16S rDNA genotyping using PCR/RFLP (restriction fragment length polymorphism) analysis among the family Vibrionaceae

Abstract

The 16S rDNA genotypes among the family Vibrionaceae were determined using PCR/RFLP analysis. Five tetrameric restriction enzymes (Hha I, Dde I, Rsa I, Sau 3AI and Msp I) were used for RFLP analysis and adequate numbers of informative bands were obtained from each enzyme. Twenty-seven genotypes were obtained from 49 type and reference strains including 35 species. Nineteen species could be assigned to specific 16S rDNA genotypes, supporting the application of this analysis for identification. Trees constructed using five endonucleases resolved groups almost identical to those inferred from 16S rRNA gene sequencing. However, the branch lengths and detailed relationships among strains within a group differed from those inferred from sequence comparisons. The results of this study should be useful for genotyping, identification and approximate classification of natural isolates belonging to the family Vibrionaceae.

Genotyping, Vibrionaceae, Vibrio, Polymerase chain reaction, Restriction fragment length polymorphism, 16S rDNA, Vibrio iliopiscarius

1 Introduction

The members of the family Vibrionaceae have been frequently isolated or detected from seawater, estuarine, and freshwater environments [1, 2]. They include several species pathogenic for humans and marine animals [3–6].

Clarifying bacterial diversity among many isolates from natural environments is required for understanding their ecology and ecological functions. In the past, many phenotypic characters had to be examined to estimate the diversity among natural isolates, in which identification was made to the genus level using selected determinative criteria[2]. In the modern taxonomy of bacteria, 16S rDNA sequence analysis has become a standard method for the investigation of their phylogenetic relationships [4, 7, 8]. DNA-DNA hybridization has routinely been required for differentiation among species[9].

However, these methods are not currently appropriate for routine identification of multiple isolates from the environment because of time and cost. Recently, comparative studies of 16S rRNA genes using polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis have been gradually increasing because of their relative simplicity and rapidity [10, 11].

In this study, we determined 16S rDNA genotypes among the family Vibrionaceae using PCR/RFLP analysis to obtain useful information for the genotyping, identification and approximate classification of natural isolates.

2 Materials and methods

2.1 Type and reference strains used

A total of 49 type and reference strains belonging to the family Vibrionaceae were used and are listed in Table 1.

1

Reference and type strains used in this study and 16S rDNA genotypes and restriction patterns obtained from RFLP analysis of 16S rRNA genes

