THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol ,262, No. 30, Issue of October 25, pp. 14479-14486.1967 Printed in U. S. A.

The Interaction and Photolabeling of Myosin Subfragment 1 with 3’(2‘)-0-(4-Benzoyl)benzoyladenosine 5’-Triphosphate*

(Received for publication, June 16, 1987)

Riaz MahmoodS, Christine Cremo, Kay L. Nakamaye, and Ralph G. Younts From the Biochemistry/Bwphysics Program, Institute of Biological Chemistry, and Department of Chemistry, Washington State University, Pullman, Washington 99164-4660

The photoprobe 3f(2’)-O-(4-benzoyl)benzoyladeno- sine 5”triphosphate (BzzATP) was used to character- ize the nucleotide-binding site of myosin subfragment 1 (SF,). Improved synthesis and purification of BzzATP are reported. ‘H NMR and ultraviolet spec- troscopy show that BzzATP is a 60:40 mixture of the 3‘(2’)-ribose isomers and that the cye1 is 41,000 M-’ cm”. Bz2ATP is hydrolyzed by SF, comparably to ATP in the presence of actin or K+, N&+, or M e + ions; and the product, Bz2ADP, has a single binding site on SF, ( E a = 3.0 X 10’ ”I). [‘H]Bz2ATP was photoincorpor- ated into SF1 with concomitant loss of K+-EDTA-ATP- ase activity. Analysis of photolabeled SF, showed that the three major tryptic peptides (23, 50, and 20 kDa) of the heavy chain fragment and the alkali light chains were labeled. The presence of ATP during irradiation protected only the 50-kDa peptide, indicating that the other peptides were nonspecifically labeled. If BzzATP was first trapped on SF1 by cross-linking the reactive thiols, SH1 and SH2, with p-phenylenedimaleimide, only the 50-kDa tryptic peptide was labeled. These results confirm and extend previous observations that [‘H]BzZATP trapped on SFI by cobalt(I1I) phenanthro- line photolabeled the same 50-kDa peptide (Mahmood, R., and Yount, R. G. (1984) J. Biol. Chem. 259,12956- 12959). Thus, the 50-kDa peptide is labeled with or without thiol cross-linking, indicating that the relative position of SH1 and SH2 does not affect the labeling pattern.

It is widely believed that energy transduction in muscle involves the cyclic interaction between actin and myosin in which the thick and thin filaments slide past each other (Huxley, 1969). The energy for muscle contraction is supplied by ATP which is hydrolyzed by myosin. Whereas significant progress has been made toward defining the kinetic steps and rate constants of ATP hydrolysis by myosin (Taylor, 1979; Adelstein and Eisenberg, 1980; Hibberd and Trentham, 1986), there has been less progress made toward elucidating the structure and composition of the ATPase site in the myosin molecule. Structural and functional studies of the active site

* This work was supported by Grant DK-05195 from the National Institutes of Health, the Muscular Dystrophy Association, and post- doctoral fellowships (to C. C.) from the Muscular Dystrophy Associ- ation and American Heart Association of Washington. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

8 Present address: Dept. of Microbiology, University of Pittsburgh, Pittsburgh, PA 15261.

8 To whom correspondence should be addressed.

of myosin will be essential to delineate the molecular mecha- nism of force generation in muscle.

In the absence of any crystallographic information, a vari- ety of methods have been used to map the topography and substructure of the myosin molecule. A major approach has been to use limited proteolysis and to assign different reactive functionalities on the myosin heavy chain to various proteo- lytic fragments. Most of this work has been done using SF1,’ the head region of the myosin molecule that contains the ATPase and actin-binding sites. The 95-kDa heavy chain of SF1 is cleaved into 23-, 50-, and 20-kDa peptides by brief trypsin digestion (Balint et al., 1978; Mornet et al., 1979; Yamamoto and Sekine, 1979). A variety of other proteolytic enzymes have been found to produce essentially the same three fragments of the SFl heavy chain as trypsin (Applegate and Reisler, 1983; Mornet et al., 1984). The N-terminal 23- kDa peptide contains the reactive lysine residue (Hozumi and Muhlrad, 1981), whereas an activity-critical carboxyl group (Korner et al., 1983) and part of the actin-binding site (Mornet et al., 1981a, 1981b; Sutoh, 1983) have been located on the 50-kDa peptide. The two reactive thiols of myosin termed SH1 and SH2 are located on the 20-kDa peptide (Balint et al., 1978). These SH groups can be cross-linked with bifunc- tional thiol cross-linking reagents in the presence of magne- sium-nucleotide (Burke and Reisler, 1977), a reaction which abolishes the ATPase activity and leads to stable noncovalent trapping of magnesium-nucleotide on the enzyme (Wells and Yount, 1979, 1982).

This trapping reaction has proven essential to obtain spe- cific photolabeling of the active site by a variety of ATP photoaffinity analogs (Nakamaye et al., 1985; Okamoto and Yount, 1985; Grammer et al., 1985; Cremo and Yount, 1987). What is not clear is whether the thiol cross-linking has an effect on the site of labeling. This is of interest because BaATP, the photoprobe which is the subject of this paper, has previously been shown to label the 50-kDa tryptic peptide when trapped by Cophen cross-linking of SH1 and SH2 (Mahmood and Yount, 1984). This result was surprising be-

l The abbreviations used are: SF,, skeletal myosin subfragment; Cophen, Co(II1) phenanthroline which is bound to SH, and SH,; HPLC, high performance liquid chromatography; Bicine, N,N- bis(2-hydroxyethy1)glycine; TEAB, triethylammonium bicarbonate; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electropho- resis; NANDP, 2-[(4-azido-2-nitrophenyl)amino]ethyl diphosphate; pPDM, p-phenylenedimaleimide; Bz, acid, 4-benzoylbenzoic acid Bz,rATP, (3’)2’-0-(4-benzoyl)benzoyl-l,P-ethenoadenosine tri- phosphate; 3’-Bz~(2‘-dtADP), 2’-deoxy-3’-0-(4-benzoyl)benzoyl- 1,N6-ethenoadenosine diphosphate. Bz,ATP, 3’(2’)-0-(4-ben- zoy1)benzoyladenosine 5”triphosphate. Williams and Coleman (1982) and subsequent workers have used BzATP as an abbreviation for 3’(2’)-0-(4-benzoyl)benzoyl-ATP. Since Bz is the recommended ab- breviation for a single benzoyl group, we feel that BzzATP (pro- nounced “buzz-2 ATP”) is a more appropriate abbreviation.

