Parallel reduction in expression, but no loss of functional constraint, in two opsin paralogs within cave populations of Gammarus minus (Crustacea: Amphipoda)
© Carlini et al.; licensee BioMed Central Ltd. 2013
Received: 31 December 2012
Accepted: 16 April 2013
Published: 23 April 2013
Gammarus minus, a freshwater amphipod living in the cave and surface streams in the eastern USA, is a premier candidate for studying the evolution of troglomorphic traits such as pigmentation loss, elongated appendages, and reduced eyes. In G. minus, multiple pairs of genetically related, physically proximate cave and surface populations exist which exhibit a high degree of intraspecific morphological divergence. The morphology, ecology, and genetic structure of these sister populations are well characterized, yet the genetic basis of their morphological divergence remains unknown.
We used degenerate PCR primers designed to amplify opsin genes within the subphylum Crustacea and discovered two distinct opsin paralogs (average inter-paralog protein divergence ≈ 20%) in the genome of three independently derived pairs of G. minus cave and surface populations. Both opsin paralogs were found to be related to other crustacean middle wavelength sensitive opsins. Low levels of nucleotide sequence variation (< 1% within populations) were detected in both opsin genes, regardless of habitat, and dN/dS ratios did not indicate a relaxation of functional constraint in the cave populations with reduced or absent eyes. Maximum likelihood analyses using codon-based models also did not detect a relaxation of functional constraint in the cave lineages. We quantified expression level of both opsin genes and found that the expression of both paralogs was significantly reduced in all three cave populations relative to their sister surface populations.
The concordantly lowered expression level of both opsin genes in cave populations of G. minus compared to sister surface populations, combined with evidence for persistent purifying selection in the cave populations, is consistent with an unspecified pleiotropic function of opsin proteins. Our results indicate that phototransduction proteins such as opsins may have retained their function in cave-adapted organisms because they may play a pleiotropic role in other important processes that are unrelated to vision.
A large number of higher taxa contain species restricted to subterranean habitats, and most of these species show convergent differentiation in a suite of features related to the absence of light [1, 2]. These features, termed troglomorphy by Christiansen 1962 , include partial to complete loss of eyes and body pigment, elaborated extra-optic sensory structures, elongated appendages, and more slender body forms compared to related surface-dwelling taxa. The elaborated and elongated features are generally assumed to be adaptations to the aphotic environment from selection for enhanced extra-optic sensory performance and efficiency of movement, and are considered as textbook examples of evolution by means of natural selection. The reduction to loss of eyes and body pigment in subterranean species is part of a universal phenomenon of vestigialization to loss of structures common to the evolutionary history of most if not all phyletic lineages. Thus, the evolution of elaborated, and of even reduced features, is not exclusive to subterranean taxa, e.g., the loss of hearing in moths on islands without bats . The study of the evolution of troglomorphy in subterranean species is interesting because both elaborated and reduced features are exhibited by the same organism, have evolved independently and repeatedly across many taxa, and are clearly related to the consequences of a single overriding environmental constraint: total darkness.
The mechanism behind the evolutionary loss of eyes and body pigment is difficult to elucidate, not because it is difficult to explain, but because many of the proposed explanations are difficult to frame into testable hypotheses [1, 4, 5]. Hypotheses regarding the mechanism of such reductive evolution in subterranean species fall into three groups: non-adaptive, directly adaptive, and indirectly adaptive. The non-adaptive hypothesis invokes neutral mutation and drift and attributes no role to selection. Simply, features such as eyes and body pigment serve no function in darkness, thus mutations of structural loci underlying these traits are neutral and would accumulate over time, resulting in the breakdown of the feature. The direct adaptation hypothesis suggests that selection directly favors individuals with smaller eyes or reduced pigment per se in darkness, although it is difficult to envision the immediate selective advantage of smaller eyes or reduced pigment. Usually, such an advantage is couched in terms of energy savings by minimizing the cost of developing or maintaining a functionless structure. Fong 1989 , however, argued that this is simply part of the indirectly adaptive hypothesis because the selective advantage of the energy saved must come from spending the energy on enhancing the performance of other features in darkness. The non-adaptive and directly adaptive hypotheses assume independence of the sets of elaborated and reduced features. The indirectly adaptive hypothesis assumes a negative genetic correlation between the elaborated and reduced features through antagonistic pleiotropy, i.e., that evolutionary reduction in features is a correlated response to selection for elaborated features.
