key: cord-0282320-6qaehskw
authors: Xu, Yun-Fei; Chen, Xu; Yang, Zhao; Xiao, Peng; Liu, Chun-Hua; Li, Kang-Shuai; Yang, Xiao-Zhen; Wang, Yi-Jing; Zhu, Zhong-Liang; Xu, Zhi-Gang; Zhang, Sheng; Wang, Chuan; Song, You-Chen; Zhao, Wei-Dong; Wang, Chang-He; Ji, Zhi-Liang; Zhang, Zhong-Yin; Cui, Min; Sun, Jin-Peng; Yu, Xiao
title: PTP-MEG2 regulates quantal size and fusion pore opening through two distinct structural bases and substrates
date: 2020-09-17
journal: bioRxiv
DOI: 10.1101/822031
sha: 1406e72225f9b3b6ee4e81fefb516f0990a94fc3
doc_id: 282320
cord_uid: 6qaehskw

Tyrosine phosphorylation of secretion machinery proteins is a crucial regulatory mechanism for exocytosis. However, the participation of protein tyrosine phosphatases (PTPs) in different exocytosis stages has not been defined. Here we demonstrated that PTP-MEG2 controls multiple steps of catecholamine secretion. Biochemical and crystallographic analyses revealed key residues that the interactions between govern the PTP-MEG2 and NSF-pY83 site, specify PTP-MEG2 substrate selectivity and modulate the fusion of catecholamine-containing vesicles. Unexpectedly, delineation of PTP-MEG2 mutants along with the NSF binding interface revealed that PTP-MEG2 controls the fusion pore opening through non-NSF dependent mechanisms. Utilizing bioinformatics search and biochemical and electrochemical screening approaches, we discovered that PTP-MEG2 regulates the opening and extension of the fusion pore by dephosphorylating the DYNAMIN2-pY125 and MUNC18-1-pY145 site. Further structural and biochemical analysis confirmed the interaction of PTP-MEG2 with MUNC18-1-pY145 or DYNAMIN2-pY125 through a distinct structural basis compared with that of the NSF-pY83 site. Our studies extended mechanistic insights in complex exocytosis processes. HIGHLIGHTS PTP-MEG2 regulates multiple steps of exocytosis. A crystal structure of the PTP-MEG2/phosphor-NSF-pY83 segment was obtained. Functional delineation of the PTP-MEG2/NSF interface led to the discovery of new PTP-MEG2 substrates. PTP-MEG2 regulates fusion pore opening and extension through the DYNAMIN2-pY125 site and MUNC18-1 pY145 site. The distinct structural basis of the recognition of substrates by PTP-MEG2 allows selective inhibitor design.

MEG2 wild type or mutants than the endogenous PTP-MEG2 were selected for electrochemical 295 studies (Supplemental Fig. 6E-F) . The PTP-MEG2 mutations did not significantly affect its 296 interaction with NSF, suggesting that residues not located within the active site of PTP-MEG2 297 may also participate in NSF associations (Supplemental Fig. 6G-6H The effects of PTP-MEG2 mutations along the PTP-MEG2/NSF-phospho-segment 328 interface on catecholamine secretion indicated that a PTP-MEG2 substrate other than NSF with 329 distinct sequence characteristics contributes to the regulation of "foot probability" (Fig. 5) . We 330 therefore utilized this key information to search for new potential PTP-MEG2 substrates by 331 bioinformatics methods (Fig. 6A) . First, we searched for the keywords "fusion pore", "secretory vesicle" and "tyrosine phosphorylation" using the functional protein association network 333 STRING and the text mining tool PubTator, which resulted in a candidate list of 54 proteins. 334 Second, we applied UniProt by selecting proteins located only in the membrane or vesicle, 335 which limited the candidates to 28 members. Third, as our experiments were carried out in the 336 adrenal gland, we used the Human Protein Atlas database for filtering to exclude the proteins 337 which is not detectable in adrenal gland, which narrowed the candidate list to 23 proteins. 338 Finally, we exploited the post-translational-motif database PhosphoSitePlus to screen candidate 339 proteins with potential phospho-sites that matched the sequence requirements at both the pY+1 MUNC18-1, VAMP7, DYNAMIN2 and SNAP25 were readily detected in the adrenal medulla, 353 DYNAMIN1 and PACSIN1 showed substantially lower expression than that in the liver and 354 brain (Supplemental Fig. 8C ). Moreover, whereas MUNC18-1, VAMP7 and DYNAMIN2 355 strongly co-localized with PTP-MEG2 in the adrenal medulla ( Fig. 6D-F) , the co-localization 356 of SNAP25 and PACSIN1 with PTP-MEG2 was relatively weak (Supplemental Fig. 8D-E) . 357 Therefore, MUNC18-1, VAMP7 and DYNAMIN2 are more likely candidate PTP-MEG2 358 substrates which were further strengthened by the fact that the high potassium chloride-or 359 AngII-stimulated tyrosine phosphorylation of these proteins in the adrenal medulla was 360 significantly dephosphorylated by PTP-MEG2 in vitro (Supplemental Fig. 9A -C, E), whereas 361 the tyrosine phosphorylation of SNAP25 had no obvious change (Supplemental Fig. 9D-E) . Moreover, the phenolic-oxygen of Y 145 forms specific hydrogen bonds with the main chain 371 amide of F 540 and the main chain carbonyl oxygens of I 539 and G 568 (Fig. 7B) . These key 372 interactions might be involved in regulating the arrangement of the arc shape of the three 373 domains of MUNC18-1, by tethering the interface between domain 1 and domain 2. The 374 phosphorylation of Y 145 likely abolishes this H-bond network and changes its ability to associate 375 with different snare complexes participating in vesicle fusion procedures (Fig. 7B) . 376

