Escherichia coli Cells Bearing a Ribosomal Ambiguity Mutation in rpsD Have a Mutator Phenotype That Correlates with Increased Mistranslation

Abstract

Escherichia coli cells bearing certain mutations in rpsD (coding for the 30S ribosomal protein S4) show a ribosomal ambiguity (Ram) phenotype characterized by increased translational error rates. Here we show that spontaneous mutagenesis increases in Ram cells bearing the rpsD14 allele, suggesting that the recently described translational stress-induced mutagenesis pathway is activated in Ram cells.

In Escherichia coli mutA cells, a mutation in a gene specifying a glycine tRNA (glyV) leads to a significant mutator phenotype (14, 25). The mutA allele differs from the wild-type (glyV⁺) tRNA gene by a single base substitution that alters the anticodon in such a way as to promote Asp→Gly mistranslation in mutA cells. Subsequent to this finding, other work revealed that mutations in many other genes that specify tRNAs (10, 26) or tRNA-modifying enzymes (32) lead to a similar phenotype presumably mediated by mistranslation. Moreover, exposure to streptomycin, an antibiotic known to promote mistranslation, induces a very similar phenotype in cells bearing certain mutations in rpsL, the gene specifying the 30S ribosomal protein S12 (6). Thus, mistranslation appears to be causally required for provoking the mutator phenotype, termed TSM for translational stress-induced mutagenesis (12). Interestingly, the mutator phenotype induced by mutA is suppressed in cells defective for RecABC-mediated homologous recombination functions (20, 21). mutA cells have an error-prone DNA polymerase activity (2) that is most likely to be a modified form of DNA polymerase III (1). The mechanism by which mistranslation leads to the TSM phenotype is not known, but two hypotheses are considered elsewhere in this communication.

Certain mutations in genes specifying the 30S ribosomal protein S4 (4, 19, 33), ribosomal protein S5 (28), 16S RNA (17), or 23S RNA (16) confer a so-called ribosomal ambiguity (Ram) phenotype characterized by decreased growth rates, increased sensitivity to streptomycin, and increased translational errors. It has been hypothesized that reduced proofreading by defective ribosomes in Ram cells leads to a generalized increase in translational errors (5). The best-studied Ram mutants are those that affect rpsD, the gene encoding S4. S4 is an RNA-binding protein that plays an essential role in the assembly of the 30S ribosomal subunit by binding to the 5′-terminal 500-nucleotide region of 16S rRNA (15, 31). Interaction with 16S rRNA may place S4 close to the ribosomal decoding center, which would be consistent with the effect of S4 mutants on ribosomal accuracy. S4 is an autoregulator that translationally represses the mRNA transcribed from the alpha operon encompassing the genes for ribosomal proteins S13, S11, S4, and L17 (13, 27, 29). Recently, S4 was shown to act also as a transcription factor that tightly binds to RNA polymerase and is involved in antitermination of Rho-dependent termination of transcription (30).

Here we examined whether cells bearing rpsD14, a widely used allele known to confer a Ram phenotype, also confer a mutator phenotype in a way that correlates with increased mistranslation. The wild-type rpsD allele codes for a 206-amino-acid-long polypeptide, whereas the rpsD14 allele codes for a 182-amino-acid-long truncated protein because of a +1 frameshift mutation (insertion of a C after G535) (9). Whereas ribosomal ambiguity mutations (such as certain rpsD mutations) increase mistranslation, ribosomal restrictive mutations (such as those in rpsL, which encodes the 30S ribosomal protein S12) increase translational accuracy and are streptomycin resistant (Str^r). We have previously isolated and characterized rpsL1408, which bears a K87E mutation in S12, with very interesting dual properties: in the absence of streptomycin, rpsL1408 cells show hyperaccurate translation (high restrictivity relative to rpsL⁺ wild-type cells), whereas in the presence of 100 μg of streptomycin/ml, rpsL1408 cells display significantly elevated mistranslation levels (6).

