Genotyping of Chromobacterium violaceum isolates by recA PCR-RFLP analysis

Abstract

Intraspecies variation of Chromobacterium violaceum was examined by comparative sequence – and by restriction fragment length polymorphism analysis of the recombinase A gene (recA-PCR-RFLP). Primers deduced from the known recA gene sequence of the type strain C. violaceum ATCC 12472^T allowed the specific amplification of a 1040 bp recA fragment from each of the 13 C. violaceum strains investigated, whereas other closely related organisms tested negative. HindII–PstI–recA RFLP analysis generated from 13 representative C. violaceum strains enabled us to identify at least three different genospecies. In conclusion, analysis of the recA gene provides a rapid and robust nucleotide sequence-based approach to specifically identify and classify C. violaceum on genospecies level.

Chromobacterium violaceum, Genotyping, recA, recA-PCR-RFLP analysis

1 Introduction

The genus Chromobacterium currently consist of a single species, Chromobacterium violaceum, a flagellated, heterotrophic, rod-shaped gram-negative organism, that produces a violet, water-insoluble pigment, hence its name. On the basis of the Judicial Opinions of the International Committee on Systematics of Prokaryotes, the correct generic name of this genus is Chromobacterium Bergonzini 1880 [1]. Phylogenetically, C. violaceum belongs to the family Neisseriaceae of b-Proteobacteria. As a saprophyte, the organism is found in a wide variety of tropical and subtropical ecosystems, predominantly in water and soil [2,3]. Although C. violaceum has a low infectious capability, some cases of severe life-threatening sepsis with metastatic abscesses similar to melioidosis and infections of various animal species have been reported [4,,6]. A characteristic feature of C. violaceum is the production of violacein, a purple pigment with antibiotic, antitumoral and anti-Trypanosoma cruzi activities [7]. However, the occurrence of pathogenic, non-pigmented variants of C. violaceum also has been described [8] and at least two other closely related genera, Janthinobacterium and Iodobacter do also produce violacein. [9,10].

Diagnosis of C. violaceum infection is currently based on culture of blood, abscess fluid, or skin exudate. Serological tests or PCR assays that allow the specific detection and differentiation among C. violaceum isolates do not exist. Recently, the entire genomic sequence (Accession Number AE016825) of the C. violaceum type strain ATCC 12472^T (LMG 1267^T) was determined [2] and is now available from http://www.brgene.lncc.br/ for detailed open reading frame (ORF) analysis.

Based on the available genomic sequence information we have developed a C. violaceum specific PCR assay using recA as the target gene. The partial recA genes of 13 C. violaceum strains of different environmental and clinical origins have been sequenced and compared to each other in multiple sequence analysis. RecA-restriction length polymorphism analysis with two endonucleases demonstrated that variations within recA can be used for subtyping of C. violaceum isolates.

2 Material and methods

2.1 Bacterial strains

A representative panel of 13 C. violaceum strains isolated from different environmental or clinical specimens and 52 strains of closely related genera were examined (Table 1). Strains or DNA were obtained from the American Type Culture Collection (ATCC, MD, USA), the National Collection of Type Cultures of the Central Public Health Laboratory (NCTC, London, UK), B. Niederwöhrmeier (WIS, Munster, Germany), J. Ellis, (Defence Science and Technology Laboratory, UK), and N. Anuntagool (Bangkok, Thailand). C. violaceum strains were exclusively obtained from the Laboratorium voor Microbiologie (BCCM/LMG, Universiteit Gent, Belgium), Strain LMG 1267 is the type strain of C. violaceum and identical to strain ATCC 12472. Bacteria were grown on standard media using the conditions recommended by the bacterial collections.

