Transcriptome analysis of the circadian regulatory network in the coral Acropora millepora.
Broadcast spawning corals utilize an extraordinarily accurate timing system to reproduce. Most species spawn on just one or two evenings per year in a time window only 20-30 min long (e.g., Willis et al., 1985; Hagman et al., 1998; Baird et al., 2002; Hughes et al., 2000; Guest et al., 2005). Although spawning in the Pacific is somewhat variable, in the Caribbean timing is extremely consistent; for some species these time windows can be predicted to within 10 min from year to year by taking into account the month, latitude, lunar phase, and solar cycle (Hagman et al, 1998; Vize et al., 2005). This extraordinary temporal fidelity is essential to achieve high fertilization rates: corals that spawn too early or too late have vastly lower rates of reproductive success due to limited spans of gamete viability and the effects of dilution (Levitan et al., 2004). The month of coral spawning is set by cycles in solar insolation and local weather patterns (van Woesik et al., 2006). Spawning occurs in July in the northern Atlantic, in August in the northern Caribbean, and in September in the southern Caribbean. The evening of the spawn is set by the lunar cycle (e.g., Harrison et al., 1984; Willis et al., 1985; Bab-cock et al., 1986; Oliver et al., 1988). In the Caribbean, most scleractinians, along with a number of other reef invertebrates, spawn on the 8th evening after the August full moon (e.g., van Veghel, 1993; Hagman and Vize, 2003; Levitan et al., 2004). Although very consistent from year to year in the Caribbean, the date of spawning is less predictable in other locations, including the Great Barrier Reef (e.g., Willis et al., 1985). The hour and minute of the spawning window is set by sunset time (Levitan et al., 2004). By taking these four factors into consideration, spawn windows for a large number of species can be accurately predicted in some locations (Vize et al., 2005).
Seasonal cycles can be entrained and controlled by long-term circannual rhythms (reviewed by Dunlap et al., 2004). Examples include plant flowering, bird migration, and reproductive cycles (Gwinner, 1981). The month of coral spawning is correlated with cycles in solar insolation (van Woesik et al., 2006) and in some cases with cycles of water temperature (Willis et al., 1985), but it is not known if this is an entrained cycle or a direct effect of local environment on gonad maturation. An entrained system could regulate a developmental process at any point in the entrained cycle and would explain why corals in different geographic regions spawn at different times of year. However, at present there is no empirical evidence of the biological basis, entrained or otherwise, of the timing systems that regulate the cycle of gamete maturation and the month of spawning in corals.
It is also not known how corals sense lunar cycles and translate this information into the sequence of physiological processes that mature gametes and determine the date of spawning, or whether these are entrained processes or are directly under the control of environmental factors. There is evidence that lunar signals act through light, as opposed to through tides or other processes driven by the orbiting moon (Jokiel et al., 1985). The brooding coral Pocillopora damicornis has been shown to release planulae in response to lunar light cycles in tanks isolated from tidal influences (Jokiel et al., 1985), and broadcast spawners may use similar lunar cues (pers. comm., B. Willis, James Cook University). Also, despite its dimness, moonlight has been shown to be strong enough to entrain circadian pacemakers in terrestrial animals (Evans et al., 2007), and corals can detect lunar intensities of light (Gorbunov and Falkowski, 2002; Levy et al., 2007).
The coral response to lunar light could be direct or entrained. If it were direct, changing lunar irradiance would have an immediate effect on the process responding to light; if it were entrained, removal of the cycle synchronizing light would result in a modest phase shift as the cycle free runs--a state than can last for a long time. The consistency from year to year of the dates of coral broadcast spawning relative to the lunar cycle, despite differences in cloud cover or weather and in patterns of lunar light from season to season, is circumstantial evidence for a circalunar entrained system, as is the transcriptional response of the circadian cry2 gene to lunar light (Levy et al., 2007). On the other hand, in a subset of P. damicornis colonies, type B, release of brooded planula can be reset by altered lunar-like irradiance over the course of only 1 month, and planula release in these individuals is also influenced by periods of heavy cloud cover. A different subtype of the same species, type Y, displays more entrained planula release patterns that continue to cycle for about 3 months in the absence of lunar illumination--strong evidence for entrainment (Jokiel et al., 1985). In summary, while there are some quantitative data in brooding species and circumstantial evidence in broadcast spawning species to support the existence of lunar-light-entrained reproductive activity, we know very little about the systems regulating these processes and nothing about the genes responsible for determining the date of spawning.
