Relationships between structure, function and stability for pyridoxalRichard Griessler, Barbara Psik, Alexandra Schwarz and Bernd Nidetzky Institute of Biotechnology and Biochemical Engi
Trang 1Relationships between structure, function and stability for pyridoxal
Richard Griessler, Barbara Psik, Alexandra Schwarz and Bernd Nidetzky
Institute of Biotechnology and Biochemical Engineering, Graz University of Technology, Austria
Using 0.4M imidazole citrate buffer (pH 7.5) containing
0.1 mM L-cysteine, homodimeric starch phosphorylase from
Corynebacterium calluane (CcStP) was dissociated into
native-like folded subunits concomitant with release of
pyridoxal 5¢-phosphate and loss of activity The inactivation
rate of CcStP under resolution conditions at 30C was,
respectively, four- and threefold reduced in two mutants,
Arg234fiAla and Arg242fiAla, previously shown to cause
thermostabilization of CcStP [Griessler, R., Schwarz, A.,
Mucha, J & Nidetzky, B (2003) Eur J Biochem 270, 2126–
2136] The proportion of original enzyme activity restored
upon the reconstitution of wild-type and mutant
apo-phos-phorylases with pyridoxal 5¢-phosphate was increased up to
4.5-fold by added phosphate The effect on recovery of
activity displayed a saturatable dependence on the
phos-phate concentration and results from interactions with the
oxyanion that are specific to the quarternary state
Arg234fiAla and Arg242fiAla mutants showed,
respect-ively, eight- and > 20-fold decreased apparent affinities for phosphate (Kapp), compared to the wild-type (Kapp 6 mM) When reconstituted next to each other in solution, apo-protomers of CcStP and Escherichia coli maltodextrin phosphorylase did not detectably associate to hybrid dimers, indicating that structural complementarity among the dif-ferent subunits was lacking Pyridoxal-reconstituted CcStP was inactive but 60% and 5% of wild-type activity could
be rescued at pH 7.5 by phosphate (3 mM) and phosphite (5 mM), respectively pH effects on catalytic rates were dif-ferent for the native enzyme and pyridoxal-phosphorylase bound to phosphate and could reflect the differences in
pKavalues for the cofactor 5¢-phosphate and the exogenous oxyanion
Keywords: apo-phosphorylase; a-glucan; glycogen; malto-dextrin; pyridoxal 5¢-phosphate
Structure–function relationship studies of a-glucan
phos-phorylases (GP) have a rich history in biochemical
litera-ture It is well established that pyridoxal 5¢-phosphate (PLP)
is the essential cofactor in all known GPs [1] PLP is bound
via a Schiff base between its aldehyde group and a
conserved lysine side chain in the active site [1,2] The
5¢-phosphate group is a main catalytic component of PLP
and is required for GP activity [2] The functional oligomeric
state of GP is dimeric [3–5] It has been shown that
dissociation of the subunits under localized denaturing
conditions exposes PLP to solvent PLP is released from the
enzyme and the activity is lost [6–8] Apo-phosphorylase can
be reconstituted, either with PLP or a range of structural
analogues thereof [2,9,10] Whereas restoration of enzyme activity upon the apofiholo conversion is determined by cofactor structure, the process of dimerization is relatively indiscriminate in respect to structural modifications of PLP Induction of structural complementarity of the interacting subunits such that they are able to recognize each other and associate to dimers is correlated with enzyme–cofactor bond formation [5,9] In a thorough investigation, Helmreich and colleagues prepared a series of hybrid phosphorylases in which one subunit contained PLP while the other was bound to an inactive cofactor analogue [5] They concluded that intersubunit contacts were also needed to elicit activity
in a potentially active holo-monomer
With very few exceptions [11,12], the results just sum-marized were obtained with a single enzyme, GP from rabbit muscle (RmGP) The activity of RmGP is under the control of allosteric and covalent regulatory mechanisms which are different or completely lacking in a large group of GPs from plants and microorganisms We therefore asked the question, what novel information might be gained by applying the same type of reconstitution experiments described for RmGP to another phosphorylase from a different source with different regulatory properties? While active-site residues are almost invariant in members of the
GP family, the dimer interfaces have been quite variable during the evolution in respect to the specific interproto-meric contacts, as revealed by comparative 3D structural
Correspondence to B Nidetzky, Institute of Biotechnology and
Bio-chemical Engineering, Graz University of Technology, Petersgasse 12/
I, A-8010 Graz, Austria Fax: +43 316 873 8434,
Tel.: +43 316 873 8400, E-mail: bernd.nidetzky@tugraz.at
Abbreviations: GP, glycogen phosphorylase; EcMalP, Escherichia coli
maltodextrin phosphorylase; CcStP, Corynebacterium callunae starch
phosphorylase; PLP, pyridoxal 5¢-phosphate; PL, pyridoxal;
RmGP, rabbit muscle GP.
Enzyme: a-glucan phosphorylase or a-1,4- D
-glucan:orthophosphate-a- D -glucosyltransferase (EC 2.4.1.1).
