Gene Expression in Escherichia coli After Amino Acid, Purine, or ...

4 downloads 0 Views 2MB Size Report
expression to amino acid insufficiency is only partially mimicked by purine or pyrimidine insufficiency. (iii) Transcription initiation atpLlo decreases in response.
JOURNAL OF BACTERIOLOGY, Apr. 1980, p. 202-211

Vol. 142, No. 1

0021-9193/80/04-0202/10$02.00/0

Gene Expression in Escherichia coli After Amino Acid, Purine, or Pyrimidine Exhaustion ROBERT M. BLUMENTHALt AND PATRICK P. DENNIS* Department of Biochemistry, Faculty of Medicine, University of British Columbia, Vancouver, British Columbia, Canada V6T 1 W5

Three strains of Escherichia coli B auxotrophic for leucine, guanine, or uracil were analyzed after exhaustion of the respective required nutrient from the growth medium. The pattern of transcription was analyzed by ribonucleic aciddeoxyribonucleic acid filter hybridization to specific deoxyribonucleic acid probes, and the pattern of translation was analyzed by autoradiography after the resolution of proteins on sodium dodecyl sulfate-polyacrylamide gels. The results obtained suggest the following conclusions. (i) Specific regulation of rpoBC transcription occurs at both the promoter (pLuo) and the putative attenuator between rpIL and rpoB. (ii) The stringent response of ribosomal protein gene expression to amino acid insufficiency is only partially mimicked by purine or pyrimidine insufficiency. (iii) Transcription initiation atpLlo decreases in response to guanine exhaustion, but in contrast increases significantly in response to uracil exhaustion. (iv) The expression of the induced lac operon is severely depressed during any of these exhaustions. These conclusions argue against simple models for regulation of ribonucleic acid polymerase production or promoter choice by the intracellular levels of its substrate nucleotides. The discovery and analysis of the stringent control system in Escherichia coli has been a major advance in understanding how cells integrate the functioning of their various physiological domains. This control system integrates the physiological domains of amino acid biosynthesis and protein synthesis through the actions of the relA gene product (stringent factor) and the unusual guanine nucleotides ppGp, ppGpp, and pppGpp (4, 41, 48). In contrast to the mutant "relaxed" strains, strains with functional stringent control systems respond to an amino acid insufficiency by sharply reducing the production of ribosome components while directly stimulating the expression of amino acid biosynthetic genes (12, 17, 48, 49). Furthermore, stringent strains have been found to differ from relaxed strains in the important respects of having lower tranalation error rates during amino acid insufficiency and of more qluickly achieving new steady states of growth after a change in the chemical composition of the growth medium (16, 40; F. S. Young and F. C. Neidhardt, personal communication). In contrast, relatively little is known about the integration of the physiological domain of the common adenine, cytosine, guanine, thymine, and uracil nucleotides with the other physiological domains of cells: there are two questions t Present address: Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724.

which stand out in particular. First, what is the relationship between the stringent control system and the pools of the common nucleotides? There are indications that the two may be intimately related (39). Second, what mechanisms govern the expression of the RNA polymerase structural genes? Do the nucleoside triphosphate (NTP) levels influence RNA polymerase production in a manner analogous to the influence of amino acid levels on ribosome production? It seems clear that any sophisticated understanding of the integration of cell physiology will require a detailed understanding of the regulation of RNA polymerase production and activity. The bulk of all known systems governing gene expression in bacteria appear to control transcript production. Some things are known about the regulation of RNA polymerase production. In particular, the genes rpoB and rpoC, which specify the /9 and ,B' subunits of RNA polymerase, are reported to be cotranscribed with the genes rplJ and rplL, which specify the ribosomal proteins L10 and L7/12; the gene order is promoter (pLjo), rplJ, rplL, rpoB, rpoC (38,55). Under normal conditions of steady-state growth, only about 20% of the transcripts which have been made as far as rpIL are continued into rpoB, implying that transcriptional attenuation occurs between rpIL and rpoB (10); this pattern of 20% read-through agrees well with the observed 1:5 ratio of RNA polymerase to ribo202

VOL. 142, 1980

AMINO ACID OR NUCLEOTIDE BASE EXHAUSTION

somes in normally growing E. coli (8, 33). What regulatory signals affect transcription initiation at pL1o, and which (if any) affect the degree of transciptional read-through from rpIL into

