Localization of the origin of transfer for Salmonella genomic island 4 from Salmonella enterica serovar I 4,[5],12:i:-

Abstract

Salmonella enterica serovar I 4,[5],12:i:- (serovar I 4,[5],12:i:-) is one of the most frequent multidrug-resistant (MDR) Salmonella serovars associated with food–animal production globally, and strains often contain Salmonella genomic island-4 (SGI-4), an integrative conjugative element (ICE) encoding metal tolerance for copper, silver, and arsenic. Horizontal gene transfer (HGT) of SGI-4 from serovar I 4,[5],12:i:- to recipient bacteria results in enhanced metal tolerance for the transconjugants; however, the origin of transfer (oriT) for SGI-4 mobilization is unknown. In this study, the oriT within SGI-4 of MDR serovar I 4,[5],12:i:- strain USDA15WA-1 was identified by (i) cloning an internal region of SGI-4 into a non-mobilizable plasmid and demonstrating HGT to a bacterial recipient, and (ii) deleting the predicted oriT region of SGI-4 from strain USDA15WA-1 and abolishing SGI-4 transfer. Sequence similarity to oriT_SGI-4 was identified in other Enterobacteriaceae, and conjugation of SGI-4 occurred from USDA15WA-1 to Salmonella serovars from Serogroups C–E as well as Escherichia coli and Citrobacter. Localization of the SGI-4 oriT enhances our understanding of a DNA region involved in HGT of an ICE in a frequent MDR Salmonella serovar, thereby providing a model to investigate HGT of SGI-4 and dissemination of metal tolerance genes in the food–animal production environment.

1. Introduction

Salmonellosis caused by non-typhoidal Salmonella (NTS) is a major foodborne illness in humans, and the Centers for Disease Control and Prevention (CDC) estimates 1.35 million cases of salmonellosis, >26,000 hospitalizations, and 420 deaths occur annually in the United States.¹  Salmonella enterica serovar I 4,[5],12:i:- (serovar I 4,[5],12:i:-), a monophasic variant of S. enterica serovar Typhimurium (serovar Typhimurium), has emerged as one of the most frequent multidrug-resistant (MDR) Salmonella serovars, is a common cause of human salmonellosis, and MDR isolates are more likely to be associated with swine.^2–5 A myriad of serovar I 4,[5],12:i:-–associated foodborne outbreaks have been reported worldwide since the late 1990s.^4,6–8 Globally, the most prevalent MDR clone of serovar I 4,[5],12:i:- is R-type ASSuT (ampicillin, streptomycin, sulfisoxazole, and tetracycline) harbouring genes bla_TEM-1, strAB, sul2, and tet(B), respectively, inserted on an ∼28 kb genetic module in the deleted fljB region of the chromosome.^9–11 In addition to antibiotic resistance genes, the MDR module encodes genes conferring tolerance to mercury.

Serovar I 4,[5],12:i:- isolates frequently possess both the MDR module and Salmonella genomic island-4 (SGI-4), an ∼80 kb integrative conjugative element (ICE) typically integrated into the chromosome at the pheR or pheV tRNA.^{2,4,9,12–15} SGI-4 encodes bacterial conjugation components and various heavy metal tolerance genes, including arsenic, copper, and silver. This genomic island was found to be present in >95% of R-type ASSuT epidemic isolates from 1993 to 2010 in the United Kingdom and Italy.⁹ Since its identification, the isolation of SGI-4-containing ASSuT clade has been reported from various parts of the world including the United States,^4,5,13 Canada,¹⁵ Japan,¹⁶ Australia,¹⁷ China,¹⁸ and Southeast Asia.¹⁹ Analysis of 13,612 serovar I 4,[5],12:i:- genomic sequences from 30 countries in the NCBI Pathogen Detection Database indicates that SGI-4 is present in 71% of isolates.² The presence of heavy metal tolerance genes on SGI-4 may provide a survival advantage to bacteria in the presence of elevated metal levels (as an antimicrobial) in animal feed. The global use of elevated concentrations of heavy metals in the swine diet for growth promotion as an alternative to antibiotics has been correlated with the selection of enhanced metal tolerance in Salmonella.²⁰

Previous studies demonstrate that SGI-4 is mobilizable from serovar I 4,[5],12:i:- to other Salmonella strains lacking SGI-4.^12,14,21 In this study, we identified a single origin of transfer (oriT) region located at the proximal end of AOL22_23465, an essential conjugation gene within SGI-4 of serovar I 4,[5],12:i:- isolate USDA15WA-1, a strain linked to a US pork outbreak in 2015.^8,13 Cloning of the oriT region into a non-mobilizable plasmid facilitated conjugative transfer in the presence of SGI-4 in trans, and deletion of the identified oriT region eliminated conjugative transfer of SGI-4. Our results show that the oriT region can efficiently conjugate SGI-4 from strain USDA15WA-1 to S. enterica serovars Choleraesuis, Infantis, Enteritidis, and Senftenberg, indicating the ability of SGI-4 to transfer from Serogroup B serovar I 4,[5],12:i:- to Salmonella Serogroup C, D, and E strains, as well as other Enterobacteriaceae (Escherichia coli and Citrobacter).

