Complex regulation of cyp26a1 creates a robust retinoic acid gradient in the zebrafish embryo

Abstract

Positional identities along the anterior-posterior axis of the vertebrate nervous system are assigned during gastrulation by multiple posteriorizing signals, including retinoic acid (RA), fibroblast growth factors (Fgfs), and Wnts. Experimental evidence has suggested that RA, which is produced in paraxial mesoderm posterior to the hindbrain by aldehyde dehydrogenase 1a2 (aldh1a2/raldh2), forms a posterior-to-anterior gradient across the hindbrain field, and provides the positional information that specifies the locations and fates of rhombomeres. Recently, alternative models have been proposed in which RA plays only a permissive role, signaling wherever it is not degraded. Here we use a combination of experimental and modeling tools to address the role of RA in providing long-range positional cues in the zebrafish hindbrain. Using cell transplantation and implantation of RA-coated beads into RA-deficient zebrafish embryos, we demonstrate that RA can directly convey graded positional information over long distances. We also show that expression of Cyp26a1, the major RA-degrading enzyme during gastrulation, is under complex feedback and feedforward control by RA and Fgf signaling. The predicted consequence of such control is that RA gradients will be both robust to fluctuations in RA synthesis and adaptive to changes in embryo length during gastrulation. Such control also provides an explanation for the fact that loss of an endogenous RA gradient can be compensated for by RA that is provided in a spatially uniform manner.

Figure 1. Responses of Embryos to Sources of RA.

Long-range induction of a rare:yfp reporter. Confocal images of live embryos, dorsal view, anterior to the left. (A) rare:yfp expression in posterior hindbrain and spinal cord at 24 hpf with an anterior boundary at r6/7. (B) Quantification of fluorescence intensities in (A). (C) Lack of YFP expression in an aldh1a2 morphant embryo. (D) Rescue of YFP expression in a morphant by somitic mesoderm (yellow) transplanted during gastrulation. (E–I) Bead implantation. RA-coated beads were implanted anterior to the first somite at 18–19 hpf and imaged at 23–24 hpf. A DMSO-coated bead (E) failed to rescue YFP expression, whereas beads soaked in either 10 μM (F) or 100 μM (G) RA induced YFP at distances of up to 200 μm. (H and I) Inhibition of both cyp26b1 and cyp26c1 in DEAB-treated embryos leads to a symmetrical response from the rare:yfp reporter to a bead soaked in 10 μM RA (compare [H] with [I]). The dotted line indicates the r6/7 boundary, the dashed line, the anterior border of somite 1. nt, neural tube; s, somites.

Figure 2. Concentration-Dependent Induction of hox Gene Expression by RA

(A) In wild-type (wt) embryos, hoxb4 expression in the neural tube (nt) extends to the r6/7 boundary, whereas the limit of hoxb5a expression is level with the first somite (s). hoxb5a is also expressed in axial, lateral (lat), and cranial (cm) mesoderm. (B and C) Beads soaked in 100 μM RA were implanted at 1–3 somites (10.3–11 hpf) into DEAB-treated embryos, fixed 3 h later, and assayed for hox expression. Dorsal views, showing in situ hybridization with probes for hoxb4 and hoxb5a at 14 hpf. (B) DMSO-coated beads do not induce hox expression. (C) RA-coated beads induce hoxb5a in lateral mesoderm (arrowhead), hoxb5a/hoxb4 double-positive cells near the bead, and hoxb4 single-positive cells further away (black arrow). (D–F) Time-course of hoxb5a induction in response to a 100 μM RA bead: no induction 30 min post-implantation (D), a few cells near the bead after 60 min (arrowhead) (E), and strong expression further away after 180 min (F). (G–L) RA-induced degradation limits the range of RA signaling. RA beads were implanted at 1–3 somites (10.3–11 hpf) into DEAB-treated embryos, fixed 3 h later, and assayed for hox expression. Dorsal views, showing in situ hybridization with probes for hoxb5a (G–I) or hoxb4 (K and L) at 14 hpf. (G) Implantation of a DMSO bead has no effect on gene expression in DEAB–treated embryos. (H and I) A 100 μM RA bead placed into a DEAB-treated, cyp26a1 morphant (I) induces hoxb5a over a much larger area than a DEAB-treated control (H). (J) Graph showing an increase in the range of hox induction in cyp26a1 morphant, DEAB-treated embryos (black bar) as compared to DEAB-treated (light grey bar). The asterisk (*) indicates significantly different from DEAB-treated controls (p < 0.05), using the Student t-test. (K and L) A 10 μM RA bead placed into a cyp26a1 morphant induces hoxb4 over a much larger area.

Figure 3. Regulation of Cyp26s by RA

(A) Loss of cyp26a1 expression in cranial mesoderm in DEAB-treated embryos and lack of induction by a DMSO-soaked bead. (B) Induction of cyp26a1 expression by an RA bead; inset, optical section showing cyp26a1 induction in both the neural tube (nt), lying above the notochord (n), and lateral mesoderm (white arrowhead). The asterisk (*) indicates the bead. (C) Weak induction of cyp26b1 (black arrowhead). (D) No induction of cyp26c1.

Figure 4. RA-Dependent Expression of cyp26a1

Wild-type (wt) and DEAB-treated embryos were stained for extended periods of time for cyp26a1 expression to investigate low-level expression. Top, dorsal view; bottom, lateral view. (A and C) Weak expression posterior to the presumptive midbrain (arrow to arrowhead). (B and D) This expression is lost in DEAB-treated embryos.

