Skip to main content
NCBI home page
Journal of Heredity logoLink to Journal of Heredity
. 2014 Feb 20;105(3):407–415. doi: 10.1093/jhered/esu008

Clock Gene Evolution: Seasonal Timing, Phylogenetic Signal, or Functional Constraint?

Trevor J Krabbenhoft 1,✉, Thomas F Turner 1
PMCID: PMC3984439  PMID: 24558102

Abstract

Genetic determinants of seasonal reproduction are not fully understood but may be important predictors of organism responses to climate change. We used a comparative approach to study the evolution of seasonal timing within a fish community in a natural common garden setting. We tested the hypothesis that allelic length variation in the PolyQ domain of a circadian rhythm gene, Clock1a, corresponded to interspecific differences in seasonal reproductive timing across 5 native and 1 introduced cyprinid fishes (n = 425 individuals) that co-occur in the Rio Grande, NM, USA. Most common allele lengths were longer in native species that initiated reproduction earlier (Spearman’s r = −0.70, P = 0.23). Clock1a allele length exhibited strong phylogenetic signal and earlier spawners were evolutionarily derived. Aside from length variation in Clock1a, all other amino acids were identical across native species, suggesting functional constraint over evolutionary time. Interestingly, the endangered Rio Grande silvery minnow (Hybognathus amarus) exhibited less allelic variation in Clock1a and observed heterozygosity was 2- to 6-fold lower than the 5 other (nonimperiled) species. Reduced genetic variation in this functionally important gene may impede this species’ capacity to respond to ongoing environmental change.

Key words: circadian rhythms, climate change, Clock, comparative biology, molecular ecology, phenology

A nearly ubiquitous feature of long-lived organisms is the seasonal timing of life-history events, such as reproduction. For many species, seasonal timing has shifted in recent years as a result of human-induced alterations to environmental conditions (e.g., climate change) (Walther et al. 2002). Not surprisingly, the genetic mechanisms underlying seasonal timing have been a target of much recent empirical and theoretical work (Stinchcombe et al. 2004; Balasubramian et al. 2006; Tauber et al. 2007; Wilczek et al. 2010). Two key questions are 1) how does natural selection shape seasonal timing? and 2) do populations possess the necessary genetic variation to respond to environmental change? For example, 2 frequent responses of organisms include altering seasonal timing or shifting latitudinal (or altitudinal) distribution. Over evolutionary time, both responses will likely require genetic adaptation in circadian or circannual rhythms.

One approach for understanding the evolution of seasonal timing is through comparative study of members of a biological community in an explicit phylogenetic context. There are 3 main advantages to this approach. First, by considering phylogenetic relationships of members of a community, variation due to phylogenetic inertia or signal can be disentangled from variation due to adaptation. Second, communities represent a natural common garden “experiment,” wherein seasonal timing can be studied in organisms experiencing shared environmental conditions over evolutionary time. Third, in seasonally variable environments, temporal resource partitioning can be strong among species with high overlap in resource needs—potentially providing a strong signal of divergent evolution in genes underlying seasonal timing.

A gene that has engendered much recent interest in studies of climate change and seasonal timing (phenology) is Circadian Locomotor Output Cycles Kaput (Clock), a transcription factor and key constituent of the core circadian oscillator. Because of its role as a transcription factor in modulating circadian rhythms, Clock is a potential target for natural selection to shape daily and perhaps seasonal rhythms. For example, in rats, CLOCK may influence reproductive timing by binding to E-box elements in the promoter of a key reproduction gene, gonadotropin-releasing hormone receptor (Resuehr et al. 2007). The transcription-activating potential of CLOCK is determined by the length of a C-terminal polyglutamine repeat domain (PolyQ) present in most CLOCK proteins (Darlington et al. 1998). In Drosophila, a deletion in the PolyQ domain greatly reduces affinity of CLOCK for its downstream targets, and thus results in longer circadian periodicity (Darlington et al. 1998), similar to the pattern observed in mice mutants (King et al. 1997).

Recently, several studies have examined the role of allele length polymorphism in circadian genes in shaping latitudinal clines in migration and reproductive seasonality within species (Costa et al. 1991, 1992; Sawyer et al. 1997; Weeks et al. 2006). One of the first studies to tentatively link Clock and seasonal reproductive timing was by Leder et al. (2006), who mapped Clock to a quantitative trait locus in rainbow trout (Oncorhynchus mykiss) that explained up to 50% of the variance in spawning time in salmon. O’Malley and Banks (2008a) subsequently demonstrated that OtsClock1b PolyQ length increases with latitude in populations of Chinook salmon (Oncorhynchus tshawytscha) and correlates with migratory run and reproductive timing, whereas OtsClock1a is highly conserved among populations. Subsequent work on several different salmon species suggested variable selection on OtsClock1b length across species, which corresponded to the extent of latitudinal variation in reproductive or migratory timing (O’Malley et al. 2010; O’Malley, Cross et al. 2013; O’Malley, Jacobson et al. 2013).

There has been little consensus on the generality of the relationship of Clock gene variation and reproductive timing when compared across a diverse phylogenetic spectrum of organisms (Weeks et al. 2006; Johnsen et al. 2007; O’Malley and Banks 2008a; Liedvogel et al. 2009, 2012; Liedvogel and Sheldon 2010; O’Malley et al. 2010; Dor et al. 2011, 2012; O’Malley, Cross et al. 2013; O’Malley, Jacobson et al. 2013). Lack of generality in the relation between Clock and seasonal timing could be due to 3 factors: 1) the difficulty of detecting small, but significant, additive genetic variation effects on a quantitative character (Liedvogel et al. 2009); 2) the function of Clock may differ among taxa with different gene or genome-duplication histories due to sub-, neo-, or nonfunctionalization of paralogs; and 3) the physiological role of Clock may differ between endotherms versus ectotherms due to Clock being coregulated by temperature. One approach for potentially finding stronger evolutionary signals of divergent selection is through comparative studies involving deeper evolutionary splits (i.e., interspecifically), particularly in a community context with strong temporal partitioning.