2. Material and methods

2.1. Bacterial strains and growth conditions

All bacterial strains were grown in Luria–Bertani (LB) medium with or without antibiotics and incubated at 37 °C unless otherwise indicated. Antibiotics were used at the following concentrations: kanamycin, 50 µg/mL; nalidixic acid, 30 µg/mL; and gentamicin, 10 µg/mL. A full list of bacterial strains and their relevant features is shown in Table 1. To isolate strains of E. coli (EBB 702), S. enterica (BBS 1552, BBS 1748, and BBS 1749) and Citrobacter youngae (BUX 82) that are resistant to nalidixic acid, parental strains were grown in LB broth cultures containing 1.5 to 6 µg/mL of nalidixic acid at 37 °C with shaking. Cultures that were turbid following overnight incubation were single colony isolated to an LB plate containing 30 µg/mL of nalidixic acid and further processed for long-term storage at −80 °C.

Table 1.

Bacterial strains.

Strain number	Strain background	Genus/serovar	Genotype	Antibiotic resistancea	Source
Escherichia coli
EXB 50	Top 10	E. coli	F- mcrA Δ(mrr- hsdRMS- mcrBC) Φ80 lacZΔM15 Δ lacX74 recA1 araD139 Δ(ara- leu)7697 galU galK rpsL endA1 nupG	S	Invitrogen
EXB 53	K-12; EK227	E. coli	Wild type		Lin et al.22
EBB 702	EXB 53	E. coli	gyrA	N	This study
Salmonella enterica
BSX 76	χ3246	Serovar Choleraesuis	Wild type	S	Bearson et al.23
BSX 120	UK-1	Serovar Typhimurium	gyrA D87G	N	Bearson et al.24
BSX 137	USDA15WA-1; FSIS1503788	Serovar I 4,[5],12:i:-	Wild type	ASSuT	Bearson et al.13
BSX 212	FSIS32105085	Serovar Infantis	Wild type		USDA, FSIS
BSX 235	FSIS12211648	Serovar Enteritidis	gyrA D87Y	N	USDA, FSIS
BSX 238	FSIS21924264	Serovar Senftenberg	Wild type		USDA, FSIS
BBS 1268	USDA15WA-1	Serovar I 4,[5],12:i:-	pKD46-Gm	ASSuTGm	Bearson et al.21
BBS 1356	USDA15WA-1	Serovar I 4,[5],12:i:-	SGI-4ΩNeo	ASSuTKn	Bearson et al.21
BBS 1552	BSX 212	Serovar Infantis	gyrA	N	This study
BBS 1614	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI658/659	ASSuTKnZ	This study
BBS 1635	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI658/670	ASSuTKnZ	This study
BBS 1636	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI672/673	ASSuTKnZ	This study
BBS 1637	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI671/659	ASSuTKnZ	This study
BBS 1648	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI680/681	ASSuTKnZ	This study
BBS 1692	USDA15WA-1	Serovar I 4,[5],12:i:-	oriT SGI-4::neo	ASSuTKn	This study
BBS 1699	USDA15WA-1	Serovar I 4,[5],12:i:-	ΔoriTSGI-4	ASSuT	This study
BBS 1714	USDA15WA-1	Serovar I 4,[5],12:i:-	ΔoriTSGI-4/p15aKanZeo-1-oBBI658/659	ASSuTKnZ	This study
BBS 1716	USDA15WA-1	Serovar I 4,[5],12:i:-	ΔoriTSGI-4/p15aKanZeo-1-oBBI672/673	ASSuTKnZ	This study
BBS 1748	BSX 76	Serovar Choleraesuis	gyrA	SN	This study
BBS 1749	BSX 238	Serovar Senftenberg	gyrA	N	This study
BBS 1861	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI671/681	ASSuTKnZ	This study
BBS 1883	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI671/840	ASSuTKnZ	This study
BBS 1885	USDA15WA-1	Serovar I 4,[5],12:i:-	p15aKanZeo-1-oBBI841/681	ASSuTKnZ	This study
BBS 1891	USDA15WA-1	Serovar I 4,[5],12:i:-	AOL22_23465::neo	ASSuTKn	This study
BBS 1893	USDA15WA-1	Serovar I 4,[5],12:i:-	ΔAOL22_23465	ASSuT	This study
BBS 1900	USDA15WA-1	Serovar I 4,[5],12:i:-	ΔAOL22_23465/p15aKanZeo-1-oBBI672/673	ASSuTKnZ	This study
Citrobacter youngae
BUX 7	SX 212	C. youngae	Wild type		This study
BUX 82	BUX 7	C. youngae	gyrA	N	This study

^aAbbreviations for antimicrobial resistance are as follows: ampicillin (A), streptomycin (S), sulfisoxazole (Su), tetracycline (T), nalidixic acid (N), gentamicin (Gm), kanamycin (Kn), and Zeocin (Z).

