Pleiotropic functions of embryonic sonic hedgehog expression link jaw and taste bud amplification with eye loss during cavefish evolution

Abstract

This study addresses the role of sonic hedgehog (shh) in increasing oral pharyngeal constructive traits (jaws and taste buds) at the expense of eyes in the blind cavefish Astyanax mexicanus. In cavefish embryos, eye primordia degenerate under the influence of hyperactive Shh signaling. In concert, cavefish show amplified jaw size and taste bud numbers as part of an adaptive change in feeding behavior. To determine whether pleiotropic effects of hyperactive Shh signaling link these regressive and constructive traits, sonic hedgehog (shh) expression was compared during late development of the surface- (surface fish) and cave-dwelling forms of Astyanax. After an initial expansion along the midline of early embryos, shh was elevated in the oral pharyngeal region in cavefish and later was confined to taste buds. The results of shh inhibition and overexpression experiments indicate that Shh signaling has an important role in oral and taste bud development. Heat shock mediated activation of an injected shh transgene at specific times in development showed that taste bud amplification and eye degeneration are sensitive to shh overexpression during the same early developmental period, although taste buds are not formed until much later. Genetic crosses between cavefish and surface fish revealed an inverse relationship between eye size and jaw size/taste bud number, supporing the linkage between oral pharyngeal constructive traits and eye degeneration. The results suggest that hyperactive Shh signaling increases oral and taste bud amplification in cavefish and suggest that selection for constructive oral pharyngeal traits may be responsible for eye loss via pleiotropic function of the Shh signaling pathway.

Introduction

Cave animals have evolved novel morphological, developmental, physiological, and behavioral phenotypes during the relatively short time since they diverged from surface dwelling ancestors (Culver, 1982). The Mexican tetra Astyanax mexicanus, which consists of a sighted surface dwelling form (surface fish) and a series of blind cave dwelling forms (cavefish), is an emerging model system for studying development and evolution of cave adapted phenotypes (Jeffery, 2008). Like many other cave-adapted animals, Astyanax cavefish have lost their eyes and pigmentation during evolution in perpetual darkness. In concert with regressive evolution, constructive traits have also evolved, including additional gustatory organs (taste buds) and changes in feeding behavior (Schemmel, 1967; 1980; Hüppop, 1967; Jeffery, 2001), which are probably adaptive in the cave environment. It has been postulated that non-visual sensory systems may have been enhanced to compensate for loss of vision during cavefish evolution (Voneida and Fish, 1984; Teyke, 1990; Jeffery et al., 2000; Jeffery, 2001) but the responsible molecular systems have not been identified. Genetic studies have revealed overlapping quantitative trait loci (QTL) governing eye size and increased gustatory organs (taste buds), which could be explained by pleiotropic tradeoffs (Protas et al., 2008). Here we address the possible pleiotropic function of sonic hedgehog (shh) in linking the gain of oral and gustatory constructive traits to the loss of eyes in blind cavefish embryos.

Despite the absence of functional eyes in adults, small eye primordia with a lens and optic cup are initially formed in cavefish embryos but subsequently arrest in development, degenerate, and sink into the orbits, where they are covered by connective tissue and epidermis (Cahn, 1958; Langecker et al., 1993; Jeffery and Martasian, 1998). As a first step in eye degeneration, the cavefish lens undergoes apoptosis (Jeffery and Martasian, 1998; Yamamoto and Jeffery, 2000; Soares et al., 2004). Later in cavefish development, the dysfunctional lens fails to induce the anterior eye chamber, iris, and cornea, although a normally layered initially develops from the optic cup. Photoreceptor cells are formed in the layered retina but subsequently degenerate (Langecker et al., 1995; Yamamoto and Jeffery, 2000). The surface fish lens can restore eye development, including the anterior sector and a retina with photoreceptor cells, after transplantation into the cavefish optic cup (Yamamoto and Jeffery, 2000), indicating that the lens has a fundamental role in sustaining eye development (Strickler et al., 2007a). Several factors have been discovered that may induce apoptosis in the cavefish lens. Two of these are the antiapoptotic factor αA-crystallin, which is downregulated in the cavefish lens (Strickler et al., 2007b) and maps near an Astyanax eye loss QTL (Gross et al., 2008), and the putative proapoptotic factor Hsp90α, which is upregulated during cavefish lens development (Hooven et al. 2005). The third is shh, which induces lens apoptosis following overexpression in surface fish embryos (Yamamoto et al., 2004).

To investigate the molecular basis of eye degeneration, we previously compared the expression of a number of eye regulatory genes in cavefish and surface fish embryos (Strickler et al., 2001; Yamamoto et al., 2004; Jeffery, 2005). These studies pointed toward genes in the Shh midline signaling system as regulators of cavefish eye regression. First, we observed that the bilateral eye domains of pax6 expression in the cavefish neural plate are reduced and separated by a larger gap along the dorsal anterior midline. Second, we showed that shhA and shhB (formerly tiggy winkle hedgehog) expression is increased along the anterior midline (prechordal plate) in early cavefish embryos. Third, we found that expression of downstream genes in the Sonic Hedgehog (Shh) signaling pathway, such as the receptor patched, nkx2.1a in the neural plate, and pax2a and vax1 expression in the optic vesicles, is also amplified, implying Shh hyperactivity along the cavefish anterior midline. Vertebrate optic vesicles are patterned by reciprocal transcriptional repression between pax6 and pax2/vax1 (Schwarz et al., 2000; Take-uchi et al., 2003), and upregulation of the latter by Shh signals is partially responsible for the small cavefish eye. Together with effects on the lens, shh mediated changes in gene expression in the optic cup suggest that the Shh signaling pathway negatively controls cavefish eye development.

