-
Proc. Natl. Acad. Sci. USAVol. 87, pp. 2675-2679, April
1990Biochemistry
Ribosomal protein L4 stimulates in vitro termination
oftranscription at a NusA-dependent terminator in theS10 operon
leader
(attenuation/autogenous control/Escherichia cola)
JANICE M. ZENGEL AND LASSE LINDAHLDepartment of Biology,
University of Rochester, Rochester, NY 14627
Communicated by Charles Yanofsky, February 1, 1990
ABSTRACT The il-gene S10 ribosomal protein operon ofEscherchia
coli is under the autogenous control of W, theproduct of the third
gene of the operon. Ribosomal protein IAinhibits both transcription
and translation of the operon. Ourin vivo studies indicated that IA
regulates transcription bycausing premature termination within the
untranslated S10operon leader. We have now used an in vitro
transcriptionsystem to study the effect ofpurified IA on expression
ofthe S10operon. We rind that the cell-free system reproduces the
in vivoobservations. Namely, in the absence of IA, most of the
RNApolymerases read through the termination site in the
S10attenuator; the addition of IA results in increased
terminationat this site. However, RNA polymerase does not terminate
atthe S10 attenuator, with or without IA, unless an
additionalfactor, protein NusA, is added to the transcription
reaction.These results suggest that the attenuator in the S10
operon isa NusA-dependent terminator whose efficiency is regulated
byribosomal protein IA.
The S10 operon of Escherichia coli contains the genes for
11ribosomal proteins (r-proteins). Expression of this operon
isunder the autogenous control of r-protein L4, the product ofthe
third gene (1-3). L4 regulates the S10 operon by
inhibitingtranslation ofthe most proximal gene (3, 4) and by
stimulatingpremature termination of transcription within the SlO
leader(5). Genetic studies indicated that L4-mediated control
oftranscription and of translation works by two
independentmechanisms (4, 6).The site of L4-stimulated
transcription termination is about
140 bases from the start oftranscription (5-7). This places
thetermination site more than 30 bases upstream of the mostproximal
structural gene, within a string of uridines on thedescending side
of a stable hairpin structure (8). We haveproposed that this
structure might function as a relativelyweak p-independent
terminator that works more efficiently inthe presence of L4 (5, 6).
To test this simple model and tobetter understand the molecular
details of LA-mediated at-tenuation, we have studied the effect of
purified r-protein L4on in vitro transcription of the S10 operon by
using a simplecell-free transcription system. We find that L4 does
indeedstimulate termination of transcription by RNA polymerase
atthe same position in vitro as in vivo. However,
attenuationrequires the addition of transcription factor NusA.
MATERIALS AND METHODSPlasmid Templates. Plasmids pLF1 (9) and
pSma2 (7) are
shown in Fig. 1. Plasmid pLL226 (7), also shown in Fig.
1,contains the S10 promoter and leader followed by a partialtRNA
gene (TDF1, see ref. 10) and a synthetic rrnC terminator
(11). Deletion derivatives of pLL226 were constructed
bylinearizing the plasmid with Sma I, treating for various
timeswith BAL-31 at 100C, digesting with Sty I, filling-in with
theKlenow fragment ofDNA polymerase I, and religating (7).Chemicals
and Enzymes. Uridine 5'-[a-[P5S]thio]tri-
phosphate (UTP[35S]) and [a-32PJUTP were from
Amersham.Nonradioactive nucleoside triphosphates were from
Pharma-cia or Sigma. RNA polymerase either was purified in
ourlaboratory by standard procedures (12) or was a gift from
T.Platt (University of Rochester) or E. Morgan (Roswell
Park,Buffalo). p protein was from T. Platt, NusA was from T.
Plattand from E. Morgan, factors NusB, NusG, and NusE werefrom J.
Greenblatt (University of Toronto), and r-proteinswere from M.
