. 2 6 8 , No. 2, Issue of’ Januau 15,pp.,867-872,1993 Prmted In U S .A.

CHEMISTRY THEJOURNAL OF BIOLOGICAL 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc

The y Subunit of the Escherichia coli ATP Synthase MUTATIONSIN THE CARBOXYL-TERMINALREGIONRESTOREENERGYCOUPLINGTO AMINO-TERMINALMUTANT yMet-23+Lys*

THE

(Received for publication, August 24, 1992)

Robert K. NakamotoS, Masatomo Maeda, and MasamitsuFutai From the Department of Organic Chemistry and Biochemistry, The Institute of Scientific and Industrial Research, Osaka University, Ibaraki, Osaka 567, Japan

The y subunit mutations, yMet-23+Lys or Arg, in the Escherichiacoli ATP synthase were previously reported to cause dramatically inefficient energy coupling between ATPase catalysis and H+ translocation (Shin, K., Nakamoto, R. K., Maeda, M., and Futai, M. (1992) J. Biol. Chem. 267, 20835-20839). In this paper, we report that second-site mutations in the y subunit can suppress the effects of yMet-23-Lys. By screening randomly mutagenized uncG (yMet-23+ Lys), eight mutations in the carboxyl-terminal region were identified; strains carrying yArg-242+Cys, yGln-269+Arg, yAla-270+Val, yIle-272+Thr, yThr273+Ser, yGlu-278+Gly, yIle-279-Thr, or yVal280-Ala in combination with yMet-23+Lys were able to grow by oxidative phosphorylation. H+ pumping assayed in membranes prepared from double mutation strains demonstrated that efficient ATP-dependent H+ transport was restored. Interestingly, the single mutations, yGln-269+Arg or yThr-273+Ser, caused reduced growth by oxidative phosphorylation; however, when these mutations were in combination with yMet23+Lys, growth was substantially increased. Furthermore, strains carryingyMet-23+Lys, yGln-269-Arg, or yThr-27343er as single mutations were temperature sensitive, whereas, strains with the double mutations, yMet-23-LyslyGln-269-Arg or yMet-23-Lysl yThr-273-Ser,were thermally stable. Taken together, these results strongly suggest that yMet-23, yArg-242, and the region between yGln-269 to yVal280 are close to each other and interact to mediate efficient energycoupling.

reconstitution of the minimal complex capable of ATP hydrolysis requires the y subunit, in addition to thenucleotidebinding a and @ subunits (8,9). McCarty and co-workers (1013) used chemical modification and protease sensitivity exsubunit of the chloroplast periments to demonstrate that ythe enzyme undergoes conformation changes relevant to catalysis and regulation. They also observed that chemical cross-linking of 2 cysteine residues within the y subunit caused proton leakiness anduncoupling. Information from sequence analysis and site-directed mutagenesis can pinpoint the regions and specific amino acids which are important for the functionssuggested by the above experiments. Alignment of all theknown y subunit sequences show that both terminalregions are highly conserved (amino acids 3-43,83-89, and 237-283, E. coli numbering, asin Refs. 14 and 15), and deletions from these parts of the subunit result in failure of the entire complex to assemble (16, 17). Site-directed mutations of conserved amino acids in the carboxyl-terminal region caused lower ATPase activities indicating that this part of the subunit is important for catalysis (17). Some of these mutations affectedenergycoupling as well. We recently reported that the amino-terminal region mua perturbation tations, yMet-23+Arg or Lys, caused profound in energy coupling (15). The mutationscaused very low levels of H’ pumping even though the enzymes had substantial levels of ATPase activity. The FOsector had normal properties, and attachmentof F, to the membranewas not affected. Clearly, the yMet-23jArg or Lys mutations caused greatly decreased efficiency of energy coupling between ATPase catalysis and proton transport. By locatinga mutation whichcausesa major defectin The Escherichia coli ATP synthase (FoF1-ATPase) can be energy coupling,we now can use the genetic method of searchseparated into two portions with well-defined functions (re- ing for suppressor mutations asa means toidentify neighborviewed in Refs. 1-5). The membrane extrinsic F1 sector con- ing amino acids. Identification of second-siteamino acid sists of five differentsubunitswiththestoichiometry, changes able to compensate for the effect of the yMet-234 a&y16,tl and contains the ATPase catalytic sites. The mem-Arg or Lysmutations would, in turn, implicate themin brane intrinsicFo domain consistsof three subunits(albsc6-lo) coupling as well. Toward this end, we screened for secondwhich mediate protonconductivity. The y subunit (286 amino site mutations in the y subunit which would confer oxidative acids) occupiesa central position in the enzyme complex. phosphorylation-dependent growth tothestraincarrying Structurally, image analyses of electron micrographs of neg- yMet-23-Lys. As a result, several suppressor mutationswere atively stained or frozen hydrated F1 suggest that the y isolated and mapped to the carboxyl-terminal region of the y subunit is positioned in themiddle of the hexameric arrange- subunit. They restored efficient ATP-dependent H’ transport ment of alternating a and p subunits (6, 7). Functionally, and energy coupling. * This research was supported in part by grants from the Ministry EXPERIMENTALPROCEDURES