Recent studies have shed light on the relative roles of selection compared to neutral mutation and drift in the reduction to loss of features in cave organisms. Jones et al. 1992  used differences in morphology of paired, reproducing individuals with sexually matured but unpaired individuals to estimate the selection gradient of body size, eye size and antenna size in both surface and cave populations in the amphipod crustacean Gammarus minus. They showed that the intensity and direction of selection were similar among cave populations and similar among surface populations, but that selection favored smaller eyes in cave populations and larger eyes in spring populations. Cave species of the freshwater crab genus Sundathelthusa exhibit equally rapid rates of reduction in eye stalk length and cornea area as rates of increase in lengths of legs and setae, consistent with selection rather than neutral mutation as the cause of reduction of the optic structures . However, the reason why selection favors smaller eyes in cave populations of G. minus and of Sundathelthusa remains unclear. In the cavefish Astyanax mexicanus loss of pigment results from mutations at a single gene, oca2, and that different populations show different mutations . Further, Protas et al. 2007  showed that in A. mexicanus the polarities of QTLs are different for eyes and pigment systems. They interpret that consistent negative polarity of QTLs for eyes is evidence of selection, but that presence of both positive and negative polarities of QTLs for pigment is evidence of neutral mutation and genetic drift. In two species of cave-dwelling planthoppers, loss of melanin pigment occurred independently and resulted from a defect at the same step in the melanin synthesis pathway, suggesting an antagonistic pleiotropic effect of molecules involved in this step on melanin production and on an unidentified adaptive trait in darkness . Upregulation of the sonic hedgehog gene (shh) leads to eye degeneration in A. mexicanus, because increased expression of shh is correlated with enhanced development of taste buds [11, 12]. Thus, loss of eyes in this cavefish is a pleiotropic response to selection for increased taste bud development, a probable highly advantageous character in darkness. Niven and Laughlin  suggested that the high energetic cost of neural tissues favors reducing sensory structures to a functional minimum to save energy, especially in environments with severe energy limitation such as caves. As stated above, however, energy conservation hypotheses are indistinguishable from hypotheses based on pleiotropy, because any selective advantage from the energy saved can only be realized by spending the energy on improving the function of other feature(s) in darkness .
G. minus is an excellent model organism to study the evolution of troglomorphy for several reasons. First, cave populations show heritable, individual variation in eye size as well as antenna size . In contrast, in most other troglomorphic species all individuals lack any manifestation of eye and body pigment. Second, it is a single species with troglomorphic and non-troglomorphic populations occupying different habitats. This is important because it avoids the problem of inappropriate comparisons as most troglomorphic species are highly endemic with no close surface relatives . Third, hydrological and genetic evidence  clearly indicate that troglomorphic populations invaded cave streams from surface spring habitats, thus showing the troglomorphies are the derived condition. Finally, evidence from allozyme variation  and from molecular sequence data  strongly indicates that multiple troglomorphic populations have evolved troglomorphy independently.
We previously conducted a survey of 15 Gammarus minus populations through analysis of DNA sequences from one mitochondrial gene, cytochrome c oxidase I (COI), and one nuclear gene, internal transcribed spacer 1 (ITS-1) . Standing variation at both loci was low within populations (< 2% for COI, < 1% for ITS-1), but in accordance with the allozyme data  a significant degree of divergence and spatial structuring of populations was observed. Average COI sequence divergences between populations ranged from 0.4% to 10.2%, with hydrologically proximate populations exhibiting low levels of divergence, regardless of morphological differentiation. For example, the average divergence between the morphologically differentiated Second Creek Springs (SCS, surface morphology) and the hydrological proximate (3.5 km) Organ Cave (OC, cave morphology) populations was 1.43%, whereas the average divergence between SCS and the hydrologically distant (18.5 km) Davis Spring (DAS, surface morphology) populations was 4.65%, a greater than three-fold difference. A second pair of hydrologically connected populations, Maiden Spring (MDS, surface morphology) and Fallen Rock Cave (FRC, cave morphology), exhibited very low levels of genetic divergence at the ITS-1 locus (0.29%) despite significant morphological differentiation, and each of these two populations differed from all other populations by > 6% ITS-1 divergence. Therefore, different replicate pairs of surface and subterranean populations represent independent derivations of troglomorphy. As such, any concordant variation associated with troglomorphy in multiple, independently derived cave populations is much more likely to reflect a true response to selection than variation observed in only a single cave population.
Given that one of the most obvious morphological differences between surface and subterranean G. minus populations is in the eye (the other two being size and pigmentation), we embarked on a study of opsin visual protein evolution. In this study we characterize molecular genetic variation and patterns of gene expression in two paralogous opsin genes in four pairs of surface and subterranean G. minus populations to determine 1) if there is a loss of functional constraint in any of the cave lineages and 2) if there are differences opsin expression between surface and cave populations. Our study provides strong evidence for persistent purifying selection acting on the opsin genes of both the surface and cave populations, while at the same time documenting significantly reduced opsin expression in the cave populations.
Sampling locations, specimen collection, and preservation
Specimens were collected live from each of nine sampling locations in Tazewell County, Virginia, and Greenbrier and Jefferson counties, West Virginia, from Fall 2005 through Spring 2012. Sampling locations were selected to overlap with those from previous studies [14, 19, 20], and included four pairs of sister surface sites and cave sites: 1) Second Creek Springs (SCS) and Organ Cave (OC), 2) US219 Spring (US) and The Hole Cave (HC), 3) Davis Spring (DAS) and Persinger’s Cave (PC), 4) Maiden Spring (MDS) and Fallen Rock Cave (FRC), as well as Vidmont Spring (VS), a hydrologically remote surface site (Figure 1). The hydrologic connection between each pair of surface and cave sites was established by dye-tracing of the groundwater [21, 22]. Specimens were either preserved immediately in 100% ethanol or transported live in coolers back to the lab, where they were maintained at 4°C in incubators containing water from the sampling location until DNA or RNA extractions were performed, usually within 2 weeks of the collection date. Both surface and cave animals that were stored prior to nucleic acid extractions were exposed to approximately 12 hours of indirect dim light per day through the glass doors of the 4°C incubators; thus, individuals from sister population pairs were treated identically.
