Distribution and seasonal differences in Pacific Lamprey and Lampetra spp eDNA across 18 Puget Sound watersheds

Pacific Lamprey and Lampetra spp quantitative PCR assay development

Sensitivity, specificity and validation of lamprey qPCR assays

The limit of detection (LOD) and limit of quantification (LOQ) for the ET_Cytb890-1015 and LACytb_890-1015 assays were determined by performing qPCR on a dilution series of each target oligonucleotide ultramer (1, 2, 3, 6, 30, 60, 300, and 600 copies/qPCR), with eight qPCR replicates per concentration. We regarded LOD as the minimum number of copies that could be detected and we regarded LOQ as the lowest number of copies that yielded positive amplification across the eight replicates (Doi et al., 2017; Takahara, Minamoto & Doi, 2013; Treguier et al., 2014). With these criteria, the LOD was determined as 1 copy/qPCR and the LOQ was determined as six copies/qPCR for each assay (Fig. S1).

To determine the specificity of the ETCytb_890-1015 and LACytb_890-1015 qPCR assays, we performed three tests. First, we tested for cross amplification and reduction of qPCR efficiency in samples containing both Pacific Lamprey and Lampetra spp genomic DNA. This goal was accomplished by testing serial dilutions (0.006 pg/μL–600 pg/μL) of Pacific Lamprey DNA in the presence of 1 pg/μL of Lampetra spp DNA and vice-versa.

Second, each assay was tested on genomic DNA from 90 Pacific Lamprey and 54 Lampetra spp (including L. ayresii, L. richardsoni, and L. hubbsi) distributed from Alaska to California (Table S2), including 13 Pacific Lamprey and 17 Lampetra spp from Puget Sound that were originally sequenced, and on genomic DNA from 39 non-target fish species commonly found in the Pacific Northwest (one individual/species) (Table S2). We used 1 ng DNA per qPCR for this test. The threshold cycle (C_t) value for each lamprey sampled outside the Puget Sound was compared to the range of C_t values for Pacific Lamprey and Lampetra spp sampled from within the Puget Sound watershed because higher C_t values from samples that have no sequence information relative to C_t values from samples with sequence information could indicate basepair mismatches within the primers and/or probe.

Third, ETCytb_890-1015 and LACytb_890-1015 primer and probe sequences were compared to published Entosphenus spp and Lampetra spp cytb sequence data sampled from Pacific coast drainages of North America (Boguski et al., 2012; Lang et al., 2009; Reid et al., 2011). We retrieved 156 cytb DNA sequences from GenBank and aligned them to the Puget Sound Pacific Lamprey and Lampetra spp sequences by using the program MEGA ver. 6 (Tamura et al., 2013). A summary of samples included in the sequence analysis are as follows: Pacific Lamprey, N = 24 (13 from our study, 10 from Boguski et al. (2012), and one from Lang et al. (2009)); E. lethophagus, E. macrostomus, E. minimus, and E. similus, N = one each from Lang et al. (2009); Lampetra spp (including L. ayresii, L. richardsoni, L. pacifica, and L. hubbsi), N = 164 (23 from our study, 135 from Boguski et al. (2012), three from Lang et al. (2009), and three from Reid et al. (2011)). In addition, primer and probe sequences were subjected to BLAST and Primer-BLAST analysis against non-target Actinopterygii (ray-finned fishes) as an additional test for specificity. The BLAST and Primer-BLAST analyses against non-target Actinopterygii species did not return any highly homologous matches.

To validate the functionality of our qPCR assays for detecting lamprey DNA in water samples collected in the field, we collected water from flow-through tanks that contained Pacific Lamprey (Umatilla Hatchery, Irrigon, OR, USA) and from a site downstream of a juvenile salmonid fish trap on Bear Creek, Washington, when Lampetra spp were known to be present in the trap. Water samples were filtered and tested by the qPCR assays following the methods described herein and results showed that only Pacific Lamprey and Lampetra DNA were detected in each respective water sample.

