Characterization of the Long-Wavelength Opsin from Mecoptera and Siphonaptera: Does a Flea See?
http://www.100md.com
分子生物学进展 2005年第5期
* Brigham Young University, Department of Integrative Biology; and Brigham Young University, Department of Microbiology and Molecular Biology
Correspondence: E-mail: kd98@email.byu.edu.
Abstract
Mecoptera and Siphonaptera represent two insect orders that have largely been overlooked in the study of insect vision. Recent phylogenetic evidence demonstrates that Mecoptera (scorpionflies) is paraphyletic, with the order Siphonaptera (fleas) nesting as sister to the family Boreidae (snow fleas), showing an evolutionary trend towards reduction in gross eye morphology within fleas. We provide the first molecular characterization of long-wavelength opsins from these three lineages (opsin gene from fleas [FL-Opsin], the Boreidae [B-Opsin], and a mecopteran family [M-Opsin]) and assess the effects of loss of visual acuity on the structure and function of the opsin gene. Phylogenetic analysis implies a physiological sensitivity in the red-green spectrum for these opsins. Analysis of intron splice sites reveals a high degree of similarity between FL-Opsin and B-Opsin as well as conserved splice sites across insect blue-green and long-wavelength opsins. Calculated rates of evolution and tests for destabilizing selection indicate that FL-Opsin, B-Opsin, and M-Opsin are evolving at similar rates with no radical selective pressures, implying conservative evolution and functional constraint across all three lineages.
Key Words: opsin ? evolution ? Mecoptera ? Siphonaptera ? insect vision
Introduction
Insects represent a group of organisms in which visual perception and the associated structural components are highly developed (Briscoe and Chittka 2001). The class Insecta is one of the largest and most diverse classes of organisms on the planet, accounting for over 70% of all known species (Resh and Carde 2003). Insects have adapted to nearly every ecological niche available, from arctic to tropic and mountaintop to river bottom. However, despite this wide diversity of habitats and lifestyles, finding links in the adaptation of visual systems to visual ecology has proven to be difficult (Briscoe and Chittka 2001). It is important, therefore, to continue to characterize the visual system for a diversity of insect groups in order to gain a better understanding of the selective pressures and constraints that guide visual evolution in insects.
The photoreceptive cells in the insect compound eye contain light-sensitive visual pigments, which consist of an opsin apoprotein, bound by a protonated Schiff's base linkage to a photosensitive chromophore molecule, usually 11-cis-retinal. The visual cascade is initiated when the chromophore absorbs a photon of light, causing it to isomerize to the trans form. A corresponding conformational change in the opsin protein, a member of the G protein–coupled receptors, initiates a transducin-mediated response, which ultimately sends a neural signal to the brain. For a review of opsin physiology, see Nathans (1987). A single individual may possess several opsin variants, each sensitive to different wavelengths of light, thus conferring a specific range of spectral sensitivity to that organism (Kochendoerfer et al. 1999). Typically, insects possess trichromatic vision, conferred by at least three opsin classes: UV-, blue-, and green-sensitive variants. All insects studied have green receptors (530 nm), most express UV (330 nm) and blue (440 nm), and only a few have been shown to contain red (>565 nm) (Briscoe and Chittka 2001). Molecular characterizations and spectral tuning studies have been performed on opsin proteins found in many insect groups, including bees (Bellingham et al. 1997), moths and butterflies (Kitamoto et al. 1998; Briscoe 2000), mantids (Towner and Gartner 1994), fireflies (Cronin et al. 2000), and fruit flies (Neufeld, Carthew, and Rubin 1991; Carulli et al. 1994).
Mecoptera and Siphonaptera represent two insect groups that have largely been overlooked in the study of vision. Mecoptera is a relatively small, holometabolous insect order with approximately 600 described species (Penny and Byers 1979). For purposes of this study, two families are of particular interest. Panorpidae (true scorpionflies) is the most speciose family of mecopterans. They have a scavenging lifestyle, feeding primarily on dead, soft-bodied arthropods and often robbing spiders' webs of the encased insects (Byers and Thornhill 1983). They have well-developed eyes and are highly visually oriented, with a maximum spectral sensitivity peak at 490–520 nm and a minor peak at 360 nm (Burkhardt and De LaMotte 1972). Boreidae (snow fleas) are wingless, alpine insects that live and feed exclusively on mosses and are most often collected on winter snow (Penny 1977). The eyes of boreids are smaller and appear slightly less well developed than those of panorpids. We have observed that boreids will often escape collection by jumping when they are too closely approached, indicating that vision plays an important role in sensing the environment. Little is known, however, of boreid visual ecology.
Siphonaptera (fleas) is a highly specialized holometabolous insect order with approximately 2,380 described species (R. E. Lewis and J. H. Lewis 1985). Fleas are strict vertebrate ectoparasites and are of tremendous economic importance as vectors of pathogens, transmitting diseases such as plague, murine typhus, and tularemia (Dunnet and Mardon 1991). Little is known about flea vision, due mainly to the radical divergence in flea eye structure from the typical insect bauplan (Crum, Knapp, and Wite 1974). Fleas show a transformation of the multifaceted eyes and ommatidia of most insects, replaced instead with heavily sclerotized, atypical ocelli, or "eyespots," or in some cases, a complete absence of any eye at all (Dunnet and Mardon 1991). The species used in our analysis exhibit a diversity of eyespot morphologies, ranging from large eyespots in Pulex irritans, to small eyespots in Ctenophthalmus agyrtes and Jellisonia sp., to no discernable eyespot in Uropsylla tasmanica. Despite the drastic reduction in eye structure, Crum, Knapp, and Wite (1974) demonstrated that cat fleas (Ctenophalides felis) and oriental rat fleas (Xenopsylla cheopis) have peak light sensitivities at 330 and 530 nm, respectively, although the light elicited different responses in the two species. Osbrink and Rust (1985) further demonstrated that visual cues seem to be among the primary factors influencing host selection in cat fleas. However, this eyespot probably cannot form an acute image but merely acts as a light sensor (Osbrink and Rust 1985; Rust and Dryden 1997; Land 2003).
Whiting (2002) recently demonstrated that Mecoptera is paraphyletic, with Siphonaptera nesting within Mecoptera as sister group to Boreidae, and the obscure family Nannochoristidae placed as sister to Boreidae + Siphonaptera (fig. 1). This phylogenetic hypothesis provides an evolutionary framework in which to evaluate shifts in visual ecology and search for possible correlates in morphological and molecular changes associated with vision in these insects. Given that nannochoristids are visual mecopterans, optimization of vision on this phylogeny implies that there was a reduction of visual acuity in the flea lineage. We compare opsin structure and function in the ectoparasitic fleas, which have a single-lens light sensor and only general light-dark sensitivity, with that of their boreid sister group, which are detritus feeders and visually oriented insects. These in turn are compared with their highly visual panorpid relatives, which have a fully developed, image-forming eye. Because nannochoristids are rare insects, we have not included them in this study.
FIG. 1.— Condensed phylogeny of Mecoptera and Siphonaptera presented by Whiting (2002) and eye morphology of Panorpidae, Boreidae, and Siphonaptera.
This study focuses on the effects of the shift to parasitism in fleas, the associated loss of visual acuity, and the molecular processes underlying the evolution of the opsin protein in these insects. Given the demonstrated photosensitivity of the fleas and scorpionflies to green light, we isolated the long-wavelength (LW) opsin gene from fleas (FL-Opsin), their close sister group, the Boreidae (B-Opsin), and the Panorpidae, a highly visual mecopteran family (M-Opsin). We provide the first molecular characterization of insect opsins from these three groups. In addition, we compare the gene structure and evolutionary history of these three opsin groups in order to assess the effects of loss of visual acuity on the structure and function of the opsin gene. We specifically test for relaxation of functional constraint on the FL-Opsin protein due to the gross reduction in eye structure and visual sensitivity.
Materials and Methods
DNA Analysis: Polymerase Chain Reaction and Sequencing
LW opsin sequences were obtained from four flea species representing four families, six panorpid species representing the single diverse genus Panorpa, and one boreid species. For a detailed list of all taxa included in the analysis, see Appendix 1 (Supplementary Material online).
Genomic DNA was extracted from specimens using the Qiagen DNeasy extraction kit using manufacturer's guidelines. Because of the low yield of DNA expected from fleas, the DNA was eluted in only 100 μl of solution. Genomic DNA vouchers and specimen vouchers are deposited in the Insect Genomics Collection, M. L. Bean Museum, Brigham Young University.
Polymerase chain reaction (PCR) primers were developed from conserved regions of arthropod LW opsins and, once sufficient sequences were obtained, redesigned for use in Mecoptera and Siphonaptera. The amplified region spans the first six of the transmembrane helices and 60% of the seventh. Primer sequences are given in table 1. PCR amplification used the following three-step cycling protocol with Amplitaq Gold polymerase (ABI, Foster City, Calif.): 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 3 min, followed by an additional 7 min of elongation at 72°C. PCR products were visualized by gel electrophoresis and purified using the Millipore Montage purification system. Initial amplifications were attempted on all taxa using LF1 and Scylla primers in order to obtain the entire target region in a single amplification. If unsuccessful, smaller overlapping segments were amplified with internal primers.
Table 1 Primers Used to Amplify LW-Opsin
Because opsin variants arose through a series of gene duplication events (Deininger, Fuhrman, and Hegemann 2000), the possibility exists that paralogous opsin variants may coamplify. In order to separate potential paralogous copies, the PCR products were cloned using the pCR 2.1-TOPO cloning kit (Invitrogen) following the manufacturer's protocol. Ten clones from each PCR product were picked and then reamplified and sequenced using the provided M13 vector primers (Invitrogen, Carlsbad, Calif.) and BigDye Terminator chemistry (ABI). Sequence data were collected using the ABI Prism 3730 capillary autosequencer in the BYU DNA Sequencing Center.
The vector ends were trimmed from all cloned products and the resulting sequence blasted in GenBank to insure that an opsin product was indeed amplified. Clones and overlapping PCR regions from the same taxa were initially included in the phylogenetic analysis as separate terminals to verify that the sequences did not represent paralogous gene copies. If they were shown to group together, the sequences were then combined, with any discrepancies coded as ambiguous or missing data, and included as a single terminal in the final analysis.