Taxon	Straina	Abbreviationb	16S rDNA genotypec	Restriction patterns of 16S rRNA gene digested with:
				Hha I	Dde I	Rsa I	Sau 3AI	Msp I
Vibrio alginolyticus	NCMB 1903T	alg1	1	1	1	1	1	1
	GIFU 1417	alg2	1	1	1	1	1	1
Vibrio parahaemolyticus	ATCC 17802T	par	1	1	1	1	1	1
Vibrio campbellii	ATCC 25920T	cam	1	1	1	1	1	1
Vibrio carchariae	ATCC 35084T	car	1	1	1	1	1	1
Vibrio harveyi	ATCC 14126T	har	1	1	1	1	1	1
Vibrio proteolyticus	NCMB 1326T	pro	1	1	1	1	1	1
Vibrio vulnificus (biotype I)	ATCC 27562T	vul1	1	1	1	1	1	1
	JCM 3726	vul2	1	1	1	1	1	1
	JCM 3727	vul3	1	1	1	1	1	1
Vibrio vulnificus (biotype II)	ATCC 33148	vul II	1	1	1	1	1	1
Vibrio aestuarianus	ATCC 35048T	aes	2	1	1	2	1	1
Vibrio diazotrophicus	ATCC 33466T	dia	3	1	2	1	1	1
Vibrio anguillarum	IFO 13266T	ang1	4	1	2	2	1	1
	GIFU 10640	ang2	4	1	2	2	1	1
	GIFU 10645	ang3	4′	1	2	2′	1′	1
Vibrio ordalii	ATCC 35509T	ord	4	1	2	2	1	1
Vibrio nigripulchritudo	ATCC 27043T	nig	5	1	3	1	1	1
Vibrio metschnikovii	IAM 1039	met	6	1	4	2	3	1
Vibrio fluvialis (biotype I)	NCTC 11327T	flu1	7	1	5	2	2	1
	JCM 3733	flu2	7	1	5	2	2	1
	GIFU 9915	flu3	7	1	5	2	2	1
	GIFU 9916	flu4	8	1	5	2	1	1
	JCM 3734	flu5	7	1	5	2	2	1
Vibrio fluvialis (biotype II)	JCM 3753	flu II	7	1	5	2	2	1
Vibrio pelagius (biotype I)	ATCC 25916T	pel	9	1	1	2	1	2
Vibrio pelagius (biotype II)	ATCC 33504	pel II	9	1	1	2	1	2
Vibrio tubiashii	ATCC 19109T	tub	9	1	1	2	1	2
Vibrio natriegens	CCM 2575T	nat	10	1	1	1	2	2
Vibrio penaeicida	IFO 15640	pen	11	1	3	2	1	2
Vibrio orientalis	ATCC 33934T	ori	12	1	1	2	1	3
Vibrio splendidus (biotype I)	ATCC 33125T	spl	13	2	1	1	2	1
Vibrio gazogenes	ATCC 29988T	gaz	14	3	5	3	1	1
Vibrio ichthyoenteri	IFO 15847	ich	15	1	1	1	4	1
Vibrio hollisae	JCM 1284	hol	16	3	6	4	1	3
Vibrio mimicus	ATCC 33653T	mim	17	3	1	3	2	4
Vibrio cholerae (serotype Ogawa)	IID 936	cho	18	4	1	3	2	4
Photobacterium damsela	ATCC 35083	dam1	19	5	1	4	5	2
	ATCC 33539T	dam2	19	5	1	4	5	2
Vibrio iliopiscarius	ATCC 51760T	ili	20	5	1	5	2	1
Photobacterium phosphoreum	IAM 12085	ppho1	20	5	1	5	2	1
	IAM 14401T	ppho2	21	6	1	5	2	1
Photobacterium leiognathi	ATCC 25521T	plei	20	5	1	5	2	1
Photobacterium angustum	ATCC 25915T	pang	22	5	1	3	2	1
Vibrio fischeri	NCTC 11327T	fis	23	7	7	6	1	2
Vibrio logei	ATCC 15382	log	24	2	8	6	6	2
Vibrio salmonicida	ATCC 43839T	sal	25	2	8	7	6	2
Vibrio marinus	ATCC 15381T	mar	26	2	9	6	2	5
Vibrio costicola	NCMB 701T	cos	27	8	10	4	7	6

^aAbbreviations: ATCC, American Type Culture Collection, Rockville, MD; CCM, Czechoslovak Collection of Microorganisms, J.E. Purkyne University, Brno, Czechoslovakia; IAM, Institute of Applied Microbiology, University of Tokyo, Tokyo, Japan; IFO, Institute for Fermentation, Osaka, Japan; IID, Institute of Medical Science, University of Tokyo, Tokyo, Japan; GIFU, Department of Microbiology, Gifu University School of Medicine, Gifu, Japan; JCM, Japan Collection of Microorganisms, The Institute for Physical and Chemical Research (RIKEN), Saitama, Japan; NCMB, National Collection of Marine Bacteria, Aberdeen, Scotland, UK.

^bAbbreviation of type and reference strain.

^cThe 16S rDNA genotypes, which are numbered from 1 to 27, represent the combination of restriction patterns obtained with five endonucleases used in this study. The 16S rDNA genotype 4′ was observed in V. anguillarum GIFU 10645. The difference between 4 and 4′ was that the smallest band in 4′ digested with Rsa I was 10 bp longer than other V. anguillarum strains, and when digested with Sau 3AI it was 25 bp shorter than other strains.

