Noonan syndrome (NS) and NS with multiple lentigines (NSML) cognitive dysfunction are linked to SH2 domain-containing protein tyrosine phosphatase-2 (SHP2) gain-of-function (GoF) and loss-of-function (LoF), respectively. In Drosophila disease models, we find both SHP2 mutations from human patients and corkscrew (csw) homolog LoF/GoF elevate glutamatergic transmission. Cell-targeted RNAi and neurotransmitter release analyses reveal a presynaptic requirement. Consistently, all mutants exhibit reduced synaptic depression during high-frequency stimulation. Both LoF and GoF mutants also show impaired synaptic plasticity, including reduced facilitation, augmentation, and post-tetanic potentiation. NS/NSML diseases are characterized by elevated MAPK/ERK signaling, and drugs suppressing this signaling restore normal neurotransmission in mutants. Fragile X syndrome (FXS) is likewise characterized by elevated MAPK/ERK signaling. Fragile X Mental Retardation Protein (FMRP) binds csw mRNA and neuronal Csw protein is elevated in Drosophila fragile X mental retardation 1 (dfmr1) nulls. Moreover, phosphorylated ERK (pERK) is increased in dfmr1 and csw null presynaptic boutons. We find presynaptic pERK activation in response to stimulation is reduced in dfmr1 and csw nulls. Trans-heterozygous csw/+; dfmr1/+ recapitulate elevated presynaptic pERK activation and function, showing FMRP and Csw/SHP2 act within the same signaling pathway. Thus, a FMRP and SHP2 MAPK/ERK regulative mechanism controls basal and activity-dependent neurotransmission strength.
Citation: Leahy SN, Song C, Vita DJ, Broadie K (2023) FMRP activity and control of Csw/SHP2 translation regulate MAPK-dependent synaptic transmission. PLoS Biol 21(1): e3001969. https://doi.org/10.1371/journal.pbio.3001969
Academic Editor: Bing Ye, University of Michigan, UNITED STATES
Received: April 1, 2022; Accepted: December 16, 2022; Published: January 26, 2023
Copyright: © 2023 Leahy et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by National Institutes of Mental Health Grant MH084989 to K.B. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: BSA, bovine serum albumin; csw, corkscrew; dfmr1, Drosophila fragile X mental retardation 1; DMSO, dimethylsulfoxide; EJC, excitatory junction current; ERK, extracellular signal-regulated kinase; FMRP, Fragile X Mental Retardation Protein; FXS, Fragile X syndrome; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GoF, gain-of-function; HFS, high-frequency stimulation; HRP, horseradish peroxidase; JNK, c-Jun N-terminal kinase; LoF, loss-of-function; LTM, long-term memory; MAPK, mitogen-activated protein kinase; mEJC, miniature excitatory junction current; NMJ, neuromuscular junction; NS, Noonan syndrome; NSML, NS with multiple lentigines; pERK, phosphorylated ERK; PPR, paired-pulse ratio; PTP, post-tetanic potentiation; PTPN11, protein tyrosine phosphatase non-receptor type 11; RIP, RNA-immunoprecipitation; RNAi, RNA interference; RRP, readily releasable pool; SHP2, SH2 domain-containing protein tyrosine phosphatase-2; TBST, Tris-buffer saline with 0.1% Tween-20; TEVC, two-electrode voltage-clamp; TRiP, Transgenic RNAi Project; VNC, ventral nerve cord
Noonan syndrome (NS) is an autosomal dominant genetic disorder caused by mutations in the mitogen-activated protein kinase (MAPK) pathway [1,2]. Missense mutations within the protein tyrosine phosphatase non-receptor type 11 (PTPN11) gene account for >50% of all disease cases . In both patients and disease models, the MAPK pathway is hyperactivated by NS gain-of-function (GoF) mutations that disrupt the auto-inhibition mechanism between the catalytic protein tyrosine phosphatase domain and N-Src homology-2 (SH2) domain of the PTPN11 encoded SH2 domain-containing protein tyrosine phosphatase-2 (SHP2; [4,5]). In the NS with multiple lentigines (NSML) disease state, PTPN11 loss-of-function (LoF) mutations decrease protein tyrosine phosphatase domain catalytic activity, but the mutants nevertheless maintain a more persistently active enzyme state with temporally inappropriate SHP2 function, causing elevated MAPK pathway hyperactivation similar to the GoF disease condition . Consequently, NS and NSML patients share a great many symptoms associated with elevated MAPK signaling, including cognitive dysfunction (approximately 30% of cases) as well as long-term memory (LTM) impairments [7,8]. The Drosophila NS (GoF) and NSML (LoF) disease models from mutation of the corkscrew (csw) homolog likewise both increase MAPK activation, with GoF and LoF also phenocopying each other [9,10]. Drosophila LTM training generates repetitive waves of csw-dependent neural MAPK activation, with the LTM spacing effect misregulated by csw manipulations . PTPN11 GoF and LoF mutations from human patients transgenically introduced into the Drosophila model provide a powerful new means to compare with csw GoF and LoF mutants in the dissection of conserved neuronal requirements .
Fragile X syndrome (FXS) is similarly well characterized by hyperactivated MAPK signaling within neurons , and the causal Fragile X Mental Retardation Protein (FMRP) RNA-binding translational regulator is proposed to directly bind PTPN11/SHP2 mRNA [14,15]. FMRP also binds many other neuronal transcripts  and could interact with SHP2 in multiple ways to coregulate the MAPK pathway. Moreover, like the NS and NSML disease states, FXS is likewise a cognitive disorder and the leading heritable cause of intellectual disability . Like NS and NSML, the Drosophila FXS disease model also manifests strongly impaired LTM consolidation [17,18]. Mechanistically, MAPK signaling is well known to modulate glutamatergic synaptic neurotransmission strength via the control of presynaptic vesicle trafficking dynamics and glutamate neurotransmitter release probability . Consistently, FMRP is also well characterized to regulate glutamatergic synaptic neurotransmission, including presynaptic release properties and activity-dependent functional plasticity . Importantly, treatment with the MAPK inhibitor Lovastatin corrects hippocampal hyperexcitability in the mouse FXS disease model and ameliorates behavioral symptoms in human FXS patients [21,22]. In the Drosophila FXS disease model, dfmr1 null mutants show elevated presynaptic glutamate release underlying increased neurotransmission strength , as well as activity-dependent hyperexcitability and cyclic increases in glutamate release during sustained high-frequency stimulation trains . Based on this broad foundation, we hypothesized that FMRP regulates PTPN11 (SHP2)/Csw translation to modulate presynaptic MAPK signaling, which, in turn, controls presynaptic glutamate release probability to determine both basal neurotransmission strength and activity-dependent synaptic plasticity.
To investigate this hypothesis, we utilized the Drosophila neuromuscular junction (NMJ) glutamatergic model synapse with the combined use of NS, NSML, and FXS disease models. We first tested both LoF and GoF conditions in both (1) csw mutants and (2) transgenic human PTPN11 lines. In two-electrode voltage-clamp (TEVC) electrophysiological recordings, all of these mutant conditions elevate synaptic transmission. We next employed cell-targeted RNAi and spontaneous miniature excitatory junction current (mEJC) recordings to find Csw/SHP2 specifically inhibits presynaptic glutamate release probability. We next tested activity-dependent synaptic transmission using high-frequency stimulation (HFS) depression assays to show that the mutants display heightened transmission resiliency, consistent with elevated presynaptic function. We discovered that both LoF and GoF mutations impair presynaptic plasticity, with decreased short-term facilitation, maintained augmentation and post-tetanic potentiation (PTP), supporting altered presynaptic function. Consistent with elevated MAPK signaling in NS, NSML, and FXS disease models, feeding with MAPK-inhibiting drugs (Trametinib and Vorinostat) corrects synaptic transmission strength in mutants. As predicted, we found that FMRP binds csw mRNA and that FMRP loss increases Csw protein levels. Both dfmr1 and csw nulls display elevated phosphorylated ERK (pERK) in presynaptic boutons. Importantly, trans-heterozygous double mutants (csw/+; dfmr1/+) exhibit presynaptic MAPK signaling and neurotransmitter release phenotypes, indicating FMRP and Csw/SHP2 operate to control MAPK/ERK signaling and synaptic function. These discoveries link previously unconnected disease states NS, NSML, and FXS via a presynaptic MAPK/ERK regulative mechanism controlling glutamatergic transmission.
Corkscrew/PTPN11 loss and gain of function mutations both increase synaptic transmission
NS and NSML patients often exhibit cognitive deficits , which we hypothesized may arise from altered synaptic transmission. To systematically test this hypothesis, we assay both Drosophila NS/NSML disease models of csw LoF and GoF [9,10,24], as well as PTPN11 mutations from human patients, including both LoF and GoF point mutants . First, we use csw5, a protein null LoF mutant , together with UAS-cswWT for wild-type Csw overexpression  and UAS-cswA72S as a constitutive GoF mutation [9,11]. Second, we use human patient mutations PTPN11N308D, PTPN11Q510E, and PTPN11Q510P to capture the range of NS/NSML disease heterogeneity [3,12]. The transgenes were driven with ubiquitous UH1-Gal4 or neuronal elav-Gal4. The NMJ glutamatergic synapse is used to assay disease model neurotransmission in all variants [26,27]. Employing TEVC recording, we compare mutants to genetic background control (w1118) and transgenic lines to driver controls (UH1-Gal4/w1118 and elav-Gal4/w1118). We test excitatory junction current (EJC) responses driven by motor nerve suction electrode stimulation (0.5 ms suprathreshold stimuli, 0.2 Hz) onto the voltage-clamped (−60 mV) ventral longitudinal muscle 6 in abdominal segments 3/4 . Each data point is the average of 10 sequentially evoked EJC responses recorded in 1 mM [Ca2+] from the same NMJ terminal. Representative recordings and quantified results for all of these comparisons are shown in Fig 1.
TEVC recordings of nerve-stimulated evoked neurotransmission in both LoF and GoF mutations of Drosophila csw and human PTPN11 mutations from NS/NSML patients. (A) Representative EJC traces for the csw mutant comparisons showing 10 superimposed evoked synaptic responses (1.0 mM Ca2+) from w1118 genetic background control, csw5 null mutant, transgenic driver control (UH1-Gal4/w1118), wild-type csw (UH1-Gal4>cswWT), and cswA72S GoF mutant (UH1-Gal4>cswA72S). (B) Quantification of the mean EJC amplitudes in all 5 genotypes using two-sided t tests. (C) Representative evoked EJC traces for the human patient PTPN11 mutations showing 10 superimposed responses in paired control (elav-Gal4/w1118) and GoF mutant (elav-Gal4>PTPN11N308D; left), and control (UH1-Gal4/w1118) and LoF mutants (UH1-Gal4>PTPN11Q510E and PTPN11Q510P; right). (D) Quantification of the mean EJC amplitudes in all 5 genotypes using two-sided t test, Kruskal–Wallis and Dunn’s multiple comparisons. The scatter plots show all of the individual data points as well as mean ± SEM. N = number of NMJs. Significance shown as: p > 0.05 (not significant, n.s.), p < 0.001 (**) and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. csw, corkscrew; EJC, excitatory junction current; GoF, gain-of-function; LoF, loss-of-function; NMJ, neuromuscular junction; NS, Noonan syndrome; NSML, NS with multiple lentigines; PTPN11, protein tyrosine phosphatase non-receptor type 11; TEVC, two-electrode voltage-clamp.
In genetic background controls (w1118), nerve stimulation causes consistent, high-fidelity neurotransmission (Fig 1A, left). In comparison, csw5 LoF mutants display highly elevated synaptic function with an obvious increase in amplitude (Fig 1A, second from left). Quantified measurements show csw5 EJC amplitudes (248.80 ± 12.51 nA, n = 14) strongly elevated compared to controls (156.30 ± 10.28 nA, n = 10), which is a significant increase (p < 0.0001, two-sided t test; Fig 1B). Since NSML (LoF) and NS (GoF) disease states manifest closely parallel phenotypes, we next examined transgenically driven wild-type csw (cswWT) and the GoF mutant (cswA72S). In transgenic ubiquitous driver controls (UH1-Gal4/w1118), nerve stimulation drives transmission comparable to the genetic background alone (Fig 1A, middle). Likewise, cswWT overexpression results in no detectable alteration in synaptic strength, with amplitudes comparable to controls (Fig 1A, second from right). In sharp contrast, the GoF mutant cswA72S exhibits a consistent elevation in transmission amplitude (Fig 1A, right). Quantification shows the UH1-Gal4/w1118 control amplitude (180.10 ± 15.74 nA, n = 18) is comparable to UH1-Gal4>cswWT (189.50 ± 12.52 nA, n = 12), with no significant difference in transmission (p = 0.671, two-sided t test; Fig 1B, middle). The cswA72S GoF mutation causes significantly elevated neurotransmission. Quantified measurements show cswA72S EJC amplitudes (233.70 ± 8.71 nA, n = 15) are strongly increased compared to UH1-Gal4/w1118 driver controls (192.10 ± 11.86 nA, n = 16), a significant elevation (p = 0.009, two-sided t test; Fig 1B). This increased neurotransmission is independent of changes in NMJ architecture (S1A Fig), including muscle size (S1B Fig), NMJ area (S1C Fig), branching (S1D Fig), and bouton number (S1E Fig), which show no significant changes. The elevated neurotransmission is also independent of changes in synapse number (S2A Fig), including active zone density (S2B Fig), postsynaptic glutamate receptors (S2C Fig), and synaptic apposition (S2D Fig), which are similarly unaltered. Expressing cswWT in the csw5 null restores neurotransmission to the control levels (S3A and S3B Fig), indicating phenotype specificity. We therefore conclude that csw LoF and GoF increase glutamatergic synaptic transmission, comparable to the phenocopy of NS/NSML disease state symptoms in human patients.
