Vol.

172,

October

No. 30,

THE

2, 1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Pages

1990

BINDING OF ATP AND INVESTIGATED

Ah4P TO ESCHERICHIA

COLI ADENYLATE

432-438

KINASE

BY ‘H AND 15N NMR SPECTROSCOPY

John Glushkaa, Octavian Barzub, Robert S. Sarfatib, Vinod K. Kansalb, and David Cowburna

aThe Rockefeller University, 1230 York Ave., N.Y., N.Y. 10021 bInstitut Pasteur, Departement de Biochimie et de GCnCtique Moleculaire, 75724 Paris, France

Received

September

10,

1990

[N6 “N]ATP and [N6 “NJAMP, complexed with E.cofi adenylate kinase (AKe), were observed with “N isotope-filtered NMR pulse sequences and ‘H{ “N} heterocorrelated experiments to determine differences between binding sites based on chemical shifts and competition by substrate analogs. The chemical shifts of the N6 amino proton and nitrogen signals changed significantly after mixing with adenylate kinase. Differences in chemical shifts between the bound ATP and AMP signals are slight. The response of these shifts to further addition of other substrates or Mg*+ supports the view that the unchelated nucleotides can bind to both the sites, whereas the metal complexed species are restricted to the MgATP/MgADP binding site. _ 1990Aca3rml;PTCSS, 1‘1,:.

Adenylate kinase is responsible for the reversible transfer of a phosphoryl group from MgATP to AMP, and from MgADP to ADP. The gene for the E.coli and overexpressed, providing

variant has been cloned

suitable material for NMR investigation (1,2).

Although the

general enzyme mechanism and features of the catalytic site are known, NMR studies (3,4), crystal structure data (5), and results from mutagenesis studies (6,7,8) do not fully agree on the precise location of the substrate binding sites. There is also disagreement over the relative affinity of substrates and analogs for the two sites(9). Investigation of bound substrates by proton NMR has been restricted to the purine ring protons (10). A synthesis of N6 “N labeled adenosine (11) has provided nucleotide probes which permit selective observation of the N6 amino protons, through the application of isotope filtered NMR pulse sequences (12). The 15N chemical shift may also provide information on hydrogen bonding of the amino group. An intact N6 amino group in the mononucleotide substrates is required for catalysis (13), and therefore is a good candidate for hydrogen bonding with suitable acceptors on the enzyme. A similar proposal has been made for the amino group in ATP (14). Using the 15N labeled nucleotides we investigated whether AMP and ATP can bind to multiple sites in AKe

Vol.

172,

No.

2, 1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

and whether the two sites can be distinguished by measuring chemical shifts of the bound labeled nucleotides. Materials and Methods Sample Preparation: E.coli adenylate kinase was prepared as described by Saint-Girons et al. (2). Solutions of AKe were prepared in 90% H20/ 10% D20 buffer containing 20mM ammonium hydrogen carbonate, 3 mM EDTA, and 0.2 mM sodium azide, at pH 7.8 - 8.0. Concentrations of protein were determined by weight and checked by absorbance at 277 nm = 5). The pH was adjusted by addition of small amounts of HCl and ammonium bicarCEl% bonate solutions. Trimethylsilyl proprionate was added as an internal standard. [N6 “N]AMP and [N6 “NJATP were obtained via synthetic methods described (11) 8-BrAMP, p,ymethyleneadenosine Y&phosphate (AMPPCP) and AMP were purchased from Sigma. Solutions (- 100 n&I) were prepared in ammonium bicarbonate buffer, and concentrations were determined spectrophotometrically at 259 nm (Eti = 15.4). Nucleotide solutions were added directly to the NMR sample tube before spectra were acquired. Protein samples were stored at 5°C and nucleotide solutions were kept frozen. NMR: All spectra were run on a GE GN500 spectrometer at 500 MHz, using 5 mm One-dimensional spectra were probes provided by Cryomagnetics, Inc. or GE-NMRI. obtained with an 15N filtered one pulse sequence (15). Typically, excitation of the water signal was avoided by using a Redfield 214 selective pulse (16) or a 1-I spin echo sequence (17). Two dimensional ‘H{ “N} heterocorrelated spectra were obtained using the pulse sequence of Bax et al (17,18). A typical one-dimensiona experiment used 4096 complex points with an acquisition time of 680 ms; 200 scans were adequate for the unrefocussed absolute value spectra, and 1000 to 2000 scans were acquired for the refocus& decoupled spectra. The absolute value two-dimensional spectra were collected in 4096 x 64 or 128 data sets, with 200 to 800 scans per tt block. The refocussed, decoupled pure absorption two-dimensional spectra required additional scans per block for adequate signal to noise. Results and Discussion The mixing of [N6 “N]AMP with a solution of AKe shifted the N6 amino proton and nitrogen resonances downfield by 0.7 ppm and 2.5 ppm, respectively ( Figure 1 and Table 1). The magnitude of the shift is sensitive to the ratio of substrate to enzyme, indicative of fast exchange (IO). Addition of increasing amounts of unlabeled AMPPCP to the binary complex shifted the AMP proton signal upfield to a value close to that of unbound AMP (Table 1). A similar upfield shift was observed with the addition of a five-fold excess of 8-BrAMP to another solution of AKe:[N6 “N]AMP. AMP is a competitive inhibitor