Detection of lamprey eDNA in Puget Sound watersheds

Sensitivity, specificity, and validation of Pacific Lamprey and Lampetra spp qPCR assays

The PCR efficiency of the Pacific Lamprey and Lampetra spp qPCR assay was each greater than 90% and r² was greater than 0.99 (Table 2). When qPCR was performed on samples that contained genomic DNA from both Pacific Lamprey and Lampetra spp, each assay amplified only the target lamprey, demonstrating specificity of the assays when samples contain both species. In addition, the presence of non-target lamprey DNA had no effect on the slope or efficiency of both qPCR assays (Table 2).

Quantitative PCR on genomic DNA from lamprey distributed across Pacific coast drainages of North America indicated the assays were highly specific to the target species (Table 3, Table S2). The range in C_t values for Pacific Lamprey sampled within the Puget Sound watershed was 21.64–27.22. No Pacific Lamprey sampled from outside Puget Sound exceeded the maximum C_t value observed within Puget Sound. For Lampetra spp, the range in C_t values from samples collected within the Puget Sound watershed was 19.43–26.97. One Lampetra spp sampled from outside the Puget Sound (Toppenish Creek, WA, C_t = 27.18) slightly exceeded the maximum C_t value observed within Puget Sound. None of the 37 non-target fish species commonly found in the Pacific Northwest amplified with either assay (Table S2).

All Pacific Lamprey cytb sequences (11 obtained from GenBank and 13 from our study) had 100% match with the target primers and the Pacific Lamprey specific probe (Fig. S2). The DNA sequence from E. macrostomus, E. minimus, and E. similis also had a complete match to the Pacific Lamprey primer and probe sequences, and E. lethophagus had one mismatch in the forward primer (Table 3). All Entosphenus spp had 5 mismatches with the Lampetra spp probe.

Across the set of Lampetra spp cytb sequence data (141 obtained from Genbank and 23 from our study), 85.4% completely matched the target primers and Lampetra specific probe (Table 3, Fig. S2). Specifically, 97.6% of Lampetra spp sequences had a complete match with the forward primer, 88.4% had complete match with the reverse primer, and 94.5% had complete match with the Lampetra spp probe. The Lampetra spp that mismatched with the primers and/or probe were from six locations: Hunter Creek, CA (five individuals sequenced and each had a single mismatch with the reverse primer); McGarvey Creek, CA (five individuals sequenced and each had a single mismatch with the reverse primer); Big River, WA (four individuals sequenced and one had a single mismatch with the reverse primer); North Fork Suislaw River, OR (five individuals sequenced and each had a single mismatch with the probe); Kelsey Creek, CA (four individuals sequenced and each had two mismatches with the forward primer and three mismatches with the reverse primer); and Mark West Creek, CA (four individuals sequenced and each had two mismatches with the reverse primer and two mismatches with the probe) (Table 3). The Lampetra from Hunter Creek, McGarvey Creek, and Big River are considered as L. richardsoni and group into the polytomy identified in Boguski et al. (2012), while the Lampetra from North Fork Suislaw River, Kelsey Creek and Mark West Creek represent genetically divergent populations. All Lampetra spp had five mismatches with the Pacific Lamprey probe, with the exception of the four individuals from Mark West Creek, CA, which had three mismatches.

Detection of lamprey eDNA in Puget Sound watersheds

Lampetra spp and Pacific Lamprey eDNA was not detected in filtration controls, negative extraction controls and no template controls, demonstrating that there was no contamination during field or laboratory procedures.

