The Molecular Basis of Adaptive Evolution of Squirrelfish Rhodopsins
http://www.100md.com
分子生物学进展 2004年第11期
Department of Biology, Emory University, Atlanta, Georgia
E-mail: syokoya@emory.edu.
Abstract
The wavelengths of maximal absorption (max) of the rhodopsins of nine squirrelfishes (N. sammara, N. argenteus, S. punctatissimum, S. microstoma, S. diadema, S. xantherythrum, S. spiniferum, N. aurolineatus, and S. tiere) and two soldierfishes (M. violacea and M. berndti) vary between 481 and 502 nm. Phylogenetic and mutagenesis analyses suggest that the common ancestor of these pigments had a max value of approximately 493 nm, and the contemporary max values were generated mostly by amino acid replacements E122M, F261Y, and A292S. The probability of observing all these amino acid replacements at specific branches of the phylogenetic tree is only 2.5 x 10–9; it is highly unlikely that these changes have occurred by neutral evolution. Because of a close association between the max values of these pigments and the wavelengths of light available to the corresponding species, the excess number of amino acid changes at specific branches in the phylogenetic tree strongly suggests that the rhodopsins have undergone adaptive changes at various stages of the holocentrid evolution.
Key Words: Squirrelfishes ? rhodopsin ? adaptive evolution
Introduction
Shallow water receives a wide range of wavelengths of light from the sun, but as we go deeper, the light is filtered by the water, and its spectrum becomes narrower at approximately 480 nm (Jerlov 1976). This change in the water environment has led to a diversity of specializations in visual systems of organisms (Lythgoe 1979). One such change can be seen in the relationship between the wavelength of maximal absorption (max) of visual pigments and depth of habitat of their possessors. Extensive survey data show that the max values of visual pigments of mesopelagic marine fishes are more blue-shifted than those of shallow-water species (e.g., Lythgoe 1979; Crescitelli 1991). Similarly, cottoid fishes of Lake Baikal that occupy progressively deeper habitats show a blue shift in the max value of their visual pigments in rods (rhodopsins) in a stepwise manner (Bowmaker et al. 1994; Hunt et al. 1996). The orthologous pigments in dolphin (Tursiops truncatus [Fasick and Robinson 1998]) and coelacanths (Latemaria chalamnae [Yokoyama et al. 1999]) also have blue-shifted max values. These max values seem to be needed for the visual adaptation to the blue water environments (Yokoyama 2000a).
Visual pigments usually consist of an opsin and the photosensitive molecule 11-cis-retinal chromophore and have been classified into five paralogous groups: (1) RH1 (consisting of mostly rhodopsins), (2) RH2 (RH1-like), (3) SWS1 (short wavelength–sensitive type 1), (4) SWS2 (SWS type 2), and (5) M/LWS (middle wavelength–sensitive and long wavelength–sensitive) pigments (Yokoyama and Yokoyama 1996; Yokoyama 2000a; Ebrey and Koutalos 2001). Among these, the molecular basis of specific wavelength sensitivity, or spectral tuning, of visual pigments has been reasonably well understood for the M/LWS pigments, where the observed max values of 510 to 560 nm can be explained fully by amino acid differences at five sites (Yokoyama and Radlwimmer 2001). The max values of the SWS1 pigments are based on at least a total of 10 amino acid sites (Yokoyama, Radlwimmer, and Blow 2000; Wilkie et al. 2000; Shi, Radlwimmer, and Yokoyama 2001; Shi and Yokoyama 2003; Fasick, Applebury, and Oprian 2002). Because of strong synergistic interactions among these pigments, however, the molecular basis of spectral tuning of the SWS1 pigments is far from clear (Shi, Radlwimmer, and Yokoyama 2001; Shi and Yokoyama 2003). Additional eight amino acid sites are also known to be involved in the max shift of RH1, RH2, and SWS2 pigments, but the mechanisms of spectral tunings of these pigments are not well understood (Yokoyama 2000a; Ebrey and Koutalos 2001; Takahashi and Ebrey 2003).
Toller (1996) cloned and sequenced the RH1 opsin cDNAs of a total nine squirrelfish species and two soldierfish species (Holocentridae: Beryciformes) and found that the max values of 481 to 502 nm of the corresponding pigments are closely associated with the depths of the holocentrid habitats (see also Munz and McFarland [1973]). These data provide a unique opportunity to explore the molecular basis not only of spectral tuning of RH1 pigments but also of possible adaptive evolution of these pigments to different water environments. Phylogenetic and mutagenesis analyses show that most of these max values were generated by amino acid replacements at sites 122, 261, and 292. The results also suggest that the RH1 pigments in the common ancestor of the holocentrids had a max value of approximately 493 nm, from which both red shifts and blue shifts in the max value have been achieved by adaptive evolution.
Materials and Methods
Background Information
Introducing various amino acid changes into the bovine RH1 pigment, H. Gobind Khorana and his colleagues and other biochemical vision scientists have elucidated many fundamental properties of the structure-function relationships of visual pigments (for reviews, see Yokoyama [2000a] and Ebrey and Koutalos [2001]). However, most mutations considered in these biochemical analyses are not found in nature, and their significance in the actual max shifts of visual pigments is not necessarily clear (Yokoyama 1995). Based on actual polymorphism data, mutagenesis experiments have also been conducted using visual pigments from a wide range of vertebrate species (Yokoyama 2000a; Shi, Radlwimmer, and Yokoyama 2001; Wilkie et al. 2000; Fasick, Applebury, and Oprian 2002; Shi and Yokoyama 2003; Takahashi and Ebrey 2003; Parry et al. 2004). From these evolutionary analyses, it is now known that specific amino acid changes at least at 23 sites have caused the variable max values of contemporary visual pigments (table 1).
Table 1 Naturally Occurring Amino Acid Changes That Shift the max of Visual Pigments More Than 5 nm
RH1 Pigments of Squirrelfishes and Soldierfishes and Sequence Data Analyses
Holocentridae, generally known as squirrelfishes, consists of subfamilies Holocentrinae (squirrelfishes) and Myripristinae (soldierfishes) (Nelson 1994). The RH1 pigments considered in this paper are those of nine squirrelfishes (denoted as N. sammara (P502), N. argent (P502), N. aurolin (P481), S. punc (P494), S. microst (P494), S. diadema (P491), S. xanther (P486), S. spinif (P490), and S. tiere (P490)) and of two soldierfishes (M. viola (P499) and M. berndti (P493)), where the numbers after P refer to max values (table 2). Based on the type of their habitats, these fish can be distinguished roughly into three groups: (1) N. sammara, N. argenteus, and S. punctatissimum and M. violacea, living mostly at the depth of 0 to 10 m (type I habitat); (2) S. microstoma, S. diadema, S. xantherythrum, S. spiniferum, S. tiere, and M. berndti, which live anywhere between 0 to 70 m (type II habitat); and (3) N. aurolineatus, living mostly at the depth lower than 60 m (type III habitat) (Toller 1996).
Table 2 RH1 Pigments of Squirrelfish and Soldierfish
To construct the rooted phylogenetic tree of the 11 holocentrid pigments, we have used the orthologous cavefish (P503) (GenBank accession number U12328), goldfish (P502) (GenBank accession number L11863), zebrafish (P501) (GenBank accession number AF109368), river lamprey (P500) (GenBank accession number M63632), clawed frog (P502) (GenBank accession number L0770), salamander (P506) (GenBank accession number U36574), pigeon (P502) (GenBank accession number AF149230), human (P500) (GenBank accession number U49742), bovine (P500) (GenBank accession number M21606), and mouse (P498) (GenBank accession number M55171) pigments as the outgroup. The topology and branch lengths of the phylogenetic tree of the visual pigments were inferred by applying the neighbor-joining (NJ) method (Saitou and Nei 1987) to the nucleotide distance data evaluated by using the two-parameter model (Kimura 1980). The reliability of the phylogenetic tree was evaluated by the bootstrap analysis with 1,000 replications (Felsenstein 1985).