Interestingly, a missense mutation of MUNC18-1 Y 145 H was found to be associated with early 377 infantile epileptic encephalopathy (Stamberger et al., 2017) (Fig. 7C) Fig. 9F-H) . These results suggested that pY 145 is one of the major sites of 387 MUNC18-1 regulated by PTP-MEG2 (Fig. 7D) . Moreover, the effects of the muations of 388 MUNC18-1 and DYNAMIN2 on protein stability were detected in response to CHX treatments. 389

Consequently, the mutations of MUNC18-1 and DYNAMIN2 had no significant influence on 390 the stability (Supplemental Fig. 9I To further dissect the mechanism underlying the phosphorylation of MUNC18-1 Y 145 as 405 well as the disease-related mutant Y 145 H in the regulation of hormone secretion, we compared 406 the interactions of wild-type and mutant MUNC18-1 with the binding partner SYNTAXIN1 407 (Lim et al., 2013) . Importantly, both the phosphorylation mimic mutant MUNC18-1-Y 145 E and 408 the disease-related mutant Y 145 H significantly impaired the interaction between MUNC18-1 409 and SYNTAXIN1 (Fig. 7H , Supplemental Figure 11C ). The effects of MUNC18-1 phospho-410 mimic Y 145 E or phospho-defective Y 145 F mutants on fusion pore dynamics are correlated with 411 the data of their interaction with the SYNTAXIN1, thus indicating that the the binding between AngII stimulation and participated in the AngII-induced catecholamine secretion. The complex structure not only revealed the structural rearrangement in PTP-MEG2 in 513 response to binding of the substrate NSF, identifying Q 559 :D 335 as the key pair for substrate 514 specificity of the pY+1 site, but also provided clues that PTP-MEG2 regulated the initial 515 opening of the fusion pore through another unknown substrate. Fortunately, we were able to 516 deduce the signature of the pY+1 and pY+2 positions of this unknown substrate by carefully 517 inspecting the PTP-MEG2/phospho-NSF-E 79 -pY 83 -K 87 complex structure and analysing the 518 functional data of the PTP-MEG2 interface mutants. Further bioinformatics studies and cellular 519 and physiological experiments enabled us to discover that PTP-MEG2 regulates the initial 520 opening of the fusion pore by modulating the tyrosine phosphorylation states of the MUNC18-521 1 at the pY 145 site and the DYNAMIN2 at the pY 125 site. Therefore, we have revealed that PTP-522 MEG2 regulates different steps of the exocytosis processes via dephosphorylating distinct 523 substrates. PTP-MEG2 regulates the vesicle size and vesicle-vesicle fusion step by 524 dephosphorylating NSF at its NSF-pY 83 site, whereas it regulates the process of LDCV fusion 525 pore initiation and expansion by controlling specific phosphorylation sites of MUNC18-1 and 526 DYNAMIN2. Moreover, our studies highlight that the combination of structural determination 527 and functional delineation of the interface mutants of the protein complex is a powerful 528 approach to characterizing the signalling events and identifying unknown downstream 529 signalling molecules. Moreover, the dephosphorylation of MUNC18-1-Y 145 was suggested to be essential in 563 maintaining the association between MUNC18-1 and SYNTAXIN1 (Lim et al., 2013) . In the 564 present study, we demonstrated that the MUNC18-1 Y 145 E phospho-mimic mutation, but not 565 the non-phosphorylated mutation Y 145 F, significantly decreased the PSF and the SAF 566 probability in cultured primary chromaffin cells. Notably, either the Y 145 E phospho-mimic 567 mutation or the epileptic encephalopathy associated Y 145 H mutant disrupted their interactions 568 with SYNTAXIN1. Structural inspection also suggested that both phosphorylation of Y 145 and 569 Y 145 H mutant could destabilize the arc shape of native MUNC18-1. Therefore, it's very likely 570 that either the association of the MUNC18-1 with the SYNTAXIN1 or the maintenance of the 571 arc shape of the MUNC18-1 actively participated in initial pore opening and expansion. Future 572 studies by solving the fusion machinery structures encompassing with MUNC18-1 at different 573 stages with high resolutions, as well as more detailed fusion pore dynamics analysis using in 574 vitro reconstitution system could provide deeper insights for these key events in proe fusion 575 processes. In the present study, the structural analysis of PTP-MEG2 in complex with 576 MUNC18-1-pY 145 and the enzymatic analysis confirmed that PTP-MEG2 regulated two 577 different substrates, the MUNC18-1-pY 145 and the DYNAMIN2-pY 125 with similar structural 578 features and generated similar effects on fusion pore dynamics. Therefore, our studies 579 exemplified how a PTPase regulated one important physiological process through two different 580 substrates, adding new information of the fusion pore regulation from another aspect. 581

