Functional importance of conserved nucleotides at the histone RNA 3' processing site

1998 4: 246-256RNA A Furger, A Schaller and D Schümperli 3' processing site.Functional importance of conserved nucleotides at the histone RNA

References http://rnajournal.cshlp.org/content/4/3/246#related-urls

Article cited in:

serviceEmail alerting

click heretop right corner of the article orReceive free email alerts when new articles cite this article - sign up in the box at the

http://rnajournal.cshlp.org/subscriptions go to: RNATo subscribe to

© 1998 RNA Society

Cold Spring Harbor Laboratory Press on July 9, 2011 - Published by rnajournal.cshlp.orgDownloaded from

http://rnajournal.cshlp.org/content/4/3/246#related-urls

http://rnajournal.cshlp.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=rna;4/3/246&return_type=article&return_url=http://rnajournal.cshlp.org/content/4/3/246.full.pdf

http://rnajournal.cshlp.org/subscriptions

http://rnajournal.cshlp.org/

http://www.cshlpress.com

Functional importance of conserved nucleotidesat the histone RNA 39 processing site

ANDRÉ FURGER,1 ANDRÉ SCHALLER, and DANIEL SCHÜMPERLIAbteilung für Entwicklungsbiologie, Zoologisches Institut der Universität Bern, Baltzerstrasse 4, 3012 Bern, Switzerland

ABSTRACT

Histone pre-mRNA 39 processing is controlled by a hairpin element preceding the processing site that interacts witha hairpin-binding protein (HBP) and a downstream spacer element that serves as anchoring site for the U7 snRNP. Inaddition, the nucleotides following the hairpin and surrounding the processing site (ACCCA 9CA) are conservedamong vertebrate histone genes. Single to triple nucleotide mutations of this sequence were tested for their ability tobe processed in nuclear extract from animal cells. Changing the first four nucleotides had no qualitative and little ifany quantitative effects on histone RNA 3 9 processing in mouse K21 cell extract, where processing of this gene isvirtually independent of the HBP. A gel mobility shift assay revealing HBP interactions and a processing assay in HeLacell extract (where the contribution of HBP to efficient processing is more important) showed that only one of thesemutations, predicted to extend the hairpin by one base pair, affected the interaction with HBP. Mutations in the nextthree nucleotides affected both the cleavage efficiency and the choice of processing sites. Analysis of these novelsites indicated a preference for the nucleotide 5 9 of the cleavage site in the order A . C . U . G. Moreover, a guanosinein the 3 9 position inhibited cleavage. The preference for an A is shared with the cleavage/polyadenylation reaction, butthe preference order for the other nucleotides is different [Chen F, MacDonald CC, Wilusz J, 1995, Nucleic Acids Res23:2614–2620].

Keywords: adenosine; cleavage site; 3 9 end formation; mutational analysis; U7 snRNP

INTRODUCTION

Metazoans generate their mRNAs from longer precur-sor mRNAs (pre-mRNAs) by two distinct types of cleav-age reactions+ The bulk of mRNAs are polyadenylatedand, here, the cleavage is tightly linked to poly(A) ad-dition, although the two steps can be dissociated in vitro(reviewed in Wahle & Keller, 1992)+ In contrast, the fam-ily of replication-dependent histone mRNAs are gener-ated by a single cleavage event (reviewed in Birnstiel &Schaufele, 1988)+ Notwithstanding many dissimilaritiesin cis-acting sequence elements and trans-acting fac-tors, both types of reactions are independent of cleav-able ATP and divalent cations (Moore & Sharp, 1985;Gick et al+, 1986;Wittop Koning, 1993) and the resultingproducts contain 39 hydroxyl and 59 phosphate groups(Moore & Sharp, 1985; Streit et al+, 1993)+ The lack of arequirement for divalent cations suggests that both re-actions may be protein catalyzed (Steitz & Steitz, 1993),

despite the fact that the U7 snRNP participates in his-tone RNA 39 processing+ However, in neither of thesesystems has the active nuclease been identified+

Two conserved sequence elements on the pre-mRNAsubstrate are important for the formation of mature his-tone mRNA 39 ends (Birnstiel & Schaufele, 1988)+ (1) Ahairpin element upstream of the cleavage site is boundby a cognate hairpin binding factor (Mowry et al+, 1989;Vasserot et al+, 1989)+ To reflect the fact that this trans-acting component has been cloned recently and henceits composition is known (Wang et al+, 1996; Martinet al+, 1997), we have renamed it hairpin binding pro-tein (HBP; Martin et al+, 1997)+ (2) A purine-rich spacerelement located several nucleotides downstream of thecleavage site interacts by base pairing with the 59 endof U7 RNA present in the U7 snRNP (Schaufele et al+,1986; Bond et al+, 1991)+

In addition to these two elements, the nucleotidesimmediately following the hairpin and encompassingthe processing site show considerable sequence con-servation (Fig+ 1)+ In vertebrate histone genes, the firstsix nucleotides following the hairpin fit the consensusACCCAC+ The seventh nucleotide is more variable, butshows a preference for A in mammals and U in non-

Reprint requests to:Daniel Schümperli,Abteilung für Entwicklungs-biologie, Zoologisches Institut der Universität Bern, Baltzerstrasse 4,3012 Bern, Switzerland; e-mail: daniel+schuemperli@zoi+unibe+ch+

1Present address: Institut für Allgemeine Mikrobiologie, Baltzer-strasse 4, 3012 Bern, Switzerland+

RNA (1998), 4:246–256+ Cambridge University Press+ Printed in the USA+Copyright © 1998 RNA Society+

246




mammalian vertebrates+Most of the 59 processing prod-ucts mapped so far terminate with ACCC or ACCCA(Gick et al+, 1986; Scharl & Steitz, 1994)+Although someof these conserved nucleotides may be important forthe binding of HBP (Williams & Marzluff, 1995) andcertain mutations in this region have been found toreduce processing efficiency (Georgiev & Birnstiel, 1985;Spycher et al+, 1994), this sequence has not been an-alyzed in detail+ We therefore have generated and an-alyzed mutations of the ACCCA9CA sequence in amouse histone H4 pre-mRNA+ Mutations in the lastthree of these nucleotides have strong effects on boththe efficiency and site of cleavage in vitro+ Our resultsconfirm previous findings indicating that cleavage oc-curs preferentially downstream of A residues (Scharl &Steitz, 1994), but also reveal the relative preference forcleavage at other nucleotides+ Preference for adeno-sines at the cleavage site is a shared feature with thecleavage involved in polyadenylation, but the order ofcleavage preferences for the other nucleotides is dif-ferent+ Finally, the nucleotide immediately following thecleaved phosphodiester bond also has an influence onthe cleavage specificity+