Pacific Lamprey eDNA was detected in 14 of the 18 Puget Sound watersheds that were surveyed (Table 4, Fig. 2A). Five watersheds sampled in the fall and 14 watersheds sampled in the spring tested positive for Pacific Lamprey eDNA. All watersheds that tested positive for Pacific Lamprey in the fall also tested positive in spring. The Pacific Lamprey eDNA detection rate was 0.26 and 0.90 in the fall and spring, respectively, across the 14 sites where Pacific Lamprey eDNA was detected. Of the five locations where Pacific Lamprey eDNA was detected in both fall and spring, the mean eDNA concentration was higher in the fall for only the Nisqually River collection and was not different between fall and spring for the Tahuya River collection (Fig. 2A). The distribution of Pacific Lamprey in Puget Sound streams inferred through eDNA sampling was concordant with trap surveys (Hayes et al., 2013) for 16 of 18 watersheds (Table 4). We detected Pacific Lamprey eDNA in two streams (Skagit River and Snoqualmie River) where they were not observed by Hayes et al. (2013).

Lampetra spp eDNA was detected in 10 watersheds in the fall and 14 watersheds in the spring (Table 4, Fig. 2B). The 10 watersheds where Lampetra spp were detected in the fall also tested positive in the spring. The Lampetra spp eDNA detection rate was 0.62 and 0.93 in the fall and spring, respectively, across the 14 sites where Lampetra spp eDNA was detected. Within watershed, Lampetra spp mean eDNA concentrations were higher on the spring collection day than on the fall collection day, with the exception of the Nooksack River (Fig. 2B). The distribution of Lampetra spp inferred through eDNA sampling was concordant with trap surveys (Hayes et al., 2013) for 17 of 18 watersheds (Table 4). We detected Lampetra spp eDNA in Salmon Creek, but individuals were not observed in juvenile salmon trapping operations (Hayes et al., 2013).

On days when water samples were collected, median stream flow was higher in the fall compared to spring for 11 of the 12 watersheds with USGS stream flow gages (Fig. 2, Table S3). The Puyallup River was the only watershed where the median stream flow was higher on the spring collection day compared to the fall collection day.

Lamprey eDNA in Puget Sound watersheds

The distribution of lamprey in Puget Sound watersheds identified through eDNA sampling was largely consistent with the distribution observed during incidental trap catch surveys in 2011 (Hayes et al., 2013). For example, we detected eDNA from target species in all watersheds where target species were found by Hayes et al. (2013). This finding is not unexpected, as results from eDNA presence/absence surveys are often consistent with results from traditional field survey methods, such as electrofishing and trapping (Baldigo et al., 2017; McKelvey et al., 2016; Rees et al., 2014a). However, the traps used by Hayes et al. (2013) may under-report Lampetra spp that do not undergo active downstream migration (i.e., Western Brook Lamprey) relative to anadromous lamprey (i.e., Pacific Lamprey and Western River Lamprey). Using eDNA, we detected Pacific Lamprey in the Skagit and Snoqualmie rivers and Lampetra spp in Salmon Creek where individuals were not observed by Hayes et al. (2013). While Pacific Lamprey are known to inhabit the Skagit River watershed (Goodman et al., 2008) and the Snohomish River watershed (J Whitney, Washington Department of Fish and Wildlife, pers. comm., 2017), which includes the Snoqualmie River, Lampetra spp in Salmon Creek were heretofore unreported. Thus our findings provide a more complete picture for lamprey distributions in Puget Sound watersheds.

Detection of lamprey eDNA differed between seasons, highlighting the importance of considering sample timing when developing an eDNA sampling strategy. For example, across the fall sample collections, we failed to detect Pacific Lamprey at nine locations and Lampetra spp at four locations where each was detected in the spring sample collections. This finding is significant because larval lamprey have an extended freshwater rearing phase and should be present in these streams year-round, suggesting that eDNA should have been consistently detected. Furthermore, seasonality appeared to have a greater effect on Pacific Lamprey eDNA detections compared to Lampetra spp. For example, the eDNA detection rate for Pacific Lamprey was 3.5 times higher in spring while the eDNA detection rate for Lampetra spp was 1.5 times higher in spring. Studies that included a temporal sampling scheme have demonstrated that eDNA abundance varies over time (Buxton et al., 2017; De Souza et al., 2016; Furlan et al., 2016; Laramie, Pilliod & Goldberg, 2015; Spear et al., 2015), which could have important conservation implications that inform seasonal activities, such as life history events, migratory behaviors, and dispersal of invasive species. Although sampling in the fall did not provide additional information on lamprey distribution in Puget Sound, it revealed that certain time periods yield better detections than others, which can ultimately influence cost effective sampling strategies for target organisms.