2.2. Mobilizable plasmid construct (p15aKanZeo-1-SGI-4_oriT)

A non-mobilizable, low copy number vector (referred to as p15aKanZeo-1) was constructed using BbsI (NEB, Ipswich, MA, USA) and T4 DNA Ligase (NEB) in a Golden Gate assembly reaction²⁵ following polymerase chain reaction (PCR) amplification of DNA fragments from pACYC177 (GenBank accession #X06402) and pCR-Blunt (Invitrogen, Carlsbad, CA, USA). The p15a origin of plasmid replication (oriV) and the kanamycin-resistance gene from pACYC177 were amplified using oBBI 595 and oBBI 596. The zeocin-resistance gene from pCR-Blunt was amplified using oBBI 597 and oBBI 598. To clone the oriT region of the SGI-4 ICE, p15aKanZeo-1 was amplified with primers oBBI 664 and oBBI 665, and several intergenic DNA regions from SGI-4 were independently cloned into the vector by Golden Gate assembly using PaqCI (NEB); the ligations were transformed into the E. coli Top10 strain by electroporation with selection for kanamycin resistance. To determine plasmid mobilization by the SGI-4 conjugation complex, candidate plasmids (p15aKanZeo-1-SGI-4_oriT) were transformed into serovar I 4,[5],12:i:- strain USDA15WA-1 by electroporation with selection for kanamycin resistance. Conjugation between kanamycin-resistant USDA15WA-1-containing p15aKanZeo-1-SGI-4_oriT (donor) and nalidixic acid–resistant BSX 120 (serovar Typhimurium strain UK-1 as a recipient) evaluated the transfer of the plasmids containing the potential SGI-4 oriT. Primers used to construct vectors and amplify oriT-specific regions of SGI-4 are listed in Table 2.

Table 2.

Oligonucleotide primers used in this study.

Primer	Sequencea	Target
oBBI 595	GGCTACGAAGACGCagtcCACCTGCAGTCtttcTGTCAGACCAAGTTTACTCATATATAC	pACYC177 vector
oBBI 596	GGCTACGAAGACGTgctaAAACTTCATTTTTACAATTCTGATTAGAAAAACTCATCGAG	pACYC177 vector
oBBI 597	GGCTACGAAGACGTtagcACGTGTCAGTCCTG	pCR-Blunt vector
oBBI 598	GGCTACGAAGACGCgactCACCTGCGACTcaatTACAATTTCCTGATGCGGTATTTTC	pCR-Blunt vector
oBBI 664	ggctatCACCTGCAGTCattgCTGTCAGACCAAGTTTACTCATATATAC	p15aKanZeo-1 vector
oBBI 665	ccgattCACCTGCgactgaaaTACAATTTCCTGATGCGGTATTTTC	p15aKanZeo-1 vector
oBBI 658	GGCTATCACCTGCAGTCtttcTTATATGGCTCTTGCAGCAC	SGI-4 oriT
oBBI 659	GGCTATCACCTGCGACTcaatAGCGTACGGCTTGATATATTTTCA	SGI-4 oriT
oBBI 670	GGCTATCACCTGCGACTcaatCAGTTTCGTGCTCGTAATGC	SGI-4 oriT
oBBI 671	GGCTATCACCTGCAGTCtttcGTTGCATTACGAGCACGAAAC	SGI-4 oriT
oBBI 672	GGCTATCACCTGCAGTCtttcATCCTGACATGAAAATATTGACGTAAG	SGI-4 oriT
oBBI 673	GGCTATCACCTGCGACTcaatGCTATACCTCCATATGACATACACAG	SGI-4 oriT
oBBI 680	GGCTATCACCTGCAGTCtttcCTTTGGAAGCAGATATCGTCAAC	SGI-4 oriT
oBBI 681	GGCTATCACCTGCGACTcaatGAGACGTTCAGACGGGTATTC	SGI-4 oriT
oBBI 689	P-GAAAGCGACCACCAATGTTCCGTGCAGTCTGTGAGGAACGCCTTCGATAGCTGAATGAGTGACGTGC	oriT KO
oBBI 690	GTCTCTGACCTGTTTGCTGCTAATGCAGAGCTGACGTTCGTCAAGC ATAGAGCAGTGACGTAGTCGC	oriT KO
oBBI 840	GGCTATCACCTGCGACTcaatCCACGCAGTGCGTGTGTT	SGI-4 oriT
oBBI 841	GGCTATCACCTGCAGTCtttcGTGGTTTTCATCGTTTTCGT	SGI-4 oriT
oBBI 842	TGTGAAAAGCAGAAGGCTGTGTATGTCATATGGAGGTATAGCGAT AGCTGAATGAGTGACGTGC	AOL22_23456 KO
oBBI 843	ACGTCCTCGTCCATCTCCATGACCTAAATCCATCGATACTAAAGC ATAGAGCAGTGACGTAGTCGC	AOL22_23456 KO

^aBold text indicates homology to the target nucleotide sequence, and underlined text denotes homology to oBBI 92/93-neo for PCR amplification of the linear knockout fragment. Italic text indicates BbsI (GAAGAC) or PaqCI (CACCTGC) restriction enzyme sites and lowercase letters denote the 4-nucleotide overhang for ligation during cloning.

