Reconstitution of Escherichia coli 30 S ribosomal subunits from purified molecular components.

Abstract Reconstitution of 30 S ribosomal subunits from 16 S RNA and a mixture of purified individual 30 S ribosomal proteins has been studied. Proteins from the 30 S ribosomal subunit of Escherichia coli were purified by a combination of phosphocellulose and DEAE-chromatography, and Sephadex gel filtration. The proteins purified correspond to the 21 proteins generally accepted as 30 S proteins, with the exception of two proteins, P3b and P3c, which correspond to the protein S6 studied by other workers. P3b and P3c are closely related, and one is probably a derivative of the other. Using a mixture of these purified proteins, reconstitution of functionally active 30 S subunits has been demonstrated. Reconstituted particles had higher activities in poly(U)-directed polyphenylalanine synthesis than reference 30 S particles in several experiments. The functional activity of reconstituted particles was also examined in several other assays; these included natural messenger RNA-directed polypeptide synthesis, poly(U)-directed Phe-tRNA binding, AUG-directed fMet-puromycin formation, AUG-directed fMet-tRNA binding, and the binding of termination codon UAA in the presence of chain termination factors. In all cases, activities comparable to reference 30 S subunits were observed. The sedimentation properties and the protein composition of reconstituted particles were also similar to 30 S ribosomal subunits. The kinetics of reconstitution using purified protein mixtures was essentially identical with those of reconstitution using unfractionated 30 S proteins. These results strongly suggest that 21 purified 30 S proteins together with 16 S RNA are sufficient to reconstitute 30 S subunits, and that no essential 30 S components were lost during the fractionation and purification of the 30 S proteins. Single component omission experiments indicated that all purified proteins, except P1(S1) and P3b,c(S6), are required for full functional activity. A requirement for P9a(S16) has been shown in some, but not all, experiments. P3b,c(S6) and P9a(S16) have been shown to be involved in the reconstitution reaction in other experiments (Mizushima, S., and Nomura, M. (1970) Nature 226, 1214; Nomura, M. (1973) Science 179, 864) and therefore are 30 S components. It is still not clear whether P1(S1) should be considered a "true" 30 S protein or a ribosomal-associated "factor."

Research, Depadme?, fs oj" Biochemistry ad Genetics, LTniuwsity oj W7isconsin, Madison, Wisconsin 53706 SUMMARY Reconstitution of 30 S ribosomal subunits from 16 S RNA and a mixture of purified individual 30 S ribosomal proteins has been studied.
Proteins from the 30 S ribosomal subunit of Escherichia coli were purified by a combination of phosphocellulose and DEAE-chromatography, and Sephadex gel filtration. The proteins purified correspond to the 21 proteins generally accepted as 30 S proteins, with the exception of two proteins, P3b and P3c, which correspond to the protein S6 studied by other workers.
P3b and P3c are closely related, and one is probably a derivative of the other. Using a mixture of these purified proteins, reconstitution of functionally active 30 S subunits has been demonstrated. Reconstituted particles had higher activities in poly(U)directed polyphenylalanine synthesis than reference 30 S particles in several experiments.
The functional activity of reconstituted particles was also examined in several other assays ; these included natural messenger RNA-directed polypeptide synthesis, poly(U)-directed Phe-tRNA binding, AUG-directed fMet-puromycin formation, AUG-directed fMet-tRNA binding, and the binding of termination codon UAA in the presence of chain termination factors. In all cases, activities comparable to reference 30 S subunits were observed.
The sedimentation properties and the protein composition of reconstituted particles were also similar to 30 S ribosomal subunits. The kinetics of reconstitution using purified protein mixtures was essentially identical with those of reconstitution using unfractionated 30 S proteins. These results strongly suggest that 21 purified 30 S proteins together with 16 S RNA are sufficient to reconstitute 30 S subunits, and that no essential 30 S components were lost during the fractionation and purification of the 30 S proteins. Single component omission experiments indicated that all purified proteins, except Pl(S1) and P3b,c(S6), are required for full functional activity.
* This is Paper l&t8 of the Laboratory of Genetics. $ Present address, I)cp:trtment of Agricultural Chemistry, Nagoya IJniversity, Nagoya Japan.
It is still not clear whether Pl(S1) should be considered a "true" 30 S protein or a ribosomal-associated "factor." The 30 S ribosomal subunit of E.schericlG coli consists of one 16 8 KNh molecule and about 20 protein molecules.