Phylogenetic Analysis
Sequences from crustaceans and the LW, blue-green (BG), blue (BL), and UV insect opsin variants were downloaded from GenBank and used to determine the phylogenetic placement of FL-Opsin, M-Opsin, and B-Opsin (see Appendix 1). The sequence of bovine rhodopsin, as the only opsin to have its tertiary structure characterized, was included as an out-group. Initial alignments were performed manually in Sequencher 4.2 (GeneCodes 2003). Introns were identified through comparison to GenBank insect cDNA sequences (fig. 2). The resulting coding sequences were translated using MacClade (W. P. Maddison and D. R. Maddison 2003) and the amino acid sequences aligned via ClustalX (Thompson et al. 1997). The alignment was further refined by performing a homology alignment of the Jellisonia sp. amino acid sequence (the most complete opsin isolated in this study) to high-resolution bovine opsin crystal structure (Palczewski et al. 2000) in the Swiss-Pdb Viewer and submitted to the SwissModel Server for homology modeling (Guex and Peitsh 1997). This allowed us to optimize our overall nucleotide alignment based on both amino acid sequence homology and structural homology.
FIG. 2.— Comparison of FL-Opsin, M-Opsin, and B-Opsin intron splice sites to other insect opsins, based on Briscoe (1999). Numbers above hatch marks indicate position of splice site relative to Drosophila melanogaster Rh1 amino acid sequence. Numbers in parentheses indicate observed size (bp) of introns (shown only for LW opsins). Regions in black were not amplified in this analysis. Intron positions and lengths determined from alignment of our novel opsin sequences with the following: Drosophila melanogaster Rh1 (X65877 [GenBank] ), Papilio glaucus Rh3 (AF098283 [GenBank] ), Bee LW1 (Bombus impatiens: AY485302 [GenBank] , Bombus terrestris: AY485301 [GenBank] , Drosophila afflicta: AY485303 [GenBank] , Drosophila rinconis: AY485304 [GenBank] , Osmia rufa: AY572828 [GenBank] ), Bee LW2 (B. impatiens: AY485306 [GenBank] , B. terrestris: AY485305 [GenBank] , D. afflicta: AY485308 [GenBank] , D. rinconis: AY485307 [GenBank] , O. rufa: AY572829 [GenBank] , Apis mellifera: XM_397398 [GenBank] .1), Anopheles LW (GPRop5, GPRop6, GPRop7; see Appendix 1, available as Supplementary Material online).
Phylogenetic analyses were performed using the nucleotide sequences after intron removal. The phylogeny was reconstructed using Bayesian methods coupled with Markov chain Monte Carlo (BMCMC) techniques as implemented in MrBayes v3.0b4, which allows for mixed model analyses (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). Because different codon positions have different structural constraints, the data set was partitioned into first, second, and third codon positions. Models for each partition were selected using the procedure implemented in Modeltest v.3.6 (Posada and Crandall 1998). Mixed models were used with unlinked parameters between partitions, treating model parameters as unknown variables with uniform default priors. Three independent BMCMC analyses were run, each consisting of 10 independent Markov chains started from a random tree and run for 3 x 106 generations, with every 1,000th generation sampled. To confirm that each separate Bayesian analysis converged and mixed well, the fluctuating value of the likelihood and all phylogenetic parameters were examined graphically. To check for congruence among independent runs, the mean likelihood score and parameter values and the posterior probabilities (pP) for individual clades were compared. All sample points prior to reaching stationary values were discarded as burn-in, and the remaining trees from each independent analysis were combined to calculate the maximum a posteriori (MAP) tree (Huelsenbeck and Imennov 2002; Huelsenbeck et al. 2002).
Opsin Evolution
The ratio of nonsynonymous versus synonymous changes (dN/dS) for the insect LW portion of the tree was estimated with maximum likelihood using the codon-based substitution model of the CODEML software package in PAML v3.14 (Yang 1998). Several different site-specific models in which selective pressure varies among different sites but the site-specific pattern is identical across all lineages were implemented (Yang et al. 2000): model M0 (null model with no variation among sites), M1 ("neutral" model, with two categories of site with fixed dN/dS ratios of 0 and 1), M2 ("selection" model with three categories of site, two with fixed dN/dS ratios of 0 and 1, and a third estimated dN/dS ratio), M3 ("discrete" model with three categories of site, where the dN/dS ratio is free to vary for each site), M7 ("beta" model—10 categories of site, with 10 dN/dS ratios in the range 0–1 taken from a discrete approximation of the beta distribution), and M8 ("beta plus omega" model—10 categories of site from a beta distribution as in model M7 plus an additional category of site with a dN/dS ratio that is free to vary from 0 to greater than 1). PAML estimates the dN/dS ratios that are free to vary under these models, as well as the proportion of sites with each ratio. All models were run twice with starting omega values of less than and greater than 1 as suggested by the PAML manual to test for entrapment in local optima (Yang 1997). Likelihood ratio tests, to determine whether more complex models provided a significantly better fit to the data than more simple models, were performed by comparing the likelihood ratio test statistic (–2[ln L1 – ln L2]) to critical values of the chi-square distribution with the appropriate degrees of freedom (Yang 1998).
Although dN/dS ratios are useful for detecting selection, they do not indicate how the identified selection affects the overall structure and function of the protein. Amino acid substitutions can have a wide range of effects on a protein depending on the difference in physicochemical properties and location in the protein structure; in order to evaluate these effects we used the insect LW portion of our Bayesian tree in TreeSAAP v3.0 (Woolley et al. 2003) to test for evidence of selection among 31 amino acid properties in the three lineages of interest. TreeSAAP allows us to identify these property changes and classify them into categories on a gradient from conservative to radical change. Based on data set–specific nucleotide substitution patterns, a neutral model of expected change for each category is generated. This null model is then compared to the observed numbers of amino acid replacements in the data set, and a z-score is calculated for each category. Categories where the observed number of amino acid replacements is significantly different than neutral expectations are considered as potentially being affected by selective pressures. In this study, we specified a scale of 20 categories of change. Greater than expected numbers of replacements (relative to the neutral model) in categories 1–3 indicate significant stabilizing selection, whereas the same situation in categories 18–20 indicate significant destabilizing selection (table 3). By mapping the identified amino acid replacements back onto the protein structure, we are able to make inferences about the influence of potential selective pressures on protein function and on the evolutionary history of particular lineages (Woolley et al. 2003).
Table 3 Conservative (Stabilizing) and Radical (Destabilizing) Amino Acid Properties Under Selection Identified from 31 Tested Amino Acid Properties in TreeSAAP Analysisa
Results and Discussion
Opsin Gene Isolation and Phylogenetic Analysis
PCR amplification of the LW opsin gene for the Mecoptera and Siphonaptera yielded 1,030- and 1,200-bp fragments, respectively. We were unable to obtain 460 bp from the 5' end of Panorpa gracilis and Boreus coloradensis. The translated sequences used in the analysis ranged between 203 and 278 amino acids. Interestingly, Siphonaptera have a unique four–amino acid–long expansion segment in transmembrane helix 6, at amino acid positions 119–222 in our alignment (fig. 3).
FIG. 3.— Two-dimensional representation of the homology model of FL-Opsin showing preservation of the seven-helix structure. Numbering begins with the first amino acid included in this analysis. Black circles represent amino acids not obtained. Boxed residues show an expansion region found only in FL-Opsin. Circles with a cross in them show key conserved residues described by Briscoe (1999). The results of the TreeSAAP analysis are mapped onto the model. Blue-colored circles show sites of conservative amino acid (AA) replacement, and red-colored circles show sites of radical AA replacement found along the branches leading to FL-Opsin, M-Opsin, or B-Opsin clades. Circles shaded in light gray indicate sites strictly conserved across the three lineages. White circles represent AA residue variation among the three lineages.
The MAP tree (ln L = –24,065.002 and pP = 0.002) used in PAML and TreeSAAP analyses is shown in figure 4. Although not entirely congruent with the species phylogeny presented by Whiting (2002), there are nonetheless several significant features about this topology. Briscoe (2000) demonstrated that opsins will cluster together primarily by physiological similarity, i.e., similar spectral sensitivities, and secondarily by species relationships. Our data strongly support these results, with opsins of physiological similarity clustering together across the topology. FL-Opsins and M-Opsins are well supported as monophyletic groups within the insect LW clade, suggesting sensitivity in the red-green spectrum. We note, however, that although support for the ordinal grouping of the FL-Opsins, M-Opsins, and B-Opsin sequences is strong, the opsin gene generally provides little or no support for the resolution of these ordinal relationships to each other.
FIG. 4.— Maximum a posteriori estimate of phylogeny (MAP) tree used in PAML and TreeSAAP analyses, ln L = –24,065.002 and pP = 0.002. Posterior probabilities >0.95 are indicated by a gray circle on the corresponding branch. A detailed list of all taxa names and their respective accession numbers can be found in Appendix 1. The shared insertion at site 156 between Boreidae, Siphonaptera, and bee LW2 is mapped onto the respective branches, indicating a gain (red rectangle) and a loss (white rectangle).
Furthermore, a recent study of the Anopheles gambiae genome revealed the presence of a higher number of opsin genes than any other characterized insect, with at least half of the identified copies being duplicates of LW-sensitive genes (Hill et al. 2002). Phylogenetic analysis of this LW gene complement indicated at least one early gene duplication within insects before the emergence of the orders Orthoptera, Mantodea, Hymenoptera, Lepidoptera, and Diptera (Spaethe and Briscoe 2004). In our analysis, the B-Opsin, FL-opsin, and M-opsin sequences do not cluster with either the A. gambiae or the bee LW2 genes (fig. 4). Analysis of our alignment using the methods described by Spaethe and Briscoe (2004) produced a tree similar to their results, with the newly characterized B-Opsin, FL-Opsin, and M-Opsin sequences grouping within the insect LW1 clade (results not shown). Although the weak support in the LW clade of our MAP tree makes inferences about opsin gene duplication and evolution difficult, the phylogenetic relationships of the opsin gene copies in A. gambiae and Hymenoptera indicate that there may be many more yet unidentified gene duplication events in the insect LW clade.