RESULTS

Definition of the cleavage site

In vitro processing reactions of histone pre-mRNA pro-duce one or two 59 fragments and a nested set of 39 prod-

ucts whose 59 ends map to several positions betweenthe 39 end of the 59 fragment and the spacer element(Gick et al+, 1986; Streit et al+, 1993; Fig+ 2A)+ Becausewe wanted to introduce mutations near the cleavagesite(s) generating the mature mRNA (i+e+, the 59 frag-ment), it was important to define precisely the site of thiscleavage+ For the 12/12 RNAused in this study,we havepreviously compared the cleavage products generatedfrom 59-and 39 end-labeled RNAs by coelectrophoresiswith the products of chemical sequencing reactions(Streit et al+, 1993)+ Since then, we have become awarethat, whereas this determination provided correct re-sults for the 39 fragments, it most likely did not for the 59fragments+ The reason is that the aniline strand scis-sion used in chemical sequencing of RNA(Peattie, 1979)only leads to complete b-elimination on the 39 side of themodified nucleoside+ The 59 fragment still carries rem-nants of the modified nucleoside and aniline ring at its39 end and therefore has an anomalous electrophoreticmobility (C+ Leumann, pers+ comm+)+ This is in contrastto the piperidine strand scission used in chemical se-quencing of DNAthat leads to complete b-elimination onboth the 39 and 59 side (Maxam & Gilbert, 1981)+

We therefore compared the histone-specific process-ing products of 59 end-labeled 12/12 RNA by coelec-trophoresis with adenosine-specific enzymatic andchemical sequencing products, as well as with an al-kaline mononucleotide ladder generated from the sameRNA (Fig+ 2B)+ Ribonuclease U2 cleaves after aden-osines, leaving a 29 or 39 phosphate group+ Becausehistone RNA processing leaves a 39-terminal hy-droxyl group (Wittop Koning, 1993), its product—should the cleavage be after an adenosine—will differby a single phosphate group, causing it to migrateapproximately 0+5 nt more slowly+ Figure 2B showsthat histone-specific cleavage (lane 2) indeed pro-duces a major 59 cleavage product migrating slightlyslower than the U2 product (lane 3) generated fromcleavage after A5, i+e+, after ACCCA+ A minor cleav-age product ending two nucleotides further down-stream, i+e+, after ACCCACA, is also observed+ Thechemical sequencing ladder (Fig+ 2B, lane 4) is re-tarded by one to two nucleotides with respect to theenzymatic one and the extent of this shift varies slightlyalong the gel+ Therefore, only the enzymatic se-quence ladder can be used to appropriately deter-mine the end of the 59 cleavage products+ Comparedto our previous study (Streit et al+, 1993), the twocleavage sites are therefore located two nucleotidesfurther downstream+ The new positions (shown inFig+ 2A) are in agreement with results of Scharl andSteitz (1994), who showed that the in vitro process-ing product of a mouse H2A pre-mRNA ended afterthe sequence ACCCA+ This new assignment of thecleavage site(s) of 12/12 RNA has already been usedin figures of recent papers from our group (Spycheret al+, 1994; Stefanovic et al+, 1995)+

FIGURE 1. Distribution of nucleotides (in absolute numbers) in thefirst seven positions following the RNA hairpin of vertebrate histonegenes+ Sequences of mammalian histone genes found in theGenBank/EMBL sequence library were from mouse (23 genes), rat(11), and man (25)+ Nonmammalian vertebrate sequences were fromrainbow trout (5 genes), Tilapia nilotica (1), Xenopus laevis (26),chicken (26), and duck (5)+ Note that one of the nonmammalianentries ends with the sixth nucleotide+ The most frequent nucleotideat each position is shown in enhanced tone+ The seventh nucleotideis the least conserved and shows a slight preference for A in mam-malian and U in nonmammalian species+ Above the compilation ofmammalian sequences, the sequence of the mouse H4-12 geneused in this work is shown with indices representing the positionsafter the hairpin+

Conserved sequence of histone RNA 39 processing site 247




The 59 fragments generated from 12-U5G6 RNA (forstructure and nomenclature of mutant RNAs, seeFig+ 2C and Materials and Methods) were similarlymapped to end with A7 (major product) and A8 (minorproduct; data not shown)+ For the other mutants, theirproducts were defined by coelectrophoresis with themapped processing products of 12/12 and 12-U5G6

RNAs on high-resolution sequencing gels+ Processingsites of all mutants are indicated by arrows in Fig-ure 2C and changes will be discussed below+

Importance of the nucleotides A 1C2C3C4

immediately following the hairpinIn vitro-transcribed RNAs carrying mutations in the firstfour nucleotides following the hairpin, i+e+, 12-U1, 12-G1A2, 12-A3G4, and 12-A3A4 (for nomenclature and se-quences, see Fig+ 2C),were tested for in vitro processingby incubation in nuclear extract from K21 mouse mas-tocytoma cells (Stauber et al+, 1990)+The wild-type 12/12RNA was used as a standard to determine quantitativeand qualitative effects of the mutations on the process-

FIGURE 2. Cleavage sites of wild-type 12/12 RNA and mutant derivatives+ A: Schematic representation of fragmentsgenerated by in vitro processing of 12/12 RNA derived from the mouse H4-12 gene+ The first 20 nt (vector/polylinkersequences) are not shown (Streit et al+, 1993)+ The 59 fragments, as mapped in B, end at the positions indicated by the black(major product) and white (minor product) vertical arrows+ The 59 end positions of the 39 fragments (arrows pointing to theright) were taken from Streit et al+ (1993)+ The spacer sequence complementary to U7 RNA is underlined+ A square bracketmarks the nucleotides shown for each construct in C+ B: Mapping of 59 fragments+ 12/12 RNA labeled at its 59 end wassubjected to partial KOH digestion (lanes 1, 5), histone-specific RNA processing in K21 cell nuclear extract (lane 2), partialA-specific RNase U2 digestion (lane 3), and partial A-specific chemical treatment (lane 4) and analyzed on a sequencinggel+ The processed 59 products were mapped by comparison with the ribonuclease U2-generated sequencing ladder+ Notethe anomalous mobility of the chemical sequencing ladder that does not allow correct mapping of the processing products(see main text)+ C: Structures and ends of processed 59 products of RNAs used in this work+ Only the nucleotides markedby the square bracket in A are shown for each RNA+ For nomenclature of mutants, see Materials and Methods+ The differentprocessing sites were arbitrarily classified as strong (black arrow), intermediate (grey arrow), or weak (open arrow)+

