Cross-linking of streptomycin to the 50S subunit of Escherichia coli with phenyldiglyoxal

[3H]Dihydrostreptomycin was covalently linked to the 50S subunit of Escherichia coli K12A19 with the bifunctional cross-linking reagent phenyldiglyoxal. The cross-linking was abolished under conditions that prevent the specific interaction of streptomycin with the ribosome. The binding primarily involved the ribosomal RNA and also a limited number of proteins, namely, L2, L6, and L17. This suggests that the binding domain for streptomycin is close to the peptidyl transferase center, in the valley between the central protuberance and the wider lateral protuberance of the 50S subunit. This domain faces the binding domain for streptomycin which we have previously characterized on the 30S subunit [Melan?on, P., Boileau, G., & Brakier-Gingras, L. (1984) Biochemistry 23, 6697-6703]. Our results indicate that the 50S subunit is involved in the binding of streptomycin to the bacterial ribosome, in addition to the 30S subunit which is generally considered as the specific target of the antibiotic. They are consistent with the occurrence of a single binding site for streptomycin on the ribosome, comprised of regions of both subunits.     

Protein-protein cross-linking of the 50 S ribosomal subunit of Escherichia coli using 2-iminothiolane. Identification of cross-links by immunoblotting techniques

We have investigated the protein-protein cross-links formed within the 50 S subunit of the Escherichia coli ribosome using 2-iminothiolane as the cross-linking reagent. The members of the cross-links have been identified by immunoblotting from one-dimensional and two-dimensional diagonal sodium dodecyl sulfate-polyacrylamide gels using antisera specific for the individual ribosomal proteins. This method also allowed a quantitation of the yield of cross-linking for each cross-link. A total of 14 cross-links have been identified: L1-L33, L2-L9, L2-L9-L28, L3-L19, L9-L28, L13-L21, L14-L19, L16-L27, L17-L30, L17-L32, L19-L25, L20-L21, L22-L32, and L23-L34. Our results are compared with those of Traut and coworkers (Traut, R. R., Tewari, D. S., Sommer, A., Gavino, G. R., Olson, H. M., and Glitz, D. G. (1986) in Structure, Function and Genetics of Ribosomes (Hardesty, B. and Kramer, G., eds) pp. 286-308, Springer-Verlag, New York). Our cross-linking data allow us to propose the approximate locations of eight proteins of the 50 S ribosomal subunit that so far have not been localized by immunoelectron microscopy and they thus contribute considerably to our knowledge of ribosome structure.     

The plastid ribosomal proteins. Identification of all the proteins in the 50 S subunit of an organelle ribosome (chloroplast)

We have completed identification of all the ribosomal proteins (RPs) in spinach plastid (chloroplast) ribosomal 50 S subunit via a proteomic approach using two-dimensional electrophoresis, electroblotting/protein sequencing, high performance liquid chromatography purification, polymerase chain reaction-based screening of cDNA library/nucleotide sequencing, and mass spectrometry (reversed-phase HPLC coupled to electrospray ionization mass spectrometry and electrospray ionization mass spectrometry). Spinach plastid 50 S subunit comprises 33 proteins, of which 31 are orthologues of Escherichia coli RPs and two are plastid-specific RPs (PSRP-5 and PSRP-6) having no homologues in other types of ribosomes. Orthologues of E. coli L25 and L30 are absent in spinach plastid ribosome. 25 of the plastid 50 S RPs are encoded in the nuclear genome and synthesized on cytosolic ribosomes, whereas eight of the plastid RPs are encoded in the plastid organelle genome and synthesized on plastid ribosomes. Sites for transit peptide cleavages in the cytosolic RP precursors and formyl Met processing in the plastid-synthesized RPs were established. Post-translational modifications were observed in several mature plastid RPs, including multiple forms of L10, L18, L31, and PSRP-5 and N-terminal/internal modifications in L2, L11 and L16. Comparison of the RPs in gradient-purified 70 S ribosome with those in the 30 and 50 S subunits revealed an additional protein, in approximately stoichiometric amount, specific to the 70 S ribosome. It was identified to be plastid ribosome recycling factor. Combining with our recent study of the proteins in plastid 30 S subunit (Yamaguchi, K., von Knoblauch, K., and Subramanian, A. R. (2000) J. Biol. Chem. 275, 28455-28465), we show that spinach plastid ribosome comprises 59 proteins (33 in 50 S subunit and 25 in 30 S subunit and ribosome recycling factor in 70 S), of which 53 are E. coli orthologues and 6 are plastid-specific proteins (PSRP-1 to PSRP-6). We propose the hypothesis that PSRPs were evolved to perform functions unique to plastid translation and its regulation, including protein targeting/translocation to thylakoid membrane via plastid 50 S subunit.     

