Working Primers and qPCR Protocols for Rapid eDNA Identification of Four Aquatic Invasive Species Found in the Lower Great Lakes with High Potential for Ballast Transport to Lake Superior

Abstract

Reliable, timely and economical target organism detection in harbors and ballast water is urgently needed to prevent the spread of aquatic invasive species (AIS) by commercial ships in the North American Great Lakes (NAGL). Inter-Great Lake ships (Lakers) transport large volumes (ca. 52 million metric tons. annually) of untreated lake water between lakes, with over 50% transported against the natural flow from the lower lakes to Western Lake Superior ports. The transport of ballast water is the number one threat of AIS spread throughout the NAGL. A relatively new tool to fight the spread of AIS is the use of eDNA for rapid detection and identification of target organisms. This technology opens doors for advancing control of ballast-mediated AIS through rapid detection. To that end, we have developed species-specific, reliable eDNA primers to target specific detection of four AIS in water samples along with qPCR protocols. Target organisms were selected based on the following criteria: (1) they are known to be invasive in the lower NAGL, (2) they are established in the lower NAGL but not in Superior, (3) they are biodegradable, and (4) they are obtainable, morphologically distinct and have existing DNA sequence information. Working primers, qPCR protocols and detection limits are provided for three invertebrate species and one alga species. These species are Daphnia lumholtzi (a water flea), Cercopagis pengoi (the fishhook water flea), Echinogammarus ishnus (a scud) and Nitellopsis obtusa (Starry Stonewort).

1. Introduction

Aquatic Invasive Species (AIS) are commonly spread to and between bodies of water such as the NAGL through the uptake and release of ballast water [1,2,3] that is utilized by commercial ships to stabilize the vessels and balance the weight differences with and without cargo. In the NAGL, the shipping routes of vessels that transit exclusively between lakes (Lakers) dictate that these ships commonly intake water from the lower NAGL, Michigan, Erie, Huron, and Ontario, and travel to the Lake Superior where they release ballast water to take on new cargo [1,2,4,5]. Additionally, the natural flow of water in the NAGL begins in Lake Superior, flowing downstream through the other lakes and to the Atlantic via the Saint Lawrence River [6]. When ballast water from the lower NAGL is released in Lake Superior it can release a multitude of AIS in Lake Superior against the natural flow of water between the NAGL, therefore increasing the rate that these AIS spread [7].

Aquatic invasive species are known to become invasive at a higher rate than terrestrial introduced organisms and current rates of invasion far exceed those of other geological time periods [8]. AIS can potentially cause severe damage to the ecosystems they have invaded through utilizing resources that would otherwise be used by the native organisms and at times outcompeting the native species that holds a similar niche [9] or becoming “environmental engineers”, causing a rapid overhaul to the established natural ecosystem [8]. Often these AIS thrive in the absence of natural predators and parasites, allowing them to rapidly increase in population size. Exponential population growth removes essential nutrients such as phosphorous from the sediment and depletes these resources, decreasing the number of native species that have access to these same resources. In addition, the negative economic impacts of AIS are predicted in the agriculture, marine shipping, and fisheries industries to range from millions to billions of dollars each year as AIS populations continue to increase [9]. In the NAGL, Rosaen et al. [10] have reported that the six industries most impacted by AIS are (1) commercial fishing, (2) power generation, (3) industrial facilities, (4) shipping related business, (5) tourism and recreation, and (6) public water supply and intake. Their report estimated that well over USD 100,000,000 annually is lost among individuals, businesses and the government to fight the impacts of AIS in the NAGL [10]. A recent analysis by Cuthbert et al. [11] determined that AIS have caused over USD 345 billion worth of damages world-wide, but they estimate that the figure is drastically underestimated due to the bias towards North America and the lack of AIS data compared to terrestrial invasive species. The group of invasive organisms causing the highest financial costs are invertebrates at 62% [11].

