Portal de Periódicos da CAPES

Mechanism of Phosphoryl Transfer in the Dimeric IIABMan Subunit of the Escherichia coliMannose Transporter

1999; Elsevier BV; Volume: 274; Issue: 10 Linguagem: Inglês

10.1074/jbc.274.10.6091

1083-351X

Regula Gutknecht, Karin Flükiger, R Lanz, Bernhard Erni,

DNA Repair Mechanisms

The mannose transporter of bacterial phosphoenolpyruvate:sugar phosphotransferase system (PTS) mediates uptake of mannose, glucose, and related hexoses by a mechanism that couples translocation with phosphorylation of the substrate. It consists of the transmembrane IICMan·IIDMan complex and the cytoplasmic IIABMan subunit. IIABManhas two domains (IIA and IIB) that are linked by a 60-Å long alanine-proline-rich linker. IIABMan transfers phosphoryl groups from the phospho-histidine-containing phospho-carrier protein of the PTS to His-10 on IIA, hence to His-175 on IIB, and finally to the 6′-OH of the transported hexose. IIABMan occurs as a stable homodimer. The subunit contact is mediated by a swap of β-strands and an extensive contact area between the IIA domains. The H10C and H175C single and the H10C/H175C double mutants were used to characterize the phosphoryl transfer between IIA to IIB. Subunits do not exchange between dimers under physiological conditions, but slow phosphoryl transfer can take place between subunits from different dimers. Heterodimers of different subunits were produced in vitroby GuHCl-induced unfolding and refolding of mixtures of two different homodimers. With respect to wild-type homodimers, the heterodimers have the following activities: wild-type·H10C, 50%; wild-type·H175C 45%; H10C·H175C, 37%; and wild-type·H10C/H175C (double mutant), 29%. Taken together, this indicates that both cis andtrans pathways contribute to the maximal phosphotransferase activity of IIABMan. A phosphoryl group on a IIA domain can be transferred either to the IIB domain on the same or on the second subunit in the dimer, and interruption of one of the two pathways results in a reduction of the activity to 70–80% of the control. The mannose transporter of bacterial phosphoenolpyruvate:sugar phosphotransferase system (PTS) mediates uptake of mannose, glucose, and related hexoses by a mechanism that couples translocation with phosphorylation of the substrate. It consists of the transmembrane IICMan·IIDMan complex and the cytoplasmic IIABMan subunit. IIABManhas two domains (IIA and IIB) that are linked by a 60-Å long alanine-proline-rich linker. IIABMan transfers phosphoryl groups from the phospho-histidine-containing phospho-carrier protein of the PTS to His-10 on IIA, hence to His-175 on IIB, and finally to the 6′-OH of the transported hexose. IIABMan occurs as a stable homodimer. The subunit contact is mediated by a swap of β-strands and an extensive contact area between the IIA domains. The H10C and H175C single and the H10C/H175C double mutants were used to characterize the phosphoryl transfer between IIA to IIB. Subunits do not exchange between dimers under physiological conditions, but slow phosphoryl transfer can take place between subunits from different dimers. Heterodimers of different subunits were produced in vitroby GuHCl-induced unfolding and refolding of mixtures of two different homodimers. With respect to wild-type homodimers, the heterodimers have the following activities: wild-type·H10C, 50%; wild-type·H175C 45%; H10C·H175C, 37%; and wild-type·H10C/H175C (double mutant), 29%. Taken together, this indicates that both cis andtrans pathways contribute to the maximal phosphotransferase activity of IIABMan. A phosphoryl group on a IIA domain can be transferred either to the IIB domain on the same or on the second subunit in the dimer, and interruption of one of the two pathways results in a reduction of the activity to 70–80% of the control. The carbohydrate transporters of the bacterial phosphotransferase system (enzymes II of the PTS) 1The abbreviations used are: PTS, phosphoenolpyruvate-dependent carbohydrate:phosphotransferase system (EC 2.7.1.69); IIABMan, hydrophilic subunit of the mannose transporter; IICMan and IIDMan, transmembrane subunits of the mannose transporter; HPr, histidine-containing phospho carrier protein of the PTS; PEP, phosphoenolpyruvate; GuHCl, guanidinium hydrochloride; MOPS, 4-morpholinepropanesulfonic acid. mediate uptake concomitant with phosphorylation of hexoses and hexitols. They consist of four functional units termed IIA, IIB, IIC, and IID that occur either as individual subunits in a protein complex or as independently folding domains of a multidomain protein. IIA and IIB sequentially transfer a phosphoryl group from the phosphoryl carrier protein HPr to the transported substrate. IIC and IID span the membrane and mediate substrate translocation. Substrate translocation is activated by the phosphorylation/dephosphorylation cycle of IIB (1Postma P.W. Lengeler J.W. Jacobson G.R. Microbiol. Rev. 1993; 57: 543-594Crossref PubMed Google Scholar, 2Postma P.W. Lengeler J.W. Jacobson G.R. Neidhardt F.C. Curtiss R. Ingraham J.L. Lin E.C.C. Low K.B. Magasanik B. Reznikoff W.S. Riley M. Schaechter M. Umbarger H.E. Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. ASM Press, Washington, D. C.1996: 1149-1174Google Scholar, 3Erni B. Int. Rev. Cytol. 1992; 137A: 127-148Crossref Scopus (22) Google Scholar, 4Meadow N.D. Fox D.K. Roseman S. Annu. Rev. Biochem. 1990; 59: 497-542Crossref PubMed Scopus (302) Google Scholar). IIA and IIB of certain transporters have regulatory activity in addition to their "energy-transducing" function. For instance, IIAGlc ofEscherichia coli, the gene product of crr, modulates the activities of adenylate cyclase (5Peterkofsky A. Seok Y.-J. Amin N. Thapar R. Lee S.Y. Klevit R.E. Waygood E.B. Anderson J.W. Gruschus J. Huq H. Gollop N. Biochemistry. 1995; 34: 8950-8959Crossref PubMed Scopus (22) Google Scholar, 6Peterkofsky A. Reizer A. Reizer J. Gollop N. Zhu P.P. Amin N. Prog. Nucleic Acid Res. Mol. Biol. 1993; 44: 31-65Crossref PubMed Scopus (46) Google Scholar), glycerol kinase (7Hurley J.H. Faber H.R. Worthylake D. Meadow N.D. Roseman S. Pettigrew D.W. Remington S.J. Science. 1993; 259: 673-677Crossref PubMed Scopus (207) Google Scholar), and of the membrane permeases for lactose and maltose (8Van der Vlag J. Van Dam K. Postma P.W. J. Bacteriol. 1994; 176: 3518-3526Crossref PubMed Google Scholar, 9Van der Vlag J. Postma P.W. Mol. Gen. Genet. 1995; 248: 236-241Crossref PubMed Scopus (12) Google Scholar, 10Osumi T. Saier Jr., M.H. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 1457-1461Crossref PubMed Scopus (75) Google Scholar, 11Saier M.H. Novotny M.J. Comeau-Fuhrman D. Osumi T. Desai J.D. J. Bacteriol. 1983; 155: 1351-1357Crossref PubMed Google Scholar, 12Dean D.A. Reizer J. Nikaido H. Saier Jr., M.H. J. Biol. Chem. 1990; 265: 21005-21010Abstract Full Text PDF PubMed Google Scholar). The IIB domains of some PTS transporters regulate the activity of antiterminator and transcription activator proteins (13Stülke J. Arnaud M. Rapoport G. Martin-Verstraete I. Mol. Microbiol. 1998; 28: 865-874Crossref PubMed Scopus (192) Google Scholar). In the absence of the cognate substrate, the IIB domain of the β-glucoside transporter (IIBCABgl) phosphorylates the antiterminator protein BglG and thereby inactivates it. This way, IIBCABglfeedback inhibits its own expression in the absence of a transportable substrate (inducer) (14Schnetz K. Rak B. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 5074-5078Crossref PubMed Scopus (75) Google Scholar, 15Schnetz K. Stülke J. Gertz S. Krüger S. Krieg M. Hecker M. Rak B. J. Bacteriol. 1996; 178: 1971-1979Crossref PubMed Google Scholar). The tertiary and quarternary structures of IIA units from different families of PTS transporters are completely unrelated. IIA occur as monomers (IIAGlc) (7Hurley J.H. Faber H.R. Worthylake D. Meadow N.D. Roseman S. Pettigrew D.W. Remington S.J. Science. 1993; 259: 673-677Crossref PubMed Scopus (207) Google Scholar), stable dimers (IIAMan) (19Nunn R.S. Markovic-Housley Z. Génovésio-Taverne J.C. Flükiger K. Rizkallah P.J. Jansonius J.N. Schirmer T. Erni B. J. Mol. Biol. 1996; 259: 502-511Crossref PubMed Scopus (61) Google Scholar), or trimers (IIALac) (20Sliz P. Engelmann R. Hengstenberg W. Pai E.F. Structure (Lond.). 1997; 5: 775-788Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar) of identical subunits. Similarly, the IIB units have different 3D structures but are monomeric (21Eberstadt M. Grdadolnik S.G. Gemmecker G. Kessler H. Buhr A. Erni B. Biochemistry. 1996; 35: 11286-11292Crossref PubMed Scopus (44) Google Scholar, 22Van Montfort R.L.M. Pijning T. Kalk K.H. Reizer J. Saier Jr., M.H. Thunnissen M.M.G.M. Robillard G.T. Dijkstra B.W. Structure (Lond.). 1997; 5: 217-225Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar, 23Schauder S. Nunn R.S. Lanz R. Erni B. Schirmer T. J. Mol. Biol. 1998; 276: 591-602Crossref PubMed Scopus (38) Google Scholar). The membrane-spanning IIC and IID subunits occur as oligomers, mostly dimers (24Meins M. Zanolari B. Rosenbusch J.P. Erni B. J. Biol. Chem. 1988; 263: 12986-12993Abstract Full Text PDF PubMed Google Scholar, 25Boer H. ten Hoeve-Duurkens R.H. Schuurman-Wolters G.K. Dijkstra A. Robillard G.T. J. Biol. Chem. 1994; 269: 17863-17871Abstract Full Text PDF PubMed Google Scholar, 26Chen Q. Amster-Choder O. Biochemistry. 1998; 37: 8714-8723Crossref PubMed Scopus (16) Google Scholar, 27Boer H. ten Hoeve-Duurkens R.H. Lolkema J.S. Robillard G.T. Biochemistry. 