(Received 25 March 2004, revised 21 June 2004,
accepted 22 June 2004)
Trang 2[13] and structure-based sequence analyses [14,15] The
overall contact pattern at the subunit interfaces of different
regulated and nonregulated GPs is however, well preserved
[13] Thus one would like to know what directs subunit
interactions towards the induction of full enzymatic activity
and optimum stability in a dimer of phosphorylase This is a
significant and central problem to the study of catalysis by
GPs and oligomeric enzymes in general where the individual
subunits seem to possess all of the requisite chemical
functions but are in a catalytically inactive and unstable
conformation The detailed examination of the steps
involved in subunit dissociation and reassociation will
contribute to a better understanding of the dimerization
process per se and the role of interprotomeric contacts to
generate a functional enzyme The utilization of a
phos-phorylase devoid of the complex regulatory mechanisms
seen in RmGP allows the analysis to be strictly focused on
catalytic activity and stability
We chose starch phosphorylase from Corynebacterium
callunae(CcStP), which has been characterized
biochemi-cally and structurally [15,16], for particular reason The
intersubunit contacts stabilizing the functional CcStP dimer
are strengthened by > 100-fold when oxyanions such as
phosphate bind to this enzyme [17] Enzyme–oxyanion
interactions occur at a protein site different from the active
site, and thermostabilization is the result of a protein
conformational change induced by the binding event
Residues involved in the structural rearrangement are
located within the predicted dimer contact region of CcStP
[15] Reversible subunit dissociation experiments should
thus be useful to explore structural requirements for the
phosphate effect on CcStP stability
We report here the preparation of apo-CcStP and the
characterization thereof in respect to structural properties
and kinetic stability The process of reconstitution with PLP
has been analyzed using CcStP and four site-specific mutants
in which amino acid replacements within the dimer contact
region have led to altered oxyanion-dependent kinetic
stabilities [15,18] The relative timing of steps involved in
dimer formation and appearance of thermostabilization by
phosphate has been examined The role of the cofactor
5¢-phosphate group in the induction of stability and
stabil-ization of the CcStP dimer has been studied Subunit
complementation experiments are reported which were
designed to detect formation of possible hybrid dimers of
CcStP and maltodextrin phosphorylase from Escherichia coli
(EcMalP) Finally, we show results from kinetic studies of
CcStP reconstituted with pyridoxal (PL), a cofactor analogue
in which the original 5¢-O-PO32–group is replaced by 5¢-O-H
Materials and methods
Enzymes, substrates and other materials
Recombinant CcStP and site-directed mutants thereof were
produced as described elsewhere [15,18] Natural CcStP was
purified by a reported procedure [16] If not stated
otherwise, recombinant CcStP was used EcMalP was
prepared according to Eis et al [19] Analytical enzymes
and enzyme substrates were specified in previous papers
[15–18] All other chemicals were of reagent grade and
obtained from Sigma and Fluka
Preparation of apo-Cc StP and apo-Ec MalP Screening for buffer conditions in which apo-CcStP could
be prepared, led to selection of 0.4Mimidazole citrate and 0.1Mcysteine hydrochloride, in short, the resolution buffer Various pH values between 5.0 and 8.0 were tested, and a
pH of 7.0 was chosen (see below) Prior to the resolution, CcStP and site-directed mutants thereof were doubly gel filtered using NAP 5 or NAP 10 columns (Amersham Biosciences) to remove phosphate from storage stock solutions to an end concentration below 0.1 mM The enzymes were incubated in the resolution buffer at 30C using protein concentrations in the range 0.5–2.0 mgÆmL)1 until the residual activity was between 1.5 and 2.5% of the original level The resolution buffer was then replaced by a
50 mMtriethanolamine buffer, pH 7.0, using gel filtration with a NAP 5 column Separate control experiments for wild-type CcStP showed that the fourfold variation in protein concentration in our experiments was not an important factor of the rate of resolution
Apo-EcMalP was prepared using a protocol developed
by Palm and coworkers (D Palm, Theodor-Boveri-Institut fu¨r Biowissenschaften, Universita¨t Wu¨rzburg, Germany; personal communication) The enzyme was diluted to
2 mgÆmL)1 in 50 mM Mes buffer, pH 7.0, containing
25 mMKCl and 2 mMdithiothreitol An equal volume of
1Mcysteine hydrochloride dissolved in the same buffer was added to give a final concentration of 0.5M Resolution was obtained by adjusting the pH with HCl to a value of 5.05 at
4C The enzyme was incubated under these conditions until the residual activity was about 1.5% of the original level Apo-EcMalP was precipitated by ammonium sul-phate at 65% saturation, and the pellet was resuspended in
50 mMpotassium phosphate buffer, pH 7.0
The time course of apo-phosphorylase formation was monitored by using a number of methods [17]: enzyme activity measurements using samples taken from the incu-bation mixture; column sizing experiments to determine the subunit association state of the protein; CD spectroscopic measurements; determination of protein-bound and disso-ciated PLP This latter measurement was performed after ultrafiltration of the sample using 30 kDa cut off micro-concentrator tubes The PLP content of the protein-containing retentate was measured using both semiquantitative fluorometric analysis and a quantitative spectrophotometric test [17] The filtrate, which was devoid
of protein, was the subject of quantitative analysis for PLP content
Apo-phosphorylases were always prepared for immediate further use and not stored for longer than about 2 h at 4C Appropriate control measurements showed that the inacti-vation of apo-enzymes was not significant under these conditions
Reconstitution of apo-phosphorylases Apo-phosphorylase of CcStP (about 0.1–0.4 mgÆmL)1) was brought to 50 mMtriethanolamine buffer, pH 7.0, contain-ing a concentration of potassium phosphate between < 0.05 and 80 mM PLP at a concentration of between 0.0 and
100 lM was added to reconstitute the holo-enzyme The reaction was carried out at 30C and typically, the time
Trang 3course of recovery of enzyme activity was monitored up to
180 min When addition of fresh PLP did not further
enhance the regain of activity, reconstitution was considered
to be exhaustive Reconstituted CcStP was characterized in
respect to its structural properties using CD spectroscopy,
cofactor fluorescence and analytical gel filtration using
Superose 12 HR 10/30 (see below) Kinetic parameters of
the direction of a-glucan phosphorolysis and synthesis were
