Printer Friendly

Field evaluation of mortality from hemolymph extraction as a source of DNA, and application to PCR-RFLP identification of threatened freshwater mussel species.

ABSTRACT Evolutionary convergence and plasticity of shell characters creates great confusion in freshwater mussel systematics and complicates field census efforts. Genetic identification offers a powerful alternative. But methods involving DNA sequencing are expensive and require tissue sampling, whose effects on survival and health of animals from natural populations have seldom been assessed. We used hemolymph sampling as a nonlethal source of tissue for DNA extraction, and developed genetic identification methods using Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis as an alternative for field surveys. We focused on two morphologically similar endemic species in Lake Waccamaw, North Carolina: the abundant Elliptio waccamawensis and the less common Lampsilis fullerkati. These served as models for surveyors who are often faced with conducting accurate census of cryptic rare species that co-occur with more common forms. Hemolymph-sampled and control individuals of these two species and Leptodea ochracea were caged together in enclosures in Lake Waccamaw. Eight-week survival was 100% and we detected no significant effect of hemolymph extraction on shell growth. PCR-RFLP analysis of the 16s rRNA gene reliably identified species and detected cases of morphological misidentification, both in these collections and from belt transects in the lake. We also developed PCR-RFLP markers that distinguished pairs of cryptic taxa from surveys of streams in southeastern North Carolina. Our results show how nonlethal tissue sampling and PCR-RFLP assays designed for a regional fauna can be useful tools in freshwater mussel conservation programs.

KEY WORDS: freshwater mussel, conservation genetics, endemic, genetic identification, hemolymph


North America boasts the highest diversity of freshwater mussels anywhere, yet almost three-quarters of these species are at risk of extinction (Williams et al. 1993, Stein & Chipley 1996, Bogan 2008). Conservation efforts, however, are hindered by poor taxonomy and phenotypic plasticity in conchological features (Campbell et al. 2005). Species identification has generally relied on external shell characters, but this approach is notoriously difficult to implement and may be unreliable. Even distantly related species can show very similar shell characters, and many shell characters are known to vary in response to environmental conditions (Baker et al. 2003).

Molecular genetic methods are a strong alternative candidate for species identification. Sequences of mitochondrial and nuclear DNA regions have been used effectively to distinguish species and evaluate genetic relationships among populations of freshwater mussels (e.g., King et al. 1999, Roe et al. 2001, Serb et al. 2003, Hughes et al. 2004). There are, however, logistical problems with this approach. First, mortality from collecting whole animals prevents biologists from studying populations at risk and hinders development of methods for distinguishing endangered from nonendangered taxa. Second, species identification that relies on DNA sequencing is considerably more expensive and slower than that, based on morphology, and it requires access to a laboratory equipped to perform the analysis. These limitations can be overcome by using nonlethal tissue sampling and rapid genetic identification methods that do not rely upon DNA sequencing.

Over the past decade or so, some nonlethal methods have been evaluated for the study of freshwater mussels. Berg et al. (1995) demonstrated through field experiments on Quadrula quadrula (Rafinesque, 1820) and Actinonaias ligamentina (Lamarck, 1819) that biopsy of a 1 [cm.sup.2] piece of mantle tissue did not cause significant mortality in treatment animals versus control animals. Naimo et al. (1998) similarly reported no detectable mortality from foot biopsies performed on Amblema plicata plicata (Say, 1817) monitored over a 1.5-y period. Tissue biopsy has been used in several genetic studies since the development of the technique (listed in Henley et al. 2006), but recent results suggest the method may have questionable application to threatened and endangered populations. For example, Henley et al. (2006) reported unpublished data showing 56.3% mortality from mantle biopsy of Epioblasma triquetra (Rafinesque, 1820), and developed a nonlethal approach of brush swabbing the foot and viscera of Quadrula pustulosa (I. Lea, 1831). They determined that extracted DNA concentrations from swab and mantle snips were comparable, and that DNA extracted from swabs could be amplified using the polymerase chain reaction (PCR) and sequenced.

Non-lethal methods of hemolymph extraction have a long history of application to marine bivalve molluscs for monitoring physiological state, isolating pathogens, and as a source for DNA sampling (Fyhn & Costlow 1975, Ford 1986, Yanick & Heath 2000). More recently, the technique has been applied to freshwater mussels. Gustafson et al. (2005) tested the effects of drawing 0.5 mL of hemolymph from the anterior adductor muscle of Elliptio complanata (Lightfoot, 1786). Animals were kept in laboratory aquaria to study the effects of hemolymph extraction on growth and survival, for which they found none. Raley et al. (2006) expanded on this by showing that PCR products of DNA extracted from hemolymph of E. complanata produced DNA sequences that were identical to those from DNA extracted from body tissues. However, mortality effects of hemolymph sampling have never, to our knowledge, been addressed under field conditions for freshwater mussels. Such an evaluation would improve confidence in the application of this method to threatened and endangered unionid populations.

This project field-tested hemolymph collection as a method for the nonlethal collection of tissues for DNA extraction and genetic identification of freshwater mussels. It had two primary objectives. The first was to evaluate the effects of hemolymph sampling on mussel growth and survival under conditions in their native environment, through field experiments. The second goal was to develop species-diagnostic DNA restriction fragment length polymorphism (RFLP) assays, employing restriction enzyme digestion of PCR amplicons. These assays are faster, cheaper, and can be performed with less expensive equipment than DNA sequence analysis. We based our work in Lake Waccamaw. This is the largest "Carolina Bay" lake in North Carolina, from which 17 species of freshwater mussels been reported to occur, either at present or historically (Bogan 2002). Not all of these species are abundant, however, so we focused our work on two described endemics, the state endangered species Elliptio waccamawensis (I. Lea, 1863) and the state threatened species Lampsilis fullerkati (R. I. Johnson 1984), both of which maintain large populations in Lake Waccamaw, but are found nowhere else. Like other freshwater mussels that are threatened with extinction, these two taxa are genetically distinct but difficult to distinguish morphologically. Hence they serve as valuable test cases for our nonlethal genetic identification methods on a fauna that is accessible and amenable to field experimentation; locally abundant so that it can withstand minor impacts imposed by sampling efforts, yet still of conservation significance.