rpoB? In an attempt to address some of these questions, we employed three strains of E. coli B auxotrophic for leucine (NF314), guanine (NF522), or uracil (U32) and examined the transcription and translation of several genes, including RNA polymerase structural genes, after exhaustion ofthe respective required nutrient from the growth medium. MATERIALS AND METHODS Bacterial strains and growth conditions. The three strains which we used are all derivatives of E. coli B. Strain NF314 (Leu-) was previously obtained from N. Fiil (15), and strain NF522 (Gua-) was obtained from L. Lindahl. Strain U32 (Ura-) was isolated by P. Dennis. All three strains are reUA+. All cultures were grown at 37°C in MOPS media (36) supplemented with 0.2% glycerol and 1 mM isopropyl-ft-D-thiogalactoside (IPTG), a gratuitous inducer for the lactose operon (45). Each required nutrient (leucine, guanine, or uracil) was added to an overnight culture at 10 times the desired final concentration; the exponentially growing culture was later diluted 1:10 into prewarmed MOPS-glycerol-IPTG medium. The final concentrations of leucine, guanine, and uracil are given in the legend to Fig. 1; the control exponential cultures were grown in the presence of 10 times those levels of leucine, guanine, and uracil. Analysis of transcription. To label cultures for an analysis of transcription, 10-nil portions were withdrawn to prewarmed flasks and pulsed for 1 min with a mixture of [2-3H]adenine (17.8 Ci/mmol; 10 ACi/ml) and one of the following required nutrients: leucine (108 jug/ml), guanine (5 ,ug/ml), or uracil (5 iLg/ml). After the pulse the RNA was prepared as described previously (11), diluted to an optical density at 260 nm of not more than 1.0 (20 to 50 yig/ml), and analyzed by hybridization by using duplicate nitrocellulose filters, each bearing about 170 fmol of specific denatured and immobilized DNAs (10, 11, 22). The hybridization probes have been described previously (see Table 1) with the exception of pSPll, which carries the 1.4% HindIII fragment of A fus; this fragment comes from the rpoA gene region and carries about 60% of that gene (S. Pedersen, personal communication). Total radioactivity in the RNA preparations was determined after precipitation in 5% trichloroacetic acid. Analysis of translation. To label cultures for an analysis of translation, 10-ml portions were withdrawn to prewarmed flasks and pulsed for 2 min with a mixture of L-[3S]methionine (1,350 Ci/mmol; 1 uCi/ ml) and the required nutrient (as in the transcription analysis). After the pulse, excess unlabeled methionine and cysteine were added, and the cells were incubated for an additional 3 min. The protein was prepared as described previously (29) and diluted with an unlabeled protein preparation (from strain NF314) so that all samples had the same raeioactivity per unit volume when checked by trichloroacetic acid precipi-

203

tation; sodium dodecyl sulfate-polyacrylamide gels were run with equal volumes, equal radioactivities, and roughly equal protein concentrations loaded for each sample. Each fixed and dried gel was subjected to autoradiography on Kodak NS2T X-ray film, and the autoradiograms were subjected to densitometry by using a Quick-Scan Jr. densitometer (Helena Instruments, Beaumont, Tex.) equipped with an integrator.

RESULTS

Growth of the strains. As Fig. 1 shows, E. coli B strains NF314 (Leu-), NF522 (Gua-), and U32 (Ura-) had similar exponential growth rates in the MOPS-glycerol-IPTG media supplemented with the respective required nutrients. Furthermore, when the required nutrient was 1

,'~ l

I

64

U32(URA-) 32

/A

0

I 16 NF522(GUA-)

=

8 ._ 0

4

_g_N ,LU NF314(LEU 1

,/

*+ 2

'I

I~~~~~~~

I

-60 -30

0

+60 +90 +120 Min. after exhaustion of required nutrient +30

FIG. 1. Effects of leucine, guanine, and uracil exhaustions on the growth of auxotrophic strains of E. coli. Cells were grown at 370 C in MOPS-glycerolIPTG medium supplemented as required to give 5.3 pg of L -leucine per ml (40.0 p), 2.0 pg ofguanine per ml (13.2 M), or 2.0 pg of uracil per ml (17.8 pM). Growth was followed for approximately 3 h before nutrient exhaustion. The accumulation of cellular mass was monitored by measuring the absorption at 460 nm, and the actual absorption at 460 nm at the apparent point of exhaustion was 0.35 for NF314, 0.29 for NF522, and 0.38 for U32. The preexhaustion doubling times were 64 min for NF314, 76 min for NF522, and 57 min for U32. The arrows indicate the times used as zero time during each exhaustion and are in all cases very close to the actual break in the growth curve. These were the same cultures which were used in the analyses of transcription and translation.

204

BLUMENTHAL AND DENNIS

J. BACTERIOL.

exhausted from a growth medium, all three strains showed similar 80 to 85% decreases in growth rate, although the transition took somewhat longer after uracil exhaustion. The continued growth at reduced rates indicates that the three strains are mildly leaky regarding their respective auxotrophies. Patterns of transcription. For each of the three strains, one sample of an exponentially growing culture and three samples taken after nutrient exhaustion were pilse-labeled for 1 min with [2-3H]adenine. In each case, the exhausted nutrient was readded concomitant with the pulse (see below). The RNA from each sample was prepared and analyzed for specific pulselabeled mRNA's by using the DNA hybridization probes listed in Table 1. The names used for the various rnRNA species are also indicated in Table 1. The results of these transcription analyses are presented in Table 2 and Fig. 2 and 3. When growing exponentially, the three strains had comparable transcription profiles, although the level of Lac mRNA was somewhat elevated in strain NF522 (Gua-). In each case, the spcl, a, and L10-L7/12 mRNA's were produced in roughly equal molar amounts, as expected (10, 11). Also, in each case the transcriptional readthrough from rpIL into rpoB (judged from the mole ratio of f,8f' mRNA to L10-L7/12 mRNA) was 15 to 25%, which is in agreement with previous measurements (10). In contrast to the similar transcription profiles of the unlimited exponential cultures, the three auxotrophic strains responded in clearly distinct ways when the respective required nutrient was exhausted from a growth medium. Strain NF314 (Leu-), which, like the others, has a functional stringent control system, gave the expected stringent response; production of spcl mRNA, L10-L7/12 mRNA, as well as a mRNA, was

