Nov 27, 1989 - measured rates of protein synthesis in isogenic stringent and ... protein synthesis rates by varying tRNA concentrations in elongation steps (3).
Proc. NatI. Acad. Sci. USA Vol. 87, pp. 1511-1515, February 1990 Biochemistry
Relationship between protein synthesis and concentrations of charged and uncharged tRNATrP in Escherichia coli (translation elongation/translation efficiency/translation regulation/codon bias/stringent control)
MUMTAZ V. RoJiANI, HIERONIM JAKUBOWSKI,
AND
EMANUEL GOLDMAN*
Department of Microbiology and Molecular Genetics, New Jersey Medical School and Graduate School of Biomedical Sciences, University of Medicine and Dentistry of New Jersey, 185 South Orange Avenue, Newark, NJ 07103
Communicated by John Abelson, November 27, 1989
We have continuously monitored TrpABSTRACT tRNATIP concentrations in vivo and, in the same cultures, measured rates of protein synthesis in isogenic stringent and relaxed strains. We have also manipulated cellular charged and uncharged [tRNATFP] by two means: (t) the strain used contains a Trp-tRNA synthetase mutation that increases the Km for Trp; thus, varying exogenous Trp varies cellular Trp-tRNATMP; and (ii) we have introduced into the mutant strain a plasmid containing the tRNATrP gene behind an inducible promoter; thus, total [tRNATrP] also can be varied depending on length of induction. The use of these conditions, combined with a previously characterized assay system, has allowed us to demonstrate that (t) the rate of incorporation of Trp into protein is proportional to the fraction of tRNATrP that is charged; for any given total [tRNAT'IJ, this rate is also proportional to the [Trp-tRNATrPI; (ii) uncharged tRNATrP inhibits incorporation of Trp into protein; and (Mi) rates of incorporation into protein of at least two other amino acids, Lys and Cys, are also sensitive to [Trp-tRNATrP] and are inhibited by uncharged tRNATrP. Our results are consistent with models of translational control that postulate modulating polypeptide chain elongation effiiciency by varying concentrations of specific tRNAs.
The modulation hypothesis (1) originally proposed that the rate of protein synthesis could be modulated by the concentrations of different aminoacyl-tRNAs. Although protein synthesis experiments in vitro (2) supported this notion, there has been little direct evidence for this in vivo, and many workers have tended to deemphasize regulation of protein synthesis at elongation, focusing instead on well-established control mechanisms at initiation. Indeed, a computer simulation of protein synthesis predicted only minor effects on protein synthesis rates by varying tRNA concentrations in elongation steps (3). The compilation of extensive sequence data showing strong codon biases in different organisms (e.g., see ref. 4), correlating roughly to concentrations of the corresponding cognate tRNAs (5), led to a renewed interest in the possibility that protein synthesis rates were regulated at elongation steps by [tRNA] (e.g., see ref. 6). Extensive codon replacements in a yeast messenger led to reduced synthesis ofthe corresponding protein (7), consistent with regulation of chain elongation. In this report, we have utilized an assay (8) that allows us to monitor concentrations of aminoacyl-tRNA in vivo and, in the same culture, measure the rate of protein synthesis. We have also varied [tRNATrP] in cells by using a strain with a TrptRNA synthetase mutation and introducing a plasmid containing the tRNATrP gene behind an inducible promoter. We demonstrate that the rate of incorporation of Trp into protein is proportional to [Trp-tRNAThp] when the total [tRNATrIp] is constant. However, when the total [tRNATrp] is varied, the The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement"
rate of incorporation of Trp into protein is found instead to be proportional to the fraction of charged tRNATrP. Uncharged tRNATr inhibits incorporation of Trp into protein. The rates of incorporation into protein of two other amino acids, Lys and Cys, are also sensitive to [Trp-tRNAT"'] and to inhibition by uncharged tRNATrP. We conclude that protein synthesis can be modulated by varying cellular concentrations of charged and uncharged tRNA.
