Circadian rhythms in the pineal organ persist in zebrafish larvae that lack ventral brain
© Noche et al; licensee BioMed Central Ltd. 2011
Received: 9 June 2010
Accepted: 13 January 2011
Published: 13 January 2011
Skip to main content
© Noche et al; licensee BioMed Central Ltd. 2011
Received: 9 June 2010
Accepted: 13 January 2011
Published: 13 January 2011
The mammalian suprachiasmatic nucleus (SCN), located in the ventral hypothalamus, is a major regulator of circadian rhythms in mammals and birds. However, the role of the SCN in lower vertebrates remains poorly understood. Zebrafish cyclops (cyc) mutants lack ventral brain, including the region that gives rise to the SCN. We have used cyc embryos to define the function of the zebrafish SCN in regulating circadian rhythms in the developing pineal organ. The pineal organ is the major source of the circadian hormone melatonin, which regulates rhythms such as daily rest/activity cycles. Mammalian pineal rhythms are controlled almost exclusively by the SCN. In zebrafish and many other lower vertebrates, the pineal has an endogenous clock that is responsible in part for cyclic melatonin biosynthesis and gene expression.
We find that pineal rhythms are present in cyc mutants despite the absence of an SCN. The arginine vasopressin-like protein (Avpl, formerly called Vasotocin) is a peptide hormone expressed in and around the SCN. We find avpl mRNA is absent in cyc mutants, supporting previous work suggesting the SCN is missing. In contrast, expression of the putative circadian clock genes, cryptochrome 1b (cry1b) and cryptochrome 3 (cry3), in the brain of the developing fish is unaltered. Expression of two pineal rhythmic genes, exo-rhodopsin (exorh) and serotonin-N-acetyltransferase (aanat2), involved in photoreception and melatonin synthesis, respectively, is also similar between cyc embryos and their wildtype (WT) siblings. The timing of the peaks and troughs of expression are the same, although the amplitude of expression is slightly decreased in the mutants. Cyclic gene expression persists for two days in cyc embryos transferred to constant light or constant dark, suggesting a circadian clock is driving the rhythms. However, the amplitude of rhythms in cyc mutants kept in constant conditions decreased more quickly than in their WT siblings.
Our data suggests that circadian rhythms can be initiated and maintained in the absence of SCN and other tissues in the ventral brain. However, the SCN may have a role in regulating the amplitude of rhythms when environmental cues are absent. This provides some of the first evidence that the SCN of teleosts is not essential for establishing circadian rhythms during development. Several SCN-independent circadian rhythms have also been found in mammalian species. Thus, zebrafish may serve as a model system for understanding how vertebrate embryos coordinate rhythms that are controlled by different circadian clocks.
Circadian rhythms are biological cycles in behavior, physiology, and biochemistry that occur approximately every 24 hours. These oscillations are present in almost every organism, from cyanobacteria to plants to humans . All circadian rhythms are regulated by a timing system composed of intracellular clocks with periods of approximately 24 hours, environmental cues, and clock-controlled outputs. An important characteristic is that circadian clocks are able to drive output rhythms even in the absence of environmental cues. However, environmental influences such as light and temperature are required to entrain or re-set the circadian oscillators so that they stay in synchronization with the organism's surroundings.
Regulation of vertebrate circadian rhythms is best understood in mammals. The SCN, a group of neurons in the hypothalamus, is among the most intensely studied cellular site of circadian oscillators. Targeted bilateral lesion of the rodent SCN abolished circadian oscillations in activity and drinking rhythms . Conversely, transplantation of a donor SCN into a host with a lesioned SCN restored circadian cycling . Further, the new locomotor activity rhythms of the host matched the circadian cycle of the donor SCN, indicating that the SCN was the major influence on the phase of the restored rhythms . The SCN controls rhythms in other tissues through a variety of mechanisms, including the secretion of peptide hormones and synaptic signaling . For instance, rhythmic production of the circadian hormone melatonin in mammals is controlled almost entirely by a multi-synaptic pathway leading from the SCN to the cells of the pineal gland .
Although it is clear that the mammalian SCN is important in controlling many rhythms in numerous organs, mammals also have circadian oscillators in other tissues as well as rhythms that are SCN-independent [6, 7]. For instance, the mammalian retina has an endogenous oscillator that controls local rhythms such as visual sensitivity and retinal melatonin synthesis . This indicates that the mammalian circadian system may have similarities to the circadian systems of many lower vertebrates, which often have many oscillators and many sites of photoreception [9–11].
