Effect of Trematode Metacercarial Infection on Walking in Larval Salamanders in the Southern Appalachian Mountains, USA

Abstract

According to the Host Manipulation Hypothesis, parasites modify the phenotype of their host to enhance host–host transmission and thereby increase fitness. Metacercarial infection of some amphibians changes host behavior, including locomotion, and thereby enhances predation by the definitive host. To further test this hypothesis, it is first necessary to determine whether a parasite actually modifies a host’s phenotype. In the southern Appalachian Mountains of the US, metacercariae of the trematode Metagonimoides oregonensis (Price, 1931) encyst in the musculature of its second intermediate host, the salamander Desmognathus amphileucus Bishop, 1941. Metacercarial infections of musculature in fish negatively affect host swimming performance. Therefore, we tested the hypothesis that infection by M. oregonensis affects walking in the aquatic larvae of D. amphileucus. We compared this mode of locomotion between infected and uninfected larvae by placing them in 1 m troughs of water and allowing them to wander freely until they stopped. Non-parametric (Kaplan–Meier) survival analysis determined that infected salamanders stopped significantly sooner than uninfected ones. Because infected salamanders move less, the presence of this parasite may contribute to genetic divergence in these salamanders by slowing dispersal and concomitant gene flow. Our results suggest that macroparasites can potentially modify a host’s behavior with biological consequences beyond enhancement of parasite transmission.

1. Introduction

Parasitic infections frequently result in the modification of host behavior. In fact, the Host Manipulation Hypothesis has emerged as an important explanation for parasite-induced changes in host behavior. The idea is that, by modifying a host’s phenotype including behavior, the parasite increases its probability of transmission and thereby its own evolutionary fitness [1]. One effect that parasites can have on host behavior is through decreasing locomotor performance, thereby increasing the probability of parasite transmission by making the host more vulnerable to predation [2]. Parasites have been shown to negatively affect locomotor performance in animal hosts ranging from monarch butterflies to roe deer [3].

While the Host Manipulation Hypothesis has received a growing amount of attention for a number of vertebrate hosts, relatively little of this research has been focused on behavioral modification in amphibians. Amphibians are the most endangered group of vertebrates [4,5] because of their unique dependence on both aquatic and terrestrial habitats as well as their reliance on cutaneous respiration, a feature that makes them particularly susceptible to chemical pollution and desiccation [6]. The endangered nature of amphibians underscores the need for a better understanding of the effects that parasites have on their amphibian hosts, including possible behavioral modification.

Parasitic infections cause a variety of physical anomalies in amphibians, including limb malformations [2,7,8,9,10,11]. Such malformations can negatively impact behavior by lowering locomotor performance such as that demonstrated in the frog Pseuacris regilla (Baird and Girard, 1852) whose swimming speed and jumping distance are decreased by limb abnormalities caused by infection by the trematode Ribeiroia ondatrae Price, 1931 [2].

Most of the studies on the effects of parasites on amphibian behavior have involved anurans, i.e., frogs and toads [1,12,13]. Not only are salamanders less studied than anurans, but the Plethodontidae, a family which comprises two-thirds of all salamander species [14], is among the least studied groups in this regard. Unlike most amphibians, plethodontids are lungless and are thus entirely dependent on cutaneous respiration to effect gas exchange [15]. Like other vertebrates, plethodontids as a group harbor a diversity of parasites [16], making these salamanders an obvious subject for research into host–parasite relationships, including possible modifications of host behavior.

In the Southern Appalachian Mountains of the southeastern United States, one of the most abundant plethodontid species is the Southern Black-bellied Salamander, Desmognathus amphileucus Bishop, 1941, formerly D. quadramaculatus (Holbrook, 1842) [17]. Semi-aquatic as adults, individuals of this species frequent smaller streams across the mountainous regions of Georgia, South Carolina, southwestern North Carolina, and southeastern Tennessee [17]. With an aquatic larval stage that lasts for 3–4 years prior to metamorphosis [18,19], this species can live for 15 or more years [20].

