Expression of Mutant Huntingtin Blocks Exocytosis in PC12 Cells by Depletion of Complexin II*

Huntington’s disease (HD) is an autosomal dominant neurodegenerative caused by an expanded CAG repeat in the HD gene. We reported recently that complexin II, a protein involved in neurotransmitter release, is depleted from both the brains of mice carrying the HD mutation and from the striatum of post mortem HD brains. Here we show that this loss of complexin II is recapitulated in PC12 cells expressing the HD mutation, and is accompanied by a dramatic decline in Ca 2+ -triggered exocytosis of neurotransmitter. Overexpression of complexin II (but not complexin I) rescued exocytosis, demonstrating that the decline in neurotransmitter release is a direct consequence of complexin II depletion. Complexin II depletion in the brain may account for some of the abnormalities in neurotransmission associated with HD.


3
Huntington's disease (HD) 1 is an autosomal dominant neurodegenerative disorder characterized by motor, emotional and cognitive dysfunction. It is caused by an expanded CAG repeat in the HD gene, which encodes the widely-expressed 348-kDa protein, huntingtin (htt; 1, 2). The expanded polyglutamine repeat is likely to cause a gain of function in HD (3)(4)(5), although the mechanism underlying the pathology is unknown. The role of htt in the cell is also not clear. Htt is a cytosolic protein that is found loosely attached to synaptic vesicles (6). It is known to interact with a number of different proteins, including HAP1 (7) and HIP1 (8,9), which themselves associate with vesicles. It has therefore been proposed that abnormal protein interactions with htt may influence neurotransmitter release or membrane recapture by endocytosis, and thereby cause neuronal dysfunction (6,10).
The R6/2 transgenic mouse expresses the first exon of the HD gene with an expanded CAG repeat (11). Although little neurodegeneration is seen in this mouse prior to its premature death (usually at 14-16 weeks), its phenotype has a number of similarities with HD, for example the dominant inheritance and the progressive nature of the neurological deficits (12,13). The mice develop normally and do not show frank motor symptoms until about 8 weeks of age (11). However, motor and cognitive deficits are apparent from about 4 weeks (12,13), and presymptomatic alterations in long-term potentiation are also seen (14). The pronounced neurological phenotype of the R6/2 mouse in the absence of neurodegeneration suggests that neuronal loss in HD is secondary to neuronal dysfunction. In support of this suggestion, motor (15,16) and cognitive deficits (17,18) can be detected in HD patients before neurodegenerative changes are seen.
To examine the possibility that abnormalities in neurotransmission underlie early events in the development of HD, we previously looked for changes in the brains of by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 4 R6/2 mice in the levels and distribution of proteins involved in neurotransmitter release (19). The proteins studied included the soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (SNAREs) synaptobrevin, syntaxin and SNAP-25, which are known to form the core of a ubiquitous membrane fusion apparatus (20), and various 'accessory proteins', such as the synaptic vesicle proteins synaptotagmin, synaptophysin and rab3A, and the cytosolic proteins α-SNAP and complexins I and II. These proteins are believed to control SNARE complex assembly and disassembly, and thereby modulate neurotransmitter release (20). We found that complexin II was specifically and progressively depleted from the brains of R6/2 mice (19). Furthermore, in 16-week old mice, complexin II appeared in a subpopulation of neuronal intranuclear inclusions, which are a characteristic feature of brains from both mouse HD models (21) and human HD patients (22). Significantly, the depletion of complexin II was also seen in the striatum of HD brains (23), the region most severely affected in HD (2). The depletion was apparent at an early stage (grade 0), before neurodegeneration is seen. Decreases in the levels of synaptobrevin 2 and rab3A were also seen in the striatum, but none of the other proteins tested was significantly affected.
In our previous studies, we established a correlation between complexin II depletion and neurological dysfunction in both HD patients and a mouse model of HD. In the present study, we sought to determine whether expression of mutant htt was able to cause complexin II depletion, and whether this effect in turn compromised Ca 2+ -triggered exocytosis.

