Groups of adenovirus type 2 mRNA's derived from a large primary transcript: probable nuclear origin and possible common 3' ends.

JOURNAL OFVIROLOGY, Mar.1978, p. 811-823 0022-538X/78/0025-0811$02.00/0

Copyright X 1978 AmericanSocietyforMicrobiology

Vol.25; No. 3

PrintedinU.S.A.

Groups of Adenovirus Type 2 mRNA's Derived from

a

Large

Primary

Transcript: Probable Nuclear

Origin

and

Possible

Common

3' Ends

J. R.NEVINSANDJ.E. DARNELL*

TheRockefeUerUniversity,NewYork, New York 10021

Received forpublication 27 September 1977

Late in adenovirus type 2infection, a number of mRNA's apparently arise by

processing

a

large

nuclear

transcript

thatrepresents the

right-hand

85% of the

genome

(smzed

in Evans et al., Cell 12:733-739, 1977). Hybridization of labeled late mRNA to a series of DNA restriction fragments representing the right-hand 70% of the genome demonstrates atleast 12 discrete mRNA's that appeartofall into five groups, each

possibly

containing acommon3'terminus. The processing necessaryto generate these mRNA's apparentlyoccurs in the

nucleus.

All the mRNA's appear to contain a sequence of -100

nucleotides

complementary

to a

fragment

with coordinates 25.5-27.9. This

fragment

contains

one of the

regions

found

by Berget

et al.

(Proc.

Natl. Acad. Sci. U.S.A. 74:3171-3175, 1977), ChowetaL

(Cell

12:1-8, 1977), and

Klessig (Cell

12:9-22,

1977)

tobe

"spliced"

ontothe 5' end of late adenovirustype 2 mRNA's.Because

the sequencestobe

spliced

exist

only

onceper

large transcript,

any of the

mRNA-specific regions might only

be

preserved

fromasmall fraction of thetranscripts.

Measurement of the

tansport efficiency

of

regions

ofthe

large

nuclear

transcript,

in

fact,

shows that

only

15 to25%ofany

particular region

is

transported

tothe

cytoplasm.

The overall conclusion of these

experiments

is that the

large late

nuclear transcript

canbe

processed

in the nucleusto

yield

anyoneof many (-12)

mRNA's;

the

unused

portions

of the

primary

tanscript

then accumulate in the

nucleus

or are

destroyed.

Information about the

synthesis

of virus-spe-cific mRNA inadenovirus-infected cells has

pro-gressed

very

rapidly

sincerestriction

endonucle-ases became availabletodivide the virus DNA

intoa

reproducible,

orderedsetof

fragments

(29,

31). The

approximate

sitesonthegenome that

correspond

to

specific early

and late mRNA's

(32, 34), the determination of which strand of the DNA is

transcribed,

and the direction of

transcription

of

early

andlate mRNA'swereall

determined

by hybridization

of unlabeled RNA

to

fragmented

labeled DNA

(for

a

review,

see

reference

9).

More

recently

(6, 26), a

detailed

mapof

mRNA

sites in

specific regions

of ade-novirus type 2

(Ad2)

DNA was obtained

by

observing

"R

loop"

formation-R

loops

(40) are

RNA-DNA

hybrids

in which onestrand of the DNA duplex is displaced by the mRNA

mole-cule-ini

specific

regionsofthegenome.

In

paralel

with themapping ofmRNAsites

on the Ad2 genome, biochemical studies of mRNA formation revealed thatAd2mRNA,in common with

celular

rmRNA,

requires

a num-ber ofnuclear

post-transcriptional

modifications of the mRNA. These include the addition of

polyadenylic

acid

[poly(A)]

to the 3' terminus

(32) and

specific methylations

(28, 37);

internal

adenylate

residues are

methylated

togive

m6A

approximately

onceevery200A's, in addition to

the formation ofa5'-terminal methylated struc-ture,

termed

a

"cap"

(35). The

primary

RNA transcripts that undergo this nuclear processing are not

simply

the same

length

as the mRNA itself. For

example,

late in infection a single large

transciptional

unit from -16 to 100 on the

rightward-reading

strand is

copied

into a

pri-mary

trnscript

-28,000 nucleotides long (the

convention

[9]

for

the description

of the Ad2

genome divides into 100 units

the

double-stranded

DNA,

each -350

nucleotide

pairs or

0.35 kilobase

[kb] long.

The

rightward

tran-scribedrstrand has0asitsleft end and100at its

right,

with thereverse true forthe leftward strand 1. The coordinates ofparticular regions of the genome for either restriction

fagments

or mRNA coordinatesisgiven by two numbers, the percentages from leftto

right;

for

example,

the 36.7-40.5

fragment

orthe 50-62 mRNA) (2, 8, 14, 43). Over 95% of the synthesis of RNA from

the

regions

ofthe genome

encoding

the major

811

on November 10, 2019 by guest

http://jvi.asm.org/

812

late proteins (8, 9, 21), themajorcore, penton, hexon, "1OOK," and fiber polypeptides,occurs as part of thislong RNA molecule. Moreover, la-beling of mRNA that originates in this region is equally sensitive to UV irradiation as the for-mation of the large nuclear transcript itself, thereby identifying the large transcript as the obligatoryprecursor to the mRNA lying within its boundaries (13, 14). Thus, the processing procedureappears toincludesite-specific cleav-age of alarge transcript as well as additionof capandpoly(A) tothe mRNA.

In studying the origin of the 5' ends of late mRNA'sinthe 50-100region, Berget etal. (4), Chowet al. (5), Gelinas and Roberts (12), and Klessig (19) have discovered a mostsurprising fact. The 5'-terminal 150-200 nucleotides ofa number of the late Ad2 mRNA's do notcome fromregions of the Ad2 genome contiguousto the remainder of the mRNA but from three distant sites. Becausethe

large

nuclear Ad2 pri-marytranscript appearsto be the mRNA pre-cursor for mRNA molecules from this

region,

they conclude that 50 to 100 nucleotides from each of three

sites,

16.7,

19.7,

and

26.0,

are

cleaved out

separately

and

ligated

onto the 5' ends of eachmRNA molecule derivedfromthe large transcript. This astounding development haspromptedus tomap asaccuratelyaspossible the number and relative amounts of various mRNAmolecules

deriving

fromthe

large

tran-script andtodetermine whether the formation ofall of the relevant mRNA's appears tobe a

nuclear event. To

satisfy

these goals, mRNA mapping by hybridization of labeled RNA is necessarybecause ofthe

difficulty

of

obtaining

quantitative

andkinetic dataon mRNA

distri-bution

by using

electron

microscopic

techniques.

Our conclusions are that at least 12 different mRNA'scanbe derivedfromnuclear

processing

of the16to100

transcript.

TheseRNAsfall into groups where two to four mRNA's may share the same 3' terminus. The map

positions

as-signed by

hybridizing

labeled molecules agree

well with the electron

microscopic

mapping (5,

6, 26). A measurement of the conservation of

various parts of the

large

nuclear transcript is described that suggests that each

long

RNA transcript

possibly gives

riseto

only

one

mRNA,

presumably

because each transcript contains

onlyone setof 5'

spliced

sequencesnecessaryto

formthemRNA.

MATERLALS AND METHODS

Celland virusgrowthaswellasthepurificationof

adenovirushave beendescribedpreviously (42).

Labeling

and extraction of RNA. HeLa cells

wereinfected at multiplicities of2,000 particles per

cell. At theappropriate time afterinfection,the

in-fectedcelLswere concentratedto 10 cells perml in

fresh medium, and [3H]uridine was added (10 to 20

itCi/ml,

20

mCi/pmol).

Afterlabeling, thecellswere

chilled and collected by centrifugation. Cytoplasmic

andnuclearextracts werepreparedasdescribed

pre-viously.The extraction of RNAaswellasthe

prepa-ration ofpoly(A)-containingRNAby polyuridylic acid

Sepharose chromatography has been described (27).

