Color Vision of Ancestral Organisms of Higher Primates

Masatoshi Nei,* Jianzhi Zhang,* and Shozo YokoyamaS *Institute of Molecular Evolutionary Genetics and Department of Biology, The Pennsylvania State University; and $Department of Biology, Syracuse University

The color vision of mammals is controlled by photosensitive proteins called opsins. Most mammals have dichromatic color vision, but hominoids and Old World (OW) monkeys enjoy trichromatic vision, having the blue-, green-, and red-sensitive opsin genes. Most New World (NW) monkeys are either dichromatic or trichromatic, depending on the sex and genotype. Trichromacy in higher primates is believed to have evolved to facilitate the detection of yellow and red fruits against dappled foliage, but the process of evolutionary change from dichromacy to trichro- macy is not well understood. Using the parsimony and the newly developed Bayesian methods, we inferred the amino acid sequences of opsins of ancestral organisms of higher primates. The results suggest that the ancestors of OW and NW monkeys lacked the green gene and that the green gene later evolved from the red gene. The fact that the red/green opsin gene has survived the long nocturnal stage of mammalian evolution and that it is under strong purifying selection in organisms that live in dark environments suggests that this gene has another important function in addition to color vision, probably the control of circadian rhythms.

Introduction

The color vision of mammals is controlled by pho- topigments, each of which consists of a protein called opsin and a chromophore 1 1-cis retinal. The diversity of color vision in mammals is due to variation in the num- ber and sequence of opsin genes. Most mammals have a short wavelength (blue)-sensitive opsin gene and a mid- dle wavelength (green)- to long wavelength (red)-sensi- tive opsin gene, so that they have dichromatic color vi- sion. In higher primates, however, trichromatic color vi- sion evolved in two different ways. Hominoids and Old World (OW) monkeys have three different color vision genes: one autosomal gene encoding the blue-sensitive opsin and two X-linked genes encoding the green- and red-sensitive opsins (Nathans, Thomas, and Hogness 1986). Most New World (NW) monkeys have only one X-linked locus in addition to the autosomal blue gene locus, but the X-linked locus is polymorphic and contains three different alleles (Mollon, Bowmaker, and Jacobs 1984; Neitz, Neitz, and Jacobs 1991). Therefore, all males are dichromatic, whereas females are either di- chromatic or trichromatic depending on the genotype. Ja- cobs et al. (1996) recently showed that one genus of NW monkeys (Alouattu; howler monkey) has two X-linked opsin genes, but this is probably due to a recent gene duplication confined to this genus and does not change the general picture of trichromacy due to polymorphism in NW monkeys. The blue and red/green genes seem to have diverged more than 500 MYA, but the green and red genes apparently diverged only after OW monkeys separated from NW monkeys (Nathans, Thomas, and Hogness 1986; Yokoyama and Yokoyama 1989). Tii- chromacy in higher primates is believed to have evolved to facilitate the detection of yellow and red fruits against dappled foliage (Mollon 1991). However, the process of

Key words: color vision, trichromacy, opsin, higher primates, an- cestral sequences, gene sharing.

Address for correspondence and reprints: Masatoshi Nei, Institute of Molecular Evolutionary Genetics, The Pennsylvania State Univer- sity, 328 Mueller Laboratory, University Park, Pennsylvania 16802. E-mail: nxm2Qpsu.edu.

Mol. Biol. Evol. 14(6):611-618. 1997 0 1997 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038

evolutionary change from dichromacy to trichromacy is not well understood (Bowmaker 1991; Mollon 1991; Ja- cobs 1993). This is partly because the color vision of ancestral species of higher primates is unknown. Al- though a number of authors have speculated on the origin of red and green opsins (Yokoyama and Yokoyama 1990; Jacobs 1993; Winderickx et al. 1993), no serious study has been done on this subject.

The purpose of this paper is to infer the color vision of ancestral higher primates and study the evolution of color vision. Since the color vision in mammals is en- tirely determined by the amino acid sequences of opsins, the problem becomes how to infer the opsin sequences of ancestral higher primates. Ancestral protein sequenc- es can be statistically inferred from present-day sequenc- es by various methods. Computer simulation (Zhang and Nei 1997) has shown that when the level of sequence divergence is low, the statistical inference is quite ac- curate. In general, once the ancestral sequence of a pro- tein is inferred, its function is studied by producing the ancestral protein artificially (e.g., Jermann et al. 1995; Chandrasekharan et al. 1996). In the case of the red and green opsins of higher primates, however, the photosen- sitivity of an opsin is determined primarily by three crit- ical amino acids, as will be mentioned later. Therefore, we can infer the color vision of ancestral primate or- ganisms without producing ancestral proteins.

In the course of this study, we reached the conclu- sion that some opsin genes probably have another im- portant function in addition to color vision. We will first discuss the color vision of ancestral higher primates and then provide evidence supporting the view that the gre- en/red genes have another function.

