Mol Gen Genet (1993) 238:49-58 © Springer-Verlag 1993 Individualization and estimation of relatedness in crocodilians by DNA fingerprinting with a Bkm-derived probe J.W. Lang 1, R.K. Aggarwal 2, K.C. Majumdar 2, L. Singh 2 1Department of Biology,University of North Dakota, Grand Forks, USA 2 Centre for Cellular and Molecular Biology, Hyderabad 500007, India Received: 15 June 1992/Accepted: 10 October 1992 Abstract. Individual-specific DNA fingerprints of crocodilians were obtained by the use of Bkm-2(8) probe. Pedigree analyses of Crocodylus palustris, C. porosus and Caiman crocodilus revealed that the multiple bands (2223 bands with Aludigest) thus obtained were inherited stably in a Mendelian fashion. Unique fingerprints permitted us to identify individuals, assign parentage, and reconstruct the DNA profile of a missing parent. Average band sharing between unrelated crocodiles was found to be 0.37. Band sharing between animals of known pedigrees increased predictably with relatedness and provided a basis for distinguishing relatives from non-relatives. Similar results obtained in other species/genera, using the same probe, suggest that this approach may be applicable to all species of crocodilians, and could facilitate genetic studies of wild and captive populations. Key words: Bkm D N A - Genetic/DNA fingerprinting Crocodilians - Paternity determination - Band sharing Introduction DNA fingerprinting utilizing multilocus and single locus probes is of proven utility for identification of individuals and for paternity determination (Jeffreys et al. 1991). Application of DNA fingerprinting to studies of wild populations has provided the behavioural ecologist with a powerful tool with which to estimate genetic relatedness among socially interacting individuals (Burke et al. 1991). Among vertebrates, D N A fingerprinting has been widely applied to studies of fishes, birds and mammals. These include demonstrations of reduced genetic variation in clonal and colonial species (Turner et al. 1990; Reeve et al. 1990), genetic parentage in species with varied mating systems (Westneat 1990; Tegelstrom et al. 1991), and genetic relatedness in species exhibiting social cooperation in breeding behaviour (Rabenold et al. Communicated by E. Bautz Correspondence to: L. Singh 1990; Amos et al. 1991; Gilbert et al. 1991). In contrast, studies in reptiles have been limited to brief reports on turtles (Jones et al. 1987; Demas et al. 1990; Demas and Wachtel 1991) and crocodilians (Demas and Wachtel 1991). In the latter study, which was carried out on five species of crocodilians, individual-specific variation in DNA fingerprints was not evident and this was attributed to low levels of heterozygosity thought to be characteristic of the group (Lawson et al. 1989). The crocodilians are an endangered group and consist largely of managed wild populations. They exhibit sophisticated communication repertoires, structured social interactions and mating systems, and extensive parental care, a feature absent or rare in other reptiles (Lang 1987). Therefore, there is a need to establish direct methods of assessing genetic relatedness for their future management in the wild as well as in captive breeding programmes. In the present investigation, we have used the approach of DNA fingerprinting using the minisatellite probe Bkm-2(8), to study genetic relatedness among individual crocodiles. The Bkm sequences were first identified and isolated as W sex-chromosome associated minor satellite D N A from the genomic D N A of the females of a poisonous Indian snake, Bungarus fasciatus (Singh et al. 1976, 1980, 1981). These sequences are found in all eukaryotes and are preferentially concentrated on the sex-chromosomes of many vertebrate species including mouse and human (Jones and Singh 1981a, b; Singh et al. 1981, 1988; Singh and Jones 1982, 1986). The Bkm2(8) probe predominantly consists of GATA repeats (Epplen et al. 1982; Singh et al. 1984), which are found scattered all over the genome (Singh and Jones 1982). This probe shows extensive restriction fragment length polymorphism (RFLP) in a wide range of higher eukaryotes (Singh and Jones 1986; Jones et al. 1987; Traut, 1987; Singh et al. 1988; Lloyd et al. 1989; Demas et al. 1990) and can be used as an efficient probe for DNA fingerprinting in humans (Singh 1991). Here we report that the Bkm-2(8) probe can be used for generating DNA fingerprints in crocodilians. Using a representative species (Crocodylus palustris) with in- 50 dividuals from known pedigrees, individual-specific, stably inherited band patterns were obtained which could be used to assign parentage and to reconstruct the fingerprint of a missing parent. Furthermore, calculations of band sharing between relatives and non-relatives provided an estimate of genetic relatedness among individual crocodiles. Our studies of several additional species (Aggarwal et al. 1992b) suggest that this technique may be applicable to all crocodilians. Materials and methods Specimens. Four males and six females of wild-caught mugger crocodiles C. palustris were sampled from the Nehru Zoological Park (Hyderabad, India) and ten males and six females from the Madras Crocodile Bank (MCB, Madras, India). These were originally caught from two geographic regions (states of Gujarat and Andhra Pradesh in India) and from four separate localities in Tamil Nadu (India) respectively. In addition, 31 males and 71 females of second generation progeny of known pedigrees, which were produced in captive breeding enclosures (Lang et al. 1989), were also sampled at MCB. Animals of known pedigrees, from a sibling group of five saltwater crocodiles (C. porosus), bred in captivity by crossing animals obtained from Tamil Nadu and Orissa (India), and a family group of eight caimans (Caiman crocodilus) from an unknown locality in Central America were also