To further test effects, we next assayed PTPN11 patient mutations. Compared to transgenic controls, all the PTPN11 mutations cause clearly strengthened synaptic function (Fig 1C). The NS PTPN11N308D, NSML PTPN11Q510E, and NSML PTPN11Q510P mutations all display consistent EJC elevations compared to the controls, similar to LoF/GoF csw animals (compare Fig 1A and 1C). For the GoF condition, the human PTPN11N308D mutation is driven only in neurons (elav-Gal4) since ubiquitous expression results in lethality complications. Quantification compared to neuronal driver control (elav-Gal4/w1118) EJC amplitude (138.70 ± 13.95 nA, n = 12) shows NS (GoF) PTPN11N308D EJC amplitude (212.20 ± 11.13 nA, n = 10) is significantly elevated (p = 0.001, two-sided t test, Fig 1D, left). The patient-derived PTPN11 LoF mutations similarly display increased transmission amplitudes, including PTPN11Q510E (227.40 ± 11.64 nA, n = 19) and PTPN11Q510P (227.90 ± 11.28 nA, n = 17) compared to the matched ubiquitous driver controls (UH1-Gal4/w1118; 178.40 ± 7.73 nA, n = 22). These changes are significant both together (p = 0.0006, Kruskal–Wallis; Fig 1D, right) and when compared individually for both PTPN11Q510E (p = 0.004, Dunn’s multiple comparison; Fig 1D) and PTPN11Q510P (p = 0.003, Dunn’s multiple comparisons; Fig 1D). The patient PTPN11 mutants are not different from each other (p > 0.999, Dunn’s multiple comparisons; Fig 1D). Additionally, PTPN11WT overexpression results in no detectable alteration in synaptic strength, with amplitudes comparable to controls (S3C and S3D Fig). Taken together, these findings indicate that both Drosophila csw and human homolog PTPN11 significantly limit neurotransmission strength. EJCs are elevated with both LoF and GoF, but not by simple overexpression. The next pressing question was to determine whether synaptic strengthening is due to increased presynaptic glutamate release, postsynaptic glutamate receptor responsiveness, or both together.
Corkscrew/PTPN11 controls presynaptic transmission by altering glutamate release probability
Our next objective was to determine where Corkscrew acts to mediate synaptic changes in neurotransmission strength. To test requirements, we knocked down csw expression through RNA interference (RNAi) driven in the different cells contributing to the NMJ, including the presynaptic motor neuron and postsynaptic muscle . We used targeted transgenic RNAi against csw (BDSC 33619; ) to test each cell-specific function. This line is from the Harvard Transgenic RNAi Project (TRiP), which provides a background control stock (BDSC 36303) containing all components except the UAS-RNAi . To test RNAi efficacy and replication of csw5 null phenotypes, we first used the ubiquitous daughterless UH1-Gal4 driver. To separate cellular requirements, we used neuronal elav-Gal4 and muscle 24B-Gal4-specific drivers, each compared to their respective driver alone transgenic controls. With each RNAi knockdown, we once again utilized TEVC recordings of evoked EJC neurotransmission to measure synaptic strength. To further test csw functional roles, we analyzed spontaneous release events by assessing changes in both frequency and amplitude with miniature EJC (mEJC) recordings . Changes in the mEJC frequency are correlated with alterations in presynaptic fusion probability, whereas changes in mEJC amplitudes indicate differential postsynaptic glutamate receptor function or altered vesicle size [32,33]. We made continuous mEJC recordings collected over 2 minutes using a gap-free configuration filtered at 10 kHz . Each data point corresponds to the mean mEJC frequency and amplitude of all the recorded release events. Representative recordings and quantified results are shown in Fig 2.
The ubiquitous transgenic driver control (UH1-Gal4/TRiP BDSC 36303 control) exhibits neurotransmission indistinguishable from the w1118 genetic background control (Fig 2A, left). Ubiquitous csw knockdown (UH1>csw RNAi) causes elevated neurotransmission closely consistent with the csw5 null mutant (Fig 2A, second from left), demonstrating RNAi efficacy as well as null phenocopy (compare to Fig 1A, left). The quantified EJC measurements show UH1>csw RNAi (233.20 ± 17.45 nA, n = 10) to be strongly elevated compared to controls (152.30 ± 15.65 nA, n = 10), which is a significant increase (p = 0.003, two-sided t test; Fig 2B). The neuronal driver control (elav-Gal4/TRiP BDSC 36303 control) compared to neuronal-specific knockdown (elav>csw RNAi) also shows strong replication of the csw5 null elevated transmission, indicating a primary csw requirement in the presynaptic neuron (Fig 2A, middle pair). Quantified measurements show elav>csw RNAi EJC amplitude (239.70 ± 19.45 nA, n = 10) also strongly increased compared with the elav-Gal4/TRiP driver controls (159.90 ± 9.68 nA, n = 12), which is significant (p = 0.001, two-sided t test; Fig 2B, middle). In contrast, targeted muscle RNAi knockdown (24B>csw RNAi) does not cause any change in evoked neurotransmission compared to the muscle driver control alone (24B-Gal4/TRiP BDSC 36303; Fig 2A, right pair), signifying that postsynaptic Csw does not detectably change synaptic function. When quantified, 24B-Gal4/TRiP (156.50 ± 11.41 nA, n = 10) is comparable to 24B>csw RNAi (170.30 ± 11.24 nA, n = 11), with no significant change in amplitude (p = 0.401, two-sided t test; Fig 2B, right). These findings indicate a primary csw requirement in presynaptic neurons regulating glutamate neurotransmitter release.
Nerve stimulation–evoked recordings based on csw RNAi expressed ubiquitously (UH1-Gal4) or targeted to neurons (elav-Gal4), or muscles (24B-Gal4). (A) Representative EJC traces showing 10 superimposed responses (1.0 mM Ca+2) from control (UH1-Gal4/TRiP) vs. csw RNAi; control (elav-Gal4/TRiP) vs. csw RNAi; and control (24B-Gal4/TRiP) vs. csw RNAi. (B) Quantification of EJC amplitudes using two-sided t tests. (C) Representative mEJC traces (1.0 mM Ca2+) in genetic background control (w1118, top) and csw5 null (bottom). (D) Quantification of the mEJC frequencies using a two-sided t test. (E) Quantification of the mEJC amplitudes using a two-sided t test. (F) Sample mEJC recordings from the driver control (UH1-Gal4/w1118; top) compared to cswA72S GoF (UH1-Gal4>cswA72S; bottom). (G) Quantification of the mEJC frequencies using a two-sided t test. (H) Quantification of mEJC amplitudes using Mann–Whitney test. (I) Sample mEJC recordings in control (elav-Gal4/w1118; top) compared to PTPN11N308D GoF (elav-Gal4>PTPN11N308D; bottom). (J) Quantification of the mEJC frequency using a Mann–Whitney test. (K) Quantification of mEJC amplitude using a Mann–Whitney test. Scatter plots show all the data points and mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.), p < 0.05 (*), p < 0.001 (**), and p < 0.0001 (***). The data underlying this figure can be found in S1 Data. csw, corkscrew; EJC, excitatory junction current; mEJC, miniature EJC; NMJ, neuromuscular junction; RNAi, RNA interference.
To further test pre- versus postsynaptic requirements, we next analyzed spontaneous mEJC release events. Compared to genetic background controls (w1118), csw5 null mutants exhibit an obvious increase in mEJC frequency, without any detectable alteration in amplitudes (Fig 2C). When quantified, mEJC frequency in csw5 nulls (1.46 ± 0.22 Hz, n = 11) is increased compared to controls (0.86 ± 0.086 Hz, n = 15), a significant elevation (p = 0.009, two-sided t test; Fig 2D). There is no significant change in mEJC amplitudes (p = 0.489, two-sided t test; Fig 2E). Like the null mutant, GoF cswA72S animals show increased mEJC frequency compared to controls, with no increase in amplitude (Fig 2F). When quantified, UH1>cswA72S (1.79 ± 0.19 Hz, n = 14) have increased mEJC frequency compared to controls (1.12 ± 0.10 Hz, n = 20), which is a significant elevation (p = 0.002, two-sided t test; Fig 2G). Quantification shows no significant change in mEJC amplitudes (p = 0.796, Mann–Whitney; Fig 2H). Similarly, patient-derived PTPN11N308D mutants display increased mEJC frequency with no change in amplitude (Fig 2I). Quantification shows PTPN11N308D frequency (1.79 ± 0.13 Hz, n = 12) increased versus controls (1.09 ± 0.09 Hz, n = 13), which is a significant elevation (p = 0.001, Mann–Whitney; Fig 2J). There is no significant change in amplitudes (p = 0.168, Mann–Whitney; Fig 2K). These findings indicate that both LoF and GoF mutations alter neurotransmission by increasing presynaptic glutamate release rate. We confirmed results further by testing mEJCs in different RNAi conditions. We find mEJC frequencies increased with ubiquitous csw RNAi (S4A and S4B Fig) and neuron-targeted csw RNAi (S4D and S4E Fig), but no change with muscle-specific RNAi (S4G and S4H Fig). None of these manipulations alter mEJC amplitude (S4C, S4F and S4I Fig). Taken together with targeted RNAi results, we conclude that a neuronal requirement regulates glutamate release from the presynaptic terminal. Quantal content determined by dividing EJC amplitude by mean mEJC amplitude shows elevated quantal content in the mutants (S5A Fig) as well as ubiquitous/neuronal csw RNAi (S5B Fig). Moreover, PTPN11 LoF patient mutations driven neuronally phenocopy all GoF defects, including elevated neurotransmission (S6A and S6B Fig) and increased presynaptic fusion (S6C and S6D Fig), but no change in mEJC amplitude (S6E Fig), consistent with the increase in quantal content (S6F Fig). This suggested that stimulation paradigms challenging neurotransmission maintenance should reveal changes in vesicle release dynamics in the absence of csw/PTPN11 function.
Corkscrew/PTPN11 regulates high-frequency stimulation synaptic depression
To further investigate how csw/PTPN11 affects presynaptic neurotransmission strength, we stimulated at a heightened frequency that has been shown to cause synaptic depression over a time course of several minutes [34–36]. Synaptic depression occurs when HFS causes synaptic vesicles to be released at a faster rate than they can be replenished in presynaptic boutons [34,37]. Based on published HFS protocols for the Drosophila NMJ [34,36,38], we compared the genetic background control (w1118), csw null LoF mutant (csw5), and patient-derived PTPN11N308D GoF mutant (elav-Gal4>PTPN11N308D) with a HFS paradigm. To determine the baseline EJC amplitudes, we first stimulated for 1 minute under basal conditions (0.5 ms suprathreshold stimuli at 0.2 Hz in 1.0 mM external [Ca2+]). We then stimulated at 100X greater frequency (20 Hz) for 5 minutes while continuously recording EJC responses. This sustained HFS train causes progressively decreased neurotransmission over time (depression). HFS transmission was quantified to analyze the synaptic vesicle readily releasable pool (RRP) and paired-pulse ratio (PPR) release probability. Representative HFS recordings and quantified results are shown in Figs 3 and S7.
Prolonged HFS drives progressive synaptic amplitude depression over several minutes of continuous recording at 20 Hz (1mM Ca+2). (A) Representative nerve-stimulated EJC traces at the basal frequency (t = 0) and indicated time points during the HFS train for genetic background control (w1118, top), csw null (csw5, middle), and PTPN11N308D GoF mutant (elav-Gal4>PTPN11N308D; bottom). (B) Quantification of cumulative EJC amplitudes over the first 100 stimulations via nonlinear regression exponential for each pair tested using extra sum-of-squares F tests. (C) Quantification of the RRP of w1118 and csw5 (two-sided t test) and elav-Gal4/w1118 and PTPN11N308D (Mann–Whitney). (D) Quantification of the PPR of w1118 and csw5 (two-sided t test) and elav-Gal4/w1118 and PTPN11N308D (Mann–Whitney). Scatter plots show all data points and mean ± SEM. N = number of NMJs. Significance: p < 0.05 (*), p < 0.001 (**), p < 0.001 (***), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. csw, corkscrew; EJC, excitatory junction current; HFS, high-frequency stimulation; NMJ, neuromuscular junction; PPR, paired-pulse ratio; RRP, readily releasable pool.