Material and Methods Background Information

The spectral sensitivity of opsins is measured by the maximum absorption wavelength (A,,,), which is 561 nm and 530 nm for the human red and green opsins, respectively. It is now well established that this differ- ence in A,,, between the red and green opsins is pri- marily controlled by three critical amino acids at sites

611

612 Nei et al.

Table 1 Red- and Green-Sensitive Opsin Sequences of Primates and Some Outgroup Species Used in this Study

Redb Greenb REFERENCES

Hominoids Human (Homo sapiens) ...................... 561 530 Nathans, Thomas, and Hogness (1986) Common chimpanzee (Pun troglodytes). ........ 560-565 530-535 Dulai et al. (1994), Deeb et al. (1994) Pygmy chimpanzee (Pun paniscus) ............ 560-565 530-535 Deeb et al. (1994) Gorilla (Gorilla gorilla). ..................... 560-565 530-535 Dulai et al. (1994), Deeb et al. (1994) Orangutan (Pongo pygmueus) ................. 560-565 530-535 Deeb et al. (1994)

Old World monkeys Macaque (Macucu fuscicularis). ............... 563 533 Dulai et al. (1994) Diana monkey (Cercopithecus diunu) .......... 563 533 Dulai et al. (1994) Talapoin monkey (Miopithecus fulapoin) ........ 563 533 Dulai et al. (1994)

New World monkeys Marmoset (Cullithrix jucchus jacchus). ......... 563, 556 543 Hunt et al. (1993) Squirrel monkey (Suimiri sciureus) ............ 561 547, 532 Neitz, Neitz, and Jacobs (1991) Tamarin (Suguinus fuscicollis). ................ 562, 556 541 Neitz, Neitz, and Jacobs (1991)

Outgroup species used Goat (Cupra hircus). ........................ 555 Unpublished data Chicken (Gallus gallus). ..................... 571 Okano et al. ( 1992) American chameleon (Anolis carolinensis) ...... 561d Kawamura and Yokoyama (1993) Gecko (Gecko gecko). ....................... 521 Kojima et al. (1992)

NOTE.-sequences less than 100 codons long are not included. a These values are approximate. The A,,, value for the same organism varies with investigator to some extent, but the variation is usually within 5 nm. b Color designation is somewhat arbitrary. c Only amino acid sequences are available. d This opsin is normally linked with 1 I-cis-3,4_dehydroretinal. To compare its A,,, with those of primate opsins, the A,,, is calculated under the assumption

that it is reconstituted with 1 I-& retinal (Yokoyama 1995).

180, 277, and 285 (Neitz, Neitz, and Jocobs 199 1; Yo- for the three OW monkeys was in the range of 0.8%- koyama and Yokoyama 1990), although four other sites (sites 116, 230, 233, and 309) have some minor effects

1.3% for the red gene and l . l %-2.7% for the green gene. Therefore, the macaque sequences were chosen as

(Asenjo, Rim, and Oprian 1994). By contrast, the blue the representatives of the OW monkey sequences in the opsin exists in almost all mammals, and its A,,, is primary analysis. The phylogenetic relationships among around 430 nm. In this study, we will therefore focus the NW monkey opsins (three species each with three on the evolutionary changes of red and green opsins, alleles) are unclear (e.g., Shyue et al. 1995), so we an- inferring the A,,, of ancestral opsins based on the three alyzed them species by species rather than simulta- critical sites. neously.

Sequences Used Phylogenetic Analysis

All available sequences of primate red and green opsin genes are listed in table 1, together with the se- quences of other higher vertebrates which were used as outgroups. The human red and green opsins have 364 amino acids, which are encoded by six exons interrupted by five introns. Unfortunately, many available sequences are not of full length. For example, the OW monkey sequences are only 126 amino acids long, comprising parts of exons 3, 4, and 5. However, six of the seven sites that determine the difference in A,,, between the human red and green opsins are located in this 126- codon region, the remaining one (site 116) being located in exon 2. Since amino acid site 116 has only a minor effect on A,,,, it is possible to infer the color vision by examining the 126 sites.

Since the divergences among the primate opsin genes are fairly low, we reconstructed the phylogeny of opsin genes by using the neighbor-joining method (Sai- tou and Nei 1987) with Jukes-Cantor distances for nu- cleotide sequences. The amino acid sequences were not very informative in this case. The tree was tested by the bootstrap method (Felsenstein 1985) with 1,000 repli- cations and the interior branch test (Rzhetsky and Nei 1992). The computer software package MEGA (Kumar, Tamura, and Nei 1993) was used in these analyses.