of Education, Science and Culture of Japan and theHuman Frontier Science Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Supported by the Japan Society for the Promotion of Science.

Materials-Oligonucleotides were synthesized with a Pharmacia

LKB (Uppsala) Gene Assembler Plus. [ c x - ~ * P ] ~ C(3000 T P Ci/mmol) was from Amersham-Japan (Tokyo). Restriction endonucleases and other DNA modifying enzymes were from Takara Shuzo Co., Kyoto, Nippon Gene Co., Toyama, Toyobo Co., Osaka, or New England

867

868

Suppression of a Coupling Mutation in ATP the

A

B ,2 4 0 .,EQA

.. . . 250

(Io

UVIYANKhaT

. . ...

260

DNGGSLIKEL

.(I .

2702 0 6

QLVYNMRQA

+

200

SITQEL'IEIV

f

f

PSI1

sac1

f. SGPIULV

T

Ram

FIG. 1. Random PCR mutagenesis of uncC(yMet-23-rLys). A , random mutagenesis was performed as described under "Experimental Procedures" using a modified PCR amplification. The PCR products were digested with the indicated restriction endonucleases and ligated into pBMG"23K. B, the carboxyl-terminal amino acid sequence of the y subunit. Residues conserved in all the known sequences (Refs. 14,15) aremarked by the asterisk. The arrows show positions of endonuclease cleavage sites used to reclone the mutations. Mutations isolated on the restriction fragments listed in Table I were ligated into pBMG"23K (yMet-23"tLys uncG) or pBWG15 (wildtype umG) using the restriction sites indicated in A .

Synthase

3.0 min at 72 "C inaPerkin-Elmer Cetus Thermocycler. Subsequently, the deficient deoxynucleotide was added to 50 p ~ and , an additional 23 cycles were performed. The products were combined and treated with SnaBI and BglrI restriction endonucleases which gave a fragment encoding the entire y subunit (including yLys-23) except the first seven amino acids (Fig. L4)or with XbaI and BglII which gave a fragment coding between residue 43 and the carboxyl terminus. The fragments were ligated into appropriately digested and alkaline phosphatase-treatedpBMG"23K. Isolation of Pseudo Revertants of uncC(yMet-23+Lys)-We previously established that strain KFlOrA (yGln-14-wmd) harboring pBMG"23K could not grow at 37 "C on succinate agar plates (15). Strain KFlOrA was transformed with the plasmids containing the mutagenized fragments and spread on succinate plates. After 5 days at 37 "C, colonies were selected and restreaked on a fresh succinate plate and on an L broth plate containing10 pg/ml tetracycline. From a total of approximately 2,000 transformants, 12 strains were able to grow on both plates. Plasmids were recovered from the strains, used to transform KFlOrA for retesting on succinate plates, and prepared for double-stranded DNA sequencing (20). Subsequently, random mutations inthe amino-terminal half of the y subunit were again screened by ligating the SnaBI toXbaI (coding amino acids 8 through 40) or XbaI to RsrII (amino acids 43-149) restriction fragments into pBMG"23K. Recloning of Suppressor Mutants-Each mutation was recloned to assure that italone was responsible for the suppression of yMet-23Lys and toeliminate the possibility of spurious mutations in regions of the plasmids not sequenced. The putative suppressor mutations were isolated on appropriate restriction fragments and ligated into pBMG"23K or pBWG15 digested with the same endonucleases (Fig. 1, Table I). Convenient restriction sites hadalready been introduced in the region coding the carboxyl terminus (17) which allowed facile separation of multiple mutations. Strain KFlOrA was transformed with the resulting plasmids and colonies were tested for growth on succinate plates and in most cases, liquid succinate medium as well. The presence of each mutation and yMet-23+Lys was confirmed by DNA sequencing. Cloning procedures were carried out by standard methods (21). Biochemical Procedures-Membrane vesicles were prepared after passing logarithmic phase cells through a French press (22). Forma' gradient wasfollowedby acridine tion of an electrochemical H orange fluorescence quenching (23). ATPase activity (22) and protein (24) were assayed by published procedures. RESULTS

