Interaction of Sulphate and Chloride with Cobalt( II) -Carbonic Anhydrase Jos& M. Moratal, Antonio Donaire, Jestis Salgado, and Maria-Jo& Martinez-Ferrer JMM,AD,JS. Department of InorgcmicChemistry, University of Valencia, Burjawot (Valencia), Spain.--M-J M-F. Centto de Estudios Universitarios S. Pablo, Valencia, Spain

ABSTRACT Theinteractionbetween Cobalt(II)-Bovine Carbonic Anhydrase II and the inhibitors sulphate and chloride have been investigated through ‘H NMR and electronic absorption spectroscopies. Both inhibitors bind to the metal ion forming a 1:l adduct and the corresponding affinity constants have been determined. These inhibitors interact weakly with CoBCA II and this ktemction only occurs at low pH values. The T, values of the me&Iii protons of the coordinated histidines have been measured. The coordination number of the metal ion in the adducts is discussed on the basis of temperature dependence of the isotropic shifts, T, , and molar absorbance values.

INTRODUCTION Carbonic Anydrase (CA) is a zinc-containing enzyme that catalyzes the simple, reversible reaction C& +H20 = HC03 - + H+ [ 1, 21. The x-ray structure is available for HCA I and HCA II at 2 A of resolution [3-71. As far as the bovine carbonic anhydrase II is concerned, its high homology with the corresponding HCA II [8, 91, suggests a close structural similarity [8] even in the absence of a high resolution crystal structure [lo]. In all cases, the catalytic zinc ion is coordinated to three imidazole groups from His-94, His-96, and His-l 19 (the numbering is relative to the CA II isoenzyme) and to one or two solvent molecules. The refined structure of CA II revealed a number of ordered water molecules in the cavity, which form a hydrogen bond network between the zinc ion and the side chain of His-t%; many hydrophillic residues are also involved in this hydrogen bond network. It is believed that the coordinated water undergoes deprotonation to provide the Address reprint requests to: Prof. J. M. Moratal, Department of Inorganic Chemistry, Faculty of Chemistry, University of Valencia, C/Dr. Moliner 50,461oO Bmjassot (Valencia), Spain. Journal of Inorganic Biochemistry, 40, 245-253 (1990) @ 1990 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010

0162-0134/90~$3~

246

J. M. Momtal et al.

active form of the enzyme [ 1, 7, 11, 121. Presently it is agreed that the enzymatic properties of Carbonic Anhydrase depend on more than one acidic group [l , 13-151 as do the spectroscopic properties of CoCA [ll, 16, 171. One acidic group is the metal coordinated water molecule and it has been proposed that His-64 (or His-208 in HCA I) will be the second ionizable group [ 181. Bertini et al. [19] have analyzed the acid-base properties of CoBCA II and CoHCA I through electronic absorption spectroscopy in terms of deprotonation of two interacting acidic groups. Carbonic Anhydrase is reversibly inhibited by a large number of simple anions. The most potent inhibitors are typical metal binding agents such as CN- and SH- , but many anions that do not normally bind to metal ions inhibit, for example, NO3 and C104-. The optical spectra of the Co(I1) enzyme suggests that different anions may bind in different ways, either displacing metal-bound water forming a four-coordinate complex or not displacing water and forming a five-coordinate complex. Five-coordinated complexes have been associated with low intensity spectra (emax< 200 M-l cm-‘) whereas four-coordinated complexes have been associated . . . with htgh mtensity spectra (emax> 300 M-t cm-‘) [ll, 201. Anionic inhibitors have been shown through activity measurements to have pH dependent affinity for the native enzyme, the affinity decreasing with increasing pH. Also, many kinetics studies have been made in the presence of anions, particularly sulphate, in order to maintain the ionic strength constant. However, the non-innocent behavior of sulphate has been confirmed [21,22]. Recently, the effect of the sulphate anion on the kinetics parameters of CQ hydration and HCOs - dehydration reactions catalyzed by carbonic anhydrase, has been examined with a stopped-flow technique [23]. From the obtained results, it is proposed that sulphate anion binds to the enzyme in a mode that inhibits both hydrase and hydrolase activities of carbonic anhydrase II. Co(I1) represents an excellent ‘H NMR spectroscopic probe for the elucidation of the structural features of the active site of zinc enzymes [24,25]. Recent ‘H NMR studies on CoCA and their adducts with several anionic inhibitors [26, 271 have allowed extensive characterization of the chemical properties of the active site and to rationalize the factors determinin g the coordination geometry of the adducts. In order to understand the binding properties of the anions, we have measured the affinity constants of sulphate and chloride, registered the ‘H NMR spectra of their adducts with CoBCA II, and determined the ‘H longitudinal relaxation times values of the coordinated histidines.

