A novel variant in Salmonella genomic island 1 of multidrug-resistant Salmonella enterica serovar Kentucky ST198

ABSTRACT Salmonella enterica serovar Kentucky ST198 is a major health threat due to its resistance to ciprofloxacin and several other drugs, including third-generation cephalosporins. Many drug-resistant genes have been identified in the Salmonella genomic island 1 variant K (SGI1-K). In this study, we investigated the antimicrobial resistance (AMR) profile and genotypic relatedness of two isolates of ciprofloxacin-resistant (CIPR) S. Kentucky ST198 from poultry in Northeastern Thailand. We successfully assembled the complete genomes of both isolates, namely SSSE-01 and SSSE-03, using hybrid de novo assembly of both short- and long-read sequence data. The complete genomes revealed their highly similar genomic structures and a novel variant of SGI1-K underlying multidrug-resistant (MDR) patterns, including the presence of blaTEM-1b, which confers resistance to beta-lactams, including cephalosporins and lnu(F) which confers resistance to lincomycin and other lincosamides. In addition, the chromosomal mutations in the quinolone resistance-determining region (QRDR) were found at positions 83 (Ser83Phe) and 87 (Asp87Asn) of GyrA and at positions 57 (Thr57Ser) and 80 (Ser80Ile) of ParC suggesting high resistance to ciprofloxacin. We also compared SSSE-01 and SSSE-03 with publicly available complete genome data and revealed significant variations in SGI1-K genetic structures and variable relationships to antibiotic resistance. In comparison to the other isolates, SGI1-K of SSSE-01 and SSSE-03 had a relatively large deletion in the backbone, spanning from S011 (traG∆) to S027 (resG), and the inversion of the IS26-S044∆-yidY segment. Their MDR region was characterized by the inversion of a large segment, including the mer operon and the relocation of IntI1 and several resistance genes downstream of the IS26-S044∆-yidY segment. These structural changes were likely mediated by the recombination of IS26. The findings broaden our understanding of the possible evolution pathway of SGI1-K in fostering drug resistance, which may provide opportunities to control these MDR strains. IMPORTANCE The emergence of ciprofloxacin-resistant (CIPR) Salmonella Kentucky ST198 globally has raised significant concerns. This study focuses on two poultry isolates from Thailand, revealing a distinct Salmonella genomic island 1 variant K (SGI1-K) genetic structure. Remarkably, multiple antibiotic resistance genes (ARGs) were identified within the SGI1-K as well as other locations in the chromosome, but not in plasmids. Comparing the SGI1-K genetic structures among global and even within-country isolates unveiled substantial variations. Intriguingly, certain isolates lacked ARGs within the SGI1-K, while others had ARGs relocated outside. The presence of chromosomal extended-spectrum β-lactamase (ESBL) genes and lincosamide resistance, lnu(F), gene, could potentially inform the choices of the treatment of CIPRS. Kentucky ST198 infections in humans. This study highlights the importance of understanding the diverse genetic structures of SGI1-K and emphasizes the role of animals and humans in the emergence of antimicrobial resistance.

insight into the spread and evolution of antibiotic resistance genes in CIP R S. Kentucky ST 198.

Phenotypic characteristics of the studied isolates
The resistance to 10 antimicrobials was determined using the disk diffusion methods.Both strains exhibited similar levels of resistance to streptomycin, ampicillin, cefotax ime, norfloxacin, nalidixic acid, ciprofloxacin, and tetracycline, as shown in Table S1.However, SSSE-01 additionally showed intermediate resistance to amoxicillin-clavulanic acid.Therefore, both isolates were considered MDR including resistance to a third-gener ation cephalosporin and ciprofloxacin.

The complete genomes of S. Kentucky SSSE-01 and SSSE-03
Both isolates were whole genome sequenced using both short-and long-read technol ogies on Illumina Nextseq500 and MinION Nanopore platforms, respectively.De novo hybrid assembly of both isolates was successfully performed using Unicycler V 0.4.8, with the sequence qualities summarized in Table S2.The hybrid assembly of both isolates resulted in four contigs each, comprising a complete-circular chromosome (4,870,588 bp and 4,869,812 bp for SSSE-01 and SSSE-03, respectively) and three circular-replicon col plasmids (4,018, 2,257, and 2,097 bp for both isolates).The genomes were annotated using the NCBI Prokaryotic Genome Annotation Pipeline and summarized in Tables S2  and S3.SSSE-01 and SSSE-03 were predicted to have 4,621 and 4,620 coding sequences (CDS), respectively.