Notably, the MUNC18-1 Y 145 H mutation is a known SNP that is associated with epileptic The design of such inhibitors will certainly help to delineate specific roles of PTP-MEG2 in 607 different physiological and pathological processes. 608

In conclusion, we have found that PTP-MEG2 regulates two different processes of 609 exocytosis during catecholamine secretion, namely, vesicle fusion and the opening and 610 extension of the fusion pore, through two different substrates with distinct structural bases. We 

Further information and requests for resources and reagents should be directed to and will be 626 The standard external solution for our amperometry measurement is as follows: 5 mM KCl, 10 693 mM glucose, 10 mM HEPES pH 7.4, 2 mM CaCl2, 150 mM NaCl and 2 mM MgCl2. We 694 analyzed all data using Igor (WaveMetrix, Lake Oswego, Oregon) and a custom-made macro 695 program. Statistical data were given as the mean ± SEM and analyzed with one-way ANOVA. 696 697

The female mice (6-8 weeks) were decapitated, and the adrenal medullas were freshly isolated 699 and cut to 150-μm-thick sections. The sections were immersed in Ringer's saline (125mM NaCl, 700 2.5mM KCl, 1.25mM NaH2PO4, 26mM NaHCO3, 10mM D-Glucose, 2mM CaCl2, 1mM MgCl2) 701

for 40 minutes at room temperature. During this period, continuous gases of 5% CO2 and 95% 702 O2 were offered to the saline to ensure the survival of the tissue slice. After 40 minutes of 703 starvation, the sections were stimulated with different conditions (control; only 100nM Ang Ⅱ 704 agonists for 1 min; only 400nM PTP-MEG2 inhibitor for 45 min; 100nM AngII agonists and 705 400nM PTP-MEG2 inhibitor for 1 min or 45 min) at 37°C respectively. These sections were 706 firstly immersed in precooled 3% glutaraldehyde and fixed at 4℃ for 2 hours, and then rinsed 707 in PBS isotonic buffer, with repeated liquid exchanges and cleaning overnight, so that the 708 samples were thoroughly rinsed and soaked in the buffer. After rinsing, the sample was fixed at 709 4℃ with 1% osmium acid for 2 hours. It was rinsed with isosmotic buffer solution at 0.1M PBS 710 for 15 minutes. The sections were dehydrated with ethanol at concentrations of 50%, 70%, 90%, 711 then ethanol at concentration of 90% and acetone at concentration of 90%, at last only acetone 712 at concentrations of 90%, 100%. We then replaced the acetone with the Epon gradually. The 713 sections were added to Epon and polymerized at 60°C for 36 hours. Ultra-thin sections were 714 performed at the thickness of 60nm by the LKB-1 ultra microtome, and then the ultra-thin 715 sections were collected with the single-hole copper ring attached with formvar film. The sample 716 was stained with 2% uranium acetate for 30 minutes, and then stained with 0.5% lead citrate 717 for 15 minutes. These prepared samples were examined by JEM-1200EX electron microscope 718 (Japan). 

