Rapid Increase of CTX-M-Producing Shigella sonnei Isolates in Switzerland Due to Spread of Common Plasmids and International Clones

The Swiss Centre for Antibiotic Resistance (ANRESIS) has recently noted an increase of extended-spectrum cephalosporin-resistant (ESC-R) Shigella sonnei isolates nationwide (3.8% in 2016 versus 37.5% in 2019). To understand this phenomenon, we analyzed 25 representative isolates (of which 14 were ESC-R) collected in Switzerland during 2016 to 2019. Whole-genome sequencing was achieved using both the Illumina and the Nanopore platforms. Both ESC-R and extended-spectrum cephalosporin-susceptible isolates belonged to sequence type 152 (ST152).

ABSTRACT

The Swiss Centre for Antibiotic Resistance (ANRESIS) has recently noted an increase of extended-spectrum cephalosporin-resistant (ESC-R) Shigella sonnei isolates nationwide (3.8% in 2016 versus 37.5% in 2019). To understand this phenomenon, we analyzed 25 representative isolates (of which 14 were ESC-R) collected in Switzerland during 2016 to 2019. Whole-genome sequencing was achieved using both the Illumina and the Nanopore platforms. Both ESC-R and extended-spectrum cephalosporin-susceptible isolates belonged to sequence type 152 (ST152). The ESC-R isolates carried bla_CTX-M-3 in IncI1-pST57 (n = 5), bla_CTX-M-15 in IncFII (F2:A-:B-) (n = 5), bla_CTX-M-15 in IncI1-pST16, and bla_CTX-M-27, bla_CTX-M-55, or bla_CTX-M-134 in other IncFII plasmids (n = 1 each). Plasmids having the same bla and Inc group exhibited high degrees of genetic identity to each other but also to plasmids previously reported in other Enterobacterales. Core-genome analysis showed that there were 4 main clusters, each of which included strains that differed by <58 single nucleotide variants (SNVs) and that consisted of both bla_CTX-M-positive and bla_CTX-M-negative isolates. Moreover, most isolates belonging to the same cluster shared an identical core-genome sequence type (cgST). For instance, cluster 1 included 4 isolates of cgST113036, of which only 3 harbored the IncI1-pST57 bla_CTX-M-3-positive plasmid. The 25 S. sonnei isolates were also subjected to phylogenetic comparison with deposited international strains. As a result, matching isolates (isolates that had the same cgST and that differed by <8 SNVs) have been reported in the United Kingdom, the United States, France, and the Netherlands. Overall, our results suggest that some common S. sonnei clusters can spread between continents and can be imported into other nations after international trips. Such clusters include, in part, isolates that do not possess bla_ESBL-harboring plasmids, indicating their tendency to acquire them from other Enterobacterales.

INTRODUCTION

Shigella flexneri is one of the most common causes of diarrhea in low- and middle-income countries and is associated with high morbidity and mortality rates. In contrast, Shigella sonnei is the leading species in high-income nations, with the majority of cases being described in returning travelers, men who have sex with men (MSM), and young children (1–3).

The emergence of antibiotic-resistant S. sonnei isolates is nowadays a matter of concern (4). The high rates of resistance to first-line options (e.g., ciprofloxacin and azithromycin) have made ceftriaxone the drug of choice for empirical treatment. However, there has also been a significant recent increase in extended-spectrum cephalosporin-resistant (ESC-R) isolates, especially in Asia (1, 5).

Usually, ESC-R S. sonnei isolates produce extended-spectrum β-lactamases (ESBL) of the CTX-M type, of which CTX-M-3, CTX-M-14, CTX-M-15, CTX-M-27, and CTX-M-55 are the most common (6). However, only a few studies have implemented whole-genome sequencing (WGS) to characterize the bla_CTX-M-carrying plasmids in detail. So far, a bla_CTX-M-3-harboring IncI1 plasmid from Italy (7), a bla_CTX-M-14-IncB/O/K/Z plasmid from China (8), a bla_CTX-M-15-harboring IncI1 plasmid from South Korea (9), a bla_CTX-M-27-harboring IncFII plasmid from the United Kingdom (10), and a bla_CTX-M-55-harboring IncI2 plasmid from China have been described in S. sonnei (11). For Switzerland, we note that the first two ESC-R S. sonnei strains (CTX-M-14 and CTX-M-15 producers) were isolated in 2009, but no WGS analyses were performed on the 8 strains detected during 2009 to 2014 (12).

Due to the ability of S. sonnei to acquire plasmids harboring multidrug resistance (MDR) genes (MDR plasmids) and the fact that it shows a higher prevalence in industrialized countries than S. flexneri (3), attention needs to be focused on the clonality of ESC-R S. sonnei isolates. Just recently (2018), in the European Union, 17 outbreaks due to S. sonnei have been documented (13).

Several authors have implemented multilocus sequence typing (MLST), which has revealed that sequence type 152 (ST152) is the most frequent lineage among extended-spectrum cephalosporin-susceptible (ESC-S) S. sonnei isolates (14–19). Other recent studies have also used core-genome analyses to investigate epidemiological events (2, 15, 20, 21), although only one survey in the United Kingdom analyzed exclusively ESBL-producing S. sonnei strains (10). Overall, these studies have shown that single nucleotide variant (SNV) analysis represents a high-resolution tool for determining clonality and tracking outbreaks at the community and global levels.

