Freshwater trematodes differ from marine trematodes in patterns connected with division of labor

Collection and identification of infected snails

We collected snails by hand or net from four sites in central Vermont during the summer months (June–August) of 2019. The sites selected for this study were chosen based on the relatively high frequency of redia-producing infections during our collections in a previous summer (not all trematode species produce rediae- some only produce sporocysts). These sites are Waterbury Reservoir (WaR; Waterbury, Washington County), Ticklenaked Pond (TP; Ryegate, Caledonia County), North Montpelier Pond (NMP; Washington County) and Shelburne Pond (SP; Shelburne, Chittenden County), and were accessed via state-maintained swimming beaches (WaR) or boat launches (TP, NMP, SP). A few additional infected snails collected after the initial survey in fall 2019 and summer 2021 were included in the analysis opportunistically because we noticed that they had interesting features while sampling for another project.

Collected snails were identified to family using the key in Thorp & Covich (2010) and placed in individual containers of native water with constant illumination overnight to promote cercarial shedding. The water around each snail was scanned for cercariae with a dissecting microscope 1–2 days after collection to identify infected snails. Cercariae were photographed for reference and a sample was collected for DNA analysis. Cercariae for DNA analysis were added to ethanol (roughly 2/3 absolute ethanol, 1/3 water and cercariae to achieve roughly 70% ethanol), allowed to settle, and had excess liquid removed before being stored frozen. If it was not possible to collect enough cercariae (>5–10) for DNA analysis, rediae were collected following dissection and also stored frozen in a small amount of ~70% ethanol. The 70% ethanol was allowed to evaporate just prior to DNA extraction.

Infections that produced cercarial types consistent with redia-producing trematode species were housed in larger individual cups with native or distilled water until they could be dissected. Infections producing cercarial types not consistent with redia-producing trematodes were dissected to confirm the presence of sporocysts and ensure no redia-producing infections were missed. Some of these non-redia infections were used as heterospecifics in the Attack Tests (below).

Species identification

Trematode species were identified using DNA barcoding of the Folmer region of COI. DNA was extracted from stored cercaria (or occasionally rediae) samples after any remaining ethanol was allowed to evaporate using an Omega EZNA Tissue kit. DNA was amplified using Dice1/11 or Dice 1/14 primer pairs (Van Steenkiste et al., 2015). Sanger Sequencing was performed at an external facility (Yale University’s W. M. Keck Biotechnology Research Laboratory or Eurofins: eurofins.com). Forward and reverse trace files were aligned and viewed using SeqTrace (Stucky, 2012) and unreliable base calls were removed. Removal of unreliable bases include trimming ends where quality scores were not consistently above 20–30, examining and editing any mismatched calls between the forward and reverse sequence, and removing (setting to ‘N’) base calls from any site with apparent polymorphisms (these were very rare). Sequences from all samples were aligned using Clustal Omega (McWilliam et al., 2013) and the resulting percent identity (PI) matrix was used to calculate sequence divergence; sequences with <5% divergence were grouped into genetic groups that likely correspond to species (termed ‘COI species’ here; Gordy et al., 2016 and others use a similar approach). One sample from each COI species was then entered into BLAST (Altschul et al., 1990) for comparison with the NCBI database to determine a tentative species identification for each COI species.

Isolation of rediae

Redia-infected snails were crushed gently and removed from their shells, then separated into three regions: apical (gonad/digestive), middle, and head/foot. Each body section was teased apart in a separate Petri dish to isolate the rediae. Most snails were dissected in distilled water, but toward the end of the summer 2019 we discovered that dissecting in dilute saline (0.1% as suggested in Schell, 1970) substantially increased the activity levels of rediae, so dilute saline was used for dissections from that time on (starting August 1, 2019). For some infections, body sections were not separated for one of several reasons including 1) the infection was not discovered until after the snail was partially dissected or 2) the snail was damaged while crushing the shell, so it was not possible to adequately separate body sections.

Morphological measurements

Once the snail sections had been teased apart, we collected a sample of rediae for morphological analysis by pipetting a sample of water from the dissected material with all rediae in a haphazardly selected area of the Petri dish onto a microscope slide. This was performed without observing what size rediae were being collected, but when individual rediae were collected with a pipette it did not seem that any size redia were more difficult to collect, so this method should be relatively unbiased. We sampled a target of 100 rediae per section, and for any body section with fewer than 100 rediae, the snail tissue and surrounding water were examined carefully to locate as many rediae as possible.

