Myosin Va interacts with the exosomal protein spermine synthase

Myosin Va (MyoVa) is an actin-based molecular motor that plays key roles in the final stages of secretory pathways, including neurotransmitter release. Several studies have addressed how MyoVa coordinates the trafficking of secretory vesicles, but why this molecular motor is found in exosomes is still unclear. In this work, using a yeast two-hybrid screening system, we identified the direct interaction between the globular tail domain (GTD) of MyoVa and four protein components of exosomes: the WD repeat-containing protein 48 (WDR48), the cold shock domain-containing protein E1 (CSDE1), the tandem C2 domain-containing protein 1 (TC2N), and the enzyme spermine synthase (SMS). The interaction between the GTD of MyoVa and SMS was further validated in vitro and displayed a Kd in the low micromolar range (3.5 ± 0.5 µM). SMS localized together with MyoVa in cytoplasmic vesicles of breast cancer MCF-7 and neuroblastoma SH-SY5Y cell lines, known to produce exosomes. Moreover, MYO5A knockdown decreased the expression of SMS gene and rendered the distribution of SMS protein diffuse, supporting a role for MyoVa in SMS expression and targeting.


Introduction
Class V myosins are processive motors that transport or tether vesicles, organelles, and macromolecules to actin filaments, and play key roles in synaptic transmission, hormone secretion, and plasma membrane homeostasis [1][2][3][4][5][6]. They can be divided into four major structural domains with specific functions: the motor domain, which contains the actin-binding site and displays ATPase activity, the lever arm that provides a large powerstroke to occur after ATP hydrolysis, the rod region, responsible for dimerization, and the globular tail domain (GTD), a protein-binding module that allows multiple roles for these molecular motors [3,4,[7][8][9][10]. In humans, three paralogous genes encode for class V myosins: MYO5A, MYO5B, and MYO5C. Defects in the MYO5A gene are related to the Griscelli syndrome type 1 (also known as Elejalde syndrome) [11,12], which is characterized by partial albinism and severe neurological disorders. The molecular mechanism behind the partial albinism involves defects on a tripartite complex between myosin Va (MyoVa), melanophilin, and Rab27a for melanosomes transport [13,14]. However, for the pleiotropic effects of MYO5A mutation in neurodevelopment, the mechanisms are still poorly understood [15].
In neuronal cells, MyoVa has been associated with organelle transport, mRNA trafficking and exocytosis of secretory vesicles [1,2]. Particularly in exocytosis, MyoVa seems to play several roles, including the capture and transport of the secretory granules in the F-actin-rich cortex, the remodeling of their membranes required for maturation, and their controlled release [16]. To date, most of the studies about MyoVa have focused on understanding how it regulates the trafficking of secretory vesicles. However, none of them have investigated whether MyoVa could also influence the internal composition of such vesicles, even though MyoVa is recurrently found in extracellular vesicles called exosomes [17].
In this work, to investigate whether MyoVa directly interacts with soluble protein components of exosomes, we performed a two-hybrid screening using the GTD of MyoVa as bait and a universal human-normalized library as prey. As envisaged, we identified the interaction of MyoVa-GTD with four proteins that compose exosomes, including the enzyme spermine synthase (SMS), which plays key roles in neurodevelopment and brain function. SMS interacts with MyoVa-GTD in vitro, with a K d in the low micromolar range, and localizes together with MyoVa in vesicles at the cytoplasm of two exosome-producer cell types. MYO5A gene silencing led to a diffuse distribution of SMS, indicating a novel role of MyoVa in the targeting of this enzyme to secretory vesicles. Moreover, either MYO5A knockdown or knockout decreased SMS gene expression, supporting that MyoVa may influence the synthesis or stability of SMS mRNA. As SMS produces the neuromodulator spermine, which is stored in secretory vesicles and released via exocytosis, our findings might have implications not only in the targeting of spermine synthase to exosomes but also in the molecular mechanisms underlying the secretion of spermine.