In this study, therefore, we used WGS to characterize the plasmids of ESC-R S. sonnei isolates detected in Switzerland. Moreover, to investigate the hidden epidemiological profile of contemporary circulating isolates, we implemented a core-genome analysis to determine the clonality of ESC-R and ESC-S S. sonnei strains.

RESULTS AND DISCUSSION

Rate of ESC-R S. sonnei and analyzed strains.

According to the Swiss Centre for Antibiotic Resistance (ANRESIS) database, 53, 39, 85, and 56 S. sonnei isolates were identified nationwide in 2016, 2017, 2018, and 2019, respectively, by participating laboratories. Of them, 2 (3.8%), 5 (12.8%), 12 (14.1%), and 21 (37.5%), respectively, were reported to be ESC-R.

Unfortunately, such results could not be compared to those of other countries. In fact, though the spread of ESC-R Shigella spp. is of concern, recent studies analyzing their trends are lacking (1, 4). Nevertheless, we note that in China the rate of ESC-R S. sonnei increased from 31.6% in 2012 to 64.3% in 2015 (22), whereas lower rates were recorded in other nations during point-prevalence surveys (e.g., in 2015, 12% in England/Wales and 0% in Nepal; in 2015 and 2016, 7.1% in New Zealand) (10, 23, 24). It is therefore difficult to interpret our data on the persistent increase in resistance to ESCs. In this context, we emphasize that the Swiss population is at greater risk of acquiring and importing MDR Shigella spp. from areas of endemicity due to its high propensity for international travel (12, 25). For this reason, in order to better understand this general epidemiological phenomenon, a molecular characterization of the strains is essential.

In the present study, we analyzed 14 ESC-R S. sonnei isolates and 11 ESC-S S. sonnei isolates collected in Switzerland during 2016 to 2019. Species identification (ID) and the antibiotic resistance phenotypes of all strains were confirmed by appropriate methods before further molecular analyses (see Materials and Methods and Table S1 in the supplemental material).

ARGs.

As shown in Table 1, both ESC-R and ESC-S S. sonnei isolates carried numerous antimicrobial resistance genes (ARGs) conferring resistance to different classes of antibiotics, including quinolones (e.g., qnrS1) and macrolides [e.g., erm(B) and mph(A)] (6). Those phenotypically resistant to ESCs mainly possessed bla_CTX-M-3 (n = 5) or bla_CTX-M-15 (n = 6) ESBL genes, but unique isolates harboring bla_CTX-M-27, bla_CTX-M-55, and bla_CTX-M-134 were also detected.

TABLE 1.

Summary of the demographic and travel-related (if any) data along with the results of WGS analyses of the S. sonnei strains analyzed in the present study^h