2.3. Deletion of the oriT region in serovar I 4,[5],12:i:- strain USDA15WA-1

SGI-4 oriT knockout was generated using the recombineering method, as described previously.²⁶ Briefly, primers (oBBI 689 and oBBI 690) with homology to upstream and downstream sequences for the predicted oriT region were used to amplify oBBI 92/93-neo encoding the neomycin resistance gene flanked by FRT sites on both sides. The purified PCR fragment was transformed into serovar I 4,[5],12:i:- strain BBS 1268 (derivative of USDA15WA-1 containing pKD46-Gm) by electroporation.²¹ The transformants were incubated on an LB kanamycin plate (for selection of neo insertion) at 37 °C overnight. The resulting colonies were screened by PCR using oBBI 672/673 to confirm removal of the oriT sequence due to insertion of neo (BBS 1692) as determined by the correct product size following gel electrophoresis. The SGI-4 ΔoriT strain (BBS 1699) was generated by deletion of the neomycin gene by FLP-catalysed recombination between FRT sites following transformation of strain BBS 1692 with temperature-sensitive pCP20-Gm.^27,28

2.4. Deletion of the conjugative gene AOL22_23465 in serovar I 4,[5],12:i:- strain USDA15WA-1

Recombineering was used to delete the conjugative gene AOL22_23465. Primers oBBI 842 and oBBI 843 were used to amplify the oBBI 92/93-neo template as described above. A purified PCR fragment was electroporated into strain BBS 1268, and kanamycin-resistant colony was selected (BBS 1891). The neo-cassette was deleted using FLP recombinase (pCP20-Gm), generating the ΔAOL22_23465 mutant (BBS 1893). Integration of the neo-cassette (BBS 1891) and its subsequent removal (BBS 1893) were verified by PCR using primers oBBI 671 and oBBI 659, and oBBI 671 and oBBI 843.

2.5. Conjugation experiments

Salmonella strains were grown overnight with antibiotic selection at 37 °C in 3 mL LB broth with shaking. The strains were diluted 1:100 and sub-cultured without antibiotics in 3 mL LB broth at 37 °C with shaking to an optical density at 600 nm (OD₆₀₀) of ∼0.4. Donor strains contained the plasmid with kanamycin resistance, while the recipient strain BSX 120 was resistant to nalidixic acid. Equal volumes of donor and recipient strains were mixed in a microfuge tube, and 10 μL of the mixture was spotted on an LB agar plate. As a control, 10 μL of the donor and recipient strains were individually spotted on LB agar. Following overnight incubation at 37 °C, cell growth from the conjugation and control plates was scraped and suspended in phosphate-buffered saline buffer. The conjugation mixture and controls were plated on LB agar medium containing both kanamycin and nalidixic acid, kanamycin alone, or nalidixic acid alone to select for transconjugants, donors, or recipients, respectively. The conjugation frequency was determined by dividing the colony-forming units (CFU)/mL of transconjugants by the CFU/mL of donors for replicates of at least 3 experiments.

2.6. BLAST analysis

BLAST analysis was performed to identify the top 5,000 hits for 3 independent query sequences derived from SGI-4 of serovar I 4,[5],12:i:- strain USDA15WA-1 (CP040686). The oriT_{SGI-4_966} sequence is a 966 bp fragment spanning nucleotides 4704810 to 4705775. The oriT_{SGI-4_212} sequence is a 212 bp fragment spanning nucleotides 4705288 to 4705499. The composite SGI-4 sequence (52,270 bp) without metal tolerance genes was derived from nucleotides 4659019 to 4680258, 4687402 to 4713938, and 4735211 to 4739704.

2.7. Quantification and statistical analysis

Data analysis for statistical significance was performed with GraphPad Prism software. P values were calculated using a 2-tailed, unpaired Student’s t-test for 2-group comparisons. P values <0.05 were considered significant. Each experiment is the average of at least 3 independent replicates.

