Mutations in Wnt2 Alter Presynaptic Motor Neuron Morphology and Presynaptic Protein Localization at the Drosophila Neuromuscular Junction

Wnt proteins are secreted proteins involved in a number of developmental processes including neural development and synaptogenesis. We sought to determine the role of the Drosophila Wnt7b ortholog, Wnt2, using the neuromuscular junction (NMJ). Mutations in wnt2 produce an increase in the number of presynaptic branches and a reduction in immunolabeling of the active zone proteins, Bruchpilot and synaptobrevin, at the NMJ. There was no change, however, in immunolabeling for the presynaptic proteins cysteine-string protein (CSP) and synaptotagmin, nor the postsynaptic proteins GluRIIA and DLG at the NMJ. Consistent with the presynaptic defects, wnt2 mutants exhibit approximately a 50% reduction in evoked excitatory junctional currents. Rescue, RNAi, and tissue-specific qRT-PCR experiments indicate that Wnt2 is expressed by the postsynaptic cell where it may serve as a retrograde signal that regulates presynaptic morphology and the localization of presynaptic proteins.


Introduction
Synapses are specialized structures that allow neurons to communicate with one another. This communication is achieved by converting the electrical signal of the axon into a chemical signal at structures called active zones (AZs). AZs are made up of a dense protein matrix that collectively participates in synaptic vesicle exocytosis (for reviews see [1,2]). Important components of this matrix include Ca 2+ channels [3], proteins involved in vesicle fusion including SNAP-25, Synaptobrevin, and Syntaxin [4,5], scaffolding proteins including Bassoon [6], Piccolo [7], CAST/ ELKS/Bruchpilot [8,9,10], cell adhesion molecules including cadherins [11] and neuroligins [12,13]. The assembly of AZs is thought to occur quickly after axonal target recognition and contact with the postsynaptic cell. AZ proteins are packaged and transported in vesicles and delivered to synaptic locations [14,15,16]. The mechanisms by which AZs are properly localized and maintained are largely unknown but likely involve transsynaptic signaling to coordinate the development of the presynaptic neuron with the postsynaptic cell.
The Wnt family of secreted glycoproteins, well characterized for their roles in several developmental processes including cell fate specification, axis patterning, and neural development (for reviews see [17][18][19]), also regulate synapse development (for reviews see [20,21]). For example, mouse Wnt7a regulates the organization of presynaptic microtubules and clustering of the presynaptic proteins synapsin I [22] and synaptophysin [22,23]. Similarly, the mouse Wnt3 protein increases axon branching, growth cone size, and synapsin I clustering in presynaptic sensory neurons [24]. Wnts 3, 5a, 7a, and 7b are expressed in the hippocampus along with Frizzled receptors where these proteins regulate synapse formation [25]. In Drosophila, Wingless (Wg) governs the development of both pre-and postsynaptic structures [26][27][28]. Finally, Drosophila Wnt5 positively regulates neuromuscular junction (NMJ) growth and the synaptic localization of active zone proteins [29].
The Drosophila larval NMJ is a well-established model system for dissecting the molecular basis of synapse formation, growth, and remodeling. These synapses are similar to mammalian central synapses in that they are glutamatergic and remodel in response to activity [41,42]. Using this system, we show in vivo for the first time that the Drosophila wnt7b ortholog, wnt2, participates in synapse development. Mutations in wnt2 result in increased branching of NMJ axons, loss of synaptobrevin, and a 50% reduction in evoked release. Rescue and RNAi data indicates that wnt2 may function in the postsynaptic muscle. We postulate that Wnt2 may serve as a postsynaptic signal that regulates the development of the presynaptic neuron.

