Analysis of the mechanism of chloramphenicol acetyltransferase by steady-state kinetics

Document Sample
Analysis of the mechanism of chloramphenicol acetyltransferase by steady-state kinetics Powered By Docstoc
					Biochem. J. (1984) 223, 211-220                                                                            211
Printed in Great Britain

          Analysis of the mechanism of chloramphenicol acetyltransferase by
                                steady-state kinetics
                                  Evidence for a ternary-complex mechanism

                           Colin KLEANTHOUS and William V. SHAW*
               Department of Biochemistry, University of Leicester, Leicester LE] 7RH, U.K.

                              (Received 19 March 1984/Accepted 25 June 1984)

            The mechanism of the enzymic reaction responsible for chloramphenicol resistance
            in bacteria was examined by steady-state kinetic methods. The forward reaction
            catalysed by chloramphenicol acetyltransferase leads to inactivation of the antibiotic.
            Use of alternative acyl donors and acceptors, as well as the natural substrates, has
            yielded data that favour the view that the reaction proceeds to the formation of a
            ternary complex by a rapid-equilibrium mechanism wherein the addition of
            substrates may be random but a preference for acetyl-CoA as the leading substrate
            can be detected. Chloramphenicol and acetyl-CoA bind independently, but the
            correlation between directly determined and kinetically derived dissociation
            constants is imperfect because of an unreliable slope term in the rate equation. The
            reverse reaction, yielding acetyl-CoA and chloramphenicol, was studied in a coupled
            assay involving citrate synthase and malate dehydrogenase, and is best described by a
            rapid-equilibrium mechanism with random addition of substrates. The directly
            determined dissociation constant for CoA is in agreement with that derived from
            kinetic measurements under the assumption of an independent-sites model.

   The biochemical mechanism of bacterial resist-         studied in. the experiments to be described in the
ance to  chloramphenicol is commonly that of in-          present paper being the most catalytically active.
activation by O-acetylation of the antibiotic, a re-      The type I and type III proteins are known to as-
action catalysed by chloramphenicol acetyltrans-          sociate with one another to form hybrid tetramers
ferase (CAT) (EC with acetyl-CoA as the         that possess properties of both (Packman & Shaw,
acyl donor (reviewed by Shaw, 1983). The enzyme           1981b), but surprisingly little is known about the
is cytoplasmic and tetrameric in all bacterial            catalytic mechanism for either variant. The type
species examined to date, consisting of four identi-      III variant of CAT was chosen for a detailed
cal catalytic subunits, the M, of which is approx.        kinetic analysis because of (a) success in obtaining
25000. The many examples of CAT characterized             protein crystals of it which were suitable for ana-
to date constitute a family of proteins that catalyse     lysis by X-ray diffraction (Liddell et al., 1978), (b)
the acetylation of chloramphenicol with variable          its catalytic efficiency, and (c) ease and efficiency
efficiency. The plasmid-associated genes for CAT          of purification in high yield by affinity chromato-
common to the enteric bacteria specify three main         graphy. Because of extended regions of sequence
enzyme types (I, II and III), the type III variant        homology among CAT variants there are good
                                                          reasons for believing that their catalytic mechan-
                                                          isms will prove to be very similar if not identical.
   Abbreviations used: CAT, chloramphenicol acetyl-       Evidence bearing on the likely importance of a
transferase; CAT,, CAT,, and CAT,,,, the three natural-   histidine residue in a highly conserved segment of
ly occurring enterobacterial enzyme variants; TSE
buffer, 50mM-Tris/HCl buffer, pH 7.5, containing          the primary structure has been summarized else-
lOOmM-NaCi and 0.1 mM-EDTA; ko, the overall cata-         where (Shaw, 1983).
lytic rate constant (also known as kcat).                    Previous studies have demonstrated that the
  * To whom
               reprint requests should be addressed.      following reactions describe the fate of chloram-
Vol. 223
212                                                                        C. Kleanthous and W. V. Shaw

phenicol in bacteria that contain CAT (see Shaw,         Preparation and purification of acyl-CoA analogues
1983):                                                      Acyl-CoA analogues were prepared by the
                                                         method of Simon & Shemin (1953), with the use of
Chloramphenicol + acetyl-CoA                             acetic anhydride, propionic anhydride and butyric
         3-acetoxy chloramphenicol + CoA (1)-            anhydride. The extent of acylation was monitored
                                                         spectrophotometrically by the reaction of residual
3-Acetoxy chloramphenicol

