Spatial scale and structure of complex life cycle trematode parasite communities in streams

Abstract

By considering the role of site-level factors and dispersal, metacommunity concepts have advanced our understanding of the processes that structure ecological communities. In dendritic systems, like streams and rivers, these processes may be impacted by network connectivity and unidirectional current. Streams and rivers are central to the dispersal of many pathogens, including parasites with complex, multi-host life cycles. Patterns in parasite distribution and diversity are often driven by host dispersal. We conducted two studies at different spatial scales (within and across stream networks) to investigate the importance of local and regional processes that structure trematode (parasitic flatworms) communities in streams. First, we examined trematode communities in first-intermediate host snails (Elimia proxima) in a survey of Appalachian headwater streams within the Upper New River Basin to assess regional turnover in community structure. We analyzed trematode communities based on both morphotype (visual identification) and haplotype (molecular identification), as cryptic diversity in larval trematodes could mask important community-level variation. Second, we examined communities at multiple sites (headwaters and main stem) within a stream network to assess potential roles of network position and downstream drift. Across stream networks, we found a broad scale spatial pattern in morphotype- and haplotype-defined communities due to regional turnover in the dominant parasite type. This pattern was correlated with elevation, but not with any other environmental factors. Additionally, we found evidence of multiple species within morphotypes, and greater genetic diversity in parasites with hosts limited to in-stream dispersal. Within network parasite prevalence, for at least some parasite taxa, was related to several site-level factors (elevation, snail density and stream depth), and total prevalence decreased from headwaters to main stem. Variation in the distribution and diversity of parasites at the regional scale may reflect differences in the abilities of hosts to disperse across the landscape. Within a stream network, species-environment relationships may counter the effects of downstream dispersal on community structure.

Fig 1. Map of landscape-level study sites.

Location of the 20 landscape-level study sites in southwestern Virginia and northwestern North Carolina within the Upper New River Basin (demarcated by dashed purple line).

Fig 2. Map and diagram of within-network study sites.

(a) Location of eight within-network study sites in the Big Reed Island Creek drainage in Carroll Co., VA, and (b) simplified stream network diagram indicating relative position of sites within network configuration (arrows indicate direction of flow). Main stem sites (MN) and headwater sites (HD, AB, BB).

Fig 3. Box and whisker plot of trematode morphotype prevalence.

Median and range of prevalence of each trematode morphotype across (a) all 20 landscape-level study sites and (b) all 8 within-network sites. Whiskers represent 1.5 interquartile range. Trematode morphotypes: Metagonimoides oregonensis (META); virgulate (VIRG); cotylomicrocercous type (COTYL); Sanguinicola sp. (SANG); and monostome type (MONO). In the within-network study (b) virgulate infections were categorized as small virgulate (SVIRG) and large virgulate (LVIRG).

Fig 4. Stacked bar chart displaying the number and relative proportion of unique haplotypes within each morphotype.

From the landscape-level study, the number of trematode samples (total N = 491 samples) per unique haplotype identified within each of the five morphotypes. Abbreviations and total number of unique haplotypes identified for each morphotype are as follows: cotylomicrocercous type (COTYL; n = 9 haplotypes); Metagonimoides oregonensis (META; n = 7 haplotypes); virgulate (VIRG; n = 8 haplotypes); Sanguinicola sp. (SANG; n = 1 haplotype); and monostome type (MONO; n = 1 haplotype). Different colors within bars show the proportion of samples per unique haplotype.

Fig 5. Plots of PCoA axes for landscape-level study morphotype and haplotype communities.

PCoA of Bray-Curtis dissimilarity from landscape-level study for (a) morphotype defined communities and (b) haplotype defined communities. Percent variance explained by each principal coordinate included in axis label. Sites are labeled with abbreviations (see S1 Table in S1 Appendix) and categorized by prevalence of cotylomicrocercous type infections defined as: High > 10%; Medium = 5–10%; and Low < 5% of snails infected. “None” indicates absence of cotylomicrocercous type.