THE JOURNAL cm BIO~GICAL CHEMISTRY Vol. 247, No. 9, Issue of May 10, pp. 2900-2904, 1972

Printed in U.S.A.

Some Effects of Calcium Ions on the Structure of

Bovine Pancreatic Deoxyribonuclease A*

(Received for publication, December 6, 1971)

THOMAS L. POULOS AND PAUL A. PRICE

From the Department of Biology, University of California, San Diego, La Jolla, California 92037

SUMMARY

The effects of divalent metal ions on the circular dichroism (CD) and optical rotatory dispersion (ORD) spectra of bovine pancreatic deoxyribonuclease have been studied. Both the CD and ORD spectra of DNase undergo significant changes below 250 nm upon the binding of calcium to the enzyme. The maximum increase in ellipticity at the 215 nm minimum in the CD spectrum is 950 deg cm2 per decimole. Fifty per cent of the maximum transition occurs near 10V5 M CaC12. MnC12 gives a smaller increase at 215 nm of 290 deg cm2 per de&role while MgClt causes no change.

Calcium is able to protect DNase completely against in- activation by trypsin. Half-maximum protection is achieved near 1.3 X lO+M CaCL. Although some protection against trypsin inactivation is given by MnCL, and less by MgCL, at no concentration is either metal able to protect completely. Calcium-induced changes in the CD spectrum and in suscepti- bility to trypsin must not involve any large changes in shape or volume of the protein since no change in the sedimentation coefficient could be observed.

Concentrations of 10m4 M to 10e2 M CaClz cause intensifica- tion of the CD bands between 310 nm and 250 nm. The same calcium concentration range induces a large ultraviolet difference spectrum in DNase. This indicates that calcium causes significant perturbations of tryptophan and tyrosine residues in DNase.

It has been established that bovine pancreatic deoxyribonu- clease requires divalent metal ions to hydrolyze the phospho- diester bonds of DNA (l-4). The binding of metal ions to nu- cleic acids has been studied (5-7) and the DNA-metal ion complex is probably the required substrate for DNase action.

Calcium ions protect one of the two disulfide bonds from re- duction by mercaptoethanol (8), and Ca++ protects DNase from inactivation by trypsin and chymotrypsin (9). Calcium also alters the extent of reaction and degree of inactivation when the c-amino groups of the 9 lysine residues are modified with trini- trobenzene sulfonic acid or potassium cyanate (10). Calcium causes a red shift in the ultraviolet adsorption spectrum of DNase

* This research was supported in part by United States Public Health Service Research Grant GM-16603.

indicating a calcium-induced perturbation of tryptophan and tyrosine residues (11). Divalent metal ions are necessary for alkylation of the active site histidine with iodoacetate (12).

Divalent metal ions also affect the mechanism by which DNase cleaves DNA. Douvas and Price’ have demonstrated that when calcium is bound to DNase, the enzyme can cleave both strands of DNA in one encounter. When magnesium alone is present, DNase cleaves exclusively by breaking only one strand of DNA per encounter. These several effects of calcium ions suggest that the binding of calcium to DNase induces some structural changes in the enzyme. We therefore wished to investigate the effects of divalent metal ions on the structure of DNase with the use of circular dichroism as our main investigative tool.

EXPERIMENTAL PROCEDURE

M&%&-Worthington DP grade DNase was treated with diisopropyl fluorophosphate to inactivate contaminant prote- olytic enzymes and purified over a column of phosphocellulose according to the method of Salnikow, Moore, and Stein (13). Four chromatographically distinct components of DNase emerge, and the last and largest peak, designated DNase A, was used in these studies. DNase A solutions were concentrated routinely by ultrafiltration. At no time was the protein concentrated by lyophilization. Worthington trypsin treated with L-l-tosyl- amido-2-phenylethyl chloromethyl ketone to inactivate any contaminating chymotrypsin was used without further purifi- cation.

DNase Assay-DNase activity was determined by the hyper- chromicity assay of Kunitz (14) as modified by Price et al. (9) against a Mn++-DNA substrate at pH 5.0.