The immense size of the NAGL, the small size of the organisms, particularly small invertebrate organisms, and the difficulty of morphological identification makes AIS discovery and tracking by traditional sampling very difficult [12]. Identification by molecular methods can be an effective alternative for traditional sampling, and environmental DNA (eDNA) provides a viable method of detection, even in large bodies of water such as the NAGL [13]. Using eDNA, AIS can be detected with greater sensitivity than active collection techniques [14] and can be identified more accurately than relying on morphological identification [15]. In a meta-analysis incorporating 535 papers that compared molecular survey methods using eDNA to traditional survey methods, eDNA methods were cheaper, more sensitive and detected more species [16]. The meta-analysis also concluded that qPCR was more effective than traditional PCR and the increased sensitivity of qPCR allowed for dilution of eDNA to decrease inhibition while maintaining the ability to identify species presence effectively [16]. In ballast water, qPCR identification tools enable rapid identification of organisms that are present in any life stage without an in-depth, physical search through the ballast water contents [2].

Four aquatic invasive species of unique interest include the Charales Nitellopsis obtusa (Desv.) J. Groves, and three Crustaceans, Cercopagis pengoi (Ostroumov), Echinogammarus ishnus (Stebbing), and Daphnia lumholtzi (Sars). Nitellopsis obtusa is a macro-alga, typically between 20 and 100 cm in size, characterized by a star shaped bulbil and orange/red antheridia [17,18]. For every whorl, the number of branchlets varies but there are usually 5–6 branchlets [18]. These dioecious organisms can reproduce under fragmentation, allowing for rapid spread in the NAGL region. Cercopagis pengoi is a water flea with both sexual and asexual reproducing individuals, characterized by its fishhook shaped tail with barbs [17]. The species is well established in the lower NAGL region and is commonly caught on commercial and recreational fishing lines [17]. Echinogammarus ishnus, also known as Chaetogammarus ishnus, is a scud found in both the lower NAGL and the Finger Lakes [19]. They range in size from 5 to 15 mm and are often found residing in algae or macrophytes for shelter and food sources [17,20]. Daphnia lumholtzi, also a water flea, is about 3 mm and identified by a sharp helmet in comparison to native water fleas. It has also been noted to outcompete native water flea species regarding food supply availability [17].

Here we report the development of species-specific primers and eDNA qPCR methods that can be used to detect four the AIS listed above that are currently at risk of being transported, via ballast water, from the lower NAGL to Lake Superior ports.

2. Materials and Methods

2.1. Organisms, Loci and Sample Collection

The four study organisms above were chosen with specific criteria in mind. First, each species was known to be invasive in the NAGL. Second, each species was established in the lower NAGL and not known to be established in Lake Superior [17] though they may have been detected on a limited basis. Third, each species was readily biodegradable thus allowing them to release DNA into the water while living and quickly biodegrade after dying. Fourth each species was obtainable, morphologically distinct, and had existing DNA sequence information genes which allowed for species level detection. More specifically, the target species needed to have existing molecular data with known sequences for the rapidly evolving CO1 and matK genes available on Genbank [21] or BOLD [22]. The organisms which fit the above criteria that were targeted for this study were the macro-alga N. obtusa using the gene matK, and the arthropods, C. pengoi [23,24,25], E. ishnus [23], and D. lumholtzi [26] using the mtDNA gene CO1. For N. obtusa, the maturase K, or matK gene was chosen for primer development. While rbcL has been a common gene for identification and classification of Characeae [27], there has been more overlapping sequence data for multiple species of Characeae using matK. Additionally, the matK gene has been shown to have more variation in its sequence across species, making it more suited for specific Characeae identification, specifically for N. obtusa [27]. For C. pengoi, E. ishnus, and D. lumholtzi, the mitochondrial gene cytochrome oxidase 1 or CO1, was chosen as this is universally known as the DNA barcoding gene for invertebrates, making it ideal for the development of species-specific primers [22].

Nitellopsis obtusa was collected at Presque Isle State Park on West Fisher Drive at the eastern boat launch (GPS locality in

Table 1. AIS target species, locations in the NAGL Region reported by GLANSIS, locations our laboratory confirmed presence, working probes, annealing temperature, limit of detection, and Cq values for a range of concentrations of DNA.

