(Received for publication, December 27, 1996, and in revised form, April 9, 1997)
From the Central Laboratories for Key Technology,
Kirin Brewery Co. Ltd., 1-13-5 Fukuura, Kanazawa-ku, Yokohama 236 Japan
and the ¶ Graduate School of Pharmaceutical Sciences, Kyushu
University, Fukuoka 812-82 Japan
A mutant hen egg white lysozyme, D52E, was
designed to correspond to the structure of the mutant T4 lysozyme T26E
(Kuroki, R., Weaver, L. H., and Matthews B. W. (1993)
Science 262, 2030-2033) to investigate the role of the
catalytic residue on the -side of the saccharide in these enzymes.
The D52E mutant forms a covalent enzyme-substrate adduct, which was
detected by electron ion spray mass spectrometry. X-ray
crystallographic analysis showed that the covalent adduct was formed
between Glu-52 and the C-1 carbon of the
N-acetylglucosamine residue in subsite D of the saccharide binding site. It suggests that the catalytic mechanism of D52E mutant
lysozyme proceeds through a covalent enzyme-substrate intermediate indicating a different catalytic mechanism from the wild type hen egg
white lysozyme. It was confirmed that the substitution of Asp-52 with
Glu is structurally and functionally equivalent to the substitution of
Thr-26 with Glu in T4 lysozyme. Although the position of the catalytic
residue on the
-side of the saccharide is quite conserved among hen
egg white lysozyme, goose egg white lysozyme, and T4 phage lysozyme,
the adaptability of the side chain on the
-side of the saccharide is
considered to be responsible for the functional variation in their
glycosidase and transglycosidase activities.
It is known that the catalytic sites of most glycosidases have
common features. For example, -glycosidases have a glutamate on the
-side of the saccharide which acts as an acid to donate a proton to
glycosidic oxygen (O-4) to initiate the catalytic reaction (1-4).
Lysozyme is one of the
-glycosidases which cleaves the glycosidic
linkage between N-acetylglucosamine and
N-acetylmuramic acid in the bacterial cell wall. There are
three lysozymes in which the tertiary structures have been solved by
x-ray crystallography (5-8). The catalytic mechanisms of these
lysozymes have been discussed from the structural point of view. Hen
egg white lysozyme is one of the enzymes in which the catalytic
mechanism has been extensively investigated (9, 10). The catalytic
mechanism of HEWL1 is considered to proceed
through an oxocarbonium ion intermediate as originally proposed by D. C. Phillips (9). Even though the configurations of the catalytic
residues are somewhat similar in both HEWL and T4L (8), the catalytic
mechanism of T4L was found to proceed via a single displacement
mechanism (11, 12), indicating a different mechanism from that of
HEWL.
Moreover, it is known that the kind of catalytic residue located on the
-side of the saccharide seems more adaptable because the acidic
residue corresponding to Asp-52 of HEWL was not found on the
-side
of the saccharide in GEWL (13), mutant HEWL (14), or mutant T4L
(15-17). Recently, we have found that the mutant T4L, Thr-26
Glu,
located on the
-side of the saccharide, resulted in the enzyme
reacting directly with a saccharide to an enzyme-substrate adduct (11).
To clarify the structural and functional features of this residue
located on the
-side of the saccharide, we chose to modify HEWL and
focused on the residue Asp-52. We mutated Asp-52 to Glu, because in
HEWL Asp-52 is located on the
-side of the bound saccharide in a
position similar to that of Thr-26 in T4L. It is shown here that this
mutation (i.e. Asp-52
Glu) in HEWL does, in fact, also
lead to the production of a covalent enzyme-substrate adduct. This
further demonstrates the overall relationship between the active sites
and the mechanism of action of T4L and HEWL.
The wild type and D52E mutant lysozymes were prepared by expression and secretion from yeast as described previously (18). A (1,4)-linked hexamer of N-acetylglucosamine ((NAG)6) was purchased from Seikagaku Kogyo (Japan). All other chemicals were analytical grade for biochemical use.
