An AFLP approach to identify genetic markers associated with resistance to vibrio vulnificus and Perkinsus marinus in eastern oysters.

ABSTRACT We conducted an arbitrary fragment length polymorphism (AFLP) analysis followed by tests of association to search for potential genetic markers associated with resistance to infection by Vibrio vulnificus and Perkinsus marinus in a sample of North Carolina oysters. We used 48 AFLP markers and found significant associations for two of these markers with the incidence of infection of P. marinus and two separate markers associated with the magnitude of infection of V. vulnificus. Further, there was evidence of epistatic interactions of genes affecting infection levels of both pathogens. This suggests existence of genes or groups of genes that are located on chromosome fragments close to the identified AFLP markers and that act directly or indirectly (through epistasis) to control the levels of infection by these microorganisms in oysters. The impact of the direct and epistatic effects of these unknown genes on infection level variability amounted to about 40% for both P. marinus and V. vulnificus. This study demonstrates the utility of the AFLP approach to identify genetic markers of pathogen/parasite resistance in eastern oysters. However, the identified AFLP markers are considered preliminary and suggestive only because of the limited sample size analyzed in this study. Further studies using this approach on a larger sample size are required to identify a set of robust molecular markers that would serve in a marker-supported breeding program designed to improve the quality of the oyster stock.

KEY WORDS: oysters, Vibrio vulnificus, Perkinsus marinus, AFLP markers, Crassostrea virginica

INTRODUCTION

The eastern oyster Crassostrea virginica (Gmelin) is an economically and ecologically important bivalve species in eastern United States estuaries that serves as a host and a vector for a variety of prokaryotic and eukaryotic parasites. Among the parasites and pathogens that have the strongest impact on the oyster industry and wild oyster populations are a bacterium, Vibrio vulnificus, and a protozoan, Perkinsus marinus. V. vulnificus is a human pathogen common in estuarine waters around the world, and oysters are major vectors in the transfer of this bacterium to humans. Ingestion of raw or undercooked oysters containing this bacterium can result in illness and even death (Oliver & Kaper 2003). Although this bacterium does not negatively affect oyster survival and health, it causes significant losses to the oyster industry because of the required warning labels and associated negative publicity (Keithly & Diop 2001) and is a serious concern for public health in the United States, particularly in Gulf of Mexico states. In contrast, P. marinus is not harmful to humans but causes devastating effects on oyster populations. Epizootics of perkinsosis or dermo disease caused by this parasite are the main source of catastrophic mortalities in oysters that can wipe out up to 100% of the living stock in affected areas, threatening both aquacultured and wild oyster populations and leading to tremendous losses in the oyster industry.

Infection levels by V. vulnificus and P. marinus in oysters are controlled by various extrinsic factors such as environmental temperature and salinity (Crosby & Roberts 1990, O'Neill et al. 1992, Chu et al. 1993, Kaspar & Tamplin 1993, Burreson & Calvo 1996, Motes et al. 1998) but these factors cannot fully explain individual variation in infection levels in oyster populations. We showed that tissue loads of V. vulnificus and P. marinus vary greatly in oysters growing in the same habitat and exposed to similar environmental conditions, including levels of pathogens in the water (Sokolova et al. 2005). This variation is likely caused by the genetic differences in resistance to infection; in fact resistance to P. marinus infection in C. virginica was earlier shown to have a significant genetic component and to depend on the host genotype and genotype-environment interactions (Oliver & Fisher 1999, Oliver et al. 2000). Overall, genetic variation in resistance to parasites and pathogens is well documented in mollusks and thus creates a potential basis for selection of parasite-resistant populations and strains (Grosholz 1994, Bushek & Allen 1996, Naciri-Graven et al. 1998, Langand & Morand 1998, Ataev & Coustau 1999, Knight et al. 1999, Wiehn et al. 2002).

Unfortunately, selective breeding applied to oysters has met with only limited success. Currently there are several oyster lines (e.g., Andrews DEBY and their descendants) that demonstrate some resistance to dermo disease (Ford et al. 1990, Calvo et al. 2003, S. K. Allen pers. comm.). Most of these lines, however, have not proven successful in growth locations outside of their sites of origin and may also suffer from the effects of inbreeding, which limits their use in aquaculture and oyster restoration programs. Further, no oyster strains with increased resistance to V. vulnificus infection are currently known. On the other hand, use of molecular genetic markers associated with resistance to parasites may strongly facilitate the current selective breeding programs and provide a noninvasive tool to detect resistant oysters from wild or cultured stock. Selection of parental stock from wild or cultured populations using molecular genetic markers would avoid the negative effects of inbreeding by diversifying the genetic background of breeders, whereas selecting for the desired parasite or pathogen resistance. However, this approach has never been applied to oyster stock, and no genetic markers associated with pathogen or parasite resistance are currently known in oysters. We therefore decided to test the feasibility of an arbitrary fragment length polymorphism (AFLP) approach to search for molecular genetic markers potentially associated with resistance to P. marinus and/or V. vulnificus infections. Here we report the details of an inexpensive technique that allows identification of such markers, and provide preliminary data on associations of AFLP marker loci with resistance of C. virginica to both V. vulnifucus and P. marinus.