Species	Reported NAGL	Source of Specimens Used	Primers	Annealing Temp (°C)	Melt Temp (°C) for qPCR	Limit of Detection	Cq Value for 1 ng/μL	Cq Value for 0.01 ng/μL	Cq Value for 0.001 ng/μL
N. obtusa	Ontario, Erie, Huron, Michigan	Lake Erie 42.153629, −80.114770	Forward: Nobt_458F CTCCTTTAATTCACCAGTTC Reverse: Nobt_695R TGAATTCACCAAATACACTA	50.0	72.0	5.60 × 10−6 ng/μL	17.71	24.66	28.14
C. pengoi	Ontario, Erie, Huron, Michigan	Lake Ontario planktonic Provided by Dr. James Watkins	Forward: Cpen_115F CAATGTAGTAGTAACAGCCCAC Reverse: Cpen_302R ACCTCCAACTAGAAGTAGAGTTAAA	53.0	79.0	1.82 × 10−3 ng/μL	26.73	33.27	36.55
D. lumholtzi	Erie, Superior	Provided by Dr. Lawrence Weider’s Lab Culture	Forward: Dlum_163F GGGTTTTGGAAATTGATTAGTT Reverse: Dlum_397R TCCCAGCCAAATGCAAAGA	51.5	78.5 and 83.0	7.60 × 10−5 ng/μL	24.29	31.83	35.60
E. ishnus	Ontario, Erie, Huron, Michigan, Superior	Lake Erie 42.156317, −80.071064	Forward: Cisc_357F GCCTCTCTCTAACTCTATAGGC Reverse: Cisc_547R TGGTAAGGACAGGAGAAGCAA	65.0	79.0	6.50 × 10−4 ng/μL	25.40	31.53	34.62

). Samples were collected directly from the shallow waters on either side of the boat launch. Samples were identified using key morphological traits and confirmed through DNA sequence data from the rbcl gene using the primers rbcl-bF-and rbcl-724R [28].

Echinogammarus ishnus was primarily collected at the North Pier of Presque Isle State Park (GPS locality in Table 1). To collect samples along the pier, a 500-micron mesh plankton net (Sea Gear Marine Supply, Cape May, NJ, USA) was used by dropping the net to substrate level and raising it along the sides of the pier. Species were morphologically identified and confirmed through DNA sequence data from the barcoding region of CO1, amplified using universal primers LCO 1490 F and HCO_2198 [29].

Daphnia lumholtzi individuals were cultured at Penn State Behrend using samples provided by Dr. Lawrence J. Weider (Department of Biology, University of Oklahoma) who has published multiple works on D. lumholtzi [30,31]. C. pengoi samples were provided by Dr. James Watkins and his research team from Cornell Biological Field Station who collected them in Lake Ontario. Dr. Watkins is an expert in zooplankton of the NAGL and is able to correctly ID C. pengoi [32].

2.2. Primer Development

qPCR primer development was completed using a combination of literature searches, published phylogenetic analysis and BLAST [33] (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 15 January 2021) to locate closely related organisms with published sequence data. Sequences were imported into Geneious 6.1 (Dogmatic) to align the sequences using Clustal Omega 1.2.4 [34] and find locations in the CO1 or matK gene with enough genetic variation to make unique primers. Candidate primer sequences consisted of forward and reverse primers that were identified by finding differences in the sequence between the target and closely related species that averaged four base pair differences within an 18–25 base pair section of the gene. The regions chosen had a minimum of 30% GC content, with ideal primer locations having 2–3 guanine or cytosine bases at the 3′ end of the sequence. Once many possible primers were created, the base pair differences and Tm values of each were plotted to determine optimal combinations between forward and reverse primers that had a Tm difference less than 5 °C and would amplify a section of the gene that was between 150 bp and 350 bp. The closely related species used as comparison organisms for D. lumholtzi was Daphnia magna based on its close phylogenetic relationship [35]. The species Lychnothamnus barbatus was used as a comparison species for N. obtusa based on BLAST searches, phylogenetic analysis [36,37], and close common co-occurrence between the two species [38,39]. For C. pengoi the comparison organism chosen was Bythotrephes longimanus, as these two cladoceran species often co-occur in the NAGL [40,41], share morphological similarity and are the most closely related species to each other in the region [42,43].

