Introduction
Wildlife translocation remains a popular tool for re-establishing declining or extirpated species and the science guiding the application of this tool has progressed (Seddon et al., Reference Seddon, Armstrong and Maloney2007, Reference Seddon, Strauss, Innes, Ewen, Armstrong, Parker and Seddon2012; Batson et al., Reference Batson, Gordon, Fletcher and Manning2015). Despite increased knowledge and improved procedures, translocations, which involve the introduction of naive hosts into new environments with novel pathogens, invariably pose an increased risk of disease (Ballou, Reference Ballou1993; Viggers et al., Reference Viggers, Lindenmayer and Spratt1993; Gerhold and Hickling, Reference Gerhold and Hickling2016). It is possible to carefully consider and mitigate the disease risk for the translocated organisms at capture, captivity and pre-release stages, but the novel host−pathogen interactions that result post-release are often problematic and difficult to predict in advance (Cunningham, Reference Cunningham1996; Ewen et al., Reference Ewen, Acevedo-Whitehouse, Alley, Carraro, Sainsbury, Swinnerton, Woodroffe, Ewen, Armstrong, Parker and Seddon2012; Hartley and Sainsbury, Reference Hartley and Sainsbury2017). This is in part due to the large number of factors that can be involved in disease outcomes which can include genetic composition of the hosts (Acevedo-Whitehouse et al., Reference Acevedo-Whitehouse, Spraker, Lyons, Melin, Gulland, Delong and Amos2006; Savage and Zamudio, Reference Savage and Zamudio2011), characteristics of parasites at the release site (LoGiudice, Reference LoGiudice2003), presence of other parasites in the host (Faria et al., Reference Faria, van Oosterhout and Cable2010), physiological stress (Parker et al., Reference Parker, Dickens, Clarke, Lovegrove, Ewen, Armstrong, Ewen, Armstrong, Parker and Seddon2012), altered host dynamics (Aiello et al., Reference Aiello, Nussear, Walde, Esque, Emblidge, Sah, Bansal and Hudson2014), host body condition (Mathews et al., Reference Mathews, Moro, Strachan, Gelling and Buller2006) and habitat variables at the release site (Raskevitz et al., Reference Raskevitz, Kocan and Shaw1991). These factors, in turn, can work singly or in combination to influence the ultimate success or failure of a restoration (Seddon et al., Reference Seddon, Armstrong and Maloney2007).
Parelaphostrongylus tenuis, commonly known as the meningeal worm, is a nematode parasite of the white-tailed deer (Odocoileus virginianus), which serves as its primary host and rarely suffers adverse effects from infection (Anderson, Reference Anderson1963). Parelaphostrongylus tenuis occurs in forested areas of eastern North America (Anderson, Reference Anderson1972; Wasel et al., Reference Wasel, Samuel and Crichton2003), and is transferred via one of many possible intermediate gastropod hosts (Anderson, Reference Anderson1963, Reference Anderson1972; Lankester and Anderson, Reference Lankester and Anderson1968). Attempts to restore elk (Cervus canadensis) to the eastern US have been hampered by disease caused by this parasite (Carpenter et al., Reference Carpenter, Jordan and Ward1973; Severinghaus and Darrow, Reference Severinghaus and Darrow1976; Eveland et al., Reference Eveland, George, Hunter, Forney, Harrison, Boyce and Hayden-Wing1979). Infection with P. tenuis is a primary source of mortality, particularly for younger age classes (Samuel et al., Reference Samuel, Pybus, Welch and Wilke1992) and can also make translocated elk susceptible to other sources of mortality including secondary infections, predation and vehicle collisions (Keller et al., Reference Keller, Montgomery, Campa, Beyer, Winterstein, Hansen and Millspaugh2015). Though most elk suffer neurological damage from P. tenuis, some individuals can tolerate low levels of infection (Samuel et al., Reference Samuel, Pybus, Welch and Wilke1992).
