Mutant huntingtin reduces vesicular zinc level by inhibiting the binding of Sp1 to ZnT3 promoter

Synaptic dysfunction caused by mutant huntingtin greatly contributes to Huntington’s disease (HD) pathogenesis. HD patients show cognitive impairment as well as uncontrolled movements. Vesicular zinc is closely linked to modulating synaptic transmission and maintaining cognitive ability. However, whether does mutant huntingtin affect zinc homeostasis in the brain or not? This will be of great signicance for further revealing the pathogenesis of HD. N171-HD82Q transgenic mice and cultured BHK cells expressing N-terminal mutant huntingtin fragment containing 160 glutamines (160Q were used to investigate the effect of mutant on zinc homeostasis and its molecular mechanisms.


Abstract Background
Synaptic dysfunction caused by mutant huntingtin greatly contributes to Huntington's disease (HD) pathogenesis. HD patients show cognitive impairment as well as uncontrolled movements. Vesicular zinc is closely linked to modulating synaptic transmission and maintaining cognitive ability. However, whether does mutant huntingtin affect zinc homeostasis in the brain or not? This will be of great signi cance for further revealing the pathogenesis of HD. Methods N171-HD82Q transgenic mice and cultured BHK cells expressing N-terminal mutant huntingtin fragment containing 160 glutamines (160Q BHK cells) were used to investigate the effect of mutant huntingtin on zinc homeostasis and its molecular mechanisms.

Results
Herein, we have demonstrated that the density of synaptic vesicular zinc decreases in the cortex, striatum and hippocampus of N171-82Q mice. Given that vesicular zinc concentration depends on the abundance of zinc transporter 3 (ZnT3) on the membrane of synaptic vesicles, ZnT3 expression is detected in the brain of N171-82Q mice and 160Q BHK cells. Mutant huntingtin leads to a dramatical decrease in ZnT3 mRNA and protein levels in the three brain regions of these mice aged from 14 to 20 weeks. Signi cantly, Sp1 activates ZnT3 transcription via its binding to the GC boxes in ZnT3 promoter. Nevertheless, mutant huntingtin inhibits the binding of Sp1 to the promoter of ZnT3 gene and down-regulates ZnT3 expression.
Furthermore, the overexpression of Sp1 ameliorates inhibition of ZnT3 gene transcription by mutant huntingtin.