During HFS, w1118 controls exhibit a steady decrease in EJC amplitudes throughout the train (Fig 3A, top). The PTPN11N308D GoF mutants and csw5 LoF nulls show stronger maintained EJC amplitudes over time and prolonged resistance to depression (Figs 3A and S7A). RRP size was calculated by dividing the cumulative EJCs during the first 100 responses by mean mEJC amplitudes . There is a sustained elevated response in both LoF and GoF mutants (Fig 3B). When compared with nonlinear regression and extra sum-of-squares, the stimulation train profiles are significantly greater for both LoF (p < 0.0001, F(2,1296) = 1064) and GoF (p < 0.0001, F(2,1996) = 705.5; Fig 3B) mutants, indicating increased resiliency to depression. The RRP size of csw5 nulls is significantly increased compared to w1118 background controls (p = 0.001, two-sided t test; Fig 3C, left). Similarly, PTPN11N308D GoF mutants exhibit an increased RRP compared to transgenic elav/+ neuronal driver controls (p = 0.047, Mann–Whitney; Fig 3C, right). PPR analyzed for both mutants shows no in change in csw5 nulls (p = 0.865, two-sided t test; Fig 3D, left) or PTPN11N308D GoF mutants (p = 0.941, Mann–Whitney; Fig 3D, right) compared to their respective controls. The depression resistance continues for 5 minutes of continuous stimulation (S7B Fig). Taken together, these results indicate mutants maintain transmission better with a HFS challenge. We therefore next turned to examining changes in activity-dependent synaptic function under both LoF and GoF mutant conditions.
Corkscrew/PTPN11 enables short-term plasticity facilitation, augmentation, and potentiation
Presynaptic activity drives numerous forms of short-term plasticity dependent on release mechanisms [40,41]. In high external [Ca2+], strong stimulation results in neurotransmission depression as above, but with reduced external [Ca2+], many forms of release strengthening are revealed, including short-term facilitation and maintained augmentation during stimulation trains, and PTP following the train [42–44]. Based on published Drosophila plasticity protocols , we compared genetic background controls (w1118 or elav-Gal4/w1118), csw LoF nulls (csw5), and PTPN11 GoF animals (elav-Gal4>PTPN11N308D) with the stimulation paradigm illustrated in Fig 4A. To determine baseline EJC amplitudes, we stimulated at the basal frequency (0.5 ms suprathreshold stimuli/0.2 Hz in 0.2 mM [Ca2+]). We then applied a 10-Hz train for 1 minute, before returning to 0.2 Hz for PTP analyses (Fig 4A). In controls, this paradigm drives strong short-term facilitation during the initial stimuli of the train, followed by maintained transmission augmentation for the full duration of the train . Following return to the basal stimulation frequency (0.2 Hz), heightened EJC amplitudes persist during the PTP period (Fig 4B; ). We normalized EJC amplitudes during and after the 10-Hz train to the initial mean EJC amplitude to show only transmission changes in response to stimulation. Quantified analyses on w1118 control, csw5 LoF, and PTPN11N308D GoF mutants were done for facilitation (<1 second), augmentation (>5 seconds), and PTP (following the HFS train). Representative short-term plasticity recordings and quantified results are shown in Fig 4.
Synaptic plasticity during and following a short-term stimulation train to measure facilitation, augmentation, and PTP. (A) Stimulation paradigm: 1 minute at 0.2 Hz (0.2 mM Ca2+), followed by 1 minute at 10 Hz, and then a return to 0.2 Hz. (B) Sample EJC traces at indicated time points during and following the 10 Hz train for control (w1118, top), GoF PTPN11N308D (elav-Gal4>PTPN11N308D; middle), and csw null (csw5, bottom). (C) Quantification of EJC amplitude during the 10-Hz train normalized to basal EJC amplitude for each genotype. The nonlinear regression exponential for each pair tested using extra sum-of-squares F test. (D-F) Quantification of facilitation (1 second, D) and augmentation (30 seconds, E) during the 10-Hz train, and PTP (10 seconds following train, F) normalized to the basal EJC amplitude for each genotype using Mann–Whitney/two-sided t tests. Scatter plots show all data points and mean ± SEM. N = number of NMJs. Significance: p < 0.05 (*), p < 0.001 (**), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. csw, corkscrew; EJC, excitatory junction current; GoF, gain-of-function; NMJ, neuromuscular junction; PTP, post-tetanic potentiation; PTPN11, protein tyrosine phosphatase non-receptor type 11.
Controls exhibit robust synaptic plasticity, including short-term facilitation (<1 second), maintained augmentation (>5 seconds), and persistent PTP (Fig 4C, top two blue lines). With HFS, w1118 controls exhibit a >3-fold amplitude increase in <5 seconds, which strengthens to a 4-fold increase by 30 seconds. After the HFS train, control animals PTP at >2-fold basal transmission. In contrast, this short-term plasticity is strongly repressed in both the csw5 LoF and PTPN11N308D GoF mutants (Fig 4C, bottom two red lines). When quantified via nonlinear regression and extra sum-of-squares, stimulation train profiles significantly differ for both LoF (p < 0.0001, F(2,662) = 38.95) and GoF (p < 0.0001, F(2,374) = 25.85; Fig 4C). During initial short-term facilitation (1 second), w1118 controls show much stronger strengthening normalized to basal amplitude (2.15 ± 0.19, n = 16) compared to csw5 LoF (1.52 ± 0.14, n = 21; p = 0.005, Mann–Whitney) and a trending decrease in PTPN11N308D GoF (1.44 ± 0.16, n = 12; p = 0.229, two-sided t test; Fig 4D). With maintained augmentation during the HFS train (30 seconds), w1118 controls are highly elevated (4.27 ± 0.70, n = 16) compared to csw5 LOF (2.67 ± 0.53, n = 21; p = 0.009, Mann–Whitney) and PTPN11N308D GOF (2.91 ± 0.53, n = 12; p = 0.015, Mann–Whitney; Fig 4E). At peak PTP after the HFS train, w1118 controls exhibit a significant increase (3.02 ± 0.45, n = 16) compared to csw5 LoF (1.63 ± 0.16, n = 21; p = 0.003, Mann–Whitney; Fig 4F). Likewise, the PTPN11N308D GoF (2.58 ± 0.33, n = 11) shows significantly decreased PTP compared to elav-Gal4/w1118 controls (4.55 ± 0.5, n = 9; p = 0.003, two-sided t test; Fig 4F). These results show a role in presynaptic release dynamics, with altered responses to evoked stimulation. To understand the mechanism of these changes, we next turned to testing the role of MAPK/ERK signaling.
Elevated Corkscrew/PTPN11 synaptic transmission corrected with pERK inhibitors
NS and NSML phenotypes are hypothesized to converge due to both LoF/GoF disease states exhibiting constitutively elevated MAPK/ERK signaling . Similarly, we hypothesize the mutant LoF/GoF neurotransmission elevation from heightened glutamate release also occurs downstream of elevated presynaptic MAPK/ERK signaling. To test this hypothesis, we used MAPK/ERK inhibitors (Trametinib and Vorinostat) to assay effects on glutamatergic synaptic function. Trametinib binds and inhibits MEK1/2 , resulting in a direct inhibition of MAPK/ERK signaling . Vorinostat acts as a HDAC inhibitor to also inhibit MAPK/ERK signaling [12,46]. Recent work using the PTPN11 mutations from human patients has highlighted these two drugs as possible treatments for a variety of different NS/NSML mutations . Both drugs are thus interesting not only for their ability to test elevated MAPK/ERK signaling upstream of neurotransmission, but also as possible future treatment avenues. We fed both drugs and then analyzed changes in EJC amplitudes using TEVC recording. For each drug, we compared the background control (w1118) without drug treatments to controls with drug treatments (Trametinib and Vorinostat), as well as the csw null mutants (csw5) without drug treatments to nulls with drug treatments (Trametinib and Vorinostat). Quantification of evoked EJC amplitudes in all 8 conditions tests whether each drug changes neurotransmission in control, as well as correction of the null csw5 elevated neurotransmission (Fig 1A). We also analyzed mEJC recordings of the same genotypes to test for correction of csw5 elevated mEJC frequency (Fig 2C and 2D). Representative EJC and mEJC traces and quantified results are shown in Fig 5.
TEVC recordings with and without two pERK inhibiting drugs (Trametinib and Vorinostat) comparing the genetic background control (w1118) and csw null mutant (csw5). (A) Representative EJC traces showing 10 superimposed responses (1.0 mM Ca+2) comparing the control (left) and csw5 null mutant (right), with and without Trametinib. (B) Quantification of mean EJC amplitudes for all 4 conditions using Kruskal–Wallis followed by Dunn’s multiple comparisons. (C) Representative EJC traces comparing the control (left) and csw5 null mutant (right), with and without Vorinostat. (D) Quantification of EJC amplitudes for all 4 conditions using one-way ANOVA followed by Tukey’s multiple comparisons. (E) Representative mEJC traces (1.0 mM Ca2+) in the w1118 control (left) and csw5 null mutant (right), with and without Trametinib. (F) Quantification of mEJC frequency in all 4 conditions using a Kruskal–Wallis followed by Dunn’s multiple comparisons. (G) Quantification of mEJC amplitudes using a Kruskal–Wallis. Scatter plots show all the data points and the mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.) and p < 0.001 (**). The data underlying this figure can be found in S1 Data. EJC, excitatory junction current; mEJC, miniature EJC; NMJ, neuromuscular junction; pERK, phosphorylated ERK; TEVC, two-electrode voltage-clamp.
Null csw5 animals fed Trametinib have clearly decreased neurotransmission compared to untreated mutants, with EJC amplitudes comparable to control animals (Fig 5A). Quantification shows untreated controls (159.10 ± 7.35 nA, n = 36) and drugged controls (161.70 ± 12.01 nA, n = 35) are not significantly different (p > 0.99, Dunn’s; Fig 5B). In contrast, csw5 EJC amplitudes (226.20 ± 9.79 nA, n = 30) are significantly increased compared to controls with (p < 0.0001, Dunn’s; Fig 5B) and without Trametinib (p = 0.001, Dunn’s, Fig 5B). Critically, csw5 nulls fed Trametinib (172.70 ± 11.37 nA, n = 27) are no longer significantly increased from controls with or without Trametinib (p > 0.99, Dunn’s) but are significantly decreased compared to the untreated csw5 nulls (p = 0.003, Dunn’s; Fig 5B). Similar results occur with Trametinib treatment of PTPN11N308D GoF mutants (S8A and S8B Fig). Similarly, Vorinostat fed csw5 nulls have EJC amplitudes restored to the control levels (Fig 5C). Quantification shows controls with (167.20 ± 7.01 nA, n = 16) and without (162.30 ± 9.46 nA, n = 20) Vorinostat are not significantly different (p = 0.994, Tukey’s; Fig 5D). In contrast, csw5 mutants (237.0 ± 14.72 nA, n = 25) are significantly increased versus controls with (p = 0.001, Tukey’s) and without (p = 0.0001, Tukey’s) Vorinostat (Fig 5D). Null csw5 fed Vorinostat (179.70 ± 11.55 nA, n = 25) are not significantly elevated compared to controls with (p = 0.897) and without (p = 0.727) Vorinostat but are significantly decreased compared to untreated csw5 nulls (p = 0.003, Tukey’s; Fig 5D). Trametinib decreases mEJC frequency in csw5 nulls compared to untreated mutants, to levels matching controls (Fig 5E). Quantification shows untreated (0.92 ± 0.12 Hz n = 26) and drugged (1.05 ± 0.103 Hz, n = 28) controls are not significantly different (p > 0.99, Dunn’s; Fig 5F). In contrast, csw5 mEJC frequency (2.13 ± 0.32 Hz, n = 21) is significantly increased compared to controls (no drug, p < 0.0001, Dunn’s; Trametinib, p = 0.003, Dunn’s; Fig 5F). Critically, csw5 nulls fed Trametinib (0.98 ± 0.14 Hz, n = 19) are no longer significantly increased from controls with or without the drug (p > 0.99, Dunn’s) but are significantly decreased compared to untreated csw5 mutants (p = 0.001, Dunn’s; Fig 5F). There are no changes in mEJC amplitude (p = 0.437, Kruskal–Wallis; Fig 5G). Thus, decreasing MAPK/ERK signaling restores presynaptic neurotransmission in csw5 animals. We therefore next aimed to identify the upstream mechanism controlling this regulation.