Infection rates for P. tenuis at elk release sites may be reduced though reductions in overlap among elk and white-tailed deer or reduced contact with intermediate hosts, which may promote the persistence of restored elk populations in the eastern US (Raskevitz et al., Reference Raskevitz, Kocan and Shaw1991; Samuel et al., Reference Samuel, Pybus, Welch and Wilke1992). However, ecological factors alone cannot explain the success of some restored populations (Bender et al., Reference Bender, Schmitt, Carlson, Haufler and Beyer2005). Some authors suggest that persistent restored elk populations occur due to decreased susceptibility to P. tenuis over time (Anderson, Reference Anderson1972; Lankester, Reference Lankester, Samuel, Pybus and Kocan2001; Larkin et al., Reference Larkin, Alexy, Bolin, Maehr, Cox, Wichrowski and Seward2003; Bender et al., Reference Bender, Schmitt, Carlson, Haufler and Beyer2005). Adaptation to the parasite may be beneficial to further elk translocation efforts since individuals that can survive at low levels of infection may have a better chance of surviving another exposure (Ewen et al., Reference Ewen, Acevedo-Whitehouse, Alley, Carraro, Sainsbury, Swinnerton, Woodroffe, Ewen, Armstrong, Parker and Seddon2012). However, infectivity and/or the severity of infection by a parasite can vary spatially due to local adaptation (Dybdahl and Storfer, Reference Dybdahl and Storfer2003). Thus, the effects of parasitism by P. tenuis on translocated elk in the eastern US are difficult to predict, in part because little is known about the spatial variation of P. tenuis.
In general, parasite distribution and geographic variation is related to the distribution of its primary host (Poulin, Reference Poulin2007). Population genetic differences of the host and environmental variation can result in spatial variation among the evolutionary strategies employed by the parasite; in other words, local adaptation (Dybdahl and Storfer, Reference Dybdahl and Storfer2003). Restricted movement and dispersal of parasites results in genetic drift. Thus, by examining the population genetic structure of P. tenuis, the possibility of local adaptation and spatial variation of virulence can be inferred. In the case of P. tenuis, its primary host, the white-tailed deer, has been translocated throughout the eastern US and shows high neutral genetic diversity and low genetic differentiation at a broad scale (DeYoung et al., Reference Deyoung, Demarais, Honeycutt, Rooney, Gonzales and Gee2003; Budd et al., Reference Budd, Berkman, Anderson, Koppelman and Eggert2018). But at a finer scale, white-tailed deer may exhibit spatial genetic structure due to habitat discontinuities (Blanchong et al., Reference Blanchong, Samuel, Scribner, Weckworth, Langenberg and Filcek2007) and/or social structure (Comer et al., Reference Comer, Kilgo, D'Angelo, Glenn and Miller2005; Cullingham et al., Reference Cullingham, Merrill, Pybus, Bollinger, Wilson and Coltman2011), which may also correlate to geographic barriers for P. tenuis (Jacques et al., Reference Jacques, Jenks, Grovenburg, Klaver and Dubay2015). Factors relevant solely to parasite dispersal, such as gastropod densities and climatic conditions, may impact geographic variation as well. For instance, certain habitat types, such as grasslands, may not support P. tenuis at early larval growth stages (Shostak and Samuel, Reference Shostak and Samuel1984), and thus may inhibit movements of P. tenuis across some ecoregions (Wasel et al., Reference Wasel, Samuel and Crichton2003).
When translocated from Kentucky to Missouri in the springs of 2011–2013 (Dent, Reference Dent2014), elk experienced high morbidity and mortality due to P. tenuis infections (Chitwood et al., Reference Chitwood, Keller, Al-Warid, Straka, Hildreth, Hansen and Millspaugh2018). In Kentucky, the persistence of some elk with P. tenuis and no clinical signs of disease (Larkin et al., Reference Larkin, Alexy, Bolin, Maehr, Cox, Wichrowski and Seward2003), and the observation of decreased mortality from P. tenuis (Slabach et al., Reference Slabach, Hast, Murphy, Bowling, Crank, Jenkins, Johannsen and Cox2018) suggest that the elk herd has become less susceptible to the local variant of P. tenuis. Though translocation-induced stress was also considered a factor, exposure to a new genetic variant of P. tenuis in Missouri was suspected to have contributed to the high mortality rate in the translocated herd (Chitwood et al., Reference Chitwood, Keller, Al-Warid, Straka, Hildreth, Hansen and Millspaugh2018).
