Phosphorylation regulates the subcellular localization of Cucumber Mosaic Virus 2b protein

The 2b protein of Cucumber mosaic virus has a role in nearly all steps of the viral cycle including cell-to-cell movement, symptom induction and suppression of antiviral RNA silencing. Previous studies demonstrated the presence of 2b protein in the nucleus and in cytoplasm as well. Phosphorylation site of 2b protein is conserved in all CMV isolates, including proposed constitute motifs for casein kinase II and cyclin-dependent kinase 2. To discern the impact of 2b protein phosphorylation, we created eight different mutants to mimic the non-phosporylated (serine to alanine) as well as the phosphorylated state (serine to aspartic acid) of the protein. We compared these mutants to the wild-type (Rs-CMV) virus in terms of symptom induction, gene silencing suppressor activity as well as in cellular localization. Here, in this study we confirmed the phosphorylation of 2b protein in vivo, both in infected N. benthamiana and in infiltrated patches. Mutants containing aspartic acid in the phosphorylation site accumulated only in the cytoplasm indicating that phosphorylated 2b protein could not enter the nucleus. We identified a conserved dual phosphorylation switch in CMV 2b protein, which equilibrates the shuttling of the 2b protein between the nucleus and the cytoplasm, and regulates the suppressor activity of the 2b protein.


Results
Phosphorylation of 2b protein in vivo. The putative phosphorylation site of CMV 2b protein (amino acids [39][40][41][42][43] located at the beginning of the second alpha helix is strictly conserved in both subgroups of the virus 26 . KSPSE has been proposed to constitute phosphorylation motifs for casein kinase II (S/TxxD/E) and cyclin-dependent kinase 2 (40SP41). To examine the in vivo phosphorylation state of the 2b protein, Western blot analysis was carried out by using phosphoserine antibodies. Since antibodies for 2b protein detection were not available, the coding region of histidine tagged 2b protein was cloned into the infectious cDNA clone of Rs-CMV RNA 2. It was previously found that attachment of a hexahistidine sequence did not alter the subcellular localization and the ability to bind sRNAs 23,28 . Nicotiana benthamiana plants were infected with the chimaera Rs-2bHis in vitro transcripts in the presence of the Rs-CMV RNA 1 and 3 transcripts. Parallel to the infection, N. benthamiana plants were infiltrated with the binary vector expressing the previously described 2bHis protein 23 . Ten days after inoculation (symptoms appeared on the upper non-inoculated leaves) and 4 days post infiltration, the accumulation and the phosphorylation of 2b protein were analysed by Western-blot using penta-his and phosphoserine antibodies (Fig. 1, Supplementary Fig. S1). The Western blot analysis proved the 2b protein phosphorylation in vivo in infected plant as well as in infiltrated patches (Fig. 1), but the mutant 2b/40-42/AAA was not phosphorylated in infiltrated patches (Fig. S1) demonstrating that the 40/42 amino acids are the exclusive phosphorylated amino acids of the 2b protein.

Mutations in the 2b protein phosphorylation domain alter symptom induction of CMV.
