Wheat Germ Cytoplasmic Ribosomes STRUCTURE OF RIBOSOMAL SUBUNITS AND LOCALIZATION OF P,W-DIMETHYLADENOSINE BY IMMUNOELECTRON MICROSCOPY*

Cytoplasmic ribosomes have been isolated from wheat germ, and the structure of ribosomal subunits has been examined by electron microscopy of negatively stained preparations. Small (40 S) subunits show structural features generally regarded as characteristic of eukaryotic particles, while large (60 S) subunits show shapes that are equally well described by models of prokaryotic 50 S particles. Small subunit 18 S RNA contains 2 residues of p,p-dimethyladenosine 19 and 20 residues from the 3'-end (Hagenbuchle, O., Santer, M., Steitz, J. A., and Mans, R. J. (1978) Cell 13, 551-563). Nucleoside analysis by high performance liquid chromatography shows no other residues of this component in the RNA. Anti-dimethyladenosine immuno- globulins were reacted with wheat germ 40 S subunits, and the resulting complexes were studied by electron microscopy in order to localize the nucleoside. In about 90% of the complexes observed, antibody-subunit con- tact was consistent with a single binding site. We place the dimethyladenosine residues at or near the end of the platform of the 40 S particle in a position nearly equivalent to that previously identified in prokaryotic and chloroplast subunits (Trempe, M. R., and Glitz, D. G. (1981) J. Biol. Chern. 256,11873-11879).

from wheat germ, a common source of experimental materials for in vitro protein biosynthesis. In wheat germ 18 S rRNA the dimethyladenosine residues occur 19 and 20 nucleotides from the 3' terminus (10). Antibodies to dimethyladenosine have allowed localization of the modified residues in E. coli (12) and pea chloroplast (13) 30 S subunits and, along with the localization of the 3'-end of the RNA (14), suggest the placement of the message-positioning Shine and Dalgarno (15) sequence of prokaryotic rRNA on the subunit platform. In this paper we describe the preparation of wheat germ ribosomal subunits and their characterization by electron microscopy. We then demonstrate the availability of the 3'terminal segment of 18 S RNA and localize the dimethyladenosine residues near the end of the platform of the 40 S particle.

MATERIALS AND METHODS
Preparation of Ribosomes and Ribosomal Subunits-Wheat germ 80 S ribosomes were prepared by a modification of the method of Treadwell et al. (16). Fifteen grams of raw wheat germ (Richards Food Corp.) were ground in a mortar on ice with an equal volume of sand for 10 min and mixed with 150 ml of extraction buffer (6 mM KHC03, 90 mM KC1, 4 mM magnesium acetate, 2 mM CaCl,, 6 mM 2-mercaptoethanol, 10% v/v glycerol). After 12 min of centrifugation at 15,000 rpm in a Sorvall SS-34 rotor, the fat layer was removed from the top and the supernatant was decanted and filtered through two layers of cheesecloth. 0.1 volume of 0.1 M magnesium acetate and 0.1 volume of 1 M Tris-HCI, pH 7.8, were added to the filtrate, and the solution was centrifuged as above for 15 min. The upper about seven-eighths of the supernatant was filtered through cheese cloth, layered over 1.5-ml pads of 20% sucrose in extraction buffer, and centrifuged in a Beckman 65 rotor for 90 min at 60,000 rpm. The resulting pellets were resuspended in buffer (20 mM Hepesl-KOH, pH 7.6,150 m M KC1,5 mM magnesium acetate, 6 mM 2-mercaptoethanol, and 10% glycerol) to give an AZrn of about 800. If not used immediately, aliquot8 of the suspension were quick-frozen and stored at -20 or -70 "C.
Ribosomal subunits were prepared by a modification of the method of Spremulli et al. (17,18). Aliquots of 120 AZrn units (about 10 mg) of ribosomes were diluted with 10 volumes of dissociation buffer (150 mM KCl, 1 mM magnesium acetate, 0.1 mM EDTA, 6 mM 2-mercaptoethanol, 50 mM Tris-HC1, pH 7.7), containing 5% sucrose. After incubation at 30 "C for 5 min, the samples were applied to 56 ml of in a Beckman SW 25.2 rotor at 19,000 rpm for 16 h at 10 "C.