We examined the hypothesis that elk translocated within the eastern US could be exposed to novel genetic variants of P. tenuis by detailing the genetic structure among P. tenuis taken from white-tailed deer and elk at a source (Kentucky) and a release site (Missouri). We employed DNA sequence data from mitochondrial and nuclear genes, both of which have proven useful for population genetic and phylogenetic studies of nematodes (Nadler, Reference Nadler1992; Blouin et al., Reference Blouin, Yowell, Courtney and Dame1998). Our study represents one of the first examinations of the genetic structure and molecular variability of P. tenuis. Furthermore, our study represents an important and often underrepresented component of restoration programmes: post-release monitoring and evaluation (Ewen et al., Reference Ewen, Acevedo-Whitehouse, Alley, Carraro, Sainsbury, Swinnerton, Woodroffe, Ewen, Armstrong, Parker and Seddon2012).
Materials and methods
Sampling in Missouri
Parelaphostrongylus tenuis samples (n = 140) were collected from 57 hunter-harvested deer (25 males, 29 females, 3 unknown sex; 134 P. tenuis) and 5 elk (6 P. tenuis) natural mortalities in 11 counties (Table 1, Supplementary Table 1) during 2015 and 2016. Carcasses were collected as soon as possible after they were reported to the Missouri Department of Conservation (MDC). If it was not feasible to preserve the entire carcass, heads and tissue samples were preserved at −20°C until necropsies could be performed. Necropsies were conducted by trained MDC personnel or by veterinary personnel at the University of Missouri College of Veterinary Medicine. At necropsy, all observed P. tenuis were carefully collected and preserved in individual vials at room temperature in a 1:1 solution of glycerol and absolute ethanol. We tested for differences between host sexes with respect to the number of P. tenuis detected at necropsy using a Mann−Whitney U test.
Samples collected and analysed for this study from white-tailed deer (O. virginianus) and meningeal worms (P. tenuis) in Missouri and Kentucky
a M = male, F = female, U = unknown.
b Two P. tenuis were removed from the 28S dataset as they did not yield readable sequence data.
c One male deer was removed from the mtDNA study, as the single sample did not yield readable sequence data.
d One deer was removed from the 28S dataset as the single sample did not yield readable sequence data.
e Four P. tenuis were removed from the mtDNA dataset and 6 P. tenuis were removed from the 28S dataset as they did not yield readable sequence data.
For the genetic study, we selected 96 P. tenuis samples collected from 28 deer (11 males, 15 females, 2 unknown sex) and 5 elk. When selecting samples, our goals were to represent all counties, maximize the number of deer with multiple parasites, and analyse all available P. tenuis from elk.
Sampling in Kentucky
Parelaphostrongylus tenuis samples (n = 71) were collected from 22 hunter-harvested white-tailed deer and 2 elk in the Elk Restoration Zone by the Kentucky Department of Fish and Wildlife Resources in December 2016 (Table 1). The collecting locations and sexes of the hosts were not provided; thus, we analysed the state as a whole. Following the same procedure that was used for the Missouri samples, P. tenuis were carefully removed from heads and preserved at room temperature in a 1:1 solution of glycerol and absolute ethanol prior to DNA extraction.
Mitochondrial DNA analyses
We extracted genomic DNA from the Missouri and Kentucky P. tenuis samples using the Qiagen DNEasy Blood and Tissue Kit with the manufacturer's protocol (Qiagen, Valencia, CA). Adult worms were removed from the preservation buffer and washed in sterile water to remove residual buffer. For small worms, we extracted DNA from the entire worm. For larger worms, we used approximately one-third of the tail. We amplified and sequenced 876 bp of the mitochondrial (mtDNA) cytochrome oxidase I (COI) gene using the primers and conditions described by Asmundsson et al. (Reference Asmundsson, Mortenson and Hoberg2008) in their study of the congeneric species Parelaphostrongylus andersoni found in western North America. Amplification products were sequenced in both directions in an ABI 3730xl DNA analyser (Applied Biosystems, Foster City, CA) at the University of Missouri DNA Core and sequences were aligned in Sequencher 5.4 (Gene Codes, Ann Arbor, MI) and collapsed into haplotypes in Fabox 1.41 (Villesen, Reference Villesen2007).
