Kismet Positively Regulates Glutamate Receptor Localization and Synaptic Transmission at the Drosophila Neuromuscular Junction

The Drosophila neuromuscular junction (NMJ) is a glutamatergic synapse that is structurally and functionally similar to mammalian glutamatergic synapses. These synapses can, as a result of changes in activity, alter the strength of their connections via processes that require chromatin remodeling and changes in gene expression. The chromodomain helicase DNA binding (CHD) protein, Kismet (Kis), is expressed in both motor neuron nuclei and postsynaptic muscle nuclei of the Drosophila larvae. Here, we show that Kis is important for motor neuron synaptic morphology, the localization and clustering of postsynaptic glutamate receptors, larval motor behavior, and synaptic transmission. Our data suggest that Kis is part of the machinery that modulates the development and function of the NMJ. Kis is the homolog to human CHD7, which is mutated in CHARGE syndrome. Thus, our data suggest novel avenues of investigation for synaptic defects associated with CHARGE syndrome.


Introduction
Synapses in the nervous system must be stable yet labile to retain existing memories and help form new memories. Several synaptic signaling molecules are important for synapse formation and/or maintenance including Wnts [1,2], bone morphogenetic proteins (BMPs) [3], and neurotrophins [4] amongst others. These signaling molecules alter intracellular signaling and target cell transcription to convert transient cellular changes to stable, functional alterations [5] that ultimately maintain long-term synaptic connections and synaptic activity. Transcriptional activation is required for long-term potentiation (LTP) [6,7], which is characterized by enhanced synaptic transmission as a result of increased synaptic activity, and learning and memory [8]. Epigenetic modifications of chromatin structure are also required for LTP and memory [9].
Epigenetics have been defined as ''the structural adaptation of chromosomal regions so as to register, signal or perpetuate altered activity states'' [10]. This regulation of chromatin structure enables transcription factors to recognize essential cis-regulatory elements and control gene expression. Chromatin remodeling enzymes like the Chromodomain Helicase DNA Binding (CHD) family of ATPases regulate transcription by altering the structure of chromatin [11] and maintaining heritable states of transcription [12]. CHD proteins are well conserved and participate in neural development in both vertebrates [13][14][15][16][17] and invertebrates [18].
Kismet (Kis) is the Drosophila ortholog of mammalian Chromodomain Helicase DNA Binding Protein 7 (CHD7). Kis was shown to localize to a majority of active transcription sites on salivary gland polytene chromosomes, suggesting that Kis may regulate the transcription of multiple target genes during development [12]. However, the functional relevance of this binding for developmental processes is unclear. Studies in fly larval salivary glands suggest that Kis functions to enable elongation by RNA Polymerase II [12,19] and functions upstream of the histone H3K4 methyltransferases Trithorax and Ash1 [19,20]. Kis is required for proper locomotion, memory, axon development, and circadian rhythms [18,21]. Ubiquitous knockdown of Kis produces flies that are unable to fly and exhibit a prominent postural defect where they hold their wings apart and below their bodies [18]. Collectively, these reports suggest that CHD proteins like Kis may regulate synapse development. Although CHD7 has been implicated in neural crest formation [14] and neurogenesis [15,22], there are no reports of Kis or CHD7 function in synapse formation or maintenance.
Here we describe the function of Kis in synapse development using the glutamatergic Drosophila NMJ. Glutamatergic synapses are highly plastic and exhibit the capacity to alter protein composition and localization to maintain excitability [23][24][25][26]. Both presynaptic activity [27] and the number and subtype of postsynaptic receptor [28] can be changed to ensure proper synaptic strength. Our results suggest that Kis function is important for motor neuron morphology, synaptic transmission, and the localization of the cell adhesion molecule FasII and postsynaptic proteins including GluRs and Dlg.

