Dissection of 16S rRNA Methyltransferase (KsgA) Function in Escherichia coli

Abstract

A 16S rRNA methyltransferase, KsgA, identified originally in Escherichia coli is highly conserved in all living cells, from bacteria to humans. KsgA orthologs in eukaryotes possess functions in addition to their rRNA methyltransferase activity. E. coli Era is an essential GTP-binding protein. We recently observed that KsgA functions as a multicopy suppressor for the cold-sensitive cell growth of an era mutant [Era(E200K)] strain (Q. Lu and M. Inouye, J. Bacteriol. 180:5243-5246, 1998). Here we observed that although KsgA(E43A), KsgA(G47A), and KsgA(E66A) mutations located in the S-adenosylmethionine-binding motifs severely reduced its methyltransferase activity, these mutations retained the ability to suppress the growth defect of the Era(E200K) strain at a low temperature. On the other hand, a KsgA(R248A) mutation at the C-terminal domain that does not affect the methyltransferase activity failed to suppress the growth defect. Surprisingly, E. coli cells overexpressing wild-type KsgA, but not KsgA(R248A), were found to be highly sensitive to acetate even at neutral pH. Such growth inhibition also was observed in the presence of other weak organic acids, such as propionate and benzoate. These chemicals are known to be highly toxic at acidic pH by lowering the intracellular pH. We found that KsgA-induced cells had increased sensitivity to extreme acid conditions (pH 3.0) compared to that of noninduced cells. These results suggest that E. coli KsgA, in addition to its methyltransferase activity, has another unidentified function that plays a role in the suppression of the cold-sensitive phenotype of the Era(E200K) strain and that the additional function may be involved in the acid shock response. We discuss a possible mechanism of the KsgA-induced acid-sensitive phenotype.

KsgA, a 16S rRNA methyltransferase, methylates two highly conserved adjacent adenosine residues in a loop region at the 3′ end of 16S rRNA (9). KsgA is conserved in all living cells so far examined, from bacteria to humans. Recently, O'Farrell et al. reported that both archaeal and eukaryotic KsgA orthologs are able to complement the KsgA methyltransferase function in bacteria, indicating that the core methyltransferase activity of this family of proteins has not evolved significantly since the last common ancestor (30). The yeast KsgA ortholog (Dim1) is essential for cell growth; however, its essentiality is not due to its methyltransferase activity but depends on its pre-18S rRNA processing activity (19). A human KsgA ortholog with rRNA methyltransferase activity originally was found as human mitochondrial transcriptional factor B1 (h-mtTFB1) (7, 26, 40). A KsgA ortholog in Arabidopsis thaliana (Pfc1) plays an important role for chloroplast biogenesis at low temperatures (46). These observations suggest that in addition to the methyltransferase activity, KsgA orthologs are recruited to play additional roles within the cells (30), although such an additional function has not been reported for Escherichia coli KsgA so far.

Era is an essential GTP-binding protein in E. coli (23, 44), Salmonella enterica serovar Typhimurium (1), Streptococcus mutans (38), and Bacillus subtilis (28). Depletion of a chicken Era homologue led to growth impairment, accompanied by an accumulation of apoptotic cells (10), suggesting that Era also plays an essential role in eukaryotic cells. The exact function of Era is not clear at present; however, it is believed to be involved in the biogenesis of 30S ribosomal subunits (14, 29), cell cycle regulation (5, 6, 11), and energy metabolism (15, 21, 33). Recently, we observed that E. coli Era can bind to 16S rRNA and 30S ribosomal subunits in vitro (39), and depletion of Era causes the accumulation of 30S and 50S ribosomal subunits relative to the amount of 70S monosomes, with concomitant accumulation of a precursor 16S rRNA (17S rRNA) (14). The structural studies of the Thermus thermophilus Era-30S ribosomal subunit complex using cryoelectron microscopy showed that Era binds to the cleft between the head and the platform of the 30S ribosomal subunits (41). A cold-sensitive Era mutant strain [Glu200 to Lys; designated Era(E200K)] originally was isolated by a PCR-prone random mutagenesis method (20). Similar to the situation for the Era-depleted strain, the Era(E200K) strain also showed accumulation of 30S and 50S ribosomal subunits with concomitant accumulation of 17S rRNA (14). Purified Era(E200K) protein binds to 30S ribosomal subunits in a nucleotide-dependent manner, like wild-type Era, and retains both GTPase and autophosphorylation activities (16). Recently, we observed that KsgA functions as a multicopy suppressor for the cold-sensitive phenotype of this strain; however, the mechanism of this suppression is not known (22).

In this study, we investigated the mechanism of suppression of the cold-sensitive phenotype of the Era(E200K) strain by KsgA. We observed that certain mutations, such as E43A, G47A, and E66A, in KsgA that severely affect its methyltransferase activity do not affect its ability to suppress the cold sensitivity of the Era(E200K) strain. On the other hand, the R248A mutation in KsgA leads to the loss of its suppressor activity, while its methyltransferase activity is unaffected. This suggests that in addition to the methyltransferase function, KsgA has another unidentified function that compensates for the Era activity at low temperatures. Interestingly, we also observed that acetate inhibits the growth of cells overexpressing wild-type KsgA but not the cells overexpressing KsgA(R248A). Furthermore, KsgA-induced cells had increased sensitivity to extreme acid conditions (pH 3.0) compared to the sensitivity of noninduced cells. These results suggest that KsgA is involved in the acid shock response, and this activity may be responsible for the suppression of Era activity at a low temperature.

MATERIALS AND METHODS

Culture media and growth conditions.

Bacteria were grown either in Luria-Bertani (LB) medium or M9 (0.4% glucose and 0.2% Casamino Acids) medium supplemented, when necessary, with 100 μg/ml ampicillin, 50 μg/ml kanamycin, or 25 μg/ml chloramphenicol.

