Use of the tna Operon as a New Molecular Target for Escherichia coli Detection

Abstract

A quantitative real-time PCR targeting the tnaA gene was studied to detect Escherichia coli and distinguish E. coli from Shigella spp. These microorganisms revealed high similarity in the molecular organization of the tna operon.

Council Directives 2006/7/EC (1) and 98/83/EC (2) on water quality and Commission Regulation 2073/2005 (7) on microbiological criteria for foodstuffs require the determination of Escherichia coli to assure hygiene control and consumer safety. Nowadays, gene probe technology provides rapid and highly sensitive techniques for the specific detection of pathogenic microorganisms (6, 8, 11, 15). E. coli is able to use tryptophan as a carbon source, and this phenotypic trait is normally employed as a test for identification in conventional culture-based techniques. The tryptophanase operon (tna) is a 3,049-bp region consisting of two major structural genes: tnaA (1,431 bp), coding for the tryptophanase that catalyzes the degradation of l-tryptophan to indole, pyruvate, and ammonia; and tnaB (1,248 bp), coding for a tryptophan permease (21). tnaA is preceded by a transcribed regulatory leader region containing a short open reading frame, tnaL, specifying a 25-residue leader peptide. In between there is a 205-bp spacer region that contains several transcription pause sites. Moreover, in silico analysis of complete genomes of eight E. coli strains available in GenBank confirmed that the tna operon is present in single copy, ranking it as a promising molecular biomarker for quantification of the target organism. The aim of this study was to develop a quantitative real-time PCR (Q-PCR) assay to detect pathogenic and nonpathogenic strains of E. coli, differentiating them from Shigella spp., which are closely related bacteria. Actually the two taxa are often difficult to distinguish by phenotypic traits, while, at genome level, some authors consider them as belonging to the same species (4, 5, 14). In the current study the molecular basis of the indole phenotype in Shigella and Escherichia species was examined by an investigation of the molecular organization of the tna operon. One-hundred eighteen E. coli strains and 100 non-E. coli strains representative of other enteric and environmental species, collected from different sources and international collections, were studied (Table 1). All the new isolates were identified by the Vitek system (bioMérieux, Marcy l'Etoile, France) and/or through the sequencing of the 16S rRNA gene operon. After a preliminary screening, the gene encoding tryptophanase (tnaA) was chosen as target for the design of E. coli-specific PCR primers and of a TaqMan MGB probe by comparing relative gene sequences of E. coli K-12 with those of Shigella flexneri 2a strain 2457^T, E. coli O157:H7 EDL 933, and E. coli CFT073 (Table 2). The rationale behind this selection was to identify genomic regions able to discriminate the two species, since published protocols are not suitable for this purpose (3, 9).

TABLE 1.

