Complete DNA Sequence and Structural Analysis of the Enteropathogenic Escherichia coli Adherence Factor Plasmid

Abstract

The complete nucleotide sequence and organization of the enteropathogenic Escherichia coli (EPEC) adherence factor (EAF) plasmid of EPEC strain B171 (O111:NM) were determined. The EAF plasmid encodes two known virulence-related operons, the bfp operon, which is composed of genes necessary for biosynthesis of bundle-forming pili, and the bfpTVW (perABC) operon, composed of regulatory genes required for bfp transcription and also for transcriptional activation of the eae gene in the LEE pathogenicity island on the EPEC chromosome. The 69-kb EAF plasmid, henceforth designated pB171, contains, besides the bfp and bfpTVW (perABC) operons, potential virulence-associated genes, plasmid replication and maintenance genes, and many insertion sequence elements. Of the newly identified open reading frames (ORFs), two which comprise a single operon had the potential to encode proteins with high similarity to a C-terminal region of ToxB whose coding sequence is located on pO157, a large plasmid harbored by enterohemorrhagic E. coli. Another ORF, located between the bfp and bfpTVW operons, showed high similarity with trcA, a bfpT-regulated chaperone-like protein gene of EPEC. Two sites were found to be putative replication regions: one similar to RepFIIA of p307 or F, and the other similar to RepFIB of R100 (NR1). In addition, we identified a third region that contains plasmid maintenance genes. Insertion elements were scattered throughout the plasmid, indicating the mosaic nature of the EAF plasmid and suggesting evolutionary events by which virulence genes may have been obtained.

Enteropathogenic Escherichia coli (EPEC), an important cause of diarrhea in the developing world (25), is characterized by infection of the small intestine and the presence of bacteria clusters attached to the epithelial surface. A similar adherence pattern—the localized adherence (LA) phenotype—is evident when EPEC is inoculated onto tissue culture cell monolayers (6). EPEC attachment induces epithelial cells to form pedestal-like structures beneath adherent bacteria and the loss of nearby microvilli; together, these features define the attaching and effacing phenotype. In vitro infection studies have shown that attached bacteria transduce signals into host cells via secretion of several EPEC effector molecules; these events are associated with cytoskeletal rearrangement and with the development of the attaching and effacing phenotype (10, 23).

Epidemiological studies of E. coli-associated diarrhea in children have shown that the LA phenotype is correlated with EPEC that harbors large plasmids estimated to range from 50 to 70 MDa, depending on serotype and strain; this family of related plasmids has been denoted EPEC adherence factor (EAF) plasmids (2). The importance of the EAF plasmids for EPEC virulence was demonstrated in a volunteer study in which a plasmid-cured EPEC strain was found to be significantly less pathogenic than the parental strain in orally challenged volunteers (26). EPEC strains cured of the EAF plasmid not only are less virulent but also do not exhibit the LA phenotype (22, 26). The distribution of the EAF plasmid among EPEC strains has been demonstrated through the use of a 1-kb DNA probe derived from the EAF plasmid of EPEC E2348/69 (O127:H6) (32, 33).

Besides being required for the LA phenotype, the EAF plasmid harbors a locus that encodes the bundle-forming pili (BFP) of the organism (44, 45). BFP are produced within adherent microcolonies of EPEC, where they form a meshwork of interbacterial fibers that appear to physically stabilize the attached colony (15). The bfp operon occupies a 12-kb region on the EAF plasmid and is composed of 14 genes including bfpA, which encodes the major pilus subunit; the 13 other genes (bfpB to bfpL) are required for BFP biogenesis and function (3, 9, 43–45). The bfp operon is a constant feature of LA phenotype-positive EPEC strains, and a probe derived from bfpA has been used in the classification of E. coli isolated during the course of epidemiological studies (14).

Located on a separate region of the EAF plasmid, the bfpTVW (perABC) operon encodes transcriptional activators for the bfpA-L operon (49); the bfpTVW (perABC) has also been reported to activate transcription of the eae gene (16), which is located on the chromosome and encodes the outer membrane protein, intimin, that is required for intimate adherence and actin condensation beneath attached bacteria (11, 20, 21). bfpT encodes a 30-kDa protein which belongs to the AraC transcriptional regulator family and binds to and transcriptionally activates the promoter region of bfpA (49). Like bfpA, a bfpT knockout mutant has been orally administered to volunteers and shown to be required for full EPEC virulence (3). Taken together, these studies demonstrate that the EAF plasmid not only harbors essential EPEC virulence determinants but may control the expression of chromosomally located genes as well.