Conclusions
Collectively, this rst study to reveal a signi cant loss of synaptic vesicular zinc and ZnT3 expression caused by mutant huntingtin in the early stage of HD. Our ndings have revealed the molecular mechanism underlying this change. Mutant huntingtin inhibits the binding of Sp1 to ZnT3 gene promoter to reduce ZnT3 expression. The imbalance of vesicular zinc homeostasis may be closely associated with synaptic dysfunction and cognitive de cits in HD. This work sheds novel mechanistic insights into the pathogenesis of HD and promises a potential therapeutic strategy for HD.
Synaptic dysfunction greatly contributes to HD pathogenesis [10,11]. Neurotransmitter release signi cantly alters in the transgenic HD mouse models [12,13]. Neuronal and synaptic dysfunction precedes cell death by many years in the HD patients [14,15] and animal models [16]. Furthermore, some pharmacological interventions focus on targeting early synaptic disturbances, which has been proven to restore synaptic function [17][18][19] and delay progression to neurodegeneration in HD transgenic mice [20,21]. Therefore, disturbed synaptic function accounts for the early symptoms of HD and triggers neuronal death in later stages of the disease [11,22,23]. It is vital to investigate the mechanism of synaptic damage in HD disease.
The divalent cation zinc in the brain contributes to e cient synaptic transmission. Approximately, 85% of total brain zinc is tightly bound to metalloproteins. 10 ∼ 15% of total brain zinc is highly localized in synaptic vesicles of excitatory glutamatergic neurons [24,25]. This pool of zinc, the ionic zinc, is either free or chelatable. It can be detected with simple histochemical method such as neo-Timm's sul de-silver method [26,27], and is thus often called histochemically reactive zinc. Neurons containing histochemically detectable zinc are present in many regions of the brain, including the neocortex, striatum, hippocampus, amygdale and olfactory bulb [28]. Vesicular zinc is closely linked to modulating synaptic transmission. It serves as a singal factor to play an important role in modifying glutamatergic neurons [29][30][31]. Zinc releases from glutamatergic neuron terminals, which may protect neurons from excitotoxicity of glutamate to attenuate the excess amount of presynaptic glutamate release [32]. Zinc de ciency affects neurogenesis and trigger neuronal apoptosis. Therefore, this can result in learning and memory de cits [32].
More improtantly, the homeostasis of zinc in the brain is tightly regulated. The zinc transporters (ZnTs) mainly function to e ux zinc out of cytoplasm or into intracellular organelles [33]. Among them, zinc transporter 3 (ZnT3), the primary vesicular zinc transporter, is located on the membrane of synaptic vesicles to transport zinc ions into presynaptic vesicles from the cytosol [34]. The concentration of vesicular zinc depends on the abundance of ZnT3 [34,35]. Targeted deletion of ZnT3 gene eliminates zinc from synaptic vesicle, which leads to age-dependent de cits in learning and memory ability [36] and neurodegeneration [36][37][38]. Consequently, ZnT3-dependent zinc homeostasis in synaptic vesicles takes an important role in maintaining synaptic function.
An imbalance of vesicular zinc homeostasis is associated with the pathogenesis of multiple neurodegenerative diseases, including Parkinson's disease (PD) [39,40], Alzheimer's disease (AD) [39,41,42,43,44] and amyotrophic lateral sclerosis (ALS) [45]. These diseases have in common with HD features and mechanisms that the misdolded proteins cause neuronal death at the late onset of the disorder. Altered homeostasis of essential elements such as iron, chromium and selenium has been observed in the HD patients [46,47] and mice model [48]. Specially, increased level of zinc is detected in the blood of HD patients, indicating that mutant Htt (mHtt) might impair zinc homeostasis [46]. Nevertheless, whether does mHtt affect zinc homeostasis in the brain or not? This will be of great signi cance for further revealing the pathogenesis of HD. Here, we have demonstrated that mHtt reduces ZnT3 expression by inhibiting the binding of Sp1 to ZnT3 gene promoter to down-regulate vesicular zinc level in the brain of N171-82Q HD transgenic mice. Disruption of vesicular zinc homeostasis will ultimately contribute to synaptic dysfunction and neurodegeneration in HD.

Materials And Methods
Huntington's disease transgenic mice B6C3-Tg (HD82Gln) 81Dbo/J (N171-HD82Q) HD transgenic mice (Jackson Laboratories) express a cDNA encoding a 171 amino acid N-terminal fragment of huntingtin containing 82 CAG (Q) repeats [48]. At the age of 4 weeks, mice were genotyped by polymerase chain reaction (PCR) of tail DNA genotyped according to the Jackson Laboratories protocol to determine hemizygosity for the HD transgene. Wildtype littermates were used as controls.

Autometallography (AMG)
The 20-week-old N171-82Q mice and age-matched WT mice (n = 3) were anesthetized and then perfused intracardially with 0.3 % sodium sul de solution in a 0.1 M sodium phosphate buffer (PB, pH 7.4) 150 ml, followed by transcardial perfusion with 250 ml 4 % paraformaldehyde in 0.1 M PB, and nally with the same sodium sul de solution again for 150 ml [50]. After perfusion, the brains were removed and postxed with 4 % paraformaldehyde in 0.1 M PB overnight. The brains were placed in 30 % sucrose overnight.
For light microscopy (LM), the samples were cut into 10 μm coronal sections in a cryostat and placed on Farmer uid cleaned glass slides and AMG developed [51]. The sections were immersed in a citrate buffered silver lactate/hydroquinone developer containing a protective colloid (gum arabic), and incubated in a 26 o C water bath shaking for 1 h, and then were stopped with a 5 % thiosulphate solution 10 min. The sections were rinsed with warm water 38 o C for 10 min in order to remove the gelatine membrane [52]. Images were taken on a Nikon microscope (Digital Camera DXM 1200).

