1 Transient-state Kinetics of Enzyme IICB , a Glucose Transporter of the Glc Phosphoenolpyruvate Phosphotransferase System of Escherichia coli: EQUILIBRIUM AND SECOND ORDER RATE CONSTANTS FOR THE GLUCOSE BINDING AND PHOSPHOTRANSFER REACTIONS Norman D. Meadow, Regina S. Savtchenko, Azin Nezami*, and Saul Roseman From the Department of Biology, The Johns Hopkins University, Baltimore, MD 21218 Running title: Kinetics of IICB from E. coli by transient-state methods Glc Address correspondence to: Dr. Saul Roseman, Department of Biology, The Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218. E-mail: [email protected]*Current address: Dana Farber Cancer Institute, Harvard Medical School, Smith 1040, 44 Binney Street, Boston, MA 02115 Acknowledgements:–We wish to thank Drs. Dimitri Toptygin and Ludwig Brand for advice on computer modeling of kinetic data that proved essential to this work; Dr. Bernhard Erni generously provided materials mentioned in the text. We also thank Matthew Ortman, Joshua Baumfeld and Ling-Mei Chen for their assistance with various phases of the work. http://www.jbc.org/cgi/doi/10.1074/jbc.M501440200 The latest version is at JBC Papers in Press. Published on October 4, 2005 as Manuscript M501440200 Copyright 2005 by The American Society for Biochemistry and Molecular Biology, Inc. by guest on June 20, 2018 http://www.jbc.org/ Downloaded from
35
Embed
Transient-state Kinetics of Enzyme IICB enolpyruvate ... · Transient-state Kinetics of Enzyme IICBGlc, a Glucose Transporter of the Phosphoenolpyruvate Phosphotransferase System
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Transient-state Kinetics of Enzyme IICB , a Glucose Transporter of theGlc
Phosphoenolpyruvate Phosphotransferase System of Escherichia coli: EQUILIBRIUMAND SECOND ORDER RATE CONSTANTS FOR THE GLUCOSE BINDING AND
PHOSPHOTRANSFER REACTIONS
Norman D. Meadow, Regina S. Savtchenko, Azin Nezami*, and Saul Roseman
From the Department of Biology, The Johns Hopkins University, Baltimore, MD 21218Running title: Kinetics of IICB from E. coli by transient-state methodsGlc
Address correspondence to: Dr. Saul Roseman, Department of Biology, The Johns Hopkins University,3400 North Charles Street, Baltimore, MD 21218. E-mail: [email protected]
*Current address: Dana Farber Cancer Institute, Harvard Medical School, Smith 1040, 44 Binney Street,Boston, MA 02115
Acknowledgements:–We wish to thank Drs. Dimitri Toptygin and Ludwig Brand for advice oncomputer modeling of kinetic data that proved essential to this work; Dr. Bernhard Ernigenerously provided materials mentioned in the text. We also thank Matthew Ortman, JoshuaBaumfeld and Ling-Mei Chen for their assistance with various phases of the work.
http://www.jbc.org/cgi/doi/10.1074/jbc.M501440200The latest version is at JBC Papers in Press. Published on October 4, 2005 as Manuscript M501440200
Copyright 2005 by The American Society for Biochemistry and Molecular Biology, Inc.
During translocation across thecytoplasmic membrane of Escherichia coli,glucose is phosphorylated by phospho-IIAGlc
and Enzyme IICB , the last two proteins inGlc
the phosphotransfer sequence of thephosphoenolpyruvate:glucosephosphotransferase system (PTS). Transient-state (rapid quench) methods were used todetermine the second order rate constants thatdescribe the phosphotransfer reactions(phospho-IIA to IICB to Glc) and also theGlc Glc
second order rate constants for the transferfrom phospho-IIA to molecularly cloned,Glc
IIB , the soluble, cytoplasmic domain ofGlc
IICB . The rate constants for the forwardGlc
and reverse phosphotransfer reactions betweenIIA and IICB were 3.9 x 10 M s and 0.31Glc Glc 6 -1 -1
x 10 M s respectively, and for the6 -1 -1
physiologically irreversible reaction between[P]IICB and Glc was 3.2 x 10 M s . FromGlc 6 -1 -1
the rate constants, the equilibrium constantsfor the transfer of the phospho-group from His90 of [P]IIA to the phosphorylation site CysGlc
of IIB or IICB were found to be 3.5 and 12Glc Glc
respectively. These equilibrium constantssignify that the thiophospho-group in theseproteins has a high phosphotransfer potential,similar to that of the phosphohistidinyl PTSproteins. In these studies, preparations ofIICB were invariably found to containGlc
Dendogenous, firmly bound Glc (estimated K '
.10 M). The bound Glc was kinetically-7
competent, and was rapidly phosphorylated,indicating that IICB has a random order, bi-Glc
bi, substituted enzyme mechanism. Theequilibrium constant for the binding of Glc wasdeduced from differences in the statisticalgoodness of fit of the phosphotransfer data tothe kinetic model.
The bacterial phosphoenolpyruvate:glycosephosphotransferase system (PTS ) comprises1
dozens of cytoplasmic and membrane proteins,most of which are the sugar-specific componentsof the system (for reviews see (1-4)). The PTShas several important functions in eubacterial cell
physiology in addition to its major role in sugartransport, especially in Gram positive bacteria (5).
When the PTS catalyzes sugar transport, thesugar is phosphorylated as it crosses themembrane, and this process requires from 3 to 6proteins, depending on the sugar. One example isthe Glc transport system in enteric bacteria, shownin Fig. 1.
The first two proteins, Enzyme I and HPr, arethe so-called general proteins of the system,meaning that they are not sugar specific. Thesecond pair of proteins, IIA and IICB are theGlc Glc
sugar specific components. In other cases, thesugar specific proteins vary from one to fourseparately encoded proteins. We emphasize thatall of the phosphotransfer reactions, except thelast step (transfer to the sugar), are physiologicallyreversible. A second important point is that thephosphorylation site residues in the PTS proteinsare generally His, but in a number of themembrane proteins, the phosphorylation site canbe a Cys residue. The phosphoryl group istransferred from PEP through a chain of Hisproteins to the membrane Enzyme II, where thephosphorylation site can be a His or a Cys (as it isin IICB ), and finally to the sugar. In IICB , theGlc Glc
B domain extends into the cytoplasm and containsthe phosphorylation site Cys.
Our long term goal is to be able to predict thekinetics of Glc uptake by intact cells, and for thispurpose it is necessary to obtain the rate constantsfor each of the reactions in Fig. 1. We establishedthe basic methodology by developing a rapidquench method for determining the rate constantsfor the reversible phosphotransfer from phospho-HPr to IIA (6), and the technique has now beenGlc
applied to the other reactions shown in Fig. 1. The B and C domains of IICB have beenGlc
genetically cloned by Buhr, et al (7) and kineticresults with the B domain, called IIB , and withGlc
intact IICB are reported below.Glc
We address two other questions. First, severalof the integral membrane sugar transporters, whenunphosphorylated, have been shown to bind theirsugar substrates (3,4) and this bound sugar can bephosphorylated, but the rate has never beenmeasured. In the present case, it was essential to
characterize and quantify the bound sugar, Glc,for accurate determination of the kinetic constants. Second, we have obtained, for the first time, theapparent equilibrium constants for thephosphotransfer reactions between the His residuein IIA and the Cys residue in domain IIB, fromGlc
which the phosphotransfer potential of a phospho-S-Cys protein can be derived. While the phospho-S-Cys structure is relatively rare, it is also foundin the protein-tyrosine-phosphatases of eukaryoticcells (8).
Experimental Procedures
Materials–All materials used for the assay ofGlc are given in Supplemental Data. L-"-dioleoylphosphatidyl-DL-glycerol (Sigma P-9664); N-lauroylsarcosine (Sigma L-5777);adenosine 5'-[(- P] triphosphate (110 TBq/m32
England Life Science Products); all buffer saltsand other reagents were of purity typical forresearch reagents from standard commercialsources. The pH of all buffers is reported at thetemperature and concentration at which they wereused.