Analysis of the two steps in polypeptide chain initiation inhibited by pactamycin.

Earlier work has shown that the inhibition by pactamycin (PM) of polypeptide chain initiation in reticulocyte extracts is associated with (1) a defect in the joining of the 60S subunit to the smaller initiation complex to form an 80S complex ("joining reaction") (Kappen, L. S., Suzuki, H., and Goldberg, I. H. (1973), Proc. Natl. Acad. Sci. U.S.A. 70, 22) and (2) a block after the synthesis of the initial dipeptide (Kappen, L. S., and Goldberg, I. H. (1973), Biochem. Biophys. Res. Commun. 54, 1083). The relative contributions of these two effects to the action of PM and their relationship to one another were evaluated in a system employing sparsomycin that permits both initiation at a certain number of initiation sites and limited oligopeptide formation without termination and release. The degree to which PM blocks the "joining reaction" and leads to the accumulation of 48S initiation complexes that either remain free or are bound to polysomes without the corresponding 60S subunit ("half-mers") was estimated by treatment of polysomes with RNase. Met-tRNAfMet binding factors are required to stabilize the RNase-generated 48S complexes. Under conditions where the initiation factor required for the "joining reaction" functions catalytically, presumably by cycling on and off initiation complexes, PM usually inhibits 80S complex formation 50-70%. Where "joining" is not limiting (presence of at least stoichiometric amounts of joining factor or high Mg2+ concentration) PM leads to the maximal accumulation of the initial dipeptide, Met-Val, in the P-site on the ribosome, indicating a block in a subsequent step in elongation. Binding studies with [3H]PM and the inability of PM to inhibit elongation of preformed Met-Val indicate that PM must interact with the ribosomes at an early stage of initiation. Taken together these data are compatible with the suggestion that PM does not interfere with the ribosomal "joining reaction" per se, but prevents the release and reuse of the joining factor, and in so doing blocks a step in elongation after formation of the initial dipeptide and its translocation to the P-site on the ribosome.     

Methylation of Ribosomal Proteins in Escherichia coli

Escherichia coli was grown in a medium containing [1-(14)C]methionine and [methyl-(3)H]methionine, and the (3)H/(14)C ratio was determined for each of the ribosomal proteins derived from the 70S ribosome. Evidence indicates that six proteins from the 50S subunit were methylated: L7, L9, L11, L12, L18, and L33. Methylation of several other 50S proteins (such as L1, L3, L5, etc.) may also occur. The methylated amino acids in protein L11 have been characterized further and found to be predominately epsilon-trimethyllysine. A small amount of a compound tentatively identified as N(G), N'(G)-dimethylarginine was also detected.     

On the role of cyclic AMP-independent protein kinases in the modification of yeast ribosomal proteins in vivo

Two proteins of yeast 40S ribosome subunit and four proteins of the 60S ribosome subunit were labelled in vivo with [32P]orthophosphate. Five of these proteins were phosphorylated by protein kinase 3, an enzyme which is cyclic AMP-independent and uses ATP and GTP as phosphoryl donors. Two proteins, belonging to the 60S ribosome subunit were phosphorylated by another, highly specific, cyclic AMP-independent protein kinase 1 B. Both in vivo and in vitro the most extensively phosphorylated protein species were acidic proteins, L44, L45 (according to the nomenclature of Kruiswijk & Planta, Molec. Biol. Rep., 1, 409-415, 1974) possibly corresponding to bacterial L7 and L12 proteins. The 40S ribosomal protein, S9, analogous to mammalian S6 protein, was phosphorylated in vivo but was not phosphorylated in vitro by either of the cyclic AMP-independent protein kinases. The obtained results clearly indicate that cyclic AMP-independent yeast protein kinases might be involved in the modification in vivo of some ribosomal proteins, in particular of the strongly acidic proteins of 60S ribosome subunit.     