quickly reduced by 70% or more. Production of Lac mRNA also decreased after leucine exhaustion, probably due in part to catabolite repression (27). In comparison, the production of /if' mRNA was only depressed by about 50%. Thus, the amount of transcriptional read-through from rpIL into rpoB must increase several-fold during the response to leucine exhaustion, possibly to 100%. Strain NF522 (Gua-) responded to guanine exhaustion with moderate 30 to 60% decreases in the production of spcl, L10-L7/12, and a mRNA's and a sharp reduction in the synthesis of Lac mRNA. The production of ,B,' mRNA was also depressed by about 50%; the transcriptional read-through from rpiL into rpoB increased relatively little. TABLE 2. Preexhaustion patterns of transcription in the three auxotrophic strains % of input radioactivity in speInput ra- cific mRNA-DNA hybrdab dioactivity - Strain 1 Stan (cpm/50 pl L10of RNA)' spcl a L7/ ffi Lac NF314 (Leu-) 92,800 1.73 0.18 0.08 0.28 0.20 NF522 (Gua-) 108,800 2.00 0.16 0.14 0.30 0.50 U32 (Ura-) 47,100 1.80 0.16 0.08 0.24 0.32 aEach hybridization value is the result of four assays in which 50-, 100-, 150-, and 200-jd amounts of the RNA preparations were used. The background values (already subtracted) ranged from 20 to 40 cpm/ filter and reflect nonspecific hybridization (see legend to Fig. 2). b The mRNA-DNA hybrids are described in Table 1, and the hybridization protocol is described in the legend to Fig. 2. The strains had been growing exponentially for several generations in the presence of 53 ,ug of L-leucine per ml, 20 ug of guanine per ml, or 20 Mg of uracil per ml, as required, when they were exposed to [3H]adenine for 1 min.

TABLE 1. Description of the DNA probes used and the mRNA species analyzed Chromo-

DNA probe

Xdspcl

mRNA hybridized'

Genes carried

Ml

(X103) b

somnal location (min) 72

Reference

rpoA, 15 genes for ribo300 14 somal proteins a 60% of rpoA pSPll 25 72 S. Pedersenc pNF1561 L10-L7/12 23 88 lOa rplJ, rplL pJC720 ,BB' rpoB, rpoC 327 88 7 Lac Aplac5 lacZ, part of lacY 120 22 47 a These names for the various mRNA species are used throughout. In the case of spcl mRNA, more than one individual mRNA species was measured, but all spcl mRNA corresponded to the genes indicated. b Mr is the approximate aggregate molecular weight of the bacterial protein(s) (or parts of proteins) coded for by the DNA of each probe. 'S. Pedersen, personal communication. spcl

AMINO ACID OR NUCLEOTIDE BASE EXHAUSTION

VOL. 142, 1980

205

r _u

z z I

0 c

0 0. C

N

c

z

v in c

F 0t

6-

I'

0

a'

a C4 m

0

15 30 45

0

15 30 45

0

15 30 45

0

15 30 45

Time (min.)

FIG. 2. Effects of leucine, guanine, and uracil exhaustions on the transcription patterns in auxotropic strains of E. coli. The three strains were grown as described in the legend to Fig. 1. At various times after apparent exhaustion of each required nutrient, both the nutrient and [3HJadenine were added to a 10-mi portion of the culture for I min. The reference (zero time) cultures were the exponential cultures described in Table 2. The hybridization probes used and mRNA species measured are described in Table 1. Each solid circle represents the results of at least four assays, each with at least duplicate filters. The data were corrected for background hybridization (20 to 40 cpm/filter; hybridization to filters bearing AC1857S7 DNA in all cases except spcl mRNA, which was also corrected for hybridization to filters bearing Xdtrk DNA [11]). The data were then divided by the total incorporation for that RNA preparation and by the Mr (Table 1) to obtain frequencies of transcription. These values were then normalized to the preexhaustion values for spcl mRNA, and the ordinates of each panel were adjusted so that the preexhaustion values for each mRNA are at the same height. The open circles represent maximum estimates of what the 9/3,' mRNA values would have been if there had been no increases in transcriptional polarity during the exhaustions. The polarity can be gauged by examining the /,B/' synthesis rate ratio (Fig. 4). Assuming that any induced polarity is evenly distributed over rpoBC and that transcriptional and translational polarity are tightly linked to one another, a maximum estimate of the true relative transcription frequency of rpoBC can be made by multiplying the measured /3,/' mRNA value by the normalized ,B/1' synthesis rate ratio for that time point. This is an approximation, as (i) the lengths of rpoB and rpoC are not quite equal (rpoB = 0.94 rpoC) and (ii) it is a linear approximation of an exponential function which can be used only at low ,8//1' synthesis rate ratios (it never exceeded 3 in this work).

Strain U32 (Ura-) responded to uracil exhaustion with a limited 10 to 40% decrease in the production of spcl and a mRNA's and a sharp reduction in Lac mRNA synthesis. In complete contrast to the reduced transcription of the other ribosomal protein genes, the production of L10-L7/12 mRNA increased significantly. The depression of ,B,B' mRNA production was strik-

ingly mild, seemingly as a result of increased initiation atpLuo rather than an increase in transcriptional read-through from rpIL into rpoB. Patterns of translation. For each of the three strains, using the same cultures that were used in the transcriptional analysis, one exponential and three postexhaustion culture samples were pulse-labeled for 2 min with [35S]me-

206

BLUMENTHAL AND DENNIS B

A

-C

J. BACTERIOL. C

-0 0

0

io 0

09

\

O

15 30 45

/

0

15

0.0%% 30 45

0

15

30

.'