We chose Gammarus fasciatus as the organism of comparison for E. ishnus. Clarifying the species and genus designation for E. ishnus in order to obtain DNA sequence data proved difficult due to discrepancies in the literature pertaining to invasive gammarids. Our lab team obtained CO1 sequence data which was identical to sequence labeled as VOUCHER DNA samples for both E. ishnus and C. ischnus showing that there is clearly some discrepancy as to the correct species name for this organism [19,44,45,46]. We have chosen to use E. ishnus because it is the prevailing descriptor for the species found in the NAGL [23,45]. Once we determined how to address the species, G. faciatus was the logical species for comparison because it is native to the NAGL and is closely related enough for our criteria [20].

2.3. Water Filtration Protocol

For each water sample, a 0.45 um MCE Nitrocellulose membrane (MF-Millipore, Burlington, MA, USA) was placed into a sterilized filter housing (Cytiva Life Sciences, Marlborough, MA, USA) using forceps. Filter housings were attached to a length of silicon tubing with a female luer lock which passed through a Masterflex peristaltic pump per manufactures directions (Avantor, Randor, PA, USA). Different treatment levels of inoculated water were pumped through the filters at a rate of 100 mL/minute until the water exiting the housing was at a slow drip. The total amount of filtered water was recorded before filters were removed from the housing and frozen until DNA extraction. Filter housings and tubing were then sterilized with 10% bleach and thoroughly rinsed and dried prior to the next filtration.

2.4. DNA Extraction Protocol

Isolation of Nitellopsis obtusa genomic DNA (gDNA) was performed using the Plant Pro Kit (Qiagen, Hidden, Germany) due to the presence of the cell wall. Protocols followed manufactures instructions with the following exception: in step 2 of the protocol we used 100% EtOH to soak the sample for 10 min, then allowing the sample to dry in the fume hood for one hour before completing step three of the kit protocol.

DNA extraction for gDNA of C. pengoi, E. ishnus, and D. lumholtzi, and eDNA of all species employed a salting out DNA extraction protocol. Briefly, 600 μL of Longmire’s buffer was pipetted into the 1.7 mL microcentrifuge tubes containing the filters or organism tissue, followed by pipetting 30 μL of 20 mg/mL Proteinase K into each tube containing the filters. The samples were incubated at 50 °C in a heat block for approximately 24 h. Samples were then extracted using Chloroform: Isoamyl Alcohol, 24:1 and nucleic acids were precipitated using Isopropanol (500 μL) and NaCl at −20 °C for at least 1 h. The precipitates were collected by centrifugation, dried and resuspended in Low EDTA TE Buffer (10 mM Tris, pH 8.5, 1 mM EDTA), and stored at −20 °C. After DNA exaction, the DNA concentration of each sample was measured using a NanoDrop Microvolume Spectrophotometer (Thermo Fisher Scientific, Pittsburgh, PA, USA) to determine if the extraction was successful. NanoDrop data was also used to calculate limits of detection for all species.

2.5. qPCR Detection and Primer Testing Protocol (eDNA and gDNA)

All primer development assays were first completed using gDNA. To determine effective annealing temperatures and primer combinations, gradient qPCR was used on the CFX96 qPCR machine (BioRad, Hercules, CA, USA).

The qPCR mastermix for each sample (made for triplicate repetitions of one sample) consisted of 40 µL AmpliTaq Gold 360 (Thermo Fisher Scientific, Waltham, MA, USA), 8 µL sterile H₂O, 4 µL each primer, and 4 µL EvaGreen fluorescent dye (Biotium, Fremont, CA, USA). 60 µL of mastermix received 20 µL of the eDNA or gDNA and was mixed, then divided into three triplicate sample tubes, and were pipetted into the designated qPCR tray location.