In order to analyse the effect of the different phosphorylated state of 2b protein, seven different mutants were created. Substitutions were integrated into the Rs-CMV RNA 2 clone in the SPS motif for analysis of the effect of alanine (to mimic the non-phosphorylated state) and aspartic acid (to mimic the phosphorylated state) substitution in single or in both serine residues (40S, 42S). The substitution of serine amino acid with negatively charged aspartate residues, known to reproduce the electrostatic effect of phosphorylation as it was previously applied for example in the case of TMV MP30 31,32 . All of the mutations were integrated into the RNA 2 infectious clone of Rs-CMV resulting in the following constructs: 2b/40-42/SAS, 2b/40-42/APS, 2b/40-42/SPA, 2b/40-42/APA, 2b/40-42/DPS, 2b/40-42/SPD and 2b/40-42/DPD. To compare these mutants with wild-type virus (Rs-CMV) in terms of symptom induction and viral accumulation, plasmids containing the full-length clones of Rs-CMV RNAs were transcribed in vitro for inoculation of N. benthamiana plants. RNA 1 and 3 transcripts were combined with the wild-type RNA 2 and either with the mutated RNA 2 transcripts. The infection was followed by visual observation and northern blot analysis. Infections of N. benthamiana with Rs-SPS/40-42/APA and Rs-SPS/40-42/ DPD were asymptomatic similarly to the previously described Rs-SPS/40-42/AAA 23 during the monitored period (six weeks). Infection with Rs-SPS/40-42/SPD and Rs-SPS/40-42/DPS induced mild symptoms: mild leaf distortion, mild systemic mosaic, and mild stunting ( Fig. 2A). N. benthamiana plants infected with Rs-SPS/40-42/APS, Rs-SPS/40-42/SAS and Rs-SPS/40-42/SPA exhibited severe symptoms, including systemic mosaic, leaf distortion, and chlorosis. However, symptoms induced by these mutants were never as severe as those induced by Rs-CMV in N. benthamiana. The northern analysis was in concordance with the symptom appearance (Fig. 2B). The identity of each of the mutants was verified by RT/PCR followed by nucleic acid sequence determination.
Gene silencing suppressor activity of the mutant 2b proteins in patch assays. Gene silencing suppressor activity is the primary function of 2b protein, so this feature was analysed in Agrobacterium-mediated co-infiltration assay. Binary vector expressing GFP reporter gene was agroinfiltrated into transgenic N. benthamiana (silenced for GFP expression) leaves together with the binary vector expressing the wild type 2b protein or the different mutant ones (2b/40-42/AAA, 2b/40-42/SAS, 2b/40-42/APS, 2b/40-42/SPA, 2b/40-42/APA, including systemic mosaic, leaf distortion, and chlorosis. However, symptoms induced by these mutants were never as severe as those induced by Rs-CMV. CMV accumulation in systemically infected leaves analysed by northern hybridization at 7 dpi. The viral RNA accumulation was in accordance with the strength of the symptoms. The radiolabelled probe was specific for Rs-CMV RNA 3. Ethidium bromide-stained rRNA from the same volume of the sample is shown below each lane. 2b/40-42/DPS, 2b/40-42/SPD, 2b/40-42/DPD) (Fig. 3A). GFP silencing was visually observed at 4 days post agroinfiltration and the accumulation level of GFP RNA in the infiltrated leaves was quantitatively measured by qRT-PCR (Fig. 3B). We found that mutants 2b/40-42/SAS, 2b/40-42/APS and 2b/40-42/SPA retained partial suppressor activity. However, mutants 2b/40-42/DPS, 2b/40-42/SPD, 2b/40-42/DPD and 2b/40-42/APA failed to suppress GFP activity similarly to the previously described 2b/40-42/AAA 23 . The accumulation of the mutant proteins were analysed by western-blot using penta-his antibody to ascertain that the different GFP levels were caused by the different suppressor activities not by the instability of the proteins (Fig. 3C).

Subcellular localization of 2b protein and its mutants monitored by BiFC.