MATERIALS AND METHODS

Animal Collection and Maintenance

Oysters were collected on June 15, 2004 from a subtidal habitat in Stump Sound, North Carolina. 50 adult oysters (55-155 mm valve height) were randomly collected from a small area (ca. 100 x 100 m) of homogenous soft-bottom habitat to ensure that all organisms used in this study were exposed to the same environmental conditions. Age of the oysters within this size range was 2-5 y as determined by the count of annual growth checks on their shells. Age determination by growth checks was verified by comparison with the growth checks on cultured oysters of known age grown subtidally in the nearby area of Stump Sound (courtesy of J. Swartzenherg, J & B Aquafood). Water temperature at the time of collection was 26[degrees]C and salinity was 31[per thousand]. Oysters were immediately placed on ice and transported to the University of North Carolina at Charlotte within 5 h of collection for further processing and analysis. Processing of oyster tissues was completed within 24 h of collection. During this time, oysters were kept on ice to prevent postharvest build-up of bacteria.

Chemicals

AFLP adapters and preselective and selective primers were purchased from Integrated DNA Technologies (Coralville, IA). Restriction enzymes and T4 DNA ligases were purchased from New England Biolabs (Ipswich, MA). TaqPro Complete PCR mixture was purchased from Denville Scientific (South Plainfield, NJ). All other reagents were purchased from Fisher Scientific (Suwanee, GA) and Sigma (St. Louis, MO) and were of analytical grade or higher.

Determination of V. vulnificus Loads and P. marinus Infection

Oysters were externally cleaned and opened with an alcohol-flamed oyster knife. The oyster contents were removed, weighed, and homogenized in sterile blender jars with an equal volume of sterile diluent (50% artificial sea water). Homogenates were diluted and plated for V. vulnificus using the cellobiose-polymyxin B-colistin (CPC) agar developed in our laboratory (Massad & Oliver 1987), which has been used by us and others for the primary isolation of V. vulnificus (Harwood et al. 2004, Oliver 2003). When colonies of appropriate color and morphology are selected, this medium has been shown to be 82% accurate in the isolation of V. vulnificus (Sun & Oliver 1995). Using these same criteria, Sloan et al. (1992) found 81% of the typical V. vulnificus colonies on CPC to be identified as this species.

Small samples of gill tissue (50-100 mg) were removed prior to homogenization and placed in DNA fixing solution for DNA extraction for AFLP analysis and PCR diagnostics of P. marinus. Remaining tissues were weighed to the nearest 0.01 g, and the maximum valve height was measured to the nearest 0.1 mm. For diagnostics of P. marinus, total DNA was isolated from 50-100 mg samples of gill tissue following an improved protocol for DNA isolation from mollusks developed by Sokolov (2000). This method allowed us to isolate total DNA, which in infected oysters contained DNA of P. marinus in addition to the host (oyster) DNA. Determination of P. marinus infection was performed using PCR with the following primers:

Pmar-F: 5' CAC TTG TAT TGT GAA GCA CCC 3'

Pmar-R: 5'GTG ACA TCT CCA AAT GAC C 3'

These primers are specific for P. marinus (Penna et al. 2001), and do not cross-amplify with oyster DNA or other parasites. Optimized PCR conditions for P. marinus detection were as follows: 25 [micro]L of reaction volume containing 1 x PCR buffer, 2 mM Mg[Cl.sub.2], 100 [micro]M of dNTPs, 0.7 U of Taq polymerase, 150 ng of each P. marinus primer and 50-100 ng of template DNA was subjected to one denaturation cycle at 94[degrees]C for 5 min, 35 cycles at 94[degrees]C for 45 s, 55[degrees]C for 30 s and 72[degrees]C for 45 s and one final extension cycle at 72[degrees]C for 7 min. P. marinus DNA obtained from monocultures of this parasite (gift of Dr. G. Vasta) was used as a positive control. Amplified DNA fragments were resolved on ethidium bromide-stained 1.5% agarose gels and screened for the presence of a ca. 304 bp product characteristic of P. marinus, which indicated infection of the oyster with this parasite. This method is highly sensitive with detection limits of 0.1 pg DNA of P. marinus, corresponding to 1 protozoan cell [g.sup.-1] oyster tissue (Penna et al. 2001).

