Proc. Natl. Acad. Sci. USA 75 (1978) 2539

Correction. In the article "Activation of Factor IX by the re-action product of tissue factor and Factor VII: Additionalpathway for initiating blood coagulation" by Bjarne 0sterudand Samuel L. Rapaport, which appeared in the December 1977issue of Proc. Natl. Acad. Sci. USA (74,5260-5264), the thirdsentence of the Abstract is not correct as it appeared. It shouldread as follows: Factor IX was not activated by thrombin, ac-tivated Factor X, or activated Factor IX itself.

Correction. In the article "Triphosphate residues at the 5' endsof rRNA precursor and 5S RNA from Dictyostelium discoi-deum" by Bambi Batts-Young and Harvey F. Lodish, whichappeared in the February 1978 issue of the Proc. Natl. Acad.Sci. USA (75, 740-744), the following undetected printer'serrors occurred. In the right-hand column of p. 741, line 4 ofthe Results should read ". . . nuclear DNA restriction fragmentcontaining the 5S RNA gene (see ... .)." In the right-hand col-umn of p. 743, the sentence starting 18 lines from the bottomof the page should read "This analysis would be even morecompelling if the 5' end of 37S RNA could be isolated andidentified as pppA-, the same as for pl7S RNA." Fig. 3 is re-printed below.

PP, [ e

PI

pGp[

pAp[

XC[

Origin,-a

PP,[ I

pGp[

pAp[

XC[

b Origin-b c d

FIG. 3. Ionophoretic separations of venom phosphodiesterasedigestion products of 5'-end residues 5-A and p17-A. Residues 5-Aand p17-A were eluted from the ionopherograms pictured in Fig. 2,and each was digested to completion with venom phosphodiesteraseand subjected to pH 3.5 ionophoresis on Whatman 3 MM paper. (band d) Radioactive PPi (and trace Pi) released in control digestionsof ['y-32P]ATP. (a and c) Fractionation patterns for digests of 5-A andp17-A, respectively. The positions of nonradioactive pAp and pGpmarkers are indicated by brackets. Any digestion product containingat least 25% of the radioactivity in the pAp fragment of p17-A shouldhave been detected as a discrete spot on these autoradiograms.

Corrections

Proc. Nati. Acad. Sci. USAVol. 75, No. 2, pp. 740-744, February 1978Biochemistry

Triphosphate residues at the 5' ends of rRNA precursor and 5S RNAfrom Dictyostelium discoideum

(polarity of rRNA transcription/transcription initiation sites/5S RNA synthesis/lower eukaryotes)

BAMBI BATTS-YOUNG* AND HARVEY F. LODISHDepartment of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Communicated by Joseph G. Gall, November 18, 1977

ABSTRACT We present here direct evidence for the pres-ervation of a transcriptional initiation sequence in a eukaryoticrRNA precursor: the 5'-end group for precursor to 17S rRNA(pl7SRNA) from Dictyostelium discoideum is identified as thetriphosphate residue pppA-. We also show that mature 5S RNAfrom Dictyostelium bears a different triphosphate residue,pppG-. In contrast, we find no evidence for more than onephosphate at the 5' end of the 25S rRNA precursor (p25S RNA).These observations indicate that synthesis of the large ribosomalRNAs of Dictyostelium begins with the 5'-terminal sequenceof the p17S RNA, and that 5S RNA transcription must be ini-tiated independently, despite the close association of the 5S andrRNA coding segments.

Although the sequence of events in rRNA production is similarin all organisms, certain characteristic variations in the pathwaydistinguish eukaryotes from prokaryotes. In bacteria, the genesencoding the 23S, 16S, and 5S RNAs are transcribed consecu-tively from a single promoter in the order 16S-23S-5S (1-3).Similarly, the larger rRNAs of eukaryotes are synthesized to-gether in a common transcription unit, with the 18S RNA se-quences proximal to the initiation site sequences (4-8). How-ever, in most eukaryotes, the 5S RNA is independently tran-scribed from DNA completely unlinked to the other rRNAgenes (9-13).The cellular slime mold Dictyostelium discoideum is a

