Highly sensitive detection of invasive shore crab (Carcinus maenas and Carcinus aestuarii) larvae in mixed plankton samples using polymerase chain reaction and restriction fragment length polymorphisms (PCR-RFLP)

The brachyuran crab genus Carcinus consists of two species, C. maenas and C. aestuarii, both of which have invaded multiple regions of the globe. C. maenas has proven a particularly adept invader, establishing introduced populations on every non-polar continent outside its native range. This species has also exhibited the capacity to spread rapidly once established and has potential for significant ecological and economic impacts throughout its introduced range. The possibility of both species invading additional coastal ecosystems, and the importance of larval dispersal—both current-driven and ballast water-mediated—to the successful establishment and expansion of introduced populations, recommend the development of rapid and cost-effective tools for detecting and monitoring Carcinus larvae in environmental samples. We have developed a PCR-RFLP approach that enables the specific and highly sensitive detection of both C. maenas and C. aestuarii in mixed plankton samples, including those drawn from ballast water. Our approach successfully identifies specimens from throughout the native and introduced ranges of both species, and excludes all non-target brachyuran species tested, including a number of species whose ranges overlap with those of the Carcinus species. Sensitivity of our PCR-RFLP assay is extremely high, allowing the detection of single stage I zoea in over 1 gram (filtered weight) of mixed non-target plankton. The assay also successfully detected single larvae in mixed plankton derived from ballast water, indicating the potential utility of this approach as a tool for targeted screening of Carcinus sp. in ship’s ballast.


Introduction
The European green crab Carcinus maenas (Linnaeus, 1758) is a notoriously successful invasive species, with established non-native populations in Australia, South Africa, Japan, Atlantic and Pacific North America, and, most recently, Argentina (Hidalgo et al 2005;Carlton and Cohen 2003).In some of these regions the species has expanded rapidly from the site of initial introduction; on the Pacific coast of North America, a population introduced to San Francisco Bay in the late 1980s had expanded over approximately 1700 kilometers of coastline by the year 2000 (Yamada 2000;Yamada and Hunt 2000).Both ballast water-mediated translocation (Carlton and Cohen 2003;Cohen et al 1995) and natural current-driven dispersal of larval crabs (Thresher et al 2003;Yamada and Hunt 2000) have been implicated in the spread of C. maenas at both global and regional scales, and are likely to be the primary contemporary vectors of introduction for the species.A number of physiological and life history characteristics contribute to the species' global success: C. maenas exhibits extremely high fecundity (Broekhuysen 1936;Cohen et al 1995), possesses long-lived feeding larval stages (Queiroga et al 2002;Dawirs 1985;Cohen et al 1995), is tolerant of wide temperature and salinity ranges (Broekhuysen 1936), and is omnivorous and opportunistic in its feeding habits (Cohen et al 1995).
In addition to its capacity to colonize novel habitats and expand rapidly on a regional scale, C. maenas has the potential to cause significant ecological and economic disturbance in recipient ecosystems.The species has been shown to be capable of considerably altering physical habitat (Davis et al 1998;Lafferty and Kuris 1996) and reducing abundance of native benthic micro-and macrofauna (Grosholz and Ruiz 1995;Grosholz et al 2000;Lafferty and Kuris 1996).Of particular concern is the implication of C. maenas in the failure of commercial clam fisheries in Atlantic North America (Lafferty and Kuris 1996;Glude 1955;Floyd and Williams 2004), and the potential for similar negative effects in other regions where the species is introduced (Walton et al 2002).In addition, C. maenas is likely to directly compete with a number of native crustaceans, including commercially important species such as Dungeness crab (Cancer magister Dana, 1852) on the Pacific coast of North America (McDonald et al 2001;Jamieson et al 1998;Grosholz et al 2000;Grosholz and Ruiz 1995).
While C. maenas has achieved deserved notoriety, its congener Carcinus aestuarii (Nardo, 1847) has exhibited less extensive success as an invasive species.A native of Mediterranean Europe and North Africa, C. aestuarii has been introduced to both Japan and South Africa (Carlton and Cohen 2003;Geller et al 1997).Morphological and genetic data suggest that C. aestuarii, rather than C. maenas, is the dominant introduced crab species in Japan's Tokyo and Dokai Bays (Geller et al 1997;Yamada and Hauck 2001), although the possibility remains that these populations may actually be of hybrid origin (Bagley and Geller 2000;Yamada and Hauck 2001).Little is known about the potential impacts of this species in its introduced range, but studies have shown it to practice the same opportunistic omnivory as its sister species (Chen et al 2004).
Here we describe the development and evaluation of a PCR-based assay for the detection of C. maenas and C. aestuarii larvae in environmental samples.We employ a two-step process for species-specific identification, involving PCR amplification using genusspecific primers followed by species-specific restriction enzyme digestion of the resulting amplicon.We show that this approach is capable of correctly identifying C. maenas and C. aestuarii specimens from throughout both the native and introduced ranges of the two species.In addition, we demonstrate specificity of the assay by testing genus-specific PCR on nontarget crab species, including a number of species whose ranges overlap with those of the Carcinus species.The sensitivity of the assay regularly enables detection of single C. maenas larvae in mixed plankton samples (including ballast water samples) even when background non-target biomass is extremely high.This specific and highly sensitive assay for detection of Carcinus larvae should provide a valuable tool for managers and researchers interested in assessing the dispersal of these species.