Strains used in this work

Species	Strain(s) (collectiona)	No. of strains isolated in this work	Source(s)
Aeromonas hydrophila	DSM 30187T
Aeromonas veronii		1	Surface water
Aeromonas agilis	ATCC 966T
Bacillus cereus	ATCC 14579T
Bacillus mycoides	ATCC 6462T
Citrobacter amalonaticus	DSM 4593T
Citrobacter braaki		1	Surface water
Citrobacter farmeri	DSM 17655T
Citrobacter freundii	DSM 30039T, CitFre1 (DiSTAM)	2	Surface water
Citrobacter koseri		1	Surface water
Citrobacter sp.		1	Surface water
Enterobacter aerogenes		1	Surface water
Enterobacter agglomerans	ENTagg4 (DiSTAM)	1	Surface water
Enterobacter amnigenus		1	Surface water
		1	Groundwater
Enterobacter cloacae	DSM 30054T	1	Surface water
		1	Groundwater
Enterobacter gergoviae	DSM 9245T
Enterobacter intermedius		1	Surface water
		1	Groundwater
Enterococcus faecalis	ATCC 19433T, ATCC 27332, NCDO 611
Enterococcus faecium	ATCC 19434T
Enterococcus hirae	ATCC 8043T
Escherichia albertii	DSM 17582T
Escherichia blattae	DSM 4481T
Escherichia coli	ATCC 11229, 11775T, 25922, 35150	19	Human stool
	DSM 682, 1576, 6255, 10650, 11250	6	Groundwater
	ED173, ED172, EF1, EF28, ED226 (ISS)	58	Surface water
	NCTC 12079	2	Wastewater
	EscCol1, EscCol2, EC43a2, EC43b, EC44, EC45, EC46, EC50, EC51, EC52, EC55, CD2, CD3, CL10, CM11, 393 (DiSTAM)	2	Treated wastewater
Escherichia fergusonii	DSM 13698T
Escherichia hermannii	DSM 4560T	2	Surface water
Escherichia vulneris	DSM 4564T	3	Surface water
Flavimonas sp.		1	Surface water
Hafnia alvei	DSM 30163T	1	Surface water
Klebsiella ornithinolytica		1	Surface water
Klebsiella oxytoca	DSM 5175T	2	Surface water
	KleOxy1 (DiSTAM)	2	Groundwater
Klebsiella pneumoniae	KlePne1 (DiSTAM)
Lactobacillus acidophilus	ATCC 4356T
Lactobacillus buchneri	CNRZ 214
Lactococcus lactis	ATCC 19435T
Leclercia adecarboxylata	DSM 5077T	1	Surface water
Listeria innocua	ATCC 33090T
Morganella morganii	DSM 30164T; CE6, MorMorg1 (DiSTAM)	2	Surface water
Pantoea agglomerans		1	Surface water
Pasteurella haemolitica		1	Surface water
Pasteurella multocida		1	Surface water
Plesiomonas shigelloides	DSM 8224T	2	Surface water
Proteus hauseri	ATCC 13315
Proteus mirabilis	DSM 4479T	1	Surface water
Providencia stuartii		1	Surface water
Pseudomonas aeruginosa	ATCC 10145T, 27853
Pseudomonas fluorescens	DSM 50106, 50148
Pseudomonas mendocina		1	Surface water
Pseudomonas putida	ATCC 12633T
Rahnella aquatilis	RahAcq1 (DiSTAM)
Salmonella enterica serovar Enteritidis	ATCC 13076T
Salmonella enterica serovar Typhimurium	ATCC 13311T
Serratia fonticola		1	Groundwater
Serratia liquefaciens		1	Groundwater
Serratia marcescens	DSM 30121T
	SERmar2, SERmar3, SERmar4 (DiSTAM)
Shigella boydii	DSM 7532T
Shigella flexneri	DSM: 4782T, ShiFlex (DiSTAM)
Shigella sonnei	ATCC 29930T; PO2, ShiSon1 (DiSTAM)
Sphingomonas paucimobilis		1	Surface water
Staphylococcus aureus	ST11 (DiSTAM)
Vibrio agarivorans	DSM 13756T
Vibrio parahaemoliticus	DSM 10027T
Yersinia enterocolitica	DSM 4780T, 13030T; YerEnt (DiSTAM)
Yersinia intermedia		1	Surface water

^aISS, Istituto Superiore di Sanitá, Rome, Italy; DiSTAM, Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Milan, Italy.

TABLE 2.

Primers and probe designed in this work and relative amplification conditions

Target gene	Primer or probe	Sequencea (5′→3′)	Position on geneb	Thermal profile of amplification (no. of cycles)	Product size (bp)
tnaA	tnaA_F	GGGGCGGTGACGCAG	464-478	95°C, 10 min (1); 95°C, 15 s (40);	136
	tnaA_R	CCTGGTGAGTCGGAATGGTG	599-580	60°C, 1 min (40)
	Probe	FAM-CGATGATGCGCGGCG-MGB	492-506
tnaL	tnaOP_F	CGAGGATAAGTGCATTATGAATATCT	−16-10	94°C, 10 min (1); 94°C, 1 min (35)	3,065
tnaB	tnaOP_R	TTAGCCAAATTTAGGTAACACGTT	3049-3026	58°C, 1 min (35); 72°C, 1 min (35)

^aFAM, 6-carboxyfluorescein.

