PCR-Ribotyping of Xenorhabdus and Photorhabdus Isolates from the Caribbean Region in Relation to the Taxonomy and Geographic Distribution of Their Nematode Hosts

Abstract

The genetic diversity of symbiotic Xenorhabdus and Photorhabdus bacteria associated with entomopathogenic nematodes was examined by a restriction fragment length polymorphism analysis of PCR-amplified 16S rRNA genes (rDNAs). A total of 117 strains were studied, most of which were isolated from the Caribbean basin after an exhaustive soil sampling. The collection consisted of 77 isolates recovered from entomopathogenic nematodes in 14 Caribbean islands and of 40 reference strains belonging to Xenorhabdus and Photorhabdus spp. collected at various localities worldwide. Thirty distinctive 16S rDNA genotypes were identified, and cluster analysis was used to distinguish the genus Xenorhabdus from the genus Photorhabdus. The genus Xenorhabdus appears more diverse than the genus Photorhabdus, and for both genera the bacterial genotype diversity is in congruence with the host-nematode taxonomy. The occurrence of symbiotic bacterial genotypes was related to the ecological distribution of host nematodes.

Xenorhabdus and Photorhabdus spp. (Enterobacteriaceae) are symbiotically associated with the entomopathogenic nematodes (EPNs). Both partners of each bacterial-helminthic complex act together to kill insect prey by producing toxins and septicemia. Nematodes reproduce in the insect cadaver, feeding on the produced bacterial biomass and the insect tissues metabolized by the bacteria (4). When nematodes are harvested freshly from soil samples, all of the bacterial isolations from the intestinal contents of infective juveniles showed the presence of Xenorhabdus spp. in Steinernema spp. and Photorhabdus spp. in Heterorhabditis spp. It has been postulated that this high specificity is mainly due to the effect of a series of antimicrobial end products excreted by the symbiont itself during the multiplication of the nematodes in the parasitized insects (12). When infective juveniles escape the insect cadaver, they harvest symbiont cells in their intestine, securing among the generations the perenniality of the symbiotic association between both partners.

Axenic nematodes and symbionts are generally entomopathogenic by themselves, but the bacterial partner requires assistance from the nematode to achieve inoculation (the 50% lethal dose is usually less than 50 cells, depending on the test insects) (16). In natural conditions, symbionts are inoculated into the insect hemolymph by their host nematode, which acts as a living syringe on the target insects. In some symbioses, both partners must participate after inoculation to achieve pathogenesis, as with, for instance, the Steinernema glaseri-Xenorhabdus poinarii symbiosis (3).

Xenorhabdus and Photorhabdus spp. appear to display a high and monophyletic diversity: five Xenorhabdus species have been described (X. nematophilus, X. poinarii, X. bovienii, X. beddingii, and X. japonicus [6, 23]), and only one Photorhabdus species has been described (P. luminescens [11]). However, within the P. luminescens species several genomic groups have been recognized by DNA-DNA hybridization (7) and suggested by 16S rDNA sequencing (20, 31). Thus, the number of species could be underestimated.

In both genera, identification of new bacterial isolates or species is difficult because most strains are phenotypically very similar and fail to give positive results in many classical tests for identification (10) and because of a lack of sufficient members per taxon. Consequently, only a few species have been described, and some of these are represented by only a few isolates (6). Ecological data relating bacterial symbionts with their nematode host or their environment remain weak. Thus, studies on the taxonomic diversity and distribution of members of Xenorhabdus and Photorhabdus spp. are needed. As a first step, we recently used a PCR-restriction fragment length polymorphism (RFLP) method applied to the 16S rRNA gene to rapidly identify new isolates of both genera of symbionts (13). This approach distinguished Xenorhabdus and Photorhabdus species and identified groups as effectively as did DNA-DNA hybridization (7, 13). The method was applied on a limited number of bacterial strains obtained after isolation from nematode collections. In the current study, a larger and more comprehensive sampling of symbionts was undertaken in order to learn more about their ecological distribution relative to host taxonomy and environmental factors.

An exhaustive sampling was performed among 14 islands in the Caribbean basin. Caribbean islands are interesting because they are assumed to present limited soil imports and are subject to the same climate and because previous studies on native EPNs are available (8, 21, 27). The sampling is based on two surveys: (i) one is an exhaustive soil collection based on a grid map covering all the seven Guadeloupe islands (Grande Terre and Basse Terre, Marie-Galante, La Désirade, Petite Terre, Les Saintes, Saint-Martin, and Saint-Barthélemy) from which nematodes were trapped to study their distribution (14), and (ii) the other is a collection of nematodes recovered at random from seven other neighboring Caribbean islands (Martinique, Saint-Vincent, Cuba, Jamaica, Puerto Rico, the Dominican Republic, and Trinidad and Tobago) (19). The purpose of this study was to examine the diversity of EPN symbionts by PCR ribotyping (13) and to correlate the results to nematode taxonomy throughout the Caribbean islands and to the sampling environment (soil type, rainfall, elevation, and vegetal covering or crops) in the Guadeloupe islands. Finally, the isolates were compared to additional Xenorhabdus and Photorhabdus strains isolated from nematodes widely distributed throughout the rest of the world and from human clinical samples in order to determine whether Caribbean bacterial symbionts represent distinct genotypes of Xenorhabdus and Photorhabdus.

