Escherichia coli mrsC Is an Allele of hflB, Encoding a Membrane-Associated ATPase and Protease That Is Required for mRNA Decay

Abstract

The mrsC gene of Escherichia coli is required for mRNA turnover and cell growth, and strains containing the temperature-sensitive mrsC505 allele have longer half-lives than wild-type controls for total pulse-labeled and individual mRNAs (L. L. Granger et al., J. Bacteriol. 180:1920–1928, 1998). The cloned mrsC gene contains a long open reading frame beginning at an initiator UUG codon, confirmed by N-terminal amino acid sequencing, encoding a 70,996-Da protein with a consensus ATP-binding domain. mrsC is identical to the independently identified ftsH gene except for three additional amino acids at the N terminus (T. Tomoyasu et al., J. Bacteriol. 175:1344–1351, 1993). The purified protein had a K_m of 28 μM for ATP and a V_max of 21.2 nmol/μg/min. An amino-terminal glutathione S-transferase–MrsC fusion protein retained ATPase activity but was not biologically active. A glutamic acid replacement of the highly conserved lysine within the ATP-binding motif (mrsC201) abolished the complementation of the mrsC505 mutation, confirming that the ATPase activity is required for MrsC function in vivo. In addition, the mrsC505 allele conferred a temperature-sensitive HflB phenotype, while the hflB29 mutation promoted mRNA stability at both 30 and 44°C, suggesting that the inviability associated with the mrsC505 allele is not related to the defect in mRNA decay. The data presented provide the first direct evidence for the involvement of a membrane-bound protein in mRNA decay in E. coli.

mRNA turnover is thought to be an important regulator of gene expression in both prokaryotes and eukaryotes. The generally accepted model for prokaryotic mRNA decay utilizes a combination of exonucleolytic and endonucleolytic cleavages (3, 11). While many laboratories have described specific endonucleolytic cleavages in mRNAs (5, 10, 40, 42, 45, 46, 57), the relatively normal mRNA decay in a polynucleotide phosphorylase (PNPase) RNase II RNase III RNase E multiple mutant (7) has suggested that the overall pathway is more complicated than previously assumed.

We thus adopted a genetic, rather than a biochemical, approach to identify new genes that encoded either RNases or regulatory proteins involved in mRNA turnover. This search led to the identification of mrsC (mRNA stability), a gene required for cell growth and normal mRNA decay (29). When we cloned and sequenced mrsC, we found that it was almost identical to ftsH (74), a gene encoding a protein homologous to a series of eukaryotic ATP-binding proteins involved in a variety of biological processes (74). Using a glutathione S-transferase (GST)–MrsC fusion, we purified the protein to homogeneity and determined a K_m for ATP of 28 μM. A constitutive promoter was identified by primer extension analysis. In addition, N-terminal amino acid sequencing showed that translation starts from a UUG codon, adding three amino acids to the N terminus of the published protein sequence (74). Changing the conserved lysine in the ATP-binding domain to glutamic acid destroyed the biological activity of MrsC, in agreement with observations by Akiyama et al. (2).

Since it has been shown that ftsH and hflB are allelic (32), we compared the phenotypic properties of the hflB29 and mrsC505 mutations. We show that mrsC505 confers a temperature-sensitive HflB phenotype, while hflB29 leads to mRNA stabilization at both 30 and 44°C. We also discuss how a protease could affect mRNA decay in the context of recent observations that HflB/MrsC can degrade a variety of proteins, including ς³² (72), λCII, SecY (35), and subunit a of proton ATPase F₀ sector (1).

MATERIALS AND METHODS

Bacterial strains, plasmids, λ phage, and growth media.

All strains and plasmids used in this study are described in Tables 1 and 2. W1485 was the Tn1000 donor strain. XL1-Blue (Stratagene) and JM83 were used as host strains for recombinant plasmids. SL215 (hflB⁺) and SL216 (hflB29) (9, 32) were obtained from Christophe Herman. K38 and plasmid pGP1-2 were kindly provided by Stan Tabor (69). Plasmid pBluescript II SK+ was purchased from Stratagene. pLG339 is a low-copy-number vector (66), and pMAK904 is a derivative of pLG339 (75a). pWSK29 is a multipurpose, low-copy-number vector (75). λCam105 (32) was graciously provided by Phillipe Bouloc.

TABLE 1.

E. coli strains used in this study

Strain	Genotype	Reference or source
K38	HfrC	69
JC1557	argG6 hisG1 leuB7 metB1 rpsL309	A. J. Clark
JC10240	recA56 srlA::Tn10 ilv-318 thr-300 thi-1 rpsE300	A. J. Clark
JM83	ara rpsL Δ(lac proAB) φ80 lacZΔM15	78
MG1693	thyA715	E. coli Genetic Stock Center
SK2732	hisG1 leuB7 metB1 rpsL309 mrsC505 mrsB1 [pVK88B (Tcrqa-2+)]	29
SK5665	thyA715 rne-1	6
SK5704	thyA715 rne-1 pnp-7 rnb-500	6
SK6501	MC4100 recA56	F. Moreno
SK6827	hisG1 leuB7 metB1 rpsL309 mrsC505	JC1557 × P1 SK2732 ArgG+ transductant
SK7012	hisG1 leuB7 metB1 rpsL309 mrsC505 recA56 srlA::Tn10	SK6827 × P1 JC10240 Tcr transductant
SK8232	thyA715 mrsC505 zgj-203::Tn10	29
SK8236	thyA715 mrsC505 pnp-7 rnb-500 rne-1 zgj-203::Tn10	29
SK8238	thyA715 mrsC505 pnp-7 rnb-500 rne-1 zgj-203::Tn10	29
SK8244	thyA715 rne-1 mrsC505 zgj-203::Tn10	29
SK8945	thyA715 hflB29 zgj-203::Tn10	This work
SK8948	thyA715 pnp-7 rnb-500 rne-1 hflB29 zgj-203::Tn10	This work
SL215	lac-3350 galK2 galT22 rpsL179 zgj-203::Tn10	32
SL216	lac-3350 galK2 galT22 rpsL179 hflB29 zgj-203::Tn10	32
XL1-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac (F′ proAB+ lacIqlacZΔM15 Tn10)	Stratagene
W1485	F+	70

TABLE 2.

