Parentage test in broad-snouted caimans (Caiman latirostris, Crocodylidae) using microsatellite DNA

In this study, microsatellite markers, developed for Alligator mississipiensis and Caiman latirostris, were used to assess parentage among individuals from the captive colony of Caiman latirostris at the University of São Paulo, in Piracicaba, São Paulo, Brazil. Many of the females in the colony were full siblings, which made maternal identification difficult due to genotypic similarity. Even so, the most likely mother could be identified unambiguously among offspring in most of the clutches studied. Two non-parental females displayed maternal behavior which would have misled managers in assigning maternity based on behavior alone. This set of variable loci demonstrates the utility of parentage testing in captive propagation programs.

Caiman latirostris is a medium-sized crocodilian that inhabits the wetlands and swamps of southeastern South America. The geographic distribution of the species covers the hydrographic basins of the Paraná and São Francisco Rivers, as well as a large number of small coastal drainage systems, from northeastern Brazil to northeastern Uruguay (Verdade, 1998;Verdade and Piña, 2006). The state of São Paulo, where this study was undertaken, is located in the center of the species range. Caiman latirostris was considered an endangered species in Brazil from 1972 to 2003 (Vanzolini, 1972;Groombridge, 1982;Bernardes et al., 1990;IBAMA 2003). The main causes for the decline in original populations were poaching for the leather trade, and habitat destruction, primarily for agricultural use (Brazaitis et al., 1988;Verdade, 1997).
Since the late 1980's, the Caiman latirostris conservation program developed by the University of São Paulo (ESALQ, Piracicaba, São Paulo, Brazil) has been successful in breeding this species in captivity (Verdade and Sarkis, 1998;. Due to the lack of informa-tion on nesting sites in the wild, and as this species is relatively common in Brazilian zoos, commercial farming operations have been proposed as the most feasible conservation strategy for the species in southern Brazil (Verdade, 1997(Verdade, , 2001. Captive propagation efforts need to be guided by well-structured genetic management of the colony to prevent possible problems, such as founder effect, genetic drift and inbreeding depression (Ballou, 1992). Genetic management in the University of São Paulo captive colony is based on the establishment of a studbook in which individual pedigrees can be assessed and reproductive groups assembled, priority being given to nonrelated or leastrelated individuals (Verdade and Kassouf-Perina, 1993).
Molecular markers have been shown to be important tools in ecological and genetic research (Palo et al., 1995;Verdade et al., 2002). Microsatellites are among the best markers for parentage identification due to their high polymorphism (Craighead et al., 1995;Garcia-Moreno et al., 1996;Davis et al., 2001a), so that with enough markers, overall exclusion probabilities of 99.8% can be obtained.
Microsatellite markers specifically developed for Alligator mississipiensis were tested with DNA from 21 species of the eight extant crocodilian genera (Glenn et al., 1996(Glenn et al., , 1998. The tested primers were more efficient when amplifying orthologous loci in the DNA of species from the Alligatorinae subfamily than those from the Crocodylidae subfamily. However, amplification of Caiman latirostris DNA was not tested, and only one set of PCR conditions (the optimal conditions for American alligators) was used. Furthermore, the amount of intra-specific species polymorphism at the amplified loci was not determined for any of the other species. Therefore, it is possible that the use of different PCR conditions could permit amplification of additional loci from other species, especially Caiman latirostris. To date, there are 13 microsatellite markers specifically developed for Caiman latirostris . Since only some microsatellite markers can be used among closely related species (Moore et al., 1991), we used microsatellite markers developed for Caiman latirostris and Alligator mississipiensis in the present study to assess parentage among individuals from the captive colony of Caiman latirostris at the University of São Paulo, in Piracicaba, São Paulo, Brazil.
Caimans build mound-nests, and females usually display parental behavior towards both the nest and hatchlings (Verdade, 1995;Thorbjarnarson, 1996). In this study, eggs were collected during the first 48 h after being laid and transferred to artificial incubators, (as described by Verdade et al., 1992). Eggs and resulting hatchlings were identified by nest. Females guarding the nest were identified and assigned as possible clutch-mothers.
In the present study we used the markers Amim8, Amim13 and Amim20 developed for Alligator mississipiensis (Glenn et al., 1998) and the markers Clam02, Clam05, Clam06, Clam07, Clam08, Clam09 and Clam10 developed for C. latirostris . Polymerase chain reaction (PCR) conditions were standardized for 25 mL with: 1 X specific buffer (Table 1, all buffers contain 300 mM Tris-HCl and 75 mM ammonium sulfate and differing concentrations of Mg 2+ and pH), 0.2 mM each of dNTP, 0.4 mM of each primer pair, 0.2 U Taq DNA polymerase, and 100 ng DNA. The thermocycle program was: (1) 94°C for 3 min, (2) 94°C for 1 min, (3) primer specific annealing temperature for 1 min, (4) 72°C for 1 min, (5) repeat steps 2, 3 and 4 for n cycles, (6) 72°C for 7 min and (7) 4°C until storage (Table 1). Products were stored at 4°C until analyzing and scoring. PCR products were loaded into a Megabace 1000 DNA sequencer system for genotyping. Primers were labeled according to Table 1 and individuals genotyped by using the Genetic profiler program.