Previous studies suggest that beside nuclear localization signals the phosphorylation state of 2b protein affects its nuclear localization since deletion of the complete phosphorylation motif (2bΔKSPSE) altered the subcellular localization of 2b protein, confining it largely to the nucleolus 27 . To determine the impact of phosphorylation state of 2b protein we monitored the subcellular localization of the wild-type and mutated 2b proteins with bimolecular fluorescent complementation assay (BiFC). Previously, self-interaction of CMV 2b protein was demonstrated in vivo in N. benthamiana using BiFC 27 , and proposed that CMV 2b protein, like TAV (Tomato aspermy virus) 2b protein, functions in the form of a homodimer or higher order of oligomer as a suppressor of RNA silencing. Both cytoplasmic and nuclear localization of the dimers were demonstrated 27,28 . To analyse the subcellular localization of 2b protein in vivo, both the N-terminal and C-terminal fragments of YFP were fused to the C-terminus of 2b protein and also to the mutant 2b proteins with highly altered phenotype, generating fusion proteins (SPS/40-42/ AAA and SPS/40-42/DPD). The 2b protein forms homodimers or higher oligomers therefore we used interaction of the wild-type 2b protein itself as a positive control applying BiFC in N. benthamiana plants. At 3 dpi, infiltrated leaves were subjected to confocal microscopy to monitor the reconstituted YFP signal. Strong YFP fluorescence in the cytoplasm and in the nucleus co-expressed Rs2b-N-YFPC and Rs2b-C-YFPC demonstrated that Rs-CMV was very apparent in cytoplasm as well as in nuclei, similarly to Gonzalez et al. 27 (Fig. 4A). No fluorescence was observed by confocal microscopic imaging in cells within leaf patches co-infiltrated with A. tumefaciens cells expressing 2b/40-42/AAA. Interestingly, fluorescence was detectable only in cytoplasm in the case of mutant 2b/40-42/DPD, but fluorescence was not detected in the nucleus, or in the nucleolus (Fig. 4A,B). This observation indicates that the phosphorylated 2b protein accumulated in cytoplasm, phosphorylation of 2b protein prevents its accumulation in the nucleus.

Subcellular localization of EGFP fused 2b protein and its mutants.
The EGFP was fused to the C-terminus of the Rs2b protein and to the 2b/40-42/APA and 2b/40-42/DPD mutants in order to visualize the subcellular localization of the monomer form of the 2b protein as well. In infiltration experiments the Rs2b-EGFP and the 2b/40-42/APA-EGFP were detected in the cytoplasm and the nuclei, while the 2b/40-42/DPD-EGFP was detected only in the cytoplasm, but not in the nuclei (Fig. 5), confirming that the phosphorylated 2b protein is excluded from the nuclei.

Nuclei isolation.
To further analyse the subcellular localization of the mutants, nuclei were purified from agroinfiltrated leaf patches as described previously 33 . The presence of the constructs in the nuclei was verified by visual observation using microscopy as well as protein extraction and staining (  (Fig. 6). These findings further demonstrate that the phosphorylated state of 2b protein precludes it from accumulating in the nucleus and restricts the location of 2b protein exclusively to the cytoplasm.

Discussion
Reversible phosphorylation is a well-studied posttranslational protein modification, which plays a fundamental role in the regulation of different protein activities in vivo. Several studies in case of different viruses suggest the physiological relevance of phosphorylation of movement protein 14,34 and coat proteins in viral cell-to-cell movement 9,35 . Limited data is available about the function of CMV 2b protein phosphorylation. Previous sequence analyses have identified potential phosphorylation sites (amino acids 39-43) for casein kinase II (CKII) (S/ TxxD/E) and CDK2 (40SP41), which are conserved in all of the cucumoviral 2b proteins, both in subgroup I and II as well as in TAV and in Peanut stunt virus (PSV) 26 . In this study, in vivo phosphorylation of 2b protein has been directly proved with transiently expressed 2b protein and in CMV infected N. benthamiana leaves (Fig. 1). Serine 40 and 42 located at the beginning of the second alpha-helix (Fig. 7B, encircled regions) was previously analysed with deletion mutants and site specific mutagenesis. It was previously described essential for nuclear accumulation and siRNAs binding to suppress PTGS 27,30 as well as for symptom induction 29 . The deletion of the entire motif (mutant 2b/DKSPSE) abolished the suppressor activity of the protein reducing its ability to bind RNAs 27 . It is consonant with our results as mutant 2b/40-42/AAA failed to suppress GFP activity (Fig. 3). Interestingly, mutants 2b/40-42/SAS, 2b/40-42/APS, 2b/40-42/SPA and retained partial suppressor activity (Fig. 3). A CMV 2b homologue protein, TAV 2b forms a pair of hook-like dimers to recognize the siRNA duplex in a sequence independent and length-preference manner 36 . Studies using the TAV 2b homologue demonstrated that in aqueous solution its homodimers or oligomers are spontaneously formed, even in the absence of siRNAs 36 . Tetramer form is not necessary for siRNA binding and suppressor activity 27 . The examined mutations are located in the forepart of the second α-helix. The hydroxyl groups of serine 40 and 42 form H-bond(s) with the phosphate oxygen atoms of the cytosine-14 (C14_PO) and/or guanosine-13 (G13_PO) (Fig. 8A,C). These H-bonds give high stability to the whole 2b tetramer -siRNA complex (Fig. 7A,B). But, in the course of the MD simulation for only very short time periods (5-50 ps) these serine side chains swing toward the solvent where a host kinase can phosphorylate these serine residues. After the phosphorylation event the phospho-serine side chains cannot turn back to the negatively charged sugar-phosphate backbone of the siRNA, since their negative charges repel the 2b protein chain from the siRNA.