Ancestral Sequence Inference

The orthologous opsin gene sequences of homi- noids are closely related to one another. The pairwise Jukes-Cantor distance for hominoid species was in the range of 0.8%-2.4% for the red gene and 1.3%-3 5% for the green gene. Therefore, we decided to use the human sequences as representatives of hominoid se- quences. Similarly, the pairwise Jukes-Cantor distance

The ancestral amino acid sequences of the opsins were first inferred by Zhang and Nei’s (1997) distance- based Bayesian method, which is a slight modification of Yang, Kumar, and Nei’s (1995) likelihood-based Bay- esian method. Computer simulation (Zhang and Nei 1997) has shown that this method gives more accurate inference than the commonly used parsimony method (e.g., Maddison and Maddison 1992) and is as good as the more time-consuming likelihood-based method. However, the parsimony method and the likelihood-

Color Vision of Ancestral Primates 613

based Bayesian method were also used to check the re- sults from the distance-based Bayesian method. The amino acid substitution model (JTT_-f model) we used for the Bayesian method was a modified version of Jones, Taylor, and Thornton’s (1992) empirical substi- tution model (JTT model). The JTT-f model is designed so as to make the equilibrium frequencies of amino ac- ids equal to the observed frequencies of the sequences under investigation (Cao et al. 1994). The Dayhoff-f model (based on Dayhoff, Schwartz, and Orcutt 1978) and the equal-input model were also used to see the effect of different substitution models. The JTT-f model and the Dayhoff-f model are based on empirical amino acid substitution data for many different proteins and therefore would represent an average substitution pattern for many proteins, whereas the equal-input model as- sumes that the probability of change of amino acid i to amino acid j is proportional to the frequency of amino acid j in the sequences.

Results and Discussion Ancestral Opsin Sequences of Higher Primates

Inference of ancestral sequences requires a reliable phylogenetic tree (topology) of the present-day sequenc- es. Figure la shows the phylogenetic tree of the red and green opsin genes for some representative higher pri- mates and outgroup species. This tree was constructed by using the nucleotide sequences for the 126-codon region of exons 3, 4, and 5 given in Figure 2. The tree topology seems to be reliable except for the branch be- tween nodes b and c, which is not statistically well sup- ported. Nevertheless, this tree topology is consistent with the current view of evolution of color vision genes in higher primates (Bowmaker 1991; Shyue et al. 1995), and the nonprimate portion of the tree is well supported by other biological information.

We inferred the amino acid sequences of the opsins at all interior nodes using this tree. As mentioned earlier, only 126 amino acids were inferred for each ancestral opsin, and the accuracy of inference of an amino acid was evaluated by the Bayesian posterior probability (Yang, Kumar, and Nei 1995; Zhang and Nei 1997). The results obtained are presented in Figure 2. In general, the accuracy of inference is quite high; only a few in- ferred amino acids of ancestral primates have posterior probabilities that are lower than 90%. Figure la shows the inferred ancestral amino acids at the three critical sites (180, 277, and 285). Node d has S (serine),Y (ty- rosine), and T (threonine) at the three positions and node e has A (alanine), F (phenylalanine), and A, as expected. Node g also has AYT with a probability of nearly 100%. For all the remaining nodes, the amino acids at sites 277 and 285 are Y and T, respectively, with a probability of 99% or higher. The amino acid at site 180 is S with a probability of about 60% and A with a probability of about 40% for nodes a, b, c, and$ but the probabilities of S and A are 92% and 8%, respectively, for nodes h and i. Therefore, our inference of amino acid at site 180 is ambiguous, but this ambiguity is understandable be- cause human red opsins have either S or A at this po-

sition (see below). We also used the likelihood-based Bayesian method to infer the ancestral opsin sequences and obtained the same results. Essentially the same re- sults were obtained for both the distance- and likeli- hood-based methods when three different models of amino acid substitution (the equal-input, Dayhoff-f, and JTT-f models) were used. The parsimony method with an unrooted tree (Maddison and Maddison 1992; Yang, Kumar, and Nei 1995) gave the same amino acids as those obtained by the Bayesian method at sites 277 and 285, but, at site 180, parsimony gave two possibilities (S or A) for nodes a, b, c, j h, and i and S for node d and A for nodes e and g. Unlike the Bayesian methods, parsimony does not use information about branch lengths and therefore does not give the probability of the inferred amino acids, but the results obtained are consistent with those from the other two methods.