DNA extraction, PCR amplification, and sequencing
Approximately one fourth to one half of the posterior end of each specimen was excised for DNA extractions using clean razor blades. Qiagen DNeasy Blood and Tissue kits were used to extract genomic DNA with the following modifications of the manufacturer protocol: 30 μl of phosphate buffered saline (PBS) + 150 μl of buffer ATL were added prior to grinding the tissue for the initial step of the protocol, and a single elution was performed using 100 μl of buffer AE for the final step of the protocol.
Opsin genes were initially amplified from individuals from surface and cave populations using degenerate pairs of PCR primers designed to amplify opsin genes from a variety of crustaceans: LWF1A: 5′ -TGG TAY CAR TWY CCI CCI ATG AA -3′ , and Scylla: 5′-TT RTA IAC IGC RTT IGC YTT IGC RAA-3′ . The forward primer LWF1A is nested within the first transmembrane domain (TM1), while the reverse primer Scylla is nested within the seventh transmembrane domain (TM7) of opsin. PCR products were cloned, and ten clones from each of two individuals from one cave (OC) and one surface population (SCS) were sequenced. Based on these comparisons we identified two divergent paralogs of the opsin gene in G. minus which differed by >20% at the nucleotide sequence level within individuals in both cave and surface populations. We designed G. minus–specific opsin primers for each of the two paralogs, which we arbitrarily designated as opsin-1 (Ops1, 735 bp, excluding PCR primers) and opsin-2 (Ops2, 747 bp, excluding PCR primers). These species–specific primers were designed from regions where there were six to eight fixed differences between the Ops1 and Ops2 paralogs, 10–30 base pairs downstream from the degenerate primers: Ops1-F (5′–ATC ACC ATT CTA GGC ATC C–3′), Ops1-R (5′–TC CAG ATG GTC AGC AAG G–3′), Ops2-F (5′–CTT CTC GGA TTT GCT ATC TTC G–3′), Ops2-R (5′–CC CAG ATG GTA AAG AGC G–3′).
To facilitate direct sequencing of PCR products on the LI-COR 4300 DNA Analysis System, M13F (−21) tails were added to the 5′ primer and T7 tails were added to the 3′ primer. This allowed us to use IRDye® labeled M13F and T7 sequencing primers (LI-COR Biosciences) in the subsequent cycle sequencing reactions. Species specific opsin primers were used in 25 μl PCR amplifications under standard conditions (50–100 ng genomic DNA, 100 nM each primer, 200 μM each dNTP, 10 mM Tris–HCl, 50 mM KCl, 1.5 mM MgCl2, 2.5 U puRETaq polymerase). A total of 40 cycles of 95°C for 30 s, 47°C for 30 s, and 72°C for 30 s, were followed by a final extension at 72°C for 7 minutes. PCR products were electrophoresed on 0.8% agarose gels, excised, and purified using minElute PCR Purification kits (Qiagen).
PCR products were sequenced directly using M13F/IRD-800 and T7/IRD-700 labeled sequencing primers (LI-COR Biosciences). Both strands of DNA were sequenced with the SequiTherm EXCEL™ II DNA Sequencing Kit (Epicentre Biotechnologies), run on a LI-COR 4300 DNA Analysis System automated DNA sequencer, and read with the eSeq™ software (LI-COR Biosciences).
Nucleotide sequences were translated into their respective amino acid sequences. Given the low levels of sequence variation within and among populations, we obtained only five unique Ops1 and six unique Ops2 protein sequences. We downloaded arthropod opsin protein sequences from GenBank (Additional file 1) representing a spectrum of opsin absorbance, along with several onychophoran and cephalopod opsins that were selected as outgroups. Downloaded opsins were aligned with our 11 G. minus opsin protein sequences using ClustalW . Both alignments were uploaded to GenBank as population sets [GenBank Ops1 Accession Numbers: JX879564 - JX879705; GenBank Ops2 Accession Numbers: JX879485 - JX879563]. The best-fit model of protein evolution to use for subsequent maximum likelihood analyses was determined with ProtTest 3 . Maximum likelihood analysis of protein sequences was conducted in PhyML 3.0 . Clade support was determined through nonparametric bootstrapping (1000 replicates) in PhyML 3.0 using fixed model parameters obtained from the original data.
Nucleotide diversity (π) within populations and nucleotide divergence between populations were calculated in MEGA5 . The PopGen modules of BioPerl were used to calculate pairwise dN/dS ratios, Tajima’s D, and to conduct coalescent simulations to test the significance of Tajima’s D . McDonald-Kreitman tests were performed via the Standard and Generalized MKT website .
We employed a codon-based maximum likelihood approach to investigate potential variation in the dN/dS ratio over the branches of a phylogenetic tree using PAML . We used a fixed input tree topology that is consistent with what is known about the hydrology  and genetic structure of the G. minus populations based on allozyme  as well as nuclear and mitochondrial gene nucleotide sequences .