A structure anatomically equivalent to the SCN has been described in zebrafish embryos and adults [12–17]. However, the function of the zebrafish SCN in regulating circadian rhythms is unknown . This question is particularly interesting as many different isolated zebrafish cells and organs, such as the pineal, eyes, heart, spleen, and kidney, have endogenous oscillators and circadian photoreceptors that are able to generate cyclic gene expression [18–21]. For example, in adult zebrafish the pineal organ contains an endogenous circadian oscillator that is sufficient to drive rhythms in melatonin synthesis as well as photoreceptive neurons that entrain this oscillator . However, whether pineal rhythms are also influenced by signals from other tissues, such as the SCN, is unknown in zebrafish and many other lower vertebrates. Interestingly, in some avian species, pineal rhythms are controlled both by an endogenous pineal oscillator and by input from the SCN, raising the possibility that multiple tissues could regulate pineal circadian rhythms in zebrafish [11, 22, 23].
Here, we provide evidence that the SCN is not required for pineal rhythms in developing zebrafish. To do this, we took advantage of zebrafish cyc mutants, which have a mutation in one of three zebrafish nodal genes [24, 25]. Lack of Cyc/Nodal signaling results in a complete absence of the hypothalamus, including the regions that give rise to the SCN, the retro-chiasmatic nucleus, and the infundibulum [17, 24–26]. Consistent with this earlier work, we find that expression of the avpl gene, which is typically expressed in and around the SCN [27, 28], is absent in cyc mutants. Despite this, expression of the putative clock component genes cry1b and cry3, as well as the structure and size of the pineal organ was indistinguishable between cyc embryos and their WT siblings. The phase of gene expression of two pineal rhythmic genes, exorh and aanat2 persisted in the mutants. However, semi-quantitative analysis suggested that the amplitude of the rhythms was slightly reduced. aanat2 mRNA levels maintained their rhythmic changes in cyc embryos placed in constant conditions. This suggests that transcriptional rhythms were not being controlled by the environment, but rather by an internal cellular clock, perhaps the endogenous clock within the pineal cells. Since cyc mutants never have an SCN, this work indicates daily rhythms within the developing zebrafish pineal can be initiated in the absence of the developing hypothalamus, including the SCN anlage.
To further characterize the ventral brain defect in cyc mutants, we examined the expression of the avpl gene. In mammals, vasopressin is expressed by the SCN neurons with a strong circadian rhythm, and has been linked to changes in hormone secretion from the pituitary, regulation of reproduction in females, and behaviors such as daily rhythms in wheel running activity in nocturnal rodents . In non-mammalian vertebrates, vasopressin is replaced by Avpl, which is expressed in many ventral brain neurons including cells in and around the SCN . In zebrafish, avpl mRNA is found in two domains of the ventral brain: the dorsal preoptic area and the anterior aspect of the ventral hypothalamus . Based on the location of the avpl positive cells in the ventral hypothalamus, these cells likely correspond to the zebrafish SCN. In cyc embryos, both avpl expression domains were undetectable (Figure 1C - 1J). This suggests that neurons expressing avpl, including those in the SCN, are missing.
To determine what happens to pineal rhythms in cyc mutants, we first assayed the embryos for expression of exorh mRNA and protein. Exorh is a putative light-sensing, G-protein coupled receptor expressed in the pineal [40–43]. exorh mRNA is synthesized with a daily rhythm, with highest levels during the night and lower levels during the day [43, 44].
Embryos were raised in a standard 14:10 h light/dark (L/D) cycle, with all other environmental parameters held constant. Rhythms in exorh mRNA were present in the pineal organs of both WT and cyc fish (Figure 2A, C). Further, the timing of the troughs and peaks of expression were similar between both sets of embryos (Figure 2A, C).