Assays for Glc and for the binding of Glc to E.coli membranes–Supplemental Data describes thefour methods used to characterize the Glc found“contaminating” both highly purified IICB andGlc
membrane suspensions, and for measuring thebinding of Glc to these membranes. These aredescribed in detail in Supplemental Data. For theroutine assay of Glc in membrane suspensionsbeing prepared for rapid quench experiments, twomethods were used. The first was the hexokinaseassay using ( P-ATP described in Supplemental32
Data. The second were the actual experimentalresults of the rapid quench assay when conductedwith a stoichiometric excess of the phospho donor,[ P]IIA . The quantity of Glc measured by the32 Glc
two methods was similar.Bacterial strains, plasmids and growth
media–All strains were grown on Luria-Bertanibroth containing the required antibiotics andinducers. E. coli strain BL21(DE3) (F ompT!
B B BhsdS (r m ) gal dcm (DE3)) (Novagen) was! !
used as the host for plasmid pJBH which encodesthe IIB -6His gene fragment (a generous gift ofGlc
Prof. B. Erni, University of Basel), and was grownas described (7). E. coli strain ZSC112)G()ptsG::cat manZ glk) from which thechromosomal gene for IICB was deleted (7) wasGlc
used as the host for plasmid pTGH11 (encodingIICB -6His) (both also gifts from B. Erni) for theGlc
preparation of membranes and for the purificationof IICB -6His. E. coli strain ZSC112)G wasGlc
also used as the host for plasmid pCB30 (encodingwild type IICB ) (9) for the preparation ofGlc
membranes. Preparation of washed membranes– Cultures of
6003 l at an O.D. of 0.6 to 1.0, were harvested bycentrifugation, washed twice with 50 mM Tris/Cl!
buffer (pH 7.5) containing 150 mM KCl, andfinally resuspended in this solution in a volumeequal to 1% of that of the original culture. Todischarge phospho-groups from the PTS proteins,10 mM methyl "-glucoside and 10 mM KF wereadded to the final cell suspension which wasincubated at room temperature for 15 min (10) andthen frozen at !80 C. After thawing, the cellso
were homogenized by two passages through aFrench Pressure cell (Spectronic Instruments) at110 MPa, the homogenate centrifuged at 5,600 × gfor 15 min, and the supernatant centrifuged at370,000 × g for 2 h. The particulate fraction wasresuspended in 10 mM Tris/glycine buffer (pH8.9), 1 mM DTT (11) to the original volume of thehomogenate and centrifuged again; the high speedcentrifugation was repeated a third time unlessotherwise indicated. The membranes wereresuspended in 1 ml of the same buffer per g oforiginal wet weight of cells, and frozen in aliquotsat !80 C.o
The final membrane preparations containedvariable quantities of Glc, and its concentrationincreased slowly upon storage or incubation andduring the 1 to 3 h required to complete a rapidquench experiment (Supplemental Data). Whilethe source of Glc was not identified, a procedurewas developed that eliminated the increase in Glcconcentration but did not significantly affect theactivity of the enzyme, reducing it by less than
20% as determined by PTS sugar phosphorylationassays. The method was to incubate thepreparation in a dialysis cassette (Pierce Slide-A-Lyzer) vs. Tris/glycine buffer, pH 8.9, 1 mM DTTor 5 mM $-mercaptoethanol, and 0.05% sodiumazide at 37 C for 15 h and continued for severalo
hours at 4 C.o
Kinetic Properties of Endogenous Glc inMembrane Preparations–The concentration ofendogenous Glc after treating the membranepreparations as described above was variable, butwas always in the same range as the concentrationof IICB . Modeling the kinetics of Reactions V,Glc
VI, and VII (Fig. 2, Scheme I) required anunderstanding of the kinetic properties of this poolof endogenous Glc, as well as decisions abouthow to model the kinetics of Reactions V(t), VI(t),and VII(t) when [ H]Glc was added to the syringe3
containing the solution of [ P]IIA . These32 Glc
methods are described in Supplemental Data onKinetics.
General Methods--Assays for the PTSproteins by measuring the rate of sugarphosphorylation were performed as described(9,12). The method of Bradford (13) was used forsoluble proteins; the reagent was from Bio-RadLaboratories (Hercules, CA). The modification ofthe method of Lowry by Markwell, et al. (14),calibrated as previously described (6) was used forIIA and for membrane proteins. Glc
The concentration of HPr or IIA wasGlc
measured by using the LDH coupled assay (12)(using homogenous auxiliary proteins) whichmeasures the quantity of protein that can accept aphospho-group from PEP. The concentrations ofHPr or IIA estimated from the specific activityGlc
of the [ P]PEP agreed with the results of the LDH32
assay within 5%. IIB was quantified by three methods: a)Glc
Solutions of the protein, purified to apparenthomogeneity, were thoroughly dialyzed andanalyzed for nitrogen by the method of Jaenicke(15). b) Quantitative phosphorylation of theprotein using an excess of [ P]PEP (of accurately32
known specific activity) and catalytic quantities ofEnzyme I, HPr, and IIA . c) The LDH assayGlc
(12). When all three methods were applied to the
same sample, the results agreed to within 10%.IICB , both purified and in membranes, wasGlc
quantified by two methods: a) The sample wasassayed for its activity in PEP driven sugarphosphorylation, using a range of concentrationsof IIA . These data were used to calculate theGlc
maxV which was converted to the concentration ofthe IICB protein by using the specific activity ofGlc
the homogenous enzyme (97 :mol sugar-P/min/mg IICB ) (9). b) By quantitativelyGlc
labeling the protein in membrane suspensions with[ P]PEP of accurately known specific activity (132
to 3 TBq/mol) (6) as follows: A mixture of 50 mMpotassium phosphate buffer (pH 7.5), 5 mM
2MgCl , 25 nmol [ P]PEP, 10 pmol Enzyme I, 732
pmol HPr, 15 pmol IIA , and approximately 150Glc
pmol of IICB in a volume of 100 :l wasGlc
incubated at room temperature for 15 min. It wasthen quenched with 50 :l of 0.6 M KOH, andanalyzed by gel filtration chromatography by thesame methods used for rapid quench samples (seebelow). When applied to three membranepreparations the two methods agreed to within15%. Membranes prepared from ZSC112)Gexhibited insignificant activity in the PEP drivensugar phosphorylation assay, as well asinsignificant labeling with [ P]PEP.32
Purification of proteins–Enzyme I, HPr, andIIA were separately overproduced in cellsGlc
carrying the relevant plasmids. The proteins werepurified by the methods used previously (6). IIB -6His was purified by the method of Buhr, et.Glc
al. (7), except that a Superose 12 HR 10/30column (Amersham Biosciences) was substitutedfor Sephadex G75. The final preparations wereapparently homogeneous as judged by SDS-PAGE. IICB -6His was purified usingGlc
modifications of the method of Waeber, et al.(11);a Superose 12 HR 10/30 column was used, butneither Glc nor methyl "-D-glucopyranoside wereadded to solutions. Before bring employed ineither rapid quench experiments or in the PTSsugar phosphorylation assay the IICB wasGlc
activated, except as noted, by mixing with anequal volume of a solution of lipid/detergentmixed micelles (5 mg/ml dioleoylphosphatidylglycerol, 1 mg/ml sodium lauroyl sarcosinate) as
described in Bouma, et al. (9)Synthesis of [ P]PEP–The enzymatic32
synthesis (16) was performed with themodifications described previously (6). Thepurification by anion exchange chromatographywas further modified by the substitution of KCl
2 3for triethylamine/H CO as the eluant. We suspectthat triethylamine or a contaminant in itoccasionally interferes with the phosphotransferreactions (data not shown). To stop the enzymaticreaction, 25 :l of 20% (v/v) Norite A suspensionwas added to the reaction mixture. The Bio-RadAG1-x8 column was equilibrated with 10 mMBisTris/Cl, pH 6.0, which was also a componentof all the eluant solutions. The reaction mixture(including the Norite A) was placed onto thecolumn which was then washed with 5 ml of the
ibuffer, and the P eluted with 5 ml of 0.1M KCl.The PEP eluted with 0.3 M KCl; 0.5 ml fractionswere collected in tubes containing 40 :l of 0.2 M
3 4 2 4Na PO , 0.2 M K PO and counted by liquidscintillation counting. The peak concentration of[ P]PEP appeared in the 5 to 10 fractions. 32 th th
Prepared by this method and stored at 4 C, theo
PEP had a spontaneous rate of hydrolysis at 4 C ofo
0.4% per day.Preparation of [ P]HPr, [ P]IIA ,32 32 Glc
[ P]IIA were prepared as described previously32 Glc
(6). [ P]IIB was prepared under the conditions32 Glc
used for the preparation of [ P]IIA , using a mol32 Glc
ratio of HPr:IIA :IIB = 1:24:24. A SuperdexGlc Glc
75 HR 10/30 column (Amersham Biosciences)was employed to fractionate the proteins.
[ P]IICB was prepared from purified32 Glc
protein which was sonicated for 30 s in a bath-type sonicator in the presence of a 10-fold molarexcess of [ P]IIA , and incubated for 15 min32 Glc
(with no added phospholipid). The proteins wereseparated on a Superdex 200 HR 10/30 column.