1995; 34: 3239-3247Crossref PubMed Scopus (16) Google Scholar, 28Meijberg W. Schuurman-Wolters G.K. Robillard G.T. J. Biol. Chem. 1998; 273: 7949-7956Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar). The multidomain composition of the PTS transporters and their dimeric structure allows for various forms of interallelic and intergenic complementation. For instance, the coexpression of two mutated IICBGlc subunits of the glucose transporter with inactive B and C domains, respectively, resulted in complementation of transport activity (29Lanz R. Erni B. J. Biol. Chem. 1998; 273: 12239-12243Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar). Complementation has also been observed between inactive mutants of IICBAMtl (30Meijberg W. Schuurman-Wolters G.K. Robillard G.T. Biochemistry. 1996; 35: 2759-2766Crossref PubMed Scopus (13) Google Scholar, 31Boer H. ten Hoeve-Duurkens R.H. Robillard G.T. Biochemistry. 1996; 35: 12901-12908Crossref PubMed Scopus (25) Google Scholar) and between two inactive mutants of the paralogous transporters for Glc and for GlcNAc (IICBGlc and IICBAGlcNAc) ofE. coli (32Vogler A.P. Lengeler J.W. Mol. Gen. Genet. 1988; 213: 175-178Crossref PubMed Scopus (25) Google Scholar, 33Vogler A.P. Broekhuizen C.P. Schuitema A. Lengeler J.W. Postma P.W. Mol. Microbiol. 1988; 2: 719-726Crossref PubMed Scopus (49) Google Scholar). It is generally assumed that complementation in vitro is because of the formation of heterodimers between two different inactive subunits and not only to transient association of different inactive homodimers. The E. coli IIABMan subunit is a homodimer (see Fig. 1 A). Each monomer comprises two independently folding domains, the A domain (residues 1–133) and the B domain (residues 156–323) connected by a 23-residue long alanine-proline-rich linker (35Erni B. FEMS Microbiol. Rev. 1989; 63: 13-23Crossref Google Scholar, 36Erni B. Zanolari B. Graff P. Kocher H.P. J. Biol. Chem. 1989; 264: 18733-18741Abstract Full Text PDF PubMed Google Scholar). The IIAMan domain contains a five-stranded β-sheet (strand order 21345) covered by helices on either face ((βα)4,αβ). Four strands are parallel, and the fifth antiparallel strand which forms one edge of the sheet is swapped between the subunits in the dimer. His-10, which is phosphorylated during phosphoryl transfer from HPr to IIB, is located at the topological switchpoint of the fold. Its imidazole ring is hydrogen bonded to Asp-67, which acts as a general base increasing the nucleophilicity of the imidazole ring (19Nunn R.S. Markovic-Housley Z. Génovésio-Taverne J.C. Flükiger K. Rizkallah P.J. Jansonius J.N. Schirmer T. Erni B. J. Mol. Biol. 1996; 259: 502-511Crossref PubMed Scopus (61) Google Scholar). The B domain contains a 180° twisted seven-stranded β-sheet (strand order 3241567, 1–6 are parallel and 7 is antiparallel) covered by helices on both faces, as deduced from the IIBLev subunit which is 47% identical to the IIBMan domain. His-175, which accepts the phosphoryl group from His-10 and transfers it to the sugar, is located on an exposed loop between the first β-strand and α-helix (23Schauder S. Nunn R.S. Lanz R. Erni B. Schirmer T. J. Mol. Biol. 1998; 276: 591-602Crossref PubMed Scopus (38) Google Scholar). Only the A domain participates in the dimer interface. The monomer-monomer interaction occurs through the interlocked β-strands and an extensive contact area of 1700 Å2 composed mainly of hydrophobic residues. This confers high stability, and the IIABMan dimer can be dissociated only concomitant with complete denaturation (37Markovic-Housley Z. Cooper A. Lustig A. Flükiger K. Stolz B. Erni B. Biochemistry. 1994; 33: 10977-10984Crossref PubMed Scopus (27) Google Scholar). The B domain interacts with the transmembrane IICMan·IIDMan complex of the mannose transporter. The IIABMan·IICMan·IIDMan complex, which can be purified intact, has a stoichiometry closest to 2:1:2 (38Rhiel E. Flükiger K. Wehrli C. Erni B. Biol. Chem. Hoppe-Seyler. 1994; 375: 551-559Crossref PubMed Scopus (15) Google Scholar, 39Huber F. Erni B. Eur. J. Biochem. 1996; 239: 810-817Crossref PubMed Scopus (37) Google Scholar, 40Erni B. Zanolari B. Kocher H.P. J. Biol. Chem. 1987; 262: 5238-5247Abstract Full Text PDF PubMed Google Scholar). The IIABMan dimer can also be purified as a soluble protein. Dissociated from the transmembrane IICMan·IIDMan complex, IIABManhas an elongated form. Ratios f/fo of 1.81 and 1.72 were calculated from the sedimentation coefficient (s 20,w = 3.7 S) determined by analytical ultracentrifugation (37Markovic-Housley Z. Cooper A. Lustig A. Flükiger K. Stolz B. Erni B. Biochemistry. 