Functional consequences of the RNase H2A subunit mutations that cause Aicardi-Goutieres syndrome

Mutations in the three genes encoding the heterotrimeric RNase H2 complex cause Aicardi-Goutières Syndrome (AGS). Our mouse RNase H2 structure revealed that the catalytic RNase H2A subunit interfaces mostly with the RNase H2C subunit that is intricately interwoven with the RNase H2B subunit. We mapped the positions of AGS-causing RNase H2A mutations using the mouse RNase H2 structure and proposed that these mutations cause varied effects on catalytic potential. To determine the functional consequences of these mutations, heterotrimeric human RNase H2 complexes containing the RNase H2A subunit mutations were prepared, and catalytic efficiencies and nucleic acid binding properties were compared with the wild-type (WT) complex. These analyses reveal a dramatic range of effects with mutations at conserved positions G37S, R186W, and R235Q, reducing enzymatic activities and substrate binding affinities by as much as a 1000-fold, whereas mutations at non-conserved positions R108W, N212I, F230L, T240M, and R291H reduced activities and binding modestly or not at all. All mutants purify as three-subunit complexes, further supporting the required heterotrimeric structure in eukaryotic RNase H2. These kinetic properties reveal varied functional consequences of AGS-causing mutations in the catalytic RNase H2A subunit and reflect the complex mechanisms of nuclease dysfunction that include catalytic deficiencies and altered protein-nucleic acid interactions relevant in AGS.     

Mechanism of action of Moloney murine leukemia virus RNA-directed DNA polymerase associated RNase H (RNase H I)

The mechanism of action of the ribonuclease H (RNase H) activity associated with Moloney murine leukemia virus RNA-directed DNA polymerase (RNase H I) and the two-subunit (alpha beta) form of avian myeloblastosis virus DNA polymerase were compared by utilizing the model substrate (A)n.(dT)n and polyacrylamide gel electrophoresis in 7 M urea to analyze digestion products. Examination on 25% polyacrylamide gels revealed that a larger proportion of the RNase H I oligonucleotide products generated by limited digestion of [3H](A)(1100).(dT)n were acid insoluble (15-26 nucleotides long) than acid soluble (less than 15 nucleotides long), while the opposite was true for products generated by alpha beta RNase H. RNase H I was capable of attacking RNA in RNA.DNA in the 5' to 3' and 3' to 5' directions, as demonstrated by the use of [3H,3'- or 5'-32P](A)(380).(dT)n and cellulose--[3H](A)n.(dT)n. Both RNase H I and alpha beta RNase H degraded [3H]-(A)n.(dT)n with a partially processive mechanism, based upon classical substrate competition experiments and analyses of the kinetics of degradation of [3H,3'- or 5'-32P](A)(380).(dT)n. That is, both enzymes remain bound to a RNA.DNA substrate through a finite number of hydrolytic events but dissociate before the RNA is completely degraded. Both RNase H I and alpha beta RNase H were capable of degrading [14C](A)n in [3H](C)n-[14C](A)n-[32P](dA)n.(dT)n, suggesting that retroviral RNase H is capable of removing the tRNA primer at the 5' terminus of minus strand DNA at the appropriate time during retroviral DNA synthesis in vitro.     

A new definition for the consensus sequence of the peroxisome targeting signal type 2

All organisms except the nematode Caenorhabditis elegans have been shown to possess an import system for peroxisomal proteins containing a peroxisome targeting signal type 2 (PTS2). The currently accepted consensus sequence for this amino-terminal nonapeptide is -(R/K)(L/V/I)X(5)(H/Q)(L/A)-. Some C.elegans proteins contain putative PTS2 motifs, including the ortholog (CeMeK) of human mevalonate kinase, an enzyme known to be targeted by PTS2 to mammalian peroxisomes. We cloned the gene for CeMeK (open reading frame Y42G9A.4) and examined the subcellular localization of CeMeK and of two other proteins with putative PTS2s at their amino termini encoded by the open reading frames D1053.2 and W10G11.11. All three proteins localized to the cytosol, confirming and extending the finding that C.elegans lacks PTS2-dependent peroxisomal protein import. The putative PTS2s of the proteins encoded by D1053.2 and W10G11.11 did not function in targeting to peroxisomes in yeast or mammalian cells, suggesting that the current PTS2 consensus sequence is too broad. Analysis of available experimental data on both functional and nonfunctional PTS2s led to two re-evaluated PTS2 consensus sequences: -R(L/V/I/Q)XX(L/V/I/H)(L/S/G/A)X(H/Q)(L/A)-, describes the most common variants of PTS2, while -(R/K)(L/V/I/Q)XX(L/V/I/H/Q)(L/S/G/A/K)X(H/Q)(L/A/F)-, describes essentially all variants of PTS2. These redefined PTS2 consensus sequences will facilitate the identification of proteins of unknown cellular localization as possible peroxisomal proteins.     