AFLP Analysis

AFLP analysis was conducted according to an AFLP gene mapping protocol described elsewhere (Vos et al. 1995, Yu & Guo 2003, Li & Guo 2004). Isolated genomic DNA (~0.5-1 [micro]g) was digested by restriction enzymes EcoRI and MseI for 1 h at 37[degrees]C. The digestion mixture (10 [micro]L) contained 0.5-1 [micro]g genomic DNA, 5 U EcoRI enzyme, 5 U MseI enzyme and 1 x MseI buffer (Buffer 2, New England Biolabs) and an appropriate amount of water. The complete digestion mixture was ligated with relevant AFLP adapters overnight at room temperature. Each ligation reaction (20 [micro]L) contained the restriction digestion product (10 [micro]L), 1 x T4 DNA ligase buffer with EDTA, 0.05 M NACl, 0.05 mg [mL.sup.-1] BSA, 6 U T4 DNA ligase, 10-pmol EcoRI adapter, 100-pmol MseI adapter and appropriate amount of water. The structure of the EcoRI-adapter was:

5'-CTCGTAGACTGCGTACC

CATCTGACGCATGGTTAA-5'

The structure of the MseI-adapter was:

5'-GACGATGAGTCCTGAG

TACTCAGGACTCAT-5'

Ligation product was diluted 20-fold and used as a template for preselective amplification.

Preselective amplification was performed using specific primers complementary to each adaptor sequence. Each preselective PCR (20 [micro]L) contained 10 [micro]L of 2 x TaqPro Complete mixture with 1.5 mM Mg[Cl.sub.2], 100 pmol of each primer, 4 [micro]L of 20-fold diluted ligation product and an appropriate amount of water. The cycling profile for preselective amplification was: one cycle of 72[degrees]C for 5 min to inactivate restriction and ligation, 30 cycles of 94[degrees]C for 40 s, 54[degrees]C for 40 s and 72[degrees]C for 2.5 min and one final cycle of 72[degrees]C for 10 min. Products from preselective. PCR were diluted 20-fold and used as templates for selective amplifications. Pairs of selective primers complementary to each adaptor sequence (except for the last three selective nucleotides added at their 3' end) were used for selective PCR.

Each selective PCR (20 [micro]L) contained 10 [micro]L of 2 x TaqPro Complete mixture with 1.5 mM Mg[Cl.sub.2], 100 pmol of each primer, 5 [micro]L of 20-fold diluted product of preselective amplification and appropriate amount of water. For selective PCR, a touch-down amplification was used: 10 cycles of 94[degrees]C for 20 s (denaturation), 66[degrees]C for 30 s (annealing), 72[degrees]C for 2 min (extension), with a 1[degrees]C decrease of annealing temperature each cycle, followed by 20 cycles of amplification at 94[degrees]C for 20 s, 56[degrees]C for 30 s and 72[degrees]C for 2 min. We analyzed six combinations of primer pairs containing core sequences complementary to EcoRI or MseI adapters with addition of the following selective nucleotide triplets on 3'-end: forward primers EcoRI-ACA and -ACA, and reverse primers MseI-CTC, -CTA, and -CAG. All amplifications were performed in an Eppendorf MasterCycler Gradient thermal cycler (Brinkmann, Westbury, NY).

DNA fragments obtained by selective amplification were resolved on 8% acrylamide gels for 7-8 h in 1 x TBE buffer at 15[degrees]C to improve resolution and were visualized by silver-staining using a method described by Sokolova and Boulding (2004). At least 3 reference samples were run on each gel as internal standards, along with 100 bp and 50 bp DNA ladders (Invitrogen, Carlsbad, CA). AFLP loci were analyzed using the Kodak EDAS 290 gel imaging system and Kodak 1D Image Analysis Software (Kodak, Rochester, NY), and scored as a presence or absence of a fragment of the respective length. Software-generated scores were verified manually. Only fragments between 100 and 350 bp were included into the analysis because shorter or longer fragments could not he reliably scored.

Originally we scored 100 AFLP markers as present or absent in a random sample of oysters, and 48 of these markers proved to be sufficiently polymorphic (neither haplotype with a frequency greater than 80%) for use. These 48 markers (designated MI through M48) were scored in 35 oysters, although 1 marker was missing in 7 oysters and 4 markers were missing in 5 oysters (all 35 markers were present in 23 of the 35 oysters). This sample of 35 contained only 3 oysters that previously had been classified as not infected with V. vulnificus (Sokolova et al. 2005), making the presence/absence of infection an unsuitable qualitative trait for analysis. Instead, we analyzed a quantitative trait, the magnitude of infection of V. vulnificus (log of CFU/g counts) among the 32 infected oysters. Because this trait was previously found to be correlated with weight of the oysters (Sokolova et al. 2005), we first used regression to adjust all these values for differences in overall weight. Our sample of 35 oysters included 7 individuals that previously were classified as not infected with P. marinus, making it possible to analyze association of AFLP markers with the presence/absence of infection by this parasite. Both dependent variables, the magnitude of V. vulnificus infection and the presence/absence of P. marinus infection, were tested for their association with the AFLP markers as described below.