1994; 33: 10977-10984Crossref PubMed Scopus (27) Google Scholar) and the diffusion coefficient (D = 4.73 10−7 cm2sec−1) determined by dynamic light scattering, respectively. The axial ratio of >10:1 derived fromf/fo (41Cantor C.R. Schimmel P.R. Biophysical Chemistry. W. H. Freeman and Co., San Francisco1980: 557-565Google Scholar) is compatible with a fully extended dimer (Fig. 1 B) composed of the central A dimer (50 Å along the major axis), the two linkers (66 Å when fully elongated), and the two B domains (35 Å average diameter). It has been shown previously (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar) that the active site mutants of IIABMan, H10C, and H175C, are completely inactive when assayed alone, but that approximately 3% of wild-type activity is recovered when the purified proteins are mixed in a 1:1 ratio. Here we show, that much higher activity is recovered when the purified mutants are mixed, completely unfolded with GuHCl, and then renatured. True heterodimers form only under these drastic conditions. Phosphoryl transfer between subunits within the dimer is very efficient, whereas transfer between different dimers is possible but inefficient. E. coli WA2127ΔHIC (manXYZ ptsHIcrr (42Mao Q. Schunk T. Flükiger K. Erni B. J. Biol. Chem. 1995; 270: 5258-5265Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar)) was transformed with derivatives of pJFL encoding wild-type and mutant IIABMan (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar). IIABMan was overexpressed and purified as described (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar). Enzyme I and HPr were purified, and membranes containing IICMan·IIDMan were prepared as described (42Mao Q. Schunk T. Flükiger K. Erni B. J. Biol. Chem. 1995; 270: 5258-5265Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar,43Erni B. Zanolari B. J. Biol. Chem. 1985; 260: 15495-15503Abstract Full Text PDF PubMed Google Scholar, 44.Deleted in proof.Google Scholar). Stock solution of purified wild-type and mutant IIABMan were adjusted to a protein concentration of 5 mg/ml. Volumes from the different stocks were mixed to achieve the desired molar ratios or molar fractions. The mixtures were then split in two aliquots. One aliquot was diluted with 8 m GuHCl to a final concentration of 4 m GuHCl (37Markovic-Housley Z. Cooper A. Lustig A. Flükiger K. Stolz B. Erni B. Biochemistry. 1994; 33: 10977-10984Crossref PubMed Scopus (27) Google Scholar), and to the other aliquot, the same volume of buffer A (10 mm MOPS, pH 7.0, 50 mm NaCl, 0.5 mm dithiothreitol) was added. Both samples were incubated for 2 h at room temperature. Both samples were then diluted 20–60-fold with buffer A to the desired a IIABMan concentration (3–125 μg/ml) and incubated for another 2 h at 4 °C. In vitrophosphorylation of [14C]Glc was assayed by ion-exchange chromatography as described (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar). 100 μl of incubation mixture contained 0.5 μg of enzyme I, 2.8 μg of HPr, and 0.5 μl of crude membranes (∼4 μg of protein) containing the IICMan·IIDMan complex. The final concentration of renatured IIABMan varied between 3 ng and 130 ng/100 μl of incubation mixture. The exact values are indicated in the figure legends. The specific activity of [14C]Glc was 1000 cpm/nmol. The rate and the extent of protein phosphorylation was measured as described (45Gutknecht R. Lanz R. Erni B. J. Biol. Chem. 1998; 273: 12234-12238Abstract Full Text Full Text PDF PubMed Scopus (12) Google Scholar). The incubation mixture (50 mm NaPi, pH 7.4, 5 mm MgCl2, 2.5 mm NaF, 2.5 mm dithiothreitol) contained, per 250 μl, 1.5 μg of enzyme I, 2.5 μg of HPr, and 85 μg of IIABMan. The phosphorylation reaction was started by adding to the incubation mixture at 24 °C [33P]PEP to a final concentration of 80 μm. Aliquots of 40 μl were withdrawn at the indicated time points and diluted into 1 ml of 80% ammonium sulfate solution at 4 °C. The protein precipitates were collected on glass microfibre filters (GF/F, Whatman) under suction, washed, and counted in a liquid scintillation counter. The background counts because of enzyme I and HPr (less than 10%) were subtracted from the counts of the complete system. Phosphorylated proteins were analyzed on 17.5% polyacrylamide gels as described (21Eberstadt M. Grdadolnik S.G. Gemmecker G. Kessler H. Buhr A. Erni B. Biochemistry. 1996; 35: 11286-11292Crossref PubMed Scopus (44) Google Scholar). 