Members of freshwater fish communities often exhibit large differences in seasonal timing of reproduction, perhaps as a means of partitioning resources and minimizing competition among larvae during the critical period of early life. In the Rio Grande in New Mexico, USA, members of the fish community exhibit partially overlapping, but distinct, spawning seasons (Turner et al. 2010; Krabbenhoft TJ et al., unpublished data). A key question is how human-induced changes to the environment (e.g., alteration of flow regimes through dams or earlier spring flooding via climate change) will affect the phenology of fish reproduction. In this study, we evaluated 3 potential explanations for variation in Clock1a in a community of 6 cyprinid fishes (Teleostei: Cyprinidae): 1) allele length variation corresponds to differences in reproductive phenology, 2) length variation is a result of phylogenetic inertia and reflects relationships among species, or 3) length variation reflects purifying selection for a certain circadian phenotype at the latitude where fishes were collected.

If Clock1a plays a role in determining reproductive timing in cyprinid fishes, then we would predict that earlier (colder) spawning species should have longer PolyQ domains. This prediction is based on the hypothesis that the greater affinity for downstream targets in longer alleles could result in shorter periodicity (King et al. 1997; Darlington et al. 1998), and thus, earlier seasonality (e.g., O’Malley and Banks 2008a; O’Malley, Cross et al. 2013; O’Malley, Jacobson et al. 2013). Alternatively, if Clock1a is not under divergent selection in these fishes but is instead under relaxed selective constraint, then length polymorphism should track phylogenetic relationships of these taxa. Finally, if Clock1a plays a role in nonreproductive circadian or circannual rhythms (e.g., adaptation to a particular latitude), then purifying selection should result in little variation among co-occurring species.

In addition to these hypotheses, we also tested whether the federally protected Rio Grande silvery minnow (Hybognathus amarus) has less genetic variation at Clock1a than the other 5 species. We based this hypothesis on independent evidence that Rio Grande silvery minnow experienced at least 1 extensive population bottleneck within the last century (Osborne et al. 2005, 2012; Turner et al. 2006), corresponding to contraction of its native range (by >90%; Bestgen and Propst 1996). In a broader context, the amount of standing genetic variation in functional genes, such as Clock1a, could play an important role in success or failure of species to adapt to future environmental change (e.g., climate change).

Materials and methods

Six species were included in this study: fathead minnow (Pimephales promelas), Rio Grande silvery minnow (H. amarus), red shiner (Cyprinella lutrensis), flathead chub (Platygobio gracilis), longnose dace (Rhinichthys cataractae), and common carp (Cyprinus carpio) (Table 1). All are native to the Rio Grande, except common carp, which is native to Asia, but was introduced to New Mexico around 1883 (Sublette et al. 1990). Fishes were collected with seine nets in the Rio Grande between Bernalillo and Los Lunas, NM, USA. This approximately 30 km stretch of river comprises a single, genetically connected population for each species (e.g., Turner et al. 2006). Individuals of all species except the endangered Rio Grande silvery minnow were sacrificed with an overdose of tricane methanesulfonate (MS-222). Rio Grande silvery minnow were anesthetized with 200mg/L MS-222 at the site of capture and a small piece of caudal fin from each individual was removed and preserved in 95% ethanol. Fish were allowed to recover in untreated river water prior to release. Total genomic DNA was isolated from caudal-fin tissues of all species via phenol–chloroform extraction (Hillis et al. 1996) for use in fragment-length analysis. Brains were dissected from a subset of individuals (n = 4 per species) and preserved in RNAlater (Ambion). RNA was extracted from brain tissues using TRIzol (Invitrogen) and converted to cDNA using Taqman Reverse Transcription Reagents (Applied Biosystems) following the manufacturer’s protocol. Complimentary DNA was used for DNA sequencing (see below).

Table 1.

Mean date of first spawn for Rio Grande cyprinids

Common name Species Mean Julian date, first appearance of YOY Standard deviation Julian date, protolarvae collection Rio Grande native?
Common carp Cyprinus carpio 126.4 20.6 Introduced
Rio Grande silvery minnow Hybognathus amarus 132.5 7.2 Native
Fathead minnow Pimephales promelas 138.5 21.2 Native
Flathead chub Platygobio gracilis 142.6 14.4 Native
Red shiner Cyprinella lutrensis 163.5 11.2 Native
Longnose dace Rhinichthys cataractae 193.5 NA Native
Open in a new tab

Dates reflect mean Julian date of first appearance of YOY fishes in samples from years 1995 and 2008–2010 (Krabbenhoft TJ et al., unpublished data). The standard deviation column refers to the estimate of standard deviation for Julian date of collection of protolarvae and is used as a proxy for length of spawning season. Standard deviation was not estimated for R. cataractae because of small numbers of protolarvae collected. NA, not available.