primitive eukaryote. Unlike higher eukaryotes, it maintains alinkage between the genes that specify 5S RNA (5S DNA) andthe genes for the larger rRNAs (rDNA) (14). We have askedwhether this linkage results in cotranscription of the 5S DNAand rDNA. Previous studies of rRNA synthesis in Dictyosteliumhave shown that the initial detectable transcription product isa 37S RNA that has the proper sequence composition for acommon precursor to mature 25S and 17S rRNAs (15, 16). Thislarge species is apparently cleaved to yield two intermediate-sized precursors, designated p25S and pl7S RNA; each of thesethen gives rise to its mature rRNA counterpart by the loss of asmall number of nucleotides (16). In order to obtain more in-formation about the arrangement of sequences in the 37Scommon rRNA transcript, and particularly to probe for thepresence of 5S RNA within this species, we have analyzed the5'-terminal structures of 5S RNA and the rRNA precursors.Through these analyses we have demonstrated that transcrip-tion from the rDNA of Dictyostelium initiates with the tri-phosphate residue pppA-, at the 5' end of the pl7S RNA se-quences, while 5S DNA transcription is initiated separately,with the residue pppG-.

MATERIALS AND METHODSMaterials. Carrier-free 32P1 and ['y-32P]ATP (used at 1-5

Ci/mol) were purchased from New England Nuclear. Carrier

tRNA from Escherichia coli K-12 was purchased fromSchwarz/Mann and further purified as described (16). RNaseA was obtained from Worthington Biochemicals, RNase Ti andRNase T2 were from Calbiochem, and bacterial alkalinephosphatase and snake venom phosphodiesterase were fromBoehringer-Mannheim. Markers of the nucleoside 5'-mono-phosphates were from Sigma; P-L Biochemicals supplied pAp;and pGp was a gift from William Haseltine of the SidneyFarber Cancer Center, Boston, MA.

Preparation of rRNA Precursors and 5S RNA. We havepreviously described conditions for labeling Dictyosteliumamoebae with 32P1 (60 mCi/2 X 108 cells) in nutrient-free liquidsuspension (16). Nuclear RNA was prepared from cells labeledfor 1 hr under these conditions; 37S, p25S, and p17S nuclearRNAs were purified as described (16). Cytoplasmic RNA wasextracted from similar cells, but after a labeling period of 10hr; the preparation of cytoplasmic polysomal rRNA and itsfractionation on sucrose gradients into 25S, 17S, and 4-5S RNAcomponents have also been described (16). The 4S and 5S RNAs,recovered together from the top fractions of the gradients byprecipitation with 2 volumes of ethanol, were further frac-tionated on a 2 X 80-cm Sephadex G-100 column equilibratedat 40 in 0.1 M NaCI/20 mM sodium acetate, pH 4.6. Materialeluted from the column in the same buffer was resolved clearlyinto distinct 4S and 5S RNA peaks. The 5S RNA was concen-trated by ethanol precipitation, dissolved in 0.5 ml of water, andpassed over a 5-ml Sephadex G-25 M column in water. RNAeluted from the column in the salt-free void volume fractionswas brought to dryness in a Iyophilizer and then was resus-pended in 50 ytl of a solution of 50% (vol/vol) deionized form-amide (MC/B), 1 mM EDTA (pH 7.6), and sufficient brom-phenol blue to act as tracking dye. The RNA suspension waslayered onto a 0.7 X 15-cm cylindrical 10% (wt/vol) polyac-rylamide gel (17) polymerized in the presence of 0.2% (wt/vol)sodium dodecyl sulfate; electrophoresis was at 7.75 mA per gelfor 5.5 hr. Afterward, the gel was sliced transversely into 1-mmslices, and the Cerenkov radiation of each slice was measured(18). Peak slices were pooled, covered with 4 ml of elutionbuffer [0.5 M sodium acetate/10mM Tris/1% (wt/vol) sodiumdodecyl sulfate/1 mM EDTA at pH 7], and shaken at 370overnight. At least 95% of the radioactive label originally in thegel slices was released into the elution buffer. The gel slices wereremoved, and RNA was precipitated from the elution bufferwith ethanol.To eliminate residual gel contaminants from it, the 5S RNA

was bound to and eluted from a 0.5-ml DEAE-cellulose(Whatman DE-52) column equilibrated in 0.1 M NaCl/20mM

Abbreviations: pl7S RNA, the immediate precursor to 17S rRNA; p25SRNA, the immediate precursor to 25S rRNA; 5S DNA, DNA thatcontains the 5S RNA gene; rDNA, DNA that contains the genes for thelarge molecular weight rRNAs.* Present address: Department of Biology, Yale University, NewHaven, CT 06520.