Bacterial strains and plasmids.

Strain JM101-ksgA19 was a kind gift from A. E. Dahlberg (Brown University). Strain CL213 [era::km F′(lacI^q lac⁺ pro⁺)] is a derivative of strain JM83 [F⁻ ara Δ(lac-pro) thi Φ80 dlacZΔM15] carrying helper plasmid pCL001. This plasmid contains the wild-type era gene and has a temperature-sensitive origin of replication (21). The strain CL213/pAC19era(E200K) was constructed by curing the helper plasmid (pCL001) from the CL213 strain harboring the pAC19era(E200K) plasmid at 42°C. To construct the pINksgA plasmid, the coding sequence of the wild-type ksgA gene was amplified by PCR with two primers, KSGAL (NdeI) (5′ GCACCATATGAATAATCGAGTCC 3′) and KSGAR (BamHI) (5′ TAGGATCCCGTTAACTCTCCTGC 3′) (sequences underlined correspond to NdeI and BamHI sites, respectively) using pUCksgA plasmid (previously named pKsgA) (22) as the template. The amplified DNA fragment was cloned into an NdeI/BamHI-digested pINIII plasmid. The resultant plasmids were designated pINksgA-S and pINksgA-L, extracted from small and large colonies, respectively. To construct pETksgA, pETksgA(E43A), and pETksgA(R248A), we amplified these coding sequences using pUCksgA, pUCksgA(E43A), and pUCksgA(R248A) plasmids as templates using two primers, KSGAL (NdeI) (5′ GCACCATATGAATAATCGAGTCC 3′) and KSGAR (BamHI) (5′ TAGGATCCCGTTAACTCTCCTGC 3′) (sequences underlined correspond to NdeI and BamHI sites, respectively). These DNA fragments, digested with NdeI and BamHI, were cloned into pET28a (Novagen) digested with the same restriction enzymes. Note that the His tag is fused to the N-terminal end of KsgA.

Site-directed mutagenesis.

Nine mutations (E43A, G47A, E66A, D91A, P115A, F181A, K223A, L229, and R248A) in ksgA were constructed as described previously (36). Plasmid pUCksgA was used as the PCR template with the use of the following mutagenic oligonucleotide primers: for E43A, 5′ CAG GCG ATG GTC GCA ATC GGC CCC GGT 3′ and 5′ ACC GGG GCC GAT TGC GAC CAT CGC CTG 3′; for G47A, 5′ GAA ATC GGC CCC GCT CTG GCG GCA TTG 3′ and 5′ CAA TGC CGC CAG AGC GGG GCC GAT TTC 3′; for E66A, 5′ CTG ACG GTC ATC GCA CTT GAC CGC GAT 3′ and 5′ ATC GCG GTC AAG TGC GAT GAC CGT CAG 3′; for D91A, 5′ ATT TAT CAG CAG GCT GCG ATG ACC TTT 3′ and 5′ AAA GGT CAT CGC AGC CTG CTG ATA AAT 3′; for P115A, 5′ TTC GGC AAC CTG GCT TAT AAC ATC TCC 3′ and 5′ GGA GAT GTT ATA AGC CAG GTT GCC GAA 3′; for F181A, 5′ CCG CCG TCA GCC GCT ACA CCA CCA CCC 3′ and 5′ GGG TGG TGG TGT AGC GGC TGA CGG CGG 3′; for K223A, 5′ AAC CAG CGT CGT GCA ACC ATT CGT AAC 3′ and 5′ GTT ACG AAT GGT TGC ACG ACG CTG GTT 3′; for L229A, 5′ ATT CGT AAC AGC GCC GGC AAC CTG TTT 3′ and 5′ AAA CAG GTT GCC GGC GCT GTT ACG AAT 3′; and for R248A, 5′ GAC CCG GCG ATG GCA GCG GAA AAT ATC 3′ and 5′ GAT ATT TTC CGC TGC CAT CGC CGG GTC 3′ (sequences underlined correspond to codons for mutated amino acid residues). PCR was performed using the cloned Pfu DNA polymerase (Stratagene, La Jolla, CA) under the following conditions: 18 cycles of 1 min at 95°C, 1 min at 50°C, and 20 min at 68°C. The PCR products, digested with DpnI to remove the template DNA (pUCksgA), were introduced into the DH5α strain, followed by being plated onto LB agar plates containing ampicillin (100 μg/ml). The nine point mutations were confirmed by DNA sequencing.

Purification of His-tagged KsgA and its mutant proteins.

A 1-liter culture of strain BL21(DE3) harboring either pETksgA, pETksgA(E43A), or pETksgA(R248A) grown at 37°C to early log phase in LB medium was induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 2 h. Cells were harvested and suspended in 30 ml of buffer A (50 mM NaH₂PO₄ [pH 8.0], 300 mM NaCl, 10 mM imidazole, 1 mM β-mercaptoethanol [β-ME]). Cells were lysed with a French pressure cell (ThermoIEC), and cell debris was removed by low-speed centrifugation (17,000 × g for 30 min at 4°C). The supernatant was passed through a 0.45-μm filter (Millipore) and applied onto a 5-ml Ni-nitrilotriacetic acid column (Qiagen). The column was washed thoroughly with 40 ml of buffer A, and proteins were eluted with 150 mM imidazole in buffer A. The samples were pooled together and dialyzed against 50 mM Tris-HCl (pH 8.0) buffer containing 150 mM NaCl, 10% glycerol, 5 mM β-ME. Purified KsgA (3.9 mg/ml), KsgA(E43A) (3.2 mg/ml), and KsgA(R248A) (3.2 mg/ml) were stored at −80°C.

Measurements of CD spectra.

