Hybrid genome assembly of Shigella sonnei reveals the novel finding of chromosomal integration of an IncFII plasmid carrying a mphA gene

Azithromycin is increasingly being used for the treatment of shigellosis despite a lack of interpretative guidelines and with limited clinical evidence. The present study determined azithromycin susceptibility and correlated this with macrolide-resistance genes in Shigella spp. isolated from stool specimens in Vellore, India. The susceptibility of 332 Shigella isolates to azithromycin was determined using the disc diffusion method. Of these, 31 isolates were found to be azithromycin resistant. The azithromycin minimum inhibitory concentration (MIC) was determined using the broth microdilution method. In addition, isolates were screened for mphA and ermB genes using conventional PCR. Furthermore, an isolate that was positive for resistance genes was subjected to complete genome analysis, and was analysed for mobile genetic elements. The azithromycin MIC for the 31 resistant Shigella isolates ranged between 2 and 16 mg l−1. PCR results showed that a single isolate of Shigella sonnei carried a mphA gene. Complete genome analysis revealed integration of an IncFII plasmid into the chromosome of S. sonnei , which was also found to carry the following resistance genes: sul1, bla DHA1, qnrB4, mphA, tetR. Mutations in the quinolone-resistance-determining region (QRDR) were also observed. Additionally, prophages, insertion sequences and integrons were identified. The novel finding of IncFII plasmid integration into the chromosome of S. sonnei highlights the potential risk of Shigella spp. becoming resistance to azithromycin in the future. These suggests that it is imperative to monitor Shigella susceptibility and to study the resistance mechanism of Shigella to azithromycin considering the limited treatment choices for shigellosis.


INTRODUCTION
Shigella species are a common cause of dysentery among children, mainly in low-and middle-income countries. Treatment of Shigella infection with antibiotics is usually recommended to reduce the complications and risk of transmission [1]. However, antimicrobial resistance (AMR) in Shigella spp. is not uncommon. Recently, increasing resistance to ciprofloxacin made third-generation cephalosporins and macrolides the drugs of choice. However, the current emergence of extended-spectrum β-lactamase-producing strains means that azithromycin is now the drug of choice, particularly for children [1][2][3].
Macrolides are being widely used for enteric pathogens, as they can achieve high intracellular concentrations and produce adequate levels in the colon [4,5]. Notably, worldwide, azithromycin is being used successfully for shigellosis, despite no clinical breakpoints being available either from the Clinical and Laboratory Standards Institute (CLSI) or the European Committee on Antimicrobial Susceptibility Testing (EUCAST). CLSI guidelines only suggest epidemiological cut-off values for Shigella flexneri and Shigella sonnei. Also, clinical evidence for the efficacy of azithromycin is limited [2].
Macrolide resistance involves a variety of mechanisms, most of which have already been reported in Enterobacteriaceae. These include: (i) target site modification by methylases encoded by erm genes, mainly ermA, ermB and ermC; (ii) OPEN ACCESS inactivation by modifying enzymes such as esterases or phosphotransferases encoded by ereA/ereB or mphA/mphB/mphD genes, respectively; (iii) acquisition of efflux pumps, mefA and msrA genes have been identified in Gram-negative organisms, though they are found essentially in Gram-positive organisms [5]. Azithromycin resistance in enteric pathogens is commonly due to two mechanisms: drug inactivation by the mphA gene, and target-site modification by the ermB gene [6]. These genes are usually carried by different incompatibility (Inc) plasmids. The most commonly reported plasmids in Shigella are the IncFII type [7]. The objective of the present study was to determine azithromycin susceptibility and to correlate this with the presence of macrolide-resistance genes in Shigella spp. isolated from stool specimens in India.

Bacterial isolates and susceptibility testing
A total of 332 Shigella spp. were obtained from stool specimens collected between the years 2017 and 2019 at Christian Medical College, Vellore, India. Genus and species identification were performed using conventional methods [8]. Serogrouping of Shigella spp. was done by slide agglutination using commercial antisera (Denka Seiken). Antimicrobialsusceptibility testing of isolates against ampicillin, trimethoprim/sulfamethoxazole, ciprofloxacin, cefotaxime, cefixime and azithromycin was performed by Kirby-Bauer disc diffusion method. The azithromycin minimum inhibitory concentration (MIC) was determined by broth microdilution for the isolates that were resistant using the disc diffusion method. The results were interpreted using CLSI 2017 guidelines [9].

