Monocistronic transcription is the physiological mechanism of sea urchin embryonic histone gene expression

Article

Reference

Monocistronic transcription is the physiological mechanism of sea urchin embryonic histone gene expression

MAURON, Alex, et al.

Abstract

We have examined histone gene expression during the early stages of sea urchin embryogenesis. The five histone genes expressed at that time are contained in tandem repetitive segments. It has been suggested that adjacent coding regions and their intervening spacer sequences are transcribed into large polycistronic messenger ribonucleic acid (RNA) precursors. We have subcloned into pBR322 deoxyribonucleic acid (DNA) sequences mapping either in the coding region, the 5' spacer, or the 3' spacer of the H2B histone gene.

These clones were used to produce radioiodinated hybridization probes. We measured the steady-state quantity of H2B messenger RNA as well as spacer-specific RNA in the total RNA from embryos taken at various stages of development from fertilization to hatching of blastulae (0 to 22 h post-fertilization). Small amounts of RNA hybridizing to both spacer probes could be found. However, we show that these RNAs form mismatched hybrids with the spacer DNA and therefore cannot originate from the spacers present in the histone genes. We conclude that there is no detectable transcription of the spacer regions on [...]

MAURON, Alex, et al. Monocistronic transcription is the physiological mechanism of sea urchin embryonic histone gene expression. Molecular and Cellular Biology, 1981, vol. 1, no. 7, p.

661-71

PMID : 9279379

Available at:

http://archive-ouverte.unige.ch/unige:85374

Disclaimer: layout of this document may differ from the published version.

1 / 1

Vol.1,No. 7 MOLECULARAND CELLULAR BIOLOGY,JUly1981, p.661-671

0270-7306/81/010661-11$02.00/0

Monocistronic Transcription Is the Physiological Mechanism of Sea Urchin Embryonic Histone Gene Expression

ALEXMAURON,SHOSHANALEVY,tGEOFFREYCHILDS,4 ANDLARRY KEDES*

HowardHughesMedicalInstituteLaboratory,*DepartmentofMedicineatStanfordMedicalSchool,and Veterans Administration MedicalCenter,PaloAlto,California94304

Received4February1981/Accepted27April1981

Wehave examined histonegeneexpressionduringtheearlystagesof^seaurchin embryogenesis. The five histone genesexpressed atthat time arecontained in tandemrepetitive segments. Ithas beensuggestedthatadjacent coding regions and theirintervening spacer sequencesare transcribed intolarge polycistronic messengerribonucleic acid (RNA) precursors. We have subcloned intopBR322 deoxyribonucleic acid (DNA)sequencesmappingeither in thecodingregion,the 5' spacer, orthe 3'spacer of the H2B histonegene. These cloneswere usedto produce radioiodinatedhybridizationprobes. We measured thesteady-statequan- tity of H2B messenger RNA aswell as spacer-specific RNA in the total RNA fromembryos takenatvariousstagesofdevelopmentfromfertilizationtohatch- ing ofblastulae (0to22hpost-fertilization). Smallamountsof RNAhybridizing tobothspacerprobes could be found.However,weshow that these RNAs form mismatchedhybrids with thespacer DNAandthereforecannot originate from thespacers present inthe histonegenes.Weconcludethatthere isnodetectable transcription of thespacerregionsoneither side of the H2Bhistone gene.The detection limit for RNAcomplementarytothe 5'spacer sequencecorrespondsto amaximum of about three RNAmoleculespercell,an amountshowntobefar less than theprojectedsteady-state pool size ofaputativepolycistronic transcript, if suchaprecursor were tobe theobligatorytranscript of the histonegenes.(This conclusionwas derived byusing the knownratesofproduction ofH2B mRNA throughout early development [R. E. Maxson and F. H. Wilt, Dev. Biol., in press].) The physiologically relevant transcript ofthe histone genes in early developmentistherefore monocistronic and probablyidenticaltothe messenger RNA itself.

The histone genes transcribed during the rapidcleavagestageof echinodermembryogen- esisaremembers ofamoderately repetitivefam- ily of tandemly arrayed genetic units each of which contains, onthe same deoxyribonucleic acid(DNA)strand, codingsequencesfor the five classes of histone messenger ribonucleic acid (mRNA)interspersed withspacer DNA(seeFig.

1;reviewedinreference 19). The DNAsequence ofnearlythe entirerepeating unitisknown for threespeciesofseaurchin(3, 33, 38,39; R.Cohn and L. H. Kedes, manuscript in preparation), and the5'and 3' endsof themRNA's havebeen determined and mapped on the DNA (17, 22, 37).

The nature of theprimarytranscripts of these genes is unknown. On one hand, it is evident

tPresent address: Stanford Research Institute, Menlo Park, CA94025.

fPresent address: Albert Einstein College ofMedicine, Bronx,NY 10461.

that theirorganization might easilylend itselfto polycistronic transcription. However, in other invertebrates, histone genes do not have the sameorganization:inDrosophila melanogaster theyareclustered butare notallthesameDNA strand (23). In yeasts, histone protein coding regions exist in unlinked pairs that are oppo- sitely oriented (18). Furthermore, ultraviolet mapping experiments in HeLa cells measured the sizeof histonegeneprimary transcripts(16) as only being able to accommodate single mRNA's. Thus, polycistronic transcription of histone genes cannotbeauniversal strategyfor theirexpression.