Single-stranded RNAwasprecipitated from2MLiCl

overnightat4°C.

Electrophoresis of RNA.3H-labeled RNA

sam-ples (including 32P -28S and -18S rRNA markers) were

dissolved in ETS (0.01 M Tris-0.01 M EDTA-0.2%

sodium dodecyl sulfate [pH 7.4]) and made 50%

(vol/vol) formamide. After heatingto65°C for2mi,

bromophenolblue andglycerolwereadded, and the

samplewasappliedto a 2.7%acrylamide-0.14%

bis-acrylamide gel (0.6 by 10 cm) prepared in Loening

buffer(23).Electrophoresiswasat6mA/gelfor3 to

4 h. Gels were sliced into 2-mm fractions, and the

RNAwaseluted in 0.5 ml of TESS buffer [0.01 M

N-tris(hydroxymethyl)methyl-2-aminomethane-sulfonic

acid(pH7.4)-0.3 M NaCl-0.01 M EDTA-0.2% sodium

dodecyl sulfate] at 650C overnight. RNA was

frag-mented with alkaliasdescribed(16).

Electrophoresis in 98%formamide-containing gels

wasperformedasdescribedby Duesberg and Vogt (7).

Hybridizations. Hybridizations of RNA to

filter-bound Ad2 DNA fragmentswere carried out as

de-scribedpreviously (2).The RNAfrom each 107 cells

wasexposedtoDNAfragments equivalentto 1,ugof

total Ad2 DNA. Excess of DNAwasprovenby

rehy-bridizingsupernatantRNAtofreshfiltersrevealingat least 80% of the hybridizable RNA to be removed in the initial exposure.

RESULTS

Experimental design. LateinAd2 infection, the majority ofAd2-specific cytoplasmic RNA ispoly(A) terminated [poly(A)+;22,33]. More-over,almost all

(>90%)

of thelabeled

poly(A)+

RNAthat reaches thecytoplasmis Ad2specific (22, 32). Thus, by collecting the labeled poly(A)+ RNA lateininfection, the total virus-specific mRNAcanbeexamined without further purification. Because the firststepinthe design of thepresent work was to separateAd2-specific molecules on the basis ofsize, the electropho-reticmigrationofthetotal poly(A)+ RNA late in infectionwascompared under two conditions: (i) electrophoresis of denatured RNA in 98% formamide and (ii) denaturation followed by electrophoresis through gels not containing formamide. With both procedures (Fig. 1), dis-cretesize classesoflabeledpoly(A)+RNA mi-grating between 18S and 28S rRNA were ob-served aspreviously reported for Ad2-infected cells (22, 30, 39).The approximate distribution of labeled RNA compared with 28S and 18S rRNA markers wassimilar, but the resolution wasconsiderablybetter in thegelswithout the formamide. This technique was therefore adoptedforthepresent work.

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

PROCESSING OF Ad2 mRNA's 813

The experiments to be described were

de-signedtomapthe variousmRNA'sindifferent

regions of the genome by first separating the mRNA according to size, followed by hybridi-zation of the labeled RNA toasufficientarray

of Ad2 DNA restrictionfragments (in DNA

ex-cess)sothatmostmRNAwouldhybridizeacross

a boundary between two fragments (Table 1). The percentage of labeled RNA ofaspecific size that hybridized to each DNAfragment would then allowassignment of theposition of the 5' and 3' termini of the mRNA exclusive of

post-transcriptionally addedsequences.Insomecases anmRNAspecies mightspan anentire fragment

andhybridize tobothright-hand and left-hand neighboring fragments. Such asituationwould

offer a measure ofthe length of mRNA

inde-pendent of its electrophoretic migration because the size of thefragmenttotallyspanned by the mRNA would be known from the DNA restric-tion fragment maps. The accuracy ofthe map

assignmentsmade with thesetechniques would

dependonthe errorinassaying the RNA:DNA

hybridization,the sizedetermination ofRNAby gel analysis, and theaccuracyoftherestriction fragment map ofthe Ad2 genome. An overall

accuracyof atleast1 to2%onthephysicalmap

would beexpected for the relative positioning of the variousmRNA's, especially when reviewed in light of the positioning of termini reported

A

10

5-10 -B

a-0 10 20 30

SLICE

FIG. 1. Electropherograms of

13HJuridine-labeled

late Ad2 RNA(0) together with 32P-rRNA (0) in98% formamide-containing polyacrylamide gel (A) and acrylamide gel preparedinaqueousbuffer (B). Elec-trophoresiswasfrom lefttoright.Aportion (20

td)

of each elutedgel slicewascounted in triton scintilla-tionfluid.

TABLE 1. DNA restriction fragmentsofAd2usedin

mapping mRNA locations

Map cooi

zyme(s)

used Name of

frag-natee" ment

25.5-27.9 Hpa I Hpa F

29.1-36.7 Sma I/BamI SmaB/BamD

36.740.5 Sma I Sma I

40.5-52.6 Sma I Sma D

40.5-42.7 SmaI/Kpn SmaD/Kpn H

42.7-47.4 Kpn Kpn I

47.4-52.6 SmaI/Kpn SmaD/KpnD

52.6-56.9 Sma I Sma H

56.9-58.5 Sma I/EcoRI SmaA/EcoA

58.5-70.7 EcoRI Eco B

70.7-75.9 EcoRI EcoK,

75.943.4 EcoRI Eco D

83.4-89.7 EcoRI Eco E

89.7-100 EcoRI Eco C

aMap coordinates of Hpa I, Bam

I,

and

EcoRI

cleavage sitesaretakenfrom datacollectedby Marc

Zabeauand distributed at the 1976 Cold Spring

-Har-bor Workshop on DNA tumor viruses. The Sma I

cleavage sitesaretaken from Carel Mulder (personal

communication). TheKpncleavage sites within the

Sma Dfragment were determined from the size of the

fragmentsproduced aswellastheir sensitivity to Sal

Idigestion.

from microscopic techniques (5, 6). The tech-nique ofhybridizing labeled mRNA

molec,ules,

ofcourse, also allows a means ofstudying the metabolismofspecificmRNA's anda quantita-tionof therelativeamountsofeachmRNA that reach the

cytoplasm.

As

pointed

out

below,

nu-cleotide analysis of RNA and DNA is finally necessary tolocate

exactly

the

region

of DNA correspondingto aparticular RNA.

General results.The datafromseveralsuch RNA:DNAhybridizationsofelectrophoitically separatedRNA from cells labeled from15 to 18

h after infection areshown in Fig. 2, 3, and4.

The general results to note iniexam iing the data of

Fig.

2 and 3 are asfollows.

(i)

mRNA hybridizations were carriedout to many

differ-ent DNA

fragments

from 36.7 to89.7, and the

size distribution of RNA that

hybridized

toeach DNA fragment was distinctive; manydifferent specific mRNA'sfromapproximately lOSto28S (about -0.8 to 5 kb) were

clearly

present.

(ii)

The reproducibility of the electrophoretic pat-ternisquite good; for example,compareFig.2A and

3A,

which represent two different

RNA-preparations

hybridized

to twodifferent

prepa-rations of the DNA fragment 36.7-40.5. (iii) Some DNA

fragments

hybridize

two or three mRNA's, each of which is greater than the length ofthe

fragment. (Note

that above the mRNA was broken with alkali to chain sizes -0.2 to 0.3kb and thatall the DNA

fragments

exceeded0.7kb andwerepresent in excess;

thus,

any

single

DNAmolecule would

actually

hybrid-VOL. 25,1978

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.501.244.437.66.237.2] [image:3.501.75.184.376.589.2]

814 NEVINS AND DARNELL

X 30* a.

u

SLICE

FIG. 2. Electropherograms oflate

rH]uridine-la-beled Ad2 RNAhybridizedto(A) Sma I(36.7-40.5)

filtersandto (B) Sma D(40.5-52.6) filters.