Opsin Gene Structure
Figure 2 shows comparative intron splice site positions of FL-Opsin, B-Opsin, and M-Opsin relative to other insect LW opsins. Within the regions amplified, the putative Siphonaptera opsin gene contains five introns, two Panorpidae, and four Boreidae. Interestingly, the B-opsin and FL-opsins (putative LW1) and bee LW2 share a unique intron at position 156, although they are not a monophyletic clade in our phylogeny. Manual mapping of this character onto the MAP tree (fig. 4) indicates a single gain of this intron site, with secondary loss in the Bee LW1 clade.
Briscoe (1999) reported three splice sites conserved across insect BG and LW opsin groups (sites 190, 239, and 332, fig. 2). Our data support this observation for both sites 190 and 239, although secondary intron deletions are observed in the bee LW1 (position 239) and Anopheles and Panorpidae (position 190) lineages. Site 332 is outside the region amplified in this study and cannot be further evaluated. Additionally, our data indicate a unique splice site at 119 that, with the exception of bee LW1 and Anopheles, appears to be characteristic of LW opsins, and we predict it will also be present in boreids.
Of the remaining intron insertions, sites 79 in Papilio and 108 in bee LW1 appear to be autapomorphies (fig. 2). Several sites (277, Papilio and bee LW1 and LW2; 296, Boreidae and Siphonaptera) appear in only a few taxa and may indicate their presence before the divergence of those lineages. All of the Anopheles LW opsin gene copies investigated (GPRop5, GPRop6, GPRop7) have lost all intron splice sites except 239, implying that patterns of intron insertion-deletion can change considerably within lineages.
Even though two groups may share a particular intron splice site, the length of the insert was observed to vary. For example, at position 296 the boreid has a long intron (537 bp) and the fleas have relatively short introns (58–84 bp). Intron lengths vary even among taxa within the same group, e.g., sequence lengths for position 119 in Jellisonia sp., C. agyrtes, and P. irritans are 99, 119, and 66 bp, respectively. These results indicate that the splice sites appear to be more phylogenetically conserved than the actual makeup of the intron.
Removal of introns reveals a preserved open reading frame in the boreid-, panorpid-, and flea-lineages. Homology modeling of the inferred amino acid sequence of the Jellisonia sp. FL-Opsin suggests that the seven transmembrane helices characteristic of G protein–coupled receptors and all known insect opsins are preserved (fig. 3). Additionally, Briscoe (1999) described seven amino acid motifs that are necessary for a functional opsin protein and that are conserved across all known insect opsins. Four of these motifs are located within the region amplified in this study, all four of which are conserved in FL-Opsin, B-Opsin, and M-Opsin (fig. 3). These residues include (1) Leu36 and (2) Asn41 found in all G protein–coupled receptors (numbered according to the first residue obtained in FL-Opsin); (3) the G protein–coupled receptor motif, E/D R, in the third transmembrane region involved in transducin activation (Asp102 and Arg103); and (4) a disulfide bridge connecting transmembrane region 3 with the second cytostolic loop at Cys78/155. As expected, many of these key residues occur in regions for which the amino acid sequence is conserved across all insect LW opsins.
Opsin Evolution
The results of the PAML analysis are summarized in table 2. For the majority of the comparisons made, we failed to reject the null hypothesis that models that allow for positive selection fit the data significantly better than models that do not. However, model M3 showed a significantly better score (P = 0.001) than the neutral model M1, indicating that selective pressure is present. Although model M3 fit the data better, no substitutions were found in categories where > 1, rejecting the hypothesis of diversifying selection. This result is confirmed by examining the individual rates of evolution found in M-Opsin, B-Opsin, and FL-Opsin and comparing them to the other lineages in the topology. If, for instance, in the flea lineage there was either strong selection away from the ancestral protein or a loss of functional constraint of the opsin protein due to the reduction in eye structure, we expect to see a greater number of nucleotide substitutions as well as a larger proportion of nonsynonymous to synonymous substitutions (Yokoyama et al. 1995). The dN/dS ratios on the branch leading to the FL-Opsin clade are not significantly different (P = 0.05) from those leading to the B-Opsin and M-Opsin clades (see table 2). These ratios are also nearly identical to the average rates of evolution across the entire topology (P = 0.05). In addition, all ratios are observed to be less than 1. As stated earlier, this indicates that any selective pressure on opsin evolution is purifying across all the insects observed such that the structural and functional properties of the protein are being conserved (Yang et al. 2000).
Table 2 Results from PAML Analysis of FL-Opsin, B-Opsin, and M-Opsin Lineages
Analysis in TreeSAAP also confirmed that a low rate of radical diversifying selection (destabilizing selection) was occurring with respect to amino acid properties. Only 9.6% of all sites showed destabilizing selection among 7 of the 31 properties tested. On the other hand, 21.4% of all sites are under strongly stabilizing selection for 22 of 31 amino acid properties tested, and 60% of the amino acid residues in FL-Opsin, B-Opsin, and M-Opsin were identical. Although several of the properties showing destabilizing selection (e.g., turn tendencies, coil tendencies, or power to be at the N-terminal) seem to be related to structural aspects of the protein, none are properties previously identified to affect spectral tuning (i.e., polarity). Additionally, all of the conserved residues as described by Briscoe (1999) also show conservation in our data set. These results are shown in figure 3 mapped onto the two-dimensional projection of FL-Opsin. See table 3 for a complete list of all properties tested. As a whole, these data not only imply functional constraint but also similar functional properties across FL-Opsin, B-Opsin, and M-Opsin.
Sensitivity to Topology
Because of the weak support of the ordinal relationships found in our analysis, the PAML and TreeSAAP analyses were also calculated after constraining the data to the Whiting (2002) species tree. The results in all cases were identical, suggesting that the data are relatively insensitive to organization of the ordinal relationships, as predicted by Yang et al. (2000). We observe that despite the overlying reduction in eye morphology and loss of visual sensitivity across these three lineages, there remains strong conservative, purifying selection on the opsin protein.
Functional Constraint of FL-Opsin, B-Opsin, and M-Opsin
Despite the considerable variation in lifestyle, habitat, and visual structure, the LW opsin gene appears to be remarkably well preserved across panorpids, boreids, and fleas. This is not entirely surprising, given that all three lineages demonstrate sensitivity and responsiveness to visual cues at some level. Perhaps with gross reduction in flea eye morphology and the incapability to perceive images, we might have expected to see some corresponding change in the mechanism of photoreception itself. However, as discussed above, not only does FL-Opsin appear to have retained functional constraint but also no evidence is found for unusual selective pressure on the amino acid sequence and structural properties. On the contrary, we observed a high degree of conservative selection, preserving amino acid properties and structure similar to other functional opsins. These results provide evidence that although vast differences in perceptual abilities exist between Panorpidae, Boreidae, and Siphonaptera, e.g., ability to form images, the underlying physiology involved in the mechanisms of photoreception appears to be preserved.
In addition, the preservation of opsin structure in Siphonaptera, combined with the apparent reduction of macroscopic eye structure may also indicate the presence of an extraretinal photoreceptor system. Indeed, preliminary transmission electron studies on flea eyespots indicate no preservation of typical microscopic adult insect eye structures (e.g., rhabdom, crystalline cone, corneal lens) as well as a heavily sclerotized layer of chitin covering the flea "eyes" (Dittmar de la Cruz, unpublished data). This structural degradation implies that photoreception may no longer occur within the eye. Extraretinal, nonvisual photoreceptors are common among insects and have been found at sites located in the central nervous system (Briscoe and Nagy 1999; Shimizu et al. 2001), the posterior margin of the compound eye (Yasuyama and Meinertzhagen 1999), and the peripheral abdominal segments (Arikawa 1997; see G. Fleissner and G. Fleissner 2003 for a review of nonvisual photoreceptors). A number of LW opsin copies have been implicated as extraretinal (Briscoe and Nagy 1999; Spaethe and Briscoe 2004), and Spaethe and Briscoe (2004) speculate that these extraretinal opsins evolved early in invertebrate evolution. Interestingly, similar to the peripheral photoreceptors found on the last abdominal segments (genitalia) of Papilio xuthus (Arikawa et al. 1980), fleas have a sensillial plate (or pygidium) on the last abdominal segment, which is packed with receptor-like structures. Unfortunately, the function of these "receptors" is still a source of speculation.
Furthermore, extraretinal photoreceptors appear to be involved in the photic entrainment of the circadian clock (Shimizu et al. 2001; Malpel, Klarsfeld, and Rouyer 2002). Koehler, Leppla, and Patterson (1989) demonstrated the presence of circadian rhythms in fleas, which could be mediated by these types of nonvisual receptors.
Other studies have also shown examples of organisms that do not have functional eye structures, yet seem to retain functional opsin proteins. R. Yokoyama and S. Yokoyama (1990), for example, presented a case where putative functional opsin genes were isolated from Astyanax fasciatus, a blind cave fish. This phenomenon was confirmed by Crandall and Hillis (1997), who demonstrated that some species of cave crayfish still possess a functional opsin with no significant differences from opsins isolated from surface species. As there is no light available, there would be little reason to maintain a traditional photoreceptor. They concluded that opsin must be involved in other pathways besides vision, although this remains highly speculative.
Conclusions
We have isolated and characterized the first LW opsin genes from two orders: Mecoptera and Siphonaptera.
Phylogenetic analysis implies a physiological sensitivity in the red-green spectrum for these opsins, consistent with previously identified spectral sensitivities.
Analysis of intron splice sites reveals the presence of two introns (190 and 239) conserved across the BG and LW groups. In addition, one intron (119) seems to be unique to LW opsins.
Flea opsin sequences showed a unique, four–amino acid–long expansion segment.
Calculated rates of evolution indicate that FL-Opsin, B-Opsin, and M-Opsin are evolving at similar rates with no radical selective pressures, implying conservative evolution and functional constraint.