Fragment-Length Analysis

In cyprinids, Clock1a (sensu Wang 2008) is orthologous with both OtsClock1a and OtsClock1b of salmon, as a result of a salmon-specific genome-duplication event (O’Malley and Banks 2008b). Conversely, cyprinid Clock1b is not orthologous with either OtsClock1a or OtsClock1b. We focused this study on Clock1a because other Clock paralogs, Clock1b and NPAS2 (=Clock2), have much shorter PolyQ domains and inferred amino acid sequences are strongly conserved across cyprinid species, based on next-generation transcriptome sequences (Krabbenhoft TJ, Turner TF, unpublished data).

Polymerase chain reaction (PCR) primers (Clk-MF, 5′-CTC GGC TTT TGC ATG GCA ACC-3′ and Clk-KR, 5′-CTG TCK GAG CGA TGA GCT G-3′) were designed to amplify an approximately 250–290bp region of Clock1a for fragment-length analysis. NCBI’s primer BLAST tool and Primer3 (Rozen and Skaletsky 2000) were used to assess primer quality, specificity, and melting temperature. Importantly, both the Clk-MF and Clk-KR lie in the last exon of Clock1a and amplify genomic DNA without interruption of introns. Clock1a was amplified using touchdown PCR in 10 μL reactions (1× buffer, 2mM MgCl2, 125 μM deoxyribonucleotide triphosphates, 5 pmol of each primer, 0.375 units of Taq DNA polymerase, and approximately 375ng template DNA) with the following conditions: 95 °C × 3min, 20 cycles of 95 °C × 1min, 59 °C × 1min (decreasing by 0.5 °C each cycle), 72 °C × 1.5min, then 21 cycles of 95 °C × 1min, 53 °C × 1min, 72 °C × 1.5min, and a final extension of 72 °C × 10min. Fragment-length analysis for Clock1a was conducted using 5′ HEX dye–labeled Clk-MF primer (and unlabeled Clk-KR primer). PCR product (1.0 μL) was combined with 10.0 μL formamide and 0.4 μL size standard (Genescan 400HD [ROX]), denatured at 95 °C for 5min, and run out on an Applied Biosystems 3100 capillary sequencer. Fragment lengths were quantified using GeneMapper (Applied Biosystems). Based on these data, within-species mean allele length (MAL), most common allele (MCA), allele frequencies, number of unique alleles (A), and observed (H o) and expected (H e) heterozygosity were determined. Rarefaction-corrected allelic richness (A R) was also calculated using HP-RARE (Kalinowski 2005) with n = 50 genes per species. Deviation from Hardy–Weinberg Equilibrium (HWE) for genotypic data was assessed using Arlequin version 3.5 (Excoffier and Lischer 2010). Finally, heterozygosity (H e) was compared across species to test the hypothesis that the federally protected Rio Grande silvery minnow has less genetic variation at Clock1a than each of the other 5 species. Sampling variance of H e was determined by generating n random genotypes drawn, with replacement, from a pool of alleles in which allele frequencies were equal to observed allele frequencies, and where n is the number of individuals of a given species sampled. For each species, 9999 random values of H e were produced to generate a distribution of expected heterozygosity values, which were used to test the null hypothesis that H ei − H e(RGSM) = 0.

DNA Sequencing

In order to assess amino acid sequence evolution and verify that Clock1a length variation is due to polyglutamine indels, we sequenced a portion of Clock1a for each species. Representative samples (4 for each species, 24 total) that were homozygous for the MCA were selected for DNA sequencing using cDNA derived from brain tissue (see above). A second forward primer (Clk-IF, 5′-ATG TGG GGC AGT TAT GGT KC-3′) was paired with Clk-KR to amplify a ~520-bp portion of the 3′ end of Clock1a, encompassing the PolyQ domain. However, because these primers are complimentary to adjacent exons that span a large (>1000bp) intron, they are only useful for amplifying cDNA. PCR products were purified using the E.Z.N.A. Cycle Pure Kit (Omega Bio-tek, Inc.) and cycle sequenced using ABI Prism Big Dye Terminators, version 1.1, following the manufacturer’s protocol and run out on an Applied Biosystems 3100. Base calls were verified with Sequencher, version 4.9 (Gene Codes, Corp.), and sequences were aligned by ClustalW and by eye using BioEdit, version 7.0.9.0 (Hall 1999). The number of substitutions per synonymous (K s) and nonsynonymous (K a) site was assessed to test for codon-level selection in Clock1a (Li 1993), using the SeqinR package for R (www.R-project.org/). DNA sequences were deposited in GenBank (accession numbers KJ002541–KJ002552). In fulfillment of data archiving guidelines (Baker 2013), we have deposited all other primary data for this research with Dryad.

Phylogenetic Relationships

Phylogenetic relationships for the 6 study species, as well as the 2 other cyprinid fishes for which Clock1a sequence data are available (zebrafish, Danio rerio, and Somalian cavefish, Phreatichthys andruzzii) were inferred from complete mitochondrial cytochrome b sequences obtained from GenBank (accession numbers: NC002333.2, EU000493.1, NC001606.1, FJ744108.1, EU811100.1, AB070206.1, EU811097.1, and GQ275158.1). Cytochrome b was used because it is the only gene with sequence data currently available for all 8 taxa and is generally well suited for reconstructing phylogenetic relationships in cyprinids (Dowling et al. 2002). DNA sequences were aligned by eye, and maximum likelihood analysis was conducted in Mega5 (Tamura et al. 2011) using the Tamura and Nei (1993) model of sequence evolution allowing for gamma-distributed rate heterogeneity among nucleotide sites. Tree support was assessed with 1000 bootstrap replicates.

