Specimens and species identification

Fifteen species belonging to 4 genera in the family Acroporidae were analyzed. Two Anacropora species and Acropora (Isopora) parifera were collected around Ishigaki Island (24N, 124E), Okinawa, Japan, and the other 12 species were collected around Akajima Island (30N, 123E), Okinawa, Japan. Species identification was based on Veron and Wallace (1984) and Veron (1986). The two Anacropora species could not be identified, and were designated Anacropora sp.1 and sp.2. Among the mass spawning species in the genus Acropora, the following four species were studied: A. digitifera, A. florida, A. gemmifera and A. nasuta. Small pieces of colonies were collected, dipped in guanidine solution (Sargent et al., 1986), and stored at room temperature. As a source of mitochondrial DNA for analysis of the entire mitochondrial genome, eggs of A. nasuta were obtained at Akajima in the manner described previously (Hatta et al., 1999), and dissolved in lysis buffer (100mM Tris-Cl pH8, 10mM EDTA, 1%SDS) containing 100μg/ml proteinase K.

DNA extraction, PCR amplification, and sequencing

Total DNA was extracted from the tissue stored solution in guanidine solution and lysis buffer by conventional phenol/chloroform extraction and ethanol precipitation.

The entire mitochondrial (mt) genome in DNA extracted from A. nasuta eggs, except for a part of the 16S ribosomal DNA (rDNA), was amplified by PCR using the primer 5′-GACAGTGAGACCCTCGTGAC-ACCATTCATA-3′ and 5′-GACTGCCAGGGGGAAACCTAGAGCAGACAC-3′, which were designed from partial sequences of the 16S rDNA of the genus Acropora (Romano and Palumbi 1996). Ten PCR cycles were performed at 94°C for 15 sec, 60°C for 30 sec, 68°C for 4 min, followed by 20 cycles at 94°C for 15 sec, 60°C for 30 sec, 68°C for 4 min with extension of 20 sec for each cycle, using Expand HiFi polymerase mix (Roche). The amplified fragment was separated by agarose gel electrophoresis, and purified using a GeneClean kit (BIO 101). The recovered fragment was digested with EcoRV, HaeIII or Sau3AI, cloned in pBluescript, and sequenced. The DNA sequences were compared with the sea anemone mtDNA sequences (Beagley et al., 1998).

The Cyt b gene was amplified by PCR using the primers described by van Oppen et al. (1999b). The ATP6 gene was amplified using primers 5′-ATGAGCGGTGCTTATTTTGATCAAT-3′ and 5′-CTAATGTAATACAATTGTATCCGCC-3′, which were designed from the A. nasuta ATP6 gene sequence. The PCR conditions were 40 cycles of 94°C for 30 sec, 60°C for 45 sec, 72°C for 45 sec, using Expand HiFi polymerase mix (Roche). The amplified products were separated by agarose gel electrophoresis, purified using GeneClean (BIO 101), and subjected to direct sequencing.

The DNA sequences of both strands were determined using a Dye Terminator Cycle Sequencing kit (ABI). The following additional primers designed from the A. nasuta sequences were used to sequence the Cyt b and ATP6 genes: for ATP6 5′-GCCAAGTGGCGC-TCCCTTG-3′ and 5′-CAAGGGAGCGCCACTTGGC-3′ for ATP6, and for Cyt b 5′-CATGCTAATGGGGCTTCT-3′, 5′-TCTGGGCTATGT-GCTACC-3′, 5′-GACGATGTGGTATTTCAT-3′, 5′-TTGGGCGATCC-AGAAAAT-3′, 5′-AGAAGCCCCATTAGCATG-3′, 5′-CTCAGGCTGAATGTGCAC-3′ and 5′-AGAAGAACAAAATTCAC-3′. The DNA sequences are available in DDBJ under accession numbers AB033171-033200.