14479

14480 Interaction of BzATP with Subfragment 1

cause a structurally analogous photoprobe containing a @- alanylarylazide esterified to the ribose hydroxyl(s) of ATP specifically labels the 23-kDa peptide (Szilagyi et al., 1979). It was of interest then to verify our previous finding by using Bz2ATP as the photoprobe without trapping. As a corollary to this study, we also assessed the effect of using a longer thiol cross-linking agent, pPDM (12-13 A), on the photola- beling kinetics and labeling pattern. This paper shows that the same 50-kDa peptide was labeled and at the same rate if Bz2ATP was bound reversibly or if it was trapped by Coph$n, a short (3-5 A) cross-linker, or by pPDM, a long (12-13 A), cross-linker. Trapping, however, was shown to be essential to obtain labeling of only the 50-kDa peptide.

EXPERIMENTAL PROCEDURES

Materials-TEAB solutions (2 M) were prepared immediately be- fore use by bubbling COz though solutions of redistilled triethylamine (J. T. Baker Chemical Co.) kept in an ice bath until the pH dropped to 7.6. pPDM and Bzp acid were from Aldrich, carbonyl-diimidazole and ammonium formate were from Sigma, and [32P]ATP was from Du Pont-New England Nuclear. All other compounds were obtained from sources mentioned earlier (Mahmood and Yount, 1984).

BzJTP Synthesis-Bz,ATP was synthesized by modifying the method of Williams and Coleman (1982). Bz, acid (5 mmol) and 1,l'- carbonyldiimidazole (10 mmol) were dissolved in 12 ml of dry N,N- dimethylformamide and stirred at room temperature. After 20 min, the sodium salt of ATP (1 mmol) in 15 ml of water was added to the reaction mixture. A thick white precipitate was formed which dis- solved after stirring overnight. The solvents were then removed by rotary evaporation, and the residue was dissolved in 8-10 ml of water. The pH was adjusted to 8.0 with NaOH, and the solution was loaded on a column (2.5 X 80 cm) of DEAE-Sephadex A-25 (HCO; form) equilibrated in water. Three major peaks of 260 nm absorbance were obtained upon elution with a linear gradient of 0-1 M TEAB (Fig. 1). The first and second peaks were identified as the salt of Bz, acid and ATP, respectively, by spectral and TLC analyses. The thlrd peak corresponding to BzZATP was pooled, diluted 8-fold with cold water, and lyophilized. Without dilution, the TEAB solution melts on the lyophilizer, CO, bubbles off, and the resulting alkaline triethylamine solution promotes the partial hydrolysis of the ester linkage between ATP and Bz, acid. The freeze-dried powder was lyophilized again with 150 ml of water to remove residual triethylamine. The resulting white powder was dissolved in 5-10 ml of water and filtered through a 0.2-pm filter. Solutions stored at -80 "C for 6 months at pH 7 showed no hydrolysis to ATP and Bz, acid by TLC analysis. This method in five separate syntheses gave a 23-27% yield of BzZATP.

[3H]Bz2ATP and [y-32P]BzzATP were synthesized as described above on a 1:10 scale utilizing [2,8-3H]ATP and [-Y-~'P]ATP, respec- tively. The products were applied on a 2.5 X 35-cm DEAE-Sephadex column and eluted with a linear gradient of 600 ml each of HZ0 and 1 M TEAB, pH 7.6. The elution profiles were essentially the same as shown in Fig. 1.

The qualitative and quantitative purity of Bz2ATP was checked by TLC and HPLC, respectively. TLC was done on cellulose plates (Analtech, Inc.), using solvent A (I-butano1:acetic acid:H,O (5:1:3)) and on silica gel plates (60 F254, EM Reagents) using solvent B (isobutyric acidNH40HH20 (75:1:24)) as the developing solvent. The RF values in solvents A and B were: ATP, 0.05 and 0.07; Bz:, acid, 0.95 and 0.95; and BzzATP, 0.30 and 0.50, respectively. With BzzATP, these RF values are averages of two poorly resolved spots corresponding to the 2'(3')-isomers.

HPLC was done on a microprocessor-controlled Altex/Beckman dual-pump set-up connected to Beckman Model 165 dual-wavelength detector and a Spherisorb 5-ODS column (0.46 X 25 cm) using 0.1 M TEAB, pH 6.6 (solvent A) and 0.1 M TEAB in ethanol (solvent B) as the eluting solvents (Mahoney and Yount, 1984).

Enzyme Inactivation by pPDM"SF1 (17 p M ) in 0.1 M KCl, 50 mM Bicine, pH 8.0, at 0 'C was incubated with 1.7 mM M&12, 85 pM ATP or Bz2ATP, and 22 p~ pPDM. After 40 min, excess reagents were removed by precipitating the enzyme twice with 2.5 volumes of saturated (NH4),S04, pH 8.0, containing 20 mM EDTA. The protein pellet obtained after centrifugation was dissolved in KCl/Bicine buffer, pH 8.0, and further purified by gel filtration over Sephadex G-25 columns (PD-10, Pharmacia Biotechnology, Inc.). The fractions containing modified SF, were pooled and assayed for ATPase activity

and trapped magnesium-nucleotide. The preparation of SF1 from rabbit skeletal muscle myosin, ATP-

ase assays, protein determination, trapping of Bz2ATP by Cophen, trypsin digestion of SFl, irradiation procedure, and SDS-PAGE were done as described previously (Mahmood and Yount, 1984).

RESULTS

Synthesis and Spectral Characterization of B z A TP-The modified synthesis of BzzATP described here routinely gave 25% yields with 295% purity as determined by HPLC (Fig. 1, inset). Preparations generally contained less than 1% ATP as a breakdown product. The synthetic method of Williams and Coleman (1982) in our hands gave BzzATP of comparable purity, but with about 12% yields. The absorbance spectra of ATP and HPLC-purified BzzATP are shown in Fig. 2. Bz2ATP shows significant absorbance above 290 nm corre- sponding to the n * n* absorption transition of the carbonyl group of benzophenone (Turro, 1978). The absorbance maxi- mum at 261 nm is consistent with a combined spectrum of ATP and the anion of Bzp acid. The molar extinction coeffi- cient of BzzATP was determined in quadruplicate on two separate preparations from the absorbance at 261 nm and the acid-labile phosphate content. Separate solutions of BzzATP and ATP were heated with 1 N HC1 in a boiling water bath for 15 min, and the amount of inorganic phosphate liberated was determined by the procedure of Rockstein and Herron (1951). This method gave an 6%' of 41,000 f 800 M" cm" for BzzATP and an 6yb9 = 15,100 f 200 M-' cm" for ATP.' This latter value for ATP agrees well with the accepted value of 15,400 M" cm". The extinction coefficient for the anion of Bzz acid, pH 8.0, was gravimetrically determined to be 20,000 f 500 M-' em" at 264 nm. The synthesis of Bz2ATP was also done using ATP and ['4C]Bz2 acid (Nakamaye and Yount, 1985) with specific activity of 5,500 cpm/nmol. The specific activity of ['4C]Bz2ATP, calculated assuming = 41,000, was found to be 5,500 cpm/nmol. Thus, two independent analytical techniques gave essentially identical extinction coefficients.

