Regulation of p68 RNA Helicase by Calmodulin and Protein Kinase C*

Human p68 RNA helicase is a nuclear RNA-dependent ATPase that belongs to a family of putative helicases known as the DEAD box proteins. These proteins have been implicated in aspects of RNA function including translation initiation, splicing, and ribosome assembly in a variety of organisms ranging from Escherichia coli to humans. While members of this family are believed to function in the manipulation of RNA secondary struc- ture, little is known about the regulation of these en-zymes. By immunological methods and sequence com- parison, we have found that p68 possesses a region of sequence similarity to the conserved protein kinase C phosphorylation site and calmodulin binding domain (also known as the I& domain) of the neural-specific proteins neuromodulin (GAe-43) and neurogranin (RC3). We report that p68 is phosphorylated by protein kinase C in vitro and binds calmodulin in a Ca2+- dependent manner. Both phosphorylation and calmodulin binding inhibited p68 ATPase activity, suggesting that the RNA unwinding activity of p68 may be regulated by dual Ca2+ signal transduction pathways through its I& domain. Neuromodulin (GAP-43) and neurogranin (RC3) are neural-specific calmodulin (CaM)' binding proteins shown to be phosphorylated by protein kinase C (PKC) in vitro and in vivo. Although the overall sequence homology between these proteins is low, they share a nearly

Human p68 RNA helicase is a nuclear RNA-dependent ATPase that belongs to a family of putative helicases known as the DEAD box proteins. These proteins have been implicated in aspects of RNA function including translation initiation, splicing, and ribosome assembly in a variety of organisms ranging from Escherichia coli to humans. While members of this family are believed to function in the manipulation of RNA secondary structure, little is known about the regulation of these enzymes. By immunological methods and sequence comparison, we have found that p68 possesses a region of sequence similarity to the conserved protein kinase C phosphorylation site and calmodulin binding domain (also known as the I& domain) of the neural-specific proteins neuromodulin (GAe-43) and neurogranin (RC3). W e report that p68 is phosphorylated by protein kinase C in vitro and binds calmodulin in a Ca2+dependent manner. Both phosphorylation and calmodulin binding inhibited p68 ATPase activity, suggesting that the RNA unwinding activity of p68 may be regulated by dual Ca2+ signal transduction pathways through its I& domain.
Neuromodulin (GAP-43) and neurogranin (RC3) are neuralspecific calmodulin (CaM)' binding proteins shown to be phosphorylated by protein kinase C (PKC) in vitro a n d i n vivo. Although the overall sequence homology between these proteins is low, they share a nearly identical region of 19 amino acids, corresponding to the overlapping site of CaM binding and PKC phosphorylation, also referred to as the I& domain (1)(2)(3)(4)(5)(6). CaM binding at the I& domain of neuromodulin has been shown to decrease the rate of phosphorylation by PKC, and phosphorylation conversely prevents binding to CaM, leading to the hypothesis that reversible phosphorylation of neuromodulin at the I& domain may regulate free CaM levels in neurons (reviewed in Ref. 7). It has recently been reported that several other proteins contain sequences related to the I& domain (reviewed in Refs. 6 a n d 8  (9). In addition, p190 (myosin-V) binds CaM at tandem repeats of the I& motif (10). It has been proposed that this sequence represents a common descriptor for Ca2+ regulation via CaM or other Ca2+-binding proteins (8).
In this study, we identify the p68 RNA helicase as a member of the I& domain-containing family. p68 belongs to a rapidly growing family of proteins (DEAD box family) which share a core region of highly conserved sequence motifs, including sites for ATP binding and hydrolysis (reviewed in Ref. 11). These proteins are involved in diverse cellular processes including RNA splicing, translation initiation, ribosome assembly, and cell growth and division. They are hypothesized to regulate RNA structure and function by unwinding double-stranded RNA or by promoting other ATP-dependent conformational changes. Members of the DEAD box family include the prototypic eukaryotic translation initiation factor eIF-4A and the yeast splicing factor Prp5 (11). Human p68 was first identified by its immunological cross-reactivity with the viral helicase, SV40 large T antigen (12). While the precise function of p68 is unknown, it has been shown to exhibit RNA-dependent ATPase activity and RNAunwinding activity in vitro (13)(14)(15). The presence of an I& domain within the primary sequence suggests that the RNA helicase activity of p68 may be regulated by Ca2+ via CaM andor by PKC. Here we report that p68 is a substrate for PKC in the absence of RNA and that phosphorylation inhibits RNA stimulation of p68 ATPase activity. In addition, CaM binds to p68 in a Ca2+-dependent manner and also blocks p68 ATPase activity. We propose that p68 helicase activity may be regulated by PKC and Ca" through its I& domain.
