Electron Microscopic Mapping of Secondary Structures in ... - NCBI

Vol. 141, No. 1

JOURNAL OF BACTERIOLOGY, Jan. 1980, p. 365-373 0021-9193/80/01-0365/09$02.00/0

Electron Microscopic Mapping of Secondary Structures in Bacterial 16S and 23S Ribosomal Ribonucleic Acid and 30S Precursor Ribosomal Ribonucleic Acid THOMAS D. EDLINDt* AND ALIX R. BASSEL Department of Microbiology, State University of New York, Upstate Medical Center, Syracuse, New York 13210

Electron microscopy revealed reproducible secondary structure patterns within partially denatured 16S and 23S ribosomal ribonucleic acid (rRNA) from Escherichia coli. When prepared with 50% formamide-100 mM ammonium acetate, 16S rRNA included two small hairpins that appeared in over 50% of all molecules. Three open loops were observed with frequencies of less than 25%. In contrast, 23S rRNA included a terminal open loop and two additional large structures in over 75% of all molecules. These secondary structure patterns were conserved in the 16S and 23S rRNA from Pseudomonas aeruginosa. The secondary structure of the 30S precursor rRNA from the ribonuclease III-deficient E. coli mutant AB105 was mapped after partial denaturation in 70% formamide-100 mM ammonium acetate. Two large open loops were superimposed on the 16S and 23S rRNA secondary structure patterns. These loops were the most frequent structures found on the precursor, and their stems coincided with ribonuclease III cleavage sites. A tentative 5'-3' orientation was determined for the secondary structure patterns of 16S and 23S rRNA from their relative locations within 30S precursor rRNA. The relation of secondary structure to ribosomal protein binding and ribonuclease III cleavage is discussed.

Escherichia coli rRNA is extensively intrastrand base paired, both in aqueous solution and within the ribosome (1, 21). A specific fraction of this secondary structure may contribute to ribosome structure. Part of this contribution would include binding sites for ribosomal proteins. For example, the E. coli ribosomal protein Li binds regions of E. coli and Bacillus stearothermophilus 23S rRNA's that show little sequence, but considerable structure, homology (24). Secondary structure has also been implicated in the processing of E. coli rRNA transcripts by the double-strand-specific endonuclease RNase III (7, 15). This enzyme is deficient in the mutant AB105, which accumulates a 30S precursor (pre-rRNA) with the composition: 5'16S, 23S, 5S-3' (3, 9, 28). The partial denaturing conditions used in electron microscope studies of HeLa cell rRNA and its precursors eliminate all or nearly all secondary structure in E. coli rRNA and pre-rRNA (14,26). On the other hand, using gene 32 protein staining, large loops have been observed in E. coli rDNA carried by a 480 transducing phage (28). We show here that conditions similar to those used to map secondary structure in phage t Present address: Department of Medical Biochemiistry, College of Medicine, Texas A&M University, College Station, TX 77843.

365

RNA (8) reveal reproducible patterns of hairpins and open loops in E. coli 16S and 23S rRNA. These patterns are largely conserved in the rRNA of Pseudomonas aeruginosa, a bacterium outside the family Enterobacteriaceae and hence only distantly related to E. coli. Highly stable loops whose stems coincide with sites of RNase III cleavage enclose the 16S and 23S rRNA patterns within E. coli 30S pre-rRNA. MATERIALS AND METHODS RNA preparation. S30 extracts were prepared as described previously (2) from E. coli strain A19 and P. aeruginosa strain 01. Ribosomal pellets were obtained by centrifugation of the S30 ribonucleoprotein at 100,000 x g for 2 h. The pellets were washed and suspended in 100 mM Tris-hydrochloride (pH 8.5). Sodium dodecyl sulfate was added to 1%, the ribosome preparation was extracted three times with phenol, and the RNA was precipitated from the aqueous phase with ethanol at -20°C. The RNA pellet was suspended in a 10 mM Tris-hydrochloride (pH 7.5)-100 mM NaCl-1 mM EDTA resuspension buffer. The RNA was centrifuged at 40,000 rpm for 7 h in a Beckman SW-40 rotor through 2.5 to 25% linear gradients of sucrose in the resuspension buffer. The 16S and 23S peak fractions were precipitated with ethanol, suspended in 10 mM phosphate buffer (pH 7.6)-10 mM EDTA, and stored at -70°C until used. E. coli strain AB105 (AB301-105) was obtained from the E. coli Genetic Stock Center, Department of Hu-

