Complexins Facilitate Neurotransmitter Release at Excitatory and Inhibitory Synapses in Mammalian Central Nervous System
Total Page:16
File Type:pdf, Size:1020Kb
Complexins facilitate neurotransmitter release at excitatory and inhibitory synapses in mammalian central nervous system Mingshan Xue*, Alicja Stradomska†, Hongmei Chen*, Nils Brose‡§, Weiqi Zhang†, Christian Rosenmund*§, and Kerstin Reim‡ *Departments of Neuroscience and Molecular and Human Genetics, Baylor College of Medicine, Houston, TX 77030; †Center for Physiology and Pathophysiology and Center for the Molecular Physiology of the Brain, Georg August University Go¨ttingen, D-37073 Go¨ttingen, Germany; and ‡Department of Molecular Neurobiology and Center for the Molecular Physiology of the Brain, Max Planck Institute of Experimental Medicine, D-37075 Go¨ttingen, Germany Communicated by Huda Y. Zoghbi, Baylor College of Medicine, Houston, TX, March 28, 2008 (received for review February 22, 2008) Complexins (Cplxs) are key regulators of synaptic exocytosis, but compared with that in fly neuromuscular junction and in vitro whether they act as facilitators or inhibitors is currently being fusion assays was suggested to be caused by the compensatory disputed controversially. We show that genetic deletion of all effects of the yet-uncharacterized CplxIII and CplxIV in CplxI/ Cplxs expressed in the mouse brain causes a reduction in Ca2؉- II-DKO mice, thus confounding the interpretation of the mouse triggered and spontaneous neurotransmitter release at both exci- CplxI/II-DKO phenotype (5, 18, 20). However, this hypothesis is tatory and inhibitory synapses. Our results demonstrate that at not compatible with the finding that overexpression of mouse mammalian central nervous system synapses, Cplxs facilitate neu- CplxIII or CplxIV rescues the phenotype of CplxI/II-DKO hip- rotransmitter release and do not simply act as inhibitory clamps of pocampal neurons (11). the synaptic vesicle fusion machinery. To stringently define the role of Cplxs in synaptic exocytosis at mammalian synapses, we genetically eliminated all Cplxs Ca2ϩ trigger ͉ exocytosis ͉ SNARE expressed in the mouse brain and studied neurotransmitter release at two major types of synapses in the central nervous ynaptic exocytosis and constitutive secretion share a similar system, excitatory and inhibitory synapses. membrane fusion machinery involving the soluble N- S Results ethylmaleimide-sensitive factor attachment protein receptors (SNAREs) and SM proteins (1, 2). Constitutive secretion occurs Generation of CplxI/II/III Triple KO Mice. We used homologous spontaneously and continuously. In contrast, Ca2ϩ-triggered recombination in embryonic stem cells to generate CplxIII-KO synaptic exocytosis is the most tightly regulated membrane mice (Fig. 1). The first coding exon of CplxIII was replaced by a fusion reaction and exhibits extremely high speed and accuracy neomycin resistance cassette (Fig. 1 a and b). Western blot (3). The molecular basis of this exquisite regulation of synaptic analyses confirmed that the expression of CplxIII was eliminated vesicle fusion is still not fully understood. One recent hypothesis in the brain of homozygous mutant mice (Fig. 1c). We bred states that even at synapses the SNARE-mediated fusion is CplxIII-KO mice with the previously generated CplxI-KO and constitutively active and that a clamping mechanism arrests the CplxII-KO mice (15) to generate CplxI/II/III triple KO (CplxI/ fusion machinery until the arrival of a Ca2ϩ-dependent triggering II/III-TKO) mice. Homozygous CplxIII-KO and CplxII/III-DKO signal, which removes the clamp and allows constitutive fusion to mice are viable and fertile, but CplxI/II-DKO and CplxI/II/III- resume (4, 5). However, it is difficult to reconcile this hypothesis TKO mice die within a few hours after birth. Western blot with the fact that constitutive secretory processes proceed by analyses confirmed that all known Cplxs are absent from the several orders of magnitude more slowly than fast neurotrans- brains of CplxI/II/III-TKO mice [supporting information (SI) mitter release (3). An alternative hypothesis is that the SNARE- Fig. S1]. mediated membrane fusion is intrinsically slow (6). To achieve the speed and accuracy of synaptic exocytosis, additional facili- Reduced Evoked Neurotransmitter Release in CplxI/II-DKO and CplxI/ tatory and inhibitory mechanisms are necessary to accelerate II/III-TKO Neurons. We used whole-cell patch clamp recording to and fine-tune synaptic vesicle fusion (3, 7). study neurotransmitter release in cultured autaptic hippocampal Complexins (Cplxs) constitute a family of SNARE complex excitatory and striatal inhibitory neurons. We first examined the binding proteins (8–11) that were suggested to be either facili- basal synaptic transmission by evoking action potentials at low tators or inhibitors of vesicle exocytosis (5, 11–22). The mouse frequency (0.2 Hz) and measuring the synaptic responses. Action genome contains four Cplx genes (CplxI–IV). CplxI, II, and III potential-evoked excitatory and inhibitory postsynaptic currents are expressed in the brain and the