Field Assessments of Growth and Survival

Two enclosures were placed in Lake Waccamaw (Columbus County, NC, map: Fig. 1) in April 2004, about 300 m offshore (Fig. 1, site 1) in the State Park at about 1.5 m water depth. Each enclosure was constructed of wire fence material onto which plastic 3/4" mesh was attached and the enclosures were anchored into the sand with rebar stakes. Each enclosed approximately 10 [m.sup.2] area of sand bottom and stood about 25 cm above the sand. All freshwater mussels were cleared from the enclosures prior to stocking them with experimental animals. To do so, we collected all of the mussels on and under the sand surface and placed them into mesh bags. We found that a few animals missed on our first subsurface search would later appear on the sand surface. Leptodea ochracea (Say, 1817), in particular, was easy to miss in this way because of an apparently greater tendency to burrow than the other two most common species. After 30 rain and again after I h, we repeated the clearing procedure until no other animals were encountered. This same search process was used on subsequent dates on which we recollected and measured the test animals within the enclosures.

Individuals of the three most abundant freshwater mussel species in the Lake: Elliptio waccamawensis, Lampsilis fullerkati, and Leptodea ochracea were selected as test species. The species of Elliptio and Lampsilis have been described as endemics in the lake (Johnson 1970, Johnson 1984). Elliptio waccamawensis occurs at a density high enough such that all of the test individuals used in an enclosure were collected from within that enclosed area, whereas it was being cleared of animals. La. fullerkati and Le. ochracea were considerably less abundant, so test individuals of these species had to be collected from a broader area (about 200-300 [m.sup.2] surrounding site 1, Fig. 1) after extensive searching. Mussels were identified to species in the field based on morphological criteria, by biologists from the North Carolina Wildlife Resources Commission (NCWRC).


Into each enclosure, we placed 20 E. waccamawensis, 16 La. fullerkati, and 20 Le. ochracea. Preliminary data revealed that the average density near the enclosures was about 5 individuals/ [m.sup.2] (test transects were 50 m long x 2 m wide and averaged 500 animals per transect). To provide adequate sample sizes for mortality and growth estimates whereas not substantially altering the ambient density of all mussels, densities were adjusted to 0.5, 5, and 30 times the natural values for Elliptio, Leptodea, and Lampsilis, respectively. We attempted to equalize sample sizes as much as possible, but after a half-day's searching, we collected only 32 Lampsilis, and so we used all of these as test animals and equalized the sample sizes of the other 2 species at 20 individuals each.

From one-half of the individuals in each of these groups, 20-100 [micro]L of hemolymph was extracted from the anterior adductor muscle. By using a thin knife to gape and tissue forceps to hold the valves open, the iridescent adductor muscle can be easily visualized. This muscle was penetrated using a 27G1/2 (0.4 mm D x 13 mm L) needle fitted to a 1 mL sterile tuberculin syringe, to which moderate suction was applied. This was the smallest gauge needle we found to reliably reach the adductor muscle for animals approximately 45-70 mm shell length; specimens 75-150 mm shell length collected from our field census required 25G5/8 needles. One to two hundred [micro]L hemolymph was collected within 30 sec, and the hemolymph was immediately placed on ice. The animals were then measured with electronic calipers and placed back into the enclosure. From the remaining one-half of the animals, no needles were inserted and no hemolymph was removed, but otherwise the animals were handled similarly, and their survivorship and growth was monitored as control animals. All animals were tagged with "bee-tags" (The Bee Works, Orillia, ON Canada) so that they could be individually identifed.

The growth and survival of all mussels was assessed at 2, 4, and 8 wk posthemolymph removal. At each sampling date, all mussels inside of the enclosures were removed (as described above) and counted. Each individual was remeasured, inspected to make sure it was alive and not moribund, and then was returned to the enclosures. Two-way analysis of variance was used to test for differences in growth, using treatments (hemolymph-sampled and controls) and species as the main effects. The two individuals that were mis-identified as La. fullerkati (see Results) were placed into E. waccamawensis for this analysis.

At the end of the experiment, all animals from the field enclosures were collected, transported on ice to UNCW, and stored frozen at -40[degrees]C. All shells, soft tissues, and portions of tissue extracts of all of these specimens are deposited at the North Carolina Museum of Natural Sciences in Raleigh, North Carolina.

Methods of Genetic Identification

DNA was extracted from 2-3 [mm.sup.3] pieces of mantle tissue from frozen specimens and/or 20-100 [micro]L of hemolymph using the PureGene DNA extraction kit (Gentra Systems, Minneapolis MN), with the kit protocol scaled to small-yield tissue samples and the RNase digestion step omitted. A StrataPrep PCR Purification Kit (Stratagene, La Jolla CA) was then used to purify the genomic DNA extracts, following the kit protocol for PCR products. Whereas purification of genomic DNA is not the intended application of this StrataPrep kit, we regularly use PCR purification spin columns to clean up genomic DNA purified by rapid sodium acetate/isopropanol precipitation methods. In our initial work, we found this additional purification to greatly improve our PCR amplification success from hemolymph DNA extracts. The eluted DNA (from the StrataPrep column) was then vacuum-dried, resuspended in 10-15 [micro]L of PCR-grade water and stored at -20[degrees]C until analysis.