^bPositions on genes are given according to the tnaL sequence of E. coli K-12 (U00096).

Real-time PCR amplification was performed in a 25-μl volume containing 1× TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA), 1 μM of primers, 200 nM of probe, and 100 ng of DNA template. Tests were performed in triplicate in an ABI Prism 7900 thermal cycler (Applied Biosystems), using MicroAmp optical tubes and caps (Applied Biosystems). Amplification conditions are reported in Table 2. One hundred five out of 118 E. coli strains (89%) gave positive signal in the Q-PCR test, while no amplification signal was observed in 98 out of 100 non-E. coli strains (98%). In particular, all Shigella sp. isolates were negative. According to ISO/FDS 16140:2003 (12), the new assay showed an accuracy of 93.1% and a specificity of 98.0%. Thirteen E. coli strains, 12 with an indole-positive phenotype, did not produce the expected fragment: CD3, DSM 682, E01, E05, E16, E22, E81 (indole negative), EC52, ED172 (O103:H2), ED226 (O113:H21), EF1 (O111:H⁻), EscCol1, and EscCol2. On the other hand, Escherichia albertii DSM 17582^T (indole negative) and Escherichia fergusonii DSM 13698^T (indole positive) strains gave a false-positive amplification signal. Then, the tna operon was entirely sequenced through a gene walking analysis using primers tnaOP_F and tnaOP_R, designed on the 5′ region of tnaL and on the 3′ region of tnaB of E. coli K-12, respectively (Table 2). The same survey was extended to Shigella boydii DSM 7532^T, Shigella flexneri DSM 4782^T, and Shigella sonnei ATCC 29930^T, SHIson1, and PO2. PCR was performed as described above, and the amplification cycles are reported in Table 2. Both strands of the amplification products of the tna operon were sequenced, analyzed by the BLAST 2 sequence program, and deposited in the GenBank database. A multiple-alignment distance, determined by the unweighted pair group method using arithmetic averages was used to draw a tree based on tna operon sequences by the Kodon software (Applied Maths, Kortrijk, Belgium). S. sonnei strains did not amplify the tna operon. E. coli E81, E. albertii DSM 17582^T, and S. boydii DSM 7532^T isolates amplified regions of 4,387 bp, 3,827 bp, and 2,814 bp, respectively. The other strains, including E. fergusonii DSM 13698^T and S. flexneri DSM 4782^T, amplified a region of 3,065 bp as expected. The dendrogram obtained by the alignment of tna operon sequences allows discrimination of two major clusters of nucleotide distance at the 72.1% level (Fig. 1). Cluster A includes all the Escherichia spp. and four strains of Shigella spp. without insertion sequences (S. boydii B12 and S. flexneri 2a strain 2457T, 5 strain 8401, and 2a strain 301); cluster B groups all the other Shigella sp. strains presenting insertion sequence (IS) elements and showing an indole-negative phenotype. E. coli K-12 W3110 has two complete IS5 elements (1,195 bp): one in the intragenic tnaL-tnaA region and the other in the tnaB gene (Fig. 2). Isolate E81, from a healthy individual, had a tnaA gene that was interrupted by a 1,329-bp IS, which showed a 100% homology with IS10 present in Salmonella enterica serovar Typhimurium (17). E. albertii DSM 17582^T has an IS of 768 bp (IS1A) in the spacing region between tnaL and tnaA; in addition, two nucleotide deletions (T₁₇₈₅ and C₂₅₈₇) and one insertion (A₁₇₁₈) with respect to E. coli K-12 were detected. As a consequence, these strains are unable to produce a functional enzymatic system for tryptophan degradation. In E. coli ATCC 11775^T, E. coli UTI89, avian pathogenic E. coli O1, E. coli 536, E. coli CFT073, and E. fergusonii DSM 13698^T (cluster A₁) the tna operon has the same organization as in E. coli K-12 but with a single nucleotide insertion (A₁₇₁₈) in the region between tnaA and tnaB. Since this insertion falls in a noncoding region, open reading frames are not affected and strains are phenotypically indole positive. With the exception of DSM 10650 and two O157:H7 isolates (EDL933 and strain Sakai), all the E. coli strains of cluster A₂ present, compared to E. coli K-12, three