E of increasing amounts of full length HMGA1 (FL), a truncated HMGA1a form (1?1), and a HMGA1a mutated form (R57,59A) for in vitro methylation assays. Methylation reactions of MIF and HMGA1a alone represent control experiments. Proteins were separated by SDS-PAGE (T = 15 ) and checked by fluorography. Experiments were repeated at least twice and a representative result is shown. (B) Blue Comassie staining was used to check for the quantification of HMGA1a proteins. (C) Schematic representation of the FL, 1?1 and R57,59A HMGA1a domain organization. doi:10.1371/journal.pone.0053750.Dimethylenastron gHMGA1a form (1?1) and a R57,59A double mutant were also included in the assay. The 1?1 HMGA1a truncated form is not able to bind PRMT6 and is not methylated by PRMT6 [11] while the R57,59A mutant is still able to bind PRMT6 (our unpublished data) but is modified by PRMT6 much less efficiently at minor methylation sites [11]. As it is possible to see from Fig. 5, the presence of a wild-type HMGA1, which is able both to interact with and to be methylated by PRMT6, strongly PHCCC enhances the PRMT6-dependent methylation of MIF; on the contrary, neither the presence of a truncated HMGA1a nor that of a not-methylatable protein exerts a relevant effect with respect to the methylation efficiency of PRMT6 towards MIF. Since the R57,59A mutant is still able to bind PRMT6, the modulatory role of HMGA1a towards PRMT6 activity seems to be linked to HMGA1a R57,59 methylation status. The possibility that interacting partners of PRMT6 could modulate the processivity of this enzyme was already envisioned [42]. Further experiments will be needed to clarify the mechanism responsible for the modulatory role of HMGA1 towards PRMT6. In summary:1. 
with our Y2H approach we were able to discover 36 new molecular partners for PRMT6. The large majority of PRMT6 interactors tested resulted to be confirmed by our in vitro and in vivo protein-protein interaction experiments. Therefore, Y2H resulted a reliable approach to fish out bona-fide cellular partners of PRMT6. Moreover, 4 new substrates for PRMT6 were discovered (see Table 1 for a summary of these results). The identification of new partners and substrates for PRMT6 and their characteristics suggests a wide 1081537 involvement of PRMT6 in the context of cell biology. A clear limitation of our study is that it is mainly based on overexpression of proteins in fusion with tags and therefore this makes difficult to assess how relevant are the interactions found in 
vivo. Once focused on selected partners/substrates other approaches should be considered to assess whether they are real PRMT6 targets. Reciprocal Co-IP experiments should be performed with the endogenous proteins and these data should be supported by orthogonal strategies, such as in vivo co-localization imaging (co-immunolocalization and/or FRET). In addition, the PRMT6-dependent methylation status of these proteins should be assessed after PRMT6 silencing/knock-outThe Protein-Protein Molecular Network of PRMTTable 1. Y2H 1st: yeast two-hybrid screening.Table 1: Summary of protein-protein interaction data and enzymatic assays Y2H 1st Med28 MTF2 CDK5RAP3 Nm23-H1 EBP1 NOB1 UTP6 hnRNP Q GRSF-1 CDK9 snRNPB PRPF39 PSMD11 PSME1 PSMB4 POMP HYPK PRDX4 SAAL1 FtL HSPB1 MIF Hint1 HPRT1 MRPL38 LDHB FH PTS QPRT COPS3 PRKX CASP6 SVEP1 TUBB2A SEPT7 HSJ-2 2nd Med28 MTF2 CDK5RAP3 Nm23-H1* EBP1 NOB1 UTP6 hnRNP Q# GRSF-1 CDK9 snRNPB PRPF39 n.s n.s n.s n.s HYPK# PRDX4 SAAL1 FtL n.s MIF Hint1 HPRT1 MRPL38* L.E of increasing amounts of full length HMGA1 (FL), a truncated HMGA1a form (1?1), and a HMGA1a mutated form (R57,59A) for in vitro methylation assays. Methylation reactions of MIF and HMGA1a alone represent control experiments. Proteins were separated by SDS-PAGE (T = 15 ) and checked by fluorography. Experiments were repeated at least twice and a representative result is shown. (B) Blue Comassie staining was used to check for the quantification of HMGA1a proteins. (C) Schematic representation of the FL, 1?1 and R57,59A HMGA1a domain organization. doi:10.1371/journal.pone.0053750.gHMGA1a form (1?1) and a R57,59A double mutant were also included in the assay. The 1?1 HMGA1a truncated form is not able to bind PRMT6 and is not methylated by PRMT6 [11] while the R57,59A mutant is still able to bind PRMT6 (our unpublished data) but is modified by PRMT6 much less efficiently at minor methylation sites [11]. As it is possible to see from Fig. 5, the presence of a wild-type HMGA1, which is able both to interact with and to be methylated by PRMT6, strongly enhances the PRMT6-dependent methylation of MIF; on the contrary, neither the presence of a truncated HMGA1a nor that of a not-methylatable protein exerts a relevant effect with respect to the methylation efficiency of PRMT6 towards MIF. Since the R57,59A mutant is still able to bind PRMT6, the modulatory role of HMGA1a towards PRMT6 activity seems to be linked to HMGA1a R57,59 methylation status. The possibility that interacting partners of PRMT6 could modulate the processivity of this enzyme was already envisioned [42]. Further experiments will be needed to clarify the mechanism responsible for the modulatory role of HMGA1 towards PRMT6. In summary:1. with our Y2H approach we were able to discover 36 new molecular partners for PRMT6. The large majority of PRMT6 interactors tested resulted to be confirmed by our in vitro and in vivo protein-protein interaction experiments. Therefore, Y2H resulted a reliable approach to fish out bona-fide cellular partners of PRMT6. Moreover, 4 new substrates for PRMT6 were discovered (see Table 1 for a summary of these results). The identification of new partners and substrates for PRMT6 and their characteristics suggests a wide 1081537 involvement of PRMT6 in the context of cell biology. A clear limitation of our study is that it is mainly based on overexpression of proteins in fusion with tags and therefore this makes difficult to assess how relevant are the interactions found in vivo. Once focused on selected partners/substrates other approaches should be considered to assess whether they are real PRMT6 targets. Reciprocal Co-IP experiments should be performed with the endogenous proteins and these data should be supported by orthogonal strategies, such as in vivo co-localization imaging (co-immunolocalization and/or FRET). In addition, the PRMT6-dependent methylation status of these proteins should be assessed after PRMT6 silencing/knock-outThe Protein-Protein Molecular Network of PRMTTable 1. Y2H 1st: yeast two-hybrid screening.Table 1: Summary of protein-protein interaction data and enzymatic assays Y2H 1st Med28 MTF2 CDK5RAP3 Nm23-H1 EBP1 NOB1 UTP6 hnRNP Q GRSF-1 CDK9 snRNPB PRPF39 PSMD11 PSME1 PSMB4 POMP HYPK PRDX4 SAAL1 FtL HSPB1 MIF Hint1 HPRT1 MRPL38 LDHB FH PTS QPRT COPS3 PRKX CASP6 SVEP1 TUBB2A SEPT7 HSJ-2 2nd Med28 MTF2 CDK5RAP3 Nm23-H1* EBP1 NOB1 UTP6 hnRNP Q# GRSF-1 CDK9 snRNPB PRPF39 n.s n.s n.s n.s HYPK# PRDX4 SAAL1 FtL n.s MIF Hint1 HPRT1 MRPL38* L.