For all hosts with multiple P. tenuis, we examined the number of COI haplotypes present. We calculated the number of polymorphic sites and the nucleotide diversity (π) in Arlequin 3.5 (Excoffier and Lischer, Reference Excoffier and Lischer2010). Because each Missouri county was represented by a small number of non-randomly collected samples, we examined the distribution of haplotypes by county but did not compute the values of F ST between counties.
To examine genetic differentiation between P. tenuis mtDNA haplotypes in Missouri and Kentucky, we calculated F ST in Arlequin 3.5. To examine relationships among all haplotypes and published sequences from an individual P. tenuis detected in Gibson Island, Maryland, USA (EF173722, Asmundsson et al., Reference Asmundsson, Mortenson and Hoberg2008) and from P. andersoni (EU052282, EU029988, Asmundsson et al., Reference Asmundsson, Mortenson and Hoberg2008), we constructed a TCS network of unique haplotypes in POPART (Clement et al., Reference Clement, Snell, Walke, Posada and Crandall2002; Leigh and Bryant, Reference Leigh and Bryant2015). Networks were manually reconstructed for clarity in Inkscape (https://inkscape.org).
Nuclear 28s ribosomal RNA analyses
We amplified and sequenced 718 bp of the nuclear 28S ribosomal RNA gene using primers Pt28SF: CGCTGATCTTTCGATGTTAATC and Pt28SR: CGCAACCTGTACGCTCTACC, which we designed from GenBank accession #EU595594 (Asmundsson et al., unpublished). The PCR was performed in 25 μL volumes including 1× Amplitaq Gold PCR buffer (Applied Biosystems), 0.4 μ m each primer, 2.0 mm MgCl2, 0.8 mm BSA, 0.5 U Amplitaq Gold polymerase (Applied Biosystems) and 15–20 ng of the extracted DNA. The PCR profile included an initial 10 min incubation at 95°C followed by 40 cycles of 95°C for 1 min, 55°C for 1 min, 72°C for 1 min and a single final incubation at 72°C for 10 min. Amplification products were sequenced in an ABI 3730xl DNA analyser (Applied Biosystems) and the resulting sequences were aligned in Sequencher 5.4 (Gene Codes, Ann Arbor, MI). Genotypes were determined by manual comparisons.
The phylogenetic relationships among P. tenuis genotypes from Missouri and Kentucky were inferred using a heuristic search with the maximum likelihood criterion in Paup* V. 4.0a166 (Swofford, Reference Swoffford DL2002). A published sequence from P. tenuis (EU595594) was added for comparison, and sequences from P. andersoni (EU595597), Parelaphostrongylus odocoilei (AY292803) and Elaphostrongylus rangiferi (EU595596) were added as outgroups for this analysis. Support for each node was assessed using 10 000 bootstrap replicates in Paup*.
Results
Prior to the genetic study, we compared the number of P. tenuis collected from male and female deer in Missouri. We detected 54 P. tenuis in 25 males (average 1.8 ± 1.3 s.d.) and 75 P. tenuis in 29 females (average 3.0 ± 2.5 s.d.). Although this does not represent a random sample of individuals in Missouri and thus we cannot extrapolate our results to the population as a whole, we found no significant difference between males and females with respect to the number of P. tenuis detected at necropsy (U = 285.5, z = −1.327, P = 0.184).
Mitochondrial DNA
Missouri
We successfully recovered COI sequences from 95 P. tenuis from 27 deer and 5 elk (Table 2, Supplementary Table 1). One sample could not be sequenced reliably and was eliminated from the mtDNA study. After confirming that all recovered sequences were most similar to published sequences of P. tenuis using a nucleotide BLAST (Basic Local Alignment Search Tool) search (https://blast.ncbi.nlm.nih.gov), we translated them in Sequencher and found no evidence of stop or nonsense codons. Twenty unique haplotypes were detected based on 31 polymorphic sites [26 Transitions (Ts):5 Transversions (Tv)]. Fifteen were found only in deer, 1 was found only in elk and 4 were found in both species.