Plasmids used in this study

Plasmid	Description	Source or reference
pBluescript II SK+	AprlacZα	Stratagene
pCIK1	Contains a 23-kb HindIII chromosomal DNA fragment in pDF41	This study
pDF41	Single-copy vector carrying immunity to colicin E1	34
pET22b	T7 promoter-derived expression vector, Apr	Novagen
pGEX-2X	GST fusion expression vector, Apr	64
pGP1-2	Carries the T7 RNA polymerase gene under λcI857, Kmr	69
pLG339	Kmr Tcr derivative of pSC101	66
pMAK900	Contains a 23-kb HindIII fragment, Kmr	This study
pMAK902	Contains a 7.5-kb EcoRI fragment, Kmr	This study
pMAK904	KmrlacZα	This study
pMAK905	Contains a 4.3-kb SalI-EcoRI fragment in pMAK904, Kmr	This study
pMAK906	Contains a 6.5-kb SalI fragment in pMAK904, Kmr	This study
pPCR52	Contains a 625-bp PCR amplified fragment in pWSK29, Apr	This study
pWK10	Contains a 2,315-bp EcoRI-PstI fragment in pGEX-2X, Apr	This study
pWK52	Contains a 2,315-bp EcoRI-PstI fragment in pMAK904, Kmr	This study
pWSK29	AprlacZα	75
pWSK29K	Same as pWSK29 but no KpnI site, Apr	This study
pWK101	Contains a 3.0-kb SalI-PstI fragment in pWSK29, Apr	This study
pWK100	Same as pWK101 in opposite orientation in pWSK29, Apr	This study
pWK210	Contains a 2.2-kb ApaI-PstI fragment in pWSK29, Apr	This study
pWK211	Contains a 1.8-kb KpnI-PstI fragment in pWSK29, Apr	This study
pWK212	Contains a 1.5-kb ClaI-PstI fragment in pWSK29, Apr	This study
pWK213	Contains a 1.6-kb SalI-ClaI fragment in pWSK29, Apr	This study
pWK214	Contains a 2.45-kb SalI-SmaI fragment in pWSK29, Apr	This study
pWK911	Contains a 6.2-kb EcoRI-PstI fragment in pMAK904, Kmr	This study
pWK912	Contains a 3.0-kb SalI-PstI fragment in pMAK904, Kmr	This study
pWK914	Contains a 3.3-kb EcoRI-BamHI fragment in pMAK904, Kmr	This study
pWK926	Same as pMAK902 but 1.0-kb ApaI-BspmII deleted, Kmr	This study
pWK936	Contains a 3.0-kb SalI-PstI fragment in pWSK29K, Apr	This study
pWK937	Same as pWK936 but Lys-201 changed to Glu-201 to give mrsC201, Apr	This study

Media used were LB and LB solidified with 2% agar (44) as well as 2× YT medium, used to grow cells to isolate plasmid DNA and overexpress the mrsC gene. Supplements were added at the following concentrations (micrograms/milliliter): thymine, 50; kanamycin, 50; ampicillin, 100; streptomycin, 20; and tetracycline, 50. Fifty microliters of 1% 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) and 20 μl of 100 mM isopropyl-β-d-thiogalactopyranoside (IPTG) were added to the solid medium for blue/white selection of recombinant plasmids. Minimal medium consisted of M56/2 buffer supplemented with glucose (0.5%) and appropriate amino acids (50 μg/ml) and antibiotics (41).

Recombinant DNA techniques.

Plasmid construction, phage DNA isolation, restriction digestions, DNA ligations, transformations, and gel electrophoresis were carried out by using standard techniques (41). Transformants of temperature-sensitive mutants were heat shocked at 37°C for 2 min and then incubated at 30°C.

Plasmid constructions.

Plasmid pCIK1 contained a 23-kb HindIII chromosomal DNA fragment in the single-copy vector pDF41 (22). Plasmids pMAK905 and pMAK906 were pMAK904 derivatives containing a 4.3-kb SalI-EcoRI fragment and a 6.5-kb SalI fragment from pCIK1, respectively. Plasmid pWK912 was constructed by cloning the 3-kb SalI-PstI fragment from pMAK905 into pMAK904. Plasmid pWK101 was generated by inserting the 3-kb SalI-PstI fragment from pWK912 into pWSK29 such that the mrsC gene was controlled by a T7 promoter. pWK100 was obtained by inserting the blunted SalI-PstI fragment into the blunted ClaI site of pWSK29 in opposite orientation to pWK101. Both pWK101 and pWK100 were used for generating nested deletions either by the ExoIII-S1 nuclease method (31) or by subcloning for DNA sequencing and genetic complementation. Other plasmids were constructed as indicated in Table 2.

Isolation of Tn1000 insertions.

Plasmid pMAK906 (Km^r) was transformed into strain W1485 (70), which carries transposon Tn1000 on a resident F′ (30). After conjugation into SK6827 (mrsC505), Km^r Sm^r transconjugants were selected at 30°C. Colonies that were unable to grow (or that grew as isolated colonies on replica plates) at 44°C were selected for further analysis. Plasmid DNAs were isolated from these colonies and subjected to restriction enzyme digestion to determine the location and direction of insertion. Representative pMAK906::Tn1000 insertions were used for maxicell analysis.

Measurement of lysogenization frequencies.

MG1693 was infected with λCam105, and lysogens were picked from the center of the turbid plaques. Individual Cm^r lysogens were grown in L broth containing chloramphenicol and induced with 1 μg of mitomycin C per ml when the culture had reached a cell density of 10⁸/ml. The cell lysates obtained were titered on MG1693. One hundred individual lysogens obtained from each lysate were tested by replica plating for Cm^r. Between 65 and 92% of the λ phage particles carried Cm^r, with the percentage varying from induction to induction. All experiments presented here were carried out with the same λCam105 lysate.