in porcine (19), yeast (20), and rabbit muscle (21)

adenylate kinase and is believed to bind to the MgATP site as well as the AMP site. In AKe, excess AMP inhibits enzyme activity, but does not show clear competitive behaviour (2). The upfield shift of the ‘H signal on addition of AMPPCP to the binary complex AKe:[N6 “N]AMP ( Figure Id and Table 1) is compatible with the displacement of AMP from the MgATP site. The additional shift observed after excess AMPPCP or 8-BrAMP was added suggests that the AMP is then displaced from the AMP site as well. This could arise by direct competition of AMPPCP for the AMP site, or indirectly, by the presence of the unchelated nucleotide triphosphate in its own site, perhaps due to electrostatic repulsion in the absence of the divalent cation. Alternatively, the observed chemical shift could also reflect 333

Vol.

172,

No.

2,

1990

b) IN6

BIOCHEMICAL

AND

BIOPHYSICAL

“N]AMP

RESEARCH

COMMUNICATIONS

k

c) [N6 “N]AMP:AKe 1.4: 1

g) [N6 “N]AMP:MgAMPPCP:AKe

I

I

8.0

7.2

‘H ppm 434

6.4

Vol.

172,

No.

2,

1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Table 1. ‘H and 15N chemical shifts of N6 NH, in [N6 15N]AMP and [N6 *‘N]ATP a substrate complex AMP AMPAKe 0.7 : 1 1.4: 1 AMP:AKe:AMPPCP 1.4 : 1 : 0.7 1.4 : 1 : 1.4 1.4: 1:5 1.4 : 1 : 14 AMP:AKe:MgAMPPCP 1.4 : 1 : 1.4 1.4 : 1 : 2.1 1.4 : 1 : 2.8 1.4 : 1 : 4.3 1.4 : 1 : 8.6

6 ‘Hb 6.83

6 15Nc 77.8

7.4 7.5 (6.89)d

80 80.5 (78.4)

AMP:AKe:8BrAMP 2.3 : 1 : 8.6

6.88

ATP ATP:AKe 0.7 : 1 1.4 : 1

6.83

77.8

7.5 7.6 (8.3)e

80 80.8 (82.0)e

MgATP:AKe 1.4: 1

6.94

7.29 7.18 6.95 6.9 7.25 7.20 7.18 7.04 (6.86) 6.98

78.6 (78.5)

a) 25”C, pH 7.5-7.8. b) c) d) e)

conformational

ppm from TSP. ppm from liquid NH3. values in parentheses are tentatively assigned to [N6 “N] adenosine. values in parentheses are tentatively assigned to [N6 15N]ADP.

change

in the region

fluorescence measurements substrates

agree with

of the bound

(14), and infrared

AMP.

‘H NMR spectra (10,22,23,24),

spectra (25) of adenylate

kinase

complexed

with

X-ray observations (26) and binding studies (13), in revealing a substrate

induced conformational change in the ternary complex. Mg’+ (10 n&I) was then added, which shifted the proton signal slightly downfield ( Figure le and Table 1). If the metal complexed nucleotide triphosphate can bind only to the MgATP site, then this spectrum should represent the AMP bound predominantly at its own

Figure 1. (a) ‘H spectrum of the 6.4 - 9.0 ppm amide region of AKe, (b)-(c) 15N filtered ‘H spectra of N6 NH, signals of IN6 “N]AMP with 1.3 mM AKe, and AMPPCP at pH 7.8 and 25°C. The ratios of nucleotides and enzyme are indicated. The full spectral width was 6024 Hz over 4096 complex points. A jump-return selective pulse was used to avoid excitation of the water signal. “N MLEV64 decoupling was applied during the acquisition, and line broadening of 10 to 20 Hz was used in processing. 435

Vol.

172,

No.

