Accepted Manuscript Title: Lack of variation in voltage-gated sodium channels of common bottlenose dolphins (Tursiops truncatus) exposed to neurotoxic algal blooms Author: Kristina M. Cammen Patricia E. Rosel Randall S. Wells Andrew J. Read PII: DOI: Reference:
S0166-445X(14)00312-9 http://dx.doi.org/doi:10.1016/j.aquatox.2014.10.010 AQTOX 3949
To appear in:
Aquatic Toxicology
Received date: Revised date: Accepted date:
17-6-2014 8-10-2014 11-10-2014
Please cite this article as: Cammen, K.M., Rosel, P.E., Wells, R.S., Read, A.J.,Lack of variation in voltage-gated sodium channels of common bottlenose dolphins (Tursiops truncatus) exposed to neurotoxic algal blooms, Aquatic Toxicology (2014), http://dx.doi.org/10.1016/j.aquatox.2014.10.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lack of variation in voltage-gated sodium channels of common bottlenose dolphins (Tursiops truncatus) exposed to neurotoxic algal blooms Kristina M. Cammena1, Patricia E. Roselb, Randall S. Wellsc, Andrew J. Reada a
Present address: School of Marine Sciences, University of Maine, Orono, ME 04469
us
1
cr
ip t
Nicholas School of the Environment, Duke University, 135 Duke Marine Lab Road, Beaufort, NC 28516, USA b National Marine Fisheries Service, Southeast Fisheries Science Center, 646 Cajundome Blvd, Lafayette, LA 70506, USA c Chicago Zoological Society, c/o Mote Marine Laboratory, 1600 Ken Thompson Parkway, Sarasota, FL 34236, USA
te
d
M
an
Corresponding author: Kristina Cammen School of Marine Sciences University of Maine Orono, ME 04469 Phone: +1 207-581-4381 Email: [email protected]
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
1
Page 1 of 25
an
us
cr
ip t
Abstract In coastal marine ecosystems, neurotoxins produced by harmful algal blooms (HABs) often result in large-scale mortality events of many marine species. Historical and frequent exposure to HABs therefore may provide a strong selective pressure for adaptations that result in toxin resistance. Neurotoxin resistance has independently evolved in a variety of terrestrial and marine species via mutations in genes encoding the toxin binding sites within the voltage-gated sodium channel gene complex. Accordingly, we tested the hypothesis that genetic variation in the putative binding site of brevetoxins in common bottlenose dolphins (Tursiops truncatus) explains differences among individuals or populations in resistance to harmful Karenia brevis blooms in the Gulf of Mexico. We found very little variation in the sodium channel exons encoding the putative brevetoxin binding site among bottlenose dolphins from central-west Florida and the Florida Panhandle. Our study included samples from several bottlenose dolphin mortality events associated with HABs, but we found no association between genetic variation and survival. We observed a significant effect of geographic region on genetic variation for some sodium channel isoforms, but this can be primarily explained by rare private alleles and is more likely a reflection of regional genetic differentiation than the cause of different levels of HAB resistance between regions. In contrast to many other previously studied neurotoxin-resistant species, we conclude that bottlenose dolphins have not evolved resistance to HABs via mutations in genes encoding the brevetoxin binding site on the voltage-gated sodium channels.
M
Keywords: common bottlenose dolphin; harmful algal bloom; brevetoxin; voltage-gated sodium channel genes
te
d
Abbreviations: HAB, harmful algal bloom; SCN, voltage-gated sodium channel; UME, unusual mortality event Highlights: We investigated common bottlenose dolphin resistance to harmful algal blooms. We sequenced the binding site of brevetoxins on voltage-gated sodium channels. Sodium channel genes are largely conserved across dolphin populations. We found no association between brevetoxin binding sites and dolphin survival.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
2
Page 2 of 25
te
d
M
an
us
cr
ip t
1. Introduction Neurotoxin resistance has independently evolved in a variety of species via mutations in genes encoding the toxin binding sites within the voltage-gated sodium channel (SCN) gene complex. Voltage-gated sodium channels are membrane-spanning pores that play a key role in the generation of action potentials and neuronal depolarization (reviewed in Catterall, 2000; Yu and Catterall, 2003) and are thus a common target of neurotoxins (Catterall, 1980). Neurotoxin binding to sodium channels may block ion transfer (e.g., tetrodotoxin, Narahashi et al., 1964; saxitoxin, Narahashi et al., 1967) or alter voltage-sensitivity and prevent channel inactivation (e.g., brevetoxin, Huang et al., 1984). However, a single amino acid change in an SCN gene can drastically change toxin binding affinity, and such mutations have been observed to induce neurotoxin resistance in a phylogenetically diverse group of organisms (Anderson, 1987; Bricelj et al., 2005; Jost et al., 2008). For example, in systems where organisms produce or sequester neurotoxins as a mechanism of predator deterrence, neurotoxin resistance is a necessity in the prey species (e.g., Kaneko et al., 1997) and a likely outcome in the predator (e.g., Geffeney et al., 2005). Additionally, in the context of human-induced evolution, many insects have evolved resistance to neurotoxic pesticides (reviewed in Dong, 2007). There are numerous examples of neurotoxin-resistant species in marine environments in which harmful algae produce neurotoxins (Anderson et al., 2005). Common algal toxins include domoic acid produced by diatoms in the Pseudo-nitzschia genus (Bates et al., 1989; Bates, 2000), saxitoxins produced by dinoflagellates in the genera Alexandrium, Gymnodinium, and Pyrodinium (Dale and Yentsch, 1978; Harada et al., 1982; Oshima et al., 1987), and brevetoxins produced by dinoflagellates in the Karenia genus (Lin et al., 1981). These toxins become a threat to marine organisms and humans that consume seafood, primarily shellfish, when the harmful algae occur in dense concentrations (>1x103 cells/L). Such harmful algal blooms (HABs) have been naturally occurring for centuries in some areas, but they are increasing in frequency and geographic range worldwide (Hallegraeff, 1993; Van Dolah, 2000). Some of this observed increase in HABs may be attributable to improved surveillance of coastal waters, but in addition, climate change (Edwards et al., 2006), eutrophication (Paerl and Whitall, 1999), and ship ballast water transfer (Hallegraeff, 1998) have likely contributed to a true increase in HABs, which may, in turn, strengthen selective pressure for the evolution of neurotoxin resistance in marine species. HABs of the dinoflagellate Karenia brevis, also known as red tides, occur almost annually along the west coast of Florida (Florida Fish and Wildlife Research Institute, 2008). Karenia brevis produces brevetoxins, a group of lipid-soluble neurotoxic compounds with a cyclic polyether backbone (Baden, 1989, Lin et al., 1981; Golik et al., 1982; Chou and Shimizu, 1982; Shimizu et al., 1986), which bind with high affinity to voltage-gated sodium channels (Poli et al., 1986). Brevetoxin binding alters voltage-sensitivity and prevents channel inactivation, thus inducing initial hyper-excitation of neurons followed by anergy (Huang et al., 1984). In humans, inhalation of aerosolized brevetoxins can cause respiratory irritation (Backer et al., 2003; Pierce et al., 2003), and ingestion of seafood containing brevetoxins can result in neurotoxic shellfish poisoning (NSP) (reviewed in Watkins et al., 2008). Exposed marine species may also exhibit neurological symptoms (e.g., manatee; O’Shea et al., 1991), but given their tendency for greater exposure to K. brevis, brevetoxin exposure in marine species is often lethal. HABs of K. brevis are often associated with large fish kills and mortalities of common bottlenose dolphins, manatees, sea birds, and sea turtles (Bossart et al., 1998; Flewelling et al., 2005; Kreuder et al., 2002; Redlow et al., 2003). Marine organisms may be exposed to brevetoxins directly in the water column (Woofter et al., 2005), through inhalation of aerosolized toxin