Molecular Phylogenetic analysis

Sequences composed of 1061 bases for Cyt b and 649 bases for ATP6 were aligned manually, and used for phylogenetic analysis using the software programs ODEN (Ina 1994) and CLUSTAL W available in DDBJ, and PHYLIP package (Felsenstein 1990). The genetic distances were calculated using a Kimura's two-parameter model (Kimura 1980). Phylogenetic trees were constructed using the neighbor-joining method (NJ; Saitou and Nei 1987) to infer genetic relationships, and the unweighted pair-group method (UPGMA; Sneath and Sokal 1973) to estimate divergence times. The bootstrap analysis was replicated 1,000 times. The view of Wells (1956) was used to give divergence times based on the fossil record.

Isolation and structure analysis of mitochondrial genes

Fig. 1 shows the gene organization of Acropora nasuta mitochondrial genome. The complete DNA sequences were determined for 8 genes and 1 non-coding region, and partial sequences for 10 genes. The full length of the mtDNA was about 16 kilobase pairs (kbp), and the distance between the identified genes were measured using the lengths of the PCR fragments between the genes. The gene organization in A. nasuta is identical to that in A. tenuis, although that was determined for only a half (van Oppen et al., 1999a), and very different from that in the sea anemone (Beagley et al., 1998). No histories of inversion or translocation between the two taxa could be identified.

Fig. 1

Gene map of the Acropora nasuta mtDNA molecule. Asterisks represent genes which were determined for the full sequences. The arrow indicates the direction of transcription.

Nucleotide substitution and genetic distance

Table 1 shows the pairwise genetic distances calculated from the Cyt b and ATP6 sequences between all 15 species. The rate of substitution within the genus Montipora was similar to the rate among the mass spawning species in the subgenus Acropora. There were no significant differences observed in the GC content and codon usage among all species. The ratio of nonsynonymous and synonymous substitution rates were 0.08 – 0.12. Transition was more than 70% of the nucleotide substitution.

Table 1

Genetic distances between species in the family Acroporidae. Values above the diagonal show distances of Cyt b, and below ATP6.

The substitution rates at the 3rd codon position of Cyt b and ATP6 for 8 species of Acropora were compared with the rate of an intron or the 3rd codon position of coding regions of the mini-collagen nuclear gene (Hatta et al., 1999) (Table 2). Among the four mass spawning species, the genetic distances of the intron varied from 0.98 to 6.09%, which presumably correspond to the neutral substitution rate in the nuclear genome. The substitution rates at the 3rd codon position of the mini-collagen gene was 0 – 6.47%. On the other hand, the genetic distances between the mt genes was small: 0% in Cyt b and 0 –0.46% in ATP6. Even between the subgenera Isopora and Acropora, the values were approximately 3 times smaller in Cyt b and 10 times smaller in ATP6 than in the nuclear intron and coding regions. There was no obvious bias of the GC content, 37% in average at the 3rd codon position of the mt genes, whereas 42% at the 3rd codon position of the mini-collagen gene and 34% in the intron.

Table 2

Comparison of genetic distances of mitochondrial genes and a nuclear gene. Only the 3rd codon position was used for Cyt b and ATP6. Acropora donei and A. tenuis spawn a few hours earlier than mass spawning species. A. florida, A. gemmifera, A. digitifera and A. nasuta take part in mass spawning events.

Phylogenetic analysis

Phylogenetic trees are shown in Fig. 2. Astreopora was used as an outgroup according to the fossil record that Astreopora occurred first in this family (Wells 1956). Branching of Astreopora became the most outside when sea anemone (Beagley et al., 1998) was used as an outgroup in Cyt b analysis (data not shown). All of the trees suggested that this family is divided in to three major lineages, although the branch length of the Astreopora lineage in the ATP6 tree became shorter than in the Cyt b tree (Fig. 2A, B). The topology of the tree produced when the Cyt b and ATP6 sequences were combined together and subjected to analysis (Fig. 2C), was the same as that of the Cyt b tree. The inconsistency of the ATP6 phylogeny with the others may be due to short DNA sequences.

Fig. 2

Molecular phylogenetic trees of (A) Cyt b, (B) ATP6 and (C) Cyt b+ATP6, constructed by the NJ method. Astreopora was used as an outgroup. Scale bars represent 1% genetic distance. Percent values of 1,000 bootstrap replicates are shown for major branches.