Sample collection and processing
Brachyuran crab tissue samples utilized in this study were obtained from a number of sources.Recently collected, preserved (95% ethanol) or frozen specimens of crab species commonly found on the Pacific coast of North America were provided by Sylvia Yamada (Oregon State University) and Greg Jensen (University of Washington).
Additional preserved (70% ethanol) tissue samples were obtained from Rob Toonen (University of Hawaii).Non-Carcinus crab samples were processed for whole genomic DNA using the DNeasy Tissue Kit (Qiagen) on either gill or leg muscle tissue.All crab DNA samples were normalized to approximately 5 ng/l.First stage zoea of Carcinus maenas, preserved in 95% ethanol, were provided by Uwe Nettlemann and Klaus Anger (Alfred Wegener Institute).

Development of Carcinus-specific PCR-RFLP assay
Partial sequences of the mitochondrial cytochrome C oxidase subunit I (COI) gene were either provided by Joseph Roman (University of Vermont), generated at the EPA Molecular Ecology Research Branch in Cincinnati, or obtained from Genbank (all non-Carcinus haplotypes).In total, 99 Carcinus haplotypes (82 C. maenas and 17 C. aestuarii) and 34 haplotypes from non-Carcinus brachyuran crab species were aligned using ClustalX (Thompson et al 1997) and scanned by eye for conserved regions within the genus Carcinus and within the species C. maenas.Three regions were identified, two which were well conserved within the entire genus and one which was conserved only within C. maenas (Table 1).The former were chosen as sites for the design of Carcinusspecific primers CF3 (5'-TTAGGAGGGCCAG Table 1.Alignment of Carcinus and non-Carcinus COI sequences at forward and reverse primer binding sites and internal EcoNI restriction site.All sequences are given from 5' to 3' in the direction of the COI open reading frame.Identities are indicated with a period (.), gaps with a dash (-).For Carcinus species, we indicate the number of individual haplotypes with the given COI sequence at forward, reverse, and internal regions.Non-Carcinus species marked with an asterisk (*) were included in specificity tests.ATATAGCTTT-3') and CR3 (5'-CTAAAACCG GCAACGATAATAATAA-3'). Given the large amount of information available on this genus and significant variability within the COI locus, it was impossible to identify primer sites that were 100% conserved across all known haplotypes.As a result, a number of known Carcinus haplotypes exhibit mismatches within the primer binding sites (see Results).Primer sites were chosen so as to minimize these mismatches, particularly at the primer 3' end, and all mismatched haplotypes were tested directly for Carcinus-specific amplification.For the reverse primer site, only a subset (n = 14) of non-target COI sequences could be aligned over the entire primer binding site.An internal EcoNI site was found to be 100% conserved within C. maenas (all haplotypes) but absent from all C. aestuarii haplotypes.This site was chosen for a species-specific diagnostic test based on restriction digestion of the Carcinusspecific amplicon.The predicted size of the Carcinus-specific PCR product is 348 basepairs for both C. maenas and C. aestuarii; digestion of this fragment by EcoNI is predicted to result in two products of 212 and 136 basepairs in C. maenas, but in an undigested 348 basepair product in C. aestuarii.We also identified two FokI restriction sites that flank the EcoNI site in C. maenas; species-specific RFLP identification was possible with this enzyme as well, and gave results equivalent to those obtained with EcoNI digestion (data not shown).