                                                         CoA with a 1 mm solution of 5,5'-dithiobis-(2-nitro-
               1-acetoxy chloramphenicol          (2)    benzoic acid) in TSE buffer at 25°C. Acylation was
                                                         judged to be greater than 99% in each case, on the
1-Acetoxy chloramphenicol + acetyl-CoA     0             basis of e412 = 13.6 x 103 M- I -cm-'. Addition of a
     1,3-diacetoxy chloramphenicol + CoA (3)             large excess of chloramphenicol (final concentra-
                                                         tion 500 gM) and CAT (20-50 units/ml) allowed the
Reactions (1) and (3) are both catalysed by CAT,         concentration of acyl-CoA to be quantified and
whereas reaction (2) is a non-enzymic and pH-de-         routinely showed agreement (± 5%) with estimates
pendent re-arrangement (Nakagawa et al., 1979;           based on the adenine absorbance at 260nm
Thibault et al., 1980). The enzymic formation of         (£260= 15.4 x 103 M- I -cm- in water).
1,3-diacetoxy chloramphenicol is slow and is not            CoA and acyl-CoA analogues were purified by a
required for inactivation of the antibiotic, since the   procedure modified from that of Frey & Utter
mono-acetoxy derivatives of chloramphenicol are          (1977), with the use of a DEAE-cellulose anion-
devoid of activity and fail to bind to ribosomes         exchange column with a stepped KCI gradient at
(Shaw & Unowsky, 1968).                                  pH4 and 4°C. The material was desalted by gel
  The studies described in the present paper were        filtration (Sephadex C-10) and eluted in water.
undertaken to gain a coherent view of the likely         Single spots, detected by their absorption of u.v.
mechanism for reaction (1) from the results of a         light, were observed for CoA and each of the acyl-
steady-state kinetic analysis of the forward and         CoA analogues after descending chromatography
reverse reactions and to extend previous kinetic         on Whatman no. 1 paper developed in butanol/
observations on the non-enzymic re-arrangement           acetic acid/water (5 :2:3, by vol.).
(reaction 2). The symbols and conventions intro-
duced by Dalziel (1957) are used for the steady-         Preparation and purification of 3-acetoxy chloram-
state kinetic analysis, wherein substrates A and B       phenicol
in the forward reaction are chloramphenicol and             Acetic anhydride (60 !l, 0.6 mmol) was added in
acyl-CoA respectively, and in the reverse reaction       portions to chloramphenicol (194mg, 0.6mmol) in
P and Q are 3-acetoxy chloramphenicol and CoA            pyridine (1 ml) at 4°C. The mixture was warmed to
respectively.                                            room temperature and extracted with ethyl acet-
                                                         ate. The combined organic extracts were dried
                                                         over MgSO4 and the solvent was removed by
Materials and methods                                    evaporation. The resulting oil was applied to a
                                                          120ml Kieselgel 60 (Merck) silica column de-
Enzymes, reagents and instruments                        veloped with chloroform/methanol (99:1, v/v).
  Type III CAT was purified from cell paste              The eluate was monitored by its absorbance at
                                                         278 nm; the diacetoxy derivative of chloram-
(stored at - 20°C) of Escherichia coli J53 (R387) ob-    phenicol was found to be eluted in the first column
tained frQm the.Microbiological Research Estab-          volume and 3-acetoxy chloramphenicol -in the
lishment, Porton Down, Wilts., U.K. Citrate syn-         second. The identity and purity of 3-acetoxy chlor-
thase, malate dehydrogenase, NADI, NADH and              amphenicol was shown by t.l.c. on Kieselgel 60
CoA were supplied by the Boehringer Corporation          F254 (Merck) silica-gel plates developed in chloro-
(London) Ltd. [14C]Chloramphenicol and                   form/methanol (19:1, v/v) (Shaw, 1967). The
[3H]CoA were purchased from New England                  extent of 1-acetoxy chloramphenicol impurity was
Nuclear. 3-Fluoro chloramphenicol was prepared           less than 1% as demonstrated by high-pressure
by Schering (Bloomfield, NJ, U.S.A.). The chlor-         liquid chromatography (see below).
amphenicol analogues were gifts from Parke,
Davis (Ann Arbor, MI, U.S.A.). All other reagents
were purchased from the Sigma Chemical Co.               Determination of the rate of isomerization of
High-pressure liquid-chromatography experi-              3-acetoxy chloramphenicol to I-acetoxy chloram-
ments were carried out on a Laboratory Data Con-         phenicol
trol System and initial rates measured on Pye-              Duplicate solutions (1 ml) of TSE buffer, aceto-
Unicam SP. 1800 and SP. 8-100 and Perkin-Elmer           nitrile (20mM) and 3-acetoxy chloramphenicol
Lambda 5 spectrophotometers.                             (0.34mM) were incubated at 25°C and, at the ap-
Mechanism of chloramphenicol acetylation                                                              213