740

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked"advertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

Proc. Natl. Acad. Sci. USA 75(1978) 741

s0 100 120

FIG. 1. Profiles of purified p25S, p17S, and 5S RNAs. Details of theRNA fractionation procedures are given in Materials and Methodsand ref. 16. Final purification of p25S (A) and p17S (B) RNAs bysedimentation through 15-30% sucrose gradients. (C) A late step in5S RNA purification, involving electrophoresis on a 10% polyacryl-amide gel. Bromphenol blue dye migrated on the gel to the positionmarked by BPB. Gel and gradient fractions were assayed for Cerenkovradiation (18), and were pooled as indicated by the bars in the fig-ure.

sodium acetate at pH 4.6. The RNA was loaded onto the columnin several milliliters of the same buffer; the column was thenwashed with 10 volumes of this buffer, and RNA was finallyeluted with a high-salt buffer (2 M NaCl/20 mM sodium ace-

tate, pH 4.6). Most of the bound 5S RNA was released from thefirst 1 ml of the high-salt eluate. Sufficient carrier tRNA was

mixed with the eluted sample to yield a total of a least 5 Mg ofRNA. The RNA was precipitated from ethanol three times torid the sample of salts; the final pellet was dried, and was re-

suspended in 25 M1 of water.The radiochemical purity of 5S RNA prepared by these

methods is illustrated in Fig. 1C. For the rRNA precursors, thepenultimate purification step involves high-resolution agarose

gel electrophoresis (16); analysis of typical gel profiles indicatesthat no more than 5% of the radioactivity in isolated p25S or

p17S RNA pools can be accounted for by trailing of materialfrom adjacent RNA species on the gel. The final fractionationof p25S and p17S RNAs yields sedimentation profiles like thoseshown in Fig. 1 A and B.

5'-End-Group Purification and Analysis. Purified RNAswere digested to completion with a mixture of RNases (T1 +T2 + A), the digests were subjected to ionophoresis at pH 3.5on DEAE-paper, and digestion products were located and re-

covered as described (19).Isolated 5'-end residues were digested with alkaline phos-

phatase under conditions described (19), except that the pH was

raised to 8.6. Complete venom phosphodiesterase digestion of

a 5'-end sample was obtained by incubating the sample for 2hr at 370 in 5 /d of a solution containing 250,gg of the enzymeper ml, 5 ,gg of pAp (to inhibit phosphatases), 10mM Tris (pH7.5), and 12.5% glycerol. As a control to test for proper func-tioning of either enzyme, [y-32P]ATP was digested at the sametime and under the same conditions used for experimentalsamples. Digests were spotted on 15 X 60-cm sheets of What-man 3 MM paper; a reserved portion of undigested 5'-endresidue was spotted to the side of the phosphatase digests, while100 ,g of nonradioactive pAp and/or pGp marker was appliedabove each phosphodiesterase digest. Ionophoresis was at pH3.5 and 35 V/cm for 1-1.5 hr, until xylene cyanole F.F. dyemoved 9-12 cm from the origin. Radioactive components weredetected by autoradiography; marker nucleotides were madevisible by illumination with a short-wave ultraviolet lamp.

Radioactive residues migrating with the nucleoside diphos-phates were eluted from the paper by repeated washing withsmall volumes of water. Each sample was mixed with 10-20,Mgof each of several nonradioactive marker nucleotides and themixture was applied in 3 Ml of water to a 20 X 20-cm glass-backed cellulose thin-layer plate (EM Laboratories). Chro-matograms were developed in the first dimension with iso-butyric acid/NH3/water (577:38:385 vol/vol) and in the second'dimension with t-butanol/HCI/water (70:15:15 vol/vol)(20).

RESULTSThe first indication that Dictyostelium 5S DNA and rDNA aretranscribed separately was provided by an annealing experi-ment in which 37S RNA failed to hybridize detectably to anuclear DNA restriction fragment containing the 5S RNA (seeref. 16 for this experiment and a discussion of its limitations).We have obtained more rigorous evidence about the arrange-ment of rRNA transcription units through a comparison of the5' ends of 5S RNA and the pl7S and p25S rRNA precursors.