The amino acid sequences of ancestral pigments were inferred using a likelihood-based Bayesian method (Yang 1997). In the inference, we considered the phylogenetic relationship of (river lamprey (P500), (((goldfish (P492), zebrafish (P501)), cavefish (P503)), ((((N. sammara (P502), N. argent (P502)), S. punc (P494), (S. microst (P494), S. diadema (P491), S. xanther (P486))), (N. aurolin, (S. spinif (P490), S. tiere (P490)))), (M. viola (P499), M. berndti (P493))))) (see Yokoyama [2000a] and Results and Discussion).
In vitro Assays of the Bovine RH1 Pigment
Point mutations were generated by using QuickChange site-directed mutagenesis kit (Stratagene). To rule out spurious mutations, the mutated opsins were sequenced by using the Sequitherm Excel II long-read kits (Epicentre Technologies, Madison, Wis.) with dye-labeled M13 forward and reverse primers. Sequencing reactions were run on a LI-COR 4200LD automated DNA sequencer (LI-COR, Lincoln, Neb.). The bovine (P500) and its mutant cDNAs in an expression vector, pMT5, were expressed in COS1 cells by transient transfection. The visual pigments were regenerated by incubating these opsins with 11-cis-retinal (Storm Eye Institute, Medical University of South Carolina, Charleston) in the dark. The resulting visual pigments were then purified by immunoaffinity chromatography by using monoclonal antibody 1D4 Sepharose 4B (The Cell Culture Center, Minneapolis, Minn.) (Yokoyama 2000b). The absorption spectra of visual pigments were recorded at 20°C using a Hitachi (Tokyo) U-3000 dual beam spectrophotometer. Recorded spectra were analyzed by using SIGMAPLOT software (Jandel, San Rafael, Calif.).
Results and Discussion
The Phylogenetic Tree of Holocentrid RH1 Pigments
Applying the NJ method to their nucleotide sequences between codon positions 4 and 332, we have constructed the rooted phylogenetic tree of 11 RH1 pigments (N. sammara (P502), N. argent (P502), S. punc (P494), S. microst (P494), S. diadema (P491), S. xanther (P486), N. aurolin (P481), S. spinif (P490), S. tiere (P490), M. viola (P499), and M. berndti (P493)) (fig. 1; see also Materials and Methods). Eight of the 10 possible nodes in figure 1 have bootstrap values greater than 0.95, and most branching patterns are highly reliable. This tree topology is consistent with those obtained by the maximum-parsimony and NJ methods of Toller (1996). These results show that holocentrid pigments are distinguished into squirrelfish and soldierfish pigments. The squirrelfish pigments are further distinguished into three major groups: (1) N. sammara (P502), N. argent (P502), S. punc (P494), S. microst (P494), S. diadema (P491), and S. xanther (P486); (2) N. aurolin (P481); and (3) S. spinif (P490) and S. tiere (P490).
FIG. 1.— The phylogenetic tree of the Holocentridae RH1 pigments. The tree was constructed by applying the NJ method to the nucleotide sequences between codon positions 4 and 332. The numbers after P refer to max values (Toller 1996). The numbers in brackets indicate the expected max values of pigments, which were calculated by assuming that E122M, F261Y, and A292S shift the max of visual pigments by –7, +10, and –10 nm, respectively. The numbers at different nodes indicate the bootstrap values. The branch lengths are not in scale.
Euteleosts and land animals separated some 400 MYA (e.g., Kumar and Hedges 1998). Taking a series of averages, the total number of nucleotide substitutions between the ancestral vertebrate pigment and the contemporary pigments is 0.174 per site. Because the corresponding value for the common ancestor of the squirrelfish and soldierfish pigments is 0.054, the two pigment groups seem to have diverged approximately 124 MYA. When the amino acid sequences are considered, the distances for the vertebrate and holocentrid ancestral pigments are 0.172 and 0.055 per site, respectively, suggesting that the contemporary holocentrid pigments diverged approximately 128 MYA. Thus, the sequence analyses suggest that the squirrelfish and soldierfish pigments have diverged some 120 to 130 MYA. For the squirrelfish pigments, the branch lengths of the common ancestor of the three pigment groups, of the pigment 1 group, and of pigment 2 and 3 groups are 0.033 (0.042), 0.019 (0.026), and 0.024 (0.033) at the nucleotide (or amino acid) level, respectively. Thus, the ancestral squirrelfish pigment goes back to approximately 75 to 100 MYA, group 1 pigments split from each other approximately 45 to 60 MYA, and the second and third groups diverged approximately 55 to 85 MYA.
Among the squirrelfish pigments, the max value of approximately 500 nm of N. sammara (P502), N. argent (P502), and M viola (P499) is suited to detect the light at the shallow water of type I habitat and that of 481 nm of N. aurolin (P481) is suited to detect the light at the much deeper type III habitat (see Materials and Methods). Most holocentrids with other pigments with max values of 486 to 494 nm live anywhere at the depth between 0 and 70 m (fig. 1). Thus, there is a close association between the max value of pigments and the wavelengths of light available to holocentrids possessing these pigments. S. punctatissimum is an exception and is supposed to live at the surface, but its RH1 pigment has a max value of 494 nm. It should be noted that not only is there some possibility of misclassification of holocentrid habitats but also some max values may be erroneous (see below). The phylogenetic tree in figure 1 also shows that the visual pigments sampled from type I and II habitats and those sampled from type II and III habitats are more closely related with each other, suggesting that the max shifts of these visual pigments have occurred in a stepwise manner.
Spectral Tuning and Evolution of Holocentrid Pigments
The max values of most holocentrid pigments reported by Toller (1996) agree with those in Munz and McFarland (1973) (table 2). In both Toller (1996) and Munz and McFarland (1973), the max value is determined by the difference between the prebleached and the fully bleached spectra. It should be cautioned, however, that this "extraction spectrophotometry (ESP)" method may not always provide the correct max value of the unbleached visual pigments. For example, the max value of the prebleached pigment in cavefish (P503) is 503 nm, but that estimated from the difference between the prebleached and fully bleached pigments is 508 nm (Yokoyama, Knox, and Yokoyama 1995). The absorption spectrum obtained by the ESP method also has a rather flat plateau with a large standard deviation and, therefore, the accuracy of the max value is not always clear-cut. Indeed, the max value of S. xanther (P486) is approximately 485 nm in Toller (1996) and Munz and McFarland (1973), but the corresponding value is 490 nm when the microspectrophotometry (MSP) method is used (Ellis R. Loew, personal communication; table 2). Thus, the max values of the holocentrid pigments must be interpreted with caution. In the future, it would be helpful to reevaluate these max values using in vitro assay (see Materials and Methods), where the prebleached (or dark) spectrum and the prebleached and bleached (or dark-light) difference spectrum can both be evaluated. Despite this uncertainty, table 2 suggests that the max values of most holocentrid pigments are repeatable.
When the RH1 pigments from a wide range of vertebrates are surveyed, most pigments have a max value of approximately 500 nm (Yokoyama 2000a; Ebrey and Koutalos 2001). Therefore, the max values of 480 to 494 nm of the squirrelfish and soldierfish RH1 pigments seem to have arisen in the past 120 to 130 Myr. Among the 23 functionally important amino acid sites (see Materials and Methods), only five sites (positions 97, 116, 164, 261, and 292) are polymorphic among the 11 pigments (fig. 2). In addition, amino acids M122 (methionine at site 122) are monomorphic among the holocentrid pigments, but those in the orthologous pigments in most other vertebrates are E122 (fig. 2; Yokoyama 2000a), showing that E122M (amino acid change from glutamic acid to methionine at site 122) occurred in the holocentrid ancestor (fig. 1). In fact, the maximum-likelihood–based Bayesian method (Yang 1997) suggests that E122M occurred in the common ancestor of the holocentrids, followed by T97S, F116S, A164G, F261Y, and A292S during holocentrid evolution.