FMRP binds csw mRNA to suppress Csw protein expression upstream of MAPK/ERK signaling
The FMRP negative translational regulator is well known to inhibit MAPK/ERK signaling in the regulation of synaptic function . Moreover, high-throughput RNA sequencing from isolated crosslinking immunoprecipitation shows FMRP binds csw homolog PTPN11/SHP2 mRNA . Therefore, we hypothesized FMRP binds csw mRNA to negatively regulate translation upstream of MAPK/ERK signaling. To test this hypothesis, we first performed RNA-immunoprecipitation (RIP) studies with tagged FMRP::YFP from larval lysates using magnetic GFP-trap beads [47,48]. We used Tubby::GFP lysates as the RIP negative control, with α-tubulin (FMRP does not bind) as the internal negative control, and futsch/MAP1B (known FMRP target) as the internal positive control . Immunoprecipitated mRNAs were reverse transcribed and tested with specific primers on 2% agarose gels. We next used western blots from larval ventral nerve cord (VNC)/brain lysates to test neuronal Csw protein levels with a characterized anti-Csw antibody . Antibody specificity was confirmed with the csw5 null and protein levels compared between the genetic background control (w1118) and FXS disease model (dfmr1 null mutants). To compare neuronal Csw protein levels in these different genotypes, we normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a housekeeping gene that we confirmed is not regulated by Csw. Normalized quantification was done to compare neuronal Csw protein levels in the w1118 controls, csw5 null mutants, and dfmr150M null mutants. Representative RIP gels, western blots, and western blot quantified data are shown in Fig 6.
(A) RIP control (Tubby::GFP, top) and FMRP (FMRP::YFP, bottom), with csw, futsch (positive control), and α-tubulin (negative control) RNAs. (B) Western blot for Csw (100 kDa, top) and GAPDH control (35 kDA, bottom) w1118 control, dfmr150M null, and csw5 null. (C) Quantification of Csw levels normalized to GAPDH using one-way ANOVA followed by Tukey’s multiple comparisons. (D) Representative NMJ images of w1118 control, csw5 null, and dfmr150M null colabeled for pERK (green) and presynaptic membrane marker anti-HRP (magenta). pERK fluorescence shown as a heat map. NMJs shown without stimulation (basal, top) and with 90 mM [K+] HFS (high K+, bottom). Scale bar: 2.5 μm. (E) Quantified normalized basal presynaptic anti-pERK fluorescence for all 3 genotypes using one-way ANOVA and Tukey’s multiple comparisons. (F) Quantified normalized stimulated presynaptic anti-pERK fluorescence using one-way ANOVA and Tukey’s multiple comparisons. (G) Quantification of normalized presynaptic pERK levels in all 3 genotypes under basal and stimulated conditions using two-sided t tests. Scatter plots show all data points and mean ± SEM. N = number of animals (C) or NMJS (E-G). Significance: p > 0.05 (not significant, n.s.), p < 0.05 (*), p < 0.001 (**), p > 0.001 (***), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. csw, corkscrew; FMRP, Fragile X Mental Retardation Protein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HFS, high-frequency stimulation; HRP, horseradish peroxidase; NMJ, neuromuscular junction; pERK, phosphorylated ERK; RIP, RNA-immunoprecipitation.
For the RIP analyses, csw, futsch, and α-tubulin mRNA bands are all present in both Tubby::GFP control and FMRP::YFP input lysates (Fig 6A). Immunoprecipitation pulls down csw mRNA from the FMRP::YFP third instar lysates, with no binding in the Tubby::GFP control (Fig 6A). Additionally, the positive control futsch mRNA is pulled down, but there is no detectable negative control α-tubulin mRNA. These results indicate FMRP binds csw mRNA, with the controls confirming binding interaction specificity. Based on this and above findings, we hypothesized FMRP partly inhibits NMJ synaptic transmission by suppressing Csw translation in neurons to decrease MAPK/ERK signaling. To test this hypothesis, western blot analyses were done to test Csw protein levels in larval brain/VNC lysates from controls (w1118), csw5, and dfmr150M null mutants. At the predicted molecular weight (100 kDa), there is a clear Csw band present in controls (Fig 6B). This band is undetectable in csw5 nulls, demonstrating specificity (Fig 6B). In the FXS disease model, there are clearly and consistently increased Csw protein levels in dfmr1 null mutants (Fig 6B). Quantified comparisons normalized to GAPDH (p < 0.0001, ANOVA) show an increase in Csw levels in dfmr1 nulls (1.55 ± 0.13) compared to controls (0.99 ± 0.029), which reveals a highly significant increase in the FXS disease model (p = 0.0008, Tukey’s, Fig 6C). There is slight background in csw5 (0.23 ± 0.06), which is very significantly decreased from controls (p < 0.0001, Tukey’s) and dfmr1 mutants (p < 0.0001, Tukey’s; Fig 6C). Thus, dfmr1 nulls have a strong increase in Csw levels in the larval neurons. Taken together, these findings show FMRP binds csw mRNA to negatively regulate Csw protein levels. We hypothesized this interaction negatively regulates presynaptic MAPK/ERK signaling.
FMRP and Csw interact to inhibit presynaptic MAPK/ERK signaling and neurotransmission
We next set forth to test MAPK/ERK signaling within presynaptic boutons in order to begin investigating how FMRP and Csw interact to control presynaptic transmission. Elevated presynaptic pERK is well known to positively regulate neurotransmitter release function . Based on this known role and our above studies, we hypothesized locally elevated pERK levels should occur in both csw and dfmr1 null synaptic boutons. To test this hypothesis, we assayed NMJ terminals double-labeled with anti-pERK  and anti-horseradish peroxidase (HRP) to mark presynaptic bouton membranes. Using HRP to delineate presynaptic boutons, we measured pERK fluorescence intensity normalized to the genetic background control (w1118). Presynaptic pERK signaling is activity-dependent [51,52]. To test this function, we compared presynaptic pERK levels in the basal resting condition to stimulation with acute (10 minute) high [K+] depolarization (90 mM; [53,54]) in w1118 control, dfmr150M null mutant, and csw5 null mutant. We hypothesized that FMRP and Csw interact to inhibit presynaptic pERK signaling-dependent transmission strength. To test this hypothesis, we assayed the double trans-heterozygous csw5/+; dfmr150M /+ mutant compared to both single heterozygous mutants alone . We first used TEVC recordings to measure stimulation evoked EJC responses and spontaneous mEJC release events. We then used pERK/HRP double-labeled imaging to measure the presynaptic pERK fluorescence intensity levels. Representative raw data of recordings and images as well as quantified results are shown in Fig 6.
Activated pERK is weakly detectable at control synapses under basal resting conditions (Fig 6D, top). In w1118 controls, pERK is localized at relatively higher levels in the presynaptic boutons, with lower levels of signaling in the adjacent muscle nuclei and very low sporadic levels throughout the muscle. Given the consistent presynaptic phenotypes above, we focused analyses on pERK signaling within presynaptic boutons. Compared to controls, both csw and dfmr1 null mutants display consistently elevated pERK levels within the presynaptic boutons (Fig 6D, top), but with similar levels of pERK fluorescence in muscle compared to the controls. Similar results occur in PTPN11 human patient mutants compared to driver controls (S9A Fig), with elevated pERK levels in all conditions (S9B Fig). This increased presynaptic pERK signaling and lack of postsynaptic changes is consistent with presynaptic perturbations in both csw and dfmr1 null mutants. Quantification of the normalized pERK fluorescent intensity within the HRP-delineated presynaptic boutons shows very highly elevated levels in both the csw (1.85 ± 0.25, n = 15) and dfmr1 (1.58 ± 0.13, n = 18) null mutants compared to controls (1.0 ± 0.12, n = 24), which is a significant increase (p = 0.001, one-way ANOVA; Fig 6E). When compared individually, there is no significant difference between dfmr1 and csw mutants (p = 0.526, Tukey’s), showing both csw (p = 0.001, Tukey’s) and dfmr1 (p = 0.024, Tukey’s; Fig 6E) nulls increase pERK signaling to a similar degree compared to controls. This elevated presynaptic pERK in both disease models fits our hypothesis that elevated MAPK/ERK signaling causes the increased presynaptic transmission in both disease models. Given the above changes in activity-dependent presynaptic function in csw null mutants, we next wanted to test whether pERK levels are dynamic and change with a stimulation challenge, and whether activity-dependent impairments occur in the two disease models.
When NMJs are strongly stimulated by acute depolarization (90 mM [K+] for 10 minutes), w1118 controls exhibit sharply increased presynaptic pERK levels compared to the basal resting condition (Fig 6D, bottom). Both dfmr1 and csw nulls show smaller pERK level increases upon stimulation. This elevation shows pERK levels can be further increased in null mutants, indicating that the mechanism behind the increase is not exhausted under basal conditions or is controlled by other mechanisms beyond activity. Quantification of presynaptic pERK fluorescent intensity levels normalized to rest (p = 0.007, one-way ANOVA) shows pERK elevation in controls (1.68 ± 0.12, n = 21), csw nulls (2.44 ± 0.22, n = 15), and dfmr1 (2.09 ± 0.14, n = 15) nulls, with csw exhibiting a significant elevation compared to controls (p = 0.005, Tukey’s; Fig 6F). When stimulated, pERK levels are similar in csw and dfmr1 (p = 0.341, Tukey’s); however, dfmr1 nulls are no longer significantly increased compared to controls (p = 0.192, Tukey’s; Fig 6F). To further assay activity-dependent changes, we directly compared the basal and stimulated pERK levels. Importantly, controls exhibit a significant activity-dependent presynaptic pERK increase when compared to rest (p = 0.0003, two-sided t test; Fig 6G). In contrast, csw nulls display only a trending elevation in stimulated pERK levels, without a significant increase from rest (p = 0.083, two-sided t test; Fig 6G). Likewise, dfmr1 nulls display a reduced activity-dependent increase in stimulated presynaptic pERK levels compared to the basal condition, albeit still significant (p = 0.014, two-sided t test; Fig 6G). We conclude that the basal elevation in pERK levels in both disease models blunts further activation in response to stimulation. This activity-dependent defect correlates with the above impaired functional neurotransmission dynamics in response to stimulation. Based on the perturbed presynaptic pERK signaling in csw and dfmr1 nulls, we hypothesized FMRP and Csw interact to inhibit synaptic MAPK/ERK signaling and transmission.
We therefore directly tested for this mechanism with csw/+; dfmr1/+ trans-heterozygotes. In TEVC recordings, these trans-heterozygotes show elevated neurotransmission compared to w1118 controls and both of the single heterozygotes (S10A Fig). Quantification reveals that the csw/+; dfmr1/+ trans-heterozygotes have higher EJC amplitudes (237.80 ± 7.5810 nA, n = 20) compared to w1118 controls (169.67 ± 8.1240 nA, n = 32), a significant increase (p < 0.0001, Dunnett’s; S10B Fig). In contrast, both csw/+ (199.10 ± 10.92 nA, n = 23) and dfmr1/+ (194.0 ± 11.36 nA, n = 18) heterozygotes display similar EJC amplitudes comparable to the w1118 control (S10A Fig), with no significant elevation (p = 0.19/0.058, Dunnett’s; S10B Fig). In mEJC recordings, double csw/+; dfmr1/+ trans-heterozygotes display a clear increase in mEJC frequency compared to both w1118 control and single heterozygotes (S10C Fig). Quantification shows trans-heterozygote mEJC frequency (2.60 ± 0.29 Hz, n = 16) elevated compared to w1118 (1.34 ± 0.15 Hz, n = 19), a significant increase (p = 0.0002, Dunn’s; S10D Fig). Both of the single heterozygotes, csw/+ (1.69 ± 0.19 Hz, n = 16) and dfmr1/+ (1.91 ± 0.26 Hz, n = 15), display a similar frequency comparable to w1118 control (S10C Fig), with no significant change (p = 0.428/0.151, Dunn’s; S10D Fig). There are no significant changes in the mEJC amplitudes (p = 0.855, Kruskal–Wallis; S10E Fig), confirming a presynaptic mechanism. Activated pERK labeling shows csw/+; dfmr1/+ trans-heterozygotes have elevated presynaptic signaling compared to w1118 control and the single heterozygotes (S10F Fig). Quantification shows increased presynaptic pERK fluorescence intensity in the trans-heterozygote (1.64 ± 0.11, n = 34) normalized to control (1.0 ± 0.07, n = 41), a significant elevation (p < 0.0001, Dunn’s; S10G Fig). Both of the single heterozygotes, csw/+ (1.02 ± 0.10, n = 31) and dfmr1/+ (1.23 ± 0.12, n = 34) have presynaptic pERK levels comparable to the control (S10F Fig), showing no significant change (p > 0.999/0.312, Dunn’s; S10G Fig). Taken together, these findings indicate FMRP and Csw interact to regulate presynaptic pERK signaling upstream of neurotransmitter release.