20-μl incubation mixtures contained 134 μm [33P]PEP, 0.15 μg of enzyme I, 0.46 μg of HPr, 10 μg of IIABMan, and 0.3 μl of IICMan·IIDMan-containing membranes. Wild-type IIABMan, H10C, H175C, and H10C/H175C double mutant were purified by phosphocellulose chromatography and gel filtration. A 1:1 mixture of purified H10C and H175C has about 5% of the specific activity of wild-type IIABMan. The activity increases nonlinearly at low concentration, and the concentration activity profile does not change after 24 h of preincubation (Fig.2). These results suggest that the activity is because of transient association between two different inactive homodimers (second order reaction) and that monomers do not exchange to form heterodimers. However, when mixtures of H10C and H175C were denatured in GuHCl and then refolded by rapid dilution, a 20-fold higher specific phosphotransferase activity was obtained (Fig.3 A). When H10C and H175C were mixed in different proportions, the activity profile was bell-shaped with a maximum at a 1:1 molar ratio (Fig. 3 B), as expected for a binomial distribution of active heterodimers and inactive homodimers. The activities of heterodimers between wild-type and mutated subunits was characterized in the following experiments. Constant amounts of wild-type IIABMan were mixed with increasing amounts of either H10C or H175C. One-half of the mixture was denatured with GuHCl and then renatured by dilution, the other was diluted only. The phosphotransferase activity remained approximately constant at all concentrations of H10C and H175C (Fig. 3 C) independently of whether 100% of wild-type IIABMan occurs as homodimer (no GuHCl) or whether only 11% of IIABMan was in homodimers and the rest in heterodimers with an inactive subunit. The activity was linearly dependent upon the concentration of wild-type IIABMan when wild-type and H10C or H175C were mixed in different molar ratios, denatured, and then renatured (Fig.3 D). This suggests that the presence of a subunit with only one inactive domain in a heterodimer has no effect on the overall phosphotransferase activity of the wild-type subunit. Mixtures between wild-type IIABMan and an excess of the H10C/H175C double mutant were prepared to characterize the phosphoryl transfer between A and B domains on the same subunit. The concentration of wild-type IIABMan was kept constant, and the concentration of the double mutant increased to a maximum of 16:1 (Fig. 3, E andF). At a concentration ratio of 16:1, when only 6% of the wild-type protein is in homodimers and 94% in heterodimers with the double mutant, the activity is still 60% of the control and identical to the activity of the nondenatured mixture. The 40% decrease of activity is because of competition of the excess of inactive homodimers (8-fold over active homo- and heterodimers) for the IICMan·IIDMan complex. Competitive inhibition becomes more pronounced when the concentration of IICMan·IIDMan is rate-limiting. Under these conditions, the phosphotransferase activity is reduced to 50% when the concentration of wild-type homodimer plus heterodimer equals the concentration of the H10C/H175C homodimer (Fig. 3 F).Figure 3Effect of GuHCl unfolding/renaturation upon phosphotransferase activity of IIABManmixtures. Open symbols, denatured/renatured mixtures; closed symbols, native mixtures. A, effect of renaturation/denaturation on the activity of a 1:1 mixture of H10C and H175C (42 ng of IIABMan/μl added). B, complementation between H10C and H175C at the indicated molar fractions (125 ng of IIABMan per assay point). C, noncomplementation between wild-type IIABMan and H10C (circles) and H175C (squares) in the presence of an excess of the mutant over wild-type IIABMan (3 ng of wild-type IIABMan per assay point at all molar ratios). D, noncomplementation between wild-type IIABMan and H10C (circles) and H175C (squares) at the indicated molar fractions (6.2 ng of IIABMan per assay point).E, negative dominant effect of the H10C/H175C double mutant over wild-type IIABMan in the presence of an excess of IICMan·IIDMan (7 ng of wild-type IIABMan per assay point at all molar ratios). F, same as panel E but with a limiting concentration of IICMan·IIDMan. Means and S.D. are of three experiments.View Large Image Figure ViewerDownload (PPT) With each experiment, a control with pure wild-type IIABManwas carried along as a reference for 100% activity and as control of refolding yield. The activity recovered after rapid dilution of wild-type IIABMan was 80 ± 30% (TableI, column IIABMan homodimer). The specific activity of heterodimers was calculated as follows. The activity contributed by IIABMan wild-type homodimers was subtracted from the total phosphotransferase activity of a mixture of all dimers. The resulting difference was then divided by the concentration of heterodimers in the mixture. The concentrations of homo- and heterodimers were calculated from the binomial distribution. The specific activities of the different dimers are summarized in TableI. The turn-over number of wild-type