|Home | About | Journals | Submit | Contact Us | Français|
Fragile X syndrome is caused by the loss of expression of the fragile X mental retardation protein, FMRP. FMRP is an RNA-binding protein that is highly expressed in neurons and undergoes multiple post-translational modifications including methylation on arginine. FMRP is methylated on the high-affinity RNA-binding motif, the RGG box, at positions 533, 538, 543 and 545 of murine FMRP. To identify the arginines important for FMRP function, we examined their role in polyribosome and mRNA association. We found that arginines 533 and 538 were required for normal FMRP polyribosome association whereas all four arginines played a role in RNA binding, depending on the identity of the RNA. The model G-quadruplex RNA sc1 required arginines 533 and 538 for normal association with FMRP, whereas AATYK mRNA did not. In vitro methylation of FMRP-bearing arginine substitutions inhibited sc1 binding but not AATYK binding. In addition, we found that PRMT1 co-immunoprecipitated with FMRP isolated from cells and that siRNAs directed against PRMT1 led to reduced FMRP methylation. Thus, two lines of experimentation demonstrate that PRMT1 acts on FMRP in cells. In summary, we provide evidence for the important role of the RGG box in polyribosome association. We also demonstrate for the first time that the different arginines of the RGG box are important for the binding of different RNAs. Finally, we show that PRMT1 methylates FMRP in cells, suggesting a model where methylation of the RGG box modulates either the quantity or the identity of the RNAs bound by FMRP.
Fragile X syndrome (FXS) is the most common form of heritable mental retardation, affecting 1/4000 males and 1/8000 females and is due to loss of expression of the fragile X mental retardation protein FMRP (1). Individuals affected by FXS display a range of characteristics including mental retardation, epilepsy and autistic-like features (2). FMRP is encoded by the FMR1 gene and has several distinct domains including a nuclear localization signal and a nuclear export signal, suggesting that FMRP shuttles between the nucleus and the cytoplasm (3–5). FMRP also has two KH domains and an RGG domain (the RGG box) (6–8). These domains have been shown to bind mRNA, and FMRP associates with ~4% of brain mRNAs (9–12). FMRP can be found in messenger ribonucleoprotein (mRNP) complexes and associated with mRNAs bound by multiple ribosomes, hereafter referred to as polyribosomes (13,14). Hence, FMRP is believed to be a translational regulator (15,16).
The RGG box is the primary mRNA-binding domain of FMRP, associating with intramolecular G-quadruplexes with nanomolar affinity (17–21). Recently, the RGG box of FMRP was shown to be unique among the fragile X family members in its ability to recognize G-quadruplex RNAs, suggesting that it plays a non-redundant role in the pathophysiology of the disease (22). Additionally, the RGG box has been shown to be methylated on arginines in cells (23). In vitro methylation assays and metabolic labeling showed that methylation specifically occurs on arginines 533, 538, 543 and 545 (23,24). Because the RGG box is the high-affinity RNA-binding domain, it was not surprising that FMRP lacking the RGG box (ΔRGG) was not distributed normally on polyribosome fractions (25). The reduced association with polyribosomes in the absence of the RGG box is likely due to the inability of FMRP to interact with mRNA efficiently.
Arginine methylation is one of many posttranslational modifications used to regulate protein function. It is a costly modification, requiring approximately 12 ATPs to add a single methyl group to a protein (26). Arginine methylation of proteins other than FMRP has been shown to affect protein–RNA interactions, protein localization and protein–protein interactions (27). The addition of methyl groups to amino acid side chains increases steric hindrance and removes amino hydrogens that might participate in hydrogen bonding (28). Arginine is a likely candidate to modulate interactions with RNA since it was among the three most common amino acids found in RNA-binding sites in surveys of 45 and 32 protein–RNA structures, respectively (29,30). Thus, methylation of arginine could either sterically hinder association with RNA or remove hydrogens that might participate in bonds with the RNA. Protein methylation has been viewed as an irreversible modification until recently when a histone arginine demethylase was described (31).