366

EDLIND AND BASSEL

man Genetics, Yale University, New Haven, Conn. Growth, chloramphenicol treatment, and RNA extraction were performed as described by Schlessinger et al. (22). The RNA was suspended in a 100 mM Trishydrochloride (pH 7.5)-5 mM EDTA-0.25% sodium dodecyl sulfate resuspension buffer and heated at 1000C for 1 min before centrifugation. The RNA was centrifuged for 6.5 h at 40,000 rpm in the Beckman SW-40 rotor through 2.5 to 25% linear gradients of sucrose in the resuspension buffer. Fractions from the 30S region were pooled and precipitated with ethanol, suspended in 10 mM phosphate buffer (pH 7.6)-10 mM EDTA, and stored at -70°C until used. Electron microscopy. RNA was prepared for electron microscopy by a minor modification of the method described previously (8). Twenty-five microliters of a solution containing 0.5 to 1 jig of RNA per ml, 2 mM phosphate buffer (pH 7.6), 2 mM EDTA, 100 mM ammonium acetate, 0.01% cytochrome c, and 50 or 70% formamide was warmed for 30 s in a 22 to 230C water bath before spreading on a hypophase of distilled water. The RNA-protein monolayers were picked up on carbon-coated grids within 2 min after spreading, dehydrated in ethanol, and rotary shadowed with uranium oxide. Negatives were taken with a Siemens Elmiskop IA at X20,000 or Elmiskop 102 at X25,000 and enlarged photographically to x100,000. Magnification was calibrated with a 2,160-lines/mm carbon grating replica. Measurements were made with a Dietzgen map measurer.

RESULTS Denaturation conditions were chosen to extend each RNA sufficiently to mi'nimize measuring errors and ambiguities. Optimal spreading and secondary structure were obtained for 16S and 23S rRNA by using 50% formamide-100 mM ammonium acetate, whereas 70% formamide was required for 30S pre-rRNA due to a higher degree of intramolecular pairing in this molecule. Spreads of 16S and 23S rRNA from 70% formamide were also done for direct comparisons to molecules of 30S pre-rRNA. All fulllength molecules, defined for each preparation as those within the range of the average length plus or minus one standard deviation (Table 1), were mapped. All measurements were normalized to the modal length of that RNA spread from 70% formamide. Electron micrographs and interpretive drawings of 16S and 23S rRNA from E. coli and P. aeruginosa and 30S pre-rRNA from E. coli AB105 are shown in Fig. 1. Labeled arrows point to the major secondary structures described below. The preparation method limited the resolution of secondary structure to about 0.03 jim

J. BACTERIOL. TABLE 1. Molecular lengths and kbase values of E. coli rRNA and pre-rRNA Length Spreading RNA conditions kbase ± (% formAvg ± SD' Modal SDb (103 amide) (,Um) (jtm) nucleotides) 50 0.42 ± 0.04 16S 70 0.41 ± 0.05 0.40 1.54 23S 50 0.78 ± 0.08 70 0.75 ± 0.09 0.75 2.9 ± 0.3 70 30S 1.33 ± 0.10 1.35 5.2 ± 0.4 a SD, Standard deviation. Average was calculated from molecules within 2 SD of the modal length. bBased on published value for E. coli 16S rRNA

(7). total length. Two types of secondary structure were mapped, hairpins (short paired regions less than 0.06 jim long) and larger structures involving pairing between more widely separated regions of the RNA strand. Large structures usually formed open loops but in some cases appeared closed or collapsed. Secondary structure maps of 10 to 20 typical molecules of E. coli 16S and 23S rRNA and 30S pre-rRNA are shown in Fig. 2. These were representative of the patterns, frequencies, and variation observed. The 5'-3' orientation was assigned as shown in all figures based on considerations described below. 16S rRNA. An electron micrograph and drawing of a typical molecule of E. coli 16S rRNA spread from 50% formamide is shown in Fig. la. Representative maps are included in Fig. 2a, and the total mapping data are summarized in the histogram of Fig. 3a. In a molecular length of 0.40 jim (Table 1), four hairpins were seen. These occurred at 0.03 ± 0.01, 0.12 ± 0.02, 0.21 ± 0.02 (labeled A in Fig. 1), and 0.32 ± 0.02 (B) ,um from the 5' end. The hairpins at 0.03 and 0.13 jim had frequencies of 25 and 10%, respectively, whereas those at 0.21 and 0.32 jim were more stable, occurring in 50% of all molecules. Three different open loops were seen. Two apparently included the 5' terminus and were 0.07 ± 0.02 (C) and 0.30 ± 0.03 (D) jim long. The third open loop was central, 0.17 ± 0.02 (E) jum long and beginning 0.13 ,um from the 5' end. The central open loop appeared twice as frequently (20%) as the two terminal loops (10% each). However, all open loops found on 16S rRNA were weak compared with those found on 23S rRNA and 30S pre-rRNA (see below). E. coli 16S rRNA spread from 70% formamide