Reproductive Timing Data

Direct observation of spawning is not possible for fishes in turbid waters such as the Rio Grande. Instead we used the first appearance of larvae (young-of-year, YOY) as a proxy for onset of spawning season (e.g., Pease et al. 2006; Turner et al. 2010). Mean Julian date of first YOY appearance was calculated based on 231 samples of more than 19 000 specimens collected across 4 years in the study reach (Krabbenhoft 2012; Krabbenhoft TJ et al., unpublished data). First appearance of a species was averaged across sites and years to provide a robust estimate of reproductive seasonality. First appearances of species were typically protolarvae—the earliest ontogenetic stage after hatching, just after absorption of egg yolk and onset of free swimming. There is a short lag between spawning and first appearance of YOY. However, the duration of the lag is similar among species because they share similar morphological and developmental trajectories across early life stages. First appearance of YOY may be a trait of ecological significance in itself, as the bulk of mortality in cyprinids occurs in early life stages. Finally, this approach assumes that the order of spawning is consistent across years, which is the case in the Rio Grande (Turner et al. 2010; Krabbenhoft 2012; Krabbenhoft TJ et al., unpublished data).

Results

Fragment-Length Analysis: Within-Species Variation

Fragment-length analysis revealed little within-species variation in PolyQ allele length (Table 2). Allelic richness ranged from just 2 to 4 in the 6 cyprinid species, with the MCA comprising between 62.9% and 96.8% of alleles in each species. Observed heterozygosity varied greatly among species: from 0.063 in Rio Grande silvery minnow to 0.419 in longnose dace and was approximately 2- to 6-fold lower in Rio Grande silvery minnow than each of the other 5 species (Table 3). Similarly, expected heterozygosity was significantly lower in Rio Grande silvery minnow than 3 of the species (P-adjusted < 0.05; longnose dace, flathead chub, and red shiner) and nearly significantly lower than the other two (P-adjusted < 0.10; fathead minnow and common carp). Exact tests did not detect departure from HWE in any of the 6 species (P > 0.05 in all species).

Table 2.

Summary of sample sizes, allelic richness (A), rarefaction-corrected allelic richness (A R), allele lengths, and allele frequencies in 6 species of cyprinid fishes in this study

Species N A A R MAL MCA Allele frequency
254 260 263 269 272 278 287
RG silvery minnow 190 4 2.02 277.0 278 0.003 0.003 0.026 0.968
Fathead minnow 74 2 1.98 286.5 287 0.061 0.939
Red shiner 68 3 2.59 276.6 278 0.088 0.015 0.897
Flathead chub 37 2 2.00 277.3 278 0.122 0.878
Longnose dace 31 2 2.00 274.2 272 0.629 0.371
Common carp 25 2 2.00 254.6 254 0.900 0.100
Open in a new tab

Table 3.

Observed (H o) and expected (H e) heterozygosity in 6 Rio Grande cyprinid fishes, as well as results of statistical comparisons of Clock1a expected heterozygosity between federally endangered Rio Grande silvery minnow (H e(RGSM)) and 5 other cyprinids (H ei)

Species H o H e s 2 (H e) H ei − H e(RGSM) Adjusted-P (Benjamini–Hochberg)
RG silvery minnow 0.063 0.062 0.0003 — —
Fathead minnow 0.122 0.115 0.0014 0.053 0.098
Red shiner 0.206 0.189 0.0022 0.125 0.011
Flathead chub 0.243 0.217 0.0046 0.152 0.014
Longnose dace 0.419 0.474 0.0082 0.404 0.000
Common carp 0.200 0.180 0.0058 0.054 0.067
Open in a new tab

Fragment-Length Analysis: Among-Species Variation

Fragment-length analysis revealed substantially longer PolyQ domains in the native Rio Grande cyprinids, with MAL ranging from 274.2 to 286.5bp, compared with zebrafish and Somalian cavefish with 257 and 233bp alleles, respectively. Interestingly, common carp specimens collected in the Rio Grande are more similar in PolyQ length (MAL = 254.6) to zebrafish (an Asian native) than to native Rio Grande species. Thus, carp have not converged on PolyQ allele lengths found in native Rio Grande cyprinids but rather possess PolyQ lengths that are similar to other old-world cyprinids.

Phylogenetic Analysis of mtDNA

Results of maximum likelihood analysis of cytochrome b mtDNA sequences generally agree with previous studies of cyprinid phylogenetics (e.g., Dowling et al. 2002; Simons et al. 2003) (Figure 1, bottom). The 5 native North American species (fathead minnow, Rio Grande silvery minnow, red shiner, flathead chub, and longnose dace) were recovered as a monophyletic clade. Phylogenetic relationships recovered within the North American clade are concordant with those of Coburn and Cavender (1992) and Simons et al. (2003). Given the extensive phylogenetic distance between common carp and the North American taxa (over 120 million years divergence time; Saitoh et al. 2011), its tetraploid genome, and it being a recently introduced species in the Rio Grande, common carp were excluded from further analyses, as historical evolutionary factors shaping Clock genes are likely different in this species.

Figure 1.

Figure 1.

Open in a new tab

Top panel: Amino acid sequences of Clock1a PolyQ domain in cyprinid fishes and corresponding allele lengths (in bp). Bottom panel: Phylogenetic relationships of 8 cyprinid fishes in this study as inferred from maximum likelihood analysis of complete mtDNA cytochrome b sequences, including the 6 Rio Grande focal species. Values at nodes represent bootstrap support (1000 replicates). The most common PolyQ allele length is plotted next to each taxon.