Drosophila Stocks
Fly stocks were maintained at 25uC on standard fly food. The kis LM27 allele was generated by EMS mutagenesis as previously described [29]. kis k13416 and all Gal4 lines were obtained from Bloomington Drosophila Stock Center. UAS-kis RNAi.b was obtained from Vienna Drosophila RNAi Center (VDRC stock #46685). UAS-Dlg and UAS-FasII were gifts from Vivian Budnik [30,31]. Kismet-GFP was a gift from Alan Spradling [32].
Behavioral Testing 25-50 wandering third instar larvae from a non-isogenic background were used for behavioral testing as previously described [33]. Larvae were washed with phosphate solution to rid them of food particles and placed on a non-nutritive agar surface for acclimatization. To measure muscle contractions, an 8.5 cm diameter agar plate was used to provide the larva with a crawling surface. Peristaltic contractions were made using a Leica Mz 125 stereomicroscope for 30 sec per larva. Three trials per larva were performed as previously described [33][34][35].
The crawling behavior of each larva was recorded using Sony DCR-SR47 Handycam with Carl Zeiss optics. The video recording was processed using iMovies software (Apple Inc.) to convert the video file into frames for manual analysis using NIH Image J software. Frame-by-frame analysis of crawling distance was performed manually for each larva. Three trials were performed for each larva. The numerical output generated was used to calculate total distance crawled by the larva.
GluR cluster sizes were measured as previously described [38] from Z-projected confocal images using NIH ImageJ. Relative fluorescence intensities were quantified by obtaining the mean fluorescence intensity for each NMJ from Z-projected confocal images using Adobe Photoshop (v. CS2) and subtracting the mean fluorescence intensity obtained from an identical area of the muscle or non-NMJ area. Brp density was calculated by counting the number of fluorescent puncta from Z-projected confocal images and dividing by the total NMJ area as indicated by HRP labeling. Apposition of GluRIIC to Brp was measured by quantifying the number of GluRIIC clusters located within 1 mm of a Brp puncta using Z-projected images. Unapposed GluRIIC clusters included any cluster that was located a distance greater than 1 mm from Brp. The unopposed clusters number was divided by the total number of GluRIIC clusters to obtain a percentage.
The number of 6/7 NMJ boutons were counted using the ''cell counter'' plugin for NIH Image J software. Axonal branches were quantified by manual counting of any bifurcation that included at least two boutons.

Electrophysiology
Two-electrode voltage clamp recordings were obtained at room temperature from muscle 6 of segments A3 or A4. Third instar larval fillet dissections were performed on Sylgard-coated cover slips in Drosophila standard saline (135 mM NaCl, 5 mM KCl, 4 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM TES, 72 mM sucrose). Muscles were clamped at 260 mV using an Axoclamp 900A amplifier (Molecular Devices). Electrodes filled with 3M KCl that had resistances of 10-20 MV were used for intracellular recordings. A 1 Hz stimulus of 10 V was delivered to segmental nerves using an electrode filled with bath saline connected to a Grass S88 stimulator and SIU5 isolation unit (Grass Technologies). Electrophysiological recordings were digitized with a Digidata 1443 digitizer (Molecular Devices) and analyzed using PClamp software (v. 10.4). 180 s of minis per larva were used for analysis. Quantal content was calculated by dividing the eEJC area (nA*ms) by the mEJC area (nA*ms). Intracellular recordings of excitatory junction potentials were also conducted as previously described [39].

Microarray Studies and data analysis
Microarray analysis was performed by Seqwright (Houston, Tx). 10 total white pre-pupae (as staged by [40]) each from genotypes Da-Gal4/+ and Da-Gal4/UAS:kismet RNAi. b were used. RNA for each sample was isolated by Seqwright from frozen tissues using Qiagen RNeasy Mini kit. The RNA quality of each sample was checked on RNA Nano chip using an Agilent 2100 Bioanalyzer. All RNA samples displayed no degradation with two sharp ribosomal peaks. Affymetrix's GeneChip One-cycle target labeling kit was used for cDNA synthesis and in vitro transcription. Affymetrix GeneChip Drosophila Genome 2.0 array was used for this study and the data were processed by Affymetrix Software Expression Console using MAS5.0 Analysis Algorithms to produce. DAT,. CEL, and CHP files of each sample following Seqwright standard procedures.
To produce gene ontology tables, gene lists identified from the microarray analysis were analyzed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) (http://www.david.abcc.ncifcrf.gov) [41,42]. Genes whose expression differed significantly (p,0.05) between Da-Gal4 and Da-Gal4/UAS:kismet RNAi.b and showed a twofold or greater change in expression (up-or downregulated) were used for subsequent DAVID analysis.