Beside phosphorylation domain, 2b protein also contains nuclear localization signals (NLS1 and NLS2) 27 . The molecular models of 2b dimers suggest that the NLS regions located at the first alpha-helix are involved not only in siRNA binding but act as interaction interface for importin protein, probably interaction with karyopherin α as suggested previously 25 . Our previous study also suggested that the less effective suppression of local gene silencing of NLS mutants is a result of the damaged structure and not the absence of nuclear localization 23 . It was also in accordance with a previous study by Gonzalez et al., as tagging GFP-2b with a NES resulted in the relocalization of the protein from the nucleus/nucleolus into the cytoplasm. The patch assays showed that NES-GFP-2b retained its ability to suppress RNA silencing 30 . The nuclear import of a subset of nuclear proteins that contain the classic nuclear localization signals is mediated by the importin α/β heterodimer (NLSs) [37][38][39] . Generally, α subunit specifically binds to basic NLSs in the cytoplasm, whereas the β subunit interacts with the NPC during the import process 39 . The nuclear import of 2b may rely on the importin-α/β transport system, and importin α could bind 2b by recognition of its nuclear localization signals (NLSs). Previously it was observed that a mutant with a deleted putative phosphorylation site was mainly localized to nucleoli, suggesting a role of phosphorylation in nuclear import 27 . Our results clearly demonstrate that mutants mimicking the non-phosphorylated state (SPS/40-42/ AAA, SPS/40-42/APA, and single mutants SPS/40-42/APS, SPS/40-42/SPA and SPS/40-42/SAS) can be transported into the nucleus since they were detectable from purified nucleoli (Fig. 6). However, we could not detect mutants which mimicked phosphorylation (SPS/40-42/DPS, SPS/40-42/SPD and SPS/40-42/DPD) from nuclei. These results were also supported by the BiFC and 2b-EGFP fusion protein localization experiment as mutant SPS/40-42/DPD was localized in the cytoplasm but not in the nucleus (Figs 4 and 5). Besides in vivo experiments, dual phosphorylated state of 2b protein at S40 and S42 was also modelled in silico. The serine 40 and 42 were mutated into aspartate residues. Aspartate is a phosphomimetic of phospho-serine and can mimic the constant phosphorylated form of the protein. The initial conformation of the phosphomimetic form of the 2b tetramer -siRNA complex was similar to the native form. Nevertheless, due to the aforementioned negative -negative charge repulsion the ribonucleoprotein complex became very instable (Fig. 7). The distances between the mutated residues (D40, D42) and the phosphate oxygen atom of the siRNA C14-PO increased from the initial 0.25 nm to the range of 1 -3 nm (Fig. 8B,D). However, during the 100 ns MD simulation the mutant 2b tetramer -siRNA complex did not completely come apart, but for example; the 2b (C, D chains) -siRNA2 dimer part were turned  almost 180 degrees from the initial state (Fig. 7B). These findings are consistent with the patch assay results as mutants 2b/40-42/DPS, 2b/40-42/SPD, 2b/40-42/DPD failed to suppress GFP activity (Fig. 3).
