Initiation of Protein Synthesis in Vitro by a Clostridial System I. SPECIFICITY IN THE TRANSLATION OF NATURAL MESSENGER RIBONUCLEIC ACIDS*

SUMMARY The homologous ribosomal systems from Escherichia co2i and Uostridium pasteurionum were compared for their ability to translate f2 RNA, formaldehyde-treated f2 RNA, T4 early messenger RNA, E. coli messenger RNA, and C. pasteurianum messenger RNA in protein synthesis assays in vitro. The E. co2i ribosomal system translated all five of these messengers, while the C. pasteurianum ribosomal system translated only the C. pusteurianum messenger RNA. The two ribosomal systems also had different characteristic magnesium profiles and exhibited different levels of endogenous activity in the protein synthesis assays. The messenger RNA responsible for the C. pasteurianum endogenous activity was shown to occur in the salt-washed ribosomes and not in the initiation factor fraction. The formaldehyde-treated f2 RNA and the C. pasteurianum messenger RNA exhibited different characteristics from the other types of messenger RNAs in their behavior with the E. coli ribosomal system in the protein synthesis assays. The two homologous ribosomal systems were also tested for their ability to bind [14C]formyl-Met-tRNA in response to different types of mRNA. The E.

and not in the initiation factor fraction. The formaldehyde-treated f2 RNA and the C. pasteurianum messenger RNA exhibited different characteristics from the other types of messenger RNAs in their behavior with the E. coli ribosomal system in the protein synthesis assays.
The two homologous ribosomal systems were also tested for their ability to bind [14C]formyl-Met-tRNA in response to different types of mRNA.
The E. coli system bound fMet-tRNA in response to synthetic poly(A,G,U), f2 RNA, T4 early mRNA, and C. pasteurianum mRNA. The feurianum ribosomal system bound fMet-tRNA sponse to poly (A,G,U) and C. pasfeurianum mRNA f2 RNA or T4 early mRNA.
C. pasin rebut not Initiation of protein synthesis by bact.erial ribosomal systems' has been a subject of intensive invest,igation in the past few years. There have been numerous recent advances in the purification of initiation factors (l-5) and in the study of the mechanism of the initiation process in vitro (4)(5)(6)(7)(8)(9)(10)(11)(12)(13). However, the mechanism by which ribosomes recognize the proper initiation sites on messenger RNA is not well understood.
The RNA triplet AUG now appears to be the initiation codon on mRNA (14,15), but it is still not known how ribosomes distinguish initiation AUG codons * This work was supported in part by Grant AM-2109 from the National Institutes of Health, United States Public Health Service.
1 In this paper the term "ribosomal system" refers to the ribosomes and initiation factors from an organism.
from the other AUG sequences on the mRNA molecule or what exactly constitutes a protein synthesis initiation site on mRNA. Secondary structure is apparently involved when the RNA from RNA-bacteriophages is used as mRNA in vitro (15-l@, but it has not yet been determined whether or not the behavior of these viral messengers can be taken as representative of the behavior of bacterial cellular mRNA. The bulk of the research on the bacterial initiation process has been carried out with Escherichia coli. Investigation in vifro of protein synthesizing systems from other procaryotic organisms has already revealed some evidence for species-specific differences in the ability of ribosomal systems t.o recognize initiation sites on the RNA from RNA-bacteriophages (19,20), but further investigations of species other than E. coli are necessary to determine which aspects of the mechanism for initiation of protein synthesis by bacterial ribosomal systems are common to all species and which are species-specific.
An amino acid incorporating system from Clostridium pasteurianum has been described previously (21). This paper describes additional properties of a modified ribosomal system derived from C. paste&unum and presents evidence for species specificity in the translation of several types of natural mRNA by homologous ribosomal systems from E. coli and C. pasieuriamm.
The ribosomal system from E. coli translates f2 RNA, formaldehyde-treated f2 RNA, T4 early mRNA, E. coli mRNA, and C. pasleurianum mRNA; the ribosomal system from C. pasteurianum translates the C. pasieurianum mRNA but none of the other four types of RNA tested.