Lysogenization measurements were performed as described by Herman et al. (32) with the exception that infectious centers were measured with MG1693 as a tester strain. Lysogenization frequencies were determined as (number of Cm^r lysogens/number of infective centers) × 100. The number of infective centers was corrected to reflect the percentage of phage particles that did not carry Cm^r.

RNA blotting analysis and mRNA half-life determinations.

Northern analyses, including RNA isolations, were carried out as described by Granger et al. (29). The half-lives of individual full-length transcripts were determined following analysis of the blots with a PhosphorImager (model 400 series; Molecular Dynamics) and linear regression analysis.

Heat shock measurements.

Cells were grown at 30°C to a Klett reading of 50 (no. 42 green filter) in L broth containing thymine. An equal volume of medium at 58°C was then added to the cells and rapidly mixed. Following addition of the hot medium, the cells were shaken in a 44°C water bath. Samples (14 ml each) were removed at various times and added to crushed frozen TM buffer (10 mM Tris [pH 7.2], 5 mM MgCl₂). RNA was isolated by using Catrimox-14 as described previously (50). groES and lpp mRNA levels were determined by RNA-DNA dot blotting (29) and quantitated with a PhosphorImager (model 400 series; Molecular Dynamics).

Analysis of plasmid-encoded proteins.

We used the procedure of Sancar et al. (59) to identify plasmid-encoded proteins in the maxicell strain SK6501 (recA56). SK6501 was transformed with various plasmids, irradiated, and labeled with [³⁵S]methionine. The labeled proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (37). After each gel was dried, the labeled protein bands were visualized by autoradiography.

DNA sequencing.

We generated a series of nested deletions of both strands from pWK101 and pWK100 for DNA sequencing (31). Plasmid DNAs were isolated and digested with an appropriate restriction enzyme to determine the deletion size. Certain regions were sequenced either after subcloning of specific restriction fragments from pWK912 and pWK101 or by using synthetic primers.

The nucleotide sequence of the mrsC gene was determined by the dideoxy-chain termination method (60), using either Sequenase version 2.0 (United States Biochemical) or a Taq DNA polymerase sequencing kit (Promega) with [α-³⁵S]dATP on either single-stranded or double-stranded DNA. Compressions in GC-rich regions were analyzed by using Taq polymerase at a reaction temperature of 70°C. Samples were resolved on 6% polyacrylamide sequencing gels.

Primer extension analysis.

Following the procedure of Williams and Rogers (76), we isolated total RNA from SK6501 that contained pWK101 (mrsC⁺). We determined the RNA concentration spectrophotometrically and ran the RNA on a 1.5% agarose gel to ensure that it was intact and contained no contaminating DNA. We end labeled the primer (5′GCCATTAGACTCGCTGGGCCCAAAGCTCTG3′) with T4 DNA polynucleotide kinase. RNA was annealed to 0.5 pmol of ³²P-labeled primer at 30°C overnight in 30 μl of hybridization buffer [40 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) (pH 6.4), 0.4 M NaCl, 1 mM EDTA, 80% formamide]. After annealing, RNA-primer hybrids were precipitated, dried, and dissolved in 25 μl of reverse transcriptase buffer containing 50 mM Tris-HCl (pH 7.6), 60 mM KCl, 10 mM MgCl₂, 1 mM dithiothreitol, and 1 mM each deoxynucleoside triphosphate. We added 50 U of avian myeloblastosis virus reverse transcriptase and incubated the mixture for 90 min at 42°C. After elongation of the primer, the synthesized fragments were digested with 1 U of RNase (DNase free) at 37°C for 30 min and then precipitated with ethanol. We dissolved the pellets in 6 μl of Tris-EDTA buffer and added 6 μl of formamide sequencing dye (98% deionized formamide, 1 mM EDTA, 0.3% xylene cyanol, 0.3% bromophenol blue). The samples were heated for 3 min at 95°C and run on a 6% sequencing gel. We obtained a sequencing ladder by using the same primer to generate size markers. The gel was fixed in 10% methanol–10% acetic acid, dried, and exposed to Kodak X-Omat AR film overnight at room temperature.

T7 expression and N-terminus sequencing of MrsC.

We overexpressed and labeled the MrsC protein by transforming plasmid pWK101 (Ap^r) into a K38 strain harboring plasmid pGP1-2 (Km^r, T7 RNA polymerase) (69). Transformants were selected on plates containing ampicillin and kanamycin. We grew the purified colonies at 30°C in LB medium with ampicillin and kanamycin to an A₆₀₀ of 0.4. Cells were harvested and washed with M9 medium. Subsequently, the cells were grown for 60 min in M9 supplemented with all amino acids except Met and Cys, and then the temperature was shifted to 42°C. After 20 min, we added rifampin (200 μg/ml; Sigma) and incubated the culture at 42°C for another 10 min. We then lowered the temperature to 37°C for 20 min. To label the protein, we added 10 μCi of [³⁵S]methionine to 1 ml of culture and incubated the mixture for 5 min. Whole-cell protein extracts were prepared and subjected to SDS-PAGE.

For protein sequencing, we grew a 100-ml induced culture for an additional 2 h at 37°C. The protein samples were run on an SDS–8% polyacrylamide gel as specified by Applied Biosystems (4a). ³⁵S-labeled proteins were loaded next to the preparative well as a marker. After electrophoresis, proteins were electroblotted onto an Immobilon-P polyvinylidene difluoride membrane. The proteins were visualized by staining in 0.2% (wt/vol) Coomassie blue R-250 in 45% methanol–10% acetic acid for 30 s and destaining with 90% methanol–7% acetic acid. We then exposed the blotted membrane to X-Omat AR film for 30 min at room temperature. The protein band that was the same size as the labeled MrsC protein was excised from the membrane and sequenced on an Applied Biosystems model 477A sequencer.

Construction of plasmids for protein expression.