2.0-

0~~~~~~~~~~,~ 45

Time(m in.) FIG. 3. Effects of leucine, guanine, and uracil exhaustions on the putative rplL-rpoB attenuator in auxotrophic strains of E. coli. The percent readthrough of the putative attenuator between rplL and rpoB was calculated from the data in Fig. 2 as 100 times the /,3/' mRNA frequency divided by the L10L7/12 mRNA frequency. This is shown as either uncorrected (0) or corrected (0) for the effects of apparent intragenic polarity with rpoBC, as described in the legend to Fig. 2. (A) NF314 (Leu-); (B) NF522 (Gua-); (C) U32 (Ura-).

thionine and chased for 3 min with excess unlabeled methionine and cysteine. In each case, as in the transcription analysis, the exhausted nutrient was readded concomitant with the pulse. After the chase period the samples were prepared for and subjected to electrophoresis on sodium dodecyl sulfate-polyacrylamide gels; autoradiograms of those gels were subjected to densitometry. The following three proteins were examined: the ,B and /3' subunits of RNA polymerase and the subunit of the homotetrameric enzyme ,B-galactosidase. The results of the translational analyses are shown in Fig. 4 and 5. When growing exponentially, the three strains again showed comparable profiles. Again, the three strains responded in clearly distinct ways after exhaustion of the respective required nutrients. Strain NF314 (Leu-) showed decreased production of (3-galactosidase. The synthesis of the RNA polymerase ,B subunit was somewhat elevated, whereas that of the /t' subunit apparently was maintained at a constant level. The resulting 3/,f' synthesis ratios showed an increase of about 50%, which is suggestive of polarity since the rpoB (,B) gene is promoter proximal to the rpoC (,/') gene. Strain NF522 (Gua-) responded to the guanine exhaustion with a decreased rate of /8-galactosidase synthesis, a delayed decrease in the production of subunit protein, and an immediate decrease in the production of the ,B' subunit of RNA polymerase. Accordingly, the /3/f,' synthesis ratio increased by about 50%, again implying an increased degree of intragenic polarity. Strain U32 (Ura-) responded to the uracil exhaustion with reduced /3-galactosidase pro-

O ls 30 45

0

0

*3.0

2.0

1.0

0

15

30

45

0 15

30

45

0

15

30

45

0

15

30

45

Time (min.)

FIG. 4. Effects of leucine, guanine, and uracil exhaustions on the syntheses of specific proteins in auxotrophic strains of E. coli. The three strains were grown as described in the legend to Fig. 1. At various times after apparent exhaustion of each required nutrient, both the nutrient and [35SJmethionine were added to a 5-mi portion of the culture. After 2 min, excess unlabeled methionine and cysteine were added, and incubation was continued for 3 min. The reference (zero time) cultures were the exponential cultures described in Table 1. Each protein preparation was resolved on sodium dodecyl sulfate-polyacrylamide gels, and autoradiograms of the gels were subjected to scanning densitometry. Equal total amounts of radioactivity were applied to the gels for each sample. 8-RNAP and /3'-RNAP refer to the ,B and8 'subunits of RNA polymerase, respectively, and ,8/,B' refers to the synthesis rate ratio of those subunits. The ./3/f' ratio, in contrast to the other data in this figure, waas not normalized. The /3-galactosidase band on the gels was not pure; gels of uninduced cells had a band in that position with 15 to 20% the intensity of the band in IPTG-induced cells.

duction, an increase in ,8 subunit synthesis, and a limited decrease in ,3' subunit synthesis. The B//B' synthesis ratio underwent a substantial 2.7fold increase, which was suggestive of a considerable increase in intragenic polarity. An examination of the overall translation patterns (Fig. 5) shows that all three strains were capable of redistributing their translational (and presumably transcriptional) commitments in response to and during the nutrient exhaustions. Several proteins showed particularly interesting responses to the three exhaustions. One, which had a molecular weight of 135,000 to 140,000 (protein a), appeared after uracil or guanine exhaustion, but not after leucine exhaustion.

d_~

.':s_S;tf.yeM4oSb *SM~ ~it

AMINO ACID OR NUCLEOTIDE BASE EXHAUSTION

VOL. 142, 1980

NF314 (LEU-) 1 2 3 4 5 1

ga

B-gaIl

S e . * *,:

.

rJ.

NF522 (GUA-) 1 2 3 4 5 1

207

U32 (URA-) 1 2 3 4 5 1 e 1~~~~~~~~~ll

.:

.:

R vf

..

..:.

...

r::::

*- -

_

:iXiBs7we

--

.:

._

':"

.?*sS ::

.1-

:'af,,S

-

we.

_

.A 3"j>S

_ w

ow

O

I..

.1,;.:i "'

Oi-'

jt

t

C.-

w.i

_E

:]:."i

:.

.....

as;;!C,

_ '_ .. w .,.

...

^$*

::e. :. St:

*§: .;:AzZ

^v:

^

3# f£s

3R :e

__*jS_. 5i 2/ff : v &S fts

FIG. 5. Effects of leucine, guanine, and uracil exhaustions on the translation patterns in auxotrophic strains of E. coli. The strains were grown and labeled with [3SJmethionine as described in the legend to Fig. 4. For each strain, the well numbers refer to samples pulsed at different times, as follows: well 1, the exponential culture described in Table 2; well 2, sample taken close to zero time (see Fig. 1); well 3, sample 15 min after exhaustion; weU 4, sample 30 min after exhaustion; well 5, sample 45 min after exhaustion. The locations of the ,B and p9' subunits of RNA polymerase and the ,B-galactosidase subunit are indicated. Also indicated are the locations of four unidentified proteins (proteins a through d) which have interesting responses and are referred to in the text. Each well received 11,500 cpm, and the autoradiogram shown was exposed for almost 2 weeks.