A portion of the 16s ribosomal RNA (rRNA) gene in the mitochondrial genome was amplified using universal primers (Lydeard et al. 1996): 16sARLMyt (5' CGACTGTTTAA CAAAAACAT 3') and 16sBRHMyt (3' ACATGTGCT GAGTTCAGAACGG 5'). Internal PCR and sequencing primers were designed from our initial freshwater mussel DNA sequences to improve success with these species: 16sUN693F (5'AGATAATGCCTGCCCAGTG 3') and 16sUNl178R (5' CGGTCTTAACTCAGCTCGTGTA 3'). PCR reactions used 1X PCR buffer with 1.5 mM Mg[Cl.sub.2] [Applied Biosystems (ABI), Foster City, CA], 0.2 mM each dNTP, 0.5 [micro]M each primer, and 1 U Taq polymerase (ABI) in a 25 [micro]L final volume. Cycling parameters were: an initial 5:00 at 94[degrees]C, followed by 35 cycles of (94[degrees]C for 1:00, 50[degrees]C for 1:00, and 72[degrees]C for 2:00), followed by a final 5:00 soak at 72[degrees]C. Reactions were carried out using a PTC- 100 Thermal Cycler (MJ Research Inc., Waltham Massachusetts).

Primers and salts were removed using the StrataPrep PCR Purification Kit. PCR products were sequenced from the forward and reverse primers using the ABI Big Dye Terminators Kit Version 3.1 and the ABI 3100 Genetic Analyzer. Sequences were edited using Sequencher (Gene Codes Corporation, Ann Arbor, Michigan), aligned using Clustal X (Thompson et al. 1994) and the alignment was imported into MacClade 4.0 (Sinauer Associates, Sunderland, Massachusetts) for final editing.

Sequence alignments of the 16s rRNA gene were screened for recognition sequences of restriction endonucleases that contained DNA substitutions that were fixed between La. fullerkati and E. waccamawensis. Hinf I (recognition sequence = GANTC), Ava II (recognition sequence = GGWCC), and Hind III (recognition sequence = AAGCTT) were chosen as candidates, based upon an initial alignment of 5 individuals per species that was expanded to include 10 La. fullerkati and 16 E. waccamawensis (see Results). Purified 16s PCR products were digested with these restriction enzymes (obtained from New England BioLabs [NEB], Beverly MA). All digests used single enzymes, and were completed in 20 [micro]L volume reactions containing 10 [micro]L of the following cocktails and 10 [micro]L of purified PCR Products. Ava II cocktails contained 1 x Buffer #4 (New England BioLabs), 8 units Ava II (10 units/[micro]L), and sterile H20. Hinf I cocktails contained 1 x Buffer #2 (New England BioLabs), 8 units Hinf I (10 units/[micro]L), and sterile [H.sub.2]O. Hind III cocktails contained 1 x Buffer #2, 20 units Hind III (20 units/[micro]L), and sterile [H.sub.2]O. The digests were incubated at 37[degrees]C for at least 16 h, then loaded onto 1.8% NuSieve 3:1 agarose gels (Cambrex Bio Science, Inc., Rockland, Maine).

As an additional application of genetic identification, we focused on two cases where cryptic taxa were revealed in field collection and phylogenetic analysis of mussels in rivers and streams in southeastern North Carolina (Sommer 2007). The first case is an unknown species that we believe to be a member of the genus Uniomerus, but that was morphologically identified in our collections as Elliptio complanata. The second case is a cryptic species of Lampsilis that we collected from the YadkinPee Dee, and other biologists (Alderman pers. com.) have collected from upper reaches of the Pee Dee and Tar Rivers. Further information on geographic source, DNA sequence and tissue depositories of these specimens are available in a companion paper (Sommer et al. submitted). DNA was extracted from hemolymph and body tissues, PCR amplification of the 16s region was performed as above, and the sequences were aligned and examined to locate species-diagnostic substitutions that would be recognized by restriction enzymes.

Field Trial of Genetic Identification Methods

In August 2004, belt transects were placed at two locations in Lake Waccamaw, a subset of animals collected from these transects were hemolymph-extracted, and our PCR-RFLP assay was used to check morphological identification of these animals. Our goal was not to provide a census of mussel diversity in the Lake, a goal clearly beyond the reach of our limited spatial sampling and outside the scope of this project. Rather, the purpose of this work was to serve as a field-based evaluation of an application of our methods. Lake Waccamaw is an ideal site in which to carry out such an evaluation, for two reasons. First, animals are abundant and easy to access, and any mortality we may have inflicted would have little impact on the large lake populations. And second, the numerically dominant species E. waccamawensis, co-occurs with a much less abundant, genetically very distinct species (La. fullerkati) that is morphologically difficult to distinguish. Hence this field trial could serve as a model for other cases that may confront field biologists surveying sites where a morphologically cryptic, threatened species co-occurs with a more abundant species.

We collected from two transects at two nearshore sites with distinct ecological characteristics. The north shore site (Fig. 1, site 2: N 34.3069[degrees], W 78.5012[degrees]) has deeper water near the shore than the area near the State Park, but we worked closer to shore so water depths were similar (1.5-2 m). The substrate at the north shore site has considerably more organic material and much more abundant emergent aquatic vegetation (Nuphar sp. and Panicum sp.) than at the State Park site. The second site was located just east of the small dam at the head of the Waccamaw River (Fig. 1, site 3: N 34.2610[degrees], W 78.5180[degrees]); again, at about 1.5-2 m depth. Dense growths of maidencane (Panicum spp.) and bald cypress (Taxodium ascendens) near the shore characterize this site.

Transects measured 30 m long and all mussels within 1 m to each side of the line were collected, identified, and counted by experienced North Carolina Dept. of Transportation (NCDOT) biologists, who are trained to survey mussel populations throughout the state. Mussels were photographed, and hemolymph was sampled from 8-10 individuals of each species (when numbers allowed). Hemolymph samples were processed as previously described and subjected to analysis by the PCR-RFLP methods described above.