However, the amplitude of exorh expression was slightly lower at almost every time point tested in the cyc embryos (Figure 2C). To rule out the possibility that this was due to smaller pineal organs in the mutants, we measured the length and width of the pineal organs labeled for expression of orthodenticle homeobox 5 (otx5), a gene that is constitutively expressed in all pineal cells . We found that the pineal size was not significantly different between WT and cyc fish (Figure 2E, F). This suggests that the decrease in the amplitude of exorh expression in cyc embryos was not due to smaller pineal organs.
aanat2 encodes the penultimate enzyme in the melatonin biosynthetic pathway . In zebrafish and many other vertebrates, aanat genes are expressed in the pineal with dramatic differences between day and night levels . Thus, aanat2 expression serves as a very sensitive readout of pineal circadian cycling. In WT and cyc embryos maintained in a L/D cycle, aanat2 transcripts cycled with indistinguishable periods and phases (Figure 2B, D). However, as was the case for exorh, the amplitude of aanat2 expression in cyc mutants was slightly diminished compared to WT siblings raised in parallel (Figure 2D). Thus, the nocturnal expression of aanat2 in the pineal was largely unaffected by the loss of the putative SCN in cyc mutant embryos.
When the embryos were exposed to a L/D cycle for only two days before transfer to constant conditions, the amplitude of the rhythmic gene expression was very low in both WT and cyc mutants, suggesting that circadian cycling was not fully established by such a short period of entrainment [Additional File 2]. In contrast, when embryos were maintained instead for three days in a L/D cycle, rhythmic expression of aanat2 persisted in WT and in cyc larva after transfer to either constant light or constant dark conditions (Figure 4). Under constant dark conditions, a slight shift in the peak of expression to the right was apparent, consistent with previous studies that showed the endogenous period of the clock that controls zebrafish pineal rhythms is slightly longer than 24 hours (Figure 4B) [10, 35, 39, 47]. However, this shift was not readily apparent in fish in constant light (Figure 4B). The persistence in cyclic changes in expression after transfer to constant lighting conditions suggests that the rhythms are not due to responses to rhythmic environmental changes, but instead due to a functioning circadian clock in the cyc embryos.
Although the overall rhythm in gene expression persisted in constant conditions, there were some significant differences between the cyc mutants and their WT siblings. As in the L/D experiments (Figure 2), the levels of mRNA were slightly lower in the cyc embryos at most time points (Figure 4). Further, after approximately 48 hours in constant conditions, for both constant light and constant dark, the aanat2 expression was notably lower in the cyc embryos than in their WT siblings (Figure 4).
To gain insight into other potential changes in cyc embryos, we examined the expression of putative clock components during embryogenesis. The vertebrate circadian clock is composed of complex positive and negative feedback loops that take approximately 24 hours to go through one cycle. Cry proteins are essential components of this clock, in which they act as transcriptional repressors in the negative feedback loop. Zebrafish have six cry genes, all of which are expressed with a daily rhythm [48, 49].
In mammals and birds, the SCN has an important role in regulating circadian rhythms throughout the organism. Retinal photoreceptors entrain circadian oscillators within the SCN, and the SCN then regulates pineal rhythms through a multisynaptic pathway . Although the retina-SCN-pineal pathway is present in fish, amphibians, and reptiles, its function is not well understood [15, 51–53]. Here we provide evidence that pineal circadian rhythms are established and have normal phase in zebrafish cyc mutants, which lack ventral brain. This suggests that neither the SCN nor the retinal-hypothalamic pathway is essential for the onset of pineal circadian rhythms during embryogenesis. However, the amplitude of the rhythms is slightly, but significantly decreased, in cyc mutants. This suggests that the SCN or other tissues that are missing in these mutants, which include other regions of the ventral brain, may have a role in regulating rhythms during development.
The SCN has been defined anatomically in zebrafish embryos and adults . However, the function of the zebrafish SCN is unknown. We used cyc mutants to test the function of the zebrafish SCN. These mutants are missing the whole ventral brain including the entire hypothalamus, and thus, the precursors to the SCN [17, 26]. We now demonstrate that expression of avpl, which is usually present in and around the SCN, is lacking in cyc mutants (Figure 1). This provides strong evidence that the SCN is completely absent in cyc mutants.
We found that rhythmic expression of aanat2 and exorh are present in cyc mutants during the first few days of development (Figure 2). The phases of expression of these genes are indistinguishable between cyc mutants and WT siblings raised in parallel in a L/D cycle. Further, the aanat2 expression rhythms persist after the embryos are transferred to constant dark or constant light, indicating they are being driven by a circadian clock (Figure 4).
The expression amplitude of aanat2 and exorh mRNA levels are slightly reduced in the cyc mutants compared to WT siblings raised in parallel. This was true for fish in a L/D cycle, and after transfer to constant dark or constant light conditions (Figures 2 and 4). In addition, the decrease in amplitude of expression became more pronounced when the fish approached the third day in constant conditions (Figure 4).
cyc mutants have many developmental defects, making it difficult to precisely identify the cause of the decrease in gene expression. It is unlikely that the change is due to defects in pineal development, as the size of the pineal was not significantly decreased in cyc mutants (Figure 2F). However, the lower amplitude of gene expression could be due in part to poor health or subtle changes in the pineal that we have not detected [24–26].