Mutations Suppressing yMet-23+Lys Map to the Carboxyl Terminus-Twelve plasmid-borne, randomly mutagenized Biolabs, Beverly, MA. Taq polymerase and deoxynucleotides were uncG(yMet-23+Lys) genes were able to confer oxidative phosphorylation-dependent growth to strain KFlOrA. DNA from Perkin-Elmer Cetus. Bacterial Strain and Growth Conditions-The y subunit-deficient sequencing revealed that all 12 had mutations resulting in E. coli strain, KFlOrA (thi, thy, recAI, uncGl0 (yGln-l4+end)), was amino acid changes between yArg-242 and yAla-284 (Table described previously and grown as before (17, 18).Minimal medium I). Two of the plasmids were found in duplicate. Two more supplemented with thymine (50 pglml), thiamine (2 pglml), and a carbon source (either 0.2% glucose or 0.4% sodium succinate) or a had multiple changes which were overlapped by others. In rich medium (L broth) were used for genetic analysis. Minimal one case, a plasmid had changes yLeu-262+Ser/rVal-280medium with 0.2% glucose was used for preparation of membranes. Ala, and yet yVal-28bAla alone (pBMG-M23K/V280A, Unless otherwise indicated, all strains were grown at 37 "C. Table I) suppressed yMet-23+Lys. Therefore, yLeu-262Random Mutagenesis of uncG-All genetic manipulations were Ser was not studied further. The same held for the plasmid done on plasmid-borne uncG containing the yMet-23+Lys mutation, containing yArg-242-+Cys/yAla-248-+Val/ySer-271+Leu pBMG-M23K, or the wild type pBWG15 (15). In order to find secondsite mutationswithin the uncG gene which could suppress the growth because the single change yArg-242+Cys (pBMG-M23K/ defect caused by yMet-BBjLys, a modifiedpolymerase chain reaction R242C) was adequate for suppression. (PCR,' Ref. 19) was employed to generate random mutations. 1 ng of From the remaining plasmids, the carboxyl-terminal mupBMG"23K (Fig. L4)was suspended with 10 pmol of the two tations were recloned into pBMG"23K to test theability of and each mutation to suppress yMet-23jLys. TWO primers. Primer 1 (5'-AGTCTCTGGTTCTGTTCGCAGCA-3') plasmids had Primer 2 (5'-CTCAGTGGAACGAAAACTCACG-3') correspond, redouble mutations within the carboxyl-terminal region (rThrspectively, to sequences 240 base pairs upstream and 55 base pairs andyMet 243+Val/yGlu-278-+ downstream of the uncG coding frame. These DNA were mixed with 273+Ser/yAla-284+Val, 1.2 units of Taq polymerase in a buffer of 10 mM Tris-HC1, 50 mM Gly). The pairs of mutations were recloned together and KC1 and 2 mM MgC12,pH 8.3. Finally, 50 p~ of three deoxynucleotides separatelyin pBMG"23K (TableI). The yThr-273-Ser and 50 nMof the fourth were added to four reactions, each one mutation conferred the same growth yield in pBMG"23K deficient in a different nucleotide. The mixtures were passed through The yAla-28hVall three amplification cycles of 1.0 rnin at 94 "C, 2.0 min at 55 "C, and as yThr-273-tSer/yAla-284+Val. yMet-23+Lys mutant did not grow on succinate plates. Likewith or without yMet 243-Val sup'The abbreviations used are:PCR, polymerase chain reaction; wise,yGlu-27&Gly pressed yMet-BS+Lys, and yMet 2434Valalone did not. Tricine, N-[2-hydroxyl-l,l-bis(hydroxymethyl)ethylJglycine.