Multilocus sequence typing profile and phylogeny of the 28 isolates of S. Kentucky ST198
To investigate the phylogenetic relationships of our samples in the broader context of S. Kentucky diversity, we compiled complete genome sequences of S. Kentucky (n = 28) from the NCBI Genome database (https://www.ncbi.nlm.nih.gov/genome/,accessed on January 2023) that have annotation status of "Annotated by NCBI RefSeq." Their metadata are provided in Table S4.Among these, eight samples were isolated from humans, while all others were isolated from animals or meat products.These isolates were from Spain (n = 10), Switzerland (n = 6), Israel (n = 2), USA (n = 2), China (n = 5), Canada (n = 1), and Indonesia (n = 2).The sequences of the samples from Spain, Israel, Switzerland, and PU131 from the US, and three samples from Anhui, China have been reported (18,(22)(23)(24)(25).For all sequences, their multilocus sequence typing (MLST) genotype was identified using PubMLST (26).Two isolates (NZ_CP051346.1 and NZ_CP022500.1)were identified as belonging to ST152, while the ST of one Indone sian isolate (NZ_CP101647.1)could not be determined, and the other sequences were identified as ST198.
Core genome single nucleotide polymorphisms (SNPs) for all sequences were identified by comparing their sequences to the S. Kentucky SSSE-01 sequence, using PhaME (27).Pairwise core genome SNP distances between SSSE-01 and SSSE-03 was 16 nucleotides.The pairwise SNP distances of the 28 sequences of S. Kentucky ST198 samples ranged from 1 to 112 nucleotides with the details shown in Table S5.The identified SNPs were used to construct a maximum likelihood phylogenetic tree.The tree shown in Fig. 1 contains only the isolates belonging to ST198 and the Indonesian isolate with unknown ST.The latter isolate was closely related to the other Indonesian ST198 isolates, with a pairwise SNP distance of seven nucleotides, and was, therefore, consid ered as belonging to the ST198 clade as well.Three isolates, PU131 from the US and two isolates (NZ_CP091997.1 and NZ_CP091998.1)from human urine from Switzerland, were early branching compared to the rest.There were three notable subclades.The first subclade included all samples from Spain and an isolate from Switzerland.The second subclade included four isolates from Anhui and Fujian in China, which were closely related to SSSE-01 and SSSE-03 from Thailand.The last subclade consisted of two isolates from Israel and two from Switzerland.