[YZJFormyl-Met-tRNA binding assays were also performed using the two homologous ribosomal systems and different types of mRNA.
The specificity exhibited in the binding assay by the two ribosomal systems toward the different types of mRNA was analogous to the specificity exhibited by the two systems in the protein synthesis assays.
The roles of the salt-washed ribosomes and the initiation factors in determining the specificity of mRNA recognition were investigated by testing the heterologous combinations of components derived from E. coli and C. pasteurianum cells and will be described in the following paper (22). Synthetic poly (A, G, U : 2,1,2) was also provided by Dr. Grunberg-Manago.

Nethods
Growth 01 Cells and Preparation of Ribosomes-E.
coli MRE 600 cells were grown in a 200.liter New Brunswick Fermacell Fermentor at 37" under forced aeration.
A 1% inoculum from an overnight culture was used. When the A 660 reached 2 (late log phase), the fermentor was chilled, and the cells were harvested on an electric Sharples Centrifuge.
The yield was 3 g per liter, wet weight. The cells were stored frozen at -90".
The salt-washed ribosomes were prepared by a modification of the method of Lelong et al. (24). All operations were carried out at 4". Two-hundred grams of frozen cells were thawed and ground with two weights of alumina (Alcoa, A305).
An S30 fraction was prepared by centrifuging the suspension for 20 min at 20,000 x g and then recentrifuging the supernatant solution twice for 30 min at 30,000 x g; the pellet was discarded each time.
The 830 fraction (about 200 ml) was dialyzed overnight against two 4-liter volumes of Buffer A and then recentrifuged for 30 min at 30,000 x g. The dialyzed S30 fraction was then centrifuged in two portions for 4 hours at 45,000 rpm in a Spinco No. 50 fixed-angle rotor (125,000 X g). The supernatant solution was poured off, and the pellets were resuspended in a total volume of 100 ml of Buffer B (10 rnxt Tris-chloride buffer, 1 M NH&l, 40 mM magnesium acetate, 2 mM EDTA, 10 mM 2-mercaptoethanol, adjusted to pH 7.5 with NHIOH) and stirred slowly overnight. The suspension was centrifuged 10 min at 30,000 X g, and the pellet was discarded.
The salt-washed ribosomes were pelleted by centrifuging 6 hours at 125,000 x g. The supernatant solution was poured off and stored frozen as the source of initiation factors.
The pellet was resuspended in 100 ml of Buffer B, stirred overnight, and centrifuged as before. The pellets were resuspended in 50 ml of Buffer C (10 mM Tris-chloride buffer, pH 7.8, 50 mM NH4CI, 10 mM magnesium acetate, 7 mM 2-mercaptoethanol), and centrifuged 3 hours at 125,000 X g. This final pellet was taken up in Buffer C to a final volume of 20 ml, clarified by a low speed spin, and frozen in small portions at -90".
The A260 of an aliquot of the final preparation was 1200 for 1 cm lightpath.
C. pasteurianum cells for culture were stored lyophilized on sterile filter paper discs under vacuum.
Cultures were started by inoculating a tube of "potato medium" with a disc (25) and were subsequently transferred to synthetic medium (25). For preparative cultures (10 to 200 liters), 0.0057, CaC12 was substituted for the CaC03 in the synthetic medium.
To grow a 200liter culture, the necessary 10 liters of log phase inoculum was prepared as follows.
Transfers were made every 12 hours, using a 5 to 10% inoculum, to successively larger volumes of medium; growth was carried out without shaking or aeration at 2530" in 0.1. to l-liter Florence flasks which were filled to the neck with medium and plugged with cotton.
