Salmonella Genomic Island 1 (SGI1), Variant SGI1-I, and New Variant SGI1-O in Proteus mirabilis Clinical and Food Isolates from China

Abstract

Salmonella genomic island 1 (SGI1) and variants (SGI1-I and the new variant SGI1-O) were mapped in five strains of Proteus mirabilis isolated from humans and food in China. Sequencing showed that SGI1 and variants were integrated at the 3′ end of the chromosomal thdF gene as previously described for Salmonella strains.

Salmonella genomic island 1 (SGI1) is a 43-kb integrative mobilizable element initially characterized in Salmonella enterica serovar Typhimurium phage type DT104 strains (3, 7). SGI1 contains a multidrug resistance region (MDR), which is a complex class 1 integron containing all the genes necessary for resistance located between a resolvase-like gene (res) and a gene of unknown function (S044) (Fig. 1). SGI1 and variants of it named SGI1-A to SGI1-N have been described for other Salmonella enterica serovars (5, 11, 12, 14, 17, 18). In Salmonella SGI1 is integrated in the last 18 bp of the thdF gene (specific attB site) (7, 14). SGI1 has been shown to be mobilized using a conjugative helper plasmid and thus transferred to Salmonella or Escherichia coli recipient strains (7). Typically SGI1 integration can be confirmed by PCR targeting the left junction region with primers U7-L12 (thdF specific) and LJ-R1 (SGI1 int specific) and/or the right junction region with primer 104RJ (SGI1 S044 specific) and either primer C9-L2 (serovar Typhimurium retron int specific) or primer 104D (yidY specific) (Fig. 1) (2, 3). Recently the presence of an SGI1 variant (SGI1-L) was described in clinical strain Proteus mirabilis 18306, the first bacterium other than Salmonella shown to harbor this element (1). In P. mirabilis 18306 the left junction region could not be detected by PCR, and it was concluded that SGI1 may be integrated elsewhere in the genome (1). In a comment on that report it was pointed out that the thdF gene of P. mirabilis HI430 is only 70% identical to the serovar Typhimurium gene and that in fact the primer U7-L12 binding site contains seven nucleotide differences and, thus, this may explain the failure of the left junction PCR (Fig. 1) (8).

FIG. 1.

Sequence alignment of the left (top) and right (bottom) junction regions of SGI1 from S. enterica serovar Typhimurium DT104 (StSGI1) and P. mirabilis B02012 (PmSGI1) and the genomic region from P. mirabilis HI4320 containing thdF and the downstream region. The nucleotide differences between the Proteus and Salmonella thdF genes are boxed (top) as are the ones in the attB site (bottom). The nucleotide differences between the Salmonella and Proteus sequences after the right junction (bottom) are not indicated, as these sequences are not homologous. Base pair coordinates for StSGI1 are from the sequence with accession no. AF261825. Primers used to detect the junction regions are indicated by arrows.

Strains characterized in this study are listed in Table 1. Primers used in this study are listed in Table 2. PCR was carried out with Amplitaq Gold (Applied Biosystems, Foster City, CA) or the FailSafe PCR System (Epicentre Biotechnologies, Madison, WI). Sequencing was by automated dideoxy cycle sequencing.

TABLE 1.

P. mirabilis strains characterized in this study

Strain	Integron cassette(s)		SGI1 type	Antibiogramc	Isolate source
	Gene(s)a	Size(s) (kb)b
B02012	aadA2, blaPSE-1	1, 1.2	SGI1	AmChCiKaNaStSuTeTm	Patient
C02035	aadA2, dfrA1-orfC	1, 1.2	SGI1-I	ChStSuTeTmd	Food
C04014	dfrA1-orfC, blaPSE-1-aadA2	1.2, 2.0	SGI1-O	AmChCiKaNaStSuTeTme	Food
C05022	dfrA1-orfC	1.2	SGI1-O	AmChKaStSuTeTm	Patient
C05023	dfrA1-orfC	1.2	SGI1-O	AmChKaStSuTeTm	Patient
P06020	dfrA7	0.75	None	(Am)ChKaStSuTeTm	Patient
P06021	dfrA7	0.75	None	AmChKaStSuTeTm	Patient

^aAs determined by PCR mapping and/or DNA sequencing.