retina, whereas Cplx IV is (EPSCs and IPSCs, respectively) from CplxIII-KO and CplxII/ present only at retinal ribbon synapses (11). CplxI and CplxII III-DKO neurons were indistinguishable from their respective double knockout (CplxI/II-DKO) mice exhibit reduced Ca2ϩ- controls (Fig. 2 a–c). However, both EPSCs and IPSCs were triggered fast neurotransmitter release at hippocampal glutama- drastically reduced in CplxI/II-DKO and CplxI/II/III-TKO neu- tergic synapses, indicating that Cplxs are positive regulators of NEUROSCIENCE transmitter release (15). In contrast, Cplxs inhibit SNARE- Author contributions: M.X., N.B., W.Z., C.R., and K.R. designed research; M.X., A.S., H.C., mediated liposome and cell fusions in vitro, which led to the C.R., and K.R. performed research; M.X., A.S., W.Z., C.R., and K.R. analyzed data; and M.X., hypothesis that Cplxs act as fusion clamps of synaptic exocytosis N.B., W.Z., C.R., and K.R. wrote the paper. (5, 18). At the larval neuromuscular junction of a Drosophila Cplx The authors declare no conflict of interest. (dmCplx) null-mutant, the spontaneous transmitter release is §To whom correspondence may be addressed. E-mail: [email protected] or rosenmun@ drastically increased (20). This observation was interpreted as bcm.tmc.edu. evidence for the fusion clamp function of Cplxs, even though This article contains supporting information online at www.pnas.org/cgi/content/full/ ϩ Ca2 -evoked release is decreased in the same mutant (20). The 0803012105/DCSupplemental. apparent difference of Cplx function in mouse neurons as © 2008 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0803012105 PNAS ͉ June 3, 2008 ͉ vol. 105 ͉ no. 22 ͉ 7875–7880 Downloaded by guest on September 25, 2021 Fig. 1. Generation of CplxIII-KO mice. (a) Maps of WT CplxIII gene, targeting vector, and mutated CplxIII gene are shown. Exons (black boxes) and restriction enzyme sites are indicated. The open bar indicates the position of the outside probe used to identify the mutant allele in Southern blot analyses (see b). Neo, neomycin resistance gene; TK, thymidine kinase gene. (b) Southern blot analysis of mouse tail DNA for the CplxIII deletion mutation. Arrows indicate the positions of WT and KO alleles. (c) Western blot analyses of CplxIII expression in the brain of adult mice using a CplxIII antibody (11). CplxIII expression is reduced in heterozygous (ϩ/Ϫ) mice and abolished in homozygous KO (Ϫ/Ϫ) mice. rons (Fig. 3 a–d). These reductions in the basal synaptic trans- and mIPSCs, respectively) were either unchanged or only slightly mission are caused by defects in the presynaptic release process reduced as compared with the controls (Table S1 and see below). because the amplitudes of miniature EPSCs and IPSCs (mEPSCs To directly assess Ca2ϩ-triggered release efficiency, we mea- sured vesicular release probability (Pvr). Pvr is defined as the fraction of vesicles in the readily releasable pool (RRP) that is released by a single action potential. We recorded synaptic responses induced by hypertonic sucrose solution to measure the sizes of RRP (23) and found no significant changes in any of the Cplx mutant neurons as compared with their controls (Table S1). This finding indicates that the numbers of fusion-competent vesicles are normal in the absence of Cplxs. Consequently, the Pvr of CplxIII-KO or CplxII/III-DKO was not changed as compared with their controls (Fig. 2d). In contrast, Pvr was reduced to Ϸ50% of the control levels in CplxI/II-DKO and CplxI/II/III- TKO neurons (Fig. 3 e and f). These results indicate that Cplxs promote Ca2ϩ-triggered synaptic exocytosis at a postpriming step. Altered Short-Term Plasticity in CplxI/II-DKO and CplxI/II/III-TKO Neu- rons. Changes in release probability often lead to alterations in short-term plasticity of synaptic transmission. In general, trains of action potentials cause synaptic depression or facilitation at synapses with high or low initial release probability, respectively (24). Consistent with the Pvr data, the synaptic responses of CplxI/II-DKO and CplxI/II/III-TKO neurons during a train of high-frequency stimulation showed marked initial facilitation, whereas those of WT, CplxIII-KO, and CplxII/III-DKO neurons showed depression (Fig. 4 a and b, Fig. S2). 2؉ Fig. 2. Evoked synaptic transmission of CplxIII-KO and CplxII/III-DKO neurons. (a Reduced Apparent Ca Sensitivity of Release in CplxI/II-DKO and 2ϩ and b) Representative traces of basal evoked EPSCs of hippocampal neurons (a) CplxI/II/III-TKO Neurons. We next studied the Ca dependence of and IPSCs of striatal neurons (b). Vertical bars represents 2-ms somatic depolar- fast transmitter release by measuring the amplitudes of the izations; depolarization artifact and action potentials were blanked. (c and d) Bar evoked synaptic responses as a function of the external Ca2ϩ graphs show summary data of evoked synaptic response amplitudes (c) and concentrations. The data were fitted with a standard Hill vesicular release probability (d). Data were normalized to the mean values of the equation to obtain the dissociation constant (Kd) as a measure respective controls (dotted line).