A more interesting possibility is that the SCN or other missing tissues in the ventral brain play a role in regulating the amplitude of pineal rhythms. Because the phenotype of cyc mutants becomes more severe as the fish get older, we kept the fish in a L/D cycle for only a short time before being assayed. Thus, one possibility is that the SCN/ventral brain functions to help establish self-sustaining rhythms during development. This role for the SCN could explain why the amplitude of rhythms were lower in the cyc pineal overall, and why the rhythms in the cyc pineal started to dampen sooner after transfer to constant conditions. Testing this hypothesis will require a mutant that more specifically affects the SCN or a way to specifically inhibit the function of SCN cells.
We do not completely understand what is driving pineal rhythms in cyc mutants. One possibility is that these rhythms are driven by clocks within the zebrafish pineal. Persistent circadian rhythms in melatonin synthesis were observed previously in isolated adult zebrafish pineals cultured for several days . This suggests that the pineal oscillators in adult zebrafish are self-sustained. Clocks in the embryonic pineal may work in the same way. In support of this, a previous study showed that light input very early in development, long before the differentiation of an SCN, is necessary for onset of pineal rhythms .
Oscillator proteins present in the pineal are also in many other zebrafish cells and tissues. Likewise, circadian rhythms have been detected in isolated cultured zebrafish embryonic and adult tissues [18–20]. Thus, it is possible that other tissues present in cyc mutants are interacting to promote circadian cycling in the pineal.
We also cannot rule out a role for the lateral eyes. The first zebrafish retinal photoreceptors differentiate at approximately 2.5 dpf . Thus, detection of environmental light conditions by the retina could explain why circadian rhythms in pineal gene expression were better established when embryos were maintained for three rather than two days in a L/D cycle (compare Figure 4 with Additional file 2). Previous work suggested that the lateral eyes are not required for establishing rhythms in the pineal. Rhythms in pineal aanat2 expression are present in zebrafish chokh (chk) mutants, which lack lateral eyes . Further, chokh mutants can re-entrain to a shifted L/D cycle . However, the persistence of pineal rhythms in chokh fish after transfer to constant lighting conditions has not yet been tested, raising the possibility that they could have defects similar to those we found in cyc mutants.
A final possibility is that pineal rhythms are present in cyc mutants because there are some residual SCN neurons. However, this is unlikely, as the whole ventral brain is absent in cyc mutants (Figure 1A - 1B) [17, 24–26, 29]. chk mutants are missing both eyes and hypothalamic avpl expression . Despite this, they have rhythms in gene expression of two clock components (clock and period4) and pineal aanat2 [56, 58]. These results are consistent with our findings that loss of neurons expressing avpl, including those in the SCN, does not abolish pineal rhythms.
Cry proteins are required for regulation of circadian rhythms in both plants and animals . These proteins were shown to play two major roles in regulation of circadian rhythms: as a photoreceptor to entrain the clock and as a repressor of Clock/BMAL-induced circadian transcription . Six Cry proteins are present in zebrafish . By heterologous expression, it was shown that Cry1b, but not Cry3, is capable of blocking Clock/BMAL-dependent transcription . This suggests that Cry1b and Cry3 have different functions. We find that cry1b and cry3 mRNAs are found in specific tissues of zebrafish embryos and larvae (Figures 5 and 6). Further, we demonstrate that cry1b and cry3 expression in developing brain, ear, and liver are present in cyc mutants (Figures 5 and 6). Although the function of the cry genes in zebrafish is not yet fully understood, the presence of cry1b and cry3 transcripts in the cyc brain suggests that the circadian clock may not be severely disrupted.
exorh was originally identified as a rhodopsin class gene expressed in the zebrafish and salmon pineal organs [40, 61]. Exorh has been predicted to be a G-protein coupled receptor with ~70% amino acid sequence identity with vertebrate Rhodopsin . Our group and others have previously demonstrated the rhythmic expression of exorh mRNA [43, 44]. In contrast to exorh mRNA, we find that Exorh protein expression does not follow a significant daily rhythm in WT embryos, nor is its expression affected in cyc mutants (Figure 3)[Additional File 1].