For the acquisition of tissue cells used in immunofluorescence, the mice were decapitated, and 732 the adrenal medullas were freshly isolated (female mice, 6-8 weeks).The isolated adrenal 733 medullas was immersed in 4% paraformaldehyde for fixation overnight at 4°C. Then the fixed 734 tissues were washed for 4 hours in PBS containing 10% sucrose at 4°C for 8 hours in 20% 735 sucrose, and in 30% sucrose overnight. Then these adrenal medullas were imbedded in Tissue-736

Tek OCT compound and then mounted and frozen them at -25°C. Subsequently, the adrenal 737 medulla was cut to 4-μm-thick coronal serial sections. The adrenal medullas sections were 738 blocked with 1% (vol/vol) donkey serum, 2.5% (wt/vol) BSA and 0.1% (vol/vol) Triton X-100 739 in PBS for 1.5 h. Then, the slides were incubated with primary antibodies against PTP-MEG2 740 by adding 120µl 1M NaOH, and the enzymatic activity was monitored by measuring the 757 absorbance at 405 nm. The activities toward phospho-peptide segment derived from NSF or 758 MUNC18-1 were measured as following: In the first column, 90μl diluted NSF/MUNC18-759 1/DYNAMIN2 phospho-peptide substrate (100μM) was added. The successive columns were 760 diluted by 1.5 times. The phospho-peptide substrates were preincubated at 37°C for 5 min. The pull-down experiment of MUNC18-1 and SYNTAXIN1 were described similar to above 787 description. The GST beads were washed five times by cold binding buffer (described as above). 788

After that, 5µg purified GST-MUNC18-1-WT or its Y 145 H, Y 145 E or Y 145 F mutant proteins were 789 added into 20μl GST-agarose and incubated at for 4 °C 2 hours with end to end rotation. PC12 790 cells transfected with FLAG-SYNTAXIN1 were lysed in lysis buffer (described as above) and 791 centrifuged to remove the pallets. The supernatants were added with 20μl GST beads/GST-792 fusion protein and the mixtures were subjected to end-to-end rotation at 4 °C for 2 hours. The 793 FLAG-SYNTAXIN1 was detected with FLAG antibody. 

The wide type and mutant proteins of His-tagged PTP-MEG2-catalytic domain were expressed 826 in BL21-DE3 Escherichia coli as previously described (Pan et al., 2013) . In brief, 0.4 mM 827 isopropyl1-thio-D-galactopyranoside (IPTG) was used to induce the expression of His-PTP-828 MEG2, and the bacteria lysates were centrifuged at 12000 rpm for 1 hour. Ni-NTA Agarose 829 was applied to bind the His-tagged PTP-MEG2 and an imidazole gradation was used to elute 830 the binding proteins. His-PTP-MEG2 was then purified with gel filtration chromatography to 831 achieve at least 95% purity. The wide type and mutant proteins of GST-PTP-MEG2 and GST-832 MUNC18-1 were also expression in E.coli in presence of 0.4 mM IPTG for 16 hours at 25℃. 833

After centrifugation and lysis, the proteins were purified by binding with GST-Sepharose for 2 834 hours and eluted by GSH. Table 1 . 854

The BRET experiment was performed to monitor the mutation effects of DYNAMIN2 on 857 AT1aR endocytosis in response to AngII (1μM) stimulation as previously described (Liu et We measured the DYNAMIN2 GTPase activity by using GTPase-Glo assay test kit (V7681) 874 from Promega Corporation according to its instructions. 875 876