The amino acid composition of FL-Opsin, B-Opsin, and M-Opsin is remarkably well preserved. Over 60% of the amino acid residues are identical across the three lineages, with an additional 22% that are under stabilizing, conservative selection. Among these are several key amino acid motifs that are necessary for proper opsin function. Panorpidae, Boreidae, and Siphonaptera exhibit vast differences in lifestyle and ecology that are reflected in the adaptations of the visual systems of each group. These lines of evidence indicate that despite the reduction in eye structure and loss of visual acuity, fleas have retained a remarkable similarity in LW opsin structure and amino acid properties. Although it is highly unlikely that fleas perceive visual images, our data support the possibility of fleas being sensitive to LW light. That the underlying mechanisms of photoreception could be preserved across visual systems of such immensely different qualities, speaks much of the flexibility, ease of adaptation, and potential of our biological world.
Supplementary Material
Supplementary data are available at Molecular Biology and Evolution online (www.mbe.oupjournals.org).
Acknowledgements
We gratefully acknowledge K. Miller, S. Cameron, and G. Svenson for their support and comments on this work. We are also grateful to D. McClellan for his invaluable insights and help with the TreeSAAP analysis, as well as to M. Pérez-Losada for his help with PAML. Thanks to J. Price for the initial design of opsin PCR primers, to M. Hastriter for his indefatigable knowledge of fleas and the use of his collection, and to R. Trowbridge for his photography. This research was funded by NSF grants DEB-9983195 (M.F.W.) and DEB-0120718 (M.F.W.), the Karl-Enigk-Foundation for Experimental Parasitology (Hannover, Germany), and the Brigham Young University Office of Research and Creative Activities.
References
Arikawa, K., E. Eguchi, A. Yoshida, and K. Aoki. 1980. Multiple extraocular photoreceptive areas on genitalia of butterfly, Papilio xuthus. Nature 288:700–702.
Arikawa, K. 1997. Hindsight by genitalia: photo-guided copulation in butterflies. J. Comp. Physiol. A. 180:295–299.
Bellingham, J., S. E. Wilkie, A. G. Morris, J. K. Bowmaker, and D. M. Hunt. 1997. Characterization of the ultraviolet-sensitive opsin gene in the honey bee, Apis mellifera. Eur. J. Biochem. 243:775–781.
Briscoe, A. D. 1999. Intron splice sites of Papilio glaucus PglRh3 corroborate insect opsin phylogeny. Gene 230:101–109.
———. 2000. Six opsins from the butterfly Papilio glaucus: molecular phylogenetic evidence for paralogous origins of red-sensitive visual pigments in insects. J. Mol. Evol. 51:110–121.
Briscoe, A. D., and L. Chittka. 2001. The evolution of color vision in insects. Annu. Rev. Entomol. 46:471–510.
Briscoe, A. D., and L. Nagy. 1999. Spatial expression of opsins in the retina and brain of the tiger swallotail Papilio glaucus. Am. Zool. 39:254B.
Burkhardt, D., and I. De LaMotte. 1972. Electrophysiological studies on the eyes of Diptera, Mecoptera, and Hymenoptera. Pp. 137–145 in R. Wehner, ed. Information processing in the visual systems of arthropods. Springer-Verlag, Berlin, Germany.
Byers, G. W., and R. Thornhill. 1983. Biology of the Mecoptera. Annu. Rev. Entomol. 28:203–228.
Carulli, J. P., D.-M. Chen, W. S. Stark, and D. L. Hartl. 1994. Phylogeny and physiology of Drosophila opsins. J. Mol. Evol. 38:250–262.
Crandall, K. A., and D. M. Hillis. 1997. Rhodopsin evolution in the dark. Nature (Lond.) 387:667–668.
Cronin, T. W., M. Jarvilehto, M. Weckstrom, and A. B. Lall. 2000. Tuning of photoreceptor spectral sensitivity in fireflies (Coleoptera: Lampyridae). J. Comp. Physiol. 186:1–12.
Crum, G. E., F. W. Knapp, and G. M. Wite. 1974. Response of the cat flea, Ctenocephalides felis (Bouché), and the oriental rat flea, Xenopsylla cheopis (Rothschild), to electromagnetic radiation in the 300–700 nanometer range. J. Med. Entomol. 11:88–94.
Deininger, W., M. Fuhrman, and P. Hegemann. 2000. Opsin evolution: out of the wild green yonder? Trends Genet. 16:158–159.
Dunnet, G. M., and D. K. Mardon. 1991. Siphonaptera. Pp. 705–716 in CSIRO, ed. The insects of Australia: a textbook for students and research workers. CSIRO and Cornell University Press, Melbourne, Australia.
GeneCodes. 2003. Sequencher. v4.2. Ann Arbor, Mich.
Guex, N., and M. C. Peitsh. 1997. SWISS-MODEL and the Swiss-Pdb viewer: an environment for comparative protein modeling. Electrophoresis 18:2714–2723.
Fleissner G., and G. Fleissner. 2003. Nonvisual photoreceptors with emphasis on their putative role as receptors of natural Zeitgeber stimuli. Chronobiology 20(4):593–616.
Hill, C. A., A. N. Fox, R. J. Pitts, L. B. Kent, P. L. Tan, M. A. Chrystal, A. Cravchik, F. H. Collings, H. M. Robertson, and L. F. Zwiebel. 2002. G protein-coupled receptors in Anopheles gambiae. Science 298:176–178.
Huelsenbeck, J. P., and N. S. Imennov. 2002. Geographic origin of human mitochondrial DNA: accommodating phylogentic uncertainty and model comparison. Syst. Biol. 51:155–165.
Huelsenbeck, J. P., B. Larget, R. E. Miller, and F. Ronquist. 2002. Potential applications and pitfalls of Bayesian inference phylogeny. Syst. Biol. 51:673–688.
Huelsenbeck, J. P., and F. Ronquist. 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17:754–755.
Kitamoto, J., K. Sakamoto, K. Ozaki, Y. Mishina, and K. Arikawa. 1998. Two visual pigments in a single photoreceptor cell: identification and histological localization of three mRNAs encoding visual pigment opsins in the retina of the butterfly Papilip xuthus. J. Exp. Biol. 201:1255–1261.
Kochendoerfer, G. G., S. W. Lin, T. P. Sakmar, and R. A. Mathies. 1999. How color visual pigments are tuned. Trends Biochem. Sci. 8:300–305.
Koehler, P. G., N. C. Leppla, and R. S. Patterson. 1989. Circadian rhythm of cat flea (Siphonaptera: Pulicidae) locomotion unaffected by ultrasound. J. Econ. Entomol. 82:516–518.
Land, M. F. 2003. Eyes and vision. Pp. 393–406 in V. H. Resh and R. T. Carde, eds. Encyclopedia of insects. Academic Press, New York.
Lewis, R. E., and J. H. Lewis. 1985. Notes on the geographical distribution and host preferences in the order Siphonaptera. J. Med. Entomol. 22:134–152.
Maddison, W. P., and D. R. Maddison. 2003. MacClade: analysis of phylogeny and character evolution. Sinauer Associates, Sunderland, Mass.
Malpel, S., A. Klarsfeld, and F. Rouyer. 2002. Larval optic nerve and adult extra-retinal photoreceptors sequentially associate with clock neurons during Drosophila brain development. Development 129:1443–1453.
Nathans, J. 1987. Molecular biology of visual pigments. Annu. Rev. Neurosci. 10:163–194.
Neufeld, T. P., R. W. Carthew, and G. M. Rubin. 1991. Evolution of gene position: chromosomal arrangement and sequence comparison of the Drosophila melanogaster and Drosophila virilis sina and Rh4 genes. Proc. Natl. Acad. Sci. USA 88:10203–10207.
Osbrink, W. L., and M. K. Rust. 1985. Cat flea (Siphonaptera: Pulicidae): factors influencing host-finding behavior in the laboratory. Ann. Entomol. Soc. Am. 78:29–34.
Palczewski, K., T. Kumasaka, T. Hori et al. (12 co-authors). 2000. Crystal structure of rhodopsin: a G protein-coupled receptor. Science 289:739–745.
Penny, N. D. 1977. A systematic study of the family Boreidae (Mecoptera). Kans. Univ. Sci. Bull. 51:141–217.
Penny, N. D., and G. W. Byers. 1979. A check-list of the Mecoptera of the world. Acta Amazon. 9:365–388.
Posada, D., and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14:817–818.
Resh, V. H., and R. T. Carde. 2003. Encyclopedia of insects. Academic Press, New York.
Ronquist, F., and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574.
Rust, M. K., and M. W. Dryden. 1997. The biology, ecology, and management of the cat flea. Annu. Rev. Entomol. 42:451–473.
Shimizu, I., Y. Yamakawa, Y. Shimazaki, and T. Iwasa. 2001. Molecular cloning of Bombyx cerebral ospin (Boceropsin) and cellular localization of its expression in the silkworm brain. Biochem. Biophys. Res. Commun. 287:27–34.
Spaethe, J., and A. D. Briscoe. 2004. Early duplication and functional diversification of the opsin gene family in insects. Mol. Biol. Evol. 21:1583–1594.
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24:4876–4882.
Towner, P., and W. Gartner. 1994. The primary structure of mantid opsin. Gene 143:227–231.
Whiting, M. F. 2002. Mecoptera is paraphyletic: multiple genes and phylogeny of Mecoptera and Siphonaptera. Zool. Scr. 31:93–104.
Woolley, S., J. Johnson, M. J. Smith, K. A. Crandall, and D. A. McClellan. 2003. TreeSAAP: selection on amino acid properties using phylogenetic trees. Bioinformatics 19:671–672.
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. K. Pedersen. 2000. Codon-substitution model for heterogeneous selection pressure at amino acid sites. Genetics 155:431–449.
Yasuyama, K., and I. Meinertzhagen. 1999. Extraretinal photoreceptors at the compound eye's posterior margin in Drosophila melanogaster. J. Comp. Neurol. 412:193–202.
Yokoyama, R., and S. Yokoyama. 1990. Convergent evolution of the red- and green-like visual pigment genes in fish, Astyanax facsiatus, and human. Proc. Natl. Acad. Sci. USA 87:9315–9318.