^bSize of amplicon produced with primers 5′-CS and 3′-CS.

^cResistant or intermediate resistance (parentheses) to drugs shown as determined by broth microdilution. Drugs tested: amikacin, ampicillin (Am), amoxicillin-clavulanic acid, cefoxitin, ceftiofur, ceftriaxone, chloramphenicol (Ch), ciprofloxacin (Ci), gentamicin, kanamycin (Ka), nalidixic acid (Na), streptomycin (St), sulfizoxazole (Su), tetracycline (Te), and trimethoprim-sulfamethoxazole (Tm).

^dBy streptomycin Etest exhibited a MIC of 8 μg/ml with colonies in the zone. A single colony from the zone exhibited a MIC of 96 μg/ml.

^eBy streptomycin Etest exhibited a MIC of 32 μg/ml with colonies in the zone. A single colony from the zone exhibited a MIC of 96 μg/ml.

TABLE 2.

Primers used in this study

Primera	Sequence (5′-3′)	Target regionb	PCR		Reference
			No.c	Size (bp)
U9-L1 (+)	TACTACAAGCAGATAACGCC	2771-3679	St1	909	13
P1-R1 (−)	TAGAAACGACAAAGCGCGTG				13
P134-L2 (+)	TGACCCAATTCCAAAGCCAC	16784-17839	St4	1,490	13
P134-R1 (−)	GTGTTTGGGCAAGATCCCAG				13
StGP21 (+)	ATAACGGCAGGTTCCGGTTC	20173-21108	St5	936	13
StGP6 (−)	CGATGAAGCGCACAAATTTG				13
StGP24 (+)	TCAAGATTCCTATCTGCAGG	24363-25201	St6	838	13
StGP28 (−)	AGAGTTACTAGACCAAGCGC				13
LJ-R1 (−)	AGTTCTAAAGGTTCGTAGTCG	int	1	500	3
PmLJ1 (+)	ACACCTACAACAAGGCTATC	thdF of P. mirabilis	1	500	This study
			23	348
104RJ (+)	CTGACGAGCTGAAGCGAATTG	S044	2	476	3
PmRJ1 (−)	GATGCACACTGAGTTGATAG	hipB of P. mirabilis	2	476	This study
			23	348
5′-CS (+)	GGCATCCAAGCAGCAAGC	5′-CS region	3	Variabled	10
3′-CS (−)	AAGCAGACTTGACCTGAT	3′-CS region	3	Variabled	10
tetG-1s (+)	GCTCGGTGGTATCTCTGCTC	tet(G)	4	468	15
tetG-2s (−)	AGCAACAGAATCGGGAACAC	tet(G)	4	468	15