2.2 DNA extraction and PCR amplification

Cultures were incubated on agar plates of half-strength ZoBell 2216E medium [12, 13]. After harvesting, cells were washed and suspended in sterile distilled water and 40 μl of the suspension was mixed with 10 μl of proteinase K (1 mg ml⁻¹) (Sigma Chemical Co., St. Louis, MO, USA) and 50 μl of 2×K buffer (40 mM Tris buffer, 0.2% Nonidet P-40, 0.2 mM EDTA, 1% Tween 20, distilled water, pH 8.0). The mixture was heat-treated at 60°C for 20 min, followed at 100°C for 15 min, then cooled rapidly on ice and centrifuged at 8000 rpm for 5 min. Primers described by Weisburg et al.[14] were used for PCR amplification. PCR amplification was done in a DNA thermal cycler (MiniCycler TM; MJ Research, Inc., Watertown, MA, USA), with the following temperature profile: an initial denaturation at 94°C for 2 min; 25 cycles of denaturation (2 min at 94°C), annealing (1.5 min at 45°C), and extension (2 min at 72°C); and a final extension at 72°C for 3 min. Amplified DNA was examined by horizontal electrophoresis on 1.0% agarose gel in TAE electrophoresis buffer (40 mM Tris, 20 mM acetate, 2 mM EDTA) with 2 μl aliquots of PCR product.

2.3 Restriction fragment analysis

The 6–8 μl of PCR products were digested with the restriction endonucleases (Toyobo Co. Ltd., Osaka, Japan) at 37°C for 90 min according to the manufacturer's instructions. The following enzymes were used: Hha I, Dde I, Rsa I, Sau 3AI, Msp I, Hae III, Hin fI and Alu I. These enzymes were chosen to obtain a good number of fragments from strains belonging to the family Vibrionaceae using the GENETYX computer program (Software Development Co., Tokyo, Japan) and the sequences of 16S rRNA gene available in DNA databases. Restricted DNAs were analyzed by horizontal electrophoresis in 4% NuSieve agarose gels (FMC Bioproducts, Rockland, ME, USA). Electrophoresis was carried out at 50 V for 170 min with a Mupid mini-gel electrophoresis apparatus (Advance Co., Tokyo, Japan) in TAE electrophoresis buffer on ice. After staining with ethidium bromide, scanning image analysis of the gel was carried out with an Atto densitograph imaging analyzer type AE-6920-MFS (Atto Co., Tokyo, Japan). The CLUSTAL W program was used for banding pattern analysis and 16S rRNA sequence data analysis[15]. Informative bands derived from restriction enzyme digestion were scored for their presence or absence, and similarity and divergence were calculated. Evolutionary distances of 16S rRNA sequences available in DNA databases were calculated using Kimura's two-parameter model[16]. The genetic distance trees were constructed from the RFLP data set and 16S rRNA sequences by the neighbor-joining method[17].

3 Results and discussion

3.1 PCR amplification and RFLP analysis on the data base

DNAs of all 49 strains tested were amplified with the ribosomal primers used in this study. All the strains produced a single band of about 1500 bp. This size corresponded to the predicted size of the 16S rRNA genes from the primer pair used here.

Moyer et al.[18] reported that the sequence data from known bacterial small subunit rRNA genes are largely incomplete as many investigators focus their efforts on partial sequences. As for the family Vibrionaceae, almost full sequences were reported by Ruimy et al.[19] but the endpoints of almost sequence data were incomplete. Thus we did not estimate the size of fragments which included endpoints of sequence data.