MAPK is well known to regulate activity-dependent signal transduction and synaptic plasticity within the nervous system . Four MAPK families have been characterized, including extracellular signal-regulated kinase 1/2 (ERK1/2), ERK5, p38 MAPK, and the c-Jun N-terminal kinase (JNK; ). These families are activated similarly through an evolutionarily conserved cascade involving initial activation of GTPases (Ras/Rac) and a subsequent three-tiered protein kinase signaling system . The best-characterized MAPK pathway, ERK1/2, has been extensively investigated within the nervous system, where ERK activation is very tightly regulated. Numerous neurological disease states display elevated ERK activity, including FXS, NS, and NSML, as well as neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease [10,13,58]. Many studies have linked such elevated ERK signaling to cognitive deficits, particularly impairment of LTM consolidation. LTM requires spaced learning sessions during which ERK is activated and then decays in a temporal cycle. In Drosophila PTPN11/SHP2 homolog csw mutants, this ERK activation timing cycle is perturbed and LTM is disrupted . Moreover, one of the targets of FMRP, a negative translational regulator, is PTPN11/SHP2 mRNA , suggesting a potential link between the FXS and NS/NSML disease states. Based on the common ERK signaling up-regulation in these disorders, we hypothesized FMRP regulates Csw translation to modulate synaptic ERK levels to control neurotransmission strength and functional plasticity.
This hypothesis provides the first proposed mechanistic connection between NS, NSML, and FXS disease conditions, through an ERK phosphorylation (pERK) signaling defect in presynaptic boutons. pERK is known to activate presynaptic function, with short-term roles in the control of neurotransmission strength and activity-dependent plasticity [49,59], and longer-term nuclear translocation roles . In the Drosophila NS/NSML disease models of csw LoF and GoF, we began with synaptic transmission assays at the NMJ glutamatergic synapse . We also tested human patient PTPN11/SHP2 mutations to confirm functional requirements . Our work reveals that all LoF/GoF mutations elevate neurotransmission strength, indicating that Csw/SHP2 is involved in inhibiting glutamatergic signaling. Consistently, previous Drosophila NS and NSML model studies also show that LoF and GoF mutations phenocopy one another, with a correlation to hyperactivated pERK signal transduction in both conditions [9,10]. Moreover, the Drosophila FXS disease model similarly increases NMJ glutamatergic synaptic transmission , consistent with the FMRP mechanistic intersection. Localized pERK signaling occurs on both pre- and postsynaptic sides [60,61], so we next used cell-targeted csw RNAi and measured spontaneous vesicle fusion events to separate these requirements. Our work reveals Csw/SHP2 has only a neuronal role in the regulation of presynaptic transmission. There is no detectable postsynaptic function. This new presynaptic Csw/SHP2 role is consistent with the abundant evidence for both MAPK/ERK and FMRP involvement in modulating glutamatergic release mechanisms.
Presynaptic vesicle fusion is a major determinant of neurotransmission strength, maintained functional resilience during strong demand, and activity-dependent plasticity . HFS trains cause the transient activation of pERK signaling in presynaptic terminals , correlating with increased vesicle fusion. To test if Csw/SHP2 similarly regulates glutamate release, we performed HFS synaptic depression assays to discover that all mutants have increased transmission resiliency under conditions of heightened demand , with elevated glutamate release from presynaptic boutons. This role is consistent with activity-dependent presynaptic MAPK/ERK signaling driving greater presynaptic glutamate release by modulating the accessible number of synaptic vesicles available for fusion in the RRP . Importantly, the mouse FXS disease model displays similar decreased short-term depression due to enhanced presynaptic glutamate release, also via up-regulation of the RRP without a change in PPR fusion . The MAPK/ERK-dependent phosphorylation of presynaptic targets is likewise known to increase short-term plasticity, and blockade of this signaling process has been shown to strongly impair facilitation, maintained augmentation, and PTP [51,63]. Our results show that all three forms of synaptic plasticity are impaired in csw null and PTPN11N308D GoF animals, which both show decreased facilitation, augmentation, and PTP, consistent with other LoF/GoF phenocopy. We hypothesize that these plasticity defects correlate to the already increased basal transmission levels that cause a decrease in range for enhancement from presynaptic pERK activation, leading to a “ceiling effect” on presynaptic function. This predicts neurotransmission defects are linked to causal changes in presynaptic MAPK/ERK signaling.
Both NS and NSML disease states exhibit elevated MAPK/ERK signaling , but there is heterogeneity in pERK activation levels and multiple pathways involved . To confirm the neurotransmission increase is due to elevated MAPK/ERK signaling, we inhibited this pathway with both Trametinib and Vorinostat, two drugs well characterized to decrease pERK signaling [46,64]. With drug treatments, the elevated neurotransmission in csw and PTPN11 mutants is restored to levels comparable to control animals, indicating that the elevated MAPK/ERK signaling is responsible for the heightened presynaptic function. This test does not rule out the possibility of other disrupted signaling pathways that may influence MAPK/ERK signaling, but does prove MAPK/ERK signaling is the cause of the elevated neurotransmission. The next task was to explore the new activity-dependent mechanism controlling this presynaptic Csw/SHP2 function. As previously discussed, NS, NSML, and FXS models/patients all display striking similarities in up-regulated MAPK/ERK signaling, synaptic phenotypes, and LTM impairments [17,18,20]. Moreover, RNA-binding FMRP is well characterized as an activity-dependent negative translational regulator of presynaptic mRNA targets . Consistently, we find that Drosophila FMRP binds csw mRNA, as suggested in a mouse FMRP screen indicating PTPN11/SHP2 binding . Additionally, we find neuronal Csw protein levels are elevated in the FXS disease model (dfmr1 null), consistent with the predicted FMRP translational repression . Finally, we find that presynaptic pERK signaling is increased in both dfmr1 and csw null mutants and that normal activity-dependent elevation in pERK signaling is impaired in both disease model conditions. The pERK enhancement levels are slightly different, but this to likely due to the relative effect of the two nulls on pERK signaling. The heightened basal presynaptic pERK signaling and repressed activity-dependent pERK signaling suggests that FMRP and Csw interact to modulate presynaptic glutamatergic neurotransmission.
One genetic test for pathway interaction employs nonallelic noncomplementation , which demonstrates that the two gene products operate within a common mechanism, in this case, the up-regulation of MAPK signaling . Both dfmr1 and csw null mutants display elevated presynaptic neurotransmission with an increased probability of presynaptic glutamate release , and trans-heterozygous dfmr1/+; csw/+ double mutants recapitulate both functional phenotypes. Importantly, both the dfmr1 and csw5 single heterozygous mutants do not display any phenotypes, despite the NSML autosomal dominant disease state. Similarly, Csw/PTPN11 overexpression does not cause any phenotypes, suggesting a change in the FXS background causes the elevated MAPK/ERK presynaptic signaling. These genetic tests indicate that FMRP and Csw/SHP2 act together to inhibit pERK signaling and presynaptic glutamate release. We propose the mechanism of mRNA-binding FMRP acting canonically as a negative translational regulator of Csw/SHP2 expression . Both the dfmr1 and csw null mutants display elevated MAPK/ERK signaling as indicated by pERK production , and we demonstrate here pERK elevation in presynaptic boutons. Consistent with a common mechanism, trans-heterozygous csw/+; dfmr1/+ mutants recapitulate this heightened presynaptic pERK signaling. We propose the mechanism of FMRP working through Csw/SHP2 phosphatase enzymatic activity to inhibit presynaptic pERK production. Given that MAPK/ERK signaling is well established to modulate presynaptic glutamatergic release , we suggest heightened presynaptic pERK signaling causes elevated glutamate release probability. We demonstrate this causal link with pharmacological treatments that block pERK production , which act to restore normal glutamatergic synaptic signaling in the disease model animals.
In conclusion, we note that there are important differences between FXS and NS/NSML disease models. Previous FXS model work has shown increased NMJ architecture and mEJC amplitudes in dfmr1 nulls , which are absent in NS/NSML model csw/PTPN11 mutants. FXS is a very complex disease state with many proteins misregulated , and there was never an expectation that all FXS phenotypes would be recapitulated in csw/PTPN11 mutants, especially for the unrelated postsynaptic changes. Nevertheless, the presynaptic parallels are striking. The mouse FXS model exhibits decreased short-term depression with no change in PPR, but an increase in RRP , matching the Drosophila results shown here. Interestingly, these phenotypes match closer than mouse H-rasG12V mutants with increased pERK signaling, which exhibit enhanced short-term synaptic plasticity , compared to the depressed plasticity shown here. Thus, although both basal transmission strength and functional plasticity properties are dependent on presynaptic MAPK/ERK signaling, there are likely other intersecting regulatory pathways. Moreover, FMRP and Csw/SHP2 could interact via multiple different mechanisms to regulate presynaptic MAPK/ERK signaling, and the elevated neurotransmission in the disease state models may not be completely dependent on presynaptic MAPK/ERK signaling. In the FXS model, Csw/SHP2 is both up-regulated and hyperactivated, and the mechanism of this activation is unknown. One possibility is decreased MAPK/ERK negative regulation, via other factors like Neurofibromin-1, which could further increase MAPK/ERK signaling [69,70]. Another possibility is that neuronal activity up-regulates and then activates Csw/SHP2 via two parallel mechanisms to increase MAPK/ERK signaling [71,72]. We have previously uncovered several other genetic mutants that likewise elevate neurotransmission and depress short-term plasticity [28,73–75], which are also candidates for furthering our understanding in future studies. The possibility for a more extensive interactive molecular network is exciting, but it can currently only be concluded that FMRP and Csw/SHP2 both control MAPK/ERK signaling and modulate neurotransmission. This presynaptic mechanism connects the previously unlinked disorders of NS, NSML, and FXS, suggesting common therapeutic targets and new treatment avenues.
Materials and methods
All the Drosophila stocks were reared on standard cornmeal/agar/molasses food at 25°C within 12-hour light/dark cycling incubators. All animals were reared to the wandering third instar stage for all experiments, with all genotypes and RNAi lines confirmed with a combination of transgenically marked balancer chromosomes, western blots, and sequencing. Due to the corkscrew gene being on the X chromosome, all experiments utilizing csw5 mutants were conducted using males only, whereas all the trans-heterozygous experiments were done using females only. All the other experiments were done on both of the sexes (males and females together). The two genetic background controls were w1118 and the TRiP RNAi third chromosome background control . The dfmr150M null mutant , csw5 null mutant , and the transgenic lines UAS-cswWT and UAS-csw RNAi [25,30] are all available from the Drosophila Bloomington Stock Center (BDSC; Indiana University, Bloomington, IN, USA). The UAS-cswA72S line  was obtained as a kind gift from Dr. Mario Rafael Pagani (Department of Physiology and Biophysics, School of Medicine, National Scientific and Technical Research Council, University of Buenos Aires, Buenos Aires, Argentina). All patient-derived UAS-PTPN11 mutant lines  were obtained as a kind gift from Dr. Tirtha Das (Department of Cell, Developmental, and Regenerative Biology, Icahn School of Medicine at Mount Sinai, New York, NY, USA). Transgenic studies were performed with neural-specific elav-Gal4 , muscle-specific 24B-Gal4 , and ubiquitous daughterless UH1-Gal4  driver lines, all obtained from BDSC. The genetic and transgenic lines used in this study are summarized below in Table 1:
Wandering third instar dissections and TEVC recordings were done at 18°C in physiological saline (in mM): 128 NaCl, 2 KCl, 4 MgCl2, 1.0 CaCl2, 70 sucrose, and 5 HEPES (pH 7.2). Staged larvae were dissected longitudinally along the dorsal midline, the internal organs removed, and the body walls glued down (Vetbond, 3M). Peripheral motor nerves were cut at the base of the VNC. Dissected preparations were imaged with a Zeiss 40× water-immersion objective on a Zeiss Axioskop microscope. Muscle 6 in abdominal segments 3 to 4 was impaled with two intracellular electrodes (1 mm outer diameter borosilicate capillaries; World Precision Instruments, 1B100F-4) of approximately 15 MΩ resistance when filled with 3M KCl. The muscles were clamped at −60 mV using an Axoclamp-2B amplifier (Axon Instruments). For evoked EJC recordings, the motor nerve was stimulated with a fire-polished suction electrode using 0.5 ms suprathreshold voltage stimuli at 0.2 Hz from a Grass S88 stimulator. Nerve stimulation–evoked EJC recordings were filtered at 2 kHz. To quantify EJC amplitude, 10 consecutive traces were averaged, and the average peak value recorded. Spontaneous mEJC recordings were made in continuous 2-minute sessions and low-pass filtered at 200 Hz. Synaptic depression experiments were performed using the above EJC recording protocol for 1 minute to establish baseline, followed by a 20-Hz HFS train for 5 minutes at the same suprathreshold voltage. RRP size was estimated by dividing the cumulative EJC amplitudes during the first 100 responses to 20 Hz stimulation by the mean mEJC amplitudes. Due to these analyses being at 20 Hz, RRP size is likely underestimated. All synaptic plasticity experiments were performed in 0.2 mM Ca2+ using 10 Hz stimulation trains for 1 minute, followed by 0.2 Hz recordings. All EJC responses within a 1-second bin were averaged, and the average value normalized to the basal EJC amplitude for each animal. Clampex 9.0 was used for all data acquisition, and Clampfit 10.6 was used for all data analyses (Axon Instruments).