StCm-L (+)	CACGTTGAGCCTCTATATGG	floR	5	888	2
StCm-R (−)	ATGCAGAAGTAGAACGCGAC	floR	5	888	2
F4 (+)	TTCCTCACCTTCATCCTACC	floR	6	599	6
F6 (−)	TTGGAACAGACGGCATGG	tetR	6	599	6
MDR-13 (+)	TCCCGATTCTGTTGCTGCTTG	tet(G)	7	1,078	This study
MDR-B3 (−)	AAGCATGGCTGCTGACAAC	orf1	7	1,078	This study
floR-R1 (−)	TCAACGTGAGTTGGATCATAG	floR	8	1,430	This study
MDR-5 (+)	TAGGTATGGGGCTCATAATTG	qacEΔ1	8	1,430	This study
104-D4 (+)	ATGCCTAGCATTCACCTTCC	sulI	9	1,183e	This study
DB-B4 (−)	ATCACATCACCCTGGAAATGG	IS6100	9	1,183e	This study
			21	2,954e
			22	2,642e
DB-T1 (+)	TGCCACGCTCAATACCGAC	IS6100	10	930	2
DB-B6 (−)	CTGTGCCTTCTTGCGAGC	S044	10	930	This study
StGP-23 (+)	CCTTGGTACGTTCGCTAATC	res	11	1,417	This study
			16	1,848
			20	2,262
int-R (−)	GCCTTGCTGTTCTTCTACGG	intI1	11	1,417	This study
AadA2-L (+)	TGTTGGTTACTGTGGCCG	aadA2	12	538	15
AadA2-R2 (−)	TGCTTAGCTTCAAGTAAGACG	aadA2	12	538	3
			15	1,316, 2,360f
MDR-20 (+)	TAGTTCAAAGTTTCAGCAAG	blaPSE-1	13	765	This study
			17	1,751, 2,608g
PSE-R2 (−)	ACAATCGCATCATTTCGCTC	blaPSE-1	13	765	3
PSE-R3 (−)	CTGAAACTTTGAACTACTTGC	blaPSE-1	14	935	This study
MDR-19 (−)	AACCGGATCAGAAATCCATGC	groEL/int	14	935	This study
			19	1,254
MDR-1 (+)	TGATCGAAATCCAGATCCTG	intI1	15	1,316, 2,360f	This study
AadA2-R3 (−)	GGTTCGAAATTTCGATGGTC	aadA2	16	1,848	This study
QS-2 (−)	TGAGTGCATAACCACCAGCC	sul1	17	1,751, 2,608g	3
dfrA1-F (+)	CGAAGAATGGAGTTATCGG	dfrA1	18	919	11
orfC-R (−)	TCTCGAATCAAGCAGGAACC	orfC	18	919	11
dfrA1-R (−)	TTAGAGGCGAAGTCTTGG	dfrA1	19	1,254	11
			20	2,262
orfC-F (+)	CATTACGAAGCGAA GCACC	orfC	21	2,954e	11
PSE-D1 (+)	AGAGCGAAATGATGCGATTG	blaPSE-1	22	2,642e	This study
Ag5 (−)	ACAAACGACAAGCCACGT	orf513	Negh	750	This study
Ag6 (+)	TATCGTCTATCGTACACTCTC	orf513	Negh	750	This study