IIABMan from experiment to experiment varies between 2500 min−1 and 1200 min−1. The H10C·H175C heterodimer has a turnover of 370 min−1. This is 37% of the activity of wild-type IIABMan measured under the same conditions. The turn-over numbers of heterodimers between a wild-type subunit and either H10C or H175C are 50 and 45% of wild-type homodimer, and the turnover-number of a heterodimer between a wild-type subunit and a H10C/H175C double mutant is 30%.Table IPhosphotransferase activity of IIABMan dimersSpecific activity (per dimer)Measured for mixture of all dimersCalculated forIIABMan homodimerActive heterodimerInactive homodimernmol Glc 6-phosphate/nmol dimer/minIIABMan · IIABManMixedaIIABMan components mixed and diluted.25501 × 2550GuHClbIIABMan components mixed, denatured and refolded by dilution.15861 × 1586 (62%)c% recovered wild-type IIABMan activity after denaturation and refolding.IIABMan · IIABManMixed12301 × 1230GuHCl9961 × 996 (81%)c% recovered wild-type IIABMan activity after denaturation and refolding.H10C · H175C (1:1)Mixed10eActivity due to phosphoryltransfer between dimers.1 × 0GuHCl1860.5 × 372 (37%)dActivity of heterodimer in % of wild-type dimer.0.5 × 0IIABMan · IIABManMixed26561 × 2656GuHCl18051 × 1805 (67%)c% recovered wild-type IIABMan activity after denaturation and refolding.IIABMan · H10C (1:8)Mixed18131 × 18138 × 0GuHCl17870.11 × 18131.77 × 896 (50%)dActivity of heterodimer in % of wild-type dimer.7.11 × 0IIABMan · IIABManMixed22461 × 2246GuHCl13871 × 1387 (62%)c% recovered wild-type IIABMan activity after denaturation and refolding.IIABMan · H175C (1:8)Mixed15801 × 15808 × 0GuHCl14320.11 × 15801.77 × 710 (45%)dActivity of heterodimer in % of wild-type dimer.7.11 × 0IIABMan · IIABManMixed11541 × 1154GuHCl14961 × 1496 (130%)c% recovered wild-type IIABMan activity after denaturation and refolding.IIABMan · H10C/H175C (1:16)Mixed8341 × 83216 × 0GuHCl9210.06 × 14961.88 × 442 (29%)dActivity of heterodimer in % of wild-type dimer.15.1 × 0The specific activities were calculated from Fig. 3, A–Eand the control experiments with pure wild-type IIABMan. The specific activities (in bold) of the homo- and heterodimers were calculated from the measured activity of the mixtures of dimers, the measured specific activity of pure wild-type IIABManhomodimers (in bold), and the concentrations of hetero- and homodimers in the mixtures derived from the binomial distribution (initalics).a IIABMan components mixed and diluted.b IIABMan components mixed, denatured and refolded by dilution.c % recovered wild-type IIABMan activity after denaturation and refolding.d Activity of heterodimer in % of wild-type dimer.e Activity due to phosphoryltransfer between dimers. Open table in a new tab The specific activities were calculated from Fig. 3, A–Eand the control experiments with pure wild-type IIABMan. The specific activities (in bold) of the homo- and heterodimers were calculated from the measured activity of the mixtures of dimers, the measured specific activity of pure wild-type IIABManhomodimers (in bold), and the concentrations of hetero- and homodimers in the mixtures derived from the binomial distribution (initalics). IIABMan is phosphorylated with [33P]PEP in the presence of enzyme I and HPr and is dephosphorylated in the presence of IICMan·IIDMan and glucose (Fig.4 A). The H175C mutant is stably phosphorylated at His-10 but cannot be dephosphorylated because His-175 is missing. The H10C mutant is weakly phosphorylated although His-10 is missing. It is dephosphorylated in the presence of IICMan·IIDMan and glucose, indicating that phosphorylation occurred at His-175. Phosphorylation of H10C is HPr-dependent but much slower than phosphorylation of wild-type IIABMan (Fig. 5). His-175 must be phosphorylated by HPr directly. Contamination of H10C by IIABMan, which could complement the IIA function, can be excluded because H10C was isolated from an E. coli strain with a chromosomal deletion of the manXYZ operon. It is likely, that phosphorylation of IIB is a consequence of high local concentration of HPr which binds to mutated IIA and then nonspecifically delivers the phosphoryl group to a nearby His-175. Phosphorylation at His-10, whether in wild-type IIABMan or in H175C results in an increased stabilization of the IIABMan dimer against dissociation by sodium dodecyl sulfate, and this effect is not reversed as a consequence of dephosphorylation by IICMan·IIDMan and mannose (Fig. 4 B).Figure 5Time course of phosphorylation of IIABMan. Purified wild-type IIABMan (circles) and H10C (squares) was incubated with [33P]PEP in the presence of catalytic amounts of enzyme I with HPr (open symbols) and without HPr (closed symbols). The reaction was stopped at the indicated