Introduction

Penguins (order Sphenisciformes, family Spheniscidae) are aquatic, flightless birds with their native habitats located mostly in the Southern Hemisphere between latitudes 45° and 60° S. Phylogenetically, they are closely related to the Procellariiformes order of seabirds such as albatrosses, shearwaters, and petrels (Ho et al. 1976; Sparks and Soper 1987; Hackett et al. 2008). The penguins, based on molecular and morphological phylogenetic evidence and fossil taxa, are believed to have originated in Gondwanaland from an ancestor of the flying seabirds about 71 mya with the emergence of the extant basal lineage Aptenodytes about 40–49 mya (Davis and Darby 1990; Sibley et al. 1988; Baker et al. 2006). The geographical expansion and radiation of most of the present-day penguin species probably did not occur until between 35 and 4 mya (Baker et al. 2006). It was proposed recently on the basis of multiple gene analysis that the modern penguins expanded from Antarctica by way of the circumpolar current to oceanic islands and eventually to the southern continents after Antarctica became ice-encrusted 35–25 mya (Baker et al. 2006), possibly involving two separate dispersals to the equatorial regions (Clarke et al. 2007).

The extant penguins are classified into six genera and 16–18 species (del Hoyo et al. 1992; Jouventin 1982; Baker et al. 2006; Clarke et al. 2007), but inconsistencies and questions still remain about their classification and phylogenetic relationships particularly with respect to the closely related species, such as the Humboldt penguin (Sphe humboldt), the Magellanic penguin (Sphe magellanicus), the African penguin (Sphe demersus), and the Galapagos penguin (Sphe mendiculus). A geographic map of wild penguin location shows that the Sphe species are distributed widely in the subantarctic and their locations range from the southern coastal regions of South Africa to South America, but also as far north as the equatorial Galapagos Islands of the Pacific Ocean (Fig. 1). Although these four species have distinctly different geographic habitats, they still share a number of common traits, including morphological similarities, living in temperate climates and nesting in burrows. They are collectively known as “the banded penguins” because of their similar coloration and external forms. Furthermore, the Sphe appears to be a penguin genus that is capable of interspecies breeding (Thumser and Karron 1994; Williams 1995; McCarthy 2006).

Fig. 1
figure 1

A global distribution map of wild penguins in their natural habitats. This map was modified from a figure by Mackintosh (1960) and Williams (1995). Breeding places of Spheniscus (Humboldt penguin, Magellanic penguin, Galapagos penguin, and African penguin), Pygoscelis (Adelie penguin, chinstrap penguin, and Gentoo penguin), and the little penguin are indicated

Full size image

The Mhc is a high-variable genomic region within and between different vertebrate species due to duplicated gene clusters involved in the gene control of the immune response and the influence of positive (balancing) selection (Hughes and Nei 1989a, b; Ellis and Ballingall 1999; Ellis 2004). The Mhc DRB-like gene is known to be a useful marker of trans-species allele distribution between closely related species of ungulates, canivores, birds, primates, and various other mammals (Bollmer et al. 2007; Cutera and Lacey 2007). Most research on the vertebrate Mhc class II DRB-like genes has been done on mammals (Trowsdale 1995), and studies on the avian Mhc DRB-like genes have been restricted primarily to gallinaceous birds (Zoorob et al. 1993; Kaufman et al. 1999; Wittzell et al. 1999; Hosomichi et al. 2006; Chaves et al. 2007; Strand et al. 2007), several passerines (Edwards et al. 1995a, b, 1998, 2000; Gasper et al. 2001; Hess et al. 2000; Aguilar et al. 2006), raptors (Alcaide et al. 2008; Burri et al. 2008), a shore bird (Ekblom et al. 2007), and some penguins (Tsuda et al. 2001; Kikkawa et al. 2005; Bollmer et al. 2007).