3. Results

3.1. Genomic location of oriT in SGI-4

Previous studies demonstrate that SGI-4 in serovar I 4,[5],12:i:- is an ICE that can transfer to a recipient Salmonella by conjugation.^12,21 However, the location of the oriT within SGI-4 is unknown. In this study, the nucleotide sequence of SGI-4 from strain USDA15WA-1 was compared against 91 oriT sequences derived from plasmids and ICEs previously identified in Ares-Arroyo et al.,²⁹ but homology between alignments was not sufficient to suggest identification of the SGI-4 oriT (data not shown). To identify the location of oriT, we hypothesized that the functional fragment was (i) located in an intergenic region and (ii) flanked by putative conjugation-associated genes (ie, not located adjacent to metal tolerance genes); 3 intergenic DNA segments within SGI-4 (4697074 to 4698508, 4670053 to 4671100, and 4704600 to 4706000 from CP040686) were independently cloned into a non-mobilizable, low copy number vector, p15aKanZeo-1 (described in the Materials and Methods section, Fig. S1). While the parent vector (p15aKanZeo-1) is not transferrable from donor to recipient bacteria, the conjugation machinery encoded by the SGI-4 ICE was predicted to mobilize in trans p15aKanZeo-1 containing the SGI-4 oriT sequence. Each candidate vector construct was transformed into strain BSX 137 (USDA15WA-1) by electroporation, and conjugation experiments were conducted to assess conjugative transfer of plasmids containing potential oriT sequences. Of the tested segments, a single 1,401 bp fragment (oBBI 658/659) spanning nucleotides 4704600 to 4706000 in the USDA15WA-1 chromosome contained a functional oriT (Fig. 1; data not shown for plasmids containing SGI-4 regions that are unable to transfer). The oriT fragment is an intergenic region between genes AOL22_23460 and AOL22_23465 encoding a putative YkgJ family cysteine cluster protein and an integrase-domain-containing protein, respectively. Conjugation between the BBS 1614 donor (kanamycin-resistant) and the BSX 120 recipient (nalidixic acid–resistant) strains resulted in plasmid transfer with an average frequency of 2.38 × 10⁻⁵ transconjugants per donor cell.

Fig. 1.

Location of SGI-4 oriT. Schematic diagram (left) of the putative oriT region (between AOL22_23460 and AOL22_23465; denoted by arrows) in SGI-4 with arrowheads indicating the relative positions of oligonucleotide primers (oBBI # listed below arrowhead) used to amplify the various lengths of the putative oriT region. The table (right) indicates the mean conjugation efficiency for transfer of plasmids containing fragments of the putative oriT region from donor strains to the BSX 120 recipient. Conjugation efficiency was calculated by dividing the CFU/mL of transconjugants divided by the CFU/mL of donors. Experiments were performed a minimum of 3 times. ^†Standard error of the mean. ^‡Transconjugants not detected (ND). Contrasting letters indicate statistically significant differences (P < 0.01).

To refine the region of a fully functional SGI-4 oriT, the 1,401 bp oriT fragment was progressively truncated at both the 5' and 3' ends, and resulting fragments were tested for transfer efficiency by conjugation (Fig. 1). Maximal transfer (2.79 × 10⁻⁵) was observed with a 966 bp fragment (oBBI 672/673) spanning nucleotides 4704810 to 4705775 and named oriT_{SGI-4_966}. Vector mobilization was significantly greater (P < 0.01) with oriT_{SGI-4_966} compared with smaller fragments evaluated including oBBI 680/681 (2.98 × 10⁻⁶; 9.4-fold lower) and oBBI 671/681 (3.82 × 10⁻⁶; 7.3-fold lower) which are internal 451 and 212 bp fragments, respectively. It is likely that truncation disrupts the optimal spacing or removes auxiliary cis-acting binding sites required for efficient nicking or relaxosome assembly. The shortest fragment that retained transfer function was the 212 bp sequence (oBBI 671/681) spanning nucleotides 4705288 to 4705499. To assess the functional contribution of each end of oriT_{SGI-4_212}, the 5' end was truncated to generate a 112 bp fragment (oBBI 841/681) and the 3' end was truncated to generate a 105 bp fragment (oBBI 671/840); neither fragment alone was sufficient to mediate plasmid transfer when cloned into p15aKanZeo-1. These findings indicate that oriT_{SGI-4_212} contains sequence elements essential for oriT functionality that are not fully present in either the oBBI 841/681 or oBBI 671/840 fragments.

Previously we demonstrated that SGI-4 marked with the kanamycin-resistance gene (SGI-4Ωneo) in serovar I 4,[5],12:i:- (BBS 1356) can be transferred to a bacterial recipient.²¹ Conjugation efficiency of SGI-4Ωneo compared with the plasmid containing oriT_{SGI-4_966} in BBS 1636 indicated that the transfer of the oriT_{SGI-4_966}-containing plasmid (2.79 × 10⁻⁵) occurred at a significantly higher frequency (∼10-fold) compared with the chromosomal SGI-4Ωneo ICE (2.92 × 10⁻⁶; Table 3).

Table 3.

Mean conjugation efficiency for donor strains with various combinations of chromosomal and plasmid containing oriT_SGI-4.^a

Donor strain	Chromosomal oriTSGI-4	Plasmid oriTSGI-4	Conjugation efficiencyb	S.e.m. c	Significance
BBS 1636	+	672/673	2.79 × 10−5	4.51 × 10−6	*
BBS 1716	−	672/673	1.44 × 10−4	1.83 × 10−5	†
BBS 1356	+	…	2.92 × 10−6	8.95 × 10−7	‡
BBS 1692	−	…	NDd	…	…

*, †, and ‡ with contrasting symbols indicate statistically significant differences (P < 0.01).

^aConjugation recipient was BSX 120.

^bMean conjugation efficiency calculated as described in Fig. 1.

^cStandard error of the mean.