3.2 16S rDNA genotyping

The five endonucleases (Hha I, Dde I, Rsa I, Sau 3AI and Msp I) gave 6–10 restriction patterns of 3–6 clear distinct bands for the total of 49 type and reference strains analyzed by PCR/RFLP (Fig. 1Fig. 2). Among the eight enzymes used, three enzymes (Hae III, Hinf I and Alu I) were omitted in the following analysis because of their ambiguous results. They produced many near-coincident double or triple bands or fragments too large to be resolved by the electrophoresis system used in the present study. Some ambiguous results were caused by unexpected size fragments, most commonly those including endpoints of sequences. Seventy-one informative characters were scored from restriction patterns obtained from the five endonucleases used. Twenty-seven different 16S rDNA genotypes were derived from restriction patterns (Table 1). Nineteen species could be assigned to a specific 16S rDNA genotype, supporting the application of this analysis for identification. Genetic distance matrices derived from 16S rRNA sequence data are listed in Table 2. In some cases, 0.1% sequence differences of species (V. cholerae and V. mimicus) were detected by PCR/RFLP (Tables 1 and 2). However, in some cases, 4.1% divergence was not detected (V. carchariae and V. vulnificus). Four pairs or groups showed the same 16S rDNA genotype and they were not separated from each other (Tables 1 and 2). They were (i) genotype 1: V. alginolyticus, V. campbellii, V. carchariae, V. harveyi, V. parahaemolyticus and V. vulnificus (similarity of 95.9–99.9%); (ii) genotype 4: V. anguillarum and V. ordalii (99.7%); (iii) genotype 9: V. pelagius and V. tubiashii (98.5%); and (iv) genotype 21: Photobacterium leiognathi, P. phosphoreum and V. iliopiscarius (96.4%). From these results, we estimate that the maximum range of the genetic distance of the 16S rRNA gene which could be included in the same genotypes by PCR/RFLP analysis using five restriction enzymes was 4.1% (Tables 1 and 2).

1

Restriction patterns of 16S rRNA genes digested with Dde I. Lanes: 1, V. vulnificus JCM 3726^T (restriction pattern 1); 2, V. anguillarum IFO 13266^T (pattern 2); 3, V. nigripulchritudo ATCC 27043^T (pattern 3); 4, V. metschnikovii IAM 1039 (pattern 4); 5, V. gazogenes ATCC 29988^T (pattern 5); 6, V. hollisae JCM 1284 (pattern 6); 7, V. fischeri ATCC 7744^T (pattern 7); 8, V. logei ATCC 15382 (pattern 8); 9, V. marinus ATCC 15381^T (pattern 9); 10, V. costicola NCMB 701^T (pattern 10). Lane M, size marker (Superladder-low, 100-bp ladder; GenSure Laboratories, Inc., Del Mar, CA, USA).

2

Restriction patterns of 16S rRNA genes digested with Hha I (A), Dde I (B), Rsa I (C), Sau 3AI (D), or Msp I (E). The lane assignments are given in Table 1 as different patterns by each five enzymes respectivity. Restriction fragments shorter than 99 bp were not used for analysis. Near-coincident double bands were identified by zone densitometoric analysis using an Atto Densitograph Imgaging Analyzer.

2

Levels of sequence similarities and evolutionary distances for 16S rDNAs of the family Vibrionaceae^a

^aThe values on the upper right are levels of sequence similarity, and the values on the lower left are corrected evolutionary distances. Accession numbers of sequences used for phylogenetic analysis: 1, V. aestuarianus (X74689); 2, V. alginolyticus (X74690); 3, V. anguillarum (X16895); 4, V. campbellii (X74692); 5, V. carchariae (X74693); 6, V. cholerae (X74694); 7, V. costicola (X74699); 8, P. damsela (X74700); 9, V. diazotrophicus (X74701); 10, V. fischeri (X74702); 11, V. fluvialis (X74703); 12, V. gazogenes (X74705); 13, V. harveyi (X74706); 14, V. hollisae (X74707); 15, V. logei (X74708); 16, V. marinus (X82142); 17, V. metschnikovii (X74711); 18, V. mimicus (X74713); 19, V. natriegens (X74714); 20, V. nigripulchritudo (X74717); 21, V. ordalii (X74718); 22, V. orientalis (X74719); 23, V. parahaemolyticus (X74720); 24, V. pelagius (X74722); 25, V. proteolyticus (X74723); 26, V. salmonicida (X70643); 27, V. splendidus (X74724); 28, V. tubiashii (X74725); 29, V. vulnificus (X74726); 30, P. angustum (D25307); 31, P. leiognathi (D25309); 32, P. phosphoreum (D25310).