Phylogeny, Allele Length, and Seasonal Timing

Mapping MCA for each taxon along the mtDNA phylogeny revealed strong phylogenetic signal in PolyQ allele length (Figure 1, bottom). For example, old-world taxa (zebrafish, carp, and Somalian cavefish) formed a monophyletic grouping and all have short MAL (range 233–254.6bp), whereas the North American species had longer MAL (range 274.2–286.5). Additionally, within the North American species, there was a trend of longer allele lengths in more recently derived species. Longnose dace, which is the most basal species, has the shortest MAL of the North American species, whereas fathead minnow and Rio Grande silvery minnow have the longest. Using zebrafish as an outgroup and assuming a stepwise mutation model, most parsimonious PolyQ allele length reconstructions revealed 10 increases in allele length and no decreases within the North American species. A binomial test revealed a significant evolutionary trend of increasing allele length in the North American species (2-tailed binomial test: n = 10, k = 10, P = 0.002), suggesting positive selection for longer alleles in derived species. PolyQ MAL was negatively correlated with reproductive timing in the 5 native cyprinids, although there was little statistical power associated with the test (MAL: Spearman’s r = −0.70, P = 0.23) (Figure 2). Earlier-spawning species have longer PolyQ domains than later-spawning species.

Figure 2.

Figure 2.

Open in a new tab

PolyQ MAL (circles) and MCA (triangles) versus mean date of first spawn in 5 native species of North American cyprinid fishes in this study. Linear regression results presented for MAL (dotted line) and MCA (solid line). There is a fairly strong, negative correlation between PolyQ MAL (Spearman’s r = −0.70) and first appearance of larvae.

Sequencing the Clock1a PolyQ Domain

DNA sequencing revealed strongly conserved Clock1a sequences in Rio Grande cyprinid fishes, particularly within and among the 5 native species (Figure 1, top). For the native species, 15 sites had synonymous nucleotide substitutions located throughout the gene, but nonsynonymous substitutions were only observed as insertions and/or deletions of glutamines in the PolyQ domain (Figure 1, top). Excluding glutamine indels in PolyQ, K a/K s = 0 in all pairwise comparisons, consistent with purifying selection on Clock1a.

Discussion

As researchers have raced to understand and predict effects of ongoing climate change, interest in phenology has burgeoned. A central question is how natural selection shapes reproductive phenology within and among species. Not surprisingly, internal clocks of organisms have become a strong focus of functional genetics studies (Sawyer et al. 1997; Balasubramian et al. 2006). Extensive progress has been made over the past decade toward understanding circadian pathways in laboratory and wild populations; however, the molecular correspondence of circadian and seasonal rhythms is poorly understood. This is somewhat surprising given a rich literature demonstrating that seasonal changes in photoperiod and temperature drive reproductive timing in ectotherms. In this study, we employed a comparative approach to elucidate mechanistic underpinnings of seasonal reproduction by studying evolutionary processes that shape a core circadian rhythm gene, Clock1a, within members of a fish community. We evaluated whether variation in the PolyQ domain in Clock1a is consistent with 1) among-species differences in reproductive timing, 2) phylogenetic inertia or signal, and/or 3) functional constraint (purifying selection). Importantly, these hypotheses are not mutually exclusive: the length and nucleotide sequence of Clock1a can be shaped by a combination of evolutionary processes. Study species exhibit temporal partitioning of YOY fish nursery habitat (Krabbenhoft 2012) that is critically important to survival, growth, and recruitment, and so we predicted a signal of divergent evolution of genes underlying seasonal timing differences across species.

We observed a general trend of longer PolyQ alleles in earlier-spawning species, consistent with patterns widely observed in salmon (O’Malley and Banks 2008a; O’Malley et al. 2010; O’Malley, Cross et al. 2013; O’Malley, Jacobson et al. 2013). In addition to differences in photoperiod, this pattern could be a result of seasonally varying water temperatures in the 2 systems. In the Rio Grande, earlier-spawning species reproduce in colder temperatures than later-spawning species. In Chinook salmon, the northern populations (which have longer PolyQ domains) presumably also reproduce in cooler temperatures (or lower degree days). It may be that in these ectothermic organisms, PolyQ allele length acts as a mechanism for temperature compensation. In ectotherms active in colder temperatures, longer PolyQ domains could compensate for temperature (and lower metabolic rates) by increasing the transactivation affinity of CLOCK for downstream targets (e.g., Period). Concordant with this hypothesis, in zebrafish, the amplitude of transcriptional activation by CLOCK is strongly temperature dependent (Lahiri et al. 2005). Future research is needed to determine how temperature and photoperiod interact to shape the evolution of Clock1a.

PolyQ Allele Length and Phylogeny

Mapping PolyQ allele length onto the mtDNA cytochrome b tree revealed obvious correspondence of phylogeny and PolyQ length, that is, phylogenetic signal. Native North American minnows have longer PolyQ domains than more ancestral, old-world cyprinids, and there is an evolutionary trend toward longer PolyQ domains in more recently derived North American cyprinid fishes. A binomial test was consistent with positive natural selection for longer allele length in more-derived taxa. Correspondence of PolyQ length, reproductive timing, and phylogeny suggests the intriguing possibility that phylogeny could drive differences in reproductive timing among species, although experimental manipulation is needed to test whether these correlations reflect causality.

Aside from variation in the number of glutamine repeats, all other amino acids were identical across native species, although 15 sites exhibited synonymous nucleotide substitutions (K a/K s = 0 in all pairwise comparisons). Amino acid conservation among species, despite abundant synonymous nucleotide variation and deep evolutionary history (time to most recent common ancestor ≈ 56.9 million years ago; Saitoh et al. 2011), strongly suggests purifying selection arising from functional constraint (Hurst 2002).

