A mutation in the general membrane trafficking machinery and hydrocephaly

Subtle changes in vesicle recycling machinery affect secretory processes in development of complex organisms.

Membrane fusion is required to deliver specific cargo molecules to distinct intracellular compartments, thereby maintaining the functional and structural organization of eukaryotic cells, as well as intercellular communication, which occurs in precisely timed spatial patterns at the plasma membrane. During embryogenesis, exocytosis is harnessed to induce differentiation and morphogenesis, as demonstrated by the work of Hong et al. in this issue of PNAS (1). Later and throughout life, the temporal control of site-specific fusion coordinates the release and processing of short and long-range messages within an organism.

SNAREs, SNAPs, and NSF: The Building Blocks of the Membrane Fusion Machinery

The basic membrane fusion process is driven by the assembly of cognate v- and t-SNAREs [soluble N-ethylmalemide-sensitive factor (NSF) attachment protein (SNAP) receptors] between two lipid bilayers (2, 3). Because SNAREs comprise a protein family with its individual members localized to distinct intracellular compartments, the distribution of cognate v- and t-SNAREs outlines the flow pattern for membrane trafficking (4). Separate SNAREs are largely unfolded, and the energy necessary to merge membranes is provided by the folding of SNAREs into a stable four-helix bundle (with the v-SNARE and t-SNARE contributing one and three helices, respectively) (5, 6). In contrast to viral fusion proteins, which are only required once during the lifetime of a virus (to invade the host cell), SNAREs are continuously recycled for further rounds of cargo transport. Thus after a successful fusion event, the fully assembled, extremely stable SNARE complex, which resides in a single lipid bilayer becomes the substrate for the general SNARE recycling machinery (7). The recycling machinery consists of two cytosolic components: (i) NSF, a hexameric ATPase that provides the catalytic function (8) and (ii) SNAPs, which act as adapters, allowing the binding of NSF to the various compartment-specific SNARE complexes (9). In mammalian cells three different SNAP homologues have been identified: α-SNAP and γ-SNAP, which are ubiquitously expressed, and β-SNAP, which is highly homologous to α-SNAP and found mainly in the adult brain (10). In yeast, only one SNAP orthologue, Sec17, exists, and its crystal structure reveals a C-terminal helical bundle and a N-terminal-twisted sheet of α-helical hairpins with an overall concave and convex surface (11, 12). Electron microscope images, site-directed mutagenesis, and analysis of the distribution of charged amino acids indicate that three SNAP molecules bind with their concave surface longitudinal to the four-helix SNARE bundle covering a large area of SNARE complex surface (13-15). Because SNAPs have to bind to each of the compartment-specific SNARE complexes, the complementary shapes of SNAPs and the SNARE complexes and electrostatic charge distributions are major determinants of this interaction. Subsequently, one NSF hexamer binds in an ATP-dependent manner to the SNAP-SNARE complex at its membrane distal end (13). The precise mechanism by which ATP hydrolysis causes conformational changes in NSF and disassembles the SNAP-SNARE complex is, however, still unknown (14). This process regenerates individual v- and t-SNARE, which are largely unfolded and energy rich, but kinetically stable. Remarkably, this step is the only known reaction where energy is invested in the overall membrane fusion process.