MATERIALS AND METHODS

Exhaustive nematode sampling in the Guadeloupe islands.

Nematodes were collected from the seven Guadeloupe islands. Exhaustive soil sampling was conducted between February and December of 1996. To harvest each sample, three portions of soil were chosen randomly in a 100-m² area. Each portion consisted of a core 5.5 cm in diameter at a 25-cm depth (ca. 0.6 dm³ of soil). The three portions were mixed, giving 1.8 dm³ from which 0.6 dm³ was used for recovering the nematodes as previously described (14). In all, 538 soil samples were collected according to a square grid with points spaced at 2-km intervals and covering the whole surface of the Guadeloupe islands. EPNs were isolated by using the Galleria trap technique (9) and identified by using morphological criteria, isozyme analysis, and satellite DNA probes (17, 19). The presence of nematodes was related to site location, elevation, rainfall, soil type, and vegetation (14).

Random sampling in other Caribbean islands.

Nematodes were also collected from the seven other Caribbean islands listed above. In contrast to the exhaustive Guadeloupe survey, these samples were collected randomly from a variety of ecosystems, including croplands, orchards, grasslands, salt marches, and forests. On Martinique, Saint-Vincent, Trinidad and Tobago, and Jamaica, samples were collected by the Institut National de la Recherche Agronomique (INRA) laboratory in Guadeloupe. From the three other Caribbean islands, nematodes were collected and identified by E. Arteaga and M. Montes (Ministerio de Agricultura, Estación Nacional de Sanidad de los Cítricos, La Habana, Cuba), W. Figueroa (University of Puerto Rico, Río Pedras, Puerto Rico), and L. Garrido and A. Carro (Universidad Autónoma de Santo Domingo, Engombe, Santo Domingo). After delivery of the biological material, the taxonomic position of this nematode collection was checked again (19) by using the methods mentioned above.

The names of the nematode species cited in this study follow recent modifications of the International Code of Zoological Nomenclature. Therefore, unlike the previous descriptions, the following names will be used in this study: S. affine, S. arenarium, S. cubanum, S. intermedium, S. puertoricense, S. rarum, S. riobrave, and H. indica instead of S. affinis, S. anomalae, S. cubana, S. intermedia, S. puertoricencis, S. rara, S. riobravis, and H. indicus, respectively (18).

Bacterial isolates and reference strains.

Individual bacterial colonies were isolated from the infective stages by the hanging-drop technique (25). We examined 77 isolates from the Caribbean area, including 31 strains from the Guadeloupe islands and 46 strains from the remaining Caribbean islands (Table 1). To identify the isolates, their phenotypic properties and RFLP patterns of amplified 16S ribosomal DNA (rDNA) were compared to those of 27 reference strains previously studied (10, 13). Thirteen other reference strains, including seven Xenorhabdus and six Photorhabdus strains, were added (Table 2). The seven new Xenorhabdus strains included strain USFL52 of X. bovienii, strain SK72 of X. poinarii, and five other Xenorhabdus spp. that originated from different parts of the world and from new species of host nematodes, i.e., Steinernema monticolum (29), S. scapterisci, S. serratum, S. kushidai, and S. riobrave (Table 2). Five opportunistic Photorhabdus spp. isolated from human patients at the Centers for Diseases Control (Atlanta, Ga.) (15) and strain Q614, which is the only known nonluminescent Photorhabdus strain, were also examined (5).

TABLE 1.

List of the 77 bacterial isolates used in this study

Island group and isolatea	Geographical origin	Host nematode origin	Source of nematodesb	16S rDNA genotypec
Guadeloupe islands
Photorhabdus isolates
FRG03-17	Grande Terre	H. indica	H. Mauléon	12
FRG19	Basse Terre	H. indica	H. Mauléon	12
FRG20	Saint-Barthélemy	H. indica	H. Mauléon	12
FRG24-25	Saint-Martin	H. indica	H. Mauléon	12
FRG28	La Désirade	H. indica	H. Mauléon	12
FRG33	Les Saintes	H. indica	H. Mauléon	12
FRG35	Marie Galante	H. indica	H. Mauléon	12
FRG01	Basse Terre	H. bacteriophora	H. Mauléon	13
FRG02	Grande Terre	H. bacteriophora	H. Mauléon	13
FRG18	Petite Terre	H. indica	H. Mauléon	27
FRG21-23	Saint-Barthélemy	H. indica	H. Mauléon	27
FRG27	Saint-Martin	H. indica	H. Mauléon	27
FRG26	Saint-Martin	Heterorhabditis sp.d	H. Mauléon	28
FRG29	Basse Terre	H. bacteriophora	H. Mauléon	28
Xenorhabdus isolates		None
Other Caribbean islands
Photorhabdus isolates
P2M	Cuba	H. indica	E. Arteaga	12
DO02-04 and -07; DO09-10, -12, and -14; and DO18-21 and 23-25	Dominican Republic	H. indica	L. Garrido	12
DO13	Dominican Republic	Heterorhabditis sp.d	L. Garrido	12
FRM03	Martinique	H. indica	H. Mauléon	12
PR02-B, PR05, PR06-C, PR14, PR16 to PR19, PR21 and PR22, PR27-A, PR27-B, PR38, PR54, PR60, and PR63	Puerto Rico	H. indica	W. Figueroa	12
DO01 and DO08	Dominican Republic	H. bacteriophora	L. Garrido	13
PR02-A	Puerto Rico	H. bacteriophora	W. Figueroa	13
TT01 and TT03	Trinidad	H. bacteriophora	H. Mauléon	13
JM12	Jamaica	H. indica	H. Mauléon	27
FRM05	Martinique	H. indica	H. Mauléon	27
Xenorhabdus isolates
CU01	Cuba	S. cubanum	E. Arteaga	3
JM26	Jamaica	S. bicornutum	H. Mauléon	25
PR06-A	Puerto Rico	S. puertoricense	W. Figueroa	19
VC01	Saint-Vincent	Steinernema sp. 1e	H. Mauléon	20
FRM16	Martinique	Steinernema sp. 2e	H. Mauléon	21