In higher primates, the A,,, value varies from 530 to 547 nm for green opsins and from 556 to 565 nm for red opsins. We call the marmoset allele P556 a red gene, because human populations are polymorphic for alleles “SYT” and “AYT” at the red gene locus, and the in- dividuals with these alleles both have normal color vi- sion (Nathans, Thomas, and Hogness 1986; Winderickx et al. 1992). In primate red opsins, the amino acid changes S+A, Y+F, and T+A at the three critical sites are known to reduce A,,, by approximately 5, 9, and 15 nm, respectively, and these effects are more or less ad- ditive (Neitz, Neitz, and Jacobs 1991). If we use this rule, we can infer that the ancestral organism of homi- noids and OW and NW monkeys had a red opsin with x max = 561 or 556 nm without any green opsin. This inference is further supported by the identity of amino acids between the sequence at node b and the goat red gene sequence at the four sites that slightly affect the x max values of primate red and green opsins (fig. 2). It is interesting to note that there are color-vision-deficient people who do not have any green genes, and they are called deuteranopes. Examining the DNA extracted from the preserved eye tissue of the famous chemist John Dalton, Hunt et al. (1995) showed that he was a deu- teranope with amino acids A, Y, and T at the three crit- ical sites of his red opsin. For deuteranopes, the color spectra that look yellow, orange, and red to normal per- sons are all of the same hue. Our primate ancestors ap- parently had a deuteranopia-like color vision. It would be interesting to reconstruct the predicted proteins by site-directed mutagenesis and examine their detailed function, as was done with artiodactyl ribonucleases (Jermann et al. 1995).

As mentioned earlier, there are four more sites (116, 230, 233, and 309) that have a minor effect (l-4 nm) on the difference in X,,, between the human red and green opsins. In vertebrates other than humans, however, the amino acids at these sites do not necessar- ily correspond to those of the human red and green op- sins (fig. 2). Therefore, the ancestral amino acids at these four sites do not seem to have much biological meaning. In our analysis mentioned above, amino acid site 116 in exon 2 was not included because of the lack of data for OW monkeys. However, we were able to

614 Nei et al.

180 277 285

a SYT 561 d Human P561 (red) S Y T AYT 556 fS\ _ 85/g8 Macaque P563 (red) S Y T

b

c

Human P561 (red)

Macaque P563 (red)

Human P530 (green)

Macaque P533 (green)

Marmoset P563 (red)

Marmoset P556 (red)

Marmoset 54.3 (green)

180277285

Human P561 (red) S Y T

Macaque P563 (red) S Y T

Human P530 (green) A F A

Macaque P533 (green) A F A

Goat P555 (red) A Y T

SYT , (4

Chicken P571 (red) SYT h SYT

Chameleon P561 (red) S y T

Gecko P521 (green) A F A

FIG. I.-Phylogenetic trees used and inferred amino acids of ancestral organisms (tree nodes) of higher primates at the three critical amino acid sites of red/green opsins. The number after P in the gene symbol refers to X,,,. Q, Ancestral amino acids inferred under the hypothesis of independent origins of tricromacy in OW and NW monkeys. The numbers below each interior branch are the bootstrap probability value followed by the confidence probability value (in percentage). A = alanine, S = serine, F = phenylalanine, Y = tyrosine, and T = threonine. b, The phylogenetic relationships of the primate red and green opsins under the single-origin hypothesis of trichromacy in higher primates. c, Ancestral amino acids inferred under the single-origin hypothesis.

infer some of the ancestral amino acids at this position human red opsin (fig. 2). However, they are the same as by excluding OW monkeys. (One hundred ninety-eight that (Y) of the goat red opsin. These results suggest that amino acid sites were used in this analysis.) The results the amino acid at this site is not important for deter- obtained show that all ancestral amino acids we inferred mining A,,,. are Y at site 116 and are the same as the amino acid Although only one species from each group of (Y) of the human green opsin rather than that (S) of the hominoids, OW monkeys and NW monkeys, was used

Color Vision of Ancestral Primates 615

116 147 Exon 3 185 204 Exon 4 240

Human-red Macaque-red Human-green Macaque-green Marmoset (563) Marmoset (556) Marmoset (543) Goat-red Chicken-red Chameleon-red Gecko-green Node-a Node-b Node-c Node-d Node-e Node-f Node-g Node-h Node-i

* ISWERWLWCKPFGNVRFDAKLAIVGIAFSWIWSAVWTA GPDVFSGSSYPGVQSYMIVLMVTCCIIPLAIIMLCYL

H Y Y Y

15 Y Y Y ?

Y Y

264

. . . . . . . . . . . . . . . . . . . . . . . . . . V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A . . . . .

................................. A .....

...... M ................ TA........A .....

...... F ........ IK..G...VA..L...L..CA ...

...... V ......... K ...... VA..V...V .......

...... F ........ IK..S....I..V...V.AWG.S.

...................... A .._........A .....

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .._t.....

................................. A .....

...... E........&E .... ..YA..V...y .......

.............. . ...... yA..V...y .......

Exon 5 313 * * +

.......................... T..S..V ....

.......................... T..T..V ....

..................... I ... FL..G..V ....

..................... I ... FL..S..V ....

..................... I ... FL..S..V ....

..................... I ... F...SV .I.. ..

......... D.R.....V.......FF.....I ....

........ DD...L.......I...F....V.L ....