wnt2 Negatively Regulates Synaptic Growth and Alters the Synaptic Distribution of Brp
We previously found that Wnt5 signals via Drl to promote synaptic growth at the Drosophila NMJ [29]. Further, Drosophila Wg affects synapses both presynaptically by regulating growth of the NMJ [27] and postsynaptically by acting on Fz2 [26]. To investigate whether wnt2 is involved in synapse development, we examined the 6/7 NMJ of wnt2 O mutant 3 rd instar larvae. The wnt2 O mutation introduces a stop codon at residue Q40 likely producing a null mutant. These mutants are viable but male sterile as previously described [43] and exhibit defects in the direct flight muscles due to the requirement of wnt2 in muscle patterning during pupation [44]. Using qRT-PCR, we observed that wnt2 RNA is expressed in the larval ventral body wall muscles (DC(t) = 9.6360.24 cycles, n = 6). Careful examination of the wnt2 O mutant larvae showed no visible defects in either the patterning or size of the muscles ( Fig. S1: WT = 2618061987 mm 2 , n = 6, wnt2 O = 2713061473 mm 2 , n = 8, p = 0.70). It's possible that, like other Wnts, wnt2 may function in the larval ventral body wall muscles to regulate synapse development.
To examine NMJ morphology, we visualized presynaptic motor neurons with the anti-HRP antibody to label neuronal membranes.
We examined the synaptic localization of the presynaptic protein, Bruchpilot/nc82 (Brp) by immunolabeling. Brp is a coiled-coil domain protein that clusters presynaptic Ca 2+ channels and promotes AZ assembly [10,45]. The total number of Brp puncta per NMJ was not significantly different in wnt2 mutants compared with controls (WT = 178.4626.32 Brp puncta per NMJ, n = 7; wnt2 O = 183.8611.91 Brp puncta per NMJ, n = 8, p = 0.85). Since the amount of Brp at active zones is variable [46,47], we quantified the mean fluorescence of Brp at the NMJ and found there was a 32% reduction in Brp immunofluorscence in wnt2 mutants (WT = 1.0060.05 a.u., n = 7; wnt2 O = 0.6860.09 a.u., n = 7, p = 0.008). We also examined the density of Brp puncta, which is calculated by dividing the total Brp puncta by the area of the presynaptic motor neuron, and found that the density of Brp is slightly but significantly reduced in wnt2 mutants compared with controls ( Fig. 1B, C; WT = 0.5160.02 puncta/mm 2 , n = 9; wnt2 O = 0.4260.03 puncta/mm 2 , n = 8, p = 0.03) Our data suggests that, although the total number of Brp puncta are similar in wnt2 mutants and controls, the Brp puncta are more dispersed at the NMJ and may contain less Brp.
To verify that wnt2 is responsible for the observed phenotype, we examined the NMJ of wnt2 O /Df(2R)BSC29 and wnt2 I [43]  To determine whether the reduction in the density of Brp affected synaptic function in wnt2 mutant animals, we recorded ionic currents from postsynaptic muscle. Muscle 6 was voltage clamped at 260 mV and the presynaptic segmental nerve was stimulated (1 Hz, 5 V) to induce synaptic activity. The amplitude of evoked excitatory junctional currents (EJCs) was reduced approximately 50% in wnt2 mutant animals (

The Morphological Change in wnt2 Mutants Occurs Early in Development
Growth of the NMJ during larval development (24-120 h AEL) involves the addition of new synaptic boutons, branches, and AZs per bouton [53][54][55]. The dramatic increase in presynaptic growth attempts to accommodate the rapidly growing larval muscles and requires transsynaptic signaling between cells (for review see [53]). To ascertain whether the increase in branch number and reduction in AZs resulted from a failure to coordinate growth of the presynaptic motor neuron and postsynaptic muscle, we examined wnt2 mutant NMJs during both 1 st and 2 nd instar larval stages (24 and 48 h AEL, respectively).  Our attempts to make an antibody that specifically recognized the Wnt2 protein were unsuccessful. Thus, we used genetic techniques to delineate the cell type expressing the wnt2 gene. First, we performed cell-type specific gene rescue experiments by expressing the UAS-wnt2 transgene specifically in neurons (using the elav-Gal4 driver), muscle (using the 24B-Gal4 driver), or glial cells (using the repo-Gal4 driver) in the wnt2 O mutant background. Expression of wnt2 in muscle rescued the number of branches to near control levels. Conversely, expression of wnt2 in neurons of wnt2 O mutants produced a further increase in branch number while expression in glia significantly reduced branch numbers (Fig. 4 A-B; control = 5.8860.46 branches, n = 16; elav.wnt2 = 8.6761.11 branches, n = 9, p = 0.01; 24B.wnt2 = 6.2560.86 branches, n = 12, p = 0.69; repo.wnt2 = 4.3060.67 branches, n = 10, p = 0.047). Expression of wnt2 in all three cell types rescued Brp density (control = 0.5160.02 puncta/mm 2 , n = 9; elav.wnt2 = 0.5360.03puncta/mm 2 , n = 9, p = 0.52; 24B.wnt2 = 0.4960.04 puncta/mm 2 , n = 10, p = 0.82; repo.wnt2 = 0.4860.04 puncta/mm 2 , n=8, p = 0.61) suggesting expression of wnt2 in any NMJ cell type is sufficient to rescue the localization of Brp.
Since Wnt2 is a secreted protein, genetic rescue could have occurred by expression from extraneous cell types. To verify and further clarify the above results, we knocked down wnt2 function in neurons or muscle. Expression of a UAS-wnt2 RNAi construct within neurons (using the Dcr2;;elav driver) produced no change in either the number of branches or Brp density compared with controls (