Once the rediae were collected, each redia was photographed for measurement. To facilitate quicker data collection, rediae were not fixed or otherwise immobilized, but most that were dissected in water were not very active and the noise introduced by redial movement is likely dwarfed by the large differences in actual body size. Measurements were taken from photos using Spot Basic software (spotimaging.com/software/spot-basic/; by MS) or ImageJ (Rasband, 1997–2018; by ATN); measurements for a single infection were always made by the same researcher/software package and a subset of measurements that were repeated by the other researcher/software package did not differ substantially. Measurements taken were: (1) length, using a segmented line down the midline of the redia, (2) width, measured at the widest point (excluding obvious protrusions like collars and appendages) and (3) pharynx diameter (longest dimension; most were roughly circular in photos).

After measurements were complete, rediae were sorted by size and we examined the photos of the smallest five and largest five rediae sampled per infection for evidence of an anterior appendage/collar and posterior appendages (to assess differences in morphology) and cercaria/embryos/germinal balls (to assess reproductive potential). Our ability to identify these structures depend on the quality of the photo (including the position of the redia in the photo), so it was not always possible to definitively determine the presence or absence of every structure. Instead, presence was scored as 0 (definitely not present), one (probably not present), two (unclear), three (probably present), or four (definitely present). We refer to this below as the ‘presence score’ for each trait.

Activity measurements

For colonies with sufficient rediae, we selected the ten largest and ten smallest rediae remaining from the dissection after the haphazard morphology sample was removed. Each redia was isolated in dissection medium (water or dilute saline) in its own well in a 48-well cell culture plate. Photos were taken 2 s apart immediately after rediae were placed in the wells and again after 15 min. When it was not possible to fit the whole redia in one microscope field, only the anterior end of the redia was photographed. Photo pairs were superimposed using ImageJ and movement distance was recorded by measuring the distance between the same point on the redia in the two successive photos. Whichever part of the anterior end of the redia that travelled the furthest during the 2 s interval was used for measurement (File S1). We also used ImageJ to measure the total length of each redia.

Attack tests

Again for colonies with sufficient rediae, we assessed the tendency of smaller vs. larger rediae to attack colony mates, conspecifics and heterospecific trematodes. For this test, we selected the 20–30 largest and 20–30 smallest rediae remaining after the previous two tests and split them into three cell culture plate wells per redia size. The first ten rediae of each size were combined with an additional ten rediae of any size from the same infection (‘colony mates’). The next ten rediae of each size were combined with ten rediae from another colony of the same species (when available; species based on morphology; ‘conspecifics’). The final ten rediae of each size were combined with ten sporocysts from a colony of a different species (‘heterospecifics’). After allowing the trematodes to interact in the well, each well was observed for at least 30 s to determine whether any rediae were attacking (attached to other rediae or sporocysts by their mouthparts). We did not standardize or record the amount of time trematodes interacted prior to being observed, but recall it being around thirty to sixty minutes.

For most colonies, several researchers worked at two camera-equipped microscopes simultaneously to collect data on morphology, activity and attacks, meaning that even though the rediae for the attack tests were selected after those for the morphological measurements and activity tests, the delay in collecting these data was not great. The total time between snail dissection and the completion of photography (for all purposes) was typically less than 3 h.

Size designations

Rediae were assigned to a size category (‘small’ vs. ‘large’) in one of two ways; for infections with a bimodal distribution of volume (see below), we identified the break (bin with fewest observations) between peaks on the histogram and used this as a cutoff for size category. Histogram bin sizes were the default in R’s hist() function and varied by sample size. For infections without a bimodal size distribution, we compared the rediae in the smallest 25% of volumes to those in the largest 25%. This allowed us to compare small and large rediae even if they do not appear to form two distinct groups.

Culture

We attempted to maintain rediae in culture for several weeks to monitor the development of small rediae. Prior to dissection, the snail shell was wiped with water to remove visible dirt, wiped twice with 70% ethanol to disinfect, and then soaked for approximately 5 min in a 0.1% saline with 100 μg/mL gentamycin (“antibiotic saline”). The snail was dissected in antibiotic saline in a sterile petri dish in a laminar flow hood. Rediae were rinsed three times in 1,000 μL antibiotic saline. 100 μL of rediae (about 7–22 rediae) were then transferred to each of 12 wells of a 24-well cell culture plate. Because we were not sure what conditions would best promote redial survival and growth, the wells contained one of six different treatments (two replicates of each): (1) redia co-cultured with snail cells (Biomphalaria galbrata embryonic cells; Bge) in a 1:1 mixture of Bge Medium and Medium F (described below), (2) redia co-cultured with snail cells in a 1:3 Bge:F media, (3) redia co-cultured with snail cells in 100% Bge Medium, (4) redia alone in a 1:1 mixture of Bge:F media, (5) redia alone in a 3:1 mixture of Bge:F media, and (6) redia alone in 100% Bge Medium. Rediae were checked every 1–2 weeks for 11 weeks (77 days) to assess survival and growth. Survival was assessed by watching each redia for 5 s. If any motion was detected, it was considered active (and alive). Growth was assessed by photographing small rediae born in culture and measuring their body length. For small rediae that were active, multiple photos were taken to reflect both their stretched and contracted length.

