Eur. J. Biochem. 113, 85-90 (1980) by FEBS 1980

Biphasic Steady-State Kinetics of Myosin Adenosine Triphosphatase Evidence for a Substrate Effector Site

David YEE, Harald WIEDNER, and Fritz ECKSTEIN

Max-Planck-Institut fur experimentelle Medizin, Abteilung Chemie, Gottingen

(Received May 20/September 24, 1980)

The steady-state kinetics of the K', Ca2+, and Mg2+-activated adenosine triphosphatase (ATPase) activities of rabbit skeletal myosin were reinvestigated in the substrate concentration range from 0.05 yM to 5 mM and found not to follow Michaelis-Menten kinetics but rather to display biphasic behavior.

The Ca2+-ATPase activity of myosin chymotryptic subfragment-1 (S-l), which has only one active site, also exhibits biphasic kinetics, thus excluding the possibility that the biphasic behavior is caused by negative cooperativity between the two active sites of myosin. Myosin K + and Mg2+- ATPase are both activated by 5'-adenyl methylenediphosphonate (AdoPP [CHZIP) in a competitive manner at high substrate concentrations; i.e. the maximal velocity observed at high substrate con- centrations is independent of the AdoPP [CH*]P concentration. This result provides evidence for substrate activation via binding to a regulatory site.

Pyrophosphate inhibits myosin ATPase in a competitive manner at low substrate concentrations and in an uncompetitive manner at high substrate concentrations, with the uncompetitive Ki being smaller than the competitive Ki; i.e. pyrophosphate binds more tightly to the effector site than to the active site.

Because of the importance of myosin in muscle contraction (see [ 1 - 31 for reviews), the ATPase ac- tivity of this protein has been studied extensively. Early work with actomyosin [4,5] and myosin [6- 101 ATPase indicated that these activities obeyed Michae- lis-Menten steady-state kinetics under a variety of conditions.

With the pioneering work of Tonomura and co- workers [l 1,121, and especially after the classic paper of Lymn and Taylor [13], the pre-steady-state kinetics of myosin have been emphasized and have led to a detailed mechanistic scheme for the hydrolysis of Mg2+-ATP by myosin [l - 31.

The success of the pre-steady-state approach in elucidating the mechanism of myosin MgZ '-ATPase has perhaps obscured some interesting observations about the steady-state kinetics of this enzyme; e.g. Hasselbach [14] reported substrate inhibition of Mg2+- ATPase at concentrations above 1 mM while Kiely and Martonosi [15] observed substrate activation of Mg2+-ATPase at concentrations above 2.5 mM. The phenomena of substrate inhibition ([ATP] > 1 mM)

Abbreviations. S-I, chymotryptic subfragment 1 ; AdoPP[CH2]P,

Enzyme. Myosin ATPase (EC 3.6.1.3). 5'-adenyl methylenediphosphonate.

and activation ([ATP] > 6 mM) were also seen for actomyosin Ca2+-ATPase (but not K+-ATPase) by Tonomura and coworkers [16,17]. In addition, Inoue et al. [18] noted that myosin Mg2+-ATPase (but not K +-ATPase) deviates from Michaelis-Menten steady- state behavior at substrate concentrations below 0.5 pM. Nihei and Filipenko [19] have also reported significant deviations from Michaelis-Menten kine- tics for myosin Mg2+-ATPase.

Bendall [20], in studying the steady-state kinetics of myofibril Mg2+-ATPase, observed biphasic behavior, and Malik suggested the existence of both positive and negative cooperativity for the ATPase activity of myosin from chicken gizzard [21] and bovine platelets [22].

In this paper we reinvestigate the steady-state kinetics of myosin K', Ca2+, and Mg2+-ATPase, confirm that they deviate significantly from Michaelis- Menten behavior, and provide evidence for a sub- strate-regulatory site.

MATERIALS AND METHODS Mutevials

ATP and [p3*P]ATP were purchased from Pharma- Waldorf and Amersham Buchler, respectively ; Na4PPi

86 Steady-State Kinetics of Myosin ATPase

was obtained from Merck (Darmstadt) ; AdoPP- [CHI2P was supplied by Boehringer (Mannheim).

Rabbit skeletal myosin was isolated and quantified as previously described [23], and its purity was checked by sodium dodecylsulphate/polyacrylamide gel elec- trophoresis according to Starr and Offer [24].