Wnt2 does not Likely Exert its Presynaptic Effects via Drl or Fz Receptors
Previous experiments have demonstrated that Wingless utilizes postsynaptic DFz2 receptors to influence Drosophila NMJ morphology [26,56,57]. We sought to determine the receptor through which Wnt2 mediated its synaptic effects by examining mutant NMJs of fz and drl mutants ( Figure 6). Previous data indicates Wnt2 binds to Fz, Fz2, and Fz3 with similar affinities [58]. If Wnt2 signals via a single receptor or a combination of receptors, then mutants for those receptors should phenocopy wnt2 mutants. fz3 G10 null mutant animals [59] most closely resembled the wnt2 O NMJ phenotype in terms of morphology ( The increase in NMJ branch numbers in both wnt2 and fz3 mutants raised the possibility that the two genes may function in the same signaling pathway to regulate synaptic growth. To test this possibility, we constructed double mutants bearing simultaneous wnt2 and fz3 mutations. The NMJ morphology of wnt2 O ; fz3 G10 double mutants was statistically different than WT and of each of the single mutants (Fig. 6, wnt2 O = 7.4060.82 branches, n = 10; fz3 G10 = 7.6760.41, n = 9, p = 0.004; wnt2 O ; fz3 G10 = 9.3660.59, n = 11, p = 0.04) suggesting that wnt2 and fz3 may each regulate synaptic growth independently of the other. Collectively, our data suggest that Wnt2 is expressed by the postsynaptic muscle and is involved in presynaptic protein localization.

Discussion
Synapse development is a complex process that requires preand postsynaptic cells to maintain constant communication with one another via transsynaptic signaling. Molecules with well established roles in this process include cell adhesion molecules [60], Ephrin ligands and Eph receptors [61], and the classical cadherins [62]. We provide evidence that Wnt2 may act as a signaling molecule that is expressed by the postsynaptic muscle where it acts on the presynaptic cell to directly or indirectly regulate size of the presynaptic motor neuron and promote protein localization.

Wnt2 Regulates NMJ Development and Localization of Presynaptic Proteins
We present several pieces of evidence to support our conclusion that Wnt2 regulates development of the NMJ. wnt2 mutations produce overgrown NMJs with an increased number of branches (Fig. 1). The significant increase in NMJ branches is present early in development as both 1 st and 2 nd instar mutant larvae also exhibit an overgrowth (Fig. 3). This could indicate that wnt2 is required shortly after synapse formation to regulate NMJ growth. Although the NMJ is enlarged in the wnt2 mutant, the number of Brp puncta remained similar to controls. The level of Brp immunfluorescence is reduced, however, suggesting that the amount of Brp protein per punctum is decreased. Since Brp is localized to active zones where it promotes Ca 2+ channel clustering [63], reduced staining of Brp puncta may indicate that functioning of the active zones are compromised. A recent paper however, reported that the majority of active zones in Drosophila rab3 mutants do not contain Brp [64].
Electrophysiological recordings from muscle 6 of wnt2 mutants showed that the amplitudes of evoked events were significantly reduced without a reduction in the frequency or amplitude of spontaneous events (Fig. 1). This intriguing finding led us to carefully examine the concentrations of presynaptic proteins including Syt, Syb, CSP, Brp. The levels of Syb and Brp were significantly reduced in the wnt2 mutant as indicated by immunocytochemistry (Fig. 2). Syb is a synaptic vesicle associated protein that assembles with syntaxin and SNAP-25 to form the SNARE complex, which renders vesicles competent for fusion (for reviews see [65,66]). The electrophysiological phenotype we observed in wnt2 mutants is consistent with both the syb and brp mutant phenotypes. Syb is required for evoked but not spontaneous transmission in Drosophila [67] and knockdown of brp in neurons reduces evoked responses while preserving spontaneous transmission [10]. Thus, our finding that the total number of Brp puncta in wnt2 mutants is unchanged coupled with the significant reduction in evoked responses, suggests that there may be a reduction in the number of functional active zones in the wnt2 mutant. Indeed, we observed that the immunolabeling of Brp puncta is reduced, suggesting that the amount of Brp protein per puncta is decreased. [51] The reduced labeling of Brp and Syb in the presynaptic motor neuron of wnt2 mutants is not likely due to changes in transcriptional mechanisms. Messenger RNA levels of both brp and syb are similar in mutant and control animals (Fig. 2). It is possible that the observed changes in Brp and Syb are due to mislocalization of mRNA. Another possibility is that the loss of wnt2 leads to mislocalization of presynaptic proteins. Rat Wnt7a, which is 77.1% similar in amino acid sequence to Wnt7b [68], when applied to hippocampal cultures, induces clustering of Syt, SV2, and increases the number of clusters containing synapto- physin [23]. Both Wnt7a and Wnt7b induce clustering of synapsin I in mouse cerebellar granule cell cultures. Treatment of culture medium with Wnt7b increased Bassoon clustering but did not increase total protein levels as indicated by Western Blots [69].

