- Email: [email protected]
Crystal structure of muconate lactonizing enzyme at 6.5 Å resolution
J. Mol. Biol. (1985) 182, 353-355
Crystal Structure of Muconate Lactonizing 65 A Resolution
Enzyme at
We have obtained crystals of Pseudomonas putida muconate lactonizing enzyme. They diffract to better than 2.4 A resolution and have two monomers in the asymmetric unit, related by a non-crystallographic 2-fold axis. The cell dimensions are 139.3 A x 139.3 A x 84.1 A, and the space group is I4. The electron density map at 6.5 A resolution shows that the enzyme is an octamer with D, symmetry.
placed in a beaker containing 250 ml of the solution 10 rnMcrystallize the protein: used to morpholinoethane sulfonate (pH 6*5), 30 miw-NaCl, 2 m&f-MnCl,, 7 mM-j?-mercaptoethanol. The reason for using the double dialysis technique is that it slows down the rate at which the salt concentration decreases. The crystallization is done in a constant temperature incubator at 18”C, and normally takes no more than a week. The crystals diffract to better than 2.4 A resolution and have cell dimensions of 139.3 A x 139.3 A x 84.1 A. Assuming a monomer molecular weight of 40,000, the V,,, value would be 2.5 A31 dalton if there were 16 molecules in the unit cell. This implies two molecules per asymmetric unit. Accordingly, the rotation function program in the Munich computing system PROTEIN (Steigemann, 1974) was run on the native Patterson map, using an inner radius of 7 A and an outer radius of 25 A. On the x = 180” section there was one unique peak, which was almost 90% of the origin peak (Fig. 1).
Muconate lactonizing enzyme from Pseudomonas is the second enzyme involved in the degradation of catechol to P-ketoadipate (Yeh & Ornston, 1980). It catalyzes the cycloisomerization of cis,cis-muconic acid to muconolactone. This reaction proceeds via syn-addition to the double bond (Avigad & Englard, 1969), and Ngai et al. (1983) have shown that the oxygen attacks the double bond before the hydrogen (i.e. the mechanism is El,,). The enzyme contains one manganese atom per monomer, which is essential for enzymatic activity. Roth sodium dodecyl sulfate polyacrylamide gel electrophoresis and gel filtration in the presence of guanidine. HCl gave a monomer molecular weight of 40,000 (Avigad et al., 1974). These authors also found an oligomeric molecular weight of 250,000 by gel filtration and sedimentation equilibrium, and therefore they proposed that MLEt was a hexamer in solution. Avigad et al. (1974) also obtained negatively stained electron micrographs of microcrystals grown in the presence of high concentrations of salt. The only symmetry element visible in their micrographs was a 4-fold, and so they thought that MLE was an octahedron in their high-salt crystals. We show here that, in crystals grown from low salt, the enzyme is in fact an octamer with D, symmetry. Dr Ornston kindly gave us MLE as a microcrystalline high-salt form already used for electron microscopy (Avigad et al., 1974). Our new crystal form, suitable for crystal structure analysis, grows from low concentrations of salt. Only by using a method we call double dialysis could we routinely obtain crystals up to 400 pm across and several millimeters long. This method involves first dissolving the protein by dialyzing it against 20 mM-Tris. HCl (pH 7.5), 0.1 m&r-MnCl,, 100 mMNaCl, 7 rnM-p-mercaptoethanol, 0.02% (w/v) NaN,, at 4°C. Then we place about 1 ml of the protein solution in a dialysis bag of diameter 0.39 in, and put it inside a 1.73 in diameter dialysis bag containing 25 ml of the buffer used to dissolve the protein. The large dialysis bag is then sealed and
P. putida MLE, when the map is rotated through 180”
t abbreviation used: MLE; muconate lactonizing enzyme.
about a vector lying in the x-y plane, plotted as a percentage of the self-correlation of the unrotated map. Psi = 0” corresponds to the vector lying along y; psi = 90”. to the vector lying along 5.
putida
0022%2&?3S/S5/06035343
$03.00/O
Angle from y OXIS ( 0 1
Figure 1. Self rotation of the native Patterson map of
353
0 1985 Academic Press Inc. (London) Ltd.
354
A. Goldman et al.
The position of the non-crystallographic 2-fold is, therefore, 24” from the x-axis. The crystals are twinned by merohedry (Buerger, 1960); the twinning fraction is normally low (around 0 to 20%), but occasionally much higher. We corrected low-twinned data sets using the method suggested by Fisher & Sweet (1980) and discarded any data set with a twinning fraction greater than 25%, because the percentage error in the corrected intensities becomes unbounded as the twinning fraction approaches 50%. A 6.5 a resolution electron density map was calculated with phases determined from two heavyatom derivatives (Table 1). The average figure of merit was 0.63. The solvent-protein boundary and the local S-fold axis could be seen reasonably clearly. We located the 2-fold axis more accurately using the rotation-translation search routines in PROTEIN (Steigemann, 1974); the axis passes through the z-axis, is parallel to the x-y plane, and bisects all the heavy-atom sites in each heavyatom derivative. The phases had converged after three cycles of: symmetry-averaging the map about the local 2-fold axis, flattening the solvent, back transformation, and phase recombination (Rossmann, 1972; Bricogne, 1976). The final average figure of merit was 0.8. We used the improved phases and observed amplitudes to calculate the electron density map shown in Figure 2. This map shows that the protein has D, symmetry; it is a tightly packed octamer. Intraoctamer contacts seem more common than interoctamer contacts. In contrast, the earlier studies (Avigad et al., 1974) found an oligomeric molecular weight in low-salt solution of about 250,000 with a monomer molecular weight of 40,000. Therefore, Avigad et al. (1974) concluded that MLE was a hexamer, and used this conclusion to interpret their electron micrographs. As their electron micrographs could resolve the MLE monomers only as spheres, they thought the micrographs showed that MLE had an octahedral oligomeric structure, because octahedrons have six subunits and contain a pseudo
Heavy-atom Reagent
Concn (mM)
Time (h)
Table 1 rejinement statistics %,,,t
No. of No. of sites reflections Fh/E$
X---G-
Figure 2. Part of the 6.5 A resolution electron density map of P. putida MLE. It extends for a complete unit cell in z and y, and from z=-6.2 A to z = +6.2 8, and shows the D, point group symmetry of the molecule; the arrows mark the position of the local 2-fold axes, and the 4-fold is perpendicular to the plane of the sections in the center of the map. The circle marks the boundary of the octamer. The map is contoured at intervals of la, starting at 1.5~7.
4-fold axis. However, our studies indicate that the molecule is a D, octamer in our low-salt crystals. and therefore it is almost certainly an octamer in low-salt solution. As none of the electron micrographs of high-salt crystals shows a 3-fold, but all show a 4-fold, we believe that the high-salt form is also composed of octamers. The most likely explanation of the discrepancy is that the oligomer molecular weight is substantially in error. We are extending our data to higher resolution. Using molecular replacement (Bricogne, 1976), we will improve our high-resolution electron density map and hope to provide a structural basis for understanding the mechanism of this enzyme. We thank Dr L. N. Ornston both for providing protein discussions. This work was supported by the United States Public Health Service grant GM22778. and for congenial
Sm(Ac), KJ’t(NO,L
0.5 0.1
24 24
4.5% 2.1%
2 6
1055 1118
1.85 2.49
The standard buffer for these crystals was 10 rn~morpholinoethane sulfonate (pH 6.5), 60 mM-N&l, 1 miw sodium adipate. The total number of native reflections collected to 65 A was 1506: this is 90% of the reflections to this resolution.
t R,,,,,= T i (‘4’$;;j1h)),where
Ii.* is the ith observation
of the reflectjbn’h: whik
Adrian Goldman, David Ollis Department of Molecular Biophysics Biochemistry, Yale University New Haven, CT 06511, U.S.A.
Ka-Leung Ngai Department of Biology Yale University New Haven, CT 06511. U.S.A.
and
Letters to the Editor
Thomas A. Steitz Department of Molecular Biophysics Biochemistry, Yale University New Haven. CT 06511. U.S.A.
and
Received 9 July 1984, and in revised form 23 November 1984
References Avigad, G. & Englard, S. (1969). Fed. Proc. Fed. Amer. Sot. Exp. Biol. 28, 345. Avigad, G., Englard, S., Olsen, B. R., Wolfenstein-Todel,
355
C. & Wiggins, R. W. (1974). J. Mol. Biol. 89, 651662. Bricogne, G. (1976). Acta Crystallogr. sect. A, 32, 832-847. Buerger, M. J. (1960). Crystal-Structure Analysis, John Wiley & Sons, New York. Fisher, R. G. & Sweet, R. M. (1980). Acta Crystallogr. sect. A, 36, 755-760. Ngai, K. L., Omston, L. N. & Kallen, R. G. (1983). Biochemistry, 22, 5223-5230. Rossmann, M. G. (1972). Editor of The Molecular Replacement Method, Gordon & Beach, New York, London and Paris. Steigemann, W. (1974). Thesis, TU, Miinchen. Yeh, W. K. & Ornston, L. N. (1980). Proc. Nat. Acad. Sci., U.S.A. 77, 5365-5369.
Edited by R. Huber
Crystal Structure of Muconate Lactonizing 65 A Resolution
Enzyme at
We have obtained crystals of Pseudomonas putida muconate lactonizing enzyme. They diffract to better than 2.4 A resolution and have two monomers in the asymmetric unit, related by a non-crystallographic 2-fold axis. The cell dimensions are 139.3 A x 139.3 A x 84.1 A, and the space group is I4. The electron density map at 6.5 A resolution shows that the enzyme is an octamer with D, symmetry.
placed in a beaker containing 250 ml of the solution 10 rnMcrystallize the protein: used to morpholinoethane sulfonate (pH 6*5), 30 miw-NaCl, 2 m&f-MnCl,, 7 mM-j?-mercaptoethanol. The reason for using the double dialysis technique is that it slows down the rate at which the salt concentration decreases. The crystallization is done in a constant temperature incubator at 18”C, and normally takes no more than a week. The crystals diffract to better than 2.4 A resolution and have cell dimensions of 139.3 A x 139.3 A x 84.1 A. Assuming a monomer molecular weight of 40,000, the V,,, value would be 2.5 A31 dalton if there were 16 molecules in the unit cell. This implies two molecules per asymmetric unit. Accordingly, the rotation function program in the Munich computing system PROTEIN (Steigemann, 1974) was run on the native Patterson map, using an inner radius of 7 A and an outer radius of 25 A. On the x = 180” section there was one unique peak, which was almost 90% of the origin peak (Fig. 1).
Muconate lactonizing enzyme from Pseudomonas is the second enzyme involved in the degradation of catechol to P-ketoadipate (Yeh & Ornston, 1980). It catalyzes the cycloisomerization of cis,cis-muconic acid to muconolactone. This reaction proceeds via syn-addition to the double bond (Avigad & Englard, 1969), and Ngai et al. (1983) have shown that the oxygen attacks the double bond before the hydrogen (i.e. the mechanism is El,,). The enzyme contains one manganese atom per monomer, which is essential for enzymatic activity. Roth sodium dodecyl sulfate polyacrylamide gel electrophoresis and gel filtration in the presence of guanidine. HCl gave a monomer molecular weight of 40,000 (Avigad et al., 1974). These authors also found an oligomeric molecular weight of 250,000 by gel filtration and sedimentation equilibrium, and therefore they proposed that MLEt was a hexamer in solution. Avigad et al. (1974) also obtained negatively stained electron micrographs of microcrystals grown in the presence of high concentrations of salt. The only symmetry element visible in their micrographs was a 4-fold, and so they thought that MLE was an octahedron in their high-salt crystals. We show here that, in crystals grown from low salt, the enzyme is in fact an octamer with D, symmetry. Dr Ornston kindly gave us MLE as a microcrystalline high-salt form already used for electron microscopy (Avigad et al., 1974). Our new crystal form, suitable for crystal structure analysis, grows from low concentrations of salt. Only by using a method we call double dialysis could we routinely obtain crystals up to 400 pm across and several millimeters long. This method involves first dissolving the protein by dialyzing it against 20 mM-Tris. HCl (pH 7.5), 0.1 m&r-MnCl,, 100 mMNaCl, 7 rnM-p-mercaptoethanol, 0.02% (w/v) NaN,, at 4°C. Then we place about 1 ml of the protein solution in a dialysis bag of diameter 0.39 in, and put it inside a 1.73 in diameter dialysis bag containing 25 ml of the buffer used to dissolve the protein. The large dialysis bag is then sealed and
P. putida MLE, when the map is rotated through 180”
t abbreviation used: MLE; muconate lactonizing enzyme.
about a vector lying in the x-y plane, plotted as a percentage of the self-correlation of the unrotated map. Psi = 0” corresponds to the vector lying along y; psi = 90”. to the vector lying along 5.
putida
0022%2&?3S/S5/06035343
$03.00/O
Angle from y OXIS ( 0 1
Figure 1. Self rotation of the native Patterson map of
353
0 1985 Academic Press Inc. (London) Ltd.
354
A. Goldman et al.
The position of the non-crystallographic 2-fold is, therefore, 24” from the x-axis. The crystals are twinned by merohedry (Buerger, 1960); the twinning fraction is normally low (around 0 to 20%), but occasionally much higher. We corrected low-twinned data sets using the method suggested by Fisher & Sweet (1980) and discarded any data set with a twinning fraction greater than 25%, because the percentage error in the corrected intensities becomes unbounded as the twinning fraction approaches 50%. A 6.5 a resolution electron density map was calculated with phases determined from two heavyatom derivatives (Table 1). The average figure of merit was 0.63. The solvent-protein boundary and the local S-fold axis could be seen reasonably clearly. We located the 2-fold axis more accurately using the rotation-translation search routines in PROTEIN (Steigemann, 1974); the axis passes through the z-axis, is parallel to the x-y plane, and bisects all the heavy-atom sites in each heavyatom derivative. The phases had converged after three cycles of: symmetry-averaging the map about the local 2-fold axis, flattening the solvent, back transformation, and phase recombination (Rossmann, 1972; Bricogne, 1976). The final average figure of merit was 0.8. We used the improved phases and observed amplitudes to calculate the electron density map shown in Figure 2. This map shows that the protein has D, symmetry; it is a tightly packed octamer. Intraoctamer contacts seem more common than interoctamer contacts. In contrast, the earlier studies (Avigad et al., 1974) found an oligomeric molecular weight in low-salt solution of about 250,000 with a monomer molecular weight of 40,000. Therefore, Avigad et al. (1974) concluded that MLE was a hexamer, and used this conclusion to interpret their electron micrographs. As their electron micrographs could resolve the MLE monomers only as spheres, they thought the micrographs showed that MLE had an octahedral oligomeric structure, because octahedrons have six subunits and contain a pseudo
Heavy-atom Reagent
Concn (mM)
Time (h)
Table 1 rejinement statistics %,,,t
No. of No. of sites reflections Fh/E$
X---G-
Figure 2. Part of the 6.5 A resolution electron density map of P. putida MLE. It extends for a complete unit cell in z and y, and from z=-6.2 A to z = +6.2 8, and shows the D, point group symmetry of the molecule; the arrows mark the position of the local 2-fold axes, and the 4-fold is perpendicular to the plane of the sections in the center of the map. The circle marks the boundary of the octamer. The map is contoured at intervals of la, starting at 1.5~7.
4-fold axis. However, our studies indicate that the molecule is a D, octamer in our low-salt crystals. and therefore it is almost certainly an octamer in low-salt solution. As none of the electron micrographs of high-salt crystals shows a 3-fold, but all show a 4-fold, we believe that the high-salt form is also composed of octamers. The most likely explanation of the discrepancy is that the oligomer molecular weight is substantially in error. We are extending our data to higher resolution. Using molecular replacement (Bricogne, 1976), we will improve our high-resolution electron density map and hope to provide a structural basis for understanding the mechanism of this enzyme. We thank Dr L. N. Ornston both for providing protein discussions. This work was supported by the United States Public Health Service grant GM22778. and for congenial
Sm(Ac), KJ’t(NO,L
0.5 0.1
24 24
4.5% 2.1%
2 6
1055 1118
1.85 2.49
The standard buffer for these crystals was 10 rn~morpholinoethane sulfonate (pH 6.5), 60 mM-N&l, 1 miw sodium adipate. The total number of native reflections collected to 65 A was 1506: this is 90% of the reflections to this resolution.
t R,,,,,= T i (‘4’$;;j1h)),where
Ii.* is the ith observation
of the reflectjbn’h: whik
Adrian Goldman, David Ollis Department of Molecular Biophysics Biochemistry, Yale University New Haven, CT 06511, U.S.A.
Ka-Leung Ngai Department of Biology Yale University New Haven, CT 06511. U.S.A.
and
Letters to the Editor
Thomas A. Steitz Department of Molecular Biophysics Biochemistry, Yale University New Haven. CT 06511. U.S.A.
and
Received 9 July 1984, and in revised form 23 November 1984
References Avigad, G. & Englard, S. (1969). Fed. Proc. Fed. Amer. Sot. Exp. Biol. 28, 345. Avigad, G., Englard, S., Olsen, B. R., Wolfenstein-Todel,
355
C. & Wiggins, R. W. (1974). J. Mol. Biol. 89, 651662. Bricogne, G. (1976). Acta Crystallogr. sect. A, 32, 832-847. Buerger, M. J. (1960). Crystal-Structure Analysis, John Wiley & Sons, New York. Fisher, R. G. & Sweet, R. M. (1980). Acta Crystallogr. sect. A, 36, 755-760. Ngai, K. L., Omston, L. N. & Kallen, R. G. (1983). Biochemistry, 22, 5223-5230. Rossmann, M. G. (1972). Editor of The Molecular Replacement Method, Gordon & Beach, New York, London and Paris. Steigemann, W. (1974). Thesis, TU, Miinchen. Yeh, W. K. & Ornston, L. N. (1980). Proc. Nat. Acad. Sci., U.S.A. 77, 5365-5369.
Edited by R. Huber