........ VEL.C..F.LT..I...FL..F..IV ...

..................... I ... F...S ..I.. ..

..................... I ... F...S ..V.. ..

.......................... T..S..V ....

..................... I ... FL..S..V ....

..................... I ... FL..S..V ....

......... D ........... I ... F.. .... I.. ..

......... D ........... I ... F.. .... I.. ..

Human-red KEVTRMVWMIFAYCVCWGPYTFFACFAAANPGYAFHPLMAALPAYFAKS Macaque-red ..... ..M .......................................... Human-green ........ ..VL.F.F.....A............P..........F .... Macaque-green ........ ..FL.F.F.....A ............................ Marmoset(563) ......... ..V ...................................... Marmoset(556) ......... ..V ...................................... Marmoset(543) ......... ..V.........A ............................ Goat-red ....... M ....... L .............. H ........ V .......... Chicken-red . ..S.......V...F.......................A .......... Chameleon-red . ..S.......I...F......V................A .......... Gecko-green R..S.......V.F.I.....AS.VS.............A .......... Node-a ........... .L.......................Y .......... Node-b ........... :.* ...................................... Node-c ......... .._L ...................................... Node-d .................................................. Node-e .......... VL.F.F.....A ............................ Node-f ........... Y ...................................... Node-g ........... Node-h ... S .......

;...;.......................~ ............. .................................

Node-i ... S ....... y F.......................A ... ..........

FIG. 2.-Partial amino acid sequences of the red and green opsins of the present-day organisms and ancestral organisms inferred according to the tree in figure la. Amino acids are denoted by single-letter codes. The numbering system of amino acid sites used is the same as that of the human red opsin. Dots indicate the identity of the amino acids with those of the human red opsin. The three critical amino acid sites (180, 277, and 285) are indicated by asterisks, whereas the four less important sites (116, 230, 233, and 309) are indicated by the symbol “+.” The other sites do not seem to affect A,,,,, in primates. The ancestral amino acids that have a probability of 90% or less are underlined. The ancestral amino acids at site 116 were inferred by using 198 amino acids (see text). The symbol “?” indicates that the amino acid is unknown.

in the above study, further analyses using other species from these groups did not change our results apprecia- bly. The only differences observed were the posterior probabilities of S and A at site 180.

The tree in figure la implies that trichromatic color vision evolved independently in NW and OW monkeys. However, Hunt et al. (1993) and Dulai et al. (1994) sug- gested that trichromacy might have evolved first as a polymorphic form before the divergence of NW and OW monkeys and that the duplicate red and green genes in OW monkeys evolved later as a result of unequal crossing over. In this hypothesis, a high degree of sim- ilarity of red and green alleles (genes) within species is attributed to sequence homogenization due to gene con- version or recombination. (Gene conversion apparently occurs in the introns of opsin genes, but the exon regions do not appear to be subject to frequent gene conversion; Shyue et al. 1994; Zhou and Li 1996). Un- der this hypothesis, the relationship of the primate opsin sequences would be something like that given in figure

lb. However, the branch lengths from nodes b to X, b to z, and x to y in figure lb are unknown. We therefore assumed as a first approximation that these branch lengths are all 0 but that the total length from node b to each present-day sequence remains the same as in figure la (see figure lc). Although this assumption appears to be extreme, the tree of figure lc is actually intermediate between the trees of figure la and b. At any rate, using this tree we inferred ancestral opsin sequences at all in- terior nodes. However, the results obtained were essen- tially the same as those for figure la, although at site 180 amino acid A now had a probability of 96% for nodes a and b. The parsimony method gave the same amino acids as those obtained by the Bayesian methods at sites 277 and 285, whereas at site 180 it gave two possibilities (A or S) for nodes h and i, S for node d, and A for nodes a, b, and e. Therefore, the two evolu- tionary hypotheses give different probabilities of A at this position, but this does not alter our conclusion about the color vision of primate ancestors. Actually, we can

616 Nei et al.

even use the tree in figure lb to infer the ancestral se- quences. When we used this tree, many branch length estimates became 0, but the ancestral amino acids ob- tained at the three critical sites were the same as those in figure lc. These results show that our inference of ancestral amino acids is not affected seriously by gene conversion or recombination.

Figure 1 suggests that the ancestral species of the goat, chicken, American chameleon, and gecko also had opsins with sequence AYT or SYT. Since the A+& change at site 180 has a relatively minor effect on color vision, it is tempting to conclude that these ancestral species also had red genes. However, since the color vision of birds and reptiles may be affected by some other factors such as the type of retinal chromophore and colored oil droplets that are lodged in photoreceptor cells, this conclusion may not be justified. Even in some nonprimate mammalian species, A,,, is not solely de- termined by the three critical amino acid positions. For example, the red/green homologous opsin in the rat Rat- tus norvegicus has AYT at the three critical sites (un- published data), although the A,,, is 510 nm (Jacobs 1993). For these reasons, it seems premature to predict color vision of nonprimate ancestral organisms by the three critical positions only, but our inference of the amino acids of important sites 277 and 285 seems quite reliable.