Numbers and geographic distribution of P. tenuis mitochondrial DNA haplotypes detected in samples from deer and elk in Missouri (MO) and Kentucky (KY)
Nucleotide diversity was 0.0082 ± 0.0043 (s.d.), and gene diversity was 0.8990 ± 0.0167 (s.d.).
Of the 27 deer, 21 had multiple parasites and 17 of those individuals had parasites from multiple mitochondrial lineages. Of the deer with multiple parasites, females (n = 10) had an average of 2.9 ± 1.2 (s.d.) mitochondrial lineages, while males (n = 9) had an average of 2.0 ± 0.7 (s.d.) mtDNA lineages. For the two deer of unknown sex, one had parasites from 2 lineages while the other had parasites from only 1 lineage. The most common haplotype was Hap8, 19 samples were found in deer and 2 in elk. This haplotype had a wide geographic distribution, being found in 7 of the 11 counties included in this study (Table 2).
Kentucky
Of the 71 P. tenuis samples provided, we successfully recovered sequences from 67 samples from 22 deer and 2 elk (Table 1, Supplementary Table 1). Four could not be sequenced reliably and were removed from this part of the study. Two sequences (KY12E and KY18D) were removed from the dataset when replicated sequences from each were translated in Sequencher and stop codons were consistently found, suggesting that they represented nuclear copies (numts, Lopez et al., Reference Lopez, Yuhki, Modi, Masuda and O'Brien1994). An additional two sequences (KY16A and KY38B, both found only in deer) were found to be only 91 and 90% similar to published sequences for this species, respectively, making them substantial outliers. Because these sequences could represent numts or parasites that had been misidentified as to species, we analysed the Kentucky mtDNA data with and without these sequences.
When the two highly divergent sequences were included, we detected 20 haplotypes in 65 samples based on 114 polymorphic sites (97 Ts: 27 Tv). Eighteen were found only in deer, 1 was found only in elk, and 1 was found in both. Without these samples, we detected 18 haplotypes in 63 samples, 16 of which were found only in deer. Only 2 haplotypes (Hap8 and Hap18) were shared with the Missouri deer. The most common haplotype was Hap2, found in 23 deer (Table 3).
Numbers and geographic distribution of P. tenuis 28S ribosomal RNA genotypes (GT) detected in samples from deer and elk in Missouri (MO) and Kentucky (KY)
Including the two divergent haplotypes, the nucleotide diversity was 0.0278 ± 0.0137 (s.d.) and gene diversity was 0.8399 ± 0.0357 (s.d.). Of the 22 deer, 15 had multiple parasites and 12 of those individuals had parasites from multiple mitochondrial lineages. Of the deer with multiple parasites, the average number was 2.3 ± 1.2 (s.d.). Without the divergent haplotypes, the nucleotide diversity was 0.0238 ± 0.0118 (s.d.) and gene diversity was 0.8295 ± 0.0372 (s.d.). One of the two elk had 2 parasites with 2 different mtDNA haplotypes. (Supplementary Table 1).
mtDNA differentiation
The pairwise value of F ST between Missouri and Kentucky was 0.180, which was found to be significant (P < 0.001) based on a permutation test (110 permutations) in Arlequin 3.5. The minimum spanning network produced using our mtDNA sequences revealed a star-like structure (Fig. 1), with many haplotypes differing by only a few base pairs (bp). It illustrates the strong differentiation between the two divergent Kentucky haplotypes, each of which differs by more than 60 bp from the most similar haplotypes, but only differ from each other by 3 bp. Three haplotypes (KY12F, KY20A and KY25B) were found to be more similar to P. tenuis sequences from Maryland and P. andersoni sequences than to other sequences from Missouri and Kentucky (Fig. 1).