Protein Isolation and Antibody Production-The cDNA encoding neurogranin was subcloned into the BamHVEcoRI sites of the bacterial expression plasmid pGEX-2T (Pharmacia Biotech Inc.). Glutathione S-transferaseheurogranin fusion protein was purified according to the method of Smith and Johnson (17). Rabbit polyclonal antibody was raised against the fusion protein, and IgG was isolated on protein Aagarose (Pierce) by standard methods. Calcineurin, protein kinase C, and CaM were purified as described (18,19).
Western Blot Using Anti-neurogrunin Antibody-Adult rat brains were homogenized in 40 l l l~ Tris-HC1, pH 7.5,l m M EDTA, 1 m M PMSF, 1 m dithiothreitol, and the homogenate was fractionated by centrifugation at 14,000 x g for 30 min at 4 "C. CHO cells were scraped from 10-cm Petri dishes and pelleted at 2000 x g for 10 min. The pelleted cells were homogenized in 10 m M Tris-HC1, pH 7.5,1 m EDTA, 1 m M PMSF on ice. Unbroken cells and nuclei were removed by centrifugation at 4000 x g for 2 min, and the cell extracts were fractionated by centrifugation at 100,000 x g for 30 min at 4 "C. The supernatants were collected, and protein concentration was assayed by the BCA method (Pierce). Proteins were separated by 12% SDS-PAGE and transferred to nitrocellulose. The membrane was treated with 3% bovine serum albumin in phosphate-buffered saline containing 0.2% Tween 20 to block nonspecific binding sites and incubated overnight at 4 "C with total rabbit anti-neurogranin IgG (3 pg/ml). Bound antibody was detected using alkaline phosphatase-conjugated secondary antibody (Cappel) according to the manufacturer's instructions (Bio-Rad). incubating the total IgG fraction with a molar excess of a biotinylated 16-amino acid peptidr corresponding to the CaM binding domain of neuromodulin ( 2 ) and immobilized avidin (Pierce) for 60 min at room temperature. The rrsin was pelleted. and the depleted I d ; pool was used for Western analysis as ahove.
I'cyfirlr S(~/uc,nc.ing ofpfiR"PC12 cells ( 5 x 10" cells) were lysed with phosphate-huffed saline, 1 msl EDTA. 1% Nonidet P-40,0.1% SDS for 10 min on ice. The rxtrnct was clarified by centrifugation at 14,000 x g for 10 min and precleared with protein A-agarose IOncogene Science) for 30 min at 4 "C. p68 was immunoprecipitated hy rotation at 4 "C with anti-neurogranin IgG and excess protein A-agarose overnight. The immunoprecipitates were washed. separated by 10% SDS-PAGE, and electrohlotted to nitrocellulose. After detection with Ponceau S, p6X was excised and digested in s i / u with trypsin 120). Cleavage fragments were separated on nn Applied Riosystems model 172 microhore high pressure liquid chromatograph and subjected to amino acid sequencing using an Applied Riosystems model 477A protein microsequencer.
I3io/irtyln/d Cnlntodrrlirt Rinding"PC12 cells ( 3 x 10" cells) were lysed with 50 mhl Tris-HCI, pH 8.0, 5 m!d EDTA, 150 mM NaCI, lC5 Nonidet P-40. 1 mh! dithiothreitol. and 1 mM PMSF for 30 min on ice and clarified hy centrifugation a t 14,000 x g for 10 min. p68 was immunopfecipitated with the monoclonal antihody PAb204 as described (15). resolved hy 10%. SDS-PAGE, and transferred to PVDF membrane. Proteins were detected by Coomassie Blue staining as described previously (21). CaM binding was assayed according to Rillingsleyef al. (22), using 4 p g h l hiotinylated CaM (Life Technologies, Inc.1 and alkaline phosphatase-conjugated avidin (Pierce]. Phosphoryla/ion of p6R-p68 was immunoprecipitated from PC12 cells with PAb204 as described above. Samples were treated with or without RNase A (0.1 mg/ml) in lysis buffer for 1 h at 4 "C, and heads were washed into PKC phosphorylation buffer. Phosphorylation assays were carried out at 30 "C with PKC in the presence or absence of effectors for times indicated according to Ape1 e1 af. (3). Reactions were stopped by the addition of Laemmli buffer, resolved by SDS-PAGE. and subjected to autoradiography.
ATPasr Assay"p6X was immunoprecipitated with PAh204 and treated with or without RNase A as described above. The samples were washed, and ATPase activity was determined using 0.1 mhf [y-"2PlATP as descrihed (15) with the following modifications. Assays were performed in 120 p1 of ATPase buffer containing 30 pl of protein G-agnrose heads bound with the PAb204-p68 complex. After a 30-min incubation at 37 "C. 1 ml of 7% activated charcoal in 50 mM HCI. 5 mM HJ'O, was added to stop the reaction. The samples were centrifuged at 14,000 x g for 10 min, and the supernatant was counted for free ."P,. Where indicated. ATPase activity was measured in the presence of exogenous total RNA from rat adrenal medulla (80 pg/ml) isolated as described (23).