Electro-phoresiswasfrom lefttoright, with thearrows

indi-catingthepositions of 28Sand 18SHeLa rRNA's.

ize lessthan its fulllength of RNA, anecessity

in filterhybridization experiments if the DNA is

not to be released from the filter [15].) For

example, Fig. 3B shows that fragment 42.7-47.4, which is1.64 kblong,hybridized three mRNA's, 27S, 23S, and 19S RNA, all larger than 18S

rRNA, which is 2.1 kb long. Clearly, several different discretemRNA's share atleastapart

of their sequences. This proved to be true in virtuallyeveryregion of the large transcript that gives risetomRNAmolecules.(iv) Eachoneof

theeight fragmentstotheleftof83.4hybridized large RNA migratingtogether with 28S rRNA,

which is5kblong. Because the DNA fragments used in thepresenthybridization workrepresent

about 18 to 20 kb intotallength,atleastthree

tofour suchlarge mRNA's mustexist. (v) The DNA fragment that hybridized the largest

amount of labeled -27Sto 28S mRNAwas in

the 52.6-56.9region where the hexon mRNA is knowntomap(9, 21). This finding is consistent with the earlierobservation that hexonprotein is themostprominently synthesized protein late in infection(45). (vi) Althoughtotalcytoplasmic poly(A)+RNAwasused for theanalyses of Fig.

2and3,poly(A)+ RNA from polysomes yielded

patterns after hybridization identical to those

shown inFig.3(datanotshown).

Sortingoutthe27S to28S mRNA's. The

large 278 to 28S mRNA's that hybridized to

virtually every DNA fragment (Fig. 2 and 3)

werepositionedasfollows.

(i)The 38-50 mRNA. A27SmRNA

hybrid-ized both to the 36.7-40.5 region and to the 40.5-52.6region.It didnot, however, hybridize

tothe29.1-36.7region (datanotshown), placing

its 5' end inthe 36.7-40.5 region. About three

10

20

C 474 52.G

20

D 52 6 56,9

80

40-2 SL

(.) E 707 75.9

15

-10

5-F 175.9 834

10

L-

~

5-G 83.4 897

40

--

20-0 10 20 30 40 SLICE

FIG. 3. Electropherogramsoflate

"Hiuridine-la-beled Ad2 RNAhybridizedto(A)Sma I(36.7-40.5),

(B) KpnI (42.7-47.4), (C)SmaD/Kpn D

(47.4-52.6),

(D) Sma H(52.6-56.9),(E)Eco F(70.7-75.9), (F)Eco

D(75.9-83.4),and(G)Eco E(83.4-89.7)DNAfilters. Electrophoresiswasfrom lefttoright.

times as much of this labeled 27S mRNA

hy-bridized

to the 42.7-47.4

region

as to the

36.7-40.5 region. Because 42.7-47.4 is 1.64 kb long and is

entirely

traversed by the large

mRNA,

then -0.5to0.6kb of the

long

mRNA

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.501.103.205.47.218.2] [image:4.501.309.408.57.546.2]

PROCESSING OF Ad2 mRNA's 815

10

0v 2 a.

8

6

4

2

a. u

0 10 20 30 40 50

SLICE

FIG. 4. Electropherogram of late

3HJuridine-la-beledAd2RNAhybridizedtoSma H(52.6-56.9) (0)

andtoEco E(83.4-89.7) (-)DNAfilters.

Electropho-resiswasfrom lefttoright;arrowsindicatepositions

ofHeLa28Sand 18SrRNA's.Inthisexperiment the

gelwasfractionated into1-mmslices rather than

2-mmslicestoafford higher resolution.

should lie to the left of 40.5 in the 36.7-40.5 region(Table 2). The 5' end of this RNA

(exclu-siveof the spliced sequences) is thus placedat

about38 to 39. If this mRNA is 4.5 to 5.0kblong and contains -0.3 kb of 5' and 3' sequences added duringprocessing, then it would occupy

about4.2 to4.7kb,or12 to 13%ofthegenome. Its encoded 3' end would then lie around 50-52. (ii) The 50-62 mRNA and the 68-80 mRNA. Thenextpossible 4.5to5.0 kbRNAto

the right of -50-52, without extensive overlap with the previous large mRNA, would begin around50and extend toaround62.

Hybridiza-tion to DNA fragments in this region, in fact,

identified such anmRNA. The 52.6-56.9

frag-mentfrom the hexon coding region (9, 21)

hy-bridized averylarge amount ofmRNA in the 27S to 28S size range as previously noted. In

additionto theveryprominent27Speakinthe hybridization patterntothe 52.6-56.9region, a

shoulder of 288 was noted. Further analysis under conditions of higher resolution demon-strated aseparation of two RNAsnominallya

27S anda28S species, withapredominance of label in the 27S species (Fig. 4). These promi-nently labeled mRNA's extend through 56.9-58.5 (a short fragment for which

hybridi-zation data is not shown inFig. 3,but isgivenin

Table2).Inaddition, the neighboringrightward fragment 58.5-70.7 also hybridized a large

amount ofRNA in the 27S to 28S size range.

Thus,itappearsthatthesetwomRNA'sbegin

around 50-52, span the 52.6-58.5 region, and

terminate in the left portion of the 58.5-70.7

fragment. The proportion of the radioactive

RNA hybridized to 58.5-70.7 that was to the 50-62 mRNA's land therefore the number of basescovered in the58.6-70.7region) couldnot

a

llcJ

IW

.

.4

.

m3

E,

IC..

.2

L..

.0

EN4

-6.

a 43

2

0. 10

.0

C-. t-.

t--

C--coi

LO

es

C--:

r-to

0._

4)

4,

z

E-CO _

oCD

c

O CwD

Q0

8

O es

0~ oe

o

LO Co

-4

qeq -4eq -4 q -4 -4 v

X

--04

A

.s

eq C 7 eN

C-.

eq

4,4

4-4

0

o

04

b.

XC.)

.

-* cx

Bz

VOL. 25, 1978

28S 18S- 4

4

4

.1

on November 10, 2019 by guest

http://jvi.asm.org/

[image:5.501.269.453.52.654.2] [image:5.501.73.218.56.196.2]

be simply calculated (see Table 2), however, because thenextfragmenttotheright, 70.7-75.9, alsohybridizedanmRNA ofapproximately27S. However, the 70.7-75.9 fragment is only 2.1 kb longitself,soalong27S mRNA mustprojecton

eitherorbothsides. The 75.9-83.4fragment did

indeedhybridizea27SRNA butalesseramount

ofradioactivitythan the 70.7-75.9fragment (Ta-ble 2). Fromthe proportion of -27S RNA hy-bridized to 70.7-75.9 and 75.9-83.4, it could be concluded thatatleast 2 kb of thelong mRNA spanningthe 70.7-75.9regionmustlie beforethe

70.5-75.9region,thatis,in the58.5-70.7region. Therefore, two different long mRNA's derive

sequences from either endofthe 58.5-70.7 re-gion.

Insummary, wecanidentify four large

-4.5-to 5-kb mRNA's, one from 38-50, two ofvery similar size in the 50-62 region, and one from

68-80.

Common 3' termini: the 36.7-62 region. As pointedoutabove,severalregionsofthe Ad2

genomeappeartoencodemultiple mRNA'sthat

sharesomeof thesamesequences.Suchagroup of mRNA's might be arranged in any of the

followingways: (i)code for thesame or apartof

the same polypeptide and share either5' or3'

termini; (ii) share a terminus, but the longer mRNAmightbe translatedonlyover aportion

of its totallengthtoproduceadifferentpeptide from the shortermRNA(reminiscentofSindbis

virus [36]);or (iii) existas "overlapping genes"

that might be translated in different-reading framesaswith 4X174(3)and simian virus40(S. Weissman, personal communication). An effort

wasmade, therefore,to calculateterminal

posi-tions ofsuchoverlappingmRNA's fromthe size

estimate affordedby electrophoretic migration

plus a determination of the fraction of such

mRNA's that hybridized to contiguous DNA

fragments.