EXPERIMENTAL Reagents Bovine carbonic anhydrase II was purchased from Sigma Chemical Co. and purified through chromatography on DEAB cellulose [28]. The cobalt(H) derivative was prepared through zinc removal by dialysis against solutions of 2,6dipicolinic acid 0.01 M at pH = 7.0 [29] and spectrophotometric titrations of the apoenzyme with a cobalt(I1) sulphate solution. CoBCA II concentrations were determined by measuring the absorbance at both 280 nm and 550 nm (~280= 5.6 x 104 and es50 = 290 at pH 6.0), the agreement between the two values was always within 10%. CoSO4 .7H@, NaNs, NaC104 . H20, RN03 , and all the other chemicals were Merck reagents analytical grade. The 99.7% D20 was obtained from the Fluka Chemical Co.

248 J. M. Momtal et al.

FIGURE 1. Blectronic spectra in the visible region of 600

700

A(nml

CoBCA II alone (-), and of their adducts with sulphate (--) aad chloride (-*-) at pH = 6.1 (MES 10 mM).

adducts with sulphate and chloride. The ‘H NMR spectrum of an aqueous solution of CoBCA II (pH = 6.1, 293 K) shows, in the downfield region, three resolved paramagnetically shifted signals at a, 65.3, b,a63.7, and c, d, 56.3 ppm, the latter being of intensity two (Fig. 2). These signals account for the four protons that are in a meta-like position with respect to the metal ion. Three of them (a, c, d) are exchangeable and are assigned to the NH protons of the three-coordinated histidines while the non-exchangeable one is assigned to the H62 of His-l 19 (Fig. 2, inset). The NH of the same histidine b has been assigned on the basis of the field dependence of the line width [26] and NOE experiments [31]. Some broad lines are observed due probably to or&-lii protons of coordinated histidines. The chemical shifts values of these signals are given in Table 1. Near the diamagnetic region we can observe one exchangeable signal at 15.9 ppm that could belong to residues not directly coordinated to the pammagnetic metal ion but close enough to experience dipolar in&actions with the unpaired electrons and then some dipolar shift [25]. This signal is missing from the spectrum of ZnBCA II. When adding sulphate, signal a moves a few ppm’s upfield whereas the signals c, d moves downfield and the position of signal b moves only slightly ( < 1 ppm) upfield. At saturating conditions of sulphate (Fig. 2) these signals collapse and the ‘H NMR spectrum of the adduct shows, in the downfield region, only one signal at 62.8 ppm. Likewise, chemical shifts of the broad lines are also changed when the sulphate adduct is formed (see Table l), and some new sharp signals, near the diamagnetic region, are observed in the ‘H NMR spectrum. The changes in shifts of the me&H signals as a function of the anion concentration are shown in Figure 2. The affinity constant estimated from this plot, 80 f 10 M-l, is consistent with the above indicated value obtained from spectral titration. The longitudinal relaxation time Tt of the signal at 62.8 ppm has been measured and the value obtained was 6.1 ms. The ‘H NMR spectra of the CoBCA II, as well as of the adduct with chloride, (pH = 6.1, 300 K) are plotted in Figure 3. When temperature is increased, only minor changes are observed in the ‘H NMR spectrum of CoBCA II. The signals corresponding to the four meta-like protons move some ppm upfield (see Table 1). The ‘H NMR spectrum of the adduct with chloride shows three resolved signals at 69.3, 66.6, and 62.0 ppm, the latter being of intensity two (Fig. 3). These signals corresponding to the four meta-like protons provide evidence that the three his&lines are still coordinated in the adduct. In fact, when the spectrum of the adduct is registered in &O, the signals at 69.3 and 66.6 ppm disappear completely whereas the signal at 62.0 ppm decreases in intensity down to about one-half, in according to the presence of three exchangeable NH protons and one H62 proton. From a ‘H NMR titration of CoBCA II-with inhibitor, the three signals at 69.3, 66.6, and 62.0 ppm

COBALT(II)-CARBONIC

i

30

60.