The optimal qPCR protocol for N. obtusa was as follows: 94 °C for 10 min (10 min hotstart to activate AmpliTaq Gold), followed by 41 cycles of denaturing at 95 °C for 30 s, annealing at 50 °C for 1 min and 30 s, and extension at 72 °C for 1 min, then followed by an extension for 5 min at 72 °C, and ended at 50 °C for 5 s. Reactions for C. pengoi and E. ishnus followed the same protocol as N. obtusa except the annealing temperatures were 53 °C and 65 °C, respectively.

The optimal qPCR protocol for D. lumholtzi was as follows: 94 °C for 10 min, followed by 49 cycles of denaturing at 94 °C for 30 s, annealing at 51.5 °C for 30 s, and extension at 72 °C for 1 min, then followed by an extension for 5 min at 72 °C, and ended at 50 °C for 5 s. This protocol required more cycles due to the small size of the organisms, small amount of gDNA after extractions and faint signal in eDNA experiments.

For all protocols, PCR can be used instead of qPCR. For these reactions, 4 μL of sterile water can be used in place of Evagreen and reactions can be tested using agarose gel electrophoresis.

qPCR protocols were also tested on repeated aquarium experiments in which 3–4 jars or aquaria and one negative control were filled with freshwater from Lake Erie, then inoculated with dead organisms for a 24 h period and then filtered following protocols established in our lab and described elsewhere (Sheehan et al. 2024 [3]). All jars/aquaria were tested for a negative result before inoculation and then tested repeatedly for the days following to look for protocol proof of concept and loss of signal. The N. obtusa microcosm was completed using four jars, each filled with 800 mL of lake water, three containing 0.4 g of dried dead organism, and the fourth jar as a control. The C. pengoi microcosm was completed using three jars, each filled with 800 mL of lake water, two containing 5 dead organisms, and the third jar as a control. The D. lumholtzi microcosm utilized four aquaria with 10 L of lake water, three aquaria were inoculated with 200 dead organisms and the fourth aquarium as a control. For E. ishnus, four aquaria filled with 10 L of lake water were used for the microcosm, and three aquaria were inoculated with 28 dead organisms and the fourth acted as a control. All experiments were monitored for temperature and kept at room temperature for the experimental duration.

2.6. In Silico and In Vitro Testing

To ensure species specificity for the primer sets and to prevent false positives, we completed in silico testing via the in silico PCR tool by PrimerDigital [47] and in vitro testing against related species when possible. To complete the in silico analysis, we selected closely related, sympatric species when possible and expanded the sampling as needed. DNA sequences were taken from NCBI [48] and confirmed in BOLD [22] because they provide geographic information from the collected specimen. If multiple sequences of a species were available, the sequences from organisms closest to the NAGL were selected. We utilized the Great Lakes Water Life Explorer tool provided by the Great Lakes Environmental Research Lab, (GLERL) [49] through NOAA for complete species lists of organisms known to be in the NAGL. For example, with C. pengoi, B. longimanus was used for the reasons explained in Section 2.2, and because there are no other species in the same family present in the NAGL. Furthermore, there are no other species in the genus Cercopagis, preventing any comparisons within the genus. Nittellopsis obtusa is the only known species from the family Characeae found in the NAGL and so we used Lychnothamnus barbatus for reasons outlined in 2.2 and added the stonewart Chara contraria for additional confirmation of primer specificity. In the family Gammaridae, the NAGL host two genera (Echinogammarus and Gammarus) and five species, E. ischnus, G. faciatus, G. lacustris, G. tigrinus, and G. pseudolimnaeus. The newly created primers were used to test each of these species.

To test D. lumholtzi, simulated PCR with the new primers was completed on all the species of the genus Daphnia found in the NAGL according to GLERL and three species that are closely related according to Cornetti et al. [35]. In total, the 12 Daphnia species from the NAGL tested included, D. ambigua, D. catawba, D. dentifera, D. dubia, D.galeata, D. laevis, D. longiremis, D. middenforffiaena, D.parvula, D. pulex, D. pulicaria, and D. retrocurva. The three closely related included were, D. sinesis, D. similis, and D. magna.