MATERIALS AND METHODS Strains and Plasmids. Escherichia coli strain trpS9969 (9), a Trp auxotroph, was obtained from C. Yanofsky (Stanford University). This strain was determined to be stringent (relA') by a plate test as described by Rudd et al. (10). The relaxed version of this strain (relA-) was constructed as follows. CR151, which contains a TnJO insertion in argA, near a relA- allele, was obtained from M. Cashel (National Institutes of Health). Phage P1vir lysates were prepared on CR151 and used for transduction into trpS9969, first selecting for tetracycline resistance, followed by screening for relAcotransductants (10). The strain was next selected for Arg prototrophy to cure it of TnJO. Both this strain, and the stringent parent, were confirmed to be relA+ or relA- by direct analysis of ppGpp production following Trp starvation (data not shown). The resulting matched isogenic relA+ and relA- trpS9969 strains were transformed with the following plasmids: (i) pCDS-110, obtained from E. Morgan (Roswell Park Memorial Institute), which is derived from pBH-16 (11) but lacks the replication origin of F and with the rrnC promoter replaced by the tac promoter (12); and (ii) pVLRR10 (13) obtained from M. Winkler (Northwestern University), which has the lacIQ gene on a compatible plasmid. Measurements of Trp-tRNAT'P and Incorporation of Trp into Protein. Cultures of trpS9969 relA+ and relA- were grown at 37°C as described (14). To induce the tRNATJp gene, isopropyl f3-D-thiogalactopyranoside (IPTG) (1 mM) was added when the cultures were at a density of either 1 x 108 or 4 X 107 cells per ml, as indicated. When the cultures reached a density of 2 x 108 cells per ml, cells were filtered through a sterile 0.2-,um Millipore filter, washed, and resuspended at the same density in warm medium containing either 0.3 ,uM (15 ,uCi/ml; 1 Ci = 37 GBq), 1 ,uM, 5 ,.M, or 50 AM (each at 30 ,Ci/ml) [3H]Trp (Amersham). This process took between 3 and 5 min. At the indicated times, incorporation in 0.5 ml of culture was stopped by adding 0.5 ml of cold 10% trichloroacetic acid; the samples were processed for incorporation into aminoacyl-tRNA and protein as described (8). The counting efficiency for 3H samples on glass fiber filters was taken to be 27%. Measurements of Total tRNATrP. Total tRNA was extracted from cultures at 2 x 108 cells per ml by a hot phenol method Abbreviation: IPTG, isopropyl,6-D-thiogalactopyranoside. *To whom reprint requests should be addressed.
in accordance with 18 U.S.C. §1734 solely to indicate this fact.
1511
1512
Proc. Natl. Acad. Sci. USA 87 (1990)
Biochemistry: Rojiani et al.
(15), which yields deacylated tRNA (16). Aminoacylation of tRNA in vitro with [3H]Trp was performed as described (16).