Obtaining the complete DNA sequence of the EAF plasmid not only offers the opportunity to identify new potential virulence determinants but also may enable comparisons between the EAF genome and the genomes of other large virulence plasmids from closely and more distantly related biotypes and species. This comparative analysis has been facilitated by the recent publication of the complete sequences of the pO157 plasmid of enterohemorrhagic E. coli (EHEC) (5, 30) and of plasmids of Yersinia pestis (19, 27). Here we report the complete sequence and annotation of the EAF plasmid of EPEC B171, henceforth designated pB171.

MATERIALS AND METHODS

Bacterial strain and plasmid.

EPEC B171-8 (O111:NM) was used for isolation of the EAF plasmid (36). The EAF plasmid, pB171, was prepared from B171-8 grown overnight at 37°C in L broth and purified by using QIAGEN tip (QIAGEN Inc.).

Subcloning for sequencing.

Since digestion of pB171 with SalI gave two equal-sized fragments, a SalI DNA fragment harboring a functional replicon was isolated as follows. Purified pB171 was digested with SalI and ligated with a kanamycin resistance gene cassette (Pharmacia), and kanamycin-resistant transformants containing the 35-kb SalI fragment of pB171 were isolated. The resulting plasmid, pB171-S, contained a SalI fragment which did not include the bfp and bfpTVW operons. DNA libraries of pB171-S were prepared by random sharing of plasmid DNA; the resulting fragments were size selected and then cloned into plasmid pUC18. After amplification of inserted fragments by PCR, sequences from the ends of fragments were determined as described by Makino et al. (30) and then assembled into a single, continuous sequence. Alternatively, libraries of pB171 were also prepared by digestion of plasmid DNA with EcoRI or BglII and the resulting fragments were cloned into pMW119 (Nippongene). After construction of a restriction map of pB171, clones containing the DNA fragments corresponding to three remaining regions of pB171 were isolated from the libraries. These include one between the bfp and bfpTVW operons, a second downstream of the bfpTVW operon to the SalI site, and the third upstream of repI to another SalI site. Series of nested deletions were created from each clone, and DNA sequence was obtained from both ends.

DNA sequence analysis and annotation.

Two continuous sequences of pB171 were previously published by our group (43, 44, 49): a 14-kb sequence encompassing the bfp region (accession no. U27184) and a 3.9-kb sequence of the bfpTVW region (accession no. L42638). These sequences were combined with sequences determined in this study, and a single continuous circular sequence of pB171 was obtained. Open reading frames (ORFs) encoding products that were at least 50 amino acids (aa) in length were identified first; then possible ORFs were selected by a combinations of database matches and by the presence of a ribosome binding site. Operons were predicted from the arrangement of ORFs. Amino acid sequences were searched against the current, nonredundant protein database of the National Center for Biotechnology Information by using BLAST software through the Internet.

Nucleotide sequence accession number.

The annotated sequence was deposited in DDBJ/GenBank/EMBL under accession no. AB024946.

RESULTS AND DISCUSSION

General overview.

Nucleotide sequences from bp 1 to 14600, which contains repI, rsv, and the bfp operon, and from bp 20564 to 24480, which contains the bfpTVW operon and ORF5 (encodes a transposase-like protein), were previously published (43, 44, 49). The entire DNA sequence of pB171 consists of 68,817 bp which form a circular plasmid. The DNA sequences of three separate regions of another EAF plasmid, pMAR, which is harbored in a different EPEC serotype, O127:H6 strain E2348/69, were reported previously (16, 32, 45). The bfp operon sequence of pMAR (accession no. Z68186) showed 99.9% similarity with the corresponding sequence of the bfp operon of pB171, and the sequence of perABC region of pMAR (accession no. Z48561) showed 99.7% similarity with the bfpTVW operon region of pB171. The third published sequence fragment of pMAR (accession no. X76137) was used as a DNA probe for detection of EAF plasmids (32). This sequence was found to be similar to two separate loci of pB171, located downstream of the bfpTVW operon (from bp 24475 to 24780 and from bp 27489 to 28244). Although the similarity of the pB171 sequence to the EAF probe DNA sequence was 100%, the corresponding sequence in pB171 was separated by insertion of nonhomologous sequence in the middle of the homologous sequence. This inserted sequence was revealed to be a new insertion sequence (IS) element as described below.