Isolation of 5'-End Residues. In the search for potential5'-end structures, we first digested each RNA to completionwith RNases T1, T2, and A. This combination of enzymes willdigest most RNA-sequences entirely to 3'-mononucleotides, butwill not cleave a phosphodiester linkage at the 3' side of a 2'-O-methylated nucleotide nor remove the 5'-phosphates froma 5'-terminal residue. Since the RNase-resistant 5'-end struc-tures and ribose-methylated oligonucleotides all contain morethan one phosphate, they are readily separated from the 3'-mononucleotides by ionophoresis at pH 3.5 on DEAE-paper(19). Ionophoretically fractionated digests of p25S, p17S, and5S RNAs are shown in Fig. 2. As expected, in each case themajor digestion products are the 3'-mononucleotides. However,there are also several radioactive species that migrate moreslowly. In lanes b and c, representing digests of p17S and p25SRNAs, respectively, there can be seen a series of faint spotsstretching from just below the mononucleotides to a positionconsiderably lower than the dye marker. These spots are mostlikely the radioactive images of the 2'-O-methylated di- andtrinucleotides. In addition, a single spot, marked p17-A, is visiblejust above the origin of the fractionation pattern of the p17SRNA digest. This material is conspicuously absent from thecorresponding pattern of separated p25S RNA digestionproducts. However, there is a comparable radioactive compo-nent, labeled 5-A, among the digestion products of 5S RNA(Fig. 2, lane a). Moving somewhat faster than 5-A is a faintstreak of radioactive material, labeled B and C. The 5S RNAdigest does not contain any radioactivity in material that movesbetween the dye marker and mononucleotides, in contrast tothe situation for p17S and p25S RNAs. As we will demonstrate,SS RNA contains no 2'-0-methylated nucleotides, so that the

A 28S 25S

3 A

2

8 B25S 17S

6 -

0

x

-A

x

E0.0

IT0

xE0.

0

x

E0.

25 30 35

C 5.8S 5S 4S BPB3 Il

2

1 [I

5 10 15 20

0 20 40 60 fFraction

Biochemistry: Batts-Young and Lodish

742 Biochemistry: Batts-Young and Lodish

*-up--99

cp

Ap-_

xcI--

C_B-

5-A *- p17-A-

a b c

FIG. 2. Segregation of 5'-end residues by ionophoresis of theRNase (T1 + T2 + A) limit digests of 5S and ribosomal precursor

RNAs. Digested samples were spotted on sheets of DEAE-paper ei-

ther (a) 60 cm or (b and c) 100 cm long, and were subjected to pH 3.5ionophoresis at 30 V/cm for (a) 2 hr or (b and c) 6 hr. Fractionationpatterns for limit digests of 5S RNA (a), p17S RNA (b), and p25SRNA (c), respectively. XC indicates the position of the blue dye, xy-

lene cyanole F.F. The discontinuity below the mononucleotide regionin lanes b and c reflects the fact that the preparation of autoradio-grams required the separation of each 100-cm-long ionopherograminto two sections.

only components other than 3'-mononucleotides that could bereleased by complete digestion with RNases TI, T2, and A are

5'-end structures.The mobilities of digestion products p17-A and 5-A indicate

that these residues contain several phosphates, making themprime candidates for 5'-end structures. Therefore, both of theseresidues, as well as B and C from the 5S RNA digest, were elutedfrom the paper for further structural analysis.

Determination of Structures of Potential 5'-End Residues.Alkaline phosphatase digestion of p17-A and 5-A released allradioactivity from each as free 32P1, indicating that neitherresidue can contain internucleotide phosphates. Therefore, eachmust be a mononucleotide, derived from the 5' end of the RNAand bearing several external phosphates.To determine the nature of these nucleotides and the number

of their extra phosphates, we digested 5-A and p17-A each withvenom phosphodiesterase and analyzed the products by paper

ionophoresis at pH 3.5 (Fig. 3). Venom phosphodiesterase is a3'-exonuclease that will also cleave between the a and , 5'-phosphates of a mononucleotide with more than one 5'-phos-phate (21). Residue 5-A was cleaved by the enzyme into twocomponents, of roughly equal radioactivity that migrated withPPi and pGp markers (Fig. 3, lane a). This distribution of di-gestion products provides strong evidence that the structure of5-A is pppGp. Similarly, venom phosphodiesterase digestionof p17-A yielded radioactive products that migrated with PPiand pAp (Fig. 3, lane c), indicating a structure of pppAp forp17-A.The ionophoretic system used for fractionating the venom

phosphodiesterase digestion products does not separate un-

modified nucleotides from most of their methylated derivatives.In order to confirm unequivocally the proposed structures forthe 5' ends of 5S and pl7s RNAs, we eluted the radioactivematerial that migrated with pGp (from 5-A) and pAp (fromp17-A) during paper ionophoresis, and fractionated each in two