Characterization of the gene content of the SGI1-K regions
To investigate the genetic structures of SGI1-K, we initially performed BLAST analysis of the complete genome sequences of all 30 S. Kentucky isolates included in this study using the 52,779-bp-long SGI1-K region of the SRC73 strain (Genbank Accession Number: AY463797.8)(29) as a query with the results shown in Fig. 2. The query sequence included the SGI1-K sequence as well as the flanking mnmE∆ (also known as trmE, a part of mnmE) on the 5′-end and yidY-yidZ-yieE∆ on the 3′-end, with the segment including mnmE∆ to yidY being 50,841 bp long in SRC73.
The S. Kentucky ST152 isolates appeared to lack the SGI1-K sequence as they lacked the S001-S027 backbone and tnp6100.The mnmE∆ and yidY were still present, with the segment including them being only 4,025 bp long as shown in Fig. 2c.
All the other isolates contained SGI1-K, based on the presence of mnmE∆ and some initial SGI1-K backbone genes.The yidY genes of some SGI1-K were partially deleted.The lengths of the DNA segments containing mnmE∆ to yidY or yidY∆ of 22 isolates ranged from 27,781 bp in SSSE-01 or SSSE-03 to 44,985 bp in NZ_CP101647.1 from Indonesia.The distance between mnmE∆ to yidY of two isolates (NZ_CP091998.1 from Switzerland and NZ_CP089788.1 from Spain) was unusually long.The yidY gene was absent from the other five isolates as shown in Table S4.
The blasted segments of each isolate were then aligned to the SGI1-K region of the SRC73 strain.Among the 20 isolates with mostly intact SGI1-K backbone from mnmE∆ to at least S025, six isolates had the complete backbone up to S027 (resG), similar to SRC73.These included the PU131 strain from US, an isolate from Switzerland (NZ_CP092009.1),a pair of isolates from Spain (NZ_CP089791.1 and NZ_CP089797.1,with a pairwise SNP distance of eight nucleotides), and a pair from Indonesia (NZ_CP101647.1 and NZ_CP101648.1,with a pairwise SNP distance of seven nucleotides).
In all, 12 isolates had partial deletions of S027 (resG).This group included eight Spanish isolates (with an average pairwise distance of 28 nucleotides) and four isolates (with an average pairwise distance of 10 nucleotides) from Israel and Switzerland.It should be noted that even though NZ_CP089788.1 had a very long distance between mnmE to yidY, it still had the same SGI1-K genetic content as the other Spanish isolates, and had only three SNP different from NZ_CP089794.1, as shown in Table S5.In addition, a pair of isolates from China and Switzerland, with a pairwise SNP distance of 31 nucleotides, had similarly more deletion of S026, as shown in Fig. 2b.
Eight isolates, including SSSE-01 and SSSE-03, had extensive deletions in the SGI-1K backbone as shown in Fig. 2c.Four isolates from Anhui and Fujian, with an average pairwise distance of 14 nucleotides, had an identical backbone deletion from S014 to S027.SSSE-01 and SSSE-03 exhibited an identical backbone deletion from the mid dle of S011 (traG) to S027 (resG).The last pair in this group were from Switzerland (NZ_CP091997.1 and NZ_CP091998.1)and had a pairwise SNP distance of only nine nucleotides.However, they had unusually long distances between mnmE∆ to yidY and different profiles of deleted regions in SGI1-K.It is worth noting that a pair of isolates with similar gene contents in the SGI1-K region generally had a small average pairwise SNP distance.

The genetic structures of the SGI1-K region in SSSE-01 and SSSE-03
The genetic structures of the SGI1-K region in SSSE-01 and SSSE-03 were identical, starting from the 5′-end with the DR-L sequence (5′-TTCTGTATTGGGAAGTAA-3′) in mnmE at the positions 4,866,037-54 and 4,865,261-78 of SSSE-01 and SSSE-03, respectively, in the complementary sequence.Compared to the original SGI1-K, the ones in SSSE-01 and SSSE-03 were substantially shorter, mainly due to the deletion of approximately 20 kb of nucleotides (from the position in AY463797.8at 8,067-27,330 bp).This deletion included a part of S011 (traG), as well as S012-S027 and IntI1, accounting for almost 30% of the mnmE∆ to IntI1 segment.The traG∆ was immediately followed by a copy of IS26, followed by a part of the segment containing ARGs of SRC73 including aac (3)-Id, aadA7, qacE∆, sul1, IS6100, mer operon, tetR(A), tet(A), pecM, and tnpA∆ but in inverted orientation, and finally followed by another copy of IS26.The segment was likely to be inverted by the recombination of IS26.Immediately downstream, the segment containing yieE-yidZ-yidY-S044∆ was also inverted compared to the original SGI1-K, as shown in Fig. 3.This resulted in the inversion of the DR-R in yidY to 5′-TTCTGTATTGGTAA GTAA-3′ at positions 4,839,421-38 and 4,838,645-62 of SSSE-01 and SSSE-03, respectively.The yieE-yidZ-yidY-S044∆ segment, which is generally considered the end of SGI1-K, was immediately followed downstream by a multi-resistance region (MRR), comprising three IS26-flanking segments: IntI1-aadA-lnu(F), IS4-aac (3)-IId-AAA, and recombinase-bla TEM-1b , respectively.Only the third segment was present partially in SRC73.The last IS26 was followed by yieF, which is normally found after yidY-yidZ-yieE in ST198.The SGI1-K region of SSSE-01 and SSSE-03 had a total of six copies of IS26, instead of two in the original SGI1-K (Fig. 3).
The four isolates from Anhui and Fujian in China were the most closely related to SSSE-01 and SSSE-03.The SGI1-K region of the isolate NZ_CP102719.1 (AH19MCS1) (25) is exemplified in Fig. 3.It had less deletion of the backbone.They contained the same ARG-containing segment as SSSE-01 and SSSE-03, which retained the same orientation as SRC73, though.The ARG-containing segment was followed by a segment containing yieE-yidZ-yidY∆, with a part of yidY and S044∆ being deleted.This segment was immediately followed by yieF.Interestingly, the Chinese isolates also contained an MRR, almost the same as SSSE-01 and SSSE-03, residing 900 kbp downstream from the yieF, as recently reported.A complete sequence of SGI1-K of another isolate of S. Kentucky CIP R ST198 isolates in Thailand has been recently reported.The isolate contained the same truncated backbone and the inverted yieE-yidZ-yidY-S044∆ segment, similar to SSSE-01 and SSSE-03.The segment containing the mer operon was inverted compared to SSSE-01 and SSSE-03, and it lacked Intl-aadA-lnu(F) and IS4-aac (3)-IId-AAA segments (30).