Optical Rotatory Diversion and Circular Dichroism-Circular dichroism spectra were measured at room temperature in the J-10 modification of the Durrum Jasco ORD/UV 5 spectro- polarimeter. Optical rotatory dispersion spectra were measured in a Durrum Jasco ORD/UV 5 instrument at room temperature. A O.lO-cm cell was used for CD2 and ORD measurements below 250 nm, and a l.O-cm cell was used for CD measurements above 250 nm. The reproducibility of both spectra and baseline were checked repeatedly.

In preparation for ORD and CD runs, DNase was either de- salted at room temperature over a Sephadex G-25 column equil- ibrated with the desired buffer or dialyzed against the desired

3 A. Douvas and P. 9. Price, manuscript in preparation. 2 The abbreviations used are: CD, circular dichroism; ORD,

optical rotatory dispersion.

2900

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from

Issue of May 10, 1972 T. L. Poulos and P. A. Price 2901

buffer at 5”. The buffers used were 1 rnM HCl for pH 3.0, 5 mM sodium phosphate for pH 5.5 and pH 6.5, and 5 mu ionic strength Tris-HCl for pH 7.5 and pH 8.5. To compare the ORD or CD of DNase with and without metal ions, a sample of DNase was divided into two equal volumes. To one was added a small volume (5 to 20 ~1) of the divalent metal ion solution. Maximum volume changes ranged from 0.2y0 to 0.5%, thus making errors in calculating ellipticity and specific rotation changes due to protein concentration differences negligible.

Ultraviolet Difference Spectra-Difference spectra were run on a Cary model 14 spectrophotometer. Two ml of 1.45 mg per ml DNase in 5 mM ionic strength Tris-HCl buffer, pH 7.5, was de- livered into each of two, l&cm path length cuvettes. After establishing a protein versus protein baseline, a small volume of CaCIZ solution was added to one cuvette and an equal volume of water to the reference cuvette. After allowing several minutes for equilibration, the difference spectrum was recorded.

Ultracentrifugation-Ultracentrifuge runs were done at 20” in 0.1 M Tris-HCI buffer, pH 8.5, in a Spinco model E ultracen- trifuge. The protein concentration was 4 to 6 mg per ml. Runs were made at 60,000 rpm. Two, 12-mm aluminum centerpiece cells were used with a 1” quartz wedge in one cell enabling the measurement of two schlieren patterns on the same photographic plate. Each cell contained DNase solutions of identical con- centration and pH except one cell contained CaClz and the other cell a sufficient amount of NaCl to keep the ionic strength con- stant. Differences in the sedimentation coefficient due to cal- cium binding could be detected in a single ultracentrifuge run with a maximum error of 1 to 2%. This error occurs in the measurement of the photographic plates.

RESULTS

Optical Rotatory Dispersion and Circular Dichroism Below 250

nm-The ORD spectrum of DNase with and without calcium is shown in Fig. 1. The minimum at 229 nm in the absence of calcium agrees with the previously published results of Cheng (15) who observed a negative minimum at 228 f 1 nm. When calcium is present the negative shoulder at 238 nm increases and becomes a definite negative minimum while the specific rotation at 229 nm decreases. Timasheff and Bernardi also have observed a double negative minimum in the ORD spectrum of DNase (16), suggesting that their samples of DNase may have contained cal- cium. Price has observed identical calcium-induced changes in the ORD spectrum of DNase with a Cary model 60 spectropolar- imeter.

The calcium-induced ORD change occurs in the spectral region where the Cotton effects attributed to the asymmetry of the peptide backbone reach their maximum negative value due to the presence of such secondary structure as the /3 pleated sheet and the a helix (17). Since an ORD spectrum change occurs in this region when calcium binds to DNase, one expects a correspond- ing change in the CD spectrum at wave lengths characteristic of peptide bond absorption. This is substantiated by the CD spectrum of DNase with and without calcium, shown in Fig. 2. The increase in the ellipticity of the peptide bond transition at the minimum indicates a significant perturbation of the asym- metry of the peptide backbone and suggests that calcium binding results in a more ordered structure of the polypeptide backbone of DNase. The CD spectrum of DNase is also characteristic

3 P. A. Price, unpublished results.

210 220 230 240 250 ii0

WAVELENGTH [ nm)

FIG. 1. Optical rotatory dispersion of DNase with and without Ca++. DNase (0.88 mg per ml); 5 mM ionic strength Tris-HCI, pH 7.5.