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.(Sean D. Taylor*, Katharin)
Correspondence: E-mail: kd98@email.byu.edu.
Abstract
Mecoptera and Siphonaptera represent two insect orders that have largely been overlooked in the study of insect vision. Recent phylogenetic evidence demonstrates that Mecoptera (scorpionflies) is paraphyletic, with the order Siphonaptera (fleas) nesting as sister to the family Boreidae (snow fleas), showing an evolutionary trend towards reduction in gross eye morphology within fleas. We provide the first molecular characterization of long-wavelength opsins from these three lineages (opsin gene from fleas [FL-Opsin], the Boreidae [B-Opsin], and a mecopteran family [M-Opsin]) and assess the effects of loss of visual acuity on the structure and function of the opsin gene. Phylogenetic analysis implies a physiological sensitivity in the red-green spectrum for these opsins. Analysis of intron splice sites reveals a high degree of similarity between FL-Opsin and B-Opsin as well as conserved splice sites across insect blue-green and long-wavelength opsins. Calculated rates of evolution and tests for destabilizing selection indicate that FL-Opsin, B-Opsin, and M-Opsin are evolving at similar rates with no radical selective pressures, implying conservative evolution and functional constraint across all three lineages.
Key Words: opsin ? evolution ? Mecoptera ? Siphonaptera ? insect vision
Introduction
Insects represent a group of organisms in which visual perception and the associated structural components are highly developed (Briscoe and Chittka 2001). The class Insecta is one of the largest and most diverse classes of organisms on the planet, accounting for over 70% of all known species (Resh and Carde 2003). Insects have adapted to nearly every ecological niche available, from arctic to tropic and mountaintop to river bottom. However, despite this wide diversity of habitats and lifestyles, finding links in the adaptation of visual systems to visual ecology has proven to be difficult (Briscoe and Chittka 2001). It is important, therefore, to continue to characterize the visual system for a diversity of insect groups in order to gain a better understanding of the selective pressures and constraints that guide visual evolution in insects.
The photoreceptive cells in the insect compound eye contain light-sensitive visual pigments, which consist of an opsin apoprotein, bound by a protonated Schiff's base linkage to a photosensitive chromophore molecule, usually 11-cis-retinal. The visual cascade is initiated when the chromophore absorbs a photon of light, causing it to isomerize to the trans form. A corresponding conformational change in the opsin protein, a member of the G protein–coupled receptors, initiates a transducin-mediated response, which ultimately sends a neural signal to the brain. For a review of opsin physiology, see Nathans (1987). A single individual may possess several opsin variants, each sensitive to different wavelengths of light, thus conferring a specific range of spectral sensitivity to that organism (Kochendoerfer et al. 1999). Typically, insects possess trichromatic vision, conferred by at least three opsin classes: UV-, blue-, and green-sensitive variants. All insects studied have green receptors (530 nm), most express UV (330 nm) and blue (440 nm), and only a few have been shown to contain red (>565 nm) (Briscoe and Chittka 2001). Molecular characterizations and spectral tuning studies have been performed on opsin proteins found in many insect groups, including bees (Bellingham et al. 1997), moths and butterflies (Kitamoto et al. 1998; Briscoe 2000), mantids (Towner and Gartner 1994), fireflies (Cronin et al. 2000), and fruit flies (Neufeld, Carthew, and Rubin 1991; Carulli et al. 1994).
Mecoptera and Siphonaptera represent two insect groups that have largely been overlooked in the study of vision. Mecoptera is a relatively small, holometabolous insect order with approximately 600 described species (Penny and Byers 1979). For purposes of this study, two families are of particular interest. Panorpidae (true scorpionflies) is the most speciose family of mecopterans. They have a scavenging lifestyle, feeding primarily on dead, soft-bodied arthropods and often robbing spiders' webs of the encased insects (Byers and Thornhill 1983). They have well-developed eyes and are highly visually oriented, with a maximum spectral sensitivity peak at 490–520 nm and a minor peak at 360 nm (Burkhardt and De LaMotte 1972). Boreidae (snow fleas) are wingless, alpine insects that live and feed exclusively on mosses and are most often collected on winter snow (Penny 1977). The eyes of boreids are smaller and appear slightly less well developed than those of panorpids. We have observed that boreids will often escape collection by jumping when they are too closely approached, indicating that vision plays an important role in sensing the environment. Little is known, however, of boreid visual ecology.
Siphonaptera (fleas) is a highly specialized holometabolous insect order with approximately 2,380 described species (R. E. Lewis and J. H. Lewis 1985). Fleas are strict vertebrate ectoparasites and are of tremendous economic importance as vectors of pathogens, transmitting diseases such as plague, murine typhus, and tularemia (Dunnet and Mardon 1991). Little is known about flea vision, due mainly to the radical divergence in flea eye structure from the typical insect bauplan (Crum, Knapp, and Wite 1974). Fleas show a transformation of the multifaceted eyes and ommatidia of most insects, replaced instead with heavily sclerotized, atypical ocelli, or "eyespots," or in some cases, a complete absence of any eye at all (Dunnet and Mardon 1991). The species used in our analysis exhibit a diversity of eyespot morphologies, ranging from large eyespots in Pulex irritans, to small eyespots in Ctenophthalmus agyrtes and Jellisonia sp., to no discernable eyespot in Uropsylla tasmanica. Despite the drastic reduction in eye structure, Crum, Knapp, and Wite (1974) demonstrated that cat fleas (Ctenophalides felis) and oriental rat fleas (Xenopsylla cheopis) have peak light sensitivities at 330 and 530 nm, respectively, although the light elicited different responses in the two species. Osbrink and Rust (1985) further demonstrated that visual cues seem to be among the primary factors influencing host selection in cat fleas. However, this eyespot probably cannot form an acute image but merely acts as a light sensor (Osbrink and Rust 1985; Rust and Dryden 1997; Land 2003).
Whiting (2002) recently demonstrated that Mecoptera is paraphyletic, with Siphonaptera nesting within Mecoptera as sister group to Boreidae, and the obscure family Nannochoristidae placed as sister to Boreidae + Siphonaptera (fig. 1). This phylogenetic hypothesis provides an evolutionary framework in which to evaluate shifts in visual ecology and search for possible correlates in morphological and molecular changes associated with vision in these insects. Given that nannochoristids are visual mecopterans, optimization of vision on this phylogeny implies that there was a reduction of visual acuity in the flea lineage. We compare opsin structure and function in the ectoparasitic fleas, which have a single-lens light sensor and only general light-dark sensitivity, with that of their boreid sister group, which are detritus feeders and visually oriented insects. These in turn are compared with their highly visual panorpid relatives, which have a fully developed, image-forming eye. Because nannochoristids are rare insects, we have not included them in this study.
FIG. 1.— Condensed phylogeny of Mecoptera and Siphonaptera presented by Whiting (2002) and eye morphology of Panorpidae, Boreidae, and Siphonaptera.
This study focuses on the effects of the shift to parasitism in fleas, the associated loss of visual acuity, and the molecular processes underlying the evolution of the opsin protein in these insects. Given the demonstrated photosensitivity of the fleas and scorpionflies to green light, we isolated the long-wavelength (LW) opsin gene from fleas (FL-Opsin), their close sister group, the Boreidae (B-Opsin), and the Panorpidae, a highly visual mecopteran family (M-Opsin). We provide the first molecular characterization of insect opsins from these three groups. In addition, we compare the gene structure and evolutionary history of these three opsin groups in order to assess the effects of loss of visual acuity on the structure and function of the opsin gene. We specifically test for relaxation of functional constraint on the FL-Opsin protein due to the gross reduction in eye structure and visual sensitivity.
Materials and Methods
DNA Analysis: Polymerase Chain Reaction and Sequencing
LW opsin sequences were obtained from four flea species representing four families, six panorpid species representing the single diverse genus Panorpa, and one boreid species. For a detailed list of all taxa included in the analysis, see Appendix 1 (Supplementary Material online).
Genomic DNA was extracted from specimens using the Qiagen DNeasy extraction kit using manufacturer's guidelines. Because of the low yield of DNA expected from fleas, the DNA was eluted in only 100 μl of solution. Genomic DNA vouchers and specimen vouchers are deposited in the Insect Genomics Collection, M. L. Bean Museum, Brigham Young University.
Polymerase chain reaction (PCR) primers were developed from conserved regions of arthropod LW opsins and, once sufficient sequences were obtained, redesigned for use in Mecoptera and Siphonaptera. The amplified region spans the first six of the transmembrane helices and 60% of the seventh. Primer sequences are given in table 1. PCR amplification used the following three-step cycling protocol with Amplitaq Gold polymerase (ABI, Foster City, Calif.): 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 3 min, followed by an additional 7 min of elongation at 72°C. PCR products were visualized by gel electrophoresis and purified using the Millipore Montage purification system. Initial amplifications were attempted on all taxa using LF1 and Scylla primers in order to obtain the entire target region in a single amplification. If unsuccessful, smaller overlapping segments were amplified with internal primers.
Table 1 Primers Used to Amplify LW-Opsin
Because opsin variants arose through a series of gene duplication events (Deininger, Fuhrman, and Hegemann 2000), the possibility exists that paralogous opsin variants may coamplify. In order to separate potential paralogous copies, the PCR products were cloned using the pCR 2.1-TOPO cloning kit (Invitrogen) following the manufacturer's protocol. Ten clones from each PCR product were picked and then reamplified and sequenced using the provided M13 vector primers (Invitrogen, Carlsbad, Calif.) and BigDye Terminator chemistry (ABI). Sequence data were collected using the ABI Prism 3730 capillary autosequencer in the BYU DNA Sequencing Center.
The vector ends were trimmed from all cloned products and the resulting sequence blasted in GenBank to insure that an opsin product was indeed amplified. Clones and overlapping PCR regions from the same taxa were initially included in the phylogenetic analysis as separate terminals to verify that the sequences did not represent paralogous gene copies. If they were shown to group together, the sequences were then combined, with any discrepancies coded as ambiguous or missing data, and included as a single terminal in the final analysis.
