Molecular studies of Callithrix pygmaea ( Primates , Platyrrhini ) based on transferrin intronic and ND 1 regions : implications for taxonomy and conservation

As classificacoes tradicionais envolvendo os macacos da infraordem Platyrrhini, principalmente baseadas em caracteristicas morfologicas, tem sido contestadas por dados moleculares recentes. A subfamilia Callitrichinae (Platyrrhine, Primates) engloba um diverso grupo de especies, muitas das quais consideradas em perigo de extincao. A presente analise de duas regioes do DNA, um gene mitocondrial (ND1) e um gene nuclear (regioes intronicas da transferrina), sugerem que Callithrix pygmaea apresenta variabilidade suficiente para justificar a existencia de subespecies ou ate mesmo de especies distintas. As arvores filogeneticas baseadas na regiao do ND1 indicam que esta especie esta relacionada mais proximamente aos marmosets amazonicos do que aos da mata Atlântica. Estes resultados reabrem a discussao sobre diversidade e programas de conservacao baseados apenas em classificacoes taxonomicas tradicionais.


INTRODUCTION
According to Schneider et al. (1996) the New World monkeys can be classified into two families: the Atelidae and the Cebidae.The latter is divided into three subfamilies: Cebinae (Cebus and Saimiri), Aotinae (Aotus) and Callitrichinae (Saguinus, Leontopithecus, Callimico and Callithrix).Some authors grouped the Callimico genus within the family Callitrichidae but allocated it to a distinct subfamily, the Callimiconinae (Napier and Napier, 1967;Ford, 1986;Kay, 1994).Dollman (1933) and Hershkovitz (1977) classified Callimico in its own family, the Callimiconidae.Thomas (1913); Simpson (1945); Stirton (1951), Cabrera (1958), and Simons (1972) included Callimico in the Cebidae family.There is a consensus that Callithrix, Leontopithecus and Saguinus all belong to the same clade but the evolutionary relationships between these genera are still controversial (Rosenberger, 1981;Ford, 1986;Schneider et al., 1996;Pastorini et al., 1998).The genus Callithrix has a wide distribution and can be divided into three distinct groups of species: the argentata group, represented by marmosets from the Amazonian forest and part of the cerrado (a type of Brazilian savanna); the jacchus group, comprising marmosets from the Atlantic forest, the cerrado and the caatinga (a dry region of stunted vegetation and brushwood) (Hershkovitz, 1977), and the pygmaea group made up of pygmy marmosets (C.pygmaea) from the Amazonian forest (Barroso et al., 1997).The pygmy marmoset was first classified as Cebuella Gray, 1866, a genus distinct from the already described Callithrix Erxleben, 1777.However, recent molecular data indicate that the pygmy marmoset should be classified as C. pygmaea (Schneider et al., 1996;Barroso et al., 1997;Porter et al., 1997).Previously, Rosenberger (1981Rosenberger ( , 1984) ) had drawn attention to morphological similarities between pygmy marmosets and species of Callithrix that could justify their classification within the same genus.
Considering the system adopted by the World Conservation Union (WCU) (see Rylands et al., 1995Rylands et al., , 1997)), several species and/or subspecies of Callitrichinae are classified as critically endangered or endangered as a consequence of habitat destruction and illegal capture or hunting.According to Rylands et al. (1995Rylands et al. ( , 1997) ) the following callitrichines may be classified as critically endangered: Leontopithecus rosalia, L. chrysopygus, and L. caissara whereas those on the endangered list are L. chrysomelas, C. flaviceps, C. aurita, Saguinus bicolor bicolor, and S. oedipus.Although C. pygmaea is affected by illegal international trade, particularly in the area of Iquitos in Peru, it is classified in the WCU's Lower Risk category (Rylands et al., 1995).Callithrix pygmaea is a small monkey and thus is not hunted for food by humans.Furthermore, it has a wide geographical distribution and is capable of existing in isolated forest patches near human settlements (see Hershkovitz, 1977;Rylands et al., 1993).
The purpose of the present study was to provide a molecular view of the relationships among the callitrichines with particular emphasis on C. pygmaea and to highlight the importance of the use of molecular studies in designing conservation programs.quenced, appear in Table I.Monkeys were anesthetized with Ketalar (10 mg/kg body weight) and blood samples collected from femoral veins.Blood samples were centrifuged and DNA extracted from the isolated leukocytes according to the protocol of Sambrook et al. (1989).
DNA sequences of the transferrin (exon 4 to exon 6) and ND1 regions were determined by direct sequencing of PCR-amplified fragments.For the ND1 fragment, a second PCR (internal) was performed on the fragment resulting from the first PCR in order to eliminate false priming products that occasionally arise in the original genomic DNA PCR.Primers for the amplification of these regions are described in Table II.The DNA sequences were determined using dye terminator cycle sequencing reactions that were subsequently loaded onto an automatic sequencer (Applied Biosystems model 373A) according to the manufacturer's protocols.Additional sequencing primers were designed as necessary.
Initial sequence alignments were performed using the Eyeball sequence editor (Cabot and Beckenback, 1989).Data were analyzed by the neighbor-joining (NJ) method using the MEGA program (Kumar et al., 1993) whereas maximum-parsimony (MP) and maximum-likelihood (ML) analyses were performed with PAUP 4.0 beta version (Swofford, 1998) using a branch and bound algorithm search (Hendy and Penny, 1982).The robustness of the phyloge-netic hypothesis obtained by MP was tested by bootstrapping (Felsenstein, 1985) and the criterion adopted to evaluate it was that suggested by Hillis and Bull (1993) who consider bootstrap values equal or superior to 70% to be statistically significant.All bootstrap analyses of DNA sequence data involved at least 2000 replications and all gaps were considered as single events.NJ analyses of the DNA sequence data were performed using the method of Jukes and Cantor (1969) for the transferrin fragment and that of Tamura and Nei (1993) for the mtDNA region (ND1).The reliability of the topologies of the NJ tree branches were tested by the standard error test (SET), using the method of Jukes and Cantor (1969) for transferrin, the Kimura twoparameter distance values for the mtDNA region (Kimura, 1980), and the MEGA program (Kumar et al., 1993).
The estimated time since divergence was calculated by means of the branch lengths of the NJ trees and the method proposed by Sarich and Wilson (1973), using the estimated time of 13 million years ago (Mya) for the emergence of the tribe Callitrichini (Goodman et al., 1998).