The columns used to fractionate thephosphorylation mixtures were equilibrated with20 mM carbonate/bicarbonate (pH 9.5) buffercontaining 1 mg/ml BSA. The specific activity ofall four phosphoproteins ranged from 10 to 40TBq/mol, with emphasis on the accuracy of thespecific activity of the [ P]PEP (6). . 32
Stability of [ P]IIB and [ P]IICB under32 Glc 32 Glc
quench conditions–[ P]IIB and [ P]IICB32 Glc 32 Glc
were isolated as described and stored at !80 C ino
20 mM carbonate/bicarbonate buffer (pH 9.5). Tests of the effect of pH on the rate of hydrolysisof the phospho-group were made by diluting 10 :lof 0.5 :M [ P]IIB solution into 3.5 ml of32 Glc
buffer, or 100 :l of 0.33 :M [ P]IICB into 4.032 Glc
ml of buffer. All equipment and vessels were pre-treated with BSA to minimize adsorption ofprotein. The buffers were: pH 2, 0.1M HCl/KCl;pH 3.8, 50 mM citric acid/Na citrate; pH 6.0 and
2 4 2 4pH 8.1, 50 mM KH PO /K HPO ; pH 10.1, 50
2 3 3 3 4mM Na CO /NaHCO ; pH 12, 25 mM Na PO ; pH13, 0.1 M NaOH; pH 14, 1 M NaOH, and pH14.3, 2 M NaOH (the latter three solutions wereprepared from fresh, commercial 2 M NaOHstandard solution). The mixtures were incubatedat 23 C for 5 min to 4 h, filtered through 23 mmo
diameter PVDF transfer membranes (Millipore) ina vacuum apparatus that allowed collection of thefiltrate; both the filter and the filtrate werecounted to ensure quantitative recovery of the
iradioactivity. Tests of the membrane with [ P]P32
showed that the background was negligible, andthat washing of the filter was not required. Othercontrols showed that protein adsorption to themembranes was quantitative.
The high rate of hydrolysis above pH 13(Results) made the use of a quench solutioncontaining 1 M KOH unsuitable for thephosphotransfer measurements involving IIB ,Glc
especially since heating of the quenched reactionis required to fully denature [ P]IIA (6). The32 Glc
conditions for quenching that were developed formaintaining the phospho-S-Cys bond intact were:0.1 M KOH, 3 M urea and heating for 5 min at55 C. These conditions yield a level of hydrolysiso
sufficiently low (less than 1% per min) to allowpreparation (with careful timing) of the quenchedsamples for separation by gel filtration (where therate of hydrolysis is negligible).
We unexpectedly found that IICB is rapidlyGlc
fragmented when heated under the conditionsdeveloped for the chromatography of [ P]IIB . 32 Glc
radioactivity in [ P]IICB or [ P]IICB -6His32 Glc 32 Glc
appears in fractions containing 10 to 20 kDamolecules. These fractions also contain the sixhistidine residues from IICB -6His as shown byGlc
using dot blots treated with anti-His antibodies(data not shown); corresponding fractions fromcontrol membranes do not bind the anti-Hisantibodies. Urea (3 M) enhances the rate offragmentation by about 30%. These resultssuggest that a peptide bond somewhere in thelinker region between the B and C domains is verylabile at high pH. Optimal conditions for attainingrapid quenching while minimizing proteincleavage and hydrolysis of the phospho-groupwere found to be 0.2 M KOH (final concentrationin the quenched reaction) with no heating beforeinjection onto the gel filtration column. Carefultiming between thawing the quenched samples andinjection produced a reproducible 13 ± 3%fragmentation, with acceptable speed ofquenching of the reaction. The raw data for theconcentration of the phosphoproteins in quenchedreactions was therefore corrected by 13%.
Rapid quench assays–The present studyemployed the rapid quench apparatus usedpreviously, and all the details for its set-up werethe same (6). Stock solutions of [ P]-labeled32
proteins were diluted with the same solution usedto fractionate the phosphorylation mixture at thetime of its preparation (see above). Stocksolutions of IIB or membrane suspensions wereGlc
diluted with 50 mM phosphate buffer (pH 7.5),0.5 mM EDTA, 0.5 mM DTT, and 1 mg/ml BSA. The phosphate buffer was pH 7.5 (rather than pH6.5 (6)) to correlate with work on the kinetics ofEnzymes II published by the time the present workwas started (e.g.,(17,18)). The rate ofphosphotransfer between HPr and IIA is notGlc
significantly affected by a change in pH from 6.5to 7.5 ((6) and unpublished data). Anothersignificant modification, described above, was ofthe conditions used for quenching the reactions. Preparation of the solutions for rapid quenchexperiments required large dilutions from stocksolutions and a change from the frozen state toambient temperature (~23 C), at which allo
experiments were performed. The diluted
solutions were therefore preincubated for an hourat ambient temperature before the experiment wasstarted.
When the phosphotransfer between IIA andGlc
IIB was studied, a Superdex 75 HR 10/30Glc
column (6) was used to separate the proteins in thequenched reactions. When the phosphotransferreactions between IIA and IICB were studied,Glc Glc
the column was a Superose 12 HR 10/30(Amersham Biosciences). This column cannotresolve Glc-6-P from inorganic phosphate, whichis always present because of hydrolysis of thephospho-donor protein during storage followingits preparation. For this purpose, a separatealiquot of each quenched reaction mixture waschromatographed on a Superdex Peptide HR10/30 column (Amersham Biosciences) that was
3 4 2 4equilibrated with 35 mM Na PO , 0.1 M Na SO .When [ H]Glc was used in rapid quench3
experiments, the [ H]Glc-6-P was isolated by3
anion exchange chromatography using amodification of the method used for PTS sugarphosphorylation assays (12). Aliquots (100 :l) ofthe quenched reactions were diluted with 900 :lof water and the pH was reduced to between 9.5and 10 by the addition of 10 :l of 0.5 M aceticacid. Inorganic phosphate and Glc-6-P (2 :Meach) were added as carriers. These samples wereapplied to 0.2 ml bed volume columns of Bio-RadAG-1 X8 (200-400 mesh) in the acetate form,washed with water, and the [ H, P]Glc-6-P3 32
eluted with 1 M NaCl, and counted by liquidscintillation counting using a double-isotopequench correction program.
Methods used to model experimental data onthe rate of phospho-group transfer–The goal ofthese experiments was to determine the rate
XXXconstants (k ) for each of the first and secondorder chemical reactions shown in Fig. 2. Themathematical model for each reaction is thedifferential equation defined by the chemicalequation. The numerical integration program,Kinsim (19) as modified by Anderson et al. (20)was used to manually fit the mathematical modelsto the experimental data. When experimental datamet the criteria for non-linear least squares fitting(21), the Fitsim module of Kinsim was used.