We identified the PTP-protein interactions by two independent bioinformatic analyses. On one 878 site, we extracted the potential PTP-interacting proteins from the STRING database 879 Rat adrenal medullas were isolated and cut into pieces in D-hanks buffer and stimulated with 894 100nM AngII for 5 minutes. The tissues were then lysed and grinded on ice in lysis buffer 895 supplemented with protease inhibitors. The lysates were centrifuged at 12000 rpm for 20 896 minutes at 4 ℃ and the pallets were removed. Before incubation with the lysates in the later 897 step, the GST beads were washed five times by cold binding buffer (20mM HEPES pH7.5, 898 1mM DTT, 1mM EDTA, and 100mM NaCl) and incubated with 5µg purified GST-PTP-MEG2-899 CD-D 470 A protein for 2 hours at 4°C. 20μl GST beads-PTP-MEG2-D 470 A protein was added 900 into 500μl supernatants and the mixtures were subjected to end-to-end rotation at 4 °C for 2 901 hours. The GST beads and their binding proteins were washed five times with cold binding 902 buffer to exclude the unspecific binding proteins. Denatured proteins were separated by 10% 903 SDS-PAGE and subjected to trypsin digestion. Phosphopeptides were analyzed by the LTQ 904

Orbitrap Elite (Thermo Scientific, Beijing Qinglian Biotech Co., Ltd). Spectra were analyzed 905 by Proteome Discoverer software, and phosphorylation sites were confirmed manually. 

All data are presented as mean ± SEM. All data were analyzed using two-tailed Student's t-test 916 or one-way ANOVA. All of the Western films were scanned, and band intensity was quantified 917 with ImageJ software (National Institutes of Health, Bethesda MD). P < 0.05 was considered 918 as statistically significant. 

Activation of GPCRs modulates quantal size in chromaffin cells through G(betagamma) and PKC

Identification of the tyrosine phosphatase PTP-MEG2 as an antagonist of 991 hepatic insulin signaling

Munc18-1 redistributes in nerve terminals in an activity-and PKC-dependent 994 manner

The control of release probability at nerve terminals

PEST control": regulation of molecular barcodes by tyrosine 998 phosphatases

Actin remodeling to facilitate membrane fusion. Biochimica et biophysica acta 1641

Control of fusion pore dynamics during exocytosis by 1002

Phosphorylation cycling of Annexin A2 Tyr23 is critical for calcium-regulated exocytosis in 1005 neuroendocrine cells

Recent new insights into the role of SNARE and associated proteins in insulin 1007 granule exocytosis

Cloning and expression of a cytosolic megakaryocyte 1009 protein-tyrosine-phosphatase with sequence homology to retinaldehyde-binding protein and yeast 1010

Munc18-1: sequential interactions with the fusion machinery stimulate vesicle 1013 docking and priming

Identification of muscarinic 1015 receptor subtypes involved in catecholamine secretion in adrenal medullary chromaffin cells by genetic 1016 deletion

Munc13-1 and Munc18-1 together prevent NSF-dependent de-priming of synaptic vesicles

Membrane fusion intermediates via directional and full assembly of the SNARE complex

Tethering the assembly of SNARE complexes

2014: mutations, PTMs and recalibrations

Possible roles for Munc18-1 domain 3a and Syntaxin1 N-peptide and C-terminal anchor in SNARE 1028 complex formation

Control of vesicle fusion by a tyrosine phosphatase

Munc18c phosphorylation by the insulin receptor links cell signaling directly to 1034 SNARE exocytosis

Dynamin-2 Stabilizes the HIV-1 Fusion Pore with a Low 1037

Oligomeric State

Regulation of dynamin family proteins by post-translational 1039 modifications

The role 1041 of Munc18-1 in docking and exocytosis of peptide hormone vesicles in the anterior pituitary

SNARE 1044 phosphorylation: a control mechanism for insulin-stimulated glucose transport and other regulated 1045 exocytic events

Crystal Structure and Substrate Specificity of PTPN12

PTPRT regulates the interaction of Syntaxin-binding 1049 protein 1 with Syntaxin 1 through dephosphorylation of specific tyrosine residue

Arrestin-biased AT1R agonism induces acute catecholamine secretion through TRPC3 coupling