248 A. Furger et al.




ing reaction+ Because cleavage of the original 12/12RNA in this extract is only reduced to approximately75% by mutations in the hairpin or by addition of un-labeled competitor RNA containing a wild-type hairpin(Streit et al+, 1993), these experiments allowed us toseparate direct effects of mutations on RNA processingfrom effects on the interaction with the HBP+

None of these mutants showed any change in thelengths of the processing products (Fig+ 3A, lanes 2–5)+Moreover, in two separate experiments with incuba-tions of either 2 h (Fig+ 3A) or 30 min (not shown), theirprocessing efficiencies (as assessed by PhosphorIm-ager analysis) did not differ by more than 10% fromthat of 12/12 RNA+ Thus, the conserved ACCC se-quence does not seem to be involved in the specifica-tion of the processing site and, at least in K21 extract,is not important for efficient processing+ Similar resultsfor this set of mutants were obtained in additional ex-periments for which no PhosphorImager quantitation isavailable+

The proximity of the sequence ACCC to the hairpinsuggested that these nucleotides may be important forthe interaction of the pre-mRNA with HBP+ In fact, Wil-liams and Marzluff (1995) have shown that sequencesimmediately downstream of the hairpin are importantfor binding of either nuclear or polysomal stem-loopbinding proteins (their name for HBP; see Discussion)+As already mentioned, processing of 12/12 RNA showsonly little dependence on an intact hairpin and HBPinteractions in K21 extract (Streit et al+, 1993) and there-fore the above experiment did not assess effects of the

mutations due to differential HBP binding+ To assay theinteraction of these mutant RNAs with HBP directly,they were incubated for 30 min in a standard process-ing assay and the formation of the hairpin-specificcomplex was analyzed by electrophoresis on a nonde-naturing agarose–polyacrylamide gel (Melin et al+, 1992)+Under these conditions, the most prominent complexformed is the one with HBP (Fig+ 4A, lane 2)+ In addi-tion, there is a U7 snRNP-specific complex that is usu-ally faint (presumably because it is rapidly turned overduring the processing reaction) and often hidden inbackground radioactivity (e+g+, Fig+ 4B), as well as anunidentified complex X (Melin et al+, 1992)+ The spec-ificity of the HBP complex is demonstrated by the factthat it can be competed by the addition of an excess ofunlabeled RNA containing a wild-type hairpin (Fig+ 4A,lanes 3, 4), but not by mutant hairpin RNA (Fig+ 4A,lanes 5, 6)+ Formation of the complex specific for HBPwas clearly less efficient for 12-U1 (Fig+ 4B, lane 2) thanfor 12/12 RNA (Fig+ 4B, lane 1), although the complex-ity of the gel precludes a precise quantitation+ In con-trast, the HBP complexes formed with 12-G1A2

(Fig+ 4B, lane 3), 12-A3G4 (Fig+ 4B, lane 4), and 12-A3A4 (Fig+ 4B, lane 5) were of very similar intensity asthe one formed with 12/12 RNA+

An additional processing experiment was performedwith the same RNAs in nuclear extract from HeLa cells+In this extract, processing of 12/12 RNA is reducedabout fourfold by mutations in the hairpin or by addingexcess competitor RNAs that can bind HBP (Streitet al+, 1993)+ The results of these experiments (Fig+ 3B,

FIGURE 3. Analysis of RNAs carrying mutations in the sequence A1C2C3C4+ A: In vitro processing in nuclear extract fromK21 mouse mastocytoma cells+ The indicated RNA substrates were incubated in K21 nuclear extract for 2 h at 30 8C andanalyzed on a sequencing gel+ Input, unprocessed substrate RNA+ The 59 and 39 processing products are indicated+ M, HpaII-digested pBR322 DNA used as size marker+ The sample containing 12-U1 RNA shows a low degree of unspecificdegradation+ A particularly active batch of nuclear extract was used in this experiment+ B: In vitro processing in nuclearextract from human HeLa cells+ Procedures were as in A, except that nuclear extract from HeLa cells (a gift of Dr+A+ Krämer,University of Geneva) was used+ Lane 1 shows in vitro processing of 12/12 RNA in K21 extract as a reference+ n+sp+,nonspecific RNA band produced in the HeLa extract+





lanes 2–6) revealed identical cleavage products for allthese RNAs to the ones observed in K21 extract(Fig+ 3B, lane 1)+ The processing efficiencies relative tothat of 12/12 RNA (Fig+ 3B, lane 2) were 71% (12-U1

RNA, Fig+ 3B, lane 3), 86% (12-G1A2, Fig+ 3B, lane 4),40% (12-A3G4, Fig+ 3B, lane 5) and 54% (12-A3A4,Fig+ 3B, lane 6), respectively+

Summarizing these results, none of the mutations inthe first four nucleotides following the hairpin affect thesite of cleavage, be it in K21 or HeLa cell extract+ Only12-U1 RNA is affected in its binding to HBP, probablybecause the 12-U1 mutation alters the secondary struc-ture by adding an A-U base pair to the base of thestem+ This may also explain the slightly reduced pro-cessing efficiency observed with 12-U1 RNA in HeLanuclear extract+ However, the approximately twofold re-duction in processing efficiency of 12-A3G4 and 12-A3A4 RNAs observed in HeLa but not K21 nuclearextract is not related to a reduced binding to HBP+

Importance of the nucleotides A 5C6A7

at the cleavage site(s)

To investigate the importance of the sequence A5C6A7,the mutants 12-G6, 12-U5G6, 12-G6U7, and 12-U5G6U7

were tested for processing in K21 extract+ In addition,the nucleotide at position 5, downstream of which mostof the wild-type RNA molecules get cleaved, was sys-

tematically changed from A to any of the other nucle-otides+ Figure 5 shows that all mutations of A5C6A7

affected the selection of cleavage sites and reducedthe processing efficiency+ The mutations 12-U5 (Fig+ 5,lane 2), 12-G5 (Fig+ 5, lane 4), and 12-U5G6 (Fig+ 5,lane 6) severely reduced cleavage at the wild-type ma-jor site (15) and the total cleavage efficiency was re-duced to 44, 49, and 58%, respectively (Fig+ 5, averageof four experiments)+ This implies that the loss of cleav-age at 15 was partly compensated by an increasedcleavage at position 17+ The fraction of products end-ing at 15 determined for 12-U5 and 12-G5 RNA (28 and17%, respectively) is probably an overestimate, be-cause the quantitation of weak bands is strongly af-fected by background variations+ A weak band endingat 15 is reproducibly visible for 12-U5; longer expo-sures are usually required to detect a correspondingband for 12-G5, but no 15 band can be detected for12-U5G6+ Instead, longer gel exposures reveal a verylow amount of cleavage of 12-U5G6 RNA at a new sitefollowing A8+

