IRF2BP2b is a nuclear protein in mouse C2
myoblasts that becomes redistributed to the cytoplasm during muscle differentiation 
. The nuclear targeting sequence has not been identified in IRF2BP2 so we generated green fluorescent protein (GFP)-IRF2BP2 fusion constructs to enable the identification of sequences necessary and sufficient for this process. The GFP protein is localized to the cytoplasm of C2
myoblasts whereas a control construct fusing GFP to the TEAD4 transcription factor is localized to the nucleus of C2
myoblasts (). The GFP-IRF2BP2a and GFP-IRF2BP2b fusion proteins were detected in the nucleus of C2
myoblasts (). Since the both isoforms localized to the nucleus, we chose IRF2BP2a for subsequent studies.
Green fluorescent protein (GFP) full-length IRF2BP2 fusion constructs are targeted to the nucleus.
Using convenient restriction sites, we found that the N-terminal sequence of IRF2BP2 encoded by exon 1 (amino acids 1-333) was not imported into the nucleus, whereas a fragment encoded largely by exon 2 (amino acids 333–587) was localized in the nucleus (). A fragment containing amino acids 101–422 was also localized in the nucleus, identifying a nuclear targeting sequence between amino acids 333 and 422. The GFP-IRF2BP2(Ex1) fusion protein was found in cytoplasmic speckles (). To determine the cellular localization of this fragment, we tested for colocalization with a mitochondrial protein (peroxiredoxin 3, PRDX3), a lysosome associated membrane protein (LAMP) and a Golgi body targeted protein Ras-related protein 11 (Rab11) using specific antibodies. We found that the N-terminal fragment of IRF2BP2 is targeted to the lysosome, and is likely degraded ().
Nuclear localization signal is contained within the C-terminal half of IRF2BP2.
Given that the C-terminal fragment of IRF2BP2 is targeted to the nucleus, we sought to identify the minimal nuclear localization sequences (NLS) within this fragment. Bioinformatic analyses () suggested two putative NLS in conserved regions of IRF2BP2, one at the N-terminus of exon 2 encoded sequence we called Ex2N, 357ARKRKP364 and the other and the C-terminus we called Ex2C, 579KVKKERD587. To evaluate whether these sites are functional, the second exon was further divided into three parts, namely 333-422, 422–587, and 575–587, fused with GFP, and then transiently overexpressed in C2C12 myoblasts. Here, the 333–422 fragment was found in the nucleus (), but the C-terminal Ex2C 575–587 fragment or the 422–587 fragment did not. Thus, the Ex2C sequence is not an NLS, nor is a cryptic NLS present in the sequences from amino acids 422 to 587. To authenticate the NLS at Ex.2N, positively charged amino acids were replaced by negatively charged aspartic acids (DDDD) using site-directed mutagenesis (). When expressed in C2C12 myoblasts, the aspartic acid mutant remained in the cytoplasm, confirming that the Ex2N NLS in IRF2BP2 is functional ().
Identification of the NLS within the C-terminal half of IRF2BP2.
A survey of phosphopeptide databases generated from large scale proteomic studies that employed mass spectrometry revealed that IRF2BP2 is phosphorylated at serine 360 (S360), two amino acids downstream of the newly identified Ex2N NLS, in many human cell types (). To test whether phosphorylation of S360 affects nuclear import of IRF2BP2, we mutagenized the serine to either an alanine (S360A) that cannot be phosphorylated or to an aspartic acid (S360D), that carries a negative charge and mimics phosphorylation. The GFP-IRF2BP2 S360A mutant was localized in the cytoplasm of C2
myoblasts. To rule out that the S360A mutant created a cryptic nuclear export signal, leptomycin B (10 nM) was used to block CRM1-dependent nuclear export, but it did not cause nuclear retention of the GFP-IRF2BP2 S360A mutant (Fig. S1
). Thus, cytosolic localization of the GFP-IRF2BP2 S360A mutant likely results from failure to import rather than a CRM1-dependent nuclear export. In contrast, the GFP-IRF2BP2 S360D mutant was located in the nucleus (). During skeletal muscle differentiation of C2
cells, we reported previously that IRF2BP2 is partially relocated to the cytoplasm 
. Thus, we asked whether the S360D mutant would enforce nuclear retention of IRF2BP2 in C2
cells 72 hours after the induction of muscle differentiation. Both the endogenous IRF2BP2 and the GFP fusion protein bearing full-length wild type IRF2BP2 were localized in both the cytoplasm and nucleus, whereas the S360D mutant was exclusively nuclear ().
Serine 360 of IRF2BP2 is phosphorylated in multiple human cell types.
Phosphorylation of serine 360 controls nuclear localization of IRF2BP2.
When co-expressed with the transcription factor TEAD1, we reported previously that IRF2BP2 strongly co-activates a mouse VEGFA promoter in African green monkey kidney CV1 cells, where endogenous IRF2BP2 levels are low
. We also observed nuclear exclusion of the S360A mutant of IRF2BP2 and nuclear retention of the S360D mutant in CV1 cells (). We next tested whether nuclear exclusion of IRF2BP2 would block and whether forced nuclear retention of IRF2BP2 would enhance mouse VEGF promoter activity in the presence of TEAD1. Compared to wildtype IRF2BP2 co-expressed with TEAD1, not only did the alanine mutant fail to co-activate the VEGFA promoter, as expected, but surprisingly the aspartic acid mutant also failed to co-activate the VEGFA promoter in CV1 cells (). We confirmed that TEAD1 and the wild type and the mutant IRF2BP2 proteins were expressed at comparable levels and were of the expected size by immunoblot analysis of transfected CV1 cells (). Thus, although the aspartic acid substitution of serine 360 forces nuclear retention of IRF1BP2, it also appears to disrupt the co-activation function of IRF2BP2.
Despite forced nuclear retention the S360D mutant did not produce a dominantly active form of IRF2BP2.
Currently, we do not know which kinase is responsible for the phosphorylation of S360. From a list of candidate kinases includes protein kinase A (PKA, consensus sequence R-R/K-X-S/T), kinase B (PKB), kinase C (PKC), kinase G (PKG), ribosomal S6 kinase (RSK), extracellular signal-regulated kianses 1/2 (ERK1/2, X-X-S/T-P), calmodulin-dependent protein kinase II (CaMKII, R-X-X-S/T), and Cdc2 (S/T-P-X-R/L), we tested various inhibitors against these kinases to determine whether they would cause cytoplasmic retention of IRF2BP2. However, in no case was IRF2BP2 excluded from the nucleus of C2
myoblasts 24 hours post-treatment (Fig. S2
The diagram shown in summarizes our finding of a single phosphorylation-modulated NLS in IRF2BP2. This sequence is conserved across the different paralogous homologs in both humans and zebrafish ().