Protein arginine methyltransferases (PRMTs) are responsible for methylating arginine residues in eukaryotic organisms. In mammals, nine PRMTs have been identified and are named PRMT1 through PRMT9 (26,32,33), and amino acid sequence analysis suggests 10th and 11th PRMTs (34,35). The PRMTs are generally divided into two categories, type 1 and type 2, based on their arginine methylation patterns. Both type 1 and type 2 PRMTs can add a single methyl group to arginine. Type 1 PRMTs (PRMTs 1, 2, 3, 4/CARM1 6 and 8) can add a second methyl group to the same guanidino nitrogen atom of arginine, resulting in asymmetric dimethylation (26). In contrast, type 2 PRMTs (PRMTs 5 and 7) can add two methyl groups to an arginine, but will use different guanidino nitrogen atoms, resulting in symmetric dimethylation (26,32,35). FMRP is both monomethylated and asymmetrically dimethylated, indicating that a type 1 PRMT is responsible for the methylation of FMRP (23). It has been shown in vitro that PRMT1, PRMT3 and PRMT4 methylate FMRP (23,24). In vitro methylation of FMRP reduces its ability to bind G-quadruplex containing RNAs (23,36).
Our goal was to establish that the RGG box is required for normal polyribosome association and to explore the role of the individual arginines. We show here that arginines 533 and 538 are required for normal FMRP polyribosome association; however, arginines 533, 538, 543 and 545—all of which have been shown to be methylated—play varying roles in RNA binding, depending on the RNA target. We also show that PRMT1 methylates FMRP in cells. Together, our data support a model whereby methylation of FMRP by PRMT1 regulates its association with target RNAs.
FMRP has been shown to associate with polyribosomes (37,38), which are a cluster of ribosomes associated with a single mRNA, indicating the translational state of the cell. Polyribosomes are visualized by UV absorption of fractionated lysates at 254 nm (39) (Fig. 1A). Because FMRP is an RNA-binding protein, we hypothesized that FMRP associates with polyribosomes through its RGG box. In fact, a previous study showed that removal of the RGG box (ΔRGG) led to a partial loss of FMRP from polyribosomes and a commensurate accumulation of ΔRGG in the lighter fractions compared with WT FMRP (25). In contrast, a second independent study suggested that the RGG box was not required for polyribosome association (40). Because we were interested in investigating the role of the RGG box on FMRP function, we attempted to resolve this issue. We performed all of our experiments in an immortalized FMR1 knockout fibroblast cell line (STEK), as was done in the first study (25) to eliminate the possibility that the introduced transgene-encoded FMRP would dimerize with the endogenous FMRP (41). We compared the distribution of WT FMRP and ΔRGG on polyribosomes (Fig. 1B). In three independent experiments, we found that a significantly greater proportion of ΔRGG was present in the first two polyribosome fractions containing free protein and mRNPs compared with WT FMRP. We also found a significantly smaller proportion of ΔRGG present in fractions four through eight compared with WT FMRP (Fig. 1B and C), suggesting a loss of ΔRGG from polyribosomes and an accumulation in the lighter fractions. Similar to the earlier study (25), our results suggested that the RGG box was indeed required for normal distribution on polyribosomes.
Arginines 533, 538, 543 and 545 were previously shown to be methylated when their substitution to lysine (533, 538), glutamine (543) and histidine (545) resulted in a 90% reduction in FMRP methylation (23). The glutamine 543 substitution occurred spontaneously during the site-directed mutagenesis reaction (23), whereas histidine 545 is a naturally occurring substitution that was found in a fragile X patient although it did not cause the disorder (42). To determine the importance of these four arginine residues for normal FMRP polyribosome association, we examined the polyribosome profile of FMRP in which arginines 533, 538, 543 and 545 were individually substituted as described earlier and hereafter referred to as 533,538,543,545m (Table 1). Similar to the polyribosome profile of ΔRGG, a significantly greater proportion of 533,538,543,545m was present in fractions two and three compared with WT FMRP, whereas a significantly reduced proportion was present in fractions six, seven and eight compared with WT FMRP (Fig. 1B and C). Thus, arginines 533, 538, 543 and 545 of the RGG box are required for normal polyribosome association of FMRP.