RESULTS Methods for Controlling Charged and Uncharged [tRNATrPJ. Plasmid pCDS-110 contains the genes for tRNAASP and tRNATrp, derived from the rrnC operon, under control of the inducible tac promoter (12). The lacIQ gene was introduced into cells on a compatible plasmid, pVL-RR10. Expression of the tRNA genes from pCDS-110 is repressed by the lac repressor produced from pVL-RR10, until an inducer such as IPTG is added to the culture. Even without induction of the plasmid gene, the basal level of tRNATIp is -3-fold higher than in cells without the plasmid. If IPTG is added when cells are at a density of 1 x 108 cells per ml, the total tRNATrP detected after one doubling is increased =10fold compared to no induction (Fig. 1 Lower). Addition of IPTG when cells are at 4 x 107 cells per ml results in a total tRNATrP almost 19-fold higher than in cultures without induction at 2 x 108 cells per ml (14). Similar levels of total tRNATrP were obtained in isogenic stringent (14) and relaxed (Fig. 1) cells. For reasons that are unclear, the yield of mature tRNAASP, as measured on acrylamide gels, is considerably lower than the yield of tRNATrp (E. Morgan, personal com-
munication). The plasmid system described above was introduced into strain trpS9969. This strain bears a mutation in the Trp-tRNA synthetase, which increases the Km for Trp such that in the absence of exogenous Trp, there is negligible aminoacylation of tRNATrp (17). We have previously reported that varying the exogenous [Trp] leads to variation in the charged
[tRNATrP] in vivo in the stringent version of the strain (without the plasmids) (18). Similar results were obtained under different levels of induction of the plasmid tRNATrp gene (14) and also in the relaxed form of the strain (Fig. 1). As the exogenous [Trp] was increased, the concentration of [3H]Trp-tRNATrP also increased continuously. Thus, we can vary both the total [tRNATrp] (by induction of the plasmid), and the charged [tRNATrp] (by the exogenous [Trp]). Measurements of [Trp-tRNATrP] and Incorporation of Trp into Protein. Fig. 2 C and D shows a time course of incorporation of Trp into protein in relA- trpS9969 cells shifted to different exogenous [Trp], along with the concomitantly measured [Trp-tRNATrP] in the same samples (Fig. 2 A and B). The amounts of protein synthesized vary with the exogenous [Trp] in the medium, as does the [Trp-tRNATrp]. As more time elapses after resuspension in different [Trp], the [Trp-tRNA Trp] declines at [Trp] of 5 ,uM and below, and further incorporation of Trp into protein also declines, presumably due to exhaustion of substrate. In a culture in which induction of the plasmid tRNATrp gene commenced at half the cell density there is considerably more aminoacylated tRNATIrp (Fig. 2B) than in the culture without induction (Fig. 2A). However, the dependence of incorporation of Trp into protein upon the exogenous [Trp] in the induced culture (Fig. 2D) does not appear to be grossly different than the dependence of incorporation in the culture without induction (Fig. 2C ). Similar time courses were obtained with the stringent version of the strain (14). Dependence of Trp Incorporation into Protein on the Fraction of Total tRNATrP That Is Charged. The data in Fig. 2 suggested that within each culture, the more Trp-tRNAITrp present, the faster the rate of Trp incorporation into protein. Nevertheless,
lUAoprn uar1-tUHbiAltUt
a,;z
8
cqJ
6111
A-
4_
2-
l-'
NO INDUCTION a--I& PTG at ix 08 cells/mal
0
6
12
5 QuM
5
zS CLi
-I0
MI
p-4
0.
EQL vO.
144_
_ Ed
fl
M
----o
0
4
8
11,'I
_
20 30 40 50
[Trp] (pm)
TOTAL
\\'
4
8
3
6_
7i
I
A
O-----4 0.3pM
-
Lx 2
0
B
5puM
o---o
.M, 23:-
re
7'
-
-
IPTG at IxO8 cells/ml
NO INDUCTION
-
4
-
I-
,
1 O0
,al::~1 8 16 12 4
--
t-
.
\
-
2
--
24
0
fo. ___ _b
12 4 -
.
8
16
11
24
a) -
CKj 4cx
CD
.c
2? 0
E
G
Q-
-II 111
IPTG at lxlO8cells/ml
FIG. 1. Manipulation of charged and total [tRNATrP] in relaxed x 108 cells per ml at which time 1PTG (1 mM) was added to a portion of the culture. Cells were harvested at 2 x 108 cells per ml, and the amount of tRNATrP was determined. Values shown are the means of several experiments for each condition. Samples from induced conditions varied 10%. Samples from uninduced conditions varied ±30%. (Upper) Induced and uninduced cells were grown in the presence of 50 ,uM Trp and then filtered, washed, and resuspended at the indicated concentrations of [3H]Trp. One minute later, incorporation was stopped by addition oftrichloroacetic acid, and the amounts of [3H]Trp-tRNATrP were determined.