The hour and minute of spawning is established by the solar cycle. Not only does sunset time allow very accurate prediction of spawning times (e.g., Vize et al., 2005), but artificially inducing early sunset for only a few days results in a corresponding early shift in those times (Knowlton et al., 1997; Levitan et al., 2004). The ability to reset spawning time so quickly argues that this is not an entrained process, but this has not yet been conclusively demonstrated.
Recently the characterization of cryptochrome genes from the broadcast spawning species Acropora millepora (Ehrenberg, 1834) demonstrated that two of these genes are transcribed in a 24-h cyclic manner and that one gene, cry2, has a transcriptional response to moonlight (Levy et al., 2007). These results show that corals share at least two of the genes known to regulate entrained genetic networks in other animals and also in plants. While cryptochromes play essential roles in circadian gene networks, they function at the terminal end of complex pathways involving dozens of other critical genes without which cryptochrome transcription cannot be entrained (Lowrey and Takahashi, 2000). Also, it has not yet been demonstrated whether cryptochromes and other circadian genes function in regulating either the lunar or solar cycle responses in corals. To explore the circadian gene landscape of corals, we made use of the recent generation of a large cDNA sequence resource developed from the larvae of Acropora millepora. This analysis allowed us to identify 24 A. millepora sequences that correspond to circadian network genes from insects and mammals. Many of these genes are very highly conserved. The presence of all members of the insect and mammalian networks implies that the same networks also function in corals. These data provide the means with which to explore the importance of entrained processes in many aspects of coral biology, including understanding the basis of how the month, evening, and hour of broadcast spawning is determined.
Materials and Methods
The 634,070 sequencing reads generated from the RNA of Acropora millepora larvae (E. Meyer, G. Aglyamova, J. Buchanan-Carter, J. Colbourne, D. Abrego, B. L. Willis, and M. Matz, University of Texas; pers. comm.) by 454 sequencing (Margulies et al., 2005) will be described in detail elsewhere and are available online (http://www.dtd.nlm.nih.gov/Traces/sra/sra.cgi?cmd=show&f=main&m=main&s=main) at the Short Read Archive of the National Center for Bioinformatics Information's trace archive (reference SRA003728). RNA used in the sequencing was collected from 5-day-old planulae (28 [degrees]C). This analysis generated 44,569 contigs covering more than 20 million bp. This sequence resource was used for analysis with the tBLASTn program of the NCBI to compare known circa-dian regulatory gene protein sequences against translated contig sequences. Circadian pathway components were identified from the Science signal transduction knowledge environment (http://stke.sciencemag.org/cm/). Genes were selected both from the suprachiasmatic nucleus pathway that entrains the mammalian master circadian pacemaker (Van Gelder, 2005b) and from the Drosophila lateral neuron pacemaker (Van Gelder, 2005a). Mammalian (Mus) sequences were chosen if possible so that Expect (E) values would all be relative to a common evolutionary distance, but when none were available, Drosophila sequences were utilized. The accession numbers for the query gene set are shown in Table 1. Because multiple related genes were used as queries (e.g., period 1 and period2), matches sometimes overlapped. When this occurred (bhlhb2/bhlhb3, cryl/cry2, perl/per2, csnkld/csnkle/dco), the gene with the lowest E value was listed as the putative ortholog and the target removed from candidacy as an ortholog for other family members. All matches are, however, included in result tables.