^aNames are given according to an international nomenclature that indicates first the international code of the home country followed by a number; strain P2M, however, is a strain designation kept due to several previous descriptions (CU, Cuba; DO, Dominican Republic; FRG, France, Guadeloupe [corresponds to HG used in Constant et al. (14)]; FRM, France, Martinique; JM, Jamaica; PR, Puerto Rico; TT, Trinidad and Tobago; and VC, Saint-Vincent). 

^bE. Arteaga, Estación de Sanidad de los Cítricos, Ministerio de Agricultura, Ciudad Habana, Cuba; L. Garrido, Universidad Autónoma de Santo Domingo, Santo Domingo, Dominican Republic; W. Figueroa, University of Puerto Rico, Río Piedras, Puerto Rico; and H. Mauléon, INRA, Unité de Recherches en Productions Végétales, Guadeloupe, France. 

^c16S rDNA genotypes are defined in Table 3. 

^dHeterorhabditis sp. not yet identified. 

^eNew species of Steinernema not yet described. 

TABLE 2.

List of the 43 Xenorhabdus, Photorhabdus, and other Enterobacteriaceae strains used as references in RFLP analysis of 16S rRNA genes

Straina	Host nematode origin	Geographical origin	Source	16S rDNA genotypeb
X. nematophilus
AN6 (ATCC 19061)T	S. carpocapsae	Georgia	R. Akhurst	1
A24	S. carpocapsae	Russia	R. Akhurst	2
F1	S. carpocapsae	France	C. Laumond	2
X. poinarii
G6 (ATCC 49121)T	S. glaseri	North Carolina	R. Akhurst	3
NC33	S. glaseri	North Carolina	R. Akhurst	3
SK72	S. glaseri	Florida	N. Simões	3
X. beddingii Q58 (UQM 2872)T	Steinernema sp.3c	Queensland, Australia	R. Akhurst	5
X. bovienii
T228 (UQM 2211)T	S. feltiae	Tasmania, Australia	R. Akhurst	6
USFL52	S. feltiae	Florida	G. Smart	6
Si	S. intermedium	South Carolina	R. Akhurst	7
F3	S. affine	France	C. Laumond	8
SK2	S. kraussei	Czech Republic	Z. Mrácek	8
F5	S. feltiae	France	C. Laumond	8
X. japonicus JP02	S. kushidai	Japan	N. Ishibashi	18
Xenorhabdus spp.
K77	S. rarum	Argentina	M. De Doucet	4
SaV	S. arenarium	Russia	R. Akhurst	9
KR1	S. monticolum	Korea	P. Stock	22
UY61	S. scapterisci	Uruguay	G. Smart	23
CN01	S. serratum	South China	J. Liu	24
USTX62	S. riobrave	Texas	E. Cabanillas	26
P. luminescens
Hb (ATCC 29999)T	H. bacteriophora	Victoria, Australia	R. Akhurst	10
Hm	Heterorhabditis sp.d	Wisconsin	K. Nealson	10
C8406	Heterorhabditis sp.d	Haï Nan Island, China	R. Akhurst	11
IS5	H. indica	Israel	I. Glazer	12
K80	Heterorhabditis sp.d	Argentina	M. De Doucet	13
HP88	H. bacteriophora	Ohio	R. Akhurst	13
X1Nach	H. megidis (NWE group)e	Russia	R. Akhurst	14
HW79	H. megidis (NWE group)e	The Netherlands	P. Westerman	14
ItH211	Heterorhabditis sp.	Italy	K. Deseo	14
HL81	H. megidis (NWE group)e	The Netherlands	P. Westerman	14
Meg	H. megidis	Ohio	R. Akhurst	15
C1 (ATCC 29304)	H. bacteriophora (H. heliothidis)	California	R. Akhurst	16
NZH	H. zealandica	New Zealand	W. Wouts	17
D1	H. indica	North Territory, Australia	R. Akhurst	27
1216-79	Clinical (ATCC 43948)	CDC, Atlanta, Ga	J. Farmer	29
2407-88	Clinical (ATCC 43952)	CDC, Atlanta, Ga.	J. Farmer	29
2617-87	Clinical (ATCC 43951)	CDC, Atlanta, Ga.	J. Farmer	29
3105-77	Clinical (ATCC 43949)	CDC, Atlanta, Ga.	J. Farmer	29
3265-86	Clinical (ATCC 43950)	CDC, Atlanta, Ga.	J. Farmer	29
Q614	Heterorhabditis sp.d	Queensland, Australia	R. Akhurst	30
Other Enterobacteriaceae
Proteus vulgaris CIP 58.60T		Institut Pasteur, Paris, France		31
Escherichia coli CIP 54.8T		Institut Pasteur, Paris, France		32
Serratia marcescens CIP 103235T		Institut Pasteur, Paris, France		33