sampled at MCB. Additional samples were obtained from two siblings of Crocodylus cataphractus at the Metro Zoo, Miami (Fla., USA), three individuals of C. moreletii at Zoo Atlanta (Ga., USA), and a single sample of C. rhombifer at the St. Augustine Alligator Farm (Fla., USA). DNA preparation. Blood (0.5 to 60 ml) was drawn from the caudal vein, brain plexus or from the heart in heparinized, sterile syringes and frozen. The frozen blood cells were freeze-fractured twice by rapid thawing at 45 ° C in TES buffer (Singh and Jones 1986). The pellet obtained after centrifugation at 8000 rpm at 4° C for 10 rain was homogenized in TES buffer (1:10 v/v), lysed with Sarkosyl NL30 (2-3%), and incubated at 37 ° C for 12-16 h in the presence of proteinase K (75 gg/ml). DNA was isolated as described by Aggarwal et al. (1992a) and dissolved in TE buffer (10 mM TRIS-HC1, 1 mM EDTA, pH 7.5). DNAfingerprinting. Genomic DNA (12-15 gg) was digested to completion with AluI or HinfI restriction enzymes according to the conditions specified by the manufacturer (New England Biolabs). Digested samples were electrophoresed in 30 cm long, 5 mm thick, 0.8% agarose gels at 60 V for 16-18 h in TPE buffer (15 mM TRISHC1, 18 mM NaH2PO~, 0.5 mM EDTA, pH 7.8). HindIII and EcoRI double digests of phage )~ DNA were used as molecular weight markers. Gel-fractionated DNA samples were transferred onto Hybond-N membrane (Amersham, UK) using a vacuum-blotting assembly at 30 mm Hg (Olszewska and Jones 1988). The membrane was baked at 80° C for 2 h under vacuum. The blots were prehybridized in 7% SDS, 0.5 M sodium phosphate buffer (pH 7.5) at 60 ° C for 2-3 h, and then hybridized with 1-2 x 1 0 6 cpm/ml of the 32p-labelled single-stranded probe (specific activity 0.7-3.0 x 108 cpm/gg) in the same but fresh buffer at 60° C for 14~16 h. The probe used was Bkm-2(8) (Patent no. 1000 DEL 88), a subclone of the clone CS 314 obtained from Bkm probing of a Drosophila genomic library. The subclone 2(8) contains a 545 bp sequence consisting of 66 copies of the GATA repeats interspersed with variable numbers of dinucleotide repeats of TA in several locations (Singh et al. 1984). Single-stranded 3Zp-labelled probe was prepared to a specific activity of 0.7-3.0 x 108 cpm/gg (Hu and Messing 1982), using 32p-dATP (specific activity 3000 Ci/mmol; Jonaki, BARC, India). The blots were washed once in 3 x SSC (1 x SSC is 0.15 M NaC1, 0.015 M sodium citrate, pH 7.2) containing 0.1% SDS at room temperature for 10 rain and thrice in 2xSSC, 0.1% SDS at 60° C for 10 min each. Several exposures, varying from 1 to 10 days were made of each hybridized membrane with intensifying screens at - 7 0 ° C, in order to score all the bands of differing intensities. Analysis of fingerprints. Molecular weight markers and duplicate samples from the same individual were run on either side of the gel as controls for mobility distortion. Samples of known individuals were run in each gel to facilitate comparison of DNA fingerprints obtained from different gels. No variation was evident in the band patterns of the same DNA sample run repeatedly on the same or on different gels. Autoradiographs were examined visually and also scanned on the Biotrac DNA fingerprinting system (Foster and Freeman Ltd., UK) using the Bioworld program to calibrate the fragments for their molecular weights. The distinct bands in each genetic profile were scored in the size range of 1.323.0 kb. All bands showing similar molecular weights (band centres within 0.5 ram) were considered to be identical for the purpose of pairwise comparisons. Band sharing was calculated following the methodology used by Burke and Bruford (1987). Mean band sharing for a particular group of individuals was calculated by averaging the band sharing values for all possible pairwise combinations. To establish baseline values for band sharing between non-relatives, independent pairwise comparisons were made (Lynch 1991). A subset of scored bands that were particularly distinct and exhibited well defined differences in hybridization intensity were further scored as either heterozygous (less intense) or homozygous (twofold increase in intensity by comparison). The differences in band intensities were inferred mainly by visual inspection but in some cases by laser densitometry. Only those bands were considered for this analysis, which repeatedly showed a fixed pattern of intensity differences in the progeny. Moreover, the differences in intensities of bands due to variation in the total amount of DNA used in the different lanes being compared were also taken into consideration. Pedigree analyses were completed following the methodology used by Jeffreys et al. (1986). The transmission frequency of 51 each of these bands was calculated as the percentage of progeny in which the band was present in each family. Results D NA fingerprints For each individual of C. palustris (45 males, 83 females) tested, an individual-specific pattern was obtained with the Bkm-2(8) probe. The total number of polymorphic bands scored in the fingerprints of all the individuals tested was 41 and 35 in the size range of 1.3-23.0 kb, with AluI and HinfI restriction enzymes, respectively. No sexspecific band was evident in any of the animals tested, including the progeny of both sexes. [ ~ (~P 1 progeny 2 3 4 5 6 7 8 9 10 I 11 In a sample of 14 individuals, originally from six different localities, the number of bands ranged from 18 to 27 (mean+SD, 23.14-2.81; AIuI digest). The HinfI digest produced specific but fewer bands (20.7± 2.62) which were readily distinguishable from the AluI fingerprints. On the other hand, in captive-bred animals the mean band number in 4 parents with AluI digest was 22.4 ± 2.75 (range, 15-26) and in 45 offspring 22.7 4- 2.44 (range, 16-28). The inheritance of parental bands in the progeny of three families (15 individuals per family) was analysed. The same male fathered these families with three different females. Without exception, bands in the offspring could be accounted for as either being inherited from the father or from the mother. Overall, the offspring shared 30% of their bands with the father, 33% with the mother, and 37% with both parents. The DNA fingerprint patterns for two of these families are illustrated in Figs. 1 and 2. 