In toto, analysis of the entire pB171 sequence predicted 80 ORFs (Fig. 1). ORFs were first selected by using DNASIS software (Hitachi Co.), and this selection was then refined as described in Materials and Methods. Of the 80 putative ORFs, 78 were predicted to encode proteins that were significantly homologous with previously described proteins; thus, only 2 ORFs had no regions of significant homology with proteins in the current database (Table 1).

FIG. 1.

Map of the entire pB171 plasmid. The outer circle shows ORFs, with their orientations denoted by their positions: outside the ring indicates clockwise, and inside the ring indicates counterclockwise. ORFs encoding previously documented virulence proteins are indicated by red boxes; ORFs encoding newly identified, putative virulence proteins are indicated by pink boxes. IS-associated ORFs are indicated by green boxes; ORFs encoding proteins related to replication and plasmid maintenance functions are indicated by yellow boxes. The inner circle shows IS elements (green) with the scale in kilobase pairs. Nomenclature of the ORFs is given in Table 1.

TABLE 1.

ORFs of pB171

ORF	Gene	Position (bp)a	Size (aa)	Homologue by BLAST	Identity/Similarity (%)	Accession no.
ORF1	repI	290–1267	325	RepI		U27184
ORF2	rsv	c1590–2126	178	Rsv		U27184
ORF3	bfpA	2714–3295	193	BfpA		U27184
ORF4	bfpG	3342–3743	133	BfpG		U27184
ORF5	bfpB	3750–5408	552	BfpB		U27184
ORF6	bfpC	5405–6613	402	BfpC		U27184
ORF7	bfpU	6615–7085	156	BfpU		U27184
ORF8	bfpD	7095–8699	534	BfpD		U27184
ORF9	bfpE	8706–9764	352	BfpE		U27184
ORF10	bfpF	9748–10743	331	BfpF		U27184
ORF11	bfpP	10750–11499	249	BfpP		U27184
ORF12	bfpH	11481–11927	148	BfpH		U27184
ORF13	bfpI	11948–12493	182	BfpI		U27184
ORF14	bfpJ	12490–13041	183	BfpJ		U27184
ORF15	bfpK	13025–13519	164	BfpK		U27184
ORF16	bfpL	13522–13971	149	BfpL		U27184
ORF17	bfpM	14259–14588	109	BfpM		U27184
ORF18		c15260–15481	73	ORF1 of stability locus of R100 (NR1)	78/89	P11907
ORF19		15526–15810	94	Putative resolvase of Y. pestis pMT1	44/71	Q58416
ORF20		16285–16878	197	Serine acetyltransferase	39/57	S71181
ORF21	trcP	c18216–18659	147	TrcA	50/67	AB010764
ORF22		c19071–19937	288	Transposase of IS3	99/99	U73857
ORF23		c19934–20233	99	Hypothetical protein	100/100	U73857
ORF24	bfpT	20757–21581	274	BfpT		L42638
ORF25	bfpV	21629–22018	129	BfpV		L42638
ORF26	bfpW	22068–22337	89	BfpW		L42638
ORF27		c22914–23366	150	Transposase of IS285		L42638
ORF28		23900–24460	186	Orf5, transposase of IS10 (C half)	36/61	L42638
ORF29		24865–25515	216	L0013	43/55	AF071034
ORF30		25515–25862	116	L0014	76/85	AF071034
ORF31		25882–27453	527	L0015	57/70	AF071034
ORF32		29084–29425	113	Transposase of IS91	94/94 (truncated)	S23782
ORF33		29581–30087	168	Transposase of IS91	92/93 (truncated)	S23782
ORF34		30507–31371	288	Transposase of IS3	99/99	U73857
ORF35		31771–33126	451	ToxB of EHEC	96/97 (truncated)	AB011549