The data suggestanumber ofregionswithin

which from two to four mRNA's seem to

ter-minate. In one of these regions, -50 on the genome,ourconclusionsaregreatlystrengthened

by the data of Ziff and Fraser (46) that the

oligonucleotidesof three different-sizedmRNA's

areindistinguishable. Because we reached the

sameconclusiononthe basisof theproportion

of hybridization of various sized mRNA's to

DNAfragmentsbetween 40 and52,wefeelmore

confidentthatwecan place3' endsreasonably accurately.Inaddition,ourdataagreewiththe

assignment of 3' ends by electron microscopic data (5, 6, 26). Nevertheless, inthe absence of

fingerprintdataonvarious-sized mRNA'sfrom

each putative 3'-coterminal site, we cannot

firmlyconclude 3'identity.

We will first consider the mRNA's

comple-J. VIROL.

mentary to fragments between positions 36.7 and 52.6.Thereare three species ofRNA, 27S, 25S, and 20S, complementary to 36.7-40.5, a fragment that is itselfonly 1.33 kblong.Thus, each of these three mRNA'sshould also hybrid-ize to DNA on one or the other side of the 36.7-40.5 region. The examination of labeled RNAcomplementary totheneighboring right-handfragment,40.5-52.6, showed thatthree dis-tinctive RNA peaks were also observed, 27S, 23S, and 19S, only the longestofwhichcoincided with the mRNA's complementaryto36.7-40.5. Because all the mRNA's in the entire region (16-100) aretranscribedtotheright(5'--3'),the twosmaller mRNAhybridizedto the36.7-40.5, therefore, mustderive sequences to the left of 36.7-40.4, i.e., from the 29.1-36.7 region, and bothmusthave their3' ends within36.7-40.5. A separatepreparation of RNAwasprepared and hybridized to the 29.1-36.7 fragment. No 28S RNAhybridized, butboth a25S and 20S RNA were detected. Thus, these two mRNA's, the 25S and 20S, have sequences in the 29.1-36.7 region and end inthe 36.7-40.5,but we cannot determine precisely their3' termini. Thelarge 27S mRNAbegins within 36.7-40.5and contin-uesinto40.5-52.6.

Todetermine moreaccurately the map posi-tions of the three RNAs in the40.5-52.6 region, further subdivision of thisDNAfragmentwith another restrictionenzyme(Kpn)wasrequired. DNA fragments with coordinates 40.5-42.7, 42.7-47.4, and47.4-52.6 wereobtained,and the proportion of radioactivity hybridizedfrom each of the 27S, 23S, and 19S mRNA's to each of thesefragmentswasdetermined(Fig.3B and C, Table 2). The 23S and 19S mRNA's failed to hybridize to the 40.5-42.7 fragment (Table 2), but both mRNA's as well as the 27S RNA hybridizedto42.7-47.4and47A-52.6 fragments (Fig. 3B and C). However, no hybridization of 23Sand 19S mRNA'soccurred to the 52.6-56.9 fragment, demonstrating that the encoded 3' ends of these molecules must lie within the 47.7-52.6region.If a27S RNAdoesbegin in the 36.7-40.5 asconcluded above, italso could not extendbeyond approximately 50-52on the ge-nome. Thus, it appears likely that three mRNA's,27S, 23S,and19S,endnearthe 50-52 region.

Ziff and Fraser (46) have identified three mRNA's of approximately the same lengths (27S, 23S, and 19S) asdescribed here whose 3' terminal 100 nucleotides, i.e., the nucleotides next topoly(A), hybridizetothe40.5-52.6 DNA fragment. Asmentioned previously,

all

three of these 3'endshave the sameoligonucleotide com-position; moreover, thesedistinctive oligonucle-otidesarealsopresent inEscherichia coliRNA

on November 10, 2019 by guest

http://jvi.asm.org/

VOL. 25, 1978

polymerasetranscripts of the47.4to52.6region. Thus,we canconclude that thesethree mRNA's (27S, 23S, and

19S)

all terminate very closeto thesamesite.

Common

terminii:

the 50-62, 68-80, and 83-92 groups. Finally, it appears that three other groups of mRNA that, within a group, maysharecommon3'-terminal regions exist in the 50-62, 67-83, and 83-92 regions of the ge-nome. As discussed above, two large RNAs, a 28S and 27S species, can be resolved by high-resolutiongel analysis in the50-62region (Fig. 4). Asdiscussedlater, if thesetwomRNA's have nonoverlapping 5'regions thatmight be neces-sary for each to be translated into different proteins, theymayhavecommon3'termini.

Next, both the 70.7-75.9 and 75.9-83.4 frag-ments(Fig. 3E, F, and G)hybridizeda27S, 22S, 19S mRNAgroup. A small -1.5kb mRNAmay

also

hybridize

to the 75.943.4 fragment. By

estimating the proportions of these various mRNA's

hybridizing

toeach ofthesetwo

frag-ments

coupled

with the gel estimates of the mRNAsizes, itseems

likely

that all four of these mRNA's have 3'-terminal sequences inthe re-gion of

79-80.

All wouldlikely terminate before the beginningof yetanothermRNAaround83 that continues across into the

neighboring

83.4-89.7region.

The80-90regionhas beenshownbyavariety

of

experiments

to encode thefiberprotein (21,

25).Aprominent 24SpeakofRNA

lying

mainly

in the 83.4-89.7 region was resolved into two mRNA's by

high-resolution

analysis (Fig. 4). These mRNA's extend into the 89.7-100 frag-ment, adistance

equal

toabout one-fifth of that whichliesinthe83.4-89.7

region

(Table

2). The larger RNA of the two may

begin

in the 75.9-83.4 fragment because a small amount of labeledRNAof thecorrectlength

hybridized

to

PROCESSING OF Ad2 mRNA's 817

both thisfragment and the 83.4-89.7 fragment (Fig.3F andG). Possibly,it isthesmallerRNA that encodes the fiberproteinbecause theregion of from -80 to 85 is nonessential for virusgrowth (9), whereas thefiberprotein is most likely an essentialgeneproduct.

Summary of mRNA detected. The tech-niquesusedfor thesehybridization studies dis-tinguishaverylargenumber of mRNA's, prob-ably asmanyas 12,that aredistinct in sizeand in map position. The earlier data on protein synthesis programmed by Ad2 mRNAhas also identifiedatleast ninespecificproteins encoded in thisarea (9, 20, 21). In additionthe in vitro protein synthesis studies identified the size of various mRNA's that gave maal program-ming for particular proteins. Thepresent work indicates many overlapping mRNA's, and it seems

possible

that

the

5'mostsection of each Ad2 mRNA isublizedintranslationas, in fact, is the case for togaviruses (18, 36) and RNA tumorvirus(17, 38) mRNA's. These RNAs con-tain two or morepotential entrysites for ribo-somes,butonly the 5' site is active.By making theassumptionof 5'regionsbeing the translated portions ofoverlapping mRNA's and using all the availabledata, themapassignments shown inFig.5 weremade.Table3givesasummary of thecalculationsemployinggel-estimated lengths andproportions of RNAhybridizing tospecific regions that have been used to determine the

map

positions

presented

in

Fig.

5. We wish to

again emphasize that only in theregion of50is thereoligonucleotide evidence ofexact 3'-coter-minal

mRNA's,

butthe dataonthis

region

ob-tained

by hybridization predicted

the coterminal natureof thesemRNA's inthat region.Our best

estimate

would

be,

therefore,

that the

other

groupsofmRNA'smayalsoterminateat

com-mon3'

poly(A)

addition sites.