Chymotryptic subfragment 1 (S-1) was prepared by chymotryptic digestion of myosin as described by Weeds and coworkers [25,26]; however, the isoenzymes were not separated. The concentration was determined spectroscopically. assuming A:;$ nm = 750 [26].

Enzyme Assu~,

All assay reactions contained in a 2-ml volume 1-500 pgiml enzyme, 25 mM Tris/HCl, pH 8.0, and 1 mM EDTA and 0.5 M KCl for K+-ATPase, 0.1 M NaCl and I0 mM CaCl2 for Ca2+-ATPase, and 0.1 M NaCl and 5 mM MgC12 for Mg2'-ATPase. The sub- strate (in the form of its sodium salt, unless other- wise noted) concentration ranged from 0.05 pM to 5 mM.

Enzymatic hydrolysis of ATP was measured by following the release of inorganic ["Plphosphate from [.y3'P]ATP [23] at 25 "C. Reactions were initi- ated by the addition of enzyme, and aliquots of the assay mixture were transferred to disposable plastic centrifuge tubes containing a slurry of 20- 120 mg activated charcoal in 1 in1 1 mM NaHzP04/0.2 M HC1 ethanol water (20: 80, v/v) solutions, which had been precooled to 0 "C. The tubes were centrifuged for about 15 min at 5000 rev./min in an MSE (Measuring Scientific Equipment, London) Chilspin centrifuge precooled to 4 'C, and aliquots of the supernatants were counted in Aquasol. No significant decomposi- tion of ATP occurred between the acid quench of an

0 5 10 15 20 v/ [s ] (ml min-1 mg-1)

Fig. 1. Eadie-Hojstee plot of myosin's K ' -A TPuse activity

assay aliquot and the withdrawal of a supernatant sample for counting.

Culculu t ions

Initial velocities were calculated from data ob- tained within the first 4 min of reaction by linear least- square analysis. Kinetic data which appeared to follow Michaelis-Menten kinetics were analyzed by linear least-squares analysis of u versus c / [ S ] (Eadie-Hofstee plot). Biphasic kinetic data were fitted to Eqn (2) using a program [27] kindly provided by Dr A. Cornish- Bowden, in which the absolute errors were assumed to be constant. Competitive inhibition constants (K, ) were determined from apparent Kn,l or Km2 values as a function of modifier concentration, and un- competitive K , values were determined from the reci- procal of the apparent Km2 value as a function of the inhibitor concentration (see Discussion).

RESULTS

Kinetics in the Presence of Vurious Cutions

The K', Ca2+ and Mg2'-ATPase activities of myosin were measured over the substrate concentra- tion range 0.05 pM to 5 mM. The only linearized kinetic plot which can visually accomodate ail the data over such a wide range is the Eadie-Hofstee plot. Biphasic kinetics of all three activities can clearly be seen in Fig. 1-3.

At substrate concentrations above 1 mM, Ca2+- ATPase but not K'-ATPase or Mg2+-ATPase ex- hibited substrate inhibition (data not shown); how- ever, the use of Ca2+-ATP in place of Na+-ATP as substrate essentially abolished this effect.

v / [ s ] (mi min-' mg-1)

Fig. 2. Eudie-Hofstee plot of' myosin's Cu2 ' - A TPuse nrtivitv

D. Yee, H. Wiedner, and F. Eckstein 87

Y / [ s ] (rnl rnin-1 rng-1)

Fig. 3. Eadie-Hqfstee plot of myosin's h4g2+-ATPase activity

log [SIliiM

Fig.4. The effect of modifiers on myosin's K'-ATPase activity. In the absence of modifiers (o---o), and in the presence of 25 mM AdoPP[CH2]P (D+), 10 mM PP; (A-A), and 25 mM PP, (v--v). For the sake of clarity, data at higher substrate con- centration and at other modifier concentrations are not shown

Studies with AdoPP]CH2]P and Pyrophosplzate

A plot of v versus log [ S ] is used to show clearly the effect of modifiers on ATPase activity. As shown in Fig.4, AdoPP[CH2]P, at concentrations up to 25 mM, activates K+-ATPase. At concentrations up to 1 mM it also activates Mg2+-ATPase but begins to inhibit a higher concentrations (data not shown), presumably because chelation of free divalent metal ion leads to a suboptimal Mg2+ concentration. The limited solubility of Ca2+-AdoPP [CH2]P precluded an investigation of the effect of AdoPP[CH2]P on the Ca2+-ATPase activity.