99 kb 23.1 - 9.42- 6.565.15_ 4.974.364.273.53- 2.322,03_ 1,901.58- Band heterozygosity In the above three families, 27 bands (2.5-23.0 kb) were resolved sufficiently to follow their segregation in the offspring. Differences in the hybridization intensity at these loci facilitated discrimination of the number of copies (single-intensity band, heterozygous; doubleintensity band, homozygous) present per band in each offspring and their parents (Figs. 1, 2). The male parent Z (Fig. 1, far left lane) was heterozygous at all 15 loci scored in this analysis. By contrast, all three females (A, B, C) had varying proportions of homozygous bands, ranging from 0.37 to 0.44 (Table 1). These females were homozygous for one band ( ~ 19 kb) which was inherited by all the 45 offspring tested (Figs. 1, 2). The absence of double intensity for this band in female B (Fig. 1 ; 9P lane) was due to the comparatively lower amount of D N A present in this lane. An addiTable 1. Mean transmission frequency (f) of different bands (homozygous vs heterozygous) to the offspring in three families ~ of Crocodylus palustris 1.33- Fig. 1. Fingerprints from a pair of mugger crocodiles, Croeodylus palustris (dr P, far left lane, individual Z; 9 P, far right lane, individual B) and their offspring (lanes 1-6 female progeny; 7-11 male progeny), obtained with the Bkm-2(8) probe, AluI digest. Segregation analysis of the resolved maternal and paternal bands in 15 progeny (4 others were done but are not shown) revealed no cosegregating bands. Note a major band of ~ 19 kb, present in the female parent, absent in the male parent, and present in all of the offspring (homozygous in the female parent; heterozygous in the progeny). Two bands visible at ~ 6.5 kb are present in the male parent, absent in the female parent and present in most, but not all of the offspring (heterozygous in parent and progeny; note lane 6). Another pair of paternal bands visible at ~ 4.25 kb appear in some of the progeny, but are absent in the female parent. Other bands are shared by both parents and are present in variable combinations in the progeny, e.g. two pairs of major bands of ~ 3 kb Total number of bands scored Male Female Z A B 15 8 12 3 7 8 3 (37) 5 (42) 7 (44) 15 1.00 5 0.986 7 0.977 9 0.498 0.774 0.513 0.835 0.400 0.783 0.532 0.703 Shared with male No. ofhomozygous bands ( + / + ) f (homozygous bands) No. of heterozygous bands ( + / - ) f (+/_ ; _/_)b f(+/--; +/--)° 0 (0) 16 Values in parentheses are % homozygous bands a The male parent is common to all the three families (number of offspring is 15 per family) b Heterozygous bands present in only one parent ° Heterozygous bands present in both the parents 52 . . i 9P progeny . . . . 1 2 3 4 . 5 . . . . 6 . I- . . . . . . . . 7 d'P 8 9 progeny 10 kb 23.1 - 9.426.565.15_ 4.974.36_ 4.273.53- 2.32tional band ( > 2.3 kb), which was homozygous in two of the females from different localities (females B and C, ~P tracks in Figs. 1, 2), was absent in the male. The mean transmission frequencies of the homozygous versus heterozygous bands were consistent with the assumption that these bands were inherited in a Mendelian pattern following a binomial distribution (Jeffreys et al. 1986). Bands transmitted to virtually all the offspring (mean transmission frequency, 0.977 to 1.00) were also judged to be homozygous on the basis of hybridization intensity. Heterozygous bands that were shared in both parents were transmitted to the offspring with a mean transmission frequency of 0.703-0.835, and those absent in one parent were transmitted with a frequency of 0.400-0.532 (Table 1). In our analysis, no bands were detected that were either allelic or cosegregating pairs. The mutation rate (mutation/band per generation), based on the difference between observed and expected transmission rates of bands judged to be homozygous, was estimated to be 1.3 x 10 -3 (Kuhnlein et al. 1990). Band sharin 9 As a baseline for comparisons among known relatives, band sharing was calculated between pairs of crocodiles from different localities in India. We have assumed that these individuals were unrelated. Among these unrelated individuals, the mean band sharing coefficient, calculated on the basis of all possible pairwise combinations, was 0.36. In contrast, wild-caught individuals from the same localities exhibited relatively high band sharing (mean coefficient, 0.59), a value closely approximating the mean observed value for half siblings (Table 2). Mean allele frequency was calculated using x = 2q - q2 (Jeffreys et al. 1985); q equalled 0.213 for the mean band sharing value (x) of 0.38 (Table 2). If shared bands represent indentical alleles and all alleles have equal frequencies, the expected band sharing between first 11 12 ......... 