Development of a species-specific qPCR assay for the detection of invasive African sharptooth catfish (Clarias gariepinus) using environmental DNA

Detection and monitoring of target species is the primary strategy in the management and control of biological invasions. Traditional methods to detect invasive species are time consuming and cumbersome, with a requisite for trained taxonomists for identification of aquatic species. Environmental DNA (eDNA)-based molecular methods offer an alternative, as they are quick, cost-effective and require minimal manpower. In this study, we design and optimize a reliable eDNA-based quantitative PCR assay to detect the African Sharptooth catfish, a highly invasive and banned species in India. Here, we delineate the step-by-step processes involved in the design and optimization of the assay, and show its performance through field-testing in selected water bodies in and around the city of Hyderabad. The present workflow can be used to design assays to detect a wide range of aquatic species.


Introduction
The intricate connection of biological invasions with increasing globalization, trade, culture, and human and climate-mediated events makes prevention and control of invasive species an exceptional challenge (Meyerson et al. 2022). With predictions that the number of alien species, as well as the intensity of biological invasions will accelerate on most continents (Seebens et al. 2021), there is an urgent need to develop and implement effective solutions for control, monitoring and management. Such strategies are particularly required in regions noted for their exceptional biodiversity and endemism, and where threats from alien species are remarkably high (see Dawson et al. 2017).
In India, a mega-diversity country, threats to biodiversity from alien invasive species are on the rise, but management and control of such species is inadequate because of insufficient research, funding and policies (Goyal et al. 2021;Mungi et al. 2019). A recent synthesis by Bang et al. (2022) revealed that the Indian economy had incurred a loss of at least US$ 127.3 billion between 1960 and 2020 due to Abstract Detection and monitoring of target species is the primary strategy in the management and control of biological invasions. Traditional methods to detect invasive species are time consuming and cumbersome, with a requisite for trained taxonomists for identification of aquatic species. Environmental DNA (eDNA)-based molecular methods offer an alternative, as they are quick, cost-effective and require minimal manpower. In this study, we design and optimize a reliable eDNA-based quantitative PCR assay to detect the African Sharptooth catfish, a highly invasive and banned species in India. Here, we delineate the step-by-step processes involved in the design and optimization of the assay, and show its performance through field-testing in selected water bodies in and around the city of Hyderabad.
invasive species, with an average annual cost of US$ 2.1 billion. Bang et al. (2022) further cautioned that these calculations could likely be gross underestimations, since they contribute to only a small fraction of the actual costs incurred.
About 7.2% of global fish diversity occurs in India (Froese and Pauly 2022). Though there is no comprehensive assessment, or a list of alien invasive fish species in India, it has been estimated that > 600 fish species have been introduced into the country, with 55 of them having established sustainable reproductive populations (Sandilyan 2022). Among those 55 species, the National Biodiversity Authority (NBA) has declared 14 inland and freshwater species as 'highly invasive' (Sandilyan et al. 2018), which includes Clarias gariepinus (African sharptooth catfish). Further this species has been banned from farming and selling by the Government of India since the year 1997 (Gopi and Radhakrishnan 2002).
Clarias gariepinus is widely regarded as one of the world's most successful aquatic invader due to its generalist (but mostly piscivorous and predatory) feeding habits, high fecundity, fast growth and eurytopic physiological traits (Booth et al. 2010). The species has a widespread presence in around 30 countries, including India, impacting native fish and other aquatic species through competition and predation. Despite its significance as an invasive alien species, only few studies have focused on the occurrence and distribution of C. gariepinus in the country (Krishnakumar et al. 2011;Singh et al. 2013;Roshni et al. 2020), and no attempt has been made so far to map its distribution in India.
Traditional methods to detect invasive fish species such as visual observations, using traps, nets, bioacoustics, etc., and also employing morphological and behavioural data can be biased, intrusive and requires some level of taxonomic expertise (Beng and Corlett 2020). An alternative method that has been highly favoured by conservationists in recent years is the use of environmental DNA (eDNA) to detect aquatic species. eDNA, which is a complex mixture of DNA obtained from different organisms in soil, sediment, water and even air (Taberlet et al. 2012), can be exploited to detect aquatic species including invasive species (Jo et al. 2021) with high accuracy at relatively low cost. With sequencing techniques becoming more affordable and easily available, eDNA based molecular methods are increasingly adopted in conservation science. With a big chasm in invasion biology research in India, detection of target species remains the crucial step in the control and management of invasive species. Here, we successfully show the validation of a quantitative PCR based assay utilising eDNA, rigorously designed, optimized, and tested to specifically detect the invasive Clarias gariepinus.