2, 1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

site. Surprisingly, further addition of MgAMPPCP shifted the AMP signal upfield to approximately 7.0 ppm. This unexpected result suggests that the MgAMPPCP can also bind at the AMP site. However, it is more likely that the increased concentration of MgAMP,

which

does not bind at the AMP site (13), caused the upfield shift. These results indicate that the presence of AMPPCP perturbs the bound AMP, displaces it from the donor-MgATP site, and at higher concentrations, displaces it from the acceptorAMP site as well. Assuming substrate site affinity, the [N6 “N]AMP:AMPPCP:AKe (1.4:1.4:1) complex with the N6 amino proton signal at 7.25 represents the AMP bound at the AMP site. The position of this signal in the binary complex suggests that the amino protons of AMP bound in the MgATP site experience a downfield shift to a value greater than 7.5 wm. Spectra of the [N6 l’N]ATP:AKe

binary complex also showed downfield shifts of both

‘H and “N signals immediately after addition of substrate to the enzyme (Table 1). A single broad signal was initially observed at 7.6 ppm and 80 ppm for proton and nitrogen signals, respectively, which presumably represents ATP in rapid exchange with both sites and the solution (4). As described in the experimental section, two X-filtered proton pulse sequences were used for the one-dimensional

spectra. The difference between the two sequences in length of delays demonstrated the short effective T, of the signals, which were weaker and broader than those obtained with [N6 15N]AMP.

The relaxation rates have contributions by

proton exchange as well as site exchange rates and so may be influenced by nearby acid or base catalysts. Because of this rapid exchange broadening, the two-dimensional spectra were of low signal to noise and low resolution (Figure 2).

[N6”N]ATP

[N6 “NIADP

J --I--8.8

7.2

8.0 ‘H PP~

Figure 2. Portion of an ‘H{15N} heterocorrelated spectrum of the [N6 “N]ATP in the presence of AKe after 12 Hrs. Also present is a downfield signal attributed to IN6 “NIADP. A 214 Redfield selective pulse was used to avoid excitation of the water signal. The data was collected in a 4096 x 64 file, with gaussian to lorenztian apcdization applied to t2, and exponential line broadening and zero-filling applied to f 1. before fourier transforming. No 15N decoupling was used during acquisition and the spectrum is presentedin absolute value mode. 436

Vol.

172,

No.

2,

1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

After addition of Mg2+, the proton signa 1 shifted upfield to 7.2 ppm. It has been proposed that the addition of Mg2+ to the binary AKe:ATP complex has the effect of “tying down” the phosphate chain (24), but in general does not alter the local mobility of the adenine ring. This infers that the position of the ring, and hence the environment of the N6 amino group are not affected. The change of chemical shift can therefore be interpreted as the reduction of the amount of ATP in the AMF site. This suggests the value of 7.2 ppm represents MgATP

in the MgATP site, and the greater downfield shift of the signal in the

absence of Mg2+ is due to ATP in both sites. An observation of the formation of adenosine tetraphosphate from MgATP and muscle adenylate kinase also supports the view that ATP can bind to both the acceptor and donor sites (27). The binary complex of AKe:ATP was stable for approximately 2 hours at room temperature, after which ATPase activity produced increasing amounts of ADP.

The 31P spectrum

taken after 4 hr (data not shown) shows a mixture of ATP and ADP, corresponding to published spectra shown for mammalian AK:ATP:AMP:ADP in absence of Mg2+ (28,29). In the ‘H spectrum, a second peak appears at 8.4 ppm and the original peak at 7.5 ppm reduces in intensity ( Figure 2 and Table 1 ). The second peak is tentatively assigned to the ADP in rapid exchange between the two sites and the solution (28,29).

The relaxation rate of this

signal is such that it is observable only with a short delay, unrefocussed X-filtered experiment. The large downfield shift of both the ‘H and 15N signals is not compatible with a simple siteexchange between the ATP and AMP sites, if bound ADP is assumed to be in a position similar to the other nucleotides. If we assume that the ternary complex AKe:MgAMPPCP:[N6

“N]AMP

contains AMY

primarily in the AMP site, and that the binary complex AKe:[NG “N]AMP contains AMP in both sites, then both bound states induce downfield shifts in the amino protons, with the MgATP

site causing a greater shift. Similarly,

AKe:MgATP

if we assume that the binary complex

contains ATP primarily in the MgATP site, and AKe:ATP contains ATP in both

sites, then the effect of the sites is opposite to that for AMP.