0

I

P[

p II pv

Initiation site

0 10 20 30 40 50 60 70 80 90 100

F Hpo I

c---I D H Sm

(Ec

Kpn H I D (Sma D)

Eco RI B F D E C

FIG. 5. Schematic representation of late adenovirus RNAs generated in region of 27-92fromthe adeno

genome.Positions of RNAs werederived fromdatapresentedin Fig. 2 through 4 andTable3. Solid bars

above RNAs indicatepossiblemaplocation of late proteins, based on approximate genome locations and

sizes of mRNA's encoding respective proteins (20). Initiation site at 16 for the major latetranscriptionunit is

fromdataofEvans etal.(8).

0

--w

on November 10, 2019 by guest

http://jvi.asm.org/

[image:7.501.107.397.484.607.2]

[image:8.501.68.459.82.232.2]

TABLE 3. Sizes andcoordinates of late adenovirus RNAs

Approximate Estimated Coded size Genome Coordinates

Fragment Svaluea length(kb)b

(kb)'

(%)d3

SmaI 25 3.8 3.5 10.0 28.5e 38.5f

20 2.7 2.4 7.0 31.5e 38.5f

SmaD 27 4.7 4.4 12.5 38.5(±0.5)' 51.0e

23 3.3 3.0 8.5 42.5e 51.0(±0.9)'

19 2.4 2.1 6.0 45.0(±0.4)eg 51.0(±0.9)'

SmaH 28 5.1 4.8 13.5 49.5e 63.0(±1.7)'

27 4.6 4.4 12.0 51.0e 63.0(±1.7)'

EcoF 27 4.7 4.4 12.5 67.Oe 79.5'

22 3.0 2.7 7.5 72.0 79.5

19 2.4 2.1 6.0 73.5 79.5

EcoD 13 1.5 1.2 3.5 76-79.5 80-83.5"

Eco E 24 3.4 3.1 8.5 83.0e 91.5

(±0.2)'

23 3.1 2.8 8.0 83.5e 91.5(±0.2)'

aNominalS value determined from gel analysis by comparison to 28S and

18S

to bethe same S value relative

torRNA insucrosegradientsforhexon mRNA(SmaHspecific)and fiber mRNA (Eco E specific).

b

Estimated

lengths given by Loening (23), assuming 28S and 18S to be 5.1 and 2.1 kb, respectively (44).

cSize from footnote b minus150nucleotides for the poly(A) segment and 150 nucleotides for the 5' noncoded

leader sequence.

dValue in footnote c dividedby 35,500 and recorded to the nearest 0.5%.

eLength of RNA fromgel analysisand one terminus calculated as in footnote e allowed calculation of second

terminus.

fAssumption has been made that where an RNA terminates and a new RNA begins, no gap or overlap exists

between thetwo.

' Determined on basis ofoverlap of RNA in two adjacent fragments (see Table 2).

hBecause thisparticularRNA liesentirelywithin the Eco D (75.9-83.4) fragment and no overlaps were

obtained,theexactcoordinatesareunknown.

Presence of RNA sequences from

25.5-27.9 invarious mRNA's. Because each

mRNA deriving mainly from sequences tothe right of30 is thought to contain three ligated oligonucleotide regions (4, 5, 12, 19), labeled

RNA was hybridized to one small DNA

frag-ment that contains one of these regions. The

25.5-27.9DNAfragment,itselfonlyabout0.8 kb

long, hybridized labeled RNA from every size

class between 2 and 5 kb, with broad peaks corresponding to the major size classes of

mRNA (Fig. 6). The total amount of labeled

mRNAhybridized tothe 25.5-27.9 DNA

com-paredtothe total of all otherfragments to the

right of30wassmall(-2 to4%) anddistinctly higherin the 2- to 3-kbrangethan in the5-kb range.Thus,the amountshybridizedto25.5-27.9

areinagreementwith thepossibilityof approx-imnately 100 nucleotides from 25.5-27.9existing ineverymRNA(Table 4).

Inaddition asmall (8S) speciesofRNAwas

found hybridizing to the 25.5-27.9 fragment (slices33 to35 in bothtotalmiRNAandRNA

hybridizingto25.5-27.9) thatwasnotfound to

hybridize to any fragment in the region from

29-92. The possibility that a specific mRNA

existsthat is encodedbytheDNA in thisregion ofthegenomeis under furtherinvestigation.

Cellular location of processing event.

TABLE4. Hpa F(25.5-27.9)hybridizationa

25.5-27.9 30 to100 % Nu-RNAspe- hybridi- hybridi- 2%55n ticdeos cies zation zation 27.9

25.5-(cpm) (cpm) 27.9

27Sto28S 860 37,650 2.3 104

22Sto24S 330 11,200 2.9 84

300 10,900 2.8

19S to 20S 250 6,800 3.7- 81

8S 120

'Onepreparationofpoly(A)+RNAwasseparated

bygelelectrophoresisandwasusedtoobtain data in

Fig.3and6.Thecountsperminutehybridizedtothe

HpaF DNA

fragment

(25.5-27.9)arecomparedin this

tabletothe total hybridizedtoall other DNA frag-mentsbetween30and 100. Slice10represents27Sto

28SRNA,15and16represent22Sto 24SRNA,and

19 represents 19S to 20S RNA. The sizes of the

mRNA's[poly(A) and 5'leader]weretaken fromgel

migrationtobe4,500,2,900, and 2,200.

Aloni et al. (1) have described results late in simianvirus 40 infectionthatsupport a possible cytoplasmic processing of a 19S into a 16S mRNA. Because so many of the mRNA's de-scribed in this reportalso overlapin sequence, asdo the simian virus 40mRNA's, the possibil-ity of cytoplasmic processing might also be raised for Ad2 mRNAformation.Infectedcells

818

NEVINS AND DARNELL J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:8.501.271.460.375.473.2]

PROCESSING OF Ad2 mRNA's

werelabeled for 45min (in contrast to3or4h

used for all previous experiments); one-half of

the culture was harvested, and the remainder wastreated ("chased") withactinomycin Dfor an additional 3 h to suspend further mRNA labeling. The mRNAwasanalyzedasbeforeby

hybridizationtothe40.5-52.6 andthe 70.7-75.9 regions from which several apparently

3'-coter-minal mRNA'sarederived (Fig. 7). All species of mRNAwere present after the 45-min label

time in approximately the same ratios as

ob-served for the 3-h labeltime (see Fig. 3 and 4), and nosignificant changewas observed in the

absence of further RNAsynthesis for 3 h. Fur-thermore, the polysomal RNA fromcelLslabeled

continuously from 12 to 28 h of infection

con-tainedvirtuallyidenticalproportionsofthe

var-ious mRNA'sineachregion (datanotshown). Next, celLs werelabeled for 12 min, and nu-clear,poly(A)+ RNAwasisolatedandsubjected

togelelectrophoresisandhybridizationtoDNA

in the 40.5-52.6region (Fig. 8). All the species previously detected in the cytoplasm, the 27S,

238,

and 19SmRNA's,werealsopresentinthe

nucleus. Itappears,therefore, thatthe process-ing of RNA to yield mRNA from the Ad2-in-fectedcells occursinthe nucleus.

Efficiency of mRNA formation from pri-marytranscripts.One oftheprimereasonsfor

obtaining detailed information onthe map po-sition ofspecific Ad2 mRNA's wasto allow a

A 28S 18S

10

10 -B

0

5

0 10 20 30 40

[image:9.501.273.415.168.351.2]

SLICE

FIG. 6. Electropherograms oftotal unhybridized late 3HJuridine-labeled Ad2 RNA (A) and RNA

hybridizedtoHpaF(25.7-27.9)DNAfilters(B).

Elec-trophoresiswasfrom lefttoright. Aportion(10 1d)of

eachelutedgel slicewascountedtoobtainthe profile in(A).

study of the efficiency of utilization as mRNA

of differentsegments ofthelarge late primary

nuclearRNA transcript. Suppose thata

partic-ular mRNAportion ofaprimarynuclear

tran-scriptwasall convertedtocytoplasmic mnRNA within 15to20min and this particular cytoplas-mic mRNAwasrelativelystable. The increase

in that fraction oflabeled nuclear RNA

corre-spondingtothesequencepresentin the

partic-0

0.