FIG. 2.— Amino acid replacements at functionally critical 23 sites. The amino acid sequence of the ancestral pigment (ancestor) was inferred using likelihood-based Bayesian method (Yang 1997). Dots indicate the identity of the amino acids with those of the ancestral pigment.
Using the orthologous bovine (P500) pigment with a max value of 500 nm, the effects of F261Y and A292S on the max shift are shown to be +10 nm (Chan, Lee, and Sakmar 1992) and –10 nm (Sun, Macke, and Nathans 1997; Lin et al. 1998; Yokoyama 2000b), respectively. At present, the effects of the other four amino acid changes on the max shift are not known. Therefore, we introduced amino acid changes T97S, F116S, E122M, and A164G into the bovine (P500) pigment. Our mutagenesis analysis shows that the respective mutant pigments have max values of 500 ± 1, 500 ± 1, 493 ± 1, and 500 ± 1 nm (fig. 3). These results strongly suggest that the variable max values in the holocentrid pigments have been generated by E122M, F261Y, and A292S (fig. 1).
FIG. 3.— The absorption spectra of the bovine RH1 pigment with mutations T97S, F116S, E122M, and A164G. The dark-light difference spectra are shown in the inset. The corresponding max values measured by dark and dark-light spectra agree with each other.
We may assume that the max shifts of the holocentrid RH1 pigments are determined by E122M, F261Y, and A292S in an additive fashion and that the RH1 pigment in the vertebrate ancestor had a max value of 500 nm (Yokoyama 2000a; Ebrey and Koutalos 2001). Under these conditions, we can estimate the max values of all ancestral and contemporary pigments; that is, the holocentrid ancestor had a max value of 493 nm, from which the 11 contemporary RH1 pigments have evolved by incorporating F261Y and/or A292S. These inferred max values are compatible with the observed values of most contemporary pigments (fig. 1). Note that the expected max value of 493 nm of S. xanther (P486) is close to the observed value of 490 nm by MSP (table 2). At present, however, the 6-nm difference between the expected and observed max values of M. viola (P499) cannot be explained.
Figure 1 shows that many contemporary holocentrids have inherited their max values of 493 nm directly from their common ancestor and live in the type II habitat (0 to 70 m), suggesting that the holocentrid ancestor was an "ecological generalist" (Toller 1996). From this ancestor, three major changes seem to have occurred during holocentrid evolution. First, although they are similar to that of the vertebrate ancestor, who lived some 400 MYA, the max values of 502 nm of N. sammara (P502) and N. argent (P502) are of much more recent origin. The red shift in the max value was caused by F261Y, and both species (N. sammara and N. argenteus) now live mostly near the surface of the ocean. Second, the common ancestor of three species (N. aurolineatus, S. spiniferum, and S. tiere) seem to have modified the max value of their RH1 pigments from 493 nm to 483 nm using A292S. Today, the direct descendant (N. aurolineatus) lives mostly at the depth below 60 m, suggesting that the common ancestor may have lived in the deeper habitat as well. Interestingly, a much more distantly related coelacanth (Latimeria chalumnae) lives at the depth of 200 m and uses RH1 pigments with a max value of 485 nm, and this blue shift was caused by E122Q and A292S (Yokoyama et al. 1999). Thus, both squirrelfish and coelacanth ancestors used the identical A292S mutation for shifting the max values of their RH1 pigments toward blue. Third, the common ancestor of S. spiniferum and S. tiere reverted the max value of its RH1 pigments from 483 nm to 493 nm using F261Y. These two species now live at the depth of 0 to 70 m. These evolutionary analyses reveal that the max values of the holocentrid pigments have been modified possibly two times, depending on the species, suggesting frequent changes in the habitat choice of squirrelfishes in the last 120 to 130 Myr.
Adaptive Evolution of Holocentrid RH1 Pigments
We have seen that the max values of most holocentrid pigments correspond to habitat depth. Thus, it is of considerable interest to evaluate whether this correlation arose because of positive selection. Such selective forces have often been inferred by demonstrating that the number of nonsynonymous changes per codon site is significantly larger than that of synonymous changes per codon site in a phylogenetic tree (e.g., Nei and Kumar 2000). Using the parsimony-based method (Suzuki and Gojobori 1999; Suzuki and Nei 2001) and likelihood-based Bayesian method (Yang 1997; Yang et al. 2000), we could not identify any positively selected codon sites. This is not surprising, because only a small number of amino acid changes are involved in the spectral tuning of visual pigments (Suzuki and Nei 2004).
To explore the possible adaptive changes in the holocentrid pigments, we shall evaluate the probability of observing E122M, F261Y, and A292S at certain branches under neutral evolution.
The probability of a specific amino acid change (AB) in a certain branch is calculated as the product of two quantities: (1) the probability () that the ancestral amino acid, A, is replaced by any other amino acid and (2) the probability (?) that the amino acid change AB occurs, given that A is replaced by another amino acid. To assess the value, we first identify all amino acid sites (N) where the vertebrate ancestor had a specific amino acid, A, and count all amino acid replacements in the entire phylogenetic tree. Note that when we trace amino acid replacements at such critical sites in an opsin tree, not only is the number of changes small but also the changes are almost always unidirectional. Then, the average number of amino acid changes per site in the entire tree (K) can be evaluated by dividing the total numbers of amino acid replacements by N. Based on the amino acid sequences of ancestral organisms (see Materials and Methods), the K values for glutamic acid, phenylalanine, and alanine are 3/19, 9/29, and 11/25, respectively (table 3). Thus, without any selective force, = 1–exp (–K x L/LT) for a specific branch of length of L and the total branch length of LT (= 0.344) for the entire tree. Table 3 shows that the values for the three nonspecific amino acid replacements are less than 3%. The ? value is simply the proportion of the transition AB among all amino acid changes from A to any other amino acids. Considering all amino acid changes from glutamic acid, phenylalanine, and alanine to others, the ? values of amino acid replacements EM, AS, and FY are given by 1/5, 8/19, and 6/18, respectively. Therefore, without any selective force, the probabilities ( x ?) of observing the specific amino acid change at branches a, b, c, and d in figure 1 are all less than 0.012 (table 3). Furthermore, the probability of observing the four independent amino acid replacements is 2.5 x 10–9.
Table 3 Probabilities of Amino Acid Changes at Sites 122, 261, and 292
We may also distinguish amino acid replacements in branches a to d (B1 branches) from those in the rest of the phylogenetic tree (B0 branches). Then, the total lengths of the B1 and B0 branches are 0.121 and 0.223 with the relative lengths of 0.35 and 0.65, respectively. Thus, under neutral evolution, the probability that at least one E122M, two F261Y, and one A292S occur independently anywhere in the B1 branches is 4.6 x 10–7. On the other hand, without any selective force, we expect that at least any one of the three amino acid changes occur in the B0 branches with the probability of 1 – exp[–(3/19 + 9/29 + 11/25) x 0.223/0.344] (= 0.44). Despite this much higher probability, we see neither such amino acid changes nor any nucleotide changes at the three sites in the B0 branches. Together, these results demonstrate that the four amino acid and nucleotide substitutions are highly specific to the B1 branches. The accelerated amino acid replacements in the B1 branches can be caused either by relaxed selective constraints or by positive selection. However, the association between the max shifts in the RH1 pigments caused by the three amino acid changes and the depth of holocentrid habitats strongly suggests that these amino acid changes have undergone positive Darwinian selection.