9b

I

I

I

I

I

I

30

ANHYDRASE

-30

0

I

I

I

249

0

I

I

6 (ppm)

T

-30

a

r

80-60

I

I

I

I

40 d (pPrr3

FIGURE 2. 200 MHz NMR spectra of CoBCA II (A) and of its dduct with sulphate (B). The variations of chemical shifts of signals a, b, c, and d vs sulpbateconccntrath a~ also reportedtogether with the best fitting curve. Solution conditions are 2.3 mM CoBCA, pH = 6.1 (MES 10 mM) and 293 K.

250 J. M. Momtal et al.

TABLE 1. Chemical Shifts and T, (as) Values of ‘H NMR Resonances for Bovine. Co@)-SuWtutcd Carbonic Anhydrase and Their Sulphate and Chloride Adducts

Signal

(TI)’

CoBCAb

b

CoBCA’ 63.7 65.3

62.8

6.1

60.6 62.1

62.0 66.6

6.0 5.9

: ;

56.3 48.5 40

62.8 -52.4

6.1

52.8 47.2 40

69.3 62.0 40 52.2

4.0 6.0

15.9 - 10 - 19

16.7 - 16.5 -26.5

15.1 -8.7 - 16.3

16.7 -5.6 - 12.4

a b

i

CoBCA(S0,‘)’

CoBCA (C1-)b

(T,)b

‘293 K. b300K.

respectively are related to signals c, a and b, d, respectively, of CoBCA (see Fig. 3). The atIinity constant estimated from this plot, b = 27 f 5 M-l, is consistent with the above indicated value obtained from spectral titraticq The T1 obtained values for the signals are indicated in Table 1. The observed isotropic shifts in paramagnetic complexes can be contact or dipolar in origin 1251. The contact term varies as T-’ and the dipolar term varie5 as Ts2, so the temperature dependence of the isotropic sbifks can be used to estimate the relative importance of these contributions. However, due to the small temperature range usually investigated, the shifts are typically inversely proportional to temperature. The usual procedure for verifying a non-T-’ dependence of the shift5 is to extrapolate the plot to infinite temperatures. FIGURE 3. 200 MHz NMR spectra of CoBCA II (A) and of it6 adduct with chloride (B) at 300 K. The variation6of chemical &ifts

60

I

90

I

I

I

60

30

I

I

I,

30

0

I,

-30

I,

0

I

-30 d (Ppn)

ofsignalsr,b,c,anddvschlorideconcentration are. also reporkd together with the best fit@ curve. Solution conditions as in Figure

m

COBALT(II)CARB

ANHYDRA!SE Wl

-25 -

-751

= 0.001

0.002

* 0.003

.

' 0.004

T-’ (K -’

)

FIGURE4.

Temperrtun-of~

i8fmpic sbifb fw

COBCA athbcts CoBCA (So3 and (b) CoBCA (Cl-).

(a)