Each transfer was made when the culture reached an A680 of 2 to 4. For growth of the IO-liter inoculum, a 15.liter New Brunswick fermentor unit was used, and the sterile medium was bubbled with purified nitrogen gas continuously before inoculation and during growth, which was at 30". The 200.liter culture was grown in a New Brunswick Fermacell Fermentor at 30" with automatic pH control, set at 6.5. Cells grown without pH control were found to be unsuitable for preparation of a system for protein synthesis in vitro.2 The cells were harvested at an AeGO of 5 to 6 (midlog phase) on an electric Sharples Centrifuge.
The yield was 6 g per liter, wet weight.
The cells were stored at -90".
C. pasteurianum salt-washed ribosomes and washed, preincubated ribosomes were prepared as follows.
They were disrupted by subjecting them to sonic oscillation from a Branson Sonifier Cell Disrupter for 3 min in a beaker kept in an ice bath.
The resulting suspension was then centrifuged twice for 30 min at 30,000 X g; the pellet was discarded each time.
The ribosome pelleting and NH4Cl washing procedures were identical to those described above for E. coli ribosomes, except that for C. pasteurianum ribosomes Buffer B contained 2 M NH,Cl and 20 nlM 2-mercaptoethanol.
To prepare salt-washed (nonpreincubated) ribosomes, the pellet from the second NH4Cl washing was taken up in 40 ml of Buffer D and repelleted by centrifuging 4 hours at 125,000 X 9. The final pellet was resuspended in Buffer D to a final volume of 4 ml; the final suspension was clarified by a low speed spin and frozen in small portions at -90".
The AW,O was 1800. To prepare salt-washed, preincubated ribosomes, the pellet from the second NH&l washing was taken up in Buffer D to a final volume of 5.0 ml. The AzOO was 1740. The suspension w-as dialyzed against 3 liters of Buffer D for 3 hours to remove residual NH&l.
These ribosomes were then used directly in the preincubation reaction.
The composition of the preincubation reaction, using the above 5.0 ml of salt-washed ribosomes, was the same as for a normal protein synthesis assay (described later in the "Methods" section), with the following modifications. The total volume of the reaction was 25 ml; the Mg++ concentration in the reaction was 15 mM; unlabeled valine was substituted for the ~42, 3-3Hz]valine; IO-formyltetrahydrofolate, initiation factors, and exogenous mRNA were omitted.
The reaction misture was brought quickly to 37", incubated at that temperature for 25 min, then cooled on ice. The mixture was then centrifuged for 10 min at 30,000 X g. The supernatant solution was layered over 20% sucrose containing Buffer D (2.5 ml in the bottom of each of four IO-ml nitrocellulose tubes) and centrifuged overnight (10 hours) at 125,000 X g. The supernatant solution was removed carefully.
The pellets were then taken up in Buffer D to a final volume of 6 ml; the final suspension was clarified by a low speed spin and frozen in small portions at -90".
The ANO was 1100.

Preparation
of Initiation Factors-The method used was the same for both E. coli and C. pasteuridnum.
To 10 ml of supernatant solution from the first NH&l washing, 5.2 g of finely ground (NH&SO4 were added slowly with stirring. The solution was stirred for 1 hour at 4". After being decanted from any residual (NH&S04, the solution was centrifuged 15 min at 20,000 x g. The pellet was taken up in Buffer E (20 mM Trischloride buffer, pH 7.25,20 mM NH&l, 2 mM magnesium acetate, 7 mM 2-mercaptoethanol, 5y0 (v/v) glycerol) to a final volume of about 3 ml with a final protein concentration of about 17 mg per ml. The solution was dialyzed against Buffer E and then frozen at -90" in small portions.
Preparations of Messenger RNA-For preparation of f2 RNA, f2 phage were grow11 on B. coli strain Q13. The bacteria were grown in 15-liter New Brunswick fermentor units at 37" in R broth (26) to an AGeO of 0.4. Phage f2 was then added at a mul-tip1icit.y of infection of about 30. Incubation and aeration were continued for 4 hours, after which the cultures were cooled. Purification of the phage from the crude lysate was accomplished by the procedure of Gesteland and Boedtker (27) except that the methanol precipitation step was omitted.