Modification of Escherichia coli ribosomes with the fluorescent reagent N-[[(iodoacetyl)amino]ethyl]-5-naphthylamine-1-sulfonic acid. Identification of derivatized L31' and studies on its intraribosomal properties

N-[[(Iodoacetyl)amino]ethyl]-5-naphthylamine-1-sulfonic acid (IAEDANS) is a fluorescent reagent which reacts covalently with the free thiol groups of proteins. When the reagent is reacted with the Escherichia coli ribosome under mild conditions, gel electrophoresis shows modification of predominantly two proteins, S18 and L31', which become labeled to an equal extent. When the native (i.e., untreated) ribosome is dissociated into 30S and 50S subunits, only the 30S ribosomal protein S18 reacts with IAEDANS despite the fact that L31' is still present on the large subunit. Upon heat activation of the subunits, a procedure which alters subunit conformation, S18 plus a number of higher molecular weight proteins is modified, but not L31'; the latter reacts with IAEDANS only in the 70S ribosome or when it is free. In contrast to the relatively stable association of L31' with native or with dissociated ribosomes, dissociation of N-[(acetylamino)ethyl]-5-naphthylaminesulfonic acid (AEDANS)-treated ribosomes weakens the AEDANS-L31'/ribosome interaction, resulting, upon gel filtration analysis, in ribosomes devoid of this derivatized protein.     

Identification of the sites of action of SrmB, a DEAD-box RNA helicase involved in Escherichia coli ribosome assembly

DEAD-box RNA-dependent ATPases are ubiquitous enzymes that participate in nearly all processes involving RNA, but their detailed molecular functions remain generally unknown. SrmB, one of the five Escherichia coli DEAD-box proteins, participates in the assembly of the large ribosomal subunit notably by facilitating the incorporation of L13, one of the ribosomal proteins that bind 23S rRNA earliest. Previously, we showed that SrmB is tethered to nascent ribosome through interactions with L4, L24 and the region from domain I of 23S rRNA that binds them. To identify the sites of action of SrmB, we have characterized rRNA mutations that bypass SrmB requirement. Five of them affect the same position from two repeated heptanucleotides in domain II of 23S rRNA, whereas two others affect a complementary hexanucleotide in 5S rRNA. Thus the sites of action of SrmB differ from its tethering site. In the mature ribosome, one of the heptanucleotides participates in a highly compact structure that contacts L13, the '1024 G-ribo wrench'. In addition, we have observed that the assembly defect of ΔsrmB cells worsens as rRNA synthesis increases. Based on these results, we propose two non-exclusive scenarios for the role of SrmB in ribosome assembly.     

GTPase activation of elongation factors Tu and G on the ribosome

The GTPase activity of elongation factors Tu and G is stimulated by the ribosome. The factor binding site is located on the 50S ribosomal subunit and comprises proteins L7/12, L10, L11, the L11-binding region of 23S rRNA, and the sarcin-ricin loop of 23S rRNA. The role of these ribosomal elements in factor binding, GTPase activation, or functions in tRNA binding and translocation, and their relative contributions, is not known. By comparing ribosomes depleted of L7/12 and reconstituted ribosomes, we show that, for both factors, interactions with L7/12 and with other ribosomal residues contribute about equally and additively to GTPase activation, resulting in an overall 10(7)-fold stimulation. Removal of L7/12 has little effect on factor binding to the ribosome. Effects on other factor-dependent functions, i.e., A-site binding of aminoacyl-tRNA and translocation, are fully explained by the inhibition of GTP hydrolysis. Based on these results, we propose that L7/12 stimulates the GTPase activity of both factors by inducing the catalytically active conformation of the G domain. This effect appears to be augmented by interactions of other structural elements of the large ribosomal subunit with the switch regions of the factors.     

Exchange and stability of HeLa ribosomal proteins in vivo.