ORF36		33144–33650	168	ToxB of EHEC	97/98 (truncated)	AB011549
ORF37		33740–33934	64	Transposase of IS3	63/66 (truncated)	AF041810
ORF38		33972–34199	75	Transposase of IS3	86/86 (truncated)	AF041810
ORF39		c34224–34637	137	IstB of IS21	92/94 (truncated)	P15026
ORF40		c34650–35198	182	IstA of IS21	97/97 (truncated)	P15026
ORF41		35255–36277	340	Transposase of IS100	98/98	U59875
ORF42		36277–36618	113	Transposase of IS100	100/100 (truncated)	U59875
ORF43		c36749–37036	95	Hypothetical RelE-like protein	87/87	Q60228
ORF44		c37033–37284	83	ORF7 of IncF plasmid	95/97	M26308
ORF45	repA1	c38249–39106	285	Replication protein (RepA1)	83/84	M26937
ORF46	copB	c39420–39668	82	CopB	98/98	S13675
ORF47	snrB	c39952–40158	68	SnrB	97/98	P13970
ORF48		c40866–41129	87	Hypothetical protein	89/92	Q99342
ORF49		41400–42050	216	L0013	43/55	AF071034
ORF50		42050–42397	115	L0014	76/85	AF071034
ORF51		42417–43988	523	L0015	57/70	AF071034
ORF52		c44385–44804	139	Unnamed protein of E. coli	76/83	AB011549
ORF53	klcA	c44851–45276	141	Antirestriction protein KlcA	85/87	AB011549
ORF54	piv	45846–46673	275	Pilin gene inverting Protein (PivML)	35/50	P20665
ORF55		c46912–48062	383	IS30 transposase	95/95	U70214
ORF56		48194–48619	141	Transposase of IS911	97/99 (truncated)	AE000133
ORF57		c48486–48917	143	Hypothetical protein	95/95	U70214
ORF58		c48832–49140	102	Hypothetical protein	100/100 (truncated)	U70214
ORF59		49226–50734	502	Reverse transcriptase-like protein	99/99	D37918
ORF60		50887–51303	138	ORFB of IS911	88/94	AF074613
ORF61		51726–52211	161	Glutamate decarboxylase	53/68 (truncated)	D90903
ORF62		52462–53922	486	Amino acid antiporter	28/48	AE001295
ORF63		53944–54765	273	Glutamate racemase	33/53	P52972
ORF64		c54884–55279	131	InsB protein of IS1	93/96	P03832
ORF65		c55306–55578	90	InsA protein of IS1	93/95	P03829
ORF66	impB	56128–56814	228	ImpB of S. flexneri	97/97	AF079316
ORF67	stbB	c57097–57489	130	StbB protein of R100 (NR1)	26/48	P11906
ORF	Gene	Position (bp)a	Size (aa)	Homologue by BLAST	Identity/Similarity (%)	Accession no.
ORF68	stbA	c57494–58465	323	StbA protein of R100 (NR1)	43/65	P11904
ORF69		58694–59338	214	ParA protein of H. pylori	27/53	AE000608
ORF70		59332–59607	91
ORF71	rsvB	c59745–60554	269	Resolvase of F plasmid	64/65	P06615
ORF72	ccdB	c60555–60881	108	CcdB	99/99	P05703
ORF73	ccdA	c60862–61080	72	CcdA	97/97	P05702
ORF74		61793–64414	873
ORF75		64613–64843	76	VagC of S. dublin virulence plasmid	90/93	S22685
ORF76		64855–65190	111	VagD of S. dublin virulence plasmid	88/92 (truncated)	S22686
ORF77		65488–66510	340	Transposase of IS100	99/99	U59875
ORF78		66507–67285	260	Transposase of IS100	100/100	U59875
ORF79		c67389–67919	176	hypothetical protein of Rhizobium plasmid	29/50	Z95210
ORF80		c67922–68210	95	hypothetical protein of Rhizobium plasmid	32/43	P55365

^aA position begin with “c” represents an ORF in the minus strand. 

New potential virulence genes.