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
3
Page 3 of 25
te
d
M
an
us
cr
ip t
(Bossart et al., 1998), or through the ingestion of K. brevis cells or contaminated prey (Flewelling et al., 2005). Brevetoxins can be potently ichthyotoxic, but certain fish are able to survive high concentrations of K. brevis and accumulate brevetoxins (Naar et al., 2007). As brevetoxins biomagnify they become a threat for upper trophic level predators in these coastal ecosystems, such as the common bottlenose dolphin (Tursiops truncatus, hereafter referred to as the bottlenose dolphin). In the past two decades, four large-scale unusual mortality events (UMEs) of bottlenose dolphins in the Gulf of Mexico have been attributed to blooms of K. brevis (Mase et al., 2000; National Marine Fisheries Service, 2004; Twiner et al., 2012). These events were considered UMEs, as defined by the Marine Mammal Protection Act, because the strandings were unexpected, impacted a large portion of the dolphin population, and deviated from the historical average number of strandings in this region (Waring et al., 2013). Each event involved the mortality of over 100 dolphins that stranded dead, often during or shortly after a red tide was observed in the area. Of those tested, many stranded dolphins were positive for brevetoxins at high concentrations in their stomach contents and livers (Twiner et al., 2012). The factors resulting in large-scale dolphin mortality associated with red tides are not well understood. Of particular interest is the observation that red tides occurring in the Florida Panhandle generally result in greater dolphin mortality than those that occur in central-west Florida. Bloom densities and toxicities are similar across the two regions (Pierce et al., 2008; Twiner et al., 2011; Twiner et al., 2012), and previous studies of brevetoxin accumulation in lower trophic levels suggest little difference in bottlenose dolphin exposure to the HAB toxins. Similar concentrations of brevetoxin were found in fish sampled during HABs in both Sarasota Bay in central-west Florida (Fire et al., 2008a; Twiner et al., 2011) and St. Joseph Bay in the Florida Panhandle (Flewelling et al., 2005; Naar et al., 2007), and the stomach contents of stranded dolphins from both areas contained fish with comparable levels of brevetoxins (Fire et al., 2007; Flewelling et al., 2005; Twiner et al., 2012). Despite these similarities in HAB characteristics, red tides appear to have different impacts on dolphin populations in the two regions. The dolphins in these two regions belong to multiple genetically differentiated populations (Waring et al., 2013). Throughout the Gulf of Mexico, estuarine bottlenose dolphins that inhabit semi-enclosed bays, sounds, and estuaries exhibit high site fidelity, which in some cases has been recorded across decades (Balmer et al., 2008; Hubard et al., 2004; Maze and Würsig, 1999; Scott et al., 1990). In addition, coastal dolphin populations that are larger in size and home range (Waring et al., 2013) inhabit waters out to roughly the 20 m isobath. Social mixing among estuarine and coastal dolphin populations has been observed (Fazioli et al., 2006; Wells, 1986), but genetic data suggest that little interbreeding occurs both between estuarine and coastal populations and among estuarine populations (Sellas et al., 2005). Limited gene flow among dolphin populations, likely as a result of strong social structure and site fidelity (Scott et al., 1990), has resulted in significant genetic differentiation at a fine geographic scale (Sellas et al., 2005). The metapopulation structure of bottlenose dolphins in the Gulf of Mexico creates an environment where individual populations may be particularly affected by localized threats, such as HABs, but also where populations may be able to locally adapt (Kawecki and Ebert 2004). Bottlenose dolphins in central-west Florida are frequently exposed to red tides, which have occurred in this region on a regular basis for centuries (Florida Fish and Wildlife Research Institute, 2008). In contrast, HABs have occurred relatively infrequently in the Florida Panhandle
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
4
Page 4 of 25
us
cr
ip t
(Steidinger et al., 1998). This difference in the strength of selective pressure could lead to local adaptive divergence between the populations, more strongly favoring genotypes that confer HAB resistance in the central-west Florida dolphin populations. Local adaptation to neurotoxins has been previously described in the cases of soft shell clam resistance to algal-produced saxitoxins (Bricelj et al., 2004; Connell et al., 2007) and garter snake resistance to toxic newts (Brodie III and Brodie Jr, 1991; Brodie Jr et al., 2002). In both cases, resistance is the result of a mutation in the genes encoding the toxin binding site in the voltage-gated sodium channels (Bricelj et al., 2005; Geffeney et al., 2005). Our objective was therefore to characterize genetic variation at all putative brevetoxin binding sites in bottlenose dolphins and test the hypothesis that variation in the voltage-gated sodium channel genes explains apparent differences among individuals or populations in resistance to brevetoxins produced by HABs of K. brevis.
te
d
M
an
2. Material and Methods 2.1. Samples Our study included DNA samples from the following groups of bottlenose dolphins: live dolphins from the Florida Panhandle, including the estuarine population found primarily in St. Joseph Bay (N=18) and the adjacent coastal population (N=19); live dolphins from central-west Florida, including the estuarine population found in Sarasota Bay (N=27) and the adjacent coastal population (N=28); dolphins that died during UMEs associated with red tides in the Florida Panhandle in 1999 (N=18) and 2004 (N=37); and dolphins that died during red tides in central-west Florida between 1992 and 2006 (N=27, including 18 dolphins that died during the 2005-2006 UME). All live dolphins were assumed to have survived at least one red tide based on their age and geographic distribution, and samples were categorized as coastal or estuarine based on photo-ID site fidelity (B. Balmer, personal communication; R. Wells, unpublished data). Dolphins that died during non-UME red tides were included if toxin analyses suggested brevetoxicosis as the putative cause of death (Fauquier et al., 2007; Fire et al., 2007). We also included eight skin samples from coastal bottlenose dolphins that stranded in North Carolina between 2002 and 2006 to represent an outgroup that has rarely, if ever, been exposed to red tides. We acquired dolphin skin and genomic DNA samples from the Chicago Zoological Society’s Sarasota Dolphin Research Program, the Mote Marine Laboratory’s Stranding Investigations Program, the North Carolina Marine Mammal Stranding Network, and the NOAA SEFSC Marine Mammal Tissue and DNA Archive. DNA was extracted from skin samples using the Wizard® Genomic DNA Purification Kit (Promega) or phenol-chloroform extraction, as described by Rosel and Block (1996) but using 250 μl volumes and Phase Lock gel tubes (5 Prime). The quality of genomic DNA was determined visually on agarose gels and quantified fluorometrically.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
2.2. Genetic analyses Brevetoxins bind to site 5 on the alpha subunit of voltage-gated sodium channels (SCNA) (Catterall and Gainer, 1985; Poli et al., 1986), which are composed of four homologous domains, each containing six transmembrane segments (Fig. 1) (Noda et al., 1984; Sato et al., 2001). The brevetoxin binding site is located on transmembrane segments in two domains (DIS6 and DIVS5) encoded by two SCNA exons (Trainer et al., 1994). We investigated genetic variation at these two exons in nine structurally different sodium channel isoforms, which exhibit unique patterns of tissue expression (Goldin, 1999). Some studies have shown tissue selectivity in