RESULTS
The transferrin sequences determined for callitrichine representatives are 1662 bp in length but only the intronic regions, corresponding to 1472 bp, were used in the phylo- genetic analyses.The trees obtained by NJ, ML and MP analysis all agree in their topology.The phylogenetic tree (Figure 1) shows both C. pygmaea individuals joining together at the same branch-point, an arrangement statistically supported by bootstrap (96%) and SET (95%) analysis.C. jacchus and C. pygmaea form a sister group, also considered significant by both bootstrap (100%) and SET (99%) analyses.The next representative to join this sister group is Callimico, supported only by bootstrap (76%) analysis.Saguinus and Leontopithecus were the most basal genera, but it was not possible to determine which one was the first to split.
One 1321-bp fragment spanning the ND1 fragment and flanking regions was sequenced, although only sequences of 951 bp that corresponded exclusively to the ND1 gene were considered in the analysis.NJ, ML and MP methods) obtained using ND1 data are not significant and no inference can be made as to the relationships between the most basal genera of the subfamily Callitrichinae and the species of the jacchus group.
Comparison between the DNA sequences of the two pygmy marmosets sampled (CPY-104 and CPY-105) showed important differences.The intronic transferrin fragments of 1472 bp (exon 4 to 6) had five transitions and two transversions.Sixty-two transitions and three transversions were detected among the ND1 sequences of the two pygmy marmosets sampled, which generated nine differences in the amino acid sequence of the ND1 protein.Comparison of the two sequences of the D-loop and adjacent region (GenBank numbers U89009, U89010) showed that there were 59 transitions and 36 transversions, along with three deletions of 1 bp and one of 15 bp.
Tables III-V show the divergence matrices obtained for representatives of the subfamily Callitrichinae.Considering the ND1 fragment, both C. pygmaea individuals showed greater differences in their sequences and, consequently, in the divergence matrix (Table IV) than were observed in comparisons between the four species of the jacchus group.A previous study (Tagliaro et al., 1997) showed that the D-loop divergence matrix (Table V) also indicated considerable divergence between the two C.  pygmaea individuals.This value is comparable to values obtained between species of marmosets of different groups.
Based on the NJ dendrograms obtained for transferrin and ND1 sequences (Figures 1 and 2) and considering an estimate of 13 Mya (Goodman et al., 1998) penicillata, C. kuhli, C. geoffroyi).This agrees with the results obtained by Canavez et al. (1996) based on karyology, and Tagliaro et al. (1997), Porter et al. (1997) and Almeida (1995), based on DNA sequences.The studies of Barroso et al. (1997) and Sena (1998), also based on DNA analysis, are inconclusive in defining whether C. pygmaea is more closely related to the argentata group or to the jacchus group.Based on detailed studies of morphological features and feeding habits, Rosenberger (1981Rosenberger ( , 1984) ) considered that the pygmy marmoset should be classified within the genus Callithrix, although this researcher believed that the pygmy marmoset was closely related to the jacchus group, a theory which contrasts with the molecular evidence.
The two C. pygmaea individuals were more different than would be expected for a monospecific group in which no subspecies have yet been described.There was a divergence of 7.4% in the ND1 divergence matrix which corresponds to values observed between marmosets of different groups.In the D-loop divergence matrix (Table V) the divergence between the two C. pygmaea individuals was 11.2%.Similar values have been found only between C. aurita and other species of the jacchus group (C.aurita representing the most basal species of this group).All other intra-group divergence values were inferior to 8.0%.
The calculated timing of separation of the two lineages of C. pygmaea ranges from 1.8 Mya (transferrin) to 5.4 Mya (ND1).In relative terms these dates are informative, being earlier than those for the emergence of the present species of the amended jacchus group in the same studies.Hershkovitz (1977) considered the possibility of the existence of subspecies of C. pygmaea, but did not agree with the subspecies C. p. niveiventris Lonneberg 1940, andVan Roosmalen andVan Roosmalen (1997) reopened this debate after observing the niveiventris morphotype in areas where it had not been thought to occur.Meireles et al. (1998), investigating protein variability, found that C. pygmaea (N = 8) was the most variable species in a study involving 13 species and subspecies of callitrichines.
Apparently Callithrix pygmaea consists of two different subspecies or perhaps even two species.However, the individuals used in this study, like those in the study by Meireles et al. (1998), were captive animals from the "Centro Nacional de Primatas".These animals and their ancestors were of unknown geographical origin.In order to obtain more conclusive evidence, further molecular studies should be undertaken with more individuals of known geographical origin.
The 21st century is going to be a critical period for the conservation of biodiversity, particularly in richly biodiverse areas such as the neotropical forests.Descriptions of new  species and subspecies of plants and animals, some of which are restricted to small geographical areas or are from peculiar habitats, continue to be published, and species already described are having their taxonomic status revised, while habitat destruction is occurring at such a rate that many species become extinct before even being described.
Molecular genetics is a new approach to taxonomic classification.It is helping to evaluate biodiversity and aid decision-making concerning to the conservation priorities of many different taxa.The recognition of groups that exhibit very little evolutionary differentiation and those that are phylogenetically distinct allows direction of conserva- tion efforts to protect the maximum biological diversity so that priority can be given to those taxa most at risk of extinction.The possibility of the existence of different species and/or subspecies of Callithrix pygmaea surely necessitates a re-evaluation of their conservation status, particularly in the light of the trade in these small monkeys in certain areas such as the Iquitos region of Peru (Rylands et al., 1993).