Nomura (University of California, Irvine) andfrom K. Nierhaus and
P. Nowotny (Max Planck Institute,Berlin).In Vitro Transcription
Reactions. The standard 20-Al reac-
tion mixture contained 40 mM Tris-HCl (pH 7.9), 10 mMMgCl2, 0.1
mM EDTA, 100-150 mM KCl, 0.2 mM dithio-threitol, 500 ,uM ATP, 500
,tM CTP, 500 ,uM GTP, 100 ,uMUTP, 20-40 ,uCi of UTP[35S] (1 Ci = 37
GBq) or 12.5 puCi of[32P]UTP, and 1-2 pug of supercoiled plasmid
DNA. RNApolymerase, r-protein L4 or S7, and other termination
factorswere added at the indicated concentrations. All
componentsexcept DNA were added on ice; the reaction was started
byaddition of DNA. After 15 min at 370C, the reaction wasstopped by
placing it on dry ice. Carrier RNA (10 ;kg of yeastRNA) was added,
and total nucleic acids were extracted withphenol and then with
chloroform/isoamyl alcohol [24:1 (vol/vol)], precipitated with
ethanol, and resuspended in 5 ,ul of 10mM Tris-HCl, pH 7.9/0.1 mM
EDTA. The RNA was mixedwith an equal volume of 95% (vol/vol)
deionized formam-ide/10 mM EDTA/0.1% bromophenol blue/0.1%
xylenecyanole and then fractionated on a standard DNA sequenc-ing
gel. Where indicated, the RNA was first subjected to a T1nuclease
mapping procedure (13) and then analyzed by gelelectrophoresis. In
vivo-labeled RNA (2, 13) was analyzed byT1 nuclease mapping (7, 8,
13) in parallel with the in vitrotranscripts. RNA size markers were
prepared as described(13).
RESULTSMapping of Attenuated and Read-Through Transcripts
Syn-
thesized in a Purified Transcription System. We first
analyzedthe effect of purified L4 on in vitro transcription of the
S10operon leader by using supercoiled plasmid pLF1 DNA astemplate
(Fig. 1) and a partially purified RNA polymerase.S10 leader RNA was
then purified using a Ti nucleasemapping procedure (7, 8, 13) in
which total transcribed RNAwas hybridized to a single-stranded DNA
probe specific for
Abbreviations: r-protein, ribosomal protein; UTP[35S], uridine
5'-[a-[35S]thio]triphosphate.
2675
The publication costs of this article were defrayed in part by
page chargepayment. This article must therefore be hereby marked
"advertisement"in accordance with 18 U.S.C. §1734 solely to
indicate this fact.
Dow
nloa
ded
by g
uest
on
July
2, 2
021
-
2676 Biochemistry: Zengel and Lindahl
AUT4 SD_
1 00 bases
S10 leader Sb' |acZ'
ATT* SD
S0 leader IlZAll* SD
:Sstl Smal Styl. 25B
* zzzzl 27B=oA///] 30B=,/i.ZZ 39B
toZZZZZZZ 26B,ZZZZZZZZZJ 7BEZZZZZZZZI 2Bvzizzzzzz]
56BLuz>,-/>> - 28BEZZZ{*/fZZZ3i lB
ih vivo in7 vitropLF1 pSma2 pLF1
~t 't st-J , X C/
+ + +
280 -*bases
150-*bases
- ATT'* ATT
FIG. 1. Maps ofS10 leader fusion plasmids. Plasmid pLF1
carriesthe beginning of the S10 operon, including the promoter
(Ps10), a172-base S10 leader, and the proximal half of the
structural gene forr-protein S10 fused in frame to lacZ'. Plasmid
pSma2 contains thefirst 165 bases of the S10 leader upstream of an
intact lacZ gene.Plasmid pLL226 also contains the first 165 bases
of the S10 leader,placed upstream ofthe rrnC terminator (t). All
three plasmids containan intact Shine-Dalgarno sequence (SD). The
site of LA-stimulatedtermination in the leader (ATT) is indicated.