Comparing the red and green genes from the hu- man and the Mexican char-acid fish Astyanax fasciatus, Yokoyama and Yokoyama (1990) suggested that the red genes in humans and fish evolved from the green gene independently through identical amino acid substitutions at a few sites. Jacobs (1993) also speculated that the prototype of higher primate red/green opsins had a A,,, of 543 nm. Similarly, the hypothesis that the red gene evolved from the green gene was supported by Winder- ickx et al. (1993). Our results suggest that the green gene evolved from the red gene unlike these specula- tions.

Opsins May Have Another Important Function in Addition to Color Vision

Until relatively recently it was believed that most mammals lost color vision due to the nocturnal living condition in the early stage of mammalian evolution and that only higher primates regained color vision (Walls 1942; Wistow 1993). Recent studies of opsin genes in- dicate that the blue and red/green genes of primates have coexisted in the genome for more than 500 Myr and, thus, the red/green gene did not become extinct during the long nocturnal stage of mammalian evolution.

The char-acid fish A. fasciatus is widely distributed in rivers of Mexico and part of Texas. However, there are populations of this species that live in dark environ- ments of Mexican caves. These populations are believed to have entered into the caves during the Pleistocene, possibly 0.5-l MYA (Barr 1968; Nei 1975). The fish in these caves are albinos and have no eyes or reduced eyes. Since there is no need for color vision for these

Table 2 Numbers of Synonymous (d,) and Nonsynonymous (&) Substitutions Per Site in Opsin Genes

Genes ds x 100

Characid fish (cave populations)

Red gene (6 alleles) . . . . . . . 1.68 (0.57) 0.11 (0.08) 15.3 Green gene 1 (4 alleles). . . . 0.68 (0.40) 0.00 (0.00) - Green gene 2 (10 alleles). . . 0.86 (0.45) 0.11 (0.07) 7.8

Characid fish (river populations)

Red gene (5 alleles) . . . . . . . 0.39 (0.28) 0.12 (0.09) 3.3 Green gene 1 (7 alleles). . . . 0.76 (0.37) 0.09 (0.07) 8.4 Green gene 2 (9 alleles). . . . 1.36 (0.41) 0.03 (0.03) 45.3

Goat and human red gene 51.06 (5.88) 5.46 (0.83) 9.4

NOTE.-The average d, and d,., and their standard errors (in parentheses) were computed by the MEGA package (Kumar, Tamura, and Nei 1993).

the DNA sequences of several alleles of these genes from cave and river fish. We therefore computed the numbers of synonymous (ds) and nonsynonymous (&) nucleotide substitutions per site to examine whether the opsin genes are still functional and whether they are under purifying selection or not. The ds and & values were computed by Nei and Gojobori’s (1986) method for each pair of polymorphic alleles from the cave and the river populations. The average values of d,‘s and dN’s are presented in table 2. In both river and cave populations, the ds/& ratio is much higher than 1, sug- gesting that the cave fish genes have been subject to purifying selection. Note also that the ds/& ratio is of the same order of magnitude as that (9.4) for the com- parison of the goat and human red genes. These results strongly suggest that the red and green genes in cave fish are still functional. Interestingly, the (partial) amino acid sequence of the red/green opsin of the blind mole rat Spalax ehrenbergi, which is believed to have lived a subterranean way of life for many millions of years (Quax-Jeuken et al. 1985), is also found to be highly conserved (Argamaso et al. 1995).

These observations indicate that a dark environ- ment, where no color vision is necessary, does not in- activate opsin genes. The fact that the red/green gene did not become extinct in dark environments suggests that this gene probably has another important function. One such possible function is the control of circadian rhythms of various cellular and physiological processes. Note that, although there is no need for an organism to know day and night in dark environments, circadian rhythms or some kind of biological clock is necessary for gene expression, cellular differentiation, embryonic development, etc. (Sassone-Corsi 1994). In the Ameri- can chameleon, the red/green gene is expressed not only in the retina but also in the pineal gland that controls circadian rhythms (Kawamura and Yokoyama 1997), and in mammals there is a strong indication that the red/green gene is involved in the control of circadian rhythms (Argamaso et al. 1995). Tosini and Menaker (1996) also showed that retinas of the golden hamster exhibit circadian rhythms of melatonin synthesis. It is

fish, one would expect that their opsin genes have be- possible that the red/green gene is another example of come nonfunctional. Yokoyama et al. (1995) published “gene sharing” (Piatigorsky and Wistow 1989).

Color Vision of Ancestral Primates 617

Acknowledgments

This work was supported by NIH grants (M.N. and S.Y.) and an NSF grant (M.N.).