The amino acid sequence deduced from ORF21, found 4.2 kb downstream of the bfp operon and 2.1 kb upstream of the bfpTVW operon, was revealed to have high homology (50% identity and 67% similarity) with the TrcA protein, which was recently found on a novel chromosomal pathogenicity island (designated LIM) of EPEC B171-8; TrcA has been shown to enhance BFP production (50). Although the ORF21 product (Orf21; 147 aa) is smaller than of TrcA protein, the amino acid sequence of Orf21 showed high homology with TrcA along its entire length (Fig. 2). Other homologues of TrcA have been found; these include Orf19 within the first-described EPEC pathogenicity island, LEE (12), and IpgB on the Shigella virulence plasmid (1, 31, 41). As expected, Orf21 of pB171 also showed homology with the same proteins: 30% identity and 52% similarity with Orf19 of LEE and 28% identity and 46% similarity with IpgB of the Shigella virulence plasmid. Functional studies of TrcA have shown it to be a chaperone-like protein which directly interacts with, and enhances the accumulation of, BfpA (50). Since Orf21 shows high homology to TrcA, it may have the same role in BFP biosynthesis as TrcA. Based on its high homology with trcA, ORF21 was denoted trcP (trcA homologue on the plasmid).

FIG. 2.

Alignment of TrcP (Orf21) with TrcA and related proteins. The predicted amino acid sequence encoded by ORF21 (trcP) was aligned by using CLUSTAL V (18) with the predicted amino acid sequences of TrcA (50), Orf19 in LEE (12), and IpgB of S. flexneri (1, 31, 41). Identical residues in all proteins are indicated by asterisks, and conserved residues are indicated by dots. Identical residues in TrcA, Orf19, and IpgB that correspond to residues in TrcP are indicated with shaded boxes.

ORF35 and ORF36 were found to be highly homologous with toxB on the large plasmid of EHEC O157:H7. The DNA sequence from positions 31673 to 33744, which includes ORF35 and ORF36, is almost identical (97%) to the corresponding sequence of the toxB region of EHEC O157:H7 (from nucleotide positions 63502 to 65581 of the AB011549 sequence). However, owing to an apparent deletion in this region of pB171 compared to the EHEC plasmid, the combined ORF35 and ORF36 sequences would encode only the C-terminal 20% of the 3,169-aa protein predicted to be encoded by toxB (97% identity) (Fig. 3). Specifically, ORF35 corresponds to aa acids 2543 to 2990 of ToxB (97% identity) and ORF36 corresponds to aa 3002 to 3169, the C terminus of ToxB (97% identity). ORF35 was found to encode a predicted N-terminal signal sequence, suggesting that the ORF35 product is exported across the cytoplasmic membrane.

FIG. 3.

Schematic representation of homologous sequences in the pB171 ORF35–ORF36 region and in the pO157 toxB region. First and third lines indicate DNA segments of pB171 and pO157, respectively. The scales (positions) are indicated above the lines according to the sequence of pB171 (this study) and pO157 (accession no. AB011549). Homologous sequences are indicated by closed boxes. These regions are surrounded by IS elements (hatched boxes). Deduced proteins encoded by these regions are indicated by arrows under the coding sequence. From the sequence of pB171 DNA, two separate protein products, Orf35 and Orf36, are predicted, while only a single, large product, 3,169 aa in length, is predicted to be encoded by the toxB region of pO157. The amino acid sequences of Orf35 and Orf36 correspond to different segments in the C-terminal part of ToxB (indicated by the closed box or arrows together with their positions in the amino acid sequence).

Other putative proteins.

Besides ORFs which showed homology with gene products of IS elements or those involved in plasmid replication or maintenance (described below), the predicted products of four ORFs showed homology to known proteins. Orf20 showed 39% identity (57% similarity) over a 117-aa span with a serine acetyltransferase of Arabidopsis thaliana and lower similarity with the same homologue of other organisms, including E. coli and Salmonella typhimurium. Serine acetyltransferase is a key enzyme in the l-cysteine biosynthetic pathway of sulfate-assimilating organisms. Orf54 shows high similarity with the pilin gene-inverting protein (Piv) of Moraxella spp., where it functions in the site-specific DNA inversion of a chromosomal segment to switch the expression from one type IV pilin gene to its alternate in Moraxella bovis, and M. lacunata. The amino acid sequence of Piv is homologous with IS transposase (24); hence, Piv is believed to possess DNA recombinase activity. Although there is no report on phase conversion of BFP expression in EPEC, it is possible that phase conversion could silence bfp expression under conditions of growth that have not yet been tested.