Field Assessments of Growth and Survival

The fraction of individuals recovered was very high and not different between the two field enclosures, so the values for both were pooled (Table 1). Only a single animal was found dead at the end of the experiment and it was a control animal. A total of 8 individuals were not recovered at the end of the experiment: 5 of these were control and 3 were hemolymph-extracted animals. The lost animals either died, escaped by burrowing underneath the fence, or they burrowed so deeply that they were missed during sampling. An observation consistent with either of the latter two explanations is that seven of the nine individuals lost were Le. ochracea, which is the species that we observed to burrow rapidly, and deeper than the other species (1 of the control animals in each species was lost: Table 1). We found no increase in mortality among the hemolymph-sampled animals.

The data on growth, measured as the increase in shell length after 8 wk, showed that mean shell length added in the control groups was reduced approximately 5 percent to 8 percent in the hemolymph extracted animals (Table 1). For example, the reduction for Le. ochracea = (3.20 3.03)/3.20 = 5.3% of the control shell growth increment. But neither the main effect of hemolymph extraction nor the interaction between hemolymph extraction and species were significant by 2-way ANOVA (Table 2). Both control and experimental groups showed an increase in size in all 3 species. Le. ochracea showed the fastest growth rate-5.1 and 7.2 times that of E. waccamawensis and La. fullerkati, respectively. Two-way ANOVA (Table 2) showed that growth rate differences between species were highly significant, and posthoc analysis (Fisher's PLSD) showed a highly significant difference between Le. ochracea and both of the other two species (P < 0.001 for both comparisons), but no significant difference between E. waccamawensis and La. fullerkati (P = 0.189).

Genetic Identification Methods

Predicted numbers of fragments and fragment sizes from "virtual digests" of sequence alignments (Table 3) were in excellent agreement with actual PCR-RFLP fragment patterns. The results from genetic identification of the Lake Waccamaw animals are treated in detail below. Among the stream-dwelling cryptic species, the first case involved the animals that were identified as Elliptio complanata. Two individuals (YPD2 and YPD3) from Morrow Mountain State Park in the Yadkin-Pee Dee River showed sequences very similar to other Elliptio (Sommer 2007). Two other individuals from the same collection at Morrow Mountain (YPD1 and YPD4) and two individuals from the Lumber River (Richland Swamp), however, showed very different DNA sequences, despite their morphological identification also as E. complanata. Mean genetic distances between these two sets of individuals was about 14% for the 16s region--a substantial difference. Field biologists often have trouble distinguishing E. complanata from Uniomerus (K. Lynch pers. com.) and our phylogenetic analysis grouped these individuals with Uniomerus carolinianus (Sommer 2007). The enzymes Tsp 5091 and Msel were predicted to produce diagnostic RFLPs between the cryptic taxon and E. complanata, and all 6 of the individuals in question produced the predicted fragment patterns (Table 3).

The second case concerns a cryptic taxon of Lampsilis (L. sp "Tar River" in Table 3). We tested samples of this taxon collected by wildlife conservation consultant J. Alderman from Lick Creek (Upper Pee Dee) and from the upper Tar River. 16s DNA sequences of these animals were very similar to sequences of two individuals (YPD1, YPD5) that we collected from Morrow Mountain and identified morphologically as L. radiata, but were 5.6% different from 16s sequences of two other L. radiata individuals (YPD2, YPD4) that we collected from the same site. In this case, the enzymes EcoR V and Rsa I were predicted to yield diagnostic fragments to distinguish the "upper Tar" cryptic taxon, and again, all 6 of the individuals in question produced the predicted fragment patterns (Table 3). Hence in both cases, we have developed good candidate RFLPs for distinguishing these problematic cryptic taxa that are sympatric with other common species.

To evaluate genetic identification methods on the Lake Waccamaw animals, individuals of both E. waccamawensis and La. Jullerkati were first used to determine the success rate of amplification from hemolymph-extracted DNA, and the quality of the resulting DNA sequences. Because of the lower yield of DNA, amplification success was lower with the hemolymph extracts, but only slightly so. In E.waccamawensis, 24/24 mantle extracts and 18/20 (90%) of hemolymph extracts, and in La. fullerkati, 20/20 mantle extracts and 15/16 (94%) of hemolymph extracts successfully amplified. For all individuals, there were no differences between DNA sequences obtained from mantle and from hemolymph, confirming that hemolymph provided a reliable source of DNA that was free of contaminating DNA, and free from PCR artifacts than can result from low concentration or poor quality of template DNA.

All mussels used in the growth and survival studies were genetically typed using PCR-RFLP assays. Agarose gels of PCR-RFLP products (two representative ones shown in Fig. 2) conformed to expected patterns (Table 2). Hinf I and Hind III each cut 16s rDNA PCR products from E. waccamawensis into two fragments, but did not cut products from La. fullerkati. Conversely, Ava II cut the La. fullerkati 16s amplicons into two fragments, but did not cut E. waccamawensis products. On the Hinf I and Ava II gels (Fig. 2), lane 2 shows the fragment patterns produced from an individual whose RFLP phenotype with both enzymes did not agree with its morphological identification as La. fullerkati. Subsequent DNA sequence analysis of this individual confirmed that it was E. waccamawensis.

Each time we encountered a discrepancy between RFLP and morphological identification, we sequenced the DNA of the PCR product from the individual in question. Morphological misidentification of E. waccamawensis as La. fullerkati occurred in a total of 2 individuals used in the enclosure experiment (Table 4). In each case, the misidentification was detected using each of the 3 restriction enzyme assays, and in each case, the individual was a control animal.


Sequence analysis was also used when the RFLP phenotype was ambiguous (Table 4, Fig. 2) or unusual. In 2 cases, digestion with the enzyme Hinf I yielded a clear RFLP result that conflicted with morphological identification, but on sequencing, we found that the 16s sequence agreed with the morphological identification of E. waccamawensis. Inspection of these sequences revealed the presence of restriction site variant in E. waccamawensis that occurred at a frequency of ~7% (3/39). Consequently, we discontinued use of Hinf I for subsequent diagnostic work on these species. Continued sampling with both Ava II and Hind III (Table 4) revealed no site polymorphism, so both of these enzymes appear to be diagnostic.

