Results Expression and purification of mutant diol dehydratases Mutant apoenzymes in which Sera224 or Lysb135 was mutated to another amino acid were expressed in Esc-herichia coli cells
Trang 1coenzyme B12-binding site in diol dehydratase catalysis Ken-ichi Ogura, Shin-ichi Kunita, Koichi Mori, Takamasa Tobimatsu and Tetsuo Toraya
Department of Bioscience and Biotechnology, Graduate School of Natural Science and Technology, Okayama University, Japan
Adenosylcobalamin (AdoCbl) is a cofactor for
enzy-matic radical reactions, including carbon skeleton
rearrangements, heteroatom eliminations, and
intramo-lecular amino group migrations [1–3] These reactions
involve the migration of a hydrogen atom from one
carbon atom of the substrate to the adjacent carbon
atom [4,5] in exchange for group X, which moves in the
opposite direction [6] The reactions are initiated by
abstraction of a hydrogen atom from substrates with an
adenosyl radical that is generated in the active site through homolysis of the cobalt–carbon (Co–C) bond
of AdoCbl [1–3,7,8] The activation and homolysis of the Co–C bond upon coenzyme binding to apoenzyme is therefore considered to be a key step for all the AdoCbl-dependent reactions Diol dehydratase (dl-1,2-pro-panediol hydrolyase; EC 4.2.1.28) is an enzyme that catalyzes the AdoCbl-dependent conversion of 1,2-diols and glycerol to the corresponding aldehydes [9,10]
Keywords
adenine anchoring; adenosylcobalamin;
coenzyme B 12 ; diol dehydratase; ion pairing
Correspondence
T Toraya, Department of Bioscience and
Biotechnology, Graduate School of Natural
Science and Technology, Okayama
University, Tsushima-naka, Okayama
700-8530, Japan
Fax: +81 86 251 8264
Tel: +81 86 251 8194
E-mail: toraya@cc.okayama-u.ac.jp
(Received 4 September 2008, revised 8
October 2008, accepted 15 October 2008)
doi:10.1111/j.1742-4658.2008.06745.x
The X-ray structure of the diol dehydratase–adeninylpentylcobalamin com-plex revealed that the adenine moiety of adenosylcobalamin is anchored in the adenine-binding pocket of the enzyme by hydrogen bonding of N3 with the side chain OH group of Sera224, and of 6-NH2, N1 and N7 with main chain amide groups of other residues A salt bridge is formed between the e-NH2 group of Lysb135 and the phosphate group of cobala-min To assess the importance of adenine anchoring and ion pairing, Sera224 and Lysb135 mutants of diol dehydratase were prepared, and their catalytic properties investigated The Sa224A, Sa224N and Kb135E mutants were 19–2% as active as the wild-type enzyme, whereas the Kb135A, Kb135Q and Kb135R mutants retained 58–76% of the wild-type activity The presence of a positive charge at the b135 residue increased the affinity for cobalamins but was not essential for catalysis, and the introduction of a negative charge there prevented the enzyme–cobalamin interaction The Sa224A and Sa224N mutants showed a kcat⁄ kinact value that was less than 2% that of the wild-type, whereas for Lysb135 mutants this value was in the range 25–75%, except for the Kb135E mutant (7%) Unlike the wild-type holoenzyme, the Sa224N and Sa224A holoenzymes showed very low susceptibility to oxygen in the absence of substrate These findings suggest that Sera224 is important for cobalt–carbon bond activa-tion and for preventing the enzyme from being inactivated Upon inactiva-tion of the Sa224A holoenzyme during catalysis, cob(II)alamin accumulated, and a trace of doublet signal due to an organic radical disap-peared in EPR 5¢-Deoxyadenosine was formed from the adenosyl group, and the apoenzyme itself was not damaged This inactivation was thus considered to be a mechanism-based one
Abbreviations
AdePeCbl, adeninylpentylcobalamin; AdoCbl, adenosylcobalamin or coenzyme B 12 ; aqCbl, aquacobalamin; CN-Cbl, cyanocobalamin; OH-Cbl, hydroxocobalamin.