Quantitative RT-PCR
For mutant analyses, total RNA was extracted from 8-12 third instar larvae using Trizol as previously described [43,44]. qRT-PCR was performed in a single step using the iScript One-Step RT-PCR Kit with Sybr Green (Bio-Rad). For RNAi analyses, total RNA was extracted from third instar larval body walls from which the CNS and ventral nerve cord had been removed. After Trizol extraction, the samples were treated with DNase and column purified. Reverse transcription was carried out using the High Capacity cDNA Reverse Transcription kit (Invitrogen). Taqman probes (Applied Biosystems) to GluRIIA, GluRIIB, GluRIIC and Act5c were used for the real-time reaction using the ''StepOne-Plus'' instrument (Applied Biosystems). Appropriate controls were maintained throughout this study. 100 ng of RNA was added to each reaction. For mutant analysis, DC(t) values were calculated by subtracting the C(t) value of the GluR-specific reaction from the C(t) value obtained for GAPDH. For RNAi analysis, Act5c was used as the reference gene and analysis was performed by calculating fold change in expression (RQ) from DDC(t) values. Each reaction was performed in triplicate and three biological replicates were performed and used for data analysis.

Statistical Analyses
Statistics were performed using Graph Pad Prism (v 5.01). Statistical comparisons included student's t-tests for analysis of kismet mutant phenotypes, and, for RNAi analysis, a one-way ANOVA with post-hoc Tukey or Games-Howell tests depending on the variation between data sets. Statistical significance in figures is represented by:* = p,0.05, ** = p,0.001, and *** = p,0.0001. Error bars represent the standard error of the mean (SEM). Summary statistics including means, SEM, number of animals, and additional information can be found in Table S1.

Kismet is localized to motor neuron and muscle nuclei
We sought to further explore the role Kis may play in synaptic development using the Drosophila NMJ as a model system. Kis has been shown to regulate RNA polymerase II-mediated transcription [12,19]. Therefore, we first examined the localization of Kis protein in third instar larvae using both immunolabeling for the long isoform of Kis (Kis-L; [12]) and a Kis-GFP protein trap stock [32]. We had previously shown that Kis is localized in the nuclei of multiple cells in the ventral nerve cord (VNC) of third instar larvae [18]. To determine first that Kis localizes to neurons, we labeled neurons in the third instar VNC by driving expression of a membrane-tagged RFP with elav-Gal4 and focusing on the midline of the VNC. We then analyzed the localization of Kis-GFP within this tissue. As previously reported [18], Kis localized to multiple nuclei within the VNC including the nuclei of multiple neurons ( Figure 1A). To verify that these neurons were motor neurons, we labeled motor neurons by driving n-syb-GFP with the motor neuron-specific D42-Gal4 driver. We observed that Kis localized to the nuclei of VNC motor neurons ( Figure S1A). Using both methods, we also observed Kis localized in multinucleated postsynaptic muscles ( Figure 1B and Figure S1B). We did not observe Kis localized outside of the nucleus in either analyses. These results are consistent with the role of Kis as a transcription regulator [19] and indicate that Kis may regulate both pre-and postsynaptic components of the NMJ via its nuclear actions.