The functional significance of this difference in temporal expression pattern is unknown. There are other cases where mRNA displays cyclic expression, while the corresponding protein does not. In the chicken pineal gland, the mRNA encoding the photopigment Pinopsin shows a rhythm that is dependent upon activation by light . In zebrafish, interphotoreceptor retinoid-binding protein (irbp) mRNA expression is circadian-regulated, while and IRBP protein levels are constant [37, 63]. In the case of IRBP, it is proposed that the higher expression of mRNA during the day is necessary to maintain the constant levels of protein, as protein turnover is much higher during the day than during the night [37, 63]. This could be the explanation for the rhythmic pattern of exorh and pinopsin mRNA expression as well. Consistent with this, Pinopsin protein and mRNA levels decline rapidly when the embryos are placed in constant darkness .
More than two decades ago, a structure anatomically equivalent to the SCN was first described in a teleost fish (the goldfish) [64, 65]. Several years later, a morphologically similar group of neurons was described in zebrafish . However, no experimental evidence exists for a pacemaker role of the SCN in zebrafish or in other teleosts . Here, we establish that the zebrafish SCN is not required to establish circadian rhythms in the embryonic pineal. However, the amplitude of circadian gene expression was slightly reduced, and so we cannot rule out a role for the SCN or other tissues missing in cyc mutants in helping to establish or maintain rhythms in developing zebrafish. This will form a strong foundation for future studies that explore the communication between different circadian tissues during development, and for comparative studies of pineal biology and circadian regulation between vertebrate species.
Zebrafish embryos and larvae were obtained by natural matings and were raised at 28.5°C at 14:10 hour (h) L/D cycle according to standard procedures . Developing fish were placed in a Sanyo MIR-153 incubator (Amsterdam, The Netherlands) with heating and cooling capabilities for maximum temperature stability. Light consisted of a single 60 watt Globe EnerSaver light bulb placed within the incubator. Stocks used were Oregon AB (WT), ZDR (WT) (Aquatica Tropicals, Plant City, FL), and cyc m294 [25, 66–68].
WISH was carried out as described by Liang et al. . Antisense RNA probes included aanat2 , exorh , shh [69, 70], cry1b , cry3 (Open Biosystems, Huntsville, AL), otx5 , and avpl .
To generate cry1b probe, cry1b cDNA in plasmid pME-18S-FL3 was subcloned into plasmid pBSK(+) using a 5' EcoRI and a 3' NotI restriction sites. Plasmid pBSK+ cry1b was linearized with EcoRI and was used for probe synthesis with T3 RNA polymerase. cry3 probe was PCR amplified from the cry3 cDNA in plasmid pME-18S-FL3 using primers and cycling conditions in ZFIN [http://zfin.org/cgi-bin/webdriver?MIval=aa-markerview.apg&OID=ZDB-CDNA-040425-188]. cry3 mRNA was made with T3 RNA polymerase.
Whole mount antibody staining was done as described . A mouse monoclonal antibody (4D2) against the N-terminus of bovine Rhodopsin [72, 73] was used to detect Exorh protein at a dilution of 1:60. A goat anti-mouse secondary antibody coupled to Oregon Green 488 (Invitrogen Molecular Probes, Carlsbad, CA) was used at a dilution of 1:2000.
Control and translation blocking morpholinos (MO) against exorh (exorh atg MO) were obtained from Gene Tools (Philomath, OR). Sequences of the MO used were: control MO, 5'-CCT CTT ACC TCA GTT ACA ATT TATA-3' and exorh atg MO, 5'-AGT TGG GTC CCT CCG TCC CGT TCAT-3'. One-cell stage embryos were injected with either 1.5 nanograms (ng) of control or exorh atg MO using a Harvard Instruments PLI-90 Pico-Injector (Holliston, MA).
To generate exorh mRNA for overexpression, the entire coding sequence of the zebrafish exorh gene (from plasmid pCR2.1-full-length exorh) [GenBank Accession Number: AB025312] was subcloned into pGEMHE plasmid [40, 74]. Donor and host plasmids were digested with EcoRI. The resulting EcoRI-digested full-length exorh coding sequence was ligated non-directionally into pGEMHE (plasmid pGEMHE-full-length exorh hereafter pGEMHE-flex) using the TAKARA DNA Ligation Kit version 1 (Madison, WI). The ligation reaction was transformed into Z-Competent E. coli cells (Zymo Research, Orange, CA). Clone orientation was verified via restriction digestion (double digestion with SacI and BanII) and DNA sequencing. DNA sequencing was done using forward and reverse sequencing primers (forward primer: 5'-TTT TTG CAG AAG CTC AGA ATA-3'; reverse primer: 5'-CAT TTT GTA AAG TGT AAG TTG GTAT-3'). DNA sequences were verified by doing a BLAST search .