1 13 14 15 ?P Fig. 2. Fingerprints from a pair of mugger crocodiles (d P, two centre lanes, individual Z; c?p, outer lanes, individual C) and their offspring (lanes 1-7 female progeny; 8-15 male progeny), obtained with the Bkm-2(8) probe, AluI digest. Segregation analysis of bands in this family, like the family in Fig. 1, showed: no cosegregating bands; a major band of 19 kb homozygous in female parent and heterozygous in all the offspring; a pair of heterozygous paternal bands of 6.5 kb, present in some of the progeny; and most of the other bands shared by both parents and present in variable combinations in the progeny, e.g. a series of bands around 3 kb in size. Male parent in this family is the same as in Fig. 1; note identical band patterns order relatives (parents-offspring; siblings) with q = 0.213 would be 0.67 (Jeffreys et al. 1985). By comparison, the observed mean band sharing (based on all pairwise combinations) for parents-offspring and for full siblings in these three families was 0.68 and 0.70, respectively (Table 2). This close agreement between the expected and the observed band sharing probabilities substantiates the assumption of independence for the bands produced with the enzyme/probe combination used in this study (Burke et al. 1991). The band sharing was higher among relatives than among non-relatives. It was the highest between parents and offspring and between full sibs with, as expected, nearly comparable mean values (see above; Table 2). Within each family, the band sharing coefficient between mother-offspring was higher, relative to the fatheroffspring value (combined mean 0.70 and 0.66, respectively; Table 2). There was, however, considerable variation in the values for mother-offspring among the three families (mean 0.63, 0.73 and 0.75 for the females A, B and C, respectively). Likewise, the father-offspring values also varied among families (mean 0.63, 0.71 and 0.64 for families from the females A, B and C, respectively). The parent-offspring values reflected the relative differences in band sharing among the parents. Band sharing between the parents of each family (fathered by the same male) was consistent with the assumption that the male and female parents were not closely related. On the other hand, two of the females, A and C, were from the same locality, and possibly were sibs. For them the band sharing coefficient was 0.58. The relatively high proportion ofhomozygous bands in the females A, B and C, and the fact that some of these bands were shared by all of these, were reflected in the band sharing values among the progeny in these families (Table 2; also see above). The D N A fingerprints from these three groups of sibs were compared within and between the families. Mean band sharing among the full sibs was higher than among the half sibs (combined mean 0.70 and 0.61, respectively, 53 Table 2. Mean band-sharing coefficient (x) among crocodilians Comparisons a Croeodylus palustris Wild population Non-relatives: Different localities (6) Different localities (6) Individuals : Same localities (5) Number Pairwise Mean_+ SD of individuals comparisons b Average relatedness 14 12 77 6c 0.36+0.08 0.37 -+0.07 - 13 14 0.59+_0.16 ? Familial analysis Parents: Overall Mother-father Mother-mother Parent-offspring: Overall Father-offspring Mother-offspring 4 4 3 6 3 3 0.45 _+0.09 0.38 _+0.07 0.52_+ 0.06 - 49 46 48 90 45 45 0.68 _+0.10 0.66 ± 0.08 0.70 ± 0.07 0.5 0.5 0.5 Full sibs (overall) Half sibs (overall) " 45 45 315 630 0.70_+0.10 0.61_+0.11 0.5 0.25 C. porosus family Full sibs (overall) Sibs-unrelated individual 5 6 10 5 0.60_+0.14 0.40 _+0.07 0.5 ? Caiman crocodilus family Full sibs (overall) Mother-offspring Mother-other female Sibs-other female Sibs-male individual 5 6 2 6 6 10 5 1 5 5 0.60_+0.12 0.65 _+0.08 0.65 0.53 _+0.09 0.41 _+0.08 0.5 0.5 ? ? a Figures in parentheses indicate number of localities b Includes all possible combinations c Independent comparisons only between putative non-relatives from different localities (Lynch 1991) Table 2). However, extensive overlap in these distributions precludes a determination of relatedness between sibs versus halfsibs, based only on the band sharing value of the two individuals. Furthermore, the band sharing values between families reflected the relatedness of the four parents. Thus in each of the three families tested, the mean band sharing coefficients for sibs and half-sibs differed numerically (0.66 and 0.61; 0.73 and 0.60; 0.71 and 0.61 in the families from the females A, B and C, respectively), and were consistently higher for the sibs in comparison to their halfsibs. Fingerprint reconstruction F o r two additional crocodile families, D N A samples were available for 12 and 11 offspring from two female parents (females B and C, which were the same female parents as for the families shown in Figs. 1, 2), but the male parent (Y) had died 4 years earlier. The band pattern of the missing male parent was reconstructed separately f r o m the fingerprints of each family (female parent and representative progeny). For this reconstruction, discernible bands within the range of 2 . 0 ~ 3 . 0 kb were used. In each family (Fig. 3), the unique bands were identified that were not assignable to the female parent, but were evident in the progeny. These consisted of 13 unique bands in family 1 (female B) and 17 bands in family 2 (female C). Within each family, the bands that were shared by both the parents were identified on the basis of their hybridization intensity and the segregation in the offspring. The n u m b e r of such shared bands was 11 in family 1 (female B) and 7 in family 2 (female C). Thus the total number of bands attributable to the male within each family was 24 (13 unique and 11 shared with female B; 17 unique and 7 shared with female C) which on comparison, were found to be identical (see legend to Fig. 3). Using the reconstructed fingerprint of the missing male Y, the band sharing coefficients were also computed for the above families which were consistent with the results obtained in the three families described earlier (Table 2). The validity of this method was confirmed by reconstructing the fingerprint pattern of the male (Z) of the three families described earlier, from the D N A fingerprints of the progeny and the female parents alone (using bands in the size range of 2.5-23.0 kb; Figs. 1, 2). A total of 15 bands were reconstructed which corresponded in number and size (molecular weight) to the actual number of bands scored for the male Z (Table 1). 54 F. . . . . . ~P kb I 2 PROGENY . . . . . . . . 