Precursors of Ribosomal RNA in the Cellular Slime Mold Dictyostelium discoideum ISOLATION AND CHARACTERIZATION*

The pathway of ribosomal RNA biogenesis in Dirtyoste- lium discoideum has been defined through identification, isolation, and characterization of the rapidly labeled nuclear RNAs which are intermediates in the process. Compar-ison of the methylation patterns, base compositions, two-dimensional oligonucleotide maps, and hybridization properties of these intermediate RNAs with those of mature rRNAs has established clearly the precursor-product se- quence relationships supporting the following for rRNA

The pathway of ribosomal RNA biogenesis in Dirtyostelium discoideum has been defined through identification, isolation, and characterization of the rapidly labeled nuclear RNAs which are intermediates in the process. Comparison of the methylation patterns, base compositions, twodimensional oligonucleotide maps, and hybridization properties of these intermediate RNAs with those of mature rRNAs has established clearly the precursor-product sequence relationships supporting the following scheme for rRNA production and processing: ,' p25 S (28 S) + 25 S 37s---+ p17S(21S) -17 s The relationship of the 37 S RNA of Dictyostelium to primary rRNA transcripts of prokaryotes and other eukaryotes is discussed.
The program by which ribosomal RNA is transcribed and processed consists of some features which are universal, and others which have been altered in the evolution of higher from lower organisms (1). Thus, in both prokaryotes and eukaryotes the genes for the two large rRNAs are arranged in tandem and are co-transcribed from a single initiation site. In eukaryotes the common rDNA' transcription unit varies in size from 4.5 x 10" for mammals to 2.6 x 10" for lower eukaryotes; a corresponding full length primary transcript RNA can be detected in cells under normal physiological conditions. At least in animals, the primary transcript rRNA does not include sequences of the small 5 S rRNA, which is transcribed independently from cistrons unlinked to the rDNA (2-5).
In contrast, the 5 S RNA gene is a part of the smaller bacterial rDNA transcription unit (molecular weight about 2.1 x 10") and is cotranscribed with the 16 S and 23 S rRNA genes (6)(7)(8) Oligo(dT)-cellulose binding buffer is 0.5 M NaCl, 10 mM Tris, pH 7.5, 5 g/liter of sodium dodecyl sulfate, and 0.1 rnM EDTA; elution buffer is 10 mM Tris, pH 7.5, 0.5 g/liter of sodium dodecyl sulfate, and 0.1 mM EDTA.
Agarose gel RNA sample solution is 5 g/liter of sodium dodecyl sulfate, 100 ml/liter of glycerol, and 0.3 g/liter of bromphenol blue dye.
Gel elution buffer is 0.5 M sodium acetate, 10 mM Tris, 10 g/liter of sodium dodecyl sulfate, and 0.1 rnM EDTA, pH 7, plus 1 to 2 /*g/ml of carrier tRNA. All solutions to be used for RNA preparations were mixed with 1 ml/liter of diethylpyrocarbonate, and then were heat-sterilized in an autoclave.
Growth, Staruation, and Radioactive Labeling of Cells-Vegetative strain AX3 cells were grown axenically to a density of 2 x 106/ml in Medium MESHL5. They were then harvested and subjected to starvation conditions, either (a) on buffered filter paper or (b) in liquid suspension: (a) Cells designated as "developing" were plated on filters 42.5 cm in diameter, as described previously ( labeled with 60 mCi of :'ePh for the periods indicated in the figure legends and in the sections below. Cell Fractionation and RNA Extraction -Cells (2 x 10') were collected by centrifugation and washed with ice cold Buffer MES-PDF. The cell pellet was frozen in dry ice/ethanol for 10 min, and then was mixed with 4 ml of Cemulsol NPT-12 lysis buffer. The mixture was gently agitated on a Vortex mixer until the pellet thawed, and then for 30 s more. All subsequent operations were at O-4". Nuclei, recovered from the lysate by centrifugation at 12,000 x g for 1.5 min, were washed twice by resuspension in 5 ml of fresh lysis buffer, followed by recentrifugation. The final nuclear pellet was resuspended in 5 ml of HMK buffer in preparation for deproteinization.