XXXconstants (k ).The convention used for numbering the
reactions in Scheme I (Fig. 2) is adopted fromRohwer, et al. (22), in which theautophosphorylation of Enzyme I from PEP iscalled Reaction I. By this convention, thephosphotransfer reaction between HPr and IIAGlc
(6) is Reaction III (see Fig. 1).
Results
Introduction–In what follows, Reactions areidentified by the Roman Numerals assigned inScheme I (Fig. 2). Representative progress curvesare shown in Figs. 3, 5-7. It is important toemphasize that in these Figures, each panelrepresents one experiment. The rate constants ofthe reactions were estimated from theexperimental progress curves by numericalintegration (Methods). The data from studying thephosphotransfer reaction between IIA and IIBGlc Glc
(with experiments using either [ P]IIA or32 Glc
[ P]IIB ) were fitted to Reaction IVa. The data32 Glc
from the phosphotransfer reactions from [P]IIAGlc
to Glc via IICB were fitted to Reactions IV, V,Glc
VI, and VII. When [ H]Glc was added to an3
experiment, Reactions Vt, VIt, and VIIt wereincluded in the model. Table I presents asummary of all of the rate constants that we reportin this work; some of the rows in the Table presentresults from global analyses of replicateexperiments, not the data from single experiments.
Rates of phosphotransfer between IIA andGlc
IIB -6His–In the E. coli Glc specific PTS, theGlc
last protein-protein phosphotransfer step in theupper pathway (Fig. 2) is from His90 in IIA toGlc
Cys421 in IICB (Reaction IV in Figs. 1 and 2). Glc
The subsequent, and final, reaction (V) is thephosphotransfer to Glc. As described below,kinetic measurements with both purified andmembrane bound IICB were complicated by theGlc
presence of endogenous Glc, a problem that couldbe avoided by using IIB as the phosphoGlc
acceptor. IIB is the cytoplasmic domain of theGlc
integral membrane protein; the cloned fragment(10,739 Da) comprises residues 1-4 of the amino
terminus of IICB , followed by residues 391-476,Glc
and terminated by a 6 His cartridge (7). Molecularly cloned IIB -6His is a soluble andGlc
readily purified protein containing thephosphorylation site (Cys421 in IICB ), but notGlc
the Glc binding site (7). Its phosphotransferreaction is designated Reaction IVa in Scheme I.
A typical progress curve and the estimatedrate constants for the reversible transfer of aphospho-group from [ P]IIA to IIB -6His are32 Glc Glc
shown in Fig. 3. The rate constants obtained froma global analysis of the data from fourexperiments, three using [ P]IIA and one using32 Glc
[ P]IIB -6His are given in Table I, Row 1. 32 Glc
There was good agreement between the constantsobtained by starting the reaction from eitherdirection. This implies that there are nosignificant concentrations of intermediatecomplexes between the two reacting proteins priorto the last step, transfer of the phosphoryl group tothe acceptor and separation of the proteins to yieldthe products. These rate constants yield anequilibrium constant of 3.5 for Reaction IVa,indicating that the thiophosphate linkage has avery high phosphate transfer potential, close tothat of phospho-IIA .Glc
Stability of [ P]IIB and [ P]IICB under32 Glc 32 Glc
instability of the phospho-group at pH 14, the pHof the quench solution developed for the phospho-His proteins (6). The stability of both phosphoproteins was therefore studied as a function of pH;the results, from pH 2 to 14, are shown in Fig. 4. Between pH 2 and pH 12 the rate constants for thehydrolysis of the phospho-group are similar inmagnitude to those published for the hydrolysis ofbutyl thiophosphate (23), cysteamine S-phosphoric acid (24), and the thiophosphopeptidesderived from IICB (25) and from II (26). Glc Mtl
Although the rate constants for hydrolysis of butylthiophosphate and cysteamine S-phosphoric acidexhibit bell shaped curves in the pH range 1-6,this was not observed with the thiophospho estersof any of the PTS proteins.
There is, however, a more importantdifference in the properties of the
thiophosphoproteins above pH 12. Butylthiophosphate is very stable in the pH range10-14, whereas the thiophospho-PTS proteins arenot. Above pH 12 the behavior of the twothiophospho-PTS proteins (Fig.4) resembles thebehavior of the mixed anhydride, $-aspartylphosphate (27). Perhaps the increase in the rate ofhydrolysis of the phospho-group above pH 13results from phospho-group migration fromCys421 to the nearby Asp419 when the protein isdenatured in strong alkali. In the previous studiesof phosphopeptides from Enzymes II, the highestpH tested was 12 (25) or 13 (26).
As a result of these findings, and also theobservation that IICB is rapidly fragmentedGlc
under highly alkaline conditions, we developedthe conditions for quenching given inExperimental Procedures.
Kinetic competence of Glc bound to IICB :Glc
the relevance of Reaction VII–We show inSupplemental Data on Glc that all of ourpreparations of [ P]IICB were “contaminated”32 Glc
with Glc that binds to the enzyme. The bindingappears to be rather tight, and the free and boundglucose are in equilibrium. For our analyses andsimulations of the kinetics of phosphotransfer, itwas essential to determine whether or not the Glcbound to the IICB is kinetically competent.Glc
The rapid quench experiment shown in Fig. 5was designed to determine this. The experimentmeasured the rate of the phosphotransfer reactionsfrom [ P]IIA to Glc via IICB . A preparation32 Glc Glc
of wild type membranes was used that was washedonly once, and neither incubated nor dialyzed, sothat the mol ratio of endogenous Glc to IICBGlc
was higher (22:1) than that in the otherexperiments reported here (2:1 to 4:1) The Figureshows only the data on the production of[ P]IICB and Glc-6-[ P], the utilization of32 Glc 32
[ P]IIA is not shown. 32 Glc
For the analysis shown in Fig. 5A it wasassumed that the Glc in IICB CGlc was notGlc
VIIkinetically competent, i. e., rate constant, k wasforced to zero. The result was a poor fit betweenthe theoretical curve and the data points but thiswas the best fit that could we could obtain. If thebound Glc is not kinetically competent, then
[ P]IICB should accumulate before any sugar32 Glc
phosphate is formed. It is clear, however, thatGlc-6-[ P] appeared more rapidly than32
[ P]IICB . Thus, as seen in Fig. 5B, when32 Glc
Reaction VII is assumed to be active and isassigned a non-zero value in the simulation, verygood agreement is obtained between thetheoretical curves and the data points. In all ofour experiments, substantially better theoreticalfits to the data were obtained when kineticallyactive IICB CGlc was included in the model.Glc
Kinetics of phosphorylation of endogenousand exogenous Glc–At the instant of mixing ofIICB with exogenous Glc added to theGlc
[ P]IIA , there are three pools of the sugar:32 Glc
exogenous Glc, Pool 3; free endogenous Glc, Pool2; and bound endogenous Glc (IICB •Glc), PoolGlc
1. The foregoing assumes that the endogenousGlc is all accessible to the IICB and that itGlc
participates in a binding equilibrium with theenzyme. This is the case as shown inSupplemental Data on Glc. Further, as shownabove, the endogenous Glc is kineticallycompetent, but what is its rate of phosphorylationrelative to the exogenous Glc? In other words,how rapidly do the exogenous and endogenousGlc pools equilibrate relative to thephosphotransfer reactions starting with phospho-IIA ?Glc
The experiment shown in Fig. 6 clearly showsthat the bound endogenous Glc is phosphorylatedmore rapidly (Reaction VII) than it equilibrateswith the exogenous pool of Glc (Reaction VI). Inthis experiment the rate of phosphotransferfrom[ P]IIA to Glc via IICB was measured,32 Glc Glc
but only the data on the production of[ H]Glc-6-[ P] and total Glc-6-[ P] are shown. 3 32 32
The experiment was conducted in two parts. Inthe first part [ H]Glc was added to the syringe3
containing the IICB and endogenous Glc. InGlc
other words, the exogenous labeled pool waspermitted to mix and equilibrate for more than 30min with the endogenous Glc before themeasurements were begun. The data points andthe fitted curve for total Glc-6- P and [ H]-32 3
containing the [ P]IIA , and came into contact32 Glc
with the endogenous unlabeled Glc and IICBGlc
only when mixed. There was a clear differencebetween the rates of phosphorylation of theendogenous Glc and the exogenous [ H]Glc for3
about the first 10 s of the progress curve (Fig. 6shows only the first 1.5 s). Complete equilibrationof the two pools took about 10 sec under theconditions used for the experiment shown in Fig.6 , while measurable phosphorylation of Glc fromIICB CGlc is seen at the first time point (25Glc
msec).Phosphorylation of IICB : kinetics of theGlc
complete system–The results described aboveestablish Reactions VI and VII, the lowerpathway in Fig. 2, as an active pathway forphosphorylating Glc. The transient-state kineticsof the upper pathway in Fig. 2 (Reactions IV andV), the pathway most often used to describe thePTS enzymes II, will now be characterized. Therate constants of all of the reactions involving[P]IIA , IICB , and Glc are summarized inGlc Glc
Table I, Rows 2-14; these include both sugarbinding and phosphotransfer reactions.