Using purified His-tagged KsgA, KsgA(E43A), and KsgA(R248A) (3 μM) in buffer B (10 mM sodium phosphate [pH 8.0], 150 mM NaCl, 10% glycerol), circular dichroism (CD) spectra were determined with an AVIV 215 spectrometer (Proterion).

Primer extension.

Primer extension was carried out as described previously (14). Total RNAs were extracted from 1.5-ml cultures with the hot phenol method as described previously (37). Oligonucleotide PE16S (5′ CGACTTGCATGTGTTAGG 3′) was labeled at the 5′ end by T4 polynucleotide kinase with [γ-³²P]ATP. Total RNAs (4 μg) were used in each primer extension experiment with 0.4 pmol of the ³²P-labeled oligonucleotide. A reverse transcriptase reaction was performed at 42°C for 1 h, and the samples were analyzed on a 6% polyacrylamide gel. The gel was dried and subjected to autoradiography. The 17S rRNA and 16S rRNA were quantitated with a Molecular Typhoon PhosphorImager (Amersham Bioscience).

Immunoblot analysis.

For immunoblot analysis, cells grown in LB medium to log phase were harvested and suspended in 50 mM Tris-HCl (pH 8.0)-1% sodium dodecyl sulfate (SDS). Equivalent protein amounts were analyzed by 12% SDS-polyacrylamide gel electrophoresis followed by immunoblotting with anti-KsgA serum. Proteins were assayed by the method of Bradford (4) using a reagent purchased from Bio-Rad with crystalline bovine serum albumin as the standard.

Acid resistance assay.

For the acid resistance assay, an overnight culture grown in M9 medium was diluted to 1:1,000 into M9 medium supplemented with 1 mM IPTG. At log phase (optical density at 600 nm [OD₆₀₀] of 0.5 to 0.6), the culture was diluted to 1:1,000 (1 × 10⁵ cells/ml) into Casamino Acids-free M9 medium (pH 3.0) and incubated at room temperature without shaking. Cells were diluted into M9 (pH 7.0) medium in the absence of glucose and Casamino Acids and were plated in duplicate on LB agar plates supplemented with ampicillin. Colonies were counted after incubation at 37°C for 24 h. The results presented are the averages from two independent experiments.

DNA microarray analysis.

For RNA extraction, W3110 cells harboring either pINIII (vector) or pINksgA-L grown in 50 ml M9 medium supplemented with 50 μg/ml ampicillin at 37°C to mid-log phase (OD₆₀₀ of 0.5) were incubated for 30 min in the presence of 1 mM IPTG. RNA samples were prepared from a 12-ml cell culture with the hot phenol method as described previously (37). Total RNA samples (200 μg) purified further with an RNeasy Mini kit (Qiagen) were treated with RNase-free DNase I, followed by phenol-chloroform extraction. One hundred units of avian myeloblastosis virus reverse transcriptase was used to label mRNAs in 80 μg of total RNAs with 0.1 mM Cy3-dUTP and Cy5-dUTP (Amersham), using 300 pmol pd(N)₆ random hexamer (Amersham) as the primer. The E. coli gene array IntelliGene E. coli CHIP, version 2 (catalog no. X003), was obtained from Takara-Bio and contains approximately 4,150 immobilized DNA fragments. This represents approximately 97% of the open reading frames of E. coli K-12 W3110. Hybridization was done as recommended by the manufacturer. A cluster of the 12 genes (the acid fitness island) and the six most strongly induced genes under heat shock (35), oxidative shock (50), and UV irradiation (8) that have been identified previously using a DNA microarray are listed in Table 2.

TABLE 2.

Ratio of KsgA-induced/control cell transcript levels^a

Stress	Gene	Description	Ratio (KsgA induced/control)
Acid shock	gadA	Glutamate decarboxylase A, pyridoxal 5′-phosphate dependent	0.33 ± 0.04
	hdeA	Stress response protein, acid resistance protein	0.22 ± 0.09
	hdeB	Acid resistance protein	0.25 ± 0.09
	hdeD	Acid resistance membrane protein	0.53 ± 0.16
	slp	Outer membrane lipoprotein	0.40 ± 0.02
	yhiD	Predicted Mg2+ transport ATPase inner membrane protein	0.44 ± 0.10
	yhiE (gadE)	DNA-binding transcriptional activator	0.16 ± 0.02
	yhiF	Predicted DNA-binding transcriptional regulator	0.66 ± 0.22
	yhiU	Multidrug resistance efflux transporter	0.61 ± 0.04
	yhiV	Multidrug transporter, RpoS dependent	0.82 ± 0.04
	yhiW (gadW)	DNA-binding transcriptional activator	0.40 ± 0.08
	yhiX (gadX)	DNA-binding transcriptional dual regulator	0.46 ± 0.07
Heat shock	ibpA	Heat shock chaperone	2.84 ± 0.09
	ibpB	Heat shock chaperone	3.23 ± 0.45
	mopB/groES	Cpn10 chaperonin GroES, small subunit of GroESL	0.90 ± 0.08
	dnaK	Chaperone Hsp70, cochaperone with DnaJ	0.92 ± 0.11
	fxsA	Inner membrane protein	1.91 ± 0.09
	dnaJ	Chaperone Hsp40, cochaperone with DnaK	1.12 ± 0.05
Oxidative shock	dps	Fe-binding and storage protein	1.07 ± 0.51
	yaiA	Predicted protein	0.69 ± 0.29
	katG	Catalase/hydroperoxidase HPI(I)	1.09 ± 0.19
	grxA	Glutaredoxin 1, redox coenzyme for ribonucleotide reductase (RNR1a)	0.70 ± 0.08
	yfiA	Cold shock protein associated with 30S ribosomal subunit	0.55 ± 0.20
	ibpA	Heat shock chaperone	2.84 ± 0.09
UV irradiation	sulA	SOS cell division inhibitor	1.00 ± 0.08
	umuC	DNA polymerase V, subunit C	0.70 ± 0.11
	umuD	DNA polymerase V, subunit D	0.72 ± 0.06
	recN	Recombination and repair protein	0.79 ± 0.01
	grxA	Glutaredoxin 1, redox coenzyme for ribonucleotide reductase (RNR1a)	0.70 ± 0.08
	yigF	Conserved inner membrane protein	0.28 ± 0.14