Primer designing and screening
We targeted the 'Cytochrome b' region of the mitochondria to design primers, and retrieved the sequences of three species of Clariids, Clarias gariepinus, C. dussumieri and C. batrachus from NCBI GenBank (accession numbers: NC_027661.1:14366-15503, NC_037193.1:14365-15502, NC_023923.1:14361-15498 respectively). These sequences were subsequently aligned using Clustal Omega (https:// www. ebi. ac. uk/ Tools/ msa/ clust alo/), and the alignment file was imported to the primer designing module of ssPRIMER (https:// www. matto rtona pps. com/ shiny/ ssPRI MER), a GUI based tool used for designing species specific primers for qPCR assays. Potential primer pairs were designed by selecting appropriate parameters (see supplementary file for the parameters and primer binding visualization). From the list of designed primers, the primer pairs that had a higher propensity to form primer dimers were omitted from consideration for qPCR assay. To further screen the primers, shortlisted primers were assessed in silico using NCBI's Primer-Blast tool, and any primers that amplified other sympatric species were excluded. The details of the finalised primer pair are shown in Table 1.

In vitro testing of primers
Since the DNA sequence databases were not complete with sequences of all extant species, in silico analysis alone does not determine specificity of the screened primer pair. Hence, we tested them for specificity in vitro against the genomic DNA extractions of C. gariepinus and a few additional non-target species by performing a PCR assay. The non-target species selected for this assay comprised phylogenetically closely related sympatric species within the genus Clarias (C. dussumieri, C. magur and C. batrachus), sympatric species within Siluriformes (Glyptothorax gracilis, Pangasianodon hypophthalmus, Horabagrus brachysoma and Plotosus canius) and a few distantly related sympatric species from orders other than Siluriformes (Labeo rohita, Labeo catla, Cyprinus carpio, Oreochromis niloticus and Tenualosa ilisha).
The PCR reactions had a total volume of 10 µl and included 1 µl of dNTP mix (10 mM each dNTP), 1 µl of 10X PCR Buffer, 0.05 µl of Taq DNA Polymerase (Takara, India), 0.2 µl of forward primer (10 µM), 0.2 µl reverse primer (10 µM), 6.55 µl of Nuclease free water and 1 µl of template DNA. The thermocycle program consisted of an activation step at 94 °C for 5 min, 35 cycles of PCR step at 94 °C for 30 s, 58 °C for 30 s, 72 °C for 30 s, and a final extension step at 72 °C for 5 min. The PCR products were then visualised in 2% agarose gel.

qPCR assay optimization
To determine efficiency of the primers and linear range, a standard curve was generated by including dilutions of standards, i.e., known copies of target amplicons comprising the target sequence. The standards were prepared in the laboratory by amplifying the target region through bulk PCR reactions (5 reactions of reaction volume 20 µl with the same components and conditions mentioned previously). Upon visualizing the PCR products in 2% agarose gel to confirm amplification, the PCR products were combined and purified using NucleoSpin® Gel and PCR Clean-up (Macherey-Nagel, Germany) kit following the manufacturer's instructions. The purified product was then quantified using Nanodrop spectrophotometer. With the known concentration of PCR product (ng/µl) and the length of the amplicon (171 bp), the number of copies of the amplicon was determined using the online calculator at http:// cels. uri. edu/ gsc/ cndna. html. After determining the copy number, the purified PCR product was then serially diluted to prepare the standards. A standard curve was generated in qPCR with the dilutions of standards ranging from 10 0 copies/reaction to 10 5 copies/reaction. To determine the Limit of Detection (lowest initial DNA concentration with 95% detection) and Limit of Quantification (lowest initial DNA concentration quantifiable with a coefficient of variation below 35%), we performed a qPCR assay using fourfold dilutions of standards ranging from 1 copy/reaction to 1024 copies/reaction. The Limits of Detection and Quantification were calculated using the LOD/LOQ calculator (Klymus et al. 2020a). The qPCR reactions had a total volume of 10 µl and included 5 µl of TB Green® Premix Ex Taq™ II-Tli RNaseH Plus (Takara (Japan), 0.2 µl of forward primer (10 µM), 0.2 µl of reverse primer (10 µM), 3.6 µl of Nuclease free water and 1 µl of template DNA. The qPCR program consisted of an initial denaturation step at 95 °C for 30 s, 40 cycles of PCR step at 95 °C for 5 s and 58 °C for 30 s, a melting step at 95 °C for 5 s, 60 °C for 60 s till reaching 95 °C, and a final cooling step at 50 °C for 30 s. All qPCR experiments were performed in Roche Lightcycler 480 II instrument.