RT-PCR
Total RNA from BHK cells, wild type and HD transgenic mice was extracted with Trizol reagent (Invitrogen). ZnT3 and Sp1 mRNA expression were ampli ed by reverse-transcription polymerase chain reaction, and β-actin mRNA was taken as an internal control.

Statistical analysis
Statistical analyses were carried out using SPSS 17.0 software for one-way ANVOA followed by Student's t test. All values were represented as mean ± S.D. Differences were considered signi cant if p < 0.05.

Results
N171-82Q mice display signi cant loss of total zinc and vesicular zinc in the brain tissue The ame atomic absorption spectrometry (FAAS) and autometallography (AMG) were respectively applied to explore the effect of mHtt on total and vesicular zinc level in the brain tissue of N171-82Q mice. Comparison of the 20-week-old N171-82Q mice with age-matched wild type (WT) mice showed that total zinc level was extremely low in the cortex, striatum and hippocampus examined in N171-82Q mice, compared to controls (Fig. 1).
AMG results showed that histochemically reactive zinc was signi cantly decreased in all three brain regions examined in the 20-week-old N171-82Q mice, compared to controls (Fig. 2). In the wild type (WT) mice, intense AMG staining was seen in the CA1, CA2 and CA3 region of hippocampus ( Fig. 2A, B), cortex (Fig. 2E) and striatum (Fig. 2E, F). However, the N171-82Q mice presented faint zinc staining in the corresponding brain regions (Fig. 2C, D, G, H).
At the ultrastructural level, the AMG reactive zinc granules were present within synaptic vesicles in striatum of the WT mice (Fig. 3A), whereas the number of positive vesicles was rarely, if at all, seen in striatum of the N171-82Q mice (Fig. 3B). These collective ndings indicates that zinc is unbalanced in the HD brain and suggests a role of zinc in HD pathogenesis.

ZnT3 expression is decreased in N171-82Q mice and BHK cells expressing mutant huntingtin
ZnT3 is required for zinc transport into synaptic vesicles and that vesicular zinc concentration is regulated by the amount of ZnT3 present on synaptic vesicle membranes [34][35][36]. Thus, we hypothesized that lower level of vesicular zinc may result from reduced ZnT3 in the N171-82Q mice. To test this assumption, we analysed ZnT3 expression in N171-82Q mice and age-matched WT mice. Light microscopic immunohistochemistry revealed that ZnT3 immunoreactivities were intensely present in the hippocampus, striatum and cortex (Fig. 4A, B) of the 20-week-old WT mice. Especially in the dentate gyrus (DG) and CA3 area of hippocampus, where are rich in axonal terminal boutons, intense ZnT3immunoreactive granules were seen. In contrast, age-matched N171-82Q mice displayed weak ZnT3immunoreactivities in the same brain regions (Fig. 4C, D).
To further identify whether ZnT3 expression was decreased in the N171-82Q mice, we investigated ZnT3 protein and mRNA level in different brain regions of the N171-82Q mice and age-matched WT mice. ZnT3 protein and mRNA levels in the cortex, hippocampus and striatum were markedly reduced in 14-, 18-, and 20-week-old N171-82Q mice, in comparison with age-matched WT controls (Fig. 5A, B). These results coincide with our assumption that mHtt disturbs ZnT3 expression in the N171-82Q mice.
Following this viewpoint, we further determined ZnT3 levels in cultured BHK cells expressing N-terminal mHtt containing 160 polyglutamine (160Q cells) or 20 polyglutamine (20Q cells). Being consistent with above ndings in the N171-82Q mice, identi ed signi cant de cit in ZnT3 expression level was detected in 160Q cells (Fig. 6). Associated with the above results, we can reasonably conclude that mHtt causes the loss of vesicular zinc by affecting ZnT3 expression.
Sp1 positively regulates ZnT3 expression through activating ZnT3 gene promotor Gene transcriptional dysregulation is greatly involved in the mechanisms of neurodegeneration in HD [53,54]. Whether does down-regulation of ZnT3 expression result from transcriptional inhibition or not in HD?
To further demonstrate Sp1 transactivates ZnT3 gene by binding GC boxes in the promoter, we pursued chromatin immunoprecipitation (ChIP) assay. Sp1 antibody precipitated more DNAs containing these ZnT3 promoter sequences with GC-1 or GC-2 than a mock immunoprecipitation with control IgG (Fig. 7C). This result indicates that Sp1 positively regulates ZnT3 transcription activity by binding directly to GC boxes.
Following this viewpoint, the role of Sp1 in the regulation of ZnT3 expression was examined in BHK cells. Sp1 overexpression promoted signi cantly the expression of endogenous ZnT3 mRNA and protein in WT BHK cells (Fig. 8). Whereas, the other transcription factors including NF-κB and WT1 did not affect ZnT3 expression in BHK cells (Fig. S1), indicating that Sp1 acts as a important transcript fact of ZnT3 gene.
In contrast, ZnT3 protein and mRNA level were signi cantly reduced in BHK cells transfected with Sp1 siRNA, but not with the non-targeting control siRNA (Fig. 9). We con rmed the reduction of Sp1 expression in the BHK cells transfected with Sp1 siRNA (Fig. S2). Consequently, these results imply that Sp1 plays a key role in transcriptional regulation of the ZnT3 gene.