ORF62 and ORF63, which appear to constitute a single operon, are homologous with a family of amino acid antiporter protein genes and with a family of glutamate racemase genes, respectively. The amino acid antiporter gene (gadC) has been shown to be necessary for glutamate-dependent acid resistance in E. coli, Shigella flexneri, and Lactococcus lactis (17, 40, 52). The genes for these antiporters were linked to genes encoding glutamate decarboxylase (gadB). Indeed, ORF61, which is located just upstream of ORF62, has homology to gadB, though ORF61 coding sequence corresponds to only one-fourth of GadB protein (accession no. M84025). Since Orf62 and GadC protein are similar in size and exhibit strong homology throughout their entire amino acid sequences, the ORF62 product could play the same role in glutamate-dependent acid resistance in EPEC. The predicted protein product of ORF63 is a homologue of a glutamate racemase which participates in the biosynthesis of d-glutamate, an essential component of the bacterial peptideglycan.

IS elements.

pB171 contains a variety of IS elements (Table 2), and in toto 29.5% of the plasmid sequence is occupied by whole or partial IS elements of many kinds. Five known intact IS elements were found: two copies of IS3 (47), one copy of IS30 (7), one copy of IS100 (35), and one copy of IS1 (νζ) (34). Other IS elements appear to be defective in transposition capacity since these sequences are fragmented due to truncation of one or both ends of the element or deletion of internal sequence or insertion of other IS elements. The distribution of many of the truncated IS elements on pB171 suggested that extensive rearrangements have occurred during the evolution of this plasmid.

TABLE 2.

IS elements and a group II intron in EPEC EAF plasmid pB171

IS element or intron	Homologya (%)	Orientationb	Coordinates to:
			EAF plasmid	Known IS
IS3 (A)	99.8	−	19041–20298	1–1258 (end)
IS21 (A)	85.7	−	20317–20344	1–28
IS1(νζ) (A)	83.2	+	20345–20499	612–766 (end)
IS630 (A)	78.0	+	22360–22527	1–178
IS285 (A)	75.7	−	22528–22667	1178–1318 (end)
IS285 (B)	73.1	−	22668–23438	1–771
IS10 (A)	52.1	+	23735–24485	579–1329 (end)
IS1491 (A1)	53.6	+	24486–24779	1882–2175
IS679 (A)	100	+	24780–27483	1–2704 (end)
IS1491 (A2)	62.5	+	27484–27743	2173–2431
IS630 (B)	83.8	−	27847–27884	1025–1062
IS630 (C)	86.1	+	27885–27969	1068–1153 (end)
IS629 (A)	97.3	−	28651–28687	1274–1310 (end)
IS91 (A)	74.3	+	28689–30144	1–1576
IS3 (B)	99.8	+	30146–31403	1–1258 (end)
IS3 (C)	82.4	+	33707–34132	329–774
IS21 (B)	95.2	−	34199–34471	1716–1983
IS21 (C)	97.3	−	34463–35168	744–1451
IS100 (A)	99.3	+	35169–36544	1–1376
IS679 (B)	99.6	+	41315–44018	1–2704 (end)
IS911 (A1)	100	+	46570–46904	1–335
IS30 (A)	99.6	−	46905–48125	1–1221 (end)
IS911 (A2)	99.5	+	48126–48554	334–762
EPEC.IntA	99.7	+	48555–50824	1–2272 (end)
IS911 (A3)	97.6	+	50825–51311	763–1250 (end)
IS1(νζ) (B)	79.8	+	51315–51502	579–766 (end)
IS1(νζ) (C)	91.6	−	54870–55635	1–766 (end)
IS100 (B)	99.4	+	65402–67355	1–1953 (end)

^aHomology with the published nucleotide sequence. See Table 3 for homology of the new IS element IS679, with related IS elements. 

^b+, clockwise; −, counterclockwise. 