'H NMR (D,O, pH 6.0,2 mM) Bz2ATP (triethylamine salt): S 8.45 (s, 0.6H, C , 3'-isomer), 8.37 (s, 0.4H, Ca, 2'-isomer), 8.15-7.3 (m, 10H, Cp+ phenyl rings), 6.3 (d, 0.42H, C1,, 2'- isomer), 6.2 (d, 0.58H, C1., 3'-isomer), 5.65 (m, IH, C2,, 2'- isomer + C3,, 3'-isomer), 5.0 (dd, 0.63H, C2., 3"isomer) (C,,, 2"isomer and C4., 3"isomer were under the H20 peak), 4.35 (broad, 0.6H, C4,, 2'-isomer), 4.20 (broad, 2H, C,,). 'H NMR spectra were measured on a Nicolet 200-MHz spectrometer at 19 "C. This 'H NMR spectrum indicates that Bz'ATP is a mixture of 3'- and 2'-monosubstituted isomers in a ratio of 60:40 (Fig. 2). The total peak integrations attributed to the benzophenone protons and the adenosine protons were con- sistent with one benzophenone group/adenosine and rule out the presence of a possible 2',3'-disubstituted compound. The relative peak areas of the resonance signals H-8, H-1', H-2', and H-3' were used to determine the percentage of 2'- or 3'-

Several workers have reported the extinction coefficient of BzZATP without detailing their analytical methods or reporting the absorption spectrum: Williams et al. (1986), 31,900 M-l cm"; Bar-Zvi and Shavit (1984), 32,000 M" cm"; Kambouris and Hammes (19851, 32,500 M" cm"; Cable and Briggs (1984), 15,400 M" cm" (later corrected in the erratum to 40,700 M" cm"). The sum of the extinc- tion coefficient of ATP (15,400 M-' cm" at 259 nm) and the anion of Bzz acid (20,000 M" cm" at 264 nm) is about 32,000 M" cm", about 22% less than our result. We have also found that 3'(2')-0-(4- benzoy1)benzoyl-l,N6-ethenoadenosine diphosphate and 2"deoxy-3'- O-(4-benzoyl)benzoyl-l,~-ethenoadenosine diphosphate (Cremo and Yount, 1987) have extinction coefficients greater than the sum of the coefficients of the reactants.

Interaction of BzATP with Subfragment 1 14481

180 -

140 -

120 - 0 (D N a loo -

80 -

280 nm 2.0 A U F

&*ATP

4 m

0 20 40 60 80

80

m 40 * 20

Fraction Number FIG. 1. Purification of BzzATP on DEAE-Sephadex. The

column (2.5 X 80 cm) was eluted at 4 “C with a linear gradient of 2.5 liters of Hz0 and 2.5 liters of 1 M TEAB, pH 7.6, a t a flow rate of 4.2 ml/min. Fractions were collected for 5 min each and assayed for absorbance at 260 nm. BzZATP was analyzed by reversed-phase HPLC as described under “Experimental Procedures.” The inset shows the HPLC profile of DEAE-purified BzZATP. A UF, absorbance units at full scale; B, solvent B.

0 220 240 280 280 300

Wavelength (nm) FIG. 2. Ultraviolet absorption spectra of BzzATP and ATP.

Bz,ATP from HPLC was pooled, diluted 5-fold with water, lyophilized twice to remove TEAB, and dissolved in water (18 p ~ , pH 6.6). The spectrum of ATP (18 p ~ , pH 6.6) is shown for comparison purposes. All spectra were recorded on a Cary-14 spectrophotometer.

isomers. Proton assignments were verified by homodecoupling experiments.

‘H NMR has been used previously to measure rates of hydrolysis and acyl migration of ribonucleoside derivatives (Griffin et al., 1966). McLaughlin and Ingram (1965a, 1965b)

found a 3‘:2’-isomer ratio of 65:35 for aminoacyladenosine using a chemical method of analysis and a 75:25 ratio for N - acetylvalyladenosine using a chromatographic analysis. Sim- ilar 3’:2’-isomer ratios have been reported for a variety of ribose-esterified nucleotide analogs (Schafer et al., 1986; Onur et al., 1983). In contrast, Williams and Coleman (1982) and Kambouris and Hammes (1985) reported that their prepara- tions of BzzATP contained only the 3’-isomer based upon ‘H NMR results. Our ‘H NMR analysis of BzzATP purified by ammonium formate elution from LH-20 (Williams and Cole- man, 1982) also indicated a 3’:2’-isomer ratio of 6040. These ’H NMR studies also revealed substantial contamination by aromatic compounds which originated from the ammonium formate. Abeijon et al. (1986) concluded that their preparation of Bz2CTP was only the 3’-isomer, based upon ’H NMR results. A reason for these discrepancies may be the pH dependence of the ‘H NMR spectrum for 0-acyl-nucleotides. At pH 7 and higher, we have found considerable signal broad- ening that prevented definitive assignments of the 2‘- and 3’- isomer signals. In addition, synthetic nucleotide preparations are prone to contamination with paramagnetic metal ions, which dramatically broaden line widths (Ts’o et al., 1969). For this reason, we routinely add EDTA and adjust the pH of nucleotide samples to 6.0 or below in preparation for ’H NMR measurements.

Binding of BzADP to SF,-The binding of Bz2ADP to SF, in the presence of Mg2+ in the dark was investigated by use of equilibrium dialysis. Scatchard analysis (Fig. 3) shows that BzzADP has a single binding site on SF, ( n = 1.06) with an apparent equilibrium association constant of 3.0 X 10‘ M-’. This value, which is about 2-30 times less than values reported for MgADP (see Watterson et al., 1983), indicates that the benzophenone grouping does not prevent relatively tight bind- ing of BzzADP to the active site.

B z A T P as a Substrate for the ATPase Activity of SF1- BzzATP serves as a substrate for the ATPase activity of SF, in the presence of different mono- and divalent cations, as

0.25 1 9) 0.20 F c t 0‘307

1 0.25 - KD=3.45 pM

n=1.08

9) 0.20 - F

57 c

2 0.15 - N N

L 0.10 - .