Analysis of Nucleosides-RNA was isolated from purified wheat germ 40 S and E. coli 30 S subunits by phenol extraction (19). The RNA was hydrolyzed to nucleosides (20) with bacterial alkaline phosphatase ribonuclease A and ribonuclease T, (Sigma) and snake venom phosphodiesterase (Worthington). Nucleosides from the hydrolysis of 0.2-1 AzM units of rRNA were quantitated by HPLC on a Supelco 5 micron reversed-phase column (250 X 4.6 mm) using a modification of the method of Buck et al. (21). A flow rate of 2 ml/ min was used in a gradient from buffer A (0.25 M ammonium acetate, pH 6.0) to buffer B (40:60 (v/v) acetonitrile/water) as follows: initial conditions, 100% A, at 10 min, 100% A; a t 45 min, 25% A via Waters curve 8; a t 50 min, 100% B via Waters curve 6. Absorbance was monitored a t 260 nm. and the data were integrated with a Shimadzu C-R3A Chromatopac. Purified nucleosides used as standards were purchased from Pharmacia LKB Biotechnology Inc.
Preparation and Characterization of Antibodies-Antibodies were induced in rabbits by toe pad injection of a covalent m2Ado-protein conjugate. The procedures used in the synthesis and characterization of nucleoside conjugates, immunization, blood collection, serum preparation and antibody purification, and characterization have been described (12,14).
Synthetic Oligodeorynuccleotides-Sequences complementary to the 3' terminus of wheat germ 18 S rRNA and E. coli 16 S rRNA were synthesized using 8-cyanoethyl phosphoramidites (22) and a Vega Biotechnologies/Dupont Coder 300 automated synthesizer. Oligomers were purified by ion exchange HPLC using a Bio-Rad TSK-DEAE-5PW column with a Waters chromatograph; products were eluted with a hyperbolic gradient (curve 7) of 0-1 M NaCl in 20 mM Tris-HCI, pH 7.5. Salts were removed by dialysis against water (Spectrapore l dialysis tubing), and the oligonucleotides were concentrated in a Speed-Vac (Savant Instruments). Radiolabeling at the 5' terminus and analysis of oligodeoxynucleotide binding with a nitrocellulose filter assay have been described (23). Poly(U)-stimulated binding of radiolabeled phenylalanyl tRNAPh' followed Zamir et al. (24).
Electron Microscopy-For the characterization of ribosomal subunit preparations, particles were diluted to an A2M of 0.1-0. negative staining with 1% uranyl acetate according to the double carbon technique (25). Electron micrographs were obtained with a JEOL lOOB microscope operated a t 80 kV and a magnification of approximately 69,000.

Wheat Germ
Cytoplasmic Ribosomes negatively stained with 1% uranyl acetate, and observed by electron microscopy. Our isolation procedure yielded particles we considered to be satisfactory in purity and apparent structural integrity. A field of such particles is shown in Fig. 1. The subunits appear quite homogeneous, and distinctive and reproducible features can be seen in the micrographs. In order to define an antibody binding site on a three-dimensional model, it is necessary to evaluate complexes in more than one two-dimensional projection. Therefore, several hundred micrographs were studied in order to define the most common and reproducible projections.
Most of the images of 40 S subunits could be clearly identified with the three representative projections (intermediate, quasi-symmetric, and asymmetric) first described by Lake (26)  The asymmetric view is shown in its two enantiomorphic projections in rows C and D. These images emphasize features considered characteristic of eukaryotic 40 S subunits: a "bill" protruding from the head, an apparent split in the platform, and additional material or lobes at the base (1-3, 6). These images and the interpretative drawings at the end of each row are consistent with previous observations of other eukaryotic 40 S subunits. Characterization of 60 S Subunits-A number of methods of preparation of 60 S subunits were evaluated. A micrograph of particles purified by two rounds of sucrose gradient sedimentation and concentration by ultracentrifugation, steps commonly used in subunit purification, is shown in Fig. 3A. Although in about 100 micrographs at least 90% of the particles appeared to be 60 S subunits, only 1-2% of the images showed a clearly defined stalk, the most useful reference feature for interpretation of micrographs and identification of the different projections. Moreover, the 60 S population did not appear very well preserved or homogeneous. If minimal handling of the ribosomes was emphasized the apparent quality of the subunit was improved. Freshly isolated 80 S ribosomes were dissociated, sedimented through sucrose gradients, and the 60 S fraction was directly prepared for microscopy. Fig. 3B shows a field of such subunits. In more than 100 micrographs about two-thirds of the particles were clearly defined 60 S subunits; the majority of the others were 40 S subunits plus some 80 S ribosomes. Even though a stalk is identifiable in a relatively low percentage of the large subunit images (about lo%), the 60 S subunits appear to be a rather homogeneous population.