In addition to these previously established IS elements, two copies of a potential IS element, located between bp 24780 and 27483 and between bp 41315 and 44018, were found. The inverted repeat sequence, 5′-GTAAGCGNNTCNNNNAACCGTNTT-3′, was found at both ends of these sequences; GGATGATC was found in the first segment of the repeated sequence and could be a target sequence of the insertion. The sequence of this potential IS element showed weak homology with IS66 (29), IS866 (4), and IS1131 (51) of Agrobacterium tumefaciens (Table 3). Moreover, sequence located at both ends of the element on the pB171 plasmid showed high similarity to each of the three IS elements: 5′-GTAAGCCNNCGGTGAAGGCC-3′ of IS66, 5′-GTATGCGNCGNCTCCNTCCCATTGATT-3′ of IS866, and 5′-GTGAGCGTCCGGNNANCNNTTT-3′ of IS1131. Thus, although similar to previously described IS elements, this potential IS element on pB171 seemed to be sufficiently unique to merit a new designation: IS679.

TABLE 3.

IS679 homologous to IS elements from A. tumefaciens^a

IS element	Size (bp)			Source	GenBank accession no.	Reference
	Total	TIR	TSD
IS66	2,548	20	8	A. tumefaciens	M10204	29
IS866	2,716	27	8	A. tumefaciens	M25805	4
IS1131	2,773	12	8	A. tumefaciens	M82888	51
IS679B	2,704	25	8	E. coli		This work

^aTIR, terminal inverted repeat. TSD, target site duplication. 

A 2,270-bp stretch of sequence lying within IS911, from bp 48555 to 58024, exhibited 99.7% similarity to an S. flexneri chromosomal group II intron-like sequence, Sf.IntA (37), which was found in the she pathogenicity island. In accord with its designation in Shigella, the sequence on pB171 was designated EPEC.IntA. As previously shown for Sf.IntA, EPEC.IntA contained an ORF which encodes a putative protein with significant homology to a family of reverse transcriptase-like proteins that are encoded within the introns of fungi (X55026, U41288, and X57546), plants (M68929), and bacteria (U77945, U50902, X98606, and X71404). The group II intron-like sequence was also found in the E. coli K-12 chromosome (AE00013) and in plasmid pMT1 of Y. pestis (AF074611 and AF053947). However, the ORFs encoding the reverse transcriptase-like proteins in these sequences were truncated. The identification of these group II intron-like sequences in EPEC, Shigella, and Yersinia raises the possibility that they are involved in the transfer of virulence determinants between different bacterial strains and species.

Replication and plasmid maintenance.

DNA sequence analysis revealed three potential plasmid replication regions: one resembles RepFIIA, another resembles RepFIB, and the third contains genes likely to be involved with plasmid maintenance. The first replication region exhibited 93% nucleotide sequence identity with RepFIIA of the IncFII plasmid R100 (NR1) (8); this region contains a complete repA1 gene (39). The same sequence showed 93% similarity to the RepFIIA replication region of plasmid pO157 of EHEC O157:H7 (5, 30) (Fig. 4A). A locus in pB171 that corresponds to the copB sequence, a gene which is involved with the control of plasmid copy number, was identified; however, the homologue in pB171 showed only low similarity with the corresponding genes of the R100 (NR1) or pO157. In contrast, it showed very high (97%) similarity with the copB gene of the IncFVII plasmid pSU233 (28). The similarity with the pSU233 copB sequence extended to downstream sequences containing a repA1 promoter, suggesting that the plasmid copy number control system of pB171 is closely related to the system to IncFVII of pSU233. In turn, this indicates that the RepFII replicon of pB171 is most likely a composite of the RepFII-related regions of IncFII and IncFVII plasmids. The ability of this sequence to function as a replicon was confirmed by the construction of an autonomously replicating plasmid that was prepared by ligating a kanamycin resistance gene to an ApaI-KpnI 2.3-kb fragment which contains the repA1 and copB genes (48).

FIG. 4.

Primary structures of RepFIIA and the ccdAB regions. (A) Comparison of the RepFIIA region of pB171 and the corresponding regions of the R100 (NR1), F, and pSU233 plasmids. ORFs are indicated by arrows under the DNA of the pB171 plasmid. Nucleotide positions are indicated on the line. Homologous sequences of NR1 (accession no. X12776), F (M12987), and pSU233 (X55893) are indicated by lines under arrows, showing homologous regions in nucleotide position in each sequence with percent identity. Note that only the sequence of pB171 corresponding to the copB region is replaced by pSU233. (B) Schematic representation of the primary structure of the plasmid maintenance region. ORFs are indicated by arrows under the DNA of pB171. The nucleotide sequence of the ccdAB region showed high homology with that of the F plasmid (M12987) and the pO157 plasmid of E. coli O157:H7 (AB011549). The impB region showed high homology with the PT110 plasmid of Salmonella typhimurium (X53528) and with the SA100 plasmid of S. flexneri (AF079316).