Because shh is a pleiotropic gene with both positive and negative roles in development (Ingham and McMahon, 2001), in addition to negative effects on eye development, Shh hyperactivity could be related to the evolution of constructive traits, such as taste buds. Taste buds are more numerous in adult cavefish than in surface fish (Schemmel, 1967; Boudriot, and Reutter, 2001 (Schemmel, 1980), and this expanded gustatory sense may be beneficial for cave life. Overexpression of shh has been previously detected in the ventral forebrain and Shh signaling domains in the developing cavefish brain (Menuet et al., 2006) but feeding structures have not been investigated. Here, we have followed shh expression during oral pharyngeal development to identify features that may be under positive control of pleiotropic Hh signaling. We found that shh expression is expanded in the oral pharyngeal region and is later expressed in taste buds. The results of functional experiments suggest that shh amplification is required for increasing taste bud number during the same developmental interval as it inhibits eye development. In addition, genetic crosses reveal an antagonistic relationship between eye size and taste bud number in Astyanax. The results support the possibility that increased oral and gustatory development may have occurred at the expense of eyes during cavefish evolution via pleiotropic effects of the Shh signaling pathway.

Materials and Methods

Animals and Embryos

Laboratory colonies of Astyanax mexicanus were derived from surface fish collected at Balmorhea Spring State Park, Texas and cavefish collected at Cueva de El Pachón, Tamaulipas, Mexico. Embryos were obtained by temperature induced spawning and reared at 25°C (Jeffery and Martasian, 1998; Jeffery et al., 2000).

In Situ Hybridization

RNA probes were generated from surface fish shh (AY661431), nkx2.1a (AY661435), and pax2a (AY661436) cDNA sequences as described previously (Yamamoto et al., 2004). Embryos or larvae were fixed in 4% paraformaldehyde-PBS (pH 7.2; PFA). In situ hybridization was done using digoxygenin-labeled RNA probes as described previously (Strickler et al., 2001; Yamamoto et al., 2004). Following in situ hybridization the specimens were post-fixed in PFA, dehydrated through an ethanol series, embedded in polyester wax, and sectioned at 10 μm. In situ hybridized specimens were viewed as whole mounts or sections and photographed.

Quantitative Real Time RT-PCR

Total RNA was extracted from 3-day post-fertilization (dpf) larvae with Ribopure kit (Ambion, Austin, TX) according to the manufacturer’s protocol. Extracted RNA was quantified and its integrity verified using the UV absorbance (260/280) bioanalyzer (Agilent Technologies, Palo Alto, CA). Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA) was used to create cDNA from 1μg of RNA according to the Introvgen protocol using an oligo (DT) primer (5′-CGGAATTCTTTTTTTTTTTTTTTTTTTTV- 3′, Sigma Genosys, The Woodlands, TX). Blank cDNA was also created with total RNA as described, but with no reverse transcriptase, to serve as a negative control for genomic contamination. mRNA levels were measured by quantitative real time RT-PCR (RT-qPCR) using 2μl of diluted cDNA (1:100) in a 20μl qPCR reaction with SYBR Green ER qPCR SuperMix using an iCycler (Invitrogen, Carlsbad, CA) and analyzed according to the manufacturer’s protocol with the iCycler iQ Real-Time PCR Detection System (Bio-Rad, Hercules, CA).

Primers were designed using Primer Express (v 2.0, Applied Biosystems) and either a known A. mexicanus sequence (see below) or the homologous region between zebrafish and Tetraodon nigroviridis cDNAs (for β-actin). The qPCR products were verified for the appropriate size by dissociation curve analysis and gel electrophoresis. Primers were 18–30 nucleotides in length with a melting temperature between 58–64 C. The primer sequences were as follows: shh (AY661431) forward primer, 5′-AGCGCTTCAAGGAGCTCATC -3′ and reverse primer, 5′-CGTGTTCTCCTCGTCCTTAAAGA-3′; vax1 (AY661437) forward primer, 5′-TCTACAGGCTGGAGATGGAGTTC-3′ and reverse primer, 5′ TTGAGTTGGCGTGCAAGCT-3′; pax2a (AY661436) reverse primer, 5′-GCACGACTTCTCCACCCGTAT-3′ and reverse primer, 5′-GATGCCGTTGATGGAGTAGGA-3′; pax6 (AY651762) forward primer, 5′-TGGCTGCCAGCAATCAGATG-3′ and reverse primer, 5′-CTTCTGAGTCCTCCCCATTTGAG-3′; α-actin (Strickler and Jeffery, unpublished) forward primer, 5′-CACGGCATCATCACCAACTG- and reverse primer, 5′-CCACACGGAGCTCGTTGTAGA-3′, and β-actin forward primer, 5′ CACACMGTGCCCATCTAYGA-3′ and reverse primer, 5′ CRGCARATCCAGACGCAGRAT-3′. The qPCR output provided a Ct value for the threshold cycle, which is representative of fluorescence derived from binding of SYBR green to the double-stranded PCR product. Data were transformed to a ΔCt value by subtracting the sample Ct value from the sample with the highest expression level in order to control for amplification efficiency. The ΔΔCt value was then calculated by normalizing gene expression to α- and β-actin using the geNorm software and methods (GeNorm v3.4, Vandesompele et. al., 2002).

All levels of gene expression were compared using a one-way ANOVA with cavefish and surface fish as the independent variables, and relative mRNA levels as the dependent variable. Values are reported as means ± SE, and p< 0.05 was required for significance. Statistica v.6.1 (StatSoft, Inc., Tulsa, OK) was used for data analysis and Graphpad was used to construct graphs (Graphpad Prism Version 4.0, Graphpad Software, Inc.).

Shh Inhibition

Shh activity was inhibited in two ways. First, Shh translation was inhibited by morpholinos. A shh MO (5′-GCCGTGGCGGAGCCGTGCGTAAAA-3′) was designed by Gene Tools Inc. (Summerton, OR) against part of the 5′ UTR of surface fish shh cDNA, a region in which cavefish and surface fish cDNAs do not differ in sequence. Embryos were injected with 2 ng shh or control (5′-CCTCTTACCTCAGTTACAATTTATA-3′) MOs at the 2–4 cell stage. In controls, MO injected embryos were subjected to in situ hybridization with probes for the shh-regulated genes nkx2.1a and pax2a. In rescue experiments, embryos were injected with 2 ng/ml shh MO and 10 pg/ml shh mRNA (see below). Second, embryos were treated with 20 μm or 200 μM cyclopamine (Sigma, St. Louis, MO) beginning at 10 and ending at 20 h post-fertilization (hpf) as described by Menuet et. al. (2007), then washed into water and allowed to develop until 5 dpf. Controls were treated with 0.1% ethanol (which was used to prepare the cyclopamine stock solution) for the same time interval.

Shh Overexpression

Shh activity was increased in two ways. First, 20–800 pg shh mRNA was injected into 2–4 cell embryos. Synthetic mRNAs were prepared and injected as described previously (Yamamoto et al., 2004). Control embryos were injected with Green Fluorescent Protein (GFP) mRNA. Second, to determine the effect of shh overexpression at different times in development, we prepared and injected embryos with the DNA expression construct hsp70:shh:GFP. This DNA construct consists of the hsp70 promoter, the coding sequence of zebrafish shh, the coding sequence of GFP (fused to the C-terminus of shh), and the SV40 polyadenylation signal, all flanked by ISceI meganuclease sites. An intermediate backbone for this vector was constructed by replacing the PstI-XbaI fragment from the vector described in Pyati et al. (2005) with a short sequence containing a BamHI restriction site. The zebrafish shh coding sequence was then amplified by PCR (Expand High Fidelity, Roche) from total zebrafish cDNA using the primers aGGATCCagccaccatgcggcttttgacgaga and aTCTAGAgcttgagtttactgacatccccaa and ligated into the BamHI/XbaI sites of the intermediate vector. The DNA construct was prepared for injection by digesting 600 ng of plasmid DNA with 10 units of I-Sce I meganuclease (New England Bio Labs) in the buffer supplied by the manufacturer for 1 h at room temperature and then stored at −20C. We injected 30 pg of hsp70:shh:GFP DNA into 2–4 cell embryos and overexpresed shh by applying 36 C heat shocks for 1 h at various times in development. The effects of heat shocks on shh overexpression were monitored by following GFP expression.

Cartilage Staining and Jaw Width Measurements

Jaw cartilages were stained with Alcian Blue at 5–6 dpf as described by Yamamoto et al. (2003). Jaw width was measured across the hinge region in whole mounts of fixed larvae. Statistical analysis of jaw measurements was performed by Student’s unpaired (independent) t test.

Taste Bud Detection and Quantification

For taste bud detection, embryos were raised to 6 dpf and fixed in two changes of fresh PFA for 2 h at room temperature. After washing in PBS, embryos were stained with calretinin antibody (1:1000 dilution; Swiss Antibodies, Bellinzona, Switzerland) and antigen-antibody complexes were detected using a biotinylated goat anti-rabbit secondary antibody (1:500 dilution) and the Vectastrain ABC Peroxidase kit (Vector Laboratories, Burlingame, CA), as described by Jeffery et al. (2000). The immunostained specimens were viewed as whole mounts and photographed.

Taste buds were counted on the upper and lower lips of calretinin stained specimens viewed under a stereoscope or compound microscope. The rosette-like shape of taste buds distinguished them from much smaller solitary mechanoreceptor cells on the lips, which also stain positively for calretinin. Statistical analysis was carried out as described above for jaw width.

Mating Experiments to Produce Small- and Large-eyed Hybrids

Cavefish were crossed with surface fish to produce an F1 generation. The F1 hybrids were interbred to produce an F2 generation, and the F2 hybrids were interbred to produce an F3 generation. The eye size of F3 hybrids was measured in living specimens at 6 dpf viewed under a compound microscope. Small-eyed F3 hybrids showed eye diameters lower than 268μm (range = 192–268 μm). Large eyed F3 hybrids showed eye diameters higher than 290 μm (range = 290–330 μm). After calretinin staining, jaw spans and taste bud numbers of small- and large-eyed F3 hybrids were determined and subjected to statistical analysis using unpaired Student’s t tests as described above.

Results

Shh Overexpressionn in the Cavefish Oral-pharyngeal Region

Previous studies showed that shh expression is expanded along the midline during early cavefish development (Yamamoto et al., 2004) and later in Shh signaling centers in the forebrain (Menuet et al., 2007). During zebrafish (Miller et al., 2000; Eberhart et al., 2006), chick (Marcucio et al., 2005; Haworth et al., 2007), and mouse (Yamagishi et al., 2006) development, shh expression is also prominent in the oral ectoderm and pharyngeal endoderm (oral-pharyngeal region). Accordingly, we asked if shh expression is also overexpressed in the cavefish oral-pharyngeal region at later stages of development.

In situ hybridization showed expanded shh expression along the cavefish anterior midline at 1 dpf (Fig. 1A, B). At 2 dpf a larger shh expression domain was observed in the oral pharyngeal region (Fig. 1C–F), outlining a wider mouth, in cavefish relative to surface fish (Fig. 1E–F). By 3 dpf shh expression in the oral-pharyngeal region was attenuated to tooth germs (Stock et al., 2006) and taste buds (Jeffery et al., 2000) (Fig. 1G–L). Taste buds can be distinguished from tooth germs by their abundance, positioning in single file along the lips, ring-like shh expression pattern (Fig. 1I, J), and staining by calretinin antibody (Fig. 1M; Jeffery et al., 2000). Sections though the oral area showed shh expression confined to the marginal (or basal) cells in each taste bud rosette (Fig. 1N). The marginal cells may be stem/precursor cells for taste receptor cells (Miura et al., 2006). No differences were apparent in the identity or number of shh-expressing cells in surface fish and cavefish taste buds.

Figure 1.