ACKNOWLEDGMENTS
We would like to thank Dr. José Augusto Pereira Carneiro Muniz (Centro Nacional de Primatas, Belém, PA, Brazil), Dr. Adelmar Coimbra-Filho and Dr. Alcides Pissinati (Centro de Primatologia do Rio de Janeiro, RJ, Brazil), Arlindo Pinto de Souza Jr., Milton Thiago de Melo, and Criadouro Barbuse Leal for the samples used in this research.We thank Dr. Colin Robert Beasley for checking the English.C.H.T. was supported by the Brazilian research agency CNPq.This research was made possible by grants to M.J.S. from the Nuffield Foundation, the Royal Society, and Northern Ireland Development and Research.
as the time of the split of the Callitrichini from the other Platyrrhini, the timing of the separation of the lineages of the two C. pygmaea individuals was estimated as 1.8 Mya for transferrin and 5.4 Mya for ND1.The estimate of the time of separation between the two C. pygmaea individuals obtained by NJ dendrogram analysis (see Figure 3) of D-loop sequences was 4.0 Mya.
the ND1 region indicate that C. pygmaea is more closely related to other Amazonian marmosets (argentata group, represented in this study by C. argentata) than to those of the Atlantic forest or the cerrado and caatinga (jacchus group; represented by C. jacchus, C.

Table I -
The origin and identification of various New World monkeys for which DNA sequences were determined.
*Animals in captivity and unknown origin.
Callithrix clade (containing representatives of species of the Callithrix genus, including C. pygmaea) was also supported by bootstrap (99%) and SET (99%) analysis.The arrangement of the dendrograms (produced by the

Table II -
Primers used for the amplification of DNA fragments of transferrin, and ND1.