SDF1 antagonism of axonal repellents requires multiple G protein components and an ADAM metalloprotease

Growth cones of developing axons navigate by interpreting signals from multiple cues. Some of these are the familiar guidance cues netrin, semaphorin, slit, and ephrin. Growth cones are also influenced by GPCR ligands, including neurotransmitters such as glutamate and chemokines such as SDF1. Previous work from our lab demonstrated that either glutamate or SDF1, acting through their receptors mGluR1 or CXCR4, respectively, can reduce growth cone responsiveness to repellent cues. This effect is pertussis toxin-sensitive, implicating Gαi/o proteins, yet dependent on increased cAMP, implicating Gαs proteins. The antirepellent effect of SDF1 could also be mimicked by inhibition of Rho, suggesting that inhibition of Rho is a component of the antirepellent pathway. Here, I demonstrate that SDF1 antirepellent activity is blocked by peptides or proteins targeting Gα i, Gαq, or Gβγ. This suggests that multiple G protein components are required for SDF1 signaling. I also show that SDF1 antirepellent activity is mimicked by constitutively active forms of Gα q, Gαi, or Gαs. This suggests that higher-than- physiological levels of individual G protein components can substitute for a combination of G protein components in antirepellent signaling. A role for Gαq in antirepellent signaling is further supported by the ability of a phospholipase C (PLC) inhibitor to block the SDF1 antirepellent effect, consistent with Gαq\u27s canonical activation of PLC. My work also reveals an alternate mechanism for SDF1-induced antagonism of repellent signaling. I show that the metalloprotease ADAM10 can cleave the repellent receptor neuropilin-1. Further, SDF1 antirepellent activity is blocked by either the metalloprotease inhibitor TAPI-2 or a dominant-negative ADAM10. Thus, inhibitory shedding of repellent receptors may contribute to the antirepellent effect. Previous work has shown that the antirepellent effect is mimicked by pharmacologically increased cAMP or blocked by a cAMP antagonist. TAPI-2 does not block the antirepellent effects of a cAMP analogue, suggesting that ADAM activation belongs to a separate pathway not downstream of cAMP. This work supports a model wherein SDF1/CXCR4 activates multiple G protein components to both increase cAMP and activate ADAM10. This would reduce sensitivity to repellents through inactivation of Rho and clearing of repellent receptors from the growth cone surface