Strain	Yr of isolation	Age (yr)/sexa	Sample	Origin of infectiona	Groupb	Coresistanceb	STc	Antimicrobial resistance genes/plasmid replicons (pMLST, approximate size)c	No. (%) of called alleles among a total of 2,513d	No. (%) of allele matches in cgSTd	cgSTd
L4094	2018	NA/NA	NA	NA	ESC-R		ST152	blaCTX-M-3, aadA1, mdf(A), dfrA1/l1 (pST57, 86 kb), Col156, Col(BS512)	2,503 (99.60)	2,496 (99.32)	cgST113036
1205-3131	2018	35/M	Stool	Unknown	ESC-R		ST152	blaCTX-M-3, aadA1, mdf(A), dfrA1/l1 (pST57, 86 kb), Col156, Col(BS512)	2,500 (99.48)	2,492 (99.16)	cgST113036
7111-69	2019	20/M	Stool	Turkey	ESC-R		ST152	blaCTX-M-3, aadA1, mdf(A), dfrA1/l1 (pST57, 86 kb), Col156, Col(BS512)	2,499 (99.44)	2,491 (99.12)	cgST113036
LC-1477-18e	2018	10/F	Stool	Albania	ESC-R	SXT	ST152	blaCTX-M-3, aadA1, aph(3'')-Ib, aph(6)-Id, mdf(A), dfrA1, sul2, tet(A)/I1 (pST57, 85 kb), Col156, Col(BS512)	2,505 (99.68)	2,498 (99.40)	cgST118753
509-1022	2019	50/F	Stool	Unknown	ESC-R	SXT	ST152	blaCTX-M-3, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/l1 (pST57, 88 kb), Col156, Col(BS512)	2,503 (99.60)	2,495 (99.28)	cgST115537
19-0822-3296	2019	5/F	Stool	Unknown	ESC-R	SXT	ST152	blaCTX-M-3, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/l1 (pST57, 88 kb), Col156, Col(BS512)	2,504 (99.64)	2,495 (99.28)	cgST115537
6607-69	2017	60/F	Stool	Sri Lanka	ESC-R	SXT	ST1503f	blaCTX-M-15, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, tet(A), sul2/I1 (pST16, 90 kb), FII, Col156, Col(BS512)	2,504 (99.64)	2,496 (99.32)	cgST64457
19-0821-3486	2019	45/M	Stool	Egypt	ESC-R	SXT	ST152	blaCTX-M-15, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, qnrS1, sul2, tet(A)/FII (F2:A-:B-, 83 kb), Col156, Col(BS512)	2,500 (99.48)	2,493 (99.20)	cgST112958
0401930105	2019	50/M	Stool	Egypt	ESC-R	SXT	ST152	blaCTX-M-15, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, qnrS1, tet(A), sul2/FII (F2:A-:B-, 83 kb), Col156, Col(BS512)	2,502 (99.56)	2,496 (99.32)	cgST112958
6904-27	2018	35/M	Stool	Local	ESC-R	SXT	ST152	blaCTX-M-15, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, qnrS1, tet(A), sul2/FII (F2:A-:B-, 83 kb), Col156, Col(BS512)	2,502 (99.56)	2,497 (99.36)	cgST112958
19-1125-3493	2019	40/F	Stool	Unknown	ESC-R	SXT	ST152	blaCTX-M-15, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, qnrS1, tet(A), sul2/FII (F2:A-:B-, 83 kb), Col156, Col(BS512)	2,496 (99.32)	2,489 (99.04)	cgST112958
19-0820-1561	2019	15/F	Stool	Nepal	ESC-R	SXT, CIP, AZT	ST152	blaCTX-M-15, aadA5, mdf(A), mph(A), dfrA1, dfrA17, qnrS1, sul1/FII (F2: A-:B-, 83 kb), Col156, Col(BS512)	2,493 (99.20)	2,477 (98.57)	cgST117387
0401952027	2019	45/M	Stool	Egypt	ESC-R	SXT, AZT	ST152	blaCTX-M-55, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), mph(A), dfrA1, sul2, tet(A)/I1, FII (F2:A-:B-, 74 kb), Col156, Col(BS512)	2,498 (99.40)	2,489 (99.04)	cgST20888
09163633	2019	50/M	Stool	Unknown	ESC-R	SXT	ST152	blaCTX-M-27, aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/FII (F2:A-:B-, 68 kb), B/O/K/Z, Col156, Col(BS512)	2,497 (99.36)	2,489 (99.04)	cgST67380
3123885	2019	30/M	Stool	Israel	ESC-R	SXT, CIP, AZT	ST152	blaCTX-M-134, aph(6)-ld, mdf(A), mph(A), dfrA1, sul2, tet(A)/FII (F35:A-:B-, 69 kb), B/O/K/Z, Col156, Col(BS512)	2,489 (99.04)	2,480 (98.69)	cgST114011
7103-58g	2018	10/M	Stool	Romania	ESC-S	SXT, CIP, AZT	ST152	blaTEM-1B, aadA5, aph(3″)-lb, aph(6)-ld, mdf(A), mph(A), erm(B), dfrA1/A17, sul1/2, tet(A)/Col156, Col(BS512)	2,489 (99.04)	2,482 (98.77)	cgST107674
7103-28g	2018	50/M	Stool	Romania	ESC-S	SXT, CIP, AZT	ST152	blaTEM-1B, aadA5, aph(3″)-lb, aph(6)-ld, mdf(A), mph(A), erm(B), dfrA1/A17, sul1/2, tet(A)/Col156, Col(BS512)	2,490 (99.08)	2,481 (98.73)	cgST107674
6407-57	2017	40/M	Stool	Local	ESC-S	SXT, CIP, AZT	ST152	blaTEM-1B, aadA5, aph(3″)-lb, aph(6)-ld, mdf(A), mph(A), erm(B), dfrA1/A17, sul1/2, tet(A)/I1, Col156, Col(BS512)	2,492 (99.16)	2,484 (98.85)	cgST108068
6110-62	2016	60/F	Stool	Brazil	ESC-S	SXT	ST152	blaTEM-1B, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA8, sul2/FII, Col156, Col(BS512)	2,499 (99.44)	2,492 (99.16)	cgST108083
6105-15	2016	35/M	Stool	Egypt	ESC-S	SXT	ST152	aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/FII, Col156, Col(BS512)	2,497 (99.36)	2,487 (98.97)	cgST37499
6101-40	2016	40/F	Stool	Western Africa	ESC-S	SXT	ST152	aadA1, aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/FII, Col156, Col(BS512)	2,492 (99.16)	2,479 (98.65)	cgST98334
6412-75	2017	50/M	Blood, stool	India	ESC-S	CIP	ST152	mdf(A), dfrA1/Col156, Col(BS512)	2,495 (99.28)	2,487 (98.97)	cgST101592
6502-32	2017	40/F	Stool	Dominican Republic	ESC-S	SXT	ST152	aph(3″)-lb, aph(6)-ld, mdf(A), dfrA1, sul2, tet(A)/I1, Col156, Col(BS512)	2,494 (99.24)	2,484 (98.85)	cgST108763
7111-23	2019	25/F	Stool	Philippines	ESC-S		ST152	sul2, dfrA14, aph(3″)-lb, aph(6)-ld, mdf(A)/I1, FII, Col156	2,495 (99.28)	2,481 (98.73)	cgST108909
7109-28	2019	30/M	Stool	Colombia	ESC-S	SXT	ST152	aadA1, aph(3″)-lb, aph(6)-ld, sul2, mdf(A), dfrA1, qnrB19, qnrB5, qnrB81, tet(A)/I1, Col156	2,499 (99.44)	2,486 (98.93)	cgST109254
7001-38	2018	50/F	Stool	Local	ESC-S	SXT	ST152	aadA1, mdf(A), dfrA1/Col156, Col(BS512)	2,503 (99.60)	2,494 (99.24)	cgST113036

^aAge and sex are based on the information provided to the clinical laboratory analyzing routine samples. Age has been approximated at ±5 years. The origin of infection has been attributed to a foreign country if symptoms (i.e., diarrhea) occurred during or after the patients returned from a specific country.

