Snails

All snails used in this study were laboratory-reared Biomphalaria glabrata of the M-line strain [28]. Snails were fed lettuce ad libitum and maintained on a 12 h light: 12 h dark schedule in artificial spring water [29]. Snails of 8–11mm shell diameter were used.

Treatment of snails with niclosamide

Niclosamide was purchased from Sigma and dissolved in dimethyl sulfoxide (DMSO) (Sigma). A volume of 2L of artificial spring water was added to each of four 3L plastic containers. Also, each container also received a total of 32–34 snails. Three niclosamide concentrations, 0.15 mg/L, 0.10 mg/L, and 0.05 mg/L, were tested. DMSO vehicle alone was added to the fourth, control container. The final DMSO concentration in each of the four containers was the same, at 1/1000 (v/v). Snails were exposed to niclosamide for 24 hours at 26–28°C with aeration. After the 24-hour exposure, twenty live snails from each container were collected and randomly divided into four replicate groups, each with five snails, for subsequent RNA extraction. For each of the three treatment groups and the control, the four replicate groups (five snails each) that were collected are considered to constitute valid biological replicates because each replicate is comprised of a distinct set of five snails. Furthermore, our design ensured that exposure conditions for a particular concentration were identical for each replicate.

We define the concentrations used as “sublethal” because the snails we sampled were alive and responsive after the 24 hour period of niclosamide exposure. It is possible that some of the snails selected for study would have died upon further observation, but had not done so at the time of sampling.

Extraction, qualification and quantitation of RNA

For a particular treatment, whole bodies of five snails dissected from their shells were pooled as a single sample, and ground in liquid nitrogen. Two sequential RNA extraction methods were applied to each such sample, first using Trizol (Invitrogen) and then the PureLink RNA kit (Ambion), following manufacturers’ instructions. Quality and quantity of RNA were checked using an Agilent Bioanalyzer 1200 and NanoDrop spectrophotometer and quality was determined to be high (sharply focused rRNA bands and A260/A280 ratio: 1.9–2.1) for all samples.

Description of microarray

The microarray used for this study contains 60-mer oligonucleotide probes that were designed using transcriptomic sequence data from B. glabrata from publically available databases in 2010. Sequences were computationally assembled to unique predicted transcripts using a method described [30] and used to design probe sequences and a microarray (duplicate probes were arranged in 8 X 60K layout) using the eArray facilities from Agilent. The details of probes, target sequences and microarray design (eArray AMADID 033677) are available at Gene Expression Omnibus (GEO) at Platform Access Number GPL20716 (www.ncbi.nih.gov/geo/). Compared to recent microarray designs (i.e., arrays with 1.1K oligonucleotide probes [31] and 5K cDNA probes [32]), this 31K microarray design provides the most comprehensive coverage of expressed sequences from B. glabrata to date.

Microarray hybridization

To provide a positive hybridization signal to as many elements of the array as possible, universal reference RNA (URR) consisted of 80% RNA from the control group and 20% RNA from experimental groups exposed to the 3 concentrations of niclosamide. All experimental samples contributed equally to the URR.

All procedures related to the microarray experiment were based on Agilent’s two-color microarray-based gene expression analysis (Version 6.6). Unless otherwise mentioned below, all reagents were purchased from Agilent Technologies and were used in accordance with the manufacturer’s protocol.

A total of 200 ng RNA was used for each biological replicate of the four unknown groups (control snails or snails exposed to the three different dilutions) to be examined. Spike A buffer mix was added to the unknown RNA sample and Spike B buffer mix was added to 200 ng of the URR. Complementary RNA (cRNA) amplification and labeling reactions with cyanine dye Cy–3 in the unknown sample and cyanine dye Cy–5 in the URR were conducted at 42°C for 2 hours. The labeled cRNA was purified using the RNeasy Mini Kit (Qiagen). The cRNA was quantified to determine the concentration and specific-labeling efficiency using a NanoDrop spectrophotometer.

A total of 300 ng of both the purified Cy–3 and Cy–5 labeled cRNA was added to 5 μl of 10x hybridization buffer and incubated at 60°C for 30 min to fragment RNA. A total of 40 μl of hybridization solution (representing a single replicate for a particular unknown sample) was added to one of the eight gasket chambers. Each of the eight available gasket chambers was in this way filled with a different sample. The gasket surrounding the eight chambers was then covered with a microarray slide such that each of chambers contacted one of the eight different identical arrays printed on that slide. Two slides were used for this study, each with eight identical arrays. The 16 arrays thereby accommodated the 16 different samples that needed to be run (4 unknowns by 4 replicates per unknown). The assembled hybridization chamber was placed in a hybridization oven at 65°C for 17 hours (rotating at 10xg).

