Two forms of purified m7G-cap binding protein with different effects on capped mRNA translation in extracts of uninfected and poliovirus-infected HeLa cells.

Eukaryotic mRNA cap binding proteins were purified from ribosomal salt wash in the presence of protease inhibitors by sucrose gradient sedimentation and m7GDP-Sepharose affinity chromatography. Rabbit reticulocyte and erythrocyte proteins with sedimentation constants of less than 6 S yielded a approximately 24,000-dalton cap binding protein. It stimulated capped mRNA translation in extracts of uninfected HeLa cells but did not restore capped mRNA function in extracts prepared from poliovirus-infected cells. Restoring and stimulatory activities both were associated with a larger, approximately 8-10 S complex that included the approximately 24,000-dalton polypeptide and several higher molecular mass components. The same two translational activities were also obtained in a slightly smaller approximately 5-7 S complex from uninfected HeLa cells but were absent from poliovirus-infected cell preparations.

Eukaryotic mRNA cap binding proteins were purified from ribosomal salt wash in the presence of protease inhibitors by sucrose gradient sedimentation and m7GDP-Sepharose affinity chromatography. Rabbit reticulocyte and erythrocyte proteins with sedimentation constants of t6 S yielded a -24,000-dalton cap binding protein. It stimulated capped mRNA translation in extracts of uninfected HeLa cells but did not restore capped mRNA function in extracts prepared from poliovirus-infected cells. Restoring and stimulatory activities both were associated with a larger, -8-10 S complex that included the -24,000-dalton polypeptide and several higher molecular mass components. The same two translational activities were also obtained in a slightly smaller -5-7 S complex from uninfected HeLa cells but were absent from poliovirus-infected cell preparations.
Eukaryotic cellular mRNAs contain a 5'-terminal cap, m7GpppN (l), a structure that is added early to initiated ends of nascent transcripts (2, 3). The cap is retained during pre-mRNA processing and becomes a distinctive feature of cytoplasmic messengers. Its presence promotes the formation of stable initiation complexes in vitro and enhances translation (4,5). How this occurs is not known, but a cap-specific binding protein (M, -24,000) has been isolated from protein synthesis initiation factors of rabbit reticulocytes by affiiity chromatography on m7GDP-Sepharose (6). This protein differentially stimulates the translation of capped mRNAs as compared to uncapped viral RNAs in HeLa cell extracts (7).
In poliovirus-infected HeLa cells, initiation of host cell protein synthesis is abruptly shut off in favor of viral protein formation (8). The molecular basis for this striking regulatory phenomenon has been studied extensively. It was shown that cellular mRNA is not degraded in poliovirus-infected cells (9), and capped mRNA extracted from them is translated in cellfree protein-synthesizing systems (10). In contrast to mRNA, protein synthesis initiation factors are altered after poliovirus infection. Factors isolated from infected cells stimulate the translation of poliovirus RNA but not cellular mRNAs (11). These observations have been interpreted on the basis that * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. &ction 1734 solely to indicate this fact. poliovirus is one of a small number of eukaryotic viruses that contain RNA with messenger activity and uncapped 5'-termini (12). Thus, if the change in initiation factors that accompanies poliovirus infection leads to a block in the activity of cap binding protein, then replacement of cellular protein synthesis by cap-independent poliovirus RNA translation presumably would be facilitated.