Rows 2-11 show the rate constants generatedby a series of global analyses of the data from fourexperiments that measured the rate of phospho-group transfer from [ P]IIA to Glc via IICB -32 Glc Glc
6His in which exogenous [ H]Glc was added to3
the [ P]IIA . Each Row shows the effects of32 Glc
Dvarying the K and/or the rate constants for the'
Glc binding reaction, and the interpretation ofthese effects is given in the next section.
Table I shows the data from only one of threeparts that were performed during each experiment. In the other two parts there was either no additionof exogenous Glc, or [ H]Glc was added to the3
membrane suspension, where it had at least 30min to equilibrate with the endogenous pools. The rate constants obtained from these additional8 time courses (data not shown) were in goodagreement with those in the Table. These resultssuggest that exogenous Glc has no effect on thekinetic properties of [ P]IIA .32 Glc
Rows 12-14 in Table I show the rate constantsobtained from three experiments using IICB ,Glc
one (Row 14) with [ H]Glc added to the3
[ P]IIA , and two with no exogenous Glc.32 Glc
Two important tests of the validity of our rateconstants are independence from the concentrationof Glc and IICB , and evidence that the His tagGlc
did not affect the kinetic properties of the enzyme.The effects of varying the concentration
IICB and Glc are seen in the two Panels of Fig.Glc
7 (the concentration of [ P]IIA was similar in32 Glc
both experiments). These experiments used eitherIICB -6His (7A) or IICB (7B); in both [ H]GlcGlc Glc 3
was added to the [ P]IIA . In Panel 7A, the total32 Glc
concentration of Glc was 107 nM and the totalconcentration of IICB -6His was 40 nM, whereasGlc
in Panel 7B, the total Glc concentration was 5 :Mand the total concentration of IICB was 132 nM. Glc
The concentration of [ P]IICB that appears32 Glc
depends on its rate of formation from [ P]IIA32 Glc
vs. the rate of decay by transfer of the phospho-group to Glc. The very different concentrations ofGlc and IICB in the two experiments would beGlc
expected to affect the concentration of [ P]IICB32 Glc
during the time course, and indeed they do. Whenthe total Glc and IICB concentrations were lowGlc
(Panel 7A), about half the total enzyme wasdetected as the phosphoenzyme, while at the highGlc and IICB concentrations (Panel 7B), theGlc
phosphoenzyme was barely detectable. Weemphasize, however, that the rate constants thatproduced the best fit to the data for thephosphorylation of IICB were the same in bothGlc
experiments, and therefore independent of theconcentrations of Glc or IICB .Glc
The second important question was whetherthe 6 His tag attached to IICB affected theGlc
kinetic behavior of the proteins. Fig. 7 shows thatIICB -6His is as catalytically efficient as theGlc
wild type protein, as does the more comprehensivesummary in Table I (cf. Rows 5-8, IICB -6His, toGlc
Rows 12-14, IICB ).Glc
eqK’ and rate constants for the binding of Glcto IICB and their effects on determination of theGlc
rate constants of the phosphotransfer reactions–Itis evident that the kinetic characteristics of thesugar binding reaction (Reaction VI) will affectthe analysis of the phosphotransfer reactions. Thebinding reaction determines the relative
concentrations of free and bound reactants presentat the initiation of the reaction as well as theirbehavior as phosphorylation proceeds. The rateconstants of the sugar binding reaction cannot bedetermined by any known method. Moreover,
eqeven the apparent binding constant (K' ) could notbe determined by flow dialysis for reasonspresented in Supplementary Data on Glc. Semi-
Dquantitative values for K ranging from 1.8 × 10' -7
to 9 × 10 were obtained from the centrifugation-8
experiments (Supplemental Data on Glc). We
Dwere, however, able to deduce likely values for K '
from the statistics of fitting the phosphotransferdata. Analysis of four data sets by the non-linearleast squares method (Table I) showed that:
D(a) The values chosen for K had a strong'
effect on the magnitude of the rate constants of thephosphotransfer steps and, importantly, on thestatistical goodness of the fit. As shown in TableI, the smallest standard error of the rate constantsof the four phosphotransfer reactions was obtained
Dwhen a K of 10 M was used. There was a very' -7
high degree of uncertainty in the phosphotransfer
Drate constants when the K was set at 10 or' -6
10 M; the standard errors were often larger than-8
the constants themselves. (b) In sharp contrast, the rate constants forbinding of Glc to IICB and dissociation of theGlc
VI !VIcomplex (k and k ) could be varied as much assix orders of magnitude without large effect on therate constants for the phosphotransfer reactions. This small effect is consistent with the progresscurves in Fig. 7 which show that the bulk of thereaction was completed in about 100s while theexperiments from the gel filtration columns
½(Supplemental Data on Glc) suggest that the t ofthe binding reaction is about 12 min.
The experiment shown in Fig. 7B permittedindependent estimates of the rate constantsassociated with Reactions VI and VII. In thisexperiment a large proportion of the IICB wasGlc
Dcomplexed even when K was designated at 10' !6
M, and both solutions contained 5 mM Glc, sothat the concentration of Glc did not change onmixing. The fit of the model to the early datapoints was, as expected, determined almostentirely by the rate constant of Reaction VII,
phosphotransfer from [ P]IIA to IICB CGlc,32 Glc Glc
which was estimated by manual fitting as 2.5 ×10 M s (Table I, Row 14), in agreement with6 -1 -1
the other estimates in Rows 5-8. The rateconstants for Reactions VI and VII were alsoestimated by the non-linear least squares method(see Supplemental Data on Kinetics), andcorroborate those shown in Table I.
Discussion
The transient-state kinetic experimentsreported here were intended to determine the rateconstants for the last two steps in thephosphorylation and uptake of Glc by E. coli cells,namely the phosphotransfer reactions fromphospho-IIA to IICB to Glc (Reactions IV andGlc Glc
V in Figs. 1 and 2). Initially, we conducted thesestudies with highly purified preparations ofIICB in lipid/detergent mixtures, but the resultsGlc
were variable, whereas natural membranescontaining active IICB gave reproducible resultsGlc
(Table I). Confirmation of the results obtained with
membranes was obtained with IIB , the soluble,Glc
homogeneous, domain of the intact protein. IIBGlc
that contains the phosphorylation site Cys ofIICB . The cloned IIB domain has kineticGlc Glc
properties that are very similar to those of thewhole protein. Both the forward and backwardrate constants of phosphotransfer between IIAGlc
and IIB are somewhat larger than thoseGlc
involving the intact membrane protein, IICB ,Glc
perhaps expected from the smaller mass of IIBGlc
and the complexity of the membrane preparations.Steady-state measurements of IICB activityGlc
(9,22,28,29) also corroborate the rate constantsreported here for Reactions IV and V. The rate
IV Vconstants k and k are equivalent to the tworespective specificity constants of IICB forGlc
cat m([P]IIA )IV V[P]IIA and Glc, i.e., k = k /K and k =Glc Glc
cat m(Glc)k /K , assuming that the mechanism of theenzyme is ping-pong (30). The agreementbetween our results and the calculated specificityconstants is good. The latter cluster around 4 ×2
cat m([P]IIA )10 M s for k /K , compared to 3.5 × 106 -1 -1 Glc 6
IVM s for k ; and around 3.2 × 10 M s for-1 -1 6 -1 -1
cat m(Glc) Vk /K , compared to 2.5 × 10 M s for k . 6 -1 -1
The advantage of the rate over the specificityconstants is that the rate constants are affected byfewer experimental errors, and no assumptions aremade concerning mechanism.