Amplified shh expression in the cavefish oral pharyngeal region. A– D. Dorsal anterior (A, B) and lateral (C, D) views of 1 dpf surface fish (A, C) and cavefish (B, D) embryos showing expanded shh expression in the cavefish anterior midline (A, B arrowheads). C–F. Lateral (C, D) and rostral (E, F) views of 2 dpf surface fish (C, E) and cavefish (D, F) embryos showing expanded shh expression in the cavefish oral epithelium (o). O: oral area. OC-P: Oral-pharyngeal cavity. G–J. Ventral views of 3 dpf surface fish (G, I) and cavefish (H, J) embryos showing shh expression in taste buds (upward pointing arrowheads in I) and primary tooth germs (oblique pointing arrowheads in I) on the lips. I, J are two- fold magnifications of G, H showing the ring-like shh expression pattern in taste buds. K–N. Sections through 3 dpf surface fish (K) and cavefish (L, N) comparing the patterns of shh expression (K, L, N) and calretinin staining (M) in taste buds. MC: marginal cells. Scale bars: A (100 μm), E (50 μm), K (20 μm); M (4 μm); magnification is the same in A–D, E–H, I and J, K and L, M and N. O. Quantification by qRT-PCR showing increased levels of shh, vax1, and pax2a mRNA and decreased levels of pax6 mRNA relative to β-actin and α-actin mRNA in 3 dpf cavefish larvae. Asterisk: p<0.05 in one-way ANOVAs comparing cave and surface fish mRNA levels (n=4).

We quantified shh expression by qPCR at 3 dpf. As shown in Figure 1O, about 3 fold higher shh RNA levels were detected in cavefish relative to surface fish embryos, which would include the sum of expression in the oral-pharyngeal region, the brain, and possibility other embryonic regions. Furthermore, vax1 and pax2a mRNA levels, which are positively controlled in eyes by Shh signaling (Ekker et al., 1995; Take-uhi et al., 2003), are increased, whereas pax6 mRNA, which is negatively controlled in eyes by Shh signaling (Macdonald et al., 1995), is decreased in cavefish embryos (Fig. 1O). The results show that shh expression is amplified in the oral pharyngeal region in cavefish relative to surface fish embryos and later expressed in taste buds.

Oral-pharyngeal Features Are Enhanced in Cavefish

We next asked whether any differences in cavefish oral-pharyngeal and taste bud development correlate with increased shh expression. We observed that the expanded cavefish mouth, which is outlined by shh expression at 2 dpf (Fig. 1E, F), presages a wider jaw span later in development (Fig. 2A–D, G; Table 1). We also compared taste buds in embryos of the two forms of Astyanax because it has been reported that cavefish adults exhibit more taste buds than surface fish, particularly in the external surface of the lower jaw (Schemmel, 1967; Bensouilah and Denizot, 1991; Boudriot, and Reutter, 2001). Previous studies showed that Astyanax embryos begin to form calretinin-positive taste buds at 3–4 dpf (Jeffery et al., 2000). At 5–6 dpf taste buds were stained with calretinin antibody throughout the oral pharyngeal region, including the upper and lower lips (Fig. 2E, F). In contrast to adults, only a few taste buds were seen on the ventral surface of the lower jaw, and their number did not differ in cavefish and surface fish embryos. Calretinin antibody also stained solitary mechanoreceptor cells on the lips and head and the cranial nerve fibers (Fig. 2C, D), as reported in another teleost (Diaz-Regueira et al., 2005), but calretinin-stained taste buds were clearly distinguishable by their rosette-like morphology (see Fig. 1M). We observed significant elevations in taste bud number on the upper and lower lips in cavefish relative to surface fish (Fig. 2H, I; Table 1). The results show that cavefish embryos have larger jaws with more taste buds than their surface fish counterparts.

Figure 2.

Constructive oral-pharyngeal features in cavefish. A–D. Dorsal (A, B) and ventral (C–F) views 6 dpf surface fish (A, C, E) and cavefish (B, D, F) showing wider jaw span (A, B; double-headed arrows), larger Alcian Blue stained mandibles (C, D), more calretinin-stained taste buds (E, F; upward pointing arrowheads), and wider oral palates (E, F; doubled headed arrows) in cavefish. Scale bars in A is 100μm; magnification is the same in A–D and E, F. G–I. Surface fish (top frames) and cavefish (bottom frames) show differences in jaw width (G, red bars) and taste bud numbers on the upper (H, blue bars) and lower (I, black bars) lips. Jaw width is indicated in units of 20 μm with unit 1 as 371–390 μm, unit 2 as 391–410 μm, and so forth.

Table 1.