869

Suppression of a Coupling Mutation inthe ATP Synthase

TABLEI Recloning of mutations found in the carboxyl-terminal region of the y subunit and testing for suppression of yMet-234Lys Plasmids are constructionsof restriction fragments carrying the indicated mutation(s)ligated into pBMG"23K. The multiple mutations Thr-273-rSer/Ala-284+Val or Met-243~Val/Gluu27bGlywere found together on the original plasmid isolates, and recloned together or separately into pBMG"23K. Growth a t 37 "C was judged on succinate plates (solid) after 3 days or in succinate medium (liquid) in which case the saturation density measured at 650 nm is reported as percent of that obtained from wild type (KFlOrA/pBWG15, ODsso = 1.4). Growth of KFlOrA/pBMG-M23K is shown for comparison. Underlined amino acids are completely conserved in all the known y subunit seauences (as in Refs. 14. 15).

%

pBMG"23K pBMG-M23K/T273S/A284V pBMG-M23K/T273S pBMG-M23K/A284V pBMG-M23K/M243V/E278G pBMG-M23K/M243V pBMG-M23K/E278G pBMG-M23K/R242C pBMG-M23K/Q269R pBMG-M23K/A270V pBMG-M23K/I272T pBMG-M23K/I279T pBMG-M23K/V280A ND. not done.

-

None e-273-rSer Ala-284-rVal B-273-Ser Ala-284dVal

ACT PstI-BglII + TCT GCG + GTG ACT + TCT GCG + GTG

Met-243-Val Glu-27kGly Met-243-Val Glu-27bGly

ATG + GTG RsrII-BgiII GAG + GGG ATG -+ GTG RsrII-PstI GAG -+ GGG PstI-BgllI

b-242"tCys Gln-269-Arg CAG Ala-270+Val Ile-272-Thr Ile-2794Thr Val-28bAla GTC

CGT -+ TGT RsrII-PstI + CGG GCA + GTA ATT + ACT ATC + ACC + GCC

PstI-Sac1 Sad-BgZII

PstI-BglII PstI-BglII PstI-BglII PstI-BglII PstI-BglII

5 81

+ +

-

80 ND"

+

80

-

+

ND 69

+ + + + + +

85 79 44 82 69 51

In total, 10 different second-site mutations were tested and A ATP CCCP eight of them were able to suppress theeffects of yMet-231 1 Lys (Table I). Interestingly, seven were found in the highly conserved region between yGln-269and yVal-280, and of these, five mutations replacedresidues identical in all the known y subunit sequences. The eighth,yArg-242+Cys, also changed a conserved residue. Because the amino-terminalhalf of the y subunit also has + Lys-23 conservative stretches (14, 151, we thought that second-site mutations in this portion might suppress effects the of yMet23+Lys, but were missed in the initialscreening. The search for suppressormutations was concentratedintheaminoterminal half by screening plasmidsmutagenized in theregion coding for amino acids 7-149. Fromapproximately 10,000 B ATP CCCP .L 1 transformants, the only plasmid-borne mutation that conferred growth on succinate to KFlOrA was a yLys-23-Thr change (codon AAA changedto ACA). Findingthatthis change was functional reinforced ourhypothesisthatthe positive charge of lysine or arginine at position 23 caused the inefficient coupling (15). Suppressor Mutations Restore Oxidative PhosphorylutiondependentGrowth and Efficient Energy Coupling-Growth dependent upon oxidative phosphorylation was substantially increased in strains carrying the double mutations (yMetFIG. 2. ATP-dependent formation ofan electrochemical 23+Lys with yArg-242-Cys, yGln-269+Arg, yAla-270", of protons in membrane vesicles from strain KFlOrA Val, yIle-272+Thr, yThr-273+Ser, yGlu-278+Gly, yIle- gradient (yGln-1 4 e n d ) harboring pBMG-MZ3K carrying the sup279"tThr, or yVaI-280-Ala) over that obtained from the pressor mutations. ATP-dependent fluorescence quenching in single mutation yMet-23-L~~ (TableI). membranes from the indicated double mutations, yMet-23-Lys The increased growth yields were accompanied by restora- alone (Lys-231,or wild type ( W T ) .100 pg of membrane vesicle protein tion of efficient energy coupling.As reported previously, mem- were suspended in 1.0 ml of 10 mM Tricine-choline, 140 mM KCl, 5 MgCL, 1 pg/ml valinomycin, and 1 p~ acridine orange, pH 8.0. branes from the yMet-23-L~~ mutant strain had a very low mM Fluorescence at 530 nm (excitation at 490 nm) was monitored at degree of ATP-dependent formation of a n electrochemical 25 "C. At the indicated times (arrows),5 91 of 0.2 M ATP (1 mM final gradient of protons despite a high level of ATPase activity concentration) or 1.5 pl of 1 mM carbonylcyanide-m-chlorophenylhy drazone (CCCP) (1.5 NM) was added. A , membranes from strains (15). In contrast, membranes from strains carrying the yMet23-Lys mutation in combination with each suppressor mu- containing the double mutations of yMet-23-L~~with yGln-269+ yAla-27bVa1,yGlu-27bGly, or yVal-28bAla. B, memtation showed increased levels of acridine orangefluorescence Arg, branes from yMet-23-Lys with yArg-242+Cys, yIIe-272+Thr, quenching (Fig. 2). yThr-273+Ser, ox yIle-279-Thr. In control experiments, membranes from the wild type and _c