5
Page 5 of 25
te
d
M
an
us
cr
ip t
brevetoxin binding (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006), but others have suggested that the evolution of resistance to neurotoxins takes place across all types of voltage-gated sodium channels (Jost et al., 2008). We identified SCNA genes using the annotated Ensembl bottlenose dolphin genome (turTru1, release 71, GenBank Assembly ID: GCA_000151865.1). In general, SCNA is encoded by 26 exons; however, the number of exons differs slightly between isoforms, and the presence of nonconventional introns (Wu and Krainer, 1999) can confuse traditional algorithms for exon identification (Widmark et al., 2011). Here, we use the exon numbers as reported in NCBI AceView (Thierry-Mieg and Thierry-Mieg, 2006) and the following literature (Ahmed et al., 1992; Chen et al., 2000; Dib-Hajj et al., 1999; Escayg et al., 2000; George Jr et al., 1993; Klugbauer et al., 1995; Plummer et al., 1998; Rabert et al., 1998; Wang et al., 1996). We designed primers (Table 1) to amplify and sequence the complete exon containing DIS6 (exon 9, 10, or 11; 198-210 bp) and a 171-186 bp fragment of the exon encoding DIVS4 and DIVS5 (exon 24, 26, 27, or 28) (Fig. 1). All PCR reactions were conducted in a final volume of 20 μl and included 12-50 ng DNA, 1x PCR buffer (20 mM Tris pH 8.8, 50 mM KCl, 0.1% Triton X-100, 0.2 mg/mL BSA NEB purified), 2-4 mM MgCl2, 0.2 mM dNTPs, 0.25 μM of each primer, and 1 unit Taq polymerase (Bioline). Only PCRs for SCN1A included 4 mM MgCl2. All PCRs were run using a touchdown temperature profile consisting of an initial step of 94°C for 3 min, followed by 30 cycles of 94°C for 15 sec, 70°C for 15 sec decreasing by 0.5°C each cycle, and 72°C for 30 sec, followed by 25 cycles with a 55°C annealing temperature, and a final step of 72°C for 5 min. 10 μl of all successful PCR products were enzymatically purified using 1 unit Exonuclease I and 0.25 units Antarctic Phosphatase (New England Biolabs) incubated at 37°C for 1 hr. Purified PCR products were sequenced in both directions. BigDye sequencing reactions were conducted in a final volume of 10 μl including 1-5.2 μl purified PCR product, 1 μM primer, 1x BigDye Buffer (40 mM Tris pH 9, 5 mM MgCl2, 1.5 M Betaine), 0.44 μl BDX64 (MCLAB), and 0.06 μl BigDye v1.1 (Duke IGSP). The sequencing profile consisted of an initial step of 94°C for 30 sec, followed by 25 cycles of 94°C for 15 sec, 50°C for 15 sec, and 60°C for 3 min. Sequencing products were purified using AmpPure magnetic beads (Agencourt) and sequenced on an Applied Biosystems ABI 3730xl automated sequencer. Resistance-inducing mutations in the voltage-gated sodium channel genes are expected to be non-synonymous mutations that result in an amino acid residue substitution and corresponding change in protein structure. Therefore, we searched resultant sequence data for polymorphisms, which we identified as two chromatogram peaks in a single position, clearly observed in both forward and reverse sequences. We manually edited and aligned sequences in CodonCode Aligner v3 (CodonCode Corporation). Genetic variation was initially analyzed in a sample set of 70 dolphins, including 8-10 samples from each of the eight groups. If nonsynonymous mutations were detected in this initial sample set, genetic variation was further analyzed in the total sample set of 182 individuals. All unique alleles that were not observed in multiple homozygous individuals were confirmed by sequencing 16-24 clones produced using pGEM®-T Easy Vectors with JM109 competent E. coli cells (Promega) following manufacturer instructions. Alleles were confirmed only if observed in multiple clones. We conducted phylogenetic analyses in MEGA v5.2 (Tamura et al., 2011) using the maximum likelihood method with the Kimura 2-parameter model (Kimura, 1980) with invariable sites (I = 0.409), the evolutionary model that best described the nucleotide substitution pattern (maximum likelihood fit with lowest Bayesian information criterion). For this analysis, we
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
6
Page 6 of 25
M
an
us
cr
ip t
combined nucleotide sequences from the DIS6-containing exon (9, 10, or 11) and the portion of the exon (24, 26, 27, or 28) encoding DIVS4-S5. For variable isoforms, we included only the most common variant. We tested for significant effects of geographic region (central-west Florida vs. Panhandle) and survival (live vs. dead) on SCNA allele frequency using multifactorial logistic models implemented in JMP Pro 11.0.0 (SAS Institute Inc., 2014). We assessed the power of this study to detect an association between genetic variation and survival using a case/control power analysis conducted in PGA (Menashe et al., 2008). Specifically, we calculated the minimum minor allele frequency given the total sample size (94 controls, 83 cases) that would provide 80% power to detect a relative risk of 2 if the prevalence of disease is between 0.01 and 0.20. Power analyses often test for the ability of a study to detect a two-fold increase in risk (Menashe et al., 2008); accordingly, we tested our ability to detect a relative risk of 2, the point at which a dolphin with a “susceptible” allele is twice as likely to die due to brevetoxin exposure than a dolphin that does not have the “susceptible” allele. We tested our power across a range of disease prevalence values because of the current lack of understanding of brevetoxicosis disease etiology and prevalence in the bottlenose dolphin populations. The range of possible disease prevalence values was estimated using the size of the studied dolphin populations (Waring et al., 2013) and the number of dolphins that died during UMEs associated with red tides (J. Litz, personal communication; Mase et al., 2000; National Marine Fisheries Service, 2004; Twiner et al., 2012), recognizing that only a portion of marine mammal carcasses are recovered (Williams et al., 2011).
te
d
3. Results and Discussion 3.1. Sodium channel isoforms We characterized genetic variation of the putative brevetoxin binding site in nine sodium channel isoforms in common bottlenose dolphins and found that variation among isoforms conformed to expectations based on the general history of SCN gene duplication. Historical gene duplications in tetrapods resulted in four groups of SCNs: 1, 2, 3, and 9; 5, 10, and 11; 4; and 8 (Widmark et al., 2011). This grouping is generally reflected in the maximum likelihood tree built using only 393 bp of each bottlenose dolphin SCNA isoform (Fig. 2). There was no evidence to suggest any of the sequences represented pseudogenes (i.e., no translated sequences contained stop codons).
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
3.2. Sodium channel variation Of the 18 SCNA exons sequenced, only two (SCN5A_exon10 and SCN11A_exon26, Table 2) contained a non-synonymous mutation in the initially screened sample set of 70 dolphins. SCN5 and SCN11 belong to the same SCN cluster (Widmark et al., 2011) and both are considered tetrodotoxin-resistant (Renaud et al., 1983; Dib-Hajj et al., 1998). Brevetoxins have been experimentally demonstrated to target the cardiac sodium channels (SCN5) (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006); however, type B brevetoxins (PbTx-2 and -3), which are the most common toxin type observed during red tides (Pierce et al., 2008), show lower relative affinity for cardiac sodium channels in comparison to skeletal (SCN4) sodium channels (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006). Studies have not yet tested brevetoxin binding affinity to SCN11, which is found primarily in dorsal root ganglion neurons (Dib-Hajj et al., 1998).