Wnt2 is Expressed by Postsynaptic Muscle
Wnts are secreted glycoproteins. An important aspect of understanding the function of Wnt2 is to determine where at the synapse it functions to regulate presynaptic motor neuron morphology and localization of proteins. Our cell-type specific cDNA expression in wnt2 mutants showed that Wnt2 may function in either presynaptic motor neurons or postsynaptic muscle. Expression in either motor neurons or muscle restored presynaptic Brp density. Expression in postsynaptic muscle also restored NMJ morphology while expression in presynaptic motor neurons caused a further increase in the number of NMJ branches (Fig. 4). Knockdown of wnt2 in muscle produced a phenotype similar to that of null mutant (Fig. 5). Our results collectively suggest that Wnt2 may be expressed by the postsynaptic muscle where it acts as a retrograde signal that negatively regulates NMJ growth and promotes the localization of presynaptic proteins. Based on our data, we cannot conclude whether wnt2 directly or indirectly regulates these synaptic characteristics.
A number of other molecules have been implicated in retrograde synaptic signaling including Ankyrin [70], nitric oxide [71], SAP97 [72], Synaptotagmin 4 [73], and secreted proteins such as Glass Bottom Boat [74], fibroblast growth factors [75], and Wnts [69,[76][77][78]. Mouse Wnt3 is secreted from motor neurons where it increases the size of growth cones and branching of incoming sensory neurons [24]. Similarly, mouse Wnt7a is expressed by cerebellar granule cells and acts on presynaptic mossy fibers to remodel axons and growth cones [22]. The receptor(s) that mediate the above effects are, as yet, unidentified but Wnt ligand binding to its receptor induces cytoskeletal changes [79,80].
We sought to determine the receptor through which Wnt2 signaled by examining mutants for drl, fz, fz2, and fz3. None of the mutants exhibited a reduction in the density of Brp. Mutations in fz3, however, led to a significant increase in NMJ branches similar to that of the wnt2 mutant. This raised the possibility that Wnt2 was signaling via Fz3 to negatively regulate NMJ growth. wnt2 O ; fz3 G10 double mutants, however, exhibited a significant increase in NMJ branches greater than that of the single mutants ( Fig. 6) suggesting wnt2 and fz3 act independently of one another to regulate synaptic growth. Wnt2 may signal via the Wnt receptors Fz4 or Smo but binding assays indicate there is no detectible binding between Wnt2 and these receptors [58]. It is also possible that we did not detect a phenotype in Frizzled mutants due to functional redundancy of these receptors. Future work will be required to uncover the receptor that mediates Wnt2 signaling.

Fly Stocks
All animals were raised at 25uC in standard fly vials with corn meal molasses medium. Wnt2 fly stocks along with the all Gal4 lines were obtained from Bloomington Drosophila Stock Center. fz D21 , fz2 C1 , fz3 G10 , and drl 2 fly stocks were generous gifts from the labs of Roel Nusse, Gary Struhl, Kaoru Saigo, and John Thomas, respectively. The UAS-wnt2 RNAi line was provided by the Vienna Drosophila RNAi Center (v38079).

Antibodies and Immunocytochemistry
For staining and microscopy, animals were dissected and fixed for 30-60 min in either Bouin's fixative (when GluRIIA or nc82 antibodies were used), or 4% paraformaldehyde (for all other immunolabeling). First and second instar larvae were dissected and fillet preparations were glued down using Sylgard-coated coverslips. Third instar larvae were dissected and fillet preparations were pinned down in Sylgard lined Petri dishes. All dissections were done in Drosophila standard saline (135 mM NaCl, 5 mM KCl, 4 mM MgCl, 1.8 mM CaCl, 5 mM TES, 72 mM sucrose) with 2 mM glutamate to preserve neuronal morphology [81] at RT.