Analysis of the Reaction Mixture of the D52E Mutant Lysozyme with (NAG)6The D52E mutant lysozyme (0.07 mM) was allowed to react with 0.5 mM of (NAG)6 at pH 5.0 and 25 °C. The reaction mixture was analyzed using reversed phase HPLC with a column of µ-Bondashere C18 (3.9 × 150 mm, 300 Å, 5 µm) (Waters). The protein was eluted with a gradient from 1 to 80% acetonitrile containing 0.05% trifluoroacetic acid for 40 min at a flow rate of 0.5 ml/min. The protein elution was monitored by the absorbance at 215 nm. Molecular mass determination of the proteins eluted from the reverse phase column was carried out on a Finnigan TSQ7000 mass spectrometer (Finnigan MAT, San Jose, CA) with an ion-spray interface and a quadruple mass analyzer with an upper mass limit of m/z = 4000 Da.
Large Scale Preparation of the D52E Mutant with a Covalent Adduct for CrystallizationOne mg of D52E mutant was allowed to react with 1 mg of (NAG)6 at pH 5.0 and 40 °C for 2 days. To remove the excess saccharide, the D52E mutant with a covalent adduct was purified by gel filtration using a column of Sephadex G-25 (1.3 × 65 cm). The eluent containing protein was collected and lyophilized to store. The lyophilized sample of D52E mutant with saccharide adduct was stable after being purified for at least 1 month at 4 °C. If the sample was kept with extra (NAG)6, the hydrolysis reaction proceeded, and the covalent adduct was decomposed within 2 weeks.
Crystallization and Tertiary Structure Analysis of the D52E Mutant Lysozyme with and without Saccharide AdductCrystallization of the apo-D52E mutant and the D52E mutant with a covalent adduct were performed using approximately the same method as that of the wild type reported previously (6, 19). Five µl of 200 mM sodium acetate buffer containing 15 mg/ml of protein was mixed with 5 µl of reservoir solution (200 mM sodium acetate buffer containing 4.0% NaCl (w/v), pH 4.5) and allowed to stand over the reservoir solution at 15 °C. In the case of the crystallization of D52E mutant with a covalent adduct, no extra (NAG)6 was provided to protect the adduct from decomposition. After 2 weeks, crystals isomorphous (P43212) to the wild type HEWL were formed (6, 19).
The co-crystallization of D52E mutant lysozyme with (NAG)6 was also attempted. One mg of (NAG)6 was dissolved in 100 µl of a protein solution containing 15 mg/ml D52E mutant lysozyme, and 5 µl of this solution were mixed with 5 µl of reservoir solution as before. The solution was allowed to stand over the reservoir solution at 15 °C. However, no crystals formed within 2 weeks. After 1 month, crystals formed that were 0.4 mm on edge and were isomorphous to that of the wild type.
The mutant structure was solved using the coordinates of wild type HEWL (1HEL) recently determined by Wilson et al. (19). The model structure was refined using the TNT refinement program package (20).
The structures of the catalytic sites of the T4L and
HEWL were compared based on the location of the residues that interact with the saccharides in subsite C and D. For this comparison, the OE2
atom of Glu-11 in T4L and the OE1 atom of Glu-35 in HEWL were chosen as
equivalent atoms because these residues act as a general acid to
hydrolyze the glycosyl bonds. Two pairs of atoms, NH of Leu-32 and CO
of Phe-104 in T4L and NH of Asn-60 and CO of Ala-107 in HEWL, were
chosen because they interacted with the N-acetyl group of
NAG bound to subsite C. The residues chosen are summarized in Table
I. Although Asp-20 in T4L has been considered to
correspond to Asp-52 in HEWL (8), these residues were excluded from the
structural comparison because the catalytic role of these aspartic
acids were recently proposed to be different (11, 12). From this
comparison, it seems that the location of Thr-26 in T4L is structurally
equivalent to the location of that of Asp-52 in HEWL. In view of this
correspondence, Asp-52 was replaced with glutamate to test for the
formation of a covalent adduct, as was observed in the mutant Thr-26
Glu in T4L.
|
The reaction mixture of mutant lysozymes with
(NAG)6 at pH 5.0 and 25 °C was analyzed by reversed
phase HPLC. The elution pattern of the reaction mixture from D52E
mutant and (NAG)6 after 1 day of reaction is shown in Fig.