trpS9969 cells. (Lower) Cells were grown to 1
±
TIME (min)
TIME (min)
FIG. 2. Incorporation of [3H]Trp into tRNATrIP and protein in relaxed trpS9969 cells as a function of time. (A and C) Values obtained in uninduced conditions. (B and D) Values obtained with IPTG added at 1 x 108 cells per ml. (A and B) Concentration of [3H]Trp-tRNATIP as a function of time. (C and D) Incorporation of [3H]Trp into protein as a function of time. Each curve follows a portion of a single culture resuspended in the indicated [Trp].
Biochemistry: Rojiani et al. roughly similar rates were observed in both induced as well as uninduced cells even though there was an order of magnitude greater [Trp-tRNATrP] in the induced cells. In other words, the increased [Trp-tRNATp] in induced cells did not accelerate the rate of Trp incorporation into protein when compared to uninduced cells for any given [Trp], even though within both induced and uninduced cells, the rate of Trp incorporation did vary with the [Trp-tRNArT]. This apparent paradox is resolved when the rate of Trp incorporation is considered as a function of the fraction of total tRNATp that is charged. This can be calculated from the measured values of charged tRNAT'r and total tRNATrp (see Fig. 1 and ref. 14). The initial rates of Trp incorporation obtained for each exogenous [TrpJ in Fig. 2 are plotted in Fig. 3 as a function of the fraction of charged tRNATrp in the same samples measured 1 min after resuspension in various [Trp]. Although the values for Trp incorporated at the 1-min point may appear at first glance to be small, the lowest amounts obtained were =150,000 cpm of [3H]Trp in protein, and '=1500 cpm of [3H]Trp in Trp-tRNATrp. There is a good correlation between the initial rate of Trp incorporation into protein and the fraction of charged tRNATrP whether the data are from uninduced or induced cells. Thus, the rate of Trp incorporation into protein depends on a single parameter, the fraction of charged tRNATrp, regardless of differences in total tRNATrP. t Inhibition of Trp Incorporation into Protein by Uncharged tRNATrP. The dependence of the rate of Trp incorporation into protein on the fraction of charged tRNATIP rather than merely on the concentration implies that stimulation of incorporation by the charged species is balanced by inhibition of incorporation by the uncharged species. This can be depicted graphically by plotting the rates of Trp incorporation as a function of only the [Trp-tRNATrP] and comparing the dependencies among different total tRNATrp conditions. Numerous experiments similar to that of Fig. 2 were performed on both stringent and relaxed trpS9969 cells. The initial rates of incorporation of Trp into protein are plotted in Fig. 4 as a function of the [Trp-tRNATrp] measured in the same samples. For reasons that are unclear, the relaxed strain generally gave lower rates of protein synthesis than the isogenic stringent strain. Within each series of measurements at a given total [tRNATrp], there is a direct proportionality between the initial rate of Trp incorporation into protein and the [Trp-tRNATrP]. However, the proportionality constants are different at different total tRNATrp levels.t In the relaxed strain, we were also able to obtain in some experiments measurements of protein synthesis rates at 2- to 3-fold higher [Trp-tRNATrp] than shown in Fig. 4 (Lower) for both uninduced and induced cultures. These protein synthetTo a first approximation, an empirical equation from the data in Fig. 3 relating the rate (R) of protein synthesis to the concentrations of charged (C) and total (T) tRNATrp can be written as R = k(C/T) + Ro where the slope k is a proportionality constant, and the intercept is the extrapolated rate of protein synthesis in the absence of R0 [3HITrpAtRNATrP. Even in the absence of Trp-tRNATrp, it is expected that ribosomes will nevertheless show some finite rate of protein synthesis at Trp codons due to mistranslation (19). From our data, the extrapolated value of Ro in the relaxed strain is =10 pmol per min per 2 x 108 cells (Figs. 3 and 4). The proportionality constant k, the slope of the line in Fig. 3, reflects the change in the rate of protein synthesis with the fraction of charged tRNATrP. From our data, the value of k is =140 pmol per min per 2 x 108 cells for Trp incorporation in the relaxed strain (Fig. 3). The data can also be examined at each constant T (Fig. 4). The empirical equation for each line in Fig. 4 is R = K(CJ + RO, which relates to our first equation in that K = kIT for each constant value of T. Thus, the slope in Fig. 4 at T1 will be kIT1; at T2, it will be kIT2, etc. Since the term CIT is the fraction of total tRNATrP that is charged, the rate of protein synthesis is proportional both to the concentration (at constant T; Fig. 4) and the fraction (at any T; Fig. 3) of charged tRNATrP.