^aRatio of transcript levels for KsgA-induced cells to transcript levels for uninduced cells. The values significantly reduced by more than 50% are shown in boldface. A cluster of the 12 genes (an acid fitness island) and the six most strongly induced genes under heat shock (35), oxidative shock (50), and UV irradiation (8) that were identified previously using DNA microarrays are shown. The results presented are the averages from two independent experiments. Gene product descriptions are taken from http://ecoli.aist-nara.ac.jp/.

Gel shift assay.

The DNA fragment (631 bp) that encodes the GTP-binding domain of Era was amplified by PCR using two oligonucleotides: 5′ GATCTCTAGAGGCTGCCGCCGAACAGG 3′ and 5′ ACTGAATTCTTACTGTGAGCGATCGGTGAT 3′ (sequences underlined correspond to XbaI and EcoRI sites, respectively). A DNA fragment (3 μM) was mixed with purified wild-type KsgA, KsgA(E43A), or KsgA(R248A) in 50 mM Tris-HCl (pH 8.0), 5 mM MgCl₂, 10 mM β-ME, 150 mM NaCl, 10% glycerol. After 5 min of incubation on ice, the DNA-protein complexes were subjected to 3.5% polyacrylamide gel electrophoresis in Tris-borate-EDTA (TBE) buffer, followed by ethidium bromide staining.

Measurement of dissociation constants.

We designed two pairs of oligonucleotides to prepare end-labeled double-stranded DNAs: for the GC-rich DNA fragment, E43A(+) (5′ CAGGCGATGGTCGCAATCGGCCCCGGT 3′) and E43A(−) (5′ ACCGGGGCCGATTGCGACCATCGCCTG 3′); for the AT-rich DNA fragment, metE(+) (5′ CATATAATTAGAGGAAGAAAAAATGAC 3′) and metE(−) (5′ GTCATTTTTTCTTCCTCTAATTATATG 3′). For end labeling, 25 μM of the E43A(+) and metE(+) oligonucleotides were labeled with [γ-³²P]ATP using T4 polynucleotide kinase. The labeled E43A(+) and metE(+) oligonucleotides annealed with an equal amount of their complementary oligonucleotides, E43A(−) and metE(−), respectively. Purified wild-type KsgA (0.45, 0.95, 1.88, 3.80, 7.50, 15.0, and 30.0 μM) was mixed with 0.35 μM of the labeled DNA fragments. After 20 min of incubation on ice, the DNA-protein complexes were subjected to 3.5% polyacrylamide gel electrophoresis in TBE buffer, followed by an autoradiogram.

RESULTS

Characterization of the ksgA19 (Ksg^r) mutation.

A mutation in the ksgA gene is known to confer resistance to the antibiotic kasugamycin. The ksgA19 allele emerged from early attempts to isolate kasugamycin-resistant strains using nitrosoguanidine mutagenesis (42). The ksgA19 mutation has been considered a standard null allele of ksgA, but it has not been characterized in detail.

In order to characterize the ksgA19 mutation, the ksgA locus was amplified from both a wild-type strain (JM101) and the ksgA19 strain (JM101-ksgA19) by PCR. The PCR product from the JM101-ksgA19 strain was longer than the wild-type JM101 strain by 800 bases, and we determined that this was due to the insertion of an IS1 element after the 11th nucleotide of the ksgA coding sequence. The insertion creates a stop codon in the frame of the ksgA coding sequence, resulting in an aberrant decapeptide in place of the full-length methyltransferase polypeptide, indicating that the kasugamycin resistance of the ksgA19 strain is indeed due to the inability of the strain to produce the active methyltransferase. This notion was confirmed by Western blot analysis with anti-KsgA serum (data not shown).

Creation of inactive KsgA mutants.

The three-dimensional structure of E. coli KsgA displays a strong similarity to the structure of ErmC′, a 23S rRNA methyltransferase from B. subtilis (31). Similar to ErmC′, E. coli KsgA consists of two domains, the N-terminal domain, with highly conserved S-adenosylmethionine-binding motifs, and the C-terminal domain. The C-terminal domain of KsgA has few conserved residues compared to the N-terminal domain (Fig. 1), suggesting that the C-terminal domain is involved in specific functions of individual KsgA orthologs from different organisms. Although the exact role of the C-terminal domain is not known, the positively charged cleft between the N- and C-terminal domains is believed to play a critical role in the binding of KsgA to RNA (31). To examine whether the methyltransferase activity of KsgA is involved in the suppression of the cold-sensitive phenotype of the Era(E200K) strain, nine point mutations were introduced into highly conserved amino acid residues of KsgA, both in the N-terminal and the C-terminal domains, by site-directed mutagenesis (Fig. 1).

FIG. 1.

Multiple-sequence alignment of KsgA orthologs in E. coli K-12, Pseudomonas putida KT2440, B. subtilis, Thermotoga maritime, Chlamydophila pneumoniae AR39, Pyrococcus abyssi GE5, Homo sapiens, and Saccharomyces cerevisiae. The alignment was generated by CLUSTAL W (45). Double-headed arrows indicate motifs (I to VIII) common to S-adenosylmethionine-dependent methyltransferases, and a dotted underline indicates the C-terminal domain of E. coli KsgA (31). Arrowheads indicate the sites of nine point mutations constructed in this study.