6

-8 c I I I , I I

190 200 210 220 230 240 250

WAVELENGTH inm 1

FIG. 2. Circular dichroism of DNase with and without Ca++. DNase (0.15 mg per ml); 5 mM ionic strength Tris-HCI, pH 7.5.

for a protein containing about equal amounts of /3 secondary structure and random coil with little a! helix (15, 16). The data in Table I show that as the pH is decreased, the change in the CD spectrum induced by lo+ M CaClz is eliminated. This would be expected if the affinity of the enzyme for binding cal- cium were greatly reduced when the amino acid side chains which bind calcium are protonated.

As can be seen in Table II, MgClt has no effect on the CD spectrum while the effect of MnClz is 70% smaller than that due to CaC12. A mixture of lop4 M CaClz and 0.05 M AkhC& or 0.05 M MgCL gives the maximum increase of 950 deg cm* per decimole. Thus, a 500-fold excess of manganese or magnesium over calcium cannot compete for the calcium-binding site responsible for the maximum change in ellipticity at 215 nm. This suggests that the binding site responsible for the CD change has at least a 500- fold greater affinity for calcium ions than either manganese or magnesium.

The effect of CaClz concentration on the change in ellipticity

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from

2902 Effects of Ca++ on Structure of DNase Vol. 247, No. 9

TABLE I Change in ellipticity at 215 nm as function of pH

The buffers used were described under “Experimental Pro- cedure.” DNase concentrations ranged from 0.12 to 0.16 mg per ml.

l -.- .--- 0.01 M CaCI;!

CaCIz concentration PH Change in ellipticity

at 215 nm in deg cm2 per decimole

M

10-s 3.0 0 5 x 10-Z 5.5 +725 2 x 10-a 6.0 +743

1.25 X 1OV 7.5 +95oa 1.25 X 1OF 8.0 +gm

0 The maximum increase of 950 deg cm2 per decimole is the average of several different experiments.

TIME OF REACTION (minutes)

FIG. 3. Proteolytic inactivation of DNase as a function of metal ion type. Tris-HCI (0.05 M), pH 7.9; 0.4 mg per ml DNase; 6.1 pg per ml trypsin; 25”.

TABLE II I I

Effect of MgClz and MnClz on ellipticity at 216 nm

The buffer used was 5 mM ionic strength Tris-HCl, pH 7.5. The DNase concentration was 0.16 mg per ml.

Metal ion type

1.0 or 5.0 X 1OF M MgClz 1 .O or 5.0 X 1OW M MnClt 5 X 1OP M MgC12 + 10e4 M Ca&.

5 X lo+ M MnCle f 10-4 M CaClz.

1o-4 M CaCL.

- I A0 at 215 nm in deg

cm2 per decimole

0 +290 +950 f950 +950

TABLE III

Change in ellipticity at 215 nm as function of CaCl, concentration

Tris-HCI buffer, 5 mM ionic strength, pH 8.5, was used. The DNase concentration was 0.13 mg per ml.

C&I* concentration A0 at 215 nm in deg cm2 per decimole

0 0 1o-5 M +485 lo-'M +950 lo@ M +950

at 215 nm is given in Table III. Half-maximum change occurs near lo+ M Calls. Price observed that half-maximum protec- tion against trypsin inactivation occurs near 1.3 X 10e5 M CaC& (18). Fig. 3 shows the effect of MnC&, MgC12, and CaC& on the rate of inactivation due to trypsin. If one plots the loga- rithm of the percentage of remaining activity veraus time a linear plot is obtained and relative pseudo first order rate constants can be calculated. The relative rates of inactivation calculated from the slopes of such plots are 10 for 0 ions, 1.5 for magnesium, and 0.26 for manganese. Even though manganese and mag- nesium slow the rate of inactivation, only calcium protects com- pletely. Thus, calcium is the most effective of the three divalent metal ions both in protecting against trypsin and inducing the CD change. Manganese is second in both effects, and mag- nesium gives no CD change and is a poor protector against tryp- sin.