Previous studies showed that capped mRNA translation could be "restored" in extracts of poliovirus-infected HeLa cells by addition of partially purified eIF-4B' (13), an initiation factor involved in mRNA binding to ribosomes. However, eIF-4B was found to contain cap binding protein (14). Subsequently, "restoring activity" was purified from rabbit reticulocyte ribosomal salt wash (15). The major polypeptide in the purified restoring activity corresponded to cap binding protein, but the most highly purified activity was labile, suggesting that other proteins might also be involved. Consistent with this possibility, we report that preparations of cap binding protein that stimulate capped mRNA translation in extracts of uninfected HeLa cells do not restore capped mRNA translation in extracts of poliov~s-infect~d cells. Restoring activity was associated with a protein complex that included cap binding protein and several higher molecular weight polypeptides. Reticulocyte lysates were prepared from phenylhydrazine-treated, anemic rabbits as described previously (18) except that the buffers contained 0.5 mM phenylmethanesuifonyi fluoride, 0.1 mg/ml of soybean trypsin inhibitor, 50 u n i~/ m l of aprotinin, and 1 mM ethylenegiycol bis(P-aminoethyl ether) N,N,N',N'-tetraacetic acid (29) in all steps prior to resuspension of the (NHdZSO, precipitates. Ribosomal salt wash was prepared from reticulocyte crude ribosomes (18). (NH4)2S04 was slowly added to attain 70% saturation (43.6 g/100 ml) followed by stirring for 60 min at 0 "C. Precipitate was collected by centrifugation (16, myosin (210,000). Position of cap binding blood was obtained from 12 New Zealand rabbits (2.5-3 kg) by cardiac puncture. Cells were pelleted (Sorvall GSA rotor, 1,500 X g, 20 min) and resuspended in four packed cell volumes of 0.14 M NaCI, 5 mM KCI, 2 mM MgCl?, and protease inhibitors as above. Washing was repeated and the cells were lysed a t 0 "C by addition of 1.5 volumes of double-distilled water. After stirring on ice for 10 min, cellular debris was collected by centrifugation (16,000 X g, 20 min). Pellets were re-extracted a t 0 "C with an equal volume of HrO. After recentrifugation, supernatant fractions were combined and slurried with DE52 (one-half the settled volume pre-equilibrated in TEM which is 20 mM Tris-HCI (pH 7.5 a t 4 "c), 0.1 mM EDTA, 5 mM 2-mercaptoethanol). After 10 min, the settled adsorbent was resuspended by gentle stirring, allowed to stand for another 10 min, and the procedure was repeated. Adsorbent was collected on a Buchner funnel under low vacuum, taking care that the DE52 was not allowed to dry. The moist resin was washed with -1.5 liters of TEM added in 50-ml aliquots to remove hemoglobin and other nonadsorbed proteins. Cap binding protein was eluted in TEM containing 0.25 M KC1 which was applied in twelve 50-ml aliquots. The 0.25 M KC1 eluate was brought to 70%. saturation with (NH.d?SO1, and the pH was maintained a t 7.5 by addition of 2 M NHnOH. After stirring for an additional 15 min, the protein solution was allowed to stand overnight a t 4 "C. Precipitate was collected (IS,ooO X g, 20 min), gently resuspended, and dialyzed overnight against 1 liter of gradient buffer with one change. Dialyzed protein was applied to 15-30'h sucrose gradients (85 mg of protein (20) per 35-1111 gradient). After centrifugation (SW-28 rotor, 28.000 rpm, 34 h, 4 "C), 60 drop fractions were collected. Fractions corresponding to the trailing edge of the major A2R,,,n, peak were pooled (upper 15-20% of each gradient) and dialyzed for affinity chromatography as above.
Polyacrylamide gel electrophoresis was performed using a modified Laemmli system ( 2 1 ) . The stacking gel contained 3% acrylamide and the resolving gel either 10 or 12.5% acrylamide as indicated in figure legends. Electrophoresis was done a t constant voltage; gels were stained, destained, and dried as described FIG . 2 (right). Stimulation of Sindbis capsid protein synthesis by affinity-purified proteins added to extracts of uninfected HeLa cells. Sindbis virus mHNA was translated as described under "Experimental Procedures," and the ,"'S-labeled products were analyzed by autoradiography after electrophoresis in 10%  Sindbis virus mRNA and 10 pCi of [""Slmethionine. After incubation for 60 min a t 37 "C, reactions were terminated by addition of 1 ml of acetone and protein was allowed to precipitate for a t least 2 h at room temperature. Precipitates were collected by centrifugation. dried at room temperature, and prepared for gel electrophoresis by incubation for 5 min at 100 "C in 25 pl of sample buffer that included 2%. sodium dodecyl sulfate and 5% /l-mercaptoethanol ( 2 1 ) .