Several reports haveshown thatlargenucleic acid molecules synthesized by sea urchin em- bryoscontain sequences that formhybrids with histone genes (20, 35). The latterreport holds that suchmoleculescontainsequences thathy- bridize with more than one histone gene and suggests that polycistronic transcription does take place. Evidence to the contrary however 661

was obtained in our previous experiments (5):

althoughwecould detect high-molecular-weight RNA that hybridizedtohistone genesimmobi- lized on a cellulose matrix, such RNAs were

foundtobe cross-hybridizing speciesrather than authentictranscripts of thosegenes.

Independent evidence against polycistronic transcription ofseaurchin histone genes came

from ultraviolet-inactivationstudies (G. Childs, unpublished data). If polycistronic transcription startedatasingle location in the histonerepeat unit,oneshould expect theappearanceof indi- vidual histone mRNA's that are differentially ultravioletsensitive, dependingontheirposition in the transcription unit (16 and references therein). Wewereunableto detectsuch a gra-

dient of histone mRNA ultraviolet sensitivity, whereas in the same experiments, the polarity of 18S and 28S rRNA could easily be demon- strated. This resultarguesagainstasinglepro-

moterperrepeatunit but itcannotexclude the existence ofoverlapping polycistronic transcrip- tion units starting in front of each of the five

genes. The work reported in this paper was

designedtodefinethe nature of thehistonegene

transcript in the seaurchin and to resolve the discrepancies of previousreports.First,wepres-

ent qualitative evidence against polycistronic transcription obtained by separating embryo RNAondenaturing gels, blottingontodiazoben- zyloxymethyl (DBM)-paper (1), and hybridizing to histone-specific probes. The main thrust of thepresentpaperis, however,ofaquantitative nature.Ifpolycistronicorhigh-molecular-weight

precursortranscriptionorboth dotakeplaceon sea urchin histonegenes, then itmustcross or

extend into thespacersegmentsflankingeachof the mRNAcoding regions. Furthermore,if tran- scriptionfrom theflanking regionsoccurs,RNA moleculescontainingthesesequencesshould be detectable,evenif the RNA moleculesareshort- lived precursors. In fact, such RNA species oughttobepresentinamountscompatiblewith the massiverate of histone mRNA production occurring in early embryos,at leastifsuch tran- scriptsaretobe ofanyphysiological significance.

The evidencewepresenthere demonstrates that thereisnosignificant transcriptionof thespacer

Hl H4

regions between the H4, H2B, and H3 histone

genes.We conclude thatthe primary transcript of the H2Bgeneand, by inference,allfive early histone genes are monocistronic and probably identical with the histone mRNA.

MATERIALS AND METHODS Cloningof restriction fragments. To obtainpure

DNAprobes,wegeneratedaseries of subclones from segments of thepreviously characterized seaurchin Strongylocentrotus purpuratus clones pSpl02 and pSpll7 (8). The HhaI-BamHI fragment ofpSpl02 constituting the 5' spacer probe was cloned, using standard recombinant DNA technology, and the re-

sultant clonewascalledpSp2b-1. This restrictionfrag- mentlies entirely within thatspacerregion: itstarts in the middle of the H4-H2B spacer and ends 37 nucleotidesupstreamofthe H2B mRNAstart(Fig.1).

The DNAsequenceofthis fragmentwasdetermined by the method ofMaxamand Gilbert(24) (Fig.2).In cloning this fragment, wefollowed the proceduresof Ullrichetal. (40) exceptwhenmentioned otherwise.

Briefly, the endsofthisfragmentwereconverted to bluntends, usingreversetranscription [conditions:0.5 mM of each of the fourdeoxynucleotide triphosphates, 50 mM tris(hydroxymethyl)aminomethane-hydro- chloride(pH 8.3),6 mMMgCl2,40mMKCI, and0.2 mMprimer-templateincubated for1hat37°C]. 32p_

end-labeledHindIIIlinkerswerethenligatedontothe fragmentwith T4ligase,and the linkedfragmentwas

treated withHindIIIandligatedinto thecorrespond- ingsite ofHindIII-cut pBR322. Ligationof^nonrecom- binant circleswasprevented by removingthe 5'phos- phatesfrom the linearizedvectorwith calfintestinal phosphatase [in 100 mM tris(hydroxymethyl)ami- nomethane-hydrochloride, pH 7.5].Transformantsof E. coli HB101werescreened, usingthenitrocellulose filtercolonyassayofGrunstein andHogness (14). The probeused in this stepwastheinsert ofpSplO2labeled with 32Pbynick translation(32).

The eucaryotic fragment ofpGC1 (the 3' ^spacer probe)is 167 basepairs longand derivesfromaregion between the H2B and H3 genes from base 1131 to base 1297 inourpublishedsequence (39).Unlikethe othersubclones,itwasnot created withHindIIIlink- ers,but since it isanEcoRI-BamHIfragment,itwas

substituted for the EcoRI-BamHI fragment of pBR322. Thisfragment starts at apositionapproxi- mately467 bases downstream of the terminus of the H2B mRNAcoding regionand ends 43 nucleotides 5' totheH3 mRNAcappingsite(Fig. 1).

