PCR-Based Identification of Common Colonization Factor Antigens of Enterotoxigenic Escherichia coli

Abstract

Colonization factor antigens (CFAs) of enterotoxigenic Escherichia coli (ETEC) have been classified into several groups based on their distinct antigenicity. We describe here a PCR-based method to detect common CFAs of ETEC, which were characterized using conventional serology. This PCR assay is simple and sensitive for the detection of expressed CFA genes.

Enterotoxigenic Escherichia coli (ETEC) infection is a major cause of infantile and traveler's diarrhea, causing substantial effects in terms of morbidity and economic consequences (1, 16, 17). Apart from heat-labile and heat-stable enterotoxins, ETEC pathogenesis is mediated by specific antigenically distinct colonization factor antigens (CFAs). Several CFAs have been identified; among these, CFA/I, CFA/II, and CFA/IV are the most common (5). CFA/II is composed of coli surface (CS) antigens CS1, CS2 and CS3, whereas CFA/IV is composed of CS4, CS5, and CS6 (10, 11, 13). In addition, several other CFAs have been identified, namely, CS12, CS14, CS17, and CS21.

CFA typing of ETEC strains is done by serology using specific monoclonal antibodies raised against these antigens. Such typing is limited to the reference laboratories and could not be used in the routine surveillance. In this study, we have developed a simple PCR-based method for the detection of CFAs and evaluated the specificity and sensitivity of the method by comparing the results with the conventional serology-based methods.

Thirty-three ETEC strains with established CFA types were included in this study. These strains were isolated from diarrheal patients in different parts of India, and the CFA typing was done at the International Centre for Diarrheal Diseases Research, Bangladesh (14). Diarrheagenic E. coli other than ETEC, along with the K-12 derivative strain DH5α, served as negative controls. All the strains were maintained at −70°C in Luria broth containing 15% glycerol. The cultures were grown on CFA agar plates (4) at 37°C for PCR experiments.

Nucleotide sequences from different CFAs were aligned to identify different CFA specific regions and were utilized to develop CFA-specific PCR primers (Table 1). Boiled bacterial cell lysate in water served as the genomic DNA template for the PCR assay. Total RNA was prepared from the ETEC strains grown on CFA agar plates (4), and the cDNA was prepared by a reverse transcription (RT) reaction as described previously (2). The PCR products were electrophoresed onto 2% agarose gels stained with ethidium bromide and digitally recorded (UVP Biospectrum AC Imaging Systems). To check the PCR specificity, amplicons were sequenced using an automated DNA sequencer (ABI Prism 3200; Applied Biosystems). PCR results revealed that most of the ETEC strains harbored genes encoding CS6 alone or in combination with other CFA types (Fig. 1 and Table 2). This prompted us to further screen the strains with an anti-CS6 antibody to correlate the phenotypes with reverse transcription-PCR (RT-PCR) results. For this, we procured an anti-CS6 antibody commercially (GenScript Corp., Piscataway, NJ), which was developed against the most antigenic region of the CssA subunit of CS6. The most antigenic region was predicted by hydropathy plot (Kyte-Doolittle plot) using the amino acid sequence of CS6 (GenBank accession no. U04844). The region from amino acid residues 131 to 144 of the CssA subunit of CS6 (YTSGDKEIPPGIYN) was selected to raise the antibody in rabbits.

TABLE 1.

Primers used in this study^a

Target antigen	Primer	Primer sequence (5′ to 3′)	Expected amplicon size (bp)	GenBank accession no.
CFA/I	1F	GCTCTGACCACAATGTTGA	364	S73191
	1R	TTACACCGGATGCAGAATA
CS1	2F	TTGACCTTCTGCAATCTGA	324	M58550
	2R	CATCTGCATGGATTGAAAG
CS2	3F	TAACTGCTAGCGTTGATCC	368	Z47800
	3R	ATTAGTTTGCTGGGTGCTT
CS3	4F	GGTGGGTGTTTTGACTCTT	264	M35657
	4R	TGTTCGTTACCTTCAGTGG
CFA/III	5F	GCCTTCTGGAAGTCATCAT	437	D37957
	5R	TGCCACATACTCCCAGTTA
CS4	6F	TTTTGCAAGCTGATGGTAG	250	X97493
	6R	TCTGCAGGTTCAAAAGTCA
CS5	7F	CGGATTGGATATACCGTTT	453	X63411
	7R	TCAACAGCAAATGTTACCG
CS6	8F	ATCCAGCCTTCTTTTGGTA	321	U04844
	8R	ACCAACCATAACCTGATCG
CS14	9F	ACTGTGACAGCCAGTGTTG	394	X97491
	9R	AAACGACGCCTTGATACTT
CS17	10F	AACCTATTCTTCGGCTTCA	290	X97495
	10R	GCGCAGTTCCTTGTGTG

^aThe primers were synthesized by Isogen Biosciences BV, Maarssen, The Netherlands, and the sequences were tested for degeneracy and cross-reactivity with other genes and CFAs of ETEC using the BLAST program. The PCR conditions were as follows: hot start at 95°C for 3 min, followed by denaturation at 95°C for 1 min, annealing at 54°C for 45 s, and extension at 72°C for 1.5 min. A 5-min final extension at 72°C was performed after 30 cycles of PCR.