Reconstitution of the vital functions of Munc18 and 1055

Munc13 in neurotransmitter release

Munc18-1-regulated stage-1057 wise SNARE assembly underlying synaptic exocytosis

Secretion of Mast Cell

Inflammatory Mediators Is Enhanced by CADM1-Dependent Adhesion to Sensory Neurons. Front Cell 1060 Neurosci 13

Phosphorylation and nitration of tyrosine residues affect functional properties of Synaptophysin and 1063

Dynamin I, two proteins involved in exo-endocytosis of synaptic vesicles

A hemi-fission intermediate links two mechanistically 1067 distinct stages of membrane fission

Tyrosine phosphorylation of Munc18-1 inhibits synaptic transmission by preventing 1070 SNARE assembly

beta-Cell Receptor Tyrosine Kinases in Controlling Insulin Secretion and 1074

Exocytotic Machinery: c-Kit and Insulin Receptor

Protein phosphatases in pancreatic 1076 islets

Cadmium is a potent inhibitor of PPM phosphatases and targets the M1 binding site

Roles of cAMP signalling in insulin granule 1081 exocytosis

Visualization of Membrane Pore in Live Cells Reveals a Dynamic-Pore Theory Governing 1084 Fusion and Endocytosis

Autoinhibition of Munc18-1 modulates synaptobrevin binding and helps to enable Munc13-1087 dependent regulation of membrane fusion

Regulated vesicle fusion generates signaling nanoterritories that control T cell 1090 activation at the immunological synapse

STXBP1 as a therapeutic target for epileptic 1092 encephalopathy

The synaptic vesicle cycle

Neurotransmitter release: the last millisecond in the life of a synaptic vesicle

Membrane fusion: grappling with SNARE and SM proteins

STRING v10: protein-protein interaction networks, integrated 1100 over the tree of life

Bidirectional modulation of exocytosis by angiotensin II 1102 involves multiple G-protein-regulated transduction pathways in chromaffin cells

A subcellular map of the human proteome

Protein tyrosine phosphatases: from genes, to function, to disease

Tyrosine phosphorylation of Munc18c regulates platelet-derived 1110 growth factor-stimulated glucose transporter 4 translocation in 3T3L1 adipocytes

Medicinal chemistry and the molecular operating environment 1113 (MOE): application of QSAR and molecular docking to drug discovery

The catalytic region and PEST domain of PTPN18 distinctly regulate the HER2 phosphorylation 1116 and ubiquitination barcodes

Exocytosis and endocytosis: modes, functions, 1123 and coupling mechanisms

Syntaxin opening by the MUN domain underlies the function of Munc13 in synaptic-vesicle priming

Substrate specificity of lymphoid-specific tyrosine phosphatase (Lyp) and identification of Src 1129 kinase-associated protein of 55 kDa homolog (SKAP-HOM) as a Lyp substrate

Regulatory Mechanisms and Novel Therapeutic Targeting Strategies for 1132

Protein Tyrosine Phosphatases

The Protein Tyrosine Phosphatase MEG2 Regulates the Transport and Signal Transduction of 1135

Tropomyosin Receptor Kinase A

Gq activity-and beta-arrestin-1 scaffolding-mediated ADGRG2/CFTR coupling are required for 1138 male fertility

A highly selective and potent PTP-MEG2 inhibitor with therapeutic potential for type 2 diabetes

Hemi-fused structure mediates and controls fusion and fission in live cells

Rapid fluctuations in transmitter release from single vesicles

AngII and without Compound 7. # indicates P<0.05 and ### indicates P<0.001. All the data 1189 were analyzed using one-way ANOVA. towards the NSF-pY 83 phospho-segment compared with wild-type PTP-MEG2. The red 1232 column indicates that the mutation sites have greater effects on the enzyme activities toward 1233 phospho-NSF segement than the PNPP. The green column indicates that the mutation sites only 1234 have significant effects on the enzyme activity toward NSF phospho-segement but not pNPP. 1235The white column stands for the mutants that have similar effects on pNPP and NSF phospho- residues of NSF and MUNC18-1 are coloured as follows: green, residues interacting with both 1367 NSF and MUNC18-1; red, residues specifically contributing to NSF recognition; blue, residues 1368 selectively contributing to MUNC18-1 interaction.