propriate times, samples (20 u1) were analysed for    system described by eqns. (4)-(6) a fall in the con-
concentrations of monoester by high-pressure          centration of oxaloacetate is instantaneously
liquid chromatography. The isomers were separ-        reflected in the reduction of NAD+ by malate
ated at room temperature on a C-18 pBondapak          and an increase in absorbance at 340nm:
reverse-phase column (HPLC Technology Ltd.)
eluted with 35% (v/v) acetonitrile in 50mM-sodium     3-Acetoxy chloramphenicol + CoA CAT
acetate buffer, pH5.7, at a flow rate of 2ml/min                   chloramphenicol + acetyl-CoA (4)
(Burke et al., 1980). Peaks were identified by ab-
sorbance at 278nm.                                    Acetyl-CoA + oxaloacetate -    cs
CA T standard assay                                                                   citrate + CoA (5)
   A modification of the procedure described by       L-Malate + NAD+ MDH
Shaw (1975) was used. The standard assay mixture                      oxaloacetate + NADH + H+ (6)
contained TSE buffer, 1 mM-5,5'-dithiobis-(2-nitro-
benzoic acid), 0.4mM-acetyl-CoA and 0.1 mM-           Each assay mixture contained TSE buffer (pH 7.5),
chloramphenicol. The reaction was initiated by the    5 mM-L-malate, 1.5 mM-NAD+, 0.1 1 mM-NADH, 2
addition of enzyme and was monitored by the in-       units (Mmol/min) of malate dehydrogenase, 4 units
crease in absorbance at 412 nm. One enzyme unit is    (pmol/min) of citrate synthase, 3-acetoxy chloram-
defined as that amount of enzyme that catalyses       phenicol, CoA and CAT in a final volume of 1 ml.
the production of 1 Mmol of product/min at 25°C.      Acetonitrile (final concentration 20mM) was incor-
Purification and specific activity of CAT             porated to solubilize 3-acetoxy chloramphenicol,
                                                      and control experiments showed it to have no
   The type III protein was purified to apparent      effect on either the activity of CAT or the coupled
homogeneity by using an affinity-chromatography       assay. None of the components of the coupled
method (Zaidenzaig & Shaw, 1976) and the proto-       assay affected the rate of the CAT forward re-
col of Packman & Shaw (1981a). Sodium dodecyl         action. Under the above conditions 0.93 equiv. of
sulphate/polyacrylamide-gel electrophoresis with      NADH was produced/equiv. of acetyl-CoA con-
molecular-mass markers and staining with Coo-         sumed. The stoichiometry is such only if NADH is
massie Blue revealed a single protein band with an    present in the starting assay mixture (Williamson
apparent M, of 25 000. Purified CAT was stored in     & Corkey, 1969).
TSE buffer containing chloramphenicol (0.2mM)            Each double-reciprocal plot was constructed
and 2-mercaptoethanol (0.2mM) at -20°C. The           from experimental points resulting from a matrix
specific activity of purified enzyme, after 18h       of four substrate and four coenzyme concentra-
dialysis at 4°C against TSE buffer, was found to be   tions, and all reactions were performed in tripli-
800+50 units/mg, on the basis of enzyme assay         cate. For the lowest chloramphenicol and 3-
under standard conditions and of amino acid ana-      acetoxy chloramphenicol concentrations, in both
lysis and the known amino acid composition of the     directions, a maximum of 15-17% depletion of sub-
enzyme (Packman, 1978).                               strate occurred, and linearity of initial rates was
Kinetic measurements                                  observed throughout. Kinetic coefficients in the
                                                      initial-rate equation:
  Forward reaction. Measurement of the initial
rates of enzymic acetylation of chloramphenicol by
acetyl-CoA for kinetic analysis by double-reci-
                                                              [El      O+ A + B           OAB]
                                                                        [A] [B]        [Al [B]
procal plots employed the same conditions as used                v
in the standard assay but over various substrate
concentrations (typically in the range 0.3 x Km to    were calculated from linear intercept and slope re-
5 x Km). The concentrations of chloramphenicol        plots from manually drawn double-reciprocal plots
and related analogues were determined from ab-        (Dalziel, 1957).
sorbance measurements at 278 nm and acyl-CoA             Km and ko for the hydrolysis of acyl-CoA ana-
compounds as described above.                         logues. Conditions were identical with those of the
  Reverse reaction. Initial rates of acetylation of   forward reaction assays except that no chloram-
CoA by 3-acetoxy chloramphenicol were measured        phenicol was added and the concentration of CAT
by coupling the CAT reverse reaction with citrate     (monomer) was typically in the range 0.5-2.0 gM.
synthase (CS) and malate dehydrogenase (MDH)          Enzyme preparations were exhaustively dialysed
and monitoring the appearance of NADH at              against TSE buffer at 4°C before use to remove
340nm, an approach adopted with success by            bound chloramphenicol. All Km and ko values were
Hersh & Peet (1977) in studies of the mechanism       determined by direct linear plots (Eisenthal &
of choline acetyltransferase. In the coupled          Cornish-Bowden, 1974).
Vol. 223
214                                                                           C. Kleanthous and W. V. Shaw

Equilibrium dialysis                                                                                                 4

   Equilibrium dialysis was performed at 26°C in
an eight-cell rotating module (Hoefer Scientific In-
struments, San Francisco, CA, U.S.A.) essentially        u,, 1 2
as described by Bennett & Shaw (1983). Each cell         -
was divided into two 0.5 ml chambers by a dialysis      i--
membrane with a nominal cut-off of M, 6000-              x
8000. Before use the membrane was boiled in a            O)     6
solution of NaHCO3 (24mM) containing EDTA
(3.8mM) and 2-mercaptoethanol (7.5mM). Mem-
branes were stored at 4°C in 50% (v/v) ethanol.
One chamber of each of seven cells received 300 .1
of a solution containing CAT in TSE buffer,                                  01          0.02           0.03             0.04
whereas the opposing chambers contained a range                            1/1Acetyl-CoAl (,UM-1)
of either ['4C]chloramphenicol or [3H]CoA dilu-         Fig. 1. Lineweaver-Burk plot for the acetylation of chlor-
tions in TSE buffer. One cell was used as a control     amphenicol by chloramphenicol acetyltransferase over a
to monitor any loss of enzyme activity during the       range of acetyl-CoA andfixed chloramphenicol concentra-
prolonged incubations (rotation at 10rev./min;                                         tions
equilibration was complete after 15 h for chloram-        The final enzyme (monomer) concentration was
phenicol and 24h for CoA). Samples (50 pl) were           0.14nM, the chloramphenicol concentrations were
taken in triplicate from each chamber and placed          98pM (1), 19.8 p1M (2), 9.9 pM (3) and 6.0M (4), and
in 4ml of either Bray's fluid (for [14C]chloram-          the incubation conditions were pH 7.5 in 0.1 M-
                                                          Tris/HCl buffer containing lO0mM-NaCl and
phenicol) or Fisofluor 1 (for [3H]CoA) for scintilla-     0.1 mM-EDTA at 25°C.
tion counting. The mean values of the ligand
present in each chamber were plotted in accord-
ance with Scatchard (1949).