AFLP markers are useful in a genome-wide search for gene or gene groups linked with the differential resistance to parasites or pathogens and do not require prior knowledge about the host DNA sequences; they are anonymous and thus can in principle be amplified from DNA of the host or the parasite. In this study, we used oyster gill tissues where the host DNA content is expected to exceed the parasite DNA content by several orders of magnitude because of the relatively low infection intensity (mean score 3.6 [+ or -] 0.24 by Mackin scale corresponding to medium infection, n = 18, Grewal & Sokolova, unpublished data) and large differences in the genome size of the host and the parasites/pathogens (700 Mb in C. virginica versus 28 Mb and 5.3 Mb in P. marinus and V. vulnificus, respectively) (Gregory 2001, Chen et al. 2003; http://www.tigr .org/tdb/e2kl/pmg/intro.shtml). Because amplification is a competitive process, the number of copies amplified from rare templates (e.g., from parasite DNA) will be negligible compared with the abundant templates (i.e., from oysters) when the same number of cycles is used, and products from rare templates are not likely to be visualized on the AFLP gel. In this study, we scored only well-defined AFLP bands of high intensity to ensure that all AFLP markers belong to the oyster DNA.

Single-Locus Associations

As a first step in testing for associations of the marker haplo-types with infection with either V. vulnificus or P. marinus, it was necessary to assess relatedness among the 48 markers themselves. This assessment was accomplished with the MAPMAKER 3.0b program (Lander et al. 1987, Lincoln et al. 1992), which tested for potential linkage groups among these markers. We reduced the default criterion of 3.0 in this program to 2.5 to be conservative with our limited sample size, and the program identified 9 linkage groups, each with 2-4 markers. However, only 22 markers were included in these groups, with the remaining 26 being unlinked and therefore considered to be independent (effective number of total linkage groups = 9 + 22 = 31).

We used 1-way analyses of variance (ANOVA; Sokal & Rohlf 1995) to test for associations of each of the 48 markers with the degree of V. vulnificus or P. marinus infections. In these analyses, the AFLP marker was the single factor (with two levels, presence or absence) and the degree of infection was the dependent variable. Each ANOVA yielded the conventional F statistic with its associated probability, with any probabilities less than 0.05 indicating conventional significance.

Because 48 ANOVAs were conducted for the analysis of V. vulnificus resistance, and another 48 ANOVAs for the analysis of P. marinus resistance, it was necessary to adjust the conventional significance level to ensure that the experiment-wise error rate did not exceed 5% (Sokal & Rohlf, 1995). To accomplish this, we used a permutation test (Churchill & Doerge 1994) involving 1,000 iterations in which the infection values (scores for P. marinus or V. vulnificus) were randomly permuted, merged with the AFLP marker data, and ANOVAs performed. These ANOVAs generated F values with their associated probabilities that were logarithmically transformed to produce LOD scores (Lander & Botstein 1989) as follows: LOD = [log.sub.10](1/Prob). For each of the 31 linkage groups, the highest LOD scores generated in each permutation run were then ranked, and the 50th and 10th highest values from each distribution represented the 5% and 1% group-wise threshold LOD scores. An experiment-wise threshold value across all linkage groups also was obtained from the 50th (5%) highest LOD scores that were observed on any linkage group during each of 1000 iterations (Churchill & Doerge 1994). If the highest LOD score calculated for a given linkage group exceeded its appropriate 5% group-wise threshold value (or especially the experiment-wise value), the test of association was considered to be significant and suggested that there is a genetic locus influencing resistance to V. vulnifucus (or P. marinus) infection on the chromosomal fragment adjacent to the respective AFLP marker.

Two-Locus Associations

We used 2-way ANOVAs to test for significance of pairs of AFLP markers for their potential interactive effects on V. vulnifucus and P. marinus resistance and/or susceptibility. Marker pairs were tested only for the 465 pairwise combinations of the 31 linkage groups because markers within linkage groups are associated. The significance of marker epistasis was indicated by the probability associated with the F value for the interaction of each pair of markers in the ANOVAs.

The multiple comparisons problem inherent in this many tests for epistasis was addressed by first calculating the effective number of independent tests for each linkage group (Cheverud 2000, Cheverud 2001). This was calculated as [M.sub.e] = M(1 - [[V.sub.[lambda]](M-1)/ [M.sup.2])]), where M is the number of markers scored (nonlinked markers were given an M value of 1), and [V.sub.[lambda]] is the variance of the eigen values of the correlation matrix of markers. The total number of independent epistasis tests then was estimated to be the sum of the crossproducts of the effective number of markers for all 465 pairs of linkage groups. This calculation yielded a sum of 814, suggesting that we might expect about 5% x 814 = 41 tests to be significant at the 5% level (8 at the 1% level and 1 at the 0.1% level) because of chance alone. Thus epistasis was indicated if the number of F tests of the interactions reaching the conventional 5% level of significance significantly exceeded 41. This procedure also allowed us to test for individual instances of epistasis by correcting our threshold level of significance via the Bonferroni procedure. Specifically, any specific two-gene interaction was considered to be significant at the 10% experiment-wise level when a given probability from the F test for the interaction of markers reached the 0.1 Bonferroni threshold level of significance of 0.1/814 = 0.000123 (Peripato et al. 2002, Leamy et al. 2005).