0.05 -

0.2 0.4 0.8 0.8 1.0

r

0.05 i r

FIG. 3. Binding of [’HIBzzADP to SF,.Equilibrium dialysis was performed as described by Wells and Yount (1979). Split cells (2.4 ml total) contained 20 mM MgCIZ, 0.1 M KCI, 50 mM Bicine, pH 8.0, and varying amounts of [3H]BzZATP on both sides and 17 p~ SF1 on one side only. The cells were agitated gently at 4 “C for 26 h. Control experiments show that all the [3H]BzzATP was hydrolyzed to [3H] BzzADP. The solutions on both sides were analyzed for radioactivity to determine r , the moles of [3H]Bz2ADP bound/mol of SF,. K D is the apparent equilibrium dissociation constant. n is the intercept on the horizontal axis.

14482 Interaction of BzzATP with Subfragment I

TABLE I Hydrolysis of ATP and BzATP by SF, at 25 "C

Assay system Bz.-ATPase ATPase Bs-ATPase:ATPase Activity ratios:

pmol P,/min/mg protein K+-EDTA" 2.2 2.6 0.85 N&+-EDTAb 10.4 11.2 0.93 M g C 0.025 0.025 1.0 F-actin + M P 0.5 1.35 0.37 "0.43 p M SF1, 0.6 M KCl, 50 mM Tris, 5 mM EDTA, 1.9 mM

0.17 p M SF], 0.56 M NH4C1, 0.23 M KCl, 36 mM EDTA, 35 mM

e 12.8 p M SF,, 0.1 M KCl, 40 mM Tris, 2 mM MgC12, 1 mM

0.77 p M SFI, 4 mM actin, 10 mM Tris, 6 mM KC1, 3 mM MgC12, 3

nucleotide, pH 7.5.

Tris, 2.3 p M nucleotide, pH 8.0.

nucleotide, pH 8.0.

mM nucleotide, pH 8.0.

C. Mahoney, unpublished results.

shown in Table I. The rate of cleavage of Bz2ATP in the presence of K+-EDTA, NHZ-EDTA, and MgZ+ was essentially identical to that of ATP. The addition of F-actin stimulated the hydrolysis of Bz2ATP to about 40% of the extent of ATP, indicating that the presence of the bulky benzophenone group did not interfere significantly with actin binding. The stimu- lation of Bz2ATP hydrolysis by F-actin suggested that BzzATP might be effective as a substrate for the support of contraction of muscle fibers. A 2 mM solution was 87% as effective as ATP in inducing tension development in glycer- inated muscle fibers?

Photolobeling of SFl by BzATP-Irradiation of SFl with UV light (>300 nm) in the presence of BzzATP caused a decrease in the K+-EDTA-ATPase activity (Fig. 4, upper). The rates of inactivation increased as BzzATP concentrations were increased beyond saturation of SFl, indicating that non- specific photoinactivation was occurring. The amount of non- specific inactivation was assessed by doing an inactivation in the presence and absence of a 100-fold molar excess of ATP over SF,. The extent of inactivation with a 5-fold molar excess of BzzATP over SFI (solid squares) decreased from 75 to 35% in 30 min when excess ATP (open squares) was added. This 40% difference is the specific ATPase inactivation, which should agree with the measured covalent incorporation for 13H]Bz2ATP into SF, under similar reaction conditions. In- deed, as shown in Fig. 4 (lower), the extent of specific covalent labeling (solid squares (no ATP), open squares (+ATP)) was about 0.35 mol of Bz2ATP/mol of SF, at 30 min. In addition, at a 10-fold excess of BzzATP over SF, (no ATP), the extent of photoincorporation was 1.4 mol of Bz,ATP/mol of SFl in 30 min, indicating that nonspecific labeling occurred readily on sites which are not necessarily activity-related. These studies emphasize the importance of using low ratios of Bz2ATP to enzyme active sites if specific photolabeling is to be attained. Other photoinactivation experiments with a 5- fold excess of the salt of Bz2 acid over enzyme gave a 20% loss of activity in 30 min with no protection by a 100-fold excess ATP over SF, (data not shown). These results indicate that Bz2 acid alone is capable of significant nonspecific label- ing, as has been shown recently by Tran and Farley (1986) for several membrane ATPases.

SDS-PAGE of SFl Photolabeled with PHlBzATP-It was of interest to determine the site(s) of photoincorporation of Bz2ATP into SF, under conditions where no thiol cross- linkers were used. It was possible that our earlier results (Mahmood and Yount, 1984) which showed labeling of only the 50-kDa tryptic peptide of the heavy chain may have been influenced by the Cophen complexation of SH1 and SH2.

I 0 5 10 15 20 25 30

Photolysis Time (mid

Photolysis Time (mid FIG. 4. Upper, photoinactivation of K+-EDTA-ATPase activity of

SF1 by Bz,ATP. SF1 (17 p ~ ) in 0.1 M KC1, 50 mM Bicine, pH 8.0, was incubated with different concentrations of magnesium-nucleotide and irradiated with W light for the indicated times. ATPase activi- ties are relative to unirradiated controls. 0, 0.17 mM MgCl,, 17 p~

p~ BzATP; W and 0, 4.25 mM MgC12, 85 p~ BQATP in absence and presence of 1.7 mM ATP, respectively. Lower, time course of photoincorporation of ['H]Bz,ATP into SFI. At 5-min intervals, samples of irradiated SF, (80 pl) were precipitated with 0.5 ml of ice- coId 5% trichloroacetic acid and centrifuged for 3 min in a Beckman microfuge. The protein pellet was washed with 0.7 ml of cold 5% trichloroacetic acid, recentrifuged, dissolved in 0.4 ml of Protoeol (Du Pont-New England Nuclear), and counted. For each time course, a nonspecific binding control aliquot was removed before irradiation. This value (<0.05 mol of BzZADP/mol of SF,) was then subtracted from each time point to give the photoincorporation values indicated. Conditions and symbols were the same as above described for A except that ['H]Bz2ATP (26,000 cpm/nmol) was used instead of cold

B-ATP; A, 0.17 mM MgCI2, 34 ptM BQATP; X, 1.7 mM MgC12, 170

BQATP.

Accordingly, SF, was irradiated for 30 min in the presence of a 5-fold molar excess of [3H]Bz2ATP as described for Fig. 4. The photolabeled enzyme was electrophoresed on polyacryl- amide gels to separate the 95-kDa heavy chain fragment and the alkali light chains, A1 and A2. The majority of the radioactivity was associated with the 95-kDa heavy chain fragment (Fig. 5A), with about 10% of the counts in A1 and A2. The presence of a 100-fold molar excess of ATP over SF1 during irradiation reduced the radioactivity incorporated into the heavy chain by about 50%. This result was in general agreement with the 40% reduction in photoincorporation measured under identical conditions in Fig. 4 (lower). In contrast, the total radioactivity incorporated into either of

Interaction of BzATP with Subfragment 1 14483

I

w [SH]BzlATP/SFl=5

0 5 10 15 20 25 30 35 40 45

Gel Slice No.