Sp1 overexpression ameliorates inhibition of ZnT3 gene transcription by mutant huntingtin
We further investigated the effect of Sp1 on ZnT3 gene transcription in 160Q BHK cells. Sp1 overexpression increased signi cantly ZnT3 mRNA levels in 160Q BHK cells, compared to 20Q BHK cells (Fig. 10A, B).

Mutant huntingtin inhibits the binding of Sp1 to ZnT3 promoter
Mutant huntingtin has been demonstrated to inhibit the binding of Sp1 to target genes in HD [56][57][58][59], which might also be the mechanism of the effect of mHtt on ZnT3 expression. In this report, we examined the binding of Sp1 to ZnT3 promoter in 160Q and 20Q BHK cells. ChIP assay displayed that there was a signi cant decrease in Sp1-immunoprecipitated DNA consisting of ZnT3 promoter in 160Q cells, compared to 20Q cells [ Fig. 11]. In consequence, mHtt inhibits ZnT3 expression by disturbing the binding of Sp1 to the ZnT3 gene.

Discussion
Impairment of synaptic function contributes to the HD pathogenesis. Vesicular zinc is of signi cance for synaptic function. Herein, we identi ed a signi cant loss of total zinc, especially vesicular zinc in the brain of the HD transgenic N171-82Q mice. Furthermore, a reduction of ZnT3 expression was observed in these mice and 160Q BHK cells. ZnT3, an important protein located on the membrane of synaptic vesicles, affects synaptic function via various mechanisms in neurons [60]. Most important of all, ZnT3 is responsible for the movement of zinc from the cytoplasm to the synaptic vesicles. Vesicular zinc level is dependent on ZnT3 protein abundance. Zinc is eliminated from synaptic vesicles in the brain of ZnT3 knockout mice [61,62]. In the present study, we found that vesicular zinc detected by AMG staining is dramatically decreased in the striatum, hippocampus and cortex of the N171-82Q mice compared to agematched controls, suggesting that the reduced ZnT3 expression in HD greatly disturbs zinc homeostasis in synaptic vesicles.
It is well-known that brain zinc homeostasis is strictly controlled to guarantee physiological function under healthy conditions. Vesicular zinc serves as a signal factor in a subclass of glutamatergic neurons, which is linked to glutamate signaling and cognitive activity [63,64]. Specially, vesicular zinc can inhibit glutamate release. In mutant mice with lacking vesicular zinc, glutamate release inhibition induced by zinc is absent [65]. Many evidences support a key role of glutamate-mediated excitotoxic cell death in HD pathogenesis [66][67][68]. Thus, it appears likely that loss of vesicular zinc in the N171-82Q mice aggravates glutamate-mediated excitotoxic neuron death. On the other side, vesicular zinc signaling meets the requirement for cognitive and emotional behavior [36]. Movement impairment and cognitive de cits occur prior to neuron degeneration in HD patients [8,69]. Our results reveal that vesicular zinc level is lower in the hippocampus of the N171-82Q mice compared to WT mice. Consequently, the decrease of vesicular zinc might be responsible for the cognitive decline in HD.
Beside affecting vesicular zinc, the reduced ZnT3 expression may also impair synaptic structures and proteins. In ZnT3 knockout mice, there is a decrease in total dendritic spines per neuron. Similarly, synaptic plasticity-related proteins, such as presynaptic synaptosome-associated protein 25 (SNAP25) and postsynaptic PSD95, are markedly decreased in the absence of ZnT3 [36]. According to the previously-reported literatures, both SNAP25 and PSD95 are also defective in HD [69,70]. Of note is that the N171-82Q mice showed signi cant reductions in ZnT3 protein and mRNA levels at a relatively early stage of this disease (about 14 weeks). Subtle alterations in synaptic function can lead to the early symptoms of HD [71]. In consequence, we conclude that reduction of ZnT3 expression in the N171-82Q mice results in synaptic dysfunction.