The effect of pyrophosphate on K+-ATPase is also shown in Fig.4. It is inhibitory at all concentrations up to 25 mM. At this latter concentration the steady- state kinetics of myosin K+-ATPase were no longer bisphasic but rather followed Michaelis-Menten kine- tics. Myosin CaZ+-ATPase was also inhibited by pyrophosphate at concentrations up to 1 mM (data not shown). Higher concentrations of pyrophosphate could not be tested in the Ca2+-ATPase assay because of the limited solubility of calcium pyrophosphate.

DISCUSSION

Despite occasional postulations of two non-identi- cal or interacting active sites, it is generally assumed that myosin has two equivalent binding sites for ATP and that the kinetics are of the Michaelis-Menten type [l - 31. However, there are reports in the litera- ture which describe deviations from this behaviour [34- 191 or which mention the difficulty in fitting the data to a single Michaelis constant [28]. During a study on the interaction of myosin ATPase with phosphorothioate analogues of ADP and ATP [29], we also observed that the kinetics were not strictly linear. This prompted us to reinvestigate the steady- state kinetics of myosin K', Ca2+ and Mg2+-ATPase. The results, summarized in Fig. 1 - 3, show clearly the biphasic nature of all three activities although the deviation from linearity varies for the three ions. The most pronounced break in the kinetics is seen for the Mg2+-ATPase. The Michaelis-Menten kinetics of myosin ATPase reported earlier were most probably caused by working in a limited range of substrate concentration as well as representing the data in a Lineweaver-Burke plot, a procedure which tends to minimize any curvature in the data [30]. The present results partially confirm those of Inoue et al. [18], who observed biphasic behavior of myosin ATPase independent of the cation. The substrate inhibition above 1 mM and activation above 2 mM ATP of Mg2+-ATPase [14,15] and Ca2+-ATPase [16,17], seen by previous workers, is most easily explained in terms of changes in the concentration of free divalent metal ion. As the substrate concentration increases, the amount of free divalent metal ion decreases to a sub- optimal level [31], leading to a decrease in the initial velocity; as the substrate concentration is further in- creased, the free divalent metal concentration ap- proaches zero, and the K+-ATPase activity is un- masked [32], resulting in apparent substrate activation [15,17]. This explanation accounts for the fact that K+-ATPase is never seen to exhibit substrate inhibition and activation at substrate concentrations above 1 mM.

In order to explain the biphasic steady-state be- havior of myosin shown in Fig. 1 - 3, we considered the followed three possibilities : (a) negative coopera-

88 Steady-State Kinetics of Myosin ATPase

tivity between the two active sites of myosin, (b) the presence of contaminating ATPase(s), and (c) sub- strate activation via a regulatory site.

If the two active sites of myosin interact in a negatively cooperative manner, then the kinetic data would be biphasic and qualitatively similar to the results obtained herein [33]. Various investigators have indeed postulated such an interaction for myosin [19,34]. In order to test this hypothesis, the steady- state kinetics of S-I, which has only one active site per molecule, were investigated and found to be like- wise biphasic in its CaZt-ATPase activity (data not shown). Thus, assuming that the two active sites of myosin are identical [35], the possibility that the two active sites interact in a negatively cooperative manner cannot explain the biphasic steady-state data. This result and conclusion are at variance with that of Nihei and Filipenko [19], who observed Michdelis- Menten kinetics with S-1 (but not with myosin). The reasons for this discrepancy are unknown.

If there is a contaminating ATPase activity in the myosin preparation, then biphasic kinetics will also result [33] . Morgan et al. [36] state that most myosin preparations are contaminated by myosin light-chain kinase and myosin light-chain phosphatase. These two enzymes, in conjunction with myosin, will appear as a contaminating CaZf-ATPase. Sleep et al. [37] have estimated that 34‘;,, of the MgZf-ATPase activity of myosin derives from a contaminating enzyme or en- zymes. If actin is present in the myosin preparation, then both myosin and actomyosin ATPase may be seen [l - 31. Myosin isoenzymes, however, seem not to differ in their ATPase activities [36,38-401. The arguments against significant contaminating ATPases rest most importantly on the independence of the biphasic kinetics with metal ion. In particular the fact that, even in the presence of 1 mM EDTA and 0.5 M KCI. the biphasic character persists argues strongly against any contamination by an ATPase since, to the best of our knowledge, myosin is the only ,ATPase which is active under such conditions.