RESULTS AND DISCUSSION
Since highly purified restoring activity as obtained previously (15) w a s not stable, it seemed possible that polypeptides intrinsic to the activity were degraded during purification. Consistent with this possibility, cap binding protein in buffers of low ionic strength, e g . 0.1 M KCI, associates with eIF-3 (15), an initiation factor containing multiple polypeptides that may result from proteolysis (22). Furthermore, pol-ypeptides of M , > 24,000 that appear to be structurally related to the cap binding protein have been detected in reticulocyte lysate and in partially purified eIF-3.? In an effort to obtain undegraded restoring activity, we purified cap binding protein from ribosomal salt wash in the presence of the four protease inhibitors as described under "Experimental Procedures." were: A, <6 S; B, 2-8 S; C, 7-10 S; and D, 8-16 S. Pools A through D were each applied to m'GDP-Sepharose, and the bound material eluted with m'GDP (6) was analyzed by polyacrylamide gel electrophoresis (Fig. 1). Pool A contained a polypeptide in the position of cap binding protein (arrow) and little material of higher molecular weight. The same pattern was obtained with cap binding protein purified from erythrocytes (Fig. 1, R B C ) . In each case, the -24,000-dalton polypeptide comprised >95% of the stained material as determined by densitometry. The predominant component in pool B was also a band of apparent M , -24,000, but two minor polypeptides of M , = 48,000 and 55,000 were evident in addition. Pool C contained a 48,000-dalton band, a lower level of the cap binding protein, and prominent polypeptides migrating near phosphorylase ( M , = 94,000) and myosin ( M , = 210,000). Similar high molecular weight bands were also present in the fast sedimenting, large complexes comprising pool D. Identity of the M , -24,000 stained band and cap binding protein was confirmed on the basis of specific, i.e. m'GDPinhibited, cross-linking to oxidized reovirus mRNA (14, data not shown). The results indicate that cap binding protein associates with higher molecular weight polypeptides to form a fast sedimenting protein complex (8-10 S) that is retained on m'GDP-Sepharose and specifically eluted with m'GDP. The 8-10 S estimate from the overlapping fractions of pools C and D was confirmed by assaying individual gradient fractions for translational activity before chromatography.
Affinity-purified cap binding protein and i t s higher molecular weight complexes were tested for the ability to enhance capped mRNA translation in vitro (7). All four pools stimulated Sindbis virus capsid protein formation in HeLa cell extracts (Fig. 2, lanes 2-8), and the same results were obtained in a separate test with erythrocyte-derived cap binding protein (Fig. 2, lanes 9-ZZ). The stimulatory activities of the different protein pools were concentration-dependent.
Although pools A and B contained cap binding protein and stimulated uninfected HeLa cell extracts, they were essentially devoid of restoring activity. As shown in Fig. 3 A , the capacity of HeLa cell extracts to translate Sindbis mRNA (lane 2 ) was abruptly shut off in extracts prepared a t 3 h after poliovirus infection (lane 4 ) . Translation was not restored by addition of cap binding protein affinity-purified from pool A (Fig. 3 B, lane 2 ) or pool B (Fig. 3 C, lanes 2 and 3 ) . Restoring activity was also absent from the erythrocyte-derived cap binding protein tested in the same way (data not shown). By contrast, pools C and D, i.e. the samples that contained the high molecular weight components in addition to cap binding protein, effectively restored Sindbis capsid synthesis (Fig. 3, B,   lane 3 and C, lanes 4 and 5, respectively) lanes 4 and 5). Arrows correspond to a molecular weight of 33,000, i.e. the position of Sindbis capsid (24). determined from migrations of standard proteins. results were obtained with rabbit globin mRNA; furthermore, translation of uncapped RNA of encephalomyocarditis virus in extracts of uninfected and poliovirus-infected HeLa cells was not stimulated by any of the affinity-purified fractions (data not shown). The purified restoring activity in pools C and D was stable to prolonged storage at -196 "C and repeated freeze-thawing. Thus, the polypeptides associated with cap binding protein in a large, 8-10 S complex may simply stabilize restoring activity; alternatively, they may have separate activities related to cap recognition. In any case, affinitypurified cap binding protein was obtained in two distinct forms: one which stimulates capped mRNA translation in extracts of uninfected HeLa cells and a more complex form which also has stable restoring activity in extracts of poliovirus-infected cells.
To determine if stimulatory and restoring activities analogous to those in rabbit blood cells are also present in other mammalian cells, initiation factors were prepared from HeLa cell cultures. The 70% (NH&SOs precipitate of ribosomal salt wash was sedimented in 1530% sucrose gradients containing 0.5 M KCI, and gradient fractions were tested for stimulatory and restoring activities in extracts of uninfected and poliovirus-infected HeLa cells, respectively. Synthesis of Sindbis virus coat protein by extracts of po~iovirus-infected cells could be restored by addition of a fast sedimenting complex (Fig.  4A, gradient fractions 18-22 correspond to -5-7 S position). The same gradient fractions effectively stimulated Sindbis RNA translation in uninfected cell extracts (Fig. 4 B ) . Similar analyses were carried out with initiation factors prepared from HeLa cells at 3 h after infection with poliovirus. As expected on the basis of previous reports (8,11,13), they had no restoring activity and did not increase the translation of capped mRNA in extracts of uninfected HeLa cells (data not shown). Although the loss of capped mRNA translation in poliovirus-infected HeLa cells is correlated with inactivation of cap binding protein translational effects, a causal relationship remains to be established. However, recent studies indicate that both stimulatory and restoring activities can be neutralized by a monoclonal antibody prepared against purified erythrocyte 24,000-dalton cap binding protein."