Loss of Kismet function slightly alters synaptic morphology
Both the actin and microtubule cytoskeletons are important for the structure, stability, and the maintenance of synapses (for review see [45]). The microtubule cytoskeleton is essential for both synaptic function and muscle morphology as it forms and provides stability to structural components of neurons and synapses [46,47].
To determine whether kis influences the dynamics of the synaptic cytoskeleton like other chromatin remodeling enzymes [48][49][50], we labeled both the actin and microtubule cytoskeletons at the NMJ using two kis mutant alleles. kis k13416 is a weak adult-viable hypomorph due to the insertion of a transposable element in the 59 end of the gene [51] while kis LM27 is a strong protein null allele [29]. Because kis LM27 nulls are early larval lethal, we used the heteroallelic kis LM27 /kis k13416 combination, which is pupal lethal. Mutations in kis do not affect gross muscle morphology as indicated by phalloidin, which labels F-actin [52], or the size of ventral body wall muscles 6 (Figures 2A-B) and 7 (data not shown, w 1118 = 1793061173 mm 2 , n= 9; kis k13416 = 169806543.5 mm 2 , n = 9, p = 0.47; kis LM27 /kis k13416 = 171506682.8 mm 2 , n=8, p = 0.61). Similarly, there were no significant differences in the levels of acetylated tubulin at the synapse or in the muscle of kis mutants compared with controls ( Figures 2C-D). Post-translational acetylation of a-tubulin occurs at stable microtubules [53].
Given the role of transcriptional regulators such as AP-1 [54] and Zfh1 [55] in NMJ development and morphology, we next examined NMJ morphology in kis mutants and in animals after ubiquitous knockdown of kis using the UAS/Gal4 system [56]. Ubiquitous knockdown of Kis was achieved using the da-Gal4 driver to express a UAS:kismet RNAi.b transgene (da.kis RNAi.b ), which was previously shown to strongly reduce Kis protein levels (roughly 90% overall) and severely reduce adult motor function [18]. While we noticed a trend indicating the numbers of boutons per 6/7 NMJ were slightly increased in both kis mutants compared with controls ( Figure 2E), this increase was not statistically significant. However, ubiquitous knockdown of kis resulted in a significant increase in the total number of boutons in da.kis RNAi.b (data not shown, 94.454565.9043, p,0.001) when compared to outcrossed controls da-Gal4/+ (57.363664.3469) and UASkis RNAi.b /+ (70.727364.8450, p,0.001). We also observed a significant increase in the total number of branches in kis LM27 / kis k13416 larvae, but not kis k13416 larvae compared to w 1118 controls ( Figure 2E). Taken together, these data suggest that kis influences NMJ morphology but does not affect ventral body wall muscle size or muscle morphology.

Kismet selectively regulates the levels of the cell adhesion molecule FasII and the postsynaptic protein Dlg
Several synaptic proteins serve as markers of synaptic plasticity at the Drosophila NMJ such as Bruchpilot (Brp), Cysteine String Protein (CSP), Discs Large (Dlg), Fasciclin II (FasII), and Synaptotagmin (Syt). To determine if the loss of kis affects other synaptic proteins, we first examined the number of active zones, sites of neurotransmitter release, by immunolabeling for the T-bar associated protein Brp [57,58]. We did not observe any significant difference in the density of Brp puncta in either kis mutant compared to control larvae ( Figure S2A). Similarly, mutations in kis did not produce a significant change in relative fluorescence for the synaptic vesicle markers CSP and Syt ( Figure S2B-C).
Interestingly, we observed a significant increase in synaptic levels of both FasII and Dlg in kis mutants ( Figure S3). Dlg is a postsynaptic scaffolding protein that stabilizes GluRIIB-containing clusters [59] while FasII is an activity-dependent cell adhesion molecule [60]. kis mutant larvae exhibited a significant increase in the relative fluorescence intensity of Dlg compared to w 1118 controls ( Figure S3A). Similarly, FasII immunoreactivity was significantly increased in kis mutants ( Figure S3B) compared to w 1118 controls. Taken together, these data suggest that Kis may serve as a negative regulator of FasII and Dlg at the synapse but does not significantly affect multiple presynaptic proteins including the functional presynaptic markers Brp, CSP, and Syt.