The genetic structures of SGI1-K of other S. Kentucky ST198
Comparison of the genetic structures of the SGI1-K in other S. Kentucky ST198 revealed considerable variations, the representatives of which are shown in Fig. 4 and 5. Interest ingly, the order of the core genes from S001-S027 was mostly consistent among all isolates, although there were variable deletions involving S011 to S027.All isolates still retained at least the segment between S001 and traG (S011).However, the regions between S027 to S044∆, which typically contained ARGs and the mer operon, exhibi ted high variability, including both insertions/deletions (in/del) and inversions.These variable segments were often flanked, and therefore, likely mediated by recombination of IS26.In addition, the ARGs themselves were not necessarily located within the conventional boundary of SGI1-K, upstream to S044∆-yidY.Instead, they were found in other genomic locations.
Manual comparison of SGI1-K revealed that three isolates, PU131, and two Spanish isolates (NZ_CP089791.1 and NZ_CP089797.1),did not contain any ARGs or mer in the SGI1-K region, although the backbones were mostly intact.In PU131, the segment between S027 and S044 was completely absent, likely due to recombination of IS26.However, it contained 10 ARGs that had been relocated into multiple sites elsewhere in the genome, as previously reported (22).The two Spanish isolates had only a single ARG, aac(6′)-Iaa, in their chromosome but not in SGI1-K.This gene, which acetylates tobramycin, kanamycin, and amikacin effectively but not gentamicin (31), was present in all studied ST198 strains and was originally identified in the chromosome of S. Typhimu rium LT2 strain.Both isolates also lacked the IS26-S044-yidY segment.Some SGI1-K elements contained only a single ARG.The four isolates belonging to the Israel/Switzerland subclade had nearly intact backbones but lacked the mer operon and contained only a single ARG, bla TEM-1 , flanked by two IS26 and located between S027 and S044∆-yidY in SGI1-K.The other ARGs typically found in SGI1-K [e.g., aadA7, aac (3)-Id, sul1, and tet(A)] were previously shown to be relocated into rbsK (26).Two other isolates from Switzerland (NZ_CP091999.1 and NZ_CP092009.1)also had only one ARG in the SGI1-K region.NZ_CP091999.1 harbored aph(3′'-Ia and lacked the IS26-S044∆-yidY segment.NZ_CP092009.1 had a highly rearranged SGI-1K between S014 and the IS26-S044∆-yidY segment, and it only harbored only bla TEM-1 in SGI1-K. All nine Asian isolates had complex ARG-containing regions in SGI1-K, which were partially homologous to the one in SRC73.A Chinese isolate, NZ_CP077680.1,had the longest ARG segment, including Intl1, aac (3)-Id, aadA7, qacE∆, sul1, IS6100, mer operon, tetR(A), tet(A), pecM and tnpA∆, aph (6)-Id, aph(3″)-Ib, IS1133, IS26, and aph(3′)-Ia in an inverted orientation as shown in Fig. 4b.The homologous segments of both Indonesian isolates (NZ_CP101647.1 and NZ_CP101648.1)were shorter, with truncation after tnpA∆ by IS26, also in the inverted orientation.The homologous segment of SSSE-01 and SSSE-03, also in the inverted orientation, was further truncated by another IS26 inserted between IntI1 and aac (3)-Id.The Anhui and Fujian isolates had a segment with a similar length to the Thai isolates but with an orientation similar to SRC73 as shown in Fig. 4.These similarities suggested that the evolution of the ARGs region of SGI-1K may have been mediated by IS26 insertions and homologous recombination.
The Spanish isolates exhibited several patterns of the SGI1-K element.Two of them had no ARGs in SGI-1K, as described above.Six isolates had the same intact backbone up to S026 followed by IS26, a segment containing IntI1-to-merEDA, another copy of IS26, and finally S044∆-yidY.These isolates retained some ARGs in the original SGI1-K, namely aac (3)-Id, aadA7, qacE∆, sul1.In another isolate, NZ_CP089795.1,an additional IS26 was inserted in S025 with the inversion of the segment between this IS26 and the one upstream of the mer genes.This resulted in the presence of S026 and S025 sequences downstream of the ARGs as shown in Fig. 4a.In the last Spanish isolate, NZ_CP089788.1,SGI1-K was scattered in the chromosome.
The yidY gene was absent in five isolates, as shown in Table S4.It was partially deleted from all four isolates from Anhui and Fujian, and this deletion was accompanied by the loss of the DR-R segment.The absence of the DNA segment was recently reported in the majority of S. Kentucky ST198 isolates from China (25).Among the other isolates, the intact yidY genes were identified mostly in the typical IS26-S044∆-yidY segment oriented similarly to SRC73, except the two Thai isolates, which were in reversed orientation.A variant of the IS26-S044∆-yidY segment was identified in seven isolates from Spain, represented by NZ_CP089798.1 and NZ_CP089795.1 in Fig. 4a.
The bla TEM-1 gene was identified in 14 isolates.The gene was in a Tn2-like structure within the typical boundary of SGI1-K in six isolates, including the Israel/Switzerland clade, NZ_CP089795.1 from Spain, and NZ_CP092009.1 from Switzerland.However, the gene was outside the typical SGI1-K boundary in the four isolates from Anhui and Fujian, two isolates from Thailand, PU131 and NZ_CP091998.1.
bla CTX-M-14b was identified in two isolates from China and Switzerland.It was previously shown to be inserted in a 2,854 bp fragment inserted in the chromoso mal region containing Type 6 Secretion System genes, downstream of the hcp1 gene (position 3,701,330 of reference genome NZ_CP126327.1).The fragment is flanked on one side by an insertion sequence ISEcp1 of the IS1380 family (32,33).The 2,854 bp fragment was absent on the chromosome of the non-bla CTX-M-14b S. Kentucky isolates.S. Kentucky harboring bla CTX-M-14b has been reported several times from China (34).
The bla CTX-M-55 was present in the chromosome but not in the SGI1-K region in all four isolates from Anhui and Fujian.A recent study in Anhui identified bla CTX-M-55 in all 10 S. Kentucky isolates (25).