^bGroups are based on the MICs obtained with the Sensititre GNX2F panel and interpreted according to the EUCAST 2019 criteria. Only key antibiotics to which nonsusceptibility is shown are reported (see Table S1 in the supplemental material for full MIC results).

^cPerformed by implementing the tools of the Center for Genomic Epidemiology (CGE), specifically, MLST (v2.0; E. coli scheme 1), ResFinder (v3.2), PlasmidFinder (v2.0), and pMLST (v2.0), when available. The main bla genes and the associated plasmids carrying these genes are indicated in bold.

^dCore-genome results obtained with cgMLSTFinder (v1.1).

^eStrain LC-1477-18 was detected in Italy (7). It was added to the analysis as a control.

^fST1503 is a single allele variant of ST152.

^gThe patients from whom strains 7103-58 and 7103-28 were isolated are relatives.

^hM, male; F, female; ESC-R, extended-spectrum cephalosporin resistant; ESC-S, extended-spectrum cephalosporin susceptible; SXT, trimethoprim-sulfamethoxazole; CIP, ciprofloxacin; AZT, azithromycin; ST, sequence type; cgST, core-genome sequence type; NA, not available.

Although studies analyzing the prevalence of specific bla_ESBL genes in S. sonnei are lacking, bla_CTX-M-15 and bla_CTX-M-3 appear to be the most frequent worldwide (6). In particular, CTX-M-3 producers were described in Turkey, Switzerland, and Italy (7, 12, 26, 27), while those with CTX-M-15 have been found in various countries, including South Korea, where an outbreak was described (9). With regard to the other ESBLs, a CTX-M-27-producing S. sonnei clone was responsible for an outbreak in 2015 among MSM in England (10) and CTX-M-55 was reported in S. sonnei isolates from China and South Korea (11, 28), while CTX-M-134 was only recently described in Escherichia coli (29).

Since bla_CTX-M genes are usually carried by plasmids that can be exchanged between different species of enterobacteria (e.g., from E. coli to S. sonnei in the human gut) (30), their characterization is crucial for understanding the expansion of ESC-R S. sonnei isolates.

IncI1 bla_CTX-M-3-carrying plasmids.

As shown in Fig. 1A, 5 S. sonnei isolates harbored 86- to 87-kb IncI1-pST57 bla_CTX-M-3-carrying plasmids with a high degree of genetic identity to each other and to the Italian pLC1477_18_1 plasmid that we recently described (conjugation frequency, 1.2 × 10⁻⁴) (7).

FIG 1.

BLAST comparisons of the sequences of S. sonnei bla_CTX-M-carrying plasmids against reference sequences. (A) Comparison of the sequences of five S. sonnei bla_CTX-M-3-carrying IncI1 plasmids against the sequence of the S. sonnei IncI1 plasmid pLC1477_18-1 (GenBank accession number CP035009) as a reference sequence. (B) Comparison of the sequences of five S. sonnei bla_CTX-M-15-carrying IncFII (F2:A-:B-) plasmids against the sequence of the K. pneumoniae IncFII (F2:A-:B-) plasmid pF93-2_1 (GenBank accession number CP026158) as a reference sequence. Rings were constructed using BRIG (BLAST Ring Image Generator; v0.95) software. Similarities with the reference plasmid are represented by the colored rings. Genome accession numbers are indicated in the key at the bottom. Red and blue arrows above the rings correspond to the gene features of interest. The delta symbol (Δ) next to the feature label corresponds to a partial/incomplete gene CDS. For each plasmid, we report the GenBank accession number, the species of isolation, the tree cluster from Fig. 2, the year, the plasmid name, and the plasmid size.

In all of these plasmids, bla_CTX-M-3 was associated with a truncated ISEcp1 in the same element reported in the Italian plasmid. Considering the strong genetic similarity of the five IncI1-pST57 bla_CTX-M-3-carrying plasmids collected in Switzerland, it is quite possible that other plasmids with similar genetic characteristics exist outside the country. For instance, an Enterobase S. sonnei strain from the United Kingdom recovered in 2019 (named strain 811053; BioSample accession number SAMN12881824) of the same core-genome sequence type (cgST) as our 509-1022 and 19-0822-3296 isolates (cgST115537) was found to harbor bla_CTX-M-3 and contained at least one pST57 plasmid (Center for Genomic Epidemiology [CGE] analysis; data not shown). Similarly, an E. coli strain from a study conducted in the Netherlands 2013 to 2015 was reported to carry an IncI1-pST57 plasmid that possessed bla_CTX-M-3 (31).

IncFII bla_CTX-M-15-harboring plasmids.

As depicted in Fig. 1B, 5 other ESC-R S. sonnei isolates harbored IncFII (F2:A-:B-) bla_CTX-M-15-carrying plasmids (83 to 89 kb) with a high degree of genetic identity to each other and to pF93-2_1 (GenBank accession number CP026158) from Klebsiella pneumoniae found in China in 2014. Globally, F2:A-:B- is the predominant F plasmid type carrying bla_CTX-M genes among Enterobacteriaceae and is highly conjugative (32).