The snail cells used in this study—Bge Cell Line, NR-40248—were provided by the NIAID Schistosomiasis Resource Center distributed through BEI Resources, NIAID, NIH (https://www.beiresources.org/Catalog/cellBanks/NR-40248.aspx; Taus et al., 2013). Prior to inclusion in this study, Bge cells were cultured in Bge Medium using the formulation provided with the cells. Co-culture with Bge cells has been shown to promote the longevity and development of a variety of freshwater trematode species (reviewed in Coustau & Yoshino, 2000). Medium F was shown to promote redial survival in a study with marine trematodes (Lloyd & Poulin, 2011), and the recipe for Medium F can be found in the supplemental materials of that article.

Analysis

All analysis was performed in R version 4.2.1 (R Core Team, 2022). Unreliable measurements, for example from damaged rediae, were removed prior to analysis. For most analyses, all rediae from a single infected snail were pooled for analysis. Analysis was performed on each infection individually; this allowed us to easily determine how consistent patterns were among infections of the same genetic group. We did not adjust our p-value cutoff for multiple comparisons (we used α = 0.05), so we expect about 5% of ‘statistically significant’ results to be false positives. Below we describe how we tested for each DOL-associated pattern.

Bimodal Volume: to test whether redia volumes showed a bimodal distribution, we first estimated the volume of each redia assuming its shape is roughly cylindrical. We then performed a Shapiro-Wilk Normality Test on the log-transformed volume. For infections that differed significantly from a normal distribution, we examined histograms of the log(volume) to determine whether there were two distinct peaks.

Appendages: to test whether smaller rediae have more pronounced appendages (anterior appendage/collar and posterior appendages) than larger rediae, we used a Wilcoxon Rank Sum Test to test the alternative hypothesis that the smallest five rediae had a higher average presence score for each appendage than the largest five redia from each infection.

Pharynx: to test whether smaller rediae have larger pharynges relative to their body size, we first estimated the volume of each pharynx assuming its shape is roughly spherical and divided pharynx volume by redia volume to obtain a relative pharynx volume. We compared relative pharynx volume between large and small rediae using a one-tailed Wilcoxon Rank Sum test. We also compared absolute pharynx volume in the same way.

Activity: before comparing the distance moved in two seconds by small vs. large rediae, we first checked that the rediae selected for analysis were indeed ‘small’ and ‘large’ in the context of all rediae measured (for a few infections, we failed to select the smallest rediae because they were rare or less obvious under the dissecting microscope). We re-classified any redia sizes that were not internally consistent (e.g. any ‘small’ redia that exceeded the length of any ‘large rediae) or not consistent with the size categories established for other analyses (see ‘Size designations’ above). We then averaged the measurements made at 0 and 15 min and compared the relative distance moved (distance moved/body length) for small vs. large rediae for each infected snail using the Wilcoxon Rank Sum Test.

Reproductive potential: we recorded the maximum presence score for embryos for the smallest five rediae in each infection.

Distribution of morphs in different body sections: we calculated the proportion (with 95% confidence interval) of small and large rediae (intermediates excluded) found in each body section that were designated ‘small’. We compared these proportions among body sections to determine whether there was a higher proportion of small redia in the anterior regions of the snail, especially the head/foot (or mid section if no rediae were found in the head/foot).

Body volume

The results examining the distribution of rediae volumes were somewhat mixed. Even though redia volumes differed from a normal distribution for 54/61 infections, the distribution of volumes in only about 23 infections appeared noticeably bimodal (Fig. 1, Table 4), and none showed the substantial gap in volumes between small and large redia that is more typical of the marine trematodes. Patterns within superfamilies and even within COI species were also quite variable.

Reproductive potential of small rediae

Because we did not conduct a detailed morphological assessment of small rediae (e.g. by fixing and staining them), it was not always clear from our photos whether each individual small redia sampled contained germinal balls or embryos. Nonetheless, we were at least relatively sure (presence score 3–4) that embryos were visible in the smallest five rediae in 43/61 infections, and for none of the infections were we confident that at least one of the smallest five rediae did not have the potential to reproduce (Table 4). For many small rediae, embryos were easily observable (e.g. Fig. 2).

Appendages

For 50/61 infections, anterior and posterior appendages were not more prominent on small rediae than on large rediae either because both small and large rediae had appendages (e.g. Fig. 2A) or because neither did (e.g. Fig. 2B). The few infections that do show an apparent difference do not appear concentrated within a single superfamily or COI species.

Pharynx size

For 47/61 infections, relative pharynx size was larger in small rediae than large rediae (Table 4). For the remaining infections, sample size may have contributed to an inability to detect a difference: four infections did not have enough data to run the statistical test, and of the ten infections that did not show a statistically significant difference in relative pharynx size, nine had five or fewer pharynx measurements for at least one redia size (Files S4 and S5). Absolute pharynx size was generally a little larger in large rediae (significantly so for 35/61 infections; Files S4 and S5). Pharynx sizes in some example rediae can be seen in Fig. 2.