For preparation of cytoplasmic polysomes, the initial cell fractionation supernatant was halved, and each 2-ml fraction was layered onto a gradient consisting of a 0.5 ml cushion of 60% (w/v) sucrose, overlaid with 2.5 ml of 7% (w/v) sucrose, all in HMK buffer. Centrifugation was for 35 min at 1" and 50,000 rpm in the Beckman SW 50.1 rotor. From each gradient the bottom 1.25 ml, containing purified polysomes and some monosomes, were recovered; the pooled polysome fractions were diluted to 5 ml with HMK buffer. For deproteinization, each preparation of nuclei or polysomes was made 0.2% (w/v) in sodium dodecyl sulfate and then 2% (v/v) in diethylpyrocarbonate.
Within 10 s after addition of the diethylpyrocarbonate, 1 volume of water-saturated phenol was mixed with the sample. Then 0.1 volume of 2 M sodium acetate, pH 7.3, was added, followed by 1 volume of chloroform:isoamyl alcohol (96:4). The mixture was agitated on a Vortex mixer for 3 min, then centrifuged at 4,000 x g for 7 min. The two phases separated by the centrifugation were each re-extracted separately: the upper aqueous phase plus interphase with 2 volumes of phenol:chloroform:isoamyl alcohol (50:48:2), and the lower organic phase with 2 ml of HMK buffer containing 0.2 M sodium acetate, pH 7.3. Both samples were agitated and centrifuged as above. The first (re-extracted) organic phase was discarded.
The second organic phase was removed from under the sample aqueous phase, and was re-extracted with the same 2.ml HMK buffer phase used for the first organic phase. After this second organic phase re-extraction, the two aqueous phases, plus interphase material, were pooled and re-extracted twice with 2.5 volumes of chloroform:isoamyl alcohol mixture (96:4) alone. The RNA in the final aqueous phase was precipitated with 2 volumes of ethanol. Cytoplasmic RNA was recovered by centrifugation at 10,000 x g for 10 min; nuclear RNA at 250,000 x g overnight.
RNA pellets were dissolved in 0.5 to 1.0 ml of oligo(dT)-cellulose binding buffer and applied to a 0.1. to 0.25-g column of oligo(dT)-cellulose, equilibrated in the same buffer. The column was washed with 10 volumes of binding buffer; the flowthrough solution contained most of the rRNA. RNA bound to the column was recovered when the column was washed with oligo(dT)cellulose elution buffer; the eluate fraction was made 0.4 M in sodium acetate, pH 5.2. Both flow-through and eluate fraction RNAs were harvested by precipitation with ethanol as described above. Agarow Gal Electrophoresis of RNA -The gel system used was a modification of that of Sharp&al. (16). The 1.4 to 1.5% (w/v) agarose (Calbiochem or Sigma electrophoresis grade) was melted in E buffer containing 10% (v/v) glycerol. Gels were cast in Pyrex glass tubes (0.6 cm inner diameter x 12 cm long). Slab gels were cast from the same buffered agarose solution, but without glycerol. Gels were prerun for 15 min under the conditions used for sample electrophoresis. Samples were suspended in a small volume (10 to 15 &cylindrical gel) of agarose gel RNA sample solution.
Sample electrophoresis was carried out at 8 mA/gel for 3.5 to 4.5 h on cylindrical gels, and at 150 V for 1.5 h on slab gels; slab gels were cooled by a fan. After electrophoresis, cylindrical gels were cut transversely with a series of razor blades into l-mm slices. For preparative gels, the Cerenkov radiation of each slice was measured directly (17). Analytical gel profiles were obtained by melting each slice in 1 ml of water, then determining radioactivity of the solubilized slice in 10 ml of Aquasol. Wet slab gels were covered with plastic wrap and exposed to film for autoradiography. Purification of Cytoplasmic The purified RNA pools were concentrated by precipitation with 2 volumes of ethanol, followed by centrifugation at 10,000 x g for 15 min. Each RNA pellet was dissolved in 0.5 ml of water and layered onto a 5-ml Sephadex G-25 M column in water. RNA was eluted from the column with water, and the salt-free void volume fractions were collected. The specific activity of the purified mature rRNAs was 1,500,OOO cpm/Fg. Purification of Nuclear rRNAs -Nuclear RNA was prepared from cells starved in suspension for 1 h and labeled at the same time with 60 mCi of :'2Pb. Usually further purification steps were carried out on RNA that flowed through oligo(dT)-cellulose, but 37 S RNA could also be purified from material that bound to the column under high salt conditions.