The procedure included ammonium sulfate precipitation, pelleting of phagc by ultracentrifugation, and finally CsCl banding. The f2 RNA was obtained from the concentrated phage suspension by extracting twice with phenol at 4" and then precipitating the RNA from the aqueous phase and washing it with ethanol.
For preparation of T4 early mRNA, phage T4Df was grown on l?. coli strain BE. The bacteria were grown and infected at 30" by Method III as described by Belle et al. (28). Incubat.ion was continued for 5 min after infection, and then the culture was rapidly chilled by pouring it over crushed, frozen M9 medium. The T4 early mRNA was prepared from the cells by the method of Salser et al. (29).
E. coli mRNA was prepared from uninfected E. coli BE cells, grown to an A 6,~ of 1 .O by the same method used for preparation of T4 mRNA above.
C. pasteurianum mRN.4 was obtained from a freshly prepared C. pasteurianum crude polysomal lysate by phenol extraction, using the same extracting procedure as for T4 mRNA above. The C. pasteurianum lysate was prepared by the method of Brodrick and Rabinowitz" C. pasteurianum ribosomal RNA was obtained by phenol extraction of salt-washed ribosomes, using the same method of extract'ion as for T4 mRNA above.
,411 RNA preparations were stored in small portions at -90". Preparation of High Speed Xupernatant Fraction-The high speed supernatant fraction (S150) was prepared from E. coli A19 cells by sonic oscillation.
The resulting solution was centrifuged at 150,000 x g for 3 hours.
The supernatant solution was dialyzed against Buffer D and stored in small portions at -90".
Assay for Protein Synthesis in Vitro-The assay for protein synthesis in vitro, used for measuring the incorporation of L-[2,3-3Hz]valine into acid-insoluble material, was based on the assay described by Nirenberg (30). The standard assay contained the following ingredients in a total volume of 0. folate; 25 ~1 of E. coli ,419 S159 solution (10 mg of protein per ml, in Buffer D); 3 A260 units of ribosomes; 0.17 mg of crude initiation factors; varying amounts of mRNA.
Except for magnesium-dependence curves, the total hIg++ concentration in the assay was 11 mM for E. coli ribosomes and 15 mnz for C. pasteurianum ribosomes (see "Results" section). The contribution to the Mg++ concentration by the supernatant fraction added to the assay has been included in these figures for the total Mg++ concentration in the assay. The assays were kept on ice during preparation.
The counting efficiency of the ~-[a, 3-3Hz]vali11e incorporated into protein and precipitated into the filter discs was determined to be 20%.4 Therefore, the specific activity of the L-[2,3-3Hs]valine in the filter discs, corrccted for counting efficiency, was calculated to be 36 cpm per pmole.
The results are expressed as picomoles of valine incorporated without subtraction of control blank values.
Time studies of the assay indicated that incorporation of ami11o acids into protein had stopped at the end of 30 min of incubation; the time course of the reactio11 was found to be the same no matter what type of ribosomal system and messenger RNA was being assayed. 5 Preparation of [%']FormyZ-JIet-tRNA-[14C]Formy1-Met-tRNA% was prepared by the method of Lelong et al. (24). I-nfractionated E. coli B tRNA was charged with unlabeled methionine in the presenceof (1)-10-[14C]formyl-tetrahydrofolate (specific activity, 50 mCi per mmole) as a formyl donor.
The source of methionyl-tRNA synthetase and transformylase enzyme was a DEAE-cellulose-treated, high speed supernatant fraction prepared from E. coli A19 or MRE 600 cells by the method of Samuel et al. (32). The supernatant fraction prepared by this method was free of contaminating tRNA and folate cofactors. The charged and formylated tRNA was purified by phenol extraction and ethanol precipitation, and then the unformylated met-tRNA in the preparation was discharged enzymatically in the presence of the DEAE-cellulose-treated supernatant fraction, pyrophosphate, and AMP. After phenol extraction and ethanol precipitation, contaminatiug nucleotides were eliminated by gel filtration, and the tRNA was reclaimed by lyophilization of the pooled radioactive fractions.