In this study, we have further examined the Mhc class II DRB-like gene as a potential evolutionary and adaptive molecular marker in four closely related penguin Sphe species. We sequenced 21 new penguin alleles of the Mhc class II DRB-like gene sequence from seven Humboldt, six African, five Magellanic, and three Galapagos penguins and examined their allele distribution and variation. Allele sharing of the DRB-like gene sequences in the genus Sphe emphasizes that they are closely related species with an allopatric distribution and that some common alleles have undergone trans-species inheritance probably because of balancing selection or recent rapid speciation.

Materials and methods

Collection of penguin samples, DNA extraction

Our field collaborators obtained blood samples from 24 captive penguins at aquariums and zoological garden in Japan and Korea and from ten wild penguins in Chile, Argentina, and the Galapagos Islands of Ecuador (Table 1). The 20 Humboldt penguin samples listed in Table 1 are the same as previously described by Kikkawa et al. (2005). A liver sample for RT-PCR was obtained from a captive Humboldt penguin at Oji Zoo in Japan.

Table 1 The DRB1 allelic details and origins for 34 captive or wild penguins of the Spheniscus genera
Full size table

DNA was extracted from peripheral blood cells of seven Humboldt penguins (six captive and one wild), six captive African penguins, five Magellanic penguins (three captive and two wild), and three wild Galapagos penguins by the standard phenol–chloroform extraction method after proteinase K treatment (Inoko et al. 1986).

Genomic and RT-PCR

Total RNA was isolated from the liver sample of a captive Humboldt penguin using the TRIZOL reagent method. The cDNA was synthesized using a GeneAmp RNA PCR kit (Applied Biosystems). The primers RT1-ex2F and RT3-ex3R were newly designed for the RT-PCR assay (Fig. 2). These primers were also used to amplify a 568 bp genomic segment of exon 2 to exon 3, including intron 2 of the Humboldt penguin DRB-like gene. The PCR amplification was performed in 20 μl volumes containing 10× Ex Taq buffer, 0.2 mM of each dNTP, 0.5 μM of each primer, 0.5 units Ex Taq (TaKaRa, Shiga, Japan), and 50–100 ng of cDNA or genomic DNA. Amplification included an initial incubation at 96°C for 3 min followed by 35 cycles with each cycle temperature profile consisting of 96°C for 30 s, 60°C for 30 s, 72°C for 1 min, and then a final cycle at 72°C for 5 min. The same DNA and cDNA sample was amplified using b-antin-F and b-actin-R primers (Shimizu et al. 2004) as a control. The PCR products were loaded on a 1.5% agarose gel with TBE buffer and run at 100 V for 45 min at room temperature.

Fig. 2
figure 2

Information and gene map of the PCR oligonucleotide primer locations used to amplify portions of the penguin Mhc class II DRB1-like gene. The previously published five primers (LP1, Lpen.hum1F, Lpen.hum2F, Lpen.hum2R, and Lpen.hum3R; Kikkawa et al. 2005) used for PCR and sequencing are included in the figure. The six newly designed primers for PCR and sequencing (RT1, RT2, Lpen.5′UTR-F, Lpen.hum1F2, Lpen.hum 2F2, and Lpen.hum 3R2) are boxed. The LP1 primer was designed on the basis of exon 1 sequences of birds, but also included a portion of the 5′UTR sequence

Full size image

Amplification and sequencing

A total of nine PCR primers were used to amplify selected portions of the Sphe DRB1-like gene by PCR for direct sequencing or sequencing after cloning the PCR products. Figure 2 shows the PCR primer sequences, the gene sequence location of the primers relative to 5′UTR and intron 3 of the DRB1 gene, and primer orientation (sense and antisense). The previously designed five primers (Kikkawa et al. 2005) and four newly designed primers (Lpen.5′UTR-F, Lpen.hum.1F2, Lpen.hum.2F2, and Lpen.hum.3R2) were used for PCR amplification.