Kismet positively regulates postsynaptic GluR localization in larval muscles
Recently our labs independently identified kis as necessary for GluR localization/expression via two separate approaches. First, through a microarray analysis of kis loss of function pupae (Tables S2-S3), we found that the expression level of the GluR subunits GluRIIB and GluRIIC were significantly reduced. Second, we also identified mutations in kis in a forward genetic screen for phenotypes that affect the localization of GluRIIA in third instar larval muscles (data not shown). To further explore the connection between kis function and GluR expression, we next examined ionotropic GluRs (iGluRs) at third instar larval NMJ in kis loss of function mutants and after ubiquitous knockdown of Kis. There are five iGluR subunits expressed in Drosophila muscles including GluRIIA, GluRIIB, GluRIIC, GluRIID, and GluRIIE. These subunits are most similar in structure and function to vertebrate AMPA and kainate receptors [37,61,62]. The five subunits form two distinct tetrameric GluR complexes, which contain either GluRIIA or GluRIIB, plus the remaining essential GluRIIC, -IID, and -IIE subunits [37,62].
Both kis loss of function mutants exhibited a significant decrease in GluRIIA cluster size compared to w 1118 controls ( Figure 3A). GluR cluster size has been shown to correlate with receptor function [38] and is used as an indicator of the number of postsynaptic receptors [59]. This corresponded to a 57.6% reduction in mean relative fluorescence intensity in kis LM27 / kis k13416 mutants compared to controls. Further, there was a small but non-significant reduction in GluRIIB cluster size in kis k13416 homozygous larvae and a significant reduction in GluRIIB cluster size in kis LM27 /kis k13416 mutant larvae compared to w 1118 controls ( Figure 3B). Although there was no significant change in GluRIIC cluster size in either kis mutants ( Figure 3C), there was a significant reduction in mean relative GluRIIC fluorescence levels in both kis mutants compared to w 1118 controls ( Figure 3C). The significant reduction in GluRIIC relative fluorescence without a corresponding reduction in GluRIIC cluster size indicates there is likely a loss of GluRIIC puncta in kis mutants. Similar to our mutant analyses, knockdown of kis in all tissues using the da-Gal4 driver also resulted in significant decreases in GluRIIA, GluRIIB, and GluRIIC cluster sizes ( Figure S4). Collectively, these data suggest that Kis positively influences synaptic levels of iGluRs.
Given the role of Kis in transcription elongation [12,19], we next analyzed the relative levels of iGluR mRNAs in both kis mutants and RNAi knockdowns. We did not observe a significant decrease in GluRIIA, GluRIIB, or GluRIIC transcript levels in kis k13416 homozygous mutant larvae, kis LM27 /kis k13416 larvae, or ubiquitous Kis knockdowns when compared to controls (data not shown). Taken together, these data suggest that Kis may be involved in the localization and clustering of postsynaptic iGluR subunits by influencing other synaptic transcripts but may not be involved in the transcriptional regulation of iGluR subunits. Thus, mutations in kis may secondarily affect iGluRs as a consequence of other synaptic perturbations.
Because we observed significant effects on both Dlg and FasII expression levels in kis mutants ( Figure S3), we sought to determine whether the increased synaptic levels of Dlg or FasII could explain the loss of postsynaptic iGluRs observed in kis mutants. We overexpressed Dlg or FasII in all tissues using the actin5c-Gal4 driver. Ubiquitous overexpression of Dlg or FasII, however, did not significantly affect GluRIIA cluster sizes ( Figure  S5) suggesting that the relative increase in either Dlg or FasII likely did not contribute to the loss of iGluRs observed in kis mutants.

Kismet positively regulates neurotransmission and quantal content
The reduction in synaptic iGluRs, increase in synaptic levels of Dlg and FasII, and change in morphology of the presynaptic motor neuron in kis mutants collectively suggest that Kis may regulate synaptic development without significantly affecting iGluR mRNA levels. We therefore examined neurotransmission using two-electrode voltage clamp and found significant reductions in both evoked excitatory junctional current (eEJC) amplitudes, miniature excitatory junctional current (mEJC) amplitudes, an indicator of spontaneous neurotransmission, and quantal content in kis mutants compared with controls ( Figure 4). There were no significant differences in mEJC frequency (data not shown, w 1118 = 3.43860.373 Hz, n = 13, kis k13416 = 3.42360.827 Hz, n = 12, p = 0.986, kis LM27 /kis k13416 = 3.01460.450 Hz, n = 12, p = 0.473). Similarly, knockdown of Kis in all tissues resulted in a significant reduction in the amplitude of excitatory junction potentials (EJPs, data not shown). These data suggest that kis function is important for neurotransmission at the Drosophila larval NMJ.

Kismet affects the alignment of postsynaptic receptors with presynaptic active zones
Proper synaptic function requires the precise localization of both pre-and postsynaptic components such that each postsynaptic iGluR cluster is directly apposed to a presynaptic active zone [63]. Because of the defects we observed in iGluR clustering ( Figure 3C) and synaptic transmission, we quantified the number of iGluR clusters (visualized by GluRIIC) that were unopposed to active zones as indicated by Brp puncta. Mutations in kis do not affect the density of Brp puncta ( Figure S2A) but may affect the number of postsynaptic GluRIIC clusters ( Figure 3C). There was a significant increase in the number of GluRIIC clusters unapposed to Brp puncta in kis mutant larvae compared to control larvae ( Figure 5) suggesting that kis mutants contain fewer functional synapses. Collectively, these data indicate that Kis, by affecting transcription of synaptic target genes, may function to regulate the relative distribution and localization of specific synaptic proteins to facilitate proper synaptic development and function.