To synthesize full-length exorh mRNA, pGEMHE-flex plasmid was linearized with SphI. Linearized pGEMHE-flex was used for mRNA synthesis with an Ambion mMESSAGE mMACHINE kit (Austin, TX). Full-length exorh mRNA was verified through RNA formaldehyde agarose gel-electrophoresis. One blastomere of 8-16 cell stage embryos were injected with 400 picograms (pg) of beta-galactosidase (control) or full-length exorh mRNA and were fixed 7 h post-injection. Injected and fixed embryos were processed for whole mount antibody staining as described above.
Embryonic and larval zebrafish samples were imaged using a Zeiss Axioplan 2 Imaging Microscope (Thornwood, NY) or Nikon Eclipse 801 Epifluorescent Microcope (Melville, NY) connected to a Spot RTke7.4 Slider digital camera together with Spot 220.127.116.11 software (Diagnostic Instruments, Sterling Heights, MI). Resulting images were processed and quantified for optical density using ImageJ 1.36b and 1.42q software [Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997-2009]. Calibration was performed following developer's instructions [http://rsb.info.nih.gov/ij/docs/examples/calibration/]. For in situ hybridization data, images were first converted to 8-bit grayscale and an oval encompassing the expression domain was drawn using the Specify tool. Optical Density (OD) was calculated by multiplying the oval area with the average intensity, both of which were obtained through the Analyze tool. Two-way Analysis of Variance (ANOVA) statistical analysis was carried out using Microsoft Excel.
For quantification of aanat2 expression in the constant light and constant dark experiments, the background value was subtracted as follows. A small circle with diameter of 20 pixels was selected close to, but not touching the anterior edge of the WISH expression signal. The circular area and optical density was determined as above. Background-corrected OD values were calculated by subtracting the (OD value of small circle × pineal oval area/small circle area) from the OD of oval around the pineal expression domain. All OD data are means ± standard deviation. Representative images with ODs closest to the mean were chosen for figures.
Fluorescent images were analyzed as above with the exception that 8-bit grayscale images were first processed using the Inverted tool prior to choosing the oval area to be quantified for OD. For Exorh protein temporal expression, fluorescent image data was tested for significance using One-way ANOVA and Tukey's test using Origin Software Version 7.5 SR4 (Northampton, MA).
To measure the dimensions of the otx5 expression domain, lines covering the length or the width of the pineal were drawn using ImageJ and the corresponding pixel length or width were converted to micrometers by calibrating the number of pixels in a 20 micrometer line.
We would like to thank Drs. Josh Gamse (Vanderbilt University, Nashville, TN), Eric Mintz (Kent State University, Kent, OH), and Neal Peachey (Cleveland Clinic, OH) for their helpful comments on this manuscript, Dr. Yoshitaka Fukada (University of Tokyo, Japan) for providing the full-length exorh cDNA, Dr. Michael Romero (Mayo Clinic, Rochester, MN) for sharing the pGEMHE plasmid and sequencing primers and advice on subcloning, Dr. Robert Molday (University of British Columbia, Canada) for providing the 4D2 antibody, Sandra Albro (Case Western Reserve University) for assistance with statistical analysis, and Drs. David Klein (National Institutes of Health), Yoav Gothilf (Tel-Aviv University, Israel), and Bernard and Christine Thisse (University of Virginia) for plasmids. Allisan Aquilina-Beck, Adelle Schuman, Jessica Clay, Stephan Brannan, Kristine DiMonte, Lain Pierce, Olga Ponomareva, Perry Hwang, Gerry Babcock, Andy Schreiner, and Brian Chen provided expert technical assistance. Supported in part by research grant No. 5-FY02-259 from the March of Dimes Birth Defect Foundation (to JOL), an Edward Mallinckrodt, Jr. Foundation grant (to JOL), 1 RO1 HD054523 (to JOL and Joshua Gamse), the Hathaway Brown Student Research Program (to LGK), and a Phi Beta Kappa Student Research Grant Award (to RRN).
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.