Further Study of the Conformation of Nuclease-(1-126) in Relation to Intrinsic Enzymatic Activity*

Nuclease-(1-126), the amino acid ordered structure of nuclease A, the H., Schechter, A. Eastlake, and C. B. Nature 251, 242-244) have demonstrated intrinsic enzymatic activity for nuclease-(1-126). To at- tempt to learn whether or not the active population of nuclease-(1-126) has the native conformation, we have examined nuclease-(1-126) with respect to enzymatic kinetics with and without the competitive inhibitor deoxythymidine 3’,5’-diphosphate (pdTp), effect of tem- perature on enzymatic activity, binding of pdTp in the presence of Ca2+ and intrinsic viscosity, Stokes radius, CD, and response to trypsin action in the presence and absence of pdTp and Caz+. The results indicate that the conformation of nuclease-(1-126) bound with pdTp in the presence of Ca2+ is partially constrained but still highly flexible below 30 “C, outside the range of ther- mal transition exhibited by the ordered elements of nuclease-(1- 126). Thus, formation or stabilization of the active site of nuclease-(1-126) (7, in contrast This may actions with residues 1 to are involved in the formation though the interactions be- residues 6 to 48 and 50-126 is not strong indicated by the binding of nuclease-T-(6-48) and nuclease-(50-126) 4

Nuclease-(1-126), although containing 89% of the amino acid sequence which folds to the ordered structure of nuclease A, is disordered and highly flexible (Taniuchi,   . To attempt to learn whether or not the active population of nuclease-(1-126) has the native conformation, we have examined nuclease-(1-126) with respect to enzymatic kinetics with and without the competitive inhibitor deoxythymidine 3',5'-diphosphate (pdTp), effect of temperature on enzymatic activity, binding of pdTp in the presence of Ca2+ and intrinsic viscosity, Stokes radius, CD, and response to trypsin action in the presence and absence of pdTp and Caz+. The results indicate that the conformation of nuclease-(1-126) bound with pdTp in the presence of Ca2+ is partially constrained but still highly flexible below 30 "C, outside the range of thermal transition exhibited by the ordered elements of nuclease-(1-126). Thus, formation or stabilization of the active site of nuclease-(1-126) by binding with ligands is not associated with cooperative folding of the entire polypeptide chain. Considering that nuclease-(1-126) does not bind to nuclease-(127-149) but does to nuclease-(111-149), the results are consistent with the idea that the specific cooperative interactions, providing extra stabilizing energy required for maintaining the polypeptide chain in the ordered state of nuclease A, may be disrupted for nuclease-(1-126) primarily due to cleavage of the peptide bond between residues 126 and 127. Then, it may be thought that binding with ligands does not compensate for this disruption.
Staphylococcal nuclease A' containing 149 residues and devoid of sulfhydryl groups and disulfide bonds (1)(2)(3) folds to a compact three-dimensional structure in which the atomic coordinates of residues 1 to 141 are defined (4)