By using the ksgA19 strain, we determined which mutations in KsgA affect its methyltransferase activity. Table 1 shows that the ksgA19 cells harboring plasmids containing the ksgA construct with the pUCksgA(E43A), pUCksgA(G47A), or pUCksgA(E66A) mutation were resistant to kasugamycin, similar to the ksgA19 cells harboring only pUC19 (vector). As described above, the resistance to kasugamycin is due to the methyltransferase activity. We conclude that the E43A, G47A, and E66A mutations result in the loss of the methyltransferase activity. Surprisingly, E43A, G47A, and E66A mutations were able to suppress the cold-sensitive cell growth of the Era(E200K) strain. Although the E66A mutation revealed a methyltransferase-deficient phenotype, we decided not to further study this mutant, as it caused severe growth defects in the JM101-ksgA19 strain.

TABLE 1.

Effects of plasmids on sensitivity to kasugamycin and low temperature

Plasmid	Kasugamycin sensitivity in strain JM101-ksgA19a	Suppression of cold-sensitive phenotype in strain CL213/pAC19era(E200K)b
pUC19	R	−
pUCksgA	S	+
pUCksgA(E43A)	R	+
pUCksgA(G47A)	R	+
pUCksgA(E66A)	R	+
pUCksgA(D91A)	S	+
pUCksgA(P115A)	S	+
pUCksgA(F181A)	S	+
pUCksgA(K223A)	S	+
pUCksgA(L229A)	S	+
pUCksgA(R248A)	S	−

^aJM101-ksgA19 cells harboring either pUC19, pUCksgA, or its derivatives carrying the mutations grown to log phase in LB medium at 37°C were spread on LB agar plates containing 0.4 mg/ml kasugamycin, and the plates were incubated at 37°C for 1 day. R, resistant to kasugamycin; S, sensitive to kasugamycin.

^bA cell suspension from a single colony of CL213/pAC19era(E200K) cells harboring either pUC19, pUCksgA, or its derivatives carrying the mutations was streaked on LB agar plates, and the plates were incubated at 25°C for 2 days. A plus sign indicates that the cells were able to grow, and a minus sign indicates that the cells did not grow.

The ksgA19 cells harboring plasmids with the ksgA construct carrying the D91A, P115A, F181A, K223A, L229A, or R248A mutation were sensitive to kasugamycin. Among them, only the R248A mutation was deficient in the suppression of the cold-sensitive phenotype of the Era(E200K) strain. The cellular content of KsgA protein was comparable in all these cells (Fig. 2), as determined by Western blot analysis, suggesting that the loss of the Era-suppressing activity of the KsgA(R248A) mutant is not due to its lower expression level. We confirmed by CD analysis that this loss of activity of the KsgA mutant is not due to its destabilization caused by the R248A mutation. The CD spectrum of the KsgA(R248A) protein resembled that of the wild-type protein (data not shown).

FIG. 2.

Cellular content of KsgA. JM101-ksgA19 cells harboring either pUCksgA, pUCksgA(E43A), pUCksgA(G47A), pUCksgA(R248A), or pUC19 (vector) were grown in LB medium to log phase at 37°C. Equivalent amounts of protein (10 μg) were analyzed with 12% SDS-polyacrylamide gels, followed by immunoblotting with anti-KsgA serum. Lane 1, wild-type KsgA; lane 2, KsgA(E43A); lane 3, KsgA(G47A); lane 4, KsgA(R248A); and lane 5, vector alone.

Defect in the processing of 16S rRNAs in the Era(E200K) strain cannot be suppressed by KsgA.

Recently, we reported that a 16S rRNA precursor (17S rRNA) accumulates in the Era(E200K) strain (14). To examine whether overexpression of KsgA suppresses the 16S rRNA processing defect, total RNAs from the CL213/pAC19era(E200K) strain harboring either pUC19 (vector), pUCksgA, pUCksgA(E43A), pUCksgA(R248A), or pJR302 (harboring the era gene) grown at 37°C and 25°C were extracted, and primer extensions were performed to quantitatively estimate the size of the 5′-end region of 16S and 17S rRNA, as described previously (14). Expression of Era inhibited accumulation of 17S rRNA in the Era(E200K) strain (Fig. 3) as observed previously, while KsgA failed to do so both at 37 and 25°C. This suggests that (i) suppression of the cold-sensitive phenotype of the Era(E200K) strain by KsgA is not related to the defective 16S rRNA maturation in this strain, and (ii) the cold sensitivity of the Era(E200K) strain is not due to the accumulation of 17S rRNA at 25°C.

FIG. 3.

Effect of KsgA on accumulation of 17S rRNA in the Era(E200K) cells. CL213/pAC19era(E200K) cells harboring either pUC19 (vector), pUCksgA, pUCksgA(E43A), or pUCksgA(R248A) grown in LB medium at 37°C to log phase were shifted to 25°C and incubated for 2.5 h. Total RNAs from both 37 and 25°C cultures were extracted by the hot phenol method as described in Materials and Methods. Total RNAs from CL213/pAC19era(E200K) cells harboring pJR302 (harboring the era gene) grown in LB medium at 37°C to log phase also were extracted. Primer expression was performed to measure the ratio of 17S rRNA to 16S rRNA as described previously (14). The amount of radioactivity in the primer extension products was quantified with a PhosphorImager. The results shown are from a single experiment.

Toxic effect of KsgA overexpression.