To understand the importance of the individual arginines of the RGG box that are possible methylation substrates, we examined the distribution of FMRP where either arginine 533 or 538 was substituted to lysine. In two independent trials, neither substitution recapitulated the WT FMRP or ΔRGG phenotype (data not shown). To determine whether multiple arginine residues are necessary for normal polyribosome association, we created two FMRP constructs with either the first pair of arginine residues substituted (533,538m; Table 1) or the last pair of arginine residues substituted (543,545m; Table 1) and examined their ability to associate with polyribosomes. In at least three independent trials, 543,545m associated with polyribosomes similar to WT FMRP (Fig. 1B and C), showing no statistically significant difference from WT FMRP. This result suggests that arginines at positions 543 and 545 are not necessary for normal polyribosome distribution. In contrast, we found that arginines 533 and 538 were required for normal polyribosome association because 533,538m significantly deviated from WT by having a significantly greater proportion of FMRP in fraction 2 and a significantly smaller proportion in fraction 8, similar to ΔRGG and 533,538,543,545m (Fig. 1B and C). Similarly, arginines must be present at both residues 533 and 538 for a WT FMRP phenotype. In two independent trials, we examined the distribution of FMRP where either arginine 533 or 538 was the only arginine present and the rest were substituted. A single arginine at position 533 or 538 was not sufficient to confer the WT FMRP phenotype (data not shown). Thus, our data indicate that both arginines 533 and 538 are key for normal polyribosome association, and, under these assay conditions, amino acids at positions 543 and 545 do not need to be arginines for the normal distribution of FMRP on polyribosomes. In fact, residue 543 does not even need to be positively charged since it was substituted to a glutamine. It is worth noting that although 533,538m was significantly different than WT, substitution of arginines 533 and 538 was not equivalent to the removal of the entire RGG box (Fig. 1B and C), suggesting that additional amino acids in the RGG box play a role in polyribosome association. In summary, these experiments demonstrate that we are able to assess the contribution of arginines in the RGG box to normal polyribosome association and that arginines 533 and 538 significantly contribute to polyribosome association.
Polyribosomes are mRNAs associated with multiple ribosomes, suggesting either active or regulated translation. It follows that if FMRP is no longer able to bind a subset of mRNAs efficiently, particularly those bearing G-quadruplex structures, there will be less FMRP associated with polyribosomes. To test the hypothesis that arginines 533 and 538 are required for normal polyribosome association because they are required for RNA binding, we began by immunoprecipitating WT FMRP, ΔRGG and the arginine substitutions from transfected cells and performing RT–PCR to detect changes in RNA association. Unfortunately, we found it difficult to obtain reproducible results, i.e. we found RNAs that were predicted to depend on the RGG box, like the FMR1 mRNA (18,43), sometimes co-immunoprecipitating with ΔRGG (data not shown). This result could be explained by the given mRNAs binding to multiple sites along FMRP, associating with domains in addition to the RGG box that have also been reported to bind mRNAs, like the N-terminus (44) and the KH domains (45,46) of FMRP. Thus, we changed our strategy to specifically examine the requirements for RNA binding to the RGG box of FMRP. We used an RNA capture assay described earlier (23,47) to examine the ability of WT FMRP, ΔRGG and the arginine substitutions to bind the high-affinity RGG box ligand, which is a stem loop G-quadruplex structure termed sc1. sc1 was identified in SELEX experiments as binding the RGG box of FMRP with high affinity (17). In fact, mRNAs bearing sc1-like sequences were found to co-immunoprecipitate with FMRP (11,48), supporting the hypothesis that sc1 reflects the cellular RNAs that associate with FMRP.