7
Page 7 of 25
us
cr
ip t
Neither non-synonymous mutation occurred in a transmembrane spanning region where brevetoxin putatively binds (DIS6 or DIVS5, Trainer et al., 1994). The non-synonymous mutation in SCN5A occurred in the intracellular loop between domains I and II, and the nonsynonymous mutation in SCN11A occurred in DIVS4 (Fig. 3). The fourth transmembrane segment of each domain regulates sodium channel voltage-sensitivity; however, this voltagesensitivity is primarily attributed to positively charged residues, mainly arginines (Catterall, 1986; Guy and Seetharamulu, 1986), and the observed non-synonymous mutation in DIVS4 involved phenylalanine and leucine, neither of which are positively charged. Five other SCNA exons contained synonymous mutations (Table 2, Fig. 3). The most frequently variable genes, SCN1A and SCN2A, belong to the same SCN cluster (Widmark et al., 2011), are broadly expressed in neurons in the central and peripheral nervous systems (Beckh, 1990; Noda et al., 1986a), and are highly tetrodotoxin-sensitive (Noda et al., 1986b). The synonymous mutations found in these SCNA exons may be linked to protein-changing mutations elsewhere in the sodium channel genes, beyond the putative brevetoxin binding site evaluated in our study.
te
d
M
an
3.3. Effects of geographic region and survival Genetic variation was rare among the sampled bottlenose dolphins, even in SCNA exons with polymorphisms (Table 2). Two minor alleles (SCN3A_exon11 and SCN8A_exon10) were only observed once in the initial sample set of Florida dolphins, and thus can have little effect on the observed differences in survival among individuals. Given the near parity of the number of live (34) and dead (28) Florida samples analyzed, we would expect to find alleles strongly associated with survival at closer to 50% frequency. A case/control power analysis suggests this study has 80% power to detect a relative risk of 2 with marker allele frequencies of 0.09-0.18. Thus, for at least SCN1A_exon9 and SCN2A_exon26, the most frequently variable SCNA exons, there should be an 80% probability of detecting an association with survival if there is one. However, we observed no significant effects of survival or the region:survival interaction term on SCNA allele frequency (Table 2). Allele frequencies varied significantly between geographic regions (central-west Florida vs. Panhandle) in three cases (SCN5A_exon10, χ2 = 12.53, p < 0.001; SCN2A_exon26, χ2 = 13.41, p < 0.01; SCN5A_exon28, χ2 = 6.69, p < 0.01, Table 2). In both SCN5A exons the difference between regions was due to rare private alleles observed only in the Panhandle, but these differences could not be attributed to linked variation between the two exons as the private alleles were not observed in the same individuals. As described above, such rare variation is unlikely to explain the difference in HAB resistance between regions. Rather, the difference in frequency of rare SCNA alleles is more likely a reflection of limited gene flow between geographic regions due to the tendency for high site fidelity in bottlenose dolphins in inshore waters of the Gulf of Mexico (Balmer et al., 2008; Hubard et al., 2004; Maze and Würsig, 1999; Scott et al., 1990). Long-term site fidelity can result in significant genetic differentiation among dolphin populations at a fine geographic scale (Sellas et al., 2005). Greater genetic differentiation is expected between dolphins from the Gulf of Mexico and the western North Atlantic (Rosel et al., 2009). However, we found very few differences in genetic variation across SCN genes between the dolphins from Florida and North Carolina. No exonic variation was observed uniquely in North Carolina, though we observed a single mutation in the intronic region upstream of SCN9A_exon10 only in North Carolina samples. SCNA minor alleles characterized in dolphins from Florida were almost never observed in dolphins from
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
8
Page 8 of 25
ip t
North Carolina (SCN8A_exon10 minor allele was observed in a single North Carolina sample), although we acknowledge that our sample size from North Carolina was small. Dolphins in North Carolina represent an outgroup that has rarely, if ever, been exposed to red tides. Yet, our results suggest similar selection pressures acting to conserve the existing voltage-gated sodium channel genetic structure across these geographically disjunct and genetically differentiated bottlenose dolphin populations.
te
d
M
an
us
cr
3.4. Discussion of lack of survival-associated SCN variation Highly conserved SCNA sequences across bottlenose dolphins along the Gulf coast of Florida likely reflect the functional importance of the voltage-gated sodium channels. In other cases of neurotoxin resistance, fitness tradeoffs are observed in conjunction with SCN amino acid substitutions (Brodie III and Brodie Jr, 1999). These costs may explain why resistanceinducing mutations are only observed in natural populations regularly exposed to neurotoxins (Brodie Jr et al., 2002; Connell et al., 2007). These costs may also explain why we do not observe SCNA mutations associated with brevetoxin resistance in bottlenose dolphins. Any change in the voltage-gated sodium channels that affects muscular control may be too costly for this species, or the fitness advantage may not outweigh the cost of the mutation. Our findings may be explained by a difference in the strength or constancy of selective pressure between the dolphin-brevetoxin system and other studied systems that have evolved neurotoxin resistance. The strength of selection for brevetoxin resistance in bottlenose dolphins may be reduced by their trophic separation from the toxin-producing organism, their long generation time and the episodic nature of HABs, and their ability to behaviorally avoid exposure. In the well-studied case of garter snakes, predators have evolved neurotoxin resistance to prey that sequester toxins as a defense mechanism (Brodie III and Brodie Jr, 1991). However, bottlenose dolphin prey do not sequester toxins as a defense mechanism, and dolphins are separated by at least one trophic level from the toxin-producing algae. Additionally, in contrast to consistently toxic prey, HABs are episodic events, though low levels of toxin can be detected in the prey fish months after the termination of a bloom (Fire et al., 2008a). Locally adapted resistance to episodic HABs has been demonstrated in other marine species (e.g., softshell clam, Connell et al., 2007; marine copepod, Colin and Dam, 2004), but these species accumulate toxin directly from the harmful algae, have shorter generation times, and are less mobile, all factors which may influence the strength of selection for algal toxin resistance. Individual or population-level variation in bottlenose dolphin resistance to HABs may alternatively be explained by non-genetic hypotheses. For example, bottlenose dolphins may behaviorally adapt to the presence of HABs such that they minimize direct (e.g., avoid areas of dense blooms) and indirect (e.g., alter foraging strategies) toxin exposure. Although highly mobile, dolphins do not entirely avoid HAB toxin exposure, as evidenced by behavioral observations during HABs (McHugh et al., 2011) and toxin concentrations measured in live bottlenose dolphins (Fire et al., 2008b). However, dolphins have been observed to alter both the size of their home range and their foraging behavior during HABs, likely in response to changes in prey abundance (Bowen, 2011; McHugh et al., 2011; Powell and Wells, 2011). Certain foraging behaviors may increase the likelihood of exposure to HABs; for example, eating discarded bycatch from commercial fishing vessels provides an avenue of exposure to toxins that originate in distant blooms (National Marine Fisheries Service, 2004). The immune system may also play a role in brevetoxin resistance. Brevetoxin exposure can cause immunosuppression in marine species (Bossart et al., 1998; Kreuder et al., 2002;