The 50 S subunit consists of one 23 S RNh molecule, one 5 S RKX molccule, and about 30 to 35 protein molecules (for review, see [1][2][3][4]. WC have previously shown that bacterial 30 S subunits can be reconstit,uted from their dissociated molecular components (5). This indicates that the information lor the correct assembly of ribosomal particlcs is contained in the struct,ure of their molecular caomponents, aud not in any nonribosomal factors.
In addition, the reconst,itutioii system has established a m&hod of analyzing functions of individual molecular components (for reviews, see 1 and 6).
Our initial reconstitution stwlies ~vere done using purified 16 8 REAL and a 30 S ribosomal protein mixture extracted from purified 30 S subunits.
In order t,o wcwmplish unambiguous identificat~ion and functional analysis of all the essential molecular components of the 30 S subunit', it is necessary to separate and purify each of the proteins contained in the 30 H subunits and then to show complctc rcconst~it~ution of functionally active particles from the 16 S RNA and a mixture of each of the purified proteins.
Preliminary reports of these cxpcriments have alrcaly been published (7,14,15). in this paper, we describe in detail our method of purification of each of the 30 S ribosomal proteins, as well as characterizatioii of the 30 S ribosomal subunits rcconstituted from t,lic pure protein components and the 16 S RNh molecule. V: 13uffer IV containing 1 11 KCl. Buffer VI: 13uffcr IV containing 0. 5 11 KCl. J3uffer VII ("Reconstitution IMfer") : Buffer IJ' cont,aining 0.33 M KCl.
Preparation of Ribosornes and Ribosomal Proteins-IS. coli strain (213 (RNase I-) was grown in a 150-liter fcrmcntor at 37". The media contained the following components per liter: 5 g of bactotr?pt.one, 2.5 g of yeast extract, 2.5 g of nutrient broth, 9.0 g of XaeHPOI, 3.0 I?; of KHsP04, 3.0 g of (NH&SOI, 50 mg of 11gS01.7H,0, 5 mg of FeS04.711Z0, 25 mg of CaCl,. 211,0, and 10 g of glucose. Esponcntially growing cells were cooled alltl harvested (the Klett reading of the cultura at the time of harvest was approximately 500 (blue filter); the Klett reading of a stationary phase culture is about 2000). Approximately 1.0 to 1.5 kg of cell paste were obtained from the 150.liter culture and stored frozen at -70" in 100-g amounts.
JTTe have used two methods to obtain cell-free extracts. The "glass bead" method has been described (16). l\Iorc rcecntly we have found that better results are obt.ained using the aluminagrinding method (17). Cell paste, 100 g, x\-as ground in a large mortar with 200 g of alumina (Fisher) at 4". The alumina and broken cells were suspended in about 300 ml of Buffer I. DNase (Worthington, RNawfree) was added to about 2 pg per ml, the suspension incubated at 4" for 20 min, and then centrifuged at 15,000 rpm for 20 min (Sorvall centrifuge, SS34 rotor).
The supernat,ant. was saved, the alumina resuspended in about. 200 ml of Buffer I, and ccntrifugcd again. The two supcrnatants were combined and centrifuged at 27,000 rpm for 30 min (Beckman 30 rotor) to remove some traces of alumina and ccl1 debris. The crude extract t.hua obtained was centrifuged at 35,000 rpm for 12 hours (Beckman 35 rotor) to sediment, the ribosomes. The pellet was suspended in 60 to 100 ml of Buffer I and stored frozen at -TO", or dialyzed against Buffer II to dissociate the i0 S ribosomcs before subunit purificat,ion.
The yield of 70 S ribosomes was generally 500 to 600 14260 units per g (wet, weight) of cells.
1Vc Irave found that t,he best yields of ribosomal protcills are obtained from 30 S ribosomes prepared from crude 70 S ribosomes. Extensive puri&ation of 70 S ribosomes by sucrose washing (16) or by sucrose-salt washing (18) appears t,o result in loss OC some 30 S ribosomal proteins (especially 1)2(S2) and P3(S3)).I Although the 30 S subunits derived from crude 70 S ribownws are not as "clean" as those prepared from sucarose-salt washed 70 S, the conlamiliatiiig proteins do not interfcrc with the subsc~quent purificatioil of the 30 S ribosomal prot,c+is. Two mctllods for separat,ioii of 30 S and 50 S subunits have been uwtl.
The cc;rrelatioll of the two nomcnclnturcs has been made previolx+ly (9). This will be filrtlrrr disrrwieti in this p:ip<tr.
The pellet was dissolvctl in approsirnatcly 15 ml of Buffer I and tlie solution dia*lgzetl agaiilst 200 volumes of 13uffcr I.