TCS network based on 876 bp of mitochondrial cytochrome oxidase I (COI) sequences in samples of P. tenuis from Missouri and Kentucky, USA. One P. tenuis sample from Maryland, USA (EF173722) and two samples of the closely related species P. andersoni (EU029988 and EU052282) were included for comparison.
Nuclear 28s ribosomal RNA
Missouri
We successfully recovered sequences from 93 samples collected from 28 deer and 6 elk (Table 1, Supplementary Table 1). We detected 41 unique genotypes based on 9 polymorphic base pairs, 35 of which were found only in deer, 3 of which were found only in elk and 3 of which were shared between species. Of the 28 deer, 21 had multiple parasites and of those 20 had parasites with differing nuclear genotypes. Of the deer with multiple parasites, females (n = 10) had an average of 3.8 ± 1.6 (s.d.) different nuclear genotypes, while males (n = 9) had an average of 2.6 ± 1.2 (s.d.) different nuclear genotypes. Two deer of unknown sex each had 2 parasites with 2 different nuclear genotypes. The most common genotype was Genotype25; 18 samples with this genotype were found in deer and 2 in elk. This genotype had a wide geographic distribution, being found in 8 of the 11 counties included in this study (Table 3).
Kentucky
We successfully recovered sequences from 62 samples collected from 20 deer and 2 elk (Table 1, Supplementary Table 1). We detected 10 genotypes based on 5 polymorphic base pairs, 8 of which were found only in deer and 2 of which were shared between deer and elk. Of the 20 deer, 15 had multiple parasites and 12 had parasites with differing nuclear genotypes. Deer with multiple parasites had 3.6 ± 1.5 (s.d.) parasites with an average of 2.5 ± 1.1 (s.d.) different nuclear genotypes. One of the two elk had only one parasite while the other had 2 parasites with different genotypes. Seven of the 10 genotypes were shared with Missouri deer and elk. The most common genotype was Genotype25, which was found in 25 deer samples and 1 elk sample (Table 3).
Nuclear 28s ribosomal RNA differentiation
Phylogenetic analyses revealed that P. tenuis did not form a monophyletic group with respect to the single P. andersoni sample included as an outgroup (Fig. 2). Although bootstrap analyses supported only the group that included both P. tenuis and P. andersoni, that group included two major sub-groups, one of which contained the majority of genotypes found in Missouri as well as two of the most common shared genotypes between states (G142B, G180D). The second clade contained all 3 of the unique Kentucky genotypes and two commonly shared genotypes (G135D, KY4A, Fig. 2).
Maximum likelihood tree based on 718 bp of the 28S ribosomal RNA gene in samples of P. tenuis from Missouri and Kentucky, USA. One published P. tenuis sequence (EU595594) was included for comparison and sequences from related species P. andersoni (EU595597), P. odocoilei (AY2928030) and Elephostrongylus rangiferi (EU595596) were included as outgroups.
Discussion
Although its definitive host is the white-tailed deer, a number of studies have confirmed that P. tenuis can parasitize many different hosts, including elk, moose (Alces alces), llamas (Lama glama), alpacas (Vicugna pacos), goats (Capra hircus), cattle (Bos taurus), horses (Equus caballus), bison (Bison bison), sika deer (Cervus nippon) and guinea pigs (Cavia porcellus; Anderson, Reference Anderson1972; Lankester, Reference Lankester, Samuel, Pybus and Kocan2001, Reference Lankester2010; Weiss et al., Reference Weiss, Sarver, Thilsted and Wolfe2008; Whitehead and Bedenice, Reference Whitehead and Bedenice2009; Gerhold et al., Reference Gerhold, Keel, Arnold, Hotton and Beckstead2010; Tanabe et al., Reference Tanabe, Gerhold, Beckstead, de Lahunta and Wade2010; Mitchell et al., Reference Mitchell, Peters-Kennedy, Stokol, Gerhold, Beckstead and Divers2011; Southard et al., Reference Southard, Bender, Wade, Grunenwald and Gerhold2012; Dobey et al., Reference Dobey, Grunenwald, Newman, Muller and Gerhold2014; Gerhold and Hickling, Reference Gerhold and Hickling2016). To date, the majority of these studies have focused on genetic confirmation of host infection and identification of morphologically indistinguishable dorsal-spined larvae in Elaphostrongyline species (Gajadhar et al., Reference Gajadhar, Steeves-Gurnsey, Kendall, Lankester and Stéen2000; Kutz et al., Reference Kutz, Veitch, Hoberg, Elkin, Jenkins and Polley2001; Gerhold et al., Reference Gerhold, Keel, Arnold, Hotton and Beckstead2010; Dobey et al., Reference Dobey, Grunenwald, Newman, Muller and Gerhold2014). Other than an unpublished study by Gerhold et al. (Reference Gerhold, Grunenwald, Muller and Su2016), we are not aware of studies of this parasite at the population or regional level.