^dTransconjugants not detected (ND).

3.2. An SGI-4 oriT_SGI-4 knockout is deficient in self-mobilization but is proficient for mobilization of a plasmid containing oriT

To exclude the possibility of having more than 1 oriT in SGI-4, a 579 bp segment of oriT_SGI-4 was deleted in serovar I 4,[5],12:i:- strain USDA15WA-1 (BBS 1268) by recombineering, creating strain BBS 1692 (SGI-4 oriT::neo). Conjugation of BBS 1692 with BSX 120 did not transfer SGI-4 to the recipient as previously shown for BBS 1356, suggesting that only a single oriT is present in SGI-4. The plasmid containing oriT_{SGI-4_966} was transformed into BBS 1699 (ΔoriT_SGI-4) to evaluate mobilization of the plasmid by the SGI-4 conjugation machinery in the absence of oriT_SGI-4 within the chromosomal ICE. Conjugation experiments using donor strain BBS 1716 with a deletion of oriT_SGI-4 had significantly enhanced plasmid transfer efficiency compared with the donor strain with wild-type SGI-4 (BBS 1636; Table 3). These results indicate that the 579 bp oriT region that was deleted does not contain accessory conjugation machinery required for horizontal gene transfer (HGT; other than oriT_SGI-4). Furthermore, competition for conjugation machinery to transfer full-length SGI-4 and the plasmid containing oriT_SGI-4 is eliminated in the ΔoriT_SGI-4 strains, resulting in significantly higher conjugation frequencies.

3.3. Deletion of AOL22_23465 abolishes SGI-4-mediated conjugation

As previously noted, the oriT_SGI-4 sequence is adjacent to AOL22_23465, encoding a putative integrase-domain-containing DNA-binding protein. Based on this context, the role of AOL22_23465 in conjugation was investigated. Deletion of AOL22_23465 from the chromosome abolished the transfer of both the SGI-4 ICE (BBS 1891) and a plasmid carrying the oriT_{SGI-4_966} sequence (BBS 1900) into the recipient strain (BSX 120), revealing its essential role in conjugative transfer (data not shown).

3.4. USDA15WA-1 can transfer SGI-4 to different Salmonella serogroups and other Enterobacteriaceae

Conjugation experiments were performed to determine transfer of SGI-4 (BBS 1356 donor) and the oriT_{SGI-4_966}-containing plasmid (BBS 1639 donor) from serovar I 4,[5],12:i:- to S. enterica serovars Choleraesuis (C₁), Infantis (C₁), Enteritidis (D₁), and Senftenberg (E₄), as well as E. coli and C. youngae (Table 4). SGI-4 and the plasmid containing oriT_{SGI-4_966} were transferred to all the bacterial recipients with varying degrees of conjugation efficiency (7.38 × 10⁻⁵ to 8.73 × 10⁻⁸). In agreement with our conjugation data with serovar Typhimurium as a recipient, the plasmid containing oriT_{SGI-4_966} transferred at a greater efficiency than full-length chromosomal SGI-4 for each bacterial recipient, ranging from 4- to 224-fold increased efficiency. Conjugational transfer of the full-length, chromosomal SGI-4 ranged from 4.92 × 10⁻⁶ for E. coli (highest) to 8.73 × 10⁻⁸ serovar Infantis (lowest), while a log difference was noted for conjugation efficiency of the oriT_{SGI-4_966}-containing plasmid into serovar Senftenberg (highest at 7.38 × 10⁻⁵) compared with serovar Infantis (lowest at 6.43 × 10⁻⁶). Thus, SGI-4 or a plasmid containing oriT_SGI-4 conjugatively transferred from Serogroup B serovar I 4,[5],12:i:- to Salmonella serovars in Serogroups C, D, and E, and other Enterobacteriaceae (E. coli and C. youngae).

Table 4.

HGT of SGI-4 or plasmid containing oriT_{SGI-4_966} to Salmonella serovars or other Enterobacteriaceae.

Recipient	Strain background	Serogroup	Conjugation efficiency of SGI-4 ICEa	S.e.m.	Conjugation efficiency of oriTSGI-4_966b	S.e.m.
BSX 120	Typhimurium	B	2.92 × 10−6	8.95 × 10−7	2.79 × 10−5	4.51 × 10−6
BBS 1748	Choleraesuis	C1	2.96 × 10−6	3.4 × 10−7	1.51 × 10−5	3.53 × 10−6
BBS 1552	Infantis	C1	8.73 × 10−8	4.83 × 10−8	6.43 × 10−6	3.94 × 10−8
BSX 235	Enteritidis	D1	4.37 × 10−6	1.73 × 10−6	2.55 × 10−5	1.23 × 10−5
BBS 1749	Senftenberg	E4	3.29 × 10−7	4.9 × 10−8	7.38 × 10−5	8.3 × 10−6
EBB 702	Escherichia coli	NAc	4.92 × 10−6	9.66 × 10−7	1.85 × 10−5	3.24 × 10−6
Bux 82	Citrobacter youngae	NAc	9.33 × 10−8	5.81 × 10−8	1.05 × 10−5	1.26 × 10−6

^aBBS 1356 donor.