The relative stabilities of individual HeLa ribosomal proteins and their capacity for exchange between ribosome-bound and -free states in the cytoplasm were examined. Most ribosomal proteins on cytoplasmic ribosomes were found to have uniform, high stability as measured by comparing the short term (12-hour) to steady state (3-day) labeling ratios determined for each ribosomal protein. This would be expected if the proteins in ribosomes either were all stable or were all degraded as a unit. The data do not rule out the possibility that individual proteins have different stabilities prior to their assembly into ribosomes. Four proteins labeled atypically. One large subunit protein (L5) had a lower than average ratio. We interpret this low ratio as being due to a large free pool of this protein. Three proteins (L10, L28, S2) had higher than average ratios, interpreted as being due to reduced protein stability. Two of these proteins (L10, L28) with high ratios were also found to exchange in vivo. The exchangeable proteins may be subject to increased degradation during the time that they spend in the exchangeable free pool. The third protein (S2) with an atypically high ratio is thought to be degraded or altered while on the ribosome, or slowly lost as ribosomes age, because exchange of this protein was not detected. These interpretations and some alternate interpretations are explained. The exchange of three large subunit proteins (L10, L19, L28) was detected by labeling of protein after ribosome synthesis had been inhibited with actinomycin D. Autoradiography of two-dimensional polyacrylamide gels showed labeling of these spots.     

Expression, purification, and crystallization of natural and selenomethionyl recombinant ribonuclease H from Escherichia coli

Ribonuclease H (RNase H) from Escherichia coli is an endonuclease that specifically degrades the RNAs of RNA:DNA hybrids. The enzyme is a single polypeptide chain of 155 amino acid residues, of which 4 are methionines. To solve the crystallographic three-dimensional structure of E. coli RNase H by the multi-wavelength anomalous diffraction technique, we have constructed methionine auxotrophic strains of E. coli that overexpress selenomethionyl RNase H. MIC88 yields about 10 mg of selenomethionyl RNase H per liter of culture, which is comparable to the overexpression of the natural recombinant protein. We have purified both proteins to homogeneity and crystallized them isomorphously in the presence of sulfate. These are Type I crystals of space group P2(1)2(1)2(1) with the cell parameters a = 41.8 A, b = 86.4 A, c = 36.4 A, one monomer per asymmetric unit, and approximately 36% (v/v) solvent. Crystals of both proteins diffract to beyond 2-A Bragg spacings and are relatively durable in an x-ray beam. On replacement of sulfate with NaCl, crystals of natural RNase H grow as Type I' (very similar to Type I) at pH between 7.0 and 8.0; at pH 8.8, crystals of Type II are obtained in space group P2(1)2(1)2(1) with a = 44.3 A, b = 87.3 A, and c = 35.7 A. Type II crystals can be converted to Type I by soaking in phosphate buffer. RNase H crystals of Type II have also been reported by Kanaya et al. (Kanaya, S., Kohara, A., Miyakawa, M., Matsuzaki, T., Morikawa, K., and Ikehara, M. (1989) J. Biol. Chem. 264, 11546-11549).     

Localization of sites of photoaffinity labeling of the large subunit of Escherichia coli ribosomes by arylazide derivative of puromycin

Previous work (Nicholson, A. W., Hall, C. C., Strycharz, W. A., and Cooperman, B. S. (1982) Biochemistry 21, 3797-3808) showed that [3H]p-azidopuromycin photoaffinity labeled 70 S Escherichia coli ribosomes and that photoincorporation into 50 S subunit proteins was in the order L23 greater than L18/22 greater than L15. In the present work we report on immunoelectron microscopic studies of the complexes formed by p-azidopuromycin-modified 50 S subunits with antibodies to the N6,N6-dimethyladenosine moiety of the antibiotic. The p-azidopuromycin-modified 50 S subunits appear to be identical to unmodified control subunits in electron micrographs. Complexes of modified subunits with antibodies to the N6,N6-dimethyladenosine moiety of p-azidopuromycin were visualized in micrographs. Individual subunits with a single bound antibody (monomeric complexes) and pairs of subunits cross-linked by a single antibody (dimeric complexes) were separately evaluated and showed similar results. Two regions of p-azidopuromycin photoincorporation were identified. The primary site, seen in about 75% of the complexes, is between the central protuberance and small projection, on the side away from the L7/L12 arm, in a region thought to contain the peptidyltransferase center. The secondary site, of unknown significance, is at the base of the subunit maximally distant from the arm. These placements are essentially identical to those we observed in analyses of puromycin photoincorporation (Olson, H. M., Grant, P. G., Cooperman, B. S., and Glitz, D. G. (1982) J. Biol. Chem. 257, 2649-2656) and quantitatively similar to evaluations of monomeric puromycin-50 S subunit complexes. The data support the placement of proteins L23, L18/22, and L15 at or near the peptidyltransferase center at the primary site and suggest, in addition, that the secondary site includes a genuine area of puromycin affinity.     