                                                              02N             H
                                                                              C-CI          H
Results                                                                       I      H    RI
                                                                             HO            R
Forward reaction: acetylation of chloramphenicol
   Double-reciprocal plots (Lineweaver & Burk,                      R
1934) over a range of fixed substrate concentra-              COCHCI2'   OH          Chloramphenicol
tions are linear (Fig. 1). Under the experimental             COCHCI2    F           3-Fluoro chloramphenicol
                                                              COCH20H    OH          2-Hydroxyacetamido chloramphenicol
conditions used, a small but reproducible 4AB term            COCH3      OH          2-Acetamido chloramphenicol
(derived from the slope re-plot) was observed; this
is consistent with a sequential mechanism in which        Fig. 2. Chloramphenicol analogues used in this study
 kAB is present in the rate equation (eqn. 7) and not
a double-displacement mechanism in which IAB iS
absent (Dalziel, 1957). The incorporation of 3- The corresponding data for propionyl-
fluoro chloramphenicol, a competitive inhibitor of      CoA (38-316 pM) and butyryl-CoA (63-960 pM) are
CAT (Ki =78 pM), into the assay medium rein-            also summarized in Table 1. In the latter experi-
forces the conclusion of a sequential mechanism by      ments the chloramphenicol concentration was
raising the apparent Michaelis constant for chlor-      varied from 5 to 115pM;
amphenicol and therefore the slopes of lines (4AB)
on a double-reciprocal plot (Slater, 1955; Koster &     Reverse reaction: acetylation of CoA
Veeger, 1968). The results of such an experiment           A steady-state kinetic analysis of the reverse re-
are recorded in Table 1.                                action needs to take account of the non-enzymic
    A kinetic approach similar to that used by          and intramolecular rearrangement of 3-acetoxy
Dalziel & Dickinson (1966) in studies with liver        chloramphenicol to the l-acetoxy isomer at alka-
alcohol dehydrogenase was employed to study the         line pH (Nakagawa et al., 1979):
effects of alternative substrates (Fig. 2) on the                                                 k3,
kinetic coefficients. The hydroxyacetamide and          3-Acetoxy chloramphenicol hka                          eno         (
acetamide analogues of chloramphenicol were                                         k,
studied over the concentration ranges 30-500 M                                    I1-acetoxy chloramphenicol (8)
and 13-254 pm respectively, and the results are col-
lated in Table 1. In both of the above cases the        To determine the time interval over which reliable
acetyl-CoA concentration was varied from 30 to          initial rates might be observed, the rate for
Mechanism of chloramphenicol acetylation                                                                                                                                             215

                                              cwn ,,
                                                                               U)                                                                           a
                                              4-sw                                                           1-           o             r-'t                          t
                                                                                                                                        -; o6 .rcl
                                                   (U                                                                                    _                 -O-

                                                                                                             00               0         IOC i
                                                                                                                                          -                  -            0
                                                                                                             W)               t-        --        t          o        00
                                              ,e           0

                 U)                           Cd4
                                                                                                             ,_o              -          o                  0         O

                                              o0 Y
                                               j   <                                                         O        N                 NCf'O)-             0-,0

                       .o                                                                  0-                OR               t-
                                                                                                                                        d-            M
                                                                                                                                                            0- , --,
                 C                                                                                           6;               Ci        -C                     * c
                                              U           r                                                  ON               IR.       --4C4               a.,~-,

                                                                                                                                        C140-                         0
                                                         .~C.)                                               +1               +1                                      +1
                                                                               ~          6                                             +1 +1               Oex

                 C.)                                                                                                                                                  C-i
                                     m Po                                                                     o0              _

                  U)            .

                  4).                         I-'-                                                                                                                    Cl
                  3                                                                                           +1               +1       ° +I                x 2 +l
                                     +        C     _d
                                                                                         NN l+
                                                                                             0%~             00                ^         <>O
                                                                                                                                   1-                                     f
                                                                                                             Cl4               N-                 %0
                                              0            0                    ~-        66                                                                     4-
                 0          .         +       A -e>~                                      O Cl
                                       11     Cd            m    .                                            +1               +1        +1+1               x S           +l
                                                   C.)                                                       If                N
                       *   .)       441I a°                      4)
                                                                                     ':i --                                              _

                 o         -o
                                              >                      C

                  '-U                              -0U
                                                                                                              _1 _1
                                                                                                             68 6 66                         _        _

                                                                                                                      Cl                 6                 o              +l
                  Cq                                                                                         u .                                 .

                                                                                                             -                 0

                     .%     )                     *d J
                  .x b"                                                             U4              =, C                                                                        4)
                                                                                                                                                                              0. .
                                                                                    la    o o :3         E
                                              <Q *. d
                                              C.)                                               C.)
                  ~-0                                                                   I-,
                                                                                     xyV°           --4S
                  0D         E
                                              O CdC: >                               C.     .   -   _-

                  4)            4                r- 4
                                                  0                                                                   <<
                                                  Y E                      L

                                                                     S.)                                      .C)     *-

                                                    Eco )                                                                 d         C                                         C.)

                                                            w                                   C~~~~~<               .