Real-time Voltammetric Measurement of Neurotransmitter Release from
Permeabilized PC12 Cells-Cells were detached and permeabilized by centrifugation at 1600 x g for 3 min; this procedure results in the permeabilization of approximately 80% of the cells, as determined by trypan blue staining (25). Cells were used immediately after permeabilization to avoid rundown in exocytosis. A cell suspension (350 µl containing ~10 7 cells), in 100 mM NaCl, 50 mM Hepes (pH 7.4), containing 1 mg/ml bovine serum albumin and 2 mM EGTA, was added to a temperaturecontrolled incubation chamber (containing Ag/AgCl reference and platinum auxiliary electrodes) set at 37°C. A glassy carbon RDE (E app = +500 mV versus the reference electrode) was rotated in the cell suspension at 3,000 rpm. Once a stable baseline was obtained (usually 1 min), a CaCl 2 solution was rapidly injected into the suspension to achieve a free Ca 2+ concentration of 100 µM. Catecholamine (predominantly dopamine) released from the cells is oxidized at the surface of the RDE, and the Hoechst 33258 dye (5 µg/ml in PBS) for 30 min at 37°C. After another five washes, coverslips were mounted on glass slides with Pro-Long mounting reagent, and examined using a Nikon Eclipse TE2000 microscope, equipped with a DXM1200 digital camera.

Amperometric Detection of Neurotransmitter Release from Single Intact PC12
Cells-Catecholamine was detected as an oxidation current using 5 µm carbon fibres polarized to 650 mV. Current was recorded at room temperature (~22°C) with a VA-

RESULTS
For our experiments we used a PC12 cell line stably expressing the first exon of htt with a 74-glutamine repeat and an N-terminal GFP tag, via the Tet-On system (26). Cells were treated for various times with doxycycline (1 µg/ml) to initiate expression of the protein. By 24 h, the cells showed a marked increase in GFP expression, which was accompanied by the appearance of GFP-containing aggregates ('inclusions') in some cells. By 48h, 62% of the cells had inclusions, and this 8 percentage remained stable thereafter (Fig. 1a). Expression of the htt fragment did not cause significant cell death, up to 120 h (data not shown). To measure neurotransmitter release, cells were permeabilized and exocytosis was triggered by addition of Ca 2+ (100 µM). Release of catecholamine neurotransmitter was detected in real time using rotating carbon disc electrode voltammetry (25). The amount of neurotransmitter released fell progressively after expression of the HD mutation, and was almost undetectable after 72 h (Fig. 1b, c). Over the same time course, there was no consistent change in the ability of the PC12 cells to take up [ 3 H]norepinephrine ( Fig. 1d), indicating that the fall in neurotransmitter release represented a defect in exocytosis rather than a reduction in the neurotransmitter content of the dense-core vesicles. In contrast to the result with the cells expressing the mutant htt fragment, exocytosis was not compromised in doxycycline-treated cells stably expressing the Tet-On vector alone (Fig. 1e).
Cells used in the neurotransmitter release assays were analysed by SDSpolyacrylamide gel electrophoresis and immunoblotting with antibodies to proteins known to be involved in exocytotic membrane fusion (20). There was no timedependent change in the overall pattern of protein expression, as revealed by Coomassie blue staining (Fig. 2a). There was also no difference between untreated cells and cells that had expressed the mutant htt fragment for 120 h in the levels of the vesicle membrane proteins synaptobrevin 2, synaptotagmin I, synaptotagmin IX and synaptophysin, or in the plasma membrane proteins syntaxin 1 and SNAP-25 (Fig.   2b). In contrast, the cytosolic protein complexin II was dramatically depleted after 9 Further, consistent with its lack of effect on exocytosis, doxycycline treatment caused no change in complexin II levels in PC12 cells expressing Tet-On vector alone (Fig.   2d).