RESULTS

V. vulnificus Infection

Three (M19, M28 and M31) of the 48 AFLP markers exhibited associations with the degree of V. vulnificus infection that reached significance at the conventional 5% level in the ANOVAs. Further, LOD scores associated with the probabilities for two of these markers (M28 and M31) exceeded the 1% group-wise significance threshold level, although neither LOD score exceeded the 5% experiment-wise threshold value of 2.85 (Table 1). The remaining marker, M19, narrowly missed significance at the 5% group-wise level. Table 1 also provides means and standard errors of V. vulnificus infection associated with the presence/absence of each of the markers. As may be seen, alleles presumably linked to each of these markers act to increase (M31) or decrease (M28) the degree of infection.

The results of the 2-way marker analyses done to test for epistatic effects on V. vulnificus infection produced 64 F values with associated probabilities less than 5%, this being significantly greater than the number (41) expected at this level by chance alone ([chi square] = 12.99, df = 1, P < 0.001). The number of F values reaching significance at the 1% level was 12, but this was not significantly greater than 8 expected at this level ([chi square] = 1.55, df = 1, P > 0.05). One marker combination, M42 with M44, reached significance at the 10% experiment-wise level (P = 0.000114), suggesting that the interaction of these two markers significantly affects V. vulnificus infection levels. In general, therefore, results of these analyses indicate that epistatic interactions of unknown genes linked to these AFLP markers have an effect on the infection level of V. vulnificus in our sample of oysters.

To assess the relative effect of individual markers and their interaction on the total variability of V. vulnificus infection levels, we ran two multiple regressions. A multiple regression of infection scores on the two markers (M28 and M31) reaching group-wise significance generated a multiple coefficient of determination of 27.8.% (or when adjusted for the number of parameters, 22.6%). This suggests that variation in these two markers alone accounts for nearly 30% of the variation in V. vulnificus infection levels. Addition of the M42-M44 significant interaction increased the amount explained to 41.8% (adjusted value = 35.4%), a significant improvement in fit ([chi square] = 5.77, P < 0.05) from the single-locus model containing the 3 markers.

P. marinus Infection

Three markers (M5, M45 and M47) exhibited significance at the conventional 5% level in the tests of associations with the presence/absence of P. marinus infection. Two of these markers (M5 and M47) reached group-wise significance, but the remaining marker, M45, did not. Table 2 shows these markers, their probabilities and LOD scores, the group-wise threshold LOD scores and the percentage of P. marinus infection for oysters with the marker present/absent. M5 has the lowest probability (highest LOD score), and oysters with this marker show nearly 100% infection with P. marinus whereas only about one-half of those oysters without the marker are infected. As was the case for V. vulnificus, however, none of these markers affecting P. marinus reach the experiment-wise level of significance (LOD = 2.85) in the association tests. In a multiple regression, these two markers accounted for 40.0% (36.2% adjusted) of the total variation in the incidence of P. marinus infection.

Results of the tests for marker interactions potentially affecting the incidence of P. marinus infection produced 57 F values with associated probabilities less than 5%, this being significantly greater than the number (41) expected at this level by chance alone ([chi square] = 6.17, df = 1, P < 0.05). The number of F values reaching significance at the 1% level was 14, and this narrowly missed being significantly greater than the 8 expected at this level ([chi square] = 3.81, df = 1, P > 0.05). No marker combinations reached significance at the 10% experiment-wise level. Thus there is some suggestion of potential interactions of unknown genes affecting the incidence of P. marinus infection, but the evidence for this is less than was seen in the V. vulnificus analysis.

DISCUSSION

This study demonstrates the feasibility of the AFLP approach to search for genetic markers associated with oyster genes potentially influencing resistance to V. vulnificus and P. marinus. In fact, the results of association tests using the AFLP markers suggest that there are genes that act directly or indirectly to control the levels of infection by these parasites in oysters. We found two markers that were significantly associated with infection levels of V. vulnificus and two others associated with the incidence of infection of P. marinus. Further, there was evidence of significant genetic interactions (epistasis) affecting the levels of both pathogens, especially of V. vulnificus. The impact of these unknown genes on variability in infection levels was impressive, accounting for about 40% of total variation in each case.

Whereas the above-described results certainly are encouraging, they should be interpreted with caution because of the relatively small sample size used in this pilot study. This is especially so because all marker associations that reached significance were at the group-wise level, and thus may be considered only as suggestive of the presence of the linked functional genes (Lander & Kruglyak 1995) responsible for the control of infection levels. To provide significant evidence for such genes, the LOD scores should have exceeded the more conservative experiment-wise threshold level (Lander & Kruglyak 1995). Therefore, our results would need to be verified with subsequent studies using larger sample sizes as well as oysters from different populations (Kramer et al. 1998). Such verification is clearly required before AFLP markers could be recommended for use in marker-supported breeding programs aimed at selecting for resistant oyster strains. Nonetheless, even with these caveats, our results clearly imply that AFLP analysis is a viable strategy to detect such markers in oyster populations and that a future study using a larger sample size may well detect genes strongly associated with resistance to these two important parasites/pathogens in oysters.