B ’ ‘50K\ 2 0 K l

9 1 1 1 ”

H ATPIBzlATP=20 w Bz2ATPISF1=5 -

3

n -

o 7

x 2 - Y P

Gel Slice No. FIG. 5. Radioactivity profile of [‘H]Bz2ATP photolabeled

SF, before ( A ) and after ( B ) trypsin digestion. SF, (17 p ~ ) in 0.1 M KC1, 50 mM Bicine, 4.25 mM MgCl,, pH 8.0, was irradiated for 30 min with 85 p~ [3H]Bz2ATP (26,000 cpm/nmol) alone (0) and in presence of 1.7 mM ATP (B). 100 pg of intact SF1 ( A ) and 150 pg of trypsin-digested SF, ( B ) (1100 w/w, 25 “C for 30 min) were analyzed by SDS-PAGE on 12% polyacrylamide gels (Laemmli, 1970). The gels were stained with 0.01% Coomassie Blue, cut into 1.5-mm pieces, solubilized, and assayed for radioactivity as described earlier (Mah- mood and Yount, 1984). Note that the 23-kDa peptide runs as if it were a 25-kDa peptide (see this figure and Fig. 8E). A , and Az refer to the two alkali light chains.

the light chains was not affected by excess ATP, indicating that under these conditions, [3H]Bz2ATP labels these peptides nonspecifically.

To determine the site(s) of photolabeling in the heavy chain, trypsin was added to cleave the chain into the charac- teristic 23-, 50-, and 20-kDa fragments (Balint et al., 1978). The radioactivity profile of an electrophoretic gel separation showed that the 50- and 20-kDa peptides were the most extensively labeled and contained 40 and 25% of the total radioactivity, respectively (Fig. 5B). Small amounts of label were found in the 75-kDa peptide (the precursor of 50- and 23-kDa peptides), in the 23-kDa peptide, and in A1 and A2. A small tryptic peptide of unknown origin, which migrated near the front of the gel, also contained about 8% of the radioactivity. The presence of excess ATP during irradiation reduced the combined radioactivity incorporated into the 50- kDa peptide and its 75-kDa precursor by about 80%. All other peptides were labeled to the same extent in the presence or

absence of ATP, indicating that their labeling was nonspe- cific. The same pattern of protection by ATP was obtained when these experiments were repeated using stoichiometric amounts of [3H]Bz2ATP and SF, (data not shown).

Inactivation of SFl and Trapping of BzATP-The rate of inactivation of SF, ATPase by cross-linking agents is greatly enhanced by the presence of magnesium-nucleotides (Burke and Reisler, 1977; Wells et al., 1980). Both ATP and BzzATP stimulated the inactivation of NH:-EDTA- and Ca*+-ATPase activities at similar rates in the presence of pPDM (data not shown). The patterns of inactivation were similar to those seen earlier with ATP and pPDM (Burke and Reisler, 1977; Wells and Yount, 1979) and with BzzATP and Cophen (Mah- mood and Yount, 1984). The pPDM-modified enzyme con- tained 0.7 & 0.03 mol of magnesium-nucleotide (ATP or Bz,ATP) trapped per mol of SF, and had less than 15% of the K+-EDTA- and Ca2+-ATPase activities. Loss of ATPase activity and trapping of BzzATP by pPDM or Cophen were linearly correlated (data not shown) in a similar manner to that shown for ADP (Wells and Yount, 1979). Stability stud- ies on the trapped [3H]Bz2ATP showed that at 0 “C, greater than 60% of the analog was still retained on SF, after 7 days (data not shown). No trapping of nucleotide or loss of ATPase activity was observed in the absence of the cross-linker. Furthermore, when enzyme inactivations were performed in the presence of [y-32P]Bz2ATP, no trapping of 32P occurred, indicating that BzzATP is hydrolyzed and trapped as Bz2ADP. This is consistent with previous experiments which showed that ATP is hydrolyzed and trapped as ADP (Wells and Yount, 1979). For simplicity, we refer to “trapping Bz2ATP,” although in each case, it is the diphosphate form which is trapped.

Evidenct. That BzATP Is Trapped at the Active Site-To show that BzzATP was trapped at the active site of SF1, competition experiments were performed in which t3H] BzzATP was trapped in the presence of differing concentra- tions of ATP. The data in Fig. 6 show that increasing concen- trations of ATP caused a linear decrease in pPDM-promoted trapping of [3H]Bz*ATP. Thus, both ATP and BzzATP appear to be trapped with the same efficiency, illustrating that they bind to SF, with similar affinities and stoichiometries.

Photoincorporation of pPDM-trapped PHjBzATP into SFl-SFI was modified bypPDM in presence of [3H]Bz2ATP,

CoPhen 0 pPDM

.3

.2

. 1

’

.

.o 0.2 0.4 0.6 0.8 1.0

ATP

ATP + Br2ATP

FIG. 6. Competition of ATP and [‘H]Bz2ATP for trapping on SF,. SF1 (17 phi) was modified by 22 p~ pPDM or Cophen in the presence of 85 pM [3H]Bz2ATP (3400 cpm/nmol) and differing con- centrations of ATP to give the [ATP]/[ATP] + [BzZATP] ratio indicated. After 40 min of inactivation, the enzyme was purified by ammonium sulfate precipitation as described under “Experimental Procedures.” The amount of trapped [3H]Bz2ATP was determined by liquid scintillation counting.

14484 Interaction of B z d TP with Subfragment 1

0.4 L

PPDM . Cophen

0.3 - L-

N

s 0.2

0- 0.1 -

Time (rnin) 10 20 30

5 10 15 20 25 30 Photolysis Time (mid

FIG. 7. Kinetics of photoincorporation of trapped [sH] BzzATP into SF1. SF1 (17 p ~ ) was modified by 22 prd pPDM or Cophen in the presence of 85 p~ [3H]Bz2ATP (3400 cpm/nmol) and purified as described under “Experimental Procedures.” Samples of pPDM- and Cophen-modified SF1 containing 0.7 and 0.8 mol of trapped [3H]Bz2ATP/mol of SF,, respectively, were irradiated under UV light for varying intervals. Covalent photoincorporation of I3H] BzZATP into SF1 was determined from the radioactivity associated with 5% trichloroacetic acid precipitates (see Fig. 4 legend). For the inset, data were plotted using the equation In a/(a - x ) = klt where a = moles of trapped [3H]BzzATP/mol of SF1 and x: = moles of [3H] BzzATP photoincorporated per mol of SF1 at time t; kl = first-order rate constant.