The other single change of the fifth nucleotide, 12-C5

(Fig+ 5, lane 3), resulted in an approximately equal fre-quency of cleavage at the 15 and 17 sites+ Interest-ingly, a similar reduction of cleavage at 15 was alsoobserved with the single mutation of the subsequentnucleotide, 12-G6, (Fig+ 5, lane 5)+ Nevertheless, the12-C5 mutation caused a stronger reduction in the over-

FIGURE 4. Gel mobility shift experiment showing the interaction of histone pre-mRNAs with the HBP+ A: Competitionexperiment showing the specificity of the HBP complex+ Radiolabeled 12/12 RNA was incubated in K21 nuclear extract for30 min at 30 8C and analyzed on a nondenaturing agarose/polyacrylamide gel+ In lanes 3–6, a 50-fold (lanes 3, 5) or500-fold (lanes 4, 6) molar excess of unlabeled 34-nt wtHP (lanes 3, 4) or mutHP (lanes 5, 6) RNA was added; mutHP RNAhas an entirely different hairpin, but identical flanking sequences (Martin et al+, 1997)+ HBP, complex of RNAs with hairpinbinding factor; U7, (faint) complex of RNAs with the U7 snRNP; X, undefined (nonspecific) complex (too faint to be seen inthis reproduction, but see panel B)+ This analysis system has been extensively described in Melin et al+ (1992)+ The positionsof the free RNA (migrating as a double band) and of the 59 cleavage product are also indicated+ Note that the 59 product ispreferentially contained in the HBP complex and released upon addition of wtHP competitor+ B: Gel mobility shift experimentwith the RNAs carrying mutations in the sequence A1C2C3C4+ The indicated RNA substrates were incubated in K21 nuclearextract and analyzed as described in A+ Only 12-U1 RNA shows a reduction in intensity of the HBP-specific complex+





all cleavage efficiency (32% compared to 59% for 12-G6)+ In 12-G6U7 (Fig+ 5, lane 7), cleavage occurred withlowered efficiency at A5 and at two novel sites, after C4

and A8, but no cleavage was detectable at the mutatedposition 17+ The total processing efficiency of this mu-tant was 44%, compared to the wild-type 12/12 RNA+In the construct 12-U5G6U7 (Fig+ 5, lane 8), cleavageonly occurred at the new cleavage site after A8 and therelative processing efficiency was 28%+

Cleavage of mutant pre-mRNAs is U 7

snRNP-dependent

To confirm that the 59-products of these mutant RNAswere not due to unspecific, i+e+,U7 snRNP-independentprocesses, four of the RNAs (12/12, 12-U5G6, 12-G6U7, and 12-U5G6U7) were incubated in nuclear ex-tract in which the U7 snRNP had been specificallyinactivated by oligonucleotide-targeted RNAse H treat-ment (Stauber et al+, 1990)+ Cleavage of the wild-type12/12 RNA at the major and minor sites (Fig+ 6, lanes 1,2), as well as cleavage of 12-U5G6 RNA at the A7 andA8 sites (Fig+ 6, lanes 3, 4), were completely preventedby preincubation of the extract with the antisense oli-gonucleotide+ Also, the new cleavage at A8 observed

with the mutants 12-G6U7 (Fig+ 6, lanes 5, 6) and 12-U5G6U7 (Fig+ 6, lanes 7, 8), as well as the band seen for12-G6U7 at the wild-type major cleavage site, wereclearly U7-dependent+ In this experiment, two addi-tional bands migrating slightly faster than the wild-typemajor product were obtained with all constructs+ Thefact that these products were not reduced by the oli-gonucleotide preincubation suggests that they may havebeen due to U7-independent RNA degradation+ Mostimportantly, however, this experiment showed that the59 cleavage products discussed above are the result ofa specific, i+e+, U7-dependent processing reaction+

Mutant RNAs with decreased processingefficiency interact normally with the U7 snRNP

We then wanted to investigate if the decrease in pro-cessing efficiency seen in the mutants 12-U5G6, 12-G6U7, and 12-U5G6U7 was correlated with an instabilityof the U7snRNP–pre-mRNA complex+ This complex canbe detected by incubation of labeled pre-mRNA in mildlyheat-treated (15 min at 50 8C) nuclear extract followedby native gel electrophoresis (Melin et al+, 1992)+ Therationale for using heat-inactivated extract [where pro-cessing is abolished due to the inactivation of a heat-

FIGURE 5. A: In vitro processing of RNAs carrying mutations in thesequence A5C6A7+ The indicated RNA substrates were incubated inK21 nuclear extract for 2 h at 30 8C and analyzed on a sequencinggel+ Input, unprocessed substrate RNAs+ The various 59 processingproducts are designated as 14 to 18 to indicate the position of theirlast nucleotide relative to the hairpin structure+ The average of fourindependent determinations of the relative processing efficiency (12/12RNA 5 100%, integrated over all processing products) is indicatedbelow the lanes+ The fraction of 59 product that is cleaved at posi-tion 5 is indicated as “% product 159+”

FIGURE 6. Dependence of processing products on functionalU7snRNPs+ The U7 snRNPs in nuclear extract from K21 cells wereinactivated by RNAse H digestion targeted with oligonucleotide cAcomplementary to nt 1–16 of U7 RNA (lanes 1; Soldati & Schümperli,1988) or the extract was mock-treated with an oligonucleotide di-rected against the 59 end of U1 RNA (lanes 2)+ The indicated RNAsubstrates were incubated in these extracts for 2 h at 30 8C andanalyzed on a sequencing gel+ Input, unprocessed substrate RNAs;5, 7, 8, 59 processing products ending at the indicated positions afterthe histone hairpin; n+sp+, bands due to unspecific degradation; M1,Hpa II-digested pBR322 DNA; M2, mononucleotide ladder obtainedby mixing four DNA sequencing reactions+





labile processing factor (Gick et al+, 1987)] rather thannative nuclear extract is that, in heat-treated extract, nocleavage of substrate or dissociation of products canoccur; thus, the amount of complex formed is only afunction of the affinity between pre-mRNA and the U7snRNP+ Although the binding behavior of U7 snRNPsmay be altered by this treatment, it has been shownthat heat-treated U7 snRNPs are still functionally ac-tive in complementation assays (Gick et al+, 1987;Lüscher & Schümperli, 1987) and that the binding be-tween pre-mRNAs and U7 snRNPs in such extractsclosely follows the predicted base pairing stabilities ofpotential pre-mRNA–U7 RNA hybrids (Melin et al+, 1992;Spycher et al+, 1994)+ In addition to the heat-labile fac-tor, the heat treatment also inactivates HBP, so that thistype of experiment measures the binding of an RNA tothe U7 snRNP in the absence of functional HBP (Melinet al+, 1992)+ In additon to the U7-specific complex,complex X (already seen with native extract, see Fig+ 4)is still formed, as well as a new unidentified complex Ythat appears only after the heat treatment (Melin et al+,1992; Fig+ 7)+