We measured the ability of biotinylated sc1 G-quadruplex RNA to capture radioactively labeled FMRP synthesized in vitro in rabbit reticulocyte lysate (23,47). As a control, we also used the sc1 mutant in which mutation of two Gs disrupts G-quadruplex formation and leads to loss of FMRP association (17). As expected, removal of the RGG box resulted in no capture by sc1 (Fig. 2). In addition, substitution of all four arginines (533,538,543,545m) resulted in only a small amount of the protein being captured (15%) compared with WT. Thus, arginines 533, 538, 543 and 545 are important for the association with sc1. Substitution of 533 and 538 (533,538m) also led to a phenotype distinct from WT FMRP; 23% of 533,538m was captured by sc1 compared with WT FMRP (Fig. 2). However, in contrast to the polyribosome distribution examined in Figure 1, arginines 543 and 545 also contributed to sc1 binding because their substitution led to only 44% of the 543,545m being captured compared with WT FMRP (Fig. 2). The results of three independent experiments are shown in Figure 2C, where all of the constructs bound sc1 RNA significantly less than WT FMRP. In addition, the capture results of 533,538m are significantly different from the results obtained with 543,545m (Fig. 2C), supporting the hypothesis that residues 533 and 538 are more important for the association with sc1 than 543 and 545. Thus, RNA capture using the sc1 ligand reflects the polyribosome data where arginines 533 and 538 are important for a normal phenotype. However, in the RNA capture experiment, arginines 543 and 545 also played a role in sc1 binding, an observation that was not evident from the polyribosome experiments.
sc1 is an in vitro ligand for the RGG box of FMRP (17). In a companion study, the first extensive list of brain RNAs that co-immunoprecipitated with the FMRP complex was described (11). They found that 67% of the mRNAs that both co-immunoprecipitated with FMRP and had an FMRP-dependent distribution on polyribosomes also contained a putative G-quadruplex (11). AATYK was the second most abundant co-immunoprecipitating brain mRNA (11) and encodes a putative G-quadruplex sequence, N(0–7)WGG-N(1–4)WGG-N(1–4)WGG-N(1–4)WGG, where N is any nucleotide and W is a U or A (17). Although it has not been experimentally established that the AATYK sequence forms a G-quadruplex, the region is guanine-rich. In fact, FMRP has been shown to bind the amyloid precursor protein (APP) mRNA through a guanine-rich region that is not predicted to form a G-quadruplex (49). This result was not surprising as in early characterization studies, FMRP was demonstrated to preferentially bind homopolymers of guanine; further, homopolymer binding was dependent on the RGG-box-containing C-terminus (7,8,47). Because we established that AATYK binds FMRP in an earlier study (23), we used it here again to examine whether the arginine substitutions interacted with AATYK the same as sc1. Similar to the results in Figure 2, removal of the RGG box resulted in the greatest loss of capture by AATYK compared with WT (21% of WT). Interestingly, substitution of all four arginines (533,538,543,545m) resulted in 67% of the protein being captured compared with WT (Fig. 3A). Although significantly different than WT, capture of 533,538,543,545m was also different from ΔRGG, suggesting that other amino acids in the RGG box are important for the association with AATYK (Fig. 3C). Surprisingly, association of FMRP with AATYK mRNA was not dependent on arginines 533 and 538, as their substitution had no effect on the efficiency of capture compared with WT (Fig. 3, 108% is not statistically different from 100%, P = 0.610). Thus, arginines 533 and 538 are not important for the association with AATYK mRNA. We cannot rule out that a positive charge is unimportant as the substitutions introduced lysines at positions 533 and 538 in 533,538m. Substitution to glutamine would examine the role of charge in association with AATYK. Like sc1, association of AATYK with arginines 543 and 545 accounted for approximately half of the binding, as their substitution resulted in a significant reduction in capture compared with WT (54% of WT). Although the values for ΔRGG, 533,538,543,545m and 543,545m are significantly different from WT, there was no significant difference between the latter two mutants (67 and 54%). Thus, binding to AATYK requires arginines at positions 543 and 545 but not at 533 and 538. In summary, binding to sc1 requires arginines at positions 533 and 538 and less so at 543 and 545, although binding to AATYK requires arginines at positions 543 and 545, suggesting that different RNAs have distinct molecular requirements for association with FMRP.
