Demographic and genetic factors in the recovery or demise of ex situ populations following a severe bottleneck in fifteen species of Hawaiian tree snails

Captive population husbandry

The captive-breeding facility for Hawaiian tree snails at the University of Hawaiʻi at Mānoa was initiated in 1991. Small numbers of snails (<15) were used to found captive populations between 1991 and 2006 (Table 2). Adults were preferentially collected when possible, but the ages of founders varied among species. Adults were identified by a “lipped” shell aperture, or a thickened edge of the shell opening, indicating sexual maturity (Kobayashi & Hadfield, 1996). The relatedness of the founding individuals of each species was unknown, but they were often collected over relatively small distances (less than 10 m). For two species, A. fulgens and A. sowerbyana, founders were collected from multiple locations, and the resulting populations were kept in separate cages until populations declined below a combined total of ten individuals, at which point a single combined population was established for each species.

Cage environments were controlled for temperature and humidity, regulated to mimic average field conditions at elevations of 600–650 m on the island of O’ahu (Hadfield, Holland & Olival, 2004). Hawaiian tree snails consume diverse microbial biofilms on leaf surfaces in the wild (Hadfield, Holland & Olival, 2004; O’Rorke et al., 2015), so leafy branches were collected from the mountains and placed in cages when they were cleaned biweekly, along with agar-cultured, calcium-supplemented mold (Cladosporium sp.) originally isolated from a native snail-host plant (Hadfield, Holland & Olival, 2004). This diet was previously shown to either increase growth rate compared with wild populations (Kobayashi & Hadfield, 1996), or at least produce an equivalent growth rate (Hadfield, Holland & Olival, 2004).

During bi-weekly cage cleanings, demographic information was collected including births, deaths, and total numbers of juveniles (<1 year-old), subadults (between 1 year and lipped), and adults (snails with a lipped shell). Individual snails were not labeled or tracked, but at death, each snail was individually preserved in 95% ethanol in a tube labeled with the species, source cage, date of preservation (within 2 weeks of death), and shell length and width.

Using biweekly reports we compiled a database of births, deaths, numbers of juveniles, subadults, and adults, and the total number of snails in each cage. We also noted any suspected cases of self-fertilization, in which an isolated adult snail produced offspring. Tree snails may store sperm, as in other hermaphroditic snails (Jordaens, Dillen & Backeljau, 2007), so in cases where a snail reached maturity when other mature snails were present in the cage, then later gave birth to offspring after some time of being isolated, we were not able to discern whether it was due to long-term sperm storage or self-fertilization. However, in some cases snails were isolated before maturity, and in these cases we could be sure any births were due to self-fertilization.

Using preserved snail samples we compiled a database of life-history information for individual snails including shell length and width at death, cage name, and date of preservation. Birth dates and an approximate date of sexual maturity were estimated for each snail based on shell growth rates obtained from lab records and field studies (Hadfield & Mountain, 1980; Severns, 1981; Hadfield & Miller, 1989; Kobayashi & Hadfield, 1996).

Tissue sample collection, DNA extraction and amplification

Tissue samples from four species, A. apexfulva, A. fulgens, A. sowerbyana, and P. variabilis, were collected from the entire set of preserved and living captive snails using sterile techniques. For live captive snails larger than 12 mm in shell length, samples were collected using nonlethal methods, cutting a very thin slice of tissue from the posterior tip of the foot, following the method of Thacker & Hadfield (2000), and storing in 100% ethanol for subsequent DNA extraction. This method has been used with both captive and wild snails for several decades, without causing any mortalities, as determined by observing captive snails post-sampling. DNA was extracted from tissue samples using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol, and eluted using two 200 µl washes of elution buffer (10 mM Tris-Cl, 0.5 mM EDTA). Live captive snails smaller than 12 mm were sampled by swabbing mucus from the snail’s body using a sterile polyester-tipped swab, and storing the tip of the swab in a sterile, dry tube at −20 °C until extraction. DNA was extracted from mucus using a QIAamp DNA Micro Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Carrier RNA was added to the cell lysis buffer according to the manufacturer’s protocol for very small amounts of DNA, and DNA was eluted in 50 µl of elution buffer.