To create an EcoRI (underlined in the MrsCU5 primer sequence) site six nucleotides upstream of the translational start codon of MrsC, we designed two primers, MRSCU5 (5′GGAATTCCCATGAGTGACATGGC3′) and MRSCU2 (5′CAGCAGCGTTTTACCGGGTACCC). We used these primers to amplify by PCR a DNA fragment from pWK101. This fragment was subsequently cloned into pWSK29. The resulting plasmid, pPCR52, was digested with EcoRI and ApaI. The resulting 80-bp DNA fragment was purified and used to replace the 600-bp EcoRI-ApaI DNA fragment in pWK912 to generate pWK52. We then cloned the EcoRI-PstI (blunt-ended) fragment containing the mrsC coding region and two additional codons at the N terminus from pWK52 into the SmaI sites of pGEX-2X (Pharmacia). The resulting plasmid, called pWK10, was used for expressing the GST-MrsC fusion protein. The junction region between the GST gene and mrsC was sequenced and shown to be translationally in frame.

Purification of the GST-MrsC fusion protein.

The fusion expression plasmids pWK10 (GST-MrsC) and pGEX-2X (GST) were transformed into either DH5α or JM83. To purify the protein, we grew 1-liter Escherichia coli cultures at 37°C in LB medium to an optical density at 600 nm of 0.6 to 1.0 and then induced them with 1 mM IPTG for another 4 h at 37°C. We harvested the cells and resuspended them in 7.5 ml of phosphate-buffered saline (0.15 M NaCl, 16 mM Na₂HPO₄, 4 mM NaH₂PO₄ [pH 7.3]) containing 1% Triton X-100 (PBST). After one cycle of freeze-thawing, the cells were lysed at 4°C for 30 min by adding 1.25 ml of 0.2 M EGTA, 0.5 ml of 5 M NaCl, 1.25 ml of lysozyme (10 mg/ml) and 25 μl of 50 mM phenylmethylsulfonyl fluoride. We adjusted the slurry to 10 mM MgCl₂–1 mM MnCl₂ before adding 10 μg of DNase I per ml of slurry, incubating it at room temperature for 20 min, and then centrifuging it for 30 min at 12,500 rpm. The supernatants from strains containing expressed GST or GST-MrsC were passed over a glutathione-Sepharose 4B (Sigma) affinity column (1 by 10 cm), extensively washed with PBST, and then eluted in 50 mM Tris (pH 8.0) containing 10 mM glutathione. The eluate was dialyzed against 20 mM Tris (pH 8.0)–0.1 mM Na₂EDTA–0.5 mM EGTA–0.1 mM phenylmethylsulfonyl fluoride–3 mM β-mercaptoethanol–50% glycerol and then stored at −20°C. We determined protein concentrations by Bradford’s method (11a). To determine the purity of GST-MrsC fusion protein fractions, we used SDS-PAGE followed by either Coomassie blue or silver staining.

ATPase activity.

ATPase activity was determined by measuring the amount of ³²P_i released from [γ-³²P]ATP. The purified protein was added to a final volume of 75 μl of assay mixture that contained 20 mM Tris-HCl (pH 8.0), 1 mM MgCl₂, 200 μg of bovine serum albumin, and 0.5 mM ATP. After incubating the mixture at 37°C for 10 min, we quenched the reaction by adding 500 μl of ice-cold activated charcoal (1.0% [wt/vol]) suspended in 0.2 N HCl–1 mM NaH₂PO₄. ³²P_i was separated from [γ-³²P]ATP (65), and the amount of ³²P_i released was quantitated by liquid scintillation counting of the resulting supernatant samples. One unit of enzyme activity is defined as the amount of enzyme required to catalyze the hydrolysis of 1 nmol of [γ-³²P]ATP in 10 min at 37°C.

Site-directed mutagenesis.

We used PCR and mismatched oligonucleotide primers to introduce a single amino acid change, from AAA (Lys-201) to GAA (Glu-201), into the consensus ATP-binding domain of MrsC. The mismatched oligonucleotide primer (MRSCU3, 5′GGTACCGGTGAAACGCTGCTGG3′) and the second PCR primer (MRSCU4, 5′CGTCGATTTCATCGATAAAGAT3′) were designed to encode unique restriction sites, KpnI and ClaI (underlined) present in the mrsC gene. We amplified a DNA fragment from 100 ng of pWK101 template DNA, using 35 cycles each consisting of 1 min at 94°C, 1.5 min at 54°C, and 1 min at 72°C. We initially cloned the amplified PCR product into the HincII site of pBluescript II SK+ and then gel purified a 180-bp KpnI-ClaI DNA fragment containing the desired mutation. Using the same enzymes, we cut out the corresponding wild-type sequence from pWK936 and replaced it with the mutagenized KpnI-ClaI DNA fragment. We confirmed the identity of the mutation as well as the correctness of the entire PCR-derived sequence within the ATP-binding site by using the dideoxy termination sequencing method (60). The biological activity of the mutated mrsC gene (mrsC201) was tested by determining if it could complement mrsC505 in SK7012 (mrsC505 recA56) at 44°C.

Computer analysis.

We analyzed the DNA sequence data by using the Wisconsin Genetics Computer Group sequence analysis program (18) and the IG program of the Biological Sequence/Structure Computational Facility at the University of Georgia. GenBank databases were searched by using the FASTA program (51).

Nucleotide sequence accession number.

The sequence reported has been assigned GenBank accession no. M93424.

RESULTS

Cloning the mrsC gene.

Since the mrsC505 mutation genetically mapped near the argG locus on the E. coli chromosome (29), we constructed a genomic library by using the single-copy vector pDF41 (22). From this library, we obtained a plasmid (pCIK1) containing a 23-kb HindIII DNA fragment that complemented both argG6 and mrsC505. A series of subclonings and genetic complementation analyses showed that pMAK906 and pMAK902 (Fig. 1A) restored the growth of SK6827 (mrsC505) after a shift to the nonpermissive temperature but that pWK926 did not (Fig. 1A).

FIG. 1.