Sedimentation Velocity Runs-If the conformational alterations observed by the increase in the 215~nm minimum of the CD

TIME (min) TIME (min)

I 0 3 mM Co”

.7960

0 8 16 24

FIG. 4. Effect of Ca++ on the sedimentation coefficient of DNase. Plot of the logarithm of the distance from the center of rotation versus time. Tris-HCl (0.1 M), pH 8.5; 4.8 mg per ml DNase (a) and 5.6 mg per ml DNase (b).

spectrum involved large alterations in shape or volume of the protein, one might expect a change in the sedimentation coeffi- cient. The technique we used involved paired sedimentation velocity runs in the analytical ultracentrifuge with two cells in the same rotor, enabling the measurement of two schlieren pat- terns on the same photographic plate. Both cells contain an identical concentration of protein at the same pH and ionic strength. One cell contains CaClz and the other sufficient NaCl to maintain constancy of ionic strength in both cells.

The results from some typical runs are shown in Fig. 4. The slopes of the plots in Fig. 4a are identical indicating that calcium causes no change relative to the enzyme with sodium. The sedimentation coefficients we obtained from one experiment to another were in good agreement with the value of 2.78 obtained by Lindberg (19). As a control to test the accuracy of our meas- urements in a single experiment, paired sedimentation runs with both samples containing calcium were made (Fig. 4b). This control enabled us to estimate our maximum error to be about 1 to 2% in the measurement of the photographic plates. A different treatment of the data can be made if one plots the dif- ference in the logarithm of the distance from the center of rota- tion between the DNase samples with and without calcium against time. The slope of such a plot is a direct measure of the change in sedimentation coefficient (20), and such plots also indicate that no change in the sedimentation coefficient has

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from

Issue of May 10, 1972 T. L. Poulos and P. A. Price 2903

4ot I 10-2M Co"

1 I I 1 I , I

250 260 270 280 290 300 310 WAVELENGTH (nml

FIG. 5. Aromatic circular dichroism of DNase at 10ep M and lo+ M Ca++. DNase (1.30 mg per ml); 5 mivr ionic strength Tris-HCI, pH 7.5.

011 I I I I _ 260 270 280 290 300

WAVELENGTH (nm)

FIG. 6. Aromatic circular dichroism difference spectrum gen- erated by 1OV M Cd-+. This plot was obtained by subtracting the ellipticity of the 0 Ca++ spectrum from the 10-z M Ca++ spec- trum at 1-nm intervals.

occurred. The calcium-induced CD change, therefore, must not represent any large alterations in size and shape of the molecule.

Circular Dichroism from 310 to %50 nm-The near ultraviolet CD of DNase has been previously published (19). Fig. 5 shows the effect of calcium on the near ultraviolet CD of DNase. The CD spectrum between 310 to 250 nm is quite complex, and de- tailed interpretation would be difficult. Nevertheless, the nega- tivc and positive bands above 290 nm can be assigned to optically active tryptophans (21). Below 290 nm, the CD bands must be due to overlap of tyrosine and tryptophan bands (21, 22). DNase contains two disulfide bonds whose inherent optical ac- tivity undoubtedly contributes to the complexity (23). Despite such difficulty in interpretation, it is clear that calcium causes significant changes in the near ultraviolet CD spectrum. Fig. 6 is the near ultraviolet CD difference spectrum and shows more clearly the maximum differences between DNase with and with- out calcium. In general, calcium leads to an intensification of the CD bands with major peaks centered at 275 and 284 nm. Low temperature CD of model tryptophan, tyrosine, and phenyl- alanine compounds leads to an intensification of the CD bands

, 004

1 /10-2M Cat+

z q

002 - y/-$ 1O-4 M Co++

O r-l I

250 270 290 310 330

WAVELENGTH (nm)

FIG. 7. Ca*-induced ultraviolet difference spectrum of DNase. DNase (1.45 mg per ml); 5 mM ionic strength Tris-HCI, pH 7.5. 1

as a result of increased asymmetry due to restricted rotational and vibrational degrees of freedom (21, 22, 24). I f calcium caused the environment of the tryptophan and tyrosine residues in DNase to become more hydrophobic, the result would be re- stricted degrees of freedom. This would also lead to increased asymmetry giving the observed increased intensity of the aro- matic CD bands. Fig. 7 shows that calcium induces a large ultraviolet difference spectrum in DNase. This shift to longer wave lengths in the spectrum is characteristic of the spectral shifts observed when aromatic amino acids are exposed to a more hydrophobic environment (25). Thus, the environment of the tryptophan and the tyrosine residues in DNase is probably be- coming more nonpolar when calcium is bound to the enzyme.