To determine whether the rpsD14 allele conferred a TSM phenotype that correlated with mistranslation, we first sought to demonstrate that cells bearing the allele showed increased translational errors by using a nonsense suppression assay. Table 1 lists the strains used in this study, and Table 2 shows the relative mistranslation levels for strains harboring the rpsD14 and rpsL1408 alleles measured as translational (phenotypic) suppression of nonsense codons in plasmid-borne lacZ alleles (16, 18) (see Table 1 for a list of plasmids). Nonsense suppression is up to threefold higher in rpsD14 cells than in wild-type cells, and this increased suppression is eliminated in rpsD14 rpsL1408 double mutants in the absence of streptomycin, as predicted from the contrasting effects of these two mutations, a finding that is consistent with similar earlier observations (11, 23). Table 3 compares the effects of rpsD14 and rpsL1408 alleles on doubling time and temperature and streptomycin sensitivities and shows that rpsD14 cells display significant growth impairment, a phenotypic feature of elevated mistranslation. The growth defects observed in rpsD14 mutants are ameliorated in rpsD14 rpsL1408 double mutants in the absence of streptomycin, a finding that is consistent with the increased translational accuracy conferred by the rpsL1408 allele. The addition of streptomycin to the growth medium reverses the ameliorating effects of the rpsL1408 allele in rpsD14 cells, as indicated by an increase in doubling time and a reacquisition of temperature sensitivity, both of which are consistent with increased mistranslation in rpsL1408 cells in the presence of streptomycin (6).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Genotype and/or phenotype	Source, derivation, and/or reference
E. coli strains
AM110	recA938::Tn9-200 (Cmr) in KH2	This laboratory (A. A. M. A1 Mamun and M. Z. Himayun, unpublished data; 24)
AM121a	lexA1 (Ind−) malF3089::Tn10 (Tetr)	This laboratory (3)
BS1408	rpsL1408 (Strr) ara Δ(proB-lac) thi F′(lacI lacZ proB+)	This laboratory (6)
BS1480b	F−rpsL1408 (Strr) ara argE Δ(proB-lac) nalA rif thi	This work (P1 BS1408 × XAc to Strr)
BS1481b	F−rpsL1408 rpsD14 ara argE Δ(proB-lac) nalA rif thi	This work (P1 BS1408 × UL478 to Strr)
BS1482	ara argE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (conjugation CC105 × XAc; Pro+ Rifr Nalr)
BS1483	rpsD14 ara argE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (conjugation CC105 × UL478; Pro+ Rifr Nalr)
BS1484	rpsL1408 ara argE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (conjugation CC105 × BS1480; Pro+ Rifr Nalr)
BS1485	rpsL1408 rpsD14 ara argE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (conjugation CC105 × BS1481; Pro+ Rifr Nalr)
BS1486c	recA938::Tn9-200 (Cmr) rpsD14 araargE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (P1 AM110 × BS1483 to Cmr)
BS1487c	lexA1 (Ind−) malF::3089::Tn10 (Tetr) rpsD14 ara argE Δ(proB-lac) nalA rif thi F′(lacI lacZ proB+)	This work (P1 AM121 × BS1483 to Tetr)
CC105	ara Δ(proB-lac) thi F′(lacI lacZ proB+)	J. H. Miller (8)
UL478	F−rpsD14 ara argE Δ(proB-lac) nalA rif thi	E. Goldman (19)
XAc	F−ara argE Δ(proB-lac) nalA rif thi	E. Goldman (7, 19)
Plasmidsd
pSG163	lacZ (UAG)	A. Dahlberg (18)
pSG853	lacZ (UAA)	A. Dahlberg (18)
pSG3/4	lacZ (UGA)	A. Dahlberg (18)

^aOther markers are thr-1 araC14 leuB6 Δ(gpt-proA)62 lacY1 tsx-33 glnV44(AS) galK2(Oc) λ-Rac-O hisG4(Oc) rfbD1 mgl-51 rpsL31 (Str^r) kdgK51 mtl-1 metB1 thi-1.