[~VJ]Formyl-&let-tRNA Binding Assay-The assay used was a modification of the one described by Nirenberg and Leder (33). Each assay contained, in a final volume of 0.05 ml: 50 mM Trischloride buffer, pH 7.4; 5.5 to 6.0 mM magnesium acetate (5.5 mM for E. coli ribosomes, 6.0 mM for C. pasteurianum ribosomes) ; 80 mM NH&l; The counting efficiency of an aqueous tritium sample counted in Bray's solution was determined by using a 3H-water standard. 5 M. R. Stallcup   the toluene scintillation fluid described above. The specific activity of the [14C]formyl-Met-tRNA on the Millipore filter, corrected for a counting efficiency of 80%, was 88 cpm per pmole. The assay results are expressed as picomoles of fMet-tRNA bound without subtraction of control blank values. The background for the scintillation counter used was 12 cpm. Time studies of fMet-tRNA binding to ribosomes at 37" indicated that the reaction was 80 to 90% complete after 5 min and was complete after 10 to 12 min.

AND DISCUSSION
Magnesium Dependence of Protein Synthesizing Xystems- Fig.  1, u and b, shows the magnesium dependence of valine incorporation into protein by the B. GO& and C. pasteurianum ribosomal systems, respectively, in response to their endogenous messengers and to four exogenous messengers, f2 RNA, T4 early mRNA, E. coli mRNA, and C. pasteurianum mRNA.
All four exogenous messengers stimulate valine incorporation by the E. coli ribosomal system far above the endogenous level (Fig. la). However, with the C. pasteurianum system, only the C. pasteurianum mRNA stimulates protein synthesis above the endogenous level; the addition of any of the other three messengers to the assay results in no increase in valine incorporation above the endogenous level (Fig. lb). Greater stimulation of amino acid incorporation by the C. pasteurianum mRNA is observed when larger amounts of the mRNA are used. Table I shows that omitting salt-washed ribosomcs or supernatant fraction from the assay reduces amino acid incorporation to a very low level.
The concentration of Mg++ for optimal amino acid incorporation with the E. coli ribosomal system varies with the different messengers.
The magnesium profiles also indicate that the C. pa.steurianum ribosomal system has a somewhat higher Mg++ optimum for protein synthesis than does the E. coli ribosomal system. For this reason, different Mg++ concentrations, 11 mM and 15 mM, were used in further experiments to assay the ribosomal systems of E. coli and C. pasteurianum, respectively.
Reduction of Level of Endogenous Activity in C. pasteurianum Ribosomes-Another difference in the results obtained with ribosomes from the two species is in the level of endogenous activity.
As shown in Fig. 1, a and b, C. pasteurianum ribosomes exhibit more than twice the level of endogenous activit,y as an equivalent amount of E. co/i ribosomes.
This relation-

Dependence of protein synthesis on ribosomes, supernatant fraction, and initiation faclors
The assay conditions were the same as in Fig. 1 ship also holds if the E. coli ribosomcs are prepared by the sonic oscillation technique normally used for preparing the C. pasteurianum ribosomes5 In order to facilitate the study of the response of the C. pasteurianum ribosomal system to exogenous messengers added to the protein synthesis assays, it was desirable to find a method for reducing the high level of endogenous activity exhibited by this system without reducing its ability to respond to the exogenous messenger.
First, it was necessary to determine whether or not the crude initiation factor fraction or the salt-washed ribosomes themselves were the source of the endogenous messenger fragments.
