Range reduction of the Oblong Rocksnail, Leptoxis compacta, shapes riverscape genetic patterns

Many freshwater gastropod species face extinction, including 79% of species in the family Pleuroceridae. The Oblong Rocksnail, Leptoxis compacta, is a narrow range endemic pleurocerid from the Cahaba River basin in central Alabama that has seen rapid range contraction in the last 100 years. Such a decline is expected to negatively affect genetic diversity in the species. However, precise patterns of genetic variation and gene flow across the restricted range of L. compacta are unknown. This lack of information limits our understanding of human impacts on the Cahaba River system and Pleuroceridae. Here, we show that L. compacta has likely seen a species-wide decline in genetic diversity, but remaining populations have relatively high genetic diversity. We also report a contemporary range extension compared to the last published survey. Leptoxis compacta does not display an isolation by distance pattern, contrasting patterns seen in many riverine taxa. Our findings also indicate that historical range contraction has resulted in the absence of common genetic patterns seen in many riverine taxa like isolation by distance as the small distribution of L. compacta allows for relatively unrestricted gene flow across its remaining range despite limited dispersal abilities. Two collection sites had higher genetic diversity than others, and broodstock sites for future captive propagation and reintroduction efforts should utilize sites identified here as having the highest genetic diversity. Broadly, our results support the hypothesis that range contraction will result in the reduction of species-wide genetic diversity, and common riverscape genetic patterns cannot be assumed to be present in species facing extinction risk.

Introduction diversity in downstream populations and isolation by distance (Hughes, Schmidt & Finn, 2009;114 Paz-Vinas et al., 2015). However, few studies have tested for such patterns in riverine species 115 that have undergone drastic range reduction, and no such study has been done for a range 116 restricted pleurocerid. 117 One pleurocerid in need of more research is the Oblong Rocksnail, Leptoxis compacta 118 (Figs. 1,2). This species is a narrow range endemic known historically from the middle Cahaba 119 River and a single tributary in central Alabama, USA ( Fig. 2; Goodrich, 1922). Until recently, 120 Leptoxis compacta was considered extinct as it had not been collected, or at least identified 121 correctly, from 1935 to 2011 (Goodrich, 1941;Johnson et al., 2006;Whelan, Johnson & Harris, 122 2012). As early as 1941, the decline of L. compacta was documented (Goodrich, 1941), and the 123 species now occupies less than 5% of its historical range ( In this study, we generated a dataset of thousands of single nucleotide polymorphisms 138 (SNPs) to answer questions about conservation and riverscape genetics of L. compacta. Given 139 the drastic range decline suffered by L. compacta, we set out to test the following hypotheses: 1) 140 Leptoxis compacta has undergone a severe genetic bottleneck and 2) genetic diversity of L. 141 compacta is considerably lower than L. ampla, a sympatric and wider ranging species. We also 142 examined how genetic diversity of L. compacta varies across its current range, specifically 143 assessing whether broad patterns seen in many other riverine taxa like isolation by distance and 144 strong genetic structure are seen in L. compacta.    After assembly, RAD-loci were filtered for missing data using the STACKS program 206 populations. In order to pass filtering steps, a RAD-locus had to be present in 75% of individuals 207 from any given collection site and also present at three collection sites. RAD-loci that had a 208 minimum minor allele frequency of less than 2.5% or heterozygosity higher than 50% were 209 removed to limit the influence of paralogy and misassembly on final datasets. Sequencing  were determined to be closely related by KING so no further dataset filtering was done.

215
After filtering, a dataset that included all SNPs per RAD-locus and a dataset with only 216 one random SNP per RAD-locus were generated. We assume that RAD-loci are unlinked and 217 that the one SNP per RAD-locus dataset had zero linkage disequilibrium. Analyses employed the 218 one SNP per RAD-locus dataset, unless otherwise noted.  We also tested for a pattern of isolation by distance by measuring the correlation between 230 pairwise FST values and geographical distance between collection sites. Pairwise FST values were 231 calculated using the Weir and Cockerham (1984) method with the R package hierfstat (Goudet,232 2005). Stream distance was measured in Google Earth by tracing paths between collection sites 233 along the Cahaba River (see Table 2). River distance was used, rather than straight line distances,  During survey work, we collected L. compacta from Cahaba River at old Marvel slab 281 upstream to Cahaba River at Booth's Ford (Fig. 3). All sites except Cahaba River at Shades

282
Creek are sites where L. compacta was not found during survey work over the least 30 years.

308
The number of private alleles at each site ranged from 28-262 (Table 1)

319
Pairwise FST values among sites ranged from 0.0-0.055 (Table 2). We found no evidence  (Fig. 3). fineRADstructure analyses corroborated ADMIXTURE analyses as two 333 semi-distinct groupings were recovered by fineRADstructure (Fig. 5, Supplementary Fig. 1).  Genetic diversity across a small landscape 364 The two most distant collection sites in this study were separated by a smaller distance 365 (9.2 km) than all but two sites sampled for L. ampla in a previous study (Whelan et al., 2019). 366 Therefore, it is difficult to make direct comparisons between genetic patterns of L. ampla and L. 367 compacta. However, we can leverage differences in geographical scale between the two studies 368 to make inferences about fine-scale versus long-distance genetic patterns in pleurocerids. FST 369 values among L. compacta collection sites (Table 2)  relevance. Overall, these data indicate that pleurocerid riverscape genetic patterns across small 374 distances will not always follow common patterns such as isolation by distance and increased 375 genetic diversity at more downstream collection sites. This is likely attributable to gene flow and 376 random drift that prevent the establishment of genetic patterns typically seen across more 377 geographically separated collection sites. From a historical standpoint, we hypothesize that L.  sites with inferred subpopulation structure ( Fig. 3; Table 2). An alternative explanation for the   decline was so drastic that L. compacta was considered extinct less than a decade ago.

418
Conservation efforts are needed to ensure the long-term survival of L. compacta as the species is 419 at risk from both chronic habitat degradation and one-time catastrophic events. Two potential 420 management strategies for L. compacta are habitat restoration and reintroduction with captively 421 reared individuals.

422
In this study, we report an 8.26 km known range extension for L. compacta. One site, is much less modified than most other systems in the southeast. Our findings suggest that 432 imperiled gastropods will benefit from water quality and habitat improvements even in relatively 433 "pristine" river systems. Improving habitat, or identifying suitable habitat, will be a necessary 434 starting point for L. compacta reintroduction efforts.

435
In addition to having a small range, L. compacta only exists along a single river path.  Even though L. compacta was considered extinct less than a decade ago, we now know 455 more about this species than most other freshwater gastropods. This is helpful for conservation 456 of L. compacta as the biggest barrier to effective management strategies for most freshwater 457 gastropods is a lack of data. Future research efforts should focus on differences in dispersal 458 dynamics among pleurocerids and causes of differences in riverscape genetic patterns seen between L. ampla and L. compacta. As more population genomic data becomes available for 460 pleurocerids, we will be better suited to develop strategies to conserve these critically important 461 components of many North American riverine ecosystems.