Another, with a molecular weight of 120,000 (protein b), was not synthesized after leucine or guanine exhaustion but increased markedly after uracil exhaustion. A protein having a molecular weight of about 60,000 (protein c) differentially increased after all three exhaustions, but particularly after guanine exhaustion. Finally, it is interesting that a fairly major band corresponding to a 35,000-dalton protein (protein d) was apparently missing in the uracil auxotroph U32, even in the presence of uracil, and this may be related to the auxotrophy of that strain. DISCUSSION Experimental protocol. The three strains used in this work, NF314 (Leu-), NF522 (Gua-),

and U32 (Ura-), are all derivatives of E. coli B. All of the data presented here indicate that in the presence of their respective required nutrients, the three strains are physiologically com-

parable. For a full analysis of these strains after exhaustion of the required nutrients, it was necessary to readd each exhausted nutrient to the culture concomitant with the pulse-labeling in order to achieve sufficient incorporation of the label. It is possible that this nutrient readdition influenced the observed responses to some degree; however, (i) our aim was to measure the distribution of RNA polymerase within and among transcription units after amino acid or nucleotide base exhaustion, and (ii) it is unlikely that this distribution changed greatly within the

208

BLUMENTHAL AND DENNIS

very short (1- or 2-min) pulse-labeling times. An additional experiment with strain U32 (Ura-), in which excess uracil was not added during the pulse, gave a distribution of incorporated label similar to that shown for strain U32 in Fig. 2 (data not shown). Because of the nature of the mRNA measurements, which are related to total incorporation of label into RNA, an mRNA species may be reported to be increasing when it is actually decreasing but doing so more slowly than total RNA synthesis. The proper and significant implication of such an increase is that the cell is devoting a greater fraction of its available transcription capacity to that gene. In making such an analysis, the only necessary assumption is that whatever label enters the cell during the pulse does get randomly incorporated into all of the RNA being made. Effects of polarity. Since the pLjo transcription unit is an unusually long one (the distance from pLuo to the end of rpoC is more than 11,000 base pairs), measurements of (3(3' mRNA are particularly sensitive to the degree of intragenic transcriptional polarity. We attempted to gauge the possible effects of nonspecific intragenic polarity on the (3(3' mRNA measurements (Fig. 2 and 3). The procedure used is described in the legend to Fig. 2; basically, if one assumes that any induced polarity is evenly distributed over the length of rpoBC (53) and that translational and transcriptional polarity are tightly linked, then an approximation of what the ,B,B' measurement would have been had there been no polarity can be made by using the (3/f synthesis rate ratio (Fig. 4). An increase in the /3/fl' ratio is suggestive of increased polarity since the rpoC (,(3) gene lies promoter distal to the rpoB (,8) gene. This approximation is not without its limitations; in the leucine exhaustion, where quite possibly transcriptional and translational polarity are not so tightly linked, the "polarity-corrected" ,8,/' mRNA values are much higher than one would predict as possible, giveii the L10-L7/ 12 mRNA data (Fig. 2). Nevertheless, this is a conservative approach in that it tends to maximize the estimated polarity effects; despite that, no such estimated effects alter the conclusion that specific regulation of rpoBC transcription occurs at both the promoter pLuo and the rpILrpoB attenuator (Fig. 3). NTP pools and the stringent response. In addition to the direct effects of the unusual guanine nucleotides ppGp, ppGpp, and pppGpp, the possibility has been raised that the levels of the NTP pools may be responsible for some features of the stringent response (18, 20). The clear implication of such a mechanism is that

J. BACTROL.

other means of reducing specific NTP pools should to some degree mimiic amino acid limitation; the data in Fig. 2 provide a fairly specific test for this. Concerning the pyrimidine nucleotides, the conversion of UMP to UTP and CTP has been found to slow greatly during a stringent response, and the resulting decline in UTP and CTP levels corresponded well with the observed decline in RNA synthesis (5, 18). On the other hand, uracil limitation does not result in the accumulation of ppGpp (6). It is worth noting that transcription from pL,o is initiated with a CTP, as is at least one species of rRNA (21, 43). A comparison of the responses to leucine and uracil exhaustions shows some parallels and some distinctions; the syntheses of spcl mRNA, a mRNA, and lac mRNA responded in parallel (but not with the same magnitude) in the two exhaustions, whereas transcription frompL10 was very different, decreasing during leucine exhaustion but increasing after uracil exhaustion. Concerning the purine nucleotides, the GTP pool has been found to shrink rapidly and substantially during a stringent response presumably due to both the production of the unusual guanine nucleotides and inhibition of the purine biosynthetic enzyme inosine monophosphate dehydrogenase by ppGpp (4, 19, 20). The Km of RNA polymerase for an NTP is higher during transcript initiation than during elongation (32); accordingly, there is in vivo evidence that during a stringent response, nascent RNAs initiated with GTP drop 25% in frequency relative to nascent RNAs initiated with ATP (23, 28). It is known that some rRNA and ribosomal protein gene transcripts are initiated with GTP (21, 42, 56). The data presented here show some parallels and some distinctions in the responses to the leucine and guanine exhaustions; the transcriptional responses (Fig. 2) are parallel but differ significantly in magnitude, and the translational responses (Fig. 4) show some clear distinctions. Here, however, the difference is more quantitative than qualitative. In summary, our data, which extend those of others (1, 13, 28, 35, 50, 52), argue against the possibility that varying the NTP pools alone can account for the transcriptional changes manifested during a stringent response, although such variation may clearly play a contributing role. Regulation of RNA polymerase production. The regulation of expression of the RNA polymerase structural genes is proving to be quite complicated, involving several levels of control. One of the major unanswered questions is whether RNA polymerase production is controlled to any extent by the levels of its substrate