Quantification of Jaw Width and Taste Bud Number

Manipulation Type	Form	N	Stage	Mean JW (μm) +/− SD	Mean Taste Bud Number +/−SD		Significance (p)
					UJ	LJ	JW	UJ	LJ
None	SF	17	6 dpf	415 +/− 21	11.6 +/− 1.8	13.3 +/− 2.0
None	CF	22	6 dpf	524 +/− 32	14.1 +/− 2.8	16.2 +/− 1.7	10.004	0.000	0.000
Control MO injection	CF	29	6 dpf	512 +/− 52	11.4 +/− 2.9	13.4 +/− 3.3
shh MO injection	CF	7	6 dpf	49 +/− 47	0.9 +/− 1.2	2.4 +/− 2.0	20.000	0.000	0.000
shh MO + shh mRNA injection	CF	16	6 dpf	174 +/− 126	2.1 +/− 4.2	9.2 +/− 6.8	30.400	0.850	0.018
Cyclopamine (control)	CF	11	5 dpf	Not measured	13.6 +/− 2.5	13.1 +/− 1.3
Cyclopamine (20μM)	CF	14	5 dpf	Not measured	11.9 +/− 2.3	14.2 +/− 1.6
Cyclopamine (100 μM)	CF	6	5 dpf	Not measured	10.9 +/−3.0	10.2 +/− 3.0
Cyclopamine (200 μM)	CF	10	5 dpf	Not measured	0	0
GFP mRNA injection	SF	38	6 dpf	411 +/− 54	11.1 +/− 2.7	13.2 +/− 2.9
shh mRNA+ injection	SF	16	6 dpf	506 +/− 75	16.1 +/− 4.1	19.2 +/− 5.8	40.000	0.000	0.000
shh mRNA++ injection	CF	27	6 dpf	Not measured	21.5 +/− 17.6	35.2 +/− 11.5
shh heat shock	SF	28	TB	408 +/− 51	12.9 +/− 2.7	16.5 +/− 3.3	50.369	0.010	0.019
shh heat shock	SF	23	1 somite	405 +/− 32	13.0 +/− 2.1	16.0 +/− 2.7	50.399	0.007	0.039
shh heat shock	SF	27	1 dpf	391 +/− 75	11.1 +/− 2.5	14.6 +/− 3.0
shh heat shock	SF	28	2 dpf	391 +/− 61	11.0 +/− 2.7	14.7 +/− 4.0
shh heat shock	SF	36	2.5 dpf	376 +/− 90	10.2 +/− 2.4	13.4 +/− 3.5
shh heat shock	SF	31	3 dpf	376 +/− 64	10.1 +/− 3.8	13.6 +/− 4.3
Small eye	F3	30	6 dpf	465 +/− 73	15.5 +/− 2.4	14.4 +/− 1.6	60.004	0.000	0.003
Large eye	F3	34	6 dpf	405 +/− 27	11.4 +/− 2.0	12.0 +/− 2.2

N sample number. JW: jaw width. UJ: upper jaw. LJ: lower jaw. SD: Standard Deviation.

SF: Surface fish. CF: Cavefish. F3: F3 hybrid progeny of SF X CF cross.

⁺20 pg shh mRNA injected.

⁺⁺800 pg shh mRNA injected.

Statistical comparisons:

¹CF versus SF;

²shhMO versus control MO injection;

³shhMO versus shhMo + shh mRNA injection;

⁴shh mRNA versus GFP injection;

⁵heat shock at the tailbud stage or 1-somite stage versus heat shock at 1 dpf;

⁶small eyed- versus large eyed hybrids.

Shh Downregulation Reduces Oral-pharyngeal Development

The possibility that jaw width and taste bud number are controlled by Shh signaling was investigated by determining the effects of manipulating shh expression levels. Shh activity was downregulated by shh morpholino injection and cyclopamine treatment (Fig. 3; Table 1). First, we injected translation-blocking shh MOs into early embryos (Nasevicius and Ekker, 2000). The effects of shh inhibition in the MO injected embryos was evaluated by monitoring the expression of nkx2.1a and pax2.1a genes in the neural plate (Fig. 3A, B). We found that shh but not control MOs blocked nkx2.1a expression, which is positively regulated by Hh signaling (Pabst et al., 2003), but had less effect on pax2.1a expression, which is independent of shh at the midbrain-hindbrain boundary, suggesting that Shh activity was downregulated. In shh MO-injected cavefish embryos, the size of the oral-pharyngeal area was reduced, shifting the mouth opening posteriorly along the longitudinal body axis (Fig. 3E, F). Morphants subsequently showed significant decreases in jaw width and taste bud number on their upper and lower lips (Fig. 3C, D, G, H, I; Table 1). Similar results were obtained with another MO directed against a splice site in the second shh intron (data not shown). Simultaneous injection of zebrafish shh mRNA with shh translation blocking MOs partially alleviated the effects on lower jaws and taste buds (Fig. 3I; Table 1). Second, Shh activity was downregulated by cyclopamine treatment (Menuet et al., 2007). In these experiments, embryos were treated with 20 or 200 μm cyclopamine beginning 15 hpf, then at 1 dpf the treated embryos were washed into water lacking the inhibitor, and at 5 dpf the effects on oral-pharyngeal features were determined (Table 1). Embryos treated with 20μm cyclopamine showed similar taste bud numbers to controls, whereas those treated with 200μm cyclopamine had very small mouths with no detectable taste buds. The results show that Shh inhibition reduces the extent of oral and taste bud development.

Figure 3.

Effect of MO-mediated shh inhibition on oral and taste bud development. A–H. Cavefish were injected with control (A, C, E, G) or shh (B, D, F, H) MOs and analyzed at the tailbud stage (A, B) or 6 dpf (C–H). A, B. In situ hybridization showing downregulation of nkx2.1a but not pax2a (asterisks) expression in shh MO injected embryos at the neural plate stage. C–F. Reduced jaw span (C, D; double headed arrows) and oral pharyngeal region (F, arrowhead) in shh MO injected larvae at 6 dpf. C, D: Ventral views. E, F: Lateral views. Double headed arrows: jaw span. G, H. Reduced numbers of calretinin-stained taste buds are formed in 6 dpf cavefish larvae injected with shh MO. G: Ventral view. H. Anterior view. Downward and upward pointing arrowheads indicate upper and lower jaws respectively. Scale bars: A (250 μm), C, E, and G (200 μm); same magnification in A and B, C and D, E and F, G and H. I. Reduced jaw span (μm; red bars) and taste bud numbers on the upper (blue bars) and lower (black bars) lips in 6 dpf cavefish injected with shh MO (middle three bars) compared to control MO (left three bars). Injection of a mixture of shh MO and zebrafish shh mRNA decreases the effects on jaw width and taste bud number (right three bars). JW: jaw width. UJ: Upper jaw. LJ: Lower jaw. Error bars indicate SE of the mean.