Proc. Natl. Acad. Sci. USA 87 (1990)
1513
80 0 NO INDUCTION
a IPTG at I x108 cells/ml
Cn
O
0
60
40
C= A)
J_j
0
20
E/ 0
0
0
20 30 40 1o FRACTION tRNATrP CHARGED 1%)
FIG. 3. Rate of incorporation of Trp into protein as a function of the fraction of tRNATIp that is charged. The initial rate of protein synthesis was taken as the rate of incorporation of Trp into protein 1 min after resuspension of cultures in the various [Trp] in Fig. 2. For the 5 and 50 ,uM samples, a rate for Trp incorporation was obtained by linear regression (as plotted in Fig. 2). For the 0.3 and 1 ,uM samples, the rate at 1 min was taken from the average of the rates of incorporation from 0-1 min and 1-2 min (see Fig. 2). The [TrptRNATrp], also measured in the same samples 1 min after resuspension, was used to calculate the fraction tRNATrP charged, dividing by the total tRNATIp level in each culture (see Fig. 1). These values have been multiplied by 100 and are graphed as percent (%). The slope and intercept were determined by linear regression.
sis rates were only slightly elevated (=10% higher) than the maximum rates indicated in the figure (data not shown). Also, increases in [Trp] from 5 to 50 ,uM result in relatively small increases in both Trp-tRNATrP and the corresponding rates of protein synthesis (Fig. 2). Thus, the maximum rates plotted in Fig. 4 would appear to be at or near the saturating level of Trp-tRNATrP for protein synthesis. Fig. 4 shows that much higher [Trp-tRNA Trp] was required to achieve comparable rates of Trp incorporation into protein when total tRNATrP was increased by induction of the plasmid tRNATrP gene. Furthermore, sensitivity of the Trp incorporation rate to variations in [Trp-tRNA1Trp] was considerably diminished in cells in which total tRNATTrp was increased. This is reflected in the slopes of the dependencies for each total tRNATrP level: in the stringent strain (Fig. 4 Upper), a slope of 120 for cultures with no induction of the tRNATrp gene, 6.6 for cultures in which the tRNATrp gene was induced for 1 doubling, and 3.0 for cultures that had 2.25 doublings in the presence of inducer; in the relaxed strain (Fig. 4 Lower), a slope of 49 for cultures with no induction of the tRNATrp gene and 4.4 for cultures in which the tRNA Trp gene was induced for 1 doubling. Since we know the total [tRNATrP] under each condition (Fig. 1 and ref. 14), it is evident that induced cells differ in containing much higher concentrations of uncharged tRNATrP than uninduced cells. Furthermore, inhibition of protein synthesis parallels total tRNATrI even at higher [Trp-RNAT'P] in induced cultures: 2.25 doublings in the presence of inducer is more inhibitory than 1 doubling (Fig. 4 Upper) and, in preliminary experiments, is also more inhibitory than 0.25 doubling (data not shown). We conclude that uncharged tRNATrp inhibits Trp incorporation into protein. It is unlikely that the observed inhibition results from increased expression of the tRNAASP gene on the plasmid, because overproduction of this species is much lower than the overproduction of tRNATrp (see above), and the level of charged tRNAASP is not varied during the Trp limitations. Effect of Trp-tRNAT'P Limitation on Incorporation of Other Amino Acids into Protein. We next asked whether incorporation of other amino acids into protein was also affected by varying the [Trp-tRNATh']. Incorporation of both Lys (Fig. 5
Biochemistry: Rojiani et al.