Variable temperature ‘H NMR spectra of CoBCA(SO4’) and CoBCA (Cl-) were recorded~n5oCaad3ooC.InFigure4theobsemdisotropicshiftsofthe meMike protons are plotted vs T-’ for both adducts. The chemical shift of the only observed signal of the sulphate adduct is very tempemtuedependent and a linear relationship is found which extrapolates through the diamnpnetic region. In a similar way the chemical shifts for two proton signals of the chloride adduct obey an equation of the type 6 = arT_‘. However, for the other two proton signals the interceptsatinfinitetemperaauewere~~,asobsemdinFigure4.Theseresults areindicativethattheshiftinthesulphateadductlnustbecontectinorigin,whereas some dipolar contribution to the isotropic shifts is deteded in the chloride addud. It has been observed that sulphate inhibits the enqmatic activity of Carbonic AnhydraseatlawpHandthee&ctwasaegligibleatpH>7.Aswith~anionic inhibitors, sulphate is competitive with respect to bicahonate and shows noncompetitive behavior with respect to C!&. In contrast, at low ionic streqth the inhibitory effects can be interpreted in terms of the electrosMc effects; at higher saltconcencrationssulphateinhibitstheenzymebybindingtotheactivecailtyandit has been suggested that sulphate binds to metal ion [22,23]. Ram NMR relaxation studiesonCoBcAinthepnsenceofsulphatesO~*-maybindtoC0(n)at1awpH [21, 32, 331. ThemetalionadivecavityofCAisvery~~,iathesensethatitbiradsmost of the molecules which in coodhtion chemdry are known as ligands. Thus, anions

252

J. M. Mot&al

et al.

that are very poor coordinating agents for metal ions in aqueous solutions, like N@ or Cl04 -, bind to cobalt(I1) in CoBCA. Sulphate and chloride interact weakly with CoBCA II, and this interaction only occurs at low pH because these anions only show sign&ant afRnity for the diprotonated form of CoBCA (Co-OH2 . . His-H+). ‘H NMR spectra of the adducts prove that the histidines remain coordinated upon binding of inhibitors, however, a spatial rearrangement of coordinated ligands takes place when the anion binds to a metal ion. According to the spectroscopic criterion [l 1, 201 the molar absorbance of the sulphate adduct (E- = 250 M-* cm-‘) suggests the presence of some fraction of the five-coordinate species whereas the chloride adduct will be essentially tetracoordinated (emax = 300 M-l cm-’ at 300 K). However, the presence of a weak temperature-dependent band at 13.900 cm-’ in this latter adduct has been interpreted in terms of an equilibrium between four- and five-coordinated species, the former species being predominant higher temperatures [21]. The Ti values of the metal-like protons provide a further criterion to assign the coordination number of CoCA derivatives. Thus, Ti values of 8-20 ms for meta-like protons have been associated with pentacoordination of the metal whereas values of 3-5 ms have been associated with tetracoordination [27,34]. The chloride complex displays Ti values of 4-6 ms for the meta-like protons whereas the sulphate complex displays a Ti average value of 6.1 ms. These Ti values are consistent with an essentially tetracoordinate geometry for Co(H) in both adducts. Indeed, the temperature dependence of the isotropic shifts for the sulphate adduct indicates that the isotropic shift is essentially contact in origin, as expected for a pseudotetrahedral geometry around the metal ion. Finally, in a support of the previously suggested [21] equilibrium between four- and five-coordinated species for the chloride adduct, we have detected a sizeable dipolar contribution to the meta-like proton shifts in this adduct. We thank the CICYT (Ministerio de Edumcidn y Ciencia, Spain) for its financial support of this work (Ptvyecto No. PB 88-0489). We are gmteful to Pmfmor I. Bertini at Florence University for many helpfu! comments.

REFERENCES 1. S. Lindskog, in Zinc Enzymes, Metal Ions in Biology, T. G. Spiro, Ed., Wiley Interscience, New York, 1983, Vol. 5, p. 78. 2. D: N. Silverman and S. Lindskcg, Act. Chem. Res. 21, 30-36 (1988). 3. K. K. Kannan, B. Notstrand, K. Fridborg, S. Loevgren, A. Holsson, and M. Petif, Pmt. Natl. Acad. Sci. USA 72, 51 (1975). 4. K, K. Kannan, in Biophysics and physioiogy of Carbon Dioxide, C. Bauer, G. Gros, and H. Bartels, Eds., Springer-Verlag, Berlin, 1980, p. 184. 5. A. Liljas, K. K. Kannan,P. C. Bergsten, I. Waara, K. Pridborg, B. Strandberg, U. Carlbom, L. Jaerup, S. Loevgren, and M. Petef, Natum (London) 235, 131 (1972). 6. K. K. Kannan, M. -am, and T. A. Jones, Ann. N. Y. Acad. Sci. 429, 49-60 (1984). 7. E. A. Eriksson, T. A. Jones, and A. Liljas, in Zinc Enzymes, I. Berthi, C. Luchiit, W. Maret, and M. Zeppezauer, Eds., Birkhauser, Boston, 1986, Chap. 23. 8. J. M. Giulian, N. Limozin, M. Charrel, G. Laurent, and Y. Derrien, C.R. Acad. Sci., Ser. D. 278, 1123 (1974).