Isolation, labeling, and strand separation of probe DNA. The techniques for growing plasmid-

H2B H3 H2A

I.

~

^{- J}

flt -t

Hha BamHIl tBamHI HaIl

HaelII

EcoRI

FIG. 1. Map of probe fragments. The thick lines represent protein coding regions. The direction of transcriptionislefttoright.The three restrictionfragmentsshownweresubcloned intopBR322and usedas

hybridizationprobesforthe H2Bcodingregion,the5',andthe 3' spacers,respectively.

MOL. CELL. BIOL.

VOL. 1,1981 HISTONE GENE EXPRESSION 663

OC1

^{< CD} <: C

OC

^<

CO CO

^{C- LOU Ow}

o<

_F-

0

%4r

WI^o CD ^{o rF oQ %0 CD}

< N Q nQ w 1W< E< m

CD C-J < < W CD 4J M i

F

rF<

^{C- CD < C)}

^C)Q

^{L.z zCD C) < C) 4-u M QX}

ol ^Zr 01

oi

^{L9 01o C-}

Cpoz

^{z o cF oQ n-,}

< W- (D N 0 n z z F- < ul 4- M

W CD < < M z w 0 4-0s %

0 o Q0 Q z z CD C"mU

trCD Q (DO z z < cP 1

C) < C) C < 0 CD 0 < CD 0 '%

co ^c< o < co qx c F-o < co W: ti 00 4-

ti 4c 4- M )QOw<Fu .

tS~~1--

~-<1 - CD L. " eg < W LA u CP -a>

QO Q o Z~~.cF F W F Xw

oF< o F < o oLi C oc onCJ CD u

CD C) < C) z z ZC 4i W

1%0CD J F- 1% z z *0 1%0 < %O cm

F~~~~~~~~ z rn ZcF Q UI +-n m

<0 z z I- ^ CD)O* St.

F ^CD

<<

^{CD J}

F ~~~z

^{C) z z C)Z C) CD 0 4JU Mco sb. tsX}

Ln< 1- L: Ln ZZ N F < F- VN 0 U t

W CD <F Z ri %t W W Ul U *st

Li CD ( W z z < z -Soti

Q o<F N ~~~z

UeF<<cn0-.

CD LA < o < + m XcS

0 < F< U CD m U ,}b

Q~~ ~~~ %t 0 U WN CD -1 0 (w 4' M _;

F~~~~~~~% < < <Fn0 u n XR X

L9 Q rXw@SteW 4-

W < M ^°ei

1- C) C) CD W C)11Qs 0>

LP < P-1 rn tn W m m 4- m

F < < F < F < 0 W +j co M w ts

F < F <

FI.

^{< o--c CP U m + eb}

Lo CDo C OQ QOU CD uCD E

o <:F o F ZC-o>o oc o UnQCD oX

C) CD :F CDN to CD W C= CD< C) mU C) m 4i E e

F < F

~~<

^{" < rn}

CDCw<

^{WN OQA}

F<Fzc F < F cc Q

Q~CD

o oc Q m C

!o

C) CD C)o- CD 1-- C) < 1- o CD i C) < C) U CM Nt

<

cQo_XwCD

^{C:-9 M S}

t~~~~C <

r ^<

t

^{n W% CD ci -. *j m}

Q~ ~~~~-

^<

<c

^1-

F-

^{+j m}

O Q ~~~< < oQ 4-c M

harboring bacteria andpurifying plasmid DNA have been described previously (8). PlasmidDNA was di- gested with the appropriate restrictionenzyme to lib- eratethe eucaryotic insert. The vector insert mixture was extracted severaltimes withphenol-chloroform, concentrated by ethanol precipitation, and labeled with "I according to Lev et al. (21). Removal of unreactediodide was also done as described by Lev et al., and the strand separation (and at the sametime, theseparation of insert and plasmid) was performed on alarge-pore polyacrylamide gel (25). Finally, resid- ual contamination by the opposite strand was elimi- nated by allowing it to reassociate and separating the single-strandedfractiononhydroxyapatite.This ma- terial was used as a probe. The reactivity of such probes was 60 to90%, and the background of self- annealingwasless than 4%.

Hybridization conditions. Hybridization and hy- droxyapatitechromatography, including the low-salt ribonuclease (RNase)control,weredoneessentially as described by Schelleretal. (34).Briefly,500to2,000 cpmof['251]DNAandvaryingamountsofRNAwere hybridized in0.4 Mphosphatebuffer(PB)-0.05%so- diumdodecylsulfate-2 mMethylenediaminetetraace- tic acidat68°Cfor18 to48h, i.e.,at aCot10 ormore timeshigherthan the DNACotI/2. Thehybridization mixture was flushed into 2 ml of0.02 M PB and divided into twoparts: 1 mlwasadjusted to0.12 M PB-0.05% sodiumdodecylsulfate andanalyzedover hydroxyapatite (measure of total duplex); the other wasincubated with RNaseA(50ug/ml)at37°Cfor1 h and then adjusted to 0.12 M PB-0.05% sodium dodecylsulfate andpassedoverhydroxyapatite.This RNase treatmenthydrolyzesRNAengagedin RNA- DNAhybridsaswellasRNA in thefreestate. The difference inhybridizationvalue found with and with- outRNase is indicative of thequantityofauthentic RNA-DNAhybrids,asopposedtoDNA-DNAhybrids whichmight form if the RNApreparationshad been contaminated with genomic DNA. With all of the RNA preparations, however, RNase treatment re- duced hybridization values to the low background levels,demonstratingthat there islittle contamination of thesingle-stranded probe byDNAof theopposite polarity. We conclude that all ofour RNA prepara- tionswerefree frommeasurableDNAcontamination.

