INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Among ATP-dependent proteases of Escherichia coli, FtsH is unique in that it is membrane integrated (1) and essential for cell viability (2, 3). Its amino-terminal region includes two transmembrane segments and the flanked periplasmic domain of 72 amino acids, whereas its central region has sequence homology to the AAA family of ATPases, followed by the carboxyl-terminal region having a Zn²⁺-metalloprotease motif (HEXXH) (1, 4). Although importance of the latter two cytoplasmic regions has been well documented (5, 6), the roles of the transmembrane and the periplasmic regions are only poorly understood.

We identified three membrane-integrated substrates of FtsH: SecY subunit of protein translocase (5, 7), subunit a of the F₁F₀ ATPase (F₀a) (8), and the YccA protein (9). Elimination of SecY and F₀a, when they failed to associate with their respective partner proteins, seems to serve as a mechanism to keep the integrity of the membrane (5, 7, 8, 10). FtsH has roles in regulation of gene expression also because it degrades some cytosolic transcription factors such as CII, which determines the lysis/lysogeny commitment of phage  (11-13), and ³², which controls the heat-shock response (14, 15).

It is also known that some ftsH mutations cause abnormal translocation of the PhoA moiety attached to the carboxyl-terminal cytoplasmic region of SecY (Std phenotype) as well as significant defects in protein export (3, 16). The tolZ allele of ftsH confers resistance to several colicins and defective membrane potential (6, 17). We showed recently that FtsH has a proteolysis-independent polypeptide-binding activity (18). These findings raise a possibility that FtsH acts like a molecular chaperone in protein assembly processes.

FtsH forms homo-oligomeric complexes in which the amino-terminal membrane region interacts (19). It also associates with the HflK·HflC complex (HflKC) (10). Unlike the prevailing view, HflKC does not itself degrade CII and resides mostly on the periplasmic side of the membrane (12). We isolated mutations hflK13 and hflC9, which interfere with the SecY degradation function of FtsH (10). These mutations stabilize SecY and F₀a but not CII (9, 12). We also identified a mutation, yccA11, that stabilizes only the membrane-bound substrates of FtsH (9). The YccA protein is a multi-spanning membrane protein that was shown to be a substrate of FtsH both in vivo and in vitro (9). The YccA11 mutant protein, having an internal deletion of 8 amino acids near the amino terminus, binds to FtsH and HflKC but is refractory to degradation. The inhibitory effect of YccA11 on SecY degradation was only observed in the presence of HflKC. The proposed function of HflKC is to differentially modulate the proteolytic activities of FtsH against different classes of substrates (9).

In this work we carried out mutational analyses of the periplasmic region of FtsH. It was found that mutational disruptions of the periplasmic region resulted in the impairment of the FtsH-FtsH and FtsH-HflKC subunit interactions, affecting the in vivo proteolytic activities of FtsH against the same spectrum of substrates as the hflK13, hflC9, and yccA11 mutations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial Strains and Plasmid-- -E. coli strains and plasmids used in this study are listed in Table I. Strain AR3291 (ftsH3::kan) was a derivative of W3110 bearing a suppressor sfhC21 with a linked transposon zad-220::Tn10.¹ pSTD233 was a derivative of pSTD120 (19) from which the two BamHI sites were eliminated (for the TnphoA/in analysis described below); site-direct mutagenesis (20) using a mutagenic primer, AATTTCGGGTCCTGAAC, eliminated the one within ftsH-his₆-myc without altering the amino acid sequence, whereas the one in the multicloning region was eliminated by digestion with BamHI, filling in with T4 DNA polymerase, and self-ligation. pSTD236 was constructed from pSTD233 using Quick Change mutagenesis kit (Stratagene) and primers CTTGGTTCTTCAGGCGGTTCCTCGAGGCCATTAGACTCGCTGGGCC and GGCCCAGCGAGTCTAATGGCCTCGAGGAACCGCCTGAAGAACCAAG. pSTD243 was constructed by replacing a 1.2-kilobase XbaI-MluI fragment of pSTD113 (19) with the corresponding fragment of pSTD236. pMW119H was constructed from pMW119 by treatment with HindIII and T4 DNA polymerase, followed by self-ligation. pSTD240 was constructed by cloning an EcoRI-HindIII fragment of ptac-cII_Y42 (12) into pSTV28 (Takara Shuzo). For constructing pSTD241, a 3.9-kilobase KpnI-EcoRI fragment containing the hflX-hflK-hflC9 genes was first cloned into pSTV29 (Takara Shuzo); then, the HindIII secY fragment of pKY3 (21) was inserted into the HindIII site on the vector.