A common parasite of D. amphileucus is the trematode currently recognized as Metagonimoides oregonensis Price, 1931, although the Appalachian form of this parasite is likely an undescribed species of a complex of cryptic species [21]. The life cycle of this trematode includes a raccoon, Procyon lotor (Linnaeus, 1758), as the definitive host and an aquatic snail, Elimia proxima (Say, 1925), as the first intermediate host [21,22]. Various stream-dwelling plethodontids can serve as second intermediate hosts, although D. amphileucus is the primary species utilized [23]. Following being shed from the snail, cercariae of M. oregonensis invade salamander tissues, particularly of the aquatic larvae, and then encyst as metacercariae within the host’s musculature [21,24]. The long larval period and long lifespan of D. amphileucus translate into years of exposure to infective cercariae, resulting in heavy infections by these trematodes [21,25]. Therefore, this salamander species seems an ideal subject for investigating host–parasite relationships.

Metacercarial infections of muscles can negatively affect the swimming ability of fish hosts [26]. Therefore, we aimed to determine whether infection by metacercariae of M. oregonensis affects locomotion in aquatic larvae of D. amphileucus.

2. Materials and Methods

We used larval D. amphileucus because their translucent venters enable the determination of metacercarial infection in a non-destructive manner. We collected 26 larval salamanders from each of two streams. One stream is located on the campus of Piedmont University in the Chattahoochee River system in Habersham County, Demorest, Georgia, USA (34.559 N; 83.543 W). It is devoid of the snail host (E. proxima) and consequently also devoid of M. oregonensis. The other stream is in Stephens County, Demorest, Georgia, USA (34.521 N; 83.388 W) within the Chattahoochee National Forest. This stream is part of the Broad River system, in which both E. proxima and M. oregonensis are abundant [27]. We collected salamanders in their primary habitat by disrupting riffle areas of streams while placing nets downstream [18]. We transported salamanders to the laboratory at Piedmont University in plastic bags of stream water, separating individuals roughly by size because of potential cannibalism by larger individuals [28]. We then inserted an aquarium bubbler in each bag to ensure sufficient aeration of the water. We kept the salamanders in the lab for several days to acclimate to the experimental temperature of 16 °C because lungless salamanders, which depend entirely on cutaneous respiration, are physiologically limited at temperatures much higher than this [8].

We designed a simple experiment to test whether M. oregonensis affects locomotion in D. amphileucus. In the laboratory, we ran locomotion trials by placing each salamander in a metal trough measuring 1 m in length, 9 cm in width, and 8 cm in height. We added water to a depth of ~5 cm. We placed the salamander in the middle of the trough and then allowed the test subject to wander freely and timed it in seconds (s) until it came to a stop and did not move again for 600 s (10 min). At this point, we subtracted the 600 s of rest from the total time to determine time spent moving. We ran each trial for a maximum of 2700 s (45 min).

Metacercariae of M. oregonensis are easily visible with magnification through the translucent ventral surface of larval D. amphieucus (Figure 1), with approximately 10% of the total metacercariae being visible [21,25]. Following each movement trial, we immobilized the salamander in a petri dish with a wet sponge and counted the number of metacercariae visible through the venter between the insertions of the fore and hind limbs under a dissection microscope. Because body size can influence movement rates in plethodontids [29], we measured head width while the salamander was under the microscope as a quick, convenient indicator of body size [25] to ensure that differences in size between salamander samples did not bias our results.

Following log-transformation of head width data, we tested for differences in mean head width between samples with ANOVA. We analyzed time spent moving using Kaplan–Meier survival analysis. Kaplan–Meier, which generates a non-parametric log-rank test, is specifically designed to test for differences in time to a specific event. A unique aspect of this analysis is that it allows for the inclusion of subjects that do not actually reach the event in question by censoring those data points. This analysis plots the probability of survival versus time [30]. We censored the individual salamanders that did not stop moving within the 45 min trial. We ran analyses in JASP ver. 19.1, a free statistical software that runs on the R platform [31] (ver. 4.4.1).