Suppression of a Coupling M %tation in the ATP Synthase

870

all the mutant strains (includingyMet-23-Lys) showed the Same extent of D-lactate-driven, respiratory fluorescence quenching(datanot shown). Furthermore,the D-lactatedriven quenching was not influenced by dicyclohexylcarbodiimide treatment which blocks the passive proton leakage of exposed Fo. These results indicate that none of the mutations caused proton leakiness through theenzyme or release of the F, complexes from the membranes. Carboxyl-terminal Suppressor MutationsLower ATPase Activity-Previously, the importance of the carboxyl-terminal region in catalysis was demonstrated by site-directed mutations replacing amino acids yGln-269, yThr-273, and yGlu275 which caused reduced ATPase activities (17). Similarly, many of the mutations between positions 269 and 280 described in this study also caused lower membrane ATPase activities (Table 11, left column). Moreover, we noted that combining each of these mutations with yMet-23-Lys resulted in ATPase activities which were further reduced by about half. This was consistent with the yMet-23jLys mutant itself which had about half the activity ofwild type (Table 11, right column). Differently, the ATPase activity of the yArg-242jCys mutant was the same, whether as the single or the double mutant with yMet-23-Lys. Temperature Sensitivity of the yMet-23+Lys, yGln-269Arg, and yThr-273-Ser Mutants-In general, strains carrying the suppressor mutations assingle mutations grew better than when combined with yMet-23-Lys (data not shown). Distinct exceptions were strains carrying yGln-269+Arg or yThr-273-Ser which grew to much lower yields a t 37 "C as single mutants than as double mutants with y M e t - 2 3 - L ~ ~ (Table 111, right column). Clearly, each mutation mutually suppressed the perturbationsof the other. The mutual suppression suggested that the effect of these mutations was to disrupt intra- and/or intersubunit interactions, but also that they did not cause gross conformational changes. Therefore, the perturbationscaused by these mutations were likely subtle,and lower temperature mayalso reverse the phenotype. This notion wasconfirmed by the higher growth levels at lower temperatures of the single mutationstrains, yGln-269+Arg, yThr-273+Ser,andyMet2 3 j L y s (Table 111).Importantly, the double mutants, yGln269+Arg/yMet-23-+Lys or yThr-273-Ser/yMet-23-Lys, had the same level of growth relative to wild type regardless of the temperature. Temperature-sensitive changes in energy coupling of the mutant enzymes paralleled the growth characteristics. The yMet-23-Lys mutantmembraneshad even lower ATPdriven acridine orange fluorescence quenching at 37 "C compared with that obtained at 25 "C (Fig. 3). Membranes from TABLE I1 ATPase activities of membranes from strain KFlOrA harboring suppressor mutations alone and in combination with y M e t - 2 3 - L ~ ~ ATPase activitiesof 10-15 figof membrane proteinwere measured in a buffer containing 20 mM Tris-HC1,4 mM ATP, and2 mM MgC12, DH 8.0. at 37 "C. Carboxyl-terminal With mutation

met-23

With yLys-23

(mutant) (wild tme) pmol PJmin . mg protein

None (wild type) Arg-2424Cys Gln-2694Arg Ala-27hVal 0.31 Ile-272+Thr Thr-273-Ser Glu-27kGly 0.71 Ile-279-Thr Val-28kAla