In addition to changing complexin II levels, expression of the mutant htt fragment caused a redistribution of the protein within the PC12 cells. In untreated cells (Fig.   3a-c), complexin II had a punctate distribution throughout the cytoplasm, whereas in cells treated with doxycycline for 72 h (Fig. 3d-f), complexin II was predominantly present in large aggregates which partially corresponded with inclusions containing GFP-tagged mutant htt fragment, indicated by the arrows in Fig. 3d, e.
To determine whether there was a causal connection between complexin II depletion and inhibition of exocytosis, we tested the ability of complexin II to 'rescue' exocytosis in PC12 cells expressing the mutant htt fragment. Initially, we repeated the experiments shown in Fig. 1, and pre-incubated the permeabilized PC12 cells for 1 min with recombinant His 6 -tagged complexin II before addition of Ca 2+ (100 µM). No increase in neurotransmitter release was seen, irrespective of the duration of expression of the mutant htt fragment (data not shown). Hence, complexin II was not able to rescue exocytosis when provided acutely. We therefore decided to overexpress complexin II over the duration of expression of the mutant htt fragment.
Cells were transiently transfected with a bicistronic vector encoding both complexin II and CFP; cyan fluorescence was used to identify transfected cells. Expression of the mutant htt fragment was initiated 24 h after transfection with complexin II/CFP, and doxycycline treatment continued for 72 h. Neurotransmitter release from single cells in response to KCl depolarization was measured by carbon-fibre amperometry, which is able to detect the release of neurotransmitter from individual dense-core vesicles (27,28). In cells expressing the mutant htt fragment, the amperometric recordings showed that vesicular release occurred infrequently (Fig. 4a, d), consistent with our voltammetry results showing reduced release in these cells (Fig. 1). Overexpression of complexin II caused a significant (approximately 3-fold) increase in the number of release events evoked by depolarization (Fig. 4c, d), and also increased the fraction of secreting cells from 50% to 85%. In contrast, overexpression of complexin I, had the opposite effect (Fig. 4b, d), reducing the cumulative spike number to about a third of that seen in control cells. However, the fraction of cells showing no secretion remained the same as controls (50%). In wild type PC12 cells, overexpression of complexin II reduced the number of evoked release events by about 50% (Fig. 4e), whereas complexin I overexpression had no effect on spike number.

DISCUSSION
The mechanism underlying polyglutamine-dependent pathogenesis in HD is unknown. Wild type htt exists predominantly in the cytoplasm (29). Mutant htt is also found in the cytoplasm; however, the N-terminal region of mutant htt, and perhaps even the full-length protein, appear to be targeted to the nucleus, at least under some circumstances (30). Interestingly, mutant htt is also recruited into abnormal aggregates of protein in nuclei of neurons, both in HD brains and in mouse models of HD. The significance of these neuronal intranuclear inclusions is disputed, and evidence has been presented supporting claims that they are toxic (21), protective (31) or merely epiphenomena (32). Nevertheless, these aggregates have been isolated and shown to contain many other proteins in addition to htt (33). In light of these findings, it has been suggested that the recruitment of proteins into the inclusions, as a consequence of abnormal interactions with htt, might impact on cell function. Recent evidence has also suggested that mutant htt dysregulates transcription in neurons (34), and the expression of many genes is known to be altered in mouse models of HD (35).
In the present study, we found that the total amount of complexin II in PC12 cells expressing mutant htt fragment initially rose (at 24 h), and then fell below normal levels (from 48 h onwards). The initial rise in complexin II levels was not accompanied by any change in neurotransmitter release; however, the later complexin II depletion was associated with a concomitant fall in neurotransmitter release. In addition to the changes in total amounts of complexin II within the cells, there was an intracellular redistribution of this protein. In particular, at later times (72 h) complexin II appeared to accumulate, at least partially, in the inclusions. Similar changes (both complexin II depletion and recruitment into inclusions) have previously been seen over a much longer time course in the brains of R6/2 mice (19). Complexin II is also depleted in human HD striatum (23). Interestingly, in hippocampus from HD brains, complexin II levels were actually found to be elevated. This effect could represent a physiological 'rescue' of synaptic transmission, since hippocampal function is not affected until late in the course of the disease.