To examine the suppression of the cold-sensitive phenotype of the Era(E200K) strain by KsgA, we routinely use the pUC-based high-copy-number plasmid carrying the ksgA gene under its own promoter (22). However, with this expression system, only partial suppression can be achieved. To examine whether higher expression levels of KsgA would show better suppressor effects, we constructed an IPTG-inducible plasmid, pINksgA, in which a strong promoter (lpp) is used to express ksgA under the control of the lac promoter-operator. When the plasmid was introduced into an E. coli strain (DH5α), we observed heterogeneous colonies (small and large colonies) on plates. Plasmids isolated from the small and large colonies were designated pINksgA-S and pINksgA-L, respectively. DNA sequencing showed that both plasmids contained the wild-type ksgA gene. However, cells harboring pINksgA-S but not pINksgA-L showed severe growth defects in the presence of IPTG. We next examined cellular amounts of KsgA in the cells harboring these two plasmids. As shown in Fig. 4, although expression levels of KsgA from pINksgA-L (lane 3) were lower than that from pINksgA-S (lane 2), the expression was inducible in the presence of IPTG (lanes 5 and 6), suggesting that the pINksgA-L plasmid gained a mutation in its promoter region leading to lower expression of ksgA. Taken together, these results suggest that overexpression of ksgA is toxic to cell growth. Although the Era(E200K) strain harboring either pINksgA-S or pINksgA-L was able to grow at a low temperature (25°C), the Era(E200K) strain harboring pINksgA-S showed slower growth than the strain harboring pINksgA-L, suggesting that the higher level of expression of KsgA did not correlate with the degree of suppression of the cold-sensitive phenotype. Note that the cellular amount of KsgA expressed from a pUC-based ksgA plasmid (pUCksgA) was significantly lower than that from pINksgA plasmids, as KsgA produced from the pUCksgA plasmid cannot be detected by Coomassie brilliant blue staining. Nevertheless, this amount was sufficient to suppress the cold-sensitive phenotype of the Era(E200K) strain.

FIG. 4.

IPTG-inducible expression of KsgA from pINksgA-L and pINksgA-S. JM101 cells harboring either pINIII (lanes 1 and 4), pINksgA-S (lanes 2 and 5), or pINksgA-L (lanes 3 and 6) grown in M9 medium at 37°C to log phase were incubated in the absence (lanes 1 to 3) and the presence (lanes 4 to 6) of IPTG. Equal amounts of cells were subjected to a 12% SDS-polyacrylamide gel, followed by Coomassie brilliant blue staining. M represents a molecular mass protein marker.

In order to examine whether the KsgA toxicity observed in cells harboring pINksgA-S can be suppressed by the R248A mutation, we constructed pINksgA(E43A)-S and pINksgA(R248A)-S by site-directed mutagenesis using pINksgA-S as the template and transformed them into the wild-type strain (W3110). As expected, cells harboring either pINksgA-S or pINksgA(E43A)-S showed a severe growth defect in the presence of 10 μM IPTG, whereas cells harboring pINksgA(R248A)-S did not show this defect (data not shown). Notably, toxicity was exerted if the cells carrying the KsgA(R248A) mutant were treated with 1 mM IPTG. These results suggest that the toxic effect of KsgA on cells is not due to its methyltransferase activity but is due to another unidentified activity, and this activity is affected by the R248A mutation.

Sensitivity to acetate, propionate, and benzoate induced by overexpression of KsgA.

As described above, overexpression of KsgA cannot reverse the 16S rRNA processing defect caused by the Era(E200K) mutation, suggesting that the Era(E200K) strain has pleiotropic functional defects. It should be noted that Era depletion causes pleiotropic phenotypes, including increased utilization of tricarboxylic acid cycle metabolites as a sole carbon source (21). On the other hand, cells overexpressing another mutant of Era, Era-dE, which carries a deletion mutation at a putative effector region in its GTP-binding domain, show severe growth defects when cells use tricarboxylic acid cycle metabolites as a sole carbon source (32). Interestingly, acetate strongly enhances this growth defect at neutral pH (15). This acetate effect is somewhat surprising, since acetate is known to be a highly toxic compound at acidic, but not neutral, pH (49). This acetate toxic effect caused by Era-dE was not observed at a mildly basic pH (pH 7.6), suggesting that the acetate toxicity is due to a direct effect on intracellular pH homeostasis. To examine whether cells harboring pINksgA exhibit a growth-defective phenotype in the presence of acetate at a neutral pH, E. coli W3110 cells harboring either pINksgA-S or pINksgA-L were incubated on M9 agar plates in the presence of 50 mM acetate (pH 7.0). Interestingly, cells harboring pINksgA-S revealed an acetate-sensitive phenotype (Fig. 5). The cells harboring pINksgA-L did not show this growth defect unless KsgA was overexpressed by IPTG (data not shown). Furthermore, cells harboring pUCksgA did not show any growth defect in the presence of acetate. These results suggest that a higher level of expression of ksgA is required to render the cells sensitive to acetate. Surprisingly, as seen from Fig. 5, this effect was not observed in the cells harboring pINksgA(R248A)-S. The cellular content of KsgA(R248A) protein was similar to that of wild-type KsgA protein (Fig. 6), ruling out the possibility of lower expression levels of the KsgA(R248A) protein leading to a lack of sensitivity of the cells to acetate. We thus conclude that the R248 residue in KsgA is involved in its toxicity and ability to render the cells sensitive to acetate.

FIG. 5.

Inhibition of cell growth of wild-type E. coli cells by KsgA overexpression in the presence of acetate. W3110 cells harboring either pINIII (vector), pINksgA-S, or pINksgA(R248A)-S were streaked on M9 agar plates in the absence (−) and the presence (+) of 50 mM potassium acetate (pH 7.0) and were incubated for 1 day.

FIG. 6.

Expression levels of KsgA from pINksgA-S and pINksgA(R248A)-S. W3110 cells harboring either pINIII (vector), pINksgA-S, or pINksgA(R248A)-S were grown in M9 medium at 37°C to log phase. Equal amounts of cells were subjected to a 12% SDS-polyacrylamide gel, followed by Coomassie brilliant blue staining.