^aNames are given according to the usual nomenclature of the strains as described in previous studies, except for the new designations, which indicate first the international code of the home country followed by a number (CN, China; JP, Japan; KR, Korea; USFL, United States, Florida; USTX, United States, Texas; and UY, Uruguay). Abbreviations: ATCC, American Type Culture Collection (Rockville, Md.); CIP, Collection de l’Institut Pasteur (Paris, France); UQM, Culture Collection of the University of Queensland Department of Microbiology (Brisbane, Australia). A superscript “T” indicates a type strain. 

^b16S rDNA genotypes are defined in Table 3. 

^cNew species of Steinernema not yet described. 

^dHeterorhabditis sp. not yet identified. 

^eNWE, Nematodes belonging to the North West European group. 

Phenotypic characterization.

To verify in a preliminary step that the isolates belonged to the Xenorhabdus and Photorhabdus genera, we used conventional phenotypic criteria (10) and compared the results with reference strains. All of the tests were conducted at 28°C. Cellular morphology and motility were assessed by microscopic examination of 24-h-old nutrient broth cultures. Dye adsorption of bromothymol blue was tested on nutrient agar supplemented with 0.004% (wt/vol) triphenyltetrazolium chloride and 0.0025% (wt/vol) bromothymol blue (NBTA medium) for Xenorhabdus isolates (2), and dye adsorption of neutral red was tested on MacConkey agar for Photorhabdus isolates (10). Antimicrobial activity was determined by the method of Akhurst (1) with Micrococcus luteus as the indicator microorganism. Other tests included the use of API 20E and API 20NE strips (Biomerieux, Craponne, France); catalase; bioluminescence; phospholipase (lecithinase); lipolysis on Tween 20, 40, 60, 80, and 85; and pigmentation (10).

Nucleic acid extraction.

Cells were grown on nutrient agar plates for 48 h at 28°C, scraped off in TE8 buffer (50 mM Tris-HCl, 20 mM EDTA; pH 8), and centrifuged in a microcentrifuge tube for 2 min at 10,000 × g. The cell pellets were washed twice in TE8 buffer and stored at −20°C. DNA extraction was performed with the nucleic acid extraction kit Isoquick (ORCA Research, Inc., Bothell, Wash.) according to the rapid DNA extraction protocol of the manufacturer. The DNA pellets were dissolved in 100 μl of pure water and diluted 20- to 100-fold to be used as templates for PCR.

16S rDNA restriction analysis.

16S rDNAs were amplified by using the primers and reaction conditions previously described (13). For each isolate, 6 to 17 μl of amplified 16S rDNA was digested overnight with 5 U of restriction endonuclease (GIBCO BRL, Cergy-Pontoise, France). PCR products of Xenorhabdus isolates and reference strains were digested separately with six tetrameric endonucleases previously found to produce polymorphic digests (13): CfoI, HinfI, DdeI, AluI, HaeIII, and MspI. Amplified DNAs from the Photorhabdus isolates were digested with three endonucleases (AluI, CfoI, and HaeIII) which were sufficient to generate all the genotypes of Photorhabdus strains previously studied (13). DNA digests were then analyzed by horizontal electrophoresis at 6 V/cm in 3% (wt/vol) Nusieve or Metaphor agarose (Tebu, Le Perray en Yvelines, France) gels in 0.5× TBE buffer (44.6 mM Tris-base; 44.6 mM boric acid; 1 mM EDTA; pH 8) containing 0.5 mg of ethidium bromide per liter. The gels were visualized under UV light with an imager (The Imager, software version 2.03; Appligene, Inc., Strasbourg, France). Genetic relationships between two amplified 16S rRNA genes were evaluated by determining the presence or absence of DNA restriction fragments of a given length. Dice’s similarity coefficient, based on the proportion of shared restriction fragments, was calculated, and the distance matrix was determined by the Nei and Li method (22). Distance values were displayed as a dendrogram by using the unweighted-pair-group method with arithmetic means (UPGMA) (28).

RESULTS

Isolation and phenotypic characterization of bacterial symbionts.

All bacterial isolates from the Guadeloupe and other Caribbean island surveys shared the common phenotypic properties with all reference strains (data not shown) and therefore belonged to Xenorhabdus and Photorhabdus genera (10).