                                                          J0                               Co Cd
                                                                                           &U                                                Qw       uo


Vol. 223
216                                                                             C. Kleanthous and W. V. Shaw

isomerization was calculated under conditions                 Double-reciprocal plots for different fixed con-
identical with those used for enzymic assay (pH 7.5        centrations of 3-acetoxy chloramphenicol (8-
and 25°C) with the use of reverse-phase high-pres-         103 M) and varied concentrations of CoA (32-
sure liquid chromatography to mneasure the con-            574 Mm) are shown in Fig. 4. Calculated 4 par-
centration of the chloramphenicol esters (see the          ameters are given in Table 1. The reverse reaction
Materials and methods section). Fig. 3 depicts the         coefficients are an order of magnitude greater than
results of such an experiment, from which the ob-          the corresponding values for the forward reaction
served rate constant (k1 +k_1 ) was calculated to          and the presence of 4pQ is unequivocal. The ratio
be 37 x 101 minw' and the equilibrium constant             of the forward and the reverse reaction velocities
Keq. to be 0.38. Derived values for k,j and k-, were       (VfIVr) is 8.8 at pH7.5 and 25°C.
calculated to be 10.4x 10-3min- I and 26.6x 10-3
min - ' respectively. Since not more than 1% of the        Reliability of 4 coefficients
3-acetoxy chloramphenicol might be expected to               Before analysis of 4 coefficient relationships for
isomerize during a min incubation, under the               particular mechanisms, the reliability of each par-
conditions employed, calculations of initial rates         ameter was tested by determining the percentage
were limited to this period.                               contribution to the overall rate equation (eqn. 7).
                                                           In nearly every case the coefficients show a major
                                                           contribution under specified conditions. As an
                                                           example, for the highest concentrations of chlor-
                                                           amphenicol (A) and acetyl-CoA (B) used (98yM
               (a)                                         and 380 gM respectively), the percentage contribu-
                                                           tion of each term to the rate equation is: 40, 75%;
                                       FI.                 OA/[A], 9%; Os/[B], 15%; OAB/[A][B], 1%. How-
  U                                                  U     ever, 4AB values for the combination of chloram-
 1-                                                 8-     phenicol with acetyl-CoA or propionyl-CoA do not
 u 50
 0         -                                        0
                                                 50 6      contribute significantly and are therefore likely to
                                                           be inherently less reliable than other parameters.
                                                           Hence minimum emphasis has been placed on
                                                           these values and parameter ratios that contain

  In   0             25     50
                          Time (min)







                          Time (min)                                0             0.05           0.10          0.15
Fig. 3. Isomerization of 3-acetoxy chloramphenicol (3AC)                1/13-Acetoxy chloramphenicoll (#M-l)
to I-acetoxy chloramphenicol (JAC) as a function of time   Fig. 4. Lineweaver-Burk plotfor the acetylation of CoA by
   (a) The procedure was as described in the Materials     CAT over a range of 3-acetoxy chloramphenicol and fixed
   and methods section. Concentrations of monoester                             CoA concentrations
   (, 3-acetoxy chloramphenicol; 0, l-acetoxy chlor-             (a) The final enzyme (monomer) concentration was
   amphenicol) were monitored by the change in peak              1.37nM, the CoA concentrations were 574pM (1),
   height after high-pressure liquid chromatography.             106 M (2), 52 M (3) and 32 M (4), and thecondi-
   (b) Determination ofthe isomerization rate constant           tions were as described in the Materials and
   by a semi-logarithmic plot. Data are taken from (a).          methods section.

Mechanism of chloramphenicol acetylation                                                                           217

Estimation of binding constants                               in the same buffer system used for the initial-rate
  Ligand-binding studies for the forward-reaction             determinations. The results, depicted in Figs. 5(a)
and reverse-reaction substrates were undertaken in            and 5(b), illustrate the binding of these substrates
order to compare directly obtained binding con-               by CAT, in the absence of the second substrate,
stants with the corresponding values derived from             with a 1 :1 stoichiometry (per monomer). The cal-
steady-state kinetic analysis. 3-Acetoxy chloram-             culated dissociation constants were 1 1.5 ( ± 0.1) 4UM
phenicol was not studied, owing to the pH-depend-             for chloramphenicol and 78 (±2)pM for CoA.
ent isomerization.                                               A direct estimation of the binding constant for
  The binding of chloramphenicol and CoA by                   acetyl-CoA by substrate-depletion techniques
CAT were studied by equilibrium dialysis at 26°C              [equilibrium dialysis, Hummel & Dryer (1962) gel
                                                              filtration, and flow dialysis (Colowick & Womack,
                                                              1969)] was made impossible by the low but signifi-
                                                              cant intrinsic thioesterase activity associated with
                                                              CAT,,,. The hydrolysis of acetyl-CoA by substrate
                 (a)                                          concentrations of CAT in the absence of chloram-
                                                              phenicol was first noted by Zaidenzaig & Shaw
                                                              (1978) for. the type I variant of the enzyme, but the
                                                              phenomenon was not examined in detail. The ap-
                                                              parent Km for hydrolysis of the three acyl-CoA
                                                              analogues are presented in Table 2. The catalytic
  S          5                                                constant (ko), in each case, was 0.01-0.02% of that
                                                              observed for each of the corresponding acylation
   az                                                         reactions.
                                                              Evidence for a sequential mechanism
                                                                 The presence and subsequent amplification of
                               0.5                  1.0       OAB by the incorporation of a competitive inhibitor
                                v                             (3-fluoro chloramphenicol) in the forward reaction
                                                              is consistent with a ternary-complex mechanism.
                                                              The corresponding terms for each of the chloram-
                                                              phenicol analogues, and that of the reverse
                                                              reaction (4pQ), are considerably greater and must
        101                                                   therefore contribute to the overall rate equation.
                                                              Zaidenzaig & Shaw (1978) noted the absence of
                                                              partial reactions between acetyl-CoA and CoA
                                                              when they were incubated with the type I variant
                                                              of CAT, an observation in accord with the steady-
                                                              state results now presented for the type III enzyme.
                                                              The formation of a ternary complex, between
  u                        0                                  enzyme and substrates, may proceed by a variety of
                                                              pathways, as outlined in Scheme 1. The present
                                                              investigation attempted to distinguish between
                                                              these possibilities.