The DNA sequences from the 5′UTR to intron 2 of the DBR1-like gene were amplified from 34 penguin genomic DNA samples (Table 1) using the previously published LP1 and Lpen.hum 2R primer pairs (Kikkawa et al. 2005). The PCR amplification was performed in 25 μl volumes containing 50–100 ng of genomic DNA, 10× NH4 buffer, 2 mM MgCl2, 0.2 mM of each dNTP, 0.12 μM of each primer, and 0.02 units of BIO Taq (BIO LINE). The PCR product of the expected target length (568 bp) was not obtained successfully with other Taq commercial products, Ampli Taq and Gold Taq (Applied Biosystems). Amplification included 96°C incubation for 3 min followed by 35 cycles with each cycle at 96°C for 30 s, 61°C for 30 s, 72°C for 2 min, and then a final cycle at 72°C for 5 min. Similarly, the 34 penguin genomic DNA samples were used to amplify the DRB1-like sequences of intron 1 to intron 3 using two other primer sets, the previously published Lpen.hum1F and Lpen.hum.3R (Kikkawa et al. 2005) and the new set Lpen.hum.1F2 and Lpen.hum.3R2 (Fig. 2). The PCR condition was the same as described above for the LP1 and Lpen.hum 2R primer pairs, except that only the annealing temperature was changed to 67°C for the Lpen.hum1F2 and Lpen.hum.3R2 primer set.

The nucleotide sequences of the PCR products were determined directly by using the Big Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems) with the LP1, Lpen primers, and the ABI PRISM 3130 automated sequencer (Applied Biosystems). The PCR products were also subcloned into the plasmid pT7Blue T vector (Novagen, CA, USA) for sequencing using the Ligation convenience kit (NIPPON GENE, Tokyo, Japan) and introduced by transformation into the competent cell DH5a (TOYOBO, Osaka, Japan). At least ten subclones per PCR product were used to determine their nucleotide sequences with all ten primers (Fig. 2).

Sequence data analysis

All sequences were assembled using a SEQUENCHER and analyzed using the GENETYX MAC software package for Apple computers. The gene trees were constructed using the neighbor-joining method with the Kimura 2 parameter settings provided by the MEGA 2 software. Sequences that were homologous to our sequences were found in the GenBank database (www.ncbi.nlm.nih.gov) by using online BLAST searches. We submitted our sequences to DDBJ and obtained accession numbers as part of the international nucleotide sequence database collaboration with EBI/EMBL and NCBI/GenBank (www.ddbj.nig.ac.jp/Welcome-j.html).

Southern hybridization

Genomic DNA (15 μg) was digested with 60 U of the restriction enzyme PstI (NEB, England) and electrophoresed on 0.8% TAE agarose gel. The DNA was transferred from the agarose gel to Hybond-N+ membrane (Amersham, England) for hybridization with the Lpen.hum1F/2R PCR product and RT1/RT2 PCR product used as probes. The probes were labeled using the PCR DIG Probe synthesis kit (Roche, Germany). The hybridizations were detected by using AlkPhos Direct Labelling and Detection System with CDP-Star (Amersham, England).

Results and discussion

Genomic- and RT-PCR and Southern hybridization

The newly designed primers RT1-ex2F and RT3-ex3R (Fig. 2) were used successfully in a separate RT-PCR and genomic PCR assay to amplify a 313 bp cDNA and a 568 bp genomic segment from the expressed liver cDNA and genomic fragment of the Humboldt penguin DRB-like gene, respectively (Fig. 3). Sequencing data of the PCR products (not shown) revealed that the 313-bp cDNA band expressed by the liver represented the nucleotide sequence of exon 2 and exon 3 of the DRB-like gene, whereas the 568 bp single band represented exon 2 to exon 3, including intron 2 that was amplified from the Humboldt penguin DRB-like gene of the genomic DNA extracted from the liver. The homogeneity of the direct sequencing results (data not shown) showed that these and other primers (Kikkawa et al. 2005) amplified sequences from a single encoding DRB1-like gene and not multiple heterologous DRB-1-like gene sequences that may or may not be present in penguins because of DRB1 gene duplications as in chickens and other birds (Shiina et al. 2004). Indeed, single band was obtained as a result of Southern hybridization (Fig. 4).