Trang 2Structure–function studies using adenine-modified
analogs of AdoCbl have shown relatively low
specific-ity of diol dehydratase for the adenine moiety in the
adenosyl group, an upper axial ligand [11–13] 1-Deaza
and 3-deaza analogs of AdoCbl are partially active
(56% and 46%, respectively) as coenzyme, whereas
7-deaza and N6,N6-dimethyl derivatives do not show
detectable coenzyme activity and act as strong
compet-itive inhibitors of AdoCbl Guanosylcobalamin is an
inactive coenzyme with low affinity for the enzyme
The nucleotide loop moiety of AdoCbl is not directly
involved in the catalytic process, but it is obligatory
for the continuous progress of catalytic cycles [14–16]
Adenosylcobinamide methyl phosphate, an analog of
AdoCbl lacking the nucleotide loop moiety, does not
show detectable coenzyme activity, but behaves as a
strong competitive inhibitor of AdoCbl [17] Upon
incubation with apoenzyme in the presence of
sub-strate, this analog undergoes irreversible cleavage of its
Co–C bond, forming an enzyme-bound
Co(II)-contain-ing species Adenosylcobinamide neither functions as a
coenzyme nor binds tightly to apoenzyme It is thus
evident that the phosphate group of the coenzyme
nucleotide loop is essential for tight binding to the
apoenzyme and therefore for subsequent activation of
the Co–C bond and catalysis
The X-ray structure of diol dehydratase showed that
the enzyme exists as a dimer of heterotrimers and
binds cobalamin in the ‘base-on’ mode, namely with a
5,6-dimethylbenzymidazole moiety coordinating to the
cobalt atom [18], as suggested by EPR studies [19,20]
The structure of the enzyme in complex with
ade-ninylpentylcobalamin (AdePeCbl), an inactive
coen-zyme analog, revealed the presence of the
adenine-binding pocket in the active site of the enzyme [21]
(Fig 1A) The adenine ring of the bound AdePeCbl is
nearly parallel to the corrin ring and faces pyrrole
ring C It is trapped in the pocket by several hydrogen
bonds with amino acid residues in the a-subunit
(Fig 1B) The overall structure of the complex is
essentially the same as that of the
enzyme–cyanocobal-amin (CN-Cbl) complex, except that the orientation of
the side chain OH group of Sera224 is largely rotated
to form a hydrogen bond with N3 of the adenine
moi-ety in the enzyme–AdePeCbl complex Sera224 is the
only residue whose side chain is hydrogen-bonded with
the adenine ring This residue is conserved between
diol dehydratases [22–24] and glycerol dehydratases
[25–27] Other residues also form hydrogen bonds with
6-NH2, N1, or N7, but through the main chain amide
groups [21] In addition, the e-NH2 group of Lysb135
forms a salt bridge with the phosphate group of
cobal-amin (Fig 1C) [18]
In this article, we report the roles of adenine anchor-ing and ion pairanchor-ing at the AdoCbl-bindanchor-ing site in diol dehydratase catalysis To study the functions of Sera224 and Lysb135 by site-directed mutagenesis, we prepared several mutant enzymes, in which either Sera224 or Lysb135 is mutated to other amino acids, and investigated their catalytic properties by kinetic and spectroscopic analyses The mechanism-based inactivation of a mutant enzyme during catalysis is also reported here
Results
Expression and purification of mutant diol dehydratases
Mutant apoenzymes in which Sera224 or Lysb135 was mutated to another amino acid were expressed in Esc-herichia coli cells and purified to homogeneity by the same procedure as that described for the wild-type enzyme [28] – that is, by extraction from crude mem-brane fractions with a buffer containing 1% Brij35 Purified preparations of the mutant enzymes were ana-lyzed by PAGE under denaturing and nondenaturing conditions Upon SDS⁄ PAGE (Fig 2A), three bands corresponding to the a-, b- and c-subunits were observed in each mutant enzyme Upon nondenaturing PAGE (Fig 2B), all the mutants were electrophoresed
as a single band that corresponds to the (abc)2complex
Catalytic activity and kinetic properties of Sera224 mutant diol dehydratases
As shown in Table 1, kcat values of the Sa224A and Sa224N mutants were 19% and 5%, respectively, of that of the wild-type enzyme, but rapid inactivation took place with the Sa224A mutant (Fig 3A) The
kcat⁄ kinact values, which show the average numbers of catalytic turnovers before inactivation [29], indicated that both Sera224 mutants were inactivated after 8000–15 000 turnovers on average, whereas the wild-type enzyme underwent inactivation after 7.5· 105 turnovers The result obtained with the Sa224A mutant suggests that the hydrogen bond donation from the side chain OH group of Sera224 to N3 of the adenine ring of AdoCbl is important for the continu-ous progress of catalytic cycles In the case of the Sa224N mutant, the hydrogen bond might not be formed, because the side chain –CH2CONH2 group of Asn in the Sa224N mutant is longer than the –CH2OH group of Ser in the wild-type enzyme
To examine the possibility that such differences might affect the affinities of the enzyme for ligands,
Trang 3apparent Km values for the coenzyme AdoCbl and a
substrate 1,2-propanediol, as well as Ki values for an
inhibitor, CN-Cbl, were determined with the mutants
Km values for AdoCbl (Table 2) and 1,2-propanediol
(Table 1) increased markedly for the Sa224N mutant,
as compared with the wild-type enzyme, whereas Ki
for CN-Cbl was not so much affected In contrast, Km
for AdoCbl and Ki for CN-Cbl decreased slightly for
the Sa224A mutant, although Km for 1,2-propanediol
did not change Thus, it became clear that the steric
crowding or inappropriate hydrogen bonding induced
by the relatively bulky side chain of Asn in the Sa224N mutant lowers significantly the affinity for the coenzyme and substrate as well as catalytic effi-ciency (kcat⁄ Km)
Catalytic activities and kinetic properties of Lysb135 mutant diol dehydratases
Table 1 indicates that kcat values of the Kb135R, Kb135A and Kb135Q mutants were 58–76% that of the wild-type enzyme at saturating concentrations of
A
Fig 1 The structure of the coenzyme-binding site in diol dehydratase (A) Stereo drawing of the hydrogen bonding and the ion pairing interactions of AdePeCbl with Sera224 (S224) and Lysb135 (K135), respectively PDO represents (S)-1,2-propanediol in this drawing Pink and green colors indicate the a- and b-subunits, respectively, darkening continuously from the N-terminal to the C-terminal sides (B) Resi-dues interacting with the adenine moiety of AdePeCbl (C) ResiResi-dues interacting with the phosphodiester group in the nucleotide loop of cobalamin.