Of the 538 soil samples collected from the seven Guadeloupe islands, 31 (5.8%) contained Heterorhabditis nematodes (Fig. 1), from which 31 Photorhabdus strains were isolated (Table 1). No Steinernema spp. were found, and therefore no Xenorhabdus spp. were obtained from the Guadeloupe islands. Of the Heterorhabditis nematodes, 27 were H. indica (87%), three were H. bacteriophora (10%), and one was not identified (3%) (Table 1). H. indica was mainly located in coastal areas and rarely inland (Fig. 1): 25 isolates originated from soil marshes, sandy beaches, and the slopes of a limestone cliff (soil pH, 8.0 to 9.3; elevation, 0 to 75 m), and 2 isolates originated from pastures (vertisol at pH 6.5 to 7.5; elevation, ≤240 m). The three H. bacteriophora were found in cropland soil (vertisol at pH 6.5 to 7.5; elevation, ≤25 m), orchard soil (sand at pH 9; elevation, ≤25 m), and rainforest soil (oxisol at pH 5.5; elevation, ≤350 m). No other inland areas provided any EPNs (Fig. 1).

FIG. 1.

Map grid showing the exhaustive soil sampling in Guadeloupe islands (small points). The sites where entomopathogenic bacterium-nematode complexes were harvested are indicated by symbols identifying the nematode species: ★, H. indica; ✷, H. bacteriophora; and ✠, Heterorhabditis sp. The symbiont strain and genotype number are shown beside each symbol.

From the seven other Caribbean islands, 46 symbionts were isolated, including 41 Photorhabdus and five Xenorhabdus spp. Photorhabdus sp. was found on all of the islands except Saint-Vincent (Table 1). Of the Photorhabdus isolates, 35 were isolated from H. indica, 5 were from H. bacteriophora, and 1 was from an unidentified Heterorhabditis sp. Of the five Xenorhabdus isolates, one originated from S. cubanum (in Cuba), one originated from S. bicornutum (in Jamaica), one originated from S. puertoricense (in Puerto Rico), and two were from two new species of Steinernema not yet described (in Martinique and Saint-Vincent) (Table 1). In the Dominican Republic and Puerto Rico, a species prevalence similar to that for the Guadeloupe islands was observed, allowing us to determine that H. indica was also dominant (15 of 18 isolates in the Dominican Republic and 16 of 17 isolates in Puerto Rico) and that H. bacteriophora was rarely represented (2 of 18 isolates and 1 of 17 isolates, respectively). However, in contrast to the Guadeloupe survey, H. indica and consequently its Photorhabdus symbiont were isolated from the inland part of Puerto Rico and the Dominican Republic and were not restricted to the coastal areas.

Amplified 16S rDNAs and RFLP data.

In all, 72 Photorhabdus, 5 Xenorhabdus, and 43 reference strains (including 20 Xenorhabdus and 20 Photorhabdus spp. and 3 other genera of Enterobacteriaceae) were further investigated by PCR ribotyping. 16S rDNA genes of all 120 strains were amplified by using PCR primers representing regions of the 16S rDNA conserved in bacteria. All of the strains produced a single band of about 1,600 bp. Polymorphic restriction patterns were obtained with the six endonucleases used. Results of the Photorhabdus and Xenorhabdus patterns are presented in Fig. 2 except for the CfoI patterns that were the same as those shown previously (13). By combining all of the restriction patterns, the 120 strains could be grouped into 33 genotypes (Table 3).

FIG. 2.

Photographic (A) and schematic (B to E) restriction patterns of PCR-amplified 16S rRNA genes from Xenorhabdus and Photorhabdus strains digested with the following enzymes: DdeI (A), HinfI (B), HaeIII (C), MspI (D), and AluI (E). The lane assignments correspond to the patterns given in Table 3. Lanes m, molecular-weight marker VIII (Boehringer Mannheim).

TABLE 3.

Genotypes and restriction patterns revealed by RFLP analysis of PCR-amplified 16S rRNA genes