LITERATURE CITED

ARGAMASO, S. M., A. C. FROEHLICH, M. A. MCCALL, E. NEVO, I. PROVENCIO, and R. G. FOSTER. 1995. Photopig- ments and circadian systems of vertebrates. Biophys. Chem. 56:3-l 1.

ASENJO, A. B., J. RIM, and D. D. OPRIAN. 1994. Molecular determinants of human red/green color discrimination. Neu- ron 12: 1131-l 138.

BARR, T. C., JR. 1968. Cave ecology and evolution of troglo- bites. Evol. Biol. 2:35-102.

BOWMAKER, J. K. 1991. Evolution of visual pigments and pho- toreceptors. Pp. 63-81 in J. R. CRONLY-DILLON and R. L. GREGORY, eds. Evolution of the eye and visual system. CRC Press, Boca Raton, Fla.

CAO, Y., J. ADACHI, A. JANKE, S. Pjia~o, and M. HASEGAWA. 1994. Phylogenetic relationships among eutherian orders es- timated from inferred sequences of mitochondrial proteins: instability of a tree based on a single gene. J. Mol. Evol. 395 19-527.

CHANDRASEKHARAN, U. M., S. SANKER, M. J. GLYNIAS, S. S. KARNIK, and A. HUSAIN. 1996. Angiotensin II-forming ac- tivity in a reconstructed ancestral chymase. Science 271: 502-505.

DAYHOFF, M. O., R. M. SCHWARTZ, and B. C. ORCUTT. 1978. A model of evolutionary change in proteins. Pp. 345-352. in M. 0. DAYHOFF, ed. Atlas of protein sequence and struc- ture. Vol. 5, Suppl. 3. National Biomedical Research Foun- dation, Washington, D.C.

DEEB, S. S., A. L. JORGENESEN, L. BATTISTI, L. IWASAKI, and A. G. MOTULSKY. 1994. Sequence divergence of the red and green visual pigments in great apes and humans. Proc. Natl. Acad. Sci. USA 91:7262-7266.

DULAI, K. S., J. K. BOWMAKER, J. D. MOLLON, and D. M. HUNT. 1994. Sequence divergence, polymorphism and evo- lution of the middle-wave and long-wave visual pigment genes of great apes and old world monkeys. Vision Res. 34:2483-249 1.

FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-79 1.

HUNT, D. M., K. S. DULAI, J. K. BOWMAKER, and J. D. MOL- LON. 1995. The chemistry of John Dalton’s color blindness. Science 267:984-988.

HUNT, D. M., A. J. WILLIAMS, J. K. BOWMAKER, and J. D. MOLLON. 1993. Structure and evolution of the polymorphic photopigment gene of the marmoset. Vision Res. 33:147- 154.

JACOBS, G. H. 1993. The distribution and nature of color vision among the mammals. Biol. Rev. 68:413-471.

JACOBS, G. H., M. NEITZ, J. E DEEGAN, and J. NEITZ. 1996. Trichromatic color vision in New World monkeys. Nature 382: 156-158.

JERMANN, T. M., J. G. OPITZ, J. STACKHOUSE, and S. A. BEN- NER. 1995. Reconstructing the evolutionary history of the artiodactyl ribonuclease superfamily. Nature 374:57-59.

JONES, D. T., W. R. TAYLOR, and J. M. THORNTON. 1992. The rapid generation of mutation data matrices from protein se- quences. Comput. Appl. Biosci. ?3:275-282.

KAWAMURA, S., and S. YOKOYAMA. 1993. Molecular charac- terization of the red visual pigment gene of the American chameleon (Anolis carolinensis). FEBS Lett. 323:247-25 1.

-. 1997. Expression of visual and nonvisual opsins in American chameleon. Vision Res. (in press).

KOJIMA, D., T OKANO, Y. FUKADA, Y. SHICHIDA, T YOSHI- ZAWA, and T. G. EBREY. 1992. Cone visual pigments are present in gecko rod cells. Proc. Natl. Acad. Sci. USA 89: 6841-6845.

KUMAR, S., K. TAMURA, and M. NEI. 1993. MEGA: molecular evolutionary genetics analysis. Version 1.01. The Pennsyl- vania State University, University Park.

MADDISON, W. I?, and D. R. MADDISON. 1992. MacClade: anal- ysis of phylogeny and character evolution. Version 3. Sin- auer, Sunderland, Mass.

MOLLON, J. 1991. The uses and evolutionary origins of primate color vision. Pp. 306319 in J. R. CRONLY-DILLON and R. L. GREGORY, eds. Evolution of the eye and visual system. CRC Press, Boca Raton, Fla.

MOLLON, J. D., J. K. BOWMAKER, and G. H. JACOBS. 1984. Variations of color vision in a New World primate can be explained by a polymorphism of retinal photopigments. Proc. R. Sot. Lond. B 222:373-399.