Origint- Origin!-a b c d

FIG. 3. Ionophoretic separations of venom phosphodiesterasedigestion products of 5'-end residues 5-A and p17-A. Residues 5-Aand p17-A were eluted from the ionopherograms pictured in Fig. 2,and each was digested to completion with venom phosphodiesteraseand subjected to pH 3.5 ionophoresis on Whatman 3 MM paper. (band d) Radioactive PPi (and trace Pi) released in control digestionsof [y-32P]ATP. (a and c) Fractionation patterns for digests of 5-A andp17-A, respectively. The positions of nonradioactive pAp and pGpmarkers are indicated by brackets. Any digestion product containingat least 25% of the radioactivity in the pAp fragment of p17-A shouldhave been detected as a discrete spot on these autoradiograms.

additional dimensions by a thin-layer chromatographic systemdesigned to resolve modified from unmodified nucleotides (22).Fig. 4 demonstrates that all material that migrated as pGpduring ionophoresis remained coincident with pGp markerduring thin-layer chromatography (Fig. 4a); similarly, thenucleotide from the p17-A digest continued to migrate exactlywith pAp marker in the chromatographic system (Fig. 4b).

Estimate of Percent of Molecules of 5S and p17S RNAsBearing 5'-Triphosphate Termini. We assume that Dictyos-telium 5S RNA, like all other 5S RNAs (23), is some 120 nu-

cleotides long; then a nucleoside tetraphosphate derived fromits 5' end should contain 3.4% of the total phosphates in themolecule. Roughly 2.0-2.4% of the 5S RNA 32P radioactivitywas actually recovered in component 5-A, indicating that about60-70% of the 5S RNA molecules carry the 5'-terminal tri-phosphate group pppG-. The other 5'-terminal structures for5S RNA are found in residues B and C (see Fig. 2); these havebeen characterized by the same methodology as was used for5-A, and their structures have been established as ppGp andpGp, respectively (see Table 1). The combined radioactivityin residues pppGp, ppGp, and pGp accounts for all of the 5'ends of 5S RNA; it also represents all of the radioactivity thatis not in 3'-mononucleotides (see Fig. 2, lane a).

In the absence of an accurate estimate for the length of p17SRNA, calculations of the yield of triphosphate end group fromthis species are necessarily approximate. However, based on itsmobility on gradients relative to other rRNAs of known mo-

lecular weight (15, 16, 24), pl7S RNA can probably be re-

stricted to a range of lengths between 2250 and 3000 nucleo-tides. The fraction of total 32P radioactivity found in p17-A(0.15%) then corresponds to 0.85-1.1 copies of pppAp per pl7SRNA molecule, provided that the RNA is uniformly labeled.In fact, we know that the a-phosphate of ATP must haveroughly the same specific activity as the a-phosphates of theother nucleotides used in synthesis of the rRNAs because, whenthe rRNA precursors studied here are digested to 5'-mononu-cleotides, the fraction of total 32p label recovered in pA (0.28;unpublished data) is virtually identical to the mole fraction ofadenosine in these molecules (16). Our calculations do requireadjustment for the excess of radioactivity in and y phosphates

IP[PPr 0L

PP[ 9

P.

pGp[r

pApF

XC,

pGp[

pAp[

XC[

Proc. Natl. Acad. Sci. USA 75 (1978)

Proc. Natl. Acad. Sci. USA 75 (1978) 743

:.

i!

pA pC

pGpGpW

pA pAp

pU ; pG . pLpGp

-11

0O2 b

FIG. 4. Two-dimensional cellulose thin-layer chromatographyof venom phosphodiesterase products of 5'-end residues 5-A andp17-A. Material migrating with pGp marker was eluted from theionopherogram pictured in Fig. 3, lane a; similarly, material migratingwith pAp was eluted from the ionopherogram of lane c, Fig. 3. Eachsample was dried in a lyophilizer, mixed with nonradioactive markernucleotides, and fractionated by thin-layer chromatography. Thepositions of marker nucleotides are indicated by broken circles. (a)pGp digestion product of residue 5-A; (b) pAp digestion product ofresidue p17-A.

of p17-A, compared to a phosphates (compare pAp and PPi,Table 1); however, the corrected yield of pppAp residues is stillsufficient to account for 60-80% of the 5' ends of p17S RNA.The other 20-40% of the 5' termini may well be variants withfewer phosphates, in analogy to the case for 5S RNA. However,a small amount of ppAp would have left too faint an impressionon the autoradiogram of Fig. 2 to be detected, and pAp wouldhave been hidden among the pl7S RNA oligonucleotides re-sistant to cleavage because of ribose methylation.