The difference in eDNA detection between the fall and spring samples within watersheds was likely due, in part, to different stream flow rates between seasons. Within watersheds, stream flow rates were typically higher when we sampled in the fall and coincided with lower observed eDNA concentrations. During periods of higher flows, cellular material shed from organisms may be diluted (Gingera et al., 2016; Jane et al., 2015), resulting in lower eDNA concentrations per sample volume or false-negative detections if eDNA concentrations fall below the LOD threshold required for a positive detection. The relationship between eDNA concentration and stream flow rate is complex because the fate of eDNA in aquatic environments is also influenced by a suite of environmental (Jane et al., 2015; Lacoursiere-Roussel, Rosabal & Bernatchez, 2016; Moyer et al., 2014; Pilliod et al., 2014; Strickler, Fremier & Goldberg, 2015), demographic (Diaz-Ferguson et al., 2014; Takahara et al., 2012; Wilcox et al., 2016), and biological (De Souza et al., 2016; Maruyama et al., 2014) factors. Further studies will be required in order to quantify relationships between lamprey eDNA abundance and flow rates.

Most of our results support the conclusion that eDNA abundance of target organisms was lower during periods of higher flow; however, a few of our results stand in contrast. For example, Pacific Lamprey eDNA in the Nisqually River and Lampetra spp eDNA in the Nooksack River was more concentrated on the fall sample collection date, coinciding with a higher stream flow rate. Furthermore, Pacific Lamprey and Lampetra spp eDNA in the Puyallup River was more concentrated on the spring sample collection date (assuming larval Pacific Lamprey were present year-round), despite a higher stream flow in the Puyallup River on the spring collection date. If stream flow was the only factor that affected eDNA abundance within watershed, then we would expect eDNA to be more concentrated on the collection date that had a lower stream flow rate. Our findings suggest that factors other than stream flow may have affected eDNA abundance.

Seasonal life history events may also have influenced lamprey eDNA abundance. Here, we address two such seasonal life history events, spawning and downstream movement of larvae and juveniles. Spawning activity and decomposition of dead adults in semelparous species would be expected to increase shedding of cells and tissues into aquatic environments. Further, increased eDNA abundance during reproductive seasons has been observed for other aquatic species, including Sea Lamprey (Petromyzon marinus) (Gingera et al., 2016), Eastern Hellbender (Cryptbranchus alleganiensis) (Spear et al., 2015), and Great Crested Newt (Triturus cristatus) (Buxton et al., 2017). We had limited information on spawn timing within our study sites, however, we observed dead Lampetra spp adults in the Bear Creek and Cedar River traps when water samples were collected in spring. In addition, higher abundance of lamprey eDNA during the spring sample collection appeared to coincide with spawn timing reported from other locations (Beamish, 1980; Stone, 2006). For example, in several streams (e.g., Dewatto River), no water samples tested positive for lamprey eDNA in the fall, whereas all three water samples tested positive in the spring. Thus, increased eDNA abundance in spring samples may be attributed to adults through the release of gametes, shedding of cells during nest building behavior, and decomposition of adults. Larval and juvenile lamprey both exhibit downstream movement. Larvae disperse downstream in late winter through spring (Potter, 1980) and juveniles outmigrate from fall through spring (Beamish & Levings, 1991; Richards & Beamish, 1981). Interestingly, the majority of lamprey captured in fish traps in the study by Hayes et al. (2013) (which was conducted mostly in April through June) were larvae, suggesting that downstream movement of larvae occurs in the Puget Sound watershed that we sampled. Downstream movement requires lamprey to emerge from sediments and enter the water column and if such movement was common in spring, but not fall, it could contribute to seasonal differences in eDNA abundance. Temporal sampling of water in fall, winter and spring could provide information on the timing of downstream movements and outmigration.