Table 1 Queried genes Source Symbol Synonym Accession Function mouse bmal1 cycle BAA76414.1 transcription mouse bhlhb2 dec1 EDK99412.1 transcription mouse bhlhb3 dec2 BAB21503.1 transcription mouse cry1 -- BAA19175.1 transcription mouse cry2 -- BAA19864 transcription mouse nrld1 rev-erbA BAE23342.1 transcription mouse per1 -- BAA22634.1 transcription mouse per2 -- AAC53592.1 transcription mouse ppm2c pdb1 AAF04507.1 transcription mouse timeless -- BAA36500.1 transcription mouse clock -- AAC53200.1 nuclear translocator mouse grin1 nmdar BAE34379.1 membrane receptor mouse vpac2 -- BAA05674.1 membrane receptor mouse csnk1d CK1 delta BAB23405.1 kinase mouse csnk1e -- BAA88082.1 kinase mouse creb -- AAL47130.1 cAMP pathway mouse Adcyaplrl paclR BAC31372.1 cAMP pathway mouse Adcyap1 pacap BAC37673.1 cAMP pathway fly ckIIalpha -- AAA28429.1 kinase fly ckIIbeta ck2 AAM29452.1 kinase fly dco cklepsilon AAD27857.1 kinase fly shaggy zw3 CAA50213.1 kinase fly slmb -- AAF63213.1 ubiquitin protein-ligase fly vri vrille ABO52821.1 transcription
Results and Discussion
The query files used to search for circadian genes were 18 mouse and 6 fly proteins (Table 1). The quality of a BLAST match between a query sequence and a target sequence is expressed as an E value, with smaller values indicating a greater and greater statistical likelihood of a nonrandom match. Only one query had a very high E value and therefore a higher chance of being a poor match--the adenylate cyclase activating polypeptide 1 (adcyapl) with a score of 0.2--all other pathway genes had stronger scores (Table 2). The highest score was for shaggy/zw3 (e-176), which displayed 364/388 (94%) similar and 299/388 (77%) identical matches over a 388 amino acid region. Even the moderately weak E values displayed by some genes such as adcyaplrl (0.0002) still displayed a high degree of relatedness to the mammalian protein (Fig. l). Although the sequences generated by 454 sequencing are more than 99% accurate (Mar-gulies et at., 2005), errors may contribute to some of the differences between query and target sequences. Alignments between gene products ranging from highly conserved to modestly conserved are presented in Figure 1. As the transcriptome coverage in the target dataset is about 3 X and the contigs are less than full length, some orthologs may have been missed and the identified targets may be homologous family members. Still, the very high confidence level of most matches indicates a high probability that many of the identified targets represent Acropora versions of the circadian regulatory genes (Table 2).
[FIGURE 1 OMITTED]
Table 2 Coral and anemone circadian genes Acropora Acropora Acropora Nematostella ** Hit Hit Hit 3 * Gene Query 1* / E 2* / E / E / E value value value value bmall BAA76414.1 14749 / 4525 / 59 / 132249 / e-103 e-60 e-46 e-41 bhlhb2 EDK99412.1 10394 / 10393 not 28948 / e-09 e-07 /e-07 found bhlhb3 BAB21503.1 10393 / 10394 / not 246249 / e-08 0.0142 0.0066 found cryl BAA19175.1 10302 / 10301 / 10300 / 203127 / 0.0 7e-136 le-130 e-46 cry2 BAA19864 10301 / 10302 / 10300 / 194898 / 0.0 2e-130 2e-135 2e-46 nrldl BAE23342.1 7574 / 4876 / 15470 / 101676 / e-23 e-24 e-23 e-18 perl BAA22634.1 59 / 1293 / 6234 / 116698 / e-11 4e-l3 2e-10 le-07 per2 AAC53592/1 59 / 1293/5e-I0 21530 / 132249 / e-12 5e-13 0.5447 ppm2c AAF04507.1 3940 / 5888 / not 87565 / e-12 1e-05 6e-05 found timeless BAA36500.1 15723 / 16388 / 15423 / 81863 / 0.0 e-58 e-09 0.0211 clock AAC53200.1 10199 / 6234 / 59 / 160110/e-62 e-55 e-23 e-20 grin1 BAE34379.1 8101 / 3818 / 5377 / 171792 / 0.0 e-22 e-17 e-07 vpac2 BAA05674.1 11562 / 10771 / 14370 / 13642 / e-18 e-05 0.0169 0.1196 csnkld BAB23405.1 10093 / 4860 / 8858 / 12115 / 0.0 e-149 e-137 e-117 csnkle BAA88082.1 4860 / 10093 / 8858 / 12115 / e-180 e-134 0-145 e-U8 creb AAL47130.1 5346 / not found not 242787 / e-33 e-22 found adcyaplrl BAC31372.1 12939 / 11562 / 17735 / 201898 / e-12 0.0002 0.0038 0.2515 adcyapl BAC37673.1 1169 / not found not not found 0.2081 found ckIIalpha AAA28429.1 3185 / 4918 / 14339 / 166091 / e-151 e-112 e-30 e-30 ckIIbeta AAM29452.1 3471 / 3470 / 102002 / 245041 /e-100 e-103 e-32 0.3816 dco AAD27857.1 4860 / 8858 / 10093 / 12115 / e-156 e-129 e-111 e-89 shaggy CAA50213.1 4326 / 5912 / 5244 / 190252 / e-172 e-176 e-46 e-41 sImb AAF63213.1 12841 / 9877 / 12495 / 228088 / 0.0 e-66 e-33 e-30 vri ABO52821.1 5888 / 3940 / 3472 / 87565 / e-07 e-07 e-06 0.0257 * SeqIndex number / E (Expect) value. ** JGI Nematostella vectensis assembly 1.0 protein model ID.