RNA extraction and quantitative real-time PCR
Quantitative real-time PCR (qRT-PCR) experiments were conducted to determine if opsin expression differed between the cave and surface populations. For an internal control gene, the 18S rRNA was amplified from G. minus individuals from several populations using degenerate “universal” primers. The 18S PCR products were cloned and sequenced, and a set of internal primers oriented to amplify a 440 bp fragment for qRT-PCR were designed (18SF: 5′–AGGAATTGACGGAAGGGC–3′; 18SR: 5′–GGACATCTAAGGGCATCACAG–3′). PCR products from two individuals each from one cave (OC) and one spring (SCS) population using these G. minus–specific 18S primers were sequenced and BLAST searched to confirm the identity of the 18S amplicon. A nested internal qRT-PCR forward primer was designed to amplify a ~300 bp fragment for the Ops1 gene, and a different internal forward primer was designed to amplify a ~300 bp fragment for the Ops2 gene. Ops1 and Ops2 qRT-PCR forward primers were designed based on the sequences obtained from the population-level analyses described above in a region where there was a fixed 10 bp difference between the two paralogs (qRT/Ops1F: 5′–T TCA TTG ATG TCA CAC AGC T–3′; qRT/Ops2F: 5′– A TCT TCT TCC TCT CAC TCC T–3′). We used the same Ops1 and Ops2 reverse primers for qRT-PCR as were used for the population sequencing (described above). PCR products using these opsin primers were sequenced from two individuals each from one cave (OC) and one spring (SCS) population to confirm the identity of the Ops1 and Ops2 amplicons.
Twenty fresh specimens from each of three pairs of hydrologically-connected populations were sampled for mRNA extractions: 1) SCS and OC, 2) HC and US, and 3) MDS and FRC. Due to the closure of PC during 2011–2012 to prevent the spread of white nose syndrome among bat populations, we were unable to obtain specimens and hence the DAS and PC population pair was not included for the gene expression analyses. Prior to mRNA extraction, individuals were photographed on both sides of the head at high resolution for subsequent measurements of eye and head surface areas. Total RNA was extracted from the heads (Ambion, RNAqueous for PCR) immediately following decapitation. To check for opsin expression outside of the head region, total RNA was extracted from the bodies of two individuals from each population. Total RNA was quantified with a microspectrophotometer (NanoVue, GE Healthcare Life Sciences), and 50 ng was used in each of two replicate oligo-dT primed cDNA synthesis reactions (AffinityScript QPCR cDNA Synthesis Kit, Agilent Technologies). Each cDNA was used to set up two replicate qRT-PCR reactions per gene using SYBR green as nonspecific fluorescent label and run on an Mx3005P QPCR System (Agilent Technologies). Ct values were calculated with the MxPro QPCR software v. 4.10, and dissociate curves for each reaction were checked to confirm that only a single PCR product was obtained. The relative expression of opsin was quantified as 2ΔCt, where ΔCt = (Ct18S – CtOps). We corrected for differences in eye size among individuals and between populations, by dividing relative expression by the proportion of eye tissue in the mRNA preparation. Because amphipods are laterally compressed, we estimated the proportion of eye tissue as the surface area of the eye divided by the total surface area of the head.
Results and discussion
Characterization of G. minusopsin proteins
Our results clearly show there are two opsin paralogs in the G. minus genome. These opsins are homologous to other crustacean long-wavelength sensitive (LWS) opsins. Although diverged from each other, the two paralogs are more closely related to each other than to any of the other crustacean opsins. They arose through a gene duplication event following their divergence from other crustacean LWS opsins. Because there are no other amphipod opsin sequences available in GenBank (as of December 2012), we cannot rule out the possibility that the duplication event preceded the speciation of G. minus.
Variation and dN/dS ratios in cave versus surface populations
Average % sequence divergence (uncorrected ρ) for pairwise comparisons within populations (diagonal elements in bold) and between populations
The ratio of nonsynonymous to synonymous substitutions, dN/dS, was calculated using the method of Stoletski and Eyre-Walker 2011 , which allows dN/dS to be calculated when dS = 0, for all pairwise comparisons within populations. For both Ops1 and Ops2, the average dN/dS ratio was < 1 for nearly all populations, the only exceptions being PC and MDS populations for Ops2, where the dN/dS ratio was very close to 1 (Table 1). No significant differences in the average pairwise dN/dS ratio of cave and surface populations were detected for Ops1 (cave = 0.17, surface = 0.35; P = 0.28, Welch’s two-tailed t-test), nor for Ops2 (cave = 0.61, surface = 0.47; P = 0.60). We also performed McDonald-Kreitman tests (MKTs) of selective neutrality for the four population pairs, and well as MKTs for each population versus VS. All MKTs for both Ops1 and Ops2 were not statistically significant. Taken together, these results indicate that neither opsin gene is under relaxed functional constraint in the cave populations.