^bBBS 1636 donor.

^cNot applicable.

3.5. The oriT_SGI-4 sequence is present in other Enterobacteriaceae

To determine whether the nucleotide sequences for oriT_{SGI-4_212} and oriT_{SGI-4_966} are present in bacterial strains other than Salmonella (taxid:590), BLAST was performed excluding this genus. Both oriT_{SGI-4_212} and oriT_{SGI-4_966} sequences were present and conserved (≥90% coverage and ≥90% identity) in numerous Enterobacteriaceae isolate sequences, although only a few strains for a given species contained the oriT_SGI-4 sequence, suggesting that the overall prevalence in the population may be low (Table S1). For example, the oriT_SGI-4 nucleotide sequences were present in 4 Citrobacter freundii strains (accessions CP119048, CP128207, CP137195, and CP186042) with 100% coverage and identity. Both oriT_{SGI-4_212} and oriT_{SGI-4_966} sequences were present in an additional 30 isolates of Enterobacteriaceae with ≥92% coverage and ≥95.87% identity, including strains of several Klebsiella species (quasipneumoniae, variicola, pneumoniae, oxytoca, and michiganensis) and 16 Edwardsiella ictaluri strains. To determine whether the oriT sequences were co-localized with SGI-4, a BLAST exclusive of Salmonella was performed with the ICE sequence (excluding the metal tolerance genes). Alignment of the entire SGI-4 ICE (with or without metal tolerance genes) from serovar I 4,[5],12:i:- strain USDA15WA-1 resulted in 99% coverage with 99.89% sequence identity to the 4 C. freundii strains, indicating the Citrobacter strains contain SGI-4 including metal tolerance genes (Table S1 and data not shown). The Klebsiella species and E. ictaluri isolate sequences have a lower coverage (≤96%) and identity (≤98.96%) to SGI-4, suggesting that sequences with lower homology may be SGI-4-like elements that are more distantly related to SGI-4 of serovar I 4,[5],12:i:-. Genomic sequences with ≤90% coverage and/or identity to oriT_{SGI-4_212} and oriT_{SGI-4_966} sequences tend to have a lower conservation of SGI-4. Numerous genomic sequences were identified with similarity to only 2 of the 3 BLAST query sequences (oriT_{SGI-4_212} and composite SGI-4, Table S2; oriT_{SGI-4_966} and composite SGI-4, Table S3; and oriT_{SGI-4_212} and oriT_{SGI-4_966}, Table S4). Most of these sequences tend to have lower conservation for 1 or both query sequences; whether the sequences with lower homology facilitate a role in conjugation is unknown, but these tables provide a foundation for future scientific inquiry into their role in microbial physiology and HGT.

4. Discussion

Genomic islands are discrete DNA segments acquired by HGT that result in bacterial diversification and often provide the host with new traits such as metabolic capacity, antibiotic resistance or enhanced virulence.³⁰ Globally, >70% of serovar I 4,[5],12:i:- strain sequences contain SGI-4, an ∼80 kb ICE encoding heavy metal tolerance genes for copper, silver, and arsenic.² SGI-4 encodes 87 open reading frames including a putative conjugation system (Type IV secretion system and mobilizable genes), indicating the potential for HGT of SGI-4 to recipient bacteria in the environment with serovar I 4,[5],12:i:-.¹² However, the location of the SGI-4 oriT was not recognized by sequence homology with 91 known oriTs published by Ares-Arroyo et al.²⁹ The goal of this study was to identify the oriT region within SGI-4.

Our experiments revealed that the SGI-4 oriT sequence of serovar I 4,[5],12:i:- strain USDA15WA-1 is located in an intergenic region between AOL22_23460 and AOL22_23465. Deletion of oriT_SGI-4 (BBS 1692) abolished transfer of SGI-4 to a recipient cell but the remaining conjugation machinery facilitated mobilization of a plasmid (p15aKanZeo-1) containing oriT_SGI-4 from BBS 1636 (Table 3), indicating a single oriT present in SGI-4. The mobilization of the oriT_SGI-4-containing plasmid from BBS 1636 occurred at a significantly higher frequency than chromosomally integrated SGI-4 in BBS 1356 (Table 3). We predict that due to its smaller size (∼3.5 kb), higher copy number (∼15/cell), and extrachromosomal location, transfer of the plasmid may have a temporal advantage compared with the excision and transfer of the ∼80 kb chromosomal ICE. Furthermore, in the absence of oriT in the chromosomal SGI-4 ICE (BBS 1716), an even greater conjugation efficiency of the plasmid containing oriT_SGI-4 was observed compared with BBS 1636 with oriT present in SGI-4 (Table 3), presumably due to elimination of competition for conjugation machinery.