DISCUSSION

Radioactive 46Ca++ has been used to study the binding of calcium ions to DNase (18). Price has shown that 2 calcium ions bind strongly to DNase at pH 7.5, with an average dissocia- tion constant of 1.4 X 10-s M. At pH 5.5 only 1 calcium ion binds strongly to DNase, with a dissociation constant of 2.2 X 10-h M. The site with high af?inity for Ca++ at pH 5.5, and one of the two sites for Ca* at pH 7.5, is specific for calcium. Even in the presence of a 500-fold excess of magnesium or manganese, a single calcium ion still binds to DNase with a dissociation con- stant of about 3.2 X 10e5 M (18).

It seems probable that the divalent metal ion-binding site responsible for the structural change observed by CD measure- ments and the protection against trypsin are associated with the calcium-specific site. The dissociation constant of the calcium- specific site, 1.4 X 10m6 M, is close to the concentration of calcium which gives half the maximum CD change, 10e5 M (Table II), and half t,he maximum protection against trypsin, 1.3 X 10-6 M (18). Neither magnesium nor manganese ions bind to the calcium-specific site (18). Neither manganese nor magnesium give the maximum change in the CD spectrum, nor do they

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from

2904 Efects of Ca++ on

protect DNase completely from inactivation by trypsin. When magnesium and manganese are present at a 500-fold excess over calcium, calcium can still bind to the calcium-specific site with high affinity (18) and calcium can still give the maximum in- crease in ellipticity at 215 nm. We therefore suggest that when calcium binds to the calcium-specific site, a conformational change results leading to a more ordered structure of the poly- peptide backbone and renders unavailable for reaction the tryp- sin-sensitive bond or bonds. Since we observed no change in sedimentation coefficient in the presence of calcium, the struc- tural change probably does not involve large alterations in the shape or volume of the protein.

Another hydrolase from t,he bovine pancreas, trypsin, is also dramatically affected by calcium (26). Changes in ORD and absorption spectra, electrophoretic mobility, stability, and ac- tivity were all observed (26). The rate of trypsin-catalyzed activation of trypsinogen is markedly increased by calcium ions (27) These structural effects of calcium ions on both pancreatic trypsin and DNase undoubtedly have physiological significance since both enzymes operate in the intestine where calcium con- centrations are high (28) and maximum stability of the proteins would be beneficial.

The spectral changes induced by calcium above 250 nm prob- ably represent different divalent metal ion-binding sites than are associated with the CD changes below 250 nm. Both the CD spectrum above 250 nm (Fig. 5) and the ultraviolet difference spectrum (Fig. 7) have not reached the maximum changes at lo-4 M calcium and continue to change up to 10e2 M calcium. The CD spectral changes below 250 nm reach saturation at lOA M calcium. It is possible that the calcium-induced changes be- tween 250 nm and 310 nm are associated with the catalytic active site. Poulos and Price (11) found that N-bromosuccinimide oxidizes a single tryptophan in DNase and abolishes enzymatic activity. Calcium could no longer induce the characteristic difference spectrum in the N-bromosuccinimide-modified enzyme (11). It is possible that the essential tryptophan is in or near the active site. I f this same tryptophan is responsible for part of the calcium-induced ultraviolet difference spectrum, then elimination of the calcium-induced difference spectrum suggests that one class of calcium-binding sites is near the essential tryp- tophan. Divalent metal ions are necessary to alkylate the active site histidine of DNase (10). Half the maximum rate of inactivation is achieved with 0.02 M manganese (10). The strict requirement for the divalent metal ion in the alkylation reaction suggested that the catalytic active site has affinity for divalent, metal ion. Therefore, it seems reasonable to suggest

Structure of DNase Vol. 247, No. 9

that at least one of the weaker divalent metal ion-binding sites represented by the complex changes in the CD spectrum above 250 nm and the ultraviolet difference spectrum is associated with the catalytic active site.