Our results also suggest that beyond single-locus effects of genes, it is important to consider screening for epistatic interactions of genes while searching for loci associated with disease/ pathogen resistance in oysters. We found evidence that gene interactions affected infection levels in the oysters, especially of V. vulnificus where the M42-M44 marker interaction reached experiment-wise significance in the association tests. This result involves two markers that were different from those (M28, M31) showing individually significant effects and thus would not have been detected had we conducted single-locus analyses only. This sort of result is not at all uncommon, and highlights the role of epistasis in various quantitative traits (Cheverud & Routman 1995, Leamy et al. 2002, Leamy et al. 2005), including disease resistance (Templeton 2000). Therefore, in future studies searching for genetic markers of disease resistance for marker-supported breeding programs, it seems important to take epistatic interactions between the loci into account.

As a corollary, the AFLP approach described in this study provides a cost-effective and relatively rapid method of screening of oyster stock for marker loci (or their epistatically interacting combinations) associated with resistance to two important pathogens affecting the oyster industry--V, vulnificus and P. marinus. A considerable advantage of this method is that identification of resistant oysters to be used for the parent stock can be performed noninvasively using a small tissue biopsy for DNA extraction and analysis and at a reasonable cost. We believe that the same approach may also be used with other parasites (e.g., Haplosporidium) or with invasive organisms such as Polydora providing a new tool to combat disease in oyster stocks. In host-parasite systems where the parasite genome size is small (such as in a protozoan P. marinus or a bacterium V. vulnificus) DNA isolated from the host tissue may be directly used in the AFLP analysis, especially if tissue with low to moderate infection intensity is selected ensuring that the parasite DNA content is negligible compared with the host DNA content. For parasites with large genomes or for very heavily infected tissues, it may be advisable to use markers that predominantly amplify from noninfected host individuals but not from the infected ones (such as M47 and M28 in this study), or to establish AFLP controls using DNA from pure parasite cultures to distinguish between the markers that are amplified from DNA of the host and the parasite.

ACKNOWLEDGMENTS

The authors thank Dr. Thomas Rosche, Ben Smith, Erin Parker and Bryn Adams for their assistance in determining the V. vulnificus levels; Dr. Gerardo Vasta for his gift of P. marinus DNA; Melanie Harrison and Danijela Bozanovic for assistance with DNA isolation and AFLP analysis; Dr. Eugene Sokolov for helpful comments on the earlier draft of the manuscript and Jim Swartzenberg of J & B Aquafood for his help with animal collection. This work was funded in part by North Carolina Sea Grant (RMG0401), and JCSU MBRS-RISE Program NIGMS 58042. Publication costs were in part covered by the UNC Charlotte.

LITERATURE CITED

Ataev, G. L. & C. Coustau. 1999. Cellular response to Echinostoma caproni infection in Biomphalaria glabrata strains selected for susceptibility/resistance. Develop. Compar. Immunol. 23:187-198.

Bushek, D. & S. K. Allen, Jr. 1996. Host-parasite interactions among broadly distributed populations of the eastern oyster Crassostrea virginica and the protozoan Perkinsus marinus. Mar. Ecol. Prog. Ser. 139:127-141.

Burreson, E. M. & L. M. R. Calvo. 1996. Epizootiology of Perkinsus marinus disease of oysters in Chesapeake Bay, with emphasis on data since 1985. J. Shellfish Res. 15:17-34.

Calvo, L. M. R., G. W. Calvo & E. M. Burreson. 2003. Dual disease resistance in a selectively bred eastern oyster, Crassostrea virginica, strain tested in Chesapeake Bay. Aquaculture 220:69-87.

Chen, C. Y., K. M. Wu, Y. C. Chang, C. H. Chang, H. C. Tsai, T. L. Liao, Y. M. Liu, H. J. Chen, A. B. T. Shen, J. C. Li, T. L. Su, C. P. Shao, C. T. Lee, L. I. Hor, S. F. Tsai. 2003. Comparative genome analysis of Vibrio vulnificus, a marine pathogen. Genome Res. 13:2577-2587.

Cheverud, J. M. 2000. Detecting epistasis among quantitative trait loci. In: J, Wolf, E. Brodie II & M. Wade, editors. Epistasis and the Evolutionary process, New York: Oxford University Press. pp 58-81.

Cheverud, J. M. 2001. A simple correction for multiple comparisons in interval mapping genome scans. Heredity 87:52-58.

Cheverud, J. M. & E. J. Routman. 1995. Epistasis and its contribution to genetic variance components. Genetics 139:1455-1461.