determined as described below Reconstitution of
apo-EcMalP was performed at 30C in 50 mM potassium
phosphate buffer, pH 7.0, and incubation was carried on
4 h after addition of 100 lMPLP
Using the conditions described above, a reconstitution
experiment was carried out in which apo-CcStP
(0.35 mgÆmL)1 of the natural enzyme) and apo-EcMalP
(1.35 mgÆmL)1) were incubated with 100 lMPLP next to
each other in solution Therefore, heterodimerization would
have been possible, and the aim was to either detect it or rule
out its occurrence under the conditions used The protein
solution was loaded on to a 5 mL Econo-Pac column of
ceramic hydroxylapatite type II (Bio-Rad) equilibrated with
50 mM potassium phosphate buffer, pH 6.8 Elution was
carried out at room temperature with a step gradient of 1M
potassium phosphate buffer, pH 6.8, at a flow rate of
40 cmÆh)1 Fractions containing protein were collected,
concentrated using ultrafiltration microconcentrator tubes,
and gel filtered using NAP 10 columns Characterization of
the fractions was carried out in respect to: the N-terminal
sequence determined by automated Edman degradation;
stability at 50C when 0.3Mpotassium phosphate (pH 7.0)
was present; and kinetic parameters for phosphorolysis of
maltohexaose (Sigma) at 30C
Enzyme kinetic measurements
Phosphorylase activity was measured in the direction of
a-glucan phosphorolysis using a continuous,
phosphoglu-comutase and NAD+-dependent glucose 6-phosphate
dehydrogenase-coupled spectrophotometric assay,
des-cribed in more detail elsewhere [16] If not mentioned
otherwise, maltodextrin 19.4 (Agrana, Gmu¨nd, Austria) was
the a-glucan substrate Initial rates of a-glucan
phosphoro-lysis and synthesis were recorded with discontinuous assays,
as reported previously [16] Linear plots of product
concentration vs time were converted into rates Kinetic
parameters were obtained from nonlinear fits of initial rate
data to Eqn (1) using theSIGMAPLOTprogram (SPSS Inc.,
Chicago, IL, USA),
v¼ kcat½E½S=ðKmþ ½SÞ ð1Þ
where v is the initial rate, kcatis the turnover number, [E] is
the molar concentration of enzyme active sites (based on the
stoichiometry of PLP and enzyme subunit), Km is an
apparent Michaelis constant, and [S] is the substrate
concentration When inhibition at high [S] was observed,
Eqn (2) was used:
v¼ kcat½E½S=ðKmþ ½S þ ½S2=KiSÞ ð2Þ
where KiSis the substrate inhibition constant
pH effects of enzyme-catalyzed initial rates were recorded
at 30C in 0.1Msodium acetate buffer in the pH range 5.0–
8.0 If not indicated otherwise, it was proved that enzyme inactivation during the time of the discontinuous assay ( 15 min) was not a source of an observable pH depend-ence of activity pH profiles were fitted to Eqn (3),
log rate¼ log½C=ð1 þ Ka=½HþÞ ð3Þ where C is the pH-independent value of the rate, Ka is a macroscopic acid dissociation constant, and [H+] is the proton concentration Equation (3) implies a pH profile that is level below pKaand decreases above pKawith a slope
of)1
Stability of apo-phosphorylase Apo-phosphorylase ( 0.2 mgÆmL)1) was incubated in 0.1Msodium acetate buffer, pH 6.9, at 22C At certain times between 0.2 and 20 h, samples were taken from the reaction mixture, PLP (40 lM) and potassium phosphate (50 mM) were added, and reconstitution was allowed to proceed for up to 4 h before recovered enzyme activity was measured The activity of the reconstituted phosphorylase
at zero incubation time served as the control A number of compounds were tested in respect to a potential stabilization
of apo-phosphorylase, and they were added in the concen-trations shown under Results Pyridoxin 5¢-phosphate was prepared by reduction of PLP with NaBH4 Control experiments were carried out in which pyridoxin 5¢-phos-phate (2 mM) was incubated at 30C with apo-phosphory-lase and regain of activity was recorded over time The total lack of recovery of activity proved that the reduction of PLP was complete
Structural characterization
CD spectroscopic measurements were carried out with a Jasco J-600 spectropolarimeter using quartz cuvettes of 0.1 cm pathlength Spectra of protein samples ( 0.1 mgÆmL)1) were recorded at 23 ± 1C in the range 200–240 nm If not mentioned otherwise, a 50 mM potas-sium phosphate buffer, pH 7.0, was used Column sizing experiments were carried out with Superose 12 HR 10/30 (22 mL bed volume) using a 50 mMpotassium phosphate buffer, pH 7.0, containing 200 mM NaCl and 0.1% (w/v) NaN3 Approximately 200 lg of protein dissolved in 0.5– 1.0 mL of buffer were loaded on to the column, and elution of protein was detected at 280 nm using an A¨ktaexplorer system (Amersham Biosciences) Fluores-cence measurements were performed with a Hitachi F-2000 spectrofluorometer using Hellma QS 101 cuvettes The excitation wavelength was set to 330 nm, and emission spectra were recorded in the range 360–600 nm Typically, a protein concentration of 0.4 mgÆmL)1 dissolved in triethanolamine buffer, pH 7.0, was used
Results
Preparation and characterization of apo-Cc StP Apo-CcStP was obtained at a practically useful rate by incubating CcStP in concentrations of between 0.5 and 2.0 mgÆmL)1in 0.4M imidazole citrate buffer, pH¼ 6.8,
Trang 4containing 0.1M L-cysteine hydrochloride at 30C Loss of
enzyme activity served as the reporter of formation of the
apo-enzyme under these conditions Semi-logarithmic plots
of the fraction of remaining active CcStP against time were
linear, suggesting that inactivation can be approximated by
a pseudo first-order model The half-life of the
holo-phosphorylase was 60 min at pH 7.0 The inactivation
rate was pH-dependent and decreased at pH values below
6.5 No significant loss of activity was observed at pH 5.0–
5.5 over 1.5 h When 50 mM potassium phosphate or
potassium sulphate was present in the buffer, pH 7.0,
formation of apo-phosphorylase was not detected over a
24 h long incubation time, indicating a half-life of 100 h or
greater Therefore, stabilization of the native dimer
struc-ture by the oxyanions must be > 100-fold (¼ 100/1), in
good agreement with previous results on the
thermostabi-lization of CcStP [15,17,18]
Column sizing experiments revealed that the
apo-phos-phorylase is a monomer It does not contain bound PLP
within limits of detection of the denaturing
spectrophoto-metric assay (± 2%) It completely lacks the characteristic
fluorescence emission of the cofactor in native CcStP which
occurs in the wavelength range 480–560 nm (see later)
Typically, apo-phosphorylases of CcStP and mutants
thereof contained equal to 2% of the original enzyme
activity which can be detected before and after the gel
filtration to replace the resolution buffer
Figure 1 shows the time course of inactivation of
apo-CcStP at 22C in the absence and presence of potential
stabilizers The half-life of apo-phosphorylase was
approxi-mately 15 h, and we observed only small effects on stability