In this study we identified a conserved dual phosphorylation switch in CMV 2b protein. Non-phosphorylated serine residues are essential in the 40 and 42 positions of the 2b protein to form stable ribonucleoprotein complex with double stranded siRNAs. Basically, two cases can be distinguished with regard to the 2b phosphorylation in relation to siRNA binding. First, if Ser-40 and Ser-42 of the monomer 2b protein is in phosphorylated state it cannot bind siRNA and obviously the ribonucleoprotein complex cannot be assembled. In the second case the stable 2b-siRNA complex (2b dimer with one double stranded siRNA or 2b tetramer with two double stranded siRNA) is phosphorylated by a host kinase and the ribonucleoprotein complex will disassemble. In accordance to these results, the impact of phosphorylation in the regulation of siRNA binding was demonstrated in the case of human Argonaute proteins, where RNA binding was strongly reduced in the phosphorylated state 40 .
Previously the nuclear-cytoplasmic partitioning of 2b protein was found determine the balance between its roles as a virulence determinant or as an RNA-silencing suppressor 28 . In this current study we have also compared  the symptoms induced by the different 2b mutant viruses to the wild-type virus (Rs-CMV). Either point mutation of the two serine residues had effects on symptom induction since infection with the different mutants caused diminshed symptoms on N. benthamiana and led to a decreased levels of viral RNA compared to the wild-type CMV-Rs isolate (Fig. 2). Previous studies suggested that mutation of serine residue 40 or 42 reduces the ability of CMV to induce symptoms without impairing viral or 2b protein accumulation 29 . Single mutation in residue 41 (Pro41Ala) in TAV 2b protein showed a significant decrease in the local gene silencing suppressor activity 36 , similarly to the 2b protein of CMV, since we found that single mutation in residue 41 results in partial loss of suppressor activity (mutant SPS/40-42/SAS, Fig. 3).
Previously the role of phosphorylation in nuclear trafficking of proteins was demonstrated in several studies. For example in the case of human dUTPase, phosphorylation abolishes its nuclear import 41 . In the case of Epstein-Barr virus nuclear antigen, phosphorylation has a key role in the regulation of nuclear import 42 . Phosphorylation of nucleoprotein (NP) controls the nuclear-cytoplasmic shuttling of Influenza A virus 43 . More functional experiments will be needed to clearly identify the biological function of the phosphorylated 2b protein retained in the cytoplasm. Now it is conceivable that the phosphorylated 2b protein is a repository that can be easily converted into functional 2b protein available for suppressing the gene silencing process. Another possibility is sequestering the 2b protein for a different function, similarly to the Akt3-mediated phosphorylation of Ago2, that was proved to be a molecular switch between target mRNA cleavage and translational repression activities of Ago2 44 .

Methods
Plasmid constructions. Description of the Rs-CMV and the infectious transcripts (pRs1, pRs2, pRs3) has been published previously 45 Supplementary Table S1. DNA constructs encoding the Rs2b protein and its mutants (2b/40-42/APA and 2b/40-42/DPD) with EGFP fusion to C-terminus were created by overlapping PCR. The sequences of primers used for overlap PCR are detailed in Supplementary Table S1. Prior to Agrobacterium transformation, the identity of the contructs were verified by nucleotide sequence determination.