The position of theleader hybridization probe used for T1 nuclease
mapping is shownabove the pLF1 map. This probe (7) is an M13
derivative containing167 bases of S10 operon DNA from 2 bases
upstream of thetranscription start site (8) to the Sst I site
within the Shine-Dalgarnosequence ofthe S10 gene. Extents ofBAL-31
deletions in derivativesof pLL226 are shown by hatched bars below
the pLL226 map.Relevant restriction enzyme sites are also
indicated.
the S10 leader, nonhybridized RNA was digested with nu-clease
T1, and the resulting protected RNAs were analyzedby gel
electrophoresis. For comparison, we also analyzedRNA synthesized in
vivo in the absence or presence ofexcessLA. The leader probe
protected three major bands with the invitro RNA, two ofwhich
migrated to the same position on thegel as the bands observed from
in vivo-labeled RNAs (Fig. 2).The upper band corresponds to
read-through transcripts-that is, RNA transcribed through the S10
attenuator regioninto the structural gene and trimmed back to the
size of theleader probe by the nuclease. The lower band (actually
adoublet band) corresponds to attenuated transcripts-that is,RNA
terminated about 140 bases into the S10 leader, here-after referred
to as ATT transcripts. Although not shown, themobility of this
doublet was not altered when the transcriptswere purified in the
absence ofT1 nuclease, as expected iftheRNA were contained entirely
within the sequence carried onthe probe. From these results we
conclude that the S10attenuator functions in vitro. Moreover, the
attenuator re-sponds as expected to L: the addition of this protein
resultedin an increase in the level of attenuated transcripts and
adecrease (visible in shorter exposures than the autoradiogramshown
in Fig. 2) in the level of read-through transcripts.Other
r-proteins, including S7 (Fig. 2), L11, and S4 (data notshown) had
no effect on attenuation.The only discrepancy between the in vitro
and in vivo
results was that in the cell-free system there was a
significantamount of transcription termination just beyond the
attenu-ator, generating transcripts (actually a series of three
bands)that were not observed in the in vivo samples and
thathereafter will be referred to as ATT' transcripts (see Fig.
2).The sizes of these ATT' RNAs correspond to terminationwithin a
series of three uridine residues just downstream ofthe four
uridines at which the LA-stimulated attenuation
FIG. 2. Mapping of in vivo and in vitro transcripts from the
S10operon. In vivo transcripts were from cells carrying plasmid
pLF1 orpSma2 pulse-labeled for 3 min with [32P]phosphate before
(lanes -)or 10 min after (lanes +L4) induction of LA oversynthesis
from asecond plasmid, pLF17 (9). In vitro transcripts were
synthesized ina 20-Al reaction mixture containing 2 Etg of RNA
polymerase(partially purified; see text), 2 ,ug of supercoiled
plasmid DNA, 12.5ILCi of [32P]UTP, and 0.9 i&g of LA, 0.6 ,.g
of S7, or the equivalentvolume of protein buffer. Transcripts were
hybridized to a leaderprobe (Fig. 1) and digested with T1 nuclease.
Protected RNAmolecules were then fractionated on a denaturing
urea/polyacryl-amide gel (13). Positions of size markers are
indicated on the left.Bands corresponding to read-through (RT) and
attenuated (AUT and,for in vitro samples, AUT') transcripts are
indicated on the right.
occurs (Fig. 3). The relevance ofthese RNAs is not clear, but,as
shown below, termination resulting in ATT' transcripts isaffected
by the source ofRNA polymerase and, under certainconditions, is
also stimulated by protein U.
Construction of Plasmid pLL226. To analyze
L4-stimulatedtermination more quantitatively, we constructed
plasmidpLL226, containing the first 165 bases of the S10 leader
fusedto a strong terminator from the rrnC rRNA transcription
unit(Figs. 1 and 3). This template allowed us to detect
read-throughtranscripts without the T1 nuclease mapping procedure,
sincemost of the transcripts extended beyond the attenuator
ter-minated at the rrnC terminator, generating read-through
mol-ecules about 285 bases long. Plasmid pLL226 is under normalin
vivo transcription regulation by LA (7).The results of a typical
transcription reaction with pLL226
are shown in Fig. 4, lanes pLL226. Total RNA synthesizedin the
cell-free system was fractionated on the gel. The majorspecies
ofRNA correspond to transcripts initiated at the S10promoter and
terminating either at the rrnC terminator (read-through RNA) or at
the attenuator (AUT and AUT' RNAs).Again, the addition of purified
L4 stimulated termination atthe S10 attenuator, resulting in a
decreased level of read-through RNAs and an increased level of
attenuated tran-scripts.