Each sample was genotyped at eleven microsatellite loci (Erickson & Hadfield, 2008; Sischo et al., in press), using the recommended amplification protocols for each primer set (Table 3). Bovine serum albumin (4 µM) was added to amplification reactions for preserved specimens that failed to amplify, likely due to the presence of hemocyanin derivatives (Akane et al., 1994). Genotyping reactions were performed by the Center for Genomic, Proteomic, and Bioinformatic Research at the University of Hawaiʻi at Mānoa. Amplification, genotyping and scoring were performed at least twice for each individual at all loci. Peakscanner version 1.0 (Applied Biosystems 2006) was used to visualize and identify alleles. Results from individual samples that failed to amplify at more than half of the loci were discarded.

Environmental data

Shapefiles containing point data, including latitude, longitude, and elevation were obtained from recent surveys of Hawaiian tree snails (Hawaii State Department of Land and Natural Resources, pers. comm., 2013). Using the software program ArcGIS (ESRI 2013, version 10.2 for desktop), annual mean temperature (°C), mean monthly precipitation for all twelve months, and annual mean precipitation (mm) were obtained for each snail-occurrence point from publicly-available layers (http://rainfall.geography.hawaii.edu, downloaded August 15, 2013; Giambelluca et al., 2013). Several variables, including minimum and maximum mean temperature, monthly precipitation, and annual precipitation were then identified for each species across the known species’ range.

Statistical analysis

Population trends

All captive tree snail populations in the captive-breeding laboratory at the University of Hawaiʻi have declined since reaching peak numbers, with the largest declines occurring since 2009 (Figs. 2A–2C). All lab populations that exceeded 100 snails have at least 25 snails remaining per species following the decline, and appear to have stabilized (Fig. 2A). All species whose peak population number was below 100 snails now have less than 10 snails (Fig. 2B), or have died out (Fig. 2C).

Fecundity

Fecundity was significantly higher in the founding individuals of captive species that recovered to at least 100 individuals in captivity, compared with the two other groups (r² = 0.16, P < 0.01; Fig. 3). When considered collectively, fecundity declined for all species over generations (r² = 0.46, P < 0.01) and over birth years (r² = 0.041, P < 0.001). The genera Partulina and Achatinella did not differ significantly in founder fecundity (t = 1.66, P = 0.10). Fecundity was not correlated with the number of adults in a cage (r² = 0.0005, P = 0.66).

Survival

In addition to having more offspring per individual, offspring in the most successful species had significantly higher survival to maturity than species that did not exceed 100 individuals in captivity (X² = 6.35, P = 0.012; Fig. 4), but did not significantly differ from species now extirpated from captivity (X² = 0.047, P = 0.83).

Genetic variation

Genetic samples were obtained from four species, A. apexfulva, A. fulgens, A. sowerbyana, and P. variabilis (Table 4), for comparison with published results from sister species A. fuscobasis (Sischo et al., in press) and A. lila (Price & Hadfield, 2014). Some correlations between heterozygosity and survival were observed within species (Fig. 5). Overall, heterozygosity was positively correlated with survival to maturity in captive-born individuals of A. sowerbyana (r² = 0.019, P = 0.014), but there was no significant trend within generations. In the Kului Gulch population of A. fulgens, there was a significant decline in observed heterozygosity between the founders and the first generation of offspring (t = 2.71, P = 0.017), and none of the offspring survived to maturity. The other populations of A. fulgens also had declines in heterozygosity between the founder and F₁ generations, but not significantly so, and only one individual in this species, from the Pia Valley population, survived to maturity. In the third generation of P. variabilis, surviving individuals had significantly higher heterozygosity than those that did not survive (t = 3.49, P < 0.001).

One species, A. apexfulva, did not show a similar correlation between heterozygosity and survival. Heterozygosity was slightly higher in the first generation of offspring than in the founders (Fig. 5), and individuals that did not survive to maturity had higher heterozygosity (0.48 ± 0.12), though not significantly so, than individuals that reached maturity (0.42 ± 0.20, t = 1.14, P = 0.26).

Inbreeding coefficients were significantly high in all of the most successful species in which we examined population genetics (Table 5). For one species that died out in the laboratory, A. sowerbyana, a decline in survival to maturity corresponded with an increase in the inbreeding coefficient over time. The first generation of adults from both populations of A. sowerbyana did not exhibit high inbreeding coefficients (Fis_{Peahinaia(F1)} = − 0.112, P = 0.87, Fis_{Pulcherrima(F1)} = 0.029, P = 0.44), but the second generation of adults did (Fis_{Peahinaia (F2)} = 0.197, P < 0.01, Fis_{Pulcherrima(F1)} = 0.238, P = 0.060).