4

2

6

4

2

10 L

[image:9.501.84.203.388.590.2]

SLICE

FIG. 7. Electropherograms of late

pH]uridine-la-beledAd2 RNAhybridizedtoSma D(40.5-52.6) DNA

filters (AandB)andtoEco F(70.7-75.9)DNAfilters

(CandD). (A)and(C)representRNAlabeledduring

a45min pulse,and(B)and(D)representRNAfrom

ceUslabeledfor45minandexposedtoactinomycin D(5pg/ml)foranadditional 3 h.

6

N

I0

0.

4

2

28S 18S

Is

i

I20

I 10 20

0

SLICE

FIG. 8. Electropherogram oflate

f3HJuridine-la-beledpoly(A)-containingnuclearAd2RNA hybrid-ized to Sma D (40.5-52.6) DNA filters. RNA was

labeledfor12 min at15 hpostinfectionandprepared

asdescribedinthetext.Electrophoresiswasfrom left

toright.Arrowsindicatepositions ofHeLa 28Sand

18S rRNA's. The nuclei in this preparation were

detergentwashedtoremoveallcytoplasm.

A 28S 18S C 28S 18S

4 4 4 4

I

-t

B 4 0 4

0 10. 20 0 10 20

819 Voi. 25,1978

3

2

5

on November 10, 2019 by guest

http://jvi.asm.org/

[image:9.501.299.394.424.573.2]

ular mRNA would be followed in 15 to20min byaparallel increase inthis particular labeled

mRNA. A timecourseofincorporation of

[3H]-uridine into nuclearandcytoplasmic RNAwas first determined for RNA complementary to

Ad2 DNA fragments 52.6-56.9, 70.5-75.9, and 83.4-89.7; all of these DNA regions are either

entirely ormostly represented in mRNA. The curveof labeled nuclear RNA for eachfragment rose rapidly, as expected, reaching a plateau

after about 40 to 60 min (Fig. 9). The labeled mRNA begantoappearapproximately linearly after about 20 min. Therate oflabeling of

nu-clear andcytoplasmic RNA specific for the three regions wasestimated from the maximalslope

of increase and is shown in Table 5. From the maximum rates of labeling of the same

se-quencesinthe nucleus andthenthecytoplasm, the utilization of thenuclearRNA foranyofthe threeregions wascalculated; about 20% of the

52.6-56.9 region and about 15% of each of the othertworegions wereconserved. Because the

nucleiwerenotstripped of cytoplasm by

deter-gent treatmentinthisexperiment, atleast 15%

of the cytoplasmic RNA might remain in the nuclear fraction (41). Therefore, the poly(A)+ nuclear RNA of thecorrectsizeformRNAwas

measured by hybridization to 52.6-56.9 and 83.4-89.7. In each case, an amount oflabeled

RNA equalto about 20to30% ofthatalready detected in the cytoplasm was found. If this

poly(A)+ RNA is added to that in the free cytoplasm,the maximum numbers forthe

trans-portefficiency wouldbebetween 20 and 26% for

these three sections of the Ad2 genome. An

eJ

20-15-8

0 '0x

10.

5

0 40 800 40 80 0 40 80

MINUTES

FIG. 9. Kineticsofsynthesis ofadenovirusnuclear

RNA and the appearance oflabeled RNA in the

cytoplasmspecific forDNAfragments52.6-56.9(A), 70.7-75.9(B), and 83.4-89.7(C). HeLa ceUs infected for 15h with adenovirus were concentrated to 10i ceUsperml,and

fIHluridine

wasadded to100

#0Ci

ml.Samples of2x107cellswereremoved at indicated

intervals, and thenuclear (0) andcytoplasmic (0)

RNAswereprepared.Each RNAsamplewas

precip-itatedwithLiCItoremoveviral DNA.Hybridizations

wereperformed with filters containingDNA

frag-ments equivalent to anamountpresentin 5 pgof adenoDNA.

TABLE5. Efficiency oftransportoflateAd2 RNAs

DNA fagment %Transport"

36.7-40.5 5

40.5-42.7 4

42.7-47.5 19

52.6-56.9 17

70.7-75.9 13

83.4-89.7 13

aDetermined from rate of

labeling

of cytoplasmic

RNA between 45 and 90 min (counts per minute

increase per45min) dividedbytherateoflabelingof

nuclearRNA between 5and 20 min (counts perminute

increaseper15min).Theincreaseinradioactivitywas

approximatelylinear duringthetimechosen for the

estimation(see Fig. 9).

analysis was also made of regions of theDNA genomewheremultiple mRNA's arise, namely, 40.5-52.6. In such regions where the primary transcriptappears tobe used for more than one type ofmRNA, the transport efficiency should vary for RNA complementary to contiguous DNAfragments, being higherwhere sequences wereshared byallmRNA'sandlower for frag-mentsthat providesequences exclusively for one ofthegroup ofmRNA's.The efficiency of trans-port of nuclear RNA from the 40.542.7

frag-ment that

hybridizes

the 27S but notthe 23S

and 19S mRNA's(seeFig. 3)wascompared with the efficiency oftransport ofthe nuclearRNA from the 42.7-47.4 region that is completely includedin two andpartially includedin athird mRNA.

Only

4% of the 40.5-42.7 region was transported compared with19% ofthe42.7-47.4 region. Another region, 36.7-40.5, which pro-vides sequencesfor several mRNA's, including a

long

27S mRNA with several shorter 3'-co-terminal mRNA's, also showed a transport of

only

5%.

These dataare mostcompatible with the in-terpretation that each

primary

transcript gives rise to a limited number ofmRNA's, possibly one; thus, any one ofthepotential mRNA re-gions existswithonlya5 to30%success rate.

DISCUSSION

Themajorfacts

established

bythis work are shown inFig.5-alarge number,perhaps 12 or more, mRNA species of distinct size (electro-phoretic

migration)

and genomepositioncanbe distinguishedfrom 29.1through100.McGrogan and Raskas (24) also have recently demon-stratedmultiple speciesof RNA inthisregion. These mRNA'saresynthesized(labeled) latein infection, and all appear to derive from one primary nuclear RNA transcript (2, 8, 13, 14, 43).All themRNA'scontaina3'

poly(A),

several have beendefinitely demonstratedtocontain a

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

[image:10.501.269.459.76.150.2] [image:10.501.65.256.438.533.2]

PROCESSING OF Ad2 mRNA's 821

5'cap(12), and from theaveragecontentofcaps

in the total late Ad2 mRNA's,most molecules

mustcontainacap(37). In each specificcase so

farexamined, the nuclear processing also

appar-ently involves breakage and reunionat16.7,19.7

and 26 in theprimary transcript toproducean

mRNA(4, 5, 12, 19). Additional supporting evi-dence is offered on this point by the datum reported here than DNA from 25-27 hybridizes

anamountoflabeled RNAequivalent to -100

nucleotide portions ofevery size class of Ad2

mRNA's.

If this picture of multiple recombinational possibilities for the late 16-100 primary

tran-script is correct, manyintriguing questionsare

posed.

(i) Do all of theseeventsoccurin the nucleus?

At leastapreliminaryanswertothisquestion is afforded in Fig. 8, where it is shown that all ofa

group ofoverlapping mRNA's observable after long labels in thecytoplasmare labeled within

12min in the nucleus. Moreover, the proportions of these mRNA's in the cytoplasm appears to

remain fixed in the absence of further RNA synthesis; the processing events, at least for

someof the Ad2 mRNA's, would appeartobe nuclear.