Determination of the 3' End Requirements for Attenuationin the
S10 Leader. r-Protein L4 regulates not only transcrip-tion but also
translation of the S10 operon. Our in vitrotranscription
experiments show that ribosomes are not re-quired for L4 regulation
of the S10 attenuator. Nevertheless,we wanted to eliminate from
plasmid pLL226 the distalregion of the leader to confirm that
sequences involved ininitiation oftranslation of the proximal gene
ofthe operon arenot required for transcription control. We were
also inter-ested in determining more precisely the 3' end
requirementsfor the S10 attenuator. For these reasons, we isolated
deriv-
Probe:
pLF1 Ado
pSma2 PSIO
pLL226 Ps10-U-f
0
0
0in
co-J
Proc. Natl. Acad. Sci. USA 87 (1990)
Dow
nloa
ded
by g
uest
on
July
2, 2
021
-
Proc. Natl. Acad. Sci. USA 87 (1990) 2677
Attenuator hairpin
A 1009CG
Ac AU AUGUA *-1B UAAU. UACG GUC G 28B GCCG
C
UG AU
UA
Cu c]56BG4U'] - ATTCG
A
UA A GCG UG GA A C U
100 SCG|4. UA CAU U UAAUACC
UACG7B+UA
150~GU
AU GC 26BA U C A
AG G U A 4-26BaU U CG AUG U CG GUCG GC .AuJ 4- ATUA AU AU4B
UUAUA
GC 39B 30B1 27B 25BGC CGU'A
GC* C AC U U UAUAAAAUAATTGGAGCUCgguacccggggaucc...PPP G CAU GC 0
I
GC CG 150
GC UA Sstuscte Sma IsiteUA §CGUUA C
CGCG ATT'
A Cs A CU U
FIG. 3. Secondary structure of the 510 leader. Determination of
the secondary structure has been described (8). Bases from the S10
leaderare in uppercase letters; bases from the plasmid carrying the
rrnC terminator are in lowercase letters. The hairpin structure
involved in LAregulation of transcription is indicated. We have not
directly determined the secondary structure of the pLL226
transcript, but we assume theleader structure through the first 150
bases is unchanged. The rest of the sequence is shown as
unstructured RNA. Termination sites within theS10 leader (ATT and
ATT') are indicated; major sites follow the boldfaced Us. Endpoints
of BAL-31 deletions are also indicated. TheShine-Dalgarno sequence
for the S1O gene (GGAG) and pertinent restriction sites are
indicated. Sequences that are also found in 23S rRNAat the proposed
L4 binding site include bases 117-125, 146-150, 152-155, and
159-163. (Inset) Lower half of the attenuator hairpin of
deletionmutant 26B. The sequence downstream of the deletion is
indicated by boldfaced letters.
atives of pLL226 deleted for various extents at the 3' end ofthe
leader (Figs. 1 and 3). The results of transcriptionreactions with
these mutant templates are shown in Fig. 4.Deletion mutants 25B,
27B, 44B, 30B, and 39B, all containingthe intact attenuator hairpin
(Fig. 3), gave results indistin-guishable from the results with the
wild-type parent plasmidpLL226. That is, addition of protein L4
resulted in anincreased level of attenuated transcript and a
decreased levelof read-through transcript. Deletion mutants 2B, 7B,
56B,1B, and 28B, which lack portions of the attenuator
hairpin,showed no prematurely terminated transcripts,
indicatingthat these mutants lack sequences necessary for
attenuation.Mutant 26B gave intermediate results (see below).