of added phosphate, sulphate, and the cofactor derivative
pyridoxin 5¢-phosphate By contrast, UDP-a-D-glucose
conferred substantial extra stability to CcStP ADP-a-D
-glucose stabilized apo-CcStP to about the same extent as
UDP-a-D-glucose (not shown) Gel filtration analysis of apo-CcStP was carried out under conditions in which UDP-a-D-glucose (1 mM) was added to the elution buffer The apo-enzyme eluted as a single protein peak and with a retention time expected for a monomer of 90–100 kDa Therefore, the stabilizing effect of UDP-a-D-glucose is clearly not due to formation of an apo-oligomer induced by the binding of the nucleotide sugar The presence of maltopentaose (5 mM) resulted in a moderate 1.5-fold increase in the half-life of apo-CcStP
Effects of mutations in the dimer contact region
on the rate of apo-enzyme formation The pseudo first-order rate constants of inactivation in resolution buffer at pH 7.0 were determined for CcStP and five mutants thereof, using straight-line fits of the data plotted as logarithmic fraction of residual activity vs time The results are summarized in Table 1 Comparison of rate constants shows that the effect of the mutation may be stabilizing (R234A, R242A), neutral (S238A, S224A), or destabilizing (R226A), compared to the wild-type Except for R226A and R242A mutants (Table 1), all enzymes were stable for 2 h in the presence of 5 mMpotassium phosphate and potassium sulphate
Reconstitutions with PLP of apo-Cc StP and mutants thereof, and characterization of the wild-type holo-enzyme
Incubation of apo-CcStP (0.2 mgÆmL)1; 2.2 lM enzyme subunits) at 30C in 50 mMtriethanolamine buffer, pH 7.0, containing 50 mM potassium phosphate led to a gradual regain of enzyme activity in a PLP concentration-dependent manner Nine levels of PLP between 2 and 100 lM were tested, and the activity recovered after a 90 min incubation (which was shown to be exhaustive) displayed a saturatable dependence on [PLP], with half-saturation being attained at
KPLP¼ 19 ± 2 lM The recovery of activity when no PLP was added was not significant within the experimental error (± 1–2%) To prevent nonspecific reactions of the aldehyde group of PLP with protein lysines other than Lys634, a concentration of 2· KPLPwas chosen for standard recon-stitution
Column sizing experiments revealed that reconstituted CcStP existed exclusively as a dimer CD and cofactor fluorescence emission spectra of native and reconstituted
Fig 1 Stability and stabilization of apo-CcStP The apo-enzyme
( 0.2 mgÆmL)1) was incubated at 22 C in 0.1 M sodium acetate
buffer, pH 6.9 Incubations were carried out without additive (d);
5 m M potassium phosphate (s); 5 m M sodium sulphate (.); 2 m M
pyridoxin 5¢-phosphate (,); and 5 m M UDP-a- D -glucose (j) Activity
in samples taken at the times indicated was measured after
reconsti-tution with 40 l M PLP and 50 m M potassium phosphate as described
under Materials and methods.
Table 1 Half-lives (t 1/2 ) of CcStP and mutants thereof in the resolution
buffer at 30 °C and pH 7.0 Stable, no inactivation with 2 h of
in-cubation.
Protein
t 1/2 (min)
No oxyanion 5 m M Sulphate 5 m M Phosphate
Trang 5CcStP and apo-CcStP are shown in Fig 2 The CD spectra
of the three proteins are very similar overall, indicating
similarity in respect to the relative composition of secondary
structural elements However, the characteristic minima in
ellipticity at 208 nm and 222 nm have greater intensities in
the native enzyme, suggesting partial loss of a-helical
structure in apo-CcStP and reconstituted holo-CcStP Data
presented in Fig 2B proves that PLP is incorporated into
apo-CcStP during reconstitution However, the intensity of
cofactor fluorescence in the reconstituted enzyme is
approximately 65% that observed in CcStP, and this
difference agrees with differences in specific activities of
native and reconstituted phosphorylase Likewise, cofactor stoichiometry is decreased from a value of 1 in the wild-type to 0.6 in the reconstituted enzyme Apparent Michaelis constants of reconstituted CcStP were determined
in 50 mM triethanolamine buffer, pH 7.0, for phosphate (4.0 ± 0.3 mM); and maltodextrin (3.9 ± 0.4 mM) in the direction of phosphorolysis; a-D-glucose 1-phosphate (1.0 ± 0.1 mM); and maltodextrin (33 ± 5 mM) in the direction of synthesis After correction of turnover numbers for the fraction of active enzyme in holo-phosphorylase, native and reconstituted CcStP are not distinguishable in regard to their kinetic properties
The time courses of recovery of enzyme activity upon reconstitution of wild-type and mutant apo-phosphorylases with 40 lM PLP were biphasic During the initial burst phase which was complete within 5 min, there appeared up
to 80% of the total enzyme activity recoverable under the conditions In the second phase, enzyme activity increased slowly to its final level and eventually decreased again Figure 3 shows typical profiles of regain of activity vs time
of reconstitution, obtained with the R226A mutant in the absence and presence of potassium phosphate In all cases except for the R242A mutant, the yield of enzyme activity (compared to the original level before resolution and expressed as a percentage thereof) was increased by added phosphate (Table 2) The effect of phosphate was composed
of two components: first, a shift of apparent equilibrium for the reconstitution reaction towards the active enzyme and second, a stabilization of the reconstituted holo-enzyme against inactivation (which was shown to be irreversible)
We compared recovery of activity of the wild-type under conditions in which phosphate (50 mM) was present from the beginning of the reconstitution or was added at the end
of the burst phase (5 min) The yield was the same in both experiments within the experimental error The recovery of activity showed a saturatable dependence on the phosphate concentration Half-saturation constants for phosphate (KdPi) were obtained from nonlinear fits of values of final
Fig 2 Comparison of spectral properties of native CcStP, apo-CcStP,
and reconstituted enzyme using CD (A) and fluorescence (B) Spectra
were recorded using approximately the same protein concentration
(0.1 mgÆmL)1± 5%) in each case (A) Spectra of the native CcStP
(j), the apo-CcStP (d), and the enzyme after exhaustive reconstitution
in the presence of 100 l M PLP (,) (B) The fluorescence emission
spectra are shown for native enzyme (––), apo-CcStP (ÆÆÆÆ), and
reconstituted enzyme (- - -) The excitation wavelength was constant at
330 nm In (A) and (B), the reconstituted enzyme showed 65% of
the original activity A 50 m M potassium phosphate buffer, pH 7.0,
was used.
Fig 3 Reconstitution of apo-enzyme of R226A mutant The assays contained 0.22 mgÆmL)1protein and used 40 l M PLP Other condi-tions are reported under Materials and methods The symbols show the different concentrations of phosphate in m , as indicated.