Fig. 3
figure 3

Gene expression analysis by RT-PCR. The DNA and cDNA isolated from the liver of a Humboldt penguin were amplified using the RT1 and RT3 primer pair and b-actin primers as control. The RT1 and RT3 primer pair amplified a PCR product of 568 bp using genomic DNA in a standard PCR and 313 bp using cDNA in a RT-PCR. The b-actin primer pair amplified a product of 887 bp using genomic DNA in a standard PCR and 353 bp using cDNA in a RT-PCR

Full size image
Fig. 4
figure 4

Southern hybridization analysis of Humboldt penguin. The left panel shows the 832 bp single band obtained with Lpen.hum1F/2R probe, and the right shows 833 and 370 bp band obtained with RT1/RT2 probe because of having one PstI recognition site in sequence

Full size image

Nucleotide and amino acid sequence diversity and allele frequency variation

About 1.5 kb of nucleotide sequence from the 5′UTR of exon 1 to intron 3 of the DRB1-like gene was determined for 34 genomic DNA samples from the four Sphe species (Table 2). A comparison of nucleotide sequence differences detected within intron 1, exon 2, intron 2, exon 3, and intron 3 of the 22 haplotypes of the DRB1-like gene of the Galapagos penguin, Humboldt penguin, African penguin, and Magellanic penguin is shown in Fig. 5. All four species had the identical nucleotide sequences at exon 1 (70 bp). The three alleles Sphu 001, 002, and 0011 of the Humboldt penguins had deletions at nucleotide position 94 in intron 1 and a “C” stretch of >11 bp (351 to 361 bp or 352 to 362 bp) where the precise copy number of cytosines was unknown. Thus, the complete lengths of intron 1 for the Humboldt penguins were estimated to be about 514 or 515 bp. There were eight or nine deletion sites within intron 1 of the Magellanic penguins, but no evidence of frameshift mutations or premature stop codons in any of the samples that we sequenced. The number of different alleles that we identified was eight for the Humboldt penguin, four for the African penguins, eight for the Magellanic penguins, and two for the Galapagos penguins. However, the sequences of the Sphu004 allele (AB301950) in a Humboldt penguin and the Spma001 allele (AB302843) in a Magellanic penguin were identical and varied by only three nucleotides from the Spma007 allele (AB303945) of a Magellanic penguin. A 198-bp portion of the exon 2 sequences for alleles Spma001, Spma004, Spme001, and Spme002 have been previously published as accession numbers (alleles) EF212010 (Spma1), EF212011 (Spma2), EF212008 (Spme2), and EF212007 (Spme1) by Bollmer et al. (2007). Overall, the nucleotide differences between the four penguin species varied from 1 (0.07%) to 27 (1.83%).

Table 2 The allele frequencies of DRB-like gene of Spheniscus
Full size table
Fig. 5
figure 5

Comparison of nucleotide sequences obtained for exon 1, intron 1, exon 2, intron 2, and exon 3 of 22 haplotypes of the DRB1-like gene of the Spheniscus species (Humboldt penguin, African penguin, and Magellanic penguin, Galapagos penguin). This figure showed only variation sites, sequence sites number (bp) above the Spme001 sequence, and determined sequence length of each gene allele on the right side. There was no variation site on exon 1 sequence. Asterisks of position 94 on intron 1 indicate the deletion sites. The number and percentage difference of nucleotides for each gene allele relative to the Sphu001 sequence is listed in the column on the right

Full size image

Figure 6 shows a comparison of the amino acid sequences translated from the nucleotide sequences of exon 2 for 22 allele variants of the DRB-like gene of Sphe and the DRB1*0101 allele sequence of the human that was included here as an outgroup reference. Three hypervariable regions (HV1 to HV3) were identified in the penguin sequences that corresponded to the well-defined hypervariable regions within the human DRB1*0101 sequence (Stern et al. 1994; Brown et al. 1993). We obtained identical amino acid sequences in exon 2 of the Magellanic penguin alleles Spma001 and Spma007 and the Humboldt penguin allele Sphu004.