After hybridization, the slides were washed in Agilent’s wash buffer I for 1 min (room temperature), then in wash buffer II for 1 min (37°C). After stabilizing against ozone deterioration in stabilization buffer, the slides were dried and shipped to the University of California San Francisco’s Viral Diagnostics and Discovery Center, where they were scanned with Agilent’s Microarray G2505C Scanner and extracted and normalized using Agilent’s Feature Extraction software.

Microarray data analysis

Microarray data were analyzed using GeneSpring GX version 12.6.1 (Agilent). For quality control, the program called “Filter probsets on data” was used. After removing all control probes, experimental probes were retained for further analyses only when they were expressed in at least 75% of all replicates from at least one biological condition (experimental or control). The features that were not positive, not significant, not above background noise, not uniform, not saturated or that were population outliers were not analyzed further. The remaining features were analyzed using volcano plots. Benjamin Hochberg FDR was applied to correct for multiple testing. Corrected P-value cut-off and fold change cut-off were set at 0.05 and 2, respectively. In accordance with MIAME guidelines, all relevant data were deposited in the GEO database at NCBI (accession GSE71223).

Blast search, gene ontogeny (GO) and heatmap analyses

The features that passed the statistical criteria (2 fold change and P< 0.05) were selected for subsequent analyses. The full-length transcripts associated with probes (see GEO platform accession GPL20716) that showed differential expression were selected for Blastx search and GO analysis (Blast2Go v.5; www.blast2go.com). The three GO term categories (biological process, molecular function and cellular component), all at the level 2, were generated using the program.

All heatmaps were done in R Studio version 0.98.953 [33–36]. The transcripts that were differentially expressed were categorized according to GO terms (http://geneontology.org/) and UniProtKB (http://www.uniprot.org/) functional descriptions. Many of the proteins shown could be placed in multiple functional categories, but for simplicity we have included each protein in only one category.

Phase I and phase II biotransformation

Based on studies with better-characterized mammalian models, cytochrome P450 (CYP) enzymes are involved in the most important phase I reactions [35]. CYPs constitute a superfamily of structurally diverse and functionally versatile enzymes with more than 15,000 known genes distributed across all biological kingdoms [41]. CYPs generally catalyze a monooxygenation reaction, in which an atom of oxygen is added to the xenobiotic [39]. In insects, the association of CYP expression with insecticide or drug resistance is well established [42–44]. For example, high expression of a single CYP allele (CYP6G1) confers resistance to DDT in Drosophila melanogaster [45] and to pyrethroids in Anopheles funestus [46]. Upregulation of CYP expression can be induced by many different xenobiotics, which generally act by combining with an intracellular receptor to form a transcription factor that then binds to xenobiotic response elements of target genes, activating their transcription [39]. Interestingly, the induced CYPs may not be involved in the metabolism of the xenobiotic that causes their upregulation.

In snails CYPs are expected to be diverse, and the array used for this study contains about 100 distinct features with similarity to CYP genes, but thus far, we know little of the overall diversity of CYP genes in B. glabrata, and their full range of functions. In this study, 9 CYPs were upregulated, and one was downregulated in niclosamide-exposed snails (Table 1 and Fig 4). Out of 16 transcripts with ≥5 fold upregulation, 4 were CYPs, indicating they are an important component of the snail transcriptomic response following exposure to niclosamide. Only one full-length cDNA of a CYP (AY922309) has been previously reported from B. glabrata, and exposure of snails to S. mansoni resulted in downregulation of its expression [47]. Since sequences of CYPs are highly diverse [41] and only partial sequences of the 10 differentially expressed CYPs observed in the study are available, a phylogenetic analysis is unlikely to reveal reliable relationships among the snail CYPs discussed. However, the complex expression pattern of the 10 CYPs presented in this study as well as altered expression of one CYP (AY922309) in response to S. mansoni suggest CYPs have diverse functions in B. glabrata.