Parameter estimate, likehoods, and likehood ration test (LRT) results for three models of coding sequence evolution of the Ops1 and Ops2 genes
One-Ratio, H0: ω0=ω1=ω2
Two-Ratio, H1: ω0=ω1≠ω2
Three-Ratio, H2: ω0≠ω1≠ω2
One-Ratio, H0: ω0=ω1=ω2
Two-Ratio, H1: ω0=ω1≠ω2
Three-Ratio, H2: ω0≠ω1≠ω2
There is no loss of functional constraint in the opsin genes of cave populations of G. minus, and this appears to be a common finding in cave-adapted organisms. Crandall and Hillis 1997  found no difference in the evolutionary rates of the rhodopsin gene between surface crayfish and eyeless crayfish species that differed in presumed durations of isolation in the subterranean habitat, and suggested that rhodopsin may be functional in these blind species. In the cave–adapted beetle Ptomaphagus hirtus highly reduced but persistent visual capacities were discovered by recovering transcripts of the phototransduction protein machinery in the adult head transcriptome , suggesting that the light detection machinery may play a role in regulation of circadian or other oscillatory processes. This beetle is similar to cave populations of G. minus in that it retains a residual lens structure in the degenerated eye and it is photonegative and thus able to detect light. Cave populations of G. minus (OC and PC) are also able to detect light but exhibit greatly reduced negative phototaxis compared to surface populations (SCS and DAS) . It is, however, possible that opsins in G. minus may play a direct or indirect role in regulating oscillatory processes as well, because the cave streams inhabited by G. minus experience seasonal influx of allochthonous organic detritus, which is assumed to be the ultimate source of energy in an aphotic environment in the absence of chemosynthesis . Teleost multiple tissue (TMT) opsin and melanopsin have been demonstrated to function as extra-retinal photoreceptors, and they may play a role in a peripheral clock in the blind Somalian Blind Cavefish Phreatichthys andruzzii. Similar to P. hirtus and G. minus, P. andruzzii also exhibits a phototactic response .
In Drosophila larvae, an opsin protein has recently been shown to function in thermodiscrimination . Interestingly, larvae containing a deletion of the opsin ninaE gene exhibited an impaired ability to discriminate between 18°C, their preferred temperature, and 24°C, whereas the thermodiscrimination phenotype was rescued upon addition of the wild-type ninaE transgene. In another recent study in Drosophila, four opsin paralogs were found to be expressed in the mechanosensory cells of the antennal ear and were required for hearing . Thus, we cannot rule out the possibility that a light-independent function of opsin is responsible for its conservation in cave populations.
It has also been suggested that genes or entire genetic pathways underlying the loss of structures may be maintained because they serve multiple functions or pathways, and can be re-expressed subsequently . Examples of the resurrection of “lost” structures include the evolution of the compound eye in myodocopid ostracod crustaceans , the induction of enamel organs in chick oral epidermis , and the development of wings in stick insects . Although a common theme, the lack of evidence for loss of functional constraint in the opsin genes of cave-adapted organisms is not universal. For example, a recent study of amblyopsid cavefishes documented at least three instances of loss of functional constraint due to indels and premature termination codons . However, the cavefish in this study are members of different species that have been diverging over the past 10 million years, a much longer time frame than the divergence of intraspecific sister population pairs in G. minus. It is possible that cave species may lose color-vision related opsins, while the other opsin paralog(s) remains conserved. In this case, one would predict that the LWS opsins would most likely remain conserved because they are maximally sensitive to low light levels.
Given the lack of evidence for loss of functional constraint in the cave populations, any observed differences between cave and surface populations in the level of opsin expression would be particularly interesting. For example, down-regulation of opsin expression in cave populations would be consistent with the notion that adaptive morphological evolution is due primarily to changes in regulatory regions, rather than changes in the coding sequences of structural genes, so as to minimize pleiotropic effects [47, 48].
Comparison of gene expression differences among sister populations and among individuals within populations
Source of variation
SRC vs OC, uncorrected
SRC vs OC, corrected for eye size
US vs. HC, uncorrected
MDS vs. FRC, uncorrected
MDS vs. FRC, corrected for eye size
SRC vs. OC, uncorrected
SRC vs. OC, corrected for eye size
US vs. HC, uncorrected
US vs. HC, corrected for eye size
MDS vs. FRC, uncorrected
MDS vs. FRC, corrected for eye size
When scaled for differences in eye surface area between the surface and cave populations, the difference in expression remained significant for the Ops1 gene in the SCS/OC comparison (P < 0.05), and for the Ops2 gene it was significant for both the SCS/OC (P < 0.001) and US/HC (P < 0.05) comparisons (Table 3). This suggests that the reduced expression in these cave populations is not simply the result of a smaller eye size (= fewer ommatidia) and thus fewer opsin expressing cells, but that these opsin genes may be differentially regulated in the surface and cave populations. However, the estimation of expression levels normalized by eye size is only approximate and may not be truly reflective of cell specific opsin expression. Compound eyes of cave populations appear disorganized compared that of surface populations. No organized rhabdomere, consisting of a lens and a rhabdom surrounded by a reflecting pigment shield, is observed in any of the cave populations. Different cave populations have lost a different set of these components, with the Organ Cave population having lost all of these components . The reduced opsin expression in cave populations, combined with a lack of evidence for loss of functional constraint, supports the view that adaptive evolution tends to proceed via changes in cis regulatory regions rather than via changes in the coding sequence of structural genes [47, 48].
The concordantly lowered expression level of both opsin genes in cave populations of G. minus compared to sister surface populations is consistent with an as of yet unspecified pleiotropic function of opsin proteins, which may require a threshold level of expression. Aspiras et al. (2012) also found a consistently lowered expression of the eye development gene hedgehog in the same cave populations of G. minus compared to the same sister surface populations, but no difference in expression of the upstream genes pax6, dachshund, and sine oculis. Their results echo the finding that upregulation of the shh gene leads to decreased eyes but increased taste bud development in the cave fish Asytanax mexicanus[11, 12]. They pointed out that although there was downregulation in an invertebrate system compared to upregulation in a vertebrate system, both involved the same homolog of shh.