The preparation of total 30 S protein (TP30)2 has lwcn described (16). For all subsequent steps in protein purification, 6 M urea which had been purificxl by ambcrlite alit1 charcoal trcatrnent (21) was used. LiCl stock solutions wcrc purified by charcoal treatment.
Four grams of charcoal (in 100 ml of H?O) wcrc added to 900 ml of LiCl solution, to give a final cow centration of 10 M LiCl, the misturc stirred for 1 hour, ant1 then filtered.
Yurijication of SO S Ribosornal Proteins-Although scvcral groups, including oursclvcs, have published procedures for purification of 30 S ribosomal proteins (7, [10][11][12][13]22), and a tentative correlation between proteins isolated by various woups has been made (9), there arc still some differences in h results obtained by various groups and some slight uncertainties in the correlation made. Iii addition, we have not tlrscribc~d in detail our own purification method, which is different froni those used by others. Fig. 1 shows our procedure with a typical result for the separation and purification of 30 8 ribosomal proteins.
The methods are sirnilar to those we have dccribcd previously (7), although some refinements have bwn made. The major changes have been the scaling up of the size of the phosphoccllulose columns and the amount of sample applied. Also, the gradient of the first phosphoccllulose column (PC-j, Fig. 1) was started at 0.15 IV{ LiCl rather than 0.2 M Li('1 as in our reported procedure (7).
WC have found that dialysis of Tl'30 against liuffcr I I I results frequently in precipitation of some proteins. l'l%(S18), for esamplc, is not very soluble in 6 M urea buffers unlew somr LiCl (or other suitable salts) is present.
It was then treatrd with 0.1 x NaOII at room temperature for 10 to 15 min, filtered, and washctl with \vatcr until the pII was about 8.0. The phosphocellulose W:I~ then suspended in water, met~hylaminc added (3 ml per liter), and the pH adjusted to 8.0 (or 6.5 wlml indicated) \T-ith concc~ntratcd phosphoric acid. The phosI~l~oct:llulosc was again filtcwtl and xashed three times with the starting buffer (0.15 x3 I,iCl ill 13uffer III).
The slurry was tlcgassed under racuu~n at room temperature for 20 to 30 rnin before pouring. The cwlumn (2.5 x 90 cm), packctl fairly tightly with the phosl~l~oc~cll~~lose, was tlwl equilibrated with starting buffer (0.15 RI LiCl il: l%uffcr III) at 4" until the pI1 of the effluent was the same as tlw buffer (pH X.0).
The sample, containing about 1 .O g of ribosomal protein (i'rom 50,000 112,, units of 30 S ribosomcn), was dialyzed, first' xgailrf approsimatcly 13 volumes of l%uffcr III overnight to rc~ll~c~: the LiCl concentration to approsimntcly 0.15 nr. The ~amplc was then dialyzed for 6 to 8 hours against 10 to 20 volumes of starting buffer before it \vas applied to the column.
After the column was washed with approximately 400 ml of the same buffer, the * The following abbreviations arc rlsed : TI'30, total 30 S proteins ext,r:tcted from 30s suhrmits with the Inxa-LiCl method (l(i).  Column fractions to be applied to another phosphocellulose with different batches of phosphocellulose. For example, column were suitably combined and dialyzed against the next PlO(S12) has not always been clearly separated from P14(S20) starting buffer. When protein fractions were to be applied to a and P15(S21). In such cases, P10(S12) was separated from Sephadex column, pertinent phosphocellulose column fractions P14(S20) and P15(S21) by a phosphocellulose column using a gradient of 0.4 to 0.7 M LiCl in Buffer III, pH 6.5. P14(S20) 3 Two proteins, P3b and P3c, have been isolated which corre-and P15(S21) were separated using the same conditions (K-5, spond to S6 (9). See the text for discussion. Fig. 1). The elution pattern of P13(S19) also varied with different batc~hcs of 1~~~ospl~ocellulosc. 1'13(SL9) sometimes eluted with 1'11 (S14) rather than l'lOn(S13) as sl~own in Fig. 1 (PC-I).
During the various purification steps, the amount of coiitamiliatiug pot&is was usually tloterminetl by gc~l clcctrophoresis of colu~ii~i fractiow.
The final recovery of a purifiwl proteilr (molts of pure proteiu per mole of starting 30 S slibuuit~) may dcpencl on severa factors, such a-; the numbc~r of *tc:ps wquirccl to obtaiil the l)ure protein, and whrtlicr tlw prottGi is feud in "fractiod" or "uiiit" amounts (24) in isolatctl 30 S subunits.
RSA and proteins were mised in the molar ratio of 1: 1.8 unless othwwise specified.