Distribution of rediae in snail

Only seven of the 37 infections for which rediae were separated by body region showed a significantly higher proportion of small rediae in the foot (Table 4). For fifteen of the remaining infections, the proportion of small rediae tended to be higher in the anterior regions of the snail, but confidence intervals overlapped.

Activity

We recorded data on the activity levels of rediae from 28 infections; 8/28 were dissected in dilute saline and the remainder were dissected in water. The distance that water-dissected rediae moved in 2 s was substantially lower than the distance moved by saline-dissected rediae in the same time: most (75%) water-dissected rediae moved less than 7 μm, while most (75%) saline-dissected rediae moved more than 13 μm. Because of this substantial discrepancy, we decided to exclude all measurements from water-dissected rediae, leaving only eight infections for analysis. Of these eight infections, most showed a tendency for large rediae to move further in 2 s than small rediae (Fig. 3A, p < 0.05 for 2/8 infections), but there did not seem to be a consistent difference in the distance moved for small and large rediae relative to their body size (p > 0.10 for all infections, Fig. 3B).

Attacks

We recorded the frequency of attacks by large vs. small rediae for only two infections dissected in dilute saline (and another twenty dissected in water). We observed attacks in one of the saline-dissected infections, with two attacks by large rediae and one by a small redia. Attacks were on a same-colony redia, a heterospecific cercaria and a heterospecific sporocyst (Table 4). We also observed attacks in two of the twenty water-dissected infections (not reported in Table 4 due to impact of osmotic stress). In one infection, we observed only a single attack by a large redia on another damaged redia from the same infection; it was unclear whether the attacking redia caused the damage or not. In a second water-dissected infection, we observed two large rediae and no small rediae attacking. Both attacks were on heterospecific sporocysts.

Culture

We monitored the rediae from only a single infection in culture, but decided to report our findings nonetheless. Rediae were taken from a snail in the family Planorbidae collected from NMP in October 2019. This infection is not included in Tables 3 and 4 because the rediae were not tested for any of the other patterns in this study (no volume distribution, pharynx size, activity, etc.); the rediae’s COI sequence was 96–98% similar to the P177 COI species from Table 3.

Rediae appeared to have higher survival when co-cultured with snail cells: by day 33, none of the approximately 84 rediae in the six wells without snail cells remained active, while all six wells with snail cells had at least some active rediae (7–35% of rediae per well, 88 rediae total). Four of the six wells with snail cells still had active rediae after 77 days (7–30% of rediae per well), after which we stopped recording data. It was unclear from our results whether the culture media had any significant influence on redia survival (File S7B).

In addition to the “adult” rediae that were initially added to each well, small “baby” rediae appeared in four of the wells (one with snail cells, three without). The number of small rediae per well increased over time for most wells (three increasing to six small rediae in the well containing snail cells; two increasing to three in two of the wells without snails cells; no increase in the final well without snail cells; File S7C). Like the larger rediae, the small rediae in wells without snail cells all ceased activity after 21 days (and some much earlier), but the small rediae in the well with snail cells all remained active throughout the 77 days over which they were monitored. We were unable to differentiate between the small rediae in each well to allow tracking of individual rediae sizes over time, but there is no apparent change in size of the small rediae when tracked collectively over 10 weeks (day 6, when they first appeared, to day 77, when we stopped keeping records; Fig. 4).

Additional observations

For a few infections collected in summer 2019 and several infections collected either fall 2019 or summer 2021, we observed one or more rediae with notably large pharynges compared with the other rediae in the infection (Fig. 5). We struggled to define exactly what constitutes a ‘large’ pharynx, but large pharynges were almost always identifiable as outliers (>[Q3 + 1.5*IQR]) from the distribution of pharynx sizes. Some infections without notably large-pharynx rediae did sometimes have one or a few rediae whose pharynx sizes were outliers, but pharynx-size outliers were more common in infections with notably large-pharynx rediae. For infections with large-pharynx redia (based on a visual inspection of photos), 5–15% of rediae had outlier pharynx sizes.

The large-pharynx rediae were not noticeably different in size from other large rediae (Fig. 6), though for some infections they did tend to be a bit wider relative to their length (e.g., Fig. 6D) and at least some appeared capable of reproducing (developing cercariae were observed in several large-pharynx rediae). The large-pharynx rediae often appeared to have a larger gut and darker body coloration than standard-pharynx rediae (Fig. 5). Large-pharynx rediae were noted in eight infections of three COI species: P153 (one infection of one sampled), P177 (two infections of two sampled), and V141 (five infections of six sampled).