Chu, F. L. E., J. F. La Peyre & C. S. Burreson. 1993. Perkinsus marinus infection and potential defense-related activities in Eastern oysters, Crassostrea virginica: salinity effects. J. Invert. Pathol. 62:226-232.

Churchill, G. A. & R. W. Doerge. 1994. Empirical threshold values for quantitative trait mapping. Genetics 138:963-971.

Crosby, M. P. & C. F. Roberts. 1990. Seasonal infection intensity cycle of the parasite Perkinsus marinus (and an absence of Haplosporidium spp.) in oysters from a South Carolina salt marsh. Dis. Aquat. Org. 9:149-155.

Ford, S. E., A. J. Figueras & H. H. Haskin. 1990. Influence of selective breeding, geographic origin, and disease on gametogenesis and sex ratios of oysters, Crassostrea virginica, exposed to the parasite Haplosporidium nelsoni (MSX). Aquaculture 88:285-301.

Gregory, T. R. 2001. Animal genome size database. Available at: http:// www.genomesize.com.

Grosholz, E. D. 1994. The effects of host genotype and spatial distribution on trematode parasitism in a bivalve population. Evolution 48:1514-1524.

Harwood, V. J., J. P. Gandhi & A. C. Wright. 2004. Methods for isolation and confirmation of Vibrio vulnificus from oysters and environmental sources: a review. J. Microbiol. Meth. 59:301-316.

Kaspar, C. W. & M. L. Tamplin. 1993. Effects of temperature and salinity on the survival of Vibrio vulnificus in seawater and shellfish. Appl. Environ. Microbiol. 59(8):2425-2429.

Keithly, W. R., Jr. & H. Diop. 2001. The demand for eastern oysters, Crassostrea virginica, from the Gulf of Mexico in the presence of Vibrio vulmficus. Mar. Fish. Rev. 63:47-53.

Knight, M., A. N. Miller, C. N. Patterson, C. G. Rowe, G. Michaels, D. Carr, C. S. Richards & F. A. Lewis. 1999. The identification of markers segregating with resistance to Schistosoma mansoni infection in the snail Biomphalaria glabrata. Proc. Natl. Acad. Sci. USA 96:1510-1515.

Kramer, M. G., T. T. Vaughn, L. S. Pletscher, K. King-Ellison, E. Adams, C. Erikson & J. M. Cheverud. 1998. Genetic variation in body weight gain and composition in the intercross of Large (LG/J) and Small (SM/J) inbred strains of mice. Genet. Moi. Biol. 21:211-218.

Lander, E. S. & L. Kruglyak. 1995. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11:241-247.

Lander, E. S. & D. Botstein. 1989. Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199.

Lander, E. S., P. Green, J. Abrahamson, A. Barlow, M. Daley, S. Lincoln, L. Newburg. 1987. MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181.

Langand, J. & S. Morand. 1998. Heritable non-susceptibility in an allopatric host-parasite system: Biomphalaria glabrata (Mollusca)-Echinostoma caproni (Platyhelminth Digenea). J. Parasit. 84:739-742.

Leamy, L. J., E. J. Routman & J. M. Cheverud. 2002. An epistasis genetic basis for fluctuating asymmetry of mandible size in mice. Evolution 56:642-653.

Leamy, L. J., M. S. Workman, E. J. Routhman & J. M. Cheverud. 2005. An epistatic genetic basis for fluctuating asymmetry of tooth size and shape in mice. Heredity 94:316-325.

Li, L. & X. Guo. 2004. AFLP-based genetic linkage maps of the Pacific oyster Crassostrea gigas Thunberg. Mar. Biotechnol. 6:26-36.

Lincoln, S., M. Daly & E. Lander. 1992. Constructing genetic maps with MAPMAKER/EXP 3.0. Whitehead Institute technical report. 3rd ed.

Massad, G. & J. D. Oliver. 1987. New selective and differential plating medium for Vibrio vulnificus and Vibrio cholerae. Appl. Environ. Microbiol. 53:2262-2264.

Motes, M. L., A. DePaola, D. W. Cook, J. E. Veazey, J. C. Hunsucker, W. E. Garthright, R. J. Blodgett & S. J. Chirtel. 1998. Influence of water temperature and salinity on Vibrio vulnificus in Northern Gulf and Atlantic Coast oysters (Crassostrea virginica). Appl. Environ. Microbiol. 64(4):1459-1465.

Naciri-Graven, Y., A.-G. Martin, J.-P. Baud, T. Renault & A. Gerard. 1998. Selecting the flat oyster Ostrea edulis (L.) for survival when infected with the parasite Bonamia ostreae. J. Exp. Mar. Biol. Ecol. 224:91-107.

O'Neill, K. R., S. H. Jones & D. J. Grimes. 1992. Seasonal incidence of Vibrio vulnificus in the Great Bay estuary of New Hampshire and Maine. Appl. Environ. Microbiol. 58(10):3257-3262.