Insertion Mutagenesis-- Insertion mutations in the periplasmic domain of FtsH were isolated by the TnphoA/in method (22). TnphoA/in was allowed to transpose onto pSTD233, and periplasmic insertions were converted into those of an in-frame 31-codons by excision of a BamHI fragment (22). Exact insertion points were determined by DNA sequencing (Fig. 4). Plasmids carrying insertion 104 and insertion 105 were designated as pINS104 and pINS105, respectively.

Media-- L medium (23), peptone medium (24), and M9 medium (23) were used. Ampicillin (50 µg/ml) and/or chloramphenicol (20 or 100 µg/ml) were added for growing plasmid-bearing strains.

Pulse-chase and Immunoprecipitation Experiments-- Cells were pulse labeled with [³⁵S]methionine and chased as described previously (7). Whereas CII was directly analyzed by 15% SDS-PAGE,² SecY was immunoprecipitated (25) and separated by 15% acrylamide, 0.12% N,N'-methylene-bis-acrylamide SDS-PAGE (26). Labeled proteins were visualized and quantified by a Fuji BAS2000 imaging analyzer.

LacZ and PhoA Activity Assays-- These enzymes were assayed by the published procedures (27, 28).

Isolation of Protein Complexes Containing a His₆-tagged Protein-- Total membranes from cells expressing FtsH-His₆-Myc or FtsH236-His₆-Myc were solubilized in buffer containing 50 mM Tris-HCl, pH 8.1, 500 mM KCl, 20 mM imidazole, 10% glycerol, 0.5% Nonidet P-40, and 10 mM 2-mercaptoethanol at 0 °C for 1 h. After removal of insoluble materials by centrifugation, proteins were incubated with Ni-NTA agarose (Qiagen), which were washed twice with 10 mM Tris-HCl, pH 8.1, 300 mM KCl, 10% glycerol, 20 mM imidazole and eluted with 250 mM imidazole in the same buffer (10).

Purification of FtsH236-His₆-Myc-- FtsH-His₆-Myc and FtsH236-His₆-Myc were purified from overproducing strains TYE024/pSTD113 and TYE024/pSTD243, respectively, as described previously (5). Samples after Ni-NTA affinity column chromatography was used as final preparations except for the gel filtration assay of FtsH aggregation (see below). ATPase activities were assayed as described previously (5).

In Vitro Degradation of SecY and CII-- In vitro activities of FtsH-His₆-Myc and FtsH236-His₆-Myc to degrade SecY were assayed as described previously (5). Briefly, SecY (40 ng) was incubated at 37 °C with either FtsH-His₆-Myc or FtsH236-His₆-Myc (500 ng) in the presence or absence of ATP (3.3 mM). SecY was visualized by SDS-PAGE and immunoblotting (29). Images were produced using ECL detection kit (Amersham Pharmacia Biotech) on x-ray films. In vitro degradation of partially purified and [³⁵S]methionine-labeled CII was similarly assayed (12) by visualization of CII with a BAS2000 imaging analyzer.

Trypsin Digestion of FtsH236-His₆-Myc-- Purified FtsH-His₆-Myc (13.2 µg protein) was treated with 5 µg/ml trypsin at 0 °C as described previously (18). A portion of the samples was treated with trichloroacetic acid. The protein patterns were then examined by 16.1% acrylamide, 0.12% N,N'-methylene-bis-acrylamide SDS-PAGE (30), followed by immunoblotting using anti-FtsH serum.