1

List of strains analyzed in this study

Strain	Source	Geographic origina	Species	+/− in PCR
RecA
LMG 1267T	Freshwater	Malaya	C. violaceum	+
ATCC 12472T	Freshwater	Malaya	C. violaceum	+
LMG 3936	Calf	Malaysia	C. violaceum	+
LMG 3940	Gibbon	Malaysia	C. violaceum	+
LMG 3942			C. violaceum	+
LMG 3947	Pond water	Malaysia	C. violaceum	+
LMG 3953			C. violaceum	+
LMG 3954	Human	USA	C. violaceum	+
LMG 3958	Bovine	USA	C. violaceum	+
LMG 3960	Freshwater	Singapore	C. violaceum	+
LMG 3962	Swine	USA	C. violaceum	+
LMG 3963	Human	Malaya	C. violaceum	+
LMG 3963vb	Human	Malaya	C. violaceum	+
LMG 6630T	Water	UK	I. fluviatilis	−
LMG 2892T	Soil	USA	J. lividum	−
LMG 1286T	Human	Zimbabwe	P. agglomerans	−
LMG 14449T	Human	USA	Bo. parapertussis	−
LMG 14455T	Human	USA	Bo. pertussis	−
ATCC 13077T	Human	USA	N. meningitidis	−
LMG 2844T	Milk		A. hydrophilia	−
ATCC 23344T	Horse	China	B. mallei	−
NCTC 10247	Human	Turkey	B. mallei	−
NCTC 10248	Human	Turkey	B. mallei	−
NCTC 03709	Horse	India	B. mallei	−
NCTC 10260	Human	Turkey	B. mallei	−
Muk	Horse	India	B. mallei	−
Bogor	Horse	Indonesia	B. mallei	−
Zagreb	Horse	Yugoslavia	B. mallei	−
ATCC 15310	Horse	Hungary	B. mallei	−
K96243	Human	Thailand	B. pseudomallei	−
ATCC 23343T			B. pseudomallei	−
NCTC 4845	Monkey	Singapore	B. pseudomallei	−
Soil 1977	Soil	Madagascar	B. pseudomallei	−
D4899/303	Environment	Venezuela	B. pseudomallei	−
7894	Human	Ecuador	B. pseudomallei	−
SID4350	Human	N Thailand	B. pseudomallei	−
SID5752	Human	UK/Thailandc	B. pseudomallei	−
SID2889	Human	UK/Bangladeshc	B. pseudomallei	−
Hainan 4	Environment	China	B. pseudomallei	−
ATCC 700388T			B. thailandensis	−
UE11	Soil	NE Thailand	B. thailandensis	−
UE17	Soil	NE Thailand	B. thailandensis	−
UE29	Soil	NE Thailand	B. thailandensis	−
E111	Soil	S Thailand	B. thailandensis	−
E217	Soil	S Thailand	B. thailandensis	−
S2	Soil	Viet Nam	B. thailandensis	−
S3	Soil	Viet Nam	B. thailandensis	−
LMG 20980	Soil	USA	B. anthina	−
LMG 18924	Maize	France	B. graminis	−
LMG 19076	Soil	UK	B. caledonica	−
LMG 2129	Sorghum bicolor	USA	B. andropogonis	−
P26			X. maltophilia	−
LMG 559	Brassica napus	UK	X. campestris	−
ATCC 10145			Ps. aeruginosa	−
P739			R. solanacearum	−

^aThe source of the isolate is given when available. N, North; S, South; NE, North-East; UK, United Kingdom, USA, United States of America; B., Burkholderia; Bo., Bordetella; Ps., Pseudomonas; X., Xanthomonas; R., Ralstonia; A., Aeromonas; J., Janthinobacterium; P., Pantoea; I., Iodobacter; N., Neisseria; C., Chromobacterium.

^bNon-pigmented variant of LMG 3963.

^cIsolated from patient after stay abroad.

2.2 DNA preparation

One distinct colony of each strain was transferred from an agar plate to 200 μl lysis buffer [5× buffer D (PCR Optimation Kit, Invitrogen, DeShelp, The Netherlands) 1:5 diluted in Aqua dest]; 0.5% Tween 20^™ (ICI, American Limited, Merck, Germany); 2 mg ml⁻¹ proteinase K (Roche Diagnostics, Germany). After incubation at 56 °C for 1 h and inactivation for 10 min at 95 °C, 2 μl of the cleared lysate were used as template in the PCR assays.