These analyses show that many holocentrids have RH1 pigments with blue-shifted max values. In fact, such species as marine eel, Conger eel, John Dory, coelacanth, chameleon, and dolphin are also known to have blue-shifted max values (table 4; Yokoyama 2000a). With the exceptions of chameleon, these species live in the ocean and the blue-shifted max values are most probably caused by their adaptation to the blue ocean environments. The chameleon pigments use the 11-cis-3, 4-dehydroretinal as the chromophore (Provencio, Loew, and Foster 1992). Visual pigments with the 11-cis-3, 4-dehydroretinal absorb more red-shifted wavelength than those with the 11-cis-retinal chromophore (Whitmore and Bowmaker 1989; Harosi 1994). Therefore, the RH1 pigments in this species must have a blue-shifted max value to readjust the absorption spectrum toward the 500-nm wavelength. Again, the changes in the max values of these RH1 pigments seem to have been affected by their photic environments. These functional changes seem to have been accomplished by a small number of amino acid replacements, such as D83N, E122Q, and A292S (table 4; Yokoyama 2000a). Thus, our results show that the molecular basis of spectral tuning in the currently characterized RH1 pigments at both amino acid and max levels can be explained by a total of seven amino acid replacements at six sites: D83N, E122Q, E122M, H211C, F261Y, A292S, and A295S.
Table 4 The Blue-Shifted max Values of RH1 Pigments in Nonholocentrids
Acknowledgements
We thank Ruth Yokoyama and Margaret McFall-Ngai for their comments on the manuscript and Kevin Moses for introducing us to the holocentrid rhodopsins. This work was supported by a grant from the National Institutes of Health.
References
Bowmaker, J. K., V. I. Gorvardovskii, S. A. Shukolyukov, L. V. Zueva, D. M. Hunt, V. G. Sideleva, and O. G. Smirnova. 1994. Visual pigments and the photic environment: the cottoid fish of Lake Bikal. Vision Res. 34:592–605.
Chan, T., M. Lee, and T. P. Sakmar. 1992. Introduction of hydroxyl-bearing amino acids causes bathochromatic spectral shifts in rhodopsin: amino acid substitutions responsible for red-green color pigment spectral tuning. J. Biol. Chem. 267:9478–9480.
Crescitelli, F. 1991. Visual pigments. The scotopic photoreceptors and their visual pigments of fishes: functions and adaptations. Vision Res. 31:339–348.
Ebrey, T., and Y. Koutalos. 2001. Vertebrate photoreceptors. Prog. Retin. Eye Res. 20:49–94.
Fasick, J. I., M. L. Applebury, and D. D. Oprian. 2002. Spectral tuning in the mammalian short wavelength sensitive cone pigments. Biochemistry 41:6860–6865.
Fasick, J. I., and P. R. Robinson. 1998. Mechanism of spectral tuning in the dolphin visual pigments. Biochemistry 37:432–438.
Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791.
Harosi, F. I. 1994. An analysis of two spectral properties of vertebrate visual pigments. Vision Res. 34:1359–1367.
Hunt, D. M., J. Fitzgibbon, S. J. Slobodyanyuk, and J. K. Bowmaker. 1996. Spectral tuning and molecular evolution of rod visual pigments in the species flock of Cottoid fish in lake Baikal. Vision Res. 36:1217–1224.
Jerlov, N. G. 1976. Marine optics. Elsevier, Amsterdam.
Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111–120.
Kumar, S., and S. B. Hedges. 1998. A molecular time scale for vertebrate evolution. Nature 392:917–919.
Lin, S. W., C. K. S. Kochendoerfer, D. Wang, R. A. Mathies, and T. P. Sakmar. 1998. Mechanisms of spectral tuning in blue cone visual pigments. J. Biol. Chem. 273:24583–24591.
Lythgoe, J. N. 1979. The ecology of vision. Clarendon, Oxford.
Munz, F. W., and W. N. McFarland. 1973. The significance of spectral position in the rhodopsins of tropical marine fishes. Vision Res. 13:1829–1874.
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York.
Nelson, J. S. 1994. Fishes of the world. John Wiley & Sons, New York.
Parry, J. W. L., S. Poopalasundaram, J. K. Bowmaker, and D. M. Hunt. 2004. A novel amino acid substitution is responsible for spectral tuning in a rodent violet-sensitive visual pigment. Biochemistry 43:8014–8020.
Provencio, I., E. R. Loew, and R. G. Foster. 1992. Vitamin A2-based visual pigments in fully terrestrial vertebrates. Vision Res. 32:2201–2208.
Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425.
Shi, Y., F. B. Radlwimmer, and S. Yokoyama. 2001. Molecular genetics and the evolution of ultraviolet vision in vertebrates. Proc. Natl. Acad. Sci. USA 98:11731–11736.
Shi, Y., and S. Yokoyama. 2003. Molecular analysis of the evolutionary significance of ultraviolet vision in vertebrates. Proc. Natl. Acad. Sci. USA 100:8308–8313.
Sun, H., J. P. Macke, and J. Nathans. 1997. Mechanisms of spectral tuning in the mouse green cone pigments. Proc. Natl. Acad. Sci. USA 94:8860–8865.
Suzuki, Y., and T. Gojobori. 1999. A method for detecting positive selection at single amino acid sites. Mol. Biol. Evol. 16:1315–1328.
Suzuki, Y., and M. Nei. 2001. Reliabilities of parsimony-based and likelihood methods for detecting positive selection at single amino acid sites. Mol. Biol. Evol. 18:2179–2185.
———. 2004. False positive selection identified by ML-based methods: examples from the Sig1 gene of the diatom Thalassiosira weissflogii and the tax gene of a human T-cell lymphotrophic virus. Mol. Biol. Evol. 21:914–921.
Takahashi, Y., and T. G. Ebrey. 2003. Molecular basis of spectral tuning in the newt short wavelength sensitive visual pigments. Biochemistry 42:6025–6034.
Toller, W. W. 1996. Rhodopsin evolution in the Holocentridae (Pisces: Beryciformes). Doctoral dissertation, University Southern California, Los Angeles.
Whitmore, A. V., and J. K. Bowmaker. 1989. Seasonal variation in cone sensitivity and short-wave absorbing visual pigments in the rudd Scadinius erythrophythalmus. J. Comp. Physiol. A 166:103–115.
Wilkie, S. E., P. R. Robinson, T. W. Cronin, S. Poopalasundaram, J. K. Bowmaker, and D. M. Hunt. 2000. Spectral tuning of avian violet- and ultraviolet-sensitive visual pigment. Biochemistry 39:7895–7901.
Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput. Appl. Biosci. 13:555–556.
Yang, Z., R. Nielsen, N. Goldman, and A.-M. Krabbe Pedersen. 2000. Codon-substitution models for heterogeneous selection pressure at amino acids. Genetics 155:431–449.
Yokoyama, R., B. E. Knox, and S. Yokoyama. 1995. Rhodopsin from the fish, Astyanax: role of tyrosine 261 in the red shift. Invest. Ophthalmol. Vision Sci. 36:939–945.
Yokoyama, S. 1995. Amino acid replacements and wavelength absorption of visual pigments in vertebrates. Mol. Biol. Evol. 12:53–61.
———. 2000a. Molecular evolution of vertebrate visual pigments. Prog. Retin. Eye Res. 19:385–419.
———. 2000b. Phylogenetic analysis and experimental approaches to study color vision in vertebrates. Methods Enzymol. 315:312–325.
Yokoyama, S., and F. B. Radlwimmer. 2001. The molecular genetics of red and green color vision in mammals. Genetics 153:919–932.
Yokoyama, S., F. B. Radlwimmer, and N. S. Blow. 2000. Ultraviolet pigments in birds evolved from violet pigments by a single amino acid change. Proc. Natl. Acad. Sci. USA 97:7366–7371.
Yokoyama, S., and R. Yokoyama. 1996. Adaptive evolution of photoreceptors and visual pigments in vertebrates. Annu. Rev. Ecol. Syst. 27:534–567.
Yokoyama, S., H. Zhang, F. B. Radlwimmer, and N. S. Blow. 1999. Adaptive evolution of color vision of the Comoran coelacanth (Latimeria chalumnae). Proc. Natl. Acad. Sci. USA 96:6279–6284.(Shozo Yokoyama and Naomi )
E-mail: syokoya@emory.edu.