Inferring species presence

Presently, there are no criteria regarding the minimum number of positive PCRs that are required for inferring species presence. eDNA studies have typically required a minimum of one positive (Laramie, Pilliod & Goldberg, 2015; McKelvey et al., 2016) or more than one positive (Rees et al., 2014a) out of several PCRs performed on a sample site. Furthermore, repeat PCRs are often performed on field replicates that yielded ambiguous results (Goldberg et al., 2013; Laramie, Pilliod & Goldberg, 2015; Rees et al., 2014a), with the aim of improving confidence in results. Goldberg et al. (2016) recommend that caution be used when inferring species presence at sample sites where only one of several PCR replicates tests positive for the target species and where the results were not replicated through repeat analysis of samples or repeated site visits. We followed this recommendation for inferring species presence and required that at least two qPCRs exceed the LOD, or when only one qPCR exceeded the LOD we required replication of this result in a repeat analysis of samples. Our criteria affected the inference of Pacific Lamprey at one location, Bear Creek, where only one of 12 qPCRs from the spring collection day and no qPCRs from the fall collection day exceeded the LOD. Repeat analysis on Bear Creek spring samples yielded no positive qPCRs. Under our criteria, the results for Bear Creek imply non-detection of Pacific Lamprey. However, application of less stringent criteria for inferring species presence, such as requiring that only a single qPCR exceed the LOD, would result in a different conclusion. Nevertheless, our goal was to apply a conservative approach for inferring species presence and our findings may warrant further evaluation for the presence of Pacific Lamprey in Bear Creek.

Application of LNAs to eDNA studies

Here, we included locked nucleic acids (LNAs) in the Taqman assay probe designs. Few studies to date have applied LNA-based PCR assays to environmental samples and these studies have generally been limited to detecting pathogens isolated from host tissues (Barkallah et al., 2016; Ruthig & DeRidder, 2012; Yamada et al., 2008), raw sewage (Alonso, Amoros & Canigral, 2011), and soil (Irenge et al., 2010). Locked nucleic acids offer an alternative chemistry for robust PCR assay design that is compatible with DNA chemistry. Therefore, LNA bases can be incorporated into primer and probe sequences allowing LNA-DNA mixmer oligonucleotides to be designed (Mouritzen et al., 2003). Several key features render LNAs well suited for eDNA assay design, including high sensitivity (Alonso, Amoros & Canigral, 2011; Josefsen et al., 2009; Letertre et al., 2003), high specificity (Letertre et al., 2003; Owczarzy et al., 2011; You et al., 2006), and good performance with low DNA concentrations (Ballantyne, Van Oorschot & Mitchell, 2008; Reynisson et al., 2006). Furthermore, LNA probes perform well in Taqman assays, as applied in our study, and produce results that are comparable to MGB probes (Letertre et al., 2003; Ruthig & DeRidder, 2012). Conserved forward primers and nearly conserved reverse primers between Pacific Lamprey and Lampetra spp LNA assays stands in contrast to the convention for MGB assays where assay specificity was found to be largely influenced by mismatches in the primers (Wilcox et al., 2013). Placement of LNAs into optimal and strategic positions increases the T_m (temperature at which half of double stranded DNA has dissociated into single strands). However, thermal stability of LNA-DNA duplexes and mismatch discrimination depends on the position of LNA bases in a sequence, the number of LNAs in sequence, and specific bases that are modified (Owczarzy et al., 2011; You et al., 2006). Thus, LNA technology has application toward eDNA assay design and LNAs may be particularly effective for differentiating closely related taxa where a limited set of base pair mismatches may constrain assay design.