Gel Filtration of FtsH-His₆-Myc and FtsH236-His₆-Myc-- FtsH-His₆-Myc or FtsH236-His₆-Myc was dialyzed against buffer containing 10 mM Tris-HCl, pH 8.1, 5 mM MgCl₂, 10% glycerol, and 0.5% Nonidet P-40. 12.5 µg of the protein was incubated in ATPase buffer (50 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 10% glycerol, 10 mM 2-mercaptoethanol, 25 µM zinc acetate, 2.5 mM Tris acetate) containing 0.5% Nonidet P-40 at 0 °C or at 37 °C for 30 min in the presence or the absence of either ATP or adenosine-5'-o-(3-thiotriphosphate) (1 mM). Samples were loaded onto Superose 6 column and eluted with ATPase buffer containing 0.5% Nonidet P-40. 0.7-ml fractions were analyzed by SDS-PAGE and immunoblotting using anti-FtsH serum.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction, Identification, and Growth-supporting Function of an FtsH Derivative Lacking the Periplasmic Domain-- We constructed a derivative of FtsH (FtsH236) in which 62 out of 72 residues in the periplasmic region had been deleted. To facilitate detection and isolation of the protein, a His₆-Myc bipartite tag was attached to its carboxyl terminus, and the gene was placed under the lac promoter control. The resulting mutant protein, FtsH236-His₆-Myc, was detected with anti-FtsH and anti-Myc antibodies as a protein of 64 kDa (data not shown; see Fig. 8 for the purified proteins). 2 h after induction, FtsH236-His₆-Myc amounted to about 1.5-2-fold the FtsH content in the wild-type cells (data not shown). FtsH236-His₆-Myc was fractionated as membrane bound (data not shown). Plasmid pSTD236 carrying ftsH236-his₆-myc did not complement the growth defect of the ftsH1 (Ts) or the zgj-525::IS1A (Cs) mutants.

In Vivo Activity of FtsH236-His₆-Myc to Degrade SecY-- SecY is rapidly degraded by FtsH when it has failed to form a complex with SecE (7, 25). The LacZ fragment can be used as a reporter of the stability of the amino-terminally attached SecY protein (7, 31). -Galactosidase activity of SecY-LacZ (in the presence of an  donor (LacZM15)) was increased more than 2-fold by the zgj-525::IS1A mutation (Fig. 1A, columns 1 and 2) at 37 °C, whereas co-expression of ftsH-his₆-myc (column 3) lowered the -galactosidase activity almost to the level in the wild-type cells. It was found that ftsH236-his₆-myc (column 4) exerted a similar effect. Stability of SecY-LacZ was directly examined by pulse-chase experiments (Fig. 1B). Consistent with the above results, degradation of SecY-LacZ in the zgj-525::IS1A mutant cells was accelerated by the co-expression of FtsH236-His₆-Myc to an extent similar to the degradation accelerated by FtsH-His₆-Myc. These results indicate that FtsH236-His₆-Myc retains almost full SecY-degrading activity.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   In vivo degradation of SecY by FtsH236-His6-Myc. A, strain AK519 (ftsH+) was transformed with pKY258 (secY-lacZ) and pMW119H (vector, lane 1). AK525 (zgj-525::IS1A) were transformed with pKY258 and pMW119H (lane 2), pSTD233 (ftsH+-his6-myc, lane 3), pSTD236 (ftsH236-his6-myc, lane 4), pINS104 (ftsH#104-his6-myc, lane 5), or pINS105 (ftsH#105-his6-myc, lane 6). Cells were grown in peptone medium and induced with 1 mM IPTG for 1 h. -Galactosidase (LacZ) activities are presented as relative activities to that of AK519/pKY258/pMW119H. Vec, WT, +, and Cs indicate a vector plasmid, the wild-type ftsH-his6-myc gene on the vector, the wild-type ftsH gene on the chromosome, and the zgj-525::IS1A mutation on the chromosome, respectively. B, cells of AK525 carrying pKY258 and pSTD236 (squares), pSTD233 (diamonds), or pMW119H (circles) were grown in M9 medium, induced with 1 mM IPTG and 5 mM cAMP for 1 h, pulse labeled with [35S]methionine for 30 s, and chased for the indicated periods. Proteins of a fixed total radioactivity were precipitated with trichloroacetic acid, and SecY-LacZ was immunoprecipitated. Radioactivities associated with SecY-LacZ were determined after SDS-PAGE. Values are reported as % of the initial (0-min chase) radioactivity for each culture.