If each myosin head had a substrate-regulatory site as well as a catalytic site, then biphasic kinetics would be expected if occupancy of the regulatory site resulted in activation. As early as 1959, Nihei and Tonomura [17] postulated the existence of a non-hydrolytic binding site for myosin. This idea has subsequently been reinvoked by many other investigators. Thus, the dependence of contraction on ATP concentration [41, 421, theeffect of PPi binding to myosin [15], the chemi- cal modification of heavy meromyosin by an ATP- analogue [43] and the fluorescence changes associated with the binding of ribose 5-triphosphate [44] have all been interpreted in terms of a second binding site.

I n order to test whether the biphasic behavior of myosin as represented in Fig. 1-3 is due to the ex- istance of such a regulatory site, we employed first

k , k2 E + S E X - E + P + k-‘ + S S

k4 k6

k-4 S E + S SEX - S E + P

Fig. 5. Proposed kinetic sclieme for niyosin A TPuse

the non-hydrolysable b, y-methylene analogue of ATP, AdoPP[CHz]P. This analogue has been shown earlier to be a weak inhibitor for myosin ATPase at concen- trations below 0.5 mM but at higher concentrations had exhibited an activating effect [45]. No detailed kinetic study on this activation had been carried out, but it was suggested that this analogue might occupy, at higher concentrations, a second weaker binding site.

It can be seen in Fig.4 that AdoPP[CH?]P at concentrations up to 25 mM does indeed activate K + - ATPase. Since the K’-ATPase is unphysiological, we attempted to carry out similar kinetic studies in the presence of Ca2+ and in particular of Mg2+. As al- ready noted by others [45] Ca2+-AdoPP[CH2]P is not very soluble; therefore, Ca2+-ATPase could not be studied in the presence of this analogue in detail. In the presence of Mg2+ only a limited study up to a con- centration of 1 mM AdoPP[CHz]P could be per- formed. Up to this concentration activation of the Mg2+-ATPase is seen, but at higher concentrations inhibition can be observed, probably because of a depletion of free Mg2+ from the system.

Pyrophosphate has also been claimed as a ligand to a second, non-hydrolytic site. It binds to myosin stoichiometrically [46] and activation has been ob- served in its presence [15]. In our study (Fig.4) it was shown to inhibit the K+-ATPase. Earlier reports claiming pyrophosphate activation of myosin ATPase can be explained simply in terms of chelation of the activating divalent metal ion by pyrophosphate, re- sulting in the unmasking of the K+-ATPase ac- tivity [17].

From the above results the kinetic scheme shown in Fig.5 is postulated to explain the biphasic steady- state kinetics of myosin. In this model [47] not only can substrate bind at the active site to yield a steady- state intermediate, EX, which then gives rise to prod- uct, it can also bind at an effector site, giving rise to the enzyme species SE and SEX. Analysis of the kinetic scheme by the method of King and Altmdn [48] yields the following steady-state rate equation :

where the coefficients A - F are combinations of vari- ous rate constants. In order to solve this equation

D. Yee, H. Wiedner, and F. Eckstein 89

Table 1. Calculated kinetic parameters jor myosin ATPase The kinetics parameters are defined by the scheme schown in Fig. 5 , with K,z = (k-5 + k6jks) and K,I = (k-1 + k z / k l )

Parameter K +- Ca2 +- Mg2+- ATPase ATPase ATPase

kb (pmolmin-'mg-') 5.1 kO.1 2.6 f 0 . 2 0.82 f0 .02 Km2 (mM) 1.0 k 0 . 4 0.19k0.05 0.10 fO.O1

k2 (pmol min-' mg-') 0.006 0.003 0.5 k 0.4 0.024 k 0.008

Kmi (PM) 0.2 kO.2 4 4 0.008 kO.008

and to obtain discrete kinetic parameters some as- sumptions have to be made.

If it is assumed that the effector site can be occupied only after the active site is occupied (i.e. there is no SE species), then Eqn (1) simplifies to Eqn (2):

where

If the additional assumption is made that kl

+ k-5 + k6, then the following kinetic parameters may be obtained from the above expressions:

kg = c; Km2 = a ; k2 = d/a and Kml = his. These four parameters for myosin ATPase activi-

ties were obtained from the above equations by fitting the biphasic data to Eqn (2). Km2, k6, K,I and k2, may be considered to be the K,,, and V values at high and low substrate concentrations, respectively. The results are summarized in Table 1. The large errors associated with the fitted parameters h and d, which in turn lead to large errors in k2 and KmI might in- dicate that one or both of the assumptions made is not valid. Thus, it might well be that there is some contribution from the existence of SE.