The 5 bla_CTX-M-15-carrying plasmids identified in the present work shared an identical genetic environment around bla_CTX-M-15, including both full (IS26 and ISKpn19) and partial (ΔTn3 and ΔIS3) transposable element-coding sequences (CDS), along with qnrS1 (Fig. 1B). The plasmid p19-0820-1561 also contained additional ARGs [mph(A), sul1, aadA5, and dfrA17] that were present only in pF93-2_1 in the form of the IS26-mph(A)-MFS transporter-tetR or acrR-IS6100-sul1-aadA5-dfrA17-intl1Δ-IS26-Tn3Δ-IS1RΔ unit. This element has been reported in multiple species, such as E. coli and K. pneumoniae (BLAST analysis; data not shown).

Other bla_CTX-M-carrying plasmids.

The remaining 4 ESC-R S. sonnei isolates possessed unique bla_CTX-M-positive plasmids (Table 1). In particular, a bla_CTX-M-15 gene associated with ISEcp1 was carried in an 89-kb IncI1-pST16 plasmid (p6607-69), but without further ARGs. This plasmid showed a high degree of identity with others found in both S. sonnei and E. coli isolates (mostly from Asia), including some expressing CTX-M-55, which is a single amino acid variant of CTX-M-15 (Fig. S1). Likewise, one of our S. sonnei isolates carried bla_CTX-M-55, but in this case, the gene was located in a 74-kb IncFII (F2:A-:B-) plasmid (p0401952027) and was flanked by two IS26 elements. This plasmid showed a high degree of genetic identity with others possessing bla_CTX-M-55 or bla_CTX-M-15 that came from E. coli or K. pneumoniae isolates detected in Europe or North America, but, interestingly, none of them cocarried the tetR or acrR-MFS trans-mph(A) unit between IS6100 and IS1RΔ (Fig. S2).

Another ESC-R S. sonnei isolate carried a 67-kb IncFII (F2:A-:B-) plasmid (p09163633) that harbored only bla_CTX-M-27 and that showed a high degree of identity with the backbone of plasmids from E. coli and S. flexneri isolates. Nevertheless, only p09163633 possessed the IS26-IS903BΔ-bla_CTX-M-27-ISEcp1Δ-IS26 unit (Fig. S3), which has been reported in multiple E. coli and K. pneumoniae isolates from Vietnam and China (BLAST analysis; data not shown) and which has also been described in a Japanese epidemic ST131 E. coli isolate (33). Moreover, it was also present in the 69-kb IncFII (F35:A-:B-) plasmid (p3123885) found in our last ESC-R S. sonnei isolate (Fig. S4), though this mobile genetic element (MGE) encoded the single amino acid variant CTX-M-134 instead of CTX-M-27 (29).

Coresistance to azithromycin.

Besides the specific ESBLs identified, 6 of the 25 S. sonnei isolates were macrolide resistant due to the presence of erm(B) and/or mph(A) ARGs (Table 1; Table S1). As mentioned above, 3 ESC-R isolates carried mph(A) in different IncFII plasmids coharboring bla_CTX-M-15, bla_CTX-M-55, or bla_CTX-M-134. Of note, two of these plasmids carried the element IS26-mph(A)-MFS trans-tetR or acrR-IS6100 (Fig. 1B; Fig. S4) and the other one carried, with a slightly different arrangement, IS6100-tetR or acrR-MFS trans-mph(A)-IS1RΔ (Fig. S2). These two very similar elements have been found in many plasmids carried by E. coli, K. pneumoniae, and Salmonella enterica (BLAST analysis; data not shown) and have also been identified in the chromosome of a CTX-M-15-producing S. enterica serovar Haardt isolate recovered from Japanese food workers (34).

Overall, these findings are epidemiologically relevant, since coresistance to azithromycin and ESCs makes the treatment of shigellosis difficult (1, 5). Such MDR plasmids have rarely been reported in S. sonnei, though an IncFII (F2:A-:B-) plasmid possessing bla_CTX-M-27, mph(A), and erm(B) was associated with the outbreak among MSM in England (10), while an IncB/O/K/Z plasmid coharboring bla_CTX-M-14 and mph(A) was linked to a waterborne outbreak in China in 2015 (8). Having observed that at least three of our IncFII plasmids carried very similar macrolide resistance elements, we speculate that under a certain antibiotic selective pressure (e.g., azithromycin), mph(A) can be acquired via integration of transposable elements [e.g., IS26-mph(A)-MFS trans-tetR or acrR-IS6100] (35).

MLST and cgMLST.

Regardless of the presence of bla_CTX-M genes, 24 S. sonnei isolates were of ST152, while 1 was of its single allele variant, ST1503 (Table 1). ST152 has been previously reported in ESC-S S. sonnei isolates in many countries (e.g., the United States [California], China, Germany, and Iran) (15–19). Recently, we also described the ST152 CTX-M-3-producing strain LC-1477-18, isolated in Italy from a girl who acquired the infection in Albania (7). Overall, since our S. sonnei isolates were acquired in different periods and/or in diverse geographic areas (Table 1), one could speculate that a unique clone (ST152) is spreading worldwide.