             0                 0.5                  1.0
                                                              Table 2. Comparison of apparent K, values (46ABA) for
                                                              acyl-CoA analogues, derivedfrom the data in Table 1, with
                                                              the corresponding Michaelis constants for the enzyme-cata-
Fig. 5. Analysis of the binding of (a) [14C]chloramphenicol                        lysed hydrolysis
and (b) [3HICoA to the type III variant of CAT by equi-          Estimates in parentheses represent unreliable
                      librium dialysis                           values.
  The procedure was as described in the Materials and
  methods section and the results were analysed in ac-                                K. (app.) (pM) Km (app.) (#M)
  cordance with Scatchard (1949). For these experi-           Acetyl-CoA               (23.6), 52-76          68
   ments the concentration of enzyme monomer was              Propionyl-CoA             (19.6)                32
   15-30pM.                                                   Butyryl-CoA                      79             65

Vol. 223
218                                                                          C. Kleanthous and W. V. Shaw

              EA                          EP               Table 3. Specificity of CAT for acyl-CoA analogues of
                                                                          various acyl-chain lengths
      A              B             QX             P                    Data are taken from Table 1.
                            k       EP                                       ko (s 1) Km (gM)      (M-1.S-1)
          E        EA
                    B                              E       Acetyl-CoA          433       74        0.59 x 107
                            k'       Q                                         709      167        0.42 x 107
      B               A'k         PA -\Q                   Butyryl-CoA         215      271        0.08 x 107

              EB                          EQ
Scheme 1. Basic representation of the possible orders of      Returning to OB, we should note that the narrow
       binding of substrates and products to CAT           range of values for this term may reflect subtle
  Key: A, chloramphenicol; B, acetyl-CoA; P, 3-            effects of variations in the structure of the ligand
  acetoxy chloramphenicol; Q, CoA. Deletion of the         occupying the chloramphenicol site on the state of
  lower (or alternatively the upper) pathway yields a      acetyl-CoA in the ternary complex.
  compulsory-order mechanism, whereas the rapid               An increase in length of the acyl-donor side
  equilibration of the two substrates and both pro-        chain by the addition of an extra methylene group
  ducts with the enzyme leads to a special case of the     (propionyl-CoA) leads to a small decrease in 00
  random-order mechanism.                                  relative to acetyl-CoA, whereas A and 4B are rela-
                                                           tively unaffected. Although a decrease in 00 shows
                                                           propionyl-CoA to be turned over at a greater rate
                                                           than acetyl-CoA, the concomitant increase in Km
Mechanistic interpretations derived from the use of        yields an efficiency quotient (kO/KB) that is essen-
alternative acyl acceptors and donors                      tially the same for both acyl donors (Table 3). The
   When compared with the natural substrate, both          appreciable increases (as compared with results for
chloramphenicol analogues yielded no change in             acetyl-CoA) that are observed in all 4 coefficients
40, a small increase in 4B and substantial increases       when butyryl-CoA is the acyl-donor show it to be a
in both OA and 4AB. The constancy of 40 indicates a        significantly poorer substrate than acetyl-CoA and
common rate-determining step for the acetylation           propionyl-CoA, an observation also reflected in
of alternative substrates, a result observed with          the corresponding ko/KB, values (Table 3). It is ap-
horse liver alcohol dehydrogenase and interpreted          parent that the catalytic constant is sensitive to
by Dalziel & Dickinson (1966) as evidence for co-          changes in acyl-chain length and that the latter
enzyme dissociation as the rate-determining step.          must affect the rate-determining step. This ob-
Furthermore, if the forward reaction, catalysed by         servation contrasts with the chloramphenicol ana-
CAT, were a compulsory-order mechanism in                  logue data, which suggest that CoA dissociation
which B leads, then 4B would be equivalent to the          may be rate-determining. Two possible explana-
association rate constant and should not vary in           tions to resolve these apparently conflicting results
spite of alterations to the structure of substrate A.      are the following. (i) Both CoA dissociation and
The relatively slight variation in OB (in comparison       interconversion of the ternary complex may contri-
with 4A) for the chloramphenicol analogues sug-            bute to the rate-determining step (see, for example,
gests that acetyl-CoA may lead in a compulsory-            the case of choline acetyltransferase; Hersh, 1982).
order mechanism. The latter condition requires             (ii) CoA dissociation alone may be rate-determin-
 kAB/OA to be the true dissociation constant               ing when acetyl-CoA and propionyl-CoA are sub-
(Dalziel & Dickinson, 1966) for acetyl-CoA and             strates, but this situation breaks down for butyryl-
should remain constant. The derived values for             CoA, and the turnover of the ternary complex be-
qPAB/OA (Table 1) appear to range from 24 gM to            comes rate-limiting. A similar situation obtains for
76 gm; however, the inherent unreliability of the          liver alcohol dehydrogenase (Dalziel & Dickinson,
4AB value determined for the parent substrates (see         1966), where the rate-determining step for the reac-
the Results section) suggests that more confidence         tion shifts from coenzyme dissociation in the re-
should be placed on the corresponding values of            duction of primary alcohols to ternary-complex
OAB/OA for the chloramphenicol analogues (53 and            interconversion for secondary alcohols.
76 pM). Subsequent studies (see below and Table 2)
on the hydrolysis of acetyl-CoA catalysed by CAT           Reverse reaction
yielded an apparent Km (and Kj) for acetyl-CoA of             For a rapid-equilibrium random-order mech-
68 pm, a value in keeping with the derived K,              anism, substrate K, values may be derived from 4
(OAB/OA) from the data for N-acyl analogues of             ratios and, if the two substrates do not affect one
chloramphenicol.                                           another's binding,   OpQ/OQ = 4p/4O'   and likewise
Mechanism of chloramphenicol acetylation                                                                   219