For the in vitro testing analysis of our primer sets and protocols, we were able to test closely related species against three of our primer sets. We made significant efforts to obtain samples that we could use but fell short with C. pengoi. For each testing protocol we followed our published protocols exactly with different species. Daphnia magna was tested with D. lumhotlzi protocols. We collected one species of macroalgae, Chara contraria (native), and two aquatic plant species Potamogeton crispus (AIS), and Utriculari macrorrhiza (native) from Lake Erie to test with N. obtusa protocols. To test E. ischnus protocols we used G. faciatus and another scud outside the Gammaeridae, but found in the NAGL, Hyella azteca.

3. Results

3.1. Limit of Detection, Cq Value Data and Microcosm Signal Duration

For each target organism, we developed a pair of primers and qPCR protocols which combine to make dependable tools for eDNA testing of these AIS (Table 1). Best fit, linear standard curves with regression line data on all species are overlayed in Figure 1. gDNA dilution assays, standard curve details and qPCR melt curves showing proof of concept are available as Supplementary Material (Figures S1–S8).

For N. obtusa, Figure S1 presents a 5.60 ng/µL gDNA sample that was serially diluted to 1:100,000 ng/µL. The qPCR protocol was able to detect a signal within all ranges. All Cq values were plotted against their concentration to create a standard curve in Figure S5, identifying an estimated concentration for each quantified Cq value.

The C. pengoi gDNA sample was serially diluted by an order of magnitude with each dilution from an 18.16 ng/µL sample to 1:10,000 ng/µL, shown in Figure S2. The qPCR protocol was able to detect a signal from each dilution. All Cq values were plotted against the corresponding concentration to create a standard curve in Figure S6, identifying an estimated concentration for each quantified Cq value.

With the gDNA D. lumholtzi sample, a 7.84 ng/μL sample was serially diluted by order of magnitude to 1:100,000 ng/μL, shown in Figure S3. The qPCR protocol was able to detect a signal from each dilution. Cq values were plotted against their concentration to create a standard curve shown in Figure S7, identifying an estimated concentration for each quantified Cq value.

As with the E. ischnus gDNA sample, it was serially diluted from 6.50 ng/μL to 1:10,000 ng/μL. The qPCR protocol was able to detect a signal from each dilution. Cq values were plotted against the corresponding concentration to create a standard curve (Supplementary Materials), identifying an estimated concentration for each quantified Cq value.

For all four species, lab-controlled microcosm experiments also demonstrated that the primers and protocols successfully detected eDNA. For N. obtusa, the microcosm indicated that a loss of signal was indicated at day 14, with a Cq at inoculation of 26.67. For D. lumholtzi, the microcosm indicated that the loss of signal occurred at 2 days with an inoculation Cq of 36.80. For C. pengoi, the microcosm indicated a loss of signal at 2 days, with an inoculation Cq of 34.20. For E. ishnus, the microcosm indicated that the loss of signal was at 12 days with an inoculation Cq of 34.26.

3.2. In Silico and In Vitro Results

In vitro primer tests with species other than our target species resulted in failed PCR amplification, confirming specificity for the species we were able to test. Results from the in silico analysis are shown in

Table 2. Results of the in silico trail for primer specificity. For each of the target AIS (bold) are the primer names and length, the additional species tested with said primers, the % match with the primers and the in silico PCR result.