Abstract
The wavelengths of maximal absorption (max) of the rhodopsins of nine squirrelfishes (N. sammara, N. argenteus, S. punctatissimum, S. microstoma, S. diadema, S. xantherythrum, S. spiniferum, N. aurolineatus, and S. tiere) and two soldierfishes (M. violacea and M. berndti) vary between 481 and 502 nm. Phylogenetic and mutagenesis analyses suggest that the common ancestor of these pigments had a max value of approximately 493 nm, and the contemporary max values were generated mostly by amino acid replacements E122M, F261Y, and A292S. The probability of observing all these amino acid replacements at specific branches of the phylogenetic tree is only 2.5 x 10–9; it is highly unlikely that these changes have occurred by neutral evolution. Because of a close association between the max values of these pigments and the wavelengths of light available to the corresponding species, the excess number of amino acid changes at specific branches in the phylogenetic tree strongly suggests that the rhodopsins have undergone adaptive changes at various stages of the holocentrid evolution.
Key Words: Squirrelfishes ? rhodopsin ? adaptive evolution
Introduction
Shallow water receives a wide range of wavelengths of light from the sun, but as we go deeper, the light is filtered by the water, and its spectrum becomes narrower at approximately 480 nm (Jerlov 1976). This change in the water environment has led to a diversity of specializations in visual systems of organisms (Lythgoe 1979). One such change can be seen in the relationship between the wavelength of maximal absorption (max) of visual pigments and depth of habitat of their possessors. Extensive survey data show that the max values of visual pigments of mesopelagic marine fishes are more blue-shifted than those of shallow-water species (e.g., Lythgoe 1979; Crescitelli 1991). Similarly, cottoid fishes of Lake Baikal that occupy progressively deeper habitats show a blue shift in the max value of their visual pigments in rods (rhodopsins) in a stepwise manner (Bowmaker et al. 1994; Hunt et al. 1996). The orthologous pigments in dolphin (Tursiops truncatus [Fasick and Robinson 1998]) and coelacanths (Latemaria chalamnae [Yokoyama et al. 1999]) also have blue-shifted max values. These max values seem to be needed for the visual adaptation to the blue water environments (Yokoyama 2000a).
Visual pigments usually consist of an opsin and the photosensitive molecule 11-cis-retinal chromophore and have been classified into five paralogous groups: (1) RH1 (consisting of mostly rhodopsins), (2) RH2 (RH1-like), (3) SWS1 (short wavelength–sensitive type 1), (4) SWS2 (SWS type 2), and (5) M/LWS (middle wavelength–sensitive and long wavelength–sensitive) pigments (Yokoyama and Yokoyama 1996; Yokoyama 2000a; Ebrey and Koutalos 2001). Among these, the molecular basis of specific wavelength sensitivity, or spectral tuning, of visual pigments has been reasonably well understood for the M/LWS pigments, where the observed max values of 510 to 560 nm can be explained fully by amino acid differences at five sites (Yokoyama and Radlwimmer 2001). The max values of the SWS1 pigments are based on at least a total of 10 amino acid sites (Yokoyama, Radlwimmer, and Blow 2000; Wilkie et al. 2000; Shi, Radlwimmer, and Yokoyama 2001; Shi and Yokoyama 2003; Fasick, Applebury, and Oprian 2002). Because of strong synergistic interactions among these pigments, however, the molecular basis of spectral tuning of the SWS1 pigments is far from clear (Shi, Radlwimmer, and Yokoyama 2001; Shi and Yokoyama 2003). Additional eight amino acid sites are also known to be involved in the max shift of RH1, RH2, and SWS2 pigments, but the mechanisms of spectral tunings of these pigments are not well understood (Yokoyama 2000a; Ebrey and Koutalos 2001; Takahashi and Ebrey 2003).
Toller (1996) cloned and sequenced the RH1 opsin cDNAs of a total nine squirrelfish species and two soldierfish species (Holocentridae: Beryciformes) and found that the max values of 481 to 502 nm of the corresponding pigments are closely associated with the depths of the holocentrid habitats (see also Munz and McFarland [1973]). These data provide a unique opportunity to explore the molecular basis not only of spectral tuning of RH1 pigments but also of possible adaptive evolution of these pigments to different water environments. Phylogenetic and mutagenesis analyses show that most of these max values were generated by amino acid replacements at sites 122, 261, and 292. The results also suggest that the RH1 pigments in the common ancestor of the holocentrids had a max value of approximately 493 nm, from which both red shifts and blue shifts in the max value have been achieved by adaptive evolution.
Materials and Methods
Background Information
Introducing various amino acid changes into the bovine RH1 pigment, H. Gobind Khorana and his colleagues and other biochemical vision scientists have elucidated many fundamental properties of the structure-function relationships of visual pigments (for reviews, see Yokoyama [2000a] and Ebrey and Koutalos [2001]). However, most mutations considered in these biochemical analyses are not found in nature, and their significance in the actual max shifts of visual pigments is not necessarily clear (Yokoyama 1995). Based on actual polymorphism data, mutagenesis experiments have also been conducted using visual pigments from a wide range of vertebrate species (Yokoyama 2000a; Shi, Radlwimmer, and Yokoyama 2001; Wilkie et al. 2000; Fasick, Applebury, and Oprian 2002; Shi and Yokoyama 2003; Takahashi and Ebrey 2003; Parry et al. 2004). From these evolutionary analyses, it is now known that specific amino acid changes at least at 23 sites have caused the variable max values of contemporary visual pigments (table 1).
Table 1 Naturally Occurring Amino Acid Changes That Shift the max of Visual Pigments More Than 5 nm
RH1 Pigments of Squirrelfishes and Soldierfishes and Sequence Data Analyses
Holocentridae, generally known as squirrelfishes, consists of subfamilies Holocentrinae (squirrelfishes) and Myripristinae (soldierfishes) (Nelson 1994). The RH1 pigments considered in this paper are those of nine squirrelfishes (denoted as N. sammara (P502), N. argent (P502), N. aurolin (P481), S. punc (P494), S. microst (P494), S. diadema (P491), S. xanther (P486), S. spinif (P490), and S. tiere (P490)) and of two soldierfishes (M. viola (P499) and M. berndti (P493)), where the numbers after P refer to max values (table 2). Based on the type of their habitats, these fish can be distinguished roughly into three groups: (1) N. sammara, N. argenteus, and S. punctatissimum and M. violacea, living mostly at the depth of 0 to 10 m (type I habitat); (2) S. microstoma, S. diadema, S. xantherythrum, S. spiniferum, S. tiere, and M. berndti, which live anywhere between 0 to 70 m (type II habitat); and (3) N. aurolineatus, living mostly at the depth lower than 60 m (type III habitat) (Toller 1996).
Table 2 RH1 Pigments of Squirrelfish and Soldierfish
To construct the rooted phylogenetic tree of the 11 holocentrid pigments, we have used the orthologous cavefish (P503) (GenBank accession number U12328), goldfish (P502) (GenBank accession number L11863), zebrafish (P501) (GenBank accession number AF109368), river lamprey (P500) (GenBank accession number M63632), clawed frog (P502) (GenBank accession number L0770), salamander (P506) (GenBank accession number U36574), pigeon (P502) (GenBank accession number AF149230), human (P500) (GenBank accession number U49742), bovine (P500) (GenBank accession number M21606), and mouse (P498) (GenBank accession number M55171) pigments as the outgroup. The topology and branch lengths of the phylogenetic tree of the visual pigments were inferred by applying the neighbor-joining (NJ) method (Saitou and Nei 1987) to the nucleotide distance data evaluated by using the two-parameter model (Kimura 1980). The reliability of the phylogenetic tree was evaluated by the bootstrap analysis with 1,000 replications (Felsenstein 1985).