In either case, nuclear RNAs were fractionated by sedimentation through 11 ml 15 to 30% (w/v) linear sucrose gradients in Buffer A; centrifugation was for 7.25 h at 23" and 37,000 rpm in a Beckman SW41 rotor. Fractions from the radioactive peaks sedimenting at roughly 21 S and 28 S were pooled separately; all material sedimenting faster than 31 S was collected in a third pool. Each RNA pool was mixed with 5 to 25 pg of carrier tRNA, and was precipitated with ethanol. The RNA pellets were dried, resuspended in gel sample solution, and fractionated further by electrophoresis on slab or cylindrical agarose gels by the procedure detailed above. Regions of the gels which contained bands of 37 S RNA, ~25 S RNA, or p17 S RNA were excised, and each sample was forced through a disposable syringe (20-gauge needle) into a centrifuge tube. A volume of gel elution buffer at least twice the gel volume was added, and the sample was mixed vigorously at room temperature for 3 min. Large gel particles were removed by centrifugation at 4,000 to 10,000 x g for 5 to 10 min, and the supernatant was passed over glass wool to filter out remaining particles. Extraction of the crushed gel was repeated one to three times, until 90% of the radioactivity had been eluted. All extraction supernatants for a given sample were pooled, and the RNA was twice precipitated with ethanol. The individual RNA species were finally purified of traces of contaminating RNAs or gel residues by sedimentation in 15 to 30% sucrose gradients identical to those used for initial nuclear RNA fractionation, except that the centrifugation of p17 S RNA was for 10.25 h. As each pure RNA represented less than 1 +g of material, 10 to 15 /*g of carrier tRNA were added before the RNA was concentrated by ethanol precipitation.
To ensure that the RNA was free of sodium dodecyl sulfate and other salts, it ~a.6 precipitated from et,hanol at least three times. The final pellets were dried and resuspended in small volumes (20 to 75 ~11 of water. Base Composition Analysis -Limit digestions of RNA with a mixture of RNase A, RNase Tl, and RNase T2 were performed as described by Rose (18). Digests were spotted on sheets of Whatman No. 3MM paper and subjected to pH 3.5 ionophoresis at 40 V/cm for 50 to 60 min, until xylene cyan01 FF dye moved 11 to 12 cm from the origin (19). The four ribonucleoside monophosphates were detected by autoradiography, excised from the paper, and assayed for radioactivity in a toluene-based scintillation fluor. Fingerprinting ofRNaseA or Tl Digests ofRNAs -The homochromatography fingerprinting method used was only slightly modified from the procedure described by Barrel1 (19

Imethyl-:'H]and "ZP-labeled
RNAs, extracted from the nuclei of aggregation stage developing cells after a labeling period of 3 h, were passed over oligo(dT)-cellulose, and the flow-through fraction was analyzed by agarose gel electrophoresis (Fig.  1). The large molecular weight RNAs were resolved clearly on the gel into live discrete species. Two of these species co-migrate with mature, cytoplasmic 25 S or 17 S rRNA.
Of the remaining RNA peaks, two are labeled "~25 S" and "~17 S," in accordance with their presumed role as imme-I Cells were starved in suspension (see "Experimental Procedures"), and were labeled for 1 h with 6G mCi of azP,, either (A) after 9.5 h of starvation, or (B) immediately after the onset of starvation.
The RNA profiles obtained did not vary in any obvious way with the time at which label was introduced, but much more label was incorporated into RNA at the earliest stage of starvation. At the end of the 1 h labeling period, cells were harvested and were extracted for nuclear RNA. The RNA was loaded onto an oligo(dT)-cellulose column in high salt binding buffer, and material which flowed through the column under these conditions was collected. The RNA which had bound to the column was then recovered by elution with low salt elution buffer (see "Experimental Procedures").