Genomic location of mrsC/ftsH/hflB. (A) Physical map of the mrsC region at 69 min on the E. coli chromosome and of the subclones and Tn1000 insertion derivatives. Arrows below the 10-kb EcoRI-SalI fragment indicate the position and direction of transcription of the mrsC gene. pMAK902, pMAK906, and pWK926 contain pMAK904, a derivative of pLG339, as the cloning vector. The arrowheads indicate the locations of Tn1000 insertions in pMAK906, determined by restriction enzyme analysis. (B) Locations and lengths of the subclones of the mrsC gene. At the right are results of complementation tests of the mrsC505 mutation in recA⁺ (SK6827) and recA56 (SK7012) isogenic strains with various plasmids. +, full complementation; −, no complementation; pc, partial complementation arising presumably by recombination between the cloned DNA fragment and the chromosomal sequence. The location of the lysine 201-to-glutamic acid substitution in pWK937 is indicated by the upward arrow. Restriction enzyme abbreviations: A, ApaI; C, ClaI; E, EcoRI; K, KpnI; P, PstI; S, SalI; Sm, SmaI.

To locate the position of the mrsC gene, we isolated Tn1000 insertions in pMAK906 (see Materials and Methods). Approximately 400 transconjugants of SK6827 (mrsC505) were grown at 30°C on LB agar plates containing kanamycin and streptomycin. Four of these were found to be inviable at 44°C. All of the insertions that abolished the complementation of the mrsC505 allele were located in the region between the ApaI and ClaI sites (Fig. 1A).

We also made a series of 5′ and 3′ deletions to generate plasmids pWK210, pWK211, and pWK212 (Fig. 1B). These subclones did not complement SK7012 (mrsC505 recA56), suggesting that deletion of as little as 600 bp from the 5′ end of the fragment disrupted MrsC activity.

Recombination between homologous sequences on the plasmid and the mrsC505 allele on the chromosome was most likely the cause of isolated colonies within the replicated patches over time at 44°C with plasmids pWK210, pWK211, and pWK212. This phenomenon has been observed with the rne/ams gene as well (8, 15). Based on these results, a 3-kb SalI-PstI DNA fragment was cloned into pMAK904 to generate pWK912. This plasmid complemented the mrsC505 allele in SK6827, SK7012, and SK8232.

Identification of the proteins encoded by mrsC.

Maxicell analysis indicated that the mrsC gene produced two protein products of 75 and 62 kDa (data not shown). To test whether the two MrsC proteins were either (i) produced from different transcripts or (ii) produced from the same transcript but different initiation codons, we used a T7 expression system to direct exclusive transcription of the mrsC gene. pWK101 carried the SalI-PstI fragment in which the T7 promoter controlled the mrsC gene. E. coli K38 containing only pGP1-2 and K38 containing pGP1-2 and pWK101 (mrsC⁺) were induced at 44°C and then labeled with [³⁵S]methionine for 5 min. We then separated the labeled proteins by SDS-PAGE. Only the 75-kDa protein was observed (data not shown), suggesting that the 62-kDa protein observed in maxicells was a proteolytic product of the 75-kDa protein.

Sequence analysis of the mrsC gene.

To sequence the mrsC gene, we cloned the 3.0-kb SalI-PstI fragment into pWSK29, a low-copy-number derivative of pBluescript (75). We then used the resulting plasmids, pWK101 and pWK100, to generate nested deletions in both orientations. The 3.0-kb SalI-PstI DNA fragment was sequenced by standard dideoxy-chain sequencing techniques (60). Our sequence analysis revealed a 1,938-bp open reading frame of 646 amino acids that could encode a protein product of 70,996 Da.

Primer extension analysis identified a single major initiation point for the mrsC transcript, beginning at a G at nucleotide 519 (Fig. 2 and 3). Upstream of the transcription start point were the sequences TATCCT at −10 and TTGAAA at −35, spaced 17 nucleotides apart, consistent with the −10 and −35 regions of other E. coli ς⁷⁰ promoters and yielding a homology score of 58.5% (47).

FIG. 2.

Primer extension analysis of the transcriptional start point of mrsC. The primer-extended products were electrophoresed in a 6% sequencing gel. Lane 1, labeled primer but no RNA; lane 2, labeled primer and 10 μg of total RNA; lane 3, labeled primer and 30 μg of total RNA; lane 4, labeled primer and 50 μg of total RNA. The DNA sequencing ladder was produced by the dideoxy-chain termination method (60), using the same primer. The G, A, T, and C lanes have been converted to C, T, A, and G as indicated so that the reading direction from bottom to top is 3′ to 5′ and the sequence is the same as the coding strand as shown on the left. The site of transcription initiation at G is indicated by an arrow.

FIG. 3.

Transcription start and N-terminal amino acid sequence of mrsC/hflB. The coding strand is shown in the 5′-to-3′ direction. Numbering of the nucleotides, starting from the first base of the SalI site, is indicated on the right. The promoter (the −35 and −10 regions) of the mrsC gene is underlined, and the transcriptional start site (G519) is indicated by a rightward arrow. The translation start point and ribosome-binding (Shine-Dalgarno [S/D]) site are shown. The first 19 amino acids confirmed by sequencing the N-terminal region of MrsC protein are underlined.

We found that two putative translational start codons, UUG (nucleotides 553 to 555) with an AAGAGG ribosome-binding site and AUG (562 to 564) with a GAG ribosome-binding site, encoded potential MrsC proteins of similar sizes (Fig. 3). To determine which one was used for the translational start codon of MrsC, we sequenced the N-terminal region of the MrsC protein (see Materials and Methods). As indicated in Fig. 3, the first 18 amino acids obtained matched exactly the predicted amino acid sequence based on the UUG initiation codon with the AAGAGG ribosome-binding site, found eight nucleotides upstream. The mrsC gene is identical to E. coli ftsH (74) except for three additional amino acids at the N terminus.