RFLP genotypea	Representative strain	Species	Restriction pattern of amplified 16S rRNA genes digested with:
			CfoI	HinfI	HaeIII	MspI	DdeI	AluI
1	AN6T	X. nematophilus	C1	Hf1	H1	M1	D4	A1
2	A24	X. nematophilus	C1	Hf1	H1	M7	D4	A1
3	G6T	X. poinarii	C1	Hf1	H8	M2	D5	A3
4	K77	Xenorhabdus sp.	C1	Hf1	H4	M2	D5	A3
5	Q58T	X. beddingii	C1	Hf1	H1	M8	D5	A2
6	T228T	X. bovienii	C2	Hf1	H11	M2	D3	A1
7	Si	X. bovienii	C2	Hf2	H11	M2	D3	A1
8	F3	X. bovienii	C1	Hf1	H11	M2	D3	A1
9	SaV	Xenorhabdus sp.	C8	Hf4	H8	M2	D11	A1
10	HbT	P. luminescens	C5	Hf3	H12	M3	D1	A5
11	C8406	P. luminescens	C5	Hf3	H12	M4	D1	A3
12	IS5	P. luminescens	C4	Hf3	H12	M3	D6	A3
13	HP88	P. luminescens	C1	Hf3	H13	M3	D14	A3
14	X1Nach	P. luminescens	C3	Hf1	H13	M4	D2	A4
15	Meg	P. luminescens	C1	Hf1	H2	M4	D7	A6
16	C1	P. luminescens	C4	Hf1	H2	M4	D7	A6
17	NZH	P. luminescens	C1	Hf1	H14	M4	D7	A4
18	JP02	Xenorhabdus sp.	C1	Hf2	H1	M1	D4	A1
19	PR06-A	Xenorhabdus sp.	C1	Hf1	H4	M2	D4	A1
20	VC01	Xenorhabdus sp.	C1	Hf1	H4	M2	D13	A1
21	FRM16	Xenorhabdus sp.	C1	Hf1	H4	M4	D8	A10
22	KR1	Xenorhabdus sp.	C1	Hf2	H19	M2	D15	A1
23	UY61	Xenorhabdus sp.	C1	Hf1	H20	M12	D16	A1
24	CN01	Xenorhabdus sp.	C1	Hf1	H21	M2	D11	A3
25	JM26	Xenorhabdus sp.	C1	Hf5	H15	M10	D12	A3
26	USTX62	Xenorhabdus sp.	C1	Hf5	H15	M10	D5	A3
27	JM12	P. luminescens	C4	Hf3	H16	M3	D6	A3
28	FRG26	P. luminescens	C1	Hf3	H12	M11	D14	A3
29	1216-79	P. luminescens	C1	Hf1	H18	M4	D1	A11
30	Q614	P. luminescens	C4	Hf3	H17	M4	D1	A4
31	CIP 5860T	Proteus vulgaris	C7	Hf2	H6	M6	D9	A8
32	CIP 548T	Escherichia coli	C6	Hf1	H5	M5	D10	A7
33	CIP 103235T	Serratia marcescens	C1	Hf1	H7	M5	D8	A9

^aRFLP genotypes numbered 1 to 17 and 31 to 33 were previously identified (13), and RFLP genotypes numbered 18 to 30 were defined in this study. 

RFLP analysis of Photorhabdus isolates and reference strains.

The 72 Caribbean Photorhabdus strains were divided into four genotypes (numbered 12, 13, 27, and 28 in Table 1) based on their RFLP patterns (Fig. 2).

A subsample of eight Photorhabdus isolates belonging to each of the four genotypes defined above was typed with three additional tetrameric enzymes (HinfI, MspI, and DdeI). No more polymorphism was observed.

RFLP patterns from the Caribbean strains were compared to those of 20 Photorhabdus reference strains, of which 14 had been previously typed (13). The six new reference strains (1216-79, 2407-88, 2617-87, 3105-77, 3265-86, and Q614) were separated into two newly defined genotypes (Table 2): genotype 29 included five opportunistic Photorhabdus clinical strains, which were all identical, and genotype 30 included the unique strain Q614 from Australia. Thus, among all the P. luminescens strains tested, 12 genotypes were defined (genotypes 10 to 17 and 27 to 30 [Table 3]), 4 of which were novel. Three restriction enzymes (AluI, CfoI, and HaeIII) were sufficient to resolve them.

Three of the four genotypes observed in Caribbean Photorhabdus isolates were also observed among Photorhabdus reference strains originating from other parts of the world: genotype 12 was represented by the reference strain IS5, a symbiont of H. indica from Israel; genotype 27 matched the genotype of strain D1, a symbiont of H. indica from Australia; and genotype 13 corresponded to those of the symbiotic strains HP88 and K80 from H. bacteriophora and Heterorhabditis sp., respectively. In contrast, genotype 28 was specific to two Photorhabdus isolates from the Guadeloupe islands and was not observed among the reference strains.

RFLP analysis of Xenorhabdus isolates and reference strains.

The amplified 16S rDNAs from the five Caribbean Xenorhabdus isolates (CU01, JM26, PR06-A, VC01, and FRM16) were analyzed with six endonucleases and were compared to the genotypes of 20 Xenorhabdus reference strains, of which 13 had been analyzed previously and shown to belong to nine 16S rDNA genotypes (13). By combining the different restriction patterns (Fig. 2; Table 3), each of the five Caribbean isolates belonged to a distinct 16S rDNA genotype, four of which were novel (Table 1). The seven newly investigated Xenorhabdus reference strains (SK72, USFL52, JP02, KR1, UY61, CN01, and USTX62) were grouped into seven different 16S rDNA genotypes (named genotypes 3, 6, 18, 22 to 24, and 26), five of which were novel (Table 2). In all, 18 Xenorhabdus genotypes were defined, half of which were novel (Table 3). The Caribbean isolate CU01, obtained from S. cubanum, and the reference strain SK72, a symbiont of S. glaseri, were identical to the type strain (G6^T) of the species X. poinarii. The new reference strain USFL52 from Florida exhibited the same pattern as T228^T, the type strain of X. bovienii.

Genetic relationships between amplified 16S rRNA genes.