Target AIS	Test Species	NCBI #	F Primer Match	R Primer Match	Amplification
Nitellopsis obtusa
	N. obtusa	AY170447	20 of 20 bp—100%	20 of 20 bp—100%	Success
Primers tested	L. barbatus	AY170448	14 of 20 bp—70%	14 of 20 bp—70%	Fail
Nobt_485F—20 bp	C. folia	MZ682285	14 of 20 bp—65%	16 of 20 bp—85%	Fail
Nobt_695R—20 bp
Cercopagis pengoi
	C. pengoi	OP830033	23 of 23 bp—100%	25 of 25 bp—100%	Success
Primers tested	B. longimanus	MH321333	14 of 23 bp—61%	16 of 25 bp—64%	Fail
Cpen_115F—23 bp	Moina sp.	LC503929	14 of 23 bp—61%	17 of 25 bp—68%	Fail
Cpen_302R—25 bp
Daphnia lumholtzi
	D.lumholtzi	AY921417	22 of 22 bp—100%	19 of 19 bp—100%	Success
Primers tested	D. ambigua	MG448989	19 of 22 bp—86%	17 of 19 bp—89%	Fail
Dlum_163F—20	D. catawba	AY380454	16 of 22 bp—73%	15 of 19 bp—79%	Fail
Dlum_397R—19	D. dentifera	MG448806	19 of 22 bp—86%	11 of 19 bp—58%	Fail
	D. dubia	AY921411	16 of 22 bp—73%	11 of 19 bp—58%	Fail
	D. galeata	MH746186	17 of 22 bp—77%	13 of 19 bp—68%	Fail
	D. laevis	MG449455	17 of 22 bp—77%	14 of 19 bp—74%	Fail
	D. longiremis	AY921413	19 of 22 bp—86%	15 of 19 bp—79%	Fail
	D. middenforffiaena	KC502301	18 of 22 bp—82%	14 of 19 bp—74%	Fail
	D. parvula	MG936477	16 of 22 bp—73%	15 of 19 bp—79%	Fail
	D. pulex	MG315907	17 of 22 bp—77%	14 of 19 bp—74%	Fail
	D. pulicara	MG448615	17 of 22 bp—77%	14 of 19 bp—74%	Fail
	D. retrocurva	OP830209	14 of 22 bp—64%	15 of 19 bp—79%	Fail
	D. sinesis	LS991517	18 of 22 bp—82%	14 of 19 bp—74%	Fail
	D. similis	MF346400	15 of 22 bp—68%	15 of 19 bp—79%	Fail
	D. magna	MG317471	16 of 22 bp—73%	14 of 19 bp—74%	Fail
Echinogammarus ischnus
	E. ischnus	FJ581620	22 of 22 bp—100%	21 of 21 bp—100%	Success
Primers tested	G. faciatus	MG734968	8 of 22 bp—36%	13 of 21—62%	Fail
Cisc_357F—22 bp	G. lacustris	MG318006	10 of 22 bp—45.5%	17 of 21—81%	Fail
Cisc_547R—21 bp	G. tigrinis	FJ581684	9 of 22 bp—41%	14 of 21—66.6%	Fail
	G. pseudolimnaeus	EU574907	10 of 22 bp—45.5%	13 of 21—62%	Fail

and confirm with all target AIS that related and sympatric species will not successfully amplify for the original primers and protocols developed in this study.

4. Discussion

4.1. Discussion of LOD Data Relative to eDNA Testing in Similar Experiments

For this study, the objectives were to collect and identify four aquatic invasive species, N. obtusa, C. pengoi, E. ishnus, and D. lumholtzi, isolate their gDNA, develop a set of primers to identify eDNA signals for each species and develop/test qPCR protocols with serially diluted gDNA and eDNA from organism treated lake water. We accomplished this with all four species as they all had a detectable range for gDNA, but as was noted in the results section, variation in Cq range and LOD numbers indicates that no two AIS are equal in the amount of their DNA present.