Quantitation of L4-Stimulated Termination of Transcrip-tion. To
quantitate the effect ofL4 on the various derivativesof pLL226, the
amount of radioactivity in the read-throughand attenuated RNAs was
measured. The results are sum-marized in Table 1. All templates
containing an intact atten-uator hairpin (pLL226 through 39B)
showed a basal level oftermination that generated ATT RNA at 10-15%
of totalRNA. Addition of L4 stimulated attenuation 2- to
3-fold,concurrent with a decrease in the amount of
read-throughtranscripts. L4 had no measurable effect on termination
thatgenerated ATT' RNA. These in vitro results suggest that
theproximal 149 bases ofthe S10 leader contained in mutant 39Bare
sufficient for L4-stimulated termination of
transcription,consistent with our in vivo analysis of this mutant
(7).Mutant 26B showed a very low basal level of termination
that generated ATT RNA (Table 1) as well as an altered band
pattern in that region of the gel (Fig. 4).
Nevertheless,addition of L4 led to increased transcription
termination inthe region of the attenuator. These results suggest
that theterminator is weakened by a deletion extending up to base
139of the leader but has residual function and can still respondto
L4. These results may be explained by the fact that thedeletion in
mutant 26B brings in a downstream DNA se-quence that restores the
two internal uridines in the 4-baseuridine string at the
attenuation point (see Fig. 3 Inset).
Deletions extending further upstream of the
L4-stimulatedtermination site, which partially or completely
destroy theattenuator hairpin, abolished the attenuator function.
Theamount of radioactivity in read-through RNAs was not
sig-nificantly decreased after addition of L4 to reactions
con-taining mutants 2B, 7B, 56B, 1B, or 28B (data not
shown).However, we were surprised that IA still had a subtle
butreproducible effect on these deletion plasmids. Namely, inthe
presence of L4 there was a small change in the pattern ofbands
corresponding to the read-through transcripts termi-nating at the
rrnC terminator, such that there was an increasein the intensity of
the lower band (Fig. 4). This effect was notobserved with deletions
leaving a functional attenuator.NusA Protein Is Necessary for in
Vitro Attenuation. Since
control of transcription termination often involves
auxiliaryproteins, we wanted to determine if any known
transcriptionfactors could influence the efficiency (and possibly
the pre-cise site) oftermination in vitro. Our initial experiments
wereperformed with a partially purified RNA polymerase. Wefound
that with a more highly purified RNA polymerase
Biochemistry: Zengel and Lindahl
Dow
nloa
ded
by g
uest
on
July
2, 2
021
-
2678 Biochemistry: Zengel and Lindahl
DNA:
L4:
RT-
cD BAL-31 deletion plasmids
n cmcom m m m m ml-j 10 ,-. '4 0) 0) wcDf En mCO c0. C) N . c C
C C'.J cN 10O- Co
-±+-+ -±-++ -+-+-+-+-++ -+
4-471
-377
-284
-210
4-150ATTLATT--
FIG. 4. Effect of LA on in vitro transcription from
plasmidpLL226 and its BAL-31 deletion derivatives. The
transcriptionreaction mixtures were as described in Fig. 2, except
each mixturecontained 4 A&g of partially purified RNA
polymerase, 30 ,uCi ofUTP[35S], and 0.25 ,ug of L4 or the
equivalent volume of proteinbuffer. Positions of read-through (RT)
and attenuated (ATT andATT') transcripts are shown on the left.
Positions of size markers areindicated on the right. Read-through
RNA from mutant 7B migratesfaster than RNAs from mutants with
similarly sized leaders becausethe deletion in this mutant extends
about 20 bases downstream of theSty I site that marks the 3'
boundary ofdeletions in the other mutants(Fig. 1).
preparation, there was no termination at the S10 attenuator,even
in the presence of L4, unless NusA protein (14) wasadded (Fig. 5).
Under the same reaction conditions, otherNus factors, including
NusB (15, 16), NusG (17), and NuOE(18, 19), had no significant
effect on transcription of the S10leader (Fig. 5). Termination
factor p (20) also had no effect(data not shown). These results
suggest that the attenuator inthe S10 operon is a NusA-dependent
terminator whoseefficiency is stimulated by r-protein L4.
Since in our initial transcription experiments (Figs. 2 and4) we
observed L4-mediated attenuation without addition ofNusA, we
surmised that our original RNA polymerase prep-aration might
contain NusA. Gel electrophoretic analysisconfirmed that this RNA
polymerase is contaminated with asmall amount of other proteins,
including NusA (data notshown). We conclude that there is a
sufficient amount ofNusA to mask the requirement for this factor. A
stableassociation between RNA polymerase and NusA during
thepurification of the enzyme has been reported (21).