To test whether or not the crude initiation factor fraction was a source of cndogenous messenger, the RNA and ot'her nucleic acid material was removed from the initiation factor preparation by passing it over a DEAE-cellulose column, according to the method of Ohta et al. (34). This procedure increased the &a: A260 ratio of the initiation factor preparation from 1.1 to 1.4. However, parallel protein synthesis assays using the DEAE-cellulose-treated initiation factors and untreated initiation factors showed no significant differences in the levels of endogenous activity or the stimulation by exogenous C. pasteurianum mRNA. This indicated that the salt-washed ribosomes themselves must bc the source of the endogenous messenger. It was found that the level of endogenous activity could be reduced by about two-thirds, without any significant reduction in the response to added C. pasteurianum mRNA, by preincubating the salt-washed ribosomes in a protein synthesis assay under conditions unfavorable for reinitiation (by omitting loformyltetrahydrofolate and initiation factors). The details of the preincubation and recovery of the ribosomes are given under "hlethods." Table II  The results for the addition of C. pasteurianum ribosomal RNA to the assay are given by Curve F (C-B); all other letters and symbols represent the same types of mRNA as in Fig. 1 factor fraction, the endogenous protein synthesis activity is dependent upon the presence of initiation factors (Table I), which suggests that reinitiation by the ribosomes on these small messenger fragments is occurring.  (26) to four different types of mRNA and to C. pasteuriunum ribosomal RNA. These curves show, more strikingly than the magnesium-dependence curves in Fig. 1, that the E. coli ribosomal system synthesizes protein in response to all four mRNA preparations, while the C. pasfeurianum system responds only to C. pasteurianum mRNA and not at all to the other three mRN,4 preparations.
C. pasteurianum rRNA serves as a control for both ribosomal systems. Its failure to stimulate amino acid incorporation indicates that rRNA is not the active component of the C. pasteurianum mRNA preparation, which is a polysomal extract and therefore contains rRNA.
The points on the ordinates of Fig. 2, a and 6 represent the levels of endogenous activity.
The f2 RNA preparation does not contain contaminating ribosomal RNA as do the other mRNA preparations, and therefore it has a much higher specific activity than the other preparations in the assay with the l?. coli ribosomal system.
In these mRNA curves, the C. pasteurianum mRNA exhibits some properties different from the other three messengers. Ill the assay with the E. coli ribosomal system, the C. pasteurianuwz mRNA is saturating at a lower level of RNA than the other two cellular RNA preparations, T4 early mRNA and E. coli mRNA.
Even more striking is the fact that C. pasteurianum mRNA stimulates protein synthesis by E. coli and C. pasteuriunum salt-washed ribosomes in the absence of any added initiation factors, whereas all of the other messengers exhibit absolute initiation factor dependence. Fig. 2 shows that C. pasteurianum mRNA is more active in protein synthesis with the E. coli ribosomal system than with the C. pasteurianum ribosomal system. This result could be attributed to (i) more efficient (faster) translation by the E. coli ribosomal system; or (ii) initiation by the E. coli system at a larger number of sites 011 the C. pasteurianum mRNA. Formaldehyde-treated f2 RNA--f2 RNA was treated with formaldehyde by the method of Lodish (17), who showed that this treatment unfolds the secondary structure of the f2 RNA molecule, with the result that initiation of protein synthesis can occur at several new positions, not accessible on the native f2 RNA. The formaldehyde-treated f2 RNA (HCHOf2 RNA) was still found to be inactive in assay with the C. pasleurianum ribosomal system, as shown in Fig. 3, which rhows valine incorporation measured as a function of mRNA concentration.
The HCHO-f2 RNA was active with the E. coli ribosomal system, but as Fig. 3 shows, its behavior with E. coli ribosomes is different from that of the untreated f2 RNA.
The HCHOf2 RNA is saturating at a much lower level than the untreated f2 RNA; and the treated RNA stimulates some protein synthesis by E. coli Fait-washed ribosomes in the absence of initiation factors, while untreated f2 RNA elicits no activity without initiation factors.
In these by the E. coli ribosomal system (a) and by the C. pasleurianum preincubated ribosomal system (6) as a function of the amount of mRNA added to the assay.
The assay conditions used were those given under "Methods" for the fMet-tRNA binding assay.
The results for addition of Poly(A,G,U) to the assay are given by curve G (V---v) ; all other letters and symbols represent the same types