The open reading frame stopped at UAA (2461 to 2463) and was followed by a putative rho-independent transcriptional terminator (13) (data not shown). Immediately upstream of the putative terminator is the sequence ACUGUAUUUG, which closely resembles sequences of two known RNase E cleavage sites: the ACAGAAUUUG sequence in 9S rRNA (4, 26) and the ACAGUAUUUG sequence in RNA I, an inhibitor of plasmid ColE1 replication (39, 71). Preliminary experiments have demonstrated a 10-fold increase in the mrsC/hflB/ftsH mRNA half-life in an rne-1 strain (39a).

ATPase activity is an intrinsic property of the MrsC protein.

Computer analysis revealed that the MrsC/HflB protein contained a consensus ATP-binding domain within a 220-amino-acid region that was homologous to a family of eukaryotic ATP-binding proteins (73). In addition to the nine proteins that Tomoyasu et al. (73) identified, significant homology also exists with S4 (19) from human cells and with Mts2 (27) and Cim5-1 (25) from yeast. Because the mrsC gene contains an ATP-binding motif, we tested the MrsC protein for ATPase activity. To facilitate our analysis, we overexpressed the MrsC protein by using Tabor and Richardson’s T7 system (69). As shown, a 2-h postinduction incubation significantly increased MrsC concentration relative to total cellular proteins (Fig. 4A). We compared the ATPase activity in crude extracts from induced strains that contained either pWK101 (mrsC⁺) or the vector alone. Initially, ATPase activity in the strain containing pWK101 increased only threefold compared with the control strain. By adding 1% Triton X-100 to the lysis buffer, we increased the recovery of soluble MrsC protein from the protein pellet by about 30%. The ATPase activity of this fraction was about 15-fold higher than that from the control strain, indicating that the MrsC protein contained an ATPase activity (data not shown).

FIG. 4.

Expression and purification of MrsC. (A) Overexpression of the MrsC protein by using a T7 system. K38 strains containing either pGP1-2 alone or pPG1-2 and pWK101 were induced by a shift to 42°C for 20 min and then grown at 30°C for an additional 2 h. Lanes 1 and 2, 20 μl of sample from induced strains containing both pGP1-2 and pWK101; lane 3, 20 μl of sample from uninduced strains with pGP1-2 and pWK101; lane 4, 20 μl of sample from induced strain containing pGP1-2 alone. Protein samples were separated by SDS-PAGE (10% gel) and stained with Coomassie blue. (B) Analysis of purified GST and GST-MrsC fractions from a glutathione-agarose column (Sigma) (see Materials and Methods). Lanes 1 and 10, molecular size marker; lane 2, 2 μl of whole-cell sample from strains overexpressing GST-MrsC fusion protein; lane 3, 2 μl of sample from pellet following centrifugation of cell lysate; lane 4, 2 μl of sample from supernatant following centrifugation of cell lysate; lane 5, 5 μl of GST-MrsC protein sample of fraction 1 eluted from glutathione-agarose column; lane 6, 10 μl of GST-MrsC protein of fraction 2 eluted from column; lanes 7 and 8, 1 μl of GST protein sample from fractions 2 and 1, respectively, collected from the glutathione-agarose column; lane 9, 1 μl of whole-cell sample from strain overexpressing GST. Proteins were separated by SDS-PAGE (10% gel), and the gel was silver stained.

To purify the MrsC protein, we constructed a GST-MrsC fusion protein (64). Expressing the protein as a GST fusion improved solubility and allowed us to purify the fusion protein in a single step under nondenaturing conditions by using a glutathione-agarose affinity column (64). The fusion protein was analyzed in both the supernatant and the cell pellet from induced cells. Although the GST-MrsC fusion protein was not completely soluble, about 45% remained in the supernatant fraction. We independently purified the GST-MrsC and GST proteins to greater than 98% homogeneity, which was determined by densitometrically scanning a silver-stained SDS–10% polyacrylamide gel (Fig. 4B; also see Material and Methods).

The ATPase activity of the GST-MrsC protein increased with increasing protein concentrations. The GST control protein, however, showed no activity, even at high protein concentrations (Fig. 5A). A time course experiment showed that for the GST-MrsC protein, the ATPase activity was linear for up to 20 min at 37°C; for the GST protein, there was no ATPase activity (Fig. 5B).

FIG. 5.

ATPase activity of GST-MrsC fusion protein. (A) ATPase activity of GST-MrsC protein at various protein concentrations. Reaction mixtures (75 μl) with different amounts of either fusion protein or GST protein were incubated at 37°C for 10 min. ATPase activity was measured as described in Materials and Methods. (B) Time course of ATP hydrolysis by GST-MrsC fusion protein. Reaction mixtures (525 μl) containing either 350 ng of fusion protein or 700 ng of GST protein and [γ-³²P]ATP at 100 μM were incubated at 37°C. At each time, 75 μl of reaction mixture was taken and ATP hydrolysis was measured for GST-MrsC fusion protein and GST protein. (C) ATP hydrolysis of GST-MrsC fusion protein at various substrate concentrations. Reaction mixtures (75 μl) containing 48 ng of fusion protein and various concentrations of [γ-³²P]ATP were incubated at 37°C for 10 min. ATP hydrolysis was measured as described in Materials and Methods.

The purified GST-MrsC protein had an apparent K_m of 28 μM and a V_max of 21.2 nmol/μg/min (Fig. 5C). The ATPase activity was unaffected by low concentrations (0 to 75 mM) of NaCl but was inhibited by higher concentrations (100 mM or above) (data not shown). The fusion protein required Mg²⁺ for activity, the optimum being 1 mM (data not shown). The ATPase activity was not stimulated by either double-stranded DNA, single-stranded DNA, RNA, or poly(A).

ATPase activity is necessary but not sufficient for the biological function of MrsC.

Since it was clear that MrsC was an Mg²⁺-dependent ATPase, we decided to determine whether the ATP-binding motif of the MrsC protein was important for in vivo activity. Mutations at the most conserved lysine consistently affect ATPase activity (16, 58, 62, 67). Based on this information, we mutated lysine 201 (AAA)—the most conserved amino acid residue within the motif (61)—to glutamic acid (GAA) by using a mismatched primer and PCR (see Materials and Methods). The A-to-G transversion was confirmed by nucleotide sequencing. We cloned the altered mrsC gene, mrsC201, into pWSK29. The resultant plasmid, pWK937 (mrsC201), did not complement the mrsC505 mutation at 44°C, while pWK936 (mrsC⁺) did.