The amino acid sequences of ancestral pigments were inferred using a likelihood-based Bayesian method (Yang 1997). In the inference, we considered the phylogenetic relationship of (river lamprey (P500), (((goldfish (P492), zebrafish (P501)), cavefish (P503)), ((((N. sammara (P502), N. argent (P502)), S. punc (P494), (S. microst (P494), S. diadema (P491), S. xanther (P486))), (N. aurolin, (S. spinif (P490), S. tiere (P490)))), (M. viola (P499), M. berndti (P493))))) (see Yokoyama [2000a] and Results and Discussion).
In vitro Assays of the Bovine RH1 Pigment
Point mutations were generated by using QuickChange site-directed mutagenesis kit (Stratagene). To rule out spurious mutations, the mutated opsins were sequenced by using the Sequitherm Excel II long-read kits (Epicentre Technologies, Madison, Wis.) with dye-labeled M13 forward and reverse primers. Sequencing reactions were run on a LI-COR 4200LD automated DNA sequencer (LI-COR, Lincoln, Neb.). The bovine (P500) and its mutant cDNAs in an expression vector, pMT5, were expressed in COS1 cells by transient transfection. The visual pigments were regenerated by incubating these opsins with 11-cis-retinal (Storm Eye Institute, Medical University of South Carolina, Charleston) in the dark. The resulting visual pigments were then purified by immunoaffinity chromatography by using monoclonal antibody 1D4 Sepharose 4B (The Cell Culture Center, Minneapolis, Minn.) (Yokoyama 2000b). The absorption spectra of visual pigments were recorded at 20°C using a Hitachi (Tokyo) U-3000 dual beam spectrophotometer. Recorded spectra were analyzed by using SIGMAPLOT software (Jandel, San Rafael, Calif.).
Results and Discussion
The Phylogenetic Tree of Holocentrid RH1 Pigments
Applying the NJ method to their nucleotide sequences between codon positions 4 and 332, we have constructed the rooted phylogenetic tree of 11 RH1 pigments (N. sammara (P502), N. argent (P502), S. punc (P494), S. microst (P494), S. diadema (P491), S. xanther (P486), N. aurolin (P481), S. spinif (P490), S. tiere (P490), M. viola (P499), and M. berndti (P493)) (fig. 1; see also Materials and Methods). Eight of the 10 possible nodes in figure 1 have bootstrap values greater than 0.95, and most branching patterns are highly reliable. This tree topology is consistent with those obtained by the maximum-parsimony and NJ methods of Toller (1996). These results show that holocentrid pigments are distinguished into squirrelfish and soldierfish pigments. The squirrelfish pigments are further distinguished into three major groups: (1) N. sammara (P502), N. argent (P502), S. punc (P494), S. microst (P494), S. diadema (P491), and S. xanther (P486); (2) N. aurolin (P481); and (3) S. spinif (P490) and S. tiere (P490).
FIG. 1.— The phylogenetic tree of the Holocentridae RH1 pigments. The tree was constructed by applying the NJ method to the nucleotide sequences between codon positions 4 and 332. The numbers after P refer to max values (Toller 1996). The numbers in brackets indicate the expected max values of pigments, which were calculated by assuming that E122M, F261Y, and A292S shift the max of visual pigments by –7, +10, and –10 nm, respectively. The numbers at different nodes indicate the bootstrap values. The branch lengths are not in scale.
Euteleosts and land animals separated some 400 MYA (e.g., Kumar and Hedges 1998). Taking a series of averages, the total number of nucleotide substitutions between the ancestral vertebrate pigment and the contemporary pigments is 0.174 per site. Because the corresponding value for the common ancestor of the squirrelfish and soldierfish pigments is 0.054, the two pigment groups seem to have diverged approximately 124 MYA. When the amino acid sequences are considered, the distances for the vertebrate and holocentrid ancestral pigments are 0.172 and 0.055 per site, respectively, suggesting that the contemporary holocentrid pigments diverged approximately 128 MYA. Thus, the sequence analyses suggest that the squirrelfish and soldierfish pigments have diverged some 120 to 130 MYA. For the squirrelfish pigments, the branch lengths of the common ancestor of the three pigment groups, of the pigment 1 group, and of pigment 2 and 3 groups are 0.033 (0.042), 0.019 (0.026), and 0.024 (0.033) at the nucleotide (or amino acid) level, respectively. Thus, the ancestral squirrelfish pigment goes back to approximately 75 to 100 MYA, group 1 pigments split from each other approximately 45 to 60 MYA, and the second and third groups diverged approximately 55 to 85 MYA.
Among the squirrelfish pigments, the max value of approximately 500 nm of N. sammara (P502), N. argent (P502), and M viola (P499) is suited to detect the light at the shallow water of type I habitat and that of 481 nm of N. aurolin (P481) is suited to detect the light at the much deeper type III habitat (see Materials and Methods). Most holocentrids with other pigments with max values of 486 to 494 nm live anywhere at the depth between 0 and 70 m (fig. 1). Thus, there is a close association between the max value of pigments and the wavelengths of light available to holocentrids possessing these pigments. S. punctatissimum is an exception and is supposed to live at the surface, but its RH1 pigment has a max value of 494 nm. It should be noted that not only is there some possibility of misclassification of holocentrid habitats but also some max values may be erroneous (see below). The phylogenetic tree in figure 1 also shows that the visual pigments sampled from type I and II habitats and those sampled from type II and III habitats are more closely related with each other, suggesting that the max shifts of these visual pigments have occurred in a stepwise manner.
Spectral Tuning and Evolution of Holocentrid Pigments
The max values of most holocentrid pigments reported by Toller (1996) agree with those in Munz and McFarland (1973) (table 2). In both Toller (1996) and Munz and McFarland (1973), the max value is determined by the difference between the prebleached and the fully bleached spectra. It should be cautioned, however, that this "extraction spectrophotometry (ESP)" method may not always provide the correct max value of the unbleached visual pigments. For example, the max value of the prebleached pigment in cavefish (P503) is 503 nm, but that estimated from the difference between the prebleached and fully bleached pigments is 508 nm (Yokoyama, Knox, and Yokoyama 1995). The absorption spectrum obtained by the ESP method also has a rather flat plateau with a large standard deviation and, therefore, the accuracy of the max value is not always clear-cut. Indeed, the max value of S. xanther (P486) is approximately 485 nm in Toller (1996) and Munz and McFarland (1973), but the corresponding value is 490 nm when the microspectrophotometry (MSP) method is used (Ellis R. Loew, personal communication; table 2). Thus, the max values of the holocentrid pigments must be interpreted with caution. In the future, it would be helpful to reevaluate these max values using in vitro assay (see Materials and Methods), where the prebleached (or dark) spectrum and the prebleached and bleached (or dark-light) difference spectrum can both be evaluated. Despite this uncertainty, table 2 suggests that the max values of most holocentrid pigments are repeatable.
When the RH1 pigments from a wide range of vertebrates are surveyed, most pigments have a max value of approximately 500 nm (Yokoyama 2000a; Ebrey and Koutalos 2001). Therefore, the max values of 480 to 494 nm of the squirrelfish and soldierfish RH1 pigments seem to have arisen in the past 120 to 130 Myr. Among the 23 functionally important amino acid sites (see Materials and Methods), only five sites (positions 97, 116, 164, 261, and 292) are polymorphic among the 11 pigments (fig. 2). In addition, amino acids M122 (methionine at site 122) are monomorphic among the holocentrid pigments, but those in the orthologous pigments in most other vertebrates are E122 (fig. 2; Yokoyama 2000a), showing that E122M (amino acid change from glutamic acid to methionine at site 122) occurred in the holocentrid ancestor (fig. 1). In fact, the maximum-likelihood–based Bayesian method (Yang 1997) suggests that E122M occurred in the common ancestor of the holocentrids, followed by T97S, F116S, A164G, F261Y, and A292S during holocentrid evolution.