Analysis of mphA and ermB genes by PCR
Total DNA was extracted using a QIAamp DNA mini kit (Qiagen). The isolates were analysed for the presence of mphA and ermB genes by PCR, as described elsewhere [5]. Known positive and negative controls were used in every run.

Whole-genome sequencing
The Shigella isolate positive for a macrolide-resistant gene was further subjected to whole-genome sequencing. Genomic DNA was extracted from an overnight culture using a QIAamp DNA mini kit (Qiagen), as per the manufacturer's instructions. DNA quality was assessed using Nanodrop spectrophotometry (Thermofisher) and quantity was assessed using the Qubit 3.0 system (Thermofisher). Whole-genome sequencing was performed using Ion Torrent with 400 bp chemistry (PGM; Life Technologies) and MinION nanopore sequencing (Oxford Nanopore Technologies), as per the manufacturers' instructions. Hybrid genome assembly was performed to get the complete genome with Ion Torrent (short reads) and MinION (long reads) using Unicycler v0.4.6, as described previously [10].
Annotation was performed using the National Center for Biotechnology Information Prokaryotic Genome Annotation Pipeline (PGAP; http://www. ncbi. nlm. nih. gov/ genomes/ static/ Pipeline. html). The downstream analysis was done using the Center for Genomic Epidemiology (CGE) server (http://www. cbs. dtu. dk/ services). The resistance gene profile and plasmids were identified using ResFinder 3.0 and Plas-midFinder 1.3, respectively [11,12]. Circular genomes of the isolates were mapped against the reference genome using CGview software [13]. Plasmid comparison was done using Gview software [14]. Additionally, integrons, prophages and insertion sequences were predicted using web tools such as phast and ISsaga [15,16].

Antimicrobial-susceptibility testing and resistance gene PCR
Among the total of 332 Shigella isolates, 31 isolates (6 S. flexneri and 25 S. sonnei) were resistant to azithromycin using the disc diffusion method as per the epidemiological cut-off value of S. flexneri. The azithromycin MIC for these Shigella isolates ranged between 2 and 16 mg l −1 . The most common multi-drug resistant (MDR) phenotypes among S. flexneri and S. sonnei isolates in this study were AMP-SXT-CIP and SXT-CIP-AZM, respectively (Table 1). Furthermore, among the 31 resistant isolates, the mphA gene was only identified in one S. sonnei isolate (FC1428) on PCR analysis. None of the other isolates harboured the gene.

Whole-genome sequencing
The S. sonnei isolate (FC1428) that carried the mphA gene, and which had an azithromycin MIC value of 16 mg l −1 , was sequenced to characterize the resistance mechanism. Sequence analysis revealed that the IncFII plasmid was integrated into the chromosome, which showed 100 % similarity to an Escherichia coli plasmid, pC15-1a (accession number AY458016.1) on blast analysis. The plasmid also carried the following AMR genes, sul1, bla DHA1 , qnrB4, mphA and tetR, as shown in Fig. 1. In addition, the dfrA17 gene was identified in the chromosome. The isolate also harboured the following mutations in the quinolone-resistance-determining region (QRDR): gyrA -S83L, D87G; and parC -S80I. Furthermore, class 1 and 2 integrons were identified within the plasmid. The isolate carried ten intact, nine incomplete and six questionable regions of the prophage. Also, a total of 465 IS elements was predicted to be present in the S. sonnei genome. The most commonly identified type was from the IS1 family, accounting for approximately 37 % of the IS elements, followed by the IS3_ssgr_IS3, IS4, IS3_ssgr_IS2 and IS21.
The plasmid region of the S. sonnei (FC1428) isolate was compared with the previously published azithromycinresistance plasmids of S. sonnei, pEG430-2 (accession number LT174531.1), and E. coli, FC853 plasmid 3 (accession number CP040922) carrying the ermC and/or ermB genes. The analysis showed that the Shigella plasmids carried additional tra genes when compared to the E. coli plasmid, as shown in Fig. 2. The compared plasmids have the repA gene for the IncFII plasmid, insertion sequences such as IS1 and IS26, hypothetical proteins, type II toxin-anti-toxin, pilus and finO genes in common. The genome sequences of the  S. sonnei isolate has been made available in GenBank under the accession number CP041322.