However, how does mHtt cause the reduction of ZnT3 expression in HD? Subsequently, the mechanism about this issue was further illuminated in this study. One of distinctive characteristics of mHtt is to form aggregates or inclusions, which directly recruit synaptic proteins [72,73]. Here, we rstly examined whether mHtt aggregates could recruit ZnT3 protein. Co-immunoprecipitation results showed that ZnT3 protein was not detected in aggregates, suggesting that down-regulation of ZnT3 does not come from the recruit of mHtt. Additionally, mHtt affects gene expression via altering the activity of transcription factors or abnormally interacting with ones [74][75][76]. Two major transcriptional pathways, namely CRE and Sp1mediated transcription, have been extensively studies in HD [53]. Increasing ndings demonstrate that several genes containing Sp1-binding motifs in their promoters are down-regulated in HD [55,77,78]. In the present study, one important nding was that Sp1 transactivated ZnT3 gene by binding the GC boxes in the promoter. We also detected whether mHtt could affect Sp1 expression to reduce ZnT3 mRNA level.
Interestingly, the N171-82Q mice displayed an increasing Sp1 expression compared to WT mice, which was consistent with previous reports [79]. The result indicates that inhibition of ZnT3 expression does not result from decreased Sp1 protein.
More importantly, Sp1 is found to interact with Htt [57,[80][81][82]. mHtt binds to C-terminal region of Sp1, which inhibits the interaction of Sp1 with targeted gene promoters [81,82]. Here, we con rmed that mHtt inhibits the binding of Sp1 to the GC boxes in the ZnT3 gene promoter so as to reduce ZnT3 expression. Similar mechanism has been veri ed in previous studies. For example, Sp1 acts through its binding to the GC boxes in the nerve growth factor receptor (NGFR) promoter, followed by activating the expression of NGFR. mHtt inhibits the binding of Sp1 to the NGFR promoter to decrease its transcription [81]. As is mentioned above, the N171-82Q mice showed the defective of vesicular zinc at a relatively early stage of this disease. Given that Sp1 disruption also occurs early in human HD pathogenesis, even in postmortem tissues of pre-symptomatic grade [77,81]. We reasonably infer that mHtt blocks the binding of Sp1 to ZnT3 promoter to decrease the expression of ZnT3, thereby disrupting vesicular zinc homeostasis. In order to further con rm this mechanism, we investigated the in uence of overexpression of Sp1 on ZnT3 mRNA level in 160Q BHK cells. As expected, overexpression of Sp1 enhanced the transcriptional activity of ZnT3 gene to up-regulate its mRNA level in our experiments. Thus, overexpression of Sp1 greatly attenuates inhibition of the binding of Sp1 to ZnT3 gene promoter by mHtt. Obviously, increasing endogenous Sp1 protein in the N171-82Q mice might be a reactive result as reduction of binding between Sp1 and ZnT3 gene promoter. According to the above data, we present the mechanism of ZnT3 downregulation and its effect on vesicular zinc in HD in Fig. 12.