Phylogenetic Analysis
Sequences from crustaceans and the LW, blue-green (BG), blue (BL), and UV insect opsin variants were downloaded from GenBank and used to determine the phylogenetic placement of FL-Opsin, M-Opsin, and B-Opsin (see Appendix 1). The sequence of bovine rhodopsin, as the only opsin to have its tertiary structure characterized, was included as an out-group. Initial alignments were performed manually in Sequencher 4.2 (GeneCodes 2003). Introns were identified through comparison to GenBank insect cDNA sequences (fig. 2). The resulting coding sequences were translated using MacClade (W. P. Maddison and D. R. Maddison 2003) and the amino acid sequences aligned via ClustalX (Thompson et al. 1997). The alignment was further refined by performing a homology alignment of the Jellisonia sp. amino acid sequence (the most complete opsin isolated in this study) to high-resolution bovine opsin crystal structure (Palczewski et al. 2000) in the Swiss-Pdb Viewer and submitted to the SwissModel Server for homology modeling (Guex and Peitsh 1997). This allowed us to optimize our overall nucleotide alignment based on both amino acid sequence homology and structural homology.
FIG. 2.— Comparison of FL-Opsin, M-Opsin, and B-Opsin intron splice sites to other insect opsins, based on Briscoe (1999). Numbers above hatch marks indicate position of splice site relative to Drosophila melanogaster Rh1 amino acid sequence. Numbers in parentheses indicate observed size (bp) of introns (shown only for LW opsins). Regions in black were not amplified in this analysis. Intron positions and lengths determined from alignment of our novel opsin sequences with the following: Drosophila melanogaster Rh1 (X65877 [GenBank] ), Papilio glaucus Rh3 (AF098283 [GenBank] ), Bee LW1 (Bombus impatiens: AY485302 [GenBank] , Bombus terrestris: AY485301 [GenBank] , Drosophila afflicta: AY485303 [GenBank] , Drosophila rinconis: AY485304 [GenBank] , Osmia rufa: AY572828 [GenBank] ), Bee LW2 (B. impatiens: AY485306 [GenBank] , B. terrestris: AY485305 [GenBank] , D. afflicta: AY485308 [GenBank] , D. rinconis: AY485307 [GenBank] , O. rufa: AY572829 [GenBank] , Apis mellifera: XM_397398 [GenBank] .1), Anopheles LW (GPRop5, GPRop6, GPRop7; see Appendix 1, available as Supplementary Material online).
Phylogenetic analyses were performed using the nucleotide sequences after intron removal. The phylogeny was reconstructed using Bayesian methods coupled with Markov chain Monte Carlo (BMCMC) techniques as implemented in MrBayes v3.0b4, which allows for mixed model analyses (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). Because different codon positions have different structural constraints, the data set was partitioned into first, second, and third codon positions. Models for each partition were selected using the procedure implemented in Modeltest v.3.6 (Posada and Crandall 1998). Mixed models were used with unlinked parameters between partitions, treating model parameters as unknown variables with uniform default priors. Three independent BMCMC analyses were run, each consisting of 10 independent Markov chains started from a random tree and run for 3 x 106 generations, with every 1,000th generation sampled. To confirm that each separate Bayesian analysis converged and mixed well, the fluctuating value of the likelihood and all phylogenetic parameters were examined graphically. To check for congruence among independent runs, the mean likelihood score and parameter values and the posterior probabilities (pP) for individual clades were compared. All sample points prior to reaching stationary values were discarded as burn-in, and the remaining trees from each independent analysis were combined to calculate the maximum a posteriori (MAP) tree (Huelsenbeck and Imennov 2002; Huelsenbeck et al. 2002).
Opsin Evolution
The ratio of nonsynonymous versus synonymous changes (dN/dS) for the insect LW portion of the tree was estimated with maximum likelihood using the codon-based substitution model of the CODEML software package in PAML v3.14 (Yang 1998). Several different site-specific models in which selective pressure varies among different sites but the site-specific pattern is identical across all lineages were implemented (Yang et al. 2000): model M0 (null model with no variation among sites), M1 ("neutral" model, with two categories of site with fixed dN/dS ratios of 0 and 1), M2 ("selection" model with three categories of site, two with fixed dN/dS ratios of 0 and 1, and a third estimated dN/dS ratio), M3 ("discrete" model with three categories of site, where the dN/dS ratio is free to vary for each site), M7 ("beta" model—10 categories of site, with 10 dN/dS ratios in the range 0–1 taken from a discrete approximation of the beta distribution), and M8 ("beta plus omega" model—10 categories of site from a beta distribution as in model M7 plus an additional category of site with a dN/dS ratio that is free to vary from 0 to greater than 1). PAML estimates the dN/dS ratios that are free to vary under these models, as well as the proportion of sites with each ratio. All models were run twice with starting omega values of less than and greater than 1 as suggested by the PAML manual to test for entrapment in local optima (Yang 1997). Likelihood ratio tests, to determine whether more complex models provided a significantly better fit to the data than more simple models, were performed by comparing the likelihood ratio test statistic (–2[ln L1 – ln L2]) to critical values of the chi-square distribution with the appropriate degrees of freedom (Yang 1998).
Although dN/dS ratios are useful for detecting selection, they do not indicate how the identified selection affects the overall structure and function of the protein. Amino acid substitutions can have a wide range of effects on a protein depending on the difference in physicochemical properties and location in the protein structure; in order to evaluate these effects we used the insect LW portion of our Bayesian tree in TreeSAAP v3.0 (Woolley et al. 2003) to test for evidence of selection among 31 amino acid properties in the three lineages of interest. TreeSAAP allows us to identify these property changes and classify them into categories on a gradient from conservative to radical change. Based on data set–specific nucleotide substitution patterns, a neutral model of expected change for each category is generated. This null model is then compared to the observed numbers of amino acid replacements in the data set, and a z-score is calculated for each category. Categories where the observed number of amino acid replacements is significantly different than neutral expectations are considered as potentially being affected by selective pressures. In this study, we specified a scale of 20 categories of change. Greater than expected numbers of replacements (relative to the neutral model) in categories 1–3 indicate significant stabilizing selection, whereas the same situation in categories 18–20 indicate significant destabilizing selection (table 3). By mapping the identified amino acid replacements back onto the protein structure, we are able to make inferences about the influence of potential selective pressures on protein function and on the evolutionary history of particular lineages (Woolley et al. 2003).
Table 3 Conservative (Stabilizing) and Radical (Destabilizing) Amino Acid Properties Under Selection Identified from 31 Tested Amino Acid Properties in TreeSAAP Analysisa
Results and Discussion
Opsin Gene Isolation and Phylogenetic Analysis
PCR amplification of the LW opsin gene for the Mecoptera and Siphonaptera yielded 1,030- and 1,200-bp fragments, respectively. We were unable to obtain 460 bp from the 5' end of Panorpa gracilis and Boreus coloradensis. The translated sequences used in the analysis ranged between 203 and 278 amino acids. Interestingly, Siphonaptera have a unique four–amino acid–long expansion segment in transmembrane helix 6, at amino acid positions 119–222 in our alignment (fig. 3).
FIG. 3.— Two-dimensional representation of the homology model of FL-Opsin showing preservation of the seven-helix structure. Numbering begins with the first amino acid included in this analysis. Black circles represent amino acids not obtained. Boxed residues show an expansion region found only in FL-Opsin. Circles with a cross in them show key conserved residues described by Briscoe (1999). The results of the TreeSAAP analysis are mapped onto the model. Blue-colored circles show sites of conservative amino acid (AA) replacement, and red-colored circles show sites of radical AA replacement found along the branches leading to FL-Opsin, M-Opsin, or B-Opsin clades. Circles shaded in light gray indicate sites strictly conserved across the three lineages. White circles represent AA residue variation among the three lineages.
The MAP tree (ln L = –24,065.002 and pP = 0.002) used in PAML and TreeSAAP analyses is shown in figure 4. Although not entirely congruent with the species phylogeny presented by Whiting (2002), there are nonetheless several significant features about this topology. Briscoe (2000) demonstrated that opsins will cluster together primarily by physiological similarity, i.e., similar spectral sensitivities, and secondarily by species relationships. Our data strongly support these results, with opsins of physiological similarity clustering together across the topology. FL-Opsins and M-Opsins are well supported as monophyletic groups within the insect LW clade, suggesting sensitivity in the red-green spectrum. We note, however, that although support for the ordinal grouping of the FL-Opsins, M-Opsins, and B-Opsin sequences is strong, the opsin gene generally provides little or no support for the resolution of these ordinal relationships to each other.
FIG. 4.— Maximum a posteriori estimate of phylogeny (MAP) tree used in PAML and TreeSAAP analyses, ln L = –24,065.002 and pP = 0.002. Posterior probabilities >0.95 are indicated by a gray circle on the corresponding branch. A detailed list of all taxa names and their respective accession numbers can be found in Appendix 1. The shared insertion at site 156 between Boreidae, Siphonaptera, and bee LW2 is mapped onto the respective branches, indicating a gain (red rectangle) and a loss (white rectangle).
Furthermore, a recent study of the Anopheles gambiae genome revealed the presence of a higher number of opsin genes than any other characterized insect, with at least half of the identified copies being duplicates of LW-sensitive genes (Hill et al. 2002). Phylogenetic analysis of this LW gene complement indicated at least one early gene duplication within insects before the emergence of the orders Orthoptera, Mantodea, Hymenoptera, Lepidoptera, and Diptera (Spaethe and Briscoe 2004). In our analysis, the B-Opsin, FL-opsin, and M-opsin sequences do not cluster with either the A. gambiae or the bee LW2 genes (fig. 4). Analysis of our alignment using the methods described by Spaethe and Briscoe (2004) produced a tree similar to their results, with the newly characterized B-Opsin, FL-Opsin, and M-Opsin sequences grouping within the insect LW1 clade (results not shown). Although the weak support in the LW clade of our MAP tree makes inferences about opsin gene duplication and evolution difficult, the phylogenetic relationships of the opsin gene copies in A. gambiae and Hymenoptera indicate that there may be many more yet unidentified gene duplication events in the insect LW clade.