Comparative study of clock-gene variation across members of a biological community is an alternative to single species studies of local adaptation and geographic variation in seasonal timing (e.g., O’Malley and Banks 2008a). The comparative approach provides insight into deeper levels of evolutionary divergence and whether observed patterns are general or species specific. The phylogenetic perspective afforded by the comparative approach is particularly useful when combined with “common garden” conditions experienced by members of a community. The strengths of the comparative approach we outline are not available in individual-based, single species studies of local adaptation across a latitudinal gradient. A trade-off associated with the comparative approach is that a large number of species (e.g., >10) are required for sufficient statistical power for some tests, such as correlation analysis, and we acknowledge a need for replicating this study with additional species occurring in other rivers. Until much more data on Clock gene variation become available, we advise appropriate caution on conclusions reached.

Potential for Adaptation to Climate Change

Although there is little within-species variation in PolyQ allele length compared with putatively neutral microsatellites or mtDNA data (e.g., Alò and Turner 2005), standing levels of genetic variation at Clock1a could prove important for future adaptation to climate change and shifting seasonal or circadian rhythms. Rare alleles can facilitate successful evolutionary responses to changing adaptive landscapes, particularly when environmental change occurs more rapidly than the accumulation of beneficial new mutations. We tested the hypothesis that the federally protected Rio Grande silvery minnow has reduced amounts of genetic diversity due to recent population bottlenecks and range contraction. As predicted, observed heterozygosity in Rio Grande silvery minnow was substantially lower (2- to 6-fold) than the other 5 species, despite larger sample sizes. Similarly, expected heterozygosity was significantly lower in Rio Grande silvery minnow than in 3 common, native species in the system (longnose dace, flathead chub, and red shiner) and slightly lower than fathead minnow, a species with high interannual variability in abundance in the Rio Grande. Genetic variation in Rio Grande silvery minnow Clock1a was even marginally lower than that observed in the nonnative common carp, a species that has experienced trans-continental introductions (and perhaps genetic bottlenecks) over the past 2 centuries.

The observation of low genetic diversity in Rio Grande silvery minnow Clock1a is concordant with previous studies of genetic bottlenecks in this species based on putatively neutral genetic markers, such as mitochondrial DNA (Osborne et al. 2005, 2012; Turner et al. 2006). Alternatively, low heterozygosity in Rio Grande silvery minnow could be because spawning season length is driven by among-individual variation in Clock1a allele length. We favor the former hypothesis, as Rio Grande silvery minnow has the most synchronous spawning season among the study species, and spawning season length is only weakly positively related (r = 0.30) with observed heterozygosity across species.

In conclusion, this study is the first (to our knowledge) to assess the evolution of Clock in a comparative, community-based context. Our analyses demonstrate that several distinct evolutionary processes shape Clock1a variation. This research provides a framework for future (experimental) studies of temporal resource partitioning in biological communities. Regardless of whether Clock1a is important for reproductive timing, it is a functionally important gene that plays a central role in circadian rhythms and could be an important determinant in whether populations are able to respond to climate change. Particularly in fragmented rivers, rare Clock1a alleles may be precisely those required for future response to climate change and large populations may be needed to prevent those rare alleles from being lost in the population, such as from genetic drift.

Funding

National Science Foundation (DEB 0717047 to T.F.T.); National Institute of General Medical Sciences of the National Institutes of Health (P20GM103452).

Acknowledgments

We thank H. Wang for providing an amino acid alignment of Clock. This research benefitted from insight provided by K. O’Malley, C. Cunningham, S. Collins, C. Bertolucci, M. Osborne, S. Platania, and C. Krabbenhoft. G. Rosenberg and the UNM Biology Molecular Facility provided valuable technical assistance. We appreciate the support of S. Loker, R. Miller, and the Center for Evolutionary and Theoretical Immunology. We also thank Robin Waples and two anonymous reviewers for helpful comments on a previous version of the manuscript. J. Kawatachi assisted with data collection and was supported by the National Institutes of Health (NIH)–funded Initiatives to Maximize Student Diversity Program at the University of New Mexico (NIH NIGMS GM-060201). Samples were collected under New Mexico Department of Game and Fish permit #3015 and US Fish and Wildlife permit #TE038055-0. This research was approved by Institutional Animal Care and Use Committee Protocol #10-100468-MCC and #10-100492-MCC. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