FIG. 2.— Amino acid replacements at functionally critical 23 sites. The amino acid sequence of the ancestral pigment (ancestor) was inferred using likelihood-based Bayesian method (Yang 1997). Dots indicate the identity of the amino acids with those of the ancestral pigment.
Using the orthologous bovine (P500) pigment with a max value of 500 nm, the effects of F261Y and A292S on the max shift are shown to be +10 nm (Chan, Lee, and Sakmar 1992) and –10 nm (Sun, Macke, and Nathans 1997; Lin et al. 1998; Yokoyama 2000b), respectively. At present, the effects of the other four amino acid changes on the max shift are not known. Therefore, we introduced amino acid changes T97S, F116S, E122M, and A164G into the bovine (P500) pigment. Our mutagenesis analysis shows that the respective mutant pigments have max values of 500 ± 1, 500 ± 1, 493 ± 1, and 500 ± 1 nm (fig. 3). These results strongly suggest that the variable max values in the holocentrid pigments have been generated by E122M, F261Y, and A292S (fig. 1).
FIG. 3.— The absorption spectra of the bovine RH1 pigment with mutations T97S, F116S, E122M, and A164G. The dark-light difference spectra are shown in the inset. The corresponding max values measured by dark and dark-light spectra agree with each other.
We may assume that the max shifts of the holocentrid RH1 pigments are determined by E122M, F261Y, and A292S in an additive fashion and that the RH1 pigment in the vertebrate ancestor had a max value of 500 nm (Yokoyama 2000a; Ebrey and Koutalos 2001). Under these conditions, we can estimate the max values of all ancestral and contemporary pigments; that is, the holocentrid ancestor had a max value of 493 nm, from which the 11 contemporary RH1 pigments have evolved by incorporating F261Y and/or A292S. These inferred max values are compatible with the observed values of most contemporary pigments (fig. 1). Note that the expected max value of 493 nm of S. xanther (P486) is close to the observed value of 490 nm by MSP (table 2). At present, however, the 6-nm difference between the expected and observed max values of M. viola (P499) cannot be explained.
Figure 1 shows that many contemporary holocentrids have inherited their max values of 493 nm directly from their common ancestor and live in the type II habitat (0 to 70 m), suggesting that the holocentrid ancestor was an "ecological generalist" (Toller 1996). From this ancestor, three major changes seem to have occurred during holocentrid evolution. First, although they are similar to that of the vertebrate ancestor, who lived some 400 MYA, the max values of 502 nm of N. sammara (P502) and N. argent (P502) are of much more recent origin. The red shift in the max value was caused by F261Y, and both species (N. sammara and N. argenteus) now live mostly near the surface of the ocean. Second, the common ancestor of three species (N. aurolineatus, S. spiniferum, and S. tiere) seem to have modified the max value of their RH1 pigments from 493 nm to 483 nm using A292S. Today, the direct descendant (N. aurolineatus) lives mostly at the depth below 60 m, suggesting that the common ancestor may have lived in the deeper habitat as well. Interestingly, a much more distantly related coelacanth (Latimeria chalumnae) lives at the depth of 200 m and uses RH1 pigments with a max value of 485 nm, and this blue shift was caused by E122Q and A292S (Yokoyama et al. 1999). Thus, both squirrelfish and coelacanth ancestors used the identical A292S mutation for shifting the max values of their RH1 pigments toward blue. Third, the common ancestor of S. spiniferum and S. tiere reverted the max value of its RH1 pigments from 483 nm to 493 nm using F261Y. These two species now live at the depth of 0 to 70 m. These evolutionary analyses reveal that the max values of the holocentrid pigments have been modified possibly two times, depending on the species, suggesting frequent changes in the habitat choice of squirrelfishes in the last 120 to 130 Myr.
Adaptive Evolution of Holocentrid RH1 Pigments
We have seen that the max values of most holocentrid pigments correspond to habitat depth. Thus, it is of considerable interest to evaluate whether this correlation arose because of positive selection. Such selective forces have often been inferred by demonstrating that the number of nonsynonymous changes per codon site is significantly larger than that of synonymous changes per codon site in a phylogenetic tree (e.g., Nei and Kumar 2000). Using the parsimony-based method (Suzuki and Gojobori 1999; Suzuki and Nei 2001) and likelihood-based Bayesian method (Yang 1997; Yang et al. 2000), we could not identify any positively selected codon sites. This is not surprising, because only a small number of amino acid changes are involved in the spectral tuning of visual pigments (Suzuki and Nei 2004).
To explore the possible adaptive changes in the holocentrid pigments, we shall evaluate the probability of observing E122M, F261Y, and A292S at certain branches under neutral evolution.
The probability of a specific amino acid change (AB) in a certain branch is calculated as the product of two quantities: (1) the probability () that the ancestral amino acid, A, is replaced by any other amino acid and (2) the probability (?) that the amino acid change AB occurs, given that A is replaced by another amino acid. To assess the value, we first identify all amino acid sites (N) where the vertebrate ancestor had a specific amino acid, A, and count all amino acid replacements in the entire phylogenetic tree. Note that when we trace amino acid replacements at such critical sites in an opsin tree, not only is the number of changes small but also the changes are almost always unidirectional. Then, the average number of amino acid changes per site in the entire tree (K) can be evaluated by dividing the total numbers of amino acid replacements by N. Based on the amino acid sequences of ancestral organisms (see Materials and Methods), the K values for glutamic acid, phenylalanine, and alanine are 3/19, 9/29, and 11/25, respectively (table 3). Thus, without any selective force, = 1–exp (–K x L/LT) for a specific branch of length of L and the total branch length of LT (= 0.344) for the entire tree. Table 3 shows that the values for the three nonspecific amino acid replacements are less than 3%. The ? value is simply the proportion of the transition AB among all amino acid changes from A to any other amino acids. Considering all amino acid changes from glutamic acid, phenylalanine, and alanine to others, the ? values of amino acid replacements EM, AS, and FY are given by 1/5, 8/19, and 6/18, respectively. Therefore, without any selective force, the probabilities ( x ?) of observing the specific amino acid change at branches a, b, c, and d in figure 1 are all less than 0.012 (table 3). Furthermore, the probability of observing the four independent amino acid replacements is 2.5 x 10–9.
Table 3 Probabilities of Amino Acid Changes at Sites 122, 261, and 292
We may also distinguish amino acid replacements in branches a to d (B1 branches) from those in the rest of the phylogenetic tree (B0 branches). Then, the total lengths of the B1 and B0 branches are 0.121 and 0.223 with the relative lengths of 0.35 and 0.65, respectively. Thus, under neutral evolution, the probability that at least one E122M, two F261Y, and one A292S occur independently anywhere in the B1 branches is 4.6 x 10–7. On the other hand, without any selective force, we expect that at least any one of the three amino acid changes occur in the B0 branches with the probability of 1 – exp[–(3/19 + 9/29 + 11/25) x 0.223/0.344] (= 0.44). Despite this much higher probability, we see neither such amino acid changes nor any nucleotide changes at the three sites in the B0 branches. Together, these results demonstrate that the four amino acid and nucleotide substitutions are highly specific to the B1 branches. The accelerated amino acid replacements in the B1 branches can be caused either by relaxed selective constraints or by positive selection. However, the association between the max shifts in the RH1 pigments caused by the three amino acid changes and the depth of holocentrid habitats strongly suggests that these amino acid changes have undergone positive Darwinian selection.