1.5 1.5 0.53 1.3 0.67 0.61 1.6 1.0 1.1

1.0 1.5 0.20 0.66 0.23 0.61 0.62

TABLEI11 Growth yields of single and double mutants by oxidative phosphorylation at different temperatures Growth yields in succinate medium are reported as percent relative to wild type growth (KFlOrA/pBWG15) at that temperature. The saturation densities measured at 650 nm obtained for wild type at each temperature are shown in the parentheses. Mutation(s) 25 "C 30 "C 37 "C 7% None (wild type) 100 (0.68) 100 (0.67) 100 (1.4) Met-23-+Lys 50

45

Gln-269-Arg Thr-273-Ser 79 76 Gln-269+Arg/Met-23+Lys Thr-273+Ser/Met-23+Lys

A

5

21 58 38 65

64 75

79 74

87

80

ATP

CCCP

J

1

I\

Arg~269+Lyr~23 Arg-269 Scr-273 Scr~Z73+Lyr~23

w

B

-

ATP

CCCP

J

.1

-

Lyr 23

FIG. 3. ATP-driven fluorescence quenching of temperature-sensitive mutations at 25 "C (A) and 37 "C ( B ) .Conditions and additions were as inFig. 2 except in B the buffer was adjusted to pH 8.0 at 37 "C and the temperatures of the cuvette contents were maintained at 37 'C. Traces with membranes from strains containing the single mutations, yMet-23+Lys (Lys-23),yGln-269-Arg (Arg269) or yThr-273+Ser (Ser-273),and the double mutations, yMet23-Lys/yGln-269+Arg (Arg-269+Lys-23) or yMet-P3+Lys/yThr273-Ser (Ser-273+Lys-23),are indicated. Quenching for wild type ( W T ) membranes are shown for comparison.

strains containingyGln-269-Arg or yThr-273-Ser as single or higher than as the double mutants had quenching the same mutants with y M e t - 2 3 - L ~ ~ a25 t "C (Fig. 3A), but lower at 37 "C (Fig. 3 B ) . This behavior seemed to contradict the ATPase activities at 37 "C which were higher for the single than for the double mutants (Table 11). These results suggested that the yGln-269-Arg or yThr-273-Ser mutant enzymes were inefficient in energy couplingas was previously observed in the case of yGln-269+Leu (17). We also note that the extent of quenching by the single mutants at 37 "C was unstable and decayed during the time of the assay(Fig. 3B).Incontrast,thequenching of the double mutants was stable. The instability was not due to release of the F, complex from Fo or proton-leaky ATPase.

Suppression of a Coupling Mutation in the ATP Synthase

871

/

At 37 "C, membranes from the single mutation strains had

Y.

stable respiratory fluorescence quenching the same as wild type and was not affected by dicyclohexylcarbodiimide treatment (data notshown). The perturbations to energy coupling caused by the mutations, yMet-23~Lys, yGln-269+Arg, and yThr-2734Ser are clearly temperature dependent, whereas the double mutations are thermally stable. DISCUSSION