In all of the trees, the two species from the genus Anacropora were included in the clade with the Montipora species. Low bootstrap values for the Anacropora branches suggest that the Anacropora species are not distinguished from Montipora genetically. The relationships within this Montipora group could not be resolved because of short genetic distances.

The subgenus Isopora formed a side branch in the Acropora lineage. The genetic distance between the subgenera Isopora and Acropora was significant (99 – 100% bootstrap values), while the relationships within the subgenus Acropora appeared to be very close. In Fig. 2A and 2C, A. donei and A. tenuis, which spawn 2–3 hours earlier than mass spawning (Hayashibara et al., 1993; Hatta et al., 1999), formed a cluster separate from the other mass spawning acroporids. This pattern is consistent with the tree inferred from the minicollagen nuclear gene (Hatta et al., 1999).

In addition to the NJ method, these results were reproduced by the maximum-likelihood method too, and we obtained the same genetic phylogeny (data not shown).

Estimation of divergence time

In Fig. 3, genetic distances were plotted on divergence times assigned by fossil records as follows. Fossils show that Montipora and Acropora first appeared during the Eocene (37– 54 million years ago; MYA) (Wells 1956), implying that the two genera diverged 54 MYA. Astreopora appeared in the late Cretaceous (about 70 MYA, Wells 1956). Mass spawning acroporids began diversification 2 MYA (Veron 1995). The genetic distances showed a linear correlation to the divergence times.

Fig. 3

The relationship between the nucleotide substitution rates and divergence times. The nucleotide substitution rate of the combined sequences of Cyt b and ATP6 was calculated for each pair of species, and an average value between two taxa was plotted on the divergence time estimated by the fossil record. Standard deviation was less than 0.1. Anacropora and Isopora were excluded from this comparison since fossil has not been found.

Figure 4 shows the molecular phylogram constructed by the UPGMA method. The divergence times were adjusted using the correlation of the genetic distances and the fossil records mentioned above (Fig. 4, arrowhead 1, 6, 7). According to the phylogram, Anacropora appeared 5 MYA (Fig. 4, arrowhead 2) and Isopora 15 MYA (Fig. 4, arrowhead 5), although no fossils of either taxa have been found. The diversification of Montipora was deduced to have occurred 6.6 MYA (Fig. 4, arrowhead 4), and the lineage of the mass spawning acroporids and early-hour spawners also diverged at about the same time (Fig. 4, arrowhead 3).

Fig. 4

Molecular phylogram constructed by the UPGMA method using the combined sequences of Cyt b and ATP6. Geological intervals are shown in the upper bar. Arrowheads with numbers indicate the divergence of lineage and corresponding time points. Shaded bar represent occurence of novel reproductive behavior, A: buoyant eggs, B: incorporation of algae into eggs, C: brooding of larvae, D: spawning at early hr.

Features of mitochondrial genes in Acroporidae

In the class Anthozoa, the complete mtDNA sequences have been determined for the sea anemone (Beagley et al., 1998) and the octocoral (Pont-Kingdon et al., 1998). The molecular phylogeny indicates that the Acroporidae corals are more closely related to the sea anemone than to the octocoral (data not shown), which is consistent with their classification into the subclasses Hexaradiata and Octoradiata. However, the gene organization of Acropora nasuta and the sea anemone is very different. Recombination with resulting rearrangement of the genes might have occurred frequently in mitochondria along the Anthozoa lineage.

This study is the first to accurately compare the mitochondrial and nuclear DNA substitution rates in Cnidaria. When the 3rd codon position of the mt genes is examined, the substitution rate was 3–10 times lower than the rate in an intron or the 3rd codon position of coding regions of a nuclear single copy gene (Table 2). No obvious biases were found in the GC content and substitution patterns between the mt genes and the nuclear locus. In a comparison with the internal transcribed spacer (ITS) of the nuclear ribosomal DNA, van Oppen et al., (1999b) concluded that the substitution rate of Cyt b is atypically low in Acropora. However, analysis using the ITS includes a serious bias, since the ITS is hypervariable corresponding to a presumed recombination hotspot in Acropora (Fukami et al., in preparation). The low substitution rate of Cyt b is not atypical even though it is significantly lower than nuclear genes.