Trang 4AdoCbl, and rapid inactivation during catalysis was
not observed with these mutants (Fig 3B) In contrast,
kcat of the Kb135E mutant was only 2% that of the
wild-type enzyme, and the kcat⁄ kinact value indicated that this mutant underwent inactivation after 5.4· 104
turnovers on average These results indicate that a positive charge in the b135 residue is not absolutely required for catalytic activity, but the introduction of
a negative charge there greatly lowers the activity The former suggests that the salt bridge between the phos-phate group of the coenzyme nucleotide loop moiety and the side chain of Lysb135 is not essential for activ-ity The latter would be probably due to the electro-static repulsion between the negatively charged phosphate group and –COO) in the side chain of the b135 residue
The effects of Lysb135 mutations on Km values for AdoCbl and 1,2-propanediol as well as Kifor CN-Cbl are summarized in Tables 1 and 2 For the Kb135R mutant, Km for AdoCbl and Ki for CN-Cbl were rather smaller than those of the wild-type enzyme This indicates that cobalamins are bound to this mutant more tightly than to the wild-type enzyme, probably because the salt bridge formation between the cobala-min phosphate group and the guanidinium group of Arg in the Kb135R mutant is appropriate For the Kb135A and Kb135Q mutants, which have a neutral side chain at the b135 residue, Km values for AdoCbl and 1,2-propanediol and Ki for CN-Cbl increased sig-nificantly These results suggest that the affinities of the enzyme for cobalamins and substrate are lowered, probably due to the inability of these mutants to form
a salt bridge with the phosphate group of cobalamins
In contrast, the Kb135E mutant, which has a negative charge at the b135 residue, showed a Km for AdoCbl and a Kifor CN-Cbl that were larger than those of the wild-type enzyme by two orders of magnitude, although Km for 1,2-propanediol was comparable to the values of the Kb135A and Kb135Q mutants It can therefore be concluded that the positive charge in the b135 residue is not essential for catalysis and is
A
B
Fig 2 PAGE analysis of the purified preparations of mutant diol
de-hydratases (A) SDS ⁄ PAGE (B) Nondenaturing PAGE Samples were
electrophoresed on 11% (A) and 7% (B) polyacrylamide gels, and the
resulting gels were subjected to protein staining with Coomassie
Brilliant Blue R-250 Molecular mass markers, SDS-7 (Sigma-Aldrich,
St Louis, MO, USA) BPB, bromophenol blue; wt, wild-type enzyme.
The bands of the a-, b- and c-subunits are indicated on the right (A).
The position of the (abc)2complexes is indicated on the right (B).
Table 1 Kinetic parameters of mutant diol dehydratases, determined at 37 C The k cat values were determined by the alcohol dehydroge-nase–NADH coupled method using 1,2-propanediol as substrate The k inact values were calculated from a change in the slope of a tangent
to the time course curve of the reaction The Km values were determined by the 3-methyl-2-benzothiazolinone hydrazone method Aver-age ± standard deviation (n = 3) The AdoCbl concentrations used were 15 l M for the wild-type enzyme, Sa224A mutant and Kb135R mutant, 45 l M for the Sa224N mutant, 30 l M for the Kb135A mutant, 57 l M for the Kb135Q mutant, and 150 l M for the Kb135E mutant.