Trang 6recovered activity to Eqn (4) and are summarized in
Table 2 They reveal marked decreases in the apparent
affinities of the R234A and R242A mutants for phosphate,
compared to wild-type
DEA¼ DEAmax½Pi=ðKdPiþ ½PiÞ ð4Þ
where DEA is the difference in recovered enzyme activity in
the presence and absence of phosphate, and DEAmaxis the
maximum value for DEA when phosphate is saturating
Reconstitutions of apo-Cc StP and apo-EcMalP next
to each other in solution
Figure 4 shows fractionation by hydroxylapatite
chroma-tography of a protein mixture obtained by reconstitutions of
apo-CcStP and apo-EcMalP under conditions that might enable subunit complementation to form a hybrid phos-phorylase Through elution with an increasing phosphate concentration, two major fractions A and C were isolated which together accounted for more than 95% of the total protein loaded on to the column It is noteworthy that fractions A and C eluted exactly as expected for native CcStP and EcMalP, respectively Likewise, CcStP and EcMalP prepared by reconstitution of the corresponding apo-phosphorylases independent of one another displayed
70% of their original phosphorylase activities and eluted exactly as the native enzymes did (data not shown) Figure 4 shows that a minor fraction B was also obtained Like fractions A and C, it contained phosphorylase activity Control experiments showed that under the conditions used, the fractionation of reconstituted EcMalP may yield a small fraction B depending on the applied amount of protein Protein fractions A–C were characterized functionally and structurally, as summarized in Table 3
Production and characterization of PL-reconstituted CcStP
PL could replace PLP in the reconstitution of apo-CcStP The formation of PL-phosphorylase after an exhaustive incubation time of 4 h showed a saturatable dependence
on PL concentration, the optimum level of PL being approximately 250 lM Addition of PLP (40 lM) after a 4 h incubation of apo-CcStP (0.3 mgÆmL)1) in the presence of
PL (250 lM) did not restore further enzyme activity, suggesting that reconstitution with PL was complete PL-phosphorylase was as stable as the native enzyme or PLP-reconstituted CcStP at 60C in 300 mM potassium phosphate buffer, pH 7.0 Therefore, the cofactor phos-phate group is not a component of oxyanion-dependent thermostabilization of CcStP
When assayed in the direction of a-glucan synthesis at
30C (using conditions described in Fig 5),
PL-phosphory-Table 2 Effect of phosphate on recovered enzyme activity during
reconstitution of apo-enzymes of wild-type CcStP and mutants thereof
with 40 l M PLP A 50 m M triethanolamine buffer, pH 7.0, was used.
K dPi is the half-saturation constant for phosphate The protein
con-centrations used varied in the range 1–4 l M of apo-enzyme (90 kDa)
and were ‡ 10· the concentration of cofactor Control experiments
carried out with the wild-type showed that the yield of reconstituted
enzyme activity did not change as result of this variation in protein
concentration The values in parentheses show the yield of recovered
enzyme activity when no phosphate was present ND, not determined,
because no significant dependence of recovered enzyme activity on
[phosphate] was seen in the range 0–80 m M
Protein K dPi (m M )
Recovered enzyme activity (%)
Fig 4 Fractionation by hydroxylapatite chromatography of a protein
mixture obtained by reconstitution of apo-CcStP and apo-EcMalP The
protein elution profile, recorded by absorbance at 280 nm, is shown.
The dashed line indicates the elution gradient used See Materials and
methods for details.
Table 3 Characterization of protein species obtained through chroma-tographic fractionation of a mixture of apo-CcStP and apo-EcMalP reconstituted with 100 l M PLP next to each other in solution Figure 4 gives details of the fractionation Fractions are labeled according to Fig 4 K mG6 and K iG6 were obtained from nonlinear fits to Eqn (2) of the initial rate data recorded at a constant saturating concentration of
50 m M P i K mG6 and K iG6 are the apparent Michaelis constant and the substrate inhibition constant for maltohexaose, respectively Half-life (t 1/2 ) incubations were carried out at 50 C in 300 m M potassium phosphate buffer, pH 7.0.
Fraction A Fraction B Fraction C
K mG6 (m M ) 2.65 ± 0.35 0.71 ± 0.07 0.76 ± 0.10
K iG6 (m M ) 360 ± 130 31.8 ± 3.1 21.9 ± 2.3
N-terminal sequence
P-E-K-Q-P-L-P-A-A a X-Q b (S)-Q-P-(I) c
a Residue Ser1 is processed off in CcStP isolated from the natural organism [15,16].bX is an unidentified amino acid.c Determin-ation of the N-terminal sequence of fraction C was not completely clear at positions 1 and 4.
Trang 7lase was inactive within the limits of detection of the
experimental procedures Addition of phosphate or
phos-phite restored phosphorylase activity, as shown in
Fig 5A,B, respectively The time course of formation of
phosphate was linear when phosphate was used as the
activator oxyanion The chosen level of phosphate (2.5 or
5 mM) did not influence the enzymic rate significantly
When phosphite was the activator oxyanion, the observed
time courses were concave upward, perhaps indicating an
autocatalytic effect of the released phosphate The reaction
rate recorded at an oxyanion concentration of 2 mMwas
4.4 times higher with phosphate than phosphite Table 4
summarizes the kinetic characterization of PL-CcStP The
restoration of activity in PL-phosphorylase by phosphate
displayed saturatable concentration dependence, and half-maximum activation was observed at 0.5 mM At
pH 7.5, about 57% of the wild-type level of activity could be recovered The Michaelis constant of the PL-enzyme for a-D-glucose 1-phosphate in the presence of phosphite was approximately 10 times that of CcStP
The pH-dependence of activity under conditions of saturation in both substrates was determined for CcStP and PL-phosphorylase in the pH range 5.0–8.0 Initial rates were recorded in the directions of a-glucan phos-phorolysis and synthesis, and assays for PL-phosphory-lase in the synthesis direction contained a saturating level
of activating phosphate (2.5 mM) Results are shown in Fig 6 In either direction of reaction, enzymatic rates which are effectively turnover numbers (kcat) decreased at high and low pH Optimum catalytic rates for phos-phorolysis were found at around pH 7.0 for both the native enzyme and PL-phosphorylase In the low pH region the pH profile of kcat for PL-phosphorylase was displaced outward by 1.0 pH unit, relative to the corresponding pH profile for CcStP The decrease in kcat (phosphorolysis; kpho) at high pH was similar for both enzymes In the synthesis direction, optimum conditions for kcat (ksyn) were observed at pH 6.0 for CcStP PL-enzyme bound to phosphate showed maximum activity at
pH 6.5–7.0 The pH profile of ksynfor PL-phosphorylase
in the presence of phosphate was displaced outward by
1.0 pH units at high pH, compared to the pH profile of
ksyn for wild-type CcStP Fits of the data to Eqn (3) yielded pKa values of 6.9 ± 0.3 and 7.9 ± 0.3 for wild-type enzyme and PL-CcStP, respectively
Discussion
Formation and characterization of apo-CcStP
A number of studies have identified prerequisites for reversible conversion of holo-GP into the apo-enzyme [2]: localized reversible denaturation promoting subunit disso-ciation; resolution of PLP through aldehyde-reactive com-pounds; and prevention of subunit aggregation In spite of these common characteristics, completely different proto-cols were needed for successful preparation of apo-enzymes
of RmGP [2], Solanum tuberosum (potato tuber) starch phosphorylase [11], and EcMalP (D Palm, unpublished data) Apo-CcStP was obtained under conditions compar-able to the ones used by Shaltiel et al [6] for resolution of RmGP; i.e using imidazole citrate and L-cysteine as structure-deforming and PLP-resolving reagents, respect-ively Interestingly, however, the pH dependence of the rate
of resolution was opposite in the two enzymes, CcStP being stable under the slightly acidic conditions It was proposed
by others [6–8] that the imidazolium ion is required for optimum resolution of RmGP at pH 6.0 In CcStP, imidazole obviously assists in locally disrupting the native structure but there was no evidence that its protonated form would be particularly effective Mutations within the dimer contact region of CcStP (Table 2; also [15,18]) had strong effects on the half-life of activity in resolution buffer Likewise, cofactor resolution was inhibited completely
in the presence of phosphate or sulphate These results are
in good agreement with the notion that weakening
Fig 5 Restoration of enzyme activity in PL-CcStP by exogenous (A)