A computer model has been developed thatcan predict the kinetic behavior of the Glc PTSunder a variety of conditions, both in vivo and invitro (22). The model was based, in part, onpreliminary results from our kinetic experiments. What we consider to be the definitiveexperimental rate constants are presented here. The effects of the new constants on thepredictions of the model will be presented in aseparate report.
From the rate constants we can calculate thecorresponding equilibrium constants for thereactions, phospho-IIA to IIB or IICB (3.5Glc Glc Glc
and 12 respectively). These equilibrium constantsappear to be the first data that permit comparisonof the standard free energies of hydrolysis of twophospho-cysteinyl PTS proteins with those of thephospho-histidinyl PTS proteins. Briefly, thephosphotransfer potential of [P]IICB isGlc
somewhat less than that of [P]IIA , but it is, likeGlc
the other phosphoproteins of the PTS, a “highenergy” phospho-compound. The implications ofthis observation will be elaborated in a futurepublication on the kinetics and thermodynamics ofthe complete pathway of the Glc specific PTS inE. coli. Whether these phosphotransfer potentialsare important for the catalytic action of anotherclass of phospho-S-Cys proteins, the protein-tyrosine-phosphatases of eukaryotic cells (8)remains to be determined. One could speculate,however, that these enzymes may transfer thephospho group to substances in addition to water,i.e., they may act as phosphotransferases as wellas phosphatases.
At the outset of this work, both the highlypurified enzyme and the membranes containingIICB were unexpectedly found to contain aGlc
“contaminant” that was phosphorylated by theenzyme when it was supplemented with[ P]IIA . The “contaminant” was identified32 Glc
(Supplemental Data) as Glc that is in equilibrium
Dwith IICB with an estimated K of 10 M; theGlc !7
Glc is kinetically competent. Our data suggestthat the sources were very low levels ofcontamination of laboratory water and reagents,and a cellular source, possibly glycogen.
Erni and co-workers purified IICB toGlc
apparent homogeneity from Salmonellatyphimuriun and E. coli and were the first tocharacterize this transporter (31), finding forinstance, that it contained a phosphorylation siteCys in the B domain similar to that found in IIMtl
(26). They found that isolated [ P]IICB could32 Glc
transfer the phospho-group to Glc, but the rate ofthe reaction was exceedingly slow relative to therate constants reported here (32). There areseveral possible explanations for this difference,e.g. the enzyme was perhaps partially detergentdenatured during its isolation (32), or perhapsthere are differences between the physical state ofthe purified enzyme in the lipid/detergent mixturescompared to its state in the natural membranes.
Garcia-Alles, et al. (18, 33) reported thatIICB from E. coli exhibits steady-state kineticsGlc
that are biphasic when a large range of sugarconcentrations (50 :M to 5 mM) is tested. Theauthors attribute this to the presence of at leastthree (and perhaps four), catalytic sites that fallinto two classes, one with higher affinity for Glc
S(K . 10 :M) but lower phosphorylation activity,
Sand the other with low affinity for Glc (K . 300:M ) but about 6 times the phosphorylationactivity of the high affinity class. II also hasMtl
high and low affinity sites that are delimited at 5:M Mtl, and the low affinity site has the highercapacity (17). Our measurements were made atGlc concentrations between 0.08 and 0.2 :M(with one instance of 5 :M) which are all wellwithin the high affinity region. We have noinformation about the kinetics of IICB in theGlc
low affinity region. The presence of multiplereactive sites and their kinetic properties will bearon interpreting the physiological significance ofthe lower branch of the mechanism of IICBGlc
(Fig. 2).The unphosphorylated forms of several
Enzymes II bind their sugar substrates (34); boundMtl is phosphorylated (35); and the enzyme waspostulated to have a random order of addition
mechanism by analysis of steady-state kinetic data(17,36). We find that Glc bound tounphosphorylated IICB is kinetically competent;Glc
therefore IICB also has a random order ofGlc
addition mechanism. It is obvious that the relativeflux through the two branches will be verydependent on the rate of binding of Glc to IICB ,Glc
VIbut k is one of the least certain of the rateconstants presented here. The lower branch of thepathway may be of physiological significanceunder conditions that deplete IICB of phospho-Glc
groups (i.e., low cellular concentrations of PEPand/or the presence of other PTS sugar substrates)in the presence of Glc.
In sum, we have analyzed an Enzyme II of thePTS with transient-state kinetic methods, and havefound that IICB has a random order of addition,Glc
bi-bi, substituted enzyme mechanism with thefollowing properties. 1) The lower branch has asmall effect on the flux through the enzyme underthe conditions used in our experiments. Since themagnitude of the effect is dependent on the
VImagnitude of k (the rate constant for Glc bindingto IICB ) and the sugar concentration, underGlc
other conditions the lower branch of the kineticmechanism could become physiologicallysignificant. 2) We have been able to estimate therate constants for the binding of Glc to IICBGlc
even though they are not directly measurable. 3)Although IICB is the fourth protein in the PTSGlc
pathway to which the phospho-group from PEP istransferred, the phospho-enzyme retains aphosphotransfer potential much higher than that ofATP. An overview of the kinetics andthermodynamics of the glucose-specific PTS willbe presented elsewhere. 4) Finally, our resultsconfirm the data used to build a kinetic model thatshowed that control of flux through the Glcspecific PTS of E. coli is exerted at the last stepsof the pathway, the phosphotransfer reactions ofIICB , in cells grown on glucose to mid-Glc
exponential phase (22). The model successfullyreplicated the flux both in vivo and in vitro whichsuggests that its extension to other sugar-specificEnzymes II will enhance our ability to predictcellular responses to a variety of physiologicalconditions.
1. Meadow, N. D., Fox, D. K., and Roseman, S. (1990) Ann.Rev.Biochem. 59, 497-5422. Postma, P. W., Lengeler, J. W., and Jacobson, G. R. (1993) Microbiol.Rev. 57, 543-5943. Robillard, G. T. and Broos, J. (1999) Biochim.Biophys.Acta 1422, 73-1044. Siebold, C., Flukiger, K., Beutler, R., and Erni, B. (2001) FEBS Lett. 504, 104-1115. Hu, K. Y. and Saier, M. H., Jr. (2002) Res.Microbiol. 153, 405-4156. Meadow, N. D. and Roseman, S. (1996) J.Biol.Chem. 271, 33440-334457. Buhr, A., Flükiger, K., and Erni, B. (1994) J.Biol.Chem. 269, 23437-234438. Guan, K. L. and Dixon, J. E. (1991) J.Biol.Chem. 266, 17026-170309. Bouma, C. L., Meadow, N. D., Stover, E. W., and Roseman, S. (1987) Proc.Natl.Acad.Sci., U.S.A.