Using the data in Table 1 and Fig.4, as well as additional data not shown, estimates of the activation constant (K, = - Ki) of AdoPP[CH2]P and the Ki value of pyrophosphate can be obtained. For AdoPP- [CH2]P the activation is competitive (i.e. kg is un- changed by the activator), and the K, value is 30 mM and 7 mM for K+-ATPase and Mg2+-ATPase respec- tively. For K+-ATPase, pyrophosphate inhibition at low substrate concentrations is competitive (i.e. k2 is unchanged) with K i = 4 mM, whereas inhibition at high substrate concentrations is uncompetitive (i.e. k6/Km2 is unchanged) with Ki = 0.7 mM.

The competitive and uncompetitive Ki values for pyrophosphate are essentially the dissociation con- stants from the active and effector sites, respectively. This indicates that pyrophosphate binds to the regula- tory site more tightly than to the active site. These data support the results of Kiely and Martonosi [15], who concluded that pyrophosphate binds to a regula- tory site, and of Nauss et al. [46] who found that it binds stoichiometrically to myosin.

Myosin thus joins a growing list of energy-trans- ducing ATPases which are known to exhibit biphasic kinetics. This list includes (Na+/K +)-transport ATPase [49], mitochondria1 (FI) ATPase [ S O ] , and Ca2+- ATPase from sarcoplasmic reticulum [51]. The bi- phase kinetics of this last enzyme, like that of myosin, have been determined, also by the use of AdoPP- [CH2]P, to result from allosteric substrate activation. It may be that such a substrate control mechanism is physiologically important for all energy-transducing ATPases. However, it is difficult at present to envisage how such a regulation can work for myosin since, at the ATP concentrations found in muscle, both the hydrolytic as well as the regulatory site will be satu- rated. Future studies on more complete systems will have to be done to answer this question.

The authors are indebted to Dr A. Cornish-Bowden (Birming- ham) for supplying the computer program used for analysis of the biphasic kinetic data, Mr M. Heidrich for his assistance in using the program, Miss G. Haag for her valuable technical assistance, and Drs B. A. Connolly, B. K. Sathyanarayana and V. W. Arm- strong for helpful discussions. The investigation was supported in part by the Deufsche Forscliungsgemeinsclzuft.

REFERENCES

1.

2.

3. 4.

5 .

6.

7.

8. 9.

10.

11.

12.

13.

14. 15.

Mannherz, H. G . & Goody, R . S. (1976) Annu. Rev. Bioclzem.

Trentham, D. R., Eccleston, J. F. & Bagshaw, C. R. (1976)

Taylor, E. W. (1979) Crit. Rev. Bioclzem. 6, 103-164. Ouellet, L., Laidler, K. J. & Morales, M. F. (1952) Arch. Bio-

Tonomura, Y., Watanabe, S. & Yagi, K. (1953) J . Biochem.

Green, I. & Mommacrts, W. F. H. M. (1954) J . Biol. Cliem.

Kielley, W. W., Kalckar, H. M. & Bradley, L. B. (1956) J .

Nanninga, L. B. (1959) Biochim. Biopkys. Acta, 36, 191 -202. Nanninga, L. B. & Mommaerts, W. F. H. M. (1960) Proc. Nut1

Tonomura, Y., Tokura, S., Sekiya, K. & Imamura, K. (1961)

Kanazawa, T. & Tonomura, Y. (1965) J . Biocliem. (Tokyo) 57,

Inoue, A. & Tonomura. Y. (1973) J . Bioclzern. (Tokyo) 73,

Lynn, R . W. & Taylor, E. W. (1970) Biochemistry, 9. 2975-

Hasselbach, W. (1952) Z . Nuturforsclz. 7h, 163- 174. Kiely, B. & Martonosi, A. (1968) J . Biol. Chem. 243, 2273-

45.427 - 465.

Q. Rev. Biophys. 9, 217-281.

cliem. Biophys. 39, 37-50; 40. 238.

(Tokyo) 40, 27 - 54.

210, 695 - 702.

Biol. Clzem. 219, 95-101.

Acad. Sci. USA, 46, 1166-1173.

Arch. Bioclzem. Bioplzys. 95, 229 - 236.

604 - 61 5.

555 - 566.

2983.

2278.