To better investigate the clonality of our S. sonnei isolates, we performed a core-genome MLST (cgMLST) analysis according to the E. coli scheme. The higher resolution of cgMLST resulted in multiple cgSTs: (i) three bla_CTX-M-3-possessing isolates and one ESC-S isolate were of cgST113036, (ii) two bla_CTX-M-3-positive isolates were of cgST115537, (iii) four bla_CTX-M-15-harboring isolates were of cgST112958, and (iv) two ESC-S isolates carrying mph(A) and erm(B) were of cgST107674. The remaining isolates showed different cgSTs, but overall, high allele matches were maintained among the 2,513 isolates analyzed (i.e., >98.5% for ESBL producers and >98.6% for those ESC-S isolates) (Table 1). These results support the hypothesis that common ESC-R clones may spread in different countries and could be imported to other nations (e.g., Switzerland) after international trips. Based on the identification of clones including both ESBL producers and nonproducers, it can be also speculated that some ESC-S S. sonnei isolates may be well predisposed to acquire MDR plasmids from other Enterobacterales.

Core-genome analyses.

To determine the distance and clonality between our S. sonnei strains, a high-resolution SNV analysis was performed (87% core-genome coverage among all strains). As shown in Fig. 2, the 4 S. sonnei clones identified by cgMLST were also confirmed to be 4 independent SNV clusters, though several additional isolates were grouped within cluster 2 and cluster 4. To summarize: cluster 1 and cluster 2 included CTX-M-3-producing or ESC-S S. sonnei isolates, cluster 3 encompassed CTX-M-15 producers, and cluster 4 included the CTX-M-134 producer and the ESC-S S. sonnei isolates. Notably, strains belonging to the same cluster differed by only a limited number of SNVs (i.e., for cluster 1, the difference was 7 to 26 SNVs; for cluster 2, the difference was 12 to 34 SNVs; for cluster 3, the difference was 5 to 19 SNVs; and for cluster 4, the difference was 2 to 58 SNVs).

FIG 2.

Analysis of the core-genome phylogeny of 25 S. sonnei isolates together with that of Italian strain LC-1477-18. For each strain, we show the strain, the collection year, the main β-lactamase (if present), and the cgST. The assembled WGS of strains is presented in a core-genome SNV tree. The differences in the SNV value (e.g., Δ = 1 SNV) corresponds to the number of nonidentical SNVs of the core genome between two strains. The four main clusters (gray boxes) were defined when the nucleotide identity across two or more strains was ≥97.5% of shared SNVs (difference of ≤65 SNVs). The cluster matrix shows the maximum nucleotide identity (in percent) between all strains across two clusters (top right corner) and the number of SNVs not shared among all strains compared. The scale bar (0.05) represents the average number of nucleotide substitutions per site. Asterisks represent identical cgSTs, as determined by CGE’s cgMLSTFinder (v1.1) program. ^a, the core genome represents the maximum total coverage (87%) of the alignment among all 26 S. sonnei conserved sequences, which corresponded to 2,608 SNVs; ^b, cluster 1 strains shared 98.77% SNVs; ^c, cluster 2 strains shared 98.01% SNVs; ^d cluster 3, strains shared 99.08% SNVs; ^e, cluster 4 strains shared 97.54% SNVs; ^f, in cluster 1 and cluster 2, bla_CTX-M-3 was consistently carried by the same IncI1-pST57 plasmid (Fig. 1A).

Together, these results corroborate the above-mentioned hypothesis on the dissemination of CTX-M-producing hyperepidemic S. sonnei clones. This is consistent with what has been observed by other authors for the fluoroquinolone-resistant international clones (e.g., global lineage III [GIII]) (36, 37). However, our data also indicate that MLST analysis alone has a limited resolution for studying the spread of such MDR pathogens. In fact, although almost all of our S. sonnei isolates were identified to be ST152, several clusters with different ARG and plasmid patterns could be differentiated using cgMLST and/or core-genome SNV analyses.

We also note that isolates included in cluster 1 and cluster 2 carried the same IncI1-pST57 bla_CTX-M-3-harboring plasmid (Fig. 1). This finding was surprising, as the two bacterial groups were genetically different (i.e., a difference of 235 SNVs; Fig. 2). We do not have a clear explanation for the independent clustering, but it can be hypothesized that ESC-S S. sonnei isolates belonging to cluster 1 and cluster 2 acquired the pST57 plasmid from a common enterobacterial ancestor, including other Shigella spp.

Link with international isolates.

To explore the lineage origins of our Swiss S. sonnei isolates, we performed a search of the Enterobase database for strains of global lineage and matching cgST. A core-genome analysis was then performed using the results of the Enterobase phylogenetic analysis as a reference. Overall, we compared our collection (n = 25) to a subset of 114 strains of global lineage (38), 16 with a matching cgST, and the Italian LC-1477-18 isolate (7). The alignment of all strains (n = 156) resulted in 4,551 SNVs and 42% core-genome coverage among all isolates.

As depicted in Fig. 3 and Table S2, all S. sonnei isolates included in cluster 1 were related to an Iranian GIII strain detected in 2003 and were almost identical to another one found in the United Kingdom in 2016 (with the same cgST and a difference of 2 to 3 SNVs). We also note that one of our isolates (isolate 7111-69) had its clinical origin in Turkey (Table 1), the area where the first CTX-M-3-producing S. sonnei isolate was described (2001), and subsequently caused epidemic events (26, 27).

FIG 3.