FIG. 1. Electron micrographs (top) and interpretive drawings (bottom) of rRNA's from E. coli (a, b, d, e) and P. aeruginosa (c, f) and AB105 30S pre-rRNA (g, h). (a-c) 16S rRNA; (d-f) 238 rRNA. (a, c, d, f) Spread from 50% formamide; (b, e, g, h) spread from 70% formamide. The 5' end of each molecule is oriented to the left. Labeled arrows point to major secondary structures described in the text. Bar = 0.1 ,im.

dT A1 I'

B --

It a

B

Al/

w

b

P t;

"-B

c

AF (t d

H G

G

f

e

I-,E D

~~~~~~~~~B

I--,

E

/~~~~~~~-K

g

h 367

368

EDLIND AND BASSEL

J. BACTERIOL.

_a

--

_

Am=_

-

_

_ _ _2 _3 _4

C

-

m -~~~~~~~~~~~~ -

_ I 0

- ~~~~~M I

.2

I

-

I

I

.4

.6

I

I

I

1.0

.8

Distance from 5'end

I

I

I

-

1.2

(/em)

FIG. 2. Representative secondary structure maps of E. coli rRNA. (a, b) 16S rRNA; (c, d) 23S rRNA; (e) 30S pre-rRNA. (a, c) Spread from 50%o formamide; (b, d, e) spread from 70%o formamide. Shaded areas above the base lines represent paired regions; open areas represent the open portions of open loops.

(Fig. lb, 2b, and 3b) retained most of the secondary structure described above. However, the hairpin at 0.32 ,im had decreased in frequency to 20%. An electron micrograph of P. aeruginosa 16S rRNA spread from 50% formamide is shown in Fig. lc, and a histogram of the mapping data is shown in Fig. 4a. Hairpins at 0.04 ± 0.02, 0.20 + 0.03, and 0.31 ± 0.03 (B) ,im from one end were seen in 10, 20, and 25%, respectively, of all molecules. Two terminal open loops 0.07 ± 0.02 (C) and 0.29 ± 0.01 pm long occurred in 10% of all molecules, whereas 20% had a central loop 0.15 ± 0.04 pum long beginning 0.15 ,um from one end. This secondary structure pattern was similar to that seen with E. coli 16S rRNA (above),

although the hairpins were seen half as frequently with P. aeruginosa 16S rRNA. 23S rRNA. As shown in Fig. ld, 2c, and 3c, E. coli 23S rRNA spread from 50% formamide had three secondary structures involving longrange pairing, as well as several hairpins. An 0.02 (labeled F in Fig. 1) pm open loop 0.13 long was located at the 5' end of the 0.75-,umlong molecule (Table 1). The two other large structures usually formed open loops, but occasionally appeared closed, and were 0.10 0.03 (G) and 0.11 0.03 (H) pum long, beginning 0.45 ±

±

±

and 0.58 ,um, respectively, from the 5' end. Hairpins could be found within any of these three loops. More often, hairpins were seen outside of the loops, at 0.25 ± 0.04, 0.40 ± 0.03, and 0.72

SECONDARY STRUCTURE OF BACTERIAL rRNA

VOL. 141, 1980

369

100

8 00-a

6;04 .0 2

6;o-

b

U) n

01)

E a 0

4 0-

2 0-

0

.1

.2

.3

.4

0

0-

Distance from 5'end ( eLm) FIG. 3. Percentage of molecules showing secondary structure in a given interval (0.02-,im width) of E. coli rRNA. (a, b) 16S rRNA; (c, d) 23S rRNA; (e) 30S pre-rRNA. (a, c) Spread from 50% fornamide; (b, d, e) spread from 70%o formamide. Hatched areas represent hairpins (paired structures