Aliquots of each RNA fraction were analyzed by agarose cylindrical gel electrophoresis.   Fig. 3 demonstrates that pure, discrete precursor rRNAs can geneous, poly(A)-containing RNAs are retained on the colin fact be isolated from nuclear RNA preparations comparable umn. Fig. 2B indicates the composition of the material that binds to the column when salt concentration is high. There is to that shown in Fig. 2A. The purification procedure is detailed under "Experimental Procedures" and involves fractionconsiderable nonspecific sticking of the putative rRNA precursors to the oligo(dT)-cellulose; however, there does not appear ation of the individual RNA species on two sucrose gradients to be preferential binding of any one of these species relative to and an agarose gel. A to C of Fig. 3 represent characteristic another. Most of the bound radioactive material is heteroge-profiles of the final purification gradients for p17 S RNA, ~25 S RNA, and 37 S RNA. D provides additional evidence that 37 S neous, and presumably represents poly(A)-containing message-like RNAs (22). A large portion of the heterogeneous RNA is not simply an aggregate of smaller RNAs. A preparation of 37 S RNA was heated at 65" in 90% forrnamide before RNA migrates only slightly faster than ~17 S RNA. Roughly 30 to 40% of total nuclear RNA binds to oligo(dT)-cellulose being layered onto the final sucrose gradient. Under these denaturing conditions, the RNA did not dissociate into smaller under the conditions used here. Thus in unfractionated nuclear RNA, heterogeneous material would obscure the profile rRNA species (Fig. 30). Peak fractions from the gradients were pooled as indicated in the figure, and were concentrated of distinct species, particularly in the region of 17 S and ~17 S RNA. This problem has made it impossible to determine from by precipitation with ethanol. From the mobility of sample peaks relative to marker RNAs on such gradients, we can previous studies whether or not the proposed rRNA species estimate the size of each discrete nuclear RNA; thus, 37 S were intact, discrete species.
RNA actually has a sedimentation coefficient somewhere be-

FIG. 4. Autoradiograms
and tracings of two-dimensional separations of oligonucleotides produced by complete RNase A digestion of p25 S, 25 S, p17 S, and 17 S RNAs. First dimension separation, on cellulose acetate, pH 3.5, is based mainly on oligonucleotide charge; while second dimension homochromatography fractionates mainly on the basis of size, with smaller oligonucleotides moving toward the top of the plate (19). The tracings are composites of the fingerprints for a precursor rRNA and the corresponding mature rRNA. Open czrcles represent those oligonucleotides common to both; filled c&es, those unique to the precursor species; and cross-hatched wcles, those found exclusively, or in far greater quantity, in the mature rRNA. Oligonucleotides outlined in dots are faint, but reproducible in many fingerprints. Most of the faint oligonucleotides near the top of the fingerprints are not reproducible, and appear more frequently when the batch of digesting enzyme is old; they have therefore been omitted from the fingerprint tracings. The regions outlined with a dashed line represent characteristic oligonucleotide groupings that distinguish the p25 S-25 S RNA pair from the ~17 S-17 S RNA pair.
tween 36 S and 38 S, and the sedimentation coefficients for ~25 S and p17 S RNAs are approximately 28 S and 21 S, respectively.
Base Compositions -The nuclear genome of Dictyostelium discoideum has a low, 22 to 23% G + C content (23-251, as does the total mRNA (26). In marked contrast is the 42 to 44% G + C content of mature ribosomal RNA (26) (see also Table I). As would be predicted for ribosomal RNA precursors, 37 S RNA, ~25 S RNA, and p17 S RNA do not differ significantly in base composition from mature ribosomal RNAs (Table I) obtained from examination of oligonucleotide map "fingerprints" of the RNase A digestion products of ~25 S and p17 S RNAs, compared to equivalent fingerprints of the mature rRNAs (Fig. 4). Because RNase A recognizes both cytidine and uridine residues as cleavage sites, a large number of small oligonucleotides are generated by its action; such small fragments are expected to be distributed commonly among RNA molecules as large as rRNAs. Thus the upper portions of all the RNase A oligonucleotide maps are almost identical, even for such relatively unrelated species as 17 S RNA and 25 S RNA. However, in the lower regions, where large unique oligonucleotides are located, the fingerprints for 17 S RNA and 17s \\ \ 17s FIG. 6. (left). Autoradiograms of Southern gel hybridization patterns: annealing of discrete nuclear RNAs and mature rRNAs to fractionated EcoRl fragments of nuclear DNA. Dzctyostelrum nuclear DNA was digested withEcoR1, fractionated on an agarose gel, and transferred to a nitrocellulose filter by the Southern blot technique as described in the text. Parallel lanes of the filter, each bearing a complete fractionated DNA image, were incubated with 32P-labeled 5 S 17 S ~17 S, 25 S, p25 S, and 37 S RNAs under the hybridization c'ondit;ons described.