84, 930-93410. Pelton, J. G., Torchia, D. A., Meadow, N. D., Wong, C.-Y., and Roseman, S. (1991) Biochem 30,10043-1005711. Waeber, U., Buhr, A., Schunk, T., and Erni, B. (1993) FEBS Lett. 324, 109-11212. Waygood, E. B. and Meadow, N. D. (1982) Methods Enzymol. 90 Pt E, 423-43113. Bradford, M. M. (1976) Anal.Biochem. 72, 248-25214. Markwell, M. A. K., Haas, S. M., Bieber, L. L., and Tolbert, N. E. (1978) Anal.Biochem. 87, 206-21015. Jaenicke, L. (1974) Anal.Biochem. 61, 623-62716. Roossien, F. F., Brink, J., and Robillard, G. T. (1983) Biochim.Biophys.Acta 760, 185-18717. Lolkema, J. S., ten Hoeve-Duurkens, R. H., and Robillard, G. T. (1993) J.Biol.Chem. 268, 17844-1784918. Garcia-Alles, L. F., Zahn, A., and Erni, B. (2002) Biochem 41, 1008619. Barshop, B. A., Wrenn, R. F., and Frieden, C. (1983) Anal.Biochem. 130, 134-14520. Anderson, K. S., Sikorski, J. A., and Johnson, K. A. (1988) Biochem 27, 7395-740621. Zimmerle, C. T. and Frieden, C. (1989) Biochem.J. 258, 381-38722. Rohwer, J. M., Meadow, N. D., Roseman, S., Westerhoff, H. V., and Postma, P. W. (2000)J.Biol.Chem. 275, 34909-3492123. Herr, E. B. Jr. and Koshland, D. E. J. (1957) Biochim Biophys Acta 25, 219-22024. Akerfeldt, S. (1960) Acta Chem.Scan. 14, 1980-198425. Meins, M., Jenö, P., Müller, D., Richter, W. J., Rosenbusch, J. P., and Erni, B. (1993) J.Biol.Chem.268, 11604-1160926. Pas, H. H. and Robillard, G. T. (1988) Biochem 27, 5835-583927. Black, S. and Wright, N. G. (1955) J.Biol.Chem. 213, 27-3828. Meadow, N. D. and Roseman, S. (1982) J.Biol.Chem. 257, 14526-1453729. Stock, J. B., Waygood, E. B., Meadow, N. D., Postma, P. W., and Roseman, S. (1982) J.Biol.Chem.257, 14543-1455230. Cornish-Bowden, A. (1995) Fundamentals of Enzyme Kinetics, Rev. Ed., Portland Press, London31. Meins, M., Zanolari, B., Rosenbusch, J. P., and Erni, B. (1988) J.Biol.Chem. 263, 12986-1299332. Erni, B. (1986) Biochem 25, 305-31233. Garcia-Alles, L. F., Navdaeva, V., Haenni, S., and Erni, B. (2002) Eur.J.Biochem. 269, 4969-498034. Robillard, G. T. and Lolkema, J. S. (1988) Biochim.Biophys.Acta 947, 493-51935. Lolkema, J. S., ten Hoeve-Duurkens, R. H., Dijkstra, D. S., and Robillard, G. T. (1991) Biochem30, 6716-672136. Lolkema, J. S. (1993) J.Biol.Chem. 268, 17850-1786037. Cleland, W. W. (1979) Methods Enzymol. 63, 103-138
In order to calculate values for the specificity constants from data in the older literature, we estimated2
the concentration of IICB from the amount of protein or dry weight used in the assays by applying theGlc
purification factors and specific activity of the pure protein that were found in later work. Since thepresent measurements were made at room temperature we also applied a correction of a factor of 0.4 tocompensate for the effect of temperature between 37 C at which the steady-state measurements wereo
Fig. 1 A diagram of the Glc specific PTS from E. coli. The phosphorylated amino acid in eachof the four proteins is indicated. There are five phosphotransfer reactions, each designated by the Romannumeral used throughout the text. The glucose permease, IICB is shown separated into its twoGlc
domains, the phosphorylation domain IIB which extends into the cytoplasm, and the sugar recognitionGlc
and binding domain IIC which is an integral membrane domain. The reactions of IICB are drawn asGlc Glc
conventionally represented, and do not illustrate the random order mechanism, presented in Results, inwhich Glc binds either to unphosphorylated or to phosphorylated IICB .Glc
Fig. 2 Phosphotransfer from [P]IIA to Glc and the random order of addition mechanismGlc
for IICB . The upper panel shows the proposed catalytic mechanism of IICB ; it is a bi-bi, randomGlc Glc
order of addition, substituted enzyme mechanism. The lower panel shows the Scheme of balanced firstand second order equations for the reactions of the bi-bi, random order mechanism, except Reaction IVa
XXXwhich applies only to the IIB domain. The rate constants of the reactions (k ) are determined (usingGlc
the kinetic simulator, Kinsim) by numerical integration of the differential equations defined by the
XXXchemical reactions. The signs of the rate constants (k ) are positive for reactions proceeding left toright. Reactions pertaining to [ H]Glc are identified by a “t.” These data are treated as separate reactions3
in the model because the appearance of [ H]Glc-6-[ P] was a separate data set, independent of the data3 32
on the appearance of total Glc-6-[ P]. In modeling with the simulator, these constants were held equal to32
the corresponding constants for unlabeled Glc; therefore they are not given in the tables or figures. Rateconstants that are omitted were assigned values of zero in the simulator because of their lowthermodynamic reversibility. In the text, Reactions IV and V are referred to as the “upper pathway,” andReactions VI and VII as the “lower pathway.”
Fig. 3 Transfer of [ P]-phospho-group from [ P]IIA to IIB -6His; data modeled by32 32 Glc Glc
Reaction IVa. Experimental and theoretical progress curves are shown for the transfer reaction. Therapid quench experiment was conducted as described in Methods using the following initialconcentrations (after mixing): [ P]IIA = 54 nM; IIA = 21 nM (produced by hydrolysis of [ P]IIA32 Glc Glc 32 Glc
during storage); IIB -6His = 45 nM. Q, [ P]IIA ; �, [ P]IIB -6His. Solid and dashed lines are theGlc 32 Glc 32 Glc
theoretical progress curves fitted by non-linear least squares (using Fitsim) to the differential equationdefined by Reaction IVa. A: The first second of the progress curves on a linear time scale. B: The fulltime course of the reaction (30 s) on a logarithmic time scale; the ordinate is to the same scale as A, andthe same symbols are used. Time points longer than 10 s were obtained after hand mixing. The rate
IVa !IVaconstants from the model are: k = 10.0 (± 0.3) × 10 M s and k = 3.7 (±0.2) × 10 M s . The6 -1 -1 6 -1 -1
eqcalculated apparent equilibrium constant for Reaction IV is therefore K' = 2.7. The data from thisexperiment are included in the global analysis shown in Table I, Row 1.
Fig. 4 Effect of pH on the rate of hydrolysis of [ P]IIB . The rate constants for hydrolysis of32 Glc
the radio-labeled phospho protein were determined at 23 C as described in Methods. F, [ P]IIB ; +,o 32 Glc
[ P]IICB32 Glc
Fig. 5 The kinetic competence of Glc bound to IICB : the relevance of Reaction VII; dataGlc
modeled by Reactions IV-VII . The rapid quench experiment measured the rate of transfer of the [ P]-32
phospho-group from [ P]IIA to Glc via IICB ; the data were modeled by the differential rate32 Glc Glc
equations defined by Reactions IV, V, VI, and VII. Only the data for the appearance of [ P]IICB and32 Glc
Glc-6- P are shown; the data for the loss of [ P]IIA are not shown. Panels A and B show the same32 32 Glc
experimental data points fitted manually in two different ways. In A , the model assumed that Glc inIICB CGlc is not capable of being phosphorylated, while in B it was assumed to be kineticallyGlc
competent. F, Glc-6- P; L, [ P]IICB . The solid and dashed lines are manually fitted theoretical32 32 Glc
curves. After mixing, the initial concentrations were: [ P]IIA = 105 nM; IIA = 20 nM (produced by32 Glc Glc
hydrolysis of [ P]IIA during storage); total (wild type) IICB = 17 nM (IICB CGlc = 15 nM,32 Glc Glc Glc
Dcalculated from a K of 10 M); total glucose = 380 nM. To increase the signal, the membranes were-7
washed only once and not incubated/dialyzed (see Methods) so that the ratio of Glc to IICB is 22 to 1,Glc
compared to the ratio of 2 or 3 to1 as in all the other experiments. In A, Reaction VII was assigned a rateconstant of zero, whereas in B it was non-zero and was fitted. Fixing the rate constants of Reaction VIIat zero, implies that all the flux of phospho-groups occurs through the upper branch of the mechanismshown in Fig. 2.