Our alignment of DNA sequences (Fig. 3 shows a subset) was based upon 10 individuals of La. fullerkati, 16 individuals of E. waccamawensis (including the 3 Hinf I site-variants), and 3 individuals morphologically identified as La. fullerkati whose RFLP patterns were consistent with E. waccamawensis. This alignment contained 419 aligned nucleotide positions, at which E. waccamawensis and La. fullerkati differ at 46 fixed substitutions. Fixed substitutions are those that are differ between species but are invariant within species. These fixed subsitutions comprised 10.9% uncorrected sequence divergence between the two species. Fixed substitutions within restriction enzyme recognition motifs would distinguish these two species by RFLP. Three out of 16 E. waccamawensis that we sequenced and whose 16s PCR products did not cut with Hinf I indeed contained a C to T transition that eliminated this site (Fig. 3: position 102). This confirmed that Hinf I digests were not reliable as species-diagnostic assays. Hind III and Ava II sites, however, were found to be fixed and species diagnostic in our entire DNA sequence data set (Fig. 3). The agreement in species identification based on morphology and RFLP phenotype at each of these enzymes for over 100 animals (Table 3) confirms that these two enzymes produce reliably diagnostic markers to discriminate the Lake Waccamaw species.

Field Trial of Genetic Identification Methods

Densities of animals were high at the north shore site (mean of 7.5/[m.sup.2]) but densities were lower (3.7/[m.sup.2]) and distributions more patchy at the dam site (Appendix). Percent composition was very similar, with rank abundances being E. waccamawensis (90%), Leptodea ochracea (6% to 12%), and Lampsilis fullerkati (~1%), with an occasional Elliptio folliculata (I. Lea, 1838). PCR-RFLP identifications with Ava II and with Hind Ill were consistent with morphological identifications in 42/43 and 41/43 cases, respectively (Table 4). Only one Hind III ambiguity was found for a single Lampsilis fullerkati, and its correct morphological ID was confirmed through sequencing. A single misidentification was discovered in the case of 1 individual identified morphologically as Lampsilis fullerkati. RFLP assays and DNA sequencing confirmed that this individual was actually E. waccamawensis. This result again demonstrates the usefulness of PCR-RFLP for checking identifications of morphologically conservative freshwater mussels in field surveys.


Our results show that hemolymph sampling provides a nonlethal and efficient method for obtaining DNA from freshwater mussels. Hemolymph removal produced no mortality and its effects on growth were small and statistically undetectable in individuals from three genera that were monitored for 8 wk in their natural environment. Routine collection of hemolymph by field biologists can be accomplished with little training, and would provide a tissue sample archive for genetic analysis with applications, for example, in long-term census of threatened populations, screening for nonnative species, or monitoring the success of reintroductions (O'Beirn et al. 1998, Raley et al. 2006).

Mantle clipping has been widely used as a nonlethal source for DNA analysis of freshwater mussels (Campbell et al. 2005, Eackles & King 2002, Grobler et al. 2006, Jones et al. 2006). Its popularity derived from field experiments on a large number of individuals (approximately 100 per species in each treatment group) that demonstrated no detectable effects on the survivorship of (Actinonaias ligamentina (Lamarck, 1819; Lampsilinae) and Quadrula quadrula (Rafinesque, 1820; Ambleminae) over one year's time (Berg et al. 1995). Henley et al. (2006), however, found that mantle clipping produced mortality in 9/16 individuals of Epioblasrna triquetra (Rafinesque, 1820) after 1.5 y of postbiopsy observation, and developed a technique of "brush-swabbing" from the viscera, foot, and mantle that they demonstrated to be an adequate source of DNA. Health effects of brush-swabbing have yet to be evaluated, but they are likely to be minimal.

Like hemolymph sampling, brush swabbing presents some challenges with details of technique application, and with DNA yield (Henley et al. 2006), that must be overcome for it to be routinely applied to large population sampling. Henley et al. (2006) encountered tissue disruption and high standard deviations in DNA yield when swabbing mantle tissue, but obtained higher DNA yields and lower standard deviation using 4 swab passes (compared with fewer passes) over the viscera and foot. In addition, they reported a general problem with high protein yield from mucous that was removed along with integumental cells by the brush swab.

A useful future study (which we have not done) would compare the methods of brush-swabbing and hemolymph sampling for ease of application and yield of PCR-quality DNA from a range of species and shell sizes, all monitored for growth and mortality effects under field conditions. A potential disadvantage of hemolymph sampling is that it is invasive, and brush swabbing appears to not be (although mantle brushing apparently does injure tissue). Although we found health effects to be absent (or slight at most), care must be taken to use adductor muscle as extraction from the pericardial sinus can kill animals (Gustafson et al. 2005).

Whereas we have no experience with the brush swabbing to allow evaluation of its merits relative to the method we chose, we suggest the following advantages of hemolymph sampling to be considered. (1) It is easy to train, and rapid (100-200 uL can be drawn in less than 10 sec); (2) It is adaptable to a wide range of shell shapes and sizes. The largest specimens can be sampled with little effort, and needle length and gauge can be adjusted to shell size and shape of the target species. Conversely, even very small specimens, yielding as little as 20 [micro]L hemolymph, generate sufficient DNA yield for PCR followed by RFLP typing and DNA sequencing. Small specimens might be a challenge for the swabbing method. (3) Hemolymph is a tissue of consistent composition that offers adequate yield of PCR-quality DNA and little contaminating protein when purified using our methods. Protein contamination is a problem for the swabbing method (Henley et al. 2006), but this can apparently be overcome. (4) Hemolymph sampling offers the option of simultaneous evaluation of hematological data for monitoring the health of threatened species (Gustafson et al. 2005).