It has been shown in vitro that PRMT1, 3 and 4 methylate FMRP (23,24). We also showed that PRMT1 methylated arginines 533, 538, 543 and 545 in vitro (23). To examine methylation of 533,538m and 543,545m by PRMT1, we transfected COS-7 cells with WT FMRP, ΔRGG and the three arginine substitution constructs. After immunoprecipitating FMRP, we performed an in vitro methylation reaction using PRMT1. As expected, ΔRGG and 533,538,543,545m showed little-to-no methylation, whereas WT FMRP; 533,538m and 543,545m were methylated (Fig. 4A). Interestingly, WT FMRP treated with BSA instead of PRMT1 was also methylated, although to a lesser degree than when incubated with PRMT1, suggesting that FMRP co-immunoprecipitated with an endogenous PRMT and that exogenous addition of PRMT1 increased its methylation.
PRMT1 has already been shown to associate with its substrates (50), so we examined whether PRMT1 was the PRMT that co-immunoprecipitated with FMRP. FMRP was immunoprecipitated from L-M(TK−) cells that stably expressed Flag-FMRP (WT) and was probed with an antibody to PRMT1. In three independent experiments, a band of the correct molecular mass was detected by the PRMT1 antibody in lysates from both the untransfected L-M(TK−) cells (VC) and WT cells, and this band was also present in the WT immunoprecipitation (IP) lane (Fig. 4B). In the lower panel of Figure 4B, the immunoprecipitated FMRP is shown in the Flag IP from WT but not from VC cells. We did not detect FMRP in the input lysate because it is not overexpressed. Together, our data suggest that enzymatically active PRMT1 associates with FMRP in cells.
We originally chose to examine the role of arginines 533, 538, 543 and 545 on FMRP function because each of these RGG box arginines is methylated (23). Arginines 533 and 538 were important in binding sc1 RNA, as these substitutions showed the greatest loss of sc1 binding compared with WT FMRP (Fig. 2). Arginines 543 and 545 played a smaller role in sc1 binding (Fig. 2). To determine the consequence of methylation of those arginines on RNA association, we examined the effect of in vitro methylation by PRMT1 of 543,545m and 533,538m on its ability to be captured by sc1. Similar to our earlier studies with WT FMRP (23), we found that in vitro methylation of 543,545m, which has arginines at positions 533 and 538, led to loss of RNA binding (Fig. 5). Likewise, in vitro methylation of 533,538m, which has arginines at positions 543 and 545, led to decreased RNA binding (Fig. 5). These data suggest that methylation of FMRP at arginines 533, 538, 543 and 545 inhibits RNA association through the RGG box.
We previously showed that methylation of FMRP inhibits association with AATYK (23), and substitution of arginines 543 and 545 reduced the ability of FMRP to be captured by AATYK (Fig. 3). We next examined whether methylation by PRMT1 of 533,538m, which has arginines at positions 543 and 545, affects association with AATYK mRNA. Because capture with AATYK was not as efficient as with sc1, we chose to methylate 35S-methionine radiolabeled FMR protein. We independently verified that we were able to methylate these FMR proteins (Fig. 5 and data not shown). In four independent experiments, we saw no statistically significant change in the capture of 533,538m after methylation (Fig. 6). Additionally, we saw that methylation of 543,545m, which has arginines at positions 533 and 538, had little-to-no effect on capture by AATYK (Fig. 6), which is not unexpected as arginines 533 and 538 played no role in association with AATYK (Fig. 3). This result suggests that the interaction between FMRP and AATYK is not regulated by methylation of arginines 533 and 538 or 543 and 545. Finally, similar to the results shown in Figure 3, we observed a significant difference between the ability of unmethylated 533,538m and 543,545m to be captured by AATYK (Fig. 6B).