Assay limitations

While the ETCytb_890-1015 and LACytb_890-1015 assays differentiated Pacific Lamprey and Lampetra spp, there are limitations to each assay that need to be considered (Table 3). Sequence comparison of ETCytb_890-1015 target sites among Entosphenus spp showed that primer and probe sites are conserved among Vancouver Lamprey (E. macrostomus, distributed in the Lake Cowichan drainage, Vancouver Island, BC), Miller Lake Lamprey (E. minimus, distributed in the Upper Klamath River drainage, OR), Klamath Lamprey (E. similis, distributed in the Klamath River drainage, OR and CA), and Pit-Klamath Brook Lamprey (E. lethophagus, distributed in the Klamath River drainage, OR and CA, and Pit River drainage, CA) (Entosphenus spp distributions from Potter et al., (2015)), with the exception that the forward primer has one mismatch with Pit-Klamath Brook Lamprey. Similar findings of conserved primer and probe sites among Entosphenus spp were reported for an eDNA assay that was developed for Pacific Lamprey in the Columbia River based on cytochrome c oxidase I (COI), with the exception that the Pit-Klamath Brook Lamprey sequence was conserved with the COI primer and probe sites and Vancouver Lamprey and Miller Lake Lamprey sequences were not available for comparison (Carim et al., 2017). It is unknown how the single mismatch in the ETCytb_890-1015 forward primer would affect amplification of Pit-Klamath Brook Lamprey DNA.

The LACytb_890-1015 assay detects L. ayresii, L. richardsoni, L. pacifica, and L. hubbsi, although the assay may have limited or no detectability of Lampetra populations in certain geographic locations (Table 3). These Lampetra populations were represented by L. richardsoni in Hunter Creek, CA, McGarvey Creek, CA, and Big River, WA, and genetically divergent Lampetra spp in North Fork Siuslaw River, OR, Kelsey Creek, CA, and Mark West Creek, CA. Boguski et al. (2012) suggested the Lampetra spp in Siuslaw River, Kelsey Creek, and Mark West Creek could represent cryptic and previously undescribed species. Because the understanding of genetic diversity in Lampetra spp is far from complete, it is possible that some unsampled Lampetra populations, particularly populations south of the Columbia River Basin, may be found containing mismatches to the LACytb_890-1015 assay. Further testing of genomic DNA on mismatched sequences would provide information on the applicability of the Lampetra spp assay for detecting eDNA at locations where individuals lack complete primer and probe match.

The ETCytb_890-1015 and LACytb_890-1015 assays are each genus-specific, but these assays may be used to infer species presence when distributional information is taken into account. For example, the ETCytb_890-1015 assay may be applied to several Entosphenus spp and the resolution of detection to species or genus would depend on whether one or more than one Entosphenus spp, respectively, is present in a particular watershed. The overlap in geographic distribution between Pacific Lamprey and other Entosphenus spp is limited; Pacific Lamprey is broadly distributed whereas E. macrostomus, E. minimus, E. similis, E. lethophagus have restricted geographic distributions. Consequently, Pacific Lamprey occur in the absence of other Entosphenus spp throughout most of the Pacific Lamprey range. Thus, eDNA may be resolved specifically to Pacific Lamprey across a large portion of their range (geographic regions where Pacific Lampey is the only known Entosphenus spp present) and resolved more generally to Entosphenus spp across a small portion of the Pacific Lamprey range (geographic regions where Pacific Lampey and other Entosphenus spp overlap in distribution). Likewise, distributional information on Lampetra spp could also be incorporated when resolving detections between genus and species. In geographic regions where Lampetra spp overlap in distribution, resource managers may simply want information on occupancy status of Lampetra spp, in general.