Yeast two-hybrid screening
The yeast two-hybrid screening was performed using the Matchmaker ® Gold yeast two-hybrid system (Takara Bio -Clontech) according to the manufacturer's instructions. All reagents used in this screen was purchased from Takara Bio -Clontech, unless stated otherwise. The MyoVa-GTD-S1651E/S1652E [9] coding sequence (residues 1448-1855; NP 000250.3) was subcloned into the EcoRI and SalI sites of pGBKT7 (pGBKT7-EE construct), and was used to transform Saccharomyces cerevisiae Y2H Gold ® strain. This phosphomimetic (EE) construct was the bait because our original aim was to identify phospho-specific interactions. The pGADT7-Prey plasmid of clones positive for the activation of all reporter genes of the system (HIS3, ADE2, AUR1-C, and MEL1) were extracted using the Easy Yeast Plasmid Isolation Kit. The purified vectors were transformed into Escherichia coli DH5α competent cells, extracted with the QIAprep Spin Miniprep Kit (Qiagen), submitted to DNA sequencing, and compared with non-redundant sequence databases using BLASTn and BLASTx [18].

Yeast two-hybrid pairwise validation
To validate the positive hits identified in the yeast two-hybrid screen and to test whether the interactions were dependent of the phosphomimetic mutation, S. cerevisiae Y2H Gold ® cells were co-transformed with a pair of pGBKT7 and pGADT7 vectors (Supplementary Table S1), and grown at 30 • C for 4 days on QDO/X/A-agar plates -a synthetic defined agar medium without tryptophan, leucine, histidine and adenine, and supplemented with 200 ng/ml aureobasidin A and 40 μg/ml 5-bromo-4-chloro-3-indolyl alpha-D-galactopyranoside (Takara Bio -Clontech). To remove potential false-positive results, we tested the activation of all reporter genes in Y2H Gold ® cells co-transformed with pGADT7-Prey and empty pGBKT7 vectors (negative controls).

Bioinformatics validation
The nucleotide sequences of preys validated in the previous step were subjected to bioinformatics analyses to filter out possible false positives still present. Initially, the sequences were analyzed using BLASTn [18] to remove those hits containing 5 or 3 UTR in frame with the GAL4 AD sequence, which would generate artificial fusion proteins. For constructs containing truncated open reading frames (ORF), only those hits containing at least one intact domain were considered as true positives, according to protein sequence analyses using the SMART server [19].

Molecular cloning
The full-length SMS ORF (NM 004595.4, 253-1353 pb) was amplified from a human cDNA library by PCR using the primers shown in Supplementary Table S2. The PCR product was purified with QIAquick PCR Purification Kit (Qiagen), digested using NdeI and XhoI restriction enzymes and inserted into the NdeI/XhoI sites of pET28a tobacco etch virus (TEV) vector [9] using T4 DNA Ligase (Promega). Positive clones were confirmed by DNA sequencing.
The MST experiment was performed using a Monolith™ NT.115 (NanoTemper Technologies) device, with a LED power of 40%, and a MST power of 60%. SMS was labeled with the His-Tag Labeling Kit RED-tris-NTA (NanoTemper Technologies), following the manufacturer's labeling protocol, and loaded into Monolith™ NT.115 MST premium-coated capillaries (NanoTemper Technologies). All assays were performed in triplicate using 50 nM His-tagged-labeled-SMS and a serial dilution of MyoVa-GTD to the maximum possible concentration in SEC buffer. The differences between the cold and hot states of each of the 16 MST profiles were used to determine the change in fluorescence intensities for each profile using the following equation: F norm = F hot /F cold · 1000, where F is the fluorescence measured in each state. The data were processed using the NTAffinity Analysis software (NanoTemper Technologies) and the mean of F norm triplicates plotted against ligand concentration were fitted with the Hill equation = ligand concentration, n = Hill coefficient, and K d = dissociation constant) using the Origin 8.0 software to estimate the K d .