Shh Upregulation Amplifies Oral-pharyngeal Development

The effects of Shh overexpression were determined by shh mRNA injection (Fig. 4). First, 20 pg of shh mRNA, a concentration known to promote eye degeneration (Yamamoto et al., 2004), was injected into surface fish embryos and the effects on jaw and taste bud development were determined. Embryos injected with shh mRNA showed lateral expansion of nkx2.1a (Fig. 4A, B), consistent with effective Shh overexpression and eye degeneration at 6 dpf (Fig. 4E). The injected surface fish embryos showed significant increases in jaw width and taste bud numbers with respect to controls (Fig. 4C-F; Table 1). Second, a large excess of shh mRNA (800 pg) was injected into cavefish embryos. When the latter were examined at 6 dpf, most of them showed large gaping mouths with lips containing 2–3 fold more taste buds than controls (Fig 4G, H, I; Table 1). The results indicate that shh overexpression increases jaw size and taste bud number.

Figure 4.

Effect of shh overexpression on oral and taste bud development. A–H. Surface fish (A–F) or cavefish (G, H) embryos were injected with shh (B, E, F–H) or GFP (A, C, D) mRNAs and analyzed at the tailbud stage (A, B) or 6 dpf (C–H). A, B. In situ hybridization showing expansion of nkx2.1a but not pax2a expression in the neural plate of cavefish embryos injected with shh MO (B). C–H. Increase in the oral pharyngeal region and calretinin-stained oral taste bud numbers (arrowheads) in shh mRNA injected surface fish (C–F) and cavefish (G, H) embryos. Lateral (C, E, G), ventral (D, F), and anterior ventral (H) views at 6 dpf. (C–F). 20 pg shh mRNA. (G, H). 800 pg shh mRNA (E, F). DE: pigmented remnant of degenerate eye. Arrowheads: calretinin stained taste buds. Doubled headed arrows: mouth opening. Scale bars: A (250 μm), C (200 μm); magnification is the same in A and B, C–F. I. Increased jaw span (red bars) and taste bud numbers on the upper (blue bars) and lower (black bars) lips in 6 dpf larvae that developed from embryos 20 pg shh mRNA (middle) or 800 pg shh mRNA (right) relative to controls injected with 20 pg GFP mRNA (left). Error bars indicate SE of the mean. JW: jaw width. UJ: Upper jaw. LJ: Lower jaw.

Shh Upregulation Positively Affects Taste Buds and Negatively Affects Eyes During the Same Early Developmental Period

To determine the developmental interval in which taste buds and eyes are sensitive to shh upregulation, surface fish embryos were injected with the hsp70:shh:GFP transgene and subsequently heat shocked at various stages of development (Fig. 5). Similar to the results obtained when shh mRNA was injected into 2–4 cell embryos (Fig. 4C–F), heat shocks at the tailbud (8 hpf) or one-somite (10 hpf) stages increased taste bud numbers on the upper and lower lips to levels resembling cavefish (Fig. 5A, C; Table 1). Similarly, heat shocks at the tailbud and one-somite stages also induced eye degeneration (Fig. 5A, C). The increases in jaw width at these stages were not significantly different from normal surface fish (Table 1), but visual inspection indicated that some heat shocked embryos showed a larger oral pharyngeal region relative to controls (Fig. 5A). In contrast, shh overexpression at 1, 2, 2.5, or 3 dpf resulted in taste bud numbers and levels of eye development resembling normal surface fish (Fig. 5B, C; Table 1). The differences observed between shh overexpression before and after 1 dpf were significant (Table 1). The results show that positive effects on taste bud development and negative effects on eye development can be induced together by shh overexpression prior to 1 dpf, although taste buds are first apparent morphologically 2–3 days later (Jeffery et al., 2000).

Figure 5.

The effects of shh overexpression on oral-pharyngeal development and eye degeneration in surface fish embryos injected with an hsp70:shh:GFP transgene and heat shocked at various stages of development. A, B. A transgene injected embryo (6 dpf) heat shocked at the tailbud (TB) stage showing a gaping mouth, enlarged forebrain (FB), and small degenerate eyes. C, D. A transgene injected embryo (6 dpf) heat shocked beginning at 1 dpf showing a normal mouth and eye. A and C: dorsal views B and D: lateral views. Scale bar in D is 200 μm.; magnification is the same in A–D. E. The shh sensitivity periods for increased taste bud development and eye degeneration in transgene injected surface fish embryos determined by heat shocking at different developmental stages. Red bars. oral width (μm). Blue bars: taste bud number on upper lips. Black bars: taste bud number on lower lips. Error bars indicate SE of the mean. Blue dots: Percentage of embryos with normal eye development. Error bars indicate SE of the mean. JW: jaw width. UJ: Upper jaw. LJ: Lower jaw. Asterisk indicates that the pooled differences between taste buds on the upper and lower jaws at the tail bud and 1-somite stage are significantly different from those at 1, 2, 2.5, and 3 dpf (p = 0.003).

Inverse Relationship Between Oral-pharyngeal Traits and Eye Development

The results described above opened the possibility that oral and taste bud development may be linked with eye development through the positive and negative effects of expanded Shh signaling. To test this hypothesis independently, we measured oral pharyngeal traits in small-and large-eyed surface fish X cavefish hybrids. To create these hybrids, the F1 progeny of surface fish X cavefish crosses were mated to produce an F2 generation, and the latter mated to produce an F3 generation. Cavefish eye regression is a multigenic trait (Wilkens, 1988). Accordingly, F3 hybrids showed a broad distribution of eye sizes, including large normal eyes resembling those of surface fish and small degenerating and sometimes de-pigmented eyes, resembling those of cavefish (Fig 6A, B). The small-eyed and large-eyed F3 progeny were fixed at 6 dpf, and jaw sizes and taste buds measured as described above. The results showed that small-eyed hybrids have significantly larger jaws and more jaw taste buds than large-eyed hybrids (Fig. 6C–E; Table 1). Thus, these experiments reveal an inverse relationship between eye size (e. g. extent of eye regression) and oral/taste bud development: hybrids with small degenerating eyes have the cavefish taste bud phenotype, whereas hybrids with large normal eyes show the surface fish taste bud phenotype.

Figure 6.