References

  1. Alò D, Turner TF. 2005. Effects of habitat fragmentation on effective population size in the endangered Rio Grande silvery minnow. Cons Biol. 19:1138–1148 [Google Scholar]
  2. Baker CS. 2013. Journal of Heredity adopts joint data archiving policy. J Hered. 104:1. [DOI] [PubMed] [Google Scholar]
  3. Balasubramian S, Sureshkumar S, Agrawal M, Michael TP, Wessinger C, Maloof JN, Clark R, Warthmann N, Chory J, Weigel D. 2006. The PHYTOCHROME C photoreceptor gene mediates natural variation in flowering and growth responses of Arabidopsis thaliana . Nat Genet. 38:711–715 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bestgen K, Propst DL. 1996. Redescription, geographic variation, and taxonomic status of Rio Grande silvery minnow. Copeia. 1996:41–55 [Google Scholar]
  5. Coburn MM, Cavender TM. 1992. Interrelationships of North American cyprinid fishes. In: Mayden RL, editor. Systematics, historical ecology, and North American freshwater fishes. Palo Alto (CA): Stanford University Press; p. 328–373 [Google Scholar]
  6. Costa R, Peixoto AA, Barbujani G, Kyriacou CP. 1992. A latitudinal cline in a Drosophila clock gene. Proc R Soc Lond Ser B Biol Sci. 250:43–49 [DOI] [PubMed] [Google Scholar]
  7. Costa R, Peixoto AA, Thackeray JR, Dalgleish R, Kyriacou CP. 1991. Length polymorphism in the threonine-glycine-encoding repeat portion of the period gene in Drosophila . J Mol Evol. 32:238–246 [DOI] [PubMed] [Google Scholar]
  8. Darlington TK, Wager-Smith K, Fernanda Ceriani M, Staknis D, Gekakis N, Steeves TDL, Weitz CJ, Takahashi JS, Kay SA. 1998. Closing the circadian loop: CLOCK-induced transcription of its own inhibitors per and tim . Science. 280:1599–1603 [DOI] [PubMed] [Google Scholar]
  9. Dor R, Cooper CB, Lovette IJ, Massoni V, Bulit F, Liljesthrom M, Winkler DW. 2012. Clock gene variation in Tachycineta swallows. Ecol Evol. 2:95–105 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dor R, Lovette IJ, Safran RJ, Billerman SM, Huber GH, Vortman Y, Lotem A, McGowan A, Evans MR, Cooper CB, et al. 2011. Low variation in the polymorphic Clock gene poly-Q region despite population genetic structure across Barn Swallow (Hirundo rustica) populations. PLoS ONE. 6:e28843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dowling TE, Tibbets CA, Minkley WL, Smith GR. 2002. Evolutionary relationships of the plagopterins (Teleostei: Cyprinidae) from Cytochrome b sequences. Copeia. 2002:665–678 [Google Scholar]
  12. Excoffier L, Lischer HEL. 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour. 10:564–567 [DOI] [PubMed] [Google Scholar]
  13. Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser. 41:95–98 [Google Scholar]
  14. Hillis DM, Mable BK, Larson A, Davis SK, Zimmer EA. 1996. Chapter 9. Nucleic acids IV: sequencing and cloning. In: Hillis DM, Moritz C, Mable BK, editors. Molecular systematics. 2nd ed. Sunderland (MA): Sinauer Associates; p. 321–381 [Google Scholar]
  15. Hurst LD. 2002. The Ka/Ks ratio: diagnosing the form of sequence evolution. Trends Genet. 18:486–487 [DOI] [PubMed] [Google Scholar]
  16. Johnsen A, Fidler AE, Kuhn S, Carter KL, Hoffmann A, Barr IR, Biard C, Charmantier A, Eens M, Korsten P, et al. 2007. Avian Clock gene polymorphism: evidence for a latitudinal cline in allele frequencies. Mol Ecol. 16:4867–4880 [DOI] [PubMed] [Google Scholar]
  17. Kalinowski ST. 2005. HP-RARE 1.0: a computer program for performing rarefaction on measures of allelic richness. Mol Ecol Notes. 5:187–189 [Google Scholar]
  18. King DP, Zhao Y, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP, Steeves TDL, Hotz Vitaterna M, Kornhauser JM, Lowrey PL, et al. 1997. Positional cloning of the mouse circadian Clock gene. Cell. 89:641–653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Krabbenhoft TJ. 2012. Reproductive phenology in fishes of the Rio Grande, New Mexico: a genes-to-community approach [PhD Dissertation]. [Albuquerque (NM)]: University of New Mexico [Google Scholar]
  20. Lahiri K, Vallone D, Babu Gondi S, Santoriello C, Dickmeis T, Foulkes NS. 2005. Temperature regulates transcription in the zebrafish circadian clock. PLoS Biol. 3:2005–2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Leder EH, Danzmann RG, Ferguson MM. 2006. The candidate gene, Clock, localizes to a strong spawning time quantitative trait locus region in rainbow trout. J Hered. 97:74–80 [DOI] [PubMed] [Google Scholar]
  22. Li W-H. 1993. Unbiased estimation of the rates of synonymous and non-synonymous substitution. J Mol Evol. 36:96–99 [DOI] [PubMed] [Google Scholar]
  23. Liedvogel M, Cornwallis CK, Sheldon BC. 2012. Integrating candidate gene and quantitative genetic approaches to understand variation in timing of breeding in wild tit populations. J Evol Biol. 25:813–823 [DOI] [PubMed] [Google Scholar]
  24. Liedvogel M, Sheldon BC. 2010. Low variability and absence of phenotypic correlates of Clock gene variation in a great tit Parus major population. J Avian Biol. 41:543–550 [Google Scholar]
  25. Liedvogel M, Szulkin M, Knowles SCL, Wood MJ, Sheldon BC. 2009. Phenotypic correlates of Clock gene variation in a wild Blue Tit population: evidence for a role in seasonal timing of reproduction. Mol Ecol. 18:2444–2456 [DOI] [PubMed] [Google Scholar]