Molecular protocols
All PCR reactions were conducted in 15 l total volume, and contained 0.5 units Taq DNA polymerase, 1x Mg-free PCR buffer, 1.7 mM MgCl2, 0.67 mM dNTPs, 1 M each forward and reverse PCR primers, and 1 L template DNA normalized to 5 ng/L.For reactions designed to control for successful DNA extraction, universal COI primers COIF-PR115 (5'-TCWACNAAYC AYAARGAYATTGG-3') and COIR-PR114 (5'-ACYTCNGGRTGNCCRAARARYCA-3') were used (Folmer et al 1994), yielding an amplicon of approximately 700 bp from all crab species tested.Control PCR cycling parameters consisted of a 5 min denaturation cycle at 94º C, followed by 35 cycles of 1 min at 94º C, 1 min at 50º C, and 1 min at 72º C , with a final extension step of 15 min at 72º C. For Carcinus-specific amplification using primers CF3 and CR3, cycling parameters were identical except annealing was conducted at 65º C instead of 50º C.
All PCR products were run on 1.5% agarose gels (unless otherwise noted) and stained with ethidium bromide; gel images were digitized with KodakGL100 software.For specificity tests, equal volumes of both universal and Carcinusspecific PCR reactions (3 L each) were run together in a single well to provide internal controls for each sample.All sensitivity tests were conducted with C. maenas larvae.
For C. maenas-specific RFLP detection, 5 L of PCR product from Carcinus-specific PCR reactions was digested for 3 hours at 37º C in a 10 L reaction including 1x buffer (NEB 4) and 1.5 units EcoNI (New England Biolabs).The entire volume of restriction digested product was loaded into a single well for visualization.

Preparation of plankton samples
Ballast water samples were collected from the vessel General Villa ported in Sacramento, CA (originating from Long Beach, CA) on August 28, 2004.Samples were collected by light trap (20 minute deployment) and preserved in 95% ethanol.Preserved plankton was filtered through 8 micron filters (Millipore) under vacuum, washed with 95% ethanol, and allowed to dry under vacuum for 3 minutes.For all sensitivity tests, between 1 and 3 mL settled plankton volume was filtered.Weights were recorded for all filtered and vacuum-dried samples before processing for DNA; these measurements are hereafter referred to as "filtered weight."For ballast plankton, experimental samples were spiked with individual C. maenas larvae (one larva per sample) after filtration.
To further explore the sensitivity of our approach, and due to limitations on the amount of plankton available from ballast samples, we generated larger scale mock plankton communities consisting of mixed, cultured zooplankton and containing up to 1.359 grams of biomass (filtered weight).These samples included Artemia salina (Linnaeus, 1758) nauplii, Daphnia pulex (Leydig, 1860), Daphnia magna (Straus, 1820), Ceriodaphnia dubia (Richard, 1894), and Hyalella azteca (Saussure, 1858) in unknown proportions.Live plankton were collected from culture and preserved in 95% ethanol.Approximately 5 mL settled plankton volume was used for each experimental sample; samples were spiked with 1, 5, 10, or 20  Carcinus maenas larvae and filtered as described above.To more closely mimic real-world attempts at detection, these samples were spiked and mixed thoroughly prior to filtration.This allows for the possibility of target organisms being lost in the filtration process.Filtered weights of were recorded for each sample before processing for DNA.