1c with the elution pattern of D52E by itself
(shown in Fig. 1a). A new peak (A) was eluted before the D52E peak, and the yield of peak A was increased
up to more than 90% when the reaction was allowed to continue for 2 days at 25 °C. To identify peak A, the molecular mass of
the peak was determined by electron ion spray-mass spectrometry. The mass spectra of peak A and the D52E mutant lysozyme eluted
from the reversed phase HPLC are shown in Fig. 2,
a and b, respectively. From the analysis, the
mass of peak A was determined to be 15,130 daltons, which is
815 daltons larger than that of D52E mutant lysozyme (14,315 daltons).
The excess molecular mass (815 daltons) was equal (within error) to
that of (NAG)4 (830.84 daltons) after subtracting the
molecular mass of one water molecule (18 daltons). Thus, the excess
molecular mass can be attributed to the saccharide adduct covalently
linked to the mutant lysozyme.
X-ray Structure of D52E Mutant Lysozyme in Three Different Forms
Three conditions were created for the structural analysis of D52E mutant lysozyme. One was the D52E mutant lysozyme by itself (apo-form) as a reference. The second was the co-crystallization with (NAG)6 expecting the slow reaction to form a covalent adduct during crystallization. The third was the crystallization of the purified sample by gel filtration after the covalent complex was formed. The covalent adduct was stable in 50 mM sodium acetate buffer at 4 °C at least 2 weeks after being purified.2 All three samples were crystallized isomorphously to the wild type lysozyme (6, 19). Tertiary structures of the D52E mutant lysozyme under these conditions were determined at 1.9, 1.8, and 2.2 Å resolutions, respectively, by x-ray crystallography. The crystallographic parameters and refinement statistics are summarized in Table II.
|
The structure of D52E mutant (apo-form) was refined to
R = 16.8% at 1.9 Å resolution. The overall structure
of the D52E mutant was almost identical to that of the wild type. The
root-mean-square differences (RMSD) of C positions from
the wild type and D52E mutant lysozymes were 0.160 Å. The mutation of
Asp-52 to glutamate did not affect the overall structure of hen egg
white lysozyme.
The D52E mutant structure co-crystallized with (NAG)6 was
determined at 1.8 Å resolution with an R value of 16.4%.
The
2Fo-Fc density map contoured at 1 value is shown In Fig.
3A. The final refined structure is superposed
on the map. It was found that the oligomer of
N-acetylglucosamine was observed to be bound from A to C
subsites in the mutant lysozyme. No saccharide was observed in the
subsite D in this condition. Although the saccharide model was not
built into the weak electron densities outside of subsite A, the bound
oligomer of the N-acetylglucosamine was considered to be a
tetramer because D52E mutant lysozyme cleaved (NAG)6 to (NAG)4 and (NAG)2 (see Fig. 2). The interaction
between the bound saccharide and the D52E mutant lysozyme observed in
the subsite A, B, and C was essentially the same as those observed in
other lysozyme-saccharide complexes reported previously (21-24). The bound saccharide at subsite C revealed a
-anomer, although more
-anomer is found (59%) in solution (25). The hydrogen bonding interaction (3.4 Å) observed between OE1 of Glu-52 and the O-1 of NAG
bound at subsite C may cause the
-anomer to be at subsite C. The
overall structure of this complex was quite similar to those both in
the wild type (RMSD = 0.201 Å) and in the apo-D52E mutant
(RMSD = 0.150 Å), respectively.
The D52E mutant having a covalent adduct was determined at 2.2 Å resolution. The
2Fo-Fc
density map contoured at 1 is shown in Fig. 3B. Fig.
3C shows the omit map
(Fo(D52E with
adduct)-Fc(D52E)) from the observed data of the D52E mutant with adduct, model data of
D52E, and phases calculated from the D52E model with the Glu-52 side
chain atoms omitted, contoured at a 2
. The final refined structure
of the D52E mutant with adduct is superimposed on the map. The tertiary
structure of the D52E mutant with a covalent adduct was finally refined
to an R value of 16.6%. The mass spectrometry data showed
that a tetrasaccharide was covalently bound to the protein. The
electron density, however, showed strong density in the sites B and C,
and weaker density in sites A and D. This is in agreement with prior
studies showing that saccharides in the A site show weak density,
probably due to the weak energy of interaction (10). Based on
refinement and on difference electron density maps, the occupancy of
the saccharide appears to be approximately 100% in A, B, and C
subsites but only about 50% in subsite D. It seems that the covalent
adduct in D52E mutant was partially decomposed during crystallization.
To allow for the possibility that the geometry of the saccharide in
subsite D might be nonstandard (cf. Ref. 11), no guide
values were applied to the angle parameters. The refinement indicated
that C-1 of N-acetylglucosamine and OE1 of Glu-52 located on
the
-side of the bound saccharide are covalently linked. In
addition, the six-membered ring of NAG at subsite D was seen to be
slightly distorted toward the sofa form but not as much as observed in
the structure of T26E mutant T4L with a covalent adduct (11). Other
interactions between the covalently bound saccharide and the mutant
lysozyme at subsites A, B, and C were similar to those of the wild type
and mutant lysozymes (21-24). The backbone structure (C
positions) of the protein was also similar to that of the wild type
(RMSD = 0.254 Å) and D52E mutant lysozymes (RMSD = 0.223 Å)
and was most similar to the protein structure of the D52E mutant and
(NAG)4 complex (RMSD = 0.174 Å).
Three catalytic mechanisms have been proposed for the mechanisms
of glycosidases (1-3). They are an inverting mechanism (Fig. 4a) and two retaining mechanisms which
proceed through either an oxocarbonium ion intermediate (Fig.
4b) or a covalent enzyme-substrate intermediate (Fig.
4c). HEWL is one of the retaining glycosidases (26). The
catalytic mechanism has been considered to involve an oxocarbonium ion
intermediate (Fig. 4b) that is stabilized by a negative
charge of the carboxylic acid at Asp-52 (9, 10). The catalytic
mechanism of T4 lysozyme was recently determined to have an inverting
mechanism creating the -anomer after hydrolyzing the
-glycosidic
linkage (12). Although T4 phage and hen egg white lysozymes have
distinct catalytic mechanisms (12), a similar glycosyl-enzyme
intermediate was observed in mutants. The location of the covalent
linkage was confirmed to be between OE1 of Glu-52 and C-1 of NAG by
x-ray crystallography as shown above, similar to the linkage seen in
the T26E mutant of T4L. This indicates that the glutamic acid replacing
Asp-52 of HEWL or Thr-26 of T4L acts as a nucleophile to attack C-1
from the
-side of the saccharide in subsite D. The mutant D52E has
already been made by at least two groups (18, 27) and was reported to
retain less than 3% activity. The existence of the covalent
enzyme-substrate adduct suggests that the catalytic action in D52E
mutant HEWL proceeds through a covalent enzyme-substrate intermediate
as shown in Fig. 4c. This also implies that the D52E mutant
HEWL has a different catalytic mechanism from that proposed for the
wild type HEWL. The structural comparison between the wild type and
D52E mutant showed that the position of the carboxylate oxygen atom
closest to the substrate binding site has shifted about 1.8 Å toward
the inside of the cleft in the mutant. This positional shift of the oxygen atom may be responsible for the creation of the covalent adduct.
Moreover, a minor species corresponding to the covalent enzyme-substrate adduct was observed in D52S mutant by electron ion
spray mass spectrometry (28). If this minor species was created during
the catalytic process, it would require a double displacement mechanism
the same as that of the D52E mutant HEWL. The tertiary structure of the
D52S mutant (24), however, does not seem to support this mechanism,
because the position of the OG atom of Ser-52 is 2.6 Å further away
from the substrate compared with the OE1 atom of Glu-52. This would
place the OG atom of Ser-52 too far away to react with the saccharide
molecule directly. To solve this discrepancy, further analysis of the
catalytic mechanism of these mutant HEWL is needed.
It is already known that there are features common to T4L, GEWL, and
HEWL. The first is the location of glutamic acid (Glu-11, Glu-35, and
Glu-73). It is well known that the acidic residue located on the
-side of the saccharide is usually glutamic acid (i.e.
Glu-11 in T4L, Glu-35 in HEWL, and Glu-73 in GEWL), which has been
shown to be essential for the activity by chemical modifications and
mutation analysis (12, 14-16, 18, 29). The glutamate is considered to
act as a general acid to donate a proton to the glycosidic oxygen
(O-4). Another common feature is the location of main chain peptides
that interact with the N-acetyl group of the saccharide
bound at subsite C. The interactions in the saccharide binding sites
corresponding to subsite C in HEWL are strongly conserved between GEWL,
HEWL, and T4L as mentioned previously (8). In contrast, it was
previously reported that the acidic residue located on the
-side of
the saccharide is not always essential for the activity as seen in hen
egg white lysozyme mutants (14, 18, 30, 31), T4 phage lysozyme mutants
(15-17), and in the wild type lysozyme from goose egg white (13). The
catalytic function of the Asp-52 in HEWL is considered to be
stabilization of the oxocarbonium ion (mechanism in Fig.
4b), while Asp-20 in T4L is thought to act as a base to help
activate the water molecule to attack the saccharide from the
-side
of the saccharide (mechanism in Fig. 4a). It is now clear
that the roles of the aspartic acids located on the
-side of the
saccharide (Asp-20 in T4L, Asp-52 in HEWL) are different. The D52E
mutant shows yet a third role in the catalytic function (mechanism in
Fig. 4c), which has already been seen in other glycosidases
(4, 32). The formation of the same covalent enzyme-substrate
intermediate indicates that the locations of glutamic acid introduced
into Asp-52 in HEWL and Thr-26 in T4L are functionally equivalent as
shown in Table I.
A careful structure comparison of the catalytic sites from three
lysozymes shows that Asp-52 is located at the position equivalent to
Thr-26 in T4L and Gly-90 in GEWL. The similar C
locations of Asp-52 in HEWL to Thr-26 in T4L and Gly-90 in GEWL suggest
that the mutation of Asp-52 to a residue which has a short side chain
such as Thr, Ser, Ala, or Gly will possibly result in changing the
catalytic mechanism toward an inverting enzyme as observed for the wild
type T4L. Indeed, a similar mutation has been made in
-glucosidase,
and the formation of the
-anomer has been reported (33). Moreover,
the D52S mutant of HEWL has been prepared, and the tertiary structure
of the D52S mutant of HEWL with substrate was determined (24). The
-anomer seen in the crystal complex with the D52S mutant of HEWL is
considered to be a product of hydrolysis. The
-anomer was
interpreted as the result of mutarotation of product following a normal
wild type HEWL retaining reaction. It would, however, be consistent with the D52S mutant of HEWL acting as an inverting enzyme as has been
observed for wild type T4L. These findings support the hypothesis that
the residue located on the
-side of the saccharide, such as Asp-52
in HEWL, Thr-26 in T4L, or probably Gly-90 in GEWL, is responsible for
the differentiation of the catalytic mechanism. Moreover, the fact that
T4L and GEWL, which have a short side chain on the
-side of the
saccharide, do not catalyze transglycosylation, while HEWL, which has
aspartate or glutamate (mutant) on the
-side of the saccharide and
catalyzes transglycosylation, may indicate that this variability is
also responsible for the functional difference between glycosidases and
transglycosidases.
We express appreciation to Drs. L. H. Weaver and B. W. Matthews in the Institute of Molecular Biology, University of Oregon for helpful advice and valuable discussions. We also express thanks to Michiko Kanai of Thermo Quest K. K. for mass analysis.