and excess nucleotide and modifying reagents were removed as described under “Experimental Procedures,” The enzyme complex, which contained 0.7 mol of [3H]Bz2ATP/mol of SF1, was irradiated for 30 min, after which 0.35-0.40 mol of [3H] Bz2ATP/mol of SF, or -50% of the trapped analog was covalently incorporated into SF1 (Fig. 7). This level of pho- toincorporation was identical to the specific photoincorpora- tion measured after 30 min of irradiation at a 5-fold excess [3H]Bz2ATP over SFl (Fig. 4, lower). The photoincorporation obeyed first-order kinetics (kl = 3.45 x lo-’ rnin”) for the first 20 min before leveling off (Fig. 7, inset). The leveling off was not due to leakage of trapped [3H]Bz2ATP from the active site as control experiments showed that less than 2% disso- ciated in 30 min under the irradiation conditions. The absence of leakage also means that all the photoincorporation occurred at the active site. The kl for photoincorporation of [3H] BzzATP trapped by Cophen was 2.9 X lo-‘ min”. The similar rates of photoincorporation which occurred with both cross- linkers indicate that the quite different thiol-thiol span did not affect the photolabeling reaction.

SDS-PAGE Analysis of pPDM-modified and Irradiated Bz,ATP-SF,-The pPDM-modified SFl containing 0.7 mol of [3H]Bz2ATP trapped per mol of SF, was irradiated for 30 min and then analyzed by SDS-PAGE. Fig. 8A shows that nearly all of the radioactivity co-migrated with the 95-kDa heavy chain. No radioactivity above background level was observed in either of the alkali light chains. Similar analysis of the same protein after limited trypsin digestion (Fig. 8B) showed that ~ 7 5 % of the radioactivity co-migrated with the 50-kDa peptide and its 75-kDa precursor. No labeling of the 23-kDa tryptic peptide was seen. The alkali light chains and the 20-kDa fragment have -5% of the radioactivity. The apparent labeling of A1 and A2 must be from co-migrating labeled tryptic peptides from the heavy chain since they were not labeled prior to trypsin digestion (Fig. 8A). SF1 was not labeled by trapped [3H]Bz2ATP in the dark (Fig. 8, open

3

A

2 m I z t? X

E

‘ 1

‘ 0 5 10 15 20 25 30 35 40 4 5 58

Gel Slice No.

Gel Slice No.

FIG. 8. SDS-PAGE analyses of pPDM-modified SF1 con- taining trapped and photoincorporated [‘H]Bz2ATP. SF1 was cross-linked with pPDM in the presence of [3H]Bz2ATP and purified as described under “Experimental Procedures.” The enzyme contain- ing 0.7 mol of [3H]BzzATP/mol SF, was irradiated for 30 min and analyzed by SDS-PAGE. 150 pg of intact SF, ( A ) and 200 pg of trypsin-digested SF, ( B ) (1100, w/w, 30 min, 25 “C) were applied on 7-14 and 16% gels (Laemmli, 1970), respectively. The gels were analyzed for radioactivity (8) as described for Fig. 6. Radioactivity in gel slices containing pPDM-modified SFl which had not been irradiated with UV light is shown (0).

DISCUSSION

Bz,ATP proved to be effective as a substrate of SF1 in the presence of different cations (Table I). Both ADP and BzzADP bind to SF, with similar affinities. Thus, the esteri- fication of the bulky carboxybenzophenone group to the 3‘(2’)-ribose hydroxyls of ATP does not significantly alter the binding and hydrolysis of Bz2ATP by SF1. Significantly, this modification also does not prevent actin activation of Mg-ATPase (MgBz2-ATPase) activity (Table I). In related experiments, Bz2ATP was found to serve as an effective substrate to support tension development in skinned muscle fibers.3 These findings are in accord with previous observa- tions in which the myosin ATPase activity was found to be quite tolerant of modifications to the ribose ring (Jeng and Guillory, 1975; Hiratsuka, 1982, 1984).

Previous work with BzzATP had shown that irradiation of SF1 containing Bz2ATP trapped by Cophen resulted in spe-

circles). cific labeling of only the 50-kDa tryptic peptide (Mahmood

Interaction of Bz2ATP with Subfragment 1 14485

and Yount, 1984). The N-terminal 23-kDa tryptic peptide, shown to be photolabeled by the ribose-modified ATP analog arylazido-8-alanyl-ATP (Szilagyi et al., 1979), was not labeled (<I%) by Bz2ATP using the Cophen-trapping conditions. This difference in peptide labeling could be due to (i) different photochemical reactivities of the ketone and azide groups, (ii) different possible conformers of the ribose-attacbed moieties while bound to the active site, or (iii) the 6-7-4 distance of the benzophenone carbonyl group versus the 10-A distance of the arylazide from the 2’- or 3‘-oxygen of the ribose ring. Perhaps more importantly, Szilagyi et al. (1979) did not stabilize their photoaffinity label at the active site by cross- linking SH1 or SH2. Thus, it was possible that the Bz2ATP labeling results, where SH1 and SH2 were bridged by Cophen, reflected a cross-link-dependent conformational change in the protein that was sensed by the photolabel at the active site.

To test this possibility and to confirm our earlier results, photoaffinity labeling experiments with Bz2ATP were re- peated without cross-linking SH1 and SH2. In addjtion, the effect of using: longer cross-linker, pPDM (12-13 A), versus Cophen (3-5 A) on the time course and pattern of photola- beling was also assessed. Irradiation of SF, in the presence of Bz2ATP, without cross-linking SH1 and SH2, inactivated the ATPase activity, where the level of inactivation depended upon the Bz2ATP concentration (Fig. 4, upper). This inacti- vation and concurrent total covalent photoincorporation (Fig. 4, lower) were partially protected by excess ATP, indicating that both specific and nonspecific labeling were occurring under these conditions. This result was confirmed by SDS- PAGE analyses of trypsinized samples of [3H]Bz2ATP-SF1 (Fig. 5 B ) , which showed that ATP protected against labeling of only the 50-kDa fragment and its 75-kDa precursor. Non- specific labeling occurred on the light chains, the 20- and 23- kDa peptides, and, to a lesser extent, the 50-kDa peptide.