Two drugs known to inhibit pERK production (Trametinib and Vorinostat) were used by feeding as published previously [12,45,46]. Both Trametinib (Cell Signaling, 62206S) and Vorinostat (Cell Signaling, 12520S) were dissolved in dimethylsulfoxide (DMSO; Fisher, 67-68-5) at 15 mM and 20 mM, respectively, to create stock solutions. Both drugs were then added to Drosophila food yeast paste and in the standard cornmeal/agar/molasses food in the final concentrations of 0.5 mM (Trametinib) and 1 mM (Vorinostat). Drosophila were induced to lay eggs on selection apple juice plates with drugged yeast paste food. Hatching first instars were selected and placed in standard vials containing Trametinib, Vorinostat, or control food with DMSO only. Larvae were reared in a 12-hour light/dark cycling incubators at 25°C and then collected as wandering third instars for TEVC studies.
Wandering third instars (20 larvae) of each genotype (UH1>FMRP-YFP or Tubulin-GFP) were homogenized in 200 μL of RNase-free lysis buffer (20 mM HEPES, 100 mM NaCl, 2.5 mM EDTA, 0.05% (v/v) Triton X-100, 5% (v/v) glycerol) with 1% β-mercaptoethanol 1× protease inhibitor cocktail (complete mini EDTA-free Tablets, Sigma, 11836170001) and 400U RNase inhibitor (Applied Biosystems, N8080119). The supernatant was collected and diluted to 300 μL to reduce nonspecific binding. Next, the samples were incubated with GFP-trap coupled magnetic agarose beads (Chromotek, GTMA20) for 3 hours at 4°C. The bound beads were washed with lysis buffer (3X, 10 minutes). The bound RNA was purified by incubating the bead-protein-RNA conjugates with a 500-μL TRIzol and chloroform mixture (Ambion, 15596026) for 10 minutes at RT, followed by centrifugation. To precipitate RNA, glycogen (1 μL) and 2-propenol (250 μL) were added to the isolated aqueous layer. Finally, the precipitated RNA was reverse transcribed into single-strand cDNA using the SuperScript VILO cDNA synthesis kit (Thermo Fisher, 11754050) and then subjected to primer-specific PCR, with 2% agarose gels used to analyze the PCR products. All primers used in this study are summarized above in Table 2.
The length of PCR products is approximately 200 bp.
Wandering third instar VNCs from 20 larvae were homogenized in 100 μL of lysis buffer (20 mM HEPES, 10 mM EDTA, 100 mM KCl, 0.1% (v/v) Triton X-100, 5% (v/v) glycerol) with a protease inhibitor cocktail (Roche, 04693132001) combined with a protease and phosphatase inhibitor cocktail (Abcam, ab201119). All samples were then sonicated and run in 4% to 15% Mini-PROTEAN TGX Stain-Free Precast Gels (BioRad, 4568083) alongside Precision Plus Protein all blue prestained protein standards (BioRad, 1610373). Next, total protein was transferred to PVDF membranes using a Trans-Blot Turbo system (BioRad). After transfer, the membrane was blocked by TBS intercept blocking buffer (LiCOR, 927–60000) for 1 hour at RT. The blocked membranes were incubated with primary antibodies overnight at 4°C. Antibodies used include rabbit anti-Csw (Lizabeth Perkins, F1088, 1:500) and goat anti-GAPDH (Abcam, ab157157, 1:2,000). The membrane was washed with Tris-buffer saline with 0.1% Tween-20 (TBST) and then incubated with secondary antibodies for 40 minutes at RT. Secondary antibodies used include Alexa Fluor 680 donkey anti-goat (Invitrogen, A21084, 1:10,000) and Alexa Fluor 800 goat anti-rabbit (Invitrogen, A32735, 1:10,000). After washing with TBST (3X, 10 minutes), the membranes were imaged using the Li-COR Odyssey CLx system.
Wandering third instars were dissected in physiological saline (see above) and fixed in 4% paraformaldehyde (EMS, 15714) diluted in PBS (Corning, 46–013-CM) for 10 minutes at RT. Preparations were then washed and permeabilized in PBS containing 0.2% Triton X-100 and 1% bovine serum albumin (BSA; 3X, 10 minutes), followed by blocking for 30 minutes at RT in the same solution. Preparations were incubated with primary antibodies overnight at 4°C. Primary antibodies used included rabbit anti-pERK1/ERK2 (Thr185, Tyr187) polyclonal antibody (Thermo Fisher, 44-680G, 1:100), goat Cy3-conjugated anti-HRP (Jackson ImmunoResearch, 123–165–021, 1:200), and goat 488-conjugated anti-HRP (Jackson ImmunoResearch, 123–545–021, 1:200). Preparations were washed (3X, 10 minutes) and then incubated with secondary antibodies for 2 hours at RT. Secondary antibodies used included: donkey 488 anti-rabbit (Invitrogen, A21206) and donkey 555 anti-rabbit (Invitrogen, A31572). Preparations were washed (3X, 10 minutes) and then mounted in Fluoromount G (Electron Microscopy Sciences) onto 25 × 75 × 1 mm slides (Fisher Scientific, 12–544–2) with a 22 × 22–1 coverslip (Thermo Fisher Scientific, 12–542-B) sealed with clear nail polish (Sally Hansen). All NMJ imaging was performed using a Zeiss LSM 510 META laser-scanning confocal microscope, with images projected in Zen (Zeiss) and analyzed using ImageJ (NIH open source). All NMJ intensity measurements were made with HRP signal-delineated z-stack areas of maximum projection using ImageJ threshold and wand-tracing tools.
All statistics were performed using GraphPad Prism software (v9.0). Data sets were subject to normality tests, with D’Agostino–Pearson tests utilized if n > 10 and Shapiro–Wilk tests if n < 10. With normal data, ROUT outlier tests with Q set to 1% were run, followed by either two-tailed Student t tests for two-way comparison with 95% confidence (2 data sets) or a one-way ANOVA followed by either a Tukey’s multiple comparison test (3+ data sets, comparing all samples) or a Dunnett’s multiple comparison test (3+ data sets, comparing to control). If data were not normal, Mann–Whitney tests (2 data sets) or Kruskal–Wallis followed by a Dunn’s multiple comparisons tests (3+ data sets) were performed. In order to fully capture changes in the datasets for experiments containing time courses, nonlinear regressions were performed followed by F extra sum of squares tests to determine if the curves were significantly different. All figures show all individual data points as well as mean ± SEM, with significance displayed as p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (****), and p > 0.05 (not significant; n.s.).
S1 Fig. NMJ architecture is unchanged in csw null and GoF mutants.
(A) Representative NMJ images of the w1118 genetic background control, csw5 null mutant, UH1-Gal4/w1118 transgenic driver control, and cswA72S GoF mutant (UH1-Gal4>cswA72S) colabeled for presynaptic membrane marker anti-HRP (magenta) and postsynaptic scaffold DLG (green). Scale bar: 10 μm. (B) Quantification of muscle length for all 4 genotypes using two-sided t tests. (C) Quantification of NMJ area for all 4 genotypes using Mann–Whitney tests. (D) Quantification of NMJ branch number for all 4 genotypes using Mann–Whitney tests. (E) Quantification of NMJ synaptic bouton number for all 4 genotypes using Mann–Whitney tests. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.). The data underlying this figure can be found in S1 Data. csw, corkscrew; DLG, Discs Large; GoF, gain-of-function; HRP, horseradish peroxidase; NMJ, neuromuscular junction.
S2 Fig. Synapse number is unchanged in csw null and GoF mutants.
(A) Representative NMJ images of the w1118 genetic background control, csw5 null mutant, UH1-Gal4/w1118 transgenic driver control, and cswA72S GoF mutant (UH1-Gal4>cswA72S) colabeled for presynaptic membrane marker anti-HRP (blue), active zone marker Brp (magenta), and postsynaptic GluRIIC (green). Scale bar: 2.5 μm. (B) Quantification of Brp puncta density for all 4 genotypes using two-sided t test/Mann–Whitney tests. (C) Quantification of GluRIIC puncta density for all 4 genotypes using two-sided t tests. (D) Quantification of the Brp:GluRIIC puncta ratio for all 4 genotypes using two-sided t test/Mann–Whitney tests. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.). The data underlying this figure can be found in S1 Data. Brp, Bruchpilot; csw, corkscrew; GluRIIC, glutamate receptor IIC; GoF, gain-of-function; HRP, horseradish peroxidase; NMJ, neuromuscular junction.
S3 Fig. Wild-type Csw/PTPN11 expression restores neurotransmission in mutants.
(A) Representative EJC traces for the csw5 null mutant rescued via expression of cswWT (csw5 UH1-Gal4>cswWT) and transgenic driver control (UH1-Gal4/w1118) showing 10 superimposed responses (1.0 mM Ca2). (B) Quantification of the mean EJC amplitudes using a two-sided t test. (C) Representative EJC traces for the wild-type PTPN11 (UH1-Gal4>PTPN11WT) and transgenic driver control (UH1-Gal4/w1118) showing 10 superimposed evoked synaptic responses (1.0 mM Ca2). (D) Quantification of the mean EJC amplitudes using a two-sided t test. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.). The data underlying this figure can be found in S1 Data. Csw, corkscrew; EJC, excitatory junction current; NMJ, neuromuscular junction; PTPN11, protein tyrosine phosphatase non-receptor type 11.
S4 Fig. Neuronal csw RNAi knockdown increases spontaneous fusion frequency.
(A) Representative mEJC traces (1.0 mM Ca+2) in driver control (UH1-Gal4/TRiP control, top) and UH1-Gal4>csw RNAi (bottom). (B) Quantification of the mEJC frequency using a two-sided t test. (C) Quantification of mEJC amplitude using a Mann–Whitney test. (D) Representative mEJC traces (1.0 mM Ca+2) in driver control (elav-Gal4/TRiP control, top) and neuronal elav-Gal4>csw RNAi (bottom). (E) Quantification of the mEJC frequency using a Mann–Whitney test. (F) Quantification of the mEJC amplitude using a Mann–Whitney test. (G) Representative mEJC traces (1.0 mM Ca+2) in driver control (24B-Gal4/TRiP, top) and muscle 24B-Gal4>csw RNAi (bottom). (H) Quantification of the mEJC frequency using a two-sided t test. (I) Quantification of the mEJC amplitude using two-sided t test. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.), p < 0.05 (*), and p > 0.001 (***). The data underlying this figure can be found in S1 Data. csw, corkscrew; mEJC, miniature EJC; NMJ, neuromuscular junction; RNAi, RNA interference.
S5 Fig. All csw/PTPN11 mutants exhibit increased synaptic quantal content release.
The quantal content at each NMJ was calculated by dividing the evoked EJC traces by the mean mEJC amplitude. (A) Quantification of the quantal content of both the csw/PTPN11 null and GoF mutants using two-sided t tests. (B) Quantification of the quantal content of csw RNAi ubiquitous (UH1), neuronal (elav), and muscle (24B) lines compared to their matched transgenic driver controls using two-sided t tests. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.), p < 0.05 (*), p > 0.001 (***), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. Csw, corkscrew; EJC, excitatory junction current; GoF, gain-of-function; mEJC, miniature EJC; NMJ, neuromuscular junction; PTPN11, protein tyrosine phosphatase non-receptor type 11; RNAi, RNA interference.
S6 Fig. Neuronal NSML model PTPN11 mutants exhibit elevated presynaptic function.