To determine whether the severe growth defect caused by KsgA is due to acetate itself or its metabolites, we examined the toxicity of KsgA using two other weak organic acids, propionate and benzoate, which also are known to influence intracellular pH (3). E. coli K-12 can utilize propionate but not benzoate as a sole carbon source. Interestingly, cells harboring pINksgA-S could not grow on M9 plates in the presence of either propionate or benzoate, suggesting that the toxicity exerted by acetate is not due to the acetate metabolism but rather is due to a property shared by these weak organic acids.

High-level sensitivity of KsgA-induced cells to acid shock.

Induction of the acid tolerance response, which induces acid survival, has been shown to protect Salmonella against weak organic acids such as acetate (2). In order to determine whether the KsgA-induced acetate sensitivity is due to a defective acid tolerance response, W3110 cells harboring either pINIII (vector) or pINksgA-L were exposed to an extreme acid condition. We first induced KsgA for 30 min in M9 medium (pH 7.0), and the cells were exposed to M9 medium (pH 3.0) for 0 to 50 min, followed by being plated on LB agar plates supplemented with ampicillin. Unexpectedly, there was no significant difference in acid sensitivity between KsgA-induced and noninduced cells (data not shown). We next induced KsgA for a longer time period before the acid shock challenge; an overnight culture of the W3110 strain harboring either pINIII or pINksgA-L was diluted 1:1,000 into fresh M9 medium (pH 7.0) in the presence of 1 mM IPTG and was incubated to log phase (OD₆₀₀ of 0.5 to 0.6), followed by acid shock treatment. As shown in Fig. 7, KsgA-induced cells had a severely increased acid sensitivity compared to that of noninduced cells. These results suggest that KsgA-induced cells have impaired functions in the acid shock response and that such a defect results in the acetate-sensitive phenotype of KsgA-induced cells.

FIG. 7.

Effect of KsgA overexpression on acid resistance. An overnight culture of W3110 cells harboring either pINIII (vector) or pINksgA-L was diluted to 1:1,000 in fresh M9 medium containing 1 mM IPTG, and the cells were grown at 37°C to log phase (OD₆₀₀ of 0.5 to 0.6). The log-phase culture was diluted to 1:1,000 into Casamino Acids-free M9 medium (pH 3.0), followed by incubation at room temperature for 0 to 50 min without shaking. No colonies were detected for W3110 cells harboring pINksgA-L at 40 and 50 min. Filled circles, W3110 cells harboring pINIII; open circles, W3110 cells harboring pINksgA-L.

Inhibition of acid-inducible genes by overexpression of KsgA.

A cluster of 12 genes, termed an acid fitness island, is located at 78.8 min on the E. coli K-12 genome (12). Most of the genes are known to be induced under acidic conditions, and mutations in some of the genes have been shown to cause an inability to survive extreme acid conditions (17, 24, 25, 47). In order to examine the expression of the 12 genes in the acid fitness island in KsgA-induced cells using the W3110 strain harboring the pINksgA-L plasmid, we performed a DNA microarray experiment as described in Materials and Methods. We found that the expression of a number of genes involved in the acid fitness island was significantly reduced after 30 min of induction of KsgA (shown in boldface in Table 2). This KsgA effect on gene expression seems to be specific to the acid shock genes, as the other types of the stress response genes, such as heat shock, oxidative shock, and UV irradiation, were not reduced significantly by KsgA (Table 2).

Binding activity of purified KsgA to a double-stranded DNA fragment.

As it has been shown that a human KsgA ortholog, h-mtTFB1, has the ability to bind DNA in a nonspecific manner (26), we found that E. coli KsgA also showed a binding activity to double-stranded DNA (Fig. 8). Interestingly, KsgA(R248A) had less binding activity than wild-type KsgA and KsgA(E43A), as 0.1 μM of both wild-type KsgA and KsgA(E43A) shifted the DNA fragment, while KsgA(R248A) did not show any gel shift. Dissociation constants of purified KsgA binding to two double-stranded 27-bp DNA fragments, one with a high AT content (74.1%) and the other with a high GC content (70.4%), were 1.29 and 1.18 μM, respectively, suggesting that the DNA-binding activity of KsgA is nonspecific, as observed for h-mtTFB1.

FIG. 8.

Interaction of KsgA with double-stranded DNA. The 631-bp DNA fragment corresponding to the GTP-binding domain of Era was incubated with 0, 0.1, 0.2, 1.0, and 10 μM of either wild-type KsgA, KsgA(E43A), or KsgA(R248A) protein for 5 min on ice. The DNA-protein complexes were subjected to 3.5% polyacrylamide gel electrophoresis, followed by ethidium bromide staining.

DISCUSSION

It has been shown that eukaryotic KsgA orthologs, such as h-mtTFB1 in humans (7, 26, 40), Dim1 in yeast (19), and Pfc1 in A. thaliana (46), have other functions in addition to their rRNA methyltransferase activities. Here, we examined if E. coli KsgA has physiologically relevant activities in addition to its methyltransferase activity. Previously, we reported that overexpression of KsgA can suppress the cold-sensitive phenotype of the Era(E200K) strain. In this study, we demonstrated that the dimethyltransferase of KsgA is not required for the suppression of the cold-sensitive phenotype of the Era(E200K) strain, suggesting that KsgA possesses activities other than the methyltransferase activity that have physiological relevance.