In Vivo Activity of FtsH236-His₆-Myc to Degrade the CII Protein-- CII is a major determinant of the lysis/lysogeny pathways of phage , and its stabilization leads to preferential lysogenization (32, 33). We used an ftsH disrupted strain (AR3291) bearing pSTD233 (ftsH-his₆-myc), pSTD236 (ftsH236-his₆-myc), or pMW119H (vector) as hosts for  phage infection (Table II). Although ftsH is essential for bacterial growth, a suppressor mutation, sfhC, in AR3291 enables cell growth in the absence of FtsH (6).¹ The strain AR3291 itself showed a lysogenization frequency of almost 100%. When FtsH-His₆-Myc was synthesized from pSTD233 in the same strain, lysogenization frequency was decreased to 4.3%. In contrast, the expression of ftsH236-his₆-myc from pSTD236 affected the lysogenization frequency only to a small extent (to about 80%). Consistent results were obtained by scoring sensitivity of these bacteria to c17, which cannot propagate in cells of increased CII contents (33). Whereas AR3291/ftsH-his₆-myc was sensitive to c17, AR3291/ftsH236-his₆-myc allowed far less plaque formation. These results, taken together, suggest that FtsH236-His₆-Myc has greatly reduced ability to degrade the CII protein.

View this table: [in this window] [in a new window]   Table II Effects of the ftsH mutations on lysogenization of

ftsH236-his

View larger version (16K): [in this window] [in a new window]   Fig. 2.   In vivo degradation of the CII protein by FtsH236-His6-Mycs. Cells of AK525 carrying pKY240 (cII) and pSTD236 (squares), pSTD233 (diamonds), or pMW119H (circles) were grown in M9 medium, induced with 1 mM IPTG and 5 mM cAMP for 1 h, and pulse labeled with [35S]methionine for 30 s, followed by chase for the indicated periods. Proteins of a fixed total radioactivity were subjected to SDS-PAGE. Radioactivities associated with CII were determined, and values are reported as % of the initial (0-min chase) radioactivity for each culture.

Effects of FtsH236-His₆-Myc on the Std Phenotype-- Impairment of the FtsH functions causes abnormal periplasmic localization of the PhoA moiety attached to the carboxyl-terminal cytoplasmic region of SecY. This phenotype, termed STD, can be assessed by measuring alkaline phosphatase activity as well as examining trypsin sensitivity of the PhoA moiety (3). The zgj-525::IS1A mutant expressing SecY-PhoA C6 fusion had the PhoA activity that was at least 2.5-fold higher than that in the wild-type strain (Fig. 3A, columns 1 and 2). Expression of either FtsH-His₆-Myc (column 3) or FtsH236-His₆-Myc (column 4) lowered the PhoA activity to the wild-type level. Trypsin treatment of the extract from the zgj-525::IS1A mutant yielded a significant proportion of the PhoA-sized fragment (Fig. 3B, lanes 5 and 6). The trypsin-resistant proportion was greatly reduced when the mutant cells contained either FtsH-His₆-Myc (lanes 3 and 4) or FtsH236-His₆-Myc (lanes 1 and 2). These results show that FtsH236-His₆-Myc retains an ability to prevent translocation of the PhoA moiety (18) of the SecY-PhoA C6 fusion protein.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Activity of FtsH236-His6-Myc to complement the Std phenotype. A, strain AK519 was transformed with pKY321 (secY-phoAc6) and pMW119H (lane 1). AK525 was transformed with pKY321 and pMW119H (lane 2), pSTD233 (lane 3), pSTD236 (lane 4), pINS104 (lane 5), or pINS105 (lane 6). Cells were grown in peptone medium and induced with 1 mM IPTG for 1 h. PhoA activities relative to that of AK519/pKY321/pMW119H were depicted. B, cells of AK525 carrying pKY321 and pSTD236 (lanes 1 and 2), pSTD233 (lanes 3 and 4) or pMW119H (lanes 5 and 6) were grown in peptone medium, induced with 1 mM IPTG for 2 h. Cells were disrupted by lysozyme-freeing-thawing and treated with or without trypsin (50 µg/ml) as indicated. Proteins were separated by SDS-PAGE and visualized with immunoblotting using anti-PhoA (5 prime  3 prime Inc.). PhoA* indicates the trypsin-resistant PhoA fragment.