The SGI-4 oriT sequence is located in an intervening region upstream of AOL22_23465 encoding an integrase-domain-containing protein. Deletion of AOL22_23465 (BBS 1891) abolished transfer of both SGI-4 and an oriT_SGI-4-containing plasmid to recipient cells, indicating a requisite role of AOL22_23465 in SGI-4 conjugative transfer including mobilization. In contrast, deletion of oriT_SGI-4 (BBS 1692) only abolished transfer of SGI-4 but not an oriT_SGI-4-containing plasmid to a recipient cell. This indicates that deletion of oriT_SGI-4 did not delete a functional promoter region for AOL22_23465 in BBS 1692 or its derivatives.

NCBI BLASTp and InterProScan analyses indicated that AOL22_23465 encodes a 300-amino-acid protein containing a putative integrase/recombinase domain with catalytic features characteristic of DNA-cleaving and rejoining enzymes. This suggests that AOL22_23465 may contribute to SGI-4 ICE excision from and/or integration into the chromosome. If AOL22_23465 is solely involved in excision/integration of SGI-4, the AOL22_23465 deletion mutant should still facilitate transfer of the plasmid containing oriT_SGI-4 to a recipient cell. However, transfer of both SGI-4 and the oriT_SGI-4-containing plasmid (BBS 1900) is abolished in a AOL22_23465 deletion mutant, suggesting that AOL22_23465 participates in the SGI-4 conjugative transfer process (potentially during initiation or resolution) and not simply excision/integration from the chromosome.

Genomic islands in Salmonella frequently have a nucleotide composition that is AT-rich compared with the core genome.^30,31 However, the percentage of GC nucleotides in oriT_{SGI-4_966} and oriT_{SGI-4_212} (50.8% and 51.9%, respectively) is similar to the entire SGI-4 ICE (51.9%) and the complete genome sequence (52%) of strain USDA15WA-1. A search for nucleotide similarity to USDA15WA-1 oriT_SGI-4 in the NCBI database using the BLAST search tool and excluding Salmonella sequences indicated that the oriT_SGI-4 region is highly conserved (90% to 100%) among several members of Enterobacteriaceae, including certain isolates of C. freundii, Klebsiella (species quasipneumoniae, variicola, pneumoniae, oxytoca, and michiganensis) and E. ictaluri (Table S1). Branchu et al. demonstrated the presence (≥75% nucleotide identity with ≥40% coverage) of SGI-4-like ICE sequences in diverse members of Enterobacteriaceae, including Erwinia tracheiphila, E. ictaluri, Enterobacter hormachei subspecies oharae and steigerwaltii, Enterobacter cloacae, and Pluralibacter gergoviae¹²; our BLAST search for the USDA15WA-1 oriT_SGI-4 sequence in the NCBI database (excluding Salmonella) demonstrated the presence (≥57% coverage with ≥78% identity) of oriT_SGI-4 in all of these isolates except E. hormachei subspecies oharae. Alignment of the entire SGI-4 ICE from serovar I 4,[5],12:i:- strain USDA15WA-1 resulted in ∼99% of SGI-4 coverage with 99.88% sequence identity for 4 C. freundii strains and lower alignment coverage for other Enterobacteriaceae, thereby expanding the investigation of Branchu et al. Identification of homologous sequences to USDA15WA-1 oriT_SGI-4 in other Enterobacteriaceae suggests an analogous role in conjugative transfer of SGI-4-like ICE. Further evidence of SGI-4 transfer amongst Enterobacteriaceae was observed with conjugation of both SGI-4 and a plasmid containing oriT_SGI-4 to Salmonella serovars Choleraesuis (C₁), Infantis (C₁), Enteritidis (D₁), and Senftenberg (E₄), as well as, E. coli and C. youngae, as has been shown by other investigations for serovars Typhimurium (B), Heidelberg (B), Thompson (C₁), Hadar (C₂), Newport (C₂), and Cerro (K).^12,14,21 Demonstration of SGI-4-dependent transfer of a plasmid containing oriT_SGI-4 to a bacterial recipient provides a model system for investigation of HGT and a platform for optimization of gene delivery in Enterobacteriaceae. In summary, our investigation identified the oriT region for a relevant ICE (SGI-4) in MDR serovar I 4,[5],12:i:- that can conjugatively transfer metal tolerance genes to other Enterobacteriaceae, thereby expanding microbial diversity and enhancing bacterial survival in metal-containing environments.

Supplementary material

Supplementary data are available at DNARES online.

Funding

This work was supported by the United States Department of Agriculture (USDA) and Agricultural Research Service (ARS) appropriated funds from CRIS project 5030-31000-007-000-D. This research utilized an appointment to the Agricultural Research Service Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the US Department of Energy (DOE) and the USDA. ORISE is managed by ORAU under DOE contract number DE-SC0014664. Funding for open access charge: USDA, ARS.

Disclaimer

Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendations or endorsement by the US Department of Agriculture. The US Department of Agriculture is an equal opportunity provider and employer.

Data availability

The data underlying this article is available at Ag Data Commons: https://doi.org/10.15482/USDA.ADC/27115081.v1³² and https://doi.org/10.15482/USDA.ADC/27111475.v1.³³