Gustafson et al. (2005) showed that hemolymph could be withdrawn from individuals repeatedly, and used circulating cell counts, protein, and ion analysis of the hemolymph as a means to demonstrate that the health of the extracted animals, maintained in laboratory aquaria, was unaffected over a 3-mo period. From this they also developed a set of reference range standards for hematological data from undisturbed populations of Elliptio complanata that can be used to evaluate health of other populations. Our study adds to this earlier one by providing mortality estimates of hemolymph extraction under field conditions, and expands the taxonomic range over which mortality has been evaluated. The suggested negative effect on growth in our study indicates that longer-term monitoring should be used in future work. For instance, Henley ct al. (2006) followed animals post-tissue biopsy over 1.5 y.


Coupled to hemolymph extraction or other nonlethal method of tissue sampling, the PCR-RFLP technique we used here shows great promise as a reliable and inexpensive method of genetic identification. Field biologists studying freshwater mussels are often faced with problems similar to the ones we approached here--a morphologically similar or cryptic taxon of conservation interest coexists with another more abundant taxon. The mitochondrial 16s region that we used contained more than enough sequence divergence between these taxa to differentiate them. Faster-evolving mtDNAs offer even greater power of discrimination between more closely related taxa in this geographic region (Sommer 2007).

The methods are rapid and require only PCR and agarose electrophoresis apparatus, and hence can be used as part of a routine survey program without the need for access to a DNA sequencing facility. We suggest the following steps be followed to develop such a program. First, a regional DNA sequence database should be accumulated to produce a hierarchical "key" of restriction enzymes that recognize target species (see below). Second, RFLP is a "first-pass" diagnostic; discrepancies between morphological and RFLP identification should be checked with DNA sequencing. This is because of the fact that ambiguous restriction fragment patterns and those that result from unexpected polymorphisms in recognition sites, such as we found, will be a regular consequence when large samples over broad geographic ranges are surveyed. And third, biologists must be prepared to deal with the discovery of novel DNA sequences, representing cryptic taxa that have not been previously encountered from a geographic region. Nonlethal sampling, of course, carries the disadvantage that the tissue archive is not associated with deposited specimens, so the approach we describe is of course not a substitute for systematic study in which DNA sequences should always be associated with vouchered shells and tissue samples. However, sacrificing animals will not be an option for many populations and our approach would minimize the need for this in routine surveys. As DNA sequence databases for populations grow, identifying unknown taxa will be simplified and in such cases, the source location of the novel sequence can be resurveyed in a targeted fashion to search for the taxon in question.

White et al. (1994) developed a PCR-RFLP key that could be used as a model for future work. They used their key to identify freshwater mussel glochidia collected off host fishes to the level of species, to determine their host specificity. By using specific primers that prevented contaminating PCR of host fish DNA, they amplified the nuclear ITS-1 region, and digested products with the enzymes Msp I, Bam HI, and Sau 96I. Fragments from these could be used to distinguish most of the freshwater mussel species in their collection (White et al. 1994). They then developed a diagnostic PCR-RFLP key that could be used to distinguish among glochidia found in French Creek, PA (White et al. 1996). The key works in a hierarchical fashion, like a morphological identification key. The authors confirmed the accuracy of their glochidium key using tissue extracts from adults that were morphologically identified. Gerke and Tiedemann (2001) and Kneeland and Rhymer (2007) developed keys to the glochidia of the six species of Europe and the 10 species known from Maine, respectively. Whereas PCR-RFLP keys to the more diverse adult fauna in many geographic regions would require more markers and larger DNA sequence databases, they would be invaluable resources for biologists.
Freshwater mussel belt transect surveys in Lake Waccamaw.

Site          Transect   Species                   N    %

North shore      1       Leptodea ochracea         41    9.1
                         Elliptio waccamawensis   399   89.0
                         Elliptio folliculata       1    0.2
                         Lampsilis fullerkati       6    1.3
                         Unknown                    1    0.2
                         Total                    448

North shore      2       Leptodea ochracea         55   12.0
                         Elliptio waccamawensis   400   87.3
                         Elliptio folliculata       3    0.6
                         Lampsilis fullerkati     458

Near dam         1       Leptodea ochracea          8    6.0
                         Elliptio waccamawensis   125   93.3
                         Elliptio folliculata       1    0.7
                         Lampsilis fullerkati     134

Near dam         2       Leptodea ochracea         20    6.4
                         Elliptio waccamawensis   286   92.0
                         Elliptio folliculata       2    0.6
                         Lampsilis fullerkati       3    1.0
                         Total                    311

              Number of Tissues
Site              Sampled

North shore          10

North shore          10

Near dam              8

Near dam              2


The authors thank Dr. R. Heise and R. Nichols of NCWRC; who collected and morphologically identified most of the specimens, led and guided field excursions, and helped with the field enclosure work in Lake Waccamaw, S. Van Horn and A. Rogers; who provided additional assistance, K. Lynch and colleagues at the North Carolina Department of Transportation (NCDOT); who helped with collection and mussel identification, C. Slaughter; who helped with earlier portions of the field and laboratory work, J. Alderman and J. Ratcliffe; who helped inspire our interest in this project back in 2001 and provided specimens of Lampsilis from the upper Tar, Flat, Waccamaw, and Pee Dee Rivers, and Dr. A. Bogan from the North Carolina Museum of Natural Sciences; who provided valuable advice, tissues from topotype specimens, and comments on the manuscript. Lake Waccamaw State Park provided a permit for work within park boundaries. Support for this project was by the NCDOT (project # 2004-09).


Baker, A. M., C. Bartlett, S. E. Bunn, K. Goudkamp, F. Sheldon & J. M. Hughes. 2003. Cryptic species and morphological plasticity in long-lived bivalves (Unionoida: Hyriidae) from inland Australia. Mol. Ecol. 12:2707-2717.

Berg, D. J., W. R. Haag, S. I. Guttman & J. B. Sickel. 1995. Mantle biopsy: a technique for nondestructive tissue-sampling of freshwater mussels. J. N. Am. Benthol. Soc. 14:577-581.