Fig. 6
figure 6

Comparison of amino acid sequences obtained for exon 2 of 22 alleles of the DRB-like gene of Spheniscus (Humboldt penguin, African penguin, Magellanic penguin, and Galapagos penguin). The bottom sequence is the DRB1*0101 allele sequence of human. The vertical boxed columns indicate the hypervariable regions (hypervariable region I (HV1), hypervariable region II (HV2), and hypervariable region III (HV3)) for human. The number and percentage difference of amino acid sites for each allele relative to the Sphu001 amino acid sequence is listed in the column on the right. Sphu004 and Spma001 have the same nucleotide sequence

Full size image

The number and percentage difference of the amino acids within the different haplotype sequences of Sphe varied from 4 (4.5%) to 15 (16.9%). In addition, the base difference between alleles of exon 2 (270 bp) within each of the Sphe species and the human DRB1 alleles was calculated. The average value of the base difference between Japanese high frequency DRB1 alleles (Hashimoto et al. 1994) was 21.2 bp (maximum 37 bp and minimum 1 bp), whereas it was 15.1 bp (21 and 5 bp) for the Magellanic penguin and 13.6 bp (19 and 1 bp) and 20.3 bp (28 and 8 bp) for the African penguin.

The allele percentage frequencies of the DRB-like gene of ten wild and 24 captive Sphe ranged between 2.5% and 75% and are shown in Table 2. The greatest number of alleles (eight each) was for the Humboldt and Magellanic penguins that both reside naturally on the southern coasts of South America (Fig. 1). Four of the five Magellanic penguins and 50% of the 20 Humboldt penguins were heterozygous when applying our PCR conditions. The homozygosity for our set of primers was high at 75% and 83% for the six African penguins and three Galapagos penguins, respectively, although there is a possibility that our PCR assay did not amplify all allelic forms and that some of these homozygotes may in fact be heterozygotes.

The sample numbers in our study are too low to be representative of actual native population frequencies, which still need to be determined for all penguin species. Therefore, we cannot conclude with certainty whether there is an unnatural increase in the homozygosity of the DRB1 gene in any of the Sphe species and whether there are fewer sequence variations within these species that are due in part to population or environmental factors or an increase in inbreeding activity. Because of the absence and/or small sample sizes for these species, we may have underestimated the actual number of DRB1 alleles and the genetic diversity among the study taxa. However, our results are consistent with the findings of Bollmer et al. (2007) who recently reported a low genetic variation for exon 2 at the DRB1 locus of 30 Galapagos penguins. Taken together, these findings confirm that the polymorphic DRB1 gene is an important marker for assessing the genetic health of these endangered species and its genetic variations may have an application for conservation management of wild animals and the efficient breeding management of captive penguins.

Nonsynonymous and synonymous nucleotide substitution rates at exons 2 and 3

The detection of nonsynonymous and synonymous substitution ratios within the Mhc class II amino acid sequences with values of greater than one provides strong evidence for positive or balancing selection acting on the gene sequence. The nonsynonymous (d N) and synonymous (d S) substitution rates and ratios for exon 2 (90 AA) and 3 (94 AA) of DRB-like gene for each species of Sphe are shown in Table 3. The d N/d S ranged in value from 11 to possibly greater than 16 for the region of exon 2 for all four species of Spheniscus, and the ratio was one and three for exon 3 of the Humboldt penguin and the Magellanic penguin, respectively, confirming the affects of balancing selection on the penguin DRB-like gene as previously indicated for exon 2 by Bollmer et al. (2007). The large nonsynonymous substitution rates at exon 2 of the Spheniscus DRB1 gene is consistent with strong positive selection forces acting on the genetic region of antigen presentation in order to cope with various infections and parasites.