Phylogenetic relationships of genera in the Acroporidae

An important finding of this study is the close relationship between the genera Montipora and Anacropora (Fig. 2). Ridley (1884) postulated that Anacropora recently speciated from Montipora, based on micromorphology characteristics; on the other hand, Veron (1995) proposed that Anacropora gave rise to Acropora, based on macromorphology. Our molecular phylogeny suggests that Anacropora diverged from Montipora very recently. Anacropora species are found in non-reef environments (Veron 1995), and inhabit quieter environments than most Montipora and Acropora in Okinawa (unpublished observation by the authors). Anacropora may have speciated by occupying niches in different environments or at different depths.

A second important finding is the significant distance between the subgenera Isopora and Acropora, which are related nevertheless (Fig. 2). Isopora arose from the Acropora lineage, however, it is clearly separated from the cluster of Acropora species in the phylogenetic trees. Isopora is morphologically distinguished from Acropora (Randall 1981), and broods planula larvae (Kojis 1986), while the other genera in the family Acroporidae spawn gametes. Combining our molecular phylogeny with the morphological and reproductive differences, we propose that Acropora and Isopora be classified as separate genera.

This study is the first report describing genetic relationships within the coral family Acroporidae, and helps to clarify the phylogenetic relationships in this family, which have been confused.

Divergence of reproductive behavior

Alteration of the reproductive feature are indicated in the phylogram in Fig. 4. Since the majority of species in the three major lineage, Astreopora, Montipora and Acropora, participate in mass spawning in the Akajima region (Hayashibara et al., 1993), mass spawning is thought to be the ancestral reproductive behavior in the family Acroporidae. The sinking eggs of Astreopora are common characteristics seen in some of the Pocillopora species in the family Pocilloporidae (Kinzie 1993), implying an ancestral feature. Buoyant eggs can be thought as a derivative feature which arose in the lineage before branching of Acropora and Montipora (Fig. 4, bar A). Buoyancy of eggs might contribute for higher chance of fertilization and dispersal to lead present prosperity of the two genera, Acropora and Montipora. The transmission of symbiotic algae into unfertilized eggs is another novelty that arose in Montipora (Fig. 4, bar B). Anacropora seems to be derived from Montipora, and it will be interesting to determine whether Anacropora eggs contain symbiotic algae. In the family Acroporidae, brooding of larvae only occurred in Isopora (Fig. 4, bar C), which evolved from the Acropora lineage. Spawning at early hours arose recently in Acropora (Fig. 4, bar D). We believe that the different reproductive systems evolved independently in different lineage in the family Acroporidae.

Evolutionary history of the Acroporidae

The molecular phylogeny of the mt genes corresponds well to the divergence times deduced from the fossil record (Fig. 3, 4). The results suggest that the present species in the genus Acropora are monophyletic and the genus began to diverge after the middle Miocene (Fig. 4), and the subgenus Acropora radiated just after the Pliocene. Speciation and diversification must have proceeded very rapidly in the subgenus Acropora, since it contains more than 150 extant species with quite varied morphology. Many other acroporids, which are found as fossils in deposits from the Eocene to Miocene, have become extinct. The genus Montipora has a similar evolutionary history. Although fossil Montipora are recorded from the Eocene, the existing Montipora species are monophyletic (Fig. 4). Their ancestor likely survived the catastrophic extinction at the Miocene-Pliocene boundary and diverged to give rise to the present species.

Climate changes, such as glacial cycles, have led repeated rounds of mass extinction and expansion from small populations of a few surviving species in a variety of organisms. Reef-building corals are no exception. The existing reef corals appeared to have spread throughout the oceans very rapidly in a short period of geological time. Our molecular phylogeny matches the geological history well, and supports this evolutionary history for corals in the Acroporidae.