OpQ/op = OQ/00'. These relations are clearly ob-          that becomes kinetically significant at high con-
served for the reverse reaction, with close agree-        centrations of added enzyme. The latter view
ment between the Michaelis constants for the sub-         seems untenable on the basis of two independent
strates and their kinetically derived dissociation        observations. First, the CAT and thioesterase
constants. This is further substantiated by the           activities have been observed to co-purify to appar-
near-identical values for goA (78.Mm), as deter-          ent homogeneity with conventional (non-affinity)
mined by equilibrium dialysis, and OpQ/op (84 MM)         and affinity purification methods (results not
from kinetic data. Re-arrangement of the above 4          shown). Secondly, covalent and stoichiometric
coefficient relations leads to the equality               modification of highly purified CAT by 3-bromo-
00'= qbp 4Q/4pQ (Dalziel, 1957), which is also ful-       acetoxy chloramphenicol is accompanied by
filled (wherein the calculated value for 4p' 4 Q/4pQ      parallel inhibition of both the acetylation and the
is 0.02s and that for 4k' is 0.021s; see Table 1).        hydrolysis reactions (C. Kleanthous & W. V.
Such an outcome is indicative of the special case of      Shaw, unpublished work).
the rapid-equilibrium random-order mechanism in              The intrinsic hydrolytic activity of CAT bears a
which there is no substrate synergism.                    superficial resemblance to that of another enzymic
                                                          group-transfer reaction, the ATPase activity of
Substrate dissociation constants                          hexokinase (Trayser & Colowick, 1961). The rate
   The apparent agreement of Km and K. for the            of hydrolysis of ATP was 0.003% that of the phos-
substrates in the reverse reaction prompted a study       phorylation rate and KATP for hydrolysis was 40-
of the binding of chloramphenicol and acetyl-CoA          fold higher than that observed for the hexokinase
to the enzyme. The dissociation constant for chlor-       reaction. Dela Fuente et al. (1970) noted that
amphenicol (11.M5M), as determined by equi-               lyxose, an inert glucose analogue, promoted the
librium dialysis, does not agree with the predicted       hydrolytic activity of hexokinase and decreased
value (3.7 gM) from the steady-state studies (assum-      the K.TP to a value equal to that obtained for phos-
ing an independent-sites model), but the unreli-          phorylation of glucose. The result was interpreted
ability of 4iAB precludes any meaningful com-             as showing substrate synergism between ATP and
parison. However, we may note the close agree-            lyxose, the latter binding first. In the case of CAT,
ment of K, for chloramphenicol with the derived           however, the Michaelis constant for acetyl-CoA
Michaelis constant (4OA/0o = 11.6 MM)                     hydrolysis approximates that observed for acetyl-
   The intrinsic thioesterase activity of CAT in-         ation. Moreover, the addition of the substrate
directly leads to an estimate of K, for the three acyl-   analogue 3-fluoro chloramphenicol had no effect
CoA analogues in the form of Michaelis constants          on the observed Km for thioester hydrolysis (results
for hydrolysis (see Table 2). The Km for acetyl-CoA       not shown). Such observations are consistent with
hydrolysis (68.Mm) lies within the predicted range        the notion of independent binding of substrates.
of reliable tAB/q5A values derived from steady-state
studies with the hydroxyacetamide and acetamide           Concluding remarks
analogues of chloramphenicol (52-76 Mm). The low            The data presented strongly favour a ternary-
value for the apparent dissociation constant for          complex mechanism for chloramphenicol acetyl-
propionyl-CoA (19.6 Mm) is mirrored in the ob-            transferase. The rate-determining step for CAT
served Km for its hydrolysis (32 pM), which is below      appears to be affected by acyl-chain length but not
those of the other acyl-CoA compounds. Since the          by the identity of the acyl-acceptor. If the intercon-
Km for butyryl-CoA hydrolysis (65 Mm) also ap-            version of the ternary complex contributes signifi-
proaches the predicted K. (79 Mm), there is a case        cantly to the rate-limiting step, then the tightness
for considering the Michaelis constant for hydro-         of chloramphenicol binding is not a contributing
lysis of the acyl-CoA analogues by CAT as a first         factor. Ligand-binding studies for the forward re-
approximation of their dissociation constants. A          action clearly show that both substrates bind inde-
basic kinetic scheme for a single substrate is pro-       pendently of each other. Furthermore, the binding
posed in which it is assumed that k+2 is the rate-        constants for chloramphenicol and acetyl-CoA
limiting step for the overall reaction. Given that        (11.5 and 68 pM respectively) show good agreement
k+ 2 iS 0.1 S- 1 (0.02% of the rate of chloramphenicol    with their respective Michaelis constants (11.6 and
acetylation) and that coenzyme dissociation rate          74 uM). This situation is apparently mirrored in the
constants are normally in the range of 1103S-l I          reverse reaction, where the kinetically derived
(Fersht, 1977), it is reasonable to assume that           values of Km and Ks, for both substrates, match
k +2 < k 1 and therefore that Km reduces to .K for        very closely, and KsoA (78 pM), as determined by
hydrolysis.                                               equilibrium dialysis, agrees well with the derived
   The above analysis is based on the premise that        value (OpQ/Op = 84 pM), assuming an independent-
the enzymic hydrolysis of the acyl thioesters is due      sites model. However, the use of different acyl-
to CAT rather than a contaminating thioesterase           acceptors, in the forward reaction, has led to the
Vol. 223
220                                                                             C. Kleanthous and W. V. Shaw