Provision of additional functional information regarding the specific CYPs induced by niclosamide or other molluscicides is important, because it will provide further needed details regarding the detoxification process in snails, and could potentially lead to an ability to select for compounds with higher specific activity for snails relative to other aquatic organisms. Also, prior to widespread application of molluscicides, it would be of interest to identify the array of CYPs present, to facilitate subsequent monitoring for the possible emergence of molluscicide-resistant snails that might express mutated CYPs as has been documented in insects. Thus far there has been no clear indication for the emergence in snails of resistance to niclosamide [48], although Sullivan et al. (1984) reported an approximately two-fold higher tolerance to niclosamide in a laboratory strain of B. glabrata after 5 generations of selection [49].

Relative to phase II (conjugation) proteins, our study revealed one glutatione S-transferase (GST) that was 8-fold overexpressed following niclosamide exposure. Like CYPs, GSTs are also markers of metabolism of xenobiotics and indicators of the presence of contaminants [50]. Based on Illumina RNA-seq analysis, Zhao et al. (2015) described the responses of the amphibious snail, Oncomelania hupensis, after challenge with two different niclosamide-based molluscicides. They showed that two CYP genes and one GST gene were upregulated following molluscicide exposure [51]. This observation coupled with our results suggests that both enzyme families are a common component of the snail transcriptomic response to molluscicides. In contrast, GSTs but not CYPs were shown to be highly expressed in a microarray study of the gills of the mussel Mytilus galloprovincialis following exposure to salts of heavy metals [52]. This suggests that molluscs can respond differently depending on the nature of the toxicant.

Phase III biotransformation

Drug transporters play a vital role in translocation of compounds such as nutrients, wastes, toxins and xenobiotics into or out of cells [53–55]. The transporters work in conjunction with drug metabolizing enzymes such as CYPs and phase II enzymes for drug elimination. They can be classified as influx and efflux transporters, which are located either at the basolateral or apical membranes.

Efflux transporters are ATP-binding cassette (ABC) transporters that belong to a superfamily including multidrug resistance proteins (MRP). These efflux pumps determine bioavailability and concentrations of many drugs. In our study, a highly expressed MRP was found (contig_14304). MRPs preferentially transport anionic compounds and compounds detoxified by cellular enzymes such as GST.

Influx transporters are members of the solute linked carriers (SLC) superfamily responsible for transporting organic anions, organic cations or oligopeptides [56]. In this study, two transcripts that encode solute carrier transporter family 28 (c27272) and organic cation transporter (c36549_rc) have also been shown to be 6- and 4-fold upregulated, respectively, following niclosamide exposure (Table 1).

In addition, we have noted that a sodium-dependent multivitamin transporter (c1938) and a cystine-glutamate transporter (c32432), two additional influx transporters, were also highly expressed. Sodium-dependent multivitamin transporter is an important transmembrane protein responsible for translocation of vitamins and other essential cofactors such as biotin. Vitamins are required for detoxification metabolism and vitamin E (α-tocopherol) is involved in repair of peroxidized lipids [40]. Cystine-glutamate transporter mediates cystine entry in exchange for intracellular glutamate in mammalian cells. Cystine is converted to cysteine, which is required for synthesis of glutathione, an antioxidant that prevents damage due to reactive electrophiles.

Transporters are well-known for their roles in drug efficacy and resistance [57]. ABC transporters are upregulated in schistosomes in response to praziquantel [58,59]. Knockdown of ABC transporters enhances susceptibility of adult and juvenile schistosomes to praziquantel [60]. In addition, ABC transporters play a critical role in diverse physiological functions in schistosomes including immune responses and reproduction [55]. The role of transporters in the response of snails to molluscicides awaits further study.

In summary, genes encoding key molecules involved in all three phases of biotransformation in B. glabrata were responsive to molluscicide treatment and provide a basis for beginning to understand the molecular basis of detoxification in freshwater snails.

Stress response

The stress response is characterized by the production of stress proteins which tend to be relatively well-conserved across both prokaryotes and eukaryotes. Often it is the presence of denatured proteins that triggers a stress response, including production of heat shock proteins (HSP). HSPs, mainly acting as molecular chaperones, are involved in protein folding, assembly, degradation, and intracellular localization. Under normal conditions, HSPs are constitutively expressed. Heightened expression is triggered by various physiological perturbations or stressors (e.g. elevated temperature, hypoxia, ischemia, heavy metals, radiation, calcium increase, glucose deprivation, pollutants, drugs, cancer, and microbial infection) [61]. Functions within the HSP superfamily are highly diverse. We found six HSPs belonging to three families (HSP20, HSP40 and HSP70) to be upregulated upon exposure to niclosamide (Table 1 and Fig 4). High molecular weight HSPs such as HSP90s were not differentially expressed.