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
9
Page 9 of 25
an
us
cr
ip t
Walsh et al., 2005; Walsh et al., 2010), but the toxin’s antigenicity and ability to induce an adaptive immune response is poorly understood. If antigenic, chronic low levels of toxin exposure may result in immunity, though alternatively, chronic low-level exposure can increase toxin sensitivity (e.g., domoic acid sensitivity in zebrafish, Lefebvre et al., 2012). In central-west Florida, low levels of brevetoxins have been measured in dolphins when HABs are not present (Fire et al., 2008b). These low levels of toxins may be explained by long-term retention of the toxin or by prolonged exposure to continual sources of toxin, such as prey species that store toxin (Fire et al., 2008a) or toxin adsorbed to plant material (Flewelling et al., 2005). Further research is needed to better understand the effects of chronic exposure to background levels (
S0166-445X(14)00312-9 http://dx.doi.org/doi:10.1016/j.aquatox.2014.10.010 AQTOX 3949
To appear in:
Aquatic Toxicology
Received date: Revised date: Accepted date:
17-6-2014 8-10-2014 11-10-2014
Please cite this article as: Cammen, K.M., Rosel, P.E., Wells, R.S., Read, A.J.,Lack of variation in voltage-gated sodium channels of common bottlenose dolphins (Tursiops truncatus) exposed to neurotoxic algal blooms, Aquatic Toxicology (2014), http://dx.doi.org/10.1016/j.aquatox.2014.10.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lack of variation in voltage-gated sodium channels of common bottlenose dolphins (Tursiops truncatus) exposed to neurotoxic algal blooms Kristina M. Cammena1, Patricia E. Roselb, Randall S. Wellsc, Andrew J. Reada a
Present address: School of Marine Sciences, University of Maine, Orono, ME 04469
us
1
cr
ip t
Nicholas School of the Environment, Duke University, 135 Duke Marine Lab Road, Beaufort, NC 28516, USA b National Marine Fisheries Service, Southeast Fisheries Science Center, 646 Cajundome Blvd, Lafayette, LA 70506, USA c Chicago Zoological Society, c/o Mote Marine Laboratory, 1600 Ken Thompson Parkway, Sarasota, FL 34236, USA
te
d
M
an
Corresponding author: Kristina Cammen School of Marine Sciences University of Maine Orono, ME 04469 Phone: +1 207-581-4381 Email: [email protected]
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
1
Page 1 of 25
an
us
cr
ip t
Abstract In coastal marine ecosystems, neurotoxins produced by harmful algal blooms (HABs) often result in large-scale mortality events of many marine species. Historical and frequent exposure to HABs therefore may provide a strong selective pressure for adaptations that result in toxin resistance. Neurotoxin resistance has independently evolved in a variety of terrestrial and marine species via mutations in genes encoding the toxin binding sites within the voltage-gated sodium channel gene complex. Accordingly, we tested the hypothesis that genetic variation in the putative binding site of brevetoxins in common bottlenose dolphins (Tursiops truncatus) explains differences among individuals or populations in resistance to harmful Karenia brevis blooms in the Gulf of Mexico. We found very little variation in the sodium channel exons encoding the putative brevetoxin binding site among bottlenose dolphins from central-west Florida and the Florida Panhandle. Our study included samples from several bottlenose dolphin mortality events associated with HABs, but we found no association between genetic variation and survival. We observed a significant effect of geographic region on genetic variation for some sodium channel isoforms, but this can be primarily explained by rare private alleles and is more likely a reflection of regional genetic differentiation than the cause of different levels of HAB resistance between regions. In contrast to many other previously studied neurotoxin-resistant species, we conclude that bottlenose dolphins have not evolved resistance to HABs via mutations in genes encoding the brevetoxin binding site on the voltage-gated sodium channels.
M
Keywords: common bottlenose dolphin; harmful algal bloom; brevetoxin; voltage-gated sodium channel genes
te
d
Abbreviations: HAB, harmful algal bloom; SCN, voltage-gated sodium channel; UME, unusual mortality event Highlights: We investigated common bottlenose dolphin resistance to harmful algal blooms. We sequenced the binding site of brevetoxins on voltage-gated sodium channels. Sodium channel genes are largely conserved across dolphin populations. We found no association between brevetoxin binding sites and dolphin survival.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
2
Page 2 of 25
te
d
M
an
us
cr
ip t
1. Introduction Neurotoxin resistance has independently evolved in a variety of species via mutations in genes encoding the toxin binding sites within the voltage-gated sodium channel (SCN) gene complex. Voltage-gated sodium channels are membrane-spanning pores that play a key role in the generation of action potentials and neuronal depolarization (reviewed in Catterall, 2000; Yu and Catterall, 2003) and are thus a common target of neurotoxins (Catterall, 1980). Neurotoxin binding to sodium channels may block ion transfer (e.g., tetrodotoxin, Narahashi et al., 1964; saxitoxin, Narahashi et al., 1967) or alter voltage-sensitivity and prevent channel inactivation (e.g., brevetoxin, Huang et al., 1984). However, a single amino acid change in an SCN gene can drastically change toxin binding affinity, and such mutations have been observed to induce neurotoxin resistance in a phylogenetically diverse group of organisms (Anderson, 1987; Bricelj et al., 2005; Jost et al., 2008). For example, in systems where organisms produce or sequester neurotoxins as a mechanism of predator deterrence, neurotoxin resistance is a necessity in the prey species (e.g., Kaneko et al., 1997) and a likely outcome in the predator (e.g., Geffeney et al., 2005). Additionally, in the context of human-induced evolution, many insects have evolved resistance to neurotoxic pesticides (reviewed in Dong, 2007). There are numerous examples of neurotoxin-resistant species in marine environments in which harmful algae produce neurotoxins (Anderson et al., 2005). Common algal toxins include domoic acid produced by diatoms in the Pseudo-nitzschia genus (Bates et al., 1989; Bates, 2000), saxitoxins produced by dinoflagellates in the genera Alexandrium, Gymnodinium, and Pyrodinium (Dale and Yentsch, 1978; Harada et al., 1982; Oshima et al., 1987), and brevetoxins produced by dinoflagellates in the Karenia genus (Lin et al., 1981). These toxins become a threat to marine organisms and humans that consume seafood, primarily shellfish, when the harmful algae occur in dense concentrations (>1x103 cells/L). Such harmful algal blooms (HABs) have been naturally occurring for centuries in some areas, but they are increasing in frequency and geographic range worldwide (Hallegraeff, 1993; Van Dolah, 2000). Some of this observed increase in HABs may be attributable to improved surveillance of coastal waters, but in addition, climate change (Edwards et al., 2006), eutrophication (Paerl and Whitall, 1999), and ship ballast water transfer (Hallegraeff, 1998) have likely contributed to a true increase in HABs, which may, in turn, strengthen selective pressure for the evolution of neurotoxin resistance in marine species. HABs of the dinoflagellate Karenia brevis, also known as red tides, occur almost annually along the west coast of Florida (Florida Fish and Wildlife Research Institute, 2008). Karenia brevis produces brevetoxins, a group of lipid-soluble neurotoxic compounds with a cyclic polyether backbone (Baden, 1989, Lin et al., 1981; Golik et al., 1982; Chou and Shimizu, 1982; Shimizu et al., 1986), which bind with high affinity to voltage-gated sodium channels (Poli et al., 1986). Brevetoxin binding alters voltage-sensitivity and prevents channel inactivation, thus inducing initial hyper-excitation of neurons followed by anergy (Huang et al., 1984). In humans, inhalation of aerosolized brevetoxins can cause respiratory irritation (Backer et al., 2003; Pierce et al., 2003), and ingestion of seafood containing brevetoxins can result in neurotoxic shellfish poisoning (NSP) (reviewed in Watkins et al., 2008). Exposed marine species may also exhibit neurological symptoms (e.g., manatee; O’Shea et al., 1991), but given their tendency for greater exposure to K. brevis, brevetoxin exposure in marine species is often lethal. HABs of K. brevis are often associated with large fish kills and mortalities of common bottlenose dolphins, manatees, sea birds, and sea turtles (Bossart et al., 1998; Flewelling et al., 2005; Kreuder et al., 2002; Redlow et al., 2003). Marine organisms may be exposed to brevetoxins directly in the water column (Woofter et al., 2005), through inhalation of aerosolized toxin