Immunocytochemistry and gene silencing assays
Immunocytochemistry assays were performed on human neuroblastoma (SH-SY5Y) and human mammary adenocarcinoma (MCF-7) cell lines cultured in Dulbecco's modified Eagle's medium (DMEM) with high glucose (GIBCO -Thermo Fisher Scientific: 12800-017) supplemented with 10% (v/v) fetal bovine serum (FBS) 100 units/ml penicillin and 100 μg/ml of streptomycin and kept in a humid atmosphere containing 5% (v/v) CO 2 at 37 • C. The Stealth RNAi™ siRNAs targeting MYO5A were purchased from Invitrogen -Thermo Fisher Scientific and are shown in Supplementary Table S3. A scramble sequence Stealth RNAi™ siRNA Negative Control High GC Duplex (Invitrogen -Thermo Fisher Scientific) was used as a control. MCF-7 cells were transfected using DharmaFECT 1 Transfection Reagent (GE Healthcare) according to the manufacturer's instructions. Cells were fixed with 2% (v/v) paraformaldehyde pH 7.4 for 20 min, and then permeabilized with 0.3% (v/v) Triton X-100, blocked with 100 mM glycine, and then 3% (m/v) BSA (adapted from Assis et al. [7]). The following antibodies were used: 1 μg/ml rabbit anti-SMS (Sigma-Aldrich: HPA029852); 2 μg/ml rat polyclonal affinity-purified anti-MyoVa Medial Tail (in house, manuscript in preparation). Secondary antibodies used were 2 μg/ml Alexa Fluor ® goat anti-rat 488 (Abcam: ab150157), 2 μg/ml Alexa Fluor ® donkey anti-rabbit 594 (Molecular Probes: A21207) IgG. The slides were mounted on ProLong ® Diamond Antifade Mountant medium with DAPI (Thermo Fisher Scientific: P36962) and the images were collected on Zeiss LSM780 Axio Observer multifotons inverted confocal microscope, with a 63× objective. The Icy BioImage open source software from the Pasteur Institute (http://icy.bioimageanalysis.org/) was used for image processing and calculation of the Pearson's coefficients with the colocalization studio plugin [22]. All the compared images were acquired with the same parameters. The raw files were used for all quantifications.

RNA isolation and quantitative PCR
Total RNA was extracted from the cells samples according to standard TRIzol protocol (Invitrogen -Thermo Fisher Scientific). Total RNA (1 μg) was reverse transcribed to cDNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems -Thermo Fisher Scientific) according to a standard manufacturer's protocol followed by amplification on the ABI 7500 Real-Time PCR System (Applied Biosystems -Thermo Fisher Scientific), using primers from the Supplementary Table S4. Samples of SH-SY5Y cells (control and differentiated) were prepared as described below. All quantitative PCR (qPCR) analyses were performed in triplicate. The expression of endogenous control (GAPDH) was used for the normalization of RNA input.
Gene expression levels were calculated by relative quantitation using the ABI 7500 Real-Time PCR SDS 1.2 software (Applied Biosystems -Thermo Fisher Scientific) and the fold expression changes were determined by 2 − C T method [23]. The data are presented as the fold change of mRNA expression in cells treated with siRNA relative to cells treated with siControl after normalization to an endogenous control (GAPDH or TBP). The qPCR data were described as mean + − standard deviation and analyzed using the Student's t test, with P≤0.01 considered statistically significant.

Neuronal differentiation
SH-SY5Y cells were cultivated in DMEM high glucose (GIBCO -Thermo Fisher Scientific: 12800-017), supplemented with 10% (v/v) FBS (GIBCO -Thermo Fisher Scientific) and 100 units/ml penicillin and 100 μg/ml of streptomycin (GIBCO -Thermo Fisher Scientific: 15140122) and kept in a humid atmosphere containing 5% (v/v) CO 2 at 37 • C. Cellular differentiation was induced as described by Encinas and coworkers [24] with some modifications. Briefly, 2.5 × 10 4 cells/cm 2 were plated on DMEM high glucose medium supplemented with 10% (v/v) FBS and 1% penicillin/streptomycin solution, in culture plates or on 13 mm 2 glass coverslips previously treated with 0.1 mg/ml poly-D-lysin (Sigma-Aldrich). The next day, the medium was removed and replaced with medium II (DMEM high glucose medium containing 1% (v/v) FBS, 1% penicillin/streptomycin solution and 10 μM retinoic acid [Abcam: ab120728]). After 3 days of growth, medium was replaced with fresh medium II supplemented with 50 ng/ml brain-derived neurotrophic factor (BDNF; Sigma-Aldrich: SRP3014) and cells were grown for further 4 days. RNA extraction and qPCR assays were performed as described above.

