XB-ART-39488Dev Biol June 15, 2009; 330 (2): 273-85.
Show Gene links Show Anatomy links
LIMK1 acts downstream of BMP signaling in developing retinal ganglion cell axons but not dendrites.
The actin cytoskeleton inside extending axonal and dendritic processes must undergo continuous assembly and disassembly. Some extrinsic factors modulate actin turnover through controlling the activity of LIM kinase 1 (LIMK1), which phosphorylates and inactivates the actin depolymerizing factor cofilin. Here, we for the first time examine the function and regulation of LIMK1 in vivo in the vertebrate nervous system. Upon expression of wildtype or kinase-dead forms of the protein, dendrite growth by Xenopus retinal ganglion cells (RGCs) was unchanged. In contrast, maintaining a low, but significant level, of LIMK1 function in the RGC axon is critical for proper extension. Interestingly, bone morphogenetic protein receptor II (BMPRII) is a major regulator of LIMK1 in extending RGC axons, as expression of a BMPRII lacking the LIMK1 binding region caused a dramatic shortening of the axons. Previously, we found that BMPRIIs stimulate dendrite initiation in vivo. Thus, the fact that manipulation of LIMK1 activity failed to alter dendrite growth suggests that BMPs may activate distinct signalling pathways in axons and dendrites.
PubMed ID: 19361494
Article link: Dev Biol
Species referenced: Xenopus laevis
Genes referenced: actl6a bmp2 bmp4 bmp7.1 bmpr2 cfl1 clstn2 limk1 limk2 myc tec
Antibodies: Limk1 Ab1 Limk2 Ab1
Article Images: [+] show captions
|Fig. 1. Expression of limk1 in the developing Xenopus retina. (A–F) Xenopus embryos processed for wholemount in situ hybridization using antisense RNA probe for limk1. Whole embryos show a restricted pattern of limk1 expression. The eye, otic vesicle, areas of the brain and the nasal placodes all express limk1 mRNA at stage 30 (A). The pattern is maintained, but stronger in the eye and brain, at stage 35/36 (B). Transverse vibratome sections through labelled embryos shows the changing pattern of limk1 in the retina. At stage 28, the optic vesicle has not yet invaginated and there is a patch of limk1 expression in the dorsal/central retina that spans the width of the tissue (C). By stage 30, the optic cup is present and limk1 is expressed in the inner layers that will become the GCL () and the inner nuclear layer (INL) (not labelled) (D). Expression has spread to the dorsal ciliary marginal zone (CMZ; white arrowhead), but is still absent from the ventral CMZ (black arrowhead). Expression in the central retina has weakened slightly at stage 33/34 (E), while the dorsal CMZ, ventral retina and ventral CMZ show robust expression. The divisions between the emerging layers are shown by the white lines in E (from lens outward is GCL, INL, ONL). limk1 expression in RGCs begins to decrease at stage 35/36 (F), but is still strong in the INL and CMZ. A, anterior; br, brain; D, dorsal; e, eye; np, nasal placodes; nr, neural retina; ot, otic vesicle; P, posterior; V, ventral. Scale bar in C is 50 μm.|
|Fig. 2. LIMK is active in developing RGCs and their processes. (A) The P-LIMK antibody detects the phosphorylated, active forms of LIMK1/2 with the molecular weight of 65 kDa. (B–D) Transverse cryosections were immunolabelled with the antibody, and stacked optical slices presented. At stage 33/34, there is label in RGC axons, shown in a section (B) where RGC axons run along the vitreal surface of the retina (arrows) and in the optic nerve head, and in a section (C) that has captured part of the bilateral optic nerves before they enter the ventral diencephalon (white arrowheads). Labelling of the brain neuropil is also evident. At stage 37/38 (D), the optic fiber layer expression is still present (white arrows), but there is now more obvious labelling of the IPL. cmz, ciliary marginal zone; Di, diencephalon; D, dorsal; IPL, inner plexiform layer; L, lens; onh, optic nerve head; PE, pigment epithelium; RGCL, retinal ganglion cell layer; V, ventral.|