Although PRMT1 has been shown to methylate FMRP in vitro and our data show that PRMT1 associates with FMRP (Fig. 4B), PRMT1 has not yet been demonstrated to methylate FMRP in cells. To determine whether PRMT1 is responsible for methylating FMRP in cells, we treated COS-7 cells with siRNAs against PRMT1 and measured endogenous FMRP methylation by labeling with 3H-methyl methionine in the presence of protein synthesis inhibitors as described previously (23,51). We expected that knockdown of PRMT1 would result in reduced FMRP methylation. Initially, we saw no effect on FMRP methylation (data not shown), likely due to the fact that we could not eliminate enough of the PRMT activity by reducing its expression using siRNAs. Similarly, when we examined the methylation state of FMRP in a PRMT1 knockdown cell line, which still had some PRMT1 activity (52), we did not see a loss of FMRP methylation (data not shown). We then tried simultaneously over-expressing FMRP in addition to administering siRNAs. We found that an ~80% reduction of PRMT1 expression reproducibly reduced methylation of FMRP by 34% in COS-7 cells over-expressing FMRP (Fig. 7A). This observation was recapitulated in HeLa cells, where a reduction in PRMT1 expression yielded a 48% reduction in FMRP methylation (Fig. 7B). We conclude that PRMT1 plays a role in methylating FMRP in cells.
In this study, we show that specific arginines within the RGG box are required for normal polyribosome association and RNA binding. Arginines 533 and 538 are required for normal polyribosome association (Fig. 1), whereas all four arginine residues play a role in RNA binding (Figs 2 and and3).3). Arginines 533 and 538 are required for sc1 binding but not AATYK association (Figs 2 and and3),3), suggesting for the first time that different arginines of the RGG box are involved in binding different target RNAs. Additionally, we show that methylation of arginine residues 533 and 538 or 543 and 545 leads to loss of sc1 RNA binding (Fig. 5). We also provide the first evidence that FMRP associates with PRMT1 (Fig. 4) and that PRMT1 methylates FMRP in cells (Fig. 7).
The dependence on different arginines for binding to sc1 and AATYK was unexpected. sc1 has been experimentally demonstrated to form a G-quadruplex, and AATYK encodes a guanine-rich configuration of nucleotides that is predicted by the RNABob algorithm to form a G-quadruplex (17); thus, we hypothesized that sc1 and AATYK would associate similarly with the RGG box. Our observation that they did not was surprising and suggests the alternative hypothesis that sc1 and AATYK form different secondary structures that interact distinctly with the RGG box. We would predict that AATYK is not bound through a G-quadruplex structure, but rather through a G-rich region analogous to the APP mRNA. Similar to the identification of the FMRP-binding site in the APP mRNA (49), it would be interesting to identify the exact sequence in the AATYK mRNA that associates with FMRP and then perform biophysical methods to query its secondary structure (19–21). This approach was used to demonstrate that another FMRP-co-immunoprecipitated mRNA, semaphorin 3A, forms a G-quadruplex that is required for FMRP binding (21).
Interestingly, methylation of arginines 533 and 538 encoded by 543,545m or 543 and 545 encoded by 533,545m had no effect on AATYK association (Fig. 6). Since arginines 533 and 538 played no role in AATYK binding, this result was not surprising; however, since 543 and 545 did participate in RNA capture (Fig. 3), it was unexpected that their methylation had no effect, especially since we showed earlier that methylation of WT FMRP inhibits AATYK binding (23). One explanation for this result is that arginines 533, 538, 543 and 545 are not as important for AATYK binding as they are for sc1 binding [note that the absence of all four arginines only yields a 33% decrease in capture (Fig. 3) compared with an 85% decrease in capture for sc1 (Fig. 2)]. Other residues in the RGG box clearly are important for the association with AATYK because its removal leads to a nearly 80% loss of capture (Fig. 3). Thus, PRMT1 may methylate other arginines in the RGG box that collectively inhibit binding of AATYK by the WT protein.
This study supports the previous observation that the RGG box is required for normal polyribosome association (25). A different study showing the opposite result where the RGG box was not required for polyribosome association may be due to the phenotype being subtle, especially when compared with the KH domain deletions and the I304N mutation (25,40); our results also required multiple trials and examining smaller volume fractions within each polyribosome experiment to determine a statistically significant difference between FMRP and ΔRGG in the early and late fractions. The latter, conflicting study also used a neuronal cell type that expresses endogenous FMRP (40). Dimerization of endogenous FMRP with transgene may obscure the reduction in transgene polyribosome association.