Hydroxyapatite columns were loaded and eluted at 60°C (single-stranded material) and then shifted to 95°C (double-stranded material). Fractions (4 ml) were scintillation counted directly by addition of Aquasol (NewEnglandNuclearCorp.).

The quantitative treatment of titration data has beenpresented bySchelleretal.(34).We definedTi Toasthe fraction of reactiveprobehybridizedwhen amixtureofRNA/proberatioequaltoRisincubated totermination. We defined1/aasthe fractionofthe RNAcomplementaryto theprobe.At low Rvalues, theplot ofTI ToversusR is linear:

T R

=- ~~~~~~(1)

To a

Tocarryoutsuchtitrations, threeconstants mustbe known: thelengthof theDNAprobe,theconcentra- tion ofRNA, and the concentration of DNA in the

reactionmixtures. The molecular length of the probe in base pairs is known directly from the DNA se- quence, whereas the RNA concentrations are deter- minedspectrophotometrically. Concentrations of the radioiodinated probe DNAs could only bedetermined by careful measurement of probe specific activities as discussed in the next section.

Specific activity determination. We used two different methods to determine the critical specific activity values. The first was the kinetic method de- scribed byScheller et al. (34). The specific activities of thetwo5'spacerprobes (the coding and noncoding strands) were determined by reannealing an equal radiochemical amount of each strand and determining the kinetics of the reaction, using Si nuclease. The half-time of the reaction allowed us to calculate the amountof DNA present, knowing the valueofCAtI/2.

This latter quantity isdirectly related to the known sequencecomplexityof the DNA and was calculated by using the known relationship between complexity and the reassociation kinetics of a standard reaction (in this case, the reassociation of4)X174DNA sheared to 300basepairs) (11).

Thespecific activity of the pGC1-derived '25Iprobe wasdetermined by titration instead of by hybridiza- tionkinetics. One of theseparatedstrands, labeled L, was incubated withincreasing amounts of cold, EcoRI- BamHI double-digested pGC1 DNA. Therefore, the excesscold component of the reactionwas amixture of linearplasmidand insert. The titration curve was expressed asTI Toversusthe ratio of micrograms of RNA/counts per minute ofprobe, R' (Fig. 3). There- fore, R' =RIS,where R is themassratio ofdriver/

probe,asinequation1.

1.01

0.6

0.4

0.2

0

0 1 2 3 4 5 86

R'X106

FIG. 3. Specificactivityofthe3'probe. Thespecific activity ofthe3'probe^wasdeterminedbytitrationof unlabeled,BamHI-EcoRIdouble-digestedpGCIwith single-stranded iodinatedDNA, asexplainedin the text.

MOL. CELL. BIOL.

HISTONE GENE EXPRESSION 665 If we consider theinitial,linearportionof this curve,

wecanagainapply equation 1:

T R R'xS R'

To a a a'

where 1/a' = 1/a x S is the initial slope of the titration curve when the abscissa represents micro- gramsof cold nucleic acidovercountsper minute of probe. We estimate that this initialslope 1/a'=2.86 x 105 cpm/,ug. On the other hand, 1/ais the mass fraction of thereactingDNA(onestrand of theinsert) in the mixture. The nonreacting components are therefore the other strand of the insert and both pBR322strands. Wecomputed 1/a=

165

165+(2x4,000) =2.02x Finally, the specific activity was calculated as:

S = 1/a'= 1.4x107cpm/Lg 1/a

Determination ofpolarity of 5' spacerprobe.

To determine the polarity of the U and L probes relative to the messenger coding strand, we used a HhaI-EcoRI fragment which encompasses the 5' spacers and the H2B codingsequence (Fig. 1).This fragment had been previouslystrandseparated, and the relationship of thesingle strandsto the mRNA was determined. We used that strand ofthe HhaI- EcoRI fragment which we knew to be identical in polaritytothe mRNA and reacted it with the U and Lprobes. The fraction ofprobe taken intohybridat terminalCotwasdetermined by theSinuclease assay performedasdescribedbyLevetal.(21).

Determination of polarity of 3' spacer probe.