Table 3 Nonsynonymous (d N) and synonymous (d S) substitutions rate of the DRB-like gene for each species of Spheniscus
Full size table

Trans-species allele sharing of the penguin DRB1 gene

The reconstruction of a gene tree by the neighbor-joining method of exons 1, 2, and 3 nucleotide sequences (622 bp) of 20 DRB-like gene sequences from the four species of Spheniscus and one chicken Mhc class II beta chain sequence (chosen as the outgroup) is shown in Fig. 7. The gene tree branching has revealed allele sharing between the Humboldt penguin and the Magellanic and Galapagos penguins and some allelic overlaps between the African penguin and the Galapagos and Magellanic penguin sequences. The Humboldt penguin sequence Sphu004 that is identical to the Magellanic penguin sequence Spma001 clustered with other Magellanic penguin sequences Spma006 and Spma007. The low bootstrap percentages (<70%) seen on the tree branches highlight the uncertainty of the branching nodal origins.

Fig. 7
figure 7

Rooted gene tree of complete exons 1, 2, and 3 sequences (622 bp) for the DRB-like gene for four species of Spheniscus. Gene tree analysis was performed by the neighbor-joining method. The chicken Mhc class II beta chain sequence was used as an outgroup (Xu et al. 1989) to root the tree. The numbers on the branches show the bootstrap values as a percentage calculated after 1,000 replications. Allele names of Spheniscus were indicated as Sphu for the Humboldt penguin, Spma for the Magellanic penguin, Spde for the African penguin, and Spme for the Galapagos penguin

Full size image

In order to view the extent of the trans-species polymorphisms of the DRB-like gene alleles in penguins, we used the neighbor-joining method to reconstruct an unrooted gene tree of nucleotide sequences (198 bp) from exon 2 for 21 representative DRB-like gene sequences from the four species of Spheniscus and 20 previously reported orthologous sequences from four other penguin species, the little penguin (Eudyptula minor), the Adelie penguin (Pygoscelis adeliae), the chinstrap penguin (Pygoscelis antartica), and the Gentoo penguin (Pygoscelis papua). The unrooted gene tree is shown in Fig. 8, and it can be seen that the gene sequences formed into distinct clades for all the species, except for the Spheniscus species. The gene alleles obtained from the genus Spheniscus formed a separate clade to Pygoscelis and Eudyptula, but the gene alleles obtained from the four species within the genus Spheniscus grouped into intermingling clusters. The gene tree also shows that the DRB1 alleles of the little penguin genus Eudyptula is intermediate between the sequences obtained from the genera Spheniscus and Pygoscelis.

Fig. 8
figure 8

Unrooted gene tree of the DRB-like gene partial exon 2 sequences (198 bp) for the four species of Spheniscus and the orthologous sequences for four other penguin species taken from Tsuda et al. (2001). Gene tree analysis was performed by the neighbor-joining method. Allele names of Spheniscus are indicated as Sphu for the Humboldt penguin, Spma for the Magellanic penguin, Spde for the African penguin, and Spme for the Galapagos penguin. The abbreviations Gentoo_N and Gentoo_S refer to the Gentoo penguins of the northern and southern Antarctic regions

Full size image

The sharing of identical Mhc class II alleles has been documented in primates (Kenter et al. 1992; Doxiadis et al. 2006; Otting et al. 2002), the European bison (Radwan et al. 2007), and in the Humboldt and Magallenic penguins (Bollmer et al. 2007 and present study). We found that the allele sequences, named Sphu004 for the Humboldt penguin and Spma001 for the Magellanic penguin, were identical from the 5′UTR end to intron 3 end of the DRB1 gene. These identical allelic sequences were homozygotes in a captive Humboldt and a Magellanic penguin, but their frequency in the wild penguin populations is not known. According to Bollmer et al. (2007), the identical alleles may have arisen because of an ancestral hybridization event as the distribution of these two species overlap in the wild.