^a(+), 5′-3′ is left to right in Fig. 2; (−), 5′-3′ is right to left in Fig. 2.

^bCoordinates are from SGI1 (accession no. AF261825), and genes are as shown in Fig. 2.

^cPCR numbers as shown in Fig. 2.

^dSize varies depending on number of cassettes in this region.

^eSize would be ∼2.9 kb larger if this region contained common region 1 (CR1) and dfrA10 as found in some SGI1 variants (2, 14).

^fSize was 1,316 bp in SGI1 but 2,360 bp in P. mirabilis C04014.

^gSize was 1,751 bp in SGI1 but 2,608 bp in P. mirabilis C04014.

^hPCR with Ag5/Ag6 primers was negative for all strains tested but would be expected to give a product of 750 bp if CR1 was present.

We recently screened 30 clinical P. mirabilis strains isolated from the environment, food, and stool of diarrheic patients in China and P. mirabilis ATCC 29906 and ATCC 29245 for the cassettes of class 1 integrons (5′-CS/3′-CS primers) and detected amplicons in seven strains. (Table 1). All seven strains were multidrug resistant, with the number of drugs ranging from four up to nine (Table 1). PCR mapping and/or DNA sequencing of the cassettes revealed that one strain harbored aadA2 and bla_PSE-1, one strain harbored aadA2 and dfrA1-orfC, three strains harbored dfrA1-orfC, and two strains harbored dfrA7 (Table 1; also see below). We used the thdF gene of serovar Typhimurium as a query sequence in a BLAST search of the completely sequenced genome of P. mirabilis HI4320 (http://www.sanger.ac.uk/Projects/P_mirabilis) to find the thdF homologue and flanking sequence and designed primers PmLJ1 and PmRJ1 (Table 2; Fig. 1). We note that the P. mirabilis HI4320 thdF gene is located 145 bp upstream from the start of hipB/hipA toxin/antitoxin homologues (Fig. 2). We conducted PCR with primers PmLJ1 and LJ-R1 (left junction) on all 30 P. mirabilis clinical strains and the two ATCC strains and obtained an amplicon of the predicted size (500 bp if SGI1 was present) only with DNA from the five strains that harbored 1- and/or 1.2-kb cassettes. The seven strains that harbored cassettes were then analyzed by PCR with primers 104RJ and PmRJ1 (right junction), and only the five strains positive for the left junction region produced an amplicon of the expected size (410 bp). The two strains harboring the dfrA7 gene (accession number X58425; 0.75-kb cassette) were not further studied, as it was concluded that although they contained integrons and were multidrug resistant, they were not associated with SGI1 (Table 1). Sequence analysis of both the left and right junction amplicons from P. mirabilis B02012 confirmed that the last 18 bp of the thdF gene corresponding to the left direct repeat (DR-L) is 100% identical to the right 18-bp direct repeat (DR-R) (Fig. 1). This result is in accordance with the sequenced genome of P. mirabilis HI4320, in which the last 18 bp of the putative thdF gene (the attB site) is identical to the previously described attP site of SGI1 (7, 8). Thus, five P. mirabilis strains appeared to have SGI1 or a variant integrated at the end of the thdF gene as in Salmonella. Further, all five strains were positive in PCRs St1, St4, St5, and St6 designed to amplify various regions of SGI1 outside of the MDR (Fig. 2) (13). We used various combinations of the primers listed in Table 2 to map the SGI1s in the five strains (Fig. 2). One strain contained SGI1 (3), one strain contained SGI1-I (11), and three strains contained a new variant named SGI1-O, which contained only the drfA1-orfC cassette. The orf513 gene from common region 1 (CR1) found in some SGI1 variants was not detected (2). Interestingly, one strain with SGI1-O, P. mirabilis C04014 isolated from chicken, was positive by PCR for bla_PSE-1 and aadA2. However, though mapping showed these to be adjacent to one another as the only cassettes in a class 1 integron, they were not linked to regions upstream (res) or downstream (IS6100) of the MDR of SGI1 (Fig. 2).

FIG. 2.

Schematic view of the SGI1 and SGI1 variants as integrated in P. mirabilis. The P. mirabilis HI4320 thdF gene and downstream region are shown at the top, and the MDRs of SGI1, SGI1-I, SGI1-L, and SGI1-O (not to scale) are shown at the bottom. Base pair coordinates are from the sequence with accession no. AF261825. PCRs carried out to map the SGI1s are indicated by thick black bars and are numbered (Table 1). SGI1 was mapped as shown, and SGI1-I was mapped as for SGI1 and also with PCRs 18, 19, and 21. The bla_PSE-1/aadA2 integron from P. mirabilis C04014 was negative in PCRs 9, 10, 11, and 16.

Thus, when present SGI1 is integrated via the specific attB site defined by the last 18 bp of the thdF gene in P. mirabilis as predicted (Fig. 1) (8).

The identification of SGI1 and its variants in P. mirabilis isolated from food and human clinical isolates from China highlights the increased dissemination of the SGI1 element both geographically and biologically. The documented potential for increased virulence in strains harboring this element in conjunction with the multidrug-resistant nature of the SGI1 element raises concerns especially since these new strains harboring SGI1 have been identified from food sources (4, 9, 16, 19).

Nucleotide sequence accession numbers.

The dfrA1-orfC cassette region from SGI1-O and the bla_PSE-1-aadA2 cassette region from P. mirabilis C04014 have been assigned accession numbers EU006710 and EU006711, respectively, and the left and right junction regions of SGI1 in P. mirabilis B02012 have been assigned accession numbers EU180707 and EU180708, respectively, in the GenBank database.