More interestingly, we found that even when a wild-type strain was transformed with pWK937 (mrsC201), cells grew very poorly at 30, 37, or 44°C. This dominant-negative effect of the altered mrsC gene, as also noted by Akiyama et al. (2), suggested that the mutated MrsC protein may form a complex either with itself or with other proteins.

To overexpress the MrsC201 protein, we cloned it into the vector pET22b such that mrsC expression was under T7 promoter control and inducible by addition of IPTG. We detected no increased ATPase activity in induced cells compared to uninduced controls (data not shown). This result indicated that the lysine-to-glutamic acid substitution in the ATP-binding motif had abolished the ATPase activity and caused the concomitant loss of biological activity. These results support the observations of Akiyama et al. (2). However, the presence of ATPase activity alone was not sufficient for biological activity. As shown above, the GST-MrsC fusion protein (N-terminal fusion) exhibited high levels of ATPase activity but did not complement the mrsC505 mutation at 44°C.

Comparison of lysogenization frequencies of hflB29 and mrsC505 mutants.

Since mrsC505 appeared to be an allele of hflB, we constructed a series of isogenic strains carrying the hflB29 (a mutation that does not cause temperature-sensitive growth) and mrsC505 alleles and measured their lysogenization frequencies by using a modification of the procedure of Herman et al. (32). As shown in Table 3, the presence of the hflB29 (SK8945) allele led to significant increases in lysogenization frequency over that of the wild-type control (MG1693) at various temperatures (470-fold at 30°C and 20-fold at 44°C). The mrsC505 strain (SK8232) showed increases in lysogenization frequency at all temperatures compared to the wild-type control (41-fold at 30°C and 15-fold at 44°C [Table 3]), which were qualitatively comparable to the observed 10-fold increase in lysogenization frequency observed with the temperature-sensitive ftsH1 mutation (32).

TABLE 3.

Lysogenization frequencies of λCam105 in various strains

Strain	Lysogenization frequency (%)a
	30°C	37°C	44°C
MG1693	0.17 ± 0.1	0.86 ± 0.3	3.0 ± 0.9
SK8232 (mrsC505)	7.0 ± 0.8	19 ± 3	47 ± 6
SK8945 (hflB29)	80 ± 10	80 ± 5	62 ± 4

^aDefined as (number of Cm^r colonies/number of infective centers) × 100, determined as described in Materials and Methods. 

Effect of the hflB29 allele on mRNA decay.

We also examined the stability of specific mRNAs in hflB29 (SK8945) and mrsC505 (SK8232) single mutants as well as in multiple mutants (SK8236 [mrsC505 pnp-7 rnb-500 rne-1] and SK8948 [hflB29 pnp-7 rnb-500 rne-1]). As shown in Fig. 6, the decay pattern for each of the specific mRNAs (trxA, secG, and lpp) was equally affected by either the hflB29 or mrsC505 allele. While the effect of the mrsC505 allele appeared to be temperature dependent, hflB29 stabilized mRNA decay at both 30 and 44°C (data not shown). The half-lives of the full-length transcripts were determined from the Northern blots and are presented in Table 4. Of particular interest was the observation of differential stabilization among the various mRNAs. The half-life of the secG mRNA increased from 16 min in the wild-type strain to over 30 min in either the hflB29 or mrsC505 single mutant. No further increase was observed in the multiple mutants (SK8236 and SK8948) (Table 4). In contrast, a large increase in half-life was observed for both trxA and lpp in the multiple mutants (6 min versus 16 to 24 min for trxA and 38 min versus 81 to 152 min for lpp) (Table 4).

FIG. 6.

Northern analysis of trxA (A), secG (B), and lpp (C) mRNAs. RNA was isolated from strains SK5704 (pnp-7 rnb-500 rne-1), SK8236 (mrsC505 pnp-7 rnb-500 rne-1), and SK8948 (hflB29 pnp-7 rnb-500 rne-1) as described in Materials and Methods and separated in 6% polyacrylamide gels containing 7 M urea. The number above each lane indicates the time (minutes) after temperature shift when the RNA was extracted. Seven micrograms of RNA was loaded in each lane. Arrowheads on the left margin indicate full-length transcripts.

TABLE 4.

Half-lives of trxA, lpp, and secG mRNAs

Strain	Genotype	Half-life (min) of full-length transcripta
		trxA	lpp	secG
MG1693	Wild type	2.8	NDb	16.0
SK8232	mrsC505	3.6	ND	35.6
SK8945	hflB29	3.3	ND	31.5
SK5704	pnp-7 rnb-500 rne-1	6.0	38	16
SK8236	mrsC505 pnp-7 rnb-500 rne-1	16.4	81	36
SK8948	hflB29 pnp-7 rnb-200 rne-1	24.7	152	32

^aDetermined as described in Materials and Methods. Each result is the average of at least two independent experiments. All determinations were for a single chromosomal copy of the gene. 

^bND, not determined. 

Comparison of the heat shock response in mrsC505 and hflB29 mutants.

Recent studies (33, 72) have demonstrated that the HflB/FtsH protein is directly involved in the cleavage of the ς³² heat shock transcription factor. Accordingly, we examined the heat shock response of the groES gene in both mrsC505 and hflB29 mutants. The data presented in Fig. 7 show that the presence of an mrsC505 mutation significantly altered the heat shock response of the groES gene, while the behavior of the hflB29 allele mimicked that of the wild-type control. This result for hflB29 agrees with the observation of Herman et al. (33). Dot blots of the same RNA probed with lpp were used as controls to ensure equal loading of samples (data not shown).

FIG. 7.

Measurement of heat shock response in mrsC505 and hflB29 mutants. Levels of groES mRNA were measured as described in Materials and Methods. The y axis shows the fold increase over steady-state levels at times after temperature shift to 44°C. ◊, MG1693 (wild type); ○, SK8232 (mrsC505); □, SK8945 (hflB29).