In the transcription reactions with the more highly purifiedRNA
polymerase, we observed less termination that gener-ated ATT'
transcript than we had observed with the originalRNA polymerase
(compare Fig. 4 with Fig. 5). Moreover, theAlT' transcripts
exhibited the same pattern as the AlT
Table 1. Quantitation of the effect of r-protein L4 onin vitro
attenuation
% total RNAs
DNA L4 AlT ATT' RTpLL226 - 10 32 59
+ 33 33 34+/- 3.3 1.0 0.58
25B - 11 29 60+ 32 29 39
+/- 2.9 1.0 0.6527B - 14 34 52
+ 30 32 38+/- 2.1 0.94 0.73
44B - 10 36 54+ 34 34 32
+/- 3.4 0.94 0.5930B - 9 31 60
+ 28 31 41+/- 3.1 1.0 0.68
39B - 13 26 61+ 29 27 45
+/- 2.2 1.0 0.7426B - 4 8 88
+ 15 6 79+/- 3.8 0.75 0.90
Radioactivity in the bands corresponding to attenuated (AUT
andATT') RNAs and read-through (RT) RNAs terminating at the
rrnCterminator was determined (1) from the gel shown in Fig. 4.
Valuesfor ATT and AlT' RNAs were corrected for background by
sub-tracting the cpm in the corresponding areas of the gel from
samplesnot terminating at the attenuator. These corrected values
and theuncorrected values for read-through RNAs were then
normalized tothe number ofuridines in the transcript (38 for AlT,
41 for AlT', and49-76 for read-through transcripts). Each
nofinalized value was thendivided by the total normalized cpm (ATT
plus ATT' plus RT) togenerate the percent ofRNA polymerases
terminating at a given siteto yield the indicated RNAs. The effect
of L4 was calculated bydividing the value in the + L4 reaction by
the value in the - L4reaction (+/-). Addition of L4 had no
significant effect on the totalnumber of transcripts from a given
template (data not shown).
transcripts: stimulation by L4 but only in the presence ofNusA.
We cannot yet account for these polymerase-dependent variations in
termination that generates ATT'transcripts, nor can we explain why
these transcripts are not
NusE(S10) + - + - + I- 1-NusB& NusG + - + -
NusA +L4 + + + + + + +I+
RT-4
ATT'_-so wATT_*
FIG. 5. Effects ofNus factors on attenuation. Each 20-/.l
reactionmixture contains 0.18 lug (20 nM) of purified RNA
polymerase, 2 ,ug(50 nM) of 39B DNA, 20 pCi of UTP[35S], and, where
indicated, 80nM NusB, 80 nM NusG, 80 nM NusE, and 160 nM L4 (+) or
S7 (-).The positions of read-through (RT) and attenuated (ATT and
ATT')transcripts are indicated.
Proc. NatL Acad. Sci. USA 87 (1990)
Dow
nloa
ded
by g
uest
on
July
2, 2
021
-
Proc. Natl. Acad. Sci. USA 87 (1990) 2679
detected with in vivo-synthesized RNA. Perhaps ATT' tran-scripts
are also formed in vivo but are rapidly shortened bynucleases to
the same size as ATT molecules.
DISCUSSIONThe S10 operon attenuator functions as a terminator in
asimple cell-free transcription system containing RNA poly-merase
and transcription factor NusA. Furthermore, thelevel of termination
is increased by the addition of r-proteinL4. The site of
termination and the stimulatory effect of L4are consistent with our
in vivo data showing that oversyn-thesis of r-protein L4 leads to
an increased level ofprematuretermination at the S10 attenuator (5,
7). The relatively simplein vitro requirements for attenuator
function together with theresults using the BAL-31 leader-deletion
mutants confirm ourconclusion from in vivo genetic and
physiological studies thattranscription control by L4 is achieved
by a mechanism thatis independent of the inhibition of translation
by L4. Eventhough there is excellent agreement between the in vitro
andin vivo experiments, it should be pointed out that the in
vitroexperiments cannot distinguish between very strong pausingat
the attenuator and actual release of the transcript.The nusA gene
product was originally identified as a factor
required for N-dependent antitermination in bacteriophage A(14).