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
3
Page 3 of 25
te
d
M
an
us
cr
ip t
(Bossart et al., 1998), or through the ingestion of K. brevis cells or contaminated prey (Flewelling et al., 2005). Brevetoxins can be potently ichthyotoxic, but certain fish are able to survive high concentrations of K. brevis and accumulate brevetoxins (Naar et al., 2007). As brevetoxins biomagnify they become a threat for upper trophic level predators in these coastal ecosystems, such as the common bottlenose dolphin (Tursiops truncatus, hereafter referred to as the bottlenose dolphin). In the past two decades, four large-scale unusual mortality events (UMEs) of bottlenose dolphins in the Gulf of Mexico have been attributed to blooms of K. brevis (Mase et al., 2000; National Marine Fisheries Service, 2004; Twiner et al., 2012). These events were considered UMEs, as defined by the Marine Mammal Protection Act, because the strandings were unexpected, impacted a large portion of the dolphin population, and deviated from the historical average number of strandings in this region (Waring et al., 2013). Each event involved the mortality of over 100 dolphins that stranded dead, often during or shortly after a red tide was observed in the area. Of those tested, many stranded dolphins were positive for brevetoxins at high concentrations in their stomach contents and livers (Twiner et al., 2012). The factors resulting in large-scale dolphin mortality associated with red tides are not well understood. Of particular interest is the observation that red tides occurring in the Florida Panhandle generally result in greater dolphin mortality than those that occur in central-west Florida. Bloom densities and toxicities are similar across the two regions (Pierce et al., 2008; Twiner et al., 2011; Twiner et al., 2012), and previous studies of brevetoxin accumulation in lower trophic levels suggest little difference in bottlenose dolphin exposure to the HAB toxins. Similar concentrations of brevetoxin were found in fish sampled during HABs in both Sarasota Bay in central-west Florida (Fire et al., 2008a; Twiner et al., 2011) and St. Joseph Bay in the Florida Panhandle (Flewelling et al., 2005; Naar et al., 2007), and the stomach contents of stranded dolphins from both areas contained fish with comparable levels of brevetoxins (Fire et al., 2007; Flewelling et al., 2005; Twiner et al., 2012). Despite these similarities in HAB characteristics, red tides appear to have different impacts on dolphin populations in the two regions. The dolphins in these two regions belong to multiple genetically differentiated populations (Waring et al., 2013). Throughout the Gulf of Mexico, estuarine bottlenose dolphins that inhabit semi-enclosed bays, sounds, and estuaries exhibit high site fidelity, which in some cases has been recorded across decades (Balmer et al., 2008; Hubard et al., 2004; Maze and Würsig, 1999; Scott et al., 1990). In addition, coastal dolphin populations that are larger in size and home range (Waring et al., 2013) inhabit waters out to roughly the 20 m isobath. Social mixing among estuarine and coastal dolphin populations has been observed (Fazioli et al., 2006; Wells, 1986), but genetic data suggest that little interbreeding occurs both between estuarine and coastal populations and among estuarine populations (Sellas et al., 2005). Limited gene flow among dolphin populations, likely as a result of strong social structure and site fidelity (Scott et al., 1990), has resulted in significant genetic differentiation at a fine geographic scale (Sellas et al., 2005). The metapopulation structure of bottlenose dolphins in the Gulf of Mexico creates an environment where individual populations may be particularly affected by localized threats, such as HABs, but also where populations may be able to locally adapt (Kawecki and Ebert 2004). Bottlenose dolphins in central-west Florida are frequently exposed to red tides, which have occurred in this region on a regular basis for centuries (Florida Fish and Wildlife Research Institute, 2008). In contrast, HABs have occurred relatively infrequently in the Florida Panhandle
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
4
Page 4 of 25
us
cr
ip t
(Steidinger et al., 1998). This difference in the strength of selective pressure could lead to local adaptive divergence between the populations, more strongly favoring genotypes that confer HAB resistance in the central-west Florida dolphin populations. Local adaptation to neurotoxins has been previously described in the cases of soft shell clam resistance to algal-produced saxitoxins (Bricelj et al., 2004; Connell et al., 2007) and garter snake resistance to toxic newts (Brodie III and Brodie Jr, 1991; Brodie Jr et al., 2002). In both cases, resistance is the result of a mutation in the genes encoding the toxin binding site in the voltage-gated sodium channels (Bricelj et al., 2005; Geffeney et al., 2005). Our objective was therefore to characterize genetic variation at all putative brevetoxin binding sites in bottlenose dolphins and test the hypothesis that variation in the voltage-gated sodium channel genes explains apparent differences among individuals or populations in resistance to brevetoxins produced by HABs of K. brevis.
te
d
M
an
2. Material and Methods 2.1. Samples Our study included DNA samples from the following groups of bottlenose dolphins: live dolphins from the Florida Panhandle, including the estuarine population found primarily in St. Joseph Bay (N=18) and the adjacent coastal population (N=19); live dolphins from central-west Florida, including the estuarine population found in Sarasota Bay (N=27) and the adjacent coastal population (N=28); dolphins that died during UMEs associated with red tides in the Florida Panhandle in 1999 (N=18) and 2004 (N=37); and dolphins that died during red tides in central-west Florida between 1992 and 2006 (N=27, including 18 dolphins that died during the 2005-2006 UME). All live dolphins were assumed to have survived at least one red tide based on their age and geographic distribution, and samples were categorized as coastal or estuarine based on photo-ID site fidelity (B. Balmer, personal communication; R. Wells, unpublished data). Dolphins that died during non-UME red tides were included if toxin analyses suggested brevetoxicosis as the putative cause of death (Fauquier et al., 2007; Fire et al., 2007). We also included eight skin samples from coastal bottlenose dolphins that stranded in North Carolina between 2002 and 2006 to represent an outgroup that has rarely, if ever, been exposed to red tides. We acquired dolphin skin and genomic DNA samples from the Chicago Zoological Society’s Sarasota Dolphin Research Program, the Mote Marine Laboratory’s Stranding Investigations Program, the North Carolina Marine Mammal Stranding Network, and the NOAA SEFSC Marine Mammal Tissue and DNA Archive. DNA was extracted from skin samples using the Wizard® Genomic DNA Purification Kit (Promega) or phenol-chloroform extraction, as described by Rosel and Block (1996) but using 250 μl volumes and Phase Lock gel tubes (5 Prime). The quality of genomic DNA was determined visually on agarose gels and quantified fluorometrically.