90 D. Yee. H. Wiedner, and F . Eckstein: Steady-State Kinetics of Myosin ATPase

16. Watanabe, S., Tonomura, Y. & Shiokawa. H. (1953) J . Bio-

17. Nihei, T. & Tonomura, Y. (1959) J . Biocliem. (Tokyo) 46,

18. Inoue, A,, Shibata-Sekiya. K . & Tonomura, Y. (1972) J . Bio-

19. Nihei, T. & Filipenho. A. (1975) Can. J . Biochem. 53, 1282-

20. Bendall, J . R. (1972) FEBS Lett. 22, 330-334. 21. Malik, M. N. (1978) Biochemistry. 17, 27-32. 22. Malik, M. N. (1979) Arch. Biochem. Biophys. 196, 147-156. 23. Wiedner. H., Wetzel, R. & Eckstein, F. (1978) J . Bid . Cliem.

24. Starr. R. & Offer, G. (1971) FEBS Lett. 15,40-44. 25. Weeds, A . (i. & Taylor, R. S. (1975) Nature (Lond.) 257,

26. Weeds. A. G. &Pope, B. (1977) J . Mol. B i d . 111, 129-157. 27. Wharton, C . W.. Cornish-Bowden, A,, Brocklehurst, K. &

Crook, t. M. (1974) Biochem. J . 141, 365-381. 28. Taylor, t:. W. (1977) Biocliernistry, 16, 732-740. 29. Yee, D. & Eckstein. F. (1980) FEBS Lett. 112, 10-12. 30. Dowd, J . E . & Riggs, D. S. (1965) J . Biol. Chem. 240, 863-

31. Mommaerts. W. F. H. M. & Green, I. (1954) J . Bid . Chem.

32. Seidel, J . c' (1969) Biorhim. Biophys. Acta, 189, 162-170. 33. Cleland. W. W. (19701 in The Enzymes (Boyer, P. D., ed.)

vol. 11, pp. 1-65. Academic Press, New York and London. 34. Walz, F. G J r (1973) J . Theor. Biol. 41, 357-373.

diem. (TokyoJ 40. 387-402.

305 - 319.

chenz. (Tokyo) 71. 115-124.

1287.

253, 2763 - 2768.

54 - 56.

869.

211, 237~-247.

35. Chock, S. P. & Eisenberg, E. (1979) J . Bid. Chem. 254, 3229-

36. Morgan, M., Perry, S. V. & Ottaway, J. (1976) Biochem. J .

37. Sleep, J. A,, Hackney, D. D. & Boyer. P. D. (1978) J . Bid.

38. Syrovy, I . (1979) Int. J . Biochem. 10, 383-389. 39. Taylor, R. S. &Weeds, A. G. (1977) FEBS Lett. 75, 55-60. 40. Wagner, P. D. (1977) FEBS Lett. 81, 81 -85. 41. Levy, H. M. & Ryan, E. M. (1966) Biochem. Z. 345. 132-147. 42. Kominz, D. R. (1970) Biochemistry, 9, 1792. 43. Wagner, P. D. & Yount, R . G. (1975) Biochemistry, 14, 5156-

44. Eccleston, J. F. (1980) FEBS Lert. 113, 55-57. 45. Yount, R. G., Ojala, D. & Babcock, D. (1971) Biochemistry,

46. Nauss, K. M., Kitagawa, S. & Gergely. J . (1969) J . Biol. Clzem.

47. Botts, J. & Morales, M. (1953) Trans. Furaday Soc. 49, 696- 707.

48. King, E. L. & Altman, C. (1956) J . P h p . Clieni. 60, 1375- 1378.

49. Kanazawa, T., Saito, M. & Tonomura. Y . (1970) J . Biochem. Tokyo) 67, 693-711.

50. Takeshige, K., Hess, B.. Bohm, M. & Zimmermann-Telschow, H. (1976) Hoppe-Seyler's Z . Physiol. Cliem. 357, 1605- 1622.

51. Taylor, J. S. & Hattan, D. (1979) d . Bid . Cliem. 254, 4402- 4407.

3235.

157, 687-697.

Chem. 253, 5235-5238.

5 162.

10,2490 - 2496.

244, 755 - 765.

D. Yee, Department of Chemistry, Harvard University, 12 Oxford Street, Cambridge, Massachusetts 02138, USA

H. Wiedner. Boehringer Mannheim GmbH, BahnhofstraBe 5. D-8132 Tutzing, Federal Republic of Germany

F. Eckstein. Abteilung Chemie, Max-Planck-Institut fur experimentelle Medizin, Hermann-Rein-StraBe 3 . D-3400 Gottingen, Federal Republic of Germany