|Fig. 3. P-cofilin1 is expressed by RGCs. In the top left is a Western blot of protein isolated from stage 37/38 embryos, labelled with an antibody that recognizes P-cofilin1. The antibody recognizes a band at approximately 21 kDa (black arrow). (A, B) Transverse cryosections of Xenopus embryos immunolabelled with the P-cofilin1 antibody. P-cofilin1 is expressed in the retina at stage 33/34, particularly in a subset of cells in the GCL (; A). In a lower power view of a labelled section through the eyes and brain of a stage 35/36 embryo, we can see that Pcofilin1 is expressed throughout the embryo, but at variable levels (B). Scale bar is 50 μm in A for A. Scale bar in B is 25 μm. br, brain; D, dorsal; L, lens; nr, neural retina; V, ventral.|
|Fig. 4. wtLIMK1 increases the phosphorylation of cofilin. Stage 27 embryos were electroporated with GFP and either wtLIMK1-MT or kdLIMK1-MT, fixed at stage 37/38, cryostat sectioned and immunolabelled for P-cofilin1. (A, C) Merged images of myc-expressing diencephalic cells in blue (arrows) and P-cofilin1 expression in red, for wtLIMK1- (A) and kdLIMK1-electroporated (C) embryos. The matching images of P-cofilin expression alone are also shown for each condition (B: wtLIMK1, D: kdLIMK1). Scale bar is 10 μm. The intensity of P-cofilin1 expression in myc+ and myc− GFP-expressing cells was scored in a blinded fashion on a scale of 0–3. (E) Graph showing the ratio of the average P-cofilin1 intensity between myc+ and myc− GFP-expressing cells for each condition. Only for wtLIMK1-expressing cells was there an increase in P-cofilin1 expression in the myc+ cells compared to the myc− cells.|
|Fig. 5. Changes to LIMK1 activity do not alter RGC dendrite growth and branching. CS2-GFP was transfected alone as a control, or with CS2-wtLIMK1-MT or CS2-kdLIMK1-MT, into the retinal primordia of stage 19 Xenopus embryos. Embryos were fixed at stage 40, and either vibratome (A–C) or cryostat sectioned (D–F), and expression of the myc-tagged LIMK1 constructs was detected by anti-myc immunohistochemistry. GFP fluorescence fills the cell and allows for the analysis of cell morphology. Scale bar is 10 μm. Stars indicate branch points. (A–C) RGCs in retinal vibratome sections expressing GFP alone (A), wtLIMK1 (B), or kdLIMK1 (C). (D–F) RGCs in cryostat sections expressing GFP alone (D), GFP plus kdLIMK1 (E), or GFP plus wtLIMK1 (F). GFP fluorescence is shown in green, the myc-immunolabelling in red, and the merge in yellow. (G–J) Graphs showing the quantification of the effects of wtLIMK1 and kdLIMK1 on RGC dendrite development. The average number of primary dendrites (G), number of branch points (H), total dendritic length (I), and width of the dendritic arbor (J) were similar in all conditions. n = the number of independent experiments, and in brackets are the numbers of cells analyzed for each condition. Error bars are S.E.M. p > 0.05, One Way ANOVA.|
|Fig. 7. wtLIMK1 overexpressing growth cones are larger than control. (A–C) Growth cones of transgene-expressing RGC axons labelled with α-GFP (A) or α-myc (B, C) antibodies in stage 40 Xenopus brains. CS2-GFP, CS2-wtLIMK1-MT or CS2-kdLIMK1-MT was electroporated into the developing retinas of stage 27 embryos, which were fixed at stage 40, dissected and processed for wholemount immunochemistry. The growth cones photographed and analyzed are those that failed to reach the tectum. Scale bar in A is 10 μm for A–C. (D) Graph showing the average area of the growth cones present at the tips of GFP, wtLIMK1-MT, or kdLIMK1-MT-expressing axons that fail to reach the tectum. n = the number of independent experiments, and in brackets are the numbers of growth cones analyzed. Error bars are S.E.M. p < 0.05, One Way ANOVA, Student–Newman–Keuls post hoc test. (E) Graph showing the average number of filopodia extended from growth cones expressing wtLIMK1-MT, dnLIMK1-MT or GFP. n's are the same as for A. Error bars are S.E.M. There is no statistical difference between the three groups. p > 0.05, One Way ANOVA.|
|Fig. 8. BMP ligands are expressed near the developing optic tract. (A–C) Stage 35/36 wholemount brains processed for in situ hybridization using antisense probes to bmp2 (A), bmp4 (B), and bmp7 (C). All three mRNAs are expressed near or within the pineal gland. bmp2 is also expressed in a spot in the mid-diencephalon (A), and bmp4 in a stripe at the back of the tectum (B). bmp7 is expressed in a large area of the ventral diencephalon and telencephalon (C). (D) Wholemount stage 40 Xenopus brain with RGC axons labelled brown following GFP electroporation and anti-GFP immunochemistry. This shows the optic pathway for comparison to the bmp expression domains. The dotted white line is at the approximate tectal border. A, anterior; D, dorsal; di, diencephalon, mb, midbrain; oc, optic chiasm; pi, pineal gland; P, posterior; tec, tectum; tel, telencephalon; V, ventral.|