2.3 RecA-PCR amplification and sequence analysis

The partial recA sequences of strains LMG 1267^T, ATCC 12472^T, LMG 3936, LMG 3940, LMG 3942, LMG 3947, LMG 3953, LMG 3954, LMG 3958, LMG 3960, LMG 3962, LMG 3963, and LMG 3963v were amplified by PCR using the known recA sequence information of the genomic sequence of C. violaceum ATCC 12472^T (Acc. No. AE016915, nucleotide position 218121–219173). The primer pair recA-viol-f (5′-AAGACAAGAGCAAGGCGCTGGC-3′) and recA-viol-r (5′-TCGAAGGCGTCGTCGGCGAAC-3′) was used to generate a 1040 bp fragment of the entire (1057 bp) recA gene. For specificity testing, a series of strains of closely related genera were used (Table 1). PCR was performed in 50 μl ready to go mastermix (Eppendorf GmbH, Germany) using 15 pmol of each primer. Amplification was carried out in a Perkin–Elmer GeneAmp2400 cycler. Thirty cycles were completed, each consisting of 30 s denaturation at 94 °C, 30 s annealing at 70 °C, and elongation at 72 °C for 1 min. A final elongation of 7 min at 72 °C completed the run. Of each PCR 7 μl were analyzed by agarose gel electrophoresis (1% w/v in TAE buffer) for the presence of a 1040 bp product.

The nucleotide sequence of each PCR product was determined by dideoxynucleotide sequencing of both strands with primers recA-viol-f and recA-viol-r. Sequencing was carried out with an ABI PRISM^? 3100 Genetic Analyzer (Applied Biosystems, USA). Nucleotide sequences were further analyzed using the Chromas 1.45 program.

2.4 RecA-RFLP analysis

Each PCR product was purified with the Qiaquick PCR purification kit (Qiagen, GmbH) and digested in separate reactions with two restriction endonucleases (PstI and HindII) that were selected on the basis of the obtained recA nucleotide sequences. Digestion was carried out with 5 U of each enzyme for 3 h using the conditions recommended by the manufacturer. Restriction fragments were separated by 6% DNA–polyacrylamide-gel electrophoresis using the MINI-PROTEAN II cell (Bio-Rad Laboratories GmbH, Germany) and 1× TBE as running buffer at a constant voltage of 130 V for 1 h. Fragments were visualized with UV light after staining in ethidium bromide (0.5 μg ml⁻¹).

2.5 Multiple sequence alignments

Multiple sequence alignments of partial recA sequences were performed with ClustalW 1.8 and prepared for publication with BOXHADE 3.21 available at http://clustalw.genome.jp/ and http://www.ch.embnet.org/software/BOX_form.html, respectively.

2.6 Accession numbers

The partial recA sequences of strains LMG 3954, LMG 3962, LMG 3963, LMG 3963v, LMG 3936, LMG 3953, LMG 3960, LMG 3942, LMG 3947, LMG 3940, LMG 1267^T, and LMG 3958 have been deposited in the EMBL Nucleotide Sequence Database under the Accession Numbers AJ871115 to AJ871126, respectively.

3 Results

C. violaceum strains were cultivated as recommended by the LMG culture collection. Under these conditions single colonies developed a violet color due to the production of violacein, a purple pigment characteristic for C. violaceum. During cultivation, however, non-pigmented colonies were observed with strain LMG 3963, which we termed LMG 3963v. The non-pigmented variant was included in subsequent recA PCR- and RFLP-analysis. Two strains, LMG 3953 and LMG 3954 showed a significantly less intensive production of violacein.

Specific amplification of recA by PCR was carried out with primers recA-viol-f and recA-viol-r, constructed from the available genomic sequence of the C. violaceum type strain ATCC 12472. A single band of 1040 bp was amplified with template DNA of all C. violaceum isolates, whereas other bacteria listed in Table 1 tested negative (not shown). In order to detect variations within recA, the nucleotide sequence of each PCR product was determined. Subsequent BLAST analysis of the 13 different PCR products confirmed the specific amplification of the recA gene. A multiple sequence alignment comprising 858 bp (position 34–892) of the entire recA (1057 bp) sequence of each strain, including the known recA sequence of the type strain ATCC 12472, was performed. The alignment revealed significant sequence variations among recA of different C. violaceum isolates (supplementary data). In total 8 different alleles among the 13 tested strains were detected. Only one strain, LMG 3963, was identical in its recA gene sequence to strain ATCC 12472^T (LMG 1267^T). The non-pigmented variant LMG 3963v differed from the pigmented variant in one nucleotide.

While most of the recA sequences differed in only 3–4 bp from recA of strain ATCC 12472 without the alteration of the corresponding primary amino acid sequence, strains LMG 3953, LMG 3954, and LMG 3947 had a significant different recA nucleotide signature with an identity of 96% and 95%, leading to the exchange of 5 (LMG 3953 and LMG 3954) and 9 (LMG 3947) amino acids compared to strain ATCC 12472^T, respectively (data not shown). Based on the sequence information of all obtained recA sequences we have selected two restriction endonucleases, HindII and PstI, for further recA-RFLP analysis. The computer calculated fragment sizes generated with PstI and HindII are listed in Table 2. The calculated sizes were finally confirmed by subsequent polyarcrylamid gelelectrophoresis of fragments obtained from HindII and PstI digestions (Fig. 1A and B).