VOL. 142, 1980

AMINO ACID OR NUCLEOTIDE BASE EXHAUSTION

NTPs. If such control existed and were analogous to the influence of aminoacyl-tRNA levels on ribosome production, one would expect both guanine and uracil exhaustions to result in sharp reductions of RNA polymerase production. This was clearly not the case after uracil exhaustion, and guanine exhaustion led to only a moderate reduction in ,B,B' mRNA synthesis (Fig. 2). On the other hand, the production of RNA polymerase does not appear to be consistently increased by nucleotide base exhaustion either, so if the regulation of RNA polymerase is not analogous to that of ribosomes, neither is it analogous to the regulation of the aminoacyl-tRNA synthetases, some of which are produced at increased rates during insufficiencies of their substrate amino acids (37). This response to uracil exhaustion must be fairly specific since the effects of guanine and uracil exhaustions on the various NTP pools are not entirely independent (26). It seems clear that specific regulation of the production of /,if' mRNA occurs at both the promoter pLuo and at the putative attenuator between rpIL and rpoB and that these two control sites respond to at least partially distinct regulatory signals (Fig. 2 and 3). Thus, during the stringent response to leucine exhaustion, initiation at pLu0 decreased, whereas readthrough of the attenuator greatly increased (31); after guanine exhaustion there was a moderate decrease in initiation atpLlo and a mild increase in attenuator read-through, and after uracil exhaustion there was a striking increase in initiation at pLuo and perhaps an increase in readthrough of the attenuator. Finally, in these three nutrient exhaustions the pattern of a mRNA synthesis was more closely related to that of spcl mRNA than to that of,B,B mRNA. Both rpoA (a) and rpoBC (flf') are cotranscribed with ribosomal protein genes (25, 38, 55), and their expression can probably be dissociated from that of the ribosomal protein genes by post-transcriptional controls (3, lOa). Nevertheless, under the conditions studied in this work, /,3/' mRNA synthesis was more independent than a mRNA synthesis of the transcription of their associated ribosomal protein genes. Regulation of lac expression. Transcription of the lac operon was unusually sensitive to all three of the exhaustions studied. The gratuitous lac inducer IPTG was included in all media, so it is unlikely that this sensitivity involved the lac repressor protein. In the case of leucine exhaustion, catabolite repression and ppGpp levels very probably affected the observed lac transcription (27, 34, 44). In addition, the presence

209

of a rho-dependent transcription termination site within the lacZ gene may make lac transcription particularly sensitive to translation restrictions (9, 24). In the cases of the guanine and uracil exhaustions, the involvement of these factors has not yet been determined, and the response of the lac operon may be a result of changes in the activity or partitioning of the RNA polymerase pool (see 2, 46, 51, 54). In summary, the responses of E. coli to leucine, guanine, and uracil exhaustions tend to argue against simple models for regulation of RNA polymerase production or specificity by the intracellular levels of the substrates (NTPs). In this regard, the regulation of production of the RNA-synthesizing machinery may differ in a very basic way from that of the protein-synthesizing machinery. ACKNOWLEDGMENTS We thank J. Chantler for use of the scanning densitometer and L. Lindahl for strain NF522. This research was supported by grant MA-6340 from the Medical Research Council of Canada and by Public Health Service grant GM-24010 from the National Institutes of Health to P.P.D. R.M.B. is a postdoctoral fellow of the American Cancer Society (grant PF-1400).

LITERATURE CMD 1. Beck, C., J. Ingraham, 0. Maaloe, and J. Neuhard. 1973. Relationship between the concentration of nucleoside triphosphates and the rate of synthesis of RNA. J. Mol. Biol. 78:117-121. 2. Blumenthal, R. M., and P. P. Dennis. 1978. Gene expression in Escherichia coli B/r during partial rifampicin-mediated restrictions of transcription initiation. Mol. Gen. Genet. 165:79-86. 3. Blumenthal, R. M., P. G. Lemaux, F. C. Neidhardt, and P. P. Dennis. 1976. The effects of the reLA gene on the synthesis of aminoacyl-tRNA synthetases and other transcription and translation proteins in Escherichia coli B. Mol. Gen. Genet. 149:291-296. 4. CasheL M. 1975. Regulation of bacterial ppGpp and pppGpp. Annu. Rev. Microbiol. 29:301-318. 5. Cashel, M., and J. Gallant. 1968. Control of RNA synthesis in Escherichia coli. I. Amino acid dependence of the synthesis of the substrates of RNA polymerase. J. Mol. Biol. 34:317-330. 6. Cashel, ML, and J. Gallant. 1969. Two compounds implicated in the function of the RC gene of Escherichia coli. Nature (London) 221:838-841. 7. Collins, J., N. P. Fiil, P. Jorgensen, and J. D. Friesen. 1975. Gene cloning of Escherichia coli chromosomal genes important in regulation of ribosomal RNA synthesis, p. 356-359. In 0. Maaloe and N. P. Kjeldgaard (ed.), Alfred Benzon Symposium IX, Control of Ribosome Synthesis. Munksgaard, Copenhagen. 8. Dalbow, D. G. 1973. Synthesis of RNA polymerase in Escherichia coli B/r growing at different rates. J. Mol. Biol. 75: 181-184. 9. de Crombrugghe, B., S. Adhya, M. Gottesman, and I. Pasta. 1973. Effect of rho on transcription of bacterial operons. Nature (London) New Biol. 241:260264. 10. Dennis, P. P. 1977. Transcription pattems of adjacent