DNA extractions
For sensitivity tests, we processed filtered ballast water plankton, with or without added C. maenas larvae, using either the DNeasy Plant Kit (Qiagen) or the PowerSoil Kit (MoBio).These kits proved more efficient than the DNeasy kit for processing samples with more than 100 mg of filtered biomass.However, since mock plankton communities consisting of cultured zooplankton (with or without C. maenas larvae) contained far more biomass than the recommended limit for these kits, we utilized the PowerMax Soil Kit (MoBio) for processing these samples.All DNA extractions from plankton samples were conducted on dried, filtered plankton according to protocols provided by the commercial suppliers.

Assay specificity
Amplification products from universal COI PCR demonstrate the presence of amplifiable mitochondrial DNA in all samples (Figure 1).Carcinus-specific PCR primers CF3 and CR3 amplified the predicted 348 basepair fragment from all tested Carcinus samples, but failed to amplify from all non-Carcinus samples at the 65º C annealing temperature (Figure 1 and Table 2).Sequence alignments suggest that our Carcinusspecific primers are unlikely to successfully amplify from any of the 34 non-Carcinus species investigated (Table 1); this prediction is confirmed for 8 species which were tested directly in specificity assays and whose COI haplotypes were included in the primer design (Figure 1).An additional 9 species were also shown to be excluded non-targets, despite not being included in the primer design stage of assay development.Carcinus samples include specimens from throughout both the native and introduced ranges of both C. maenas and C. aestuarii (Table 2).Carcinus-specific amplification was also successful at 65ºC for rare haplotypes that showed nucleotide mismatches within the conserved priming sites (Figure 1 and Table 2); amplification was successful in all such cases attempted, though not all data is shown here.
Carcinus-specific amplification products shown in Figure 1 were subjected to EcoNI digestion.Digestion of C. maenas products resulted in generation of the predicted 212 and 136 basepair fragments, whereas all C. aestuarii products remained undigested after 3 hours at 37º C (Figure 2 and Table 2).In a more demanding test of sensitivity, larvae were added to larger scale mock communities consisting of up to 1.359 grams (filtered weight) of mixed, cultured zooplankton.Even in these experiments, we were able to detect a single larva in over 1 gram of non-target biomass; detection was successful in all four experiments with 1, 5, 10, or 20 C. maenas larvae (Figure 3C and Table 3).The amount of final amplification product appeared to decrease with the number of target organisms in these experiments.For example, in the single larva experiment, the final product was considerably weaker than for the spiked ballast experiments.

Discussion
The European green crab C. maenas-and, to a lesser extent, it's congener C. aestuarii-has demonstrated its ability to successfully establish invasive populations that pose significant potential threats to recipient ecosystems.A number of vectors have been implicated in the anthropogenic translocation of both species beyond their native ranges (Cohen et al 1995;Carlton and Cohen 2003).Probably one of the most important contemporary vectors is the transport of larvae in ballast water.Given the duration of larval stages for C. maenas, it is likely that larvae could survive even lengthy transoceanic voyages (Cohen et al 1995;Carlton and Cohen 2003).Genetic evidence for the recent establishment of a C. maenas population in Nova Scotia suggests that the opening of new shipping lanes between northern Europe and the Strait of Canso Superport may have enabled the introduction of C. maenas to this region (Roman 2006); ballast water would thus be the most likely vector for this invasion.Similarly, ballast water discharge has been cited as the most probable source of invasive C. maenas populations in Argentina (Hidalgo et al 2005).Considering the impressive fecundity of C. maenas females (Broekhuysen 1936), entrainment, translocation, and discharge of larvae in ballast water could be a significant source of propagules for seeding introduced populations throughout the globe.Moreover, the natural dispersal of C. maenas larvae by offshore currents likely has contributed to the regional spread of the species within its introduced ranges.In Pacific North America, enhanced northward currents and warmer ocean temperatures accompanying periodic El Niño events almost certainly have facilitated the rapid expansion of the C. maenas population from San Francisco Bay to Vancouver Island (Yamada and Hunt 2000).A similar mechanism may have led to the expansion of C. maenas populations from southeastern Australia to Tasmania, although coastal shipping may provide an alternative explanation for this event (Thresher et al 2003).
It is important to note that the established introduced range of C. maenas is considerably smaller than the potential global range based on the environmental requirements of the species (Carlton and Cohen 2003).Continued transportation of C. maenas in ballast water and by other vectors thus has the potential to result in additional invasions across the globe.This fact, together with the potential importance of natural and anthropogenic larval dispersal to range expansion of established introduced populations, highlights the importance of detection and monitoring of C. maenas larvae in environmental samples as a tool for assessing and managing future risks associated with this species.