The gallery of 60 S subunits in Fig. 4 is arranged to show characteristic projections that we consistently observed. They will be described using the Lake nomenclature (1,2, 26). Row A shows both enantiomorphic projections of one of the quasisymmetric views, characterized by three protuberances: the

Wheat Germ Cytoplasmic Ribosomes
stalk, the headlike central protuberance, and the shoulder. In this projection of the subunit the stalk is seen at about a 45" angle from the vertical axis. Row B shows both enantiomorphs of a quasi-symmetric view which corresponds to a different orientation of the particle in space; the stalk projects horizontally with respect to the central protuberance. Rows C and D show images that lack the stalk. In row C the particles appear symmetric about the vertical axis. They could.correspond to either of the two views described above. In the images in row D, a region of accumulated stain can be seen between the central protuberance and the shoulder. These images were observed quite frequently in the micrographs. They could represent a projection in which the subunit orientation is such that the stalk is obscured by the particle. Row E shows another quasi-symmetric projection of the particle in which the stalk is clearly identifiable. Row F shows the asymmetric view.
Our interpretation of the structure of wheat germ 60 S particles is consistent with a common basic morphology of large ribosomal subunits, they resemble E. coli 50 S particles in overall shape but show a deeper notch between the central protuberance and the body, and/or the protuberance is a t a different (sharper) angle. Consistent with observations of Boublik et al. (6) we do not see the "eukaryotic lobe" or the "bulge" described by Lake (1, 2).
Quantitation of Base-methylated Adenosines in Wheat Germ 18 S rRNA-Wheat germ 18 S rRNA and E. coli 16 S rRNA, used as a control, were hydrolyzed to nucleosides and analyzed by reversed-phase HPLC. The integrated Azo was used to determine the amounts of each nucleoside in the hydrolysates. A level of 1.65 mol of m:Ado/mol of RNA was found in wheat germ 18 S rRNA, and E. coli 16 S rRNA hydrolysates gave 1.7 mol of m!Ado/mol of RNA. Wheat germ 18 S rRNA was also found to contain NG-monomethyladenosine at about 1 mol of modified nucleoside/mol of rRNA. No m'Ado was detected in E. coli 16 S rRNA hydrolysates.
Binding of a Complementary Oligodexoynucleotide to the 40 S Ribosomal Subunit-The oligonucleotide dCAATGATC-CTTC, which complements the 3"terminal nucleotides of wheat germ small subunit RNA (lo), was chemically synthesized, purified by HPLC, and labeled with '*P at its 5"end.
Varying amounts of oligomer were incubated with a fixed quantity of 40 S subunits, and oligonucleotide-subunit complexes were quantitated. Binding reached a plateau of about 0.3 mol of oligomer/mol of 40 S in the presence of a 2-fold molar excess of nucleotide; a similar level of binding of a 3'terminal complementary oligonucleotide was observed with E. coli 30 S subunits (see also Ref. 23), and in each instance the level of oligodeoxynucleotide binding was slightly greater than the poly(U)-stimulated binding of ["Clphenylalanyl tRNAPhe to active 30 S subunits. We interpret the results to indicate that the 3"terminal region of 18 S rRNA is available for binding on the surface of the 40 S subunit.
Localization of Dimethyludenosine-Purified wheat germ 40 S ribosomal subunits were incubated with anti-m$Ado IgG.
After removal of unreacted antibodies by size exclusion HPLC, the ribosomal subunits were negatively stained and examined in the electron microscope. Two types of antibodysubunit complexes were observed monomers, in which a single subunit has one IgG molecule bound, and dimers, in which two subunits are linked by an antibody. preferentially orient with respect to the carbon layer in the asymmetric projection, in antibody-subunit complexes the particles are most commonly seen in the quasi-symmetric projection. Bound antibody appears to introduce a constraint on the way the particles are adsorbed to the carbon layer.
The location of dimethyladenosine in the 40 S particle was approximated by analysis of the apparent point of contact of antibody with the subunit in each characteristic projection. Only those complexes that showed an identifiable antibody in contact with a structurally intact subunit were used in the analysis. About 70 micrographs containing approximately 7 x 10:' small subunits were used, and 549 antibody-subunit interactions were identified and evaluated. In 429 monomers 88% of the contact points were consistent with .a single binding site. In antibody-linked subunit dimers, 90% of the contact points were consistent with the same binding site, as described below.
A gallery of ribosomal subunit-antibody monomers is shown in Fig. 6. The subunits are arranged according to the characteristic projections. In the intermediate view, row A, the antibody molecules are seen attached at or near the end of the platform. The cleft between the subunit body and the platform was often obscured, but the lack of basal lobes and the bill allows differentiation of this view from the asymmetric projection. In the asymmetric view, row B, antibodies bind to the convex side of the particle. This projection was particularly useful in placement of the binding site on the platform far from the bill. In the quasi-symmetric view, rows C and D, the point of antibody contact is at or slightly below the partition between the upper and lower sections of the subunit. These images were useful in placing the contact site on the vertical axis, even when the antibody approach is from the top (e.g. row D, frames 1 and 2 ) . Fig. 7 shows a gallery of antibody-linked subunit dimers in which each characteristic projection is represented. Antibody contact occurs at the same site described above. For example. in row A, frame 2, and row C, frame 3 bound at or near the end of the platform in the intermediate view; in row A, frame 1, and row B, frame 2, the site of contact is seen at the convex side of the subunit in the asymmetric view. Geometric constraints appear to limit the frequency with which antibody appears to be fully extended (e.g. frame A-2, in which both subunits are in the intermediate projection); in most instances the IgG molecule is partially obscured by one or both subunits.