210

BLUMENTHAL AND DENNIS

segments on the chromosome of E8cherichia coli containing genes coding for four 508 ribosomal proteins and the ,B and #9' subunits of RNA polymerase. J. Mol. Biol. 165:603-625. lOa.Dennis, P. P., and N. P. Fill. 1979. Transcriptional and post-translational control of RNA polymerase and ribosomal protein genes cloned on composite ColEl plasmids in the bacterium Escherichia coli. J. Biol. Chem. 254:7540-7547. 11. Dennis, P. P., and M. Nomura. 1975. Regulation of the expression of ribosomal protein genes in Escherichia coli. J. Mol. Biol. 97:61-76. 12. Dennis, P. P., and M. Nomura. 1975. Stringent control of the transcriptional activities of ribosomal protein genes in E. coli. Nature (London) 255:460-465. 13. Erlich, IL, J. Gallant, and R. A. Lazzarini. 1975. Synthesis and turnover of ribosomal ribonucleic acid in guanine-starved celis of Escherichia coli. J. Biol. Chem. 250:3057-3061. 14. Fiandt, M., W. Szybalski, F. Blattner, S. R. Jaskunas, L. Lindahl, and M. Nomura. 1976. Organization of ribosomal protein genes in Escherichia coli. I. Physical structure of DNA from transducing X phages carrying genes from the aroE-str region. J. Mol. Biol. 106:817835. 15. Fiill, N., K. von Meyenburg, and J. D. Friesen. 1973. Accumulation and turnover of guanosine tetraphosphate in Escherichia coli. J. Mol. Biol. 71:769-783. 16. Friesen, J. D., N. P. Fiil, and K. von Meyenburg. 1975. Synthesis and turnover of basal level guanosine tetraphosphate in Escherichia coli. J. Biol. Chem. 250:305309. 17. Furano, A. V., and F. P. Wittel. 1977. Effect of the relA gene on derepression of amino acid biosynthetic enzymes in growing Escherichia coli depends on the pathway being derepressed. J. Bacteriol. 132:352-355. 18. Gallant, J., and M. Cashel. 1967. On the mechanism of amino acid control of ribonucleic acid biosynthesis. J. Mol. Biol. 25:545-553. 19. Gallant, J., H. Erlich, B. Hall, and T. Lafmler. 1970. Analysis of the RC function. Cold Spring Harbor Symp. Quant. Biol. 35:397-405. 20. Gallant, J., and B. Harada. 1969. The control of ribonucleic acid synthesis in Escherichia coli. III. The functional relationship between purine ribonucleoside triphosphate pool sizes and the rate of ribonucleic acid accumulation. J. Biol. Chem. 244:3125-3132. 21. Gilbert, S. F., H. A. de Boer, and M. Nomura. 1979. Identification of initiation sites for the in vitro transcription of rRNA operns rrnE and rmnA in E. coli. Cell 17: 211-224.

22. Gillespie, D. 1968. The formation and detection of DNARNA hybrids. Methods Enzymol. 12B:641-668. 23. Ginsburg, D., and J. A. Steitz. 1975. The 30S ribosomal precursor RNA from Escherichia coli. A primary transcript containing 23S, 16S, 5S sequences. J. Biol. Chem.

250:5647-5654. 24. Hansen, M. T., P. AL Bennett, and K. von Meyenburg. 1973. Intracistronic polarity during dissociation oftranslation from transcription in Escherichia coli. J. Mol. Biol. 77:589-604. 25. Jaskunas, S. R., R. R. Burgess, and M. Nomura. 1975. Identification of a gene for the a-subunit of RNA polymerase at the str-spc region of the Escherichia coli chromosome. Proc. Natl. Acad. Sci. U.S.A. 72:50365040. 26. Jenson, K. F. 1979. Apparent involvement of purines in the control of expression of Salmonella typhimurium pyr genes: analysis of a leaky guaB mutant resistant to pyrimidine analogs. J. Bacteriol. 138:731-738. 27. Kennell, D., and C. Simmons. 1972. Synthesis and decay of messenger ribonucleic acid from the lactose operon of Escherichia coli during amino acid starvation. J.