Bz2ATP was shown to be stably trapped at the active site on SF, by cross-linking SH1 and SH2 with pPDM. ATP and Bz2ATP were trapped by pPDM with a similar efficiency at the active site (Fig. 6). Excess nucleotide and pPDM were removed from the stable enzyme-nucleotide complex before repeating the same photoincorporation experiments described earlier for untrapped nucleotide. Photoincorporation was sat- urable (Fig. 7) with the same time course as the specific labeling observed in the absence of trapping (Fig. 4, lower). The final levels of specific photoincorporation for the two experiments (Figs. 4, lower, and 7) were nearly identical. SDS- PAGE of the irradiated enzyme showed incorporation of radioactivity only into the 95-kDa heavy chain, with no light chain labeling. After limited trypsin cleavage, essentially all of the label was found in the 50-kDa peptide and its 75-kDa precursor. Thus, the trapping procedure, which allows for removal of all unbound nucleotide, eliminated virtually all the nonspecific labeling observed in experiments without trap- ping. Nevertheless, the labeling patterns for either method, with or without trapping, are identical, Le. the specific labeling with Bz2ATP modifies the 50-kDa peptide and occurs with the same time course of photoincorporation. The labeling of the 50-kDa was the same using no cross-linker, a 3-5-A cross- linker (Cophen), or a 12-13 A cross-linker (pPDM). Thus, the photoaffinity labeling results do not appear to be sensitive to the relative positions of SH1 and SH2 within the enzyme.

The lack of stoichiometric covalent incorporation of [3H] BzzATP into SF, (Fig. 6) was surprising since the triplet state of the benzophenone carbonyl group is known to be inert toward reaction with water (Ledger and Porter, 1972). Thus, in theory, the benzophenone group has multiple chances to react with nearby amino acids and, in this case, should give

nearly stoichiometric incorporation of trapped nucleotide into SF,. Although photoincorporation appeared to be saturable, we consistently found that only 50-60% of the trapped BzzATP reacted covalently with SF,. With or without trap- ping, ~ 3 5 % of the total protein was specifically covalently modified by Bz2ATP. The same result was observed by Cremo and Yount (1987) using the fluorescent analogs of BzZATP, BzztATP and 3’-Bz2(2‘-dtADP). Since the benzophenone tri- plet state is well known to insert into C-H bonds, it was possible that the remaining analog was reacting with the Bicine buffer. However, irradiation of SF, containing trapped [3H]Bz2ATP in 50 mM sodium phosphate buffer, pH 8.0, still resulted in only 50% labeling in 30 min. In addition, the Bz2ATP which did not photoincorporate was spectrally iden- tical t o BZ~ATP.~ This observation ruled out the possibility that the carbonyl of some of the benzophenone moieties was reduced to the alcohol by introduction of a double bond into adjacent amino acid side chains. Such a photocatalyzed oxi- dation-reduction reaction is known to occur with steroids linked to benzophenone moieties (Baldwin et al., 1970; Bres- low et al., 1973). Two other possible reasons for the less than stoichiometric incorporation are (i) the SF1 containing bound analog is heterogeneous, and only one form reacts covalently with Bz,ATP; and (ii) BzzATP contains Bz, acid esterified at the 2’- or 3’-hydroxyls of ATP in a ratio of 40:60 (see “Results”), and only the 3’-isomer photoinserts. This latter explanation requires that both the 2’- and 3”isomers be trapped and that the rate of interconversion of the 2’- to 3’- isomer is slow when Bz2ATP is trapped on the enzyme. If it is rapid, then again stoichiometric labeling would be expected. At present, we have no evidence to distinguish between these two possible explanations.

The 23-kDa tryptic peptide, previously labeled with three other photoaffinity analogs, was not labeled by Bz2ATP. Szilagyi et al. (1979) photolabeled the 23-kDa peptide with an ATP analog containing an arylazido-P-alanyl group esterified at the 3’(2’)-OH of ribose. The 23-kDa peptide has also been labeled by NANDP, a photoaffinity analog of ADP (Naka- maye et al., 1985; Okamoto and Yount, 1985), and by 2-azido- ADP (Grammer et at., 1985): In addition, direct photoaffinity labeling with UTP of myosin I1 from Acantharnoeba also labeled specifically Glu-185 in the analogous 23-kDa peptide (Atkinson et al., 1986). Because these latter studies all used photoprobes which would label the SF, purine-binding site for ATP, it is perhaps not surprising that they incorporate into a different part of the heavy chain.

Of immediate interest to this work is the recent observation that a selective thrombin-catalyzed cut between Lys-560 and Ser-561 in the 50-kDa region dramatically altered both ATP hydrolysis and actin binding (Chaussepied et al., 1986a, 1986b). Subsequently, a 30-kDa fragment representing the peptide segment from Ser-561 to the C terminus of the SF, heavy chain was isolated and renatured. The renatured pep- tide bound tightIy to actin in an ATP-dependent manner (Chaussepied et al., 1986c), whereas the C-terminal 20-kDa fragment-actin complex was not affected by ATP. These results suggested the IO-kDa peptide from the C terminus of the 50-kDa fragment provided the polyphosphate-binding site for ATP. If true, then portions of both the 23- and 50-kDa peptides must be present at the active site, as suggested in this and our earlier work (Mahmood and Yount, 1984) and more recently by Hiratsuka (1986) using fluorescent reporter groups. It should be cautioned, however, that at present, there is no incontrovertible evidence that the 50-kDa peptide di-

‘ R. Mahmood, unpublished results. H. Kuwayama and R. Yount, unpublished results.

14486 Interaction of B z d T P with Subfragment 1

rectly participates in ATP binding or hydrolysis. After these studies were completed, Srivastava et al. (1986)

showed that Bz2ATP photolabeled predominantly the 50-kDa tryptic peptide from the heavy chain of gizzard myosin subfragment 1. However, significant labeling of both the 29- and 25-kDa tryptic peptides occurred and, in this case, as in the studies presented here, it is clear that some type of nucleotide trapping must be used if there is any hope of isolating specific active site peptides photolabeled by Bz2ATP. Such studies are underway for both skeletal and smooth muscle myosins.

Finally, it has become customary to designate ribose ester derivatives of ATP (and other ribonucleotides) as the 3'-ester (see e.g. Schafer et al., 1986). However, it is likely that all such analog preparations are mixtures of the 2'- and 3"esters (Onur et al., 1983), and future work should note the possible complications such isomeric mixtures may introduce. Specif- ically, where 3'(2')-Bz2-nucleotides are used as photolabels, if both the 2'- and 3"isomers bind to an enzyme with similar affinity, the actual photolabeling pattern may be much more heterogeneous than if a single isomeric species was present.

Acknowledgments-We thank Charles Mahoney for his help with HPLC separations and for the determination that BzzATP supported tension development in skinned muscle fibers. We are indebted to Doug Cole for the ATPase assays in Table I and to Don Appel for the 'H NMR spectral determinations.

REFERENCES

Abeijon, C., Capasso, J. M., Tal, D., Vann, W. F., and Hirschberg, C.