1514
CYSTEINE
rel A+
140r-
o-o NO INDUCTION - IPTG at 1XI08 cells/ml o-- -a IPTG at 4x107 cells/ml
120
1001_
0f 0
U)
o
cr
0
2
8
6
4
10
12
.2.
14
rel
140
A
ct0 _j E 120
-o
o
&----6
:
v--
,
0 2 4 6 810 16 TIME (min)
24
B IPTG at I x8 cells/ml
7000
2400 2000 1600
E
m
/
2800-
[3H -Trp -t RNATrPI (pmoles /2x108 cells)
.c
5IP/M 0.3 M
--
400
.E
1
50uM
_
o---o
0 0.5
INDUCTION
800
C.5
0
LYSINE
[Trp]
1200 -D
o
20
co
Co
LUL
1,,-__
o0
0 o
-
cn 2
5c
_
I.,
0
OD
X -;
A NO
1600
0
0
-
40 en U)
3200 2800 2400 2000
0
80 60
Proc. Natl. Acad. Sci. USA 87 (19%)
NO INDUCTION IPTG at lx108 cells/mI
5000
1200_/
3000
800 _
2000
/___o
0 2 4 6 810 16 TIME (min)
80
IPTG at lxO8cells/ml
4000
400
100 - IF
E
6000
1000 0o
24
/
)246810
0
16
_-
24
TIME (min)
0
aI,A /~~~~~~~~~
_ 60 >
en
-400 In -F 350
-
-'Q
M OD
40
Gto
-° 300
I-C
LE
a
°
a.
20 0
?250
=200 -
II
) 0
0.5
1 [3H
0
2
4
6
-rrp -tRNATrs] ()pmoles /2 x 108
8
7 150
10 0
cells )
J- cor a rLo
E
r-
FIG. 4. Rate of incorporation of Trp into protein as a function of the [Trp-tRNATrM] at different total tRNATIP levels. (Upper) Stringent (relA') cells. (Lower) Relaxed (relA-) cells. Samples with induction of the plasmid tRNATIP gene are as indicated. Each point represents an individual measurement in separate experiments, under various exogenous [Trp]. The initial rate of protein synthesis was measured as described in Fig. 3. [Trp-tRNATp] was also measured in the same samples. The slopes and intercepts of each line were determined by linear regression.
D and E) and Cys (Fig. 5 A and B) diminishes (with some variation as to extent) as the exogenous [Trp] decreases. Plotting the initial rates of incorporation from these experiments vs. the measured [Trp-tRNATr ] (in analogy with the data treatment of Fig. 4), suggests that for both Lys (Fig. SF) and Cys (Fig. 5C), rates of incorporation into protein are dependent on Trp-tRNA Trp levels at suboptimal [Trp], for a given total tRNATrP condition. Furthermore, as was seen for Trp incorporation, uncharged tRNATrI diminishes the rates and sensitivity to [Trp-tRNAT'1] for both Lys and Cys incorporation. For Lys, the slope of the dependency without induction of tRNATrP was 85 vs. a slope of 8.2 for 1 doubling in the presence of inducer, and, for Cys, the slope without induction was 20 vs. a slope of 1.4 for 1 doubling in the presence of inducer.
DISCUSSION We have shown that the rate of protein synthesis can be modulated by the fraction charged of a specific tRNA. Our data also show that uncharged tRNA can inhibit incorporation of amino acids into protein. We were able to obtain these results by virtue of our ability to manipulate the concentrations of both charged and total tRNATm over at least an order of magnitude, as well as our ability to measure
PR
50 aw 0
o
1o
o--