Utility of DNA-based methods for sensitive detection of targets
The important role of larval dispersal in the spread of marine invasive species has already prompted the development of several DNA-based tools for the rapid and sensitive detection of larvae and other propagules in environmental samples.This task necessitates the design of assays capable of discriminating target species from non-targets in a background that is potentially both diverse in biotic composition and overwhelming in terms of non-target biomass.The sensitivity of the PCR-based assay described here compares favorably with other assays reported in the literature for monitoring invasive species in environmental samples.Our ability to detect single first stage larvae in up to 178 mg of mixed plankton derived from ballast water (Figure 3) is comparable to the detection limits reported for other similar assays.Patil et al. (2005a) recently described the development of species-specific PCR assays for the detection of the toxic dinoflagellate Gymnodinium catenatum (Graham, 1943) in both ballast water and environmental plankton samples.Using this approach, as few as 5 G. catanatum cysts could be detected in approximately 131 mg of plankton (filtered weight), the equivalent of nearly 75 liters of filtered ballast water.Similar success was achieved in developing PCR-based approaches for detecting larval forms of Pacific Oyster Crassostrea gigas (Thunberg, 1793) and the seastar Asterias amurensis (Lütken, 1871).In the case of C. gigas, specific detection of 5 Dhinge larvae or 50 earlier-stage larvae (ciliated blastulae) was possible in a background of approximately 150 mg mixed plankton (Patil et al 2005); for A. amurensis, detection limits were as low as one larva in 200 mg plankton (Deagle et al 2003).In another study, the availability of such tools allowed researchers to recognize the existence of a free-living larval form of a relatively poorly studied invasive gastropod species (Gunasekera et al 2005).
Given the facility with which we were able to detect single larvae in these samples, we pursued more demanding tests by spiking mock plankton samples containing several-fold higher levels of non-target biomass.These experiments demonstrate the ability of our assay to detect at extremely low levels, as low as a single larva in over 1 gram of mixed plankton (Figure 3).The sensitivity demonstrated in these experiments is significantly higher than published sensitivity estimates for other invasive species detection assays, likely reflecting both the specificity of PCR and the ease with which Carcinus DNA is recovered using standard, commercially available extraction methods.The composition of our mock plankton samples is unlikely to mimic any realistic environmental sample, being drawn from cultured stocks of both marine and freshwater zooplankton.However, the success of these experiments suggests that our assay is capable of specifically detecting C. maenas larvae in very large amounts of background biomass.The generation of weaker amplification products in these more demanding tests-a particularly weak band is observed when detecting a single larva in 1.184 grams of plankton (Figure 3)-indicates that we are likely approaching the detection limits of the assay.