3 4 5 6 7 1 8 F. . . . . SP ZP SP 1 PROGENY . . . . . . . 2 3 4 5 6 23.10 9.42 6.56 5.15 4-97 4.36 4.27 3.53 2.32 2.03 1.90 1.58 FAMILY I 2 3 4 7 8 SP Fig. 3. Fingerprint reconstruction for missing male parent Y (~ P, line drawing in centre) based on Bkm-2(8) probe-generated AluI fingerprints, of two mugger families from females C (far left lane) and B (far right lane). On the basis of segregation analysis of the families, 11 bands were unique to the male parent (unmarked bands in the line diagram), and 5 bands (marked by dots) were shared by all parents. In addition, 6 bands (marked with arrowheads' at left side) were detected as unique paternal bands in family 1, but as shared (with female parent B) in family 2. Likewise, 2 more bands (marked with arrowheads at right-hand side) appeared as unique paternal bands in family 2, but as shared among parents in family 1. Thus in total, the reconstructed band pattern of the missing male consists of 24 bands (11 unique to the male in both the families, 5 shared by all parents and 8 unique/shared in the two families) FAMILY II I r--Progeny--1 GM ~P kb 7 r--Progeny ---i SP GM dP I 2 3 4 9P dP 25.10 9.42 6.56 5.15 4.97 4.36 4.27 3.53 2.32 2.03 1.90 1.58. 1.33 I FAMILY I II FAMILY 11 Paternity determination The families discussed above were sired by lone males (Z and Y) living in separate breeding enclosures with multiple females. In an adjacent pen there were 10 males housed with 18 females. One of the males (X) was clearly I Fig. 4. Paternity determination in two crocodile families based on DNA fingerprinting with Bkm-2(8) probe, after AluI digestion. Left panel shows family 1 and right panel family 2 (lane identification: GM, grandmother; d P, putative male parent X; 1-4, progeny; ~ P, female parent; single far right lane (3' P, shown separately, is putative male parent W). The fingerprints of progeny 3 and 4 in family 1 and of all the progeny in family 2 contain a total of 12 unique bands (indicated by arrowheads) which are absent in the female parents and male X, but present in male W. By contrast, the fingerprints of progeny 1 and 2 in family 1 lack these bands, but show bands (at 4.5 and 6.0 kb) which are evident in the putative male parent X (indicated by dots) dominant in mating behaviour, whereas others were subordinate in social rank and were rarely observed to mate. The parentage of the two sib groups from the latter breeding enclosure was tested by fingerprinting the progeny along with the putative father (dominant male, X) and the known mother. 55 In one family, the putative father was judged to be the genetic parent of offspring 1 and 2 (Fig. 4; family 1, lanes 1 and 2) on the basis of bands unique to male X (Fig. 4; left panel, lane c? P). However, progeny 3 and 4 (Fig. 4; family 1, lanes 3 and 4) exhibited other bands which were not assignable to male X, the putative parent. Similarly all the four offspring in family 2 (Fig. 4) showed many bands not attributable to the putative father X. These paternal bands (not accounted for by the female parents; shown by small arrowheads in Fig. 4) in offspring 3 and 4 in family 1 and in offspring 1-4 in family 2 were found to be identical, indicating that the unidentified male parent in both the families was the same. Subsequent examination of the D N A fingerprints of 10 candidate males from this breeding enclosure, clearly identified one of the subordinate males W (Fig. 4; right hand lane d' P) as the father of these offspring. The mean number of paternal bands per offspring inherited from W was 6 . 3 i 1.8 (range 4 9). A total of 12 such bands unique to W were identified from the fingerprints of 6 progeny individuals (Fig. 4; lanes 3-4 of family 1; 1-4 of family 2). Thus in family 1, offspring 1 and 2 were fathered by X and offspring 3 and 4 by W. In this family the four progeny individuals were, in fact, produced from two clutches of eggs laid in different years by the same female (offspring 1 and 2 from one clutch, and offspring 3 and 4 from another clutch 1 year later). These individuals were, therefore, not clutch-mates; thus, this is not an instance of multiple paternity within a single clutch of eggs. Fingerprinting other croeodilians In a separate study (Aggarwal et al. 1992b), samples from 18 of the 23 extant species of crocodilians were used to generate D N A fingerprints, specific at the individual, species, and generic levels, using Bkm-derived probe 2(8). These included species in the genera Alligator, Caiman, Paleosuchus, Gavialis, Tomistoma, Osteolaemus, and Crocodylus. In the three Crocodylus species tested in the present study, band patterns differed significantly from each other between the species but the overall profiles were similar to those of C. palustris (Fig. 5). In the saltwater crocodile, C. porosus, five sibs and one unrelated animal were tested with the Bkm-2(8) AluI digest. The resultant band patterns were individual-specific and revealed a number of shared bands among relatives. The mean number of bands per individual was 18.2:k2.93 (range 14-23) of 1.3 23.0 kb in size. Mean band sharing was 0.60 :k 0.07 among sibs and 0.40 + 0.07 between the single unrelated animal and the sibs (fingerprints not shown). For the spectacled caiman, Caiman crocodilus, samples were available for five sibs, the female parent, the putative male parent, and another adult female from the same breeding group in which the sibs were produced. As in C. porosus, the resultant band patterns were unique for each individual and higher band sharing was evident in related animals (fingerprints not shown). The number of bands in HinfI profiles per individual aver- Cr.m. kb 23.10 Cr.r, Cr.c. I i r"'] 1 2 3 4 I - - 1 5 Cr.p. I 6 I 7 8 9,42 6.56 5.15 4.97 4.36 4.27 3.53 2.32 2.03 1.90 1.58 1.33 Fig. 5. Representative fingerprints of four species of Crocodylus, obtained with