^bTransfer of the rpsL allele was confirmed by PCR amplification and DNA sequence analysis of the transductants.

^cAntibiotic-resistant transductants were tested for UV sensitivity by streaking on Luria-Bertani agar alongside wild-type cells and exposing the streaks to UV.

^dPlasmids were derived from pACYC184 (18).

TABLE 2.

Relative mistranslation in rpsD14 and rpsL1408 cells

Plasmid	Strain	Relevant genotype	β-Galactosidase activity (Miller units)a
			−STR	+STR
pSG163 (UAG)	XAc	Wild type	38.6 ± 1.7
	UL478	rpsD14	115.7 ± 12.2
	BS1480	rpsL1408	18.6 ± 1.1	36.7 ± 3.2
	BS1481	rpsD14 rpsL1408	25.3 ± 2.4	67.9 ± 8.7
pSG853 (UAA)	XAc	Wild type	19.6 ± 1.5
	UL478	rpsD14	27.3 ± 3.8
	BS1480	rpsL1408	9.7 ± 2.4	14.8 ± 1.4
	BS1481	rpsD14 rpsL1408	15.2 ± 3.4	31.9 ± 9.4
pSG3/4 (UGA)	XAc	Wild type	106.5 ± 6.0
	UL478	rpsD14	249.2 ± 23.8
	BS1480	rpsL1408	17.2 ± 1.9	56.2 ± 9.9
	BS1481	rpsD14 rpsL1408	30.5 ± 6.6	113.2 ± 8.6

^aValues shown are averages ± standard deviations of results from three experiments carried out as described previously (6). −STR, without streptomycin; +STR, with streptomycin.

TABLE 3.

Effects of rpsD14 and rpsL1408 alleles on growth parameters and streptomycin sensitivity

Strain	Relevant genotype	Generation timea (min)		Growth at 44°Cb		STR sensitivity (MIC)c (μg/ml)
		−STR	+STR	−STR	+STR
BS1482	Wild type	27		+	−	6
BS1483	rpsD14	49		−	−	1
BS1484	rpsL1408	42	44	+	−	128
BS1485	rpsD14 rpsL1408	35	41	+	−	64

^aGeneration times were determined from viable-cell counts of two culture aliquots taken at the beginning and the end of the logarithmic phase of growth in Luria-Bertani medium at 37°C with (+) or without (−) streptomycin (STR). The following standard equation was used: G = t/[3.3 log(b/B)], where G is the generation time, t is the time interval in minutes, B is the cell count at the beginning of the time interval, and b is the cell count at the end of the time interval used.

^bCells (from fresh overnight colonies grown on Luria-Bertani agar at 37°C) were streaked on Luria-Bertani agar and incubated overnight at 44°C. +, growth; −, no growth.

^cMICs of streptomycin (STR) were determined by using the Etest (AB Biodisk, Solna, Sweden) in accordance with the procedures described by the manufacturer.

The TSM mutator phenotype can be conveniently detected by colony papillation based on a lacZ→ lacZ⁺ reversion on P-Gal (phenyl-β-d-galactoside) minimal plates containing X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) (6). The lacZ allele used in the present investigation reverts only by AT→TA transversions at a specific base pair (8). Figure 1A shows that the wild-type strain (BS1482; see Table 1 for strain derivation) displays a normal low level of papillation characterized by a small number of dark lacZ⁺ colonies growing on a colorless lawn of lacZ cells. In contrast, BS1483, a strain carrying an rpsD14 allele in an isogenic background, shows numerous papillae, a hallmark of a mutator strain. Introduction of the rpsL1408 allele, as demonstrated above, should increase translational accuracy in rpsD14 cells, which in turn should suppress the mutator phenotype of rpsD14 cells in the absence of streptomycin, as indeed was observed for BS1485 (rpsD14 rpsL1408) cells. When streptomycin is included in the agar (Fig. 1B), both BS1485 (rpsD14 rpsL1408) and BS1484 (rpsL1408) cells display a mutator phenotype (6), which is consistent with the reversal of the rpsL1408 effect by streptomycin and parallels similar reversals in growth effects as noted above. Table 4 summarizes spontaneous mutation rates in rpsD14 and rpsD14 rpsL1408 cells, confirming that rpsD14 confers a mutator phenotype that is eliminated by rpsL1408 in the absence of streptomycin. Therefore, the mutator phenotype displayed by rpsD14 cells is strongly correlated with mistranslation.