Opsin Gene Structure
Figure 2 shows comparative intron splice site positions of FL-Opsin, B-Opsin, and M-Opsin relative to other insect LW opsins. Within the regions amplified, the putative Siphonaptera opsin gene contains five introns, two Panorpidae, and four Boreidae. Interestingly, the B-opsin and FL-opsins (putative LW1) and bee LW2 share a unique intron at position 156, although they are not a monophyletic clade in our phylogeny. Manual mapping of this character onto the MAP tree (fig. 4) indicates a single gain of this intron site, with secondary loss in the Bee LW1 clade.
Briscoe (1999) reported three splice sites conserved across insect BG and LW opsin groups (sites 190, 239, and 332, fig. 2). Our data support this observation for both sites 190 and 239, although secondary intron deletions are observed in the bee LW1 (position 239) and Anopheles and Panorpidae (position 190) lineages. Site 332 is outside the region amplified in this study and cannot be further evaluated. Additionally, our data indicate a unique splice site at 119 that, with the exception of bee LW1 and Anopheles, appears to be characteristic of LW opsins, and we predict it will also be present in boreids.
Of the remaining intron insertions, sites 79 in Papilio and 108 in bee LW1 appear to be autapomorphies (fig. 2). Several sites (277, Papilio and bee LW1 and LW2; 296, Boreidae and Siphonaptera) appear in only a few taxa and may indicate their presence before the divergence of those lineages. All of the Anopheles LW opsin gene copies investigated (GPRop5, GPRop6, GPRop7) have lost all intron splice sites except 239, implying that patterns of intron insertion-deletion can change considerably within lineages.
Even though two groups may share a particular intron splice site, the length of the insert was observed to vary. For example, at position 296 the boreid has a long intron (537 bp) and the fleas have relatively short introns (58–84 bp). Intron lengths vary even among taxa within the same group, e.g., sequence lengths for position 119 in Jellisonia sp., C. agyrtes, and P. irritans are 99, 119, and 66 bp, respectively. These results indicate that the splice sites appear to be more phylogenetically conserved than the actual makeup of the intron.
Removal of introns reveals a preserved open reading frame in the boreid-, panorpid-, and flea-lineages. Homology modeling of the inferred amino acid sequence of the Jellisonia sp. FL-Opsin suggests that the seven transmembrane helices characteristic of G protein–coupled receptors and all known insect opsins are preserved (fig. 3). Additionally, Briscoe (1999) described seven amino acid motifs that are necessary for a functional opsin protein and that are conserved across all known insect opsins. Four of these motifs are located within the region amplified in this study, all four of which are conserved in FL-Opsin, B-Opsin, and M-Opsin (fig. 3). These residues include (1) Leu36 and (2) Asn41 found in all G protein–coupled receptors (numbered according to the first residue obtained in FL-Opsin); (3) the G protein–coupled receptor motif, E/D R, in the third transmembrane region involved in transducin activation (Asp102 and Arg103); and (4) a disulfide bridge connecting transmembrane region 3 with the second cytostolic loop at Cys78/155. As expected, many of these key residues occur in regions for which the amino acid sequence is conserved across all insect LW opsins.
Opsin Evolution
The results of the PAML analysis are summarized in table 2. For the majority of the comparisons made, we failed to reject the null hypothesis that models that allow for positive selection fit the data significantly better than models that do not. However, model M3 showed a significantly better score (P = 0.001) than the neutral model M1, indicating that selective pressure is present. Although model M3 fit the data better, no substitutions were found in categories where > 1, rejecting the hypothesis of diversifying selection. This result is confirmed by examining the individual rates of evolution found in M-Opsin, B-Opsin, and FL-Opsin and comparing them to the other lineages in the topology. If, for instance, in the flea lineage there was either strong selection away from the ancestral protein or a loss of functional constraint of the opsin protein due to the reduction in eye structure, we expect to see a greater number of nucleotide substitutions as well as a larger proportion of nonsynonymous to synonymous substitutions (Yokoyama et al. 1995). The dN/dS ratios on the branch leading to the FL-Opsin clade are not significantly different (P = 0.05) from those leading to the B-Opsin and M-Opsin clades (see table 2). These ratios are also nearly identical to the average rates of evolution across the entire topology (P = 0.05). In addition, all ratios are observed to be less than 1. As stated earlier, this indicates that any selective pressure on opsin evolution is purifying across all the insects observed such that the structural and functional properties of the protein are being conserved (Yang et al. 2000).
Table 2 Results from PAML Analysis of FL-Opsin, B-Opsin, and M-Opsin Lineages
Analysis in TreeSAAP also confirmed that a low rate of radical diversifying selection (destabilizing selection) was occurring with respect to amino acid properties. Only 9.6% of all sites showed destabilizing selection among 7 of the 31 properties tested. On the other hand, 21.4% of all sites are under strongly stabilizing selection for 22 of 31 amino acid properties tested, and 60% of the amino acid residues in FL-Opsin, B-Opsin, and M-Opsin were identical. Although several of the properties showing destabilizing selection (e.g., turn tendencies, coil tendencies, or power to be at the N-terminal) seem to be related to structural aspects of the protein, none are properties previously identified to affect spectral tuning (i.e., polarity). Additionally, all of the conserved residues as described by Briscoe (1999) also show conservation in our data set. These results are shown in figure 3 mapped onto the two-dimensional projection of FL-Opsin. See table 3 for a complete list of all properties tested. As a whole, these data not only imply functional constraint but also similar functional properties across FL-Opsin, B-Opsin, and M-Opsin.
Sensitivity to Topology
Because of the weak support of the ordinal relationships found in our analysis, the PAML and TreeSAAP analyses were also calculated after constraining the data to the Whiting (2002) species tree. The results in all cases were identical, suggesting that the data are relatively insensitive to organization of the ordinal relationships, as predicted by Yang et al. (2000). We observe that despite the overlying reduction in eye morphology and loss of visual sensitivity across these three lineages, there remains strong conservative, purifying selection on the opsin protein.
Functional Constraint of FL-Opsin, B-Opsin, and M-Opsin
Despite the considerable variation in lifestyle, habitat, and visual structure, the LW opsin gene appears to be remarkably well preserved across panorpids, boreids, and fleas. This is not entirely surprising, given that all three lineages demonstrate sensitivity and responsiveness to visual cues at some level. Perhaps with gross reduction in flea eye morphology and the incapability to perceive images, we might have expected to see some corresponding change in the mechanism of photoreception itself. However, as discussed above, not only does FL-Opsin appear to have retained functional constraint but also no evidence is found for unusual selective pressure on the amino acid sequence and structural properties. On the contrary, we observed a high degree of conservative selection, preserving amino acid properties and structure similar to other functional opsins. These results provide evidence that although vast differences in perceptual abilities exist between Panorpidae, Boreidae, and Siphonaptera, e.g., ability to form images, the underlying physiology involved in the mechanisms of photoreception appears to be preserved.
In addition, the preservation of opsin structure in Siphonaptera, combined with the apparent reduction of macroscopic eye structure may also indicate the presence of an extraretinal photoreceptor system. Indeed, preliminary transmission electron studies on flea eyespots indicate no preservation of typical microscopic adult insect eye structures (e.g., rhabdom, crystalline cone, corneal lens) as well as a heavily sclerotized layer of chitin covering the flea "eyes" (Dittmar de la Cruz, unpublished data). This structural degradation implies that photoreception may no longer occur within the eye. Extraretinal, nonvisual photoreceptors are common among insects and have been found at sites located in the central nervous system (Briscoe and Nagy 1999; Shimizu et al. 2001), the posterior margin of the compound eye (Yasuyama and Meinertzhagen 1999), and the peripheral abdominal segments (Arikawa 1997; see G. Fleissner and G. Fleissner 2003 for a review of nonvisual photoreceptors). A number of LW opsin copies have been implicated as extraretinal (Briscoe and Nagy 1999; Spaethe and Briscoe 2004), and Spaethe and Briscoe (2004) speculate that these extraretinal opsins evolved early in invertebrate evolution. Interestingly, similar to the peripheral photoreceptors found on the last abdominal segments (genitalia) of Papilio xuthus (Arikawa et al. 1980), fleas have a sensillial plate (or pygidium) on the last abdominal segment, which is packed with receptor-like structures. Unfortunately, the function of these "receptors" is still a source of speculation.
Furthermore, extraretinal photoreceptors appear to be involved in the photic entrainment of the circadian clock (Shimizu et al. 2001; Malpel, Klarsfeld, and Rouyer 2002). Koehler, Leppla, and Patterson (1989) demonstrated the presence of circadian rhythms in fleas, which could be mediated by these types of nonvisual receptors.
Other studies have also shown examples of organisms that do not have functional eye structures, yet seem to retain functional opsin proteins. R. Yokoyama and S. Yokoyama (1990), for example, presented a case where putative functional opsin genes were isolated from Astyanax fasciatus, a blind cave fish. This phenomenon was confirmed by Crandall and Hillis (1997), who demonstrated that some species of cave crayfish still possess a functional opsin with no significant differences from opsins isolated from surface species. As there is no light available, there would be little reason to maintain a traditional photoreceptor. They concluded that opsin must be involved in other pathways besides vision, although this remains highly speculative.
Conclusions
We have isolated and characterized the first LW opsin genes from two orders: Mecoptera and Siphonaptera.
Phylogenetic analysis implies a physiological sensitivity in the red-green spectrum for these opsins, consistent with previously identified spectral sensitivities.
Analysis of intron splice sites reveals the presence of two introns (190 and 239) conserved across the BG and LW groups. In addition, one intron (119) seems to be unique to LW opsins.
Flea opsin sequences showed a unique, four–amino acid–long expansion segment.
Calculated rates of evolution indicate that FL-Opsin, B-Opsin, and M-Opsin are evolving at similar rates with no radical selective pressures, implying conservative evolution and functional constraint.