Although the sharing of DRB1 gene variants between species may imply interspecific hybridization between members of different species despite their allopatric isolation, recent rapid speciation events, gene conversion, or maintenance of trans-species alleles because of balancing selection provide alternative explanations (Meyer-Lucht et al. 2008). Allele or haplotype sharing between species suggests an inherent closeness and recent speciation or divergence events. Speciation within Spheniscus based on the reconstructed phylogeny of nuclear and mitochondrial DNA sequences is considered to have occurred mainly within the last 4 myrs with the African and Magellanic penguin sister groups and the Galapagos and Humboldt penguin sister groups emerging on the islands and coastal land masses of the Atlantic Ocean and Pacific Ocean, respectively (Fig. 1), well after the Spheniscus and Eudyptula split about 25 mya (Baker et al. 2006). Therefore, recent speciation is a reasonable explanation for the trans-species allele sharing of the DRB1 genes within Spheniscus, particularly if speciation was completed before fixation of the segregating alleles. With recent, rapid speciation events, closely related species can be expected to share alleles or sequences more frequently than more distantly related species as supported by the results of our reconstructed gene tree for the DRB1 sequences of the four closely related Spheniscus species and the more distantly related four Eudyptula and Pygoscelis species.

Numerous previous studies have implicated balancing selection in the retention of Mhc alleles between closely related species for much longer periods than expected under neutral evolution due to a population's interaction with infectious agents (Klein 1987; Takahata 1993; Cutrera and Lacey 2008). However, it is often difficult to distinguish between the effects of balancing selection, gene conversion, and rapid speciation on allele retention between species, which in reality probably depends on one or all of these processes to some degree (Meyer-Lucht et al. 2008).

Conclusion

We have developed a PCR amplification method to sequence RT-PCR products of the expressed DRB1 gene and overlapping portions of the DRB-like gene from the end of the 5′UTR to intron 3, including exons 1, 2, and 3 and introns 1 and 2 in four species of Spheniscus. These are the longest Mhc class II B gene sequences (946 to 1,586 bp) obtained so far in a study of Spheniscus. In the sequence analysis, we confirmed that there was little nucleotide sequence variation within the exon 1 and exon 3 regions of the gene, but that there were many polymorphic or variable sites and a high rate of nonsynonymous substitutions within exon 2 that corresponds to the hypervariable regions of the human DRB1 gene. It is known that the hypervariable regions of the human DRB1 gene are binding sites that interact with a variety of short antigenic peptide sequences involved in the adaptive immune regulatory system (Klein 1986). In addition, balancing selection is believed to contribute to the allele sequence variation of the DRB1 genes at Mhc loci where the rates of nonsynonymous base pair substitutions are usually greater than expected for neutral loci (Hughes and Nei 1988, 1989a, b). On the basis of these analyses and the comparison of the Spheniscus sequences to the human DRB1 gene, we conclude that the penguin sequences using our PCR amplification method represent a classical single copy class II gene of the Mhc. The specificity of our PCR method however does not exclude the possibility that penguins have duplicated copies of the DRB1-like sequence that have yet to be identified using other detection methods. The reconstruction of the penguin DRB1 gene tree could be greatly improved in future studies by incorporating greater sample numbers and providing a more representative analysis of all 16–18 extant penguin species. The addition of the DRB1-like sequence to other gene sequences with neutral evolutionary histories for a multigenic approach to molecular phylogeny also will be important in further analysis and assessments of penguin evolutionary history and Mhc molecular dating, but possibly not as a reliable marker for the phylogenetic classification of penguins.