phosphate and (B) phosphite Incubations were carried out at 30 C in
0.1 M sodium acetate buffer, pH 7.6, containing 30 lgÆmL)1protein.
The substrate levels were constant at 80 gÆL)1maltodextrin and 50 m M
a- D -glucose 1-phosphate The levels of exogenous activator oxyanion
are indicated by symbols and given in m M In (A) the concentrations of
released phosphate were sufficient to allow an accurate determination
of the activity in spite of the added phosphate The possible inhibition
of the enzymatic reaction by phosphate is compensated using a high
concentration of a- D -glucose 1-phosphate.
Trang 8subunit-to-subunit interactions in CcStP [15,17,18] is a key factor driving the resolution of the holo-enzyme
Like apo-RmGP, apo-CcStP is monomeric and displays
no enzyme activity A number of observations indicate that
it retains native-like tertiary structure Stabilization of apo-CcStP by UDP-a-D-glucose and ADP-a-D-glucose is par-ticularly relevant because it suggests the preservation of
a cofactor–substrate binding scaffold in apo-CcStP The nucleotide-activated sugars structurally resemble the noncovalent complex of PLP and a-glucose 1-phosphate that is formed at the phosphorylase active site in the course
of the enzymatic reaction [20,21] The available evidence from gel filtration analysis excludes the occurrence of a transient apo-dimer lacking phosphorylase activity, induced
by the presence of the stabilizing UDP-a-D-glucose
UDP-a-D-glucose at a level of 5 mMinhibits the reaction of native CcStP to less than 15%, suggesting the absence of a high-affinity effector site for nucleotide sugars in the active holo-phosphorylase dimer Furthermore, it does not retard the resolution of the cofactor in CcStP (data not shown), indicating that the observed stabilizing effect is specific to the apo-enzyme
Now, given that PLP resolution caused only minor denaturation of CcStP tertiary structure, it was especially interesting that thermostabilization of the holo-enzyme by phosphate was lost in apo-CcStP; and recovered fully upon reconstitution This result could indicate that in apo-CcStP (a) the actual oxyanion binding site was disrupted, or (b) a conformational change that accompanies oxyanion binding
in the holo-enzyme cannot take place Whatever was truly responsible, the data suggest that dimerization is required for restoration of oxyanion-dependent thermostabilization
of CcStP (see below)
Reconstitution of the holo-enzyme Reconstitution experiments were designed to address two specific questions of phosphorylase recognition First, do apo-phosphorylases of CcStP and EcMalP associate in solution to form hybrid dimers? Secondly, is there a role of interactions between protein and oxyanion during the apofiholo conversion of CcStP?
Complementation of phosphorylase apo-protomers in solution has obvious advantages over working with immo-bilized subunits, as described by others [5,7] However, it
Table 4 Kinetic characterization of PL-CcStP in the presence of activator oxyanion Initial rates were recorded in 50 m M Tris-acetate buffer, pH 7.5, using a discontinuous assay in which samples were taken after 20, 40 and 60 min of incubation The rates were calculated from linear plots of [P i ] released against the reaction time When phosphate was the activator oxyanion, initial rates were calculated from the difference between the concentrations of total phosphate at a certain incubation time and phosphate initially present In all cases this difference was sufficient to allow accurate determination of the enzymatic rate The values of v max for the native phosphorylase determined in the presence and absence of 10 m M
phosphite were identical within the experimental error, indicating weak (if any) inhibition by the added oxyanion Glc1P, a- D -glucose 1-phosphate;
MD, maltodextrin (dextrin equivalent 19.4).
Glc1P (m M ) or MD (gÆL)1) Activator oxyanion (m M ) v max (UÆmg)1) K m (m M ) PL-phosphorylase
Native phosphorylase
Fig 6 pH profiles in the direction of a-glucan synthesis (A) and
phosphorolysis (B) catalyzed by wild-type CcStP (d) and PL-CcStP (s)
activated by exogenous phosphate ions (A) Results were obtained in
0.1 M sodium acetate buffer containing 2.5 m M P i The substrate levels
were 80 gÆL)1maltodextrin and 50 m M a- D -glucose 1-phosphate Solid
lines are nonlinear fits of the data to Eqn (3) For PL-CcStP the
cata-lytic rate at pH 8 was not included in the calculation because its value
reflects the effects of pH on both rate and enzyme stability (B) Results
were obtained in 50 m M potassium phosphate buffer containing
80 gÆL)1maltodextrin The lines indicate the trend of the data.