DISCUSSIONThe data we have presented here on the properties of Dicty-ostelium rRNA and rRNA precursors, combined with ourprevious observations (16), provide strong support for threemajor conclusions. One is that the genes for the 17S and 25SrRNAs are part of a common transcription unit which initiatesRNA synthesis at the 5' end of the pl7S RNA sequences. Sec-ond, the primary transcription product of this unit is almostcertainly the 37S RNA, which is evidently unprocessed at leastat the 5' end. Finally, it is apparent that 5S RNA must be

Table 1. Distribution of radioactivity in venom phosphodiesterasedigestion products of isolated 5'-end residues

from 5S and p17S RNAs

% of total radioactivity in each digestion product5' end pAp pGp Pi PP,

5-A 44 56B -67* -33*C 95

p17-A 35 65

Each 5'-end structure was digested and fractionated as illustratedin Fig. 3. Radioactive components, detected by autoradiography, wereexcised from the paper and assayed for Cerenkov radiation (18). Thedistribution of radioactivity did not change when assayed in a tolu-ene-based scintillation fluor. All sample readings were corrected forthe radiation in blank strips from the autoradiogram.* Determined by inspection only.

transcribed independently of the other rRNAs, despite thelinkage of the genes encoding 37S and 5S RNA.

Central to these conclusions is the demonstration that thenucleoside 5'-triphosphates pppA- and pppG- represent themajority of the 5' ends for Dictyostelium p17S and 5S RNA,respectively. The remaining, alternate 5' ends of 5S RNA arepartially dephosphorylated derivatives of the triphosphate, andit is likely that the same is true for p17S RNA. it is possible thatthe loss of phosphates from some of the 5' ends is caused byprocessing in vivo. However, the large proportion of both 5Sand p17S RNA molecules bearing the full complement of 5'-phosphates (compare to refs. 21 and 25-28) suggests that thetriphosphate residue may represent the only 5'-end structurefor these species in vivo. In fact, small losses of external phos-phates might well be expected during the many preliminary5'-end purification steps, especially during the initial extractionof RNA from cells and during the digestion of purified RNAswith the mixture of RNases T1, T2, and A.

In contrast to p17S RNA, p25S RNA contains no detectableradioactive component that migrates as pppNp or ppNp (seeFig. 2). Thus, the 5'-end group for this large molecular weightrRNA is probably either pNp, which would be hidden amongthe ribose-methylated oligonucleotides, or simply Np, whichwould be inseparable from all the internal mononucleotides.

Both prokaryotic and eukaryotic RNA polymerases constructRNA chains in vitro such that the first nucleotide incorporatedretains its 5'-triphosphate moiety; usually the initiating nucle-oside triphosphate is either GTP or ATP (29-S1). Thus, theend-group analyses we have reported suggest that the 5' endof pU7S RNA represents a site for the initiation of rRNA syn-thesis, while the 5' end of p25S RNA is probably derived bypost-transcriptional cleavage from a larger precursor. Since thegenes for 17S RNA and 25S RNA are very closely linked in therDNA of Dictyostelium (14, 24), it is now possible to proposea scheme for rRNA synthesis in which transcription begins withthe 5'-terminal portion of p17S RNA and continues withoutinterruption to the end of the p25S RNA sequences and possiblysomewhat beyond. The initial product of rRNA transcription,then, would be a large molecule containing sequences of bothpU7S and p25S RNAs. Previous work from this laboratory hasshown that Dictyostelium does in fact manufacture such amolecule, the 37S RNA (16). The primary rRNA gene producttherefore is presumably identical to 37S RNA and is organizedsuch that the triphosphate residue of p17S RNA is coincidentwith its 5' end and the p25S RNA sequences are at or near its3' end. If there exists an extremely transient rRNA precursorlarger than 37S RNA, it could differ from this general structuralpattern only by having extra sequences at its 3' end.

This analysis would be even more compelling if it could beisolated and identified as the 5' end of 37S RNA pppA-, thesame as for pU7S RNA. This has not been possible so far: theyield of pure radioactive 37S RNA from cells labeled with asmuch as 60 mCi of 32Pi has never been greater than about 'Aoothe amount necessary for even a perfunctory 5'-end analysis.