(ii) Canaprimarytranscript give risetoonly

onemRNA?Iftheputative ligation sitesat16.9, 19.7, and 26.0 arenot repetitive at those sites (and thiscanonlybetoldbysequenceanalysis), then one transcriptcould donate the 5' 150 to

200 nucleotides to only one mRNA. Such a

mechanism wouldallowanyspecific mRNA

re-gion toexitto the cytoplasm withafrequency

sustantially less thanone. Anattempt to

mea-suretransportefficiencyofseveralregionsof the primary transcript known to encode major mRNA speciesshowsapproximately 15to20% conservationfor three of the fourand 20 to 30%

conservationforone (thehexonregion).These

resultsareinaccord,butclearlydo notestablish

the proposal that one transcript can be

con-verted intoonlyonemRNA.

(iii) Whatlocatestheprimary cleavage sites?

The moststriking generalfact aboutthe

distri-butionofcleavagesites(or mRNA termini) con-cerns the 3' termini. There may be only five

different 3' regions that exist for 12 different mRNA's-one group of mRNA's (ending at

'50) share theexact same3'regions (46), and

theremaybe fourothersuchcommon 3' sites.

However, each of the 12 mRNA's may have differentsequencesin its5'portions, excluding, ofcourse,theterminal -200nucleotidesthatare

presumablycommontoallmRNA's.

Themeaning ofthisarrangementof presumed coding regions may possibly parallel the mode

of polypeptide synthesis in certain togaviruses

(18,36),the retroviruses (17, 38),andthe papo-vaviruses

(S.

Weissman,

personal

communica-tion). For

example,

with

Sindbis

virus (35) and Rous sarcoma virus

(38),

mRNA's of

various

lengths exist that are translated into a single polypeptide. Thus, we might expect the 28S mRNA from 50-62 to encode the

small

pVI

protein from its5'region andthe 27SmRNA,a slightly shorter molecule, to lack the pVI se-quence and encode the hexon protein. In this case nooverlappingcodingsequenceswould oc-cur.Lewisetal. (21) suggested suchapossibility

several

years agobut didnotattempt toeffecta

complete

separation

ofthetwomRNA'stoprove thepoint.

Intheregion that encodes thepentonprotein, 40.5-52.6,theoverlapping mRNA'sandthe pro-teins encoded by eachpresent greater

difficul-ties. The penton protein is -70,000 daltons, whichrequiresatleast2,300nucleotidesorabout 7.0% of the genome.The overlapping mRNA's in this region appear not to leave that much space free at the 5' end of any of the three mRNA's in the region. Thus, these mRNA's might be readin adifferent frameasis thecase for certainregionsin

4X174

(3) andsimianvirus

40

(Weissman,

personal

communication).

Alter-natively,thesemRNA'smight encode polypep-tides thatwereidenticalattheirCOOH-terminal segments but

possessed

different

NH2

termini. (iv) Whatarethe relative

proportions

ofthe various mRNA's and how are these decisions mediated?

The one clear fact is that themRNA's from different

regions

neitherexitnor

exist

at

steady

state in

equal

concentration in the

cytoplasm;

e.g., there ismuch more labeled hexon mRNA than any other mRNA.

However,

Flint and Sharphave shownmRNA from70-76tobe the

most

populous

at

steady

state

(10).

Does this

meanthat the

cleavage

schedule favors certain regions over others? Not

necessarily.

If there

werefive

equally probable

initial

cleavages,

say

to createfivedistinct 3'

termini,

followed

by

a

subdivision within these groups based on the number of mRNA's within each

section,

thenno

primary

cleavage

mechanism would be

required

todivide the

primary

transcript

unequally.

Such amodel of

processing

accords

with,

but

by

no

means is

proved

by,

the present data. In four

differentregions,RNAappearstobe conserved with about 20%efficiency.Thegreater

accumu-lation of hexonmightthen be

explained

bythe existence of

only

twomRNA's fromthisregion. Clearly this discussion of these

complicated

eventsshould

only

betaken asan introduction

to the

problem.

For

example,

no mention has

yet been made of

cytoplasmic stability

of differ-entmRNA's,possiblyamost

important

pointin

Voi. 25,1978

on November 10, 2019 by guest

http://jvi.asm.org/

822 NEVINS AND DARNELL

controlling relative concentrations of the differ-ent mRNA's. The point worth making is that eventhougharemarkably complex setof

post-transcriptional

events is involved in producing

thewelter of Ad2 mRNA's from the large tran-script, there is as yet no need to hypothesize that these are coordinated into a regulatory function. Suchsteps maybenecessaryand de-liver different amounts of primary transcript into differentmRNA's, but thismay occurwith afixed

probability

andnotbea

regulated

setof events

capable

of

responding

to change. Only moreexperimentscansettle thisissue.

ACKNOWLEDGMENTS

WethankMarianneSalditt-Georgieff and Patricia White forexcellent technical assistance and NigelFraser for the generousgiftofKpnDNA restriction fragments.

This work wassupportedby Public HealthServicegrant CA 16006-04from theNationalCancer InstituteandAmerican CancerSocietygrant NP213G. J.R.N. isaFellow of the Jane Coffin Childs MemorialFund for MedicalResearch.

LITERATURE CITED

1.Aloni, Y., M.Shani, and Y. Reuveni. 1975. RNAs of simian virus40inproductivelyinfectedmonkeycells: kinetics offormationanddecay in enucleate cells. Proc. Natl. Acad. Sci. U.S.A. 72:2587-2591.

2. Bachenheimer, S., and J. E.Darnell.1975.Adenovirus 2mRNA istranscribed aspart of ahigh molecular weight precursor RNA. Proc. Natl. Acad. Sci.U.S.A. 72:4445-4449.

3.Barrell, B. G.,G. M.Air,and C. A. Hutchison m. 1976. Overlapping genes in bacteriophageOX174. Na-ture(London)264:34-41.

4. Berget,S.M., C.Moore,and P. A.Sharp.1977.Spliced segments at the 5' terminus of adenovirus 2late mRNA. Proc.Natl. Acad.Sci. U.S.A. 74:3171-3175.

5. Chow,LT.,R. E.Gelinas,T. R.Broker,and R. J. Roberts.1977.Anamazingsequence arrangementat the 5' ends of adenovirus2messenger RNA. Cell12:1-8. 6. Chow, LT.,J. M.Roberts, J. B.Lewis,and T. R. Broker.1977. Amap ofcytoplasmicRNAtranscripts fromlytic adenovirus type 2, determined by electron microscopy of RNA:DNAhybrids.Cell11:819-836. 7. Duesberg,P.H.,and P. K.Vogt.1973.Gel

electropho-resis of avianleukosis andsarcomaviral RNA in form-amide:comparisonwith other viral andcellularRNA species.J. Virol. 12:594-599.

8. Evans, R.,N. W.Fraser,E.Ziff,J.Weber,M.Wilson, andJ. E.Darnell.1977.The initiation sites for RNA transcriptioninAd2 DNA.Cell12:733-739.

9. Flint,J. 1977.Thetopographyandtranscriptionof the adenovirusgenome.Cell10:153-166.

10.Flint, S.J.,andP.A.Sharp.1976.Adenovirus transcrip-tion. V.Quantitationof viral RNAsequencesin AD-2 infected and transormed cells. J. Mol. Biol. 106:749-771.

11.Flint,S.J., Y.Werwerka-Lutz,A.S.Levine, J.

Sam-brook,and P. A.Sharp.1975.Adenovirus

transcrip-tion.II.RNAsequencescomplementarytosimianvirus 40and adenovirus2DNA inAd2+NDr-and Ad2+ND3-infectedcells.J.Virol.16:662-673.

12.Gelinas, R.E., and R. J. Roberts.1977.Oneprominent 5'-undecanucleotide in adenovirus 2 late messenger RNAs.Cell11:533-544.

13.Goldberg, S.,J.NeAins,and J. E.Darnell. 1978. Evi-dence from UVtrnscriptionmappingthat late

adeno-virus type 2 mRNA is derived from alarge precursor molecule.J. Virol.25:806-810.