Trang 9requires methods which select for true hybrids Mixtures of
reconstituted CcStP and EcMalP were separated by using
hydroxylapatite chromatography [19] Conditions were
used in which a hybrid would be clearly detectable if it
displayed intermediate binding properties, compared to
wild-type CcStP (weak binding) and EcMalP (strong
binding) The observed elution pattern from the
hydroxyl-apatite column was not consistent with the formation of
hybrids in substantial amounts However, a small protein
fraction was detected that eluted before and after the peaks
clearly assigned to native or reconstituted EcMalP and
CcStP, respectively This fraction contained enzyme activity
and obviously, it could be a phosphorylase hybrid
Furthermore, we had to consider the possibility that
heterodimers escape detection because the different subunits
interact with hydroxylapatite independently of one another
Therefore, the three protein fractions obtained (A–C) were
characterized by N-terminal sequencing and two parameters
of enzyme function distinguishing sensitively between CcStP
and EcMalP: (a) apparent substrate affinity and substrate
inhibition in the direction of phosphorolysis of
maltodex-trins; and (b) kinetic stability at 50C The results showed
that, within limits of detection of the fractionation
proce-dure (5%), only wild-type enzymes were present and no
hybrid dimers formed The observed small protein peak
(fraction B) very likely contains reconstituted EcMalP, and
its occurrence can be explained by an incomplete retention
of reconstituted EcMalP by the hydroxylapatite column It
seems that the structural complementarity between
pro-tomers of CcStP and EcMalP was not sufficient for the
different subunits to recognize each other This finding is
interesting because the packing of hydrophobic residues
dispersed over the main part of the dimer interface is highly
conserved among known a-glucan phosphorylases [22]
including EcMalP and, by sequence similarity, CcStP It
suggests that interfacial contacts mediated by polar groups
must be different in EcMalP and CcStP
We were interested to examine the relative timing of
steps involved in dimer formation and the appearance of
oxyanion-dependent stabilization of activity during
recon-stitution of apo-CcStP Analysis of time courses of
recovery of enzyme activity in the absence and presence
of phosphate showed that the yield but not the rate at
which the activity was regained was strongly dependent
on the added phosphate These observations are novel
and consistent with a mechanism in which the active
dimer is formed first, and enzyme–oxyanion interactions
that are lacking in the monomer are utilized to shift the
equilibrium towards the catalytically competent enzyme
(Scheme 1) The data are in excellent agreement with the
proposed pathway of thermal denaturation of CcStP [17]
and contribute to an improved understanding of the
effect of phosphate binding on the dimer stability of
CcStP The evidence presented here and summarized in
Scheme 1 significantly advances the mechanism
underly-ing oxyanion-dependent dimer stabilization because it was
possible for the first time to investigate the properties of
the native-like folded apo-monomer of CcStP Because
of its low conformational stability under conditions of
thermally induced dissociation of the CcStP subunits, the
apo-monomer usually escaped detection in the previous
studies of CcStP stability [17,18]
Reconstitution of mutant apo-enzymes yielded results that were fully consistent with recent comparisons of thermoinactivation rates of the same mutants [15,18] After correction for differences in protein concentration used, the level of activity recovered during the burst phase was similar among wild-type and all mutants when no phosphate was present Therefore, this implies that the mutations did not cause changes in the association rate of the phosphorylase subunits Altered kinetic stabilities of the mutants, compared to wild-type, are therefore likely due to changes in protomer dissociation rate The effect
of phosphate on the recovery of activity was sensitive to mutations in the dimer contact region R234A had lost much of the apparent affinity of the wild-type for phosphate, and a phosphate effect on activity recovery was lacking completely in R242A under the conditions used The data reinforce the conception [15] that the side chains of Arg234 and Arg242 have key roles in the mechanism by which phosphate binding induces a kinetically stabilized conformation of CcStP (Scheme 1)
Restoration of enzyme activity in PL-reconstituted phosphorylase by exogenous phosphate
The characterization of CcStP reconstituted with PL in place of the natural cofactor PLP yielded results that are relevant in the context of function of the 5¢-phosphate group
in phosphorylase catalysis [1], as follows A number of studies using PL-RmGP have shown that the otherwise inactive PL-phosphorylase recovered up to 19% of wild-type activity when exogenous oxyanions were present Among a series of compounds tested phosphite was the most powerful activator anion of PL-RmGP [23,24] Using PL-CcStP, phosphate was 4.5-fold more effective than phosphite, and in saturating concentrations of 3 mM it restored 60% of the original enzyme activity at pH 7.5 The data suggest that phosphate binds to the cleft vacated in PL-CcStP through removal of the original cofactor 5¢-phos-phate group; and the positions of the dissociable phos5¢-phos-phate
in PL-CcStP and the covalently bound phosphate in the native enzyme are probably similar
Scheme 1 Formation of active dimers of CcStP during reconstitution of the apo-phosphorylase with PLP in the absence and presence of phos-phate M is the native-like folded monomer; M¢ is an irreversibly denatured monomer; D is the PLP-containing, active dimer; D* is the stabilized dimer bound to phosphate; M aggr is aggregated protein All monomeric forms lack enzyme activity The denaturation of D as shown is supported by evidence published elsewhere [17].
Trang 10The direct comparison of pH profiles for the catalytic
rates of CcStP and the complex PL-phosphorylase and
phosphate can arguably provide mechanistic information
because enzyme systems were analyzed whose active sites
differed only by a minimal modification However, any
interpretation must be tempered considering that in
RmGP, slightly different binding modes for
cofactor-bound and mobile phosphate groups have been detected
by X-ray crystallography [25] The question of interest
was whether differences in pKa values for covalent and
noncovalent phosphate (pKa¼ 7.2 [23]) groups are
mirrored in the corresponding pH-rate profiles The
pKa values of the cofactor phosphate in unliganded
EcMalP and the EcMalP–arsenate complex are 5.6 [26]
and 6.7 [27], respectively The pKa of the 5¢-phosphate
group in a model Schiff base is 6.2 [26] The available
evidence for EcMalP defines a range of plausible pKa
values for CcStP because residues interacting with the
5¢-phosphate group in EcMalP are completely conserved
in CcStP Log ksynfor the wild-type decreased above an
apparent pKa of 6.9 whereas a pKa value of 7.9 was
calculated from the pH profile of log ksyn for PL-CcStP
bound to phosphate Unfortunately, the activity of
PL-CcStP in the presence of activator phosphite was too low
to permit determination of a reliable pH profile The
observed DpKaof 1.0 pH units would agree reasonably
with DpKa¼ 1.2 predicted on the basis of pKavalues of
phosphate and the cofactor 5¢-phosphate in a model
compound These data are consistent with a
pH-depend-ent mechanism in which the cofactor phosphate must be
protonated so that catalysis to a-glucan synthesis occurs
[1,28,29] The pH profiles of log kpho for wild-type and
PL-CcStP decreased above an apparent pKa value of
7.3 It is not possible to assign this pKa value to the
pH-dependent ionization of a group on the reactive
enzyme–substrate complex; obviously it could reflect the
ionization of the substrate phosphate
Acknowledgements
The financial support from the Austrian Science Funds (P15118 and
P11898 to B.N.) is gratefully acknowledged We thank Dr Dieter Palm
for communicating a protocol for the preparation of apo-EcMalP.