Design of PCR-based detection assays
The development of the assay described here takes advantage of the considerable genetic information available for the target species.The ability to design highly species-specific DNAbased assays depends crucially on the amount of obtainable sequence data.Basing speciesspecific assay design on limited genetic data, though often necessary, raises the possibility of false negative results in the case of populations exhibiting unknown nucleotide variants not recognized by the assay.Previous genetic studies on Carcinus have generated abundant sequence data from the mitochondrial COI locus (Roman and Palumbi 2004), greatly facilitating assay development.Additional mtDNA sequencing (Darling, et al. unpublished data) provided us with a total of 99 Carcinus haplotypes from almost every known region within the genus' native and introduced ranges.The availability of such extensive sequence data is unusual for invasive species, and generates additional confidence in the utility of our assay for detecting Carcinus across the globe.In addition, the frequent adoption of COI as an informative locus for phylogenetic analysis and, more recently, DNA barcoding (Hebert et al 2003) increased the availability of multiple non-target DNA sequences necessary for development of the PCR assay.Sequence alignments indicate that our Carcinus-specific primers are unlikely to amplify from any of the non-Carcinus species investigated (Table 1); this is confirmed by direct testing of a number of non-target crab species (Figure 1 and Table 2).Importantly, many of the species tested exhibit ranges that overlap with that of the Carcinus species.In particular, we have tested many of the crab species likely to coexist with C. maenas along the Pacific coast of North America.
Our assay is capable of successfully discriminating between C. maenas and C. aestuarii in all tested cases (Figure 2), and sequence alignments suggest that the presence of the EcoNI site within the Carcinus-specific amplicon is truly diagnostic of C. maenas (Table 1).PCR-RFLP is an ideal approach for detecting multiple target species, and its utility has been repeatedly demonstrated.Weathersbee et al. (2003) recently adopted PCR-RFLP to distinguish between morphologically cryptic eggs of two closely related root weevils, the regulated invasive Diaprepes abbreviatus (Linnaeus, 1758) and the minor native pest Pachnaeus litus (Germar, 1824).In some cases, underlying variation has been sufficient even to target populations from specific geographic origins.Saltonstall et al. (2003), for instance, were able to develop a rapid and inexpensive means of distinguishing invasive and noninvasive haplotypes of the common reed Phragmites australis (Cav.(Trin.)ex Steud.) in North America.In another study, speciesspecific restriction sites and genus-specific PCR primers allowed identification of both European and Asian varieties of introduced gypsy moths Lymantria dispar (Linnaeus, 1758) (Pfeifer et al. 1995).Our study thus contributes to a growing literature indicating the utility of the PCR-RFLP approach for the specific detection of invasive and pest species.
Given the number of different COI haplotypes that have been found within the genus, it was not possible to find any universally conserved regions large enough to design genus-specific PCR primers.This variation thus necessitated the design of primers that possess known nucleotide mismatches to certain target haplotypes.This problem was particularly pronounced for the reverse priming site.Every effort was made in the primer design process to limit these mismatches to the 5' end of the primer, while at the same time maintaining non-target nucleotide mismatches in the 3' end (see Table 1).This approach ensured that PCR amplification of all targets was possible even under the relatively stringent reaction conditions sufficient to prevent recognition of non-target template.Direct testing indicates that our assay successfully amplifies COI from Carcinus individuals possessing these mismatched haplotypes (Figure 1 and Table 2).This success demonstrates the possibility of developing specific PCR-based assays even when high levels of nucleotide variation preclude identification of universally conserved regions for primer design, a situation most likely to arise in the case of species for which considerable sequence information is available.Still, this difficulty underlines the possibility of this or any similar PCR-based assay encountering unrecognized haplotypes and generating false negative results.Due to the number and geographic range of available C. maenas haplotypes, it is likely that our dataset provides an excellent sampling of the existing genetic diversity for that species; the only unrepresented region was the species' putative Atlantic African range.In the case of C. aestuarii, it is more likely that additional mismatches may occur, as the native population is more poorly sampled.