Given previous suggestions that P. tenuis has little intraspecific diversity or population structure (Gerhold et al., Reference Gerhold, Grunenwald, Muller and Su2016), we were surprised to find high levels of diversity at both mitochondrial and nuclear DNA in Missouri and Kentucky. While we found that 2 of the 38 mtDNA haplotypes and 7 of the 44 nuclear genotypes we detected were shared between states, there were significant frequency differences between states. These frequency differences and the presence of unique mtDNA haplotypes and nuclear genotypes in each state contributed to the high levels of diversity within and the high level of differentiation between states. Further, we found that the previously published mtDNA haplotype of P. tenuis from Maryland, USA, differed at 30 or more bp from the majority of our Missouri and Kentucky haplotypes and that three haplotypes detected in Kentucky clustered with the Maryland haplotype. Taken as a whole, our results support previous suggestions that genetic differences among populations across the range of this parasite may have contributed to differences in susceptibility to infection between hosts that have acquired immunity to parasite lineages in their home ranges (Chitwood et al., Reference Chitwood, Keller, Al-Warid, Straka, Hildreth, Hansen and Millspaugh2018). Translocation of such locally adapted individuals risked exposing them to strains to which they were immunologically naive.
Our finding of two highly divergent mtDNA haplotypes is especially interesting. While we cannot rule out the possibility that these haplotypes represent nuclear copies of the COI gene, we found no evidence of stop codons and a NCBI-BLAST search found that they are most similar to previously published sequences of P. tenuis. However, the sequences of haplotype KY16A and KY38B were found to be only 90–91% similar to published sequences (EF173722-23) for this species. In the haplotype network, they differ from other P. tenuis sequences by a minimum of 62 bp, or 7% of the COI sequence. The 28S sequences for these samples were found to match two of the most common genotypes in both states, G142B and G180D, supporting their identification as P. tenuis. Collectively, our mitochondrial and nuclear results suggest there may be previously unrecognized lineages present in Parelaphostrogylus in Missouri and Kentucky, similar to the novel protostrongylid described by Dobey et al. (Reference Dobey, Grunenwald, Newman, Muller and Gerhold2014), who detected the DNA in a goat with lesions that suggested infection with P. tenuis.
Although we found a relatively large number of nuclear genotypes, they were based on only 9 polymorphic base pairs. As in Carreno et al. (Reference Carreno, Caporaso, Beade, Marull, Uhart, Markwardt and Nadler2012), our nuclear data did not clearly resolve differences between species as it could not differentiate the P. andersoni sequence we included as an outgroup from our P. tenuis sequences. Thus, our results support the recommendations of Blouin (Reference Blouin2002), who demonstrated that mitochondrial DNA had much higher levels of sequence divergence within nematodes than the nuclear ITS-1 and ITS-2 regions and suggested that the high substitution rate of mtDNA makes it a much more potent tool for detecting cryptic lineages of nematodes than nuclear DNA.