NATHANS, J., D. THOMAS, and D. S. HOGNESS. 1986. Molecular genetics of human color vision: the genes encoding blue, green, and red pigments. Science 232:193-202.

NEI, M. 1975. Molecular population genetics and evolution. North-Holland, Oxford.

NEI, M., and T. GOJOBORI. 1986. Simple methods for estimat- ing the numbers of synonymous and nonsynonymous nu- cleotide substitutions. Mol. Biol. Evol. 3:418-426.

NEITZ, M., J. NEITZ, and G. H. JACOBS. 1991. Spectral tuning of pigments underlying red-green color vision. Science 252: 97 l-974.

OKANO, T., D. KOJIMA, Y. FUKADA, Y. SHICHIDA, and T. YOSH- IZAWA. 1992. Primary structures of chicken cone visual pig- ments: vertebrate rhodopsins have evolved out of cone vi- sual pigments. Proc. Natl. Acad. Sci. USA 89:5932-5936.

PIATIGORSKY, J., and G. J. WISTOW. 1989. Enzyme/crystallins: Gene as an evolutionary strategy. Cell 57:197-199.

QUAX-JEUKEN, Y., S. BRUISTEN, E. NEVO, H. BLOEMENDAL, and W. W. DEJONG. 1985. Evolution of crystallins: expres- sion of lens-specific proteins in the blind mammals mole (T&pa europaea) and mole rat (Spalax ehrenbergi). Mol. Biol. Evol. 2:279-288.

RZHETSKY, A., and M. NEI. 1992. A simple method for esti- mating and testing minimum-evolution trees. Mol. Biol. Evol. 9:945-967.

SAITOU, N., and M. NEI. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425.

SASSONE-CORSI, I? 1994. Rhythmic transcription and autore- gulatory loops: winding up the biological clock. Cell 78: 361-364.

SHYUE, S.-K., D. HEWETT-EMMETT, H. G. SPERLING, D. M. HUNT, J. K. BOWMAKER, J. D. MOLLON, and W.-H. LI. 1995. Adaptive evolution of color vision genes in higher primates. Science 269: 1265- 1267.

SHYUE, S.-K., L. LI, B. H.-J. CHANG, and W.-H. LI. 1994. Intronic gene conversion in the evolution of human X-linked color vision genes. Mol. Biol. Evol. 11:548-551.

TOSINI, G., and M. MENAKER. 1996. Circadian rhythms in cul- tured mammalian retina. Science 272:4 19-42 1.

WALLS, G. L. 1942. The vertebrate eye and its adaptive radi- ation. Cranbook Press, Bloomsfield, Mich.

WINDERICKX, J., L. BATTISTI, Y. HIBIYA, A. G. MOTULSKY, and S. S. DEEB. 1993. Haplotype diversity in the human red and green opsin genes: evidence for frequent sequence exchange in exon 3. Hum. Mol. Genet. 2:1413-1421.

618 Nei et al.

WINDERICKX, J., D. T. LINDSEY, E. SANOCKI, D. Y. TELLER, A. G. MOTULSKY, and S. S. DEEB. 1992. Polymorphism in red photopigment underlies variation in color matching. Nature 356:43 l-433.

WISTOW, G. 1993. Lens crystallins: gene recruitment and evo- lutionary dynamism. Trends Biochem. Sci. l&301-306.

YANG, Z., S. KUMAR, and M. NEI. 1995. A new method of inference of ancestral nucleotide and amino acid sequences. Genetics 141:1641-1650.

YOKOYAMA, R., and S. YOKOYAMA. 1990. Convergent evolu- tion of the red- and green-like visual pigment genes in fish, Astyanax fasciatus, and human. Proc. Natl. Acad. Sci. USA 87:9315-9318.

YOKOYAMA, S. 1995. Amino acid replacements and wave- length absorption of visual pigments in vertebrates. Mol. Biol. Evol. 12:53-61.

YOKOYAMA, S., A. MEANY, H. WILKENS, and R. YOKOYAMA. 1995. Initial mutational steps toward loss of opsin gene function in cavefish. Mol. Biol. Evol. 12:527-532.

YOKOYAMA, S., and R. YOKOYAMA. 1989. Molecular evolution of human visual pigment genes. Mol. Biol. Evol. 6:186- 197.

ZHANG, J., and M. NEI. 1997. Accuracies of ancestral amino acid sequences inferred by the parsimony, likelihood, and distance methods. J. Mol. Evol. 44:S139-S146.

ZHOU, Y.-H., and W.-H. LI. 1996. Gene conversion and natural selection in the evolution of X-linked color vision genes in higher primates. Mol. Biol. Evol. 13:780-783.

MARCY K. UYENOYAMA, reviewing editor

Accepted February 14, 1997