That is, ATP experiences a

greater downfield shift in the AMP site than in the MgATP site. This would suggest that the ATP and AMP adenine rings sit in different locations when in the different binding sites. The behaviour of [N6 “N]ADP also suggests a location different from both AMP and ATP. If the downfield peak in figure 2 represents ADP primarily in the Am site (due to presence of ATP in the MgATP site), then it would suggest the ADP is in a location similar to that of ATF in the AMP site, i.e. where the amino protons are shifted downfield. Ackrwwledgmenl Supported by NM grant DK-20357. References 1. 2.

Brune, M., Schumann, R., and Wittinghofer, A. (1985) Nucleic Acid Res. 13, 71397151. Saint-Girons, I., Gilles, A.-M., Margarita, D., Michelson, S., Monnot, M., Fermandjian, S., Danchin, A., and Btizu, 0. (1987) J. Biol. Chem. 262, 622-629. 437

Vol.

172,

3. 4.

Fry, D. C., Kuby, S. A., and Mildvan, A. S. (1987) Biochemistry 26, 16451655. Fry, D. C., Kuby, S. A., and Mildvan, A. S. (1986) Proc. Natl. Acad. Sci. USA 83, 907-911. Muller, C. W. and Schulz, G. E. (1988) J. Mol. Biol. 202, 909-912. Reinstein, J., Gilles, A.-M., Rose, T., Wittinghofer, A., Saint-Girons, I., Barzu, O., Surewicz, W. K., and Mantsch, H. H. (1989) J. Biol. Chem. 264, 8107-8112. Haase, G. H. W., Brune, M., Reinstein, J., Pai, E. F., Pingoud, A., and Wittinghofer, F. (1989) J.Mol.Biol. 207, 151-162. Kim, H. J., Nishikawa, S., Tokutomi, Y., Takenaka, H., Hamada, M., Kuby, S. A., and Uesugi, S. (1990) Biochemistry 29, 1107-1111. Shyy, Y. -J., Tian, G., and Tsai, M. -D. (1987) Biochemistry 26, 64116415. ROsch, P., Klaus, W., Auer, M., and Goody, R. S. (1989) Biochemistry 28, 4318-4324. Sarfati, S.R. and Kansal, V.K. (1988) Tetrahedron 44, 6367-6372. Griffey, R. K. and Redfield, A. G. (1987) Quart. Rev. Biophysics 19, 51-82. Hamada, M., Palmieri, R. H., Russell, G. A., and Kuby, S. A. (1979) Arch, Biochem. Biophys. 195(l), 155-177. Tomasselli, A. G. and Noda, L. H. (1983) Eur. J. Biochem. 132, 109-115. Freeman, R., Mareci, T. H., and Morris, G. A. (1981) J. Magn. Res. 42, 341-345. Redfield, A. G., Kunz, S., and Hurd, T. (1975) J. Magn. Res. 19, 114-117. Sklenar, V. and Bax, A. (1987) J. Magn. Res. 74, 469-479. Bax, A., Griffey, R. H., and Hawkins, B. L. (1983) J. Magn. Res. 55, 301. Yazawa, M. and Noda, L. (1976) J. Biol. Chem. 251, 3021-3026. Ito, Y., Tomasselli, A. G., and Noda, L. H. (1980) Eur. J. B&hem. 105, 85-92. Rhoads, D. G. and Lowenstein, J. M. (1968) J. Biol. Chem. 243, 3963-3972. McDonald, G. G., Cohn, M., and Noda, L. (1975) J. Biol. Chem. 250( 17), 6947-6954. Kalbitzer, H. R., Marquetant, R., Rt)sch, P., and Schirmer, R. H. (1982) Eur. J. B&hem. 126, 531-536. Sanders II, C. R., Tian, G., Tsai, M.-D., Brenner, S. L., Mitchell, R. S., Morrical, S. W., Neuendorf, S. C., Schutte, B. C., and Cox, M. M. (1989) J. Biol. Chem 262, 401 I-4016. Arrondo, J. L. R., Gilles, A.-M., B$rzu, O., Fermandjian, S., Yang, P.W., and Mantsch, H.H. (1989) B&hem. Cell. Biol. 67, 327-331. Schulz, G. E., MUller, C. W., and Diederichs, K. (1990) J. Mol. Biol. 213, 627-630. Kupriyanov, V. V., Ferretti, J. A., and Balaban, R. S. (1986) Biochimica et Biophysics Acta 869, 107- 111. Rao, B. D. Nageswara, Cohn, M., and Noda, L. (1978) J. Biol. Chem. 253(4), 11491158. Vasavada, K. V., Kaplan, J. I., and Rao, B. D. Nageswara (1984) Biochemistry 23(5), 961-968.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

No.

2,

1990

BIOCHEMICAL

AND

438

BIOPHYSICAL

RESEARCH

COMMUNICATIONS