2b-His infectious clone. The 2b-His infectious clone was created from the previously described Rs2
GAF/108-110/AAA mutant and the pGEM-2bHis clones, which was generated in pRs2 plasmid DNA containing the full-length cDNA of CMV-Rs RNA 2 23 as follows. The pGEM-2bHis clone was digested with BamHI (located downstream of the STOP codon) and blunt-ended with T4 DNA polymerase, and digested with StuI (at 2660 nt, in the middle of the 2b gene), and the insert was isolated. This fragment was ligated into the Rs2 GAF/108-110/ AAA mutant digested with PstI and blunt ended with T4 DNA polymerase (at/prior the STOP codon) and StuI in the middle of the 2b gene. The resulting clone contained the complete 2b gene, the His-tag region and the STOP codon. The identity of pRs2b-His was verified by nucleotide sequence determination.

Plant material, growth and inoculation. Nicotiana benthamiana plants were kept in environmentally
controlled growth chambers at cycle of 14 h of light (23 °C) and 10 h of dark (18 °C). The plants were mechanically inoculated at the fourth-leaf stage with in vitro mutated RNA 2 transcripts in the presence of wild type RNA 1 and 3 transcripts or RNA 1, 2 and 3 of the wild type Rs-CMV transcripts. Symptom development was observed and recorded continuously during the following 2 weeks.

Analysis of plants.
Total RNA was extracted from 200 mg systemically infected leaves 7 days after inoculation 46 . Virus RNA accumulation was followed by northern blot analysis. Approximately 100 ng total RNA was denatured with formamide and separated in formaldehyde-containing agarose gels and blotted on to nylon membranes 47 . Northern blot hybridization analysis was performed with random-primed DIG-labeled DNA fragments specific for the Rs-CMV RNA3 sequence (Roche). RT-PCR/DNA sequence determination was performed to analyse the stability of the mutant virus. It was carried out with the Qiagen OneStep RT-PCR kit according to the manufacturer's instructions, using primers flanking of 2b coding region (forward 5′-GTTTGCCTGGTGTTACGACACCGA-3′, reverse 5′-GCGGATCCTGGTCTCCTTTTGGAGGCCC-3′). PCR products were purified by High Pure PCR product Purification Kit (Roche) prior to nucleotide sequence determination. Construction of YFP fusion proteins for BiFC analysis. Agrobacterium-based binary vectors pGWBc-nYFP and pGWBc-cYFP for BiFC assay were constructed as described previously 48 . CMV 2b and mutant 2b ORFs were amplified by PCR using the following primers: 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCATGGAATTGAACGTAGGTGCAA -3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCGTGATGATGATGATGATGGAAAGCA-3′. All constructs were created by Gateway Technology (Invitrogen). The orientation of each insert was confirmed by sequencing.
Agroinfiltration and confocal microscopy. Agrobacterium-mediated transient expression on N. benthamiana leaves was conducted by pressure infiltration as described previously 49 . Agrobacterium culture of p14-expressing strain was adjusted at 600 nm (OD 600 ) to a final optical density of0.4 and the strains expressing the various 2b mutants to 0.2. The plants were maintained for 3 days at 22 °C in growth chamber (14 h light/10 h dark) before single discs of infiltrated N. benthamiana leaves were assayed for fluorescence by CLSM using a Leica TCS SP8 confocal microscope. The images were captured digitally and processed using the Leica confocal software.
To analyse the subcellular localization of the monomer Rs2b protein and its mutants, N. benthamiana plants were infiltrated with A. tumefaciens cell harboring plasmids pRs2b-EGFP, p2bAPA-EGFP and p2bPD-EGFP. A. tumefaciens cells were diluted to final optical density at 600 nm to 0.5. N. benthamianna plants were assayed 3 days after the infiltration using ZEISS LSM 410 Laser Scan Microscope and Olympus BH-2 Fluorescence Microscope. 488 nm laser was used for excitation and 525/50 barrier filter for GFP detection.