Figure 5. cyp26a1 Is Required Outside of the Anterior Neural Domain

(A and D) hoxb1b expression in wild-type (wt) embryos. (A–C′) Lateral view, anterior to top, dorsal to right. (D–F′) Dorsal view, anterior to top. (B, C′, E, and F′) Merged brightfield and fluorescent images show positions of transplanted cyp26a1-deficient cells labeled with lysine-fixable rhodamine dextran (red). (B–C′) Transplants of cyp26a1-deficient cells located within the endogenous hoxb1b domain cause a nearly cell-autonomous decrease in hoxb1b expression (black arrowheads). (E–F′) Transplants of cyp26a1-deficient cells also cause changes in hoxb1b expression in adjacent host cells (white arrowheads).

Figure 6. Regulation of cyp26a1 by RA and Fgfs

(A and B) Anterior and marginal expression domains of cyp26a1 expression are unchanged by 10 μM DEAB. Anterior is to the left in all panels. (C and D) RA beads (black arrowheads) induce cyp26a1 expression in DEAB-treated embryos far from (white arrowhead and arrow in [D]), but not adjacent to the bead. (E–H) Fgf signaling sets the posterior boundary of cyp26a1 expression. (E) cyp26a1 expression in wild type (wt) (lateral view, anterior to the left). (F and G) Expression in SU5402-treated embryos shifts posteriorly in a concentration-dependent manner (compare brackets in [E], [F] and [G]). (H) The posterior shift caused by blocking Fgf signaling is abolished by DEAB treatment.

Figure 7. Modeling the RA Morphogen Gradient

A one-dimensional mathematical model of the gastrula-stage zebrafish embryo was developed that incorporates RA synthesis, diffusion, cell permeation, degradation, and signaling; a stable Fgf gradient; and expression of cyp26a1 under the control of RA and Fgf signaling, as well as other position-specific cues (see Materials and Methods). In each panel, anterior is to the left and posterior to the right. On the abscissa, zero represents the posterior boundary of the domain of high, anterior cyp26a1 expression (approximately the r1/r2 border), which is set independently of RA. Values on the ordinate are in arbitrary units. Fgf concentration is shown in green, Cyp26a1 in blue, extracellular RA ([RA]_out) in black, and RA signaling (a function of intracellular RA) is shown in red. Parameter values common to all panels are given in Materials and Methods. (A) Typical patterns of RA signaling and cyp26a1 expression generated by the model. Note the presence of low-level cyp26a1 expression that declines from anterior to posterior. (B and C) Feedback and feedforward regulation of cyp26a1 expression leads to robustness to RA synthesis rates. In both panels, the dashed curves show the effects of a 2-fold decrease in RA production. In (B), cyp26a1 is taken to be constant over the interval 0 < x < 360 μm, whereas in (C), it is regulated as prescribed by the model. Note the dramatic increase in robustness in (C). (D and E) Effect of embryo elongation on Fgf and RA gradients. In both panels, solid lines are calculated for a posterior margin at 400 μm; dashed lines for a posterior margin at 450 μm. In (D), the dashed line is calculated using a rate of Fgf synthesis 2.75 times higher than for the solid line; this is to the degree of increase that would be required to compensate for the rearward movement of the Fgf source. (E) shows that, using the same sets of parameters, the RA gradient automatically compensates for the rearward movement of the RA source, without any need for readjustment of RA synthesis. For comparison, the dotted line shows what would happen to the RA gradient were there no such compensation. (F and G) RA supplied in a completely delocalized manner can produce a relatively normal morphogen gradient. (F) is a modification of (A), in which the endogenous posterior source of RA (from 360 < x < 400) has replaced by a constant influx of RA at all locations. For an appropriate rate of exogenous RA influx, relatively normal patterns of RA signaling and cyp26a1 expression ensue, despite the lack of an extracellular RA gradient. This is particularly true at anterior locations, as shown in (G), in which the RA signaling gradients from (A) and (F) are compared (solid and dashed curves, respectively). (H) RA signaling gradients produced by delocalized RA are robust to rates of RA influx. Like gradients produced by endogenous RA (C), gradients produced by exogenous, delocalized RA are also predicted to be robust (solid curve is RA signaling from [F]; dashed curve shows the effect of a 2-fold decrease the in the rate of RA influx). Values of parameters that differed between panels were β = 1 ([A–E] and solid curve in [G]) or 0.5 ([F], [H], and dashed curve in [G]); x _f = 400 μm ([A–C] and [F–H]) or 450 μm (D–E); f ₀ = 0 (B), 400 ([A], [C], [F–H], and solid curves in [D] and [E]), or 1,100 (dashed curves in [D] and [E]); k _deg = 500 sec⁻¹ ([A] and [C–G]) or 0.03 sec⁻¹ (B); k _max = 0.001 ([A] and [C–G]) or 4 (B); V(x) = 5 for x > x _f – 40, zero otherwise ([A], [D], [E], and solid curves in [B], [C], and [G]), 2.5 for x > x _f – 40, zero otherwise (dashed curves in [B] and [C]), 0.2 for all x ([F], dashed curve in [G], and solid curve in [H]), or 0.1 for all x (dashed curve in [H]).

Figure 8. Model of Hindbrain Patterning by RA

(A) Diagram of interactions between RA, Fgf, and cyp26a1. Only those shown in blue are included in the mathematical model; dotted lines are extrapolated from other animal models. (B) Schematics illustrate dorsal views of gastrulating embryos and corresponding gradients, anterior to the left. RA is produced by Aldh1a2 in somitic mesoderm (red) and diffuses through the neurectoderm. Cyp26a1 (blue) degrades RA at differing rates to produce a gradient across the hindbrain that specifies rhombomere fates. The regulation of cyp26a1 by both Fgf and RA signaling creates a robust RA gradient that grows appropriately as the A–P axis of the embryo elongates.