Sampling for field testing
To test the performance of the assay in field conditions, we randomly selected 11 lakes in and around Hyderabad, Telangana State, India, for the pilot study (Suppl. Figure 2). Two litres of water were collected in triplicate from each site at the shore without disturbing the sediment, during the month of January 2021 and filtered in the laboratory on the same day. 250 ml of water from each sample was filtered using disposable 50 ml syringes through mixed cellulose ester membrane of 47 mm in diameter and of 0.45 µm pore size (Merck life science Pvt. Ltd.). After filtration, the filter paper was cut into two halves, with one half utilised for DNA isolation and the other stored in − 30 °C. Besides the 11 lakes, we also included a sample from a pond located inside the Nehru Zoological Park in Hyderabad as the positive control, where the presence of C. gariepinus was visually confirmed. Since we could not reliably identify any natural water body in and around Hyderabad where C. gariepinus is confirmed to be absent, we included a sample from our laboratory aquarium as the environmental negative control. Between filtration of each sample, the filter assemblies were bleached with 4% Sodium hypochlorite solution to prevent contamination between samples. eDNA was extracted from the filters by the standard Phenol-Chloroform-Isoamyl alcohol method. The eDNA filter was placed in a 2 ml microcentrifuge tube and 1 ml of tissue lysis buffer (pH = 8.0), 100 µl of SDS (20%) and 20 µl of Proteinase K (20 mg/ml) were added and vortexed for 1 h. The tube was placed in a rotating wheel at 56 °C for 2 h for lysis. The filter paper was then removed and 700 µl of Phenol:Chloroform:Isoamyl alcohol mixture (25:24:1 proportion) was added to the aqueous content and kept in rotating wheel for 10 min for thorough mixing. The contents were then centrifuged at 10,000 rcf and the aqueous layer was transferred to a new centrifuge tube. 700 µl of Phenol:Chloroform:Isoamyl alcohol mixture was added again to the transferred aqueous layer and the previous steps of mixing and centrifugation were repeated. The aqueous layer was then transferred to a new centrifuge tube, and 700 µl of Chloroform:Isoamyl alcohol mixture (24:1) was added to the transferred aqueous layer and the previous steps of mixing and centrifugation were repeated. The aqueous layer was then carefully transferred to a new 1.5 ml centrifuge tube and 50 µl of 5 M NaCl and 700 µl of chilled isopropanol were added. The contents were mixed well by gently inverting the tubes and stored at 4 °C for 2 h for precipitation. The tubes were centrifuged at 10,000 rcf for 30 min and the aqueous contents were discarded without disturbing the pellets. 500 µl of 70% molecular grade ethanol was added to the pellet and centrifuged at 10,000 rcf for 10 min and ethanol was discarded without disturbing the pellets. The ethanol wash step was repeated with 100% molecular grade ethanol. Upon discarding ethanol, the pellet was air dried till complete removal of any trace ethanol. 100 µl of 1X TE buffer was then added to dissolve the pellet by placing the tubes in dry bath at 56 °C for 30 min and the DNA was stored at 4 °C for further analysis. After eDNA extraction, the DNA concentrations of all samples were adjusted to 20 ng/µl for the subsequent qPCR assay, and no adjustments were made for samples having concentrations less than 20 ng/µl. qPCR detection of eDNA Each eDNA sample was loaded in three technical replicates, along with negative controls, and the assay was run in Roche Lightcycler 480 II. The copy numbers of the target gene in each eDNA sample were calculated with the help of the standard curve generated during the qPCR assay optimization stage. For the reactions where primer dimers were observed through the Melt curve analysis, the assay was repeated for the respective samples. The qPCR assays did not involve any additional probes, since the designed primers were highly specific to the target species. All qPCR assays were performed in a separate laboratory space in a different floor/section of the building dedicated for qPCR experiments to avoid contamination. The qPCR products were purified using NucleoSpin® Gel and PCR Clean-up (Macherey-Nagel, Germany) kit and sequenced on ABI 3730XL DNA sequencer platform using the BigDye Terminator (version 3.1) Cycle Sequencing Kit and POP-7 Polymer separation matrix (Applied Biosystems, Inc.). Upon trimming the sequences in Chromas (Version 2.6.6) software (the sequence lengths were ranging from 96 to 114 bp), the sequences were then analysed in NCBI's nucleotide BLAST tool to verify the species identity of the amplified product.