1.

2.

MIYAJI, T., AND GREENSTEIN, J. P. (1951) Arch. Biochem. Biophys. 32, 414-423

LASKOWSKI, M., BOYER, P. D., LARDY, H., AND MYRBACK, M. (Editors) (1961) The enzymes, Vol. 5, p. 123, Academic Press, New York.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

:5”. 16:

17.

SHACK, J., AND BYNUM, B. S. (1964) J. Biol. Chem. 239, 3843-3848

EICHHORN, G. L., CLARK, P., AND TARIEN, E. (1969) J. Biol. Chem. 24, 937-942

EISINGER, J., FAUZ-ESTRUP, F., AND SCHULMAN, N. G. (1965) J. Chem. Phys. 42, 43

SCHULMAN, N. G., STERNLICHT, H., AND WYLUDA, B. J. (1965) J. Chem. Ph,ys. 43, 3116

EICHHORN, G. L., AND SHIN, A. Y. (1968) J. Amer. Chem. Sot. 90, 7323

PRICE, P. A., STEIN, W. H., AND MOORE, S. (1969) J. Biol. Chem. 244, 929-932

PRICE, P. A., LIU, T.-Y., STEIN, W. H., AND MOORE, S. (1969) J. BioZ. Chem. 244, 917-923

PLAPP, B. V., MOORE, S., AND STEIN, W. H. (1971) J. BioZ. Chem. 246, 939-945

POULOS, T. L., AND PRICE, P. A. (1971) J. BioZ. Chem. 246, 4041

PRICE, P. A., MOORE, S., AND STEIN, W. H. (1969) J. BioZ. Chem. 244, 924-928

SALNII<OW, J., MOORE, S., AND STEIN, W. H. (1970) J. BioZ. Chem. 246, 5685-5690

KUNITZ, M. (1950) J. Gen. Physiol. 33, 349 CHENG, P. Y. (1965) Proc. Nat. Acad. Sci. U. S. A. 56, 1535 TIMASHEFF, S. N., AND BERNARDI, G. (1970) Arch. Biochem.

Biophys. 14, 53-58

:9”: 20. 21.

YANG, J. T. (1968) in G. D. FASMAN (Editor), PoZy a-amino acids: protein models for conformational studies, P. 239-291. Marcei Dekker, New kork -

-

PRICE. P. A. (1972) J. BioZ. Chem. 247. 2895-2899 LIND&RG, U.‘ (1967) Biochemistry 6, 335 SCHUMAKER, V. (1968) Biochemistry 7, 3427 STRICKLAND, E. H., HORWITZ, J., AND BTLLUPS, C. (1969) Rio-

chemistry 8, 3205 22. STRICKLAND, E. H., HORWITZ, J., AND BILLUPS, C. (1970) J.

Amer. Chem. Sot. 92, 2119 23. 24.

BEYCHOK, S. (1965) Proc, Nat. Acad. Sci. U. S. A. 63,999 HORWITZ, J., STRICKLAND, E. H., AND BILLUPS, C. (1969) J.

Amer. Chem. Sot. 91, 184 25. YANARI, S., AND BOVEY, F. A. (1960) J. BioZ. Chem. 236, 2818 26. SIPOS, T., AND MERKEL, J. R. (1970) Biochemistry 9, 2766 27. ABITA, J. P., DELAAGE, M., LAZDUNSKI, M., AND SAVRDA, L.

28. (1969) Eur. J. Biochem. 8, 314

KODICEK, E. (1967) Proc. Nutr. Sot. 26, 67

REFERENCES

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from

Thomas L. Poulos and Paul A. PriceDeoxyribonuclease A

Some Effects of Calcium Ions on the Structure of Bovine Pancreatic

1972, 247:2900-2904.J. Biol. Chem. 

  http://www.jbc.org/content/247/9/2900Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/247/9/2900.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on January 29, 2018http://w

ww

.jbc.org/D

ownloaded from