COBALT(II)-CARBONIC ANHYDRASE

253

9. J. M. Giulian, N. Limozin, B. Mallet, J. Di Costanzo, and M. Chad, Biuchimie 59, 293 (1977). 10. V. Kumar, K. Sankann, and K. K. Kannan, J. Mol. Biol. 190, 129 (1986). 11. I. Beztid and C. Luchat, Act. Chem. Res. 16,272 (1983). 12. S. Lii and J. E. Coleman, Pnx. Nati. Aaad. Sci. USA 70, 2502 (1973). 13. S. Lindskog, S. A. lbrahim, B. H. Jonsson, and I. Sinsson, in The Coordination Chemistry of Metalloenqymes, I. Be&i, R. S. Drago, and C. Luchinat, Eds., D. R&de1 Publish@ Co., Dodedlt, Holland, 1982, p. 49.

S. Lindskog, Adv. Znorg. Biochem. 4, 115 (1982). D. N. Silverman and B. H. Vicent, CRC Crit. Rev. Biochem. 14,207 (1983). I. Bertini, C. Lwhinat, and A. Scozzafava, Ino@. Chim. Acta 46, 85 (1980). I. Berth& C. Luchinat, and A. Scozzafava, Struct. Bonding (Berlin) 48, 45 (1982). I. Sinmsson and S. Lindskog, Ear. J. B&hem. 123,29 (1982). 19. I. Bextini, A. Dei, C. Lwhinat, and R. Monnanni, Znorg. Chem. 24, 301 (1985). 20. R. C. Rosenberg, C. A. Root, R. H. Wang, M. Cerdonio, and H. B. Gray, Pm. Natl. 14. 15. 16. 17. 18.

Acad. Sci. USA 70, 161 (1973). 21. I. Bertini, G. Canti, C. Luchinat, and A. Scozzafava, J. Am. Chem. Sot. 100, 4873 (1978).

22. S. H. Keening, R. D. Brown, and G. S. Jacob, in Biophysics and Physiology of Carbon Dioxide, C. Bauer, G. Gras, and H. Bartels, Eds. Springer-Verlag, Berlin, 1980, p. 238. 23. Y. Packer and C. H. Miao, Biochem&y 26, 8481 (1987). 24. I. Bertini and C. Luchinat, Adv. Znorg. Biochem. 6, 71 (1984). 25. I. Bertini and C. Luchinat, NMR of Pammagnetic Molecules in Biological Systems, Benjamin Cummings, Boston, 1986. 26. L. Banci, I. Bertini, C. Luchinat, R. Monnanni, and J. Moratal, Gau. Chim. Ital. 119, 23 (1989). 27. L. Banci, I. Berthi, C. Luchinat, A. Donaire, M. J. Martinez, and J. M. Moratal, &mm. Znorg. Chem. 9(5), 245 (1990). 28. S. Lindskog, Biochim. Biophys. Acta 39, 218 (1960). 29. J. B. Hunt, M. J. Rke, and C. B. Storm, Anal. Biochem. 79, 614 (1977). 30. J. Hodmum and H. Kellerhals, J. Magn. Reson. 38, 23 (1980). 31. G. N. La Mar (submitted for publication). 32. I. Beti, G. Canti, C. hchinat, and A. Scozzafava, B&hem. Biophys. Res. Commun. 78, 158 (1977). 33. M. E. Fabry, J. Biol. Chem. 253, 3568 (1978).

34. I. hfi, C. Luchi~t, and R. Monmumi, in Carbon Dioxide as a Source of Carbn, M. Aresta and G. Forti, Ids., D. Riedel Publishing Company, 1987, p. 161. Race&d February 28, 1990; acceptadMarch 21, 1990