Thepolarityof the U and L strandsrelative to the codingstrandwasdeterminedby usingaDNA frag- mentencompassing both the 3' spacer andacoding sequence,as wasdone with the 5' spacer. In this case, however,wehadaclone derived frompSpll7in which theeucaryoticpiecehas beenrecloned into the vector M13mp2 (13), a single-stranded phage. The phage used contained the seaurchin sequence identical to the H3 and H2A messenger sequences aswellasthe sequenceinpGC1.Since the coding strand is thesame forall fivemRNA's, thepGC1strandhybridizingto thissingle-strandedDNAprobes for spacer transcripts of thesamepolarityasthemRNA.Wehybridizedthe U and L strand of thepGCl probetothis DNA under the usual conditions andpassed both hybridizations overhydroxyapatite.Theresult shows that Lhybrid- izedtothe cloned single-stranded DNA, whereas U didnot.

Preparation ofembryo RNA Sea urchin em- bryosweregrownaspreviouslydescribed (15). Total RNA wasprepared, usingtheguanidine hydrochloride method ofStrohman et al. (36) as previously described (6). In addition, residual contamination of the RNA byDNAwas eliminated by treating with pancreatic deoxyribonuclease, proteinase K, and phenol-chloro- form,asdescribedby Woldetal. (41).For this purpose, RNase-freedeoxyribonuclease(Worthington Diagnos- tics)wasfurtherpurified by affinity chromatography on agarose 5'-(p-aminophenylphosphoryl)uridine- 2'(3')-phosphate (Miles-Yeda) (27).

RESULTS

Detection of histone-specific RNAs on Northern blots. We separatedhistone-specific RNAsequencesbyelectrophoresis to detectany high-molecular-weight species differinginlength from histone mRNA's. Weprepared nuclearas well as polysomal RNA from 7-h-old embryos andseparated thembyelectrophoresis onden- aturing agarose gels (28). Afterwe blotted the gel onto DBM-paper (1), we searched for his- tone-specific RNAsbyhybridizingthe blot with histonegeneprobeslabeledbynicktranslation.

Figures 4A to D show the autoradiography of such blots hybridized to pSplO2 and pSpll7, which contain theH4, Hi,and H2Bgenesand the H3and H2Agenes,respectively, and which together encompass an entire histone gene re- peat unit. The only reproducible and specific RNA bands apparent onthesedeliberatelyov- erexposed autoradiographsaremessengersized, both innuclear and inpolysomal RNA. No high- molecular-weight histone-specificRNAwas de- tected.

Very small amounts ofapolycistronic tran- script, whichmustalsoincludespacersequences, mightberevealed with greatersensitivity with a spacer-specific probe. We therefore probed similar RNA blots with nick-translated DNA from pGCl, a subclone ofpSpll7 specific for partof thespacerbetween H2B andH3 (see Fig.

1). No high-molecular-weight RNA was detect- able (Fig. 4, lanes E and F). There is faint

A B

C

D E F

U.

FIG. 4. Northern blot hybridization ofhistone-spe- cific RNAs. Nuclear and polysomal RNAfromearly embryoswasseparated on a denaturing gel, blotted onto DBM-paper and hybridized with histone-spe- cific, 32P-labeledprobes. (A, C, and E) Polysomal RNA. (B, D, and F) Nuclear RNA. The probe for A andB waspSplO2;for Cand D, it was pSpll 7, and forE andF, itwasthespacer-specific probepGC1.

VOL. 1,1981

666 MAURON ET AL.

hybridization at a position corresponding to a histone mRNA. This signal is moreintense in polysomal than in nuclear RNA,and webelieve thatit is duetocross-hybridization betweenthe probe and H4 mRNA since theyhave a short base pair sequence in common. We conclude that thequalitativeRNAblot experimentsdem- onstrate nospecific, high-molecular-weight his- tone gene transcripts. Nevertheless, to assess further the validity of this negative result, we turned to the quantitative approach described below.

Strategyforthe measurement of histone spacer transcripts. Our overall strategy was todetermine the absolute concentration of RNA molecules per cell or per embryo capable of forming RNA-DNA hybrids with histone gene spacerDNAfragments.Themethodweusedfor titrating single-stranded DNA probes against RNA-containingfractionswastheprobeexcess titration technique described by Schelleret al.

(34). Briefly, the fraction ofprobe annealed at hybridization termination was determined for increasing RNA concentrations but undercon- ditions ofprobeexcessrelativetothehybridizing RNAsequences.Theplotof thisfractionversus theRNA/DNA ratio is linear if onlyonestrand isrepresented in the RNA and isinitially linear, evenincasesofsymmetrical representation (34).

From the slope, the concentration ofreacting RNAcanbecomputeddirectly.

Hybridizationof thespacerprobestosea urchin RNA. Weprepared^totalcellular RNAs frommatureeggs andfromembryosharvested at various developmental stages. We then hy- bridized these RNAstotheseparatedstrands of the 5' and 3'spacer probes. First, titrations of RNA from 11-h-old embryos were carried out with eachprobe.The 5'spacerprobeshybridized toRNA presentin 11-h-oldembryos,andtran-

scripts of both polarities were found (Fig. 5A).

Since the titration curves are linear over the range ofDNA/RNAratiosemployed,wecould use a singleratio in this range toevaluate the amounts ofhybridizingtranscripts inthe other RNA samples takenatvarious stages in devel- opment.

Single hybridization mixtures containing RNA from various developmental stages were

prepared at a relatively constant total DNA/

RNA ratio of 106. These mixtures were incu- batedinhybridizationconditions,and thedegree ofhybridization wasmeasuredbyhydroxyapa- tite chromatographywiththeappropriate con-

trols,asexplainedabove.