Bkm-2(8) probe after AluI digestion. Individualspecific bands are evident in all the fingerprints, primarily in the 2.0-20.0 kb size range, except in the case of C. rhombifer (Cr.r., lane 4) for which only one sample was tested. Among individuals (putative siblings) of C. moreIetti (Cr.m., lanes 1-3) and of C. cataphractus (Cr.c., lanes 5-6), the banding patterns are similar but not identical. Individuals of C. palustris (Cr.p., lanes 7-8) are from different geographic regions. In these species, band patterns in the size range of 1.3 to 3.0 kb are distinctive, and appear to be species-specific aged 24.9:t:3.44 (range 19-30) of 1.3-23.0 kb in size. Mean band sharing was 0.60 i 0.10 among the sibs, and 0.65 + 0.07 between the offspring and the female parent. It was evident from the fingerprint patterns that many bands in the sibs were shared among siblings and with the mother. In contrast, few bands were shared between siblings and the putative father. Mean band sharing between this male and the sibs was 0.41 :~ 0.07. The lack of correspondence between band patterns of the putative father and offspring, and the relatively low mean band sharing value, indicate that this male was not the parent. Unfortunately, other candidate fathers living in the same breeding enclosure were not tested. Further analysis of these Caiman suggested that the two females (mother and the other adult female) were closely related (band sharing coefficient, 0.65). As expected, the mean band sharing between the sibs and this female was higher (0.53 i 0.07) than between the non-parent male and the sibs. 56 Discussion Utilizing the Bkm-2(8) probe, individual-specific DNA fingerprints have been produced in 18 species belonging to seven of the eight extant genera of the order Crocodilia (Caiman, Crocodylus, Alligator, Gavialis, Osteolaemus, Paleosuchus, and Tomistoma; this study and Aggarwal et al. 1992b). Our results, based on a large data set within the species and also across genera, contradicts a recent report in which identification of individual crocodilians was not possible from Bkm-generated band patterns, although species-specific differences were noted (Demas and Wachtel 1991). This discrepancy may be due to the different combination of restriction enzymes used by these authors. They suggested that the absence of individual-specific fingerprints may be due to the high levels of homozygosity noted previously in crocodilian populations (Lawson et al. 1989). On the other hand, we have demonstrated band heterozygosity in the size range of 1.3 to 23.0 kb in the DNA fingerprints, and this is comparable to those observed in other species with commonly used probes (Burke et al. 1991 ; Jeffreys et al. 1991 ; Singh 1991). A captive breeding population was studied of known pedigree mugger crocodiles, originating from wild animals collected at various localities in India. For each animal tested, the DNA fingerprint was unique and reproducible. In a sample of non-relatives, on an average 23 bands (n) per individual were resolved and the mean band sharing coefficient was 0.37. Thus the probability of an exact match, calculated according to Jeffreys et al. (1985), between any two mugger crocodiles is x", or 0.3723. The assumption of linkage equilibrium implicit in this approach (Burke et al. 1991) was tested by comparing the DNA profiles of three large families of crocodiles of known pedigree. All of the parental bands were transmitted to representative progeny in varying combinations. Within each family, frequencies of individual bands were predictable and consistent with the assumption of band independence. This was also confirmed by segregation analyses on the basis of discernible differences in hybridization intensity of specific bands. Thus for the crocodile, we have confirmed by a detailed pedigree analysis that DNA fingerprints generated with the enzymes HinfI or AluI and Bkm-2(8) probe combinations are especially useful for individual identification and paternity determination. Our study failed to reveal any sex-specific fingerprint pattern or sex-specific inheritance of any discernible band in any of the crocodilian species. This suggests that the Bkm-related sequences in these species are probably not preferentially concentrated in a specific sex-determining region of a chromosome as in the mouse (Singh and Jones 1982), or distributed along a specialized sex chromosome as in snakes (Singh et al. 1980). This is in agreement with the cytogenetic finding of the universal absence of heteromorphic sex chromosomes among crocodilians (Cohen and Gans 1970; Singh and Ray-Chaudhuri 1973; King et al. 1986). This however, is in sharp contrast with the observation of Demas et al. (1990) that Bkm-2(8) probe gives sex-specific fingerprint patterns in sea turtles, Chelonia mydas and Lepidochelys kempi, which like crocodilians also have temperature-dependent sex determination (Morreale et al. 1982; Shaver et al. 1988). It is possible that in crocodilians, stretches of Bkm repeats associated with the sex-determining region are too small to be detected as a sex-specific band; or such sex-specific fragment(s), if generated, may be small and hidden in the smears in the low molecular weight region of the fingerprints. It is also possible that the sex-determining region may not be polymorphic for AluI or HinfI used in the present study but may reveal sex-specific restriction polymorphism with other restriction enzymes. The large number of bands unique to each parent and the differing band intensities at specific sites facilitated the reconstruction of fingerprints and the assignment of parentage. Using samples from the maternal parent and a number of offspring, it was possible to reconstruct the paternal band pattern and, importantly, to identify which bands were unique and which were shared in each parent. In addition, paternity testing of