Alignments were also performed against the genome of the sea anemone Nematostella vectensis (Stevenson, 1935), build 1.0 (http://genome.jgi-psf.org/Nemvel/Nemvel.home. html). Once again, matches were identified to all queried genes with, as in Acropora, the exception of adcyapl. The Nematostella matches had E values ranging from 0 (zero chance of random match) to e-07 for vri, which is still a strong match. Scores were in general better to Nematostella genes than to Acropora. This is due to the anemone genes being derived from a sequenced genome and therefore full-length sequences are available for all genes, whereas the Acropora contigs are shorter and only rarely full length. Once again, the most highly conserved genes were kinases, and the most poorly conserved were members of the cAMP pathway. Together, these data indicate that cnidarian genomes contain and express genes orthologous to those regulating circadian pathways in mammals and insects.
In Figure 2 the circadian gene network of the mammalian suprachiasmatic nucleus is illustrated along with the confidence scores of corresponding Acropora candidate orthologs. The period gene perl plays a central role in the integration of inputs from multiple signaling pathways (reviewed by Dunlap, 1999). All the nuclear proteins in this pathway downstream of perl, with the exception of dec2/ bhlhb3, have Acropora sequences with E values of less than [10.sup.9] and likely represent coral orthologs of the mammalian genes. The proteins generating input signals for perl, such as adcyapl, vpac, and nmdar, are less conserved. This may represent a greater tolerance in these proteins for amino acid substitutions or may simply be due to the queried dataset containing only homologous and not orthologous targets. As we have no data indicating which signal transduction pathways synchronize coral circadian networks, it is of course possible that the mammalian neuropeptide and cAMP pathways in which these proteins function are provided by alternatives in cnidarians.
[FIGURE 2 OMITTED]
Because we know that moonlight and sunlight are potential entraining signals, we also performed a search for Acropora photoreceptors. Coral rhodopsin-like genes have already been described from A. millepora (Anctil et al., 2007). In many animals the nonvisual photopigment melanopsin/opn4 entrains circadian networks, but this gene has not been described in cnidarians. Contig 15327 matches mammalian opn4 with an E value of 3e-20 (Fig. 3) and may, like its vertebrate equivalents, function in synchronizing the circadian gene expression. The best matches for the conceptual protein encoded by 15327 in the nr collection at the National Center for Biotechnology Information are various Nematostella predicted opsins, followed by the chick, human, pig, and mouse opn4 genes. As melanopsin uses calcium as a second messenger (Panda et al, 2005; Kumbalasiri et al, 2007) and calcium can entrain circadian rhythms (Love et al, 2004), linkage between this photoreceptor and the circadian regulatory network may be a valuable avenue of future exploration.
[FIGURE 3 OMITTED]
As with the suprachiasmatic nucleus pathway, the pacemaker circadian gene network of Drosophila is also represented and very highly conserved in corals (Fig. 4). All classes of proteins in this pathway--kinases, transcription factors, and a ubiquitin protein ligase--are highly conserved. Once again, the most conserved family members are kinases (sgg and ckl/dco) followed by transcription factors. As all members of both of these networks are present in corals, it is possible that the interactions between network members are also conserved.