Culture of FO and RO fibroblasts
FO cells are human fibroblasts isolated from skin fragments of a patient carrying a mutation in the MYO5A gene, which renders these cells null for MyoVa protein (Griscelli Syndrome type I/Elejalde syndrome). RO cells are normal human fibroblasts age-paired with FO, also isolated in the same laboratory, and used as control. The cells were cultured in DMEM high glucose (GIBCO -Thermo Fisher Scientific: 12800-017) supplemented with 10% (v/v) FBS (GIBCO -Thermo Fisher Scientific), at 37 • C, 5% (v/v) CO 2 . Total RNA was extracted using standard TRIzol protocol and used for cDNA synthesis and qPCR, as described above. All methods presented here were performed in accordance with the relevant guidelines and regulations approved by

Results
Using MyoVa-GTD as bait, we obtained 54 clones in a yeast two-hybrid screen of a human-normalized cDNA library. These clones represented 35 different genes, according to DNA sequence analyses of the prey plasmids. To further validate these results, we performed a pairwise two-hybrid assay resulting in 21 positive clones. However, nine of them were from noncoding regions, such as 5 and 3 UTRs of mRNA, and three encoded zinc finger proteins or heat shock proteins, which are recurrent false positives in yeast two-hybrid assays [25] (Table 1). After the exclusion of truncated single-domain proteins (probably misfolded), four clones remained as potential binding partners of MyoVa-GTD: SMS, WDR48, CSDE1, and TC2N ( Figure 1A,B). Interestingly, these four proteins are known components of exosomes (Table 2). Further, we evaluated whether these interactions could be regulated by the MyoVa-GTD phosphorylation at Ser1652. However, the four proteins interacted with both the GTD-EE (S1651E/S1652E) and GTD-AA (S1651A/S1652A) constructs in a pairwise two-hybrid assay, indicating that their binding to GTD is independent of the phosphorylation under the tested conditions ( Figure 1A). The Ala mutant was used to mimic the nonphosphorylated state of MyoVa-GTD and preserves the structure and dynamic behavior of the wild-type construct produced in E. coli [9].
After analyzing the functional data available for SMS, WDR48, CSDE1, and TC2N in the literature, we decided to further characterize the interaction between MyoVa-GTD and SMS due to the key role of this enzyme in neurodevelopment [26,27]. For in vitro affinity assays, MyoVa-GTD and SMS proteins were expressed and purified to homogeneity ( Supplementary Figures S1 and S2), the His-tag of MyoVa-GTD was removed, and the purified SMS protein was labeled with the RED-tris-NTA dye (Supplementary Figure S3). According to MST experiments, MyoVa-GTD and SMS formed a complex in vitro with a dissociation constant of 3.5 + − 0.5 μM (Figure 1C), supporting SMS as a novel MyoVa-binding partner.
In breast cancer (MCF-7) and neuroblastoma (SH-SY5Y) cell lines, SMS localized together with a subset of MyoVa-labeled vesicles at the cytoplasm (Figure 2). We also noticed that MYO5A knockdown induced a more diffuse distribution of SMS, in contrast to the fewer, but sharper puncta observed at the control ( Figure 3A), indicating that MyoVa may play a role in the targeting of SMS to vesicles. RNA-binding protein [43] Potential MyoVa-binding partners are highlighted in bold. * The final pb identified in our sequencing, not the final pb in the prey.   [19] and IBS [41]. (C) MST assays showing that MyoVa-GTD binds to SMS with a K d in the low micromolar range. F norm = normalized fluorescence. Data are presented as mean + − SD (error bars) from triplicates. and SMS (red) labeling. The correlation is based on the average of 16 and 18 independent cells, for MCF-7 and SH-SY5Y cells, respectively, with standard deviations shown. The correlation between MyoVa (rotated 180 • ) and SMS channels was used as negative control (random colocalization). Nucleus were stained with DAPI (Blue). Scale bar: 10 μm (field), 1 μm (zoom). **** P≤0.001 from two-tailed Wilcoxon's non-parametric rank test in C (because one of the populations were not normally distributed) and two-tailed paired t test in F).
According to qPCR analyses, the mRNA expression levels of SMS decreased upon MYO5A-silencing ( Figure  3B,C), which correlated with a slight decrease in SMS protein content assessed by Western blot (Supplementary Figure S4). To investigate a possible correlation between depletion of MyoVa protein and SMS transcription, we also evaluated the expression of SMS in MYO5A-null primary fibroblasts isolated from patients with Griscelli Syndrome Type 1/Elejalde syndrome. As expected, the SMS expression was lower in cells lacking functional MyoVa compared with normal primary fibroblasts, indicating a role for MyoVa in the synthesis or stability of SMS mRNA ( Figure 3D). In contrast, when the expression of MYO5A was stimulated upon neuron differentiation, the expression level of SMS also increased, further supporting a correlation between the expression of these two genes ( Figure 3E-G).