Eight different mutations in the y subunit were able to restore efficient energy couplingto theyMet-23+Lys mutant. Seven of the mutations were changes between yGln-269 and yVal-280, and one replaced yArg-242. DNA sequencing of several mutagenized genes revealed that mutations were introduced by PCR throughout uncG however, only mutations in the carboxyl-terminal region were able to suppress the effects of yMet-B3+Lys. The large number of suppressor mutations found in this CATALYSIS TRANSPORT study strongly suggests that yMet-23, yArg-242, and yGlnFIG. 4. Working modelof the arrangement of a-helical seg269 to yVal-280 are close to each other andcapable of interments containing yMet-23, ~ A r g - 2 4 2and , ~ G l n - 2 6 9to ?Valacting. This hypothesis coincides with the demonstration of 280. The a-helical segmentswere predicted by two different secondintra-subunit interactions in the chloroplast y subunit. Illu- arystructurealgorithms (34, 35). The amino acids in ovals are mination in the presence of o-iodosobenzoateinduced for- replaced by the mutations described in this paper. Enzymes with mation of a disulfide bond between Cys-322 of the carboxyl replacements of the shaded residues were temperature-sensitive. terminus andCys-89 which falls in the thirdof the y subunit Combinations of yMet-23-LyslyGln-269-Arg or yMet-234Lysl conserved regions (14,25). These resultscombined with those yThr-273-Ser were thermally stable. presented heresuggest that all three of the conserved stretches found in the y subunit (amino acids 3-43, 83-89, and 237- yThr-273-Ser mutations showed somewhat inefficient cou283; E. coli numbering) appear to be close to each other. We pling but in a less obvious manner because these mutations propose that these regions of the y subunit form a functional perturbed catalysis aswell. domain which mediates energycoupling in the manner of The distribution of seven suppressormutations over a "indirect" or conformational coupling as hypothesized by stretch of 12 amino acidsin an a-helixsuggests that effect of several authors (26-28). the yMet-23+Lys mutation is to disrupt interactionsbetween Of the 8 different residues which were implicated by the regions of the enzyme complex as opposed todisrupting suppression, our attention was drawn to the changes of the specific interactions between itself and other amino acids. conserved residues yGln-2694Arg and yThr-273-Ser. First, Thisnotionisstrengthened by the observation that the replacementsinthesepositions were also deleteriousto characteristics of the types of amino acid changes do notfall growth, and their effects were in turn suppressed by yMet- into a pattern (e.g. large to small or polar to non-polar). The 23-Lys. Second, the perturbations in energy coupling caused data limit the number of possible ways in which the yMetby replacements in positions yMet-23, yGln-269, and yThr- 23+Lys mutation can perturbenergy coupling. One possibil273 were temperature-dependent. Either lower temperature ity is thata positive charge at position 23 forces the carboxylor combining the mutations (yMet-23+Lys/yGln-269+Arg terminal region into a deleterious role and any one of the or yMet-23-LyslyThr-27343er) partially reversed the de- suppressor mutations mitigates its influence on the enzyme. fects. Interestingly, the double mutations restored efficient However, this is unlikely because of the apparent importance energy coupling but not higher ATPase activities. Previous of the conserved carboxyl terminus to ATPase function. Sinstudies have shown that the y subunit has roles in catalysis gle amino acid replacements greatly reduced ATPase activity, (17,25,29,30) and transport (10-13,15,25,31-33); we suggest and furthermore, deletion of this region resulted in defective that yGln-269 and yThr-273 are intimately involved in both assembly of the complex (17). A more favorable explanation of those roles. is that yMet-23+Lys destabilizes an interaction between the Fig. 4 illustrates ourworking model in which yMet-23 is in carboxyl-terminal region and another region in the enzyme an a-helix proximal to helices containing yGln-269 to yVal- complex (which may be the amino-terminal region of the y 280 on one side and yArg-242 on the other. According to subunit or in the other subunits). The destabilizationof this secondary structural predictions (34,35), the two helices (one interaction, which is clearly important for energy coupling, is containing yArg-242, and the other, yGln-269 to yVal-280) likely to be subtle because it is partially reversed by lower are separated by random coil and turn structures. Therefore, temperature, or by one of the several second-site mutations. In summary, the second-site suppressor mutations identisituating thetwo regions on separatehelices is reasonable and consistent with the difference in behavior between the yArg- fied in the conserved carboxyl-terminal region serve to define 2 4 2 j C y s suppressor mutation and the others. The yArg- a functional domain within the y subunit of the ATP syn242jCys mutation was capable not only of suppressing the thase. The domain,which likely includes the amino-terminal yLys-23 effects on coupling, but ATPase catalysis as well. region and the conserved region around ycys-87 as well, has Furthermore, we placed residues 269 and 273 on the side of roles in both catalysis and thecoupling of catalysis to transthe helix facing away from residue 23 because even though port. mutations in these positions mutually suppressed each other, REFERENCES their effects on theenzyme were different. Thismodel predicts M., and Kanazawa, H. (1983)Microbiol. Rev. 47, 285-312 that carboxyl-terminal mutations with behavior similar to the 2.1. Futai, Walker, J. E., Saraste, M., and Gay, N. J. (1984) Biochim. Biophys. Acto y M e t - 2 3 - L ~ ~would be expected. In fact, yGln-269-+Arg and 768,164-200

f

872

SuppressionCoupling of a

Mutation in the ATP Synthase

3. Futai, M., Noumi, T., and Maeda, M. (1989) Annu. Rev. Biochem. 58,111136 4. Fillingame, R. H. (1990) in The Bacteria (Krulwich, T.A,, ed)Vol. XII, pp. 345-391, Academic Press, Inc., New York 5 . Senior, A. E. (1990) Annu. Reu. Biophys. Biophys. Chem. 19,7-41 6. Gogol, E. P., Aggeler, R., Sagermann, M., andCapaldi, R. A. (1989) Biochemistry 2 8 , 4717-4724 7. Boekema, E. J., Xiao, J., and McCarty,R. E. (1990) Biochim. Biophys. Acta 1020,49-56 8. Futai, M. (1977) Biochem. Biophys. Res. Commun. 103,604-612 9. D u m , S. D., and Futai, M. (1980) J. Biol. Chem. 255,113-118 10. Weiss, M. A,, and McCarty, R. E. (1977) J. Biol. Chem. 252,8007-8012 11. Moroney, J. V., and McCarty, R. E. (1979) J. Biol. Chem. 254,8951-8955 12. Moroney, J. V., Andreo, C. S., Vallejos, R. H., and McCarty, R. E. (1980) J. Biol. Chem. 255,6670-6674 13. Moroney, J. V., and McCarty, R. E. (1982) J. Biol. Chem. 257,5915-5920 14. Miki. J.. Maeda., M... Mukohata.. Y... and Futai, M. (1988) FEES Lett. 2 3 2 , 22i-226 15. Shin, K., Nakamoto, R. K., Maeda, M., and Futai, M.(1992) J. Biol. Chem. 267,20835-20839 16. Kanazawa. H.. Hama, H.,Rosen, B. P., and Futai,M. (1985)Arch. Biochem. Biophys.' 241,3641370 17. Iwamoto, A,, Miki, J., Maeda,M., and Futai, M.(1990) J. Biol. Chem. 2 6 5 , 5043-5048 18. Miki, J., Takeyama, M., Noumi, T., Kanazawa, H., Maeda, M., and Futai, M. (1986) Arch. Biochem. Biophys. 251,458-464 19. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G .

T., Mullis, K. B., and Erlich, H. A. (1988) Science 239,487-491 20. Sanger, F., Coulson, A.R., Barrell, B. G., Smith, A. J. H., and Roe, B. A. (1980) J. Mol. Biol. 1 4 3 , 161-178 21. Sambrook, J., Fritsch, E. F., and Maniatis, T.(1989) Molecular Cloning: a Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Flarhor N Y

"__I_-(

22. Futai, M., Sternweis, P. C., and Heppel, L. A. (1974) Proc. Natl. Acad. Sci. U. S. A. 7 1,2725-2729 23. Kanazawa, H. Miki, T., Tamura, F., Yura, T., and Futai, M. (1979) Proc. Natl. Acad. h i . U. S. A. 76,1126-1130 24. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J . Biol. Chem. 193, 265-275 25. Moroney, J. V., Fullmer, C. S., and McCarty, R. E. (1984) J. Biol. Chem. 2 5 9 , 7281-7285 26. Boyer, P. D. (1975) FEES Lett. 5 8 , l - 6 27. Harris, D. A. (1978) Biochim. Biophys. Acta 463,245-273 28. Jencks, W. P. (1980) Adu. Enzymol. 51,75-106 29. Nalin, C. M., and McCarty, R. E. (1984) J. Biol. Chem. 259,7275-7280 30. Ketcham, S. R., Davenport, J. W., Warncke, K., and McCarty,R. E. (1984) J. Biol. Chem. 259,7286-7293 31. Yoshida M. Okamoto, H., Sone, N., Hirata, H., and Kagawa, Y. (1977) Proc. h a t i Acad. Sci. U. S. A. 74,936-940 32. Humhert, R., and Altendorf, K. (1989) J . Bacteriol. 1 7 1 , 1435-1444 33. P a a, S , Guerrieri, F., Zanotti, F., Fiermonte, M., Capozza, G., and Jirillo, (1990) FEES Lett. 272,117-120 34. Chou, P. Y., and Fasman, G. D. (1974) Biochemistry 13,212-245 35. Gamier, J., Osguthorpe, D. J., and Robson, B. (1978) J. Mol. Biol. 1 2 0 , 97-120

E.