Enterobase SNV tree dendrogram of the Swiss S. sonnei strains (n = 25), the global lineage of strains (n = 114), strains with matching cgST (n = 16), and the Italian LC-1477-18 strain as a reference. The combined single nucleotide polymorphism profiles of all 156 strains mapped to the reference sequence are represented in a RAxML tree, corresponding to a total of 9,850 SNVs. Country labels are represented by colored circles (missing country labels correspond to strains IBESS820 from the Netherlands, Ss046 from China, and 53G from South Korea). Holt lineages I, II, III, GIII, and IV are presented in colored boxes. Dashed black brackets with lines correspond to zoom-in sections of the tree where the present study’s strains are clustered. For the Swiss isolates, we show the cluster (if any)/ESBL (if any), while for the international strains with a matching cgST, we show only the ESBL (if any). Among these strains, those detected in 2019 are indicated with an asterisk. The scale bar represents the average number of nucleotide substitutions per site. See Table S2 in the supplemental material for more information regarding the Parsnp program SNV analysis results for zoom-in sections 1 to 7: section 1, compared to the Swiss isolates possessing the same cgST, strains 811053, 266979, 175609, 152507, CFSAN091705, and 795376 show 0 to 1, 0 to 4, 1 to 5, 2, 5, and 4 SNVs, respectively; section 2, compared to the Swiss isolate, strain 526163 has the same cgST and shows 2 SNVs; section 3, compared to the Swiss isolates, strain 299890 has the same cgST and shows 2 to 3 SNVs; section 4, compared to the Swiss isolates possessing the same cgST, strains 190807, 201907857, 638735, IBESS820, 821179, and PNUSAE013040 show 6, 1, 1, 0, 8, and 1 SNVs, respectively; section 5, 61 SNVs are found between the Swiss isolate and the one isolate found in Egypt; section 6, compared to the Swiss isolate, strain 191891 has the same cgST and shows 2 SNVs; section 7, compared to the Swiss isolate, strain 524350 has the same cgST and shows 4 SNVs.

The cluster 2 and cluster 3 isolates were part of a large group shared by two GIII strains of Egyptian origin reported in 2005 and 2006, LC-1477-18, and four isolates detected in the United Kingdom: one CTX-M-3 producer isolated in 2019, two CTX-M-15 producers detected in 2015 and 2016, and one ESC-S isolate found in 2015 (Fig. 3). These UK strains showed differences in SNVs of ≤5 compared to the sequences of cgST-matching isolates from Switzerland, indicating their commonality. Further evidence of their possible origin could be seen in isolates 6105-15, 0401930105, 19-0821-3486, and 0401952027, the origin of which was Egypt (Table 1), suggesting that these strains may have originated in that geographic area.

S. sonnei cluster 4 isolates were grouped with recently detected ESC-S strains in the United Kingdom, the United States, and France. Two additional CTX-M-15 producers, one from the United Kingdom (821179) and one from the Netherlands (IBESS820), were also highly related to the cluster 4 isolates (Fig. 3). The latter was identified in 2017 during a cross-sectional multicenter study (39) and was genetically identical to our ESC-S S. sonnei 6412-75 strain (they had the same cgST and a difference of 0 SNPs). In that study, the patient from whom strain IBESS820 was recovered was reported to have a history of travel to India, as did the Swiss patient with an infection caused by 6412-75 (Table 1). Despite these similarities, the Swiss S. sonnei isolate was ESBL negative; nevertheless, this finding highlights the great capacity of certain clones to acquire bla_CTX-M-harboring plasmids.

In total, 12 of the 16 international S. sonnei strains that had the same cgST as our Swiss isolates were detected in the United Kingdom, while the remaining 4 were isolated in the United States, France, and the Netherlands. This indicates that common S. sonnei lineages have been circulating in Europe at least since 2015 and are now expanding in Switzerland. In fact, we note that all 10 Swiss ESC-R S. sonnei isolates detected in 2019 were linked to isolates detected in the same year in the United Kingdom and France (Table 1 and Fig. 3), and most of these were producers of CTX-M-3 or CTX-M-15, as in the case of those in Switzerland.

It can be speculated that the same common plasmids described in the present work (e.g., the IncI1-pST57 bla_CTX-M-3-positive plasmid) are also carried by contemporary non-Swiss ESC-R S. sonnei isolates (as demonstrated for LC-1477-18). However, since the matching cgST isolates identified from the Enterobase database are in the form of whole-genome shotgun assemblies generated from short-read sequencing data, without the full characterization of bla_CTX-M-carrying plasmids with long-read sequencing data, as in our study, this hypothesis cannot be fully corroborated.

Conclusions.

In this work, we present the first detailed molecular investigation of S. sonnei isolates detected in Switzerland. Hybrid WGS assemblies were implemented to accurately describe the bla_CTX-M-harboring plasmids, while core-genome and phylogenetic analyses were used to study the clonality of the strains.

Based on our results, we conclude that most of the contemporary Swiss ESBL-producing S. sonnei isolates carry identical bla_CTX-M-positive plasmids that often have their counterparts in other reported Enterobacterales worldwide. More importantly, due to transnational travel, common international clones of MDR S. sonnei are emerging in Switzerland, and this limits our therapeutic armamentarium. Overall, our findings underline the importance of continuously conducting epidemiological surveys using the WGS approach and linking the results with those from other countries (40).

MATERIALS AND METHODS

Epidemiological data.