Oliver, J. D. 2003. Culture media for the isolation and enumeration of pathogenic Vibrio species in foods and environmental samples. In: J. E. L. Corry, G. D. W. Curtis, & R. M. Baird, editors. Culture media for food microbiology, 2nd ed. Amsterdam: Elsevier Science. pp. 249-269.

Oliver, J. L., P. M. Gaffney, S. K. Allen, Jr., M. Faisal & S. L. Kaattari. 2000. Protease inhibitory activity in selectively bred families of eastern oysters. J. Aquat. Anim. Health 12:136-145.

Oliver, L. M. & W. S. Fisher. 1999. Appraisal of prospective bivalve immunomarkers. Biomarkers 4:510-530.

Oliver, J. D. & J. B. Kaper. 2003. Vibrio species. In: Food microbiology: fundamentals and frontiers. Washington, DC: ASM Press. pp. 263-300.

Penna, M. S., M. Khan & R. A. French. 2001. Development of a multiplex PCR for the detection of Haplosporidium nelsoni, Haplosporidium costale and Perkinsus marinus in the eastern oyster (Crassostrea virginica, Gmelin, 1971). Mol. Cell. Probes 15:385-390.

Peripato, A. C., R. A. de Brito, T. T. Vaughn, L. S. Pletscher, S. R. Matioli & J. M. Cheverud. 2002. Quantitative trait loci for maternal performance for offspring survival in mice. Genetics 162:1341-1353.

Sloan, E. M., C. J. Hagen, G. A. Lancette, J. T. Peeler & J. N. Sofos. 1992. Comparison of five selective enrichment broths and two selective agars for recovery of Vibrio vulnificus from oysters. J. Food Prot. 55:356-359.

Sokal, R. R. & J. F. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. New York: Freeman.

Sokolov, E. P. 2000. An improved method for DNA isolation from mucopolysaccharide-rich molluscan tissues. J. Mol. Stud. 66:573-575.

Sokolova, I. M. & E. G. Boulding. 2004. Length polymorphisms in an intron of aminopeptidase N provide A useful nuclear DNA marker for Littorina species (Caenogastropoda). J. Mol. Stud. 70:165-172.

Sokolova, I. M., L. Leamy, M. Harrison & J. D. Oliver. 2005. Intrapopulational variation in Vibrio vulnificus levels in Crassostrea virginica (Gmelin 1971) is associated with the host size but not with disease status or developmental stability. J. Shellfish Res. 24:503-508.

Sun, Y. & J. D. Oliver. 1995. The value of CPC agar for the isolation of Vibrio vulnificus from oysters. J. Food Prot. 58:439-440.

Templeton, A. R. 2000. Epistasis and complex traits. In: J. B. Wolf, E. D. Brodie III & M. J. Wade, editors. Epistasis and the evolutionary process. New York: Oxford University Press. pp. 41-57.

Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Homes, A. Frijters, J. Pot, J. Peleman, M. Kuiper & M. Zabeau. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23:4407-4414.

Wiehn, J., K. Kopp, S. Rezzonico, S. Karttunen & J. Jokela. 2002. Family-level covariation between parasite resistance and mating system in a hermaphroditic freshwater snail. Evolution 56:1454-1461.

Yu, Z. & X. Guo. 2003. Genetic linkage map of the eastern oyster Crassostrea virginica Gmelin. Biol. Bull. 204:327-338.

INNA M. SOKOLOVA, * JAMES D. OLIVER AND LARRY J. LEAMY

Department of Biology, University of North Carolina at Charlotte, 9201 University City Blvd., Charlotte, North Carolina 28223

* Corresponding author. E-mail: [e-mail]

TABLE 1.

AFLP markers significantly associated with the degree of infection of
V. vulnificus in the sample of oysters. Listed are the markers and
their associated LOD scores, 5% and 1% group-wise threshold LOD scores,
and sample sizes, means, and standard errors for V. vulnificus
infection for the two haplotypes at each marker.

                               Threshold LOD         Marker present

Marker      P         LOD       5%      1%       N    Mean    St. Error

M28       0.0114    1.94 **    1.23    1.74     12    3.79      0.218
M31       0.0032    2.49 **    1.62    2.24     24    4.37      0.125

               Marker absent

Marker     N    Mean    St. Error

M28       20    4.48        0.139
M31        7    3.62        0.355

** = P < 0.01

TABLE 2.

AFLP markers significantly associated with the degree of infection
of P. marinus in the sample of oysters. Listed are the markers and
their associated LOD scores, 5% and 1% group-wise threshold LOD
scores, and infection percentages for oysters for which the marker is
present/absent.

                            Threshold LOD        Marker

Marker      P      LOD        5%      1%    Present   Absent

M5       0.0021   2.68 **    1.21    1.86    95.5%     53.9%
M47      0.0339   1.47 *     1.36    1.81    55.5%     88.5%

* = P < 0.05: ** = P < 001