Enzyme
k cat , s)1 (%)
K m for 1,2-propanediol (m M )
k cat ⁄ K m · 10)6 (s)1Æ M )1)
k inact (min)1) k cat ⁄ k inact · 10)4
Trang 5only moderately important for cobalamin binding, but
the introduction of a negative charge largely prevents
the enzyme–cobalamin interaction
Spectral changes of AdoCbl upon incubation with
mutant enzymes in the presence of substrate
Figure 4 shows the spectral changes of AdoCbl upon
incubation with mutant apoenzymes in the presence of
1,2-propanediol AdoCbl underwent a spectral change
to cob(II)alamin (B12r) upon incubation with the
wild-type or Sa224A apoenzyme for 5 min – that is, the
absorbance at 525 nm decreased and a new peak at
478 nm appeared (Fig 4A,B) These spectra reflect the
steady-state concentrations of cob(II)alamin during
catalysis With the wild-type enzyme, it was not so much different from the spectrum obtained at 30 min
of incubation In the case of the Sa224A mutant, how-ever, the peak at 478 nm increased gradually upon prolonged incubation, and the spectrum obtained at
30 min of incubation resembled the typical spectrum
of cob(II)alamin As the Sa224A holoenzyme was completely inactivated by 10 min of incubation, this spectrum should be that of the completely inactivated holoenzyme of this mutant The cob(II)alamin-like spe-cies in the inactivated Sa224A holoenzyme was stable even under aerobic conditions, but underwent oxida-tion to aquacobalamin (aqCbl) upon denaturaoxida-tion of the complex with guanidine-HCl under acidic condi-tions The spectrum thus obtained no longer changed upon photoillumination, suggesting that the Co–C bond of the coenzyme had been completely and irre-versibly cleaved upon incubation with the Sa224A mutant for 30 min in the presence of substrate In con-trast, the spectral change of AdoCbl upon incubation with the Sa224N apoenzyme was rather small even at
30 min of incubation (Fig 4C) The spectrum observed after denaturation of this mutant holoenzyme was sim-ilar to that of free AdoCbl and changed to that of aqCbl upon photoillumination Relatively low activity and a small kinact of the Sa224N mutant (Table 1) would account for the lower steady-state concentration
of cob(II)alamin species and the slower rate of irre-versible cleavage of the Co–C bond with this mutant When AdoCbl was incubated with the Kb135A mutant in the presence of substrate, a similar spectral change was observed within 5 min (Fig 4D) However, the peak at 478 nm then decreased gradually, and the absorbance at 356 nm and 530 nm increased with time of incubation These absorption peaks are characteristic of the diol dehydratase-bound
Fig 3 Time courses of 1,2-propanediol dehydration by mutant diol
dehydratases (A) Sera224 mutants (B) Lysb135 mutants The
alco-hol dehydrogenase–NADH coupled method was used The reaction
mixture consisted of 60 ng of the wild-type or 600 ng of a mutant
apoenzyme, 0.1 M 1,2-propanediol, 50 lg of yeast alcohol
dehydro-genase, 0.2 m M NADH, 0.04 M potassium phosphate buffer
(pH 8.0), and AdoCbl, in a total volume of 1.0 mL The reaction
was started by adding AdoCbl at a concentration given in the
legend to Table 1 The absorbance changes (DA at 340 nm) per
60 ng of enzyme are shown here.
Table 2 Binding affinities of mutant diol dehydratases for AdoCbl and CN-Cbl, determined at 37 C The k cat values were obtained as described in the legend to Table 1 Apparent Kmand Kivalues were determined by the 3-methyl-2-benzothiazolinone hydrazone method, followed by Lineweaver–Burk plots Average ± standard deviation (n = 3).
Enzyme
Kmfor AdoCbl (l M )
kcat⁄ K m (AdoCbl) · 10)8 (s)1Æ M )1)
Kifor CN-Cbl (l M )
Trang 6hydroxocobalamin (OH-Cbl) [30], suggesting that the
oxidation of cob(II)alamin accompanies the
inactiva-tion of the Kb135A holoenzyme during catalysis
Upon denaturation of the complex after 30 min of
incubation, the spectrum of enzyme-bound OH-Cbl
was at least partly converted to a free aqCbl-like
spec-trum When the mixture was then photoilluminated,
the spectrum underwent a further change to that of
free aqCbl It is therefore likely that AdoCbl was
mostly converted to the enzyme-bound OH-Cbl by
30 min of incubation, but a fraction of the coenzyme
still remained as AdoCbl at this time
EPR spectra obtained with the Sa224A mutant
diol dehydratase
When the wild-type holoenzyme was incubated with
1,2-propanediol at 4C for 1 min under anaerobic
conditions, the typical EPR spectrum of reacting
holo-enzyme was obtained (Fig 5A) The characteristic
high-field doublet signal with a splitting 14.3 mT
was assigned to the 1,2-propanediol-1-yl radical
(substrate-derived radical) [31], and the low-field broad signal to the low-spin Co(II) of cob(II)alamin Such a spectrum arises from weak coupling in the Co(II)– organic radical pair [32–34] The intensity of the dou-blet signal decreased within 3 min of incubation at
25C, and a new small peak with a g-value of 2.1 appeared upon further incubation for 30 min The latter low-field signal might be due to the inactivated holoenzyme being formed In contrast, only a trace of the typical doublet signal of reacting holoenzyme was observed with the Sa224A mutant upon incubation with substrate at 4C for 1 min (Fig 5B) This indi-cates that the steady-state concentration of an organic radical intermediate is very low with the Sa224A mutant, which is consistent with the low catalytic activity of this mutant Upon further incubation at
25C for 3 min, the trace of doublet signal disap-peared, and new signals with g-values of 2.08 and
2.2 appeared Although the radical species giving the g = 2.08 signal has not yet been identified, the relative intensity of this signal increased with time of incubation, and the signal became predominant after
Fig 4 Spectral changes of AdoCbl upon aerobic incubation with mutant diol dehydratases in the presence of 1,2-propanediol Apoenzyme (5 nmol) of the wild-type (A), Sa224A mutant (B), Sa224N mutant (C) or Kb135A mutant (D) was incubated at 30 C with 4.5 nmol of AdoCbl
in 0.01 M potassium phosphate buffer (pH 8.0) containing 1.3 M 1,2-propanediol and 1% Brij35, in a total volume of 1.0 mL Spectra were taken at 5 min (thick solid lines) and 30 min (thin solid lines) after the addition of AdoCbl Enzymes were then denatured by adding 6 M gua-nidine–HCl and 0.06 M citric acid After incubation at 37 C for 10 min, the mixture was neutralized to pH 8 by adding 200 lL of 1 M potas-sium phosphate buffer (pH 8.0) and 40 lL of 5 M KOH After the spectral measurement (broken lines), the mixture was photoilluminated in
an ice-water bath for 10 min with a 300 W tungsten light bulb from a distance of 20 cm, and the spectra were taken again (dotted lines) Spectra are corrected for dilution The spectra of apoenzymes were subtracted from the spectra obtained.