The molecular signature of both opsin paralogs did not reveal any obvious relaxation of selective constraint, such as premature termination codons, indels, or elevated dN/dS ratios in the cave populations. Any or all of the three following explanations may account for our observations: 1) the sister populations have not been diverging long enough for such substitutions to accrue in the cave populations, 2) these opsin genes perform an alternate function unrelated to vision, and/or 3) these opsins retain some visual function in peripheral photoreceptors. Given the signature of persistent purifying selection, combined with the reduced but measureable expression of both paralogs in multiple independently derived cave populations, our results suggest that phototransduction proteins such as opsins may retain their function in cave-adapted organisms because they play an important role in peripheral photoreceptors or in other as of yet unidentified cellular processes and functions. In principle, these non-visual functions could be identified through experiments using RNAi to eliminate opsin expression in surface populations to determine what affect this would have on non-visual functions such as oscillatory behavior or temperature preference.
Our finding that there were differences in expression levels, but not in the coding sequences, of cave versus surface populations makes sense when considering the pleiotropic effects of mutations in coding versus regulatory regions. A mutation in a coding region of a gene will affect the function of the encoded protein in all tissues, whenever it is expressed. In contrast, a mutation in a single cis-regulatory element will only affect the expression of that gene in the domain under the control of that regulatory element. In this way, the diversification of individual regulatory elements allows expression to be modulated in a tissue– and/or temporal–specific manner without altering gene function. Given the lack of evidence for loss of functional constraint in both G. minus opsin genes, it appears possible that cis regulatory changes underlie the morphological differences between the cave and surface populations.
Our results are similar to the finding that the enzyme catalyzing reactions downstream of the defect in the melanin synthesis pathway in cave adapted insects remain functional because they also play an important role in the innate immune response and in wound healing in arthropods [50–52]. The reduced expression of opsin paralogs in cave populations, perhaps driven by increased hh expression, may be a pleiotropic consequence of selection for other cave-adapted structures similar to the pleiotropic effects of shh expression on eye size and taste bud development in cave fish [11, 12]. Interestingly, structures analogous to a lateral line system have been described from another species of Gammarus amphipod . Comparison of such structures in surface and cave populations of G. minus may point to a potential pleiotropic function of lowered opsin expression.
We thank Ben Hutchins, Katie Schneider, and Megan Porter for their assistance with specimen collection, and Ariel Aspiras, Mobola Oyefule, and Laura Lee for their assistance in the laboratory. We also thank Mara Laslo for her helpful advice and comments on a draft of our manuscript. This research was supported by a research grant from the Cave Conservancy of the Virginias awarded to DBC and DWF, as well as a College of Arts and Sciences Mellon Research Grant from American University awarded to DBC.
- Culver DC, Pipan T: The biology of caves and other subterranean habitats. 2009, New York: Oxford University PressGoogle Scholar
- Romero A: Cave biology. 2009, Cambridge: Cambridge University PressView ArticleGoogle Scholar
- Christiansen K: Proposition pour la classification des animaux cavernicoles. Spelunca. 1962, 2: 75-78.Google Scholar
- Fong DW, Culver DC, Kane TC: Vestigialization and loss of nonfunctional characters. Ann Rev Ecol Syst. 1995, 26: 249-268.View ArticleGoogle Scholar
- Fong DW: Gammarus minus as a model system for the study of adaptation. Encyclopedia of Caves. Edited by: Culver DC, White WB. 2012, Burlington MA: Elsevier Academic Press, 341-347.View ArticleGoogle Scholar
- Fong DW: Morphological evolution of the amphipod Gammarus minus in caves: quantitative genetic analysis. Amer Midl Naturalist. 1989, 121: 361-378. 10.2307/2426041.View ArticleGoogle Scholar
- Jones RD, Culver DC, Kane TC: Are parallel morphologies of cave organisms the result of similar selection pressures?. Evolution. 1992, 46: 353-365. 10.2307/2409856.View ArticleGoogle Scholar