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
2.2. Genetic analyses Brevetoxins bind to site 5 on the alpha subunit of voltage-gated sodium channels (SCNA) (Catterall and Gainer, 1985; Poli et al., 1986), which are composed of four homologous domains, each containing six transmembrane segments (Fig. 1) (Noda et al., 1984; Sato et al., 2001). The brevetoxin binding site is located on transmembrane segments in two domains (DIS6 and DIVS5) encoded by two SCNA exons (Trainer et al., 1994). We investigated genetic variation at these two exons in nine structurally different sodium channel isoforms, which exhibit unique patterns of tissue expression (Goldin, 1999). Some studies have shown tissue selectivity in
5
Page 5 of 25
te
d
M
an
us
cr
ip t
brevetoxin binding (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006), but others have suggested that the evolution of resistance to neurotoxins takes place across all types of voltage-gated sodium channels (Jost et al., 2008). We identified SCNA genes using the annotated Ensembl bottlenose dolphin genome (turTru1, release 71, GenBank Assembly ID: GCA_000151865.1). In general, SCNA is encoded by 26 exons; however, the number of exons differs slightly between isoforms, and the presence of nonconventional introns (Wu and Krainer, 1999) can confuse traditional algorithms for exon identification (Widmark et al., 2011). Here, we use the exon numbers as reported in NCBI AceView (Thierry-Mieg and Thierry-Mieg, 2006) and the following literature (Ahmed et al., 1992; Chen et al., 2000; Dib-Hajj et al., 1999; Escayg et al., 2000; George Jr et al., 1993; Klugbauer et al., 1995; Plummer et al., 1998; Rabert et al., 1998; Wang et al., 1996). We designed primers (Table 1) to amplify and sequence the complete exon containing DIS6 (exon 9, 10, or 11; 198-210 bp) and a 171-186 bp fragment of the exon encoding DIVS4 and DIVS5 (exon 24, 26, 27, or 28) (Fig. 1). All PCR reactions were conducted in a final volume of 20 μl and included 12-50 ng DNA, 1x PCR buffer (20 mM Tris pH 8.8, 50 mM KCl, 0.1% Triton X-100, 0.2 mg/mL BSA NEB purified), 2-4 mM MgCl2, 0.2 mM dNTPs, 0.25 μM of each primer, and 1 unit Taq polymerase (Bioline). Only PCRs for SCN1A included 4 mM MgCl2. All PCRs were run using a touchdown temperature profile consisting of an initial step of 94°C for 3 min, followed by 30 cycles of 94°C for 15 sec, 70°C for 15 sec decreasing by 0.5°C each cycle, and 72°C for 30 sec, followed by 25 cycles with a 55°C annealing temperature, and a final step of 72°C for 5 min. 10 μl of all successful PCR products were enzymatically purified using 1 unit Exonuclease I and 0.25 units Antarctic Phosphatase (New England Biolabs) incubated at 37°C for 1 hr. Purified PCR products were sequenced in both directions. BigDye sequencing reactions were conducted in a final volume of 10 μl including 1-5.2 μl purified PCR product, 1 μM primer, 1x BigDye Buffer (40 mM Tris pH 9, 5 mM MgCl2, 1.5 M Betaine), 0.44 μl BDX64 (MCLAB), and 0.06 μl BigDye v1.1 (Duke IGSP). The sequencing profile consisted of an initial step of 94°C for 30 sec, followed by 25 cycles of 94°C for 15 sec, 50°C for 15 sec, and 60°C for 3 min. Sequencing products were purified using AmpPure magnetic beads (Agencourt) and sequenced on an Applied Biosystems ABI 3730xl automated sequencer. Resistance-inducing mutations in the voltage-gated sodium channel genes are expected to be non-synonymous mutations that result in an amino acid residue substitution and corresponding change in protein structure. Therefore, we searched resultant sequence data for polymorphisms, which we identified as two chromatogram peaks in a single position, clearly observed in both forward and reverse sequences. We manually edited and aligned sequences in CodonCode Aligner v3 (CodonCode Corporation). Genetic variation was initially analyzed in a sample set of 70 dolphins, including 8-10 samples from each of the eight groups. If nonsynonymous mutations were detected in this initial sample set, genetic variation was further analyzed in the total sample set of 182 individuals. All unique alleles that were not observed in multiple homozygous individuals were confirmed by sequencing 16-24 clones produced using pGEM®-T Easy Vectors with JM109 competent E. coli cells (Promega) following manufacturer instructions. Alleles were confirmed only if observed in multiple clones. We conducted phylogenetic analyses in MEGA v5.2 (Tamura et al., 2011) using the maximum likelihood method with the Kimura 2-parameter model (Kimura, 1980) with invariable sites (I = 0.409), the evolutionary model that best described the nucleotide substitution pattern (maximum likelihood fit with lowest Bayesian information criterion). For this analysis, we
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
6
Page 6 of 25
M
an
us
cr
ip t
combined nucleotide sequences from the DIS6-containing exon (9, 10, or 11) and the portion of the exon (24, 26, 27, or 28) encoding DIVS4-S5. For variable isoforms, we included only the most common variant. We tested for significant effects of geographic region (central-west Florida vs. Panhandle) and survival (live vs. dead) on SCNA allele frequency using multifactorial logistic models implemented in JMP Pro 11.0.0 (SAS Institute Inc., 2014). We assessed the power of this study to detect an association between genetic variation and survival using a case/control power analysis conducted in PGA (Menashe et al., 2008). Specifically, we calculated the minimum minor allele frequency given the total sample size (94 controls, 83 cases) that would provide 80% power to detect a relative risk of 2 if the prevalence of disease is between 0.01 and 0.20. Power analyses often test for the ability of a study to detect a two-fold increase in risk (Menashe et al., 2008); accordingly, we tested our ability to detect a relative risk of 2, the point at which a dolphin with a “susceptible” allele is twice as likely to die due to brevetoxin exposure than a dolphin that does not have the “susceptible” allele. We tested our power across a range of disease prevalence values because of the current lack of understanding of brevetoxicosis disease etiology and prevalence in the bottlenose dolphin populations. The range of possible disease prevalence values was estimated using the size of the studied dolphin populations (Waring et al., 2013) and the number of dolphins that died during UMEs associated with red tides (J. Litz, personal communication; Mase et al., 2000; National Marine Fisheries Service, 2004; Twiner et al., 2012), recognizing that only a portion of marine mammal carcasses are recovered (Williams et al., 2011).
te
d
3. Results and Discussion 3.1. Sodium channel isoforms We characterized genetic variation of the putative brevetoxin binding site in nine sodium channel isoforms in common bottlenose dolphins and found that variation among isoforms conformed to expectations based on the general history of SCN gene duplication. Historical gene duplications in tetrapods resulted in four groups of SCNs: 1, 2, 3, and 9; 5, 10, and 11; 4; and 8 (Widmark et al., 2011). This grouping is generally reflected in the maximum likelihood tree built using only 393 bp of each bottlenose dolphin SCNA isoform (Fig. 2). There was no evidence to suggest any of the sequences represented pseudogenes (i.e., no translated sequences contained stop codons).
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
3.2. Sodium channel variation Of the 18 SCNA exons sequenced, only two (SCN5A_exon10 and SCN11A_exon26, Table 2) contained a non-synonymous mutation in the initially screened sample set of 70 dolphins. SCN5 and SCN11 belong to the same SCN cluster (Widmark et al., 2011) and both are considered tetrodotoxin-resistant (Renaud et al., 1983; Dib-Hajj et al., 1998). Brevetoxins have been experimentally demonstrated to target the cardiac sodium channels (SCN5) (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006); however, type B brevetoxins (PbTx-2 and -3), which are the most common toxin type observed during red tides (Pierce et al., 2008), show lower relative affinity for cardiac sodium channels in comparison to skeletal (SCN4) sodium channels (Bottein Dechraoui and Ramsdell, 2003; Bottein Dechraoui et al., 2006). Studies have not yet tested brevetoxin binding affinity to SCN11, which is found primarily in dorsal root ganglion neurons (Dib-Hajj et al., 1998).