Trang 733 min It might be due to the same radical species as appeared with the wild-type enzyme upon prolonged incubation The signal with a g-value of 2.2 might
be assigned to Co(II) of cob(II)alamin These results indicate that the Sa224A mutant undergoes rapid and irreversible inactivation during catalysis by the extinc-tion of an organic radical intermediate through unde-sirable side reaction(s)
Fate of the adenosyl group of AdoCbl in inactivation of a mutant holoenzyme during catalysis
To study the fate of the adenosyl group, the upper axial ligand of AdoCbl, in the inactivation of the Sa224A holoenzyme during catalysis, adenosyl group-derived product(s) from AdoCbl were identified After
30 min of incubation of the Sa224A holoenzyme with 1,2-propanediol, the inactivated holoenzyme was dena-tured, and product(s) formed from the coenzyme were extracted and analyzed by HPLC on a reversed-phase column The only nucleoside product derived from the adenosyl group was identified as 5¢-deoxyadenosine The retention time of 5¢-deoxyadenosine was 8 min under the conditions employed The formation of adenine, adenosine, 4¢,5¢-anhydroadenosine, 5¢,8-cyclic adenosine or adenosine 5¢-aldehyde was not observed
at all It is therefore evident that the inactivation of this mutant enzyme during catalysis is a mechanism-based one, because the hydrogen abstraction from sub-strate by the coenzyme adenosyl radical takes place as the initial event of catalysis The amount of 5¢-deoxy-adenosine formed from 4.6 nmol of the Sa224A mutant and 15 nmol of AdoCbl was 4.7 nmol, which corresponds to approximately one mol per mol of enzyme As diol dehydratase exists as a dimer of heterotrimers, this result suggests that only one of the two heterotrimeric units is involved in the formation
of 5¢-deoxyadenosine
Recovery of active apoenzyme by resolution of
an inactivated mutant enzyme
A typical result of resolution experiments is shown in Table 3 The Sa224A mutant was completely inacti-vated by incubation with 1,2-propanediol for 30 min, followed by dialysis After resolution by acid ammo-nium sulfate treatment, 56% of the original specific activity of the mutant enzyme was recovered The reso-lution of cobalamin by this procedure was not com-plete, and the resolved enzyme still contained OH-Cbl The cobalamin recovered in the supernatant was aqCbl, and the extent of cobalamin resolution was
Fig 5 EPR spectra observed upon incubation of the wild-type (A)
and Sa224A mutant (B) holodiol dehydratases with 1,2-propanediol.
The arrows correspond to g = 2.0 Holoenzymes were formed
under an argon atmosphere by incubating 1.9 mg (9.2 nmol) of
sub-strate-free wild-type and Sa224A apoenzymes at 25 C for 3 min
with 50 nmol of AdoCbl in 0.65 mL of 0.05 M potassium phosphate
buffer (pH 8.0) containing 18 m M sucrose monocaprate The
enzyme reaction was started by adding 50 lmol of 1,2-propanediol
in 0.05 mL After 1 min at 4 C, the reaction mixture was rapidly
frozen in an isopentane bath that had been previously cooled to
approximately )160 C, and then in a liquid nitrogen bath EPR
spectra were taken at )130 C After the first measurement, the
mixture was incubated at 25 C for 3 min and frozen again, as
described above, for the second measurement The mixture was
then incubated at 25 C for an additional 5 min and 25 min for the
third and fourth measurements.