These results are consistent with a placement of the two dimethyladenosine residues at or near the end of the platform, as illustrated in Fig. 8. This site is roughly equivalent to that previously seen with both E. coli (12) and chloroplast (13) subunits. Therefore, the location of the dimethyladenosine residues is essentially conserved in the three-dimensional structure of the small ribosomal subunit as well as in the primary structure of the RNA.

Ribosome and Subunit Preparation and Characterization-
Several purification schemes were tested to obtain preparations of wheat germ ribosomal subunits suitable for study by electron microscopy, i.e. that appeared homogeneous and structurally intact. No single purification procedure was acceptable (by this criterion) for the preparation of both 40 and 60 S subunits. Highly purified 40 S subunits were fully satisfactory for the studies reported here. In contrast, extensive purification of 60 S subunits resulted in the loss of characteristic features (e.g. the stalk) and apparent quality (e.g. reproducibility of images). The disappearance of the stalk of the wheat germ 60 S particle upon centrifugation can be compared with the loss of the stalk from the E. coli 50 S subunit upon treatment with NH,Cl and ethanol (27) or the loss of ribosome integrity upon purification noted by Boublik and Ramagopal (28). Our results re-emphasize the effects that preparation methods can have on observed structures and the need for caution in evaluating structural characteristics seen in micrographs, as stressed by van Heel and Stoffler-Meilicke (29).
Nucleotide Analysis and Auailubility-HPLC analysis of wheat germ small subunit RNA showed no more than 2 m!Ado residues/molecule. Terminal sequence a~alysis (10) showed that 2 residues of mgAdo occur 19 and 20 nucleotides from the 3'-end. Hence, these must be the only residues of this modified component in the RNA. Binding of an oligodeoxynucleotide complementary to the 3"terminal sequence suggests that this region of the RNA is on the subunit surface and likely to be available for interaction with antibody. The equivalent region of E. coli 16 S RNA is similarly capable of binding a complementary oligomer (23,30) and functions in the alignment of mRNA on the small subunit (15). Although a similar message placement function in eukaryotes has been postulated, it is not clear that the 3"region of 18 S functions in this manner (31); our data indicate only that such a function remains possible.
Localization of Dimethyladenosine in 40 S subunits-Antibodies of well established specificity (12) were used to place W,W-dimethyladenosine on the wheat germ 40 S subunit. This antibody preparation strongly binds the modified nucleoside, and interaction with unmethylated (or monomethylated) adenosine is negligible. (Since wheat germ 18 S RNA contains 1 m'Ado residue, a possible interference of the monomethylated compound had to be t.aken into account.) Our interpretation of the electron micrographs of 40 S subunits reacted with anti-W,W-dimethyladenosine IgG is that in the great majority of the complexes (about 90%) the antibody is bound at a single area at or near the end of the platform. This placement of mgAdo within the wheat germ 40 S subunit shows the modified nucleosides in a conserved position in an eukaryotic cytoplasmic particle relative to their location in either E. coli or chloroplast ribosomes (12,13). The two adjacent W,W-dimethyladenosine residues that are nearly universally conserved in the sequence at the 3'-end of small subunit RNA thus appear to be very highly conserved in the three-dimensional structure of the ribosome as well.
The location of the mzAdo residues in all three types of ribosomes is on the platform of the small subunit, i.e. in an area of the particle that is involved in subunit-subunit interaction and in translation. Several studies have attempted to determine a specific function of the mzAdo residues (11); the methyl groups have minor effects on the interaction of the ribosomes with molecules involved in the initiation cycle and upon accuracy in translation in uiuo. Distinct characteristics are seen when conformational and thermodynamic effects of the methylated groups on RNA secondary structure are investigated, and it is postulated that the methylated bases, both through their hydrophobicity and their effect on RNA conformation, may control the "fine tuning" of the initiation process. The location of the m!Ado residues in wheat germ 40 S particles virtually assures a platform location for the 3'terminal sequence of the 18 S rRNA; cDNA binding shows that sequence to be at the subunit surface and further suggests some role in the function of the small subunit.