FIG. 1.

Papillation phenotypes of rpsD and rpsL strains with and without additional mutations. The strains were streaked on minimal A agar medium containing 0.2% glucose, 500 μg of P-Gal/ml, and 40 μg of X-Gal/ml and incubated for 4 to 5 days at 37°C (6). Plate B was supplemented with 100 μg of streptomycin/ml (+Str), whereas plates A and C had no streptomycin (−Str).

TABLE 4.

Effects of rpsD14 and rpsL1408 alleles on spontaneous lacZ→lacZ⁺ reversion rates^a

Strain	Relevant genotype	Mutation rate (10−8)
		−STR	+STR
BS1482	Wild type	1.22
BS1483	rpsD14	2.51
BS1484	rpsL1408	1.02	3.75
BS1485	rpsD14 rpsL1408	0.89	3.77

^aReversion rates were determined as described previously (6) by using the episomal lacZ allele from the CC105 strain (8), which can revert only by an AT→TA transversion. +STR, with 100 μg of streptomycin/ml; −STR, without STR.

It is interesting that whereas the rpsD14 mutator effect is nullified by rpsL1408 in the absence of streptomycin, the mutator effect of streptomycin in rpsL1408 rpsD14 cells is unaffected, an expected result because in the presence of streptomycin rpsL1408 promotes rather than restricts such translational errors. In a sense, the rpsL1408 effect is dominant over that of rpsD14 in both the presence and absence of streptomycin.

We have previously shown that expression of the TSM mutator phenotype does not require SOS induction in mutA cells (20) or in rpsL1408 cells exposed to streptomycin (6). Figure 1C shows that the mutator phenotype of rpsD14 cells persists in lexA1 (Ind⁻) cells, showing that SOS induction is also not required in rpsD14 cells. However, in contrast to the situation in mutA cells (14, 20, 21), miaA cells (32), and streptomycin-exposed rpsL1408 cells (6), the rpsD14 mutator phenotype is not suppressed in recA cells. This last finding is similar to the lack of a recA requirement for the mutator phenotypes displayed by alaV^Glu and glyV^Glu mutator tRNAs (10) and argues that there is no absolute requirement for RecBC-mediated recombination genes for the expression of the mutator phenotype.

Two hypotheses to explain how translational errors can lead to a mutator phenotype have been proposed. The epsilon subunit mistranslation hypothesis suggests that in mutA cells, low-level Asp→Gly mistranslation creates a small fraction of inactive epsilon protein (the proofreading subunit of DNA polymerase III holoenzyme), which when recruited into a functional holoenzyme leads to a transient mutator phenotype (25, 26). The induced pathway hypothesis generalizes the mechanism to many types of mistranslation beyond Asp→Gly misreading and postulates that increased translational errors generate a signal (increased protein turnover or creation of gain-of-function mutant proteins) that induces a pathway that ultimately modifies DNA polymerase III (1, 2, 6, 10, 12, 20, 21). Thus, both hypotheses invoke a modification of DNA polymerase III, a prediction that has recently been verified (1, 22). However, the two hypotheses differ in the proposed mechanisms for the modification, and experimental evidence available to date cannot be unequivocally interpreted as supporting one mechanism over the other.

At present, no experimental basis is available to interpret the apparent requirement for homologous recombination functions for the expression of the TSM phenotype under certain mistranslation conditions but not in others. A speculative possibility is that RecBC-mediated recombination functions are required to overcome a replication defect caused by high levels of misreading at particular codons. From this perspective, the differences in the requirement for recombination functions under different mistranslation conditions may reflect underlying qualitative or quantitative differences in the misreading of specific codons and their consequent effects on replication.