The second replication region was found to be 83% identical to the RepFIB replication region of enterotoxigenic plasmid p307 (42) and 89% identical with plasmid pFM82139 of Salmonella enteritidis (38). The RepFIB replication region of S. enteritidis has been proven to be a functional replicon (38), and the corresponding region of pB171 also contains all of the essential features of a replicon, including the repeating sequence B through J, with 5′-ANATAAGCTGTAGNNNGNAAA-3′ (44). In addition to these features, a DnaA binding sequence was found in the upstream region, from bp 68704 to 68712. Taken together, these features suggest that the second replicon may also be a functional replication origin.

The third replication region contains genes necessary for stable maintenance of the plasmid (Fig. 4B). Three ORFs with high similarity to the ccdA, ccdB, and rsv (protein D) genes in the RepFIA region of the F plasmid (accession no. M12987) were found. The same region was found on the pO157 plasmid of EHEC O157:H7 (5, 30). The sequence downstream of this region was not homologous with a downstream region of the F plasmid, which encodes repE and sopAB genes. On pB171, the corresponding region did not contain a repE homologue but instead contained an ORF exhibiting homology to parA (function unknown) of Helicobacter pylori (AE000608). Two ORFs which flank this region, ORF67 and ORF68, would encode proteins homologous with the predicted protein products of stbA and stbB, within the stability locus of the R100 (NR1) plasmid (46) or the R1 plasmid (13). Although these amino acid sequence homologies indicate that Orf68 and Orf67 could be functional homologues of the stbA and stbB gene products, the nucleotide sequence of this region showed low identity with the stb region of the R100 (NR1) plasmid. This suggested that the incompatibility group of pB171 is likely to be different from R100 (NR1). ORF66, which is downstream of the stbAB operon, has high nucleotide sequence homology with impB on virulence plasmids of S. flexneri or Salmonella typhimurium. impB is the last gene in the impCAB operon in Shigella and Salmonella, whereas only the impB homologue was found in this region of pB171. The impCAB operon of Salmonella typhimurium has been shown to be involved in UV protection and mutation.

In addition to these loci, nucleotide sequence from 64613 to 65165 showed high similarity (90%) to the vagCD region of a virulence plasmid of Salmonella dublin, where it is believed to be involved in plasmid maintenance. In this region of pB171, ORF75 and ORF76 would correspond with vagCD, but ORF76, which has homology to vagD, has apparently been frameshifted by a one-base deletion in the middle of the coding sequence and thus would encode a protein with VagD homology only in the first 25 aa out of total length of 111 aa.

Base distribution.

The mosaic nature of pB171 suggests that the development of the EAF plasmid has occurred through the acquisition of DNA elements and parts of DNA elements from various sources. This hypothesis is further supported by the results of a base composition analysis of the plasmid. Although average G+C content of the plasmid was 46%, regions of the EAF plasmid containing the bfp operon (38.4%) and the bfpTVW virulence regulatory operon (29.9%) were significantly different in G+C content from the surrounding regions of DNA (Fig. 5). At least three other regions of the plasmid also were found to be lower in G+C content than surrounding regions: kbp 17.5 to 19, which is an intervening region between the bfp operon and the bfpTVW operon; kbp 31.5 to 33.7, a region which contains ORF35 and ORF36, showing high homology to toxB ORF of EHEC plasmid pO157; and kbp 51.5 to 54.8, which contained ORF61, ORF62, and ORF63, exhibiting high homology to the glutamate decarboxylase, amino acid antiporter, and glutamate racemase genes. All of these low-G+C regions as well as the bfpTVW operon were surrounded by intact or fragmented IS elements, suggesting that the plasmid acquired these regions through the horizontal transfer of a mobile element.

FIG. 5.

Base composition of pB171. The plots showing G+C content were derived by using the DNASIS program, which also shows selected ORFs by hatched boxes to the correct scale. The scale on each boxes indicates its position in the plasmid in kilobase pairs.