Phenotypic data regarding the S. sonnei isolates detected in Switzerland during 2016 to 2019 were retrieved from the Swiss Centre for Antibiotic Resistance (ANRESIS) database (http://www.anresis.ch/), which collects information from 30 Swiss clinical laboratories. Strains were categorized as ESC-R when nonsusceptible (i.e., intermediate or resistant) to ceftazidime, ceftriaxone, and/or cefepime according to the criteria implemented for Enterobacterales by the routine clinical laboratories during the corresponding years. The research project was exempted from the requirement for ethical approval because no health-related personal data were used, while age, gender, and trip-related information (if available) were retrieved from the laboratory databases.

Strains, ID, and antimicrobial susceptibility tests.

All ESC-R and ESC-S S. sonnei isolates available at −80°C and collected during 2016 to 2019 at the Institute for Infectious Diseases, MCL Medizinische Laboratorien, and labormedizinisches zentrum Dr. Risch were analyzed.

The initial identification, obtained by implementing matrix-assisted laser desorption ionization–time of flight mass spectrometry (Bruker), was confirmed with the Type Strain Genome Server (TYGS) tools using the Genome BLAST Distance Phylogeny approach (https://tygs.dsmz.de/), based on genome data (see below). MICs were obtained by implementing Sensititre GNX2F microdilution panels (Thermo Fisher). For isolates possessing the erm(B) and/or mph(A) gene, MICs for azithromycin were obtained using the Etest (bioMérieux). Results were interpreted according to the EUCAST 2019 criteria (41).

Whole-genome sequencing (WGS).

WGS was performed using both the NovaSeq-6000 (Illumina) and the MinION (Oxford Nanopore) sequencing platforms as previously described (42–45). In brief, Illumina raw reads were quality filtered with the Trimmomatic (v0.36) trimmer, followed by whole-genome shotgun assembly with the SPAdes (v3.12.0) program. Adaptors from Nanopore raw reads were trimmed with the Porechop (v0.2.4) program and quality filtered with the Filtlong (v0.2.0) program. Long-read assemblies were done with the Canu (v1.7) program. The final hybrid assemblies were generated by aligning the paired-end Illumina reads to the Canu assemblies with the Bowtie2 (v2.3.4.1) program, followed by multiple rounds of polishing with the Pilon (v1.22) program.

Illumina SPAdes assemblies were used for whole-genome ID and analysis with the tools of the Center for Genomic Epidemiology (CGE; http://www.genomicepidemiology.org/): ResFinder, MLST with E. coli scheme 1, PlasmidFinder, and pMLST. Hybrid assemblies were used to characterize the bla_ESBL-carrying plasmids.

Annotations of both the Illumina and the hybrid assemblies were carried out by the NCBI Prokaryotic Genome Annotation Pipeline. All annotated features presented in Fig. 1 and Fig. S1 to S4 in the supplemental material were manually curated with the UniProt (https://www.uniprot.org/blast/) and ISfinder (https://isfinder.biotoul.fr/) tools and annotated accordingly.

Core-genome analyses.

All S. sonnei isolates underwent cgMLST with the CGE cgMLSTFinder (v1.1) program, using Illumina raw reads as the input and the species database set to E. coli Enterobase. These isolates also underwent core-genome SNV analysis as previously done (46). Briefly, the core-genome alignment was performed with the Parsnp (v1.2) program. All strains were treated as curated genomes (−C parameter), and the Italian ST152 CTX-M-3-producing S. sonnei strain (strain LC-1477-18; GenBank accession number JAATWD000000000) was used as a reference (7). The −C parameter was set to 200, and the other parameters were left at the default. Variants with no flags (PASS) were determined to be reliable and used for downstream SNV analysis with a custom R (v3.6.2) script. The Parsnp program-generated core-genome SNV phylogenetic tree was visualized with the FigTree (v1.4.4) program and set to midpoint rooted and nodes by decreasing order (Fig. 2).

An SNV tree dendrogram of the Swiss S. sonnei collection versus global lineage and matching cgST strains was created in the Enterobase Escherichia/Shigella database (https://enterobase.warwick.ac.uk/species/index/ecoli) (Fig. 3). The analyzed strains consisted of 156 total strains, of which 114 were of global lineages (38), 16 had matching cgSTs, 25 were from Switzerland, and 1 was the Italian LC-1477-18 strain, used as a reference. The following search queries were used to find the global lineage strains (date, 10 April 2020) in Enterobase: species equals “Shigella sonnei”; comment contains “Holt lineage”; and to find matching cgST strains (date, April 21, 2020), the experiment type was cgMLST V1 + HierCC V1 and ST were 108909, 108083, 64457, 108763, 117387, 101592, 108068, 114011, 107674, 113036, 67380, 109254, 98334, 20888, 37499, 118753, 115537, and 112958. The Illumina raw reads of our 25 S. sonnei isolates were uploaded to Enterobase for processing. The resulting genomes assembled by Enterobase were used to create an SNV project of 156 strains with default settings (minimum percentage of sites present, 95). The tree was visualized with the web-based browser.

An independent core-genome analysis (Table S2) with the Parsnp program was used to analyze the strain clusters identified in Fig. 3, which also included the 156 S. sonnei assemblies. As described above, the genome of the Italian LC-1477-18 strain was used as the reference genome, the Parsnp −C parameter was set to 300, and the rest of the parameters were set as the default.

Data availability.

The Illumina SPAdes assemblies were deposited in GenBank under BioProject accession number PRJNA578838. The hybrid assemblies (bla_ESBL-carrying plasmids and the corresponding chromosomes) were deposited in GenBank under BioProject accession number PRJNA578858.