Conclusions
Given the clear risks posed by C. maenas and the uncertainty surrounding the invasive capacity of C. aestuarii, detection and early monitoring of both species is clearly warranted for those regions at high risk.These include areas possessing environmental conditions conducive to green crab colonization and connected to already established populations either by currentdriven dispersal (e.g.much of the Pacific coast of North America and South Australia), or by transoceanic shipping (e.g.Pacific South America and mainland east Asia) (Carlton and Cohen 2003;Cohen et al 1995).In many of these areas, such monitoring programs already exist.On the Pacific coast of North America, for example, management plans recommend biweekly or monthly sampling of uninvaded embayments (Grosholz and Ruiz 2002).Generally, however, such monitoring is limited to trapping postlarval juveniles, or "young of the year" crabs.The ability to detect the presence of larval crabs, either in the water column of uninvaded estuaries or in ballast water being released into those estuaries, should greatly improve forecasting and enable more direct assessment of the propagule pressure on at-risk ecosystems.
The general need for rapid, inexpensive, in situ monitoring tools for invasive species has prompted the development of DNA-based methods for specific and sensitive detection of target species in environmental samples.Tools such as the PCR-based assay described in this work represent only the first generation in the development of DNA-based technologies appropriate for invasive species management.They provide the foundation for exploration of more advanced approaches such as real-time PCR for quantification of target species abundance, microarray-based assays for detection of multiple targets in a single sample, or PCR-independent technologies appropriate for true "lab-on-a-chip" applications (Darling and Blum 2007).Even in their present form, however, assays such as that described here will enable early detection of potentially damaging invasions and monitoring of likely vectors and pathways of introduction, and will improve predictive models and risk assessments.In addition, by providing novel means of assessing larval transport such tools may prove valuable to researchers seeking to better understand the population dynamics of invasive species establishment and spread.larvae used in sensitivity tests.We also thank Joe Roman for providing COI haplotype sequences for C. maenas and C. aestuarii.Monaca Noble provided ballast samples as part of a contractual arrangement with the US EPA, and Mark Smith of the US EPA aquaculture facility kindly aided in collection of cultured zooplankton.Mike Blum and Greg Toth provided helpful editorial comments on previous versions of the manuscript.Although this work was reviewed by US EPA and approved for publication, it may not necessarily reflect official Agency policy.

Figure 1 .
Figure 1.Specificity of Carcinus-specific PCR.Top panel: C. maenas and C. aestuarii DNA.Bottom panel: non-Carcinus crab DNA.Numbering is as in Table 2. High molecular weight band (~700 bp, top arrow) represents universal COI control; lower molecular weight band (348 bp, bottom arrow) represents Carcinus-specific COI product.M, 100 base pair ladder.

Figure 2 .
Figure 2. RFLP identification of C. maenas and C. aestuarii.EcoNI digests of Carcinus-specific products shown in Figure 1.Products were run on a 2% agarose gel to increase resolution for smaller fragments.Topmost arrow indicates undigested 348 bp Carcinus-specific PCR amplicon; lower two arrows indicate 212 and 136 bp C. maenas-specific digestion products.M, 100 base pair ladder.Samples correspond to C. maenas samples 1 through 9 and C. aestuarii samples 23 to 31 in Figure 1 and Table2.

Figure 3 .
Figure 3. Sensitivity of Carcinus-specific PCR.Universal COI control PCR (high molecular weight band, top arrow) and Carcinus-specific PCR (low molecular weight band, bottom arrow) run for each sample.A, PowerSoil mini extractions; B, DNeasy Plant Kit mini extractions; C, PowerMax Soil Kit maxi extractions; M, 100 base pair ladder.Sample numbers correspond to Table 3. Unloaded lanes are shown between experiments for clarity of presentation.

Table 2 .
Samples used in specificity analysis.For non-Carcinus samples 5-17 and for all Carcinus samples, location indicates actual collection location.For non-Carcinus samples 1-4, location indicates only known native range.Specificity tests were either positive (+), negative (-), or not done (ND).Asterisks (*) indicate haplotypes with one or more mismatches in Carcinusspecific primer binding sites.

Table 3 .
Description of sensitivity tests.Sample IDs are as shown in figure3.Plankton biomass was measured as filtered weight.NA, not applicable.