Kismet exhibits tissue-specific effects on iGluR clustering and locomotion
Kis is strongly expressed in both presynaptic motor neuron nuclei as well as in muscle nuclei (Figure 1). To determine the tissue-specific effects of kis loss of function on NMJ development, we used two drivers including Dcr2;;elav-Gal4 to knockdown Kis in neurons and Dcr2;;24B-Gal4 to knockdown Kis in postsynaptic muscles. To validate the effects we observed with Kis knockdown using da-Gal4, we utilized a second Gal4 line (actin5c-Gal4) as a positive control for ubiquitous knockdown. We focused our analysis on GluRIIB since it exhibited the most severe reduction at kis LM27 /kis k13416 mutant NMJs ( Figure 3B). Consistent with our previous results using da-Gal4, ubiquitous Kis knockdown with actin5c-Gal4 resulted in a significant reduction in GluRIIB cluster size ( Figure 6B). Knockdown of Kis in neurons but not muscles produced both a significant decrease in GluRIIB cluster size   Figure 6B) and an increase in motor neuron branches (Figure 6C). In contrast, knockdown of Kis in muscles alone resulted in a significant increase in synaptic boutons ( Figure 6C). Taken together, our data suggests that Kis may regulate synaptic gene products that function in the motor neuron to help properly localize postsynaptic iGluRs. Kis gene products function in both motor neurons and muscle to regulate NMJ morphology.
To extend our findings into functional behavior and to further analyze the tissue-specific effects of Kis, we next analyzed the effects of ubiquitous and tissue-specific knockdown of Kis on larval locomotive behavior by examining the frequency of muscle contractions and total crawling distance. Larval locomotion is the result of central pattern generators producing bursts of neuronal activity [64], which lead to contractions of both dorsal and ventral body wall muscles [65]. Thus, larval locomotion relies, at least in part, on proper glutamatergic neurotransmission [66]. There was a significant reduction in both the frequency of muscle contractions ( Figure 7A) and distance crawled ( Figure 7B) by larvae after ubiquitous knockdown of Kis compared to controls. Presynaptic knockdown of Kis, but not muscle-specific knockdown of Kis, also led to a significant reduction in the frequency of muscle contractions produced and total distance crawled by these larvae compared to controls (Figure 7). Taken together, these data suggest that Kis influences both the function of and development of the glutamatergic NMJ.

Discussion
Our results are the first to demonstrate a role for Kis in synaptic development and function of the Drosophila NMJ. The functional role is likely related to Kis influencing transcription of target genes involved in iGluR localization and apposition to presynaptic release sites. Kis mRNA is expressed ubiquitiously early in embryogenesis [67] and Kis protein has been detected in larval salivary glands [12], central nervous system [18], retina [29], and in the adult wing [68]. Our results indicate that Kis is localized to ventral motor neuron and muscle nuclei (Figure 1, Figure S1). Although we did not find that Kis directly regulated iGluR transcript levels, the proposed role of Kis in transcription elongation [19] implies that Kis-mediated transcriptional regulation of target genes relevant to synaptic function could collectively regulate synapse development. Indeed, our microarray results from kis loss of function pupae identified many potential target genes that encode synaptic proteins and transcription factors that we have not tested in this particular study (Tables S2-S5).

Kismet regulates synapse function and morphology
Kis contains an ATPase domain, two chromodomains, and a BRK domain [67,69] suggesting that it promotes chromatin remodeling. Chromatin remodeling enzymes including those that regulate DNA methylation, histone modifications, and nucleosome positioning may positively or negatively influence transcription in response to developmental cues and neural activity [70]. In mammals, chromatin modifying enzymes including nBAF complexes containing BAF53b [71] and histone deacetylase 2 (HDAC2) [72] regulate hippocampal dendritic density and synaptic activity. We find that reduced kis levels result in an overgrowth of the presynaptic motor neuron ( Figure 2) and a significant reduction in the amplitudes of both evoked and spontaneous (i.e. mEJCs) synaptic currents and quantal content (Figure 4). The reduction in mini amplitudes may reflect the reduction in synaptic iGluRs ( Figure 3) we observe in kis mutants. The endocytosis mutants brain tumor [73] and dap160 [74,75] exhibit increases in the number of synaptic boutons and decreases in evoked EJPs. Although mEJP amplitudes are significantly increased in these mutants [73,75], mutations that affect endocytosis differentially influence spontaneous activity. The Gactin sequestering protein, twinfilin, positively regulates endocytosis and localization of GluRIIA but does not influence spontaneous neurotransmission [76]. Our collective data suggest that mutations in kis lead to a defect at the level of vesicular release, vesicular loading, or synaptic vesicle endocytosis, which over time may lead to reduced receptor clustering, larval motor behavior, and other synaptic changes.