HSP20s protect other proteins against heat-induced aggregation or denaturation. HSP20 was identified as a biomarker for environmental stress in the disk abalone, Haliotis discus discus, and its expression could be induced by extreme temperatures, salinities, heavy metals and microbial infection [62]. HSP90, HSP70, HSP24.1 and sequestosome–1 were also highly expressed in the marine bivalve Mytilus galloprovincialis exposed to toxic metals [54]. The latter gene encodes a protein involved in ubiquitin binding and is therefore related to proteasome degradation. We also found this gene to be upregulated in B. glabrata exposed to niclosamide. HSP40s, also referred to as DnaJ/Hsp40, stimulate the ATPase activity of chaperone proteins, HSP70s, by stabilizing their interactions with protein substrates [63].

In addition to stress responses, HSPs are also involved in detoxification, immune responses, pathogenesis and cancer development [64,65]. In Biomphalaria, an increased expression of HSPs has been linked to susceptibility of snails to schistosome parasites [66–68]. The relationships between the HSPs discussed in these studies and the 6 HSPs presented in this study are unclear because complete cDNA of the HSPs we studied are lacking. Eventual comparisons of full-length HSP cDNAs will help resolve this matter.

Intracellular and transmembrane trafficking

The most highly expressed gene we noted following niclosamide exposure was the gene encoding ADP-ribosylation factor 1 (ARF1). ARF1 is a member of the family of GTPases, and a key regulator of intracellular vesicle trafficking at the Golgi apparatus and endosomes. At the Golgi complex, ARF1 facilitates membrane recruitment of many cytoplasmic coat proteins to allow sorting of membrane proteins for transport. It also stimulates the activity of enzymes that modulate the lipid composition of the Golgi, and that assemble cytoskeletal scaffolds on the Golgi [69–71]. A recent study demonstrated that ARF1 upregulation, apparently coupled to ARF4 downregulation, enables Golgi secretory pathway activity to continue even in the presence of inhibitors [72]. This effect may involve an ARF-CREB signaling pathway, which may be affected by a CREB3 regulatory factor-like isoform, which acts as a negative regulator of the ER unfolded protein stress response. The upregulation of CREB3 regulatory factor-like isoform we observed may then interfere with the production of some ARF molecules like ARF4, which could be compensated for by the production of others, like ARF1 [72].

In eukaryotic cells membrane compartments are connected through cargo-selective vesicle trafficking, thereby mediating the exchange of components between different organelles. This exchange is essential to maintain structural integrity and specific composition [73]. A fundamental regulatory step in vesicle formation is the activation of small ARF GTPases by exchanging their bound GDP for GTP, which is a prerequisite for ARF-mediated effector recruitment [74]. In our study, in addition to ARF1, four additional GTPases (c17735, c28906, c37491_rc, and BGC03483), all with a conserved ABC-ATPase domain, were over-expressed in niclosamide-treated snails.

Other recent studies also suggest that ARF1 plays an important role in modulating the morphology and function of mitochondria [75–76]. Mitochondria and ER are found in close proximity to each other and it is thought that they maintain contact sites to facilitate exchange of molecules. ARF1 is well known for its essential role in the generation of coatomer protein I-coated vesicles in the Golgi complex, which are important for maintenance of mitochondrial function, possibly at ER-mitochondrial contact sites. Indeed, our study has revealed that two coatomer genes were over-expressed following niclosamide treatment (Table 1). Thus, the upregulation of ARF1, coatomer, and GTPase genes may be an important compensatory response to toxic damage, allowing maintenance of Golgi and mitochondrial function.

In contrast to a potential role in preserving organelle function, GTPases can be involved in a process involving mitochondrial fragmentation. Mitochondria form a highly dynamic network throughout the cell, which is maintained through constant fission and fusion of mitochondrial tubules. These fission and fusion events are regulated by two types of GTPases in mammalian cells [77]. Park et al. (2011) showed that niclosamide was a potent inducer of mitochondrial fragmentation in human HeLa cells, resulting in the disruption of mitochondrial membrane potential, reduction of adenosine triphosphate (ATP) levels, and cell death, and that this fragmentation was mediated by the GTPase Drp1 [78]. Whether the 4 GTPases upregulated in our study have a protective or cytotoxic role requires further study.