(A) Representative EJC traces for the transgenic driver control (elav-Gal4/w1118), and PTPN11 patient mutants PTPN11Q510E (elav-Gal4>PTPN11Q510E) and PTPN11Q510P (elav-Gal4>PTPN11Q510P) showing 10 superimposed evoked synaptic responses (1.0 mM Ca2+). (B) Quantification of the mean EJC amplitudes in all 3 genotypes using one-way ANOVA and Tukey’s multiple comparisons. (C) Representative mEJC traces (1.0 mM Ca2+) in above driver control (top), PTPN11Q510E (middle), and PTPN11Q510P (bottom). (D) Quantification of the mEJC frequency using one-way ANOVA and Tukey’s multiple comparisons. (E) Quantification of mEJC amplitude using a Kruskal–Wallis test. (F) Quantification of quantal content using one-way ANOVA and Tukey’s multiple comparisons. Scatter plots show all the individual data points as well as mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.), p < 0.05 (*), p < 0.001 (**), and p > 0.001 (***). The data underlying this figure can be found in S1 Data. EJC, excitatory junction current; mEJC, miniature EJC; NMJ, neuromuscular junction; NSML, NS with multiple lentigines; PTPN11, protein tyrosine phosphatase non-receptor type 11.
S7 Fig. HFS transmission depression ameliorated in csw nulls.
Prolonged HFS at 20 Hz (1 mM Ca+2) drives progressive synaptic amplitude depression over several minutes of continuous recording. (A) Representative evoked nerve-stimulated EJC traces at the basal frequency (t = 0) and indicated time points during the HFS train for the genetic background control (w1118, top) and the csw null mutant (csw5, bottom). (B) Quantification of normalized EJC amplitudes at the indicated time points during the HFS train using two-sided t tests. Scatter plots show all data points and mean ± SEM. N = number of NMJs. Significance: p < 0.05 (*), p < 0.001 (**), and p < 0.001 (***). The data underlying this figure can be found in S1 Data. csw, corkscrew; EJC, excitatory junction current; HFS, high-frequency stimulation; NMJ, neuromuscular junction.
S8 Fig. Reducing ERK signaling restores NS model PTPN11 synaptic function.
TEVC recordings with and without the pERK inhibiting drug Trametinib comparing the driver control (elav-Gal4/w1118) and NS GoF patient mutant (elav-Gal4>PTPN11N308D). (A) Representative EJC traces showing 10 superimposed responses (1.0 mM Ca+2) comparing the control (left) and PTPN11N308D mutant (right), with and without Trametinib. (B) Quantification of mean EJC amplitudes for all 4 conditions using one-way ANOVA and Tukey’s multiple comparisons. Scatter plots show all the data points and the mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.) and p < 0.05 (*). The data underlying this figure can be found in S1 Data. EJC, excitatory junction current; ERK, extracellular signal-regulated kinase; GoF, gain-of-function; NMJ, neuromuscular junction; NS, Noonan syndrome; pERK, phosphorylated ERK; PTPN11, protein tyrosine phosphatase non-receptor type 11; TEVC, two-electrode voltage-clamp.
S9 Fig. PTPN11 LoF and GoF mutants exhibit elevated presynaptic pERK levels.
(A) Representative NMJ images of the driver control (UH1-Gal4/w1118, top left), the GoF mutant (elav>PTPN11N308D; top right), and two LoF mutants (UH1-Gal4>PTPN11Q510E, bottom left, and UH1-Gal4>PTPN11Q510P; bottom right) colabeled for presynaptic membrane marker anti-HRP (magenta) and pERK (green). Scale bar: 2.5 μm. (B) Quantified presynaptic anti-pERK fluorescence for all 5 genotypes using a two sided t test (PTPN11N308D) and one-way ANOVA and Tukey’s multiple comparisons (PTPN11Q510E/ PTPN11Q510P). Scatter plots show all data points and mean ± SEM. N = number of NMJs. Significance: p < 0.001 (**), p > 0.001 (***), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. GoF, gain-of-function; HRP, horseradish peroxidase; LoF, loss-of-function; NMJ, neuromuscular junction; pERK, phosphorylated ERK; PTPN11, protein tyrosine phosphatase non-receptor type 11.
S10 Fig. Trans-heterozygous csw/+; dfmr1/+ recapitulate disease model phenotypes.
(A) Representative evoked EJC traces showing 10 superimposed TEVC recordings in background control (w1118), single heterozygotes (csw5/+ and dfmr150M/+), and the trans-heterozygote (csw5/+; dfmr150M/+). (B) Quantification of mean EJC amplitudes for all 4 genotypes using one-way ANOVA and Dunnett’s multiple comparisons. (C) Representative mEJC traces from the same 4 genotypes. (D) Quantification of mEJC frequency for all 4 genotypes using Kruskal–Wallis and Dunn’s multiple comparisons. (E) Quantification of mEJC amplitude for all 4 genotypes using Kruskal–Wallis. (F) Representative NMJ images from the same 4 genotypes colabeled for anti-pERK (green) and presynaptic membrane anti-HRP (magenta). pERK also shown as a heat map. Scale bar: 2.5 μm. (G) Quantification of normalized synaptic pERK fluorescence for all 4 genotypes using Kruskal–Wallis and Dunn’s multiple comparison tests. Scatter plots show all data points and the mean ± SEM. N = number of NMJs. Significance: p > 0.05 (not significant, n.s.), p < 0.001 (**), p > 0.001 (***), and p < 0.0001 (****). The data underlying this figure can be found in S1 Data. EJC, excitatory junction current; HRP, horseradish peroxidase; mEJC, miniature EJC; NMJ, neuromuscular junction; pERK, phosphorylated ERK; TEVC, two-electrode voltage-clamp.
S1 Data. Excel document detailing raw data for all analyses.
S1 Raw Images. Original uncropped gel and blot of Fig 6.
We are most grateful to Dr. Tirtha Das (Icahn School of Medicine at Mount Sinai, New York, USA) for providing the transgenic UAS-PTPN11 lines, Dr. Mario Rafael Pagani (University of Buenos Aires, Argentina) for the cswA72S line, and Dr. Daniela Zarnescu (University of Arizona, Tucson, Arizona, USA) for the UAS-YFP-dfmr1 line. We thank the Bloomington Drosophila Stock Center (Indiana University, Bloomington, Indiana, USA) for critical genetic lines and the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA, USA) for antibodies. We thank Broadie lab members for their constructive input throughout the course of this study.
- 1. Noonan JA. Hypertelorism with Turner phenotype. A new syndrome with associated congenital heart disease. Am J Dis Child. 1968 Oct;116(4):373–80. pmid:4386970
- 2. Gelb BD, Tartaglia M. Noonan syndrome and related disorders: dysregulated RAS-mitogen activated protein kinase signal transduction. Hum Mol Genet. 2006 Oct 15;15:R220–6 pmid:16987887
- 3. Tartaglia M, Gelb BD, Zenker M. Noonan syndrome and clinically related disorders. Best Pract Res Clin Endocrinol Metab. 2011 Feb;25(1):161–79. pmid:21396583
- 4. Barford D, Neel BG. Revealing mechanisms for SH2 domain mediated regulation of the protein tyrosine phosphatase SHP-2. Structure. 1998 Mar 15;6(3):249–54. pmid:9551546
- 5. Hof P, Pluskey S, Dhe-Paganon S, Eck MJ, Shoelson SE. Crystal Structure of the Tyrosine Phosphatase SHP-2. Cell. 1998 Feb 20;92(4):441–50. pmid:9491886
- 6. Rauen KA. The RASopathies. Annu Rev Genomics Hum Genet. 2013;14:355–369. pmid:23875798
- 7. Johnson EM, Ishak AD, Naylor PE, Stevenson DA, Reiss AL, Green T. PTPN11 Gain-of-Function Mutations Affect the Developing Human Brain, Memory, and Attention. Cereb Cortex 2019 Jul 1;29(7):2915–23. pmid:30059958
- 8. Pierpont EI. Neuropsychological Functioning in Individuals with Noonan Syndrome: a Systematic Literature Review with Educational and Treatment Recommendations. J Pediatr Neuropsychol. 2015 Sep 1;2(1):14–33.
- 9. Oishi K, Gaengel K, Krishnamoorthy S, Kamiya K, Kim IK, Ying H, et al. Transgenic Drosophila models of Noonan syndrome causing PTPN11 gain-of-function mutations. Hum Mol Genet. 2006 Feb 15;15(4):543–53. pmid:16399795
- 10. Oishi K, Zhang H, Gault WJ, Wang CJ, Tan CC, Kim IK, et al. Phosphatase-defective LEOPARD syndrome mutations in PTPN11 gene have gain-of-function effects during Drosophila development. Hum Mol Genet. 2009 Jan 1;18(1):193–201. pmid:18849586
- 11. Pagani MR, Oishi K, Gelb BD, Zhong Y. The Phosphatase SHP2 Regulates the Spacing Effect for Long-Term Memory Induction. Cell. 2009 Oct 2;139(1):186–98. pmid:19804763
- 12. Das TK, Gatto J, Mirmira R, Hourizadeh E, Kaufman D, Gelb BD, et al. Drosophila RASopathy models identify disease subtype differences and biomarkers of drug efficacy. iScience. 2021 Apr 23;24(4):102306. pmid:33855281
- 13. Wang X, Snape M, Klann E, Stone JG, Singh A, Petersen RB, et al. Activation of the extracellular signal-regulated kinase pathway contributes to the behavioral deficit of fragile x-syndrome. J Neurochem. 2012 May;121(4):672–9. pmid:22393900
- 14. Darnell JC, Van Driesche SJ, Zhang C, Ying K, Hung S, Mele A, et al. FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell. 2011 Jul 22;146(2):247–261. pmid:21784246
- 15. Hale CR, Sawicka K, Mora K, Fak J, Kang JJ, Cutrim P, et al. FMRP regulates mRNAs encoding distinct functions in the cell body and dendrites of CA1 pyramidal neurons. Elife. 2021 Dec 23;10:e71892 pmid:34939924
- 16. Razak KA, Dominick KC, Erickson CA. Developmental studies in fragile X syndrome. J Neurodev Disord. 2020 May 2;12(1):13. pmid:32359368
- 17. Zhang YQ, Bailey AM, Matthies HJG, Renden RB, Smith MA, Speese SD, et al. Drosophila fragile x-related gene regulates the MAP1B homolog Futsch to control synaptic structure and function. Cell. 2001 Nov 30;107(5):591–603. pmid:11733059
- 18. Bolduc F V., Bell K, Cox H, Broadie KS, Tully T. Excess protein synthesis in Drosophila Fragile X mutants impairs long-term memory. Nat Neurosci. 2008 Oct;11(10):1143–5. pmid:18776892
- 19. Kushner SA, Elgersma Y, Murphy GG, Jaarsma D, Woerden GM van, Hojjati MR, et al. Modulation of Presynaptic Plasticity and Learning by the H-ras/Extracellular Signal-Regulated Kinase/Synapsin I Signaling Pathway. J Neurosci. 2005 Oct 19;25(42):9721–34. pmid:16237176
- 20. Deng PY, Sojka D, Klyachko VA. Abnormal Presynaptic Short-Term Plasticity and Information Processing in a Mouse Model of Fragile X Syndrome. J Neurosci. 2011 Jul 27;31(30):10971–82. pmid:21795546
- 21. Çaku A, Pellerin D, Bouvier P, Riou E, Corbin F. Effect of lovastatin on behavior in children and adults with fragile X syndrome: an open-label study. Am J Med Genet A. 2014 Nov;164A(11):2834–42. pmid:25258112
- 22. Muscas M, Louros SR, Osterweil EK. Lovastatin, not Simvastatin, Corrects Core Phenotypes in the Fragile X Mouse Model. eNeuro. 2019 May;6(3):ENEURO.0097-19.2019.