Nuclei isolation. Agrobacterium-mediated transient expression on Nicotiana benthamiana leaves was conducted by pressure infiltration as described above. Agrobacterium culture of GFP-expressing strain was adjusted to a final optical density at 600 nm (OD 600 ) 0.4 and the strains expressing the various 2b mutants to 0.2. Four days post-infiltration, nuclei were purified from agroinfiltrated leaf patches as described previously 33 . Thirty leaf discs from each 2b protein constructs were used for purification, nuclei were resuspended in 250 µl nuclei storage buffer. The presence of the purified nuclei of each construct has been verified by visual observation prior to DAPI staining (1 µl of nuclei in 7 µl of DAPI, 1 µg/ml) with an Olympus bx-51 microscopy using U-MNU2 filter, as well as by protein extraction and staining (1:1 with Laemmli buffer).
Protein analysis, SDS-PAGE, and immunoblotting. Protein extracts from N. benthamiana leaves were prepared from leaf samples (20 mg, fresh weight). Leaf discs were ground and homogenized in an ice-cold mortar in Laemmli solution, heated at 95 °C for 5 min, and centrifuged (5 min at 10,000 g) to remove insoluble material. Aliquots of the supernatant (1 to 10 µL) were separated by SDS-PAGE on 20% gels. After electrophoresis, proteins were transferred to a Hybond-C membrane (GE Healthcare Bio-Sciences) and subjected to immunoblot analysis with Penta·His HRP Conjugate Kit following the manufacturer's instructions (Qiagen). For phosphoserine detection we used Phosphoserine antibody (Qiagen), HRP conjugated mouse anti-Rabbit IgG as secondary antibody (Millipore) and for detection Pierce Western blotting substrate was used following the manufacturer's instructions (Thermo Fischer Scientific).
Modelling of 2b protein-siRNA complex. The three-dimensional structure of the CMV 2b was generated with I-TASSER 50 . The model was built using the Rs-CMV 2b sequence. The NCBI/GenBank accession number is AJ517801. The main template was the X-ray structure of TAV 2b (PDB ID code: 3CZ3) to create the alpha helical regions (aa 1-69). The biologically active tetramer form with siRNA duplexes (sequence of the 19nt siRNA is the same as in the 3CZ3 structure) was built with the Schrödinger Suite 51 molecular modelling software package. The SPS/40-42/DPD mutant complex was derived from the native model using the "mutate residue" option of the molecular modelling software. The completed tetramer-siRNA ribonucleoprotein complexes were refined with energy minimization to eliminate the steric conflicts between the protein and RNA atoms.
Molecular dynamics and graphics. Prior to MD simulations the Desmond system building panel (incorporated into Schrödinger Suite) was used to construct the solvated siRNA-2b tetramer ribonucleoprotein complexes. 0.15 M NaCl salt concentrations were used in both simulations. The OPLS-AA/2005 force field was used. Before starting the 100 nanosecond production runs the systems were prepared with energy minimizations and short restrained MD simulations: first, only the position of the solvent molecules was minimized and then the potential energy of the whole system without any restraints. In the subsequent step on an NVT ensemble, a short 12 ps position-restrained MD simulation at 10 K was carried out using a Berendsen thermostat to ensure the proper solute settlement around the protein. After this, 12 ps of position restrained MD simulation was performed on an NPT ensemble at 10 K and 1 bar using a Berendsen thermostat and pressure coupling. Then a 12 ps restrained NPT MD simulation warmed the solute to 300 K at 1 bar. Finally, a 24 ps non restrained NPT MD simulation prepared the systems for the production run at 300 K and 1 bar.
The MD simulations were initiated on the NPT ensemble 300 K and 1 bar using the Desmond 52 molecular dynamics software application. The simulations were evaluated by the Desmond "simulation event analysis" application package of the Schrödinger Suite. The computing capacities were around 4-6 ns/day for the 2b tetramer-siRNA ribonucleoprotein complex simulations. The MD simulations were performed on the MTA Cloud supercomputing centre using two virtual computers (instances) with 16 VCPUs, respectively. The VMD version 1.9.3 53 was used to create spectacular molecular graphics.