Antibiotic resistance gene profiles of SSSE-01 and SSSE-03
S. Kentucky ST198 strains have become highly resistant to ciprofloxacin by accumulating various combinations of mutations in the QRDRs of gyrA, encoding a subunit of DNA gyrase and parC, encoding a subunit of DNA topoisomerase IV (15).All 28 isolates had Ser83Phe substitution in GyrA.All three mutations affecting Asp87 were present among the 28 isolates.All the five early branching isolates in the phylogenetic tree had Asp87Gly while Asp87Tyr was unique to the four-isolate clade from Israel and Switzerland.The other 19 isolates, including SSSE-01 and SSSE-03, had Asp87Asn.All isolates had the quinolone resistance associated with mutations of both Thr57Ser and Ser80Ile in ParC.
ResFinder identified various ARGs, ranging from 1 to 18 per isolate as summarized in Table S6 and Fig. 1.The aac(6′)-Iaa gene was identified in all isolates, similar to a previous report (35), while aadA7, aac (3)-Id, and sul1 were present in the majority.The tet(A) and bla TEM-1B genes were identified in about half of the isolates.SSSE-01 and SSSE-03 additionally harbor aadA17 and lnu(F).The isolates from Anhui and Fujian harbored the most numbers of ARGs, including almost all of the mentioned ARGs.They, as a group, uniquely harbored bla CTX-M-55 , rmtB, and floR.Some of them also uniquely harbored fosA3, qnrS1, aac (3)-Id, aac (3)-IV, and aph (4)-Ia.They also harbored aadA17, aac (3)-Iid, and lnu(F) similar to SSSE-01 and SSSE-03, and aph(3′)-Ia similar to other isolates.A single colistin resistance gene, mcr-1.1 (chromosomal position at nt 4,160,060-4,161,685), was identified in the sample NZ_CP077680.1,isolated in 2020 from chicken at an unspecified location in China (16).