The volume ratio of Huffcr IV to Buffer VI in the final reconstitution misture was 1: 2, giving an optimal KC1 conceutration for recoiistitution of 0.33 31 (28). The reconstitution misture, containing between 3.0 and 10 rln~o units of 16 S RXA per ml, n-as then heated for 60 min at 40" (or as indicated).
The misture was cooled on ice, and reconstituted particles were isolated by centrifugation in the rcconstitution buffer ("isolated particles"). Rltcrnativcly, the particles were purified by sucrose gradient ccntrifugation (5 to 205, in Buffer VII) ; the peak fractions were pooled and particles were recovered by ceiitrifugatioii ("purified particles"). The particles wrc suspcnt1ed in lluffer I, and their functional activities determined (see below).
Iii this case, the assay method described previously (16) n-as motlified slightly ,so that the amount of reconstitution clurillg the assay coultl bc minimizetl.
Other . The reaction misturr n-as incubated at 30" for 5 mill, tcrminatccl by addition of 1.0 ml of cold buffer (10 ma1 Tris.IICl, 1111 7.4 at. 24"; 10 mu magnrsium acetate; 50 mx XII&l), ant1 filtered slowly through Mllipore filters (IIAl\Vl' 2400) follo~etl by n-aahi1l.g with the same buffer three times. The filters were then dried and counted in a liquid scintillation counter with a toluene-based scintillation fluid. Purified initiation Factors IF-1 and IF-2 were generously supplied by Dr. J. W. Hershey (cf. 34). All assays were tested and found to be linear with respect to ribosome concentrations used. Reference 30 S subunits were preheated at 40" for 20 min in Buffer I before assay. Preincubation in Buffer VII (reconstitution buffer) gave similar activities. One-dimensional polyacrylamide electrophoresis was done at pH 4.5 as described previously (16). Two-dimensional polyacrylamide electrophoresis of ribosomal proteins was done according to the method of Kaltschmidt and Wittmann (19). l'rotein concentration was measured by the method of Lowry et al. (35), using crystalline bovine serum albumin as a standard. The color yield of the bovine serum albumin in this reaction was compared with that of TP30, purified 30 S ribosomal proteins P4a(S5) and P4b(S8).
At protein concentrations where usual assays were done, TP30 and 1'4a(S5) gave the same color yield as bovine serum albumin, whereas P4b(S8) gave a color yield about 7 % less than bovine serum albumin.

RESULTS
Purijied SO S Ribosomal Proteins Used for Reconstitution-As described under ",Materials and Methods," we have purified 23 proteins from 30 S subunits ( Fig. 1 and Table I).
The purity of these proteins was examined by one-dimensional polyacrylamide electrophoresis at pH 4.5 (10% polyacrylamide). As shown in Fig. 2(a), most of the proteins showed a single major band. Those proteins which migrate together under the standard conditions were then examined for possible cross-contamination using other methods.
For example, P3(S3), P3b(S6), and PACT migrate together in the standard gel electrophoresis at PI-I 4.5 (7, 16), but are separable in 8% acrylamide at pH 8.6. (The latter are the conditions used in the first dimension in the two-dimensional electrophoresis (19)). As can be seen in Fig (c) 8% gels, pH 8.5. The samples (15 pg each) were applied at the centers of gels. The condition is identical with that used for the first dimension in two-dimensional gel electrophoresis. Anode is at the left, and cathode is at the right.
It should br llotcltl here that both PlO(Sl2) and Pgb(S17) usually give two +ots in typo-tlirncilsional clectrophoresis. Electrophoresis of PlO(S12) and 1'9b(S17) under "oxidized" and "reduced" colltlitiolls shown in Fig. 3(b) indicates the faster moving spot (of PlO(S12) or I'gb(S17)) is due to oxidation of the protein during polymerization of the sample gel, or during elcctrophorcsis, or both. The elcctrophoretic mobility difference bctwecn the ositlizeti and the reduced forms of Pgb(S17) was prcviousl\ noted by Iiardy el al. (IO).
In illcx ('asc of the gel clec~trol)horetic alialysis me1ltioried above, hevrral known amounts of proteins wcrc applied to g(>ls, alltl the tlegrccl of contamillat,ioil was determined \-isually by conipariug -lained inteilsity of coirtaminating proteins on gels witI1 the intalsit,y of the standards. With these methods, all of the purified protein preparations (except possibly 1'1 (Sl)) wwc estimatctl to be at leaqt 90y0 pure, and in most casts, more than 95c/, plU'C.
I'3a was used in the reconstitution studies cleL+cribrtl in our earlier report.