[FIGURE 4 OMITTED]
It is interesting that all of these genes are transcribed in 5-day-old planulae, the source of the RNA used in sequencing the Acropora transcriptome, as this indicates that circadian cycles are established very early in larval development. Further studies analyzing the expression of each of these genes during solar and lunar cycles may help explain how these systems are regulated. The availability of DNA and conceptual protein sequences for circadian genes can now be used to design quantitative mRNA measurement protocols and generate anti-peptide antibody-based protein assays that will allow these questions to be addressed at the molecular level. This will enable us to understand the role of entraining signals, transduction pathways, and transcription in regulating reproductive behavior and other cyclical aspects of coral biology such as skeletal deposition, metabolism, growth, and feeding. Although the larvae used in this study lacked zooxanthellae, in adult tissues the metabolic activity of the coral host and their symbiotic zooxanthellae will be interfacing, and this may have profound effects on circadian systems. The identification of circadian orthologs provides a molecular toolkit with which such interactions can now be studied.
This analysis was made possible by the generous early distribution of 454 data by Drs. M. Matz and E. Meyer (University of Texas) prior to its publication. I also thank Kevin Synder for scripting the BLAST searches.
Anctil, M., D. C. Hayward, D. J. Miller, and E. E. Ball. 2007. Sequence and expression of four coral G protein-coupled receptors distinct from all classifiable members of the rhodopsin family. Gene 392: 14-21.
Babcock, R. C, G. D. Bull, P. L. Harrison, A. J. Heyward, J. K. Oliver, C. C. Wallace, and B. L. Willis. 1986. Synchronous spawnings of 105 scleractinian coral species on the Great Barrier Reef. Mar. Biol. 90: 379-394.
Baird, A. H., P. A. Marshall, and J. Wolstenholme. 2002. Latitudinal variation in the reproduction of Acropora in the Coral Sea. Pp. 385-389 in Proceedings of the 9th International Coral Reef Symposium, Vol. 1, 23-27 October 2000, Bali, Indonesia.
Dunlap, J. C. 1999. Molecular bases for circadian clocks. Cell 96: 271-290.
Dunlap, J C., J.L. Loros, and P. J. DeCoursey. 2004. Chronobiology: Biological Timekeeping. Sinauer Associates, Sunderland, MA.
Evans, J. A., J. A. Elliot, and M. R. Gorman. 2007. Circadian effects of light no brighter than moonlight. J. Biol. Rhythms 22: 356-367.
Gorbunov, M. Y., and P. G. Falkowski. 2002. Photoreceptors in the cnidarian hosts allow symbiotic corals to sense blue moonlight. Limnol. Oceanogr. 47: 309-315.
Guest, J, R, A. H. Baird, B. P. Goh, and L. M. Chou. 2005. Seasonal reproduction in equatorial reef corals. Invertebr. Reprod. Dev. 48: 207-218.
Gwinner, E. 1981. Circannual systems. Pp. 391-410 in Handbook of Behavioral Neurobiology, J. Aschoff, ed. Plenum, New York.
Hagman, D. K., and P. D. Vize. 2003. Mass spawning by two brittlestar species Ophioderma rubicundrum and O. Squamosissiomum, at the Flower Garden Banks, Gulf of Maxico. Bull, Mar. Sci. 72: 871-876.
Hagman, D. K., S. R. Gittings, and K. J. Deslarzes. 1998. Timing, species participation, and environmental factors influencing annual mass spawning at the Flower Garden Banks (northwest Gult of Mexico). Gulf Max. Sci. 16: 170-179.
Harrison, P. L., R. C. Babcock, G. D. Bull, J. K. Oliver, C. C. Wallace, and B. L. Willis. 1984. Mass spawning in tropical reef corals. Science 223: 1186-1189.
Hughes, T. P., A. H. Baird, E. A. Dinsdale, N. A. Moltschaniwskyj, M. S. Pratchett, J. E. Tanner, and B. L. Willis. 2000. Supply-side ecology works both ways: the link between benthic adults, fecundity, and larval recruits. Ecology 81: 2241-2249.
Jokiel, P. L., R. Y. Ito, and P. M. Liu. 1985. Night irradiance and synchronization of luar release of planula larvae in the reef coral Pocillopora domicornis. Mar. Biol. 88: 167-174.