Discussion
In this work, we reveal the direct interaction between MyoVa-GTD and four components of exosomes -WDR48, CSDE1, TC2N, and SMS -using a highly stringent yeast two-hybrid system. Furthermore, we validate the interaction between MyoVa-GTD and the enzyme SMS in vitro, and provide primary evidence about a new role for MyoVa in the expression and targeting of SMS into secretory vesicles.
MyoVa-GTD and SMS form complexes in vitro with a dissociation constant in the low micromolar range (K d = 3.5 + − 0.5 μM), which is typical of transient complexes and is similar to the affinity of MyoVa-GTD to other binding partners, such as the C2 domains of RPGRIP1L (3-9 μM, also determined by MST but using a covalent probe) [7]. The protein SMS has been found in the cytosol, in nuclear bodies and in exosomes [28][29][30]. This enzyme converts spermidine into spermine (EC 2.5.1.22) [31][32][33], a polyamine that acts as second messenger in neurotransmission, targeting receptors in postsynaptic membranes [34][35][36]. The vesicular storage of spermine and spermidine involves an active transporter from the SLC18 family, but the mechanisms coupling spermine synthesis to secretion are still elusive [37]. In this context, and based on our results showing the physical interaction between MyoVa-GTD and SMS, the colocalization of SMS and MyoVa in a subset of cytoplasmic vesicles and the more diffuse distribution of SMS protein upon MYO5A gene silencing, it is tempting to hypothesize that the actin-based motor MyoVa targets the enzyme that produces spermine to secretory vesicles for spermine secretion via exocytosis [37,38], and/or SMS release via exosomes [28,29]. Moreover, whether and how the interaction between MyoVa and SMS influence the spermine synthase activity should be further evaluated in future studies.
Besides our RNAi studies show lower levels of SMS mRNA upon MYO5A silencing, other evidences support a correlation between MYO5A and SMS expression. In MYO5A-null fibroblasts, SMS transcripts are less abundant than in normal cells, indicating that functional MyoVa is required for the usual expression of SMS gene. Moreover, MYO5A and SMS are up-regulated in differentiated neurons and the proteins they encode are less abundant in human brains with Huntington's disease, a progressive neurodegenerative disorder [39]. Phenotypic traits further support a correlation between MyoVa and SMS, since patients harboring loss-of-function mutations in SMS or MYO5A genes share some neurological symptoms, including intellectual disability, seizures, and hypotonia [12,27,40].
In summary, the new interaction between MyoVa and exosomes components, including the enzyme SMS, and the co-occurrence of MyoVa and SMS in cytoplasmic vesicles bring new perspectives about the roles of this molecular motor in exocytic pathways, especially in the filling of exosomes and secretion of the neuromodulator spermine.

Availability of materials and data
Materials, data, and associated protocols will be promptly available to readers without undue qualifications in material transfer agreements.