Comparison of the restriction profiles obtained with Photorhabdus and Xenorhabdus isolates revealed 30 distinctive genotypes. The three additional genotypes in Table 3 correspond to the other Enterobacteriaceae. To estimate the genetic relationships between PCR-amplified 16S rDNAs, we calculated a matrix of pairwise genetic distances for the 33 genotypes defined with the six restriction enzymes (Table 3). A mean of 33 restriction fragments per genotype was analyzed. The distance matrix was used to construct a dendrogram based on a UPGMA algorithm (Fig. 3). Twenty-three of the 16S rDNA genotypes were represented by only one to two strains, and the remaining seven were represented by multiple strains (Fig. 3). Two major groups (I and II) were delineated at a genetic distance of 0.038 and corresponded to Xenorhabdus and Photorhabdus genera, respectively. As expected, the three other Enterobacteriaceae genera branched apart from these two groups.

FIG. 3.

Cluster analysis (UPGMA) of the 33 PCR-RFLP genotypes of 16S rDNA defined in Table 3. The name of the representative strain and the number of strains which had the same genotype are in parentheses.

Within group I, three subgroups (I-a, I-b, and I-c) were shown (Fig. 3). The first one, I-a, comprised 20 of the 25 Xenorhabdus strains examined, including the 5 described species. Strain JP02, from S. kushidai, was grouped with the X. nematophilus species at a very low genetic distance of 0.005. Xenorhabdus isolates PR06-A, from S. puertoricense (Puerto Rico), and VC01, from Steinernema sp. (Saint-Vincent), clustered near the type strain of X. poinarii and strain K77 from S. rarum. Strain KR1, from S. monticolum, clustered near the X. bovienii. Strain FRM16, a Xenorhabdus isolate from Martinique, could not be grouped specifically with any other strain. The second subgroup, I-b, included three genotypes, each represented by a single strain isolated from the nematodes: S. scapterisci, S. serratum, and S. arenarium. A relatively high genetic divergence of 1.3% was found among them. The third separate subgroup, I-c, branched far apart from subgroups I-a and I-b at the limit of the genus Xenorhabdus. It was composed of only two strains: JM26, from S. bicornutum (Jamaica), and USTX62, from S. riobrave (Texas), which were closely related (0.35% divergence).

Group II of the Photorhabdus bacteria was divided into two subgroups, II-a and II-b (Fig. 3). Within subgroup II-a the type strain of P. luminescens clustered with the four genotypes encountered in the Caribbean basin (genotypes 12, 13, 27, and 28). Among them, genotypes 12 and 27, from symbionts of H. indica, were closely related (0.2% divergence). Genotypes 13 and 28 from symbionts of H. bacteriophora were less related (0.6% divergence). Subgroup II-b encompassed more divergent genotypes, including symbionts of H. megidis and H. zealandica (genotypes 14 to 17), the clinical strains (genotype 29), and the strain Q614 (genotype 30).

Bacterial genotype distribution in relation to host nematodes and geography.

We isolated 72 Photorhabdus and 5 Xenorhabdus spp. from the Caribbean basin. Among the Photorhabdus spp. four genotypes were identified. All of the 63 isolates of the genotypes 12 and 27 originated from H. indica. The two reference strains IS5 and D1, isolated from H. indica, also shared these genotypes. Genotype 12 was therefore the most prevalent (56 of 63 isolates) of the Photorhabdus genotypes, and it occurred throughout the Caribbean region. It was restricted to the coastal areas in the Guadeloupe islands (Fig. 1), but it also occurred inland in the Dominican Republic and Puerto Rico. Genotype 27 was rare and found only in Petite Terre, Jamaica, Martinique, and in the northern Guadeloupe islands (Saint-Martin and Saint-Barthélemy).

All seven isolates of genotype 13 originated from H. bacteriophora and matched the reference H. bacteriophora strain HP88. Three other genotypes (genotypes 10, 16, and 28) also were found in isolates from H. bacteriophora (strains Hb, C1, and FRG29). Genotype 28 was new, was restricted to Guadeloupe, and was not detected among the reference strains collected in the rest of the world.

Steinernema spp. were very scarce in the Caribbean basin and therefore could not be related to geographical origins. The five Xenorhabdus isolates corresponded to five different genotypes, and each of them was isolated from a different species of Steinernema: S. cubanum, S. bicornutum, S. puertoricense, Steinernema sp.1, and Steinernema sp.2, the two latter being new species not yet described.

DISCUSSION

Based on RFLP analysis of 16S rDNA from bacterial symbionts, host nematode characterization, and the geographical distribution of genetic bacterial groups, we described here the genetic composition of EPN symbiotic bacteria in the Caribbean basin and compared them to symbiotic strains collected at various localities throughout the world.

The sampling intensity allowed us to estimate a higher degree of diversity compared with the Xenorhabdus and Photorhabdus strains studied previously. Thirteen new genotypes were detected and successfully added to the 17 previously defined ones (13); this was done without substantially altering the phylogenetic relationships established previously by clustering analysis. Thus, PCR-RFLP analysis applied to 16S rDNA proved to be a rapid and sensitive typing method for distinguishing strains of the Xenorhabdus and Photorhabdus genera. Because we found new genotypes and new restriction patterns among Xenorhabdus and Photorhabdus spp., the number of endonucleases required to generate all of the genotypes has to be reconsidered. HaeIII, CfoI, and AluI have to be used to differentiate all of the Photorhabdus genotypes, and five restriction enzymes (CfoI, HinfI, MspI, HaeIII, and DdeI) are necessary in order to distinguish all of the Xenorhabdus genotypes.