As in previous experiments (Melin et al+, 1992; Spy-cher et al+, 1994), wild-type 12/12 RNA formed a U7-specific complex (Fig+ 7, lane 1) that could be shifted tothe pocket of the gel by addition of the monoclonalanti-Sm antibody Y-12 (Fig+ 7, lane 3), but not by acontrol monoclonal antibody (Fig+ 7, lane 2)+ Also inagreement with our previous results, the additional com-

plexes X an Y were not affected by either antibody+Most importantly, however, the U7-specific complex wasalso formed with the three mutant RNAs, 12-U5G6

(Fig+ 7, lanes 4, 5), 12-G6U7 (Fig+ 7, lanes 6, 7), and12-U5G6U7 (Fig+ 7, lanes 8, 9)+ Because the intensity ofthese U7-specific complexes did not differ significantlyfrom that obtained with the wild-type 12/12 RNA, weconclude that the severely decreased processing effi-ciency observed with these mutants was not due to adisruption of the U7 snRNP–pre-mRNA interaction+

DISCUSSION

In addition to the two well-characterized hairpin andspacer elements, which interact with two processingfactors, the HBP and the U7 snRNP, respectively, thenucleotides immediately following the hairpin have alsobeen conserved among vertebrate histone genes(Fig+ 1)+ This suggested that the sequence ACCCA9CAmight play an important role in histone RNA 39 pro-cessing and/or in further steps of histone RNA metab-olism+ It seemed unlikely that the high degree ofsequence conservation could be explained entirely bythe need for proper formation of the hairpin structureand/or interaction with HBP+ Rather, the fact that thecleavage event occurs within this sequence and ma-ture mRNAs mostly terminate with ACCCA (Gick et al+,1986; Scharl & Steitz 1994; this work) suggested thatsome nucleotides within this sequence might play arole in cleavage site selection+ The present mutationalanalysis demonstrates that this is indeed the case forthe nucleotides A59C6A7, but not for the first four nu-cleotides, A1C2C3C4+

Putative roles of the conserved ACCC feature

Because none of the mutations of the sequenceA1C2C3C4 show any change in the length of 59 prod-ucts (Fig+ 3), we can safely rule out a role of thesenucleotides in cleavage site selection+ Although it wasnot our primary goal to investigate the specificity ofinteraction between the histone pre-mRNA (or mRNA)and HBP, our mutagenesis of the nucleotides A1C2C3C4

sheds some new light on this interaction+ Previous stud-ies (Pandey et al+, 1991, 1994; Williams & Marzluff,1995) have shown that the sequence of the stem andflanking sequences 59 and 39 of the hairpin are criticaldeterminants for HBP binding+ A deletion of all se-quences 39 of the hairpin [mutant SL(D39)] or a substi-tution of the first three nucleotides by Gs [mutantSL(39GGGCA)] reduced the binding to 5% and ,5%,respectively (Williams & Marzluff, 1995)+ This sug-gested that, in addition to the physical requirement foran oligonucleotide extension 39 of the hairpin, the in-teraction had some specificity for these nucleotides+The fact that neither of the three mutants 12-G1A2,12-A3G4, and 12-A3A4 shows a reduction in HBP bind-

FIGURE 7. Gel mobility shift experiment showing the interaction ofmutant RNAs with the U7 snRNP+ The indicated RNA substrateswere incubated in heat-treated K21 nuclear extract for 30 min at30 8C and analyzed on a nondenaturing agarose/polyacrylamide gel(Melin et al+, 1992)+ U7, complex of RNAs with the U7snRNP; X, Y,undefined (nonspecific) complexes+ Lanes 1, addition of 2+4 mg mono-clonal anti-Sm antibody Y-12 retains the U7-specific complex in thegel pocket; lane c, 2+4 mg control monoclonal antibody added (seeMaterials and Methods); lanes 2, no antibody added+ In lane 9, theshift of the U7 complex by the Y-12 antibody is incomplete+





ing (Fig+ 4) now indicates that this specificity cannot bevery stringent+Only one mutant, 12-U1, binds HBP lessefficiently+ However, this mutation is predicted to addan A-U base pair to the base of the hairpin and maytherefore affect HBP binding by altering the hairpin struc-ture itself+ Hence, it is more similar to the previouslyreported mutant [SL(GC)3], in which an additional G-Cbase pair was added to the foot of the stem and whichshowed ,5% binding to HBP (Williams & Marzluff,1995)+ It is therefore possible that the binding of HBPonly reacts to extensive sequence changes in the 39flanking nucleotides+Alternatively, the strongly reducedbinding of mutant SL(39GGGCA) in the study by Wil-liams and Marzluff might have been due to the RNAassuming an alternative secondary structure+

Three of the mutations in the A1C2C3C4 sequenceshow some effect on processing efficiency in HeLa(Fig+ 3B) but not K21 extract (Fig+ 3A)+ The moderate(29%) reduction in processing efficiency seen with 12-U1

RNA in HeLa extract is most likely related to the re-duced HBP binding discussed above, because process-ing of the wild-type 12/12 RNA is known to show somehairpin dependence in nuclear extract from HeLa butnot from K21 cells (Streit et al+, 1993)+ However, in theother two mutants, 12-A3G4 and 12-A3A4 (approximate-ly 50% reduction of processing), the binding to HBP isunaffected+ The cause for the reduction in processingefficiency in HeLa cells seen with these mutants re-mains obscure, especially because it is not accompa-nied by a change in cleavage specificity+