point mutations in the region from which primer tnaA_F was designed: G₄₇₂ to A, G₄₇₅ to C, and G₇₈₈ to A. Moreover, apart from strains ED226 (O113:H21), ED172 (O103:H2), and EF1 (O111:H⁻), they also present two point mutations in correspondence to primer tnaA_R: C₅₈₀ to T and C₅₈₃ to T. All these mutations take place on the third base of the codon, and, since the amino acid does not change, they do not affect the translation into a functional tryptophanase. However, they decrease the annealing efficiency during PCR amplification, giving rise to a false-negative signal. It is noteworthy that the same point mutations are present in Shigella strains grouped in this cluster. Shigella boydii B12 has an indole-negative phenotype since, as reported by Rezwan et al. (20), it presents an IS before the region we have investigated that disrupts the expression of the tna operon. Cluster B groups Shigella spp. whose tna operons are affected by insertion elements causing the indole-negative phenotype. In particular, S. boydii strains in cluster B₁ show one partial 192-bp IS1 sequence followed by a full IS1 sequence; these sequences determine a deletion of 49 bp, including 21 bp of tnaL, and they present the same point mutations in tnaA primers annealing regions already described for E. coli strains of cluster A₂. S. boydii and S. flexneri strains of cluster B₂ displayed instead a full IS1 of 768 bp starting from base 55 of tnaL; the insertion was followed by the deletion of the complete 205-bp interspace region between tnaL and tnaA and of the first 777 bp in the 5′ start sequence of the tnaA gene; therefore, 235 bp is missing. Shigella dysenteriae strains (cluster B₃) presented the same 768-bp IS, IS1, at base 55 of tnaL of cluster B₂, but without any deletion. Moreover, strain D3 showed the presence of another IS1 located in an opposite orientation on the tnaA gene, giving rise to a lower nucleotide homology (71.0%) compared to other strains (>99.9%) of the same cluster. Furthermore, these isolates presented, in the regions from which primers were drawn, the same point mutations detected on cluster B₁ and A₂ strains. The high frequency of IS elements is known to mediate various genetic rearrangements, including inversions and deletions, that could play an important role in the evolution of the taxa. Besides, the occurrence of gene transfer by conjugation, transduction, and formation of recombinants between S. flexneri and E. coli, particularly for pathogenic serotypes, has already been demonstrated (16, 18). IS elements have often been associated with negative phenotypes since they can disrupt the functionality of the genes (20); in our study this trait was highlighted through the sequence analysis of the E81 strain, the sole E. coli isolate with a negative indole phenotype. The other E. coli strains that do not gave amplification signals in this Q-PCR assay reveal the same point mutations and the same organization of the operon as some S. boydii and S. flexnerii strains, corroborating the hypothesis that the Shigella pathotype arose from E. coli ancestors (7, 19). Anyway, it is not possible to find other regions in the tna operon that could allow the discrimination between the species that we considered in this work. Although E. fergusonii and E. albertii gave a false-positive signal, this fact does not invalidate the test since they have the same habitat as E. coli (10). Finally, this PCR assay can be employed as a rapid preliminary tool, even if it should be integrated with phenotypic results.

FIG. 1.

Tree obtained by the unweighted-pair group method using arithmetic averages (UPGMA) for tna operon sequences of Escherichia and Shigella species. Bootstrap values are indicated on each node of the tree (1,000 pseudoreplicates). *, sequences obtained in this work; nd, not determined.

FIG. 2.

Comparison of gene organizations of tna operons from Escherichia and Shigella species.

Nucleotide sequence accession numbers.

Nucleotide sequences obtained in this work have been deposited in the GenBank database under accession numbers EF445878 to EF445895.