Relationship between oral pharyngeal traits and eye size in the F3 hybrid progeny of a surface fish X cavefish cross. A. Examples of small- (A) and large- (B) eyed hybrid. The eye(s) of small-eyed hybrids are sunken into the orbit and unpigmented, resembling those of cavefish, whereas the eyes of large-eyed hybrids are exposed and pigmented, resembling those of surface fish. C–E. Differences in jaw width (red bars) and taste bud numbers on the upper (blue bars) and lower (gray bars) lips of small- and large-eyed F3 hybrids. Jaw width is indicated in units of 20 μm with unit 1 as 331–350 μm, unit 2 as 351–370 μm, and so forth.

Discussion

The present investigation has revealed a link between constructive oral pharyngeal development and eye regression via the pleiotropic Shh pathway in the blind cavefish Astyanax mexicanus. The results support the following general conclusions. First, the expansion of shh expression, previously reported along the embryonic midline (prechordal plate) in early cavefish embryos (Yamamoto et al., 2004) continues in the oral-pharyngeal region and taste buds later in cavefish development. Second, jaw size and oral taste bud numbers are increased in cavefish embryos and these constructive traits can be manipulated by Shh inhibition or overexpression. Third, eye degeneration and increased taste buds show the same shh sensitive periods during early development, although taste buds do not appear until much later. Finally, genetic crosses have revealed an inverse relationship between jaw size/taste bud number and eye size. The results suggest that hyperactive Shh signaling is responsible for increased oral pharyngeal traits in cavefish embryos and support an evolutionary model in which selection for larger jaws and more taste buds occurs at the expense of eyes via a pleiotropic Shh signaling.

Shh Expression in the Oral Pharyngeal Region and Taste Buds

The domain of shh expression is wider along the embryonic midline in tailbud stage cavefish embryos compared to their surface fish counterparts (Yamamoto et al., 2004), as well as in classical Shh signaling centers in the forebrain and hypothalamus later in development (Menuet et al., 2006). Here we demonstrate that shh expression is also expanded in the cavefish oral pharyngeal region. Quantification by qRT-PCR showed an approximate 3-fold increase in shh transcripts in cavefish relative to surface fish at 3 dpf, which at least in part reflects the oral pharyngeal increase. Shh signaling has been implicated in regulating pharyngeal development in other vertebrates (Moore-Scott and Manley, 2005). For example, shh is expressed in the zebrafish (Miller et al., 2000; Eberhart et al., 2006) and chick (Haworth et al., 2007) oral pharyngeal endoderm, where it controls condensation of skeletal elements in the developing pharyngeal arches and cranium. Accordingly, enhanced shh expression could be indirectly related to development of the distinctive cavefish craniofacial skeleton (Yamamoto et al., 2003). In zebrafish, oral-pharyngeal shh expression is induced by earlier Shh signals emanating from the ventral forebrain (Eberhart et al., 2006). The ventral forebrain is also a site of expanded shh expression in cavefish (Menuet et al., 2006). Consequently, elevated Shh signaling from the cavefish ventral forebrain could also induce the enlarged oral-pharyngeal expression domain. Alternatively, increased oral-pharyngeal shh expression might simply be a continuation of earlier expansion of the midline signaling domain in the prechordal plate.

As development proceeds shh expression is downregulated in most of the oral-pharyngeal epithelium except for strong foci in the tooth germs and marginal cells in the taste buds. Taste buds are under continuous renewal in vertebrates, and the marginal cells may be stem/precursor cells involved in their replenishment (Miura et al., 2001; 2006). The precise role of Shh in taste bud development is unclear, however, and may differ among various vertebrate species. In axolotl, taste buds are specified and appear early in the pharyngeal epithelium (Barlow, 2001), as they do in Astyanax (Jeffery et al., 2000) and zebrafish (Hansen et al., 2002). In contrast to our results in Astyanax, however, neither shh mRNA or Shh protein have been detected in the axolotl pharyngeal epithelium during taste bud formation (Parker et al., 2004). The situation is different in mammals, in which taste bud formation is preceded by the development of taste papillae on the emerging tongue. Expression of shh is initially uniform in the mammalian oral pharyngeal and prelingual areas, then becomes progressively restricted to the tongue, the taste papillae, and finally to the taste buds (Hall et al., 1999; Jung et al., 1999; Miura et al., 2001; 2003; Liu et al., 2004). The mammalian situation is temporally similar to that in Astyanax embryos, although the latter form taste buds directly from the oral pharyngeal epithelium.

Oral and Taste Bud Development in Cavefish Embryos

Taste buds begin to develop in Astyanax embryos between 3 and 4 dpf (Jeffery et al., 2000), and shh expression is detected in taste bud primordia as soon as they protrude above the oral pharyngeal epithelium. The timing of taste bud development is similar in Astyanax and zebrafish embryos (Hansen et al., 2002). We have demonstrated that cavefish embryos have a larger number of taste buds on both their upper and lower lips than their surface fish counterparts. The mouth, and later the jaw span, are also wider in cavefish. These constructive changes may remodel the cavefish mouth into a shovel-like structure that is effective for scooping sediment from the bottom of cave ponds.

The increase in taste buds observed in cavefish embryos we describe here is more modest than the 5- to 7-fold elevation reported in cavefish adults (Schemmel, 1967). There are several possible explanations for this difference. First, calretinin antibody could recognize only a subset of developing taste buds in Astyanax, as appears to be the case in amphibians (Barlow et al., 1996). This explanation is unlikely because all structures distinguishable as taste buds by their typical rosette-shaped morphology stained positively with calretinin antibody. Further, calretinin-stained taste buds were closely packed on the lips, leaving little or no room for additional taste buds between them. Second, some the structures originally described as taste buds by electron microscopy (Schemmel, 1967) might actually be other types of sensory organs, such as mechanosensory cells. If so, the difference in taste bud numbers between adult cavefish and surface fish would be inflated when assayed by electron microscopy. Third, differences in the number of external taste buds may appear later during cavefish development and thus would not be represented in our analysis. Supporting the last possibility, external taste buds appear much later in zebrafish and catfish development than oral-pharyngeal taste buds (Hansen et al., 2002; Northcutt, 2005). Our results indicate that oral taste buds are more numerous in cavefish embryos during the period in which shh expression in the oral pharyngeal region.