progeny produced in a multi-male breeding enclosure revealed which progeny were sired by a particular male. In all these analyses, approximately one-third of the bands in the progeny were shared only with the mother, another third shared with only the father, and the remaining third shared with both the parents. The expected number of paternal bands was 8.0. Based on a mean allele frequency (q) of 0.213, the observed paternal bands in the three families initially analysed was 6.7. The probability that an unrelated male has all the paternal fragments by chance (Jeffreys et al. 1985) is x", or approximately 0.378=3.5x 10 -4, giving an exclusion probability of 0.99965. Maternal exclusion probability also is equally robust. The unique, individual-specific feature of DNA fingerprints has fostered the prospect that this technique may be extended to estimate genetic relationships beyond first-order relatives (Lynch 1991). The most direct approach is to establish baseline data on the empirical relationship between band sharing and relatedness for particular populations and/or species of interest. To serve as a reference point for comparisons of band sharing among closely related crocodiles, we determined the mean band sharing coefficient between non-relatives. The observed band sharing between known pedigree crocodiles varied predictably with relatedness. A clear rank order was apparent in the sample population studied. Between first-order kin, band sharing was highest; between non-relatives, band sharing was lowest. Overlap in the distribution of band sharing values between firstand second- order kin, e.g. sibs and halfsibs, blurred the distinction in these categories. As a result, an estimate of relatedness between any two crocodiles on the basis of band sharing alone was possible only at the level of relative versus non-relative. Nevertheless, band sharing provides a useful first approximation of relatedness, particularly in assessing the status of known groupings of crocodiles, e.g. when comparing clutches suspected to be sired by the same male. Further refinement of this approach will be dependent on developing an empirically determined calibration curve of kinship versus band 57 sharing. Additional values f r o m third-order kin and f r o m non-relatives will be needed to determine whether the calibration curve is linear or curvilinear. Examples o f b o t h types o f curves have been described recently in birds (Jones et al. 1991) and m a m m a l s , even between populations o f the same species (Gilbert et al. 1991). Fingerprint analyses o f two additional species, saltwater crocodiles and caiman, indicated that the metho d o l o g y outlined here for the m u g g e r crocodile m a y be usefully extended to other species o f crocodilians. Firstorder kin were clearly identified on the basis o f b a n d i n g patterns. The sibling and the parent-offspring groups exhibited levels o f b a n d sharing c o m p a r a b l e to those observed in the m u g g e r crocodile families (Table 2). Genetic analysis using D N A fingerprinting has the potential to reveal the breeding structure o f wild p o p u l a tions as well as elucidate the genetic relationships underlying specific social interactions, e.g. parental care o f hatchlings, This technique is equally applicable to captive breeding p o p u l a t i o n s o f rare a n d / o r endangered species o f reptiles wherein there is a need to verify individual identity, establish paternity, and to determine pedigree. Finally, preliminary fingerprinting o f two additional species suggests that these techniques will be applicable n o t only to the crocodilians included in this study, but also to other m e m b e r s o f the order Crocodilia and to other reptiles such as turtles, lizards, and snakes. F u r t h e r studies (unpublished) revealed that Bkm-derived p r o b e 2(8) can be used as a universal p r o b e for D N A fingerprinting in eukaryotes. Acknowledgements. We thank S. Bhaskar for her excellent technical assistance and P.M. Bhargava for going through the manuscript, both of CCMB, Hyderabad, India. We are grateful to M. Wise at St. Augustine Alligator Farm, Florida; B. Ziegler at Miami Metro Zoo, Florida~ H. Hunt at Zoo Atlanta, Georgia (all in USA); and P. Kumar at Nehru Zoological park, Hyderabad, India, for providing crocodilian samples. We are also grateful to R. Whitaker and H. Andrews for encouragement, field assistance, and for providing access to specimens and records at the Madras Crocodile Bank. Support for J.W.L. was provided by Smithsonian Institution, National Science Foundation, and National Geographic Society of USA. References Aggarwal RK, Lang JW, Singh L (1992a) Isolation of high-molecular-weight DNA from small samples of blood having nucleated erythrocytes, collected, transported and stored at room temperature. Genet Anal: Techn Applic 9 (2):54~57 Aggarwal RK, Majumdar KC, Lang JW, Singh L (1992b) Generic affinities among crocodilians as revealed by DNA fingerprinting using a Bkm-derived probe. Proc Natl Acad Sci USA (submitted) Amos B, Barrett J, Dover GA (1991) Breeding behaviour of pilot whales revealed by DNA fingerprinting. Heredity 67:49-55 Burke T, Bruford MW (1987) DNA fingerprinting in birds. Nature 327:149-152 Burke T, Hanotte O, Bruford MW, Cairns E (1991) Multilocus and single locus minisatellite analysis in population biological studies. In: Burke T, Doff G, Jeffreys A J, Wolff R (eds) DNA fingerprinting: approaches and applications. Birkhauser Vcrlag, Basel, pp 154-168 Cohen MM, Gans C (1970) The chromosomes of the order Crocodilia. Cytogenetics 9:81-105 Demas S, Wachtel S (1991) DNA fingerprinting in reptiles: Bkm hybridization patterns in Crocodilia and Chelonia. Genome 34: 472-476 Demas S, Duronslet M, Wachtel S, Caillouet C, Nakamura D (1990) Sex-specific DNA in reptiles with temperature sex determination. J Exp