However, our 5'-end group data do provide clear and directevidence for the preservation of a transcriptional initiationsequence in a eukaryotic rRNA precursor. Of all the otherrRNA precursors studied, only the 30S RNA that accumulatesin RNase III mutants of E. coli exposed to chloramphenicol isknown to bear a triphosphate 5'-end group, pppA- (2, 3). At-tempts have been made to isolate polyphosphate 5'-terminifrom rRNA precursors in Xenopus and Novikoff hepatoma cells(32, 33), but have not been successful. Thus our results may wellrepresent the best evidence obtained so far for a natural rRNAprecursor that is a true primary transcription product.

There is an additional observation consistent with our con-

Biochemistry: Batts-Young and Lodish

744 Biochemistry: Batts-Young and Lodish

clusion about the ordering of species within the rRNA tran-scription unit: indirect evidence, based on hybridization analysisof the sequences that remain in a central rDNA restrictionfragment after partial digestion with a 5'-exonuclease, hassimilarly suggested that 17S RNA sequences are located to the5' side of 25S RNA sequences (24). Thus, the polarity of therRNA transcription unit in Dictyostelium is similar to thatdeduced for bacteria (1), as well as for yeast (8) and for highereukaryotes (4-7). However, in higher animals (4), and probablyalso in E. coli (1), some material is clipped from the 5' end ofthe initial rRNA transcript very early in the maturation process,before the sequences for the two large rRNAs are separatedfrom each other, while in Dictyostelium we have shown thatthe initiation segment is preserved for a considerably longerperiod, so that it remains associated with the 5' end of the p17SRNA until mature 17S RNA is finally excised from this pre-cursor (24, 34).

Another distinctive feature of the rRNA synthetic sequenceof Dictyostelium is also revealed by the 5'-end group studies.We have demonstrated that the 5'-terminal residue for 5S RNAfrom this organism is pppG-. Retention of the triphosphategroup indicates that if 5S RNA is derived from a larger pre-cursor, it must be excised from the extreme 5' end of that pre-cursor. Since it appears to be the 5'-terminal pppA- of p17SRNA that lies at the 5' end of the primary rDNA transcriptionproduct, we conclude that the 5S RNA cannot be synthesizedas part of a precursor common to the other rRNAs.The 5'-terminal triphosphate residue of the 5S RNA, as well

as its independent mode of synthesis, appear to be characteristicproperties that distinguish all eukaryotic from prokaryotic 5SRNAs. Thus, bacterial 5S RNAs generally bear a 5'-terminalpU- (or, rarely, pC-) (35), consistent with the processing of theseRNAs from larger precursors. In contrast, the 5'-end groups ofseveral eukaryotic 5S RNAs have been shown to be triphosphateresidues, almost always pppG- (26-28, 36), although at least onealgal 5S RNA is terminated with pppA- (37). In animals (9-12),and also in the protozoan Tetrahymena (13), independenttranscription of 5S RNA is insured by the complete lack oflinkage between 5S DNA and rDNA. However, for the lowereukaryote yeast, as for Dictyostelium, the 5S RNA genes remaininterspersed with the larger rRNA genes (38). In yeast, analysesof the kinetics of 5S RNA production and examination of thecoding strand of 5S DNA have provided evidence that the 5SRNA cannot be cotranscribed with the other rRNAs, despitethe close association of genes (8, 39; R. Kramer, personal com-munication). The results reported in this paper indicate thatthe same is true for Dictyostelium. In this respect, then, therRNA synthetic apparatus for these two primitive eukaryotesis intermediate in form between the prokaryote and animalsystems.

We are grateful to Alan Weiner for the original suggestion that itmight be interesting to examine the 5' ends of the rRNA precursors;to Jack Rose, for his invaluable advice on 5'-end determination pro-cedures; to Nancy Maizels for her patience and insight in reading thismanuscript and in proposing ways to improve it; and to Naomi Cohen,for her cheerfulness and efficiency in providing a ready supply ofhealthy Dictyostelium. This research was supported by GrantPCM74-04869-A02 from the National Science Foundation.

1. Pace, N. R. (1973) Bacteriol. Rev. 37, 562-603.

2. Ginsburg, D. & Steitz, J. A. (1975) J. Biol. Chem. 250, 5647:-5654.

3. Hayes, F., Vasseur, M., Nikolaev, N., Schlessinger, D., SriWidada,J., Krol, A. & Branlant, C. (1975) FEBS Lett. 56,85-91.