Subunit arrangement in V-ATPase from Thermus thermophilus

The V0V1-ATPase of Thermus thermophilus catalyzes ATP synthesis coupled with proton translocation. It consists of an ATPase-active V1 part (ABDF) and a proton channel V0 part (CLEGI), but the arrangement of each subunit is still largely unknown. Here we found that acid treatment of V0V1-ATPase induced its dissociation into two subcomplexes, one with subunit composition ABDFCL and the other with EGI. Exposure of the isolated V0 to acid or 8 m urea also produced two subcomplexes, EGI and CL. Thus, the C subunit (homologue of d subunit, yeast Vma6p) associates with the L subunit ring tightly, and I (homologue of 100-kDa subunit, yeast Vph1p), E, and G subunits constitute a stable complex. Based on these observations and our recent demonstration that D, F, and L subunits rotate relative to A3B3 (Imamura, H., Nakano, M., Noji, H., Muneyuki, E., Ohkuma, S., Yoshida, M., and Yokoyama, K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 2312-2315; Yokoyama, K., Nakano, M., Imamura, H., Yoshida, M., and Tamakoshi, M. (2003) J. Biol. Chem. 278, 24255-24258), we propose that C, D, F, and L subunits constitute the central rotor shaft and A, B, E, G, and I subunits comprise the surrounding stator apparatus in the V0V1-ATPase.     

The proteins of the messenger RNA binding site of Escherichia coli ribosomes.

Oligo(U) derivatives with [14C]-4-(N-2-chloroethyl-N-methylamino)benzaldehyde attached to 3'-end cis-diol group via acetal bond, p(Up)n-1UCHRCl as well as with [14C]-4-(N-2-chloroethyl-N-methylamino)benzylamine attached to 5'-phosphate via amide bond, ClRCH2NHpU(pU)6 were used to modify 70S E. coli ribosomes near mRNA binding centre. Within ternary complex with ribosome and tRNAPhe all reagents covalently bind to ribosome the extent of modification being 0.1-0.4 mole/mole 70S. p(Up)n-1UCHRCl alkylates either 30S (n=5,7) or both subunits (n=6,8). rRNA is preferentially modified within 30S subunit. ClRCH2NHpU(pU)6 alkylates both subunits the proteins being mainly modified. The distribution of the label among proteins differ for various reagents. S4, S5, S7, S9, S11, S13, S15, S18 and S21 are found to be alkylated within 30S subunit, proteins L1, L2, L6, L7/L12, L19, L31 and L32 are modified in the 50S subunit. Most proteins modified within 30S subunit are located at the "head" of this subunit and proteins modified within 50S subunit are located at the surface of the contact between this subunit and the "head" of 30S subunit at the model of Stoffler.     

Immune electron microscopic localization of dinitrophenyl-modified ribosomal protein S19 in reconstituted Escherichia coli 30 S subunits using antibodies to dinitrophenol

Escherichia coli small ribosomal subunits have been reconstituted from RNA and high performance liquid chromatography-purified proteins including protein S19 that had been modified at its amino-terminal proline residue with 1-fluoro-2,4-dinitrobenzene. As detailed in the accompanying paper (Olah, T. V., Olson, H. M., Glitz, D. G., and Cooperman, B. S. (1988) J. Biol. Chem. 263, 4795-4800), dinitrophenyl (DNP)-S19 was efficiently incorporated into the site ordinarily occupied by S19. Antibodies to DNP bound effectively to the reconstituted subunits and did not cause dissociation of the modified protein from the subunit. Electron microscopy of the immune complexes was used to localize the modified protein on the subunit surface. More than 95% of the antibody binding sites seen were consistent with a single location of protein S19 on the upper portion or head of the subunit, on the surface that faces the 50 S particle in a 70 S ribosome, and in an area relatively distant from the subunit platform. The S19 site is close to the region in which 30 S subunits are photoaffinity labeled with puromycin. Protein S19 is thus near protein S14 in the small subunit and in proximity to the peptidyl transferase center of the 70 S ribosome.