Genes downregulated following niclosamide exposure

One of the most conspicuously downregulated genes with known homologs in GenBank was FREP4. FREP4 is involved in immune responses to trematodes [79] as well as in ontogenesis [80]. FREPs belong to a gene family with many members, and the functions of its members are expected to be diverse [81–83]. Other studies have also noted underexpression of immune genes in animals subjected to toxin or xenobiotic pressure. Varotto et al. (2013) found that several transcripts associated with immunity including fibrinogen-containing protein gene 7 were under expressed in M. galloprovincialis exposed to toxic metals [52]. In strains of the mosquito Aedes aegypti resistant to Bacillus thuringiensis israelensis toxins, genes involved in immune responses were generally under-transcribed [84], and exposure of sphingid caterpillars to toxic plant compounds was shown to weaken their melanization immune responses [85].

In planorbid snails like B. glabrata, hemoglobin plays the primary role in transporting oxygen and is an abundant hemolymph protein, so it is noteworthy that a four-fold reduction in hemoglobin gene expression was observed. It is clear that snails like lymnaeids that rely on hemocyanin as a respiratory pigment are also affected by niclosamide [7], so it seems unlikely that depressed hemoglobin synthesis per se could be the only mechanism of niclosamide toxicity in B. glabrata. As noted above, early investigations demonstrated that niclosamide affects oxygen uptake in B. glabrata, with higher concentrations resulting in the inhibition of oxygen intake [13]. Downregulation of hemoglobin gene expression may be one reason for decreased oxygen uptake noted in such studies, which in turn could contribute to toxicity.

Direct effects of niclosamide versus indirect toxic effects

Hypothetically, at the organismal level changes in gene expression following exposure to a toxicant can result from direct as well as indirect mechanisms. Specifically, changes can occur in target cells that directly interact with a toxicant, or can occur in cells that instead are responding to non-toxicant molecules (e.g., from necrotic, inflammatory, or neuroendocrine cells) or to toxicant-altered physiological parameters (e.g., hemolymph pH, osmolarity, or O₂ concentration). If concentration increases to a level causing wholesale death of target cells, then we would expect indirect mechanisms to become more pronounced. With increasing concentrations of niclosamide, we observed an increase in the number of both upregulated and downregulated genes, with a predominance of upregulated genes. Some of the changes noted at higher concentrations may indeed be the result of indirect toxic effects However, the known functions of responsive genes and the overall patterns observed are clearly suggestive of the involvement of protective biological processes that are affected by niclosamide. The responsive genes comprise a list of candidate targets for niclosamide and offers leads for future development of novel molluscicides.

Conclusion

Our data suggest that niclosamide alters expression of several genes involved in biotransformation, stress responses, intracellular organelle trafficking and oxygen transport. The aggregate toxicological significance of these altered expression patterns is not clear, but it is reasonable to speculate that some of these transcriptomic responses are adaptive, allowing enhanced elimination of niclosamide and perhaps maintenance of protein and organelle function and repair, whereas others may lead to enhanced toxicity, e.g., by suppressing immune function, lowering oxygen carrying capacity of hemolymph and contributing to mitochondrial damage. Future studies of the effects of niclosamide on snails need to address which of the transcriptomic responses are key to understanding its mechanism of toxicity. Such responses can then serve as useful biomarkers for testing new candidate molluscicides. Also, it will be important to distinguish between direct effects of niclosamide on target cells and indirect effects, and to assess the role of these effects in niclosamide toxicity at low and high concentrations. Finally, genes expressed in snails during recovery from exposure may provide insights on protective detoxification and repair mechanisms that can serve as future targets.

A long-term goal of molluscicide research is to develop chemicals that have more specific effects on targeted snail species, without widespread toxicity to nontarget organisms such as fish, which are killed by niclosamide. Previous studies have shown that mammalian and tapeworm mitochondria differ markedly in their responses to niclosamide, thereby accounting for the selective toxicity of this drug against tapeworms and its low toxicity in mammals [86]. This observation raises the possibility that mitochondrial or other cellular functions in snails might also be different and potentially more vulnerable to more specific molluscicides that leave aquatic vertebrates unaffected. Deciphering such differences with the use of molecular tools should help in the rational design of a next generation of highly specific molluscicides.