High levels of genetic diversity in nematodes have been attributed to high substitution rates and large effective population sizes (Blouin et al., Reference Blouin, Yowell, Courtney and Dame1995, Reference Blouin, Yowell, Courtney and Dame1998). But it should also be acknowledged that the history of introductions of the white-tailed deer hosts in Missouri and Kentucky may have contributed to the high levels of P. tenuis genetic diversity observed in both Missouri and Kentucky. In Missouri, the host population declined to near extirpation by the late 1800s. A combination of hunting limitations, predator eradication and translocations from Wisconsin, Michigan and Minnesota between 1925 and 1957 restored the Missouri population, which has expanded to an estimated size today of 1.4 million (Bennitt and Nagel, Reference Bennitt and Nagel1937; McDonald and Miller, Reference McDonald and Miller2004; Budd et al., Reference Budd, Berkman, Anderson, Koppelman and Eggert2018). Similarly, the Kentucky population was supplemented by deer from Wisconsin in the 1920s, then driven down to approximately 100 individuals by around 1935. Between 1935 and 1953, the population recovered through natural recruitment as well as the translocation of deer from Wisconsin, Oklahoma and Tennessee (Gassett, Reference Gassett, Maehr, Noss and Larkin2001). Today, white-tailed deer in both Missouri and Kentucky have high levels of genetic diversity and low levels of differentiation across the landscape (Doerner et al., Reference Doerner, Braden, Cork, Cunningham, Rice, Furman and McElroy2005; Budd et al., Reference Budd, Berkman, Anderson, Koppelman and Eggert2018), consistent with rapid population growth and the mixing of multiple lineages derived from recovering native populations and translocated individuals. Our data suggest that these processes also resulted in high levels of genetic diversity within state populations and genetic differentiation between states in the parasitic nematode P. tenuis.
At the time of the white-tailed deer translocations, managers were not yet fully aware of the risks of translocating parasites and pathogens along with their hosts (Mathews et al., Reference Mathews, Moro, Strachan, Gelling and Buller2006). After translocation, parasites are subject to the same evolutionary pressures as their host, including population bottlenecks, hybridization between locally adapted lineages, and adaptation to novel environmental conditions, including new host species. With their large population sizes and rapid generation times, parasites have the capacity for rapid evolution. The star-like pattern we detected in mtDNA, as well as the different frequencies of nuclear genotypes we detected in P. tenuis suggest both the introduction of new parasite lineages in Missouri and Kentucky and the rapid diversification of those lineages since introduction. This combination could have resulted in P. tenuis lineages in Missouri to which translocated elk from Kentucky were unable to mount an effective immune response.
Since P. tenuis is as ubiquitous in eastern North America as its primary host and is an important source of disease and mortality for elk (Keller et al., Reference Keller, Montgomery, Campa, Beyer, Winterstein, Hansen and Millspaugh2015) and moose (Lankester, Reference Lankester2010), we suggest that conservation and restoration efforts for these ungulates would benefit from an expanded understanding of the spatial genetic variability of P. tenuis. Our study has provided evidence that spatial variability exists in a central region of the eastern US, and we suspect that a broader survey of the parasite would be informative. In addition to a broader knowledge base, we recommend that translocation plans incorporate analyses of P. tenuis strains when evaluating the suitability of a release site. Our study has provided a framework for such an evaluation. If a large amount of genetic differentiation exists among source and release sites, the potential for increased disease risk from exposure to a novel strain of P. tenuis should be considered in the decision-making process of a reintroduction or restoration.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0031182020001912.
Acknowledgements
M. Colter Chitwood, Sheri Russell and Kelly Straka initiated foundational discussions and helped shape this project. Gabe Jenkins and Joe McDermott with the Kentucky Department of Fish and Wildlife Resources collected samples from Kentucky. David Vasquez Jr., Christine Sholy, Madison Harris and Kaitlin Sulkowski assisted with laboratory analyses. Joe Gunn and David Vasquez Jr. assisted with data analyses and interpretation.
Financial support
Funding for laboratory work and data analyses was provided by a grant from the Missouri Department of Conservation to LSE and a NSF-GRF to KB. Funding for field work was provided by grants from the U.S. Fish and Wildlife Service Wildlife Restoration Grant, the Missouri Department of Conservation and the Rocky Mountain Elk Foundation to JJM.
Conflicts of interest
None.
Ethical standards
None, the project did not concern vertebrates.