14.Goldberg,S., J. Weber, and J. E. Darnell. 1977. The definition of alarge viral transcription unit late in Ad2 infection of HeLacells:mapping by effect of ultraviolet irradiation. Cell 10:617-621.

15.Haas, M., M. Vogt, and R. Duibecco. 1972. Loss of simian virus 40 DNA-RNAhybrids fromnitrocellulose membranes; implications for the study ofvirus-host DNA interactions. Proc. Natl. Acad. Sci. U.S.A. 69:2160-2164.

16.Jelinek,W.,G.Molloy, R. Fernandez-Munoz, M.

Sal-ditt, and J. E. Darnell. 1974. Secondary structure in hnRNA:involvement of regions from repeated DNA sites. J. Mol. Biol.82:361-370.

17.Joho, R.H., M. A.Billetes, and C.Weisaman.1975. Mapping of biological function on RNA of avian tumor vinres:location of regions required fortransformation anddetermination of host range. Proc. Natl. Acad. Sci. U.S.A. 72:47724776.

18.Kennedy, S. I. T. 1976. Sequencerelationships between thegenome and theintracellularRNAspecies of stan-dard anddefective-interfering Semliki forestvirus.J. Mol.Biol. 108:491-512.

19. Klessig,D. F. 1977. Twoadenovirus messenger RNAs have a common 5'terminal leader sequence which is encodedatleasttenkilobasesupstream from the main codingregions for these messengers.Cell12:9-22. 20. Lewis, J. B., C. W. Anderson, J. F. Atkins, and R. F.

Gesteland.1974.Theorigin and destiny of adenovirus proteins. Cold Spring Harbor Symp. Quant. Biol. 39:581-590.

21. Lewis,J. R., J. F. Atkins, C. W. Anderson, P. R.

Baum,andR.F. Gesteland. 1975. Mapping of late adenovirusgenesby cell-free translation of RNA se-lected by hybridization to specific DNA fragments. Proc.Natl.Acad. Sci. U.S.A.72:1344-1348.

22. Lindberg, U., T. Persson, and L Philipson. 1972.

Isolation andcharacterizationofadenovirus messenger ribonucleic acid in productive infection. J. Virol. 10:909-919.

23. Loening,U. E.1967.The fractionationofhighmolecular weightribonucleicacidbypolyacrylamide gel

electro-phoresis.Biochem.J. 102:251-257.

24. McGrogan,M.,and IL J.Raskas.1977.Species identi-ficationand genomemappingofcytoplasmicadenovirus type 2RNAs synthesized late ininfection. J. Virol. 23:240-249.

25.Mautner,V.,J.Williams,J.Sambrook,P.A.Sharp, andT.Grodzicher. 1975.The location of the genes codingfor hexon and fiberproteinsinadenovirus DNA. Cell 5:93-99.

26. Meyer,J., P. D.Neuwald, S.-P. Lai,J. V.Maizel,and ELWestphal. 1977.Electronmicroscopyof late ade-novirus type2mRNAhybridizedtodouble-stranded viral DNA. J.Virol.21:1010-1018.

27. Molloy,G.R.,W.Jelinek,M.Salditt,and J. E.

Dar-nell. 1974. Arrangement ofspecific oligonucleotides within poly(A) terminated hnRNA molecules. Cell 1:43-53.

28.Moss, B., andF.Koczot.1976.Sequenceofmethylated nucleotides atthe 5'-terminus of adenovirus-specific RNA.J.Virol. 17:385-392.

29. Mulder,C.,J.R.Arand,W.Keller,U.Petterseon, R.J.Roberts,and P. A.Sharp.1974.Cleavagemaps of DNA from adenovirus types2and5byrestriction endonucleasesEcoRI andHpa1.ColdSpringHarbor Symp.Quant. Biol.39:397400.

30. Parsons,J.T.,J.Gardner,andM.Green.1971.

Bio-chemical studies on adenovirus multiplication. XIX. Resolutionoflate viral RNAspeciesinthe nucleus and

cytoplasm.Proc.Natl.Acad.Sci.U.S.A.68:557-560.

31. Pettersson,U.,C.Mulder,H.Delius,andP. A.Sharp.

on November 10, 2019 by guest

http://jvi.asm.org/

VOL 25,1978

1973.Cleavage of AD-2 DNA into sixuniquefragments byendonuclease R.R1. Proc. Natl. Acad. Sci. U.S.A. 70:200-204.

32. Philipson, L, U. Petteron, U. Lindberg, C. Tib-betta, B. Venstrom, and T. Persson. 1974. RNA synthesis andprocesinginadenovirus infected cells. ColdSpringHarborSymp. Quant. Biol.39:447-456. 33.Philipson,L,R.Wall,G.Glickman,and J. E.Darnell.

1971.Addition ofpolyadenylatesequences to virus-spe-cific RNA during adenovirus infection. Proc. Natl. Acad.Sci.U.S.A. 68:2806-2809.

34.Sharp,P.A.,P. H.Gadlimore, and S. J. Flint. 1974. Mapping of adenovirus 2 RNA sequences inlytically infected cells and transformed cell lines. ColdSpring HarborSymp. Quant.BioL39:457-474.

35. Shatkin,A.J. 1976.Cappingofeukaryotic mRNAaCell 9:645-664.

36. Simmo D.T.,and J.H. Strauss.1974.Translation ofSindbis virus 26S RNA and 49S RNA inlysatesof rabbitreticulocytes.J.Mol. Biol. 86:397-409. 37. Sommers,S., MLSalditt-Georgieff,S.Bachenheimer,

J.E.Darnell,Y.Furuichi, AL Morgan,and A. J. Shatkin.1976.Themethylationofadenoviuns-specific nuclear and cytoplasmic RNA. Nuc. Acids Res. 3:749-766.

38. Stacey,D.W.,V. G.Affrey,andH. Hanasa 1977. Microinjection analysisofenvelopeglycoprotein mes-senger activitiesof avian leukosis viral RNAs Proc.

PROCESSING OF Ad2 mRNA's

823

Natl. Acad. Sci. U.S.A.74:1614-1618.

39. Tal, J., E. A. Craig, S. Zimmer, and H. J.Raskas. 1974.Localization ofadenovirus2 mRNAs tosegments of the viralgenome defined by endonuclease EcoR.R. Proc.Natl. Acad. Sci. U.S.A.71:4067-4061.

40.Thomas, AL, R. Iv White, and R. W. Davis. 1976. Hybridization of RNA to double-strandedDNA: for-mation of R-loops. Proc. Natl. Acad. Sci U.S.A. 73:2294-2298.

41. Vaughan,ML H.,J.R. Warner, and J. E. Darnell. 1967. Ribosomalprecursor particles in the HeLacell nucleus. J. Mol. Biol.25:235-251.

42.Wall, R., L Philipson, and J. E.Darnell. 1972. Proc-essing ofadenovirus-specific nuc!ear RNA during virus replication. Virology 50:27-34.

43.Weber, J., W.Jelinek,and J. E. Darnell. 1977. The definition of alargeviraltranscriptionunit late in AD-2infection of HeLacells: mapping of nascent RNA molecules labeled inisolatednuclei. Cell 10:612-617. 44.Wellauer, P. K., and I. B. Dawid. 1973. Secondary structuremaps ofRNA:processingofHeLaribosomal RNA. Proc. Natl. Acad. Sci.U.S.A. 70:2827-2831. 45.White,D.D.,MLD.Scharff,and J. V.MaizeL1969.

The polypeptides ofadenovirus. HI. Synthesis in in-fectedcells.Virology38:3956406.

46. Zi E.,and N.Fraser. 1978. Adenovirus type 2 late mRNA's:structuralevidence for3Tcoterminal species. J.Virol.25:897-906.

on November 10, 2019 by guest

http://jvi.asm.org/