It has since been shown in a variety of transcriptionsystems that
NusA enhances RNA polymerase pausing and,depending on the
transcription unit, causes an increase or adecrease in termination
(for review, see ref. 22). Thesecomplex effects of NusA suggest it
may function as atranscription "fidelity" factor, interacting with
RNA poly-merase and other termination factors to modulate the
enzymeresponse to pause and termination sites (23). Although
NusAhas not been shown to be absolutely essential for in
vivoregulation of any bacterial gene, our in vitro
experimentssuggest that NusA is required for termination (or a very
stablepause in elongation) at the S10 attenuator and, therefore,
isa necessary cofactor in L4-mediated regulation of transcrip-tion
of the S10 operon. Several lines of evidence suggest thatthe boxA
sequence, CGCTCTTA, is involved in NusA reg-ulation of A gene
expression (24, 25). Even though similarsequences have been found
in other operons whose tran-scription is affected by NusA,
including rRNA transcriptionunits (26, 27), the precise role of
boxA-like sequences inNusA function is still not clear. In fact, in
some cases boxAseems to be dispensable for NusA activity (see,
e.g., refs.28-31). The S10 leader has no boxA sequence, although
itdoes have a sequence, AACAAT (bases 61-66; Fig. 3), whichis
homologous to the recently extended portion ofboxA (32).Whether
that sequence plays any role in NusA regulation ofthe S10 operon
needs to be determined.The BAL-31 deletion analysis suggested that,
in the ab-
sence of a functional attenuator, L4 still affected
terminationof transcription, but at a terminator further
downstream. Thestimulation of attenuation and the effect at the
rrnC termi-nator seem to be mutually exclusive, since we observed
thedownstream effect only when the deletion abolished attenu-ation.
One model to explain these results is that r-protein L4programs a
fraction of the RNA polymerase molecules toterminate at the S10
attenuator. If the attenuator is defectiveor deleted, the same
modified RNA polymerases respond ina different way to the rrnC
terminator. This model needs tobe tested experimentally, but it
does raise two interestingpossibilities. (i) The target for L4 may
lie upstream of thedeletion endpoint in mutant 1B, that is,
upstream of base 119.This implies that L4 recognition does not
require the hairpinstructure involved in attenuation or the leader
sequenceshomologous to sequences in the L4 binding domain of
23SrRNA (see Fig. 3), which were originally thought to benecessary
for L4 control (6, 33). (ii) r-Protein L4 can mod-
ulate termination at not only the S10 attenuator but also
therrnC terminator. Whether this effect is simply an
artifactresulting from the fusion of the S10 leader to a
foreignterminator or an indication that L4 plays a more general
rolein regulating transcription of other operons in E. coli
remainsto be seen. In any event, it is interesting that a
singler-protein, L4, plays three important roles in the cell: as
acomponent ofthe 50S subunit ofthe ribosome, as aregulatoryprotein
causing decreased translation of its own transcript,and as a
termination factor regulating transcription of its ownoperon.
We thank T. Platt, M. Nomura, J. Greenblatt, K. Nierhaus,
P.Nowotny, and E. Morgan for generous gifts of proteins, and
W.McClain for the TDF1 plasmid. We also thank E. Grayhack
forcritical reading of this manuscript. This research was supported
bya grant to L.L. from the National Institute of Allergy and
InfectiousDiseases.
1. Lindahl, L. & Zengel, J. M. (1979) Proc. Natl. Acad. Sci.
USA 76,6542-6546.
2. Zengel, J. M., Mueckl, D. & Lindahl, L. (1980) Cell 21,
523-535.3. Yates, J. L. & Nomura, M. (1980) Cell 21, 517-522.4.
Freedman, L. P., Zengel, J. M., Archer, R. H. & Lindahl, L.
(1987)
Proc. Natl. Acad. Sci. USA 84, 6516-6520.5. Lindahl, L., Archer,
R. & Zengel, J. M. (1983) Cell 33, 241-248.6. Lindahl, L. &
Zengel, J. M. (1988)-in Genetics of Translation, eds.