Zool 253:319-324 Epplen JT, McCarrey JR, Sutou S, Ohno S (1982) Base sequence of a cloned snake W-cfiromosome DNA fragment and identification of a male-specific putative mRNA in the mouse. Proc Natl Acad Sci USA 79:3798-3802 Gilbert DA, Packer C, Pusey AE, Stephens JC, O'Brien SJ (1991) Analytical DNA fingerprinting in lions: parentage, genetic diversity, and kinship. J Hered 82:378-386 Hu N, Messing J (1982) The making of strand-specific M13 probes. Gene 17:271-277 Jeffreys A J, Brookfield JFY, Semeonoff R (1985) Positive identification of an immigration test-case using human DNA fingerprints. Nature 317:818-819 Jeffreys A J, Wilson V, Thein SL, Weatherall D J, Ponder BAJ (1986) DNA "fingerprints" and segregation analysis of multiple markers in human pedigrees. Am J Hum Genet 39 : 11-24 Jeffreys AJ, Royle NJ, Patel I, Armour JAL, MacLeod A, Collick A, Gray IC, Neumann R, Gibbs M, Crosier M, Hill M, Signer E, Monckton D (1991) Principles and recent advances in human DNA fingerprinting. In: Burke T, Dolf G, Jeffreys AJ, Wolff R (eds) DNA fingerprinting: approaches and applications. Birkhauser Verlag, Basel, pp 1 19 Jones CS, Lessells CM, Krebs JR (1991) Helpers-at-the-nest in European Bee-eaters (Merops apiaster): a genetic analysis. In: Burke T, Dolf G, Jeffreys A J, Wolff R (eds) DNA fingerprinting: approaches and applications. Birkhauser Verlag, Basel, pp 169 192 Jones KW, Singh L (1981a) Conserved sex-associated repeated DNA in vertebrates. In: Dover G, Flavell R (eds) Genome evolution. Academic Press, London, pp 135-154 Jones KW, Singh L (1981b) Conserved repeated DNA sequences in vertebrate sex chromosomes. Hum Genet 58:46-53 Jones KW, Olszewska E, Singh L (1987) Rapidly evolving Bkm DNA is associated with hypervariable domains. In: Stahl A, Luciani JM, Vagner-Capodano AM (eds) Chromosomes today, vol 9. Allen and Unwin, Suffolk, UK, pp 22-29 King M, Honeycutt R, Contreras N (1986) Chromosomal repatterning in crocodiles: C, G and N-banding and the in situ hybridization of 18S and 26S rRNA cistrons. Genetica 70:191-201 Kuhnlein U, Zadworny D, Dawe Y, FairfulI RW, Gavora JS (1990) Assessment of inbreeding by DNA fingerprinting: development of a calibration curve using defined strains of chickens. Genetics 125:161-165 Lang JW (1987) Crocodilian behaviour: implications for management. In: Webb GJW, Manolis SC, Whitehead PJ (eds) Wildlife management: crocodiles and alligators. Surrey Beatty and Sons, Sydney, pp 273 294 Lang JW, Andrews H, Whitaker R (1989) Sex determination and sex ratios in Crocodylus palustris. Am Zool 29:935-952 Lawson R, Kofron C, Dessauer HC (1989) Allozyme variation in a natural population of the Nile crocodile. Am Zoo129:863-871 Lloyd MA, Fields MJ, Thorgaard GH (1989) Bkm minisatellite sequences are not sex associated but reveal DNA fingerprint polymorphisms in rainbow trout. Genome 32:865-868 Lynch M (1991) Analysis of population genetic structure by DNA fingerprinting. In : Burke T, Doff G, Jeffreys AJ, Wolff R (eds) DNA fingerprinting: approaches and applications, Birkhauser Verlag, Basel, pp 113-126 Morreale SJ, Ruiz GJ, Spotila JR, Standora EA (1982) Temperature-dependent sex determination: Current practices threaten conservation of sea turtles. Science 216:1245-1247 Otszewska E, Jones KW (1988) Vacuum blotting enhances nucleic acid transfer. Trends Genet 4: 92-94 58 Rabenold PP, Rabenold KN, Piper WH, Haydock J, Zack SW (1990) Shared paternity revealed by genetic analysis in cooperatively breeding tropical wrens. Nature 348:538-540 Reeve HK, Westneat DF, Noon WA, Sherman PW, Aquadro CF (1990) DNA "fingerprinting" reveals high levels of inbreeding in colonies of the eusocial naked mole-rat. Proc Natl Acad Sci USA 87: 24965500 Shaver D J, Owens DW, Chancy AH, Caillouet CW Jr, Burchfield P, Marquez R (1988) Styrofoam box and beach temperatures in relation to incubation and sex ratios of Kemp's riddley sea turtles. National Oceanic and Atmospheric Administration, Technical Memorandum NMFS-SEFC 214:103-108 Singh L (1991) DNA profiling and its applications. Curr Sci 60:580-585 Singh L, Jones KW (1982) Sex reversal in mouse (Mus musculus) is caused by recurrent non-reciprocal crossover involving the X and an aberrant Y chromosome. Cell 28:205-216 Singh L, Jones KW (1986) Bkm sequences are polymorphic in humans and are clustered in pericentric regions of various acrocentric chromosomes including the Y. Hum Genet 73 : 304-308 Singh L, Ray-Chaudhuri SP (1973) DNA replication pattern in the chromosomes of Crocodylus palustris (Lesson). Nucleus 16: 33-37 Singh L, Purdom IF, Jones KW (1976) Satellite DNA and evolution of sex chromosomes. Chromosoma 59 : 43-62 Singh L, Purdom IF, Jones KW (1980) Sex chromosome associated satellite DNA: evolution and conservation. Chromosoma 79:137-157 Singh L, Purdom IF, Jones KW (1981) Conserved sex chromosomeassociated nucleotide sequences in eukaryotes. Cold Spring Harbor Syrup Quant Biol 45:805 813 Singh L, Phillips C, Jones KW (1984) The conserved nucleotide sequences of Bkm which define Sxr in the mouse are transcribed. Cell 36:111-120 Singh L, Winking H, Jones KW, Gropp A (1988) Restriction fragment polymorphism in the sex-determining region of the Y chromosomal DNA of European wild mice. Mol Gen Genet 212: 440-449 Tegelstrom H, Searle J, Brookfield J, Mercer S (1991) Multiple paternity in wild common shrews (Sorex araneus) is confirmed by DNA fingerprinting. Heredity 66:373-379 Yraut W (1987) Hypervariable Bkm DNA loci in a moth, Ephistia kuehniella." Does transposition cause restriction fragment length polymorphism ? Genetics 115 : 493-498 Turner BJ, Elder JF, Laughlin TF, Davis WP (1990) Genetic variation in cional vertebrates detected by simple-sequence DNA fingerprinting. Proc Natl Acad Sci USA 87:5653 5657 Westneat DF (1990) Genetic parentage in the indigo bunting: a study using DNA fingerprinting. Behav Ecol Sociobiol 27:67 76