DISCUSSION

In this report, we have physically characterized the mrsC gene and shown that it is the same as hflB/ftsH. While our nucleotide sequence agreed with that of Tomoyasu et al. (74), the presence of a strong ribosome-binding site at nucleotides 539 to 544 (Fig. 3) predicted a UUG start codon (Fig. 3), rather than the AUG that they had suggested from visual inspection of the DNA sequence (74). Our prediction was confirmed when we sequenced the first 18 amino acids of the gel-purified MrsC protein. This result added three amino acids to the amino terminus (Fig. 3).

Only a few other E. coli genes use UUG for initiation. Interestingly, three genes involved in mRNA decay—those encoding RNase D, PNPase, and poly(A) polymerase I—utilize this start codon (12, 56, 80). While the significance of this observation is unclear, one explanation is that the UUG codon reduces MrsC translation. The presence of a UUG codon in rnd and cya (encoding adenylate cyclase) limits their expression (55, 79). That the mrsC gene could not be cloned into a high-copy-number vector (data not shown) indicates that high levels of protein harm the cell. The UUG codon in the mrsC gene may therefore serve to maintain the MrsC/HflB concentration at acceptable levels under normal growth conditions.

A computer search demonstrated that MrsC significantly resembled (35 to 47% identity in a segment of about 220 amino acids) 12 distinct eukaryotic proteins: valosin-containing protein (VCP), P97, CDC48, Sec18, N-ethylmaleimide-sensitive fusion protein, Pas1, TBP-1, MSS1, SUG1, S4, Mts2, and Cim5 (20, 21, 23, 25, 27, 28, 36, 49, 52, 63, 68). Although these proteins are associated with different biological functions, including transcription activation (68), proteolysis (19), and secretion (20, 77), a common feature of all 12 proteins is that they either function as a component of a multiprotein complex or interact with cellular membranes. Since they all have highly conserved ATP-binding domains, ATP hydrolysis must be important for their collective biological functions. However, the absence of sequence homology among these proteins beyond the 220-amino-acid domain containing the ATP-binding motif suggests that the remaining regions of each protein must govern their actual substrate specificity.

We therefore used a two-pronged approach to determine the biochemical features necessary for MrsC function in vivo. We purified a GST-MrsC fusion protein and showed that it encoded an Mg²⁺-dependent ATPase (Fig. 5) with a K_m for ATP of 28 μM. In addition, changing the conserved lysine 201 residue in the ATP-binding domain to glutamic acid (mrsC201) resulted in a protein that was no longer biologically active. These results are in general agreement with the observations of Tomoyasu et al. (72), although they obtained a K_m of 80 μM. The difference in the measured K_m for ATP could arise from the fact that our determination was done with a GST fusion protein. The fact that an amino-terminal GST-MrsC fusion protein was no longer biologically active supports the importance of the membrane attachment observed directly by Tomoyasu et al. (73).

How is the HflB/FtsH/MrsC protein involved in mRNA decay? One possibility is that it is an ATP-dependent RNase. There is one example of such a protein in HeLa cell nuclear extracts (48). Although we did not detect nuclease activity with our purified GST-MrsC fusion protein (data not shown), since it was not active in vivo, it is still remotely possible that it has nucleolytic activity in its native form.

Our finding that the mrsC201 allele exhibited a dominant-negative phenotype, along with comparable observations by Akiyama et al. (2), clearly indicates that in vivo the HflB/FtsH/MrsC protein could act either as a homodimer or as part of a multiprotein complex. Furthermore, there are examples in the yeast Saccharomyces cerevisiae of ATP-binding proteins being involved in both mRNA turnover (38) and tRNA splicing (17). Thus, another possibility is that HflB/MrsC is a subunit of an RNA-processing complex.

One such known complex involves RNase E, an endonuclease, PNPase, an exonuclease, and the RhlB helicase, specifically associating to form an RNA-processing complex (14, 43, 53, 54). This complex might contain additional proteins, such as MrsC, that could serve as a membrane anchor. The assembly of such a complex might be carried out in an ATP-dependent manner. This hypothesis is not particularly attractive since the genetic data presented by Granger et al. (29) strongly suggest that mrsC/hflB and rne are in different pathways. However, it is worth noting that the involvement of MrsC/HflB in mRNA decay is the first indication that the bacterial membrane may be required in this process. It thus appears that cellular localization may also play a role in mRNA decay in E. coli.

In light of the recent observation that the HflB/FtsH/MrsC protein acts as an ATP-dependent protease in the degradation of ς³² (72) and other proteins (1, 35), the hypothesis that we currently favor is that the HflB/MrsC protein activates an RNase or one or more protein subunits of a different mRNA processing complex by proteolysis. It is therefore possible that the effect on mRNA decay observed in mrsC and hflB mutants results from either the failure to activate an RNase through proteolytic processing or the inability to proteolytically degrade a protein that normally stabilizes mRNAs. While there is no direct evidence to support either of these hypotheses, the recent report of an mRNA protection complex containing the GroEL protein (24) provides a mechanism by which the inactivation of the HflB/FtsH/MrsC protein could lead to stabilization of mRNAs. Under normal conditions, one or more of the components of the mRNA protection complex would be proteolytically degraded in an ATP-dependent reaction. Inactivation of the protease would lead to an accumulation of the protection complex and mRNA stabilization.

It is also worth noting that the stabilization of mRNA is not the cause for the temperature-sensitive growth phenotype of mrsC505 mutants (29), since mRNA stability was comparable in mrsC505 and hflB29 mutants (Fig. 6), and hflB29 does not cause conditional lethality. In addition, the observation of the differential effect on the heat shock response by the two alleles (Fig. 7), along with the data of Herman et al. (33), further suggests that if the MrsC/HflB protein’s only activity is as a protease, it is possible to differentially inactivate its ability to degrade its various protein substrates. It is thus possible that MrsC/HflB is a multifunctional protein.