S RNA differ significantly;
some of the distinguishing oligonucleotide groupings are outlined in the fingerprint tracings. The fingerprints and tracings of Fig. 4 demonstrate clearly that all the unique oligonucleotides characteristic of 25 S RNA are also found in p25 S RNA. In fact, except for four extra large oligonucleotides in p25 S RNA, the fingerprints of ~2.5 S RNA and 25 S RNA appear identical. Essentially the same type of relationship is seen between the fingerprints of p17 S RNA and 17 S RNA, although 17 S RNA contains one distinctive oligonucleotide, probably representing one end of the intact RNA, which is not found in ~17 S RNA.
S, p25 S, and 37 S RNAs were allowed to hybridize with the DNA on parallel strips of the nitrocellulose filter, each derived from one slot of the same gel. As shown in Fig. 6, the p25 S RNA hybridizes specifically to those EcoRl DNA restriction fragments complementary to 25 S RNA; similarly, the pattern of p17 S RNA hybridization is identical with that for 17 S RNA. Purified 37 S RNA anneals with all three of the rRNAspecific restriction fragment bands. Thus, this large RNA species contains the proper sequence complement required for a common precursor to 25 S and 17 S rRNAs.
Tl RNase, with a greater degree of specificity than RNase A, cleaves next to guanosine residues only. Consequently, a Tl RNase digest contains many more large, distinctive oligonucleotides than does an RNase A digest. Fig. 5 illustrates the many sequences distinguishing 17 S RNA from 25 S RNA which are revealed in Tl oligonucleotide maps. The striking correlation of Tl oligonucleotide patterns in ~2.5 S RNA compared to 25 S RNA, and in p17 S RNA compared to 17 S RNA (fingerprints and tracings of Fig. 51, therefore provides strong substantiating evidence for the precursor-product relationships suggested by the RNase A fingerprints.
As is demonstrated by the tracings of Figs. 4 and 5, the RNase Tl fingerprints also expose more clearly than RNase A fingerprints those extra oligonucleotides which distinguish each precursor from its mature rRNA counterpart.
Hybridization with EcoRl Fragments Which Code for rRNA -When Dictyostelium nuclear DNA is digested with EcoRl restriction endonuclease and is fractionated by agarose gel electrophoresis, nine discrete fragment bands, representing reiterated DNA sequences, are resolved above a background of heterogeneous fragments. One of these DNA bands (actually a doublet) hybridizes specifically with 25 S rRNA; a second band hybridizes with 17 S rRNA; and a third band hybridizes with both rRNAs (20,27). We have used these rRNA-specific restriction fragments as probes for the detection of rRNA sequences in proposed precursor rRNAs. DISCUSSION We have characterized three discrete, rapidly labeled RNA species, all extensively methylated, from the nuclei of starved and developing Dictyostelium discoideum. These RNAs correspond in mobility and methylation properties to very transient radioactive species observed by Iwabuchi et al. (13) in sucrose gradient profiles of whole cell RNA from vegetative NC-4 Dictyostelium (NC-4 is parent to the axenic strain we have used). We assume that the comparable vegetative and developing cell RNA intermediates are the same. The one possible exception is a 30 S RNA intermediate, observed as a shoulder in Iwabuchi's gradients of vegetative RNA, which we do not detect in high resolution agarose gel profiles of nuclear RNA from starved or developing cells. Kinetics of RNA labeling in the vegetative cells suggested that the short lived intermediates were probably rRNA precursors. We have exploited the special properties of starved and developing cells to demonstrate further that each intermediate can be isolated as a discrete species, uncontaminated by heterogeneous material, incomplete chains, or aggregates of smaller rRNAs. It has not been possible to establish for Dictyostelium an effective pulsechase procedure for demonstrating clearly the flow of label from precursor to product RNAs (131. We have instead applied fingerprinting and hybridization techniques, in order to obtain unequivocal evidence for the derivation of mature rRNAs from the sequences of the proposed percursors. For the hybridization analysis, total nuclear DNA was di-The evidence presented in this paper suggests that the 25 S gested with EcoRl, and the digest was fractionated on an and 17 S rRNA genes in Dictyostelium are associated in a agarose slab gel as described under "Experimental Proce-large common transcription unit, which produces a 37 S RNA dures." The entire fractionated DNA pattern was then transferred from the gel to a nitrocellulose membrane by the proce-primary transcription product that includes sequences of both dure devised by Southern (21)