- Klaus S, Mendoza JCE, Liew JH, Plath M, Meier R, Yeo DCJ: Rapid evolution of troglomorphic characters suggests selection rather than neutral mutation as a driver of eye reduction in cave crabs. Biol Lett. 2013, 9: 20121098-10.1098/rsbl.2012.1098.PubMed CentralPubMedView ArticleGoogle Scholar
- Protas ME, Hersey C, Kochanek D, Zhou Y, Wilkens H, Jeffery WR, Zon II, Borowsky R, Tabin CJ: Genetic analysis of cavefish reveals molecular convergence in the evolution of albinism. Nat Genet. 2006, 38: 107-111. 10.1038/ng1700.PubMedView ArticleGoogle Scholar
- Protas M, Conrad M, Gross JB, Tabin CJ, Borowsky R: Regressive evolution in the Mexican cave tetra, Astyanax mexicanus. Curr Biol. 2007, 17: 452-454. 10.1016/j.cub.2007.01.051.PubMed CentralPubMedView ArticleGoogle Scholar
- Jeffery WR: Adaptive evolution of eye degeneration in the Mexican blind cavefish. J Hered. 2005, 96: 185-196. 10.1093/jhered/esi028.PubMedView ArticleGoogle Scholar
- Jeffery WR: Evolution of eye degeneration in cavefish: the return of pleiotropy. Subterranean Biology. 2005, 3: 1-11.Google Scholar
- Niven JE, Laughlin SB: Energy limitation as a selective pressure on the evolution of sensory systems. J Exp Biol. 2008, 211: 1792-1804. 10.1242/jeb.017574.PubMedView ArticleGoogle Scholar
- Culver DC, Kane TC, Fong DW: Adaptation and natural selection in caves: the evolution of Gammarus minus. 1995, Cambridge MA: Harvard University PressView ArticleGoogle Scholar
- Glazier DS, Horne MT, Lehman ME: Abundance, body composition, and reproductive output of Gammarus minus (Crustacea: Amphipoda) in ten cold springs differing in pH and ionic content. Freshwater Biology. 1992, 38: 149-163.View ArticleGoogle Scholar
- Ford D, Williams P: Karst Hydrogeology and Geomorphology. 2007, New York: John Wiley and SonsView ArticleGoogle Scholar
- Holsinger JR, Culver DC: Morphological variation in Gammarus minus Say (Amphipoda, Gammaridae), with emphasis on subterranean forms. Postilla. 1970, 146: 1-24.Google Scholar
- Barr TC, Holsinger JR: Speciation in cave faunas. Ann Rev Ecol Syst. 1985, 16: 313-337. 10.1146/annurev.es.16.110185.001525.View ArticleGoogle Scholar
- Kane TC, Culver DC, Jones RT: Genetic structure of morphologically differentiated populations of the amphipod Gammarus minus. Evolution. 1992, 46: 272-278. 10.2307/2409822.View ArticleGoogle Scholar
- Carlini DB, Manning J, Sullivan PG, Fong DW: Molecular genetic variation and population structure in morphologically differentiated cave and surface populations of the freshwater amphipod Gammarus minus. Mol Ecol. 2009, 18: 1932-1945. 10.1111/j.1365-294X.2009.04161.x.PubMedView ArticleGoogle Scholar
- Sarbu SM, Kane TC, Culver DC: Genetic structure and morphological differentiation: Gammarus minus (Amphipoda, Gammaridae) in Virginia. Amer Midl Naturalist. 1993, 129: 145-152. 10.2307/2426443.View ArticleGoogle Scholar
- Jones WK: Hydrology of limestone karst in Greenbrier County, West Virginia. 1973, Morgantown, WV: West Virginia Geologic and Economic SurveyGoogle Scholar
- Chang BSW, Ayers D, Smith WC, Pierce NE: Cloning of the gene encoding honeybee long-wavelength rhodopsin: a new class of insect visual pigments. Gene. 1996, 173: 215-219. 10.1016/0378-1119(96)00165-5.PubMedView ArticleGoogle Scholar
- Taylor SD, Dittmar de la Cruz K, Porter ML, Whiting MF: Characterization of the long-wavelength opsin from Mecoptera and Siphonaptera: does a flea see?. Mol Biol Evol. 2005, 22: 1165-1174. 10.1093/molbev/msi110.PubMedView ArticleGoogle Scholar
- Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG: ClustalW and ClustalX version 2. Bioinformatics. 2007, 23: 2947-2948. 10.1093/bioinformatics/btm404.PubMedView ArticleGoogle Scholar
- Darriba D, Taboada GL, Doallo R, Posada D: ProtTest 3: fast selection of best-fit models of protein evolution. Bioinformatics. 2011, 27: 1164-1165. 10.1093/bioinformatics/btr088.PubMedView ArticleGoogle Scholar
- Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O: New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol. 2010, 59: 307-321. 10.1093/sysbio/syq010.PubMedView ArticleGoogle Scholar
- Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S: MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol. 2011, 28: 2731-2739. 10.1093/molbev/msr121.PubMed CentralPubMedView ArticleGoogle Scholar
- Stajich JE, Hahn MW: Disentangling the effects of demography and selection in human history. Mol Biol Evol. 2005, 22: 63-73.PubMedView ArticleGoogle Scholar
- Egea R, Casillas S, Barbadilla A: Standard and generalized McDonald-Kreitman Test: a website to detect selection by comparing different classes of DNA sites. Nucleic Acids Res. 2008, 36: W157-W162. 10.1093/nar/gkn337.PubMed CentralPubMedView ArticleGoogle Scholar
- Yang Z: PAML: A program package for phylogenetic analysis by maximum likelihood. Appl Biosci. 2007, 13: 555-556.Google Scholar