Bogan, A. E. 2002. Workbook and key to the freshwater bivalves of North Carolina. North Carolina Freshwater Mussel Conservation Partnership, Raleigh, North Carolina. pp. 101.

Bogan, A. 2008. Global diversity of freshwater mussels (Mollusca, Bivalvia) in freshwater. Hydrobiologia 595:139-147.

Campbell, D.-C., J.-M. Serb, J.-E. Buhay, K.-J. Roe, R.-L. Minton & C. Lydeard. 2005. Phylogeny of North American amblemines (Bivalvia, Unionoida): prodigious polyphyly proves pervasive across genera. Invertebr. Biol. 124:131-164.

Eackles, M. S. & T. L. King. 2002. Isolation and characterization of microsatellite loci in Lampsilis abrupta (Bivalvia: Unionidae) and cross-species amplification within the genus. Mol. Ecol. Notes 2:559-562.

Ford, S. E. 1986. Effect of repeated hemolymph sampling on growth, mortality, hemolymph protein and parasitism of oysters, Crassostrea virginica. Comp. Biochem. Physiol. A 85:465-470.

Fyhn, H. J. & J. D. Costlow. 1975. Anaerobic sampling of body fluids in bivalve molluscs. Comp. Biochem. Physiol. A 52:265-268.

Gerke, N. & R. Tiedemann. 2001. A PCR-based molecular identification key to the glochidia of European freshwater mussels (Unionidae). Conserv. Genet. 2:287-289.

Grobler, P. J., J. W. Jones, N. A. Johnson, B. Beaty, J. Struthers, R. J. Neves & E. M. Hallerman. 2006. Patterns of genetic differentiation and conservation of the slabside pearlymussel, Lexingtonia dolabelloides (Lea, 1840) in the Tennessee River drainage. J. Molluscan Stud. 72:65-75.

Gustafson, L. L., M. K. Stoskopf, A. E. Bogan, W. Showers, T. J. Kwak, S. Hanlon & J. F. Levine. 2005. Evaluation of a nonlethal technique for hemolymph collection in Elliptio complanata, a freshwater bivalve (Mollusca: Unionidae). Dis. Aquat. Organ. 65:159-165.

Henley, W. F., P. J. Grobler & R. J. Neves. 2006. Non-invasive method to obtain DNA from freshwater mussels (Bivalvia: Unionidae). J. Shellfish Res. 25:975-977.

Hughes, J., A. M. Baker, C. Bartlett, S. Bunn, K. Goudkamp & J. Somerville. 2004. Past and present patterns of connectivity among populations of four cryptic species of freshwater mussels Velesunio spp. (Hyriidae) in central Australia. Mol. Ecol. 13:3197-3212.

Johnson, R. I. 1970. The systematics and zoogeography of the Unionidae (Mollusca:Bivalvia) of the southern Atlantic slope region. Bull. Muse. Compar. Zool. 140:263450.

Johnson, R. I. 1984. A new mussel, Lampsilis (Lampsih's') fullerkati (Bivalvia:Unionidae) from Lake Waccamaw, Columbus County, North Carolina, with a list of the other unionid species of the Waccamaw River system. Occasional papers on mollusks 4. pp. 305-319.

Jones, J. W., R. J. Neves, S. A. Ahlstedt & E. M. Hallerman. 2006. A holistic approach to taxonomic evaluation of two closely related endangered freshwater mussel species, the oyster mussel Epioblasma capsaeformis and tan riffleshell Epioblasma florentina walkeri (Bivalvia: Unionidae). J. Molluscan Stud. 72: 267.

King, T.-L., M.-S. Eackles, B. Gjetvaj & W.-R. Hoeh. 1999. Intraspecific phylogeography of Lasmigona subviridis (Bivalvia: Unionidae): Conservation implications of range discontinuity. Mol. Ecol. 8:S65-S78.

Kneeland, S. C. & J. M. Rhymer. 2007. A molecular identification key for freshwater mussel glochidia encysted on naturally parasitized fish hosts in Maine, USA. J. Molluscan Stud. 73:279-282.

Lydeard, C., M. Mulvey & G.-M. Davis. 1996. Molecular systematics and evolution of reproductive traits of North American freshwater unionacean mussels (Mollusca: Bivalvia) as inferred from 16S rRNA gene sequences. Philos. Trans. R. Soc. Lond. B Biol. Sci. 351:1593-1603.

Naimo, T. J., E. D. Damschen, E. M. Monroe & R. G. Rada. 1998. Nonlethal evaluation of the physiological health of unionid mussels: methods for biopsy and glycogen analysis. J. N. Am. Benthol. Soc. 17:121-128.

O'Beirn, F.-X., R.-J. Neves & M.-B. Steg. 1998. Survival and growth of juvenile freshwater mussels (Unionidae) in a recirculating aquaculture system. American Malacological Bulletin.

Raley, M. E., J. F. Levine & A. Bogan. 2006. Hemolymph as a nonlethal and minimally invasive source of DNA for molecular systematic studies of freshwater mussels. Tentacle. 14:33-34.

Roe, K. J., P. D. Hartfield & C. Lydeard. 2001. Phylogeographic analysis of the threatened and endangered superconglutinateproducing mussels of the germs Lampsilis (Bivalvia: Unionidae). Mol. Ecol. 10:2225-2234.

Serb, J. M., J. E. Buhay & C. Lydeard. 2003. Molecular systematics of the North American freshwater bivalve genus Quadrula (Unionidae: Ambleminae) based on Initochondrial NDI sequences. Mol. Phylogeneh Evol. 28:1-11.

Sommer, K. 2007. Genetic identification and phylogenetics of Lake Waccamaw endemic freshwater mussel species. M.S., University o1" North Carolina, Wilmington, Wilmington, North Carolina. 107 Pp.