Plasmids of S. Kentucky strains
Both SSSE-01 and SSSE-03 contained three plasmids, with the same sizes of 4,018 bp, 2,257 bp, and 2,097 bp, and nearly identical sequences in both isolates as summarized in Table S3.These plasmids did not contain any ARGs.PlasmidFinder identified the plasmids pSSSE01a and pSSSE03a as having a similar replicon (91.6%) to Col(pHAD28) (accession number KU674895), as shown in Fig. S1.The replicons of plasmids pSSSE01b and pSSSE03b were similar (88.1%) to Col(MP18) (accession number NC013652).The replicons of plasmids pSSSE01c and pSSSE03c were highly similar (98.96%) to ColpVC (accession number JX133088), which was associated with various mobile genetic elements and is known to carry multiple antibiotic resistance genes.

DISCUSSION
S. Kentucky is commonly reported from poultry worldwide (32,(36)(37)(38)(39).The serovar is polyphyletic, consisting of strains belonging to several genotypes, including ST152, ST198, ST314, etc. (33).Human diseases caused by S. Kentucky are relatively uncommon, and to our knowledge, there have been no reports of human diseases caused by S. Kentucky in Thailand yet.However, salmonellosis caused by ST198 has been reported in several European countries, occasionally with travel histories to Africa and Asia (8,40).Recent studies have also indicated the increased public health importance of S. Kentucky (41).For example, S. Kentucky was identified in 48% of NTS isolated from hospitalized diarrheal cases around Delhi, India (42).ST198, in particular, poses significant public health implications, while ST152 does not raise the same level of concern.Overall, these findings underscore the emerging importance of S. Kentucky as a potential health threat.
A clade of S. Kentucky ST198, which exhibits resistance to ciprofloxacin (CIP R ST198), is of particular concern, as the drug is commonly used to treat salmonellosis.The clade also usually harbors extended-spectrum beta-lactamase (ESBL) genes, making it resistant to common alternative drugs, third-generation cephalosporins (17,43).In this study, we reported the complete genome sequences of two isolates of CIP R S. Kentucky ST198, which carried an ESBL gene, bla TEM-1b , and also lnu(F), which made them resistant to cephalosporins and lincosamide.Therefore, these strains would be difficult to treat if necessary.
While ARGs in Enterobacteriaceae, including Salmonella, are often located in plasmids, which facilitate their transmission across several species and genotypes, resistance to ciprofloxacin the CIP R ST198 clade is due to the mutations in chromo somal target genes, gyrA and parC (11,12).Moreover, several ESBL genes, including bla TEM-1b , bla CTX-M-14 , and bla CTX-M-55 , were identified in the chromosome, among which only bla TEM-1b was identified in SGI1-K (43)(44)(45)(46).Therefore, focusing surveillance on the transmission of the CIP R ST198 clade would be useful in controlling the spread of the MDR strains.
In this study, we also identified a previously unreported isolate from China, NZ_CP077680.1,harboring intact mcr-1.1, which confers resistance to colistin.The presence of mcr may lead to the more challenging treatment of the strain in the future (16).
SGI1 is known as a hotspot for the integration of ARGs in many species of Enterobac teriaceae (13)(14)(15)(16).SGI1-K was initially discovered in S. Typhimurium and might have transferred to S. Kentucky ST198 shortly before 1990 (17).Previous studies using NGS short-read data have shown that the region in SGI1-K between resG (S027) and S044∆ is the primary site in the Salmonella genome where most ARGs are located (9).However, by analyzing the complete genome sequences, mostly from the hybrid assembly, we demonstrated significant variation in SGI1-K, which was previously sporadically reported (17).Some isolates from Europe and the USA completely lost the ARG-containing segment from SGI1-K or retained only a single ARG.By contrast, isolates from Asia mostly retained some ARGs originally reported in SGI1-K of the SRC73 strain.In addition, both groups occasionally carried more ARGs in other chromosomal regions.The role of SGI1-K in the formation of the multidrug resistance gene regions outside itself is not clear, as there is evidence suggesting that the integration of some ESBL genes to the chromosome may not require the SGI1-K region (47).
The substantial genetic structural variations of SGI1-K in CIP R ST198 are not limited to the ARG-containing region.While the order of the backbone genes of SGI1-K is mostly conserved, with the segment from mnmE to S010 always present, the rest of the backbone can be deleted or inverted, primarily mediated by transposition and homologous recombination of IS26.The 3′-end of SGI1-K exhibited high variability.The IS26-S044∆-yidY segment is not always present; it can be inverted or completely absent.Even when the IS26-S044∆-yidY segment was present, ARGs may not necessarily be located in the segment between it and mnmE.Further study is needed to understand how the evolution of the SGI1-K backbone affects its function as a hotspot of integration of ARGs.
In conclusion, even with a limited number of samples included in this study, their complete genomes revealed vast genetic structural variations of SGI1-K among CIP R S. Kentucky ST198.Resistance to third-generation cephalosporins was mediated by a variety of bla TEM1 or bla CTX genes located in many sites in the chromosome including SGI1-K.The exact location of them may provide a method for recognizing various circulating clades of the bacteria.The identification of lnu(F) and mcr-1.1 signifies the existence of lincosamides and colistin-resistant clones which require close monitoring.

Collection of S. Kentucky strains
Two S. Kentucky isolates, SSSE-01 and SSSE-03, were isolated from a chicken slaughterhouse in Mukdahan, a province in Northeast Thailand bordering the Lao People's Democratic Republic as a part of a surveillance project in 2016.The pro tocol of the sample collection was approved by the Faculty of Veterinary Science-Animal Care and Use Committee at Mahidol University, under protocol number MUVS-2016-09-38.Both isolates were identified as serovar Kentucky using a serolog ical test at the Ministry of Science in Thailand.WGS of both were done with Illu mina NextSeq 500 using paired-end reads.The sequencing data has been deposited under the PRJNA841788 project, with accession numbers AAEXAW010000000 and AAEXAP010000000.The serovar designation and MLST were confirmed in silico using the Enterobase database (https://enterobase.warwick.ac.uk/species/index/senterica).In addition, complete genome sequences of 28 S. Kentucky isolates collected were retrieved from NCBI (www.ncbi.com)as shown in Table S4.These isolates were confirmed to be S. enterica by the analysis of WGS and the sequence types were identified using the PubMLST database (26).

Genomic DNA extraction and Illumina sequencing
S. Kentucky isolates were cultured on trypticase soy agar (TSA) at 37°C for 24 h.Genomic DNA (gDNA) was extracted using the QIAmp DNA mini kit (Qiagen, Hilden, Germany).The quality and quantity of DNA were assessed using Nanodrop DenoVix for absorbance value and the Qubit 3.0 fluorometer for double-strand DNA quantity.Purified gDNAs of SSSE-01 and SSSE-03 were sequenced at the Advanced Genomic Technologies Cluster New York State Department of Health/Wadsworth Center, NY by Illumina NextSeq500 instrument, using the Nextera XT library preparation protocol and the NextSeq 500/550 Mid Output Kit v2.

Oxford nanopore library preparation and sequencing
Genomic DNA samples that met the following criteria were used for library construction: (i) A260/280 between 1.8 and 1.9, and (ii) A260/230 between 2.0 and 2.2.Approximately 400 ng of total input DNA was used for each flow cell (FC).Electrophoresis was per formed to separate and determine the sizes of linear DNA fragments.
The gDNA samples underwent ligation (SQK-LSK110) following the library prepara tion protocol provided by Oxford Nanopore Technologies (ONT).The libraries were then sequenced using qualified FLO-MIN106 flow cells (R9.4.1, with an active pore number >800) on the MinION (ONT, Oxford, UK) for approximately 48 h.Real-time basecalling was performed using Guppy, with a modified basecalling model for 6 mA dam/5 mC dcm and CpG, which was integrated into the MinKNOW software v3.5.40 installed on MinION.

Multiple sequence alignment and phylogenetic tree analysis
To provide context for our newly assembled sequences, we searched in the NCBI database to identify previously reported sequences of S. Kentucky.We retrieved a total of 30 genome sequences that were designated as Annotated by NCBI RefSeq by NCBI and were accessible in January 2023.Then, PhaME (27) was used for sequence alignment and core genome SNP identification using the SSSE-01 as a reference sequence, because of its high read depths and geographic proximity to many unreported samples, which are mostly in Asia.The resulting core genome SNP alignment was then used to compute pairwise SNP difference using snp-dists (57), and maximum likelihood phylogenetic reconstruction using IQ-TREE 2 (28).A comparison of SGI1-K was done by Easyfig (58).

FIG 1 A
FIG 1 A maximum likelihood phylogeny of 28 isolates of Salmonella Kentucky ST198.The tree was inferred from a core genome SNP alignment (44,140 SNP sites) using IQ-TREE2 (28).The tree was rooted using the S. Kentucky ST152 (Accession numbers NZ_CP022500.1,NZ_CP051346.1)as outgroups (not shown here).The clade support values were computed using the ultrafast bootstrap (UFBoot) method (1,000 replicates), implemented in IQ-TREE2.For each internal node, black and orange points denote UFBoot values of 100 and >80, respectively.A scale bar indicates the branch length in the unit of substitutions per site.Panels from left to right show the country of isolation and the presence of some selected drug resistance genes.The rightmost panel shows patterns of amino acid changes in the quinolone resistance determining regions (QRDRs) within GyrA (the first two mutations) and ParC (the last two), including pattern A (black) representing the S83F, D87N, T57S, and S80I mutation, pattern B (green) representing the S83F, D87Y, T57S, and S80I mutation, and pattern C (orange) representing the S83F, D87G, T57S, and S80I.

FIG 2
FIG 2 Genome content comparison of SGI1-K in 30 isolates to the 52,779-bp-long S. Kentucky SRC73 (AY463797.8)as reference.The presence of each DNA segment of each isolate was inferred from the result of BLAST using the reference sequence as the query.The results are categorized into three groups.Group (a) consists of isolates with long SGI1-K backbones.Group (b) includes the Spanish isolates.Group (c) includes isolates with short SGI1-K backbones, as well as two ST152 isolates, NZ_CP022500.1 and NZ_CP051346.1.The uppermost row of each figure indicates the positions of genes in the reference sequence AY463797.8.Genes are color-coded as follows: purple for CDS, yellow for IntI1, pink for drug resistance genes, and blue for insertion sequences.

FIG 3
FIG 3 Comparison of SGI1-K regions of three isolates: SRC73 (AY463797.8) in the upper panel, SSSE-01 in the middle panel, and a representative isolate from Anhui, China (NZ_CP102719.1) in the lower panel.

FIG 4
FIG 4 Comparison of the genetic structures of SGI1-K of several representative samples of S. Kentucky ST198.(a) The SGI1-K of NZ_CP077680.1,which was the longest among this group, is shown with labeled genes.(b) The SGI1-K of several representative isolates from mnmE to yidY.In cases where yidY is missing, the regions were shown to the last IS26.A group of isolates with very similar genetic structures is shown only once.The figure is ordered to correspond to the order in the phylogenetic tree in Fig. 2. The last isolate, NZ_CP026327.1, was PU131.