RESULTS AND DISCUSSION
An Antibody against Neurogranin Cross-reacts with p6R RNA Helicase-In an attempt to identify proteins containing the IQ domain, polyclonal antibodies were raised against neurogranin as a fusion protein with glutathione S-transferase. Western blot analysis of rat brain extracts using anti-neuro-TAIUX I

Sequence of prpfidrn grnrrnlrd hs In s r / n digrslron of liX.hI)n proldvn
The 68-kDa protein was immunoprecipitatcd from PC12 c(.IIs. rlrctrohlotted, and digested in s i f u with trypsin. Six of thr tryptic prptidrs separated by high pressure liquid chromatography wrw sc.qurncrtl and are listed with the corresponding position in the. drduced human p6X RNA helicase cDNA srquence (13, 24 1. granin I& showed cross-reactivity with nruromodulin and a higher apparent molecular mass species of approximatcly 68 kDa, also detectable in CHO cell extracts (Fig. 1). Thr crossreactivity with neuromodulin was presumably dur to thr conserved 19-amino acid sequence which contains thr IQ domain. the site of CaM binding and phosphorylation by PKC. Thr only sequence homology between neurogranin and neuromodulin lies within this common CaM binding domain. Sprcific removal of antibodies against the IQ domain of neurogranin and nruromodulin eliminated the cross-reaction writh nruromodulin and the 68-kDa protein (Fig.  1). indicating that a similar epitope may reside within the higher molrcular mass protein. This was further supported by the finding that immunoprecipitation of the 68-kDa protein from PC12 cclls with anti-ncurogranin I& was blocked in the prrsence of a molar excess of peptide corresponding to the IQ domain (data not shown).
In order to identify the cross-reacting protrin, tbr 68-kDa protein was immunoprecipitated from PC12 cells and subjectrd to in situ tryptic digestion. Subsequent sequcnce analysis of six of the resultant peptides revealed simificant homology with the deduced amino acid sequence of the human p68 RNA helicase (Table I ) (13,24). Western blot analysis and immunoprecipitation using the p68-specific antibody PAb204 (12) confirmed the identity of this protein (data not shown). Furthrr examination of the primary sequence indicated that p68 contains a six-amino acid regwn, ""IQTSFR, corrrsponding to IQASFR found within the consensus domains of neurogranin and neuromodulin. Phosphorylation of t h r nrurospecific proteins occurs on the serine within this region (3, 5 ) . In p6R this site is COOH-terminal to the epitope for PAh204 sharrd with the SV40 large T antigen (13) and presumably constitutes the site of recognition for the anti-neurogranin antibody.
p68 RNA Helicase Binds CaM-The presence of the IQ domain in p68 suggested that the enzyme may be regulated by Ca"-dependent processes via CaM and/or PKC. Binding of CaM to p68 was examined with a CaM overlay procedure in which p68 was immobilized on PVDF membranes (Fig. 2). Both p68 and the CaM-dependent phosphatase calcineurin, used as a positive control, bound biotinylated CaM in a Ca2+-dependent fashion, as evidenced by the lack of binding in the presence of EGTA. This is in contrast to neuromodulin and neurogranin, which bind CaM preferentially in the absence of Ca2+ (5,25) but similar to calcium vector protein target, which exhibits Ca2+dependent binding to its target (9). p 6 8 Is Phosphorylated by PKC in the Ahsence of RNA-The conserved serine within the IQ domain of p68 corresponding to the phosphorylation site in neuromodulin and neurogranin led us to examine the phosphorylation of p68 by PKC in vitro (Fig.  3 A ) . When immunoprecipitated from PC12 cells without further treatment, p68 was a poor substrate for PKC. However, p68 co-immunoprecipitates with RNA when cells are lysed at physiological salt concentrations (15). Removal of bound RNA with RNase A resulted in a marked increase in Ca2+-and phos-  Fig. 3. The samples werr wnshed, nnd p6X wns incuhntrd with PKC in the presence or nhsencr of (h'* nnd phosphnliplds fnr 60 min a t 30 "C. After additional washing of thca hcwls. ATPnw nrtivlty was determined in the presence or tlhscmcr nf exoprnous totnl RNA n~ described under "Experimrntal Procedures." H . inhlhition hy ('nS1 hlnding. RNase-treated pfi8 was preincuhnted with AT1';lsr huffear containing 1 msr CaCI, in the presence or zlhwnce of CnM 1 3 p v ) for 2 h and assayed for ATPase activity as ahove. All rrnctinns wrrr prrformrd In triplicate; each data point 1s the mean t S.1). l<~sults shown arr rrprvsentative of at least four rxperiments correctrd for spnntnnrous ATP hydrolysis.