|Fig. 9. Blocking the interaction between BMPRII and LIMK1 disrupts RGC axon extension. (A–D) Lateral views of stage 40 wholemount Xenopus brains showing transgene-expressing RGC axons labelled in brown with α-GFP (A, B) or α-myc antibodies (C, D). The developing retinas of stage 27 Xenopus embryos were electroporated with plasmids encoding GFP alone (A, B) or GFP plus BMPRIIδLBR-MT (C, D). Embryos were fixed at stage 40 and their brains processed for wholemount immunochemistry. B and D are higher magnification views of the tracts shown in A and C. The majority of RGC axons expressing GFP alone have grown through the diencephalon, made a caudal turn in the mid-diencephalon () and innervated the optic tectum by stage 40 (dotted line shows the approximate anterior border of the optic tectum). Many of the axons expressing BMPRIIδLBR-MT fail to reach their target by stage 40, and their growth cones (arrows) can be detected at various locations within the optic tract. Scale bar in A is 50 μm for A, C. Scale bar in B is 50 μm for B, D. A, anterior; D, dorsal; hb, hindbrain; pi, pineal gland; P, Posterior; ot, optic tract; tec, tectum; tel, telencephalon; V, ventral. (E) Graph showing the percentage of axons expressing either GFP or GFP/BMPRIIδLBR that failed to reach the tectum at the time of fixation. n = the number of independent experiments and in brackets are the numbers of axons analyzed. Error bars are S.E.M. p < 0.05, paired Student's t-test. (F) Graph showing the distribution of transgene-expressing growth cones along the optic pathway, averaged over three experiments. The numbers reflect the average percentage of total transgene-expressing axons with growth cones in progressively more distal sections of the optic tract. n's are the same as in E. Error bars are S.E.M. (G) Schematic showing a proposed model for the pathways that act downstream of BMPR signalling in RGC dendrites and axons. An unknown signalling pathway functions in dendrites, while the LIMK pathway is active in axons. Too much or too little LIMK activity changes the balance of actin polymerization–depolymerization regulated by ADF/cofilin, and impairs RGC axon extension.|
|Fig. 6. Altered LIMK1 activity prevents RGC axons from reaching their target. (A) Lateral views of stage 40 wholemount Xenopus brains showing RGC axons labelled with α-GFP (A, B) or α-myc antibodies (C). The developing retinas of stage 27 Xenopus embryos were electroporated with plasmids encoding CS2-GFP alone (A, B), CS2-GFP plus CS2-wtLIMK1-MT (C, D) or CS2-GFP plus CS2-kdLIMK1-MT (E, F). Embryos were fixed at stage 40 and their brains processed for wholemount immunochemistry. B, D, and F are higher magnification views of the tracts shown in A, C, and E, respectively. (A, B) The vast majority of RGC axons expressing GFP alone have grown through the diencephalon, made a caudal turn in the mid-diencephalon (⁎) and innervated the optic tectum by stage 40 (dotted line shows the approximate anterior border of the optic tectum). (C, D) Many of the axons expressing wtLIMK1-MT fail to reach their target by stage 40, and their growth cones (arrows) can be detected at various locations within the optic tract. (E, F) Some of the kdLIMK1-expressing axons also fail to reach the tectum (arrow; F). Scale bar in A is 50 μm for A, C, E. Scale bar in B is 50 μm for B, D, F. A, anterior; D, dorsal; hb, hindbrain; pi, pineal gland; P, Posterior; ot, optic tract; tec, tectum; tel, telencephalon; V, ventral. (G) Graph showing the percentage of transgene-expressing axons that fail to reach the tectum by stage 40. n=the number of independent experiments, and in brackets are the numbers of axons analyzed. Error bars are S.E.M. ⁎⁎⁎pb0.001, One Way ANOVA, Studentewmaneuls post hoc test. (H) Graph showing the locations of the growth cones for the GFP, wtLIMK1-MT or kdLIMK1-MT-expressing axons. Growth cones were categorized as being in the ventral diencephalon, mid-diencephalon, or dorsal diencephalon, and then graphed as the percentage of the total number of expressing axons, including those that reached the tectum, which were found in each region. n's are the same as in G.|