For logical reasons, such as the movement of individuals being restricted to individual enclosures, statistics were estimated by considering enclosures as though they were sampling units, as described above, ARN1 (N = 12) with one known parent (the father), three candidate parents (the possible mothers) and eight offspring from two clutches, ARN3 (N = 18) with one known parent (the father), five candidate parents (the possible mothers) and twelve offspring from three clutches, ARN4 (N = 10) with one known parent (the father), five candidate parents (the possible mothers) and four offspring from one clutch. The CERVUS 2.0 (Marshall et al., 1998) program was used for calculating exclusion power and null allele frequencies for each locus ( Table 2). The overall probability of exclusion for the maternity test by enclosure was computed with none parent known (Excl(1)) or with one parent known (Excl(2)) as shown in Table 2. CERVUS 2.0 was also used to assign maternity to possible mothers of offspring from the clutches in each enclosure, by employing the observed allele frequencies for enclosed populations to determine the statistical significance of the D value. This parameter was calculated by a simulation procedure that takes into account typing error rates and incomplete sampling for each possible mother, considering a given known father and offspring. At the end of this step, the possible mothers of each offspring were discriminated by D value and CI, e.g. the confidence interval, which could be either 80% or 95%, and corresponds to relaxed and restricted settings for CI, respectively, as shown in the last two columns of Table 3. 876 Parentage test Caiman latirostris Exclusion power and null allele frequency estimates, for each locus and by enclosure, are presented in Table 2. The overall probability of exclusion for the maternity test, by enclosure and considering one parent known (Excl(2)), that is the case for this study, since the offsprings' father is always known as there was one single male by enclosure, was 99,1% for ARN1 (clutches 1 and 5), 96,4% for ARN3 (clutches 2, 3 and 6) and 96,3% for ARN4 (clutch 4).
According to the parentage test (Table 3) and on comparing genotypes (Table 4), the indicated mother for Clutch 1 is 4-CL106, in disagreement with the classification of female 2-CL25 as clutch-mother based solely on maternal behavior displayed by this individual and not the former. Nevertheless, the female 2-CL25 was excluded from maternity by six microsatellite markers, Amim13, Clam02, Clam05, Clam06, Clam08 and Clam10, and the other possible mother, 3-CL53, by five microsatellite markers, Amim13, Clam05, Clam06, Clam08 and Clam10 (Table 4).
The behaviorally assigned mother of clutch 4, 83-CL9, was excluded from maternity of this clutch by microsatellite markers Clam05, Clam06, Clam07 and Clam08 (Table 4), whereas of the remaining females, 84-CL2 was excluded by Clam08 and 86-CL4 by Clam07 and Clam08. Female 85-CL3 could be neither excluded from maternity, nor indicated as the mother through parentage testing. Female 87-CL19 could not be excluded from maternity (Table 4), but was assigned as mother through parentage testing (Table 3). This was another case in which the molecularly assigned mother (87-CL19) was different from the behaviorally assigned (83-CL9).
In clutch 6, female 35-CL5 was assigned as mother of 142-CL479 (Table 3), but was excluded from maternity of the remaining hatchlings by markers Clam02, Clam07 and Clam09 (Table 4). Female 36-CL13 was indicated as mother of 144-CL481 and 146-CL483 (Table 3), and could not be excluded from the remaining hatchlings by comparison among genotypes (Table 4). Female 34-CL10 was excluded as mother by Clam07, whereas female 37-CL14 was from maternity by Amim13, Clam06 and Clam09 and female 38-CL70 as mother by markers Amim13, Amim20, Clam02, Clam06, Clam07, Clam08 and Clam09 (Table 4). Based on the above, female 36-CL13 was assigned as mother of the clutch through microsatellite analysis, which was also in accordance with behavioral displays.
With the set of markers used, it was possible to identify a single mother for all the offspring: clutches 1 (4-CL106), 2 (34-CL10), 3 (35-CL5), 4 (87-CL19), 5 (3-CL53) and 6 (36-CL13). Surprisingly, two of the females (33%) that displayed maternal behavior were not confirmed as actual mothers: 2-CL25 and 83-CL9. A display of maternal behavior by nonmothers can be explained as either a behavioral malfunction caused by the captive environment or species social adaptation as described in other vertebrates (Wrangham and Rubestein, 1986). Both hypotheses can be tested in future studies.
Farming operations are based on captive breeding and generally involve a small number of founders. Therefore, they require effective genetic management, in order to prevent genetic disorders as inbreeding depression (Foose, 1980). Assignment of mothers based exclusively on behavioral displays can lead to errors when assembling a Studbook and in establishing individual pedigrees. Under these circumstances microsatellite markers might be useful. In addition, these markers can also be useful in demographic and behavioral ecological studies in which the mating system and dispersal pattern are assessed based on parentage among individuals (e.g., Verdade et al., 2002).