The results of these titration measurements arepresentedinTable 1andFig.6.Thevalues inTable 1for numbers ofmoleculesper eggor

0.4 I A

0.2

0

0 1 2 3 4 5 6 7 8

RNA/DNA x10-5

0.6 B

0.4-

0.2

0

0 1 2 3 '4 5 6 7 8

RNAIDNAx1-

FIG. 5. Titration of spacersequences in 11-h-old embryo RNA. (A) Total RNAfrom 11-h-oldembryo washybridized to both strands of 5'spacer-specific, iodinated DNA (pSp2b-1). The polarity of theprobes relative to the coding strand has beendetermined as described in the text. Symbols: 0, titration with L strand (coding strand); A, titration with Ustrand;

0, L strand and A, U strand hybridized at 60°C (lowerstringency). (B) Similar experiments were done usingthe 3' spacer probes(pGC1).Symbols:0, titra- tion with L strand (coding strand); A,titration with Ustrand.

embryowerecalculatedasindicated in the foot- notes toTable 1. These datademonstrate the presence inembryos of RNA molecules ableto hybridize to the spacer sequence flanking the H2B histone gene on the 5' side. Moreover, RNAs ofbothpolaritiescould bedetected.Each of thesetwoRNAclasses exhibitedadistinctive timecourseofaccumulation. For thepurposeof comparison,Fig.6also shows the timecourseof accumulation of H2B mRNAsequences.These data were obtained byhybridizing these same RNA samples to a radiolabeled probe specific for the H2B coding region (unpublished data).

Thequantitiesofthe5'spacer RNAsarelow,in theorderof1/2,000theamountsof H2BmRNA presentatthesamedevelopmentalstage.

Theresultsregardingthe 3'spacerprobesare MOL. CELL. BIOL.

HISTONE GENE EXPRESSION 667 TABLE 1. Quantitation of 5'spacer-specific RNA at various stages ofdevelopment

ReactingRNA Timeofem-

~RNA/DNA

^{mas TT}

bryodevelop- Strand ratio'm Fraction oftotal Ribonucleotides No. of mole- ment(h)

(rbonucleotides

(pg)per egg or em- culesper egg

bryod orembryo Mature egg U 1.64X 106 0.004 <10-" <3x 10" <100

L 1.89x106 0 <10-8 <3x10" <100

5.7 U 4.00x 105 0.070 1.75 x10-7 5.08x 10`0 1,780

L 3.74 x105 0.010 2.67x 10-8 7.74 x10-" 271

9 U 8.76 x105 0.505 5.76 x 10-7 1.67 x10-9 5,840

L 8.26x 105 0.108 1.31 x10-7 3.80x 10-10 1,330

11 U 1.42x 105 0.272 1.92x 10-7 5.54x 10`s 1,940

L 1.47x 106 0.540 3.67x 10-7 1.06x 10-9 3,710

13 U 2.10 x106 0.129 6.14x 10-8 1.78 x1010 622

L 3.03 x 106 0.544 1.80 x10-7 5.22 x10`0 1,830

16 U 1.53 x106 0.055 3.59x 108 1.04 x1010 364

L 1.66 x106 0.197 1.19x10-7 3.45x 10-'O 1,210

22 U 1.22 x 106 0.119 9.75 x 10-8 2.83x1010 990

L 1.20 x106 0.161 1.34 x10-7 3.89 x10`0 1,360

a RNAquantity obtained from determination of absorbanceat260nm.DNAquantitywascomputedfrom input counts perminute,andprobe specific activitywasmeasuredkineticallyasexplainedin thetext.

b This measures authentic RNA-DNA hybrid as determined by the low-salt RNase test. The data are corrected forreactivityandself-annealing; therefore,the valuerangeis0to1.

'Calculated as 1/a = (T/To)/R orfrom theslope of thecomplete titration curveof RNA from 11-h-old embryos (Fig.5A).

dAssumes2.9 x 10-3utg of total RNA per embryo throughout early development (L. Washburn, Ph.D.

dissertation,UniversityofCalifornia, Berkeley, 1971).

somewhat different: spacer transcription was

6 _ 12 asymmetric; i.e., only one of the two probe

strands(correspondingtothe strand fromwhich

A/

^{\ the mRNA was}transcribed) formed detectable

5 -lO°R hybrids. Typical titration data derived with

cc l a RNA extracted with 11-h-old embryos are

^' 4 _ / B - 8 o shownin Fig. 5B. The initial slope of thetitra-

o tion with the L strand corresponds to about 3

cr / 2\ \

Xx 104

molecules^per

embryo.

This is^oneorderof

co

co~~~I

~ ~ -"-'th ocntainf

w3 /u

agni_6tude

6 greater thantheconcentrationofthe

Ki / , 5' spacerl

\|

transcriptsatthesamedevelopmental

z stage.Wealso

investigated

thepresenceofRNA

D2 - ! tt 4 D

homologous

to the 3' spacer at other

develop-

o / 01\ \ \ .* mental stages. Asmallamount of such RNA was

E / o7 \ > 2 detectable in egg RNA, whereas it seemed to

U-

^{2 have}

entirely disappeared by

the 22-h stage (datanotshown).