Trang 8estimated to be 61% from the cobalamin contents of
the 1,2-propanediol-inactivated Sa224A holoenzyme
and of the resolved enzyme This value of cobalamin
resolution is in good agreement with the recovery of
enzyme activity It was therefore concluded that
resolved apoenzyme recovered from the inactivated
mutant holoenzyme could be reconstitutable to fully
active holoenzyme – that is, the Sa224A apoenzyme
itself did not undergo damage in the mechanism-based
inactivation by the substrate 1,2-propanediol
Inactivation of mutant holoenzymes by O2in the
absence of substrate
The holoenzyme of diol dehydratase undergoes
irre-versible inactivation by O2 in the absence of substrate
[35] This inactivation is accompanied by the
irrevers-ible Co–C bond cleavage of the enzyme-bound
coen-zyme, forming OH-Cbl It is thus believed that the
inactivation is caused by the reaction of the activated
Co–C bond with O2, and thus reflects the extent of
Co–C bond activation upon the coenzyme binding to
apoenzyme in the absence of substrate As shown in
Table 4, the inactivation followed pseudo-first-order
reaction kinetics, with a rate constant (kinact,O2) of
0.20 min)1 for the wild-type holoenzyme To examine
the contributions of enzyme–coenzyme interactions at
the Sera224 and Lysb135 residues to Co–C bond
acti-vation, rates of O2inactivation of mutant holoenzymes
in the absence of substrate were determined The rate
constants of O2 inactivation for the Lysb135 mutants
were in the range 0.13–0.17 min)1, which is slightly
slower but almost comparable to that of the wild-type
holoenzyme In contrast, the rate of O2inactivation of mutant holoenzymes was very slow for the Sa224A mutant, and inactivation was not observed with the Sa224N mutant These results suggest that the appro-priate hydrogen bonding between the side chain OH group of Sera224 and N3 of the adenine ring of the coenzyme adenosyl group is important for activation
of the Co–C bond in the absence of substrate as well
On the other hand, ion pairing between the e-NH2 group of Lysb135 and the phosphate group of the coenzyme nucleotide loop seems not to be essential for Co–C bond activation This would be reasonable, because the cobalamin moiety of AdoCbl is accommo-dated to the cobalamin-binding site of the enzyme through multiple interactions with many amino acids
Discussion
The homolytic fission of the Co–C bond of enzyme-bound AdoCbl leads to the introduction of an adenosyl radical, a catalytic radical, into the active sites This is
an essential early event in all of the AdoCbl-dependent enzymatic reactions [1–3,7,8] We synthesized various coenzyme analogs in which one of the structural compo-nents is substituted by a closely related group, and used them as probes to investigate the mechanism of enzy-matic activation (labilization) of the coenzyme Co–C bond as well as the role of each structural component of the coenzyme in the interaction with diol dehydratase [1,11–17] It was demonstrated that the cobalamin moi-ety [14–17,29] and the adenosyl group [11–13] are required for its tight binding to the apoenzyme and for activation of the Co–C bond, respectively, and that the
‘adenine-attracting effect’ of the apoenzyme is a major element that weakens the Co–C bond [36,37] Later, the X-ray structures of the diol dehydratase–AdePeCbl complex revealed that the enzyme has a cobalamin-bind-ing site [18] and an adenine-bindcobalamin-bind-ing pocket [21] for Ado-Cbl A modeling study using X-ray structures suggested that the tight binding of AdoCbl to both of these sites induces marked distortions, including both angular strains and tensile force, that inevitably lead to Co–C bond cleavage [21,38] We proposed this ‘steric strain model’ as the molecular mechanism for the enzymatic activation of the coenzyme’s Co–C bond
The X-ray structures of diol dehydratase show that the coenzyme adenine moiety is anchored in the pro-tein by hydrogen bonding of N3 with the side chain
OH group of Sera224, of 6-NH2 and N7 with main chain amide groups of other residues, and of N1 with
a water molecule [21] Cobalamin is accommodated to
a space that is mainly surrounded by hydrophilic groups [18] Five amide groups out of six peripheral
Table 3 Resolution of 1,2-propanediol-inactivated Sa2424A mutant
holoenzyme by acid ammonium sulfate treatment.
Specific activity, UnitsÆmg)1(%)
B 12 bound,
l M ( %)
Inactivated holoenzyme 0.0 (0) 3.1 (100)
Table 4 O 2 inactivation of mutant holo-diol dehydratases in the
absence of substrate, determined at 37 C.
a No inactivation was observed for at least 10 min.