The functional importance of the sequenceA5C6A7 for histone RNA 39 processing

Our results demonstrate that nt 5–7 following the hair-pin play an important role in specifying the cleavagesite+ Normally, cleavage occurs mostly after A5 and, ina minority of events, after A7 (Fig+ 2)+ In all cases whereone of these two As has been mutated, the cleavage isstrongly or completely redirected to other sites (Fig+ 5;for a summary, see Fig+ 2C)+ Most of the cleavageevents recorded in this work occur downstream of Aresidues+ The only exceptions are the inefficient cleav-ages after C4 in 12-G6U7 and after C5, U5, and G5 in thedifferent RNAs with mutations at position 5+ Thus, anadenosine is by far the favored nucleotide 59 of thecleavage site+ The same conclusion was alreadyreached by Scharl and Steitz (1994), albeit in a differ-ent experimental context+ They inserted variable num-bers of C residues between the spacer and hairpinelements and found that, as this distance increased,the cleavage site(s) were moved in 39 direction by anaccording number of nucleotides+ This led to the pro-posal that the U7 snRNP serves as a “molecular ruler”defining the cleavage site at a more or less fixed dis-tance upstream of the pre-mRNA–U7 RNA hybrid+How-ever, in most of the insertion mutants, cleavage (now

occurring within the inserted Cs) was inefficient+ Thiseffect could be alleviated partly by positioning an Aresidue at an appropriate distance upstream from thepre-mRNA–U7 RNA hybrid+ In some cases, cleavagealso became redirected to a nearby but not ideally po-sitioned A residue+ A main difference between thesetwo studies is that we have altered the nucleotides atthe cleavage site in their natural context,whereas Scharland Steitz (1994) did so in an “extended configuration,”where productive interactions within the processing com-plex were somewhat disturbed, as is indicated by thefact that processing efficiency in the A replacement con-structs did not reach wild-type levels+

Based on nucleotide sequence comparisons, theadenosine at position 5 appears to be a widespreadfeature of animal histone genes+ Screening of theGenBank/EMBL data bank revealed that, in 75+5% of122 analyzed histone genes from fish, frogs, chicken,duck, mouse, rat, and man, the fifth position followingthe hairpin was an A residue; 19+5% of the genes con-tained a C and 5% a U at this position+ Most genescontaining a C or U at the fifth position had an A atposition 4 or 6+ This strongly suggests that the adeno-sine at position five or at one of its neighbors generallyspecifies the cleavage site in metazoan replication-dependent histone genes+

Having studied all possible nucleotide variations atposition 5 (Fig+ 5), we are able to derive a hierarchy ofcleavage preferences+ A cytidine residue is toleratedwith perhaps a threefold reduction in cleavage at po-sition 5+ The other two changes, to U or G, obviouslyhave stronger effects than the change to a C+ If onelooks at the ratio of 59 products cleaved at position 5(mutated) versus 7 (wild-type A), the preference is A .C . U . G, although the difference between U and Gmay be negligible+Several other replacements also clas-sify uridine as a highly unfavorable residue+ Changingthe fifth position from 12-G6 to 12-U5G6 completely elim-inates cleavage at position 5+ Furthermore, the cleav-age at position 7 is completely eliminated if one altersit either from 12-G6 to 12-G6U7 or from 12-U5G6 to12-U5G6U7+

The above order defines the hierarchy of cleavage atposition 5, but, interestingly, a different order (A . U,G . C) is obtained if one looks at the total processingefficiency+ This means that the nucleotide at position 5may also influence to what extent the loss of cleavageat position 5 can be compensated for by increasedcleavage at the unaltered position 7+

It has been noted previously (Scharl & Steitz, 1994)that the preference for cleavage downstream of aden-osines is a common feature of the cleavage event as-sociated with polyadenylation and histone RNA 39processing+ Polyadenylation cleavage was originally re-ported to occur preferentially at A’s (Sheets et al+, 1990),but, in a recent analysis, the complete hierarchy of pref-erences was derived to be A. U . C .. G (Chen et al+,





1995)+ Thus, although the two cleavage activities ap-pear to share the most (adenosine) and least (guano-sine) preferred nucleotides, they clearly differ in therelative order of cytidines and uridines+

In addition to the nucleotide preceding the cleavedphosphodiester bond, the one following it also appearsto contribute to the cleavage specificity+ We have notsystematically changed the nucleotide at position 6,but, in two different contexts, replacement of C6 by Greduced the frequency of cleavage at its 59 phosphate+Going from the wild-type 12/12 RNA to 12-G6 resultedin a shift from 89% to 55% relative utilization of this site(in absolute amounts, the cleavage at this site wasreduced approximately threefold); going from 12-U5 to12-U5G6, the reduction was from 28% relative utiliza-tion to undetectable levels+ Most interestingly, the mu-tant 12-G6U7 caused not only a reduction in cleavageat position 5, but also a partial redirection of cleavageactivity to the phosphodiester bond following C4, al-though A5 was not changed in this mutant+

Our results also demonstrate the range within whichthe cleavage site can be located+ In the “molecularruler” concept of Scharl and Steitz (1994), the U7snRNP, by base pairing to the spacer element of thepre-mRNA, defines a window within which cleavagecan occur and within which the most suitable phospho-diester bond is selected+ Scharl and Steitz (1994) ob-served a maximal variation over four phosphodiesterbonds in one of their insertion mutants+ In our study, wehave observed cleavages after C4 (in 12-G6U7) andafter A8 (in several mutants), but not at positions furtherremoved from the natural site+ Because the cleavagesat these positions were always inefficient (but the lowefficiency was not due to an inability of the pre-mRNAsto bind U7 snRNPs; Fig+ 7), these positions appear torepresent the true limits for this window, i+e+, our resultsconfirm a maximal range of four nucleotides+

MATERIALS AND METHODS

Nomenclature and construction of mutants

The original 12/12 construct has 20 nt of pSP65 polylinkerand 65 nt of the 39 untranslated region from the H4-12 gene(Meier et al+, 1989) containing the hairpin and spacer ele-ments (Fig+ 2A) inserted upstream of a Hind III site into themodified BlueScript vector OT7 (Spycher et al+, 1994)+ Themutants constructed for this study are presented in Fig-ure 2C+ In the nomenclature used, the letters correspond tothe exchanged nucleotides with respect to the wild-type se-quence+ The indices indicate the positions of the mutations,starting with the first nucleotide after the hairpin+ The cloningstrategy consisted of two steps+ First, the sequence ACCCACAwas changed to ACCTGCCA by cloning of double-strandedoligonucleotides+ This generated a BspM I cleavage site (ACCTGC) with a nonpalindromic 59 overhang a few nucleotidesfurther downstream+ Note that the 12-BspM I construct islonger by one nucleotide+

To produce the final mutants, individual 41-nt bottom strandoligonucleotides containing the specific mutations were hy-bridized with a common 24-nt top strand oligonucleotide (end-ing one nucleotide before the region containing the mutations)+This resulted in partly double-stranded oligos with appropri-ate 59 and 39 overhangs that allowed ligation to the Sac I andBspM I sites of OT7 12-BspM I+ The gap of nine nucleotideswas filled in by T4 DNA polymerase+ By replacement of theupstream part of the histone insert, the BspM I recognitionsequence was removed and the original length of the insertwas restored+ Most mutations were selected to createtransversions+