FIG. 1.

PCR results obtained with representative ETEC strains. Specific CFAs based on the amplicon size on the agarose gel are shown: lane 1, CFA/I; lane 2, CS1; lane 3, CS2; lane 4, CS3; lane 5, CFA/III; lane 6, CS4; lane 7, CS5; lane 8, CS6; lane 9, CS14; and lane 10, CS17. M, 100-bp size marker. Representative strains used are shown here with respective strain numbers on the right side of the panel. The top eight panels are from RT-PCR analyses; the bottom panel (strain 933) is from a genomic PCR analysis.

TABLE 2.

PCR assay and serological results with ETEC and non-ETEC strains

ETEC straina	Antigen(s) detected by:
	Serologyb	Genomic DNA	RT-PCR	Dot blot ELISA
1571	CS6	CS5, CS6	CS6	+
2654	CS2, CS3	CS2, CS3, CS6, CFA/III	CS2, CS3	-
3539	CS17	CS6, CS17	CS17	-
4660	CS4, CS6	CS4, CS6	CS4, CS6	+
6547	CS3	CFA/I, CFA/III, CS3, CS5, CS6	CS3	-
169	CS5, CS6	CS5, CS6	CS5, CS6	+
4266	CS5, CS6	CS6	CS6	+
2812	CS4, CS6	CS4, CS6	CS4, CS6	+
2924	CS5, CS6	CS5, CS6	CS5, CS6	-
3019	CS4, CS6	CS4, CS6	CS4, CS6	-
5349	CS5, CS6	CS5, CS6	CS6	+
3020	CS6	CS6	CS6	+
6300	CS5, CS6	CS5, CS6	CS5, CS6	-
3023	CS4, CS6	CS4, CS6	CS4, CS6	+
5348	CS5, CS6	CS5, CS6	CS5, CS6	+
881	CS6	CS6, CFA/I	CS6	+
3401	CS1, CS3	CS1, CS3	CS1, CS3	-
1453	CS1, CS3	CS1, CFA/III	CS1, CFA/III	-
4407	CS3	CS3	CS3	-
5607	CS5, CS6	CS5, CS6	CS5, CS6	+
5065	UT	CS5, CS6	CS5, CS6	+
831	UT	CS6	CS6	-
933	UT	CFA/I, CFA/III, CS6	CS6	-
4147	UT	CS6	CS6	+
4552	UT	CS6	CS6	+
507	UT	CS6, CS14, CFA/III	CS6, CS14, CFA/III	+
4582	UT	CS6	CS6	+
E-3	UT	CS5, CS6, CS3, CFA/III	CS5, CS6	+
E-7	UT	CS4	None	-
E-9	UT	CS3, CS4, CS5, CS14	CS3, CS4, CS5	-
E-13	UT	CS4, CS5, CS6	None	-
E-16	UT	CS4, CS5, CS6	CS4, CS5, CS6	+
E-17	UT	CS3, CS5, CS6	CS3, CS5, CS6	+

^aFor the non-ETEC strains tested—EPEC 11044, EAEC 042, STEC VT3, E. coli BL21, and E. coli DH5α—no CS antigens were detected.

^bConfirmed at the International Centre for Diarrheal Diseases Research, Bangladesh. UT, untypeable ETEC strains by serology.

To check its recognition specificity, this antibody was evaluated by dot blot enzyme-linked immunosorbent assay (ELISA) using heat-saline extract (HS) prepared from strains that gave CS6-specific amplicons in the RT-PCR assay (11). In the dot blot assay, equal concentrations of HS extracts, measured by the Bradford method, were blotted onto a nitrocellulose membrane (Bio-Rad). CS6 was detected by anti-CssA antibody, followed by alkaline phosphatase-conjugated secondary antibody, and developed according to a standard protocol. As a negative control, HS extract from six randomly chosen CS6-positive ETEC strains grown at 17°C was used, since CS6 was not expressed at this temperature (8, 11).