time points by ammonium sulfate precipitation. Protein precipitates were collected on filters and counted.View Large Image Figure ViewerDownload (PPT) IIABMan consists of two domains, IIA and IIB, that sequentially transfer a phosphoryl group from the phosphoryl carrier protein HPr to the transported sugar. IIABMan is a homodimer. The subunits are tightly linked through mutual exchange of β-strands between the β-sheets of IIA (19Nunn R.S. Markovic-Housley Z. Génovésio-Taverne J.C. Flükiger K. Rizkallah P.J. Jansonius J.N. Schirmer T. Erni B. J. Mol. Biol. 1996; 259: 502-511Crossref PubMed Scopus (61) Google Scholar). The B domains, in contrast, neither interact with each other nor strongly interact with the IIA domains to which they are, however, covalently linked via 60-Å long alanine-proline-rich linker (Fig. 1, A andB). Phosphoryl groups can be transferred from IIA to IIB on the same subunit (cis), on different subunits (trans), or both. Our results indicate that cisand trans pathways are of comparable efficiency. Wild-type IIABMan with four sites and four pathways (twocis and two trans) per dimer has the highest specific activity. The heterodimer between wild-type and H10C or H175C with three active sites and only two pathways (one cis and one trans) has 50% specific activity. The active monomer in this heterodimer retains its full activity. Heterodimers with only one functional A and one functional B domain and only one pathway (cis or trans) retain between 30 and 40% activity. Taken together, this indicates that both cis andtrans pathways contribute to the maximal phosphotransferase activity of IIABMan. A phosphoryl group on a IIA domain can be transferred either to the IIB domain on the same or on the second subunit in the dimer, and interruption of one of the two pathways results in a reduction of the activity by 20% to 30% of the control. The results confirm our previous observation of interallelic complementation (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar) and similar observations by others (26Chen Q. Amster-Choder O. Biochemistry. 1998; 37: 8714-8723Crossref PubMed Scopus (16) Google Scholar, 31Boer H. ten Hoeve-Duurkens R.H. Robillard G.T. Biochemistry. 1996; 35: 12901-12908Crossref PubMed Scopus (25) Google Scholar, 46Van Weeghel R.P. Meyer G. Pas H.H. Keck W. Robillard G.T. Biochemistry. 1991; 30: 9478-9485Crossref PubMed Scopus (24) Google Scholar,47Van Weeghel R.P. van der Koek Y.Y. Pas H.H. Elferink M.K. Robillard G.T. Biochemistry. 1991; 30: 1768-1773Crossref PubMed Scopus (34) Google Scholar). But in the case of IIABMan, the interpretation has changed. The weak complementation was because of phosphoryl transfer between randomly colliding homodimers. IIABMan monomers do not exchange, as evident from the structure of the IIA dimer (19Nunn R.S. Markovic-Housley Z. Génovésio-Taverne J.C. Flükiger K. Rizkallah P.J. Jansonius J.N. Schirmer T. Erni B. J. Mol. Biol. 1996; 259: 502-511Crossref PubMed Scopus (61) Google Scholar). However, the long linker (Fig. 1 B) allows sterically unconstrained interaction between IIA and IIB domains on different dimers. The linker allows the IIA dimer to dock on the IIBMan·IICMan·IIDMan complex in either of two orientations (Fig. 1 C). The cisorientation is presented in Fig. 1 A. A IIABMan mutant with His-86 on the IIA domain replaced by Asn was described to have the same properties as H175C mutant with an inactive IIB domain (34Stolz B. Huber M. Markovic-Housley Z. Erni B. J. Biol. Chem. 1993; 268: 27094-27099Abstract Full Text PDF PubMed Google Scholar, 36Erni B. Zanolari B. Graff P. Kocher H.P. J. Biol. Chem. 1989; 264: 18733-18741Abstract Full Text PDF PubMed Google Scholar). However, the x-ray structure of IIA showed that His-86 is in a surface-exposed loop and far from the active site. In addition, His-86 is not conserved in any of the homologous proteins (see below). Both observations make His-86 an unlikely target for mutations with a strong phenotype. Resequencing showed the supposed H86N mutation to contain the H175C mutation. We conclude that the H86N mutant is neutral and that vectors must have been exchanged by mistake. Bacillus subtilis, Klebsiella pneumoniae, Vibrio furnissii,and Lactobacillus casei express transporters homologous to the mannose transporter of E. coli except that IIA and IIB are expressed as separate proteins subunits and not as two domains connected by an alanine-proline-rich linker (48Veyrat A. Monedero V. Perez-Martinez G. Microbiology (U. K.). 