Selected primer pair
After rigorous screening of the designed primers and their in silico analysis, we finalised a primer pair (Table 1) for the study.
In vitro primer specificity assay Figure 1 provides details of the PCR products of tissue DNA samples from the target and non-target species amplified with the selected primer pair. A crisp band was observed only in the PCR product of C. gariepinus DNA sample at the expected range of 171 bp, while no other bands were observed for the closely related and other non-target species.
qPCR assay standardization To estimate the absolute copy number of the target gene in eDNA samples, we generated the standard curve with the R 2 value of 0.9973, efficiency of 96.5% and the y-intercept value (the predicted Cp of a reaction with 1 copy of the target sequence) at 34.76 cycles (Fig. 2). Through LOD/LOQ assay, the limit of detection was found to be four copies and the limit of quantification was found to be nine copies.

eDNA detection in representative samples
Of the 12 lake-water samples, 11 produced amplifications, including the positive sample from the pond inside the Nehru Zoological Park (Fig. 3). Only one sample (site KC) was negative. The copy numbers calculated for all the positive samples using the standard curve were above the limit of detection and quantification, confidently indicating the presence of C. gariepinus. No amplification was observed in the environmental negative control (ENC) as well as in the no template controls. When the sequences of qPCR products were analysed in NCBI's nucleotide BLAST tool, the input sequences showed significant alignments with multiple sequences of C. gariepinus species with percentage identity ranging from 100 to 93% and query coverage from 100 to 97%. We also observed possible hits with other species namely, Clarias anguillaris, Bathyclarias gigas, Bathyclarias ilesi, Bathyclarias nyasensis and Bathyclarias worthingtoni with percentage identity ranging from 98 to 93%. Since, these species are either endemic to Africa Fig. 1 In vitro PCR validation of primers against the target species C. gariepinus and 12 additional non-target species or not present in India, the identity of the analysed sequences can be confirmed that of C. gariepinus.

Discussion
While real-time quantitative PCR assays have been used in various fields involving detection and quantification of specific nucleotide sequences (Kubista et al. 2006), its potential as a tool for environmental DNA studies is only recently emerging. In this study, we designed a cost-effective eDNA based qPCR assay for the detection of North African sharptooth catfish in natural aquatic ecosystems. Before embarking on a large-scale eDNA based study to map the distribution of any target species, a pilot study is recommended to standardize the design, including the development and validation of assay, as well as considerations for contamination and suitable analysis methods (Goldberg et al. 2016). Such a pilot study also enables reoptimization and validation of the assay when applied in different geographical regions.
Mitochondrial gene sequences are preferred as the target sequence for eDNA studies as it increases the chances of detection because of the high copy number in cells (Rees et al. 2014), despite their inability to distinguish hybrids (Evans and Lamberti 2018). Also, incorporation of specificity (i.e., detection of only the target species) and sensitivity (i.e., detection of target DNA at low quantities) assessments is vital to make the assay more reliable (Klymus et al. 2020b). While the in silico and in vitro validation of primers against the non-target species informs about the specificity of the assay, the limit of Detection (LOD) and Limit of Quantification (LOQ) assessments inform about sensitivity.
To ensure that the primers are specific to the target species, it is imperative to include the phylogenetically closely related species and distantly related sympatric species in the in vitro specificity assay. In addition, verification of the positive detections by sequencing the PCR products adds another layer to assay integrity. Keskin (2014) and Elberri et al. (2020) have previously demonstrated the importance of qPCR based eDNA studies to detect C. gariepinus. However, the primers used in their studies were amplifying other Indian congeneric Clarias species (e.g., C. magur, C. dussumieri) as well. Hence, our assay was optimized to detect C. gariepinus in India with high specificity, by including all native species of the genus Clarias, and selected species of other closely related families in the in vitro validation. For future studies outside India, we suggest revalidation of the specificity of our assay by including other co-occurring closely related species in the geographical range of interest. Since our assay does not include any additional probes, this has reduced the cost involved and can be used by any laboratory equipped with a basic qPCR machine.
The goal of developing, optimizing and testing a species-specific qPCR assay for eDNA based studies with stringent quality control measures is to detect the target species. This pilot study will serve as a foundation to map the distribution of invasive C. gariepinus, and also as a useful tool to inform management authorities for timely control and regular monitoring of this species. Finally, the workflow employed in this study can also serve as a template to design and optimize eDNA based assays to detect other invasive and/ or threatened species for improved aquatic management and conservation.