Conclusions
In conclusion, this work has identi ed signi cant reductions in vesicular zinc and its molecular mechanism in HD. ZnT3 expression is decreased in the brain at the early stage of the N171-82Q mice, indicating that altered neuronal zinc homestasis is an early event in HD pathogenesis. Sp1 serves as one Availability of data and material The data and materials are available from corresponding author on reasonable request.

Ethics approval
All procedures for animal use were approved by the institutional Animal Care and Committee of Tongji Medical College, Huazhong University of Science and Technology.

Consent for publication
All authors have approved of the contents of this manuscript and provided consent for publication. Supplementary Information Legends Total zinc level in N171-82Q (TG) mice brain. Total zinc decreases in cortex, striatum and hippocampus of the 20-week-old TG mice compared to age-matched wild type (WT) mice. n = 3. *, p < 0.05 compared to WT mice.

Figure 2
Free zinc level in TG mice brain. AMG stain is abundant in hippocampus and striatum of the 20-week-old WT mice, but poor in age-matched TG mice brain.   Immunohistochemistry determination of ZnT3 expression in TG mice brain. The 20-week-old TG mice display weak ZnT3-immunoreactivities in the brain compared to age-matched WT mice. A, B, C and D are hippocampus of WT mice, striatum of WT mice, hippocampus of TG mice, and striatum of TG mice, respectively. DG: dentate gyrus; ml: molecular layer. Scale bar: 500 μm. ZnT3 expression level in TG mice brain. ZnT3 protein expression and mRNA level decrease in cortex, striatum and hippocampus of TG mice aged from 14 to 20 weeks compared to age-matched WT mice. A1 and A2 are Western blots analysis of ZnT3 expression and the quantitative representation of ZnT3 band intensity normalized to GAPDH, respectively. B1 and B2 are RT-PCR analysis of ZnT3 mRNA level and the relative amounts of ZnT3 mRNA normalized to β-actin, respectively. Hip: hippocampus. n=4. *, p < 0.05 compared to WT mice. to GAPDH, respectively. B1 and B2 are RT-PCR analysis of ZnT3 mRNA level and the relative amounts of ZnT3 mRNA normalized to β-actin, respectively. n=4. *, p < 0.05 compared to 20Q BHK cells.    ChIP assays of Sp1 binding to ZnT3 promoter in 160Q BHK cells. mHtt inhibits the binding of Sp1 to ZnT3 promoter containing GC-1 box or GC-2 box. A and B display ChIP-qRT-PCR analysis of Sp1 binding to ZnT3 promoter containing GC-1 box or GC-2 box, respectively. The relative amounts of ZnT3 were normalized to input DNA. n=3. *, p < 0.05 compared to BHK cells transfected with 20Q Htt.

Figure 12
The mechanism of ZnT3 down-regulation and its effect on vesicular zinc in HD. B shows inhibition of the binding of transcription factor Sp1 to ZnT3 gene promoter by mHtt in HD. Down-regulated ZnT3 expression reduces the transport of zinc into synaptic vesicles in the axonal terminal. Zinc dyshomestasis contributes to synaptic dysfunction. Instead, A shows the relationships between Sp1, ZnT3 and vesicular zinc in neuron. Sp1 binds to ZnT3 gene promoter and up-regulates its expression.
Vesicular zinc is dependent on ZnT3 on the membrane of synaptic vesicles. SV, synaptic vesicle.