Inclusion of CaM in the assay blocked phosphorylation of p68. with no effect on PKC autophosphorylation. CaM also inhibits the PKC phosphorylation of hoth neuromanin and nruromodulin (2,5). In addition, a 20-amino acid peptide corresponding to residues 549-568 of p68 was phosphorylated in a (:a"-and phospholipid-dependent manner hy PKC (data not shown).
These data indicate that a site of PKC phosphorylation of p68 lies within the conserved IQ domain; it is not known whcther additional sites may be suhject to PKC phosphorylation, Recause p68 used in these assays was immunoprccipitatcd. stoichiometry of phosphorylation could not hr drtrrmincd. Phosphorylation of p68 was saturable, rraching maximal IrvrI by approximately 1 h (Fig. 3R).
Phosphorylation and CaM Binding Inhibit ATl'nsr Actir~itv nf p68-Since p68 helicase is a substrate for PKC in the ahscnce of RNA in vitro, it was of interest to detcrminr whcthrr phosphorylation affected the RNA-dependent ATPase activity of the enzyme. Immunoprecipitated PC12 p68 exhibited ATPase activity due to the presence of hound RNA which co-immunoprrcipitates with the protein ( 1 5 ) (Fig. 4A ). ATPasr activity was almost completely abolished by RNase trcatmcnt hut was re- stored with subsequent addition of exogenous RNA (80 pg/ml). However, when p68 was phosphorylated following RNase treatment, ATPase activity was not fully recovered by the addition of exogenous RNA (-40% of maximal), illustrating that PKC phosphorylation blocked RNA stimulation of ATPase activity.

Regulation
This may result from reduced affinity of p68 for RNA; at higher concentrations of exogenous RNA (ie. 0.2 mg/ml), PKC phosphorylation had little or no effect on RNA-stimulated ATPase activity (data not shown). Interestingly, Ca2+/CaM also significantly inhibited the ATPase activity of the enzyme (Fig. 4B), whereas Ca2+ or CaM alone had no effect (data not shown). Ca2+lCaM markedly reduced ATPase activity of enzyme co-immunoprecipitated with bound RNA and prevented activation of p68 by addition of exogenous RNA (80 pg/ml). The inhibition of RNA-stimulated ATPase activity by CaM was concentration-dependent ( Fig. 5) and suggests that the affinity of p68 for CaM lies in the low micromolar range. Although relatively high concentrations of CaM were required for inhibition of p68 ATPase activity, nuclear CaM levels have been estimated to be as high as 500 I~M (26)(27)(28). Since ATPase activity is required for p68 unwinding activity, the RNA helicase activity of p68 may likewise be inhibited by PKC and/or by Ca2+lCaM.
The molecular mechanism for PKC or CaM inhibition of p68 RNA-stimulated ATPase activity has not yet been established, although there are several possibilities. For example, conformational changes caused by phosphorylation or by CaM binding may indirectly disrupt RNA binding, ATP binding, and/or hydrolysis. Alternatively, the phosphorylated residue(s) may reside within or near the RNA-binding site, resulting in electrostatic repulsion of nucleic acid. While the RNA-binding site of p68 is currently undefined, it has been proposed that the basic region at the carboxyl terminus of p68 is a specific RNAbinding motif (29). Moreover, the HRIGXXR region conserved in DEAD box proteins has been reported to be required for RNA binding and ATP hydrolysis of family member eIF-4A (29).
These data represent the first report of PKC/Ca2+ regulation of a member of the DEAD box family and strongly suggest that 68 RNA Helicase regulation is mediated through the IQ domain. Sequence examination of the DEAD box proteins reveals that other family members including Dbp73D (301, vasa (31), and Prp5 (32) also contain sequences similar to the IQ domain, suggesting that these proteins may also be modulated by PKC or CaM binding. Whereas the specific functions of p68 are unknown, the enzyme has been shown to undergo dramatic changes in nuclear localization during the cell cycle. It is localized in the nucleoplasm during interphase and translocates t o the nucleoli during telophase, suggesting a role for p68 in nucleolar assembly (33). This observation together with our results suggests that the RNA helicase activity of p68 may be regulated by PKC and/or CaM during the cell cycle. Interestingly, CaM has been implicated as the mediator of Ca2+-dependent regulation of cell cycle progression, especially at the G,IS, GJM, and metaphase1 anaphase transitions (for review see Ref. 34). Furthermore, CaZ+/CaM has been shown to inhibit nuclear events such as transcription through direct protein-protein interaction with transcription factors of the basic helix-loop-helix group (28). In conclusion, we have defined a potential mechanism for regulation of RNA secondary structure by Ca2+ signal transduction pathways.