The polyribosome experiments in Figure 1 demonstrate that arginines 533 and 538 are important for normal FMRP association, yet the RNA capture experiments demonstrated that arginines 533 and 538 are important only for sc1 association and not for AATYK binding. Further, arginines 543 and 545, which play no discernible role in polyribosome association, interacted with both RNAs. A possible explanation for this disparity is that the RNA capture is a more sensitive assay, i.e. polyribosome association does not allow elucidation of somewhat compromised RNA binding like the 44% of 543,545m captured by sc1 compared with the severely compromised binding of 533,538m (24% captured). An alternative explanation is that additional factors, like the methylation state of FMRP, that occur in cells but not in the capture experiments may contribute to polyribosome association. For example, if methylation inhibits association with RNAs and arginines 543 and 545 are constitutively methylated, then in cells, they are functionally unavailable for RNA binding; therefore, their substitution would have no effect on polyribosome association.
We have also provided evidence that FMRP is methylated by PRMT1 in cells. Because we were unable to achieve complete loss of FMRP methylation using siRNAs against PRMT1, it is possible that other PRMTs methylate FMRP, as has been suggested (24). At this point in time, we do not know whether methylation occurs to inhibit further RNA binding in the cell or whether it occurs on selective arginines to control the identity of the RNAs bound. Regarding the first possibility, the resulting loss of RNA binding by the RGG box may lead to the KH domains becoming more important in RNA binding, altering the potential targets for FMRP RNA binding. It has also been shown that protein methylation occurs in a distributive manner, so proteins can be partially methylated and released from the enzyme to re-associate later and become fully methylated (53). With this in mind, a second possible scenario is that partial methylation of FMRP by PRMT1 alters the identity or quantity of RNA associated with FMRP. Characterization of the methylation state of the individual arginines of FMRP in vivo will give insight into this question. Cellular FMRP is not completely methylated, as it can be further methylated in vitro (Fig. 4) (23). Thus, only a portion of the FMRP present in a cell is methylated at any one time. It will be informative to determine whether FMRP exists in two pools, completely methylated and unmethylated, or whether it exists as a collection of proteins with distinct methylated arginines. If the latter were true, methylation would allow the cell to express FMRPs with varying RNA-binding abilities.
All media and supplements were purchased from GIBCO-BRL, unless otherwise stated. All cultured cell lines were grown at 37° in 5% CO2 in Dulbecco's modified Eagle's medium containing 10% fetal calf serum supplemented with 10 mm HEPES, 1× non-essential amino acids (Biowhittaker) and 100 units/ml of penicillin/streptomycin (complete media). L-M(TK−) cells expressing the empty RSV.5 vector or Flag-FMRP were maintained in complete media supplemented with 6 μg/ml of mycophenolic acid and 252 μg/ml of xanthine (MAX) (Sigma, St Louis, MO, USA) (54). The transient transfections were performed using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) following manufacturer's instructions. The STEK FMR1 KO cell line was provided by Dr Eduard Khandjian (55). The EGFP-Flag-FMRP construct used to generate the arginine substitution constructs along with the mutagenesis primers used to make each substitution are described in Stetler et al. (23). For synthesizing Flag-FMR protein from rabbit reticulolysate (RRL), the EGFP-Flag-FMRP, EGFP-Flag-ΔRGG and EGFP-Flag-tagged arginine substitution constructs were used as the templates for PCR reactions using primers described in Stetler et al. (23). The Flag-FMRP PCR products were then ligated into pGEMT-Easy using the TA cloning protocol (Invitrogen).