Isolation and Characterization of Insertion Mutations Affecting the Periplasmic Region of FtsH-- We isolated a series of 31 amino acid-insertion mutations in the periplasmic region of FtsH-His₆-Myc using the TnphoA/in method (22). We identified four different insertions within the periplasmic region from a total of 10 independent isolates (Fig. 4). They were classified into two groups: class I complementing the ftsH1 mutation and class II without complementing activity. It was found that class I mutations had occurred after the 26th or 34th codon, whereas the class II mutations had occurred after the 42nd or 60th codon (Fig. 4). All the insertion derivatives were identified by immunoblotting experiments using anti-FtsH or anti-Myc antibodies as proteins of slightly slower electrophoretic mobilities than FtsH-His₆-Myc (data not shown). We chose 105 (class I) and 104 (class II) for further analyses.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Deletion and insertion mutations of ftsH constructed in this study. The 236 mutation is indicated by a horizontal and bidirectional arrow. Insertion mutations are indicated by vertical arrowheads. Insertions indicated by a filled arrowhead complemented ftsH1, but those indicated by an open arrowhead did not. TM1 and TM2 indicate the first and the second transmembrane segments of FtsH. The amino acid residue numbers from the amino terminus methionine (34) are indicated on the left.

Self-interaction of FtsH Is Impaired by the Periplasmic Deletion-- FtsH forms a homo-oligomeric complex, and this FtsH-FtsH interaction is mediated by the amino-terminal region (16, 19). We examined interaction between FtsH236-His₆-Myc and FtsH. To minimize possible complication arising from the FtsH-HflKC interaction, we used the hflKC-deleted strain AK990 in which either FtsH-His₆-Myc or FtsH236-His₆-Myc had been expressed. Nonidet P-40-solubilized membrane proteins were subjected to Ni-NTA affinity chromatography. The sample from the FtsH-His₆-Myc expressing cells gave two anti-FtsH reacting protein bands in the imidazole-eluate (Fig. 5A). As reported previously (10), the faster migrating band contained the chromosomally encoded FtsH as well as a carboxyl-terminally truncated form of FtsH-His₆-Myc (FtsH'), both of which were brought down because of their interaction with the tagged derivative. In contrast, the chromosomally encoded FtsH was not detected in the eluate fraction when the sample from the cells expressing FtsH236-His₆-Myc was used, although FtsH236-His₆-Myc itself was normally recovered in this fraction (Fig. 5B). These results indicate that the deletion of the periplasmic region impairs the self-interacting property of FtsH.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Impaired FtsH-FtsH interaction by the periplasmic deletion. Membranes were prepared from cells of AK990 (hflK-hflC::kan) carrying pSTD233 (A) or pSTD236 (B), solubilized with 0.5% Nonidet P-40, and subjected to Ni-NTA column chromatography. Columns were washed two times and were eluted with 250 mM imidazole. Proteins in load (lane 1), flow through (lane 2), first wash (lane 3), second wash (lane 4), and eluate (lane 5) fractions were separated by SDS-PAGE and visualized by immunoblotting using anti-FtsH antiserum. FtsH' indicates a carboxyl-terminally cleaved product of FtsH-His6-Myc (19).

FtsH-HflKC Interaction Is Impaired by the Periplasmic Deletion-- FtsH forms a complex with HflKC (10), which can also be isolated using His₆-tagged FtsH (Fig. 6A). The ftsH strain was used as a host to avoid any contribution from the chromosomally encoded FtsH. FtsH236-His₆-Myc failed to bring down HflKC (Fig. 6B). Thus, the periplasmic domain of FtsH seems to be important for the FtsH-HflKC association.