suggestion that acetyl-CoA binds before chloram-             Dela Fuente, G., Lagunas, R. & Sols, A. (1970) Eur. J.
phenicol.                                                      Biochem. 16, 226-233
  Andersson et al. (1984) have reconciled similarly          Eisenthal, R. & Cornish-Bowden, A. (1974) Biochem. J.
conflicting observations for liver alcohol dehydro-            139, 715-720
genase. Both substrates bind independently of one            Fersht, A. (1977) Enzyme Structure and Mechanism, p.
another, but, under the conditions normally chosen              130, W. H. Freeman, San Francisco
                                                             Frey, W. H. & Utter, M. F. (1977)J. Biol. Chem. 252, 51-
for kinetic analysis, the E-NADH pathway to the                56
ternary complex is preferred. Reaction flow by this          Hersh, L. B. (1982) J. Biol. Chem. 257, 12820-12825
route is due to a decreased rate of coenzyme dis-            Hersh, L. B. & Peet, M. (1977) J. Biol. Chem. 252,4796-
sociation from the ternary complex as compared                 4802
with that observed for the aldehyde. Their investi-          Hummel, J. P. & Dreyer, W. J. (1962) Biochim. Biophys.
gation highlights the limitations of rate equations            Acta 63, 530-532
in describing the order of ligand binding for ter-           Koster, J. F. & Veeger, C. (1968) Biochim. Biophys. Acta
nary-complex systems when the individual associ-                151, 11-19
ation and dissociation rate constants are not                Liddell, J. M., Swan, I. D. A. & Shaw, W. V. (1978) J.
known. Therefore, although a preferred-order                   Mol. Biol. 124, 285-286
                                                             Lineweaver, H. & Burk, D. (1934) J. Am. Chem. Soc. 56,
pathway for the forward reaction seems a likely                658-666
conclusion for CAT,,,, this may not obtain for all           Nakagawa, Y., Nitahara, Y. & Miyamura, S. (1979)
concentrations of chloramphenicol or reaction                  Antimicrob. Agents Chemother. 16, 719-723
conditions different from those employed in the              Packman, L. C. (1978) Ph.D. Thesis, University of
present study.                                                 Leicester
                                                             Packman, L. C. & Shaw, W. V. (198 la) Biochem. J. 193,
  We are deeply indebted to Dr. Paul Engel for his ad-
                                                             Packman, L. C. & Shaw, W. V. (1981b) Biochem. J. 193,
vice in the early stages of this work and his criticism of      541-552
the manuscript. We also thank Dr. George Miller              Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-672
(Schering Corporation) for the gift of the 3-fluoro          Shaw, W. V. (1967) Antimicrob. Agents Chemother. 1966
analogue of chloramphenicol, This work was supported            221-226
by a research project grant to W. V. S. from the Medical     Shaw, W. V. (1975) Methods Enzymol. 43, 737-755
Research Council.                                            Shaw, W. V. (1983) CRC Crit. Rev. Biochem. 14, 1-46
                                                             Shaw, W. V. & Unowsky, J. (1968)J. Bacteriol. 95, 1976-
References                                                   Simon, E. J. & Shemin, D. (1953) J. Am. Chem. Soc. 75,
Andersson, P., Kvassman, I., Oden, B. & Pettersson, G.       Slater, E. C. (1955) Disc. Faraday Soc. 20, 231-240
  (1984) Eur. J. Biochem. 139, 519-527                       Thibault, G., Guitard, M. & Daigneault, R. (1980)
Bennett, A. D. & Shaw, W. V. (1983) Biochem. J. 215,            Biochim. Biophys. Acta 614, 339-349
  29-38                                                      Trayser, K. A. & Colowick, S. P. (1961) Arch. Biochem.
Burke, J. T., Wargin, W. A. & Blum, M. R. (1980) J.             Biophys. 94, 161-168
  Pharmacol. Sci. 69, 909-912                                Williamson, J. R. & Corkey, B. E. (1969) Methods
Colowick, S. P. & Womack, F. C. (1969) J. Biol. Chem.           Enzymol. 13, 501-506
  244, 774-777                                               Zaidenzaig, Y. & Shaw, W. V. (1976) FEBS Lett. 62,
Dalziel, K. (1957) Acta Chem. Scand. 11, 1706-1723             266-271
Dalziel, K. & Dickinson, F. M. (1966) Biochem. J. 100,       Zaidenzaig, Y. & Shaw, W. V. (1978) Eur. J. Biochem. 83,
   34-46                                                        553-562


Shared By: