Heparan sulfate proteoglycan-dependent induction of axon branching and axon misrouting by the Kallmann syndrome gene kal-1

Abstract

Kallmann syndrome is a neurological disorder characterized by various behavioral and neuroanatomical defects. The X-linked form of this disease is caused by mutations in the KAL-1 gene, which codes for a secreted molecule that is expressed in restricted regions of the brain. Its molecular mechanism of action has thus far remained largely elusive. We show here that expression of the Caenorhabditis elegans homolog of KAL-1 in selected sensory and interneuron classes causes a highly penetrant, dosage-dependent, and cell autonomous axon-branching phenotype. In a different cellular context, heterologous C. elegans kal-1 expression causes a highly penetrant axon-misrouting phenotype. The axon-branching and -misrouting activities require different domains of the KAL-1 protein. In a genetic modifier screen we isolated several loci that either suppress or enhance the kal-1-induced axonal defects, one of which codes for an enzyme that modifies specific residues in heparan sulfate proteoglycans, namely heparan-6O-sulfotransferase. We hypothesize that KAL-1 binds by means of a heparan sulfate proteoglycan to its cognate receptor or other extracellular cues to induce axonal branching and axon misrouting.

Figure 1

The Kallmann syndrome gene kal-1 in C. elegans. (A) Schematic domain structure of the Kallmann syndrome proteins in humans (Hs), worms (Ce), and flies (Dm). Pairwise similarities of domains between proteins were calculated by aligning the domains as predicted by SMART. WAP, whey acidic protein domain; Fn(III), fibronectin III motif; SS, signal sequence. (B) Schematic drawing of the gene structure of C. elegans kal-1 (located on cosmid K03D10) and the transcriptional reporter fusion with green fluorescent protein (gfp) and the unc-54 3′ untranslated region. (C) Expression pattern of the kal-1∷gfp reporter construct in transgenic animals (otIs33IV).

Figure 2

Cell-specific expression of kal-1 causes axon branching. (A) Anti-KAL-1 antibody staining of transgenic animals that overexpress kal-1 in the AIY interneuron class (otIs35) and in wild-type N2 animals. Arrows point to axonal branches induced by KAL-1 protein (the unbranched nature of the AIY interneurons can be observed with gfp labeling; see B). See supporting information for more comments, micrographs, and technical notes on antibody staining. (B and C) Data were obtained from extrachromosomal (Ex) and integrated (Is) mis/overexpressing lines as indicated. Numbers in parentheses give the number of extrachromosomal lines. The number of defective animals of all extrachromosomal lines were added up to yield percentages; it is specifically noted if the spread of percentages of individual lines was significant. (B) Overexpression phenotype in AIY interneurons. Scoring criteria for the branching defects are detailed in the supporting methods. White arrows point to axonal branches and white arrowheads show arborized branches in lines overexpressing the kal-1mWAP construct. All experiments were performed in a mgIs18 (Is[ttx-3∷gfp]) gfp reporter background unless indicated otherwise. kal-1ΔC and kal-1S241K recapitulate mutations that were found in patients with Kallmann syndrome (18). kal-1mWAP contains two point mutations, C134S and C135S, that abolish two integral disulfide bonds in the WAP domain. Note that the branching effect is unrelated to the presence of gfp in the cell because KAL-1 antibody staining shows similar branches (A). ^a, results with a different Is[ttx-3∷gfp], termed mgIs32, are virtually identical. (C) Mis/overexpression phenotypes in AFD sensory neurons. A white arrow points to an axonal branch at the ventral portion of the main axon. All experiments were performed in an oyIs17 (Is[gcy-8∷gfp]) reporter background. ^a, results varied between 48% and 80% per line.

Figure 3

Panneuronal expression of kal-1 causes axon misrouting and hypodermal defects. (A) Panneuronal phenotypes in AIY interneurons and AFD sensory neurons. AIY “short stops” are defined as the failure of the two bilaterally symmetric axons to reach the dorsal midline. Shown Left are AFD wild-type morphology (black arrowheads point to the amphid commissure), and the amphid-misrouting phenotype exemplified by the absence of the amphid commissure [(Center) unilateral; (Right) bilateral absence, marked by white arrowheads)] as visualized with DiI staining (Center) or oyIs17, which exclusively labels AFD (Right). Animals were scored as defective if they displayed unilateral or bilateral defects, respectively. ^a, results varied between 4% and 44% per line. AIY morphology was examined in an mgIs18 background, and AFD morphology was examined in an oyIs17 background. Note that amphid commissure defects were 20% penetrant if otIs81 was scored in an N2 background with DiI filling (Table 1). (B) Variable abnormal (Vab) phenotypes as observed in lines overexpressing kal-1 panneuronally.

Figure 4

Mutations that modify the KAL-1-dependent phenotype. (A) Genetic interaction of the kal-1(gf)-inducing axon-branching defect with mutants described as being involved in axonal patterning and other aspects of neuronal function (34, 35) and mutants retrieved from a modifier screen. Percent suppression does not refer to qualitative changes in branch appearance but penetrance of the defect. All experiments were done in a mgIs18 background to label AIY and, if not indicated otherwise, with the ttx-3∷kal-1 overexpressor otIs76IV. The moderately penetrant suppression by spectrins may be explained by their proposed role in localizing cell adhesion/signaling proteins (36). ^a, ttx-3∷kal-1 overexpressor otIs35X. ^b, ttx-3∷kal-1 overexpressor otIs77II. ^c, strain was marked with lon-2(e678). ^d, enhancement was arbitrarily defined as branches exceeding 20 μm (for quantification see Fig. 8, which is published as supporting information on the PNAS web site). (B) Representative example of the ttx-3∷gfp reporter control (mgIs18) (Upper Left), the ttx-3∷kal-1 overexpressor (otIs76 mgIs18) (Lower Left), the ot16 suppressor mutation in combination with the overexpressor (Upper Right), and the enhancer ot21 that shows considerably longer branches than the overexpressor alone (Lower Right). (C) hst-6 cDNA structure was experimentally determined by PCR and contained SL1 splice leader sequences. An alignment of two N-terminal portions of C. elegans HST-6 with its human (Hs), mouse (Mm), and fly (Dm) orthologs is shown. 5′PSB, binding site for the 3′-phosphoadenosine 5′-phosphosulfate cofactor (37).

Figure 5

Heparin-dependent binding of C. elegans KAL-1∷Fc protein fusions to cell lines. (A) Expression constructs. Supernatants of HEK-293T cells expressing the individual constructs were immunoblotted with anti-Fc antibodies. (B, C, and D) Binding KAL-1∷Fc to individual cell lines. U-25 and U-3731, human glioma; T986G, human glioblastoma; U-1248 MG, human glioblastoma; A172, human glioblastoma; 293, human kidney fibroblasts. (C) Binding of KAL-1∷Fc to cells can be inhibited by adding exogenous heparin or high salt. (D) Domain requirement for KAL-1 binding to cells. mWAP (C134S/C135S) and S241K are point mutations introduced into full-length KAL-1∷Fc (see text).