1994; 140: 1141-1149Crossref PubMed Scopus (55) Google Scholar, 49Wehmeier U.F. Lengeler J.W. Biochim. Biophys. Acta. 1994; 1208: 348-351Crossref PubMed Scopus (36) Google Scholar, 50Bouma C.L. Roseman S. J. Biol. Chem. 1996; 271: 33457-33467Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar, 51Martin-Verstraete I. Débarbouillé M. Klier A. Rapoport G. J. Mol. Biol. 1990; 214: 657-671Crossref PubMed Scopus (114) Google Scholar). Using the Basic Local Alignment Search Tool (BLAST) program, IIAB homologs with alanine-proline-rich or Q-linkers (52Wootton J.C. Drummond M.H. Protein Eng. 1989; 2: 535-543Crossref PubMed Scopus (163) Google Scholar, 53Perham R.N. Biochemistry. 1991; 30: 8501-8512Crossref PubMed Scopus (366) Google Scholar) were found in bacterial genomes 2http://www.ncbi.nlm.nih.gov/BLAST/unfinishedgenome.html. (complete and in progress) of: Yersinia pestis, Actinobacillus actinomycetemcomitans, Enterococcus faecalis, Clostridium acetobutylicum, Streptococcus pneumoniae, and Streptococcus pyogenes. Why two forms of IIAB units and what function if any does the linker have? All things being equal, binding of the IIABMan dimer to the IICMan·IIDMan complex must be much stronger than binding of a monomeric IIB subunit because the dimer forms two contacts per molecule, whereas a IIB monomer forms only one (54Fersht A.R. Enzyme Structure and Mechanism. W. H. Freeman and Company, New York1985: 308Google Scholar). Although not covalent in the chemical sense, binding might become very strong, and IIAB remain membrane-bound for most of the time. Untying of IIB from the IIC·IID complex is necessary whenever IIB has a regulatory function and must diffuse to other targets. For example, monomeric IIBLev of B. subtilis is not only a subunit of the fructose transport complex, but it also can phosphorylate and thereby inactivate the transcriptional activator LevR (16Martin-Verstraete I. Débarbouillé M. Klier A. Rapoport G. J. Mol. Biol. 1994; 241: 178-192Crossref PubMed Scopus (51) Google Scholar, 17Débarbouillé M. Martin-Verstraete I. Klier A. Rapoport G. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2212-2216Crossref PubMed Scopus (78) Google Scholar, 18Martin-Verstraete I. Charrier V. Stülke J. Galinier A. Erni B. Rapoport G. Deutscher J. Mol. Microbiol. 1998; 28: 293-303Crossref PubMed Scopus (96) Google Scholar). An analogous situation is observed in E. coli. The transporter for Glc and GlcNAc (IICBGlc·IIAGlc and IICBAGlcNAc) are homologous, but whereas IICBAGlcNAc is a three-domain protein, IIAGlc and IICBGlc are independent subunits. IIAGlc plays a pivotal role in regulation of catabolite repression and inducer inclusion, and it has been shown to interact with glycerol kinase, the transporters for lactose and maltose, and adenylate cyclase (5Peterkofsky A. Seok Y.-J. Amin N. Thapar R. Lee S.Y. Klevit R.E. Waygood E.B. Anderson J.W. Gruschus J. Huq H. Gollop N. Biochemistry. 1995; 34: 8950-8959Crossref PubMed Scopus (22) Google Scholar, 6Peterkofsky A. Reizer A. Reizer J. Gollop N. Zhu P.P. Amin N. Prog. Nucleic Acid Res. Mol. Biol. 1993; 44: 31-65Crossref PubMed Scopus (46) Google Scholar, 7Hurley J.H. Faber H.R. Worthylake D. Meadow N.D. Roseman S. Pettigrew D.W. Remington S.J. Science. 1993; 259: 673-677Crossref PubMed Scopus (207) Google Scholar, 8Van der Vlag J. Van Dam K. Postma P.W. J. Bacteriol. 1994; 176: 3518-3526Crossref PubMed Google Scholar, 9Van der Vlag J. Postma P.W. Mol. Gen. Genet. 1995; 248: 236-241Crossref PubMed Scopus (12) Google Scholar, 10Osumi T. Saier Jr., M.H. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 1457-1461Crossref PubMed Scopus (75) Google Scholar, 11Saier M.H. Novotny M.J. Comeau-Fuhrman D. Osumi T. Desai J.D. J. Bacteriol. 1983; 155: 1351-1357Crossref PubMed Google Scholar, 12Dean D.A. Reizer J. Nikaido H. Saier Jr., M.H. J. Biol. Chem. 1990; 265: 21005-21010Abstract Full Text PDF PubMed Google Scholar). These interactions with soluble and membrane-bound target proteins require that IIAGlc can freely diffuse through the cell. The structural stability of the IIAB dimers and their mechanism of phosphoryl transfer might be unique among the different families of dimeric PTS transporters. Nevertheless, it indicates that interactions between different subunits within a dimer (first order reaction) as well as interactions between different dimers (second order reaction) have to be taken into consideration when weak interallelic complementation is observed. The ease with which stable heterodimers can be generated by reversible unfolding will facilitate the characterization by fluorescence energy transfer of domain motions that might occur during phosphorylation and transport of mannose. We thank S. Mukhija (ARPIDA AG, Münchenstein) for the gift of [33P]PEP, S. D. Snyder (Protein Solutions Inc., Charlottesville) for determining the diffusion coefficient by dynamic light scattering, and S. Schauder for the help with preparing Fig. 1.