4. Perry, R. P. (1976) Annu. Rev. Biochem. 45,605-629.5. Dawid, I. B. & Wellauer, P. K. (1976) Cell 8,443-448.6. Reeder, R. H., Higashinakagawa, T. & Miller, O., Jr. (1976) Cell

8,449-454.7. Schibler, U., Hagenbuchle, O., Wyler, T., Weber, R., Boseley,

P., Telford, J. & Birnstiel, M. (1976) Eur. J. Biochem. 68,471-480.

8. Trapman, J. & Planta, R. J. (1975) Biochim. Biophys. Acta 414,115-125.

9. Brown, D. D. & Weber, C. S. (1968) J. Mol. Biol. 34,661-680.10. Pardue, M. L., Brown, D. D. & Birnstiel, M. L. (1973) Chromo-

soma 42, 191-203.11. Wimber, D. E. & Steffensen, D. M. (1971) Science 170, 639-

641.12. Aloni, Y., Hatlen, L. E. & Attardi, G. (1971) J. Mol. Biol. 56,

555-563.13. Tonnesen, T., Engberg, J. & Leick, V. (1976) Eur. J. Biochem.

63,399-407.14. Maizels, N. (1976) Cell 9,431-438.15. Iwabuchi, M., Mizukami, Y. & Sameshima, M. (1971) Biochim.

Biophys. Acta 228, 693-700.16. Batts-Young, B., Maizels, N. & Lodish, H. F. (1977) J. Biol. Chem.

252,3952-3960.17. Bishop, D. H. L., Claybrook, J. R. & Spiegelman, S. (1967) J. Mol.

Biol. 26, 373-387.18. Clausen, T. (1968) Anal. Biochem. 22,70-73.19. Rose, J. K. (1975) J. Biol. Chem. 250, 8098-8104.20. Walker, R. T. & RajBhandary, U. L. (1975) Nucleic Acids Res.

2,61-78.21. Roblin, R. (1968) J. Mol. Biol. 31,51-61.22. Nishimura, S. (1972) Prog. Nucleic Acid Res. Mol. Biol. 12,

49-85.23. Erdmann, V. A. (1976) Progr. Nucleic Acid Res. Mol. Biol. 18,

45-90.24. Frankel, G., Cockburn, A. F., Kindle, K. L. & Firtel, R. A. (1977)

J. Mol. Biol. 109,539-558.25. Dahlberg, J. (1968) Nature 220, 548-552.26. Hatlen, L. E., Amaldi, F. & Attardi, G. (1969) Biochemistry 8,

4989-5005.27. Hindley, J. & Page, S. M. (1972) FEBS Lett. 26, 157-160.28. Soave, C., Nucca, R., Sala, E., Viotti, A. & Galante, E. (1973) Eur.

J. Biochem. 32,392-400.29. Bremer, H., Konrad, M. W., Gaines, K. & Stent, G. S. (1965) J.

Mol. Biol. 13,540-553.30. Maitra, U., Cohen, S. N. & Hurwitz, J. (1966) Cold Spring Harbor

Symp. Quant. Biol. 31, 113-122.31. Chambon, P., Gissinger, F., Mandel, J. L., Jr., Kedinger, C.,

Gniazdowski, M. & Meihlac, M. (1970) Cold Spring HarborSymp. Quant. Biol. 35,693-707.

32. Slack, J. M. W. & Loening, U. (1974) Eur. J. Biochem. 43,59-67.

33. Choi, Y. C. & Busch, H. (1970) J. Biol. Chem. 245, 1954-1961.

34. Sakuma, K., Kominami, R. & Muramatsu, M. (1976) Eur. J.Biochem. 63, 339-350.

35. Fox, G. E. & Woese, C. R. (1975) Nature 256,505-507.36. Wegnez, M., Monier, R. & Denis, H. (1972) FEBS Lett. 25,

13-20.37. Jordan, B. R., Galling, G. & Jourdan, R. (1974) J. Mol. Biol. 87,

205-225.38. Rubin, G. M. & Sulston, J. E. (1973) J. Mol. Biol. 79,521-530.39. Aarstad, K. & 0yen, T. B. (1975) FEBS Lett. 51, 227-231.

Proc. Natl. Acad. Sci. USA 75 (1978)