Fig. 6 Comparison of the rate of phosphorylation of exogenous [ H]Glc when added to3
different reagent solutions. Data from two parts of a rapid quench experiment that measured the rate oftransfer of the [ P]-phospho-group from [ P]IIA to Glc via IICB . The data from the progress curves32 32 Glc Glc
were modeled by the differential rate equations defined by Reactions IV, V and Vt, VI and VIt, and VIIand VIIt. . Only the data for the two radioactive forms of Glc are shown; the data for the phospho-proteins are not shown. Panel A: Expt. 1, [ H]Glc was added to the IICB -6His solution (F, +), and3 Glc
allowed to equilibrate with the endogenous Glc for 30 min. After mixing with the labeled phosphoIIA , the initial concentrations were: [ P]IIA = 138 nM; IIA = 25 nM (produced by hydrolysis ofGlc 32 Glc Glc
D[ P]IIA during storage); total IICB -6His = 97 nM (IICB CGlc = 47.4 nM calculated from a K of32 Glc Glc Glc
10 M); total Glc = 95 nM (85 nM of endogenous and 10 nM [ H]Glc). Panel B: Expt. 2, [ H]Glc was-7 3 3
added to the [ P]IIA solution (�,�) and mixed with IICB containing endogenous Glc. After32 Glc Glc
mixing, the initial concentrations were: [ P]IIA = 128 nM; IIA = 35 nM (produced by hydrolysis of32 Glc Glc
D[ P]IIA during storage); total IICB -6His = 97 nM (IICB CGlc = 43 nM calculated from a K of 1032 Glc Glc Glc -7
M); total Glc = 110 nM (85 nM endogenous, 25 nM [ H]Glc from the [ P]IIA solution). The data3 32 Glc
were fitted manually; the solid line is a theoretical curve. Expt. 1: F, total Glc-6-[ P]; +,32
[ H]Glc-6-[ P]. Expt. 2: �, total Glc-6- P; �, [ H]Glc-6-[ P]. In Expt. 2, the specific activity of the3 32 32 3 32
[ H]Glc changes continuously for the first several seconds (see Supplemental Data on Kinetics). After3
the rate constants for the experiment had been determined, Kinsim was used to simulate the time courseof the change in the specific activity. This time course was then used to calculate the exact concentrationof [ H]Glc-6-[ P] at the specific time points. Therefore, the dotted line is drawn through the data points,3 32
i.e., it is not a theoretical fit. These data are included in the global analyses shown in Table I, Rows 2-11.Fig. 7 Transfer of [ P] from [ P]IIA to Glc via IICB : comparison of IICB and32 32 Glc Glc Glc
IICB -6His at different concentrations as well as the effect of Glc concentration on the rateGlc
constants. Data from two rapid quench experiments that measured the rate of transfer of the [ P]-32
phospho-group from [ P]IIA to IICB and then to Glc. The data from the progress curves were32 Glc Glc
modeled by the differential rate equations defined by Reactions IV, V and Vt, VI and VIt, and VII andVIIt. The time scales are logarithmic and the symbols are the same for both experiments: L,[ P]IICB -6His (Panel A) or [ P]IICB (Panel B) (each in membranes); Q, [ P]IIA ; F, total Glc-6-32 Glc 32 Glc 32 Glc
[ P]; �, [ H]Glc-6-[ P]; solid and dashed lines are the theoretical progress curves obtained by manual32 3 32
fitting. Time points longer than 10s were obtained after hand mixing. Panel A: [ H]Glc added with3
[ P]IIA . The specific activity of [ H]Glc used in this Panel was calculated as if the exogenous [ H]Glc32 Glc 3 3
mixed instantly with the total Glc, not just the free Glc (see Supplemental Data on Kinetics). Since, inthis experiment, 21% of the total Glc was initially sequestered as IICB CGlc this specific activity wasGlc
about 20% too low at the first time point, but this error decreased continually and disappeared by about30s. Therefore, the curve for the concentration of [ H]Glc-6-[ P] generated by Kinsim is 20% too high at3 32
the initial time points, but it becomes correct by about 30 s. After mixing, the initial concentrationswere: [ P]IIA = 117 nM; IIA = 17 nM (produced by hydrolysis of [ P]IIA during storage); total32 Glc Glc 32 Glc
Dwild type IICB = 40 nM (IICB CGlc = 22 nM calculated from a K of 10 M); total glucose = 107Glc Glc -7
nM; [ H]Glc (added with [ P]IIA solution) = 25 nM. The estimated rate constants are (all M s ×3 32 Glc -1 -1
IV !IV V VII VI !VI10 ): k = 5; k = 0.7; k = 2.5; k = 1.5; k and k were fixed for simulation. Panel B: Both-6
syringes of the rapid quench apparatus contained 5 :M exogenous glucose, which was tritiated in thesyringe containing the [ P]IIA . The specific activity of the tritiated glucose after mixing with the32 Glc
unlabeled glucose in the syringe containing the IICB was used to calculate the concentration of theGlc
[ H]Glc-6-[ P] produced during the time course. Because the fraction of total Glc that was sequestered3 32
by IICB was very small, the change in the specific activity of the [ H]Glc during the time course wasGlc 3
insignificant. After mixing, the initial concentrations were: [ P]IIA = 168 nM; IIA = 33 nM32 Glc Glc
(produced by hydrolysis of [ P]IIA during storage); total IICB = 132 nM (IICB -6HisCGlc = 12932 Glc Glc Glc
DnM calculated from a K of 10 M); total Glc = 5 :M . The manually estimated rate constants are (all-7
IV !IV V VII VI !VIM s × 10 ): k = 3.5; k = 0.33; k = 2.7; k = 2.5; k and k were fixed for simulation. The data-1 -1 -6
from the experiment shown Panel A are included in the global analyses shown in Table I, Rows 2-11; thedata from the experiment shown in Panel B are included in Table I, Row 14.
Table IRate constants of the phosphotransfer reactions from [ P]IIA to IIB -6His or through IICB to glucose32 Glc Glc Glc
The constants were estimated either by manual fitting of the data from individual experiments using Kinsim (19,20), or by non-linear least
Dsquares fitting of the data from groups of experiments using Fitsim (21). The analyses were performed by first choosing a K for the sugar'
D Dbinding reaction; the Table shows the results obtained for K = 10 , 10 , or 10 M. The K was used to calculate, from the total concentration!6 !7 !8 '
Dof Glc and IICB before mixing, the concentrations of free Glc, free IICB , and IICB CGlc present in the syringe. For each value of the K’ ,Glc Glc Glc
VI !VIat least 7 pairs of values for k and k were chosen (the results from only three or four of these pairs are shown in the Table). Finally, the
VI !VIsimulation was performed keeping k and k fixed while fitting the rate constants for the phosphotransfer reactions.
Fixed constants Fitted constants1,2 2
D VI !VI IV IVa !IV !IVa V VIIK k k k (k ) k (k ) k k SD' 3
M M s s M s (± SE ) × 10 x 10-1 -1 -1 -1 -1 4 -6 9
Global, non-linear least squares analysis of four experiments using IIB -6His (116 data points) RowGlc
(8.0 ± 0.3) (2.3 ± 0.2) 1.7 (1)Global, non-linear least squares analysis of four experiments using IICB -6His in membranes (160 data points)5 Glc
Dconstant itself (21,37), and in the fits with a K = 10 M, the standard errors range from about one-half to about one-tenth the magnitude of the-7
rate constant with which they are associated. In these experiments, [ H]Glc was added to the [ P]IIA solution. Two additional time courses were obtained, one which had no Glc added,5 3 32 Glc
the other which had Glc added to the IICB solution. The rate constants derived from the additional 8 data sets agreed well with those presentedGlc
in the Table.
D Varying the K for Glc binding to IICB from 10 to 10 M had essentially the same effect as shown for IICB -6His. The best fit was6 Glc -6 -8 Glc
Norman D. Meadow, Regina S. Savtchenko, Azin Nezami and Saul Rosemansecond order rate constants for the glucose binding and phosphotransfer reactions
andphosphoenolpyruvate phosphotransferase system of Escherichia coli: Equilibrium Transient-state kinetics of enzyme IICBGlc, a glucose transporter of the
published online October 4, 2005J. Biol. Chem.
10.1074/jbc.M501440200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here