Knowlton, N., J. L. Mate, H. M. Guzman, R. Rowan, and J. Jara. 1997. Direct evidence for reproductive isolation among the three species of the Montastraea annularis complex in Central America (Panama and Honduras). Mar. Biol. 127: 705-711.
Kumbalasiri, T., M. D. Rollag, M. C. Isoldi, A. M. Castrucci, and I. Provencio. 2007. Melanopsin triggers the release of internal calcium stores in response to light. Photochem photobiol. 83: 273-279.
Levitan, D. R., H. Fukami, J. Jara, D. Kline, T. M. McGovern, K. E. McGhee, C. A. Swanson, and N. Knowlton. 2004. Mechanisms of reproductive isolation among sympatric broadcast spawning corals of the Montastraea annularis species complex. Evolution 58: 308-323.
Levy, O., L. Appelbaum, W, Leggat, Y. Gothlif, D. C. Hayward, D. J. Miller, and O. Hoegh-Guldberg. 2007. Light-responsive cryptochromes from a simple multicellular animal, the coral Acropora mille-pora. Science 318: 467-470.
Love, J., A- N. Dodd, and A. A. Webb. 2004. Circadian and diurnal calcium oscillations encode photoperiodic information in Arabidopsis. Plant Cell 16; 956-966.
Lowrey P. L., and J. S. Takahashi. 2000. Genetics of the mammalian circadian system: photic entrainment, circadian pacemaker mechanisms, and posttranslational regulation, Annu. Rev. Genet. 34: 533-562
Margulies, M., M. Egholm, W. E. Altman, S. Attiya, J. S. Bader, L. A. Bemben, J. Berka, M. S. Braverman, Y.-J. Chen, Z. Chen, et al. 2005. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437: 376-380.
Ovlier, J. K., R. C. Babcock, P. L. Harrison, and B. L. Willis. 1988. Geographic extent of mass coral spawning: clues to ultimate causal factors. Pp. 803-810 in Proceedings of the 6th International Coral Reef Symposium, Vol. 2, 8-12 August 1988, Townsville, Australia.
Panda, S., S. K. Nayak, B. Campo, J. R. Walker, J. B. Hogenesch, and T. Jegla. 2005. Illumination of the melanopsin signaling pathway. Science 307: 600-604.
Van Gelder, R. N. 2005a. Drosophila Circadian Pathway. Sci. Signal. (Connectons Map in the Database of Cell Signaling) [Online] Available: http://stke.sciencemag.org/cgi/cm/stkecm;CMP_13010 [2008 June 5].
Van Gelder, R. N. 2005b. Murine Circadian Pathway. Sci. Signal. (Connections Map in the Database of Cell Signaling) [Online] Available: http://stke.sciencemag.org/cgi/cm/stkecm; CMP_13010 [2008 June 5].
van Veghel, M. L. 1993. Multiple species spawning on Curacao. Bull. Mar. Sci 52: 1017-1021.
van Woesik, R., F. Lacharmoise, and S. Koksal. 2006. Annual cycles of solar insolation predict spawning times of Caribbean corals. Ecol. Lett. 9: 390-398.
Vize, P. D., J- A. Embesi, M. Nickell, D. P. Brown, and D. K. Hagman. 2005. Tight temporal consistency of coral mass spawning at the Flower Garden Banks, Gulf of Mexico, from 1997-2003. Gulf Mex. Sci. 23: 107-114.
Willis, B. L., R. C. Babcock, P. L. Harrison, and J. K. Oliver. 1985. Patterns in the mass spawning of corals on the Great Barrier Reef from 1981 to 1984. Pp. 343-348 in Proceedings of the 5th International Coral Reef Congress, Vol. 4, 27 May-1 June 1985, Tahiti.
PETER D. VIZE
Department of Biological Sciences, University of Calgary, Calgary, Algerta T2N1N4, Canada
Received 3 September 2008; accepted 9 January 2009.
|Printer friendly Cite/link Email Feedback|
|Author:||Vize, Peter D.|
|Publication:||The Biological Bulletin|
|Date:||Apr 1, 2009|
|Previous Article:||The canonical network of autotrophic intermediary metabolism: minimal metabolome of a reductive chemoautotroph.|
|Next Article:||Cyclic guanosine monophosphate signaling cascade mediates pigment aggregation in freshwater shrimp chromatophores.|