The anti-FMRP antibody, MAb1C3, was obtained from Dr Jean-Louis Mandel at the Institute of Genetics in Ilkirch, France, and was used as a hybridoma supernatant at a 1/10 dilution. The anti-FMRP antibody, KI, was provided by Dr Andre Hoogeveen at Erasmus University in Rotterdam and used at a 1/500 dilution. The PRMT1 antibody was purchased from Abcam (Cambridge, MA, USA) and used at a 1/500 dilution. The 7G1–1 antibody is described in Brown et al. (11). The eIF5 antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA) and used at 1/10 000 dilution. Antibody reactivity was visualized using an anti-mouse HRP conjugate (KPL, Gaithersburg, MD, USA) for MAb1C3 and anti-PRMT1 or anti-rabbit HRP conjugate (Amersham) for anti-eIF5 and KI and developed with ECL (Amersham).
COS-7 or HeLa cells were plated at 3 × 106 cells/100 mm dish and transiently transfected with the FMRP constructs. Twenty-four hours later, the cells were expanded into 150 mm dishes and allowed to grow 24 h before metabolic labeling with [3H-methyl]-methionine (Amersham, Piscataway, NJ, USA) in the presence of translation inhibitors was performed as described in Stetler et al. (23). The supernatant was precleared with the anti-class I antibody 16-1-11N (ATCC) and then FMRP was immunoprecipitated with antibody 7G1-1. IPs were washed twice in ice-cold lysis buffer, resuspended in sample buffer, boiled for 5 min and resolved on a 7.5% SDS–PAGE. For autoradiography, the gels were soaked for 20 min in 10% acetic acid and 20% methanol, 20 min in double-distilled water and 30 min in Fluor-Hance (Research Products International, Mount Prospect, IL, USA) and then dried at 75° on a BioRad gel dryer.
STEK cells were plated at 1 × 106 cells/150 mm dish and transiently transfected with the FMRP constructs. Twenty-four hours later, each dish was expanded into three dishes and harvested 24 h later. Before the cells were harvested, they were treated for 15 min with cycloheximide (100 μg/ml). Linear 15–45% sucrose gradients were prepared with buffer containing 100 mm KCl, 20 mm Tris (pH 7.5), 5 mm MgCl2 using a gradient maker (BioComp). Cells were washed in PBS, lysed in the buffer described earlier supplemented with 0.3% Igepal CA-630 and centrifuged for 10 min at 20 000× g at 4°C. Postnuclear supernatants were overlaid on the gradient and centrifuged for 75 min at 188 000g at 4°C. Each gradient was fractionated into 650 μl fractions by bottom displacement using a gradient fractionator (Isco) with the ribosomal profile monitored at 254 nm.
Flag-FMRP constructs expressed in the pGEMT-Easy vector were translated in an RRL kit (Promega, Madison, WI, USA) as per manufacturer's instructions. In the binding experiment, 3–5 μl of the FMR protein reaction was used, and the experiments were performed as described previously (23). For PRMT1-labeled FMR protein and mock-treated FMR protein, 25 μl of each reaction was used in the binding experiments, the amounts of the other reagents scaled up to match the larger sample volume and FMR protein was labeled as described subsequently.
FMR protein was synthesized in an RRL or immunoprecipitated with antibody 7G1-1 from ~5 × 106 cells. FMR protein was labeled for 1 h at 37°C with 1.0 μg of PRMT1 and 0.5–2 μCi of [3H]-SAM as described previously (23).
In our study, 0.5−1.0 × 109 cells were lysed and Flag-FMRP was immunoprecipitated with 50 μl of mouse anti-flag M2 agarose (Sigma). On a 12% SDS–PAGE, 20 μg of lysate was run along with the IP reactions, and western blotting was performed using anti-PRMT1.
This work was supported by the National Institutes of Health, University of Illinois at Urbana-Champaign Developmental Psychobiology and Neurobiology Training Grant (2T32HD007333-21 to E.B.) and the Spastic Paralysis Research Foundation of the Illinois-Eastern Iowa District of Kiwanis International (to S.C.) and in very small part by the National Institute of Child Health and Human Development (HD41591 to S.C.).
We would like to thank fellow laboratory members, Dr Deepa Venkitaramani and Dr Craig Mizzen, for their critical reading of this manuscript.
Conflict of Interest statement. None declared.