References
1 Palm, D., Klein, H.W., Schinzel, R., Buehner, M & Helmreich,
E.J.M (1990) The role of pyridoxal 5¢-phosphate in glycogen
phosphorylase catalysis Biochemistry 29, 1099–1107.
2 Graves, D.J & Wang, J.H (1972) a-Glucan phosphorylases –
chemical and physical basis of catalysis and regulation Ann Rev.
Biochem 7, 435–482.
3 Feldmann, K., Zeisel, H.J & Helmreich, E.J.M (1976)
Com-plementation of subunits from glycogen phosphorylases of frog
and rabbit skeletal muscle and rabbit liver Eur J Biochem 65,
285–291.
4 Tu, J.-I & Graves, D.J (1973) Association-dissociation properties
of sodium borohydride-reduced phosphorylase b J Biol Chem.
248, 4617–4622.
5 Feldmann, K., Zeisel, H & Helmreich, E (1972) Interactions
between native and chemically modified subunits of matrix-bound
glycogen phosphorylase Proc Natl Acad Sci USA 69, 2278–
2282.
6 Shaltiel, S., Hedrick, J.L & Fischer, E.H (1966) On the role of pyridoxal 5¢-phosphate in phosphorylase II Resolution of rabbit muscle phosphorylase Biochemistry 5, 2108–2116.
7 Hedrick, J.L., Shaltiel, S & Fischer, E.H (1969) Conformational changes and the mechanism of resolution of glycogen phosphor-ylase b J Biol Chem 8, 2422–2429.
8 Pan, P., Schinzel, R., Palm, D & Christen, P (1993) Reaction of imidazole-citrate-deformed glycogen phosphorylase with amino acids Eur J Biochem 215, 761–766.
9 Pfeuffer, T., Ehrlich, J & Helmreich, E (1972) Role of pyridoxal 5¢-phosphate in glycogen phosphorylase II Mode of binding of pyridoxal 5¢-phosphate and analogs of pyridoxal 5¢-phosphate to apophosphorylase b and the aggregation state of reconstituted phosphorylase proteins Biochemistry 11, 2136–2145.
10 Shaltiel, S., Hedrick, J.L., Pocker, A & Fischer, E.H (1969) Reconstitution of apophosphorylase with pyridoxal 5¢-phosphate analogs Biochemistry 8, 5189–5196.
11 Shimomura, S., Emman, K & Fukui, T (1980) The role of pyr-idoxal 5¢-phosphate in plant phosphorylase J Biochem 87, 1043– 1052.
12 Tagaya, M., Shimomura, S., Nakano, K & Fukui, T (1982) A monomeric intermediate in the reconstitution of potato apopho-sphorylase with pyridoxal 5¢-phosphate J Biochem 91, 589–597.
13 Watson, K.A., Schinzel, R., Palm, D & Johnson, L.N (1997) The crystal structure of Escherichia coli maltodextrin phosphorylase provides an explanation for the activity without control in this basic archetype of a phosphorylase EMBO J 16, 1–14.
14 Hudson, J.W., Golding, G.B & Crerar, M.M (1993) Evolution of allosteric control in glycogen phosphorylase J Mol Biol 234, 700–721.
15 Griessler, R., Schwarz, A., Mucha, J & Nidetzky, B (2003) Tracking interactions that stabilize the dimer structure of starch phosphorylase from Corynebacterium callunae Eur J Biochem.
270, 2126–2136.
16 Weinha¨usel, A., Griessler, R., Krebs, A., Zipper, P., Haltrich, D., Kulbe, K.D & Nidetzky, B (1997) a-1,4- D -glucan phosphorylase
of gram-positive Corynebacterium callunae: isolation, biochemical properties and molecular shape of the enzyme from solution X-ray scattering Biochem J 326, 773–783.
17 Griessler, R., D’Auria, S., Tanfani, F & Nidetzky, B (2000) Thermal denaturation mechanism of starch phosphorylase from Corynebacterium callunae: oxyanion binding provides the glue that efficiently stabilizes the dimer structure of the protein Protein Sci.
9, 1149–1161.
18 Nidetzky, B., Griessler, R., Pierfederici, F., Psik, B., Scire, A & Tanfani, F (2003) Mutagenesis of the dimer interface region of Corynebacterium callunae starch phosphorylase alters the oxy-anion ligand-dependent conformational relay that enhances oli-gomeric stability of the enzyme J Biochem (Tokyo) 134, 599–606.
19 Eis, C., Griessler, R., Maier, M., Weinha¨usel, A., Bo¨ck, B., Hal-trich, D., Kulbe, K.D., Schinzel, R & Nidetzky, B (1997) Efficient downstream processing of maltodextrin phosphorylase from Escherichia coli and stabilization of the enzyme by immobilization onto hydroxyapatite J Biotechnol 58, 156–166.
20 Oikonomakos, N.G., Acharya, K.R., Stuart, D.I., Melpidou, A.E., McLaughlin, P.J & Johnson, L.N (1988) Uridine (5¢)diphospho(1)-a- D -glucose A binding study to glycogen phos-phorylase b in the crystal Eur J Biochem 173, 569–578.
21 Holm, L & Sander, C (1995) Evolutionary link between glycogen phosphorylase and a DNA modifying enzyme EMBO J 14, 1287–1293.
22 Lin, K., Hwang, P.K & Fletterick, R.J (1997) Distinct phos-phorylation signals converge at the catalytic center in glycogen phosphorylases Structure 5, 1511–1523.
23 Chang, Y.C., McCalmont, T & Graves, D.J (1983) Functions of the 5¢-phosphoryl group of pyridoxal 5¢-phosphate in