-1lt.hough the 1)ossibility cannot be csclutled that 1'3a is a ribosomal protein alIt is important for sotne fun&ion we have not tc*ted, we have omitted this protein from the standard protein mixture for reconstitution experiments described in this paper. Two proteitls (P3b and 1'3~) corresponding to S6 (Wittmann's nomcnclaturc') have been purified ill our laboratory from E. coli strains K-12 (see above, and Fig. 2 (c)) and 1111E600,4 whereas other laboratories have reported only 011~ protein (9). ,%mino acid sequence analysis of the two proteins from MRE600 indrates that the primary sequences of the first several NEIs-terminal residues arc identical." Also, reconstitution erperimciits indicate that 1'3b and P3c arc functionally intclrchangeable (7, 36). Both P3b and 1'3~ correspond to Protein S6 isolated by \I'ittmann et al., a5 judged by clectrophoretic mobility as well as immunological behavior (9). Possibly, one of these proteins is a derivative of the other.
Complete sequence analysis should provide the answer.
In the present reconstitution experiments, a misture of equal amounts of both 1'3b and 1'3~ was used and treated an a single protein (called P3b,c or S6).
Althougll wc have some doubts about tho validity of c>alling Pl(S1) a 30 S ribosomal protein (see below), we included l'l(S1) in the standard 30 S protein mixture.
Thus, omitt,ing P3a and counting 1'3b and 1'3~ as one protein, we used altogether 21 purified proteins in the reconstitution experiments described in this paper.
Reconstitution with Standard .Ilixtltre of PuriJied Proteins and Ifi X RXd-Twenty-one purified protc& wvcre mixed in an equal molar ratio as described under "Materials and Methods" (called the standard purified protein mixture).
Various amounts of this standard purified protein mixture were adclcd to a constant amount of 16 S RNA and incubated at 40" for I hour under optimal reconstitution conditions. The particles were then assayed directly for their activity in poly(U)-directed ['Qphenylalanine incorporation.
The results are shown in Fig. 4. It can be seen that maximum activity is reached whet1 the molar rat'io of the standard protein mixture to 16 S RNA is about 1.7 to 2.4, indicating some protein or proteins are present in less than the calculated amount. This may be due to errors in our estimate of the amount of som~~ prot,eins using the Lowry reaction (see "Materials and Illethods"). Alternatively, it may be due to selcctivc loss of some proteins during several rnanipulations inclutling dialysis (adsorl)tion to dialysis tubes or glass walls), partial inactivation of some proteins during purification, or combinations of these possibilities.
In all experiments rem ported here, 1.8 to 2.0 molar equivalents of proteins were used per mole of 16 S R?;A.
The general shape of the curve shown in Fig. 4 is similar to that obtained in previous experiments using unfractiollated total 30 S proteins (37). In thr previous experiments, the results were interpret.ed to indicate a high degree of cooperativity in the as;sembly of 30 S particlcs.
However, because of the above melitioned uncertainty in the exact amount of individual proteins used in the present experiments, it is difficult to make a definite interpretation of the shape of the curve in this case. The high cficiency of recon&ution shown in Fig. 4 suggests that ow standard purifirtl protein mixture contains all the cssent.ial protein components of the 30 S subunits.
Kinetics of Reconstitution-Wc have compared the kinetics of reconstitution in the present system with reconstitution using unfractionated 30 S ribosomal proteins (TP30). Fig. 5 shows that the kinetic data of reconstitution at 40 and 30" are essent,ially the same whether using TP30 or purified 30 S proteins. is almost completely absent in the purified reconstituted particles as in 30 S subunits extensively purified by high salt washing. Various other assays related to 30 S functions were also used to assess the functional activity of the reconstituted particles. Purified reconstituted particles were used in this case. As shown in Table III, the reconstituted particles showed activities comparable to reference 30 S subunits in the fillet-puromycin reaction (in the presence of 50 S subunits), the binding of termination codon UAX (in the presence of 50 S subunits and chain termina-FIG. 7. Two-dimensional polyacrglamide g,el electrophoresis of protein extracted from purified reconstit,uted particles @Pi) and 30 S subunits. Protein was extracted from 30 S subunits or purified reconstituted particles by RNase digestion method (38) and analyzed by two-dimensional polyacrylamide electrophoresis. tion Factors RF-l or RF-2) as well as the binding of Phe-tRNA directed by poly(U) in the presence of 50 S subunits. The binding of fMet-tRN-4 in the presence of purified initiation factors, IF-1 and IF-2, by the reconstituted particles was somewhat lower (48 to 80%) than the reference 30 S subunits. Since our reference 30 S subunits are not salt-washed, it is possible that some non-ribosomal protein factors which stimulate fleet-tRNA binding (such as IF-l or IF-3) are present in the reference 30 S subunits. Such factors are not present in the reconstituted particles assembled from purified known components. The lower activity of the reconstituted particles could be explained on this basis. AUG-directed fMet-tRNA binding was also assayed in the presence of IF-2 only. The activity of the reconstituted particles relative to the reference 30 S subunits in this assay was significantly lower than in the presence of both IF-l and IF-2. These results are consistent with the explanation given above. Either IF-1 or IF-3, in the presence of IF-2, is known to stimulate fMet-tRNA binding (34,39,40). It is possible that IF-1 (or IF-3, or both) is present in the reference 30 S subunits in significant amounts.
The data shown in Table III also suggest that purified reconstituted particles are somewhat less active thau the isolated (unpurified) reconstituted particles (cf . Table II), both in poly(U)directed polyphenylalanine synthesis and phage RNA-directed polypeptide synthesis. It is possible that the extra purification steps remove some important proteins or cause inactivation of a small fraction of the reconstituted particles. However, further studies are required to establish this conclusion.
Single Component Omission Experiments-The data described so far indicate that 21 purified proteins together with 16 S RNA are su$cient to reconstitute 30 S particles with properties similar to the original reference 30 S subunits. We have then esamined the question of whether all of the 21 proteins used are required for reconstitution of functional 30 S particles. For this purpose, single component omission experiments were performed; the rcconstitution r?as performed using a mixture of purified proteins with a single protein omitted. Activity of the reconstituted particles in various functional assays was then analyzed either directly or after isolation and purification of the reconstituted particles. Preliminary experiments of this kind have been described in an earlier paper (7). However, in the earlier experiments, protein-deficient particles were prepared using several protein fractions (with known protein composition) in combination with several purified proteins. This was done in order to conserve purified protein preparations.
Moreover, as mentioned before, the uncharacterized Fraction PM containing PlOb(S15) was used instead of purified PlOb(S15). In addition, a mixture of P9a- ture of purified 30 S proteins with one protein omitted as indicltcd and hratcd at 40" for 1 hour under the st,:tndard reconstitrltion conditions. An aliqrlot was then taken for measurement of activity in poly(U)-dire&cd phenylalaninc incorporation (direct assay). The remaining particles were pru-ified on 5 to 20':; SIIcrose gradients in reconstit\ltion buffer. The peak fractions were pooled, concentrated by centrifugation, s\lspcnded in Bluffer I, and assayed for activity in poly(U)-directed phenylalxnine incorporation (:rssa~-after purification).
Figrn-es are per cent of the activity with all proteins.
IXfferent preparations of plu-ified 30 S proteins were used in the two experiments. (S16) and 1'9b(Sli) was used, ant1 the rcquirfmcnt of es& alone n-as not stutlicd.
\Tc have rcpcatcd the same kind of csperimcllts using the 21 purifictl proteins dccribtttl above. The results obtained in two indcpcntlcllt espcrimelits using different. protein preparations are shon-11 ill Tablr IV. In this table, the results 011 requirements of in&\-itlual prot,cGns for poly(U)-directed pal\-pheliylalaniile synthesis are givcu.
The activity in various ot,lier func~tional assays lv'iits also dctcrminctl in similar csperimeiits, but will be reported in a .scparatc pal)c~-.
'Tn-0 aqsaps were pcrformetl in each of the tn-0 csperiments dcscsribcd in Table IV: (u) tlircct assay without isolation of particles; ant1 ib) assay after I)urifica,tion of the particles.
For compari--on, the data show11 in the earlier report (7) is also included. Sever:11 conclusio~is can be drawn from the data.
We have found that in the abscncc of both Pga(S16) and Pgb(S17) or I'Sa(S16) alonc, the rate of reconstitution is very slow compared to the complete system containing; these proteins.
The optimum conditions for rrconstitution jn the abscncc of these protctins arc tlifferent from the standard condit~ioiis.
The activit,ies of these protein-tleficient particles relative to control particles can vary considerably depcntling on the duration of incubation or changes in the conditions of rccaonstitution.
The unique role of these proteins in the assembly reaction will be described separately.6 zlnother factor may be variability in the tlcgrc~~ of inactivation or purity of individual proteins in different protein preparatious, or loss of some proteins in the process of making various protein mixtures from individual proteins.
Such possibilities have been mentioned before in this paper.
It is our espericnce that more reproducible results mere obtained when csperiments were repeatcd using the same protein preparations.
3. Several proteia-deficient reconstitutctl particles (e.g. par& clcs reconstituted in the absence of Pl2(Sl8)) showed a hi& activity when assayed directly, but only a IT-eak activity when assayed aft,cr purification of the particles in the same esperimcnt It is probable that P12(S18), for example, is not directly required for function but is required for the binding of some other proteins. These proteins may be weakly bound to the reconstituted particles in the absence of 1'12(S18), but lost during purification, resulting in a drastic &crease in the activity of the reconstituted particles.
Interdependence of protein bindings during assembly process has been st.utlied, including the rolr of 1'12(Sl8) in the assembly process (36).6

Stoichionzetry 01 Inclividual
Proleins Required jar Mazivzum Activity in Reconstitution-Since we now can reconstitute 30 S subunits from 16 S RNA and a mixture of 21 purified proteins, T\-e should be able to determine the amount of each protein needed for marimum reconstitution activity 11cr unit amount of 16 S RNA.
In these cspcrimeiits, all proteins, except the protein being tested, were added in 2.0 molar equivalents relative to 16 S RNA1. Various amounts of the protein to be tested were then added, and reconstitution was performed.
In the case of PY(Sll), a slightly higher than l-eq amount, (1.2 to I .4 eq) was required, whereas somewhat less than 1 eq (about 0.8 eq) of PE(S21) was required for maximum activity. Ilowevcr, these numbers are preliminary.
Again, the main uncertainty may come from errors in estimating the amount of individual proteins by the Lowry's reaction using serum albumin as a standard (however, P4a(S4) was shown to give almost the same color yield as serum albumin (see "Materials and Methods")), and the possible loss of proteins during dialysis to remove urea before reconstitution. IHowever, with further experiments, these possible errors could be avoided, and it should be possible to dct,errnine accurate stoichiometry for most of the proteins required for reconstitution of active 30 S subunits.

DISCUSSION
Ribosomal proteiiis from E. coli 30 S subunits have been isolat.ed and purified in several laboratories.
Correlation has been made between proteins from different laboratorics and 21 proteins, designated as Sl to S21, have been gcncrally accepted as 30 S ribosomal proteins (9). t,eins obtained with the present method is satisfactory for most of the experiments designed for functional analysis of these proteins.
The experimental results presented in this paper indicate that the 21 purified proteins together with 16 S RN;1 molecule are sufficient to reconstitute 30 S particles with functional and physical properties nearly identical with the original reference 30 S subunits from which the proteins were isolated.
It is important to note in this connection that our reference 30 S subunits are very active as judged by the rate of poly(U)-directed polyphenylalanine synthesis as well as their activity in LUG-directed fMet-tRNA binding.
For example, in the presence of purified IF-l and IF-2, 3.2 pmoles of the 30 S subunits bound 1.36 pmoles of radioactive fMet-tRNA molecules. ilssuming one fhlet-tRr\'X binding site ("initiation site") per 30 S particle, the data indicate that at least 42% of the 30 S ribosomal particles are in an active state with respect to the initiation function. 4 Since there might be some noa-ribosomal stimulatory factors missing in our assay system and the experimental conditions used may not be optimum for the binding, 427, may be a minimum figure.
Thus, highly efficient reconstitution relative to the reference 30 S subunits in the present system means that we have probably not missed any ribosomal protein with an important function.
However, this does not necessarily mean that all the 21 proteiiis used for the reconstitution experiments arc "true" ribosomal proteins, especially since several of them are present in amounts less than a single copy per particle in isolated ribosomal preparations ("fractional proteins") (24, 41). Single component. omission experiments have shown that most of the proteins listed arc, in fact, required for full activity of ribosomes in various functional assays (the data shown in Table IV and other unpublished experiments, as well as the data published previously (7)). The main exception is Pl(S1) and PYb,c(S6).
The situation F&h 1'1 (Sl), howcrcr, is diffcrent. Omission of l'l(S1) from the reconstitution misturc does not cause a reduction in any of the ribosomal functions t,ested (Table IV and other unpublished  espcrimcnts; pee also (7)). This protein usually fails to get incorporated into the rcconstit.uted ribosomc under reconstitution conditions, which illcluclc the use of high ionic strength buffers. As shown in the present paper, purified reconstituted particles lack 1'1 (Sl) almost. completely (see also (36)).
Isolated 30 S ribosomal particles contain only about 0.1. to 0.3 copy of this protein per 30 S parbicle (41). The chemical properties of l'l(S1) arc also tliffcrcnt front all otller ribosomal proteins.
In many csperiments we have found that reconstitutccl particles have polypeptide-synthesizing activity higher thrill the rcference 30 S subunits.
As discussed under "Results" in conncction with variability of the activity data given in Table II, reconstituted particles with activity higher than reference 30 S subunits may reflect the real situation. This is consistent w&h the previous observation that the isolat,ed E. coli 30 S subunits The reconstitution mixture contained 1G S RNA, a two molar excess of all the purified 30 S proteins except the protein being examined (indicated in the figure), and the indicated amounts (expressed as moles of protein per mole of 16 S RNA) of the protein omitted.
After incubation at 40" for GO min, aliquots (0.2 APeO unit) were assayed for their activity in poly(U)directed polyphenylalanine synthesis using the direct assay procedure.
In this paper, we have first described in detail our own purification method of 30 S ribosomal proteins.
TTT-o differences are noted from the generally accepted list of 30 S proteins.
First, TT-e have isolated two proteins, 1'3b and 1'3c, which correspond to S6. hs discussed under "Results, " it appears that one protein is a derivative of the other.
The significance of this observation is not clear and must await further chemical studies on these proteins.
Secondly, w-e have isolated protein P3a I\-hich does not correspond to any of the proteins other workers have isolated. Ss already mentioned, our previous rcconstitutioii experiments failed to reveal any significant role of this protein in reconstitution. Thus, this protein has been omitted from our list of 30 S ribosomal proteins in accordance wit,h other laboratories.
Purity of the separated individual proteins is important fox functional analysis of the ribosomal proteins using the present reconstitution system. Purity was judged by polyacrylamidc gel electrophorctic analysis using various systems and by immunochemical methods.
& described under "Results," most of the proteins (except possibly l'l(S1)) used for the present studies were more than 95y0 pure.
In addition, sequence analysis of the 30 S proteins (15 out of 21 so far analyzed) purified by the present mcthotl has shown that the sequence of the first 20 to 40 amino acid residues from the NH, terminus is unique in all cases and is different from one another (except 1'3b and I'3c which have identical scqucnccs, see above).
No significant cross-contamination has been 0bservcd.j However, small amounts of contaminating proteins IT-ere occasionally observed.
For example, the protein 1'3(S3) preparation in Fig. 2 shows a second weak band corresponding to 1'5(S7). The conOamination XT-as estimated to be about 570. Since we added 1.8 to 2 moles of each prot'ein per mole of RN-4 in the single component omission experiments, this contamination would mean that our (-P5(S7)) protein mixtures might hare cont,ained up to about 0.1 mole of P5(S7) per mole of 16 S RNL Thus, some weak residual activity observed with (-1'5(S7)) protrin mixtures (see Table IV) could be explained, at, least partly, on this basis. Nevcrthelcss, purity of the pro-are partially deficient in some ribosomal proteins ("fractional proteins") (24,41). Stimulation of the activity of isolated 30 S subunits upon addition of extra ribosomal proteins (24,43) is also in agreement with the above interpretation.
In this connccbion, it may be interesting to study the stoichiometry of each of t,he protein components in the isolated reconstituted 30 S particles obtained under the optimum conditions which ensure the masimum reconstitution efficiency. As mentioned above, the observed weak "residual activity" of the protein-deficient particles described in the single component omission experiments (Table IV) could be explained, in some cases (such as the omission of P5(S7)), by presence of the omitted protein in other preparations as contaminants. However, in several other cases, the residual activity is too high to be accaounted for by contamination of the omitted protein in other protein preparations.
Whenever the possibility of such contamination existed in these cases, the protein mistures prepared by omitting the pertinent protein were analyzed.
The omitted protein in the protein mixtures was either completely absent or t,oo small in amounts to account for the activity.
High residual activity observed after omission of some proteins may suggest that the omitted protein is not directly involved in the function of assembled 30 S subunits, but has a role in the assembly process. We have obtained data indicating that Protein PSa(S16) belongs to this category.6 Alternatively, the omitted protein may be important to maintain an "active center" in a proper configura tion in the ribosome structure; without that prot.ein only a fraction of the reconstituted particles may have the "active ceuter" in a correct structure, and hence, the reconstituted particles are partially active.
Other possibilities also exist and have been clis-(Bussed previously (I, 6). The present reconstitution system may be useful in analysis of the role of each of the molecular components in the aysernbly process, as well as in ribosomc function (1,(6)(7)(8)36). In addition, the present system offers an analytical method for detecting a protein which is identical or functionally equivalent to a particular 30 S ribosomal protein.
We have already used the present system successfully to show that most, if not all, of the E. coli proteins have functionally equivalent count,erparts among protcills extracted from Bacillus sfearotlrer,?2op/2il1ts 30 S subunits (15).