Trang 9side chains of the corrin ring form hydrogen bonds
with five amino acids in the a-subunit and three in the
b-subunit The phosphate group of cobalamin forms a
salt bridge between the e-amino group of Lysb135, in
addition to hydrogen bonds with the two residues in
the b-subunit In this study, the impacts of adenine
anchoring and ion pairing on catalysis were evaluated
by site-directed mutagenesis at Sera224 and Lysb135
The Sa224A mutant, which cannot form a hydrogen
bond with N3 of the adenine moiety, showed a relative
activity (kcat) of 19% and decreased sensitivity of the
holoenzyme to O2 in the absence of substrate,
suggest-ing the importance of hydrogen bondsuggest-ing between N3
of the adenine moiety and the side chain OH group of
Sera224 for activation of the coenzyme Co–C bond
On the other hand, the Sa224N mutant had only 5%
relative activity, which was much lower than that of
the Sa224A mutant Its complex with AdoCbl did not
show the sensitivity to O2 in the absence of substrate
One possibility is that the hydrogen bond might not be
formed or be formed at an improper position with the
Sa224N mutant The other possibility is that it might
be due to the coenzyme-binding problems caused by
the Sa224N mutation, regardless of the hydrogen
bonding interaction with N3
The O2inactivation of the holoenzyme occurs only in
the absence of substrate The inactivation mechanism
remains unclear at present, and the identification of the
inactivation products would provide important clues to
solve this problem It was well established from the
spec-tral changes that OH-Cbl is formed upon the
inactiva-tion [35] However, products from the adenosyl group of
AdoCbl have not yet been definitely identified [35]
(M Yamanishi, S Yamanaka & T Toraya,
unpub-lished results) Product(s) from O2 also remain to be
identified O2 inactivation can be considered to be
clo-sely related to catalysis, because only the complexes with
active coenzyme analogs undergo this inactivation,
except for the complexes with 3-deazaAdoCbl [12],
neb-ularylcobalamin (deamino analog of AdoCbl) [11], and
aristeromycylcobalamin (carbocyclic analog of AdoCbl)
[11,39] The substrate facilitates Co–C bond cleavage by
inducing conformational changes that increase the steric
strain of the Co–C bond of enzyme-bound coenzyme
[38] Therefore, these results suggest that the Co–C bond
activation in the absence of substrate is not sufficient
with coenzyme analogs that lack hydrogen bonding
interactions with the enzyme
The kcat⁄ kinact value is a good measure of the
resis-tance of holoenzymes to mechanism-based inactivation
These ratios for the Sa224A and Sa224N mutants
indi-cate that these enzymes underwent inactivation after
only 8000 and 15 000 turnovers, respectively, on
aver-age We have previously reported that the wild-type enzyme shows a small kcat⁄ kinact value when 3-dea-zaAdoCbl, a coenzyme analog that cannot form a hydrogen bond with Sera224, is used as coenzyme [12]
It can thus be concluded that the proper hydrogen bonding between N3 of the adenine moiety and Sera224 plays an essential role in protecting highly reactive radi-cal intermediate(s) from undesired side reactions, proba-bly through stable anchoring of the adenine moiety to the adenine-binding pocket Upon the inactivation of Sa224A, a cob(II)alamin-like spectrum was observed The EPR spectrum suggested that cob(II)alamin and an unidentified radical species were formed from AdoCbl and accumulated upon prolonged incubation with the Sa224A mutant A stoichiometric amount of 5¢-deoxy-adenosine was formed upon inactivation from the enzyme-bound coenzyme, but the apoenzyme itself was not damaged Thus, the inactivation of the Sa224A mutant during catalysis was concluded to be a mecha-nism-based one, as shown below:
E Ado CblIII
SH E Ado
SH
•
•S
reaction
E Cbl•II
where E is enzyme, SH is substrate, Cbl is cobalamin, and Ado is 5¢-deoxy-5¢-adenosyl
Mutant enzymes in which Lysb135 was substituted with Arg, Ala or Gln showed relatively high activity and relatively large kcat⁄ kinact values, whereas the Kb135E mutant possessed a trace of activity and a small kcat⁄ kinact value Binding affinities for AdoCbl and CN-Cbl were strengthened when Lysb135 was substituted with Arg, being almost the same as the those of the wild-type enzyme upon the Ala substitu-tion, and slightly lowered upon the Gln substitution at the b135 residue This might be due to hydrogen bond-ing and interactions other than the ion pairbond-ing bebond-ing strong enough to maintain the tight binding of cobala-min In contrast, the Kb135E mutant showed more than 100-fold lower affinity for both cobalamins These results indicate that a salt bridge between the phosphate group and the side chain of the b135 resi-due is not essential for either catalysis or cobalamin binding, but the introduction of a negative charge in the b135 residue destroyed the affinity for cobalamins and lowered the enzyme activity to 2% even when AdoCbl was used at a concentration higher than its
Km It was also evident from the decreased kcat⁄ kinact
value with the Kb135E mutant that the electrostatic repulsion between the cobalamin phosphate group and the side chain –COO) of Glub135 also results in the destabilization of the reactive radical intermediate(s) during catalysis
Trang 10Experimental procedures
Materials
Crystalline AdoCbl was a gift from Eisai Co Ltd (Tokyo,
Japan) Crystalline CN-Cbl was obtained from Glaxo
Research Ltd (Greenford, UK) Other chemicals were
analyti-cal grade reagents and were used without further purification
Construction of expression plasmids for mutant
diol dehydratases
The mutations described in this article were introduced into
the diol dehydratase genes (pddABC) of Klebsiella oxytoca
(formerly Aerobacter aerogenes) ATCC8724, using a
Quik-Change site-directed mutagenesis kit (Stratagene, La Jolla,
CA, USA) pUSI2E(DD) [22], an expression plasmid for the
wild-type enzyme, was used as a template The mutagenic
sense primers used were 5¢-ctacgccgaaaccatcgccgtctacggcac-3¢
for Sa224A, 5¢-ctacgccgaaaccatcaacgtctacggcac-3¢ for
Sa224N, 5¢-ggcatccagtcgagaggcaccacggtgatc-3¢ for Kb135R,
5¢-ggcatccagtcggcaggcaccacggtgatc-3¢ for Kb135A, 5¢-ggcatc
cagtcgcaaggcaccacggtgatc-3¢ for Kb135Q, and 5¢-gcatc
cagtcggaaggcaccacggtgatc-3¢ for Kb135E, in which
underlin-ing indicates amino acid substitutions The oligonucleotides
having the complementary sequences in the opposite
direc-tion were used as the respective antisense primers It was
con-firmed by sequencing of the DNA region encompassing the
entire diol dehydratase genes and tac promoter that no
unin-tended mutations had been incorporated during mutagenesis
Expression and purification of mutant diol
dehydratases
E coli XL1-Bluecells were transformed with the
above-men-tioned expression plasmids Recombinant E coli cells were
grown aerobically in LB medium containing 0.1%
1,2-pro-panediol and ampicillin (50 lgÆmL)1), and induced by 1 mm
isopropyl b-d-1-thiogalactopyranoside, as described
previ-ously [22] Mutant apoenzymes were purified from
over-expressing E coli cells, essentially as described previously for
the recombinant wild-type enzyme [28] The DEAE–cellulose
chromatography step was omitted, as the enzymes extracted
from crude membrane fractions were found to be almost
homogeneous However, the DEAE–cellulose-purified
apoenzymes were used in the EPR experiments
Substrate-free apoenzymes
Substrate-free apoenzymes for the measurements of Km
values for 1,2-propanediol and rates of inactivation of
holoenzymes in the absence of substrate were obtained by
dialysis at 4C for 36 h against 100 volumes of 50 mm
potassium phosphate buffer (pH 8.0) containing 0.1%
Brij35 with two buffer changes
Enzyme and protein assays Diol dehydratase activity was routinely measured by the 3-methyl-2-benzothiazolinone hydrazone method, using 1,2-propanediol as substrate [11] One unit is defined as the amount of enzyme activity that catalyzes the formation of
1 lmol of propionaldehyde⁄ min at 37 C under the stan-dard assay conditions Time courses of the diol dehydratase reaction were measured by the alcohol dehydrogenase– NADH coupled method [29]
The protein concentration of purified enzyme was deter-mined by measuring the absorbance at 280 nm The molar absorption coefficient at 280 nm, calculated by the method
of Gill & von Hippel [40], for diol dehydratase is
120 500 m)1Æcm)1[41]
PAGE PAGE analyses of purified mutant enzymes were performed under nondenaturing conditions as described by Davis [42],
in the presence of 0.1 m 1,2-propanediol [22], and under denaturing conditions as described by Laemmli [43] Pro-teins were stained with Coomassie Brilliant Blue R-250
EPR measurements The wild-type and the Sa224A mutant apoenzymes purified
as described previously [28] were used Substrate-free apoen-zyme solution [1.9 mg of protein in 0.6 mL of 50 mm potas-sium phosphate buffer (pH 8.0) containing 20 mm sucrose monocaprate] was mixed at 0C with AdoCbl solution (50 nmol in 0.05 mL) in a quartz EPR tube (outside diameter
5 mm) stoppered with a rubber septum After replacement of the air in the tube with argon by repeated evacuation and flushing with argon three times, holoenzymes were formed, reacted with 1,2-propanediol, and rapidly frozen as described
in the legend to Fig 5 The frozen sample was transferred to the EPR cavity and cooled with a cold nitrogen gas flow con-trolled by a Eurotherm B-VT 2000 temperature controller EPR spectra were taken as described previously [44,45] at )130 C on a Bruker ESP-380E spectrometer modified with
a Gunn diode X-band microwave unit EPR microwave frequency was 9.484–9.488 GHz, modulation amplitude was
1 mT, modulation frequency was 100 kHz, and microwave power was 10 mW
Fate of the adenosyl group of AdoCbl in inactivation
of mutant holoenzymes during catalysis The adenosyl group-derived product(s) formed from AdoCbl
in the inactivation of a mutant holoenzyme during catalysis was identified as described previously [46] Substrate-free apoenzyme (1.0 mg, 4.6 nmol) was incubated at 37C for
30 min in the dark with 15 lm AdoCbl in the presence
of 0.1 m 1,2-propanediol The enzyme protein was then