3. Results

Infected salamanders had a mean ± 1 SE of 136.7 ± 12.4 metacercariae visible through the ventral surface of the trunk. As expected, salamanders from the stream with no snails had no visible metacercariae. As a convenient measurement of size, mean ± 1 SE for head width of infected salamanders was 6.28 ± 1.73 mm. Mean head width of uninfected salamanders was 6.92 ± 1.81 mm. There was no significant difference between the two samples in means of log-transformed head width (F_1,49 = 1.702; P = 0.198) and thus no difference in salamander size.

During 45 min movement trials, 7 of 26 (27%) infected salamanders did not stop moving and were therefore censored for survival analysis, whereas 14 of 26 (54%) uninfected salamanders were censored because they did not stop moving. Kaplan–Meier analysis indicated a significant difference between the two groups in time spent moving (logrank χ² = 4.713; P = 0.0299), with uninfected salamanders spending more time moving (Figure 2).

4. Discussion

We cannot completely rule out differences in mobility between the two study populations that are not related to parasitism. However, our results indicate that larval D. amphileucus infected with metacercariae of M. oregonensis stop moving sooner than uninfected larvae, an outcome that is not surprising given that metacercariae randomly encyst within salamander musculature [21]. Metacercarial infections of fish musculature can negatively affect swimming ability [26], and as a result, fish spend less time moving, possibly because of inflammation of muscle tissue and the consequent interference with oxygen consumption at the cellular level [32].

Infection by metacercariae of the trematode Ribeiroia ondatrae causes morphological malformations in frogs [33] that lead to lower survivorship, in part because of the reduced ability of malformed frogs to escape predators [2]. The result is that infected frogs are more likely to be eaten, thus enhancing transmission of trematodes to the definitive host. Salamanders have two different modes of locomotion. Swimming, which involves undulations of the trunk and tail, is important to escape predation [34,35]. In walking, the mode used when salamanders are in contact with the substrate [36], only the limbs are engaged. Because D. amphileucus larvae are largely benthic, walking is their primary mode of locomotion and the only mode used by our study animals. Given that the vast majority of metacercariae encyst within the trunk and tail of host D. amphileucus [21], it is possible that heavy metacercarial infection impairs a salamander’s ability to escape. However, we did not specifically test for the effect of parasitic infection on swimming, and further testing is necessary to determine whether behavioral modification by M. oregonensis enhances trematode transmission to the definitive host by increasing the host’s vulnerability to predation.

Even in the absence of an effect on swimming performance, the reduction in walking as a result of trematode infection can have a significant impact on the biology of these salamanders. Because walking is the mode of locomotion used during simple wandering, less time spent walking would likely have a negative impact on dispersal. In dispersal-limited organisms, random genetic changes in subpopulations can result in isolation by distance, even in the absence of isolating barriers [37,38]. Isolation by distance can in turn promote speciation [39]. Plethodontid salamanders frequently undergo genetic divergence over relatively small distances because of their low dispersal rates [40,41,42] and are commonly characterized by isolation by distance [43,44,45].

Until recently, D. amphileucus was a cryptic member of what was thought to be a single species that ranged across the southern Appalachian Mountains from Georgia to West Virginia [17]. Because the snail E. proxima occurs across much of this region [46], where it serves as the first intermediate host for M. oregonensis [47], infection by this parasite may have played a role in speciation in this group of salamanders by restricting dispersal. Parasites have been hypothesized to drive host divergence through several ways [48], including differential host adaptation across ecologically unique parasite communities [49]. Dispersal restriction would be an additional mechanism to promote host divergence.

We do not know if the relationship between M. orgonensis and D. amphileucus is peculiar to this host–parasite system or is more broadly characteristic of other systems. Regardless, our results suggest that the relationship between M. oregonensis and Desmognathus salamanders will prove fruitful for investigations into the role of parasites in host divergence, as well as other aspects of host–parasite relationships.

5. Conclusions

Our study demonstrates that larvae of the salamander D. amphileucus walk less when infected by metacercariae of the trematode M. oregonensis than salamander larvae that are not infected. Although we did not specifically test for the effect of parasitic infection on host swimming performance, reduced walking likely decreases dispersal and gene flow, which can significantly influence genetic divergence in these salamanders.