The mechanism underlying the changes in complexin II levels seen in the present study is unclear. There was no difference between complexin II mRNA levels in untreated cells and in cells treated with doxycycline for 120 h (data not shown), indicating that effects at the level of transcription are unlikely, and that the depletion of the protein at later times reflects an increased rate of degradation. It is possible that the gradual recruitment of complexin II into inclusions, together with the fall in total amount of cellular complexin II would have resulted in a depletion of this protein from the cytosol and a consequent reduction in its availability at sites of exocytotic membrane fusion. The fact that exocytosis was substantially rescued by co-expression 12 of complexin II supports the idea that exocytosis was compromised by a lack of functional complexin II.
Complexin II binds rapidly and with high affinity to the SNARE complex (36,37); however, its precise role in membrane fusion is still undetermined. It was originally proposed to be a negative regulator of exocytosis, based on the ability of an anti-complexin II antibody to stimulate neurotransmitter release from neurons in the Aplysia buccal ganglion, and of recombinant complexin II to inhibit release (38). In support of these findings, it was shown recently that overexpression of complexin II inhibited exocytosis of small synaptic-like vesicles in PC12 cells (39) and of large dense-core vesicles in a related system, the adrenal chromaffin cell (40). In contrast, the deficiencies in long-term potentiation in complexin II-deficient mice (41), and the reduced Ca 2+ -sensitivity of neurotransmitter release in mice lacking both complexins I and II (42), suggest that the complexins have a positive role in exocytosis.
Furthermore, complexin II has recently been shown to enhance an interaction between the complementary transmembrane regions of syntaxin and synaptobrevin, suggesting a mechanism by which it could promote membrane fusion (43). In our experiments, complexin II reduced exocytosis in control PC12 cells but enhanced it in cells expressing the mutant htt fragment. Interestingly, spike numbers in response to overexpression of complexin II were similar in the wild type cells and in the cells expressing the mutant htt fragment. This result suggests that although complexin II is required to support exocytosis, there is an optimal cytosolic concentration, above which exocytosis becomes inhibited, perhaps because other crucial factors are titrated out.
Complexins I and II have distinct, but overlapping distributions within the brain (38,42,44). Mice lacking complexins I and II have distinct phenotypes, which might Complexin I-deficient mice develop a strong ataxia and suffer from sporadic seizures, whereas complexin II-deficient mice show no obvious phenotypic abnormalities (42).
In fact, defects in neurotransmission, particularly long-term potentiation, have been identified in complexin II-deficient mice (41). Furthermore, we have shown recently that the complexin II-deficient mice, while appearing outwardly normal, in fact show progressive deficits in a number of complex behaviors. These include abnormalities in exploration, socialization, motor co-ordination and learning, suggesting that complexin II is essential for the maturation of higher cognitive functions (Morton; unpublished results). It is particularly interesting that similar deficits are found in R6/2 mice (12,13), which also show a progressive depletion of complexin II (19).
The close sequence similarity between complexins I and II has led to the implicit assumption that they are likely to play similar roles. However, in our study, there was a sharp contrast between the functional effects of overexpression of complexins I and II, suggesting that the two proteins have distinct molecular properties. Few studies have directly compared the properties of the two proteins, although it has been shown that complexin II binds to and dissociates from the SNARE complex more rapidly than complexin I (37). Whether this relatively minor biochemical difference can account for the significant difference in the function of the two complexins is unclear.
We have shown that expression of the HD mutation in PC12 cells causes a specific depletion of complexin II, and that this depletion causes a reduction in the capacity of these cells to release neurotransmitter. Our results suggest a molecular mechanism that might account for the deficiencies in synaptic transmission that are known to occur in HD (14), and which might in turn underlie the neurological symptoms that characterize the disease (2).