The yeast KsgA ortholog, Dim1, can methylate the E. coli 16S rRNA (18). In order to determine if Dim1 can suppress the cold-sensitive phenotype of the Era(E200K) strain, an IPTG-inducible plasmid (pINdim1) expressing the yeast ksgA ortholog (dim1) was introduced into both the Era(E200K) strain and the JM101-ksgA19 strain. Although the kasugamycin-resistant phenotype of the ksgA19 strain was reverted to the kasugamycin-sensitive phenotype by Dim1 as reported previously (18), it was not able to suppress the cold-sensitive phenotype of the Era(E200K) strain (data not shown). This result further confirms that the methyltransferase activity is not responsible for the suppression of the cold-sensitive phenotype of the Era(E200K) strain. It should be noted that the sequence similarity of the C-terminal domains between E. coli KsgA and Dim1 is very poor (Fig. 1). Since the mutation (R248A) in the C-terminal domain of KsgA results in the loss of its Era-suppressing activity while retaining the methyltransferase activity, one can conclude that the C-terminal domain of KsgA is essential for this additional function. The C-terminal domain of E. coli KsgA consists of four α-helices, having five highly conserved residues (F218, R211, R222, K223, and R248) that are spatially clustered (31). Not only the R248 residue but also the other highly conserved residues may be important for the KsgA suppressor effect.

Recently, we observed that a precursor 16S rRNA (17S rRNA) accumulates in the Era(E200K) strain (14). In the present study, we found that the defect in the processing of 16S rRNA cannot be suppressed by overexpression of KsgA, suggesting that the suppression of this strain by KsgA is not associated with ribosome biogenesis. Since Era possesses multiple functions, such as cell cycle regulation (5, 6, 11), energy metabolism (15, 21, 33), and ribosome biogenesis (14, 29), it is reasonable to propose that the Era(E200K) mutation confers two independent functional defects; one is associated with ribosome biosynthesis, and the other is an unidentified function essential for cell growth at a low temperature. We propose that the suppression of this strain by KsgA likely is related to this unidentified role of Era.

What could be the additional unknown function of KsgA? In this study, we found that acetate became a highly toxic compound even at neutral pH in cells overexpressing KsgA. Such a severe growth defect also was observed in the presence of other weak organic acids, such as propionate and benzoate, suggesting that this effect is not directly related to acetate metabolism but rather is related to the effect of the addition of weak acids to the culture medium. These chemicals are known to enter cells in uncharged (protonated) forms and to be deprotonated in the cells, lowering the intracellular pH (3). Therefore, one can speculate that overexpression of KsgA affects intracellular pH homeostasis so that cells become incapable of efficiently adapting to such chemicals. Indeed, we observed that KsgA-induced cells were highly sensitive to an extreme acid condition (pH 3.0). Since 30 min of induction of KsgA was not enough to reveal this acid-sensitive phenotype, KsgA overexpression itself does not seem to cause the acid-sensitive phenotype, and therefore a cellular factor(s) may be involved in this phenotype. In this study we showed that the expression of a number of genes involved in the acid fitness island was significantly reduced by KsgA. One can speculate that KsgA overexpression inhibits the gene expression responsible for the acid shock response, leading to the acid-sensitive phenotype.

Recently, it has been shown that the heat-stable nucleoid-structuring (H-NS) protein directly inhibits gadA and gadX transcription by binding to the promoter regions of these genes, which are known to be the major components of the acid resistance system. GadA is one of the isozymes for glutamate decarboxylase, and GadX is one of the primary activators of the gad system (48). Deletion of the hns gene encoding H-NS protein increases resistance to acid shock compared to that of its parental strain (13). In this study, we observed that wild-type KsgA has the ability to bind double-stranded DNA, while the KsgA(R248A) mutant has weaker activity. Since the acetate toxic effect induced by KsgA was not observed in cells overexpressing KsgA(R248A), we can speculate that the DNA-binding activity of KsgA has an important role in negative regulation of the genes responsible for the acid shock response. Interestingly, H-NS, like KsgA, reveals a severe growth defect when overexpressed (43).

h-mtTFB1, a human KsgA ortholog, originally was found as a mitochondrial transcription factor (26) and recently was demonstrated to have a methyltransferase activity modifying tandem adenine residues in the small rRNA subunit of ribosomes (40). It has been shown that the methyltransferase of h-mtTFB1 is not likely to play an essential role in its transcription function (27), suggesting that h-mtTFB1 is a dual-function protein. It should be noted that h-mtTFB1 has a sequence-nonspecific DNA-binding activity, as we observed in KsgA in this study. The nonspecific DNA-binding activity of KsgA might be involved in the regulation of the acid-inducible genes. Further analysis is necessary to identify the physiological relevance of the DNA-binding activity of KsgA.

It has been reported that mutants of S. enterica lacking polyphosphate kinase (ppk) grow poorly in the presence of weak organic acids such as acetate, propionate, and benzoate (34). These authors speculated that the growth defect likely is due to unfolded MetA protein (the first enzyme in the methionine biosynthesis pathway), as this sensitivity can be suppressed by either the addition of 0.3 mM methionine or overexpression of the metA gene. However, the acetate-sensitive phenotype of cells overexpressing KsgA observed in the present study was not suppressed by methionine (even at 100 mM), suggesting that the effect of overexpression of KsgA is not related to MetA destabilization. In order to determine other multicopy suppressor genes besides the ksgA gene, we repeated suppressor screening for the Era(E200K) cold-sensitive phenotype using a pUC-based genomic library. However, a significant number of suppressor genes identified were either the era gene or the ksgA gene. This emphasizes the role of KsgA in suppressing function for Era at a low temperature.

Although in the present study we unambiguously demonstrated that KsgA has an activity, in addition to its 16S rRNA methyltransferase activity, that is physiologically relevant, further analyses are necessary to understand the manifestations and exact mechanism of this activity. Analyses of protein factors that interact with wild-type KsgA, but not with KsgA(R248A), using yeast-two hybrid analyses are currently under way in our laboratory.