O.

, ^{L L} Spacer RNAs are not

authentic

^tran-

o 5.7 9 11 13 16 19 22 scripts.Since S.purpuratushistonegenes are TIMEPOST-FERTILIZATION (hours) asymmetrically transcribed (8), only RNA tran- FIG. 6. Timecourse ofspacer-specific RNA accu- scripts homologous to one or the other of the mulation. Dataare taken from Table1. Symbols:0, two spacerstrandscanbeconsideredasevidence RNAspecifictoL strand of 5' spacer probepSp2b-1;

to space str

^{n e}

coNsideredra Ince

A, RNA specific to Ustrand;, accumulation curve for apossible histone mRNAprecursor. Inthe of H2B mRNA, determined by titration hybridiza- case of both5' and 3' spacer probes, the RNAs tions with an H2B coding region probe. Note the complementarytothe Lstrandsareof thecor- differentscalesfor the two data sets. rectpolarity;i.e., they aretranscribed from the VOL. 1,1981

668 MAURON ET AL.

same strand as the mRNA's. To investigate whether these RNAs are indeed transcribed from this regionofthehistone genes, we deter- mined the melting profileofthe hybrids. RNA- DNAhybridswereformedat68°C in 0.4 MPB, using RNA from 11-h-old embryos and either theUorLstrand of

"2'I-labeled

spacer DNA. As acontrol for themelts, these probes weresimi- larly incubated withan excessof cold, homolo- gous DNA.The melting profileof these prepa- rationswasdeterminedbyhydroxyapatitechro- matography in0.12MPB,starting at600C(Fig.

7).

Themeltingcurvesshowthatallof the RNA- DNA hybrids melted over a broad range of temperaturesandattemperaturessignificantly lower than those of the reassociated DNA. The differences inTm's for the 5'spacerprobeswere 150C for both the U and the L strands. In addition, the RNA-DNA melting profile sug- gestssignificantheterogeneityof thehybrids,as evidenced by the shallower slopes. Although there is anintrinsic difference between the Tm

ui c 1.0 4

w 0.9

A.

C,,I

4 0.7

a

0.6 z

_ 0.5

LU

Z 0.4 0

-i LU

Z 0.2

z o 0.1

N

04

o 0.

60 65 70 75 80

T°C

FIG. 7. Thermalanalysis ofhj DNAhybridsprepared usingtot RNAand 5'spacer-specific['2I

on hydroxyapatite at 60°C in 0.

melting profilewasdetermined.

sociated[1251I/DNA ^{was run}on

apatite column. Symbols: A, RN hybrid; ^A, simultaneous DNA-l

versusL strandhybrid;0,corresg melt. Thereassociatedprobe DN, atureconsistentwith itsguaninej and intactfragment length.

of RNA-DNAand DNA-DNA melting, this dif- ference is small in aqueous solutions. Further- more, since the DNA sequences of the histone genesand spacers within the repeating cluster, and fromindividual to individual, are virtually identicalto oneanother(unpublisheddata),the hybridizing RNAscannotbe homologous to any member of the principal histone gene repeat family.We conclude that the bulk of the hybrid- izing RNAis not an authentic transcript from the histonegenespacerDNA.Assuming that we mightnotdetectperfectlymatched hybrids that amountedtoless than 10%ofthe totalhybrid- izingmaterial,themaximumnumber of authen- tic 5' spacertranscripts ofthe correct polarity could be 370 molecules per 11-h-old embryo, correspondingtoanaverageof threemolecules percell. Duplexes formed with the3' Lstrand were even less stable, with a Tm difference of 230C. We estimate that no more than 5% of authentictranscript would have escaped detec- tioninthismeltingcurve. Wecanthereforeset the maximum number of3' spacer transcripts presentin 11-h-old embryosat 1,500molecules perembryo.

DISCUSSION

A polycistronic primary transcript is veryunlikely.Ourexperimentswereaimedat establishing the existence ofspacer transcripts of the

early

histone genes, a necessary conse-

quenceofpolycistronic transcription.Usingtwo

independent approaches, hybridization

^toRNA blots andtitration-hybridization,wewereunable tofindanysuchtranscripts.Obviously,itwould be impossible to argue from these results that

I

^suchtranscriptsneverexist, since thisnegative findingisonlyascertain asthe finite detection limit ofourexperiments. However,thepurpose of this discussion is to argue a quite different conclusion, namely,thatnophysiologically sig- nificant transcription of spacer sequences of earlyhistonegenestakesplaceduringseaurchin embryogenesis, atleastthroughthe first22hof development. We shall first examine the data pertainingtothe 5'spacer of the H2B gene and

85 90 95 100 show thatthey constitute compelling evidence thataspacertranscriptcannotbe thephysiolog- ybridfidelity.RNA- ically

obligatory transcript

of most histone lal, 11-h-old embryo genes.

r]DNAwereloaded Letusconsider the kinetics of the H2B gene .12 MPB, and the expression process: ifa colinear 5'spacer-gene Ineachcase, reas- transcript (such as a polycistronic transcript)

aparallelhydroxy- werepartofthe

obligatory

pathwayofH2B gene IA versus Ustrand

)NA

melt; 0, RNA expressioninearly development,such molecules pondingDNA-DNA would have to be present in amounts that are Ameltsat atemper- compatiblewith the rate of H2B synthesis dur-

iluscytosinecontent ingthat time. More

specifically,

thesmallerthe poolsize of suchatranscript,thefaster it would MOL. CELL. BIOL.