- 23. Repicky S, Broadie K. Metabotropic Glutamate Receptor–Mediated Use–Dependent Down-Regulation of Synaptic Excitability Involves the Fragile X Mental Retardation Protein. J Neurophysiol. 2009 Feb;101(2):672–87. pmid:19036865
- 24. Perrimon N, Engstrom L, Mahowald AP. A pupal lethal mutation with a paternally influenced maternal effect on embryonic development in Drosophila melanogaster. Dev Biol. 1985 Aug;110(2):480–91. pmid:3926563
- 25. Johnson Hamlet MR, Perkins LA. Analysis of Corkscrew Signaling in the Drosophila Epidermal Growth Factor Receptor Pathway During Myogenesis. Genetics. 2001;159(3):1073–1087. pmid:11729154
- 26. Van Vactor D, Sigrist SJ. Presynaptic Morphogenesis, Active Zone Organization and Structural Plasticity in Drosophila. Curr Opin Neurobiol. 2017 Apr;43:119–29. pmid:28388491
- 27. Chou VT, Johnson S, Long J, Vounatsos M, van Vactor D. dTACC restricts bouton addition and regulates microtubule organization at the Drosophila neuromuscular junction. Cytoskeleton (Hoboken). 2020 Jan;77(1–2):4–15. pmid:31702858
- 28. Kopke DL, Leahy SN, Vita DJ, Lima SC, Newman ZL, Broadie K. Carrier of wingless (Cow) regulation of drosophila neuromuscular junction development. eNeuro. 2020 Mar;7(2):ENEURO.0285-19.2020. pmid:32024666
- 29. Brink DL, Gilbert M, Xie X, Petley-Ragan L, Auld VJ. Glial Processes at the Drosophila Larval Neuromuscular Junction Match Synaptic Growth. PLoS One. 2012 May 29;7(5):e37876. pmid:22666403
- 30. Ni JQ, Zhou R, Czech B, Liu LP, Holderbaum L, Yang-Zhou D, et al. A genome-scale shRNA resource for transgenic RNAi in Drosophila. Nat Methods. 2011 May;8(5):405–7. pmid:21460824
- 31. Perkins LA, Holderbaum L, Tao R, Hu Y, Sopko R, McCall K, et al. The Transgenic RNAi Project at Harvard Medical School: Resources and Validation. Genetics. 2015;201(3):843–852. pmid:26320097
- 32. Harris KP, Littleton JT. Transmission, Development, and Plasticity of Synapses. Genetics. 2015 Oct;201(2):345–75. pmid:26447126
- 33. Karunanithi S, Marin L, Wong K, Atwood HL. Quantal Size and Variation Determined by Vesicle Size in Normal and Mutant Drosophila Glutamatergic Synapses. J Neurosci. 2002;22(23):10267–10276. pmid:12451127
- 34. Kauwe G, Isacoff EY. Rapid feedback regulation of synaptic efficacy during high-frequency activity at the Drosophila larval neuromuscular junction. Proc Natl Acad Sci U S A. 2013 May 28;110(22):9142–7. pmid:23674684
- 35. Newman ZL, Hoagland A, Aghi K, Worden K, Levy SL, Son JH, et al. Input-Specific Plasticity and Homeostasis at the Drosophila Larval Neuromuscular Junction. Neuron. 2017 Mar 22;93(6):1388–1404.e10. pmid:28285823
- 36. Bhimreddy M, Rushton E, Kopke DL, Broadie K. Secreted C-type lectin regulation of neuromuscular junction synaptic vesicle dynamics modulates coordinated movement. J Cell Sci. 2021 May 1; 134 (9):jcs257592. pmid:33973638
- 37. Dobrunz LE, Stevens CF. Heterogeneity of Release Probability, Facilitation, and Depletion at Central Synapses. Neuron. 1997 Jun;18(6):995–1008. pmid:9208866
- 38. Stevens RJ, Akbergenova Y, Jorquera RA, Littleton JT. Abnormal synaptic vesicle biogenesis in Drosophila synaptogyrin mutants. J Neurosci. 2012 Dec 12;32(50):18054–67. pmid:23238721
- 39. Baccino-Calace M, Schmidt K, Müller M. The E3 ligase Thin controls homeostatic plasticity through neurotransmitter release repression. Elife. 2022 Jul 711:e71437. pmid:35796533
- 40. Vandael D, Borges-Merjane C, Zhang X, Jonas P. Short-Term Plasticity at Hippocampal Mossy Fiber Synapses Is Induced by Natural Activity Patterns and Associated with Vesicle Pool Engram Formation. Neuron. 2020 Aug 5;107(3):509–521.e7. pmid:32492366
- 41. Xue R, Ruhl DA, Briguglio JS, Figueroa AG, Pearce RA, Chapman ER. Doc2-mediated superpriming supports synaptic augmentation. Proc Natl Acad Sci U S A. 2018 Jun 12;115(24):E5605–E5613. pmid:29844163
- 42. Regehr WG. Short-Term Presynaptic Plasticity. Cold Spring Harb Perspect Biol. 2012 Jul 1;4(7):a005702. pmid:22751149
- 43. Zucker RS, Regehr WG. Short-term synaptic plasticity. Annu Rev Physiol. 2002;64:355–405. pmid:11826273
- 44. Kalkstein JM, Magleby KL. Augmentation increases vesicular release probability in the presence of masking depression at the frog neuromuscular junction. J Neurosci. 2004 Dec 15;24(50):11391–403. pmid:15601945
- 45. Smith JA, Mayeux PR, Schnellmann RG. Delayed MEK/ERK Inhibition by Trametinib Attenuates Systemic Inflammatory Responses and Multi-Organ Injury in Murine Sepsis. Crit Care Med. 2016 Aug;44(8):e711–20.
- 46. Zhong HM, Ding QH, Chen WP, Luo RB. Vorinostat, a HDAC inhibitor, showed anti-osteoarthritic activities through inhibition of iNOS and MMP expression, p38 and ERK phosphorylation and blocking NF-κB nuclear translocation. Int Immunopharmacol. 2013 Oct;17(2):329–35.
- 47. Cziko AM, McCann CT, Howlett IC, Barbee SA, Duncan RP, Luedemann R, et al. Genetic Modifiers of dFMR1 Encode RNA Granule Components in Drosophila. Genetics. 2009 Aug;182(4):1051–60. pmid:19487564
- 48. Levi O, Arava Y. mRNA association by aminoacyl tRNA synthetase occurs at a putative anticodon mimic and autoregulates translation in response to tRNA levels. PLoS Biol. 2019 May 17;17(5):e3000274. pmid:31100060
- 49. Giachello CN, Fiumara F, Giacomini C, Corradi A, Milanese C, Ghirardi M, et al. MAPK/Erk-dependent phosphorylation of synapsin mediates formation of functional synapses and short-term homosynaptic plasticity. J Cell Sci. 2010 Mar 15;123(6):881–93. pmid:20159961
- 50. Linke R, Pries R, Könnecke M, Bruchhage KL, Böscke R, Gebhard M, et al. The MEK1/2-ERK1/2 pathway is activated in chronic rhinosinusitis with nasal polyps. Arch Immunol Ther Exp (Warsz). 2014 Jun;62(3):217–29. pmid:24609540
- 51. Vara H, Onofri F, Benfenati F, Sassoè-Pognetto M, Giustetto M. ERK activation in axonal varicosities modulates presynaptic plasticity in the CA3 region of the hippocampus through synapsin I. Proc Natl Acad Sci U S A. 2009 Jun 16;106(24):9872–7. pmid:19487674
- 52. Engholm-Keller K, Waardenberg AJ, Müller JA, Wark JR, Fernando RN, Arthur JW, et al. The temporal profile of activity-dependent presynaptic phospho-signalling reveals long-lasting patterns of poststimulus regulation. PLoS Biol. 2019 Mar;17(3):e3000170. pmid:30822303
- 53. Ramaswami M, Krishnan KS, Kelly RB. Intermediates in synaptic vesicle recycling revealed by optical imaging of Drosophila neuromuscular junctions. Neuron. 1994 Aug;13(2):363–75. pmid:8060617
- 54. Kopke DL, Broadie K. FM dye cycling at the synapse: Comparing high potassium depolarization, electrical and channelrhodopsin stimulation. J Vis Exp 2018 May 24135:57765. pmid:29889207
- 55. Medina JH, Viola H. ERK1/2: A Key Cellular Component for the Formation, Retrieval, Reconsolidation and Persistence of Memory. Front Mol Neurosci. 2018 Oct 5;11:361. pmid:30344477
- 56. Cargnello M, Roux PP. Activation and Function of the MAPKs and Their Substrates, the MAPK-Activated Protein Kinases. Microbiol Mol Biol Rev. 2011 Mar;75(1):50–83. pmid:21372320
- 57. Plotnikov A, Zehorai E, Procaccia S, Seger R. The MAPK cascades: Signaling components, nuclear roles and mechanisms of nuclear translocation. Biochim Biophys Acta. 2011 Sept;1813(9):1619–33. pmid:21167873
- 58. Kim EK, Choi EJ. Pathological roles of MAPK signaling pathways in human diseases. Biochimica et Biophysica Acta. 2010 Apr 1;1802(4):396–405. pmid:20079433
- 59. Sweatt JD. Mitogen-activated protein kinases in synaptic plasticity and memory. Curr Opin Neurobiol. 2004 Jun;14(3):311–7. pmid:15194111
- 60. Corvol JC, Valjent E, Toutant M, Enslen H, Irinopoulou T, Lev S, et al. Depolarization Activates ERK and Proline-rich Tyrosine Kinase 2 (PYK2) Independently in Different Cellular Compartments in Hippocampal Slices. J Biol Chem. 2005 Jan 7;280(1):660–8. pmid:15537634
- 61. Boggio EM, Putignano E, Sassoè-Pognetto M, Pizzorusso T, Giustetto M. Visual stimulation activates ERK in synaptic and somatic compartments of rat cortical neurons with parallel kinetics. PLoS ONE. 2007 Jul 11;2(7):e604. pmid:17622349
- 62. Jackman SL, Regehr WG. The Mechanisms and Functions of Synaptic Facilitation. Neuron. 2017 May 3;94(3):447–64. pmid:28472650
- 63. Cheng Q, Song SH, Augustine GJ. Molecular Mechanisms of Short-Term Plasticity: Role of Synapsin Phosphorylation in Augmentation and Potentiation of Spontaneous Glutamate Release. Front Synaptic Neurosci. 2018 Oct 30;10(1):33. pmid:30425632
- 64. Liu L, Mayes PA, Eastman S, Shi H, Yadavilli S, Zhang T, et al. The BRAF and MEK Inhibitors Dabrafenib and Trametinib: Effects on Immune Function and in Combination with Immunomodulatory Antibodies Targeting PD-1, PD-L1, and CTLA-4. Clin Cancer Res. 2015 Apr 1;21(7):1639–51. pmid:25589619
- 65. Richter JD, Zhao X. The molecular biology of FMRP: new insights into fragile X syndrome. Nat Rev Neurosci. 2021 Feb 19;22(4):209–22. pmid:33608673
- 66. Darnell JC, Klann E. The translation of translational control by FMRP: therapeutic targets for FXS. Nat Neurosci. 2013 Nov;16(11):1530–6. pmid:23584741
- 67. Kidd T, Bland KS, Goodman CS. Slit is the midline repellent for the robo receptor in Drosophila. Cell. 1999 Mar 19;96(6):785–94. pmid:10102267
- 68. Bagni C, Zukin RS. A Synaptic Perspective of Fragile X Syndrome and Autism Spectrum Disorders. Neuron. 2019 Mar 20;101(6):1070–88. pmid:30897358
- 69. Casingal CR, Kikkawa T, Inada H, Sasaki Y, Osumi N. Identification of FMRP target mRNAs in the developmental brain: FMRP might coordinate Ras/MAPK, Wnt/β-catenin, and mTOR signaling during corticogenesis. Mol Brain. 2020 Dec 16;13(1):167.
- 70. Greenblatt EJ, Spradling AC. Fragile X mental retardation 1 gene enhances the translation of large autism-related proteins. Science. 2018 Aug 17;361(6403):709–12. pmid:30115809
- 71. Tsang B, Arsenault J, Vernon RM, Lin H, Sonenberg N, Wang LY, et al. Phosphoregulated FMRP phase separation models activity-dependent translation through bidirectional control of mRNA granule formation. Proc Natl Acad Sci U S A. 2019 Mar 5;116(10):4218–27. pmid:30765518
- 72. Napoli I, Mercaldo V, Boyl PP, Eleuteri B, Zalfa F, de Rubeis S, et al. The Fragile X Syndrome Protein Represses Activity-Dependent Translation through CYFIP1, a New 4E-BP. Cell. 2008 Sep 19;134(6):1042–54. pmid:18805096
- 73. Parkinson WM, Dookwah M, Dear ML, Gatto CL, Aoki K, Tiemeyer M, et al. Synaptic roles for phosphomannomutase type 2 in a new Drosophila congenital disorder of glycosylation disease model. Dis Model Mech. 2016 May 1;9(5):513–27. pmid:26940433
- 74. Dani N, Zhu H, Broadie K. Two Protein N-Acetylgalactosaminyl Transferases Regulate Synaptic Plasticity by Activity-Dependent Regulation of Integrin Signaling. J Neurosci. 2014 Sep 24;34(39):13047–65. pmid:25253852
- 75. Dear ML, Dani N, Parkinson W, Zhou S, Broadie K. Two classes of matrix metalloproteinases reciprocally regulate synaptogenesis. Development. 2016 Jan 1;143(1):75–87. pmid:26603384
- 76. Osterwalder T, Yoon KS, White BH, Keshishian H. A conditional tissue-specific transgene expression system using inducible GAL4. Proc Natl Acad Sci U S A. 2001 Oct 23;98(22):12596–601. pmid:11675495
- 77. Barwell T, DeVeale B, Poirier L, Zheng J, Seroude F, Seroude L. Regulating the UAS/GAL4 system in adult Drosophila with Tet-off GAL80 transgenes. PeerJ. 2017 Dec 14;5(12):e4167. pmid:29259847
- 78. Kopke DL, Lima SC, Alexandre C, Broadie K. Notum coordinates synapse development via extracellular regulation of wingless trans-synaptic signaling. Development. 2017 Oct 1;144(19):3499–510. pmid:28860114