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Impaired FtsH-HflKC interaction by the periplasmic deletion. Membranes were prepared from cells of AR3291 (ftsH3::kan) carrying pSTD233 (A) or pSTD236 (B), solubilized with 0.5% Nonidet P-40, and subjected to Ni-NTA column chromatography. Columns were washed two times, and eluted with 250 mM imidazole. Proteins in load (lane 1), flow through (lane 2), first wash (lane 3), second wash (lane 4), and eluate (lane 5) fractions were separated by SDS-PAGE and visualized by immunoblotting using anti-FtsH antiserum (upper panels) or anti-HflKC (lower panels).

ftsH236-his

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Suppression of the HflC9 phenotype by the FtsH236-His6-Myc mutation. Cells of AK990 (hflKC::kan) carrying pKY241 (secY, hflK-hflC9) and pSTD236 (squares), pSTD233 (diamonds), or pMW119H (circles) were grown in M9 medium and induced with 1 mM IPTG and 5 mM cAMP for 1 h. They were pulse labeled with [35S]methionine for 30 s and chased for the indicated periods. Proteins of a fixed total radioactivity were subjected to SDS-PAGE. Radioactivities associated with SecY were determined, and values are reported as % of the initial (0-min chase) radioactivity for each culture.

In Vitro Activities of FtsH236-His₆-Myc-- We purified FtsH236-His₆-Myc (Fig. 8A). It was found that FtsH236-His₆-Myc (0.71 µmol P_i released/mg protein/h) was only about as active as the wild-type protein (4.1) in ATP hydrolysis.

View larger version (44K): [in this window] [in a new window]   Fig. 8.   In vitro activities of FtsH236-His6-Myc to degrade SecY and CII. A, purified preparations of FtsH-His6-Myc (0.475 µg; lane 2) and FtsH236-His6-Myc (0.55 µg; lane 3) were visualized by SDS-PAGE and staining with Coomassie Brilliant Blue. Lane 1 was for molecular weight standards. Note that the faster migrating protein in lane 2 was a carboxyl-terminally truncated form of FtsH-His6-Myc (FtsH') that had lost the bipartite tag (19). B, a purified preparation of SecY (40 ng) was incubated at 37 °C with FtsH-His6-Myc (500 ng; lanes 1-10) or FtsH236-His6-Myc (500 ng; lanes 11-20) in the presence (lanes 6-10 and 16-20) or absence (lanes 1-5 and 11-15) of 3.3 mM ATP. Portions of the samples were withdrawn at 0 (lanes 1, 6, 11, and 16), 7.5 (lanes 2, 7, 12, and 17), 15 (lanes 3, 8, 13, and 18), 30 (lanes 4, 9, 14, and 19), and 60 (lanes 5, 10, 15, and 20) min. Proteins were separated by SDS-PAGE and visualized by immunoblotting using anti-SecY serum. C, partially purified and [35S]methionine-labeled CII was incubated at 37 °C with FtsH-His6-Myc (500 ng; lanes 1-8) or FtsH236-His6-Myc (500 ng; lanes 9-16) in the presence (lanes 5-8 and 13-16) or absence (lanes 1-4 and 9-12) of 3.3 mM ATP. Portions of the samples were withdrawn at 0 (lanes 1, 5, 9, and 13), 0.5 (lanes 2, 6, 10, and 14), 1 (lanes 3, 7, 11, and 15) and 2 (lanes 4, 8, 12, and 16) h. Proteins were separated by SDS-PAGE and visualized by BAS2000 imaging analyzer. SecY' indicates an amino-terminal fragment of SecY present in the purified sample (5).

View larger version (80K): [in this window] [in a new window]   Fig. 9.   Trypsin digestion patterns of FtsH236-His6-Myc in the presence or absence of ATP. FtsH236-His6-Myc (13.2 µg) was incubated with 5 µg/ml trypsin at 0 °C for 0 (lanes 1 and 6), 2 (lanes 2 and 7), 4 (lanes 3 and 8), 8 (lanes 4 and 9), and 16 (lanes 5 and 10) min, in the presence (lanes 6-10) or absence (lanes 1-5) of 1 mM ATP. Proteins were separated by 16.1% SDS-PAGE and visualized by immunoblotting using ant-FtsH. The arrowhead indicates a trypsin-resistant 33-kDa fragment of FtsH.