Tuite, M. F., Picard, M. & Bolotin-Fukuhara, M. (Springer,
Berlin),pp. 105-115.
7. Zengel, J. M. & Lindahl, L. (1990) J. Mol. Biol. 211, in
press.8. Shen, P., Zengel, J. M. & Lindahl, L. (1988)
NucleicAcids Res. 16,
8905-8924.9. Freedman, L. P., Zengel, J. M. & Lindahl, L.
(1985) J. Mol. Biol.
185, 701-712.10. McClain, W. H., Guerrier-Takada, C. &
Altman, S. (1987) Science
238, 527-530.11. Normanly, J., Masson, J.-M., Kleina, L. G.,
Abelson, J. & Miller,
J. H. (1986) Proc. Nat!. Acad. Sci. USA 83, 6548-6552.12.
Burgess, R. R. & Jendrisak, J. J. (1975) Biochemistry 14,
4634-
4638.13. Zengel, J-. M., McCormick, J. R., Archer, R. H. &
Lindahl, L.
(1990) in Ribosomes and Protein Synthesis: A Practical
Approach,ed. Spedding, G. (Oxford Univ. Press., Oxford), in
press.
14. Friedman, D. I. & Gottesman, M. (1983) in Lambda II,
eds.Hendrix, R. W., Roberts, J. W., Stahl, F. W. & Weisberg, R.
A.(Cold Spring Harbor Laboratory, Cold Spring Harbor, NY),
pp.21-51.
15. Strauch, M. & Friedman, D. I. (1981) Mol. Gen. Genet.
182,498-501.
16. Swindle, J., Ajioka, J. & Georgopoulos, C. (1981) Mol.
Gen. Genet.182, 409-413.
17. Horwitz, R. J., Li, J. & Greenblatt, J. (1987) Cell 51,
631-641.18. Friedman, D. I., Schauer, A. T., Baumann, M. R., Baron,
L. S. &
Adhya, S. L. (1981) Proc. Natl. Acad. Sci. USA 78, 1115-1119.19.
Das, A., Ghosh, B., Bank, S. & Wolska, K. (1985) Proc.
Nat!.
Acad. Sci. USA 82,4070-4074.20. Platt, T. (1986) Annu. Rev.
Biochem. 55, 339-372.21. Schmidt, M. C. & Chamberlin, M. J.
(1984) J. Biol. Chem. 259,
15000-15002.22. Yager, T. D. & von Hippel, P. H. (1987) in
Escherichia coli and
Salmonella typhimurium: Cellular andMolecularBiology, ed.
Neid-hardt, F. C. (Am. Soc. Microbiol., Washington), pp.
1241-1275.
23. Greenblatt, J. & Li, J. (1981) Cell 24, 421-428.24.
Friedman, D. I. & Olsen, E. R. (1983) Cell 34, 143-149.25.
Schauer, A. T., Carver, D. L., Bigelow, B., Baron, L. S. &
Fried-
man, D. I. (1987) J. Mol. Biol. 194, 679-690.26. Holben, W. E.
& Morgan, E. A. (1984) Proc. Natl. Acad. Sci. USA
81, 6789-6793.27. Li, S., Squires, C. L. & Squires, C.
(1984) Cell 38, 851-860.28. Chen, C.-Y. A. & Richardson, J. P.
(1987) J. Biol. Chem. 262,
11292-11299.29. Schmidt, M. C. & Chamberlin, M. J. (1987) J.
Mol. Biol. 195,
809-818.30. Sigmund, C. D. & Morgan, E. A. (1988)
Biochemistry 27, 5628-
5635.31. Zuber, M., Patterson, T. A. & Court, D. L. (1987)
Proc. Nat!.
Acad. Sci. USA 84, 4514-4518.32. Morgan, E. A. (1986) J.
Bacteriol. 168, 1-5.33. Olins, P. 0. & Nomura, M. (1981) Cell
26, 205-211.
Biochemistry: Zengel and Lindahl
Dow
nloa
ded
by g
uest
on
July
2, 2
021