Limit of detection (LOD) numbers (Table 1) for these samples are comparable to others found in the literature. Examples of reported LOD numbers found in similar studies using qPCR for eDNA species detection include the following: Blackman et al. [50] found an LOD of 1 × 10⁻⁴ ng/μL in their study on effective eDNA detection of quagga mussel, Dressena rostriformis bugensis [50]. In a study testing the sensitivity of nested PCR, conventional PCR and qPCR to detect eDNA of a non-native fish, Pseudorasbora parva, in England, investigators found the LOD in qPCR to be 3.34 × 10⁻⁶ ng/μL or approximately 76 copies/ul. They found that their qPCR assay had a 100% detection rate at 1 × 10⁻⁴ which was less sensitive than their nested or conventional PCR assays [51] contrary to what was cited in the introduction as found by Fediajevaite et al. [16]. Wozney and Wilson [52] provided working primers and qPCR protocols for four species of Asian Carp that are potential invaders in the NAGL along with a protocol for multiplex assays. They reported an LOD between 1 and 10 copies/reaction with 10 copies providing 100% detection and 1 copy providing an 83% success rate [52]. Similarly to our study, Hernandez et al. [53] provided primers and protocols for invasive, threatened or exploited organisms in Quebec, Canada. They only reported LOD for a portion (18) of their study species, and the LOD for these species had a range of 2–20 gene copies/reaction using CO1 from the mtDNA [53]. Mauvisseau et al. [54] compared known protocols using CO1 and 16S for eDNA detection of the Freshwater Pearl Mussel, Margaritafera margaritafera, and found the CO1 assay to be more reliable. They reported the LOD at 10 × 10⁻⁴ which was similar to previously published work on the mussel [54] but was judged as high compared to eDNA from other organisms such as the Red Swamp Crayfish and crested newt [55]. The cited work on Redswamp Crayfish, Procambus clarkii, reported a detection limit of 10 × 10⁻⁸ ng/μL [56] and the LOD for crested newt, Triturus cristatus, was reported as 10 × 10⁻⁷ [57]. In reviewing the species specific eDNA qPCR LOD for a wide range of aquatic species we note that our findings of LOD numbers, which vary by orders of magnitude, are not outside the range found in various studies performing important work to aid in our understanding of eDNA species detection.

4.2. Conservation Relevance of the Target Research

The availability of working eDNA qPCR primers and protocols for N. obtusa, C. pengoi, E. ishnus, and D. lumholtzi, is a major step towards development of a rapid identification protocol of these AIS in ballast water transport. Similar protocols have become the basis for large-scale microcosm experiments involving eDNA of Hemimysis anomala and other aquatic invasive species. Species specific primers and qPCR protocols could prove beneficial in ballast water transport regulation because they can help determine if the eDNA signal indicates present organisms or that of legacy eDNA [3]. These protocols have also been used to compare the efficacy of a targeted qPCR approach vs. a metabarcoding approach to test for AIS presence and preliminary data shows the qPCR approach to be very effective (Gruwell and Welsbacher, unpublished data). Indeed, the Great Lakes Commission and the US Fish and Wildlife Aquatic Nuisance report [58,59] highlight rapid AIS detection as a critical step in fighting the invasive species problem. eDNA monitoring techniques such as the protocols we have provided, have repeatedly been suggested as less expensive and more reliable than traditional monitoring techniques [13,14,16,60,61].

We have demonstrated that our qPCR methods are quantitative over several orders of magnitude of DNA target concentrations and have used them in replicated microcosm experiments to show that eDNA signal degrades rapidly to levels below detection limits within days (as in Sheehan et al. [3]). These data are consistent with other studies using large organisms where the organisms were removed from microcosms and eDNA signal rapidly decreased (e.g., Kirtane et al. [62]). Our experiments are unique and particularly relevant to ballast water-mediated transport because we conducted the experiments with planktonic invertebrates and algae, organisms that are small enough to be taken up in ballast tanks and where removal experiments are impractical. Next steps in this research, which are currently underway, include testing these protocols in regular sampling efforts in the NAGL to determine their efficacy in locating these species so that the protocols can be used in the efforts to track the spread of our target AIS in the NAGL and other aquatic systems.

5. Conclusions

Our goal in this project was to provide rapid and reliable eDNA qPCR methods for detecting four invasive species from the lower NAGL. Each of these species has been identified by GLANSIS as invasive and a potential threat to becoming established in Lake Superior. If used, the eDNA protocols for testing water samples outlined in this paper, will aid in the rapid identification of these AIS in the NAGL and potentially aid in slowing their distribution.