- Porter ML, Bok MJ, Robinson PR, Cronin TW: Molecular diversity of visual pigments in Stomatopoda (Crustacea). Vis Neurosci. 2009, 26: 255-65. 10.1017/S0952523809090129.PubMedView ArticleGoogle Scholar
- Stoletzki N, Eyre-Walker A: Estimation of the neutrality index. Mol Biol Evol. 2010, 28: 63-70.PubMedView ArticleGoogle Scholar
- Bielawski JP, Yang Z: Maximum likelihood methods for detecting adaptive protein evolution. Statistical Methods in Molecular Evolution. Edited by: Nielsen R. 2005, New York, NY: Springer, 103-124.View ArticleGoogle Scholar
- Crandall KA, Hillis DM: Rhodopsin evolution in the dark. Nature. 1997, 387: 667-668. 10.1038/42628.PubMedView ArticleGoogle Scholar
- Friedrich M, Chen R, Daines B, Bao R, Caravas J, Rai PK, Zagmajster M, Peck SB: Phototransduction and clock gene expression in the troglobiont beetle Ptomaphagus hirtus of Mammoth cave. J Exp Biol. 2011, 214: 3532-3541. 10.1242/jeb.060368.PubMedView ArticleGoogle Scholar
- Vawter AT, Fong DW, Culver DC: Negative phototaxis in surface and cave populations of the amphipod Gammarus minus. Stygologia. 1987, 3: 83-88.Google Scholar
- Cavallari N, Frigato E, Vallone D, Fröhlich N, Lopez-Olmeda JF, Foà A, Berti R, Sánchez-Vázquez FJ, Bertolucci C, Foulkes NS: A blind circadian clock in cavefish reveals that opsins mediate peripheral clock photoreception. PLoS Biol. 2011, 9: e1001142-10.1371/journal.pbio.1001142.PubMed CentralPubMedView ArticleGoogle Scholar
- Tarttelin EE, Frigato E, Bellingham J, Di Rosa V, Berti R, Foulkes NS, Lucas RJ, Bertolucci C: Encephalic photoreception and phototactic response in the troglobiont Somalian blind cavefish Phreatichthys andruzzii. J Exp Biol. 2012, 215: 2898-2903. 10.1242/jeb.071084.PubMed CentralPubMedView ArticleGoogle Scholar
- Shen WL, Kwon Y, Adegbola AA, Luo J, Chess A, Montell C: Function of rhodopsin in temperature discrimination in Drosophila. Science. 2012, 331: 1333-1336.View ArticleGoogle Scholar
- Senthilan PR, Piepenbrock D, Ovezmyradov G, Nadrowski B, Bechstedt S, Pauls S, Winkler M, Möbius W, Howard J, Göpfert MC: Drosophila auditory organ genes and genetic hearing defects. Cell. 2012, 150: 1042-1054. 10.1016/j.cell.2012.06.043.PubMedView ArticleGoogle Scholar
- Marshall CR, Raff EC, Raff RA: Dollo’s law and the death and resurrection of genes. Proc Natl Acad Sci. 1994, 91: 12283-12287. 10.1073/pnas.91.25.12283.PubMed CentralPubMedView ArticleGoogle Scholar
- Oakley TH: On homology of arthropod compound eyes. Integr Comp Biol. 2003, 43: 522-530. 10.1093/icb/43.4.522.PubMedView ArticleGoogle Scholar
- Kollar EJ, Fisher C: Tooth induction in chick epithelium: expression of quiescent genes for enamel synthesis. Science. 1980, 207: 993-995. 10.1126/science.7352302.PubMedView ArticleGoogle Scholar
- Whiting MF, Bradler S, Maxwell T: Loss and recovery of wings in stick insects. Nature. 2003, 421: 264-267. 10.1038/nature01313.PubMedView ArticleGoogle Scholar
- Niemiller ML, Fitzpatrick BM, Shah P, Schmitz L, Near TJ: Evidence for repeated loss of selective constraint in rhodopsin of amblyopsid cavefishes (Teleostei: Amblyopsidae). Evolution. 2012, 67: 732-748.PubMedView ArticleGoogle Scholar
- Carroll SB: Evolution at two levels: on genes and form. PLoS Biol. 2005, 3: 1159-1166.Google Scholar
- Davidson EH, Erwin DH: Gene regulatory networks and the evolution of animal body plans. Science. 2006, 311: 796-800. 10.1126/science.1113832.PubMedView ArticleGoogle Scholar
- Aspiras AC, Prasad R, Fong DW, Carlini DB, Angelini DR: Parallel reduction in expression of the eye development gene hedgehog in separately derived cave populations of the amphipod Gammarus minus. J Evol Biol. 2012, 25: 995-1001. 10.1111/j.1420-9101.2012.02481.x.PubMedView ArticleGoogle Scholar
- Bilandzija H, Cetkovic H, Jeffery WR: Evolution of albinism in cave planthoppers by a convergent defect in the first step of melanin biosynthesis. Evol Dev. 2012, 14: 196-203. 10.1111/j.1525-142X.2012.00535.x.PubMedView ArticleGoogle Scholar
- Cerenius L, Lee BL, Soderhall K: The proPO-system: pros and cons for its role in invertebrate immunity. Trends Immunol. 2008, 29: 263-271. 10.1016/j.it.2008.02.009.PubMedView ArticleGoogle Scholar
- Kuballa AV, Elizur A: Differential expression of profiling of components associated with exoskeleton hardening in crustaceans. BMC Genomics. 2000, 9: 575-View ArticleGoogle Scholar
- Platvoet D, Song Y, Shuqiang L, Van der Velde G: Description of the lateral line organ of Dikerogammarus villosus (Sowinsky, 1894), with discussion on its function (Peracarida, Amphipoda). Crustaceana. 2007, 80: 1373-1392. 10.1163/156854007782605619.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.