A genetic distance tree constructed from the RFLP data set is shown in Fig. 3A. This tree, constructed using five endonucleases (Hha I, Dde I, Rsa I, Sau 3AI and Msp I), defined groups nearly identical to those inferred from 16S rRNA gene sequencing (Fig. 3B) and reported previously[8]. This method provided a more correct phylogenetic relationship between V. anguillarum and V. ordalii than partial sequence data[8]. However, the branch lengths and specific relationships among strains included in the same group were not in full accord with full and partial sequence data [8, 19]. We also obtained almost the same results using the smaller number of four restriction enzymes (Hha I, Dde I, Rsa I and Sau 3AI) (data not shown). It is important to determine the minimum number of restriction enzymes to examine the phylogenetic affiliation of the many isolates from environments using RFLP analysis of PCR amplified 16S rDNA. Laguerre et al.[10] reported that the four tetrameric restriction enzyme combinations were the minimum needed to discriminate among Rhizobium. Moyer et al.[18] reported that a combination of three tetrameric restriction enzymes (Hha I, Rsa I and Bst UI) gave good resolution based on the phylogeny of their computer-simulated groups. From these data, four or three endonuclease combinations are necessary to estimate approximate phylogenetic relationships among 16S rRNA sequences amplified from various environmental samples. In the present study, PCR/RFLP analysis using five tetrameric restriction enzymes was demonstrated to approximate phylogenetic relationships among members of the family Vibrionaceae.

3

Genetic distance trees based on the RFLP data set (A) or DNA sequence data (B) by the neighbor-joining method. ^aNumbers indicate 16S rDNA genotypes and abbreviations indicate the names of bacterial species (see Table 1). ^bNaming of the RFLP groups are corresponding to the 16S rRNA sequence groups constructed by Kita-Tsukamoto et al.[8].

The 16S rRNA sequences of two species pathogenic for marine animals, V. penaeicida and V. ichthyoenteri, have not been previously reported [5, 6]. In this study, the approximate phylogenetic position of these two species, estimated by our analysis, was in the main group of the genus Vibrio (Fig. 3A). V. iliopiscarius was previously suggested to be closely related to V. salmonicida based on partial 16S rRNA sequence[4]. To our surprise, V. iliopiscarius showed the same restriction pattern as P. phosphoreum and P. leiognathi, suggesting that it should be assigned to the genus Photobacterium. This is now supported by the full 16S rRNA sequence (accession number AB000278). Thus, this PCR/RFLP method could serve as a rapid tool to estimate the approximate phylogenetic relationship of isolates, without need for 16S rRNA sequencing.

3.3 Intraspecies divergence

Different restriction patterns among strains of a defined species were observed for V. fluvialis, V. anguillarum and P. phosphoreum. From these results, intraspecific relationships may need to be considered when interpreting PCR/RFLP analyze of 16S rRNA genes.

We obtained twenty-seven 16S rDNA genotypes from 35 species in this study, even though the 16S rRNA gene is a highly conserved one (Table 1). In recent years, PCR/RFLP analysis on the basis of the 16S-23S spacer region has also been increasingly used to resolve closely related isolates[20], since this spacer region is more variable than the 16S rRNA genes. Thus, it is likely that PCR/RFLP analysis of the region will resolve genotypes that cannot be distinguished by 16S rDNA.

3.4 Conclusions

RFLP analysis linked with DNA databases should be particularly helpful in studies requiring rapid examination of numerous isolates from different environments. However, two problems remain. First, we emphasize that many 16S rRNA sequences available in DNA databases are not complete and lack endpoints. Second, our data suggest that there may be significant intraspecies divergence. Thus, the 16S rDNA genotypes defined in this study (using almost full-length 16S rDNA) should contribute to more accurate genotyping, identification and simple classification of natural isolates belonging to the family Vibrionaceae.

Acknowledgements

We would like to thank Dr. A. Hiraishi for helpful suggestions and information.