Initially, 20 serologically CFA-typed ETEC strains were used to validate the data generated by the newly developed PCR-based CFA detection assay. Comparative analysis revealed that five of them (1571, 2654, 3539, 6547, and 881) gave amplicons for additional CFAs, along with those detected by serological assay. Such an observation can be explained by the fact that the expression of CFAs is dependent on the culture condition (3), and PCR-based detection using genomic DNA depends on the presence of the gene per se but not its functional expression in vitro. Therefore, the analysis was extended by repeating the PCR assay using RNA converted to cDNA instead of genomic DNA since it relied on the expressed CFA genes, which might mimic the serological CFA typing. In this study, RT-PCR and the serological typing results matched perfectly for 17 of the 20 strains, except for strains 4266, 1453, and 5349 (Table 2). For strains 4266 and 5349, CS5 was detected by serology but not by RT-PCR. Strain 4266 was also negative for CS5 in the genomic PCR. Similarly, strain 1453 showed the presence of CS1 and CFA/III by PCR instead of CS3 as detected by serology. Screening of CFA types by serological assay has been reported to give false-positive results in certain cases (5), and this may be true in the case of two ETEC strains (4266 and 1453). Strain 5349, though produced CS6-specific amplicon along with CS5 in the genomic PCR assay, failed to generate CS5-specific amplicon in the RT-PCR assay. The authenticity of the amplicons was confirmed by DNA sequencing (data not shown). These data indicate that the tested culture condition may not be favorable for the expression of CS5 by the strain 5349 and raise the possibility that such differential expression of CFA in vitro may be strain specific.

In order to expand the scope of this study, 13 clinical ETEC strains identified as untypeable by serological CFA typing were analyzed. CFA encoding genes were detected in all of them, but two strains (E-7 and E-13) did not express CFA genes as detected by RT-PCR. This can be attributed either to antibody specificity (15) or to the lack of corresponding CFA transcripts under the tested laboratory condition. The dot blot ELISA results obtained with anti-CssA antiserum are presented in Table 2. We found that of 28 strains positive for the CS6-encoding gene, 24, 14, and 19 were positive by RT-PCR, serology, and dot blot assay, respectively. We assume that different culture conditions could give a better correlation with the phenotypic assay. Ten of the ETEC strains were untypeable by serology, and six were untypeable by dot blot, but all were determined to be CS6 positive by RT-PCR. Of 33 ETEC strains analyzed, 28 were positive by RT-PCR, whereas 20 were detected by the available antibody-based method. Two strains remained untypeable by both methods. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of RT-PCR compared to the serological results were 85, 15.4, 60.7, and 40%, respectively (Table 3). The likelihood ratio suggested that the RT-PCR positive result is almost onefold more likely to occur than the serology. The RT-PCR-based detection of CFAs was also statistically more sensitive (P < 0.001, MacNemer chi-square test) than serology. Thus, the PCR-based detection of CFAs in ETEC strains is more sensitive and as specific as the existing serology method. A comparative analysis of RT-PCR and serology results for 28 strains that were positive for CS6-encoding gene was also performed. The sensitivity, specificity, PPV, and NPV were 100, 28.6, 58.3, and 100%, respectively (Table 3). The newly developed antibody in the dot blot assay gave slightly better specificities and PPV values compared to serology (Table 3).

TABLE 3.

Comparison between RT-PCR and antibody-based detection

Methods compareda	% Sensitivity	% Specificity	PPV (%)	NPV (%)	LRb
Serology vs RT-PCR (n = 33)	>85.0	15.4	16.7	40	1.0
Serology vs RT-PCR (n = 28)*	100	28.6	58.3	100	1.4
Dot blot vs RT-PCR (n = 28)*	100	37.5	79.2	100	1.6

^a*, strains were positive for CS6-encoding gene. n, number of strains.

^bLR, likelihood ratio.

In conclusion, it appears that an antibody typing method alone may not be sufficient to classify ETEC strains because many factors, such as the antigenic variation of colonization factor antigens (18), the differential expression of CFAs (3, 6, 12), and gene silencing due to transcriptional control, may influence the conventional assay results. Considering these factors, the PCR assay seems to be simple and perform well with several advantages compared to the conventional methods for the detection of CFAs. Moreover, the RT-PCR method developed here is based on CFA gene expression under established culture conditions. The results presented here suggest that this method is as sensitive as serological assays and very handy when an isolate is untypeable with the available antibodies. The DNA hybridization assay reveals the existence of genes in the genome but does not give any information about phenotypes (7). The other disadvantage of using a probe hybridization assay is the cross-reactivity among the genes encoding CFAs. Antibody-based typing in slide agglutination and detection of antibody in patients' immune systems also has not proved very successful due to cross-reactivity and sequence variations (9, 15, 18). Our PCR assays can detect the presence of common CFA genes, as well as their expression in the same strain. To expedite the identification of ETEC CFAs and its epidemiology, a multiplex PCR system can be developed in the future with specific combinations of primers and PCR conditions.