characterized for study 2 and study 3 isolates and used to assign a putative MLST based on previously published associations between spa types and MLSTs.[2,3]

Selection of S. Aureus CC9 Isolates for WGS Analysis

A total of 236 putative or MLST-confirmed S. aureus CC9 isolates were recovered from IHO pigs (n = 91) and humans (n = 145) in North Carolina during 2011–2016 (Appendix). For this study, a convenience sample of 49 isolates from North Carolina were subjected to WGS analysis, including 10 isolates from pigs raised on 4 different IHOs, 34 isolates from 25 IHO workers, 1 isolate each from 3 IHO minors, and 1 isolate each from 2 community resident adults (Appendix). For comparative purposes, we also included an international collection of 32 S. aureus CC9 genomes available as of August 1, 2018, from the National Center for Biotechnology Information (NCBI) Reference Sequence Database (, which included information on source, geographic location, and collection year.

WGS and Bioinformatic Analyses

We prepared DNA for multiplexed, paired-end sequencing by preparing libraries using either the Nextera XT DNA Library Preparation Kit (Illumina, Inc.), according to manufacturer instructions, or the Kapa Hyper Prep Kit (Kapa Biosystems, Inc., and uniquely barcoded adaptors from NEXTFLEX-96 Unique Dual Index barcodes (Bioo Scientific, We prepared equimolar pools of S. aureus libraries at a concentration of 2 nmol and sequenced on a MiSeq (Illumina, Inc., at 2 × 300 bp. WGS data are available in the NCBI Sequence Read Archive (; BioProject no. PRJNA574434).

We used SPAdes[13] to generate de novo assemblies and compared these against the S. aureus MLST database[14] to assign MLSTs. We used ABRicate ( to search the ResFinder database for antimicrobial-resistance (AMR) genes.[15] We used BLASTN ( to detect genes in the phage-associated immune evasion cluster (IEC), including scn, chp, sak, sea (GenBank accession no. NC_009641), and sep (GenBank accession no. BA000018).[16]

We used the NASP pipeline[17] to map sequence reads against the de novo-assembled genome of North Carolina isolate IHOW6.1 (BioProject accession no. PRJNA574434) and to perform single-nucleotide polymorphism (SNP) calling, as described previously.[8] We used Gubbins version 2.3.1[18] to remove recombination from the SNP alignment and used the remaining SNPs in the core genome to construct a midpoint-rooted maximum-likelihood tree by using PhyML[19] with a general time-reversible model of nucleotide substitution and 100 bootstrap replicates.[20] We used the same methods to perform a separate SNP analysis of the cluster containing the North Carolina isolates (clade 3) to improve the resolution of the transmission analysis. We calculated pairwise SNP differences by using MEGA5.[21] To define a SNP-based threshold for assigning isolates into putative transmission clusters, we used the maximum within-farm pairwise SNP distance among S. aureus CC9 isolates from IHO-1, in which all isolates were collected from the same IHO at the same sampling time.

Antimicrobial Susceptibility Testing

Isolates in the North Carolina collection previously were assessed for susceptibility to a panel of antimicrobial drugs by using the Phoenix Automated Microbiology System (Becton Dickinson, or the Kirby-Bauer disk diffusion method (Appendix Table 2). Testing was completed by the Clinical Microbiology Laboratory at the Johns Hopkins Hospital based on Clinical Laboratory Standards Institute (CLSI; guidelines specified in the source studies[1–3] (Appendix Table 2). We defined MDRSA as S. aureus isolates resistant to ≥3 classes of antimicrobial drugs.[22] We defined MRSA as S. aureus harboring the mecA gene.

Statistical analysis

We used the χ 2 test to compare AMR and IEC genes between groups. We performed all statistical analyses by using Stata version 14.2 (StataCorp LLC,