Electrophysiology
All electrophysiology was performed on ventral body wall muscle 6. Larval recordings were performed on third instar larvae 110-120 hr AEL. Muscle 6 was voltage-clamped at 260 mV. Standard two-electrode voltage clamp techniques were used, as previously described [82]. Data were acquired and analyzed using an Axopatch amplifier and pClamp9 (Axon Instruments, Union City, CA). All dissections and recordings were done in standard Drosophila saline at 19C.

qRT-PCR
wnt2 O and control animals were homogenized and RNA was extracted using TRIreagent (Sigma, St. Loius, MO). RNA was obtained from 1 st instar (24-26 h after egg laying (AEL)), 2 nd instar (48-50 h AEL), and 3 rd instar larvae (110-120 h AEL). Reverse transcription of RNA was performed using Qiagen's Quantifast Sybr Green RT-PCR kit (Valencia, CA) with mRNA specific primers for actin (forward primer: GCACCACACCTTCTA-CAATGAGC, reverse primer: TACAGCGAGAGCACAGCC-TGGATG), Brp (forward primer: GCAAGAGGATTAAACGA-ACGAG, reverse primer: TAGCGGGTTCTTGGATAGTC), Syb (forward primer: GCACATTGTCAAGCAAATTCAC, reverse primer: TGTTGTTCCTGATTTGATGGTC), and Wnt2 (forward primer: ATTGTGGAACTGTGGAACTG, reverse primer: GCTGGACACTAATCTTATTTCC). Primers were designed against exon-intron borders and no primer sets yielded products when DNA was used as a template. qRT-PCR reactions were run using a Stratagene MX3000P. To obtain DC(t) values, the cycle threshold (C(t)) for Brp, Syb, and Wnt2 along with an actin control were measured for each sample. The difference between the mRNA-specific primer and the actin C(t) were calculated to determine the DC(t) value. DC(t) values were normalized to WT by dividing each DC(t) value by the mean DC(t) value for WT to yield the mRNA levels relative to WT. There was no significant differences in the actin C(t) values between WT and wnt2 mutants. For the qRT-PCR experiments examining wnt2 expression in nervous system and ventral body wall muscles, nervous systems and ventral body wall muscles were extracted from third instar larvae. The brain, ventral nerve cord, and intersegmental nerves were first extracted followed by the ventral body wall muscles. To minimize contamination of other tissue types in the ventral body wall muscles, only ventral body wall muscles between A3-A5 were used for RNA extraction.

Data Acquisition and Statistics
The total number of boutons and branches were acquired from 6/7 NMJs of hemisegments A3 or A4 of all animals. Branches were defined as an extension of the presynaptic motor neuron that included more than one bouton. The density of Brp labeling was quantified by counting the total number of Brp puncta in a projected Z-image and dividing by the total NMJ area as indicated by Brp labeling using ImageJ (NIH) software. We quantified immunoreactivity for all other synaptic proteins by measuring the mean fluorescence intensity of the NMJ using Adobe Photoshop software and subtracting the mean non-NMJ background over an identical area of the neighboring muscle membrane. For DLG and muscle acetylated tubulin, the average background from a nonsynaptic, non-muscle area was used.
Statistics were performed using GraphPad Prism (v. 4.01., 5.01). All statistical comparisons were made using unpaired students ttests. Unless otherwise noted, control animals used were WT-Berlin, elav-Gal4,24B-Gal4, wnt2 O ;elav/+, wnt2 O ;24B/+, and wnt2 O ; UAS-wnt2. There was no significant difference in bouton or branch numbers or Brp puncta density between genotypes. Therefore, the data were combined into one control group where appropriate. Statistical significance in figures is represented as follows: * = p,0.05, ** = p,0.01, and *** = p,0.001. All error bars represent S.E.M. Figure S1 Muscle size in wnt2 mutants is similar to that of controls. A: Representative confocal micrographs show the 6/7 NMJ labeled with HRP (green) to visualize neuronal membranes and phallotoxin to label F-actin (magenta). Scale bar = 20 mm. B: Quantification of muscle sizes in controls and wnt2 mutants. C: Representative confocal micrographs show the 6/7 NMJ immunolabeled with HRP (magenta) and acetylated tubulin (green). Scale bar = 20 mm.