DISCUSSION
Antibiotics are believed to reduce the duration of illness caused by Shigella spp. Azithromycin is often used to treat MDR, including fluroquinolone-resistant, pathogens [1]. Although azithromycin is being used worldwide for shigellosis, the lack of interpretative guidelines and only limited clinical data leads to uncertainty about the clinical outcome. Recently, the emergence of decreased susceptibility to azithromycin (DSA) has been reported in Shigella spp. [17,18]. There are also reports of azithromycin-treatment failure, particularly in men who have sex with men and human immunodeficiency virus patients with shigellosis [19,20]. This indicates that the determination of clinical breakpoints and their correlation with the clinical outcome in azithromycin-treated patients is urgently required.
In this study, the correlation of the MIC with a resistance gene showed that the MIC ranged between 2 and 16 mg l −1 among the study isolates, and that the mphA gene was only found in the S. sonnei isolate with a MIC of 16 mg l −1 . Similarly, a study by Darton et al., showed that S. flexneri isolates with a MIC of 16 mg l −1 were positive for either mphA or ermB genes [2]. In contrast, another study by Salah et al. showed that the isolates with a MIC of 16 mg l −1 were found to be negative for all tested macrolide-resistance genes. However, both S. flexneri and S. sonnei isolates with an azithromycin MIC of ≥32 mg l −1 have been found to be positive for the mphA gene [3]. There are also reports of E. coli isolates with a MIC of ≥256 mg l −1 carrying the mphA gene, followed by ermB and mphB in isolates with a MIC of >1024 and 128 mg l −1 , respectively [5].
Generally, azithromycin-resistance genes such as mphA and ermB were reported to be carried in plasmids [2]. Remarkably, in this study, the plasmid carrying the mphA gene was found to be integrated into the chromosome of S. sonnei. However, the phenomenon can be reversible if the integration is through homologous recombination, and this might be possible due to the presence of various insertion sequences and other mobile elements in Shigella spp. Apart from this unique finding, genome analysis revealed the presence of multiple resistance genes that were expected to be present. The mutations we observed in the QRDR are commonly reported to be associated with fluoroquinolone resistance in Shigella spp., as reported in previous studies [7]. The genome also contained various mobile genetic elements that are reported to play a significant role in AMR dissemination in Shigella spp.
Earlier, a study by Nguyen et al. showed that E. coli acts as a reservoir for macrolide-resistance genes from which resistant Shigella spp. might have emerged through horizontal gene transfer. The phenomenon has been previously demonstrated with E. coli donating mphA to S. sonnei [5].
In this study, we looked for the occurrence of a similar event among the studied isolates. We compared the IncFII plasmid profile of Shigella and E. coli carrying the macrolideresistane gene to identify the backbone similarity. Though the analysis revealed several genes in common, the Shigella plasmid harboured an additional tra operon compared to E. coli, which might have been acquired due to evolution over time. The tra genes had always been the only genomic factors that make the plasmid conjugative. Interestingly, Benz et al., in their study, showed that the plasmid transfer is mainly based on the functional tra (transfer) genes rather than the plasmid types [20]. Furthermore, earlier studies have shown that plasmids carrying required functional tra genes can spread even without antibiotic selection pressure [20]. These highlight the potential risk of plasmids with resistance genes carrying functional tra genes being transferred by natural conjugation.
The widespread emergence of MDR Shigella with changing AMR patterns has also being reported. Generally, in Shigella spp., acquired resistance is more common. β-Lactam resistance is mainly due to the presence of OXA-type β-lactamases, followed by TEM and CTX-M. Trimethoprim/sulfamethoxazole resistance is encoded by dhfr1A and sul genes. Quinolone resistance involves the accumulation of mutations in QRDR and plasmidmediated quinolone resistance (PMQR) genes. Furthermore, resistance to tetracycline, chloramphenicol and streptomycin has been shown to be due to the presence of tetA/B, catA1 and of either strA/B or aadA1 genes or both [7]. These show the species ability in acquiring AMR determinants.

Conclusion
To conclude, the novel finding of an integrated plasmid in this study indicates the potential risk of Shigella isolates becoming resistant to azithromycin in the future. Characterization of the functional tra genes is important to track the potential dissemination of macrolide resistance. The study also highlights the significance of the hybrid assembly approach in the complete genome analysis. These findings suggest that it is imperative to monitor Shigella susceptibility and to study the resistance mechanism of Shigella against azithromycin, considering the limited treatment choices for shigellosis.

Ethical statement
This study was approved by the Institutional Ethics Committee of Christian Medical College, Vellore, India (83-i/11/13). No human subjects nor animals were involved in this study.