7
Page 7 of 25
us
cr
ip t
Neither non-synonymous mutation occurred in a transmembrane spanning region where brevetoxin putatively binds (DIS6 or DIVS5, Trainer et al., 1994). The non-synonymous mutation in SCN5A occurred in the intracellular loop between domains I and II, and the nonsynonymous mutation in SCN11A occurred in DIVS4 (Fig. 3). The fourth transmembrane segment of each domain regulates sodium channel voltage-sensitivity; however, this voltagesensitivity is primarily attributed to positively charged residues, mainly arginines (Catterall, 1986; Guy and Seetharamulu, 1986), and the observed non-synonymous mutation in DIVS4 involved phenylalanine and leucine, neither of which are positively charged. Five other SCNA exons contained synonymous mutations (Table 2, Fig. 3). The most frequently variable genes, SCN1A and SCN2A, belong to the same SCN cluster (Widmark et al., 2011), are broadly expressed in neurons in the central and peripheral nervous systems (Beckh, 1990; Noda et al., 1986a), and are highly tetrodotoxin-sensitive (Noda et al., 1986b). The synonymous mutations found in these SCNA exons may be linked to protein-changing mutations elsewhere in the sodium channel genes, beyond the putative brevetoxin binding site evaluated in our study.
te
d
M
an
3.3. Effects of geographic region and survival Genetic variation was rare among the sampled bottlenose dolphins, even in SCNA exons with polymorphisms (Table 2). Two minor alleles (SCN3A_exon11 and SCN8A_exon10) were only observed once in the initial sample set of Florida dolphins, and thus can have little effect on the observed differences in survival among individuals. Given the near parity of the number of live (34) and dead (28) Florida samples analyzed, we would expect to find alleles strongly associated with survival at closer to 50% frequency. A case/control power analysis suggests this study has 80% power to detect a relative risk of 2 with marker allele frequencies of 0.09-0.18. Thus, for at least SCN1A_exon9 and SCN2A_exon26, the most frequently variable SCNA exons, there should be an 80% probability of detecting an association with survival if there is one. However, we observed no significant effects of survival or the region:survival interaction term on SCNA allele frequency (Table 2). Allele frequencies varied significantly between geographic regions (central-west Florida vs. Panhandle) in three cases (SCN5A_exon10, χ2 = 12.53, p < 0.001; SCN2A_exon26, χ2 = 13.41, p < 0.01; SCN5A_exon28, χ2 = 6.69, p < 0.01, Table 2). In both SCN5A exons the difference between regions was due to rare private alleles observed only in the Panhandle, but these differences could not be attributed to linked variation between the two exons as the private alleles were not observed in the same individuals. As described above, such rare variation is unlikely to explain the difference in HAB resistance between regions. Rather, the difference in frequency of rare SCNA alleles is more likely a reflection of limited gene flow between geographic regions due to the tendency for high site fidelity in bottlenose dolphins in inshore waters of the Gulf of Mexico (Balmer et al., 2008; Hubard et al., 2004; Maze and Würsig, 1999; Scott et al., 1990). Long-term site fidelity can result in significant genetic differentiation among dolphin populations at a fine geographic scale (Sellas et al., 2005). Greater genetic differentiation is expected between dolphins from the Gulf of Mexico and the western North Atlantic (Rosel et al., 2009). However, we found very few differences in genetic variation across SCN genes between the dolphins from Florida and North Carolina. No exonic variation was observed uniquely in North Carolina, though we observed a single mutation in the intronic region upstream of SCN9A_exon10 only in North Carolina samples. SCNA minor alleles characterized in dolphins from Florida were almost never observed in dolphins from
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
8
Page 8 of 25
ip t
North Carolina (SCN8A_exon10 minor allele was observed in a single North Carolina sample), although we acknowledge that our sample size from North Carolina was small. Dolphins in North Carolina represent an outgroup that has rarely, if ever, been exposed to red tides. Yet, our results suggest similar selection pressures acting to conserve the existing voltage-gated sodium channel genetic structure across these geographically disjunct and genetically differentiated bottlenose dolphin populations.
te
d
M
an
us
cr
3.4. Discussion of lack of survival-associated SCN variation Highly conserved SCNA sequences across bottlenose dolphins along the Gulf coast of Florida likely reflect the functional importance of the voltage-gated sodium channels. In other cases of neurotoxin resistance, fitness tradeoffs are observed in conjunction with SCN amino acid substitutions (Brodie III and Brodie Jr, 1999). These costs may explain why resistanceinducing mutations are only observed in natural populations regularly exposed to neurotoxins (Brodie Jr et al., 2002; Connell et al., 2007). These costs may also explain why we do not observe SCNA mutations associated with brevetoxin resistance in bottlenose dolphins. Any change in the voltage-gated sodium channels that affects muscular control may be too costly for this species, or the fitness advantage may not outweigh the cost of the mutation. Our findings may be explained by a difference in the strength or constancy of selective pressure between the dolphin-brevetoxin system and other studied systems that have evolved neurotoxin resistance. The strength of selection for brevetoxin resistance in bottlenose dolphins may be reduced by their trophic separation from the toxin-producing organism, their long generation time and the episodic nature of HABs, and their ability to behaviorally avoid exposure. In the well-studied case of garter snakes, predators have evolved neurotoxin resistance to prey that sequester toxins as a defense mechanism (Brodie III and Brodie Jr, 1991). However, bottlenose dolphin prey do not sequester toxins as a defense mechanism, and dolphins are separated by at least one trophic level from the toxin-producing algae. Additionally, in contrast to consistently toxic prey, HABs are episodic events, though low levels of toxin can be detected in the prey fish months after the termination of a bloom (Fire et al., 2008a). Locally adapted resistance to episodic HABs has been demonstrated in other marine species (e.g., softshell clam, Connell et al., 2007; marine copepod, Colin and Dam, 2004), but these species accumulate toxin directly from the harmful algae, have shorter generation times, and are less mobile, all factors which may influence the strength of selection for algal toxin resistance. Individual or population-level variation in bottlenose dolphin resistance to HABs may alternatively be explained by non-genetic hypotheses. For example, bottlenose dolphins may behaviorally adapt to the presence of HABs such that they minimize direct (e.g., avoid areas of dense blooms) and indirect (e.g., alter foraging strategies) toxin exposure. Although highly mobile, dolphins do not entirely avoid HAB toxin exposure, as evidenced by behavioral observations during HABs (McHugh et al., 2011) and toxin concentrations measured in live bottlenose dolphins (Fire et al., 2008b). However, dolphins have been observed to alter both the size of their home range and their foraging behavior during HABs, likely in response to changes in prey abundance (Bowen, 2011; McHugh et al., 2011; Powell and Wells, 2011). Certain foraging behaviors may increase the likelihood of exposure to HABs; for example, eating discarded bycatch from commercial fishing vessels provides an avenue of exposure to toxins that originate in distant blooms (National Marine Fisheries Service, 2004). The immune system may also play a role in brevetoxin resistance. Brevetoxin exposure can cause immunosuppression in marine species (Bossart et al., 1998; Kreuder et al., 2002;
Ac ce p
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
9
Page 9 of 25
an
us
cr
ip t
Walsh et al., 2005; Walsh et al., 2010), but the toxin’s antigenicity and ability to induce an adaptive immune response is poorly understood. If antigenic, chronic low levels of toxin exposure may result in immunity, though alternatively, chronic low-level exposure can increase toxin sensitivity (e.g., domoic acid sensitivity in zebrafish, Lefebvre et al., 2012). In central-west Florida, low levels of brevetoxins have been measured in dolphins when HABs are not present (Fire et al., 2008b). These low levels of toxins may be explained by long-term retention of the toxin or by prolonged exposure to continual sources of toxin, such as prey species that store toxin (Fire et al., 2008a) or toxin adsorbed to plant material (Flewelling et al., 2005). Further research is needed to better understand the effects of chronic exposure to background levels (