The Fusion Machinery, Physiological Regulation, and Disease

Because membrane trafficking is essential for cell homeostasis and many physiological processes, a complete functional loss of the recycling machinery (NSF, SNAPs) or most SNAREs leads to cell death or embryonic lethality. However, partial protein inactivation, e.g., by a point mutation, resulting in a specific pathophysiology can provide important information about the role of that protein in a select physiological process. For example, although a homozygous knockout of syntaxin 4 (a plasma membrane t-SNARE component) is embryonically lethal, heterozygous mice that are characterized by a reduced syntaxin 4 level develop muscle insulin resistance (16). This phenotype is consistent with a role of syntaxin 4 in the fusion of glucose transporter containing transport vesicles with the plasma membrane and the regulation of the blood glucose level by the muscle. It also illustrates how the copy number of a specific membrane trafficking protein can be more or less critical for certain physiological processes. This principle is obeyed throughout nature, and thus numerous physiological reactions are controlled by the availability/activity of proteins through alterations in their expression level, postranslational modifications, or through direct binding of regulatory factors. A key feature of physiological regulation is that it is temporally and in many cases spatially restricted, in contrast to mutants that introduce permanent changes and often result in pathophysiology. For example, the reversible down-regulation of NSF activity by S-nitrosylation demonstrates how the temporal and spatial regulation of a general factor involved in SNARE recycling specifically inhibits certain types of regulated exocytosis, without shutting down vesicular trafficking in general (17). On the other hand, temperature-sensitive mutants of NSF/Sec18 and SNAP/Sec17 in yeast result in vesicle accumulation and a general halt in membrane trafficking at the restrictive temperature (18, 19). In comatose, a temperature-sensitive mutant of Drosophila NSF, the primary effect of a shift to the restrictive temperature is a paralysis of the flies (20). These temperature-sensitive mutants, which cause a dramatic inactivation of the corresponding proteins, clearly demonstrate a general role of SNAP and NSF in vesicular trafficking and confirm NSF's physiological relevance for continuous neurotransmitter release. These observations raise the question of whether subtle, but permanent, changes in the activity of the general SNARE recycling machinery have effects on secretory processes important for the development of complex organisms.

An α-SNAP Point Mutation Linked to Hydrocephaly

In this issue of PNAS, Hong et al. (1) show that a point mutation in α-SNAP causes hydrocephaly with hop gait (hyh) in mice. Hydrocephaly with hop gait is a specific form of a broader class of congential hydrocephalies characterized by impaired brain development and an accumulation of the cerebrospinal fluid in cerebral cavities as a result of mutations in various genes (21). Fine-resolution mapping, DNA sequencing, extensive sequence comparisons with WT inbred mice, and database searches identified α-SNAP as the only mutated protein among seven candidates in the hyh locus. The point mutation is a conservative change of methionine in position 105 to isoleucine, resulting in reduced α-SNAP protein levels.

The mutated methionine in position 105 is conserved among several mammalian species and in Drosophila, but it is replaced by a glutamate in the yeast orthologue Sec17 and in Caenorhabditis elegans. In the crystal structure of Sec17, this methionine is located on the convex surface that is involved in SNARE complex disassembly and NSF binding. However, the mutation does not result in a major impairment of α-SNAP function, because mutant α-SNAP forms complexes with SNAREs and NSF, which are disassembled upon ATP hydrolysis. Although it cannot be excluded that a subtle change in α-SNAP activity could cause the pathological hyh phenotype, the reduced level of α-SNAP appears to be the critical determinant. There seems to be a particularly high demand for the general membrane trafficking protein α-SNAP during brain development, suggesting that efficient vesicular transport of certain signaling factors, receptors, or cell adhesion proteins is essential for proper brain morphogenesis. This issue is addressed in independent studies by Walsh and colleagues (22), who report results similar to those of Takahashi et al. and further demonstrate that the expression of WT α-SNAP rescues the hyh phenotype. Extensive morphological analyses show that the reduced size of cerebral cortex in hyh mice correlates with a premature production of cerebral cortical neurons accompanied by depletion of the progenitor pool. This alteration in cell fate appears to be caused by an abnormal localization of apical and adherens junction proteins, including those implicated in cell fate determination. Interestingly, the authors also report a decreased apical localization of Vamp-7, a v-SNARE involved in apical transport, which could provide the direct link to α-SNAP function and membrane trafficking. It remains to be determined whether thorough analysis of the mutant α-SNAP hyh mice will reveal additional defects in other physiological processes or unveil α-SNAP functions other than SNARE recycling.