The addition of new genotypes in the 16S rDNA clustering analysis revealed an unusually high level of genetic diversity within the Xenorhabdus genus compared to previous descriptions (13, 20, 26, 30, 31). This development likely resulted from the large number of studied strains originating from 16 identified Steinernema species and from various localities worldwide. For instance, S. arenarium, S. bicornutum, S. scapterisci, and S. serratum, whose symbionts were not previously typed, proved to harbor divergent Xenorhabdus symbionts that were distantly related to described species. Most of the genotypes were so divergent that they may represent new species. However, due to a lack of similar bacterial isolates, new Xenorhabdus species could not be described. Compared to 16S ribosomal sequencing studies (20, 30, 31), the phylogenetic position of the symbiont of S. kushidai, which is closely related to X. nematophilus, was corroborated, whereas the phylogenetic position of the symbiont of S. riobrave was different. Because novel Xenorhabdus strains were detected, the complete 16S rRNA genes should be sequenced in order to refine the phylogenetic tree of the Xenorhabdus genus.

Clustering analysis of the Photorhabdus genotypes revealed two major subgroups corresponding to the host nematodes and their ecological data. The first subgroup, II-a, included symbionts of H. indica and H. bacteriophora, which were found in the Caribbean and other tropical regions, whereas the second subgroup, II-b, included symbionts of H. megidis and H. zealandica, which were limited to the temperate regions. Previous 16S rDNA sequencing analyses corroborate the delineation between symbionts of H. indica and H. bacteriophora and the symbionts of H. megidis (20, 31). Moreover, within subgroup II-a, the similarity between symbionts of H. indica and H. bacteriophora is also corroborated by ribosomal sequencing (20).

Because of a higher number of isolates and a precise identification of their symbiotic nematodes that were not available in our previous data (13), a clear relationship between 16S rDNA genotypes and Heterorhabditis species origins was detected. Thus, Photorhabdus genotypes 12 and 27 were exclusively associated with H. indica, whereas Photorhabdus genotypes 13 and 28 were only associated with H. bacteriophora. Yet in two cases (strains Hb^T and C1) an inconsistency was observed. Hb^T and C1 were isolated from nematodes initially named H. bacteriophora but which are now known to differ from the typical H. bacteriophora represented by HP88 (18). The native host-nematodes of Hb^T and C1 may belong to two distinct species or subspecies that are different from the species of H. bacteriophora associated with genotypes 13 and 28. Because the identification of Heterorhabditis spp. is difficult, the characterization of their bacterial symbionts may help resolve some difficult taxonomic questions regarding their hosts.

Geographical grouping of the Photorhabdus genotypes was linked to nematode distribution. Bacterial genotypes associated with H. indica are restricted to tropical areas as is their host H. indica (24), whereas genotypes associated with H. bacteriophora seem to be more homogeneously distributed, as is their host H. bacteriophora (18). Furthermore, in the Guadeloupe islands, most of the H. indica isolates were found in coastal sandy soils, and all of the H. bacteriophora were found in the vertisols of croplands and the oxisols of forests. These results agree with previous studies indicating that Heterorhabditis spp. mainly occur in coastal areas (18). However, some H. indica nematodes were isolated from inland soils in Puerto Rico and the Dominican Republic, possibly as a result of soil material transfer on these relatively developed islands. Both genotypes associated with H. indica (genotypes 12 and 27) were spread throughout the Caribbean basin, suggesting that the host species is the predominant determinant of geographic distribution. To evaluate more accurately the selective pressure applied by the host nematode versus those that might be applied by soil factors on symbiont populations, further studies on the possible occurrence of free-living Xenorhabdus and Photorhabdus isolates in soil are required.

The high degree of Xenorhabdus diversity is congruent with the wide diversity of associated Steinernema nematodes and, with only one exception, the genotypes reflect the nematode host species. This exception is represented by the genotype of X. poinarii, which is associated with two nematode species: S. glaseri and S. cubanum. These two nematode species are closely related because they share morphological and ITS-based similarities (18). A complementary study of the symbionts by phenotypic characterization and DNA-DNA hybridization is in progress to confirm this finding. If verified, this would be the second reported case of a Xenorhabdus species associated with different Steinernema species, X. bovienii associated with S. feltiae, S. kraussei, and S. affine that occur in the same region and environment (11).

Molecular tools, such as 16S rDNA PCR-RFLP analysis for bacterial typing, along with satellite DNA probes and isozyme analysis for EPN identification, are fast and accurate ways of comparing bacterium-nematode associations on a large geographical scale. A high level of taxonomic congruence has been detected by using this approach between symbiont pairs, a finding that supports an early coevolution of these symbioses. The perenniality of the association may have resulted from the vertical transmission of symbiotic bacteria during monoxenic sepsis produced during parasitism and from an early intestinal contamination of infective juveniles escaping insect cadavers. Each species of nematode seems to secure a very restricted microbial niche that is more or less specific to a particular Xenorhabdus or a Photorhabdus species.