Environment

Minimum and maximum mean annual temperatures (°C) within the species’ natural ranges were significantly lower in species that are now extirpated from captivity than in species that remain in captivity ($X_{\min }^2=7.71$ P_min = 0.021; $X_{\max }^2=8.94$, P_max = 0.012; Fig. 6). Monthly ($X_{\mathrm{monthly}\hspace{0.17em}\min }^2=0.075$, P_{monthly min} = 0.96; $X_{\mathrm{monthly}\hspace{0.17em}\max }^2=0.18$, P_{monthly max} = 0.91) and annual precipitation ($X_{\mathrm{annual}\hspace{0.17em}\min }^2=0.067$, P_annual min = 0.96; $X_{\mathrm{annual}\hspace{0.17em}\max }^2=0.022$, P_annual max = 0.90; Fig. 7) varied substantially among and even within species, but was not predictive of survival in the captive breeding facility.

Self-fertilization

Self-fertilization or long-term sperm storage was observed in multiple species from two achatinelline genera in captivity. Individuals of A. fulgens, A. mustelina, and P. mighelsiana, isolated for multiple years before maturity, gave birth to offspring. Individual Achatinella decipiens, A. sowerbyana, Partulina physa, and P. semicarinata gave birth after being isolated for 10 months or more.  In contrast, for several other species that died out in captivity, adults were isolated for years and never gave birth. The lone A. apexfulva that remains in the lab, the last known individual of its species, reached maturity in 2012, as indicated by forming a “lip” on the shell aperture, but has not produced any offspring.

Implications for in situ and ex situ conservation efforts

Nonmarine mollusks have experienced very high extinction rates (Lydeard et al., 2004), yet little information is available to provide management targets for many of the species that remain. The present study provides several benchmarks for both ex situ and in situ populations of Hawaiian tree snails. Populations may be predicted to decline to extinction if numbers fall below two offspring per adult per year and 20% survival to maturity.

Our 20-year study provides empirical support for management-recovery targets of greater than 450 individuals for the remaining populations of rare Hawaiian tree snails in the wild, to buffer against stochastic threats such as disease or hurricanes. In this study, only those captive populations that recovered to more than 450 individuals prior to the recent declines were able to persist in numbers that suggest long-term stability. These recovery targets are necessary in wild populations to retain genetic diversity and maintain evolutionary potential as conditions change. The species with the most successful recovery in captivity, Achatinella lila, experienced a loss of allelic diversity over time (Price & Hadfield, 2014), perhaps due to widespread selective sweeps typical in captive populations as they adapt to conditions in captivity, or due to genetic drift, as is common in small populations (Frankham, 2012). For wild populations, increasing variability in temperature and precipitation may produce similar selective sweeps as they adapt to warmer, drier temperatures (Boutin & Lane, 2014; Charmantier & Gienapp, 2014; Crozier & Hutchings, 2014).

Invertebrates have historically received less funding than charismatic vertebrates, but are foundational to the survival of ecosystems and the maintenance of biodiversity. Ex situ programs for invertebrates must be maintained with the same rigor given to programs for charismatic species, with careful attention given to issues of disease, environmental conditions, and genetic stability. Prior to the recent declines, many species in the ex situ tree snail program at the University of Hawaiʻi were on a demographic trajectory toward recovery and success (Figs. 2A–2C). Today, more than twenty years after its initiation, only a few species are present in substantial numbers, and by many benchmarks, the ex situ effort for Hawaiian tree snails has not been successful. However, the source populations are now extinct for two species with sizeable ex situ populations remaining in captivity, A. fuscobasis and A. lila (Table 1). Furthermore, the ex situ effort has preserved the largest populations that exist today for these two species. A third species, A. apexfulva, no longer exists in the wild and has not been observed there for over 15 years. The one individual in the captive breeding facility represents the last of its species. In these three cases, snails exist today that would not have remained without the ex situ effort. Future efforts may be more successful if larger numbers of snails are brought to the captive-breeding facility to initiate captive populations, environmental-chamber settings closely mimic conditions in the wild for each individual species, fitness criteria are used to optimize husbandry, and genetic diversity is maintained.