Stein, B. A. & R. M. Chipley. 1996. Priorities for conservation: 1996 annual report card for United States plant and animal species. Arlington, Virginia: The Nature Conservancy.

Thompson, J. D., D. G. Higgins & T. J. Gibson. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl Acids Res. 22:4673-4680.

White, L. R., B. A. McPheron & J. R. Stauffer, Jr. 1994. Identification of freshwater mussel glochidia on host fishes using restriction fragment length polymorphisms. Mol. Ecol. 3:183-185.

White, L. R., B. A. McPheron & J. R. Stauffer, Jr. 1996. Molecular genetic identification tools for the unionids of French Creek, Pennsylvania. Malacologia 38:181-202.

Williams, J. D., M. L. J. Warren, K. S. Cummings, J. L. Harris & R. J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18:6-22.

Yanick, J. F. & D. D. Heath. 2000. Survival and growth of mussels subsequent to hemolymph sampling for DNA. J. Shellfish Res. 19:991-993.

MICHAEL A. MCCARTNEY,* KRISTINE SOMMER AND AMI E. WILBUR, Department of Biology and Marine Biology, Center for Marine Science, University of North Carolina Wilmington, 5600 Marvin Moss Lane, Wilmington, North Carolina 28409

* Corresponding author. E-mail:
Growth and survivorship of hemolymph-extracted and control
animals over 8 wk in Lake Waccamaw field enclosures.

                          Number of Mussels   Number of Mussels
Genus         Treatment        Week 0              Week 8

Elliptio      Control            22                  21
              Extracted          20                  20
Lampsilis *   Control            14                  13
              Extracted          16                  16
Leptodea      Control            20                  16
              Extracted          20                  17
Total         Control            56                  50
              Extracted          56                  53

              Mean [DELTA] Shell   Standard Error
Genus            Length (mm)         of the Mean

Elliptio            0.62                0.11
                    0.57                0.08
Lampsilis *         0.45                0.08
                    0.42                0.08
Leptodea            3.20                0.16
                    3.03                0.17

* Two of the individuals from the Lampsilis control group were
subsequently identified as E. maccamawensis and were excluded.

Two-Way ANOVA on species and treatment differences in
growth in the field enclosures.

                               Sum of    Mean
Factor                    df   Squares   Square      F

Species                    2   149.2     74.60    299.1 ***
Treatment                  1     0.192    0.192     0.772 ns
(hemolymph extracted
  /not extracted)
Interaction                2     0.082    0.041     0.165 ns
  (Species X Treatment)
Error                     96    23.94     0.249

df= degrees of freedom; *** P < 0.001; ns = not significant (P > 0.05).

Diagnostic PCR-RFLP of the 16s rRNA gene.

                Lake Waccamaw endemic species

Enzyme     Morphological species   Number Assayed

Hinf I       La. fullerkati              43
"            E. waccamawensis            69
Ava II       La. fullerkati              43
"            E. waccamawensis            69
Hind III     La. fullerkati              30
"            E. waccamawensis            39

                Lake Waccamaw endemic species

Enzyme     Number of Fragments   Fragment Size (bp)

Hinf I              1                    440
"                   2                 120, 320
Ava II              2                 280, 160
"                   1                    440
Hind III            1                    440
"                   2                 238, 202

              Cryptic species from NC streams

Enzyme               Taxon              Number Assayed

Tsp 5091   E. complanata                      2
"          E. complanata-like cryptic         4
Mse I      E complanata                       2
"          E. complanata-like cryptic         4
EcoR V     Lampsilis radiata radiata          2
"          Lampsilis sp. "Tar River"          4
Rsa I      Lampsilis radiata radiata          2
"          Lampsilis sp. "Tar River"          4

              Cryptic species from NC streams

Enzyme     Number of Fragments     Fragment Size (bp)

Tsp 5091            2                    84, 388
"                   4               84, 56, 160, 172

Mse I               3                  83, 316, 73
"                 5 or 6         315, 251, 395, 117, 573

EcoR V              3                 118, 315, 66
"                   2                    433, 66
Rsa I               3                 124, 264, 111
"                   4               124, 93, 171, 111

Results of PCR-RFLP identifications of Lake Waccamaw mussels.


                                                 RFLP & Morphology
Enzyme    Morphological Species   Number Typed      Consistent

Ava II      E. waccamawensis           39               38
             La. fullerkati            30               28
Hinf I       E. waccamwensis           39               36
             La. fullerkati            30               27
Hind III    E. waccamawensis           39               37
             La.fullerkati             30               28


          RFLP & Morphology                    RFLP Site
            Inconsistent      RFLP Ambiguous    Variant

Enzyme            0                 1              0
                  2                 0              0
Ava II            0                 1              2
                  2                 1              0
Hinf I            0                 2              0
                  2                 0              0
Hind III

                                                 RFLP & Morphology
Enzyme    Morphological Species   Number Typed      Consistent

Ava II      E. waccamawensis           30               30
             La. fullerkati            13               12
Hind III    E. waccamawensis           30               30
             La. fullerkati            13               11


          RFLP & Morphology                    RFLP Site
Enzyme      Inconsistent      RFLP Ambiguous    Variant

Ava II            0                 0              0
                  1                 0              0
Hind III          0                 0              0
                  1                 1              0
COPYRIGHT 2009 National Shellfisheries Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:polymerase chain reaction-restriction fragment length polymorphism
Author:McCartney, Michael A.; Sommer, Kristine; Wilbur, Ami E.
Publication:Journal of Shellfish Research
Article Type:Report
Geographic Code:1USA
Date:Apr 1, 2009
Previous Article:Histological studies on gametogenesis, hermaphroditism and the gametogenic cycle of Anodonta gabillotia pseudodopsis (Locard, 1883) in the Lake...
Next Article:Brown muscle disease and Manila clam Ruditapes philippinarum dynamics in Arcachon Bay, France.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters