Role of the Proximal Ligand in Peroxidase Catalysis - The Journal of ...

6 downloads 0 Views 1MB Size Report
M Tris acetatemes buffer using a Cary 4 UVNisible spectrophotometer. The lifetime of the oxy-fen-yl center ... (Fowler and Bright, 1935). Enzyme concentrations ...
Vol. 269, No. 32, Issue of August 12, PP. 20239-20249, 1994 Printed in U.S.A.

JOURNAL OF BIOLOGICAL CHEMISTRY b 1994 by The American Society for Biochemistry and Molecular Biology, Inc. THE

Role of the Proximal Ligandin Peroxidase Catalysis CRYSTALLOGRAPHIC, KINETIC, AND SPECTRAL STUDIES OF CYTOCHROME c PEROXIDASE PROXIMAL LIGAND MUTANTS* (Received for publication, May 13, 1994, and in revised form, June 6, 1994)

Kalidip ChoudhurySO, Munirathiram Sundaramoorthyi, Alison Hickmannll, Takashi Yonetani**, Eilika WoehlSS, Michael F. DunnSS, and ThomasL. PouloslOO From the $Departments of Molecular Biology & Biochemistry and Physiology & Biophysics, University of California, Irvine, California 9271 7-3900, the **Department of Biochemistry and Biophysics, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6089, the llCenter for Advanced Research in Biotechnology, Maryland Biotechnology Institute, University of Maryland, Shady Grove Campus, Rockville, Maryland 20850, and the Department of Biochemistry-015, University of California, Riverside, Riverside, California 92521-0129

The role of the proximal histidine ligand in peroxidase Site-directed mutagenesis has shed considerable light on the function was studied by replacing the His side chain in CCP mechanism. For example, conversion of His-52 to Leu cytochrome c peroxidase with Gln, Glu, or Cys. In addi- results in a lo5 drop in t h e rate of Compound I formation tion, a double mutant was prepared where His-175 is (Erman et al., 1993) which supports the proposed roleof this converted to Gln and site the of free radical formation in residue as an acidhase catalyst (Poulos and Finzel, 1984). Compound I, Trp-191 (Sivaraja,M., Goodin, D. B., Smith, Our own laboratory has studied the role of the proximal M., and Hoffman, B. M. (1989) Science 245, 738-740), is ligand (Sundaramoorthy et al., 1991; Choudhury et al., 1992) converted to Phe. With the exception of the His-175 -* using site-directed mutagenesis. Replacement of the proximal Cys mutant, the proximal ligand mutants retain high Gln has little effect on the x-ray structure ligand, His-175, with levels of enzyme activity. Stopped flow studies show outside that the site of mutation, and the Gln-175 mutant exhibits replacing the His ligand with Gln has only a modest the same activity as t h e wild type enzyme. Mauroet a l . (1988) I formation demonstraeffect on the rate of Compound ting that the precise nature of the proximalisligand not earlier had shown that another proximal side mutant, Trp-191 of peroxide 0-0bond + Phe, is inactive because this mutant can no longer accept important in achieving a high rate cleavage. The double mutant, His-175 -+ GWl‘rp-191+ electrons from ferrocytochrome c. However, a double mutant where His-175 is converted to Gln and Trp-191 to Phe retains Phe, also forms Compound I rapidly but the initial prod20% wild type steady state activity (Choudhury et a l . , 1992). P cation uct formedis very likely a long-lived porphyrin radical that slowly converts to a species more closely Therefore, changing the proximal ligand recovers activity from an otherwise inactive enzyme. We now have extended these resembling the heme oxyferryl center of wild type Compound I. The relevance of these studies to the cyto- studies to include determination and functional characterizachrome c peroxidase-cytochromec electron transfer sys- tion of two additional proximalligand mutants (His-175+ Glu tem are discussed. and His-175 4 Cys) including solution of the crystal structures, steady state and stopped flow kinetics kinetics, reduction potential measurements, and spectral characterization including Cytochrome c peroxidase or CCP’ catalyzes the peroxide de- electron spin resonance. pendent oxidation of ferrocytochrome c in t h e following multiMATERIALSANDMETHODS

step reaction (Yonetani, 19761, CCP H,O,

Mutagenesis-Oligonucleotide site-directed mutagenesis was carried out following the methods of Kunkel et al. (1987)in theCCP gene cloned into the PT7-7 vector (Danvish et al., 1991).The PT7-7 (Danvish et al.,

CCP(Fe4+).

CCP(Fe4+).+ 2cyt.c(Fe2+) + CCP + 2cyt.c(Fe3+)

1991) system was used to prepare wild type and two variants: His-175

-,Glu and His-175 Cys. Single stranded, uracil-encoded plasmidwas

REACTION 1

isolated by growing the PT-7 vector in anappropriate host (Rz1032,gift from Dr.E. Esienstein, University of Maryland) and used as a template for annealing, extension, and ligation by the appropriate mutant oligonucleotide in vitro. The reaction mixture was then transformed into Dr. E. Esienstein, University host Escherichia coli cells (JV30, gift from of Maryland) which has a mechanism for deleting uracil-encoded DNA. * This work was supported by National Science Foundation Grants The efficiency of mutagenesis was about 80%. The His-175 + Gln and MCB-9249307 ( t o T. L. P.), MCB 921890 (to M.F. D.), National Insti- His-175 + Gln/Trp-lgl + Phe mutants were prepared according to tutes of Health Grants GM 42614 (to T.L. P.), HL 14508 (to T.Y.), and Sundaramoorthy et al. (1991)and Choudhury et al. (19921,respectively. GM 48130 (to T. Y.). The costs of publication of this article were defrayed Bakers’ yeast CCP differs in sequence from the laboratory yeast in partby the payment of page charges. This article must therefore be strain used to construct our expression system at 2 amino acid posihereby marked “advertisement”in accordance with 18 U.S.C. Section tions: 53 and 152. Bakers’yeast CCP contains Thr andAsp at positions 1734 solely to indicate this fact. 8 Present address: Institute of Biochemistry and Cell Biology, Syntex 53 and 152 (Takioet al., 1980),respectively, whilethe yeast strain has Ile and Gly at these positions (Kaput et al., 1982). Sincea majority of Pharmaceutical, 3401 Hillview Ave., Palo Alto, CA 94303, 11 Present address: NIDDK, National Institutes of Health, Bldg. 5, Rm. work on CCP overthe years has been with bakers’ yeast CCP, we felt it was important to use the bakers’ yeast sequence as our recombinant 306, National Institutes of Health, Bethesda, MD 20892. 88 To whom correspondence should be addressed: Dept. of Physiology wild type enzyme. Therefore, site-directed mutagenesis was used to & Biophysics, University of California, Irvine, CA 92717. Fax: 714-856- replace Ile-53 and Gly-152 with Thr and Asp. N-terminal sequencing 8540; E-mail [email protected]. has shown that there is no N-terminal Met in our wild type CCP and The abbreviations used are: CCP, cytochrome c peroxidase; Mes, that the firstseveral N-terminal residues are those expected forbakers’ 4-morpholineethanesulfonicacid; EPR, electron spin resonance. yeast CCP (Darwish et al., 1991). H,O, oxidizes CCP to an intermediate state termed Compound I which consists of a ferry1 (Fe*+) iron and an amino acidcentered free radical (denoted by the . in the above scheme).

20239

MutantsLigand CCP Proximal

20240

TABLE I Data collection summary Mutant His-175 A 2.00 Maximum resolution Total observations 59,206 0.082R," 0.074 Percent data collected to

I/ 41) at

indicated resolution

--j

Glu

His-I75 1.87 A

-

TABLE I1 Crystallographic refinement summary Cys

68,462 28,063

23,883

3.39 3.63 A 98%

A 94%

2.88 98% 2.52 A 91% 2.32 81% 2.13 A 72% 1.87 2.00 A 55%

A

2.70 A 94% 2.35 8, 91% 2.14 A 91% 0.019 1.98 A 82% A 39%

2.29 A 5.33

2.14 A 3.41

2.13 A 3.30 2.00 A 1.98

1.98 A 1.70 1.87 A 0.76

Mutant Resolution 10.0-1.8 range (A) Reflections measured Reflections 19,509 used" 0.167 R-factor* Rms deviation of Bond distances (A) Bond angles 0.041

(A,

-

His-175 His-175 Glu 8.0-2.1

--f

Cys

18,495 0.210 0.028 0.030

Reflections with F > 2dF). b R = IIFo-FcI/ZFo.

dye and CCP. All measurements were carried out at room temperature and in the dark when possible. Gln mutant was deterThe reduction potential of the His-I75 Protein Expression, Purification, and Crystallization-Purification of mined in a similar manner except that indigo carmine was used as the recombinant CCP was according to Fishel et al. (1987). The proteins indicator dye rather than riboflavin. The 2.5-ml final cuvette volume indigo carmine and 0.15 PM riboflavin.The were twice crystallized by dialysis against distilled water. The yields typically contained 15 were typically80 mg of crystalline CCP per liter of cell culture. Diffrac- reduction of indigo carmine was followed at 406 nm, a protein isosbestic tion quality crystals were prepared from 2-methyl-2,4-pentanediol ac- point, and protein reduction a t 430 nm after subtracting the contribucording to Edwards and Poulos (1990)with the modifications described tion due to indigo carmine. The reduction potential was determined by calculating the potential, by Sundaramoorthy et al. (1991) and Choudhury et al. (1992). Data Collection and Refinement-X-ray intensity data were collected E, at each point according tothe Nernst equation, Cys and His-175 Glu from a single crystal each of the His-175 E = E, + 29.8 logfdye]oddye]red (Eq. 1) mutants as described earlier for the His-175 .--f Gln and His-175 + Gln/Trp-l91+ Phe mutants (Sundaramwrthy et al., 1991).Asummary where E, at pH 5.5 for riboflavin is -133 mV (Draper and Ingraham, of data collection is provided in Table I. 1968) and -40mVfor indigo carmine (Meyer and Treadwell, 1952). Refinement of the His-175 + Cys mutant was straightforward and Plots of log [CCPlox/[CCPlred uersus E yielded both the number of converged to R = 0.17 using restrained least squares procedures (Hen- electrons, n, transferred during reduction (slope = n), and the protein drickson and Konnert, 1980). The wild type CCP structure first was redox potential, E, from the x intercept. refined against the mutant data set and then theCys ligand side chain EPR Spectra-For EPR spectroscopy, wild type and mutant CCPs fit to F, - F, and 2F, - F, electron density maps. Residual electron were converted to Compound I at 0 "Cby addition of an equimolar density around the Cys sulfur atom indicated that the Cys had been amount ofH,O,. Samples were transferred into quartz EPR sample oxidized to cysteic acidand that one of the cysteic acidoxygen atoms is tubes and frozen within 30 s after theaddition of peroxide by immersing bonded to the iron atom. Modeling a cysteic acid-iron bond eliminated the tubes into liquid nitrogen. EPR spectra were measured with a Varian X-band EPR spectrometer with 100-kHz field modulation. An all residual F, - F, difference density around the ligand. Refinement of the His-I75 --* Glu mutant was more problematic Air-Products Helitran (LTD3-110) liquid helium cryostat was used to using restrained least squaresprocedures. However, switching to simu- maintainthe EPR samples at cryogenic temperatures. Microwave lated annealing worked well. Using X-PLOR (Brunger, 19921, the wild power saturation was performed at 77 K. type CCP structure, with no side chain at position 175, was refined Steady State Kinetics-The steady state oxidation of ferrocytochrome against the His-I75 + Glu data setusing positional refinement for 100 c was determined as described previously(Choudhuryet al., 1992).The cycles followed by one round of simulated annealing refinement. The reactions were carried out at 25 OC in 0.1 M Tris a c e t a t a e s buffer a t final R factor was 0.21. Refinement statistics are presented in Table 11. pH 5.5. Horse heart cytochrome c was reduced with sodium dithionite Optical Spectra-All UVNlsible spectra were collecteda t 25 "Cin 0.1 and separated from excess dithionite with a small Sephadex G-25 colM Tris acetatemes buffer using a Cary 4 UVNisible spectrophotometer. umn. All kinetics were performedwith 40 p cytochrome c, which was The lifetime of the oxy-fen-yl center in CCP CompoundI was estimated greater than 90% reduced, using E,,, nm = 27.6 m?.-l cm". The 30% by adding a molar equivalent of H,O, to the enzyme and following the hydrogen peroxide stocks were periodically standardized with KMnO, visible absorption spectrum as a function of time. (Fowler and Bright, 1935). Enzymeconcentrations ranged from 0.07 to 1nM, depending on the activity of the sample. In addition to measuring Reduction PotentialSince the mutants were unstable at alkaline pH, the reduction potential determinations were measured only at pH the kinetics at pH 5.5, the steady state oxidation of ferrocytochrome c 5.5. The His-175 + Glu mutant proved unstable during photochemical was also measured as a function of pH, from 4.0 to 8.0, in the buffer titration so only the reduction potential of the His-175 + Gln mutant system described above. Someof the mutants defective in their ability could be determined with any degree of accuracy. The reduction poten- to oxidize cytochrome c were tested for their ability to oxidize a small tial of wild type CCP at pH 5.5 was determined by photochemical nonphysiologicalsubstrate, ferrocyanide. Asakura and Yonetani (1969) titration using a 3-ml anaerobic cuvette essentially as described by had found that there were certain modifications of the heme that alMauro et al. (1988). Typically, the reduction was carried out in a final tered the ability of CCP to oxidize cytochromec but not hydrogen peroxide or ferrocyanide.Thus it would be possible to uncouplethe peroxvolume of 2.5 ml consisting of 0.1 M potassium phosphate or sodium acetate buffer, pH5.5,30 PMriboflavin, 10mM EDTA, and 3 PM protein. idase and oxidase activities of CCP. Steady state assays of the peroxidedependent oxidation of potassium ferrocyanidewere carried out in 0.1 M All solutions were filtered before use.The protein was initially placed in H,O, at 420nmby assuming an extinction a side arm and the other components deaerated by vigorous bubbling KPO,, pH 6.0, with 500 coeficient of 1.0 mM-l cm" for ferricyanide. with scrubbed argon for at least 90 min. The protein then was mixed It should also be noted that CCP doesnot saturate in this assay and with the buffer. Photochemical reduction was carried out by illuminating the cuvette with a small light source forincreasing lengths of time, that Eadie-Hosfteeplots are not linear (Goodin etal., 1991) overa large followed by stirring for 2 or 3 min to allow the system to reach redox range of peroxide concentrations. For determining values of kc,, and Ifrn, equilibrium. After each illumination, the absorption spectrum was the concentration range of peroxide used gives reasonably linear Eadiemeasured from 380 to 700 nm. Riboflavin reduction was monitored at Hosftee plots and was chosen to be comparable with other published 466 nm, a protein isosbestic point, and reduction of wild type CCP was values using the same assay procedure. For the pH-dependent studies, followed at 410 nm after subtracting thecontribution due to riboflavin. the peroxide concentration used was much higher and the V,, (initial Once illumination ceased to cause further reduction, the cuvette was velocityienzyme concentration) values reported are higher than those opened tothe air and solid Na,S,O, added to completely reduce both the reported in Table I11 from the Eadie-Hofstee plots. It R,, = 1I Ii-(Ii) I EIi, where Ii = intensity of the ithobservation and (Ii) = mean intensity.

--j

-

-

CCP Proximal Ligand Mutants

20241

FIG. 1.Stereo views of the omit electron density maps aroundthe two proximal ligand mutants. Top, ZF,,- F, electron density map contoured at lo of the His-175 + Glu mutant. The Glu-175 atoms were excluded prior to running several cycles of refinement followed by calculation of the map. Middle, F,, - F, electron density map contoured at 30 of the His-175 + Cys mutant. The oxygen atoms of the cysteic acid but not the sulfur were excluded fromthe refinement and phase calculation. Bottom, W , - F, electron density map contoured at l o of the His-175 Cys mutant. The cysteic acid 175 atoms were excluded from the refinement and phase calculation.

-

Stopped Flow-Experiments were carried out using a Hi-Tech Scientific SF-51 Stopped-Flow Spectrophotometer using a 10-mm path length. Reactant reservoirs, drive syringes, mixer, and the observation chambers were thermostated at 20 "C bya circulating water bath. Final protein concentration was 2.5 PM,while the range of H,O, concentration used was 2.5-20 p.All studies were carried out in 10 m~ potassium phosphate and 1m~ EDTA at pH 6.0 with an ionic strength of 0.1 M. Transient spectra were recorded with a rapid scanning stopped flow spectrophotometercomprised of a Durrum Dl10 stopped flow spectrometer and a Princeton Applied Physics (EG & G) OMA-I11 multichannel analyzer with a 1463 detector controller card and a 1214 photodiode array detector (Brzovic and Dunn, 1993). The experiments reported herein used 600 pixels and a repetitive scan rate of 10 ms/scan with a

wavelength resolution of d . 5 nm. Scan times in Fig. 9 refer to the time interval between stopping of flow and initiation of a scan. RESULTS

Crystal Structures-Omit electron density mapsfor both the His-175 + Glu and His-175 + Cys mutants are presented in Fig. 1. The Glu-175 mutant shows strong electron density between the side chain and iron, indicatingcoordination. No assumptions were made during the refinement regarding ligand bond distances and angles and these parameterswere not restrained during therefinement.

19 1

D235

D235

His 175

GIn 175

D23S

GInl75~hel91

D23S

D235

Glu 175

Cys 175

FIG.2. Molecular modelsof the 4 proximal side mutants andof wild type CCP.Dark spheres are oxygen, Zightly shaded spheres are figures were prepared withthe program MOLSCRIPT carbon, andheavily shaded spheres are nitrogen.Thin lines denote hydrogen bonds. These (Kraulis,1991).

The Cys-175 mutant refined well but residualF, - F electron density persisted to surround the cysteine sulfur atom (Fig. 1, ~ i ~ d l eThis ) . residual electron density shows three regions approximately tetrahedrally disposed about the sulfur atom. This finding strongly suggested that Cys-175had been oxidized to cysteic acid. Assuming cysteic acid followed by refinement showed no residual Po- F, electron density around the ligand. The final 2 F o - F, omit electron density map shows continuous density between a cysteic acid oxygenand the iron, indicating oxygen-iron coordination. The ligand-iron distance is approximately 2.4 A, a value which is significantly longer than the2.0 A found in the Glu-175 mutant. Molecular models for all four proximal ligand mutants are shown in Fig. 2 and for comparison the wild type structure is included. The Glu-175 and Gln-175 side chains adopt slightly different conformations,but both are close to the iron atom, 2.0 A and 2.2 A for the Glu and Gln side chains, respectively. Both ligands also are close enough to Asp-235 for hydrogen bonding interactions. We have assumed that in the Gln-175 mutant, the side chain oxygen atom coordinates with the iron, leaving the side chain nitrogen available for hydrogen bonding with Asp235. Glu-175 also is about 2.7 A from Asp-235, so such a carboxylate-c~boxylateinteraction suggests that these two side chains may share a proton. Both mutants also have a new water molecule in the proximal cavity close enough for inter-

actions with the mutant side chain. This new water occupies space left by the His ligand, since Gln and Glu have slightly smaller volumes than His. The His-175 Cys mutant also forms a bond with the iron atom (approximately 2.4 A), but in this case it isa side chain oxygen ofcysteic acid. Apparently, Cys-175 is oxidized to cysteic acid during purification and reconstitution of the apoenzyme with heme. Optical Spectra-Theoptical spectrum of themutants closely resembles that of wild type CCP as reported earlier (Choudhury et dl., 1992) at pH 5-6. A pH-independent spectrum ischaracteristic of active, pentacoordinate and high-spin CCP (Yonetaniand Anni, 1987).Although the mutants are active, the spectra are extremely pH-sensitive. Above pH 6.5, all proximal ligand mutants exhibit a large decrease in the Soret maximum indicating loss of heme or significant destabilization of the heme pocket. Addition of 1 H,O, eq leads to the formation of a spectral intermediate that closely resembles that of wild type Compound I in the Gln-175 and Glu-175 mutants. The Cys-175 mutant exhibits much smaller spectral shifts. As noted earlier (Choudhury et at., 1992), the oxyferryl center in the proximal ligand mutants is much less stablethan inwild type CCP. For wild type CCP, the characteristic Compound I spectrum returns to the resting state spectrum over a period of hours with "-f

20243

CCP Proximal Ligand Mutants wild type

Pi==!

wild type

FIG.3. X-band EPR spectra of wild typeandmutant CCPs: (a)ferric resting state enzymes and ( b ) Compound I aftertheaddition of one equivalentof Hs02.

~

His175Glu

0.0

the proximal ligand mutants, a decrease in theSoret maximum occurs within minutes indicating that the oxyfenyl center is much less stable in theproximal ligand mutants. EPR-Fig. 3 illustrates X-band EPR spectra of wild type and mutant CCPs a t 4.2 K. The wild-type enzyme showed predominantly rhombically distorted axially symmetric high-spin signals with a minor axially symmetric high-spin signal indicating that wild-type recombinant CCP is essentially 5-coordinate even in the absence of glycerol (Yonetaniand Anni, 1987). No freezing-induced low-spin signal was detected. The His-175 Glu and His-175 .+ Gln mutants exhibited only axially symmetric high-spin EPR signals of a 6-coordinate high-spin type. The axial high-spin signals of wild type, His-175 Glu, and His-175 -+ Gln were partially reduced and replaced by the rhombic signals in the presence of 60% glycerol. Wild type Compound I exhibited the axially symmetric, free radical-type signal with absorption extreme a t gll = 2.037 and g i = 2.004 (gil > g l ) , which is indistin~ishablefrom Compound I for CCPisolated from bakers' yeast (Hori and Yonetani, 1985). %e His-175 -+ Gln and His-175 Glu mutants, on the other hand, exhibited another axially symmetric EPR signal with inverted g values ( g i = 1.99 < pll = 2.00). These two g values were so close that the EPR signals gave an almost isotropic appearance. Despite the difference in line shape, the EPR signals of both mutants showed microwave powersaturation characteristics (Fig. 4) similar to those of wild type CCP, indicating that they are not a typical, isolated protein free radical. Reduction Potentials-The reduction potential for wild type and theHis-175 Gln mutant were determined from the plots shown in Fig. 5. The redox potential of wild type CCP was determined to be -102 * 3 mV ( n = 1.0 2 0.1)at pH 5.5, similar to the value of -108 mV previously determined by Conroyet al. (1978). Although the photoreduction titration of the mutant was not as clean as for the wild type, it is clear that thereduction potential of CCP Gln-175 increased by at least +75 mV relative to the wild type enzyme. Steady State Actiuities-In Table I11 are presented the steady state parameters for the various mutants. The His-175 -+ Gln kcatis very nearly the same as wild type CCP, while the His-175 .+ Glu mutant is about 7 times as active. The His-175 Cys exhibits about 7% wildtype activity, while the double mutant, His-175 -+ GlnlTrp-191+ Phe, has about 20% wild type activity. The pH rate profiles are shown in Fig. 6. The pH depen-

2.85

4.0 1

2.0 Magnetic Field (kG)

3.65

3.25

Magnetic Field (kc)

I

I

-0-

''.

'\

I

H175H (Wild type)

*

H17SE

-U

H175Q

-+

-+

--j

-j

-j

3

cl

I

-1

I

I l l

I

0

I

Ill

I

+ l

I

I

II

+2

Log Power (mmw)

FIG.4. Microwave saturation characteristics of wild typeCCP and two axial ligand mutants at 77 K using a liquid nitrogen Dewar. Amplitude = peak to peak amplitude farbitrazy units);power = microwave powerin milliwatts. For comparison, the saturation behavior of a free radical also is shown.

dence is very muchthe same for all mutantswith the exception of the double mutant. Here the activity does not peak between pH 5 and 6 as do the others but insteadcontinues to increase as the pH decreases. When ferrocyanide is used as the substrate rather than cytochrome c, the mutantsexhibit wild type levels of activity with the exception of the His-175 -+ Cys mutant which is 4 % as active. Stopped Flow Kinetics-Stopped flow traces of the wild type, His-175 Gln, and His-175 Gln/Trp-191+ Phe mutants are shown in Fig. 7. Since the resting state andCompound I spectra of the mutants were slightly different, it was necessary to follow the reaction of the mutants at a wavelength different than 424 nm usually used for wild type CCP, This was empirically determined for each mutant, and 414 nm was a common wavelength that could be used for all mutants.Similar to wild type CCP, the His-175 Gln mutant exhibits a fast increase in absorption that is H,O,-dependent followed by a slow,H,O,independent increase in absorption (Balny et al., 1987). The His-175 Gln/Trp-191 .+ Phe double mutant, however, behaved quite differently. There is first a rapid decrease in absorption followed by a slow increase in absorbance. -+

-+

-+

--j

CCP Proximal Ligand Mutants mutant is at least as fast and probably faster than wild type CCP. The stopped flow traces for the His-175 ”-f Glu mutant resembled that of the His-175 + Gln mutant except the initial fast phase was, like the double mutant, too fast to obtain accurate rate constants. The double, His-175 + GlrdTrp-191 ”-f Phe, mutant stopped flow trace (Fig. 7) is very similar to thatobserved for the single Trp-191+ Phe mutant (Erman et al., 1989). For the Trp-191 ”-f Phe mutant the initial decrease in absorbance was attributed to formation of a porphyrin rr cation radical which slowly converts toa species whose spectrum is characteristicof wild type CCP Compound I (Erman et al., 1989). We therefore carried out rapid scan experiments to if see the early intermediatesformed in thereaction betweenthe double mutant andperoxide has an absorption spectrum characteristicof a porphyrin rr cation radical. The results are displayed in Fig. 9. Only the double mutant shows an early intermediate witha large and clearly defined decrease in the Soret maximum. These traces are very similar to those obtained by Erman et al. (1989) for the single Trp-l9l+ Phe mutant, suggesting that double the mutant also forms a porphyrin rr cation radical. The double mutant spectrum “relaxes” back to a species which more closely resembles that of wild type CCP Compound I. That is, the Soret maximum exhibits a red shift with new bands appearingbetween 500 and 600 nm. The His-175 + Gln mutant exhibits changes in the Compound I spectrum between the first (10 ms) and last (828 ms) spectrum obtained. This change probably represents the H,O,-independent phase of the reaction. The last spectrum obtained again resembles what is expected for Compound I. The His-175 + Glu mutant spectra aremore difficult to interpret. The last spectrumobtained at 2,000 ms looks somewhat like Compound I with new bands appearing between 500 and 600 nm.

Reduction potential(mV)

-80

-60

-40

-20

0

Reduction potential(mV)

FIG.5. Plot of log [CCP(Fe3+)I/[CCP(FeZ+)l uersus redox potential for the photochemicaltitration of (a) wild type CCP and the ( b )His-175 + Gln mutant.

TABLE I11 Steady state activity parameters for wild type and mutant CCPs Steady state parameters were determined from Eadie-Hofstee plots as described under “Methods and Materials.” Enzyme

-

k,,

KM

s”

w

1463Wild type His-175 63.6 1673 Gln His-175 + Glu 11174 His-175 + Cys 53.9 96 His-175 283+ Gln Trp-191+ Phe

6.7 240.3 23.2

Wild type activity

k,JKM

%

w” s”

100 114 764 6.6 19.3

218.4 26.3 46.5 1.8 12.2

Values of the observed rate constants for the reaction with H,O, were obtained by fitting the rateof change of the absorbance to a two exponential empirical equation AA = A A A exp(-k,t)

+A A , exp(-k,t) + c

(Eq.2)

A nonlinear least squares regression analysis was used to determine values of the rate constants and pre-exponential factors. The reported values are the mean of at least four determinations. The second-order rate constants were estimated from the plots shown in Fig. 8. The value estimated for wild type CCP, 3.9 x lo7 M - ~s-l, agrees well with various published values for recombinant CCP, 3.03 x lo7 M - ~s-’ (Goodin et al., 1991) and 4.5 x lo7 M - ~s-l (Erman et al., 1993). For the double mutant, a majorityof the absorbance decreaseat 414 nm occurs during the mixing time of the instrument, which precludes accurate estimations of the second-order rate of reaction with H,O,. Therefore, the formation of Compound I in the double

DISCUSSION

Crystal Structures All three proximal ligand mutants form a bond with the heme iron atom. The Cys mutant, however, was oxidized to cysteic acid. In contrast, the same mutant inmyoglobin where the proximal His is converted t o Cys has been reported to form an S-Fe bond (Adachi et al., 19931, as evidenced by a number of spectral probes including resonance Raman spectroscopy. Interestingly, the absorption spectrum of the myoglobin mutant exhibits a band at 390 nm, which owing t o its similarity to P450, was taken as evidence for a n S-Fe bond. With CCP, the His-175 + Cys mutant also exhibits a Soret band near396 nm immediately after the apoenzyme is reconstituted with heme, but within several minutes this band shifts to400 nm. Moreover, there is a slight increase in activity concomitant with the spectral shift which plateaus at about 7% wild type activity. The most likely explanation for these spectral changesis that the Cys is oxidized to cysteic acid, and coordination of the cysteic acid side chain oxygen atom results in a shift of the Soret band from 396 to 400 nm. This behavior would indicate that inCCP, the hemeand/or protein is more rigidly fixed than in myoglobin which prevents the requiredmovements needed t o form a stable bond with the smallerCys side chain.aNevertheless, theCys sulfur atomwould remain within 3-4 A of the iron atom, and, in the presence of dissolved oxygen, the iron atom could mediatethe oxidation of sulfur. Moreover, the smaller Cys residue could allow 0, into theproximal pocket for direct reaction with the sulfur atom. That in the His-175 + Glu and His-175 ”-f Gln mutants, there issufficient room for a new water molecule in the proximal pocket lends credence to this suggestion.

CCP Proximal Ligand Mutants

20245

4000

I -J-

3200 3000

h

r

n F I

I

0

2800

0 Q) v)

Q) v) v

W

2600

Q)

2400

\

2200

>

Q)

\

0

>

0

2000

1800 3

6

5

4

7

,

1 loo

n

r

u

4

7

6

5

8

PH

PH

10000

3

9

8

n

9000

v-

La

Q) v)

80

J

-

Q) v)

U

8000

60-

Q)

\

Q)

0

>

>o \

7000

40-

20 3

4

5

6

7

9

8

! 3

.

I

4

PH

.

I

5

.

I

6

.

, . , 7

.

8

PH

F

His175Gln T r p l 9 1 Phe

400 Q) v)

300

3

4

5

6

7

8

PH FIG.6. pH rate profilesfor wild type and mutant CCPs. Each experimental point was determined a t a single concentration of substrates: 40 p ferocytochrome and 180 p H202 in 0.1 M Tris acetate/MES buffer a t the indicated pH values. The data are plotted as V,, (initial velocity/enzymeconcentration) uersus pH.

EPR The EPR spectra of wild type Ccp indicates a predominantly &“rdinate high-spin species, while the His-175 + Glu and His-175 + Gln mutants are B-coordinate in phosphate buffer,

pH 7.0, at 4.5 K. The His-175 + Gln EPR spectrum is consistent with the Crystal structure where there is continuous electron density between the iron and theaxial water, ligand in the distal pocket suggesting coordination at a distance of approxi-

MutantsLigand Proximal CCP

20246

His175Gin/TrplOlPho st 414nm

Hi11175GInat 414nm

I ‘

I

I

0.19

1

0.16

-

0.17

-

4

0.13

~~

0.04

0.08

0.12

0.16

0.12

~

0.20

-

lime in seconds

I

I

I

I

I

-

0.14

l.W

I

-

1

0

1

2

1

1

5

6

lime in seconds

typical isolated protein free radical but instead may be magnetically coupled to the heme iron (S = 1). Recently Houseman et al. (1993) have provided a n interesting explanation for these unusual features of the CCP Compound I free radical in terms 600 of weak exchange between the ferry1 iron (S = 1)and the Trpr J / 191 free radical (S = 1/2) with a distribution of exchange cou0 pling parameters. The EPR featuresof CCP Compound I were 8 400 reasonably interpreted by a distribution of two components, antiferromagnetically (J < 0) and ferromagnetically (J > 0) coupled Trp-191 free radical. Goodin and McRee (1993) reported thatCCP Compound I of the Asp-235 + Glu mutant exhibited an EPR spectrum with a significantly altered anisotropy. Subsequent EPR and ENDOR examinations (Houseman et al., 1993) showed that the EPR spectrum of the Asp-235 + Glu Compound I could be adequately represented by a ferromagnetically (J > 0) coupled Trp-191 radical. Asp-235 hydrogen bonds with both the proximal histidine ligand, His-175, and the indole ring nitrogen of Trp-191. The crystal structure of the Asp-235 + Glu mutant shows a slight alteration in the hydrogen bonding geometry and a partial rotation of His-175, the proximal ligand (Goodin and McRee, 1993). Such a change could be one reason why the normal distribution of exchange coupling and elimination of the antiferromagnetic (J < 0) coupling. The EPR spectra of Compound I for the His-175 + Glu and His-175 + Gln mutants arepractically identicallyto thatof the 0 4 . , . , . I . 1 . 1 . 1 Asp-235 + Glu mutant. Therefore, it is likely that the EPR 0 2 4 6 8 10 12 centers of the two proximal ligand mutants both consist of [H202] F M ferromagnetically coupled Trp-191radical. In seeking some FIG.8. Pseudo first-order rateof reaction in the formationof common grounds for why the EPR properties of the His-175 Compound I as a function of peroxide concentration.The second- Gln mutant is so similar t o that of the proximal ligand muthe order rates were estimated fromthe slope of the best fit line through tants, we note that in theAsp-235 + Glu mutant theproximal data. ligand undergoes a slight rotation but Trp-191 does not move mately 2 A. However, the His-175 + Glu mutant does not (Goodin and McRee, 1993). Hence, the T stacking interaction exhibit continuous electron density between the iron and water, between His-175 and Trp-191 has been altered in theAsp-235 indicating a pentacoordinate iron. Since the crystals are pre- + Glu mutant. The proximal ligand mutants also have lost the pared in 30% methyl pentanediol, this inconsistency with the interaction formed by stacking of the proximal histidine imidEPR data could reflect the known effects of alcohols and tem- azole ring parallel to the Trp-191 indole ring. Moreover, the changes in hydrogen bonding geometry between Glu-235 and perature (Yonetani and Anni, 1987) on the ligation geometry. Glu mutant probably weakens this The free radical species of CCP Compound I (Yonetani, 1976) His-175 in the Asp-235 has been interpreted to have axial symmetry with gll = 2.037 interaction which could, in part, explain why the Asp-235 and g l = 2.00 (Hori and Yonetani, 1985). EPR and ENDOR Glu mutant exhibits a higher redox potential than wild-type studies of wild type recombinant and mutant CCPs indicate CCP (Goodin and McRee, 1993). Taken together, these obserthat the radical iscentered on Trp-191 (Sivaraja et al., 1989). vations implicate the proximal ligand in coupling the Trp-191 However, the unusual EPR characteristic of this radical signal, radical to theferry1 center. Communication between the proximal His-175 and Trp-191 also is implicated by the crystal strucsuch as the large g anisotropy, broad line shape, line shape (i.e. harder topower- ture of the complex between nitric oxide (NO) and CCP (Edvariability and, short spin relaxation time wards and Poulos, 1990). Here it was found that among the saturate), suggested that the free radical (S = 1\21 is not a wild type CCP

-

800

Y

1

I

4

V

--j

-

-

CCP Proximal Ligand Mutants

20247

8.6

8.7

8.6

8.5

8.5 8.4

8.5 Y

C

a

V

C

8.4

a

f p"

al Y e

8.4

0

0

0

U

r

8.3

g 8.2

a

E

E

8.3

a 8.3

8.2

8.2 8.1

8.1

8. I

8.8

8.8

488

588

688

Wauelength(nm1 Wauelength (nm)

8.8 488

588

688

488

588

688

Wauelengthlnm)

FIG.9. Rapid spectral scanning of the reaction between the various mutants and an equimolar amount of H,O,. The scan times shown are for His-175 + Glu: 1 = 0 , 2 = 131, 3 = 787 ms, and 4 = 2019 ms; His-175 + Gln: 1 = 0 , 2 = 10, 3 = 141, and 4 = 384 ms; His-175 + Gln/Trp-191+ Phe: I = 0 , 2 = 2 1 2 , 3 = 787 ms, 4 = 1211, and 5 = 2019 ms.

various ligands to CCP (fluoride, cyanide, carbon monoxide, most likely due t o changes in electron transfer rates. One posand NO), NO is unique inits ability to cause structural pertur- sible reason why the Glu-175 mutant is so active is that the bations around Trp-191. It was argued that this is due to the negative charge on the Glu side chain provides additional elecdelocalization of the unpaired spin on NO t o the proximal trostatic stabilization to Fe4+.This effect would then increase pocket including Trp-191 viathe His-175 ligand (Edwards and the thermodynamic drivingforce of electron transfer andhence Poulos, 1990). Changes in this interaction might also be ex- increase the rate. pected to significantly alter the EPRproperties of the Trp-191 Compound I Formation-The stopped flow work shows that radical. Although further EPR and ENDOR studies areneeded the precise nature of the proximal ligand is not important for for definitive identification of the EPR center in the proximal high rates of peroxide 0-0 bond cleavage. These data, coupled ligand mutants,it is reasonable to conclude, given the current with the mutagenesiswork of Erman et al. (1993), demonstrate EPR data, that the free radicalCompound in I in theproximal that the distal peroxide binding pocket and especially the distal ligand mutants remains associated with Trp-191, but with an His, and not theproximal ligand, are responsible for high rates altered coupling to the ferryl iron. of Compound I formation. Nevertheless, the proximal ligand is important in determining the stabilityof Compound I. This is Reduction Potential not surprising since replacing the proximal ligand disrupts a Although attempts were made to determine the reduction set of interactions involving the ligand,His-175, and two other potential of all mutants,only the His-175 + Gln and wild type residues, Trp-191 and Asp-235. Asp-235 H-bonds to both HisCCP yielded decent titration curves. The reduction potential of 175 and Trp-191, while the indole ring of Trp-191 and the wild type CCP at pH 5.5 (-102 mV) agrees well with published imidazole ring of His-175 form a parallel T stacking interacvalues (-108 mV, Conroy et al. (1978)). The His-175 4 Gln tion. Thisstructure is thought tobe important instabilizing the mutant, however, exhibits a reduction potential at least +75 freeradical of CCP Compound I. Mutagenesis of Asp-235 mV higher than wild type CCP. This result was anticipated coupled with crystal structures and spectral work clearly show based on previous suggestions on what helps to control the how sensitive the His-Asp-Trp triad is to subtle perturbations heme redox potential. In both CCP and lignin peroxidase (Pou- (Goodin and McRee, 1993). 10set al.,19931, the proximal His ligandis H-bonded to a buried Why Is the Double Mutant Active?-The most surprising reAsp side chain. This interaction is thought to impart a partial sult from our studies on proximal side mutants is that the imidazolate character to the Hisligand which helps to stabilize double mutant, His-175 + Gln/Trp-191 + Phe, retains 20% Fe3+relative to Fez+, thus lowering the redox potential relative wild-type activity (Choudhury et al., 1992) while the single to myoglobin, which lacks a His-Asp interaction. Evidence to Trp-191 + Phe mutant is essentially inactive (Mauro et al., support thisview stems from mutagenesis work where conver- 1988). Trp-191 has been implicated as an essential partof the sion of this Asp in CCP to eitherAsn or Ala increases theredox electron transfer circuit between ferrocytochrome c and CCP potential (Goodin and McRee, 1993). The Gln ligand does not (Pelletier and Kraut, 19921, and theevidence clearly supports a have a dissociable proton as does His and thus theHis-175 + crucial role for Trp-191 (Mauro et al., 1988). Nevertheless, it Gln mutant is less likely to carry a partial negative charge. also is clear that even though a Trp at position 191 is imporTherefore, a Gln could not as effectively stabilize the lower tant, a Trp-191 radical is not. Ho et al. (1983) have shown that redox state and hence, the observed increase in reduction po- CCP can be oxidized to 1 eq above the resting ferric state, tential. giving an oxyferryl center butno free radical. This form of CCP accepts an electron from ferrocytochrome c at rates about 1.6 Enzyme Activity and Roleof the Proximal Ligand times slower than CCP Compound I containing both the oxySteady State Actiuity-With the exception of the His-175 + ferryl and radical centers (Hazzardet al.,1987), demonstrating Cys species, all the mutants exhibit high levels of activity in that a Trp-191 radical is not essential for electron transfer. steady state assays,while the His-175 + Glu mutant ishyper- Furthermore, electron transfer between Trp-191 and the oxyactive. Hyperactivity in CCP mutants has been observed ear- ferryl center appearsto be too slow to account for the observed lier (Goodin et al., 1991). Since the stopped flow work shows turnover rates in the steady state reaction (Ho et al., 1984; that the proximal ligand mutants still react with H,O, as fast Summers and Erman, 1988). These observations indicate that as wild type enzyme, the rate-limiting step is very likely elec- Trp-191 is partof an importantelectron transfer path, but that tron transfer (Goodin et al., 1991), so changes in activity are the radicalplays only a secondary role in controlling rates and

20248

CCP Proximal MutantsLigand

specificity of electron transfer. One possibleexplanation for whythe double mutant exhibits such unexpectedly high levels of activity, eventhough Trp-191 is thought to play such a critical role, is that a new path of electron transfer has opened up, thus by-passing the need for Trp-191. The rapid scan data show that the double mutant reacts with peroxide to give a porphyrin rn cation radical with an estimated half-life -700 ms. From the data of Erman et al. (19891, we estimate the half-life for the same porphyrin T cation radical in the single Trp-191 + Phe radical to be 20 ms. Thus, the longer lifetime of the heme-centered radical in the double mutant provides sufficient time for ferrocytochromec to bind and deliver its electron to the porphyrin radical via some alternate site than thefavored route through Trp-191. Attractive and simple as thishypothesis is, however, it does not quite fit all the kinetic data. The turnover number for the double mutant is about 280 s-', which means that if we assume electron transfer israte-limiting, electron transfer takesabout 3-4 ms. Therefore, the half-life of the porphyrin radical in theTrp191 + Phe mutant (20 ms, Erman et al. (1989))is long enough for electron transfer to take place. It then follows that if the porphyrin T cation radical is an effective recipient of electrons from ferrocytochromec, then thesingle Trp-191+ Phe mutant should be active, yet it clearly is not (Mauro et al., 1988). While a long-lived porphyrin radical may beone part of why the double mutant is active, additional factors must contribute. The instability of Compound I in all proximal ligand mutants indicates that Compound I in the mutants is more reactive which could be due to changes in either kinetic or thermodynamic barriers, or both. In our earlier work (Choudhury et al., 1992),we estimated, based on the dataof Conklin and McLendon (19881, that the increased rate of electron transfer in the double mutant would necessitate a 0.4 V increase in driving force. This would require that theredox potential of Compound I increase t o the level of H202itself (1.5 V). Thus, an increase in thermodynamic driving force couldpotentially explain why the double mutant ismore active. The dependence of activity on pH also could be of relevance here. Only the double mutant exhibits a continued increase in activity as thepH is lowered. While the precise explanation for why this occurs is open to speculation, especially since we do not know if electron transfer continues t o be rate-limiting at low pH, it could reflect a fundamentally different electron transfer mechanism in the double mutant. For example, at low pH, surface carboxylates become protonated which would be expected t o effect the electrostatic CCPqtochrome c complex. If this is the explanation for the pH dependence, then the double mutant may be more effective at using an alternate, secondary site for electron transfer. Another possibility is that lowering the pH may alter the electrostatic environment of the proximal pocket by protonating Asp-235. This in turn should significantly alter theenergetics of electron transfer. Relevance to Other Redox Systems-Similar changes in driving force and reorganization energy are well understood for at least one system: cytochromes P450. Cytochrome P450,, will accept electrons from its naturalredox partner, putidaredoxin, only when P450,,, is in thesubstrate-bound form (Sligar and Gunsalus, 1976). Substrate binding displaces several solvent molecules in theactive site including an aqua ligand (Poulos et al., 1986)which leads to an increase in redox potential (Sligar and Gunsalus, 1976). A comparison of the substrate-free and bound x-ray structures shows that there is very little movement in protein atoms and the primary change is in the axial aqua ligand and nearby solvent structure (Poulos et al., 1986). The redox potentials are now such that thetransfer of an electron from putidaredoxin to substrate-bound P450,,, is thermo-

dynamically favored. In addition, the change in active site solvent content should have a significant influence on the reorganization energy (Marcus and Sutin, 1985). The same principles may be operating in the CCP/cytochrome c system. By changing the proximal ligand, there could well be large changes in both the redox potential of Compound I and the reorganization energy. Thus the mutants could still use the favoredTrp-191 pathway suggested by Pelletier and Kraut (19921, even with a Phe at position 191. It is important t o emphasize here that, as discussed earlier, a radical at position 191 appearsnot t o be important in supporting electron transfer. Thus, there is no need to postulate the unlikely formation of a Phe-191 radical in the double mutant. Despite the fact that the double mutant exhibits a long-lived porphyrin rn cation radical, our data do not allow us to choose between a route of electron entry that uses the same path as wild type CCP, or a new route involving the porphyrin radical. Since the porphyrin T cation radical is long-lived, it may be possible to study electron transfer to the porphyrin radical using rapid reaction methods to help resolve this issue. In summary, the key findings in this work are asfollows. 1) The precise nature of the proximal ligand is not critical in achieving high rates of peroxide 0-0 bond cleavage but the ligand is importantfor controlling the stability of the oxyferryl center once formed. 2) Interactions between the proximal ligand and Trp-191 are important for controlling coupling between the Trp-191radical and heme iron unpaired spin centers. 3) Changing the proximal ligand to a neutral residue that cannot readily carry a partial negative charge (His to Gln) results in an increase in reduction potential. An increase is expected if simple electrostatic interactions help to control reduction potentials. 4) It is possible, using site-directed mutagenesis, to alter the location and lifetime of the Compound I free radical center. This finding could be important in engineering peroxidases to operate on molecules other than natural substratesby providing alternate sites of oxidation. Acknowledgments-We thank Dr. Kame1Danvish for development of the CCP expression and mutagenesis system, Joel Hoskins (Center for Advanced Research in Biotechnology (CARB)) for preparation of oligonucleotides, and Dr. Matt Mauro and Dr.Keith McKenney for valuable advice. REFERENCES Adachi, S., Nagano, S., Ishimori, R, Watanabe, Y., Morishima, Egawa, T., Kitagawa, T., and Makino, R. (1993) Biochemistry 32,241-252 Asakura, T., and Yonetani, T.(1969) J. Biol. Chem. 244,4573-4579 Balny, C., Ami, H., and Yonetani, T.(1987) FEBS Lett. 221,349-354 Brunger, A. (1992) X-PLOR 3.Z, Yale University Press, New Haven, CT Brzovic, P. S., and D u n , M. F. (1993) in Bioanalytical Instruments. Methods of Biochemical Analysis (Sueleter, C. M., ed) pp. 191-273, J. Wiley & Sons, Inc., New York Choudhury,R, Sundaramoorthy,M., Mauro, J. M., and Poulos, T. L. (1992)J. Biol. Chem. 267,25656-25659 Conklin, K. T., and MeLendon, G . (1988) J . Am. Chem. SOC.11,33454350 Conroy, C. W., Tyma,P., Daum, P., and Erman, J. E. (1978)Biochim. Biophys. Acta 537,619-622 Darwish, K., Li, H., and Poulos, T. L. (1991) Protein Eng. 4, 701-708 Draper, R. D., and Ingraham, L. L. (1968)Arch.Biochim. Biophys. 126,802-808 Edwards, S. L., and Poulos, T.L. (1990) J. Biol. Chem. 266,2588-2595 Erman, J.E., Vitello, L. B., Mauro, J. M., and Kraut, J. (1989) Biochemistry 28, 7992-7995 Erman, J. E., Vitello, L. B., Miller, M. A,, Shaw, A,, Brown, K. A., and Kraut, J. (1993) Biochemistry 32,9798-9806 Fishel, L. A., Villafranca,J. E., Mauro, J.M., and Kraut, J. (1987)Biochemistry 26, 351-360 Fowler, R. M., and Bright, H. A. (1935)J . Res. Natl. Bur: Standards 16,493-515 Goodin, D. B., and McRee, D.E. (1993) Biochemistry 32,33134324 Goodin, D. B., Davidson, M. G., Roe, J. A,, Ma&,A. G., and Smith, M. (1991) Biochemistry 30,49534962 Hazzard, J. T., Poulos, T. L., and Poulos, G . (1987) Biochemistry 26, 2836-2848 Hendrickson, W. A,, and Konnert, J. H. (1980) in Computing in Crystallography (Diamond, R., Ramaseshan, S., and Venkatesan, K., eds) pp. 13.1-13.23, Indian Institute of Science, Bangalore, India Ho, P. S., Hoffman, B. M., Kang, C. H., and Margoliash, E. (1983) J . Biol. Chem. 268,4356-4363 Ho, P. S., Hoffman, B. M., Solomon, N., Kang, C. H., and Margoliash, E. (1984)

LigandProximal CCP Biochemistry 23,41224128 Hori, H., and Yonetani, T. (1985)J. Biol. Chem. 260,349-355 Houseman, A. L. P., Duoan, P.E., Goodin, D. B., and Hoffman, B. M. (1993) Biochemistry 32,443043 Kaput, J., Goltz, S., and Blobel, G . (1982)J. Biol. Chem. 257, 15054-15058 Kraulis, P. J. (1991)J. Appl. Crystallogr. 24, 946-950 Kunkel, T. A,, Roberts, J. D., and Zokour, R. A. (1987)Methods Enzymol. 164, 367382 Marcus, R. A,, and Sutin, N. (1985)Biochim. Biophys. Acta 811, 265-322 Mauro, J.M.,Fishel, L. A,, Hazard, J. T., Meyers, T.E., 'Ibllin, G., Cusanovich, M. A., and Kraut, J. (1988)Biochemistry 27,6243-6256 Meyer, H. W., and Treadwell, W.D.(1952)Helu. Chim. Acta 36, 1444-1460 Pelletier, H., and Kraut, J. (1992)Science 258, 1748-1755 Poulos, T. L., and Finzel, B.C. (1984)in Peptide and Protein Reviews(Hearn, M. T. W., ed) Vol. 4,pp. 115-171, Marcel Dekker, Inc., New York

Mutants

20249

Poulos, T. L., Finzel, B. C., and Howard, A. J. (1986)Biochemistry 26,5314-5322 Poulos, T. L, Edwards, S. L., Wariishi, H., and Gold, M.H. (1993)J . Biol. Chem. 268,4429-4440 Sivaraja, M.,Goodin, D. B., Smith, M., and Hoffman, B. M.(1989)Science 245, 738-740 Sligar, S. G., and Gunsalus, I. C. (1976)Proc. Natl. Acad. Sci. U. S . A . 73, 10781082 Summers, F. E., and Erman, J. E. (1988)J. Biol. Chem. 263,14267-14275 Sundaramoorthy, M., Choudhury, K., Edwards, S. L., and Poulos, T. L. (1991)J . Am. Chem. Soc. 113,7755-7757 Takio, K., Titani, K., Ericson, L. H., and Yonetani, T.(1980)Arch. Biochem. Biophys. 203,615-629 Yonetani, T. (1976)The Enzymes 13,346-361 Yonetani, T., and Anni, H. (1987)J . Bid. Chem. 262,9547-9554