These analyses show that many holocentrids have RH1 pigments with blue-shifted max values. In fact, such species as marine eel, Conger eel, John Dory, coelacanth, chameleon, and dolphin are also known to have blue-shifted max values (table 4; Yokoyama 2000a). With the exceptions of chameleon, these species live in the ocean and the blue-shifted max values are most probably caused by their adaptation to the blue ocean environments. The chameleon pigments use the 11-cis-3, 4-dehydroretinal as the chromophore (Provencio, Loew, and Foster 1992). Visual pigments with the 11-cis-3, 4-dehydroretinal absorb more red-shifted wavelength than those with the 11-cis-retinal chromophore (Whitmore and Bowmaker 1989; Harosi 1994). Therefore, the RH1 pigments in this species must have a blue-shifted max value to readjust the absorption spectrum toward the 500-nm wavelength. Again, the changes in the max values of these RH1 pigments seem to have been affected by their photic environments. These functional changes seem to have been accomplished by a small number of amino acid replacements, such as D83N, E122Q, and A292S (table 4; Yokoyama 2000a). Thus, our results show that the molecular basis of spectral tuning in the currently characterized RH1 pigments at both amino acid and max levels can be explained by a total of seven amino acid replacements at six sites: D83N, E122Q, E122M, H211C, F261Y, A292S, and A295S.
Table 4 The Blue-Shifted max Values of RH1 Pigments in Nonholocentrids
Acknowledgements
We thank Ruth Yokoyama and Margaret McFall-Ngai for their comments on the manuscript and Kevin Moses for introducing us to the holocentrid rhodopsins. This work was supported by a grant from the National Institutes of Health.
References
Bowmaker, J. K., V. I. Gorvardovskii, S. A. Shukolyukov, L. V. Zueva, D. M. Hunt, V. G. Sideleva, and O. G. Smirnova. 1994. Visual pigments and the photic environment: the cottoid fish of Lake Bikal. Vision Res. 34:592–605.
Chan, T., M. Lee, and T. P. Sakmar. 1992. Introduction of hydroxyl-bearing amino acids causes bathochromatic spectral shifts in rhodopsin: amino acid substitutions responsible for red-green color pigment spectral tuning. J. Biol. Chem. 267:9478–9480.
Crescitelli, F. 1991. Visual pigments. The scotopic photoreceptors and their visual pigments of fishes: functions and adaptations. Vision Res. 31:339–348.
Ebrey, T., and Y. Koutalos. 2001. Vertebrate photoreceptors. Prog. Retin. Eye Res. 20:49–94.
Fasick, J. I., M. L. Applebury, and D. D. Oprian. 2002. Spectral tuning in the mammalian short wavelength sensitive cone pigments. Biochemistry 41:6860–6865.
Fasick, J. I., and P. R. Robinson. 1998. Mechanism of spectral tuning in the dolphin visual pigments. Biochemistry 37:432–438.
Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791.
Harosi, F. I. 1994. An analysis of two spectral properties of vertebrate visual pigments. Vision Res. 34:1359–1367.
Hunt, D. M., J. Fitzgibbon, S. J. Slobodyanyuk, and J. K. Bowmaker. 1996. Spectral tuning and molecular evolution of rod visual pigments in the species flock of Cottoid fish in lake Baikal. Vision Res. 36:1217–1224.
Jerlov, N. G. 1976. Marine optics. Elsevier, Amsterdam.
Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111–120.
Kumar, S., and S. B. Hedges. 1998. A molecular time scale for vertebrate evolution. Nature 392:917–919.
Lin, S. W., C. K. S. Kochendoerfer, D. Wang, R. A. Mathies, and T. P. Sakmar. 1998. Mechanisms of spectral tuning in blue cone visual pigments. J. Biol. Chem. 273:24583–24591.
Lythgoe, J. N. 1979. The ecology of vision. Clarendon, Oxford.
Munz, F. W., and W. N. McFarland. 1973. The significance of spectral position in the rhodopsins of tropical marine fishes. Vision Res. 13:1829–1874.
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York.
Nelson, J. S. 1994. Fishes of the world. John Wiley & Sons, New York.
Parry, J. W. L., S. Poopalasundaram, J. K. Bowmaker, and D. M. Hunt. 2004. A novel amino acid substitution is responsible for spectral tuning in a rodent violet-sensitive visual pigment. Biochemistry 43:8014–8020.
Provencio, I., E. R. Loew, and R. G. Foster. 1992. Vitamin A2-based visual pigments in fully terrestrial vertebrates. Vision Res. 32:2201–2208.
Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–425.
Shi, Y., F. B. Radlwimmer, and S. Yokoyama. 2001. Molecular genetics and the evolution of ultraviolet vision in vertebrates. Proc. Natl. Acad. Sci. USA 98:11731–11736.
Shi, Y., and S. Yokoyama. 2003. Molecular analysis of the evolutionary significance of ultraviolet vision in vertebrates. Proc. Natl. Acad. Sci. USA 100:8308–8313.
Sun, H., J. P. Macke, and J. Nathans. 1997. Mechanisms of spectral tuning in the mouse green cone pigments. Proc. Natl. Acad. Sci. USA 94:8860–8865.
Suzuki, Y., and T. Gojobori. 1999. A method for detecting positive selection at single amino acid sites. Mol. Biol. Evol. 16:1315–1328.
Suzuki, Y., and M. Nei. 2001. Reliabilities of parsimony-based and likelihood methods for detecting positive selection at single amino acid sites. Mol. Biol. Evol. 18:2179–2185.
———. 2004. False positive selection identified by ML-based methods: examples from the Sig1 gene of the diatom Thalassiosira weissflogii and the tax gene of a human T-cell lymphotrophic virus. Mol. Biol. Evol. 21:914–921.
Takahashi, Y., and T. G. Ebrey. 2003. Molecular basis of spectral tuning in the newt short wavelength sensitive visual pigments. Biochemistry 42:6025–6034.
Toller, W. W. 1996. Rhodopsin evolution in the Holocentridae (Pisces: Beryciformes). Doctoral dissertation, University Southern California, Los Angeles.
Whitmore, A. V., and J. K. Bowmaker. 1989. Seasonal variation in cone sensitivity and short-wave absorbing visual pigments in the rudd Scadinius erythrophythalmus. J. Comp. Physiol. A 166:103–115.
Wilkie, S. E., P. R. Robinson, T. W. Cronin, S. Poopalasundaram, J. K. Bowmaker, and D. M. Hunt. 2000. Spectral tuning of avian violet- and ultraviolet-sensitive visual pigment. Biochemistry 39:7895–7901.
Yang, Z. 1997. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput. Appl. Biosci. 13:555–556.
Yang, Z., R. Nielsen, N. Goldman, and A.-M. Krabbe Pedersen. 2000. Codon-substitution models for heterogeneous selection pressure at amino acids. Genetics 155:431–449.
Yokoyama, R., B. E. Knox, and S. Yokoyama. 1995. Rhodopsin from the fish, Astyanax: role of tyrosine 261 in the red shift. Invest. Ophthalmol. Vision Sci. 36:939–945.
Yokoyama, S. 1995. Amino acid replacements and wavelength absorption of visual pigments in vertebrates. Mol. Biol. Evol. 12:53–61.
———. 2000a. Molecular evolution of vertebrate visual pigments. Prog. Retin. Eye Res. 19:385–419.
———. 2000b. Phylogenetic analysis and experimental approaches to study color vision in vertebrates. Methods Enzymol. 315:312–325.
Yokoyama, S., and F. B. Radlwimmer. 2001. The molecular genetics of red and green color vision in mammals. Genetics 153:919–932.
Yokoyama, S., F. B. Radlwimmer, and N. S. Blow. 2000. Ultraviolet pigments in birds evolved from violet pigments by a single amino acid change. Proc. Natl. Acad. Sci. USA 97:7366–7371.
Yokoyama, S., and R. Yokoyama. 1996. Adaptive evolution of photoreceptors and visual pigments in vertebrates. Annu. Rev. Ecol. Syst. 27:534–567.
Yokoyama, S., H. Zhang, F. B. Radlwimmer, and N. S. Blow. 1999. Adaptive evolution of color vision of the Comoran coelacanth (Latimeria chalumnae). Proc. Natl. Acad. Sci. USA 96:6279–6284.(Shozo Yokoyama and Naomi )