J. BACTERIOL. Mol. Biol. 70:451-464. 28. Konrad, M., J. Toivonen, and D. P. Nierlich. 1972. Initial nucleotide frequencies of bacterial RNA synthesized during amino acid starvation or changes of carbon source. Nature (London) New Biol. 238:231-233. 29. Laemmnli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 30. Lazzarini, R. A., and A. E. Dahlberg. 1971. The control of ribonucleic acid synthesis during amino acid deprivation in Escherichia coli. J. Biol. Chem. 246:420-429. 31. Maher, D. L., and P. P. Dennis. 1977. In vivo transcription of E. coli genes coding for rRNA, ribosomal proteins and subunits of RNA polymerase: influence of the stringent control system. Mol. Gen. Genet. 155:203211. 32. Mangel, W. F., and M. J. Chamberlin. 1974. Studies of ribonucleic acid chain initiation by Escherichia coli ribonucleic acid polymerase bound to T7 deoxyribonucleic acid. I. An assay for the rate and extent of ribonucleic acid chain initiation. J. Biol. Chem. 249:29953001. 33. Matzura, H., B. S. Hansen, and J. Zeuthen. 1973. Biosynthesis of the ,B and P' subunits of RNA polymerase in Escherichia coli. J. Mol. Biol. 74:9-20. 34. Nakada, D., and B. Magasanik. 1964. The roles of inducer and catabolite repressor in the synthesis of #9galactosidase by Escherichia coli. J. Mol. Biol. 8:105127. 35. Nazar, R. N., and J. T. F. Wong. 1972. Nucleotide changes and the regulation of ribonucleic acid accumulation during growth rate shifts in Escherichia coli J. Biol. Chem. 247:790797. 36. Neidhardt, F. C., P. L. Bloch, and D. G. Smith. 1974. Culture medium for enterobacteria. J. Bacteriol. 119: 736-747. 37. Neidhardt, F. C., J. Parker, and W. G. McKeever. 1975. Function and regulation of aminoacyl-tRNA synthetases in prokaryotic and eukaryotic cells. Annu. Rev. Microbiol. 29:215-250. 38. Newmanl, A. J., T. G. inn, and R. S. Hayward. 1979. Evidence for co-transcription of the RNA polymerase genes rpoBC with a ribosomal protein gene of Escherichia coli. Mol. Gen. Genet.; 169:195-204. 39. Nierlich, D. P. 1978. Regulation of bacterial growth, RNA, and protein synthesis. Annu. Rev. Microbiol. 32:

393-432. 40. O'Farrell, P. H. 1978. The suppression of defective translation by ppGpp and its role in the stringent response.

Cell 14:545-557. 41. Pao, C. C., and J. Gallant. 1979. A new nucleotide involved in the stringent response in Escherichia coli. J. Biol. Chem. 254:688-692. 42. Post, L. E., A. E. Arfsten, F. Reusser, and M. Nomura. 1978. DNA sequences of promoter regions for the str and spc ribosomal protein operons in E. coli. Cell 15:

215-229.

43. Post, L. E., G. D. Strycharz, M. Nomura, H. Lewis, and P. P. Dennis. 1979. Nucleotide sequence of the ribosomal protein gene cluster adjacent to the gene for RNA polymerase subunit #9 in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 76:1697-1701. 44. Primakoff, P., and S. W. Artz. 1979. Positive control of lac operon expression in vitro by guanosine 5'-diphosphate 3'-diphosphate. Proc. Natl. Acad. Sci. U.S.A. 76:

1726-1730. 45. Rickenberg, H. V., G. N. Cohen, G. Buttin, and J. Monod. 1956. La galactoside-permease d'Escherichia coli. Ann. Inst. Pasteur Paris 91:829-857. 46. Sanzey, B. 1979. Modulation of gene expression by drugs affecting deoxyribonucleic acid gyrase. J. Bacteriol.

138:40-47. 47. Shapiro, J., L. MacHattie, L. Eron, G. Ihler, K. Ippen,

VOL. 142, 1980

AMINO ACID OR NUCLEOTIDE BASE EXHAUSTION

and J. Beckwith. 1969. The isolation of pure lac DNA. Nature (London) 244:768-774. 48. Stent, G., and S. Brenner. 1961. A genetic locus for the regulation of ribonucleic acid synthesis. Proc. Natl. Acad. Sci. U.S.A. 47: 2005-2014. 49. Stephens, J. C., S. W. Artz, and B. N. Ames. 1975. Guanosine 5'-diphosphate 3'-diphosphate (ppGpp): positive effector for histidine operon transcription and general signal for amino-acid deficiency. Proc. Natl. Acad. Sci. U.S.A. 72:4389-4393. 50. Thom, G. A., N. F. Varney, and K. Burton. 1970. Nucleic acid synthesis and nucleotide pools in purnnedeficient Escherichia coli. Biochem J. 120:117-124. 51. Travers, A. 1976. RNA polymerase specificity and the control of growth. Nature (London) 263:641-646. 52. Varney, N. F., G. A. Thomas, and K. Burton. 1970. Synthesis of ribonucleic acid in purine-deficient Eschoperon

53.

54. 55.

56.

211

erichia coli and a comparison with the effects of amino acid starvation. Biochem. J. 120:125-132. von Meyenburg, K., L D. Nielsen, K. Johnsen, S. Molin, B. Svenningsen, and G. Miozzari. 1978. Reevaluation of the mode of action of streptolydigin in Escherichia coli: induction of transcription termination in vivo. Antimicrob. Agents Chemother. 13:234-243. Wanner, B. L, R. Kodaira, and F. C. Neidhardt. 1977. Physiological regulation of a decontrolled lac operon. J. Bacteriol. 130:212-222. Yamamoto, M., and M. Nomura. 1978. Cotranscription of genes for RNA polymerase subunitsa and ,8' with genes for ribosomal proteins in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 75:3891-3895. Young, R. A., and J. A. Steitz. 1979. Tandem promoters direct E. coli ribosomal RNA synthesis. Cell 17:225234.