Population genomic molecular epidemiological study of macrolide resistant Streptococcus pyogenes in Iceland, 1995-2016: Identification of a large clonal population with a pbp2x mutation conferring reduced in vitro β-lactam susceptibility

Resistance to macrolide antibiotics is a global concern in the treatment of Streptococcus pyogenes (Group A Streptococcus, GAS) infections. In Iceland, since the detection of the first macrolide-resistant isolate in 1998, three epidemic waves of macrolide-resistant GAS infections have occurred with peaks in 1999, 2004, and 2008. We conducted whole genome sequencing of all 1,575 available GAS macrolide-resistant clinical isolates of all infection types collected at the national reference laboratory in Reykjavik from 1998 to 2016. Among 1,515 erythromycin-resistant isolates, 90.3% were of only three emm types: emm4 (n = 713), emm6 (n = 324), and emm12 (n = 332), with each being predominant in a distinct epidemic peak. The antibiotic efflux pump genes, mef(A) and msr(D), were present on chimeric mobile genetic elements in 99.3% of the macrolide-resistant isolates of these emm types. Of note, in addition to macrolide resistance, virtually all emm12 isolates had a single amino acid substitution in penicillin-binding protein PBP2X that conferred a two-fold increased penicillin G and ampicillin MIC among isolates tested. We conclude that each of the three large epidemic peaks of macrolide-resistant GAS infections occurring in Iceland since 1998 was caused by the emergence and clonal expansion of progenitor strains, with macrolide resistance being conferred predominantly by inducible Mef(A)-Msr(D) drug efflux pumps. The occurrence of emm12 strains with macrolide resistance and decreased beta-lactam susceptibility was unexpected and is of public health concern. Repositories Genomic sequencing data for all 1,515 macrolide/erythromycin-resistant isolates were deposited into the National Center for Biotechnology Information Sequence Read Archive under bioproject accession PRJNA614628 and assembled sequences for composite elements Φ29854, Φ29862 and Φ29661 were deposited in Genbank under accessions ###, ### and ### respectively.

transposon into a prophage (19)(20)(21)(22). Resistance to macrolides at low frequency can also 91 spontaneously arise via mutations in the 23S rRNA and in ribosomal proteins L4 and L22, 92 encoded by genes rplD and rplV, respectively (14). 93 Since the first reports of macrolide-resistant GAS in England in the late 1950s (23), 94 resistance has disseminated worldwide, and its prevalence has been reported to vary profoundly 95 geographically (i.e. between countries/regions at a point in time) and temporally (i.e. in the same 96 country/region over time) (15,24). In many instances, an increase in the prevalence of resistant 97 isolates clearly corresponded with increased antibiotic usage, consistent with the influence of 98 antibiotic selective pressure (25). However, in some cases precipitous changes in resistance 99 prevalence has occurred in association with a change in predominant GAS clone or mechanism 100 of resistance, but independent of any perceived change in antibiotic usage (26). In Iceland, 101 erythromycin susceptibility testing was performed on at least 100 GAS isolates per year and the 102 first macrolide-resistant isolate was not detected until early 1998. Over the next year the monthly 103 proportion of macrolide-resistant GAS precipitously increased from 0% in March 1998 to 56% 104 in March 1999 (27). Among 367 erythromycin-resistant GAS isolates collected through July 105 considered to be the same strain if they were collected twice or more, ≤7 days apart from the 137 same patient. When antibiotic resistant susceptibilities were inconsistent between isolates taken 138 from the same patient, the isolate from the more invasive infection sample was used. Isolates 139 were grown on tryptic soy agar with 5% sheep blood (Benton Dickson) or with 5% horse blood 140 (Oxoid) at 37°C and 5% CO2. 141

Whole Genome Sequencing 142
All viable GAS isolates that tested resistant to the macrolide antibiotic erythromycin (n = 143 1,575) within the collection were sent to the Center for Molecular and Translational Human 144

Infectious Diseases Research, Department of Pathology and Genomic Medicine, Houston 145
Methodist Research Institute (Houston, Texas) for whole genome sequencing. Genomic DNA 146 extraction and multiplexed library preparation was performed as previously described (33). 147 Paired-end, 150 nucleotide-long sequencing reads were obtained using an Illumina NextSeq 500 148 sequencer. Sequence data preprocessing (i.e. artifact and adapter trimming, quality filtering, and 149 base call error correction) and de novo assembly for each isolate was done as previously 150 described (33). 151

Initial genetic typing and gene content profiling 152
Multi-locus sequence type (MLST), emm type, and antibiotic resistance gene content was 153 determined for each isolate from the sequencing reads relative to publicly available reference 154 databases using SRST2 software (34) as previously described (33). Mobile genetic element 155 typing was determined relative to a published database of S. pyogenes phage and ICE-encoded 156 integrase and virulence factor genes as previously described using SRST2 (35). The pbp2x gene 157 was identified in and retrieved from isolate genome assemblies using blastn and bedtools-158 getfasta respectively. 159