Role of Shh Signaling in Jaw and Taste Bud Development

The results of overexpression experiments suggest that shh is sufficient to promote the differences in oral and taste bud development we have seen between cavefish and surface fish embryos. Two key points are emphasized concerning these results. First shh mRNA injection in surface fish embryos can increase the number of taste buds to levels typical of cavefish embryos while also inducing defective eye development. Previous results showed that shh overexpression in surface fish promotes eye degeneration by inducing lens apoptosis (Yamamoto et al., 2004), which occurs naturally in cavefish embryos (Jeffery and Martasian, 1998). Second, upregulation of shh at specific times during surface fish development by temperature induced activation of the hsp70:shh:GFP DNA construct showed sensitive periods for eye and taste bud development prior to 1 dpf, although of taste buds do not appear morphologically until 2–3 days later. The results suggest that both eye and taste bud development can be affected by expanded Shh signaling along the early cavefish midline.

It seems likely that shh expression is also necessary for jaw and taste bud development. Although Shh inhibition with MOs did not completely suppress taste bud development, and co-injection of shh mRNA did not entirely rescue the effects of shh MOs, complete inhibition of taste bud development did occur after cyclopamine treatment. Teleosts contain two paralogous shh genes, shhA, the gene we have focused on in these studies, and shhB (formerly tiggy winkle hedgehog) (Ekker et al., 1995). Both shh genes are expanded along the cavefish anterior midline (Yamamoto et al., 2004), and it is possible that they are functionally redundant, requiring a double knockdown to completely affect taste bud formation. However, cyclopamine can inhibit the function of both genes because it acts downstream of ShhA/B by binding to the Smo protein (Chen et al., 2002). There also may be functional redundancy between Shh and other signaling ligands and transcription factors involved in taste bud development. For example, Notch, Bmp, Fgf (Jung et al., 1999; Seta et al., 2003), Prox1, Mash1, Nkx2.2, and NeuroD (Jeffery et al, 2000, Suzuki et al., 2002; Miura et al., 2003) are expressed in vertebrate taste buds. Except for shh, which is required for taste bud development in the mouse (Mistretta et al., 2003; Liu et al., 2004), little is known about the roles of these molecules and how they may interact during taste bud development.

Modularity of Sense Organs, Pleiotropic Tradeoffs, and Evolution of Eye Degeneration

It has been proposed that sensory organs are organized as developmental modules in Astyanax and that natural selection can affect developmental interactions between them, resulting in tradeoffs (Franz-Odendaal and Hall, 2006). Further, regulatory genes could force a sensory module into a specific type of differentiation, and if these genes are pleiotropic, there can be concerted positive and negative consequences on development of other sensory modules. Accordingly, our results suggest that the Astyanax eye module may be linked to the oral-taste bud module by pleiotropic effects of Shh signaling. A summary of the known pleiotropic activities of Hh signaling along the cavefish midline based on current knowledge of genes involved in oral/taste bud development and eye regression is shown in Figure 7.

Figure 7.

Diagram illustrating the relationship between Shh signaling, oral pharyngeal constructive traits, and eye development along the Astyanax embryonic midline. A. Surface fish. B. Cavefish indicating the effects on oral pharyngeal, lens, and optic cup development of enhanced Shh signaling. Letter size indicates relative increase or decrease in cavefish compared to surface fish. See text for other details.

Negative effects of Shh on eye development (Ekker et al, 1995; Yamamoto et al., 2004) and corresponding positive effects on oral and taste bud development suggest a developmental tradeoff between eyes and feeding organs. This possibility is supported by two important observations. First, the sensitive periods for eye degeneration and taste bud enhancement both occur during early development, prior to the appearance of taste buds. Second, independently of the shh results, genetic crosses show an inverse relationship between eye size and the extent of oral and taste bud development. An inverse relationship between these traits is consistent with offsetting positive and negative effects of shh overexpression in the concerted evolution of these two sensory modules. Third, genetic studies have revealed overlapping quantitative trait loci (QTL) governing eye size and increase in taste buds (Protas et al., 2008). One way of explaining this overlap would be to postulate a single pleiotropic gene within the OTL that controls both traits.

Cavefish have evolved a specialized bottom feeding behavior relative to surface fish, which normally feed in the water column using visual cues (Schemmel, 1980). Efficient bottom feeding requires posture at an angle in which the mouth can sample substrate in cave pools. Thus, increase in jaw size and taste bud number could have evolved as an adaptation to the challenges of searching for and sampling the quality of limited food in the cave environment (Schemmel, 1967; Hüpopp, 1987). Whereas feeding efficiency may be one of the traits driving eye regression through Shh signaling, it is not the only possibility. For example, cavefish also have an enlarged ventral forebrain controlled by an expanded Hh signaling center in the floor plate, which may lead to the production of more olfactory inter-neurons (Menuet et al., 2006). Together, strong Shh signals from the floor plate (Rétaux et al., 2008) and the prechordal plate resulting in the development of multiple beneficial traits may synergistically drive rapid evolution of eye degeneration in blind cavefish.

Although shh appears to have major role in eye degeneration and enhancement of constructive traits, it is not the gene that is mutated in cavefish. Genetic analysis has shown that none of the multiple QTL underlying cavefish eye regression are located near a known hedgehog gene locus (Protas et al., 2007). Furthermore, the expression domains of upstream regulators of the Shh midline pathway, such as nodal and goosecoid, are also expanded in cavefish (Yamamoto unpublished). Thus, further progress in understanding the amplification of shh-dependent phenotypes in cavefish will require identification of the genes upstream of shh that have been mutated to cause hyperactivity of the midline signaling system.