Kismet regulates the localization of synaptic proteins including Dlg, FasII, and iGluRs
Synaptic transmission is also affected by mutations that disrupt the apposition of GluRs to Brp. Microtubule star (mts), which encodes the catalytic subunit of protein phosphatase 2A [77], and rab3 mutants [78] both exhibit a significant increase in the percentage GluRIIC clusters unopposed to Brp similar to kis mutants ( Figure 5). mts positively regulates evoked neurotransmission and mEJP amplitudes [77] similar to Kis. Conversely, rab3 is important for short-term facilitation but does not influence mEJP amplitudes or quantal content [78]. These data indicate that iGluR apposition to Brp is affected by disparate synaptic mechanisms and may be a general indicator of altered synaptic signaling. Since we observe a significant decrease in GluRIIC fluorescence without a significant decrease in GluRIIC cluster size in the kis mutants we analyzed, we hypothesize that 1) there are fewer GluRIIC clusters (decreased fluorescence) and 2) the GluRIIC clusters that are present are not properly aligned with presynaptic active zones.
Changes in synaptic function may lead to compensatory changes in the synapse over time. We observe an increase in the relative fluorescence of Dlg and FasII in kis mutants ( Figure S3). An increase in the levels of FasII over time would promote the formation of additional synapses through an increase in the total number of boutons [30], which is what we observe in kis mutants. While loss of kis in third instar larvae results in a reduction in iGluR cluster sizes (Figure 3, Figure S4), we did not observe a significant effect on GluRIIA, GluRIIB, and GluRIIC transcripts in kis mutants or as a result of ubiquitous Kis knockdown.
Although this result appears contradictory to our microarray results (Table S3), iGluR transcript levels were measured in third instar mutant larvae while Kis knockdown pupae were used for the microarray. Given that Kis is involved in transcription elongation and epigenetic control of transcription [12,19], our data suggests that the loss of GluRs from the synapse may be secondary to other synaptic changes instead of a decrease in Kis-mediated iGluR transcription. Indeed, our tissue-specific analyses suggest that the decreased iGluR levels in postsynaptic muscles may be due to defective signaling from presynaptic motor neurons ( Figure 6).

Synaptic mechanisms underlying CHARGE Syndrome
kis is the Drosophila homolog of CHD7 in humans [18]. Haploinsufficiency of CHD7 is associated with CHARGE syndrome, an autosomal dominant disorder characterized by multiple developmental symptoms that include hypotonia, limb abnormalities, and cranial nerve defects [79][80][81][82][83]. Individuals with CHARGE syndrome can display delayed or impaired motor coordination and adaptive motor skills, as well as musculoskeletal anomalies [79,84].
Mechanisms underlying the symptoms associated with CHARGE syndrome are currently not clear. We have previously shown that kis loss of function leads to defective adult motor behavior. Ubiquitous knockdown of kis produces flies that are unable to fly and exhibit a prominent postural defect where they hold their wings apart and below their bodies [18]. This postural defect is reminiscent of defects associated with muscles and neuromuscular synaptic transmission and is consistent with hypotonia related postural problems often observed in CHARGE individuals [84]. Our current data suggest that defective synaptic function translates to defective motor behavior in animals with decreased kis function (Figure 7). Loss of kis function leads to defects in motor coordination as assayed through larval crawling. The aberrations observed in synaptic transmission and synaptic development in kis mutant larvae may also be present in mammals with decreased CHD7 function, including CHARGE individuals.
In summary, we describe, for the first time, the function of the CHD7 homolog, Kis, in synaptic development. Our results suggest that Kis function is important for proper synaptic transmission, localization of synaptic proteins, motor neuron morphology, and larval motor behavior. We suggest that the aberrant synaptic activity in kis mutants leads to decreased levels of and altered localization of postsynaptic GluRs over time, further affecting motor behavior and NMJ function.