Adelstein, R. S., and Eisenberg, E. (1980) Annu. Reu. Biochem. 49,

Applegate, D., and Reisler, E. (1983) Proc. Natl. Acad. Sci. U. S. A.

Atkinson, M. A. L., Robinson, E. A., Appella, E., and Korn, E. D.

Baldwin, J. E. Bhatnagar, A. K., and Harper, R. W. (1970) Chem.

Balint, M., Wolf, I., Tarcsafalvi, A., Gergely, J., and Sriter, F. A.

Bar-Zvi, D., and Shavit, N. (1984) Biochim. Biophys. Acta 765, 340-

Breslow, R., Baldwin, S., Flechtner, T., Kalicky, P., Liu, S., and

Burke, M., and Reisler, E. (1977) Biochemistry 16,5559-5563 Cable, M. B., and Briggs, F. N. (1984) J. BioL Chem. 259, 3612-

Chaussepied, P., Mornet, D., Audemard, E., Derancourt, J., and

Chaussepied, P., Mornet, D., Barman, T. E., Travers, F., and Kassab,

Chaussepied, P., Mornet, D., and Kassab, R. (1986~) Biochemistry

Cremo, C., and Yount, R. G. (1987) Biochemistry, in press Grammer, J. C., Czarnecki, J. J., and Yount, R. G. (1985) Biophys. J.

Griffin, B. E., Jarman, M., Reese, C. B., Sulston, J. E., and Trentham,

Hibberd, M. G., and Trentham, D. R. (1986) Annu. Reu. Biophys.

B. (1986) J. Biol. Chem. 261, 11374-11377

921-956

80, 7109-7112

(1986) J. Biol. Chem. 261, 1844-1848

Commun. 659-661

(1978) Arch. Biochem. Biophys. 190, 793-799

346

Washburn, W. (1973) J. Am. Chem. SOC. 95,3251-3262

3615; Correction (1984) J. Biol. Chem. 259, 14315

Kassab, R. (1986a) Biochemistry 25, 1134-1140

R. (1986b) Biochemistry 25, 1141-1149

25,6426-6432

47, 306 (abstr.)

D. R. (1966) Biochemistry 5, 3638-3649

Biophys. Chem. 15, 119-161

Hiratsuka, T. (1982) Biochim. Biophys. Acta 719, 509-517 Hiratsuka, T. (1984) J. Biochem. (Tokyo) 96, 147-154 Hiratsuka, T. (1986) J. Biol. Chem. 261, 7294-7299 Hozumi, T., and Muhlrad, A. (1981) Biochemistry 20, 2945-2950 Huxley, H. E. (1969) Science 164, 1356-1366 Jeng, S. J., and Guillory, R. J. (1975) J. Supramol. Struct. 3, 448-

Kambouris, N. G., and Hammes, G. G. (1985) Proc. Natl. Acad. Sci.

Korner, M., Thiem, N. V., Cardinaud, R., and Lacombe, G. (1983)

Laemmli, U. K. (1970) Nature 227, 680-685 Ledger, M. B., and Porter, G. (1972) J. Chem. SOC. Faraday Trans.

Mahmood, R., and Yount, R. G. (1984) J. Biol. Chem. 259, 12956-

Mahoney, C. W., and Yount, R. G. (1984) Anal. Biochem. 138, 246-

McLaughlin, C. S., and Ingram, V. M. (1965a) Biochemistry 4, 1442-

McLaughlin, C. S., and Ingram, V. M. (1965b) Biochemistry 4, 1448-

Mornet, D., Pantel, P., Audemard, E., and Kassab, R. (1979) Biochem.

Mornet, D., Bertrand, R., Pantel, P., Audemard, E., and Kassab, R.

Mornet, D., Bertrand, R., Pantel, P., Audemard, E., and Kassab, R.

Mornet, D., Ue, K., and Morales, M. F. (1984) Proc. Natl. Acad. Sci.

Nakamaye, K. L., and Yount, R. G. (1985) J. Labelled Cmpd. Radi-

Nakamaye, K. L., Wells, J. A., Bridenbaugh, R. L., Okamoto, Y., and

Okamoto, Y., and Yount, R. G. (1985) Proc. Natl. Acad. Sci. U. S. A.

Onur, G., Schafer, G., and Strotman, H. (1983) 2. Naturforsch. Teil

Rockstein, M., and Herron, P. W. (1951) Anal. Chem. 23, 1500-1507

468

U. S. A. 82, 1950-1953

Biochemistry 22,5843-5848

68,539-553

12959

251

1447

1456

Biophys. Res. Commun. 89,925-932

(1981a) Biochemistry 20, 2110-2120

(1981b) Nature 292, 301-306

U. S. A. 81, 736-740

ophurm. 22,607-613

Yount, R. G. (1985) Biochemistry 24,5226-5235

82,1575-1579

C Biochem. Biophys. Biol. Virol. 38, 49-59

Schafer, G., Lucken, U., and Lubben, M. (1986) Methods Enzymol. 126,683-712

Srivastava, S., Cable, M. B., and Driska, S. P. (1986) Eur. J. Biochem.

Sutoh, K. (1983) Biochemistry 22, 1579-1585 Szilagyi, L., Balint, M., Sreter, F. A., and Gergely, J. (1979) Biochem.

Taylor, E. (1979) CRC Crit. Reu. Biochem. 6, 103-164 Tran, C. M., and Farley, R. A. (1986) Biochim. Biophys. Acta 860,

Ts'o, P. 0. P., Kondo, N. S., Schweizer, M. P., and Hollis, D. P.

Turro, N. J. (1978) Modern Molecular Photochemistry, Beniamin/

156,447-451

Biophys. Res. Commun. 87, 936-945

9-14

(1969) Biochemistry 8,997-1029 - .

Cummings, Reading, MA Watterson. J. G.. Foletta, D.. Kunz, P. A.. and Schaub. M. C. (1983)

Eur. J. Biocheh. 131,'89-96, and references therein' . .

Wells, J. A., and Yount, R. G. (1979) Proc. Natl. Acad. Sci. U. S. A .

Wells, J. A., and Yount, R. G. (1982) Methods Enzymol. 95, 93-116 Wells, J . A,, Sheldon, M., and Yount, R. G. (1980) J. Biol. Chem.

Williams, N., and Coleman, P. S. (1982) J . Biol. Chem. 257, 2834-

Williams, N., Ackerman, S., and Coleman, P. S. (1986) Methods

Yamamoto, K., and Sekine, T. (1979) J. Biochem. (Tokyo) 86,1855-

76,4966-4970

255, 1598-1602

284 1

Enzymol. 126, 667-682

1862