The amino acid composition of FL-Opsin, B-Opsin, and M-Opsin is remarkably well preserved. Over 60% of the amino acid residues are identical across the three lineages, with an additional 22% that are under stabilizing, conservative selection. Among these are several key amino acid motifs that are necessary for proper opsin function. Panorpidae, Boreidae, and Siphonaptera exhibit vast differences in lifestyle and ecology that are reflected in the adaptations of the visual systems of each group. These lines of evidence indicate that despite the reduction in eye structure and loss of visual acuity, fleas have retained a remarkable similarity in LW opsin structure and amino acid properties. Although it is highly unlikely that fleas perceive visual images, our data support the possibility of fleas being sensitive to LW light. That the underlying mechanisms of photoreception could be preserved across visual systems of such immensely different qualities, speaks much of the flexibility, ease of adaptation, and potential of our biological world.
Supplementary Material
Supplementary data are available at Molecular Biology and Evolution online (www.mbe.oupjournals.org).
Acknowledgements
We gratefully acknowledge K. Miller, S. Cameron, and G. Svenson for their support and comments on this work. We are also grateful to D. McClellan for his invaluable insights and help with the TreeSAAP analysis, as well as to M. Pérez-Losada for his help with PAML. Thanks to J. Price for the initial design of opsin PCR primers, to M. Hastriter for his indefatigable knowledge of fleas and the use of his collection, and to R. Trowbridge for his photography. This research was funded by NSF grants DEB-9983195 (M.F.W.) and DEB-0120718 (M.F.W.), the Karl-Enigk-Foundation for Experimental Parasitology (Hannover, Germany), and the Brigham Young University Office of Research and Creative Activities.
References
Arikawa, K., E. Eguchi, A. Yoshida, and K. Aoki. 1980. Multiple extraocular photoreceptive areas on genitalia of butterfly, Papilio xuthus. Nature 288:700–702.
Arikawa, K. 1997. Hindsight by genitalia: photo-guided copulation in butterflies. J. Comp. Physiol. A. 180:295–299.
Bellingham, J., S. E. Wilkie, A. G. Morris, J. K. Bowmaker, and D. M. Hunt. 1997. Characterization of the ultraviolet-sensitive opsin gene in the honey bee, Apis mellifera. Eur. J. Biochem. 243:775–781.
Briscoe, A. D. 1999. Intron splice sites of Papilio glaucus PglRh3 corroborate insect opsin phylogeny. Gene 230:101–109.
———. 2000. Six opsins from the butterfly Papilio glaucus: molecular phylogenetic evidence for paralogous origins of red-sensitive visual pigments in insects. J. Mol. Evol. 51:110–121.
Briscoe, A. D., and L. Chittka. 2001. The evolution of color vision in insects. Annu. Rev. Entomol. 46:471–510.
Briscoe, A. D., and L. Nagy. 1999. Spatial expression of opsins in the retina and brain of the tiger swallotail Papilio glaucus. Am. Zool. 39:254B.
Burkhardt, D., and I. De LaMotte. 1972. Electrophysiological studies on the eyes of Diptera, Mecoptera, and Hymenoptera. Pp. 137–145 in R. Wehner, ed. Information processing in the visual systems of arthropods. Springer-Verlag, Berlin, Germany.
Byers, G. W., and R. Thornhill. 1983. Biology of the Mecoptera. Annu. Rev. Entomol. 28:203–228.
Carulli, J. P., D.-M. Chen, W. S. Stark, and D. L. Hartl. 1994. Phylogeny and physiology of Drosophila opsins. J. Mol. Evol. 38:250–262.
Crandall, K. A., and D. M. Hillis. 1997. Rhodopsin evolution in the dark. Nature (Lond.) 387:667–668.
Cronin, T. W., M. Jarvilehto, M. Weckstrom, and A. B. Lall. 2000. Tuning of photoreceptor spectral sensitivity in fireflies (Coleoptera: Lampyridae). J. Comp. Physiol. 186:1–12.
Crum, G. E., F. W. Knapp, and G. M. Wite. 1974. Response of the cat flea, Ctenocephalides felis (Bouché), and the oriental rat flea, Xenopsylla cheopis (Rothschild), to electromagnetic radiation in the 300–700 nanometer range. J. Med. Entomol. 11:88–94.
Deininger, W., M. Fuhrman, and P. Hegemann. 2000. Opsin evolution: out of the wild green yonder? Trends Genet. 16:158–159.
Dunnet, G. M., and D. K. Mardon. 1991. Siphonaptera. Pp. 705–716 in CSIRO, ed. The insects of Australia: a textbook for students and research workers. CSIRO and Cornell University Press, Melbourne, Australia.
GeneCodes. 2003. Sequencher. v4.2. Ann Arbor, Mich.
Guex, N., and M. C. Peitsh. 1997. SWISS-MODEL and the Swiss-Pdb viewer: an environment for comparative protein modeling. Electrophoresis 18:2714–2723.
Fleissner G., and G. Fleissner. 2003. Nonvisual photoreceptors with emphasis on their putative role as receptors of natural Zeitgeber stimuli. Chronobiology 20(4):593–616.
Hill, C. A., A. N. Fox, R. J. Pitts, L. B. Kent, P. L. Tan, M. A. Chrystal, A. Cravchik, F. H. Collings, H. M. Robertson, and L. F. Zwiebel. 2002. G protein-coupled receptors in Anopheles gambiae. Science 298:176–178.
Huelsenbeck, J. P., and N. S. Imennov. 2002. Geographic origin of human mitochondrial DNA: accommodating phylogentic uncertainty and model comparison. Syst. Biol. 51:155–165.
Huelsenbeck, J. P., B. Larget, R. E. Miller, and F. Ronquist. 2002. Potential applications and pitfalls of Bayesian inference phylogeny. Syst. Biol. 51:673–688.
Huelsenbeck, J. P., and F. Ronquist. 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17:754–755.
Kitamoto, J., K. Sakamoto, K. Ozaki, Y. Mishina, and K. Arikawa. 1998. Two visual pigments in a single photoreceptor cell: identification and histological localization of three mRNAs encoding visual pigment opsins in the retina of the butterfly Papilip xuthus. J. Exp. Biol. 201:1255–1261.
Kochendoerfer, G. G., S. W. Lin, T. P. Sakmar, and R. A. Mathies. 1999. How color visual pigments are tuned. Trends Biochem. Sci. 8:300–305.
Koehler, P. G., N. C. Leppla, and R. S. Patterson. 1989. Circadian rhythm of cat flea (Siphonaptera: Pulicidae) locomotion unaffected by ultrasound. J. Econ. Entomol. 82:516–518.
Land, M. F. 2003. Eyes and vision. Pp. 393–406 in V. H. Resh and R. T. Carde, eds. Encyclopedia of insects. Academic Press, New York.
Lewis, R. E., and J. H. Lewis. 1985. Notes on the geographical distribution and host preferences in the order Siphonaptera. J. Med. Entomol. 22:134–152.
Maddison, W. P., and D. R. Maddison. 2003. MacClade: analysis of phylogeny and character evolution. Sinauer Associates, Sunderland, Mass.
Malpel, S., A. Klarsfeld, and F. Rouyer. 2002. Larval optic nerve and adult extra-retinal photoreceptors sequentially associate with clock neurons during Drosophila brain development. Development 129:1443–1453.
Nathans, J. 1987. Molecular biology of visual pigments. Annu. Rev. Neurosci. 10:163–194.
Neufeld, T. P., R. W. Carthew, and G. M. Rubin. 1991. Evolution of gene position: chromosomal arrangement and sequence comparison of the Drosophila melanogaster and Drosophila virilis sina and Rh4 genes. Proc. Natl. Acad. Sci. USA 88:10203–10207.
Osbrink, W. L., and M. K. Rust. 1985. Cat flea (Siphonaptera: Pulicidae): factors influencing host-finding behavior in the laboratory. Ann. Entomol. Soc. Am. 78:29–34.
Palczewski, K., T. Kumasaka, T. Hori et al. (12 co-authors). 2000. Crystal structure of rhodopsin: a G protein-coupled receptor. Science 289:739–745.
Penny, N. D. 1977. A systematic study of the family Boreidae (Mecoptera). Kans. Univ. Sci. Bull. 51:141–217.
Penny, N. D., and G. W. Byers. 1979. A check-list of the Mecoptera of the world. Acta Amazon. 9:365–388.
Posada, D., and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14:817–818.
Resh, V. H., and R. T. Carde. 2003. Encyclopedia of insects. Academic Press, New York.
Ronquist, F., and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574.
Rust, M. K., and M. W. Dryden. 1997. The biology, ecology, and management of the cat flea. Annu. Rev. Entomol. 42:451–473.
Shimizu, I., Y. Yamakawa, Y. Shimazaki, and T. Iwasa. 2001. Molecular cloning of Bombyx cerebral ospin (Boceropsin) and cellular localization of its expression in the silkworm brain. Biochem. Biophys. Res. Commun. 287:27–34.
Spaethe, J., and A. D. Briscoe. 2004. Early duplication and functional diversification of the opsin gene family in insects. Mol. Biol. Evol. 21:1583–1594.
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24:4876–4882.
Towner, P., and W. Gartner. 1994. The primary structure of mantid opsin. Gene 143:227–231.
Whiting, M. F. 2002. Mecoptera is paraphyletic: multiple genes and phylogeny of Mecoptera and Siphonaptera. Zool. Scr. 31:93–104.
Woolley, S., J. Johnson, M. J. Smith, K. A. Crandall, and D. A. McClellan. 2003. TreeSAAP: selection on amino acid properties using phylogenetic trees. Bioinformatics 19:671–672.
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. K. Pedersen. 2000. Codon-substitution model for heterogeneous selection pressure at amino acid sites. Genetics 155:431–449.
Yasuyama, K., and I. Meinertzhagen. 1999. Extraretinal photoreceptors at the compound eye's posterior margin in Drosophila melanogaster. J. Comp. Neurol. 412:193–202.
Yokoyama, R., and S. Yokoyama. 1990. Convergent evolution of the red- and green-like visual pigment genes in fish, Astyanax facsiatus, and human. Proc. Natl. Acad. Sci. USA 87:9315–9318.
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.(Sean D. Taylor*, Katharin)