Enzymatic reactions and cloning in Escherichia coli, strainHB101, were performed by standard procedures (Sambrooket al+, 1989)+ All nucleotide sequences were verified by DNAsequencing (Sanger et al+, 1977) using the Sequenase Ver-sion 2+0 DNA Sequencing Kit (USB)+

In vitro transcription

Templates were linearized by digestion with Hind III, succes-sively extracted twice with phenol, once with chloroform-isoamylalcohol (24:1;CIA), and precipitated with ethanol+Onemicrogram linearized template was incubated for 1 h at 37 8Cin presence of 0+05 mM rGTP, 0+5 mM each of rATP, rCTP,rUTP, 25 mCi a-32P-rGTP (800 Ci/mmol, New England Nu-clear), 20 mM Tris-HCl, pH 7+5, 10 mM MgCl2, 1 mM dithio-threitol, 10 U T7 (or SP6, for mutant 12-G4U5G6) RNAPolymerase (New England Biolabs), and 40 U RNasin (Pro-mega)+ After 1 h, a further 10 U of RNA polymerase wereadded and the samples incubated for another 30 min+ Theresulting RNAs were purified on a 6% polyacrylamide/8+3 Murea gel, excised, and eluted in 300 mL of 0+3 M NaCl, 0+1 MEDTA, 10 mM Tris-HCl, pH 7+5, for 1–2 h+ RNAs extractedtwice with phenol and once with CIA were precipitated withethanol+ Unlabeled RNA was transcribed as described above,but in the presence of 0+5 mM rGTP+

In vitro processing

Individual reactions contained, in a final volume of 12 mL,20,000 cpm (approximately 30 fmol) universally labeled RNA,3 mg yeast tRNA, 16 mM EDTA, and 5 mL nuclear extractfrom K21 mouse mastocytoma cells (Stauber et al+, 1990; agift of R+ Mital)+

The reaction mixture was pre-incubated without the pre-mRNA for 10 min on ice, followed by the addition of radio-labeled pre-mRNA, and further incubation for 2 h at 30 8C+The reaction was terminated by adding 2 mg of heparinand incubating for 10 min on ice+ Following electrophoresison a 6% polyacrylamide/8+3 M urea gel, the RNA wasdetected by autoradiography or quantitated by PhosphoIm-ager 425 (Molecular Dynamics) analysis+ Processing effi-ciency was determined after multiplying the integratedvolumes (intensities) of unprocessed and processed RNAswith correction factors, taking into account the different num-bers of labeled guanosines+

In some experiments, the U7 snRNP was inactivated priorto processing by oligonucleotide-targeted RNAse H treat-ment (Stauber et al+, 1990)+ Briefly, 40 mL nuclear extract(which contains endogenous RNAse H) was incubated in the





presence of 2+7 mM MgCl2 and 400 ng synthetic DNA oligo-nucleotide cA (Soldati & Schümperli, 1988) for 15 min at20 8C+ Oligo cA is complementary to nt 1–16 at the 59 end ofmouse U7 RNA+ The digestion was stopped by adding20 mM EDTA+As a control, the same reaction was performedin the presence of 400 ng 0ligo U1a, which is complementaryto the 59 end of the U1 RNA+

Gel retardation assays

Processing reactions were performed as described above(30 min at 30 8C) with native nuclear extract or extract thathad been heat inactivated by preincubation for 15 min at50 8C (Stauber et al+, 1990)+ In some reactions, the U7-specific complex was “super-shifted” by adding 2+4 mg mono-clonal anti Sm antibody (a gift of I+ Haussmann) followed byincubation for 10 min on ice+ As control, a monoclonal anti-body for the yeast SEP1 protein (kindly provided by W+-D+Heyer) was used+ In other reactions, the formation of a com-plex with the HBP was prevented by the addition of an ex-cess of unlabeled 34-nt wtHP or mutHP RNAs (Martin et al+,1997)+ Reactions were stopped by adding 2 mg heparin andincubating for 10 min on ice+ The complexes were analyzedon composite agarose–polyacrylamide gels, as described (Me-lin et al+, 1992)+

59 End-labeling and sequencing of RNA

Approximately 20 pmol unlabeled RNA treated with calf in-testine phosphatase was incubated with 25 mCi g-32P-ATP,1 U RNAsin (Promega), 50 mM Tris-HCl, pH 9, 10 mM MgCl2,8 mM DTT, 5% glycerol, and 20 U T4 polynucleotide kinase(Boehringer Mannheim) for 45 min at 37 8C+ The 59-labeledRNA was gel-purified as described above for uniformly la-beled RNAs+

For mononucleotide ladders, 50,000–100,000 cpm of 59end-labeled RNA was incubated with 12 mM KOH in a totalvolume of 205 mL for 8 or 4 min at 42 8C+ The samples wereput on ice immediately and the reaction was terminated byadding 7 mL acetic acid (100%)+ One microgram glycogenwas added as carrier and the RNA precipitated with ethanol+Chemical sequencing reactions involving modifications of Aresidues by diethyl pyrocarbonate and strand scission by an-iline were performed as described (Peattie, 1979)+ An enzy-matic adenosine-specific cleavage was obtained by incubating30,000–50,000 cpm 59 end-labeled RNA with 8 mM sodiumcitrate, pH 3+5, 0+4 mM EDTA, 1+75 M urea, 1+5 mg yeasttRNA, and 0+1 U U2 ribonuclease (Pharmacia) for 12 min at55 8C+ The reaction was terminated by placing the sampleson ice and electrophoresed immediately+ The RNA sampleswere analyzed on 6% polyacrylamide/8+3 M urea gels+

ACKNOWLEDGMENTS

We thank Branko Stefanovic, Andreas Gruber, and IrmgardHaussmann for technical and other advice, Isabel Roditi andBerndt Müller for critical comments on the manuscript, andToni Wyler for preparing the photographs+ We acknowledgefinancial support by the State of Bern and by grants 3100-27753+89 and 3100-41751+94 from the Swiss National Sci-ence Foundation+

Received September 19, 1996; returned for revisionOctober 18, 1996; revised manuscript receivedDecember 23, 1997

REFERENCES

Birnstiel ML, Schaufele FJ+ 1988+ Structure and function of minorsnRNPs+ In: Birnstiel ML, ed+ Structure and function of major andminor small nuclear ribonucleoprotein particles+ Berlin: SpringerVerlag+ pp 155–182+

Bond UM, Yario TA, Steitz JA+ 1991+ Multiple processing-defectivemutations in mammalian histone premessenger RNA are sup-pressed by compensatory changes in U7 both in vivo and in vitro+Genes & Dev 5:1709–1722+

Chen F, MacDonald CC,Wilusz J+ 1995+ Cleavage site determinantsin the mammalian polyadenylation signal+ Nucleic Acids Res14:2614–2620+

Georgiev O, Birnstiel ML+ 1985+ The conserved CAAGAAAGA spacersequence is an essential element for the formation of 39 termini ofthe sea urchin H3 histone mRNA by RNA processing+ EMBO J4:481–489+

Gick O, Krämer A, Keller W, Birnstiel ML+ 1986+ Generation ofhistone mRNA 39 ends by endonucleolytic cleavage of the pre-mRNA in a snRNP-dependent in vitro reaction+ EMBO J 5:1319–1326+

Gick O, Krämer A, Vasserot A, Birnstiel ML+ 1987+ Heat-labile regu-latory factor is required for 39 processing of histone precursormRNAs+ Proc Natl Acad Sci USA 84:8937–8940+

Lüscher B, Schümperli D+ 1987+ RNA 39 processing regulates his-tone mRNA levels in a mammalian cell cycle mutant+ A process-ing factor becomes limiting in G1-arrested cells+ EMBO J 6:1721–1726+

Martin F, Schaller A, Eglite S, Schümperli D, Müller B+ 1997+ Thegene for histone RNA hairpin-binding protein is located on humanchromosome 4 and encodes a novel type of RNA-binding protein+EMBO J 16:769–778+

Maxam AM, Gilbert W+ 1981+ Sequencing end labelled DNA withbase specific chemical cleavages+ Methods Enzymol 65:499–560+

Meier VS, Böhni R, Schümperli D+ 1989+ Nucleotide sequence of twomouse histone H4 genes+ Nucleic Acids Res 17:795+

Melin L, Soldati D,Mital R, Streit A, Schümperli D+ 1992+ Biochemicaldemonstration of complex formation of histone pre-mRNA withU7 small nuclear ribonucleoprotein and hairpin binding factors+EMBO J 11:691–697+

Moore CL, Sharp PA+ 1985+ Accurate cleavage and polyadenylationof exogenous RNA substrate+ Cell 41:845–855+

Mowry KL, Oh R, Steitz JA+ 1989+ Each of the conserved sequenceelements flanking the cleavage site of mammalian histone pre-mRNAs has a distinct role in the 39 end processing reaction+ MolCell Biol 9:3105–3108+

Pandey NB, Sun JH, Marzluff WF+ 1991+ Different complexes areformed on the 39 end of histone mRNA with nuclear and poly-somal proteins+ Nucleic Acids Res 19:5653–5659+

Pandey NB, Williams AS, Sun J, Brown VD, Bond U, Marzluff WF+1994+ Point mutations in the stem-loop at the 39 end of mousehistone mRNA reduce expression by reducing the efficiency of 39end formation+ Mol Cell Biol 14:1709–1720+

Peattie DA+ 1979+ Direct chemical method for sequencing RNA+ ProcNatl Acad Sci USA 76:1760–1764+

Sambrook J, Fritsch EM, Maniatis T+ 1989+ Molecular cloning: A lab-oratory manual, 2nd ed+ Cold Spring Harbor, New York: ColdSpring Harbor Laboratory Press+

Sanger F, Nicklen S, Coulson AR+ 1977+ DNA sequencing with chainterminating inhibitors+ Proc Natl Acad Sci USA 74:5463–5467+

Scharl EC, Steitz JA+ 1994+ The site of 39 end formation of histonemessenger RNA is a fixed distance from the downstream elementrecognized by the U7 snRNP+ EMBO J 13:2532–2440+

Schaufele F, Gilmartin GM, Bannwarth W, Birnstiel ML+ 1986+ Com-pensatory mutations suggest that base pairing with a small nu-clear RNA is required to form the 39 end of H3 messenger RNA+Nature 323:777–781+

Sheets MD, Ogg SC,Wickens MP+ 1990+ Point mutations in AAUAAAand the poly (A) addition site: Effects on the accuracy and effi-





ciency of cleavage and polyadenylation in vitro+ Nucleic AcidsRes 18:5799–5805+

Soldati D, Schümperli D+ 1988+ Structural and functional character-ization of mouse U7 small nuclear RNA active in 39 processing ofhistone pre-mRNA+ Mol Cell Biol 8:1518–1524+

Spycher C, Streit A, Stefanovic, B, Albrecht, D, Wittop Koning TH,Schümperli D+ 1994+ 39 End processing of mouse histone pre-mRNA: Evidence for additional base-pairing between U7 snRNAand the pre-mRNA+ Nucleic Acids Res 22:4023–4030+

Stauber C, Soldati D, Lüscher B, Schümperli D+ 1990+ Histone-specific RNA 39 processing in nuclear extracts from mammaliancells+ Methods Enzymol 181:74–89+

Stefanovic B, Wittop Koning TH, Schümperli D+ 1995+ A synthetichistone pre-mRNA–U7 small nuclear RNA chimaera undergoingcis-cleavage in the cytoplasm of Xenopus oocytes+ Nucleic AcidsRes 23:3152–3160+

Steitz TA, Steitz JA+ 1993+ A general two-metal-ion mechanism forcatalytic RNA+ Proc Natl Acad Sci USA 90:6498–6502+

Streit A, Wittop Koning T, Soldati D, Melin L, Schümperli D+ 1993+

Variable effects of the conserved RNA hairpin element upon 39end processing of histone pre-mRNA in vitro+ Nucleic Acids Res21:1569–1575+

Vasserot AP, Schaufele FJ, Birnstiel ML+ 1989+ Conserved terminalhairpin sequences of histone mRNA precursors are not involvedin duplex formation with the U7 RNA but act as a target site for adistinct processing factor+ Proc Natl Acad Sci USA 86:4345–4349+

Wahle E, Keller W+ 1992+ The biochemistry of 39 end cleavage andpolyadenylation of messenger RNA precursors+ Annu Rev Bio-chem 61:419–441+

Williams AS, Marzluff WF+ 1995+ The sequence of the stem andflanking sequences at the 39 end of histone mRNA are criticaldeterminants for the binding of the stem-loop binding protein+Nucleic Acids Res 23:654–662+

Wittop Koning TH+ 1993+ Histone pre-mRNA 39 processing: A closerlook into the molecular mechanism [thesis]+ Bern: University ofBern+





Functional importance of conserved nucleotides at the histone RNA 3' processing site

Documents