2

Computer calculated PstI–HindII restriction patterns of recA from various C. violaceum isolates

Strains	Length (bp)
PstI	HindII
LMG 3953/LMG 3954	814/226	720/320
LMG 3947	900/140	519/326/195
ATCC 12472T and other Chr. Violaceum isolates listed in Table 1	226/267/547	102/320/618

1

HindII (A) and PstI (B)-recA restriction profiles of PCR products amplified with primers recA-viol-f and and recA-viol-r. DNA size-standard (lane 1), LMG 1267^T (lane 2), LMG 3936 (lane 3), LMG 3940 (lane 4), LMG 3942 (lane 5), LMG 3958 (lane 6), LMG 3960 (lane7), LMG 3962 (lane 8), LMG LMG 3963 (lane 9), LMG 3953 (lane 10), LMG 3954 (lane 11), and LMG 3947 (lane 12).

4 Discussion

C. violaceum is the only species listed within the genus Chromobacterium. Formerly, the name Chromobacterium was used as a collective term for a series of violet pigment producing organisms. Some of these species like Chromobacterium lividum and Chromobacterium fluviatile, have been reclassified and assigned to new genera, namely Janthinobacterium (Type species: J. lividum) and Iodobacter (type species: I. fluviatilis), respectively.

Identification of C. violaceum is currently based on biochemical analysis with API 20 NE and the production of the purple pigment violacein. However, both methods have major drawbacks. Inglis et al. [11] and Lowe et al. [12] reported that the most common misidentification of Burkholderia pseudomallei by biochemical analysis with API 20 NE was C. violaceum. This is of particular importance for clinicians since the clinical picture of a C. violaceum infection may mimic melioidosis, the disease caused by B. pseudomallei[4,5]. This is even strengthened by the occurrence of non-pigmented variants of C. violaceum that might be confused with B. pseudomallei. In 1977, Sivendra and Tan [8] demonstrated that non-pigmented variants can not be distinguished from pigmented strains in terms of their virulence. By the same authors non-pigmented variants were sent to three independent laboratories for identification. Non of the laboratories did identify them correctly, but classified them as Aeromonas hydrophilia or Pseudomonas sp. Furthermore, the production of violacein is not restricted to Chromobacterium since at least two other closely related genera, Janthinobacterium and Iodobacter do also produce this pigment. [9,10].

Due to these diagnostic uncertainties, the aim of this study was to develop a PCR-based assay to specifically detect and to differentiate among different isolates of C. violaceum. For this purpose a representative panel of C. violaceum strains originating from different environmental and clinical sources (Table 1) was selected for further analysis. The usefulness of recA as a target gene for the development of species specific PCR assays followed by adjacent RFLP analysis with a high discriminative power of inter- and intraspecies level differentiation was demonstrated for several bacteria of different genera [13,,,16].

With the primers used we were able to specifically amplify recA of all C. violaceum strains, whereas other closely related bacteria tested negative (Table 1). Therefore, this assay might serve as a tool with potential diagnostic value to specifically detect and to differentiate C. violaceum from other closely related organisms. Comparison of the obtained recA sequences in multiple sequence analysis revealed a total of 9 different recA alleles among the 13 strains tested. Due to this high variability, recA might serve as an important target gene in multi locus sequence analysis (MLST). Based on the obtained sequence information two restriction endonucleases, HindII and PstI, were selected for further recA-RFLP analysis. Both enzymes enabled the separation of three distinct genospecies within the 12 different strains tested. Although recA gene sequencing has a higher discriminative power, recA-RFLP analysis can be used as a first approach for intraspecies differentiation.

In conclusion, the developed recA-based PCR-RFLP assay provides a rapid and robust nucleotide sequence-based approach to specifically identify and to classify C. violaceum on genospecies level. The application of this assay to an extended panel of C. violaceum isolates will provide more details on the complexity of the genus Chromobacterium.

Acknowledgements

We are grateful to C. Lodri and I. Patzwald for excellent technical assistance.

Appendix A Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.femsle.2005.02.005.