Phylogenetic inference and population structure 171
Concatenated SNP sequences used for evaluation of genetic relationships among isolates 172 were generated using Prephix and Phrecon (www.github.com/codinghedgehog). To limit 173 phylogenetic inferences to primarily vertically inherited core chromosomal SNPs, mobile genetic 174 element (phage and ICE) encoded regions were excluded and regions of horizontal transfer and 175 recombination were identified and excluded using Gubbins (38). Phylogeny among isolates was 176 inferred by the Neighbor-Joining method using SplitsTree (39), and phylograms were generated was used to introduce the Met593Thr substitution into pbp2x of MGAS27213-L601P. Primers 183 pbp2x-5'fwd (CAATTGTACAAAACCGTTACGATCC AAG) and pbp2x-5'rev 184 (TAGTAACATACATCAAAAAGTCTGGTTTATC) were used to amplify the pbp2x 5' end. 185 Primers pbp2x-T593-3'fwd (CTTTTTGATGTATGTTACTACGACTAAACCAC) and pbp2x-186 3'rev (GTGAATACATATCAGTATTTGTGGGTCATC) were used to amplify the pbp2x 3'end 187 introducing a single A to C nucleotide change in pbp2x condon 593. Primers pBBL740-fwd 188 (GTAACGGTTTTG TACAATTGCTAGCGTAC) and pBBL740-rev 189 (AAATACTGATATGTATTCACGAACGAAAATC) were used to amplify and linearize 190 suicide plasmid pBBL740 by inside-out PCR. The pbp2x 5'-end and 3'-end amplicons were 191 spliced with the linearized pBBL740 amplicon using NEBuilder HiFi kit (New England 192 Biolabs  To genetically characterize the cohort, all 1,575 available viable erythromycin-resistant 223 GAS isolates were whole genome sequenced to an average 214-fold depth of coverage (range: 18 224 to 1859´) using Illumina paired-end sequencing. Based on the sequence data, 60 of the isolates 225 were excluded from the investigation, for reasons such as the isolate was not S. pyogenes, was a and their epidemiological and genetic characteristics are listed in Table S1. Sequence reads for 228 the isolates assembled on average into 67 contigs summing to 1.82 Mbp with a G+C content of 229 38.4%, values which are consistent with closed genomes of S. pyogenes. 230 The 1,515 macrolide-resistant isolates were comprised of 27 emm types (Table S1 and 231  (Table 1). One or more ARG was 239 detected in 1,471 (97.1%) of the isolates, and no macrolide-resistant gene was found in 44 240 isolates of 13 different emm types (Table S1)

Potential for altered beta-lactam antibiotic susceptibility 296
Recently it has been shown that many S. pyogenes clinical isolates with nonsynonymous 297 (amino acid substituting) nucleotide changes in the penicillin-binding protein 2X gene (pbp2x) 298 are associated with reduced susceptibility in vitro to one or more members of the beta-lactam 299 family of antibiotics (12, 13). Among the 1,515 macrolide-resistant isolates, 25 pbp2x alleles 300 encoding for 10 PBP2X variants were identified ( The susceptibility to penicillin G, ampicillin, and erythromycin of the three predominant 311 PBP2X variants present in the emm 4, 6, and 12 isolates was tested for five isolates of each emm 312 type ( Table 2). The isolates were selected to represent the temporal spread of each emm type 313 corresponding with the three peaks of macrolide-resistant infections. All five emm4 isolates 314 having the PBP2X WT variant were fully susceptible to the beta-lactam antibiotics penicillin G 315 and ampicillin. Despite the five emm6 isolates having a PBP2X variant with three amino acid 316 substitutions relative to the PBP2X WT, they were also fully susceptible to the beta-lactam antibiotics. In contrast, all five emm12 isolates with a Met593Thr substitution PBP2X variant 318 had approximately two-fold increased MICs for both penicillin G and ampicillin. To 319 unambiguously determine if the PBP2X Met593Thr substitution alters beta-lactam susceptibility, 320 we constructed an isogenic PBP2X Thr593 substitution derivative using the parental strain 321 MGAS27213-PBP2X-L601P (12). Importantly, whole genome sequencing confirmed that the 322 constructed derivative strain, MGAS27213-PBP2X-L601P,M593T, only differs from the parent 323 strain by a single nucleotide change in codon 593 (ATG to CTG) of pbp2x. As anticipated, the 324 parental strain had fully susceptible PBP2X WT penicillin G and ampicillin MIC levels. In 325 contrast, the isogenic PBP2X Met593Thr derivative had two-fold increased MICs (Table 2)       Alignments were made with a 20 nt moving window at 95% identity. Core chromosomal, Tn1270.1 and phage gene content of the Φ10394.4 gene map are colored as indicated in the index.