View larger version (60K): [in this window] [in a new window]   Fig. 10.   Gel filtration profiles of FtsH-His6-Myc and FtsH236-His6-Myc. FtsH-His6-Myc (A-D) or FtsH236-His6-Myc (E-H) was incubated at 0 °C (A, B, E, and F) or at 37 °C (C, D, G, and H), in the presence or the absence of 1 mM ATP as indicated, and fractionated by Superose 6 gel filtration chromatography. A portion of each fraction was subjected to SDS-PAGE followed by immunoblotting using anti-FtsH. The molecular mass markers (indicated above panel A) used were thyroglobulin (670 kDa), bovine -globulin (158 kDa), chicken ovalbumin (44 kDa), and equine myoglobin (17 kDa). WT, wild type.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We have shown here that the periplasmic domain of FtsH is important for its function. The growth-supporting function of FtsH requires the intact periplasmic domain. The periplasmic region of FtsH is not essential for the in vivo proteolytic activity of FtsH against SecY, but it is required for in vivo degradation of CII. It was shown that FtsH-HflKC interaction is disrupted by the lack of the periplasmic domain. Although the homo-oligomeric interaction of FtsH is also disrupted, the substrate-specific inactivation of FtsH236-His₆-Myc may primarily be ascribed to the lack of interaction with HflKC. This is because the hflKC mutant, in which the homo-oligomeric interaction of FtsH is maintained,³ exhibits similar deviation in substrate preference of FtsH.

In apparent contradiction to the in vivo results, FtsH236-His₆-Myc in detergent extracts can degrade both SecY and CII. Thus, FtsH236-His₆-Myc retains intrinsic proteolytic activity even against CII. We suspect that the lack of interaction with HflKC may be responsible for the in vivo inability of FtsH236-His₆-Myc to degrade CII. In contrast, in vitro conditions in which any topological segregation has been compromised seem to allow the mutant protein to act against CII. We proposed the following mechanism about the hflKC effect in vivo (12). HflKC normally acts from the periplasmic side as a negative modulator of proteolysis of membrane-bound substrates. In its absence, FtsH is directed more to the membrane-bound substrates, and soluble protein substrates might escape from FtsH. This "balance shift" model may also explain the phenotypes of FtsH236-His₆-Myc, which cannot interact with HflKC. Other explanations may also be possible. For instance, anchoring of FtsH to the membrane may restrict accessibility of substrates to FtsH from the cytoplasmic side. In this case, HflKC will somehow make FtsH overcome such a restriction.

HflK and HflC have large periplasmic domains of about 35 kDa (12), whereas the cytoplasmic tail of HflK is only 79 residues long, and HflC has essentially no cytoplasmically exposed region (12). Deletion of the cytoplasmic region of HflK does not abolish the FtsH-HflKC interaction.³ Thus, the cytoplasmic domains of FtsH and HflKC do not significantly contribute to their interaction. It is interesting to know whether the transmembrane segments of FtsH and HflKC interact. Our results indicate that transmembrane interaction alone, if any, is not sufficient.

Although FtsH236-His₆-Myc was only about 20% active in ATP hydrolysis, it exhibited almost full proteolytic activities in vitro. The ATP-induced conformational change is unaffected by the 236 mutation (Fig. 9). The results of Ni-NTA affinity isolation and gel filtration experiments suggest that FtsH236-His₆-Myc in isolation in the absence of ATP exists as a monomer. However, ATP induces some oligomerization of FtsH236-His₆-Myc at 37 °C. Thus, although the periplasmic region of FtsH is important for the stable FtsH-FtsH interaction, some other parts in FtsH can undergo ATP-dependent self-association. It is not known whether FtsH can have proteolytic activity in a monomeric state. Unlike the wild-type FtsH that makes extensive aggregates upon incubation at 37 °C in the absence of ATP (Fig. 10) (18), FtsH236-His₆-Myc did not form such aggregates. Thus, in the wild-type protein, ATP may be partly utilized to prevent the formation of the aggregates, but this ATP requirement is alleviated for FtsH236-His₆-Myc, providing a possible explanation for its low ATPase activity accompanied by high protease activity. It should be noted that FtsH236-His₆-Myc still requires ATP hydrolysis to catalyze proteolysis. Because neither binding nor hydrolysis of ATP is required for the substrate binding of FtsH (18), the role of ATP hydrolysis will be in a post-binding processes, in which substrate proteins are presented to the proteolytic active site.