  26. O’Malley KG, Banks MA. 2008a. A latitudinal cline in the Chinook salmon (Oncorhynchus tshawytscha) Clock gene: evidence for selection on PolyQ length variants. Proc R Soc Lond Ser B Biol Sci. 275:2813–2821 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. O’Malley KG, Banks MA. 2008b. Duplicated Clock genes with unique polyglutamine domains provide evidence for nonhomologous recombination in Chinook salmon (Oncorhynchus tshawytscha). Genetica. 132:87–94 [DOI] [PubMed] [Google Scholar]
  28. O’Malley KG, Cross TF, Bailie D, Carlsson J, Coughlan JP, Dillane E. 2013. Circadian clock gene (OtsClock1b) variation and time of ocean return in Atlantic salmon Salmo salar . Fish Manage Ecol. 10.1111/fme.12058 [Google Scholar]
  29. O’Malley KG, Ford MJ, Hard JJ. 2010. Clock polymorphism in Pacific salmon: evidence for variable selection along a latitudinal gradient. Proc R Soc Lond Ser B Biol Sci. 277:3703–3714 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. O’Malley KG, Jacobson DP, Kurth R, Dill AJ, Banks MA. 2013. Adaptive genetic markers discriminate migratory runs of Chinook salmon (Oncorhynchus tshawytscha) amid continued gene flow. Evol Appl. 10.1111/eva.12095 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Osborne MJ, Benavides MA, Turner TF. 2005. Genetic heterogeneity among pelagic egg samples and variance in reproductive success in an endangered freshwater fish, Hybognathus amarus . Environ Biol Fish. 73:463–472 [Google Scholar]
  32. Osborne MJ, Carson EW, Turner TF. 2012. Genetic monitoring and complex population dynamics: insights from a 12-year study of the Rio Grande silvery minnow. Evol Appl. 5:553–574 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pease AA, Davis JJ, Edwards MS, Turner TF. 2006. Habitat and resource use by larval and juvenile fishes in an arid-land river (Rio Grande, New Mexico). Freshw Biol. 51:475–486 [Google Scholar]
  34. Resuehr D, Wildemann U, Sikes H, Olcese J. 2007. E-box regulation of gonadotropin-releasing hormone (GnRH) receptor expression in immortalized gonadotrope cells. Mol Cell Endocrinol. 278:36–43 [DOI] [PubMed] [Google Scholar]
  35. Rozen S, Skaletsky HJ. 2000. Primer3 on the WWW for general users and for biologist programmers. In: Krawetz S, Misener S, editors. Bioinformatics methods and protocols: methods in molecular biology. Totowa (NJ): Humana Press; p. 365–386 [DOI] [PubMed] [Google Scholar]
  36. Saitoh K, Sado T, Doosey MH, Bart HL, Jr, Inoue JG, Nishida M, Mayden RL, Miya M. 2011. Evidence from mitochondrial genomics supports the lower Mesozoic of South Asia as the time and place of basal divergence of cypriniform fishes (Actinopterygii: Osteriophysi). Zool J Linn Soc. 161:633–662 [Google Scholar]
  37. Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, Costa R, Kyriacou CP. 1997. Natural variation in a Drosophila clock gene and temperature compensation. Science. 278:2117–2120 [DOI] [PubMed] [Google Scholar]
  38. Simons AM, Berendzen PB, Mayden RL. 2003. Molecular systematics of North American phoxinin genera (Actinopterygii: Cyprinidae) inferred from mitochondrial 12S and 16S ribosomal RNA sequences. Zool J Linn Soc. 139:63–80 [DOI] [PubMed] [Google Scholar]
  39. Stinchcombe JR, Weinig C, Ungerer M, Olsen KM, Mays C, Halldorsdottir SS, Purugganan MD, Schmitt J. 2004. A latitudinal cline in flowering time in Arabidopsis thaliana modulated by the flowering time gene FRIGIDA. Proc Natl Acad Sci USA. 101:4712–4717 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sublette JE, Hatch MD, Sublette M. 1990. The fishes of New Mexico. Albuquerque (NM): University of New Mexico Press [Google Scholar]
  41. Tamura K, Nei M. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol. 10:512–526 [DOI] [PubMed] [Google Scholar]
  42. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol. 28:2731–2739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Tauber E, Zordan M, Sandrelli F, Pegoraro M, Osterwalder N, Breda C, Daga A, Selmin A, Monger K, Benna C, et al. 2007. Natural selection favors a newly derived timeless allele in Drosophila melanogaster . Science. 316:1895–1898 [DOI] [PubMed] [Google Scholar]
  44. Turner TF, Krabbenhoft TJ, Burdett AS. 2010. Reproductive phenology and fish community structure in an arid-land river system. In: Gido K, Jackson D, editors. Community ecology of stream fishes: concepts, approaches, and techniques. American Fisheries Society Symposium 73. Bethesda (MD): American Fisheries Society; p. 427–446 [Google Scholar]
  45. Turner TF, Osborne MJ, Moyer GR, Benavides MA, Alò D. 2006. Life history and environmental variation interact to determine effective population to census size ratio. Proc R Soc Lond Ser B Biol Sci. 273:3065–3073 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Walther G, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin J-M, Hoegh-Guldberg O, Bairlein F. 2002. Ecological responses to recent climate change. Nature. 416:389–395 [DOI] [PubMed] [Google Scholar]
  47. Wang H. 2008. Comparative analysis of teleost fish genomes reveals preservation of different ancient clock duplicates in different fishes. Mar Genom. 1:69–78 [DOI] [PubMed] [Google Scholar]
  48. Weeks AR, McKechnie SW, Hoffmann AA. 2006. In search of clinal variation in the period and clock timing genes in Australian Drosophila melanogaster populations. J Evol Biol. 19:551–557 [DOI] [PubMed] [Google Scholar]
  49. Wilczek AM, Burghardt LT, Cobb AR, Cooper MD, Welch SM, Schmitt J. 2010. Genetic and physiological bases for phenological responses to current and predicted climates. Philos Trans R Soc Lond Ser B Biol Sci. 365:3129–3147 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Heredity are provided here courtesy of Oxford University Press

RESOURCES