HISTONE GENE EXPRESSION 669 haveto turn overtoprovideagivenrateof H2B

mRNAproduction; it happens thatwe arein a positiontoknowjusthow fast this hastobe.

Themaximumamountof5' spacertranscript present in 11-h-old embryos and which could have gone undetected in our experiments is equal to 370molecules perembryo. Let usas- sume that all of these 370molecules represent suchspacer-containingprecursorsand thatthey are converted to functional H2B mRNA with 100%efficiency.Therefore,370molecules is the maximumpoolsize of theH2BmRNAprecursor at apointindevelopmentatwhichsynthesisof histone mRNA ismaximal (26). These authors have determined that the rate ofsynthesis of H2A+H2B +H3+H4messengerRNAspeaks atapproximately83fg/minperembryoat10h ofdevelopment. Wecan assumethatone-quar- ter of this value is H2B mRNA synthesis, i.e., 20.8 fg/min per embryo, or 75,000 molecules.

Thismustalso be the minimumrateofproduc- tion of H2Bprimary transcripts, ming100%

efficientprecursor-productconversion. The fol- lowing equation relates thepoolsize ofaputa- tive RNA precursor C,itsrateofsynthesis,k.,

anditsrateof first orderdecay, kd (10):

dC-d=k.-kdC dt

Thesteady-statepoolsizeCo is obtained for d C/

dt=0:

Co= k

kd

Assuming that the precursor poolhas reached itssteadystatein the 11-h-oldembryo, we can calculate thehalf-life of suchaprecursor.

kd = 75,000 molecules CO

xmin-' per 370molecules=203min-1 tl,2= - =3.4x 10-3 min = 0.2 s

The rate of RNA polymerizationin seaurchin embryoshas been showntobe5to8nucleotides persecond (2). Thus, our calculated precursor half-life of0.2 scorrespondstothe time it takes topolymerize only 1to 2nucleotides. Thisun- realistically low half-life simply shows that a precursorpool size of370moleculesperembryo isutterlyinadequate tosustain thehigh rate of H2B mRNA synthesis which is

occuring

in these embryos. Therefore, the true precursor mustbe ofanother type which does not include the 5'spacer sequence.

Onecan, of course, argue that thehalf-lifeof the putativeH2B precursorpool is longer than we calculate because the pool has not yet reached asteady-state level. But noreasonable

correction forsteady-state levelscanstretch the half-lifetovalues compatible with the existence ofapolycistronic transcript.

Asimilar calculation can be made for the 3' spacer sequences. We haveseenthatthe maxi- mumnumberof transcripts from the pGCl re- gion is1,500moleculesperembryoatthatsame 11-hstage.Thehalf-life calculatedasdescribed above isequalto 0.8 s.

Inthecaseof the 3'spacer,independentcon- firmation isavailable that itstranscripts, if there areany,haveanexceedinglyshorthalf-life. Ex- periments have been done in this laboratory (Childs et al., unpublished data) in which the nuclear RNA ofembryos labeled for5min was hybridized topGC1 DNA on nitrocellulose fil- tersand the percentage of RNase-resistantra- dioactivity was determined. In these experi- ments, plasmids pSp2 and pRC9, the Hi, H4, and H2B genes and the H4 gene, respectively, were usedas hybridization controls. No values significantly above background were obtained, and the maximum proportion of total nuclear RNA synthesis related to pGC1-specific se- quences wasdetermined tobe0.002%.Again, if we assumethat thismaximumvaluerepresents sequencesthatarepartofaprimarytranscript, then thehalf-life of the transcriptmustbe con- siderably shorter than1min.

We have demonstrated the absence of any significant amountsoftranscripts ofspacer se- quences present oneither side of the H2B his- tone gene. The most straightforward interpre- tation of this evidence is that transcription of the H2B gene occurs as a monocistronic unit.

The only model ofpolycistronic transcription compatible withourdata would beoneinvolving atightlycoupled transcription-processing com- plex that would degrade newly synthesized stretches ofspacer transcripts before they ob- tainedasignificant length, i.e.,morethanafew nucleotides.Suchamodel would beatvariance with what is known about other eucaryotic genes, in which primary transcripts are stable enough to be completed and then processed endonucleolytically.

There are additional arguments against a polycistronic mechanism ofhistone gene tran- scription. The 5' end of the H2B mRNA has been mapped and located in relation to a se- quence, PyCATTCPu, found next to it in the H2B gene (22).Thisisthe"consensus" form of a sequence which is found 50 to 70 bases up- streamofallof the otherhistone coding regions and nowhereelse inthehistone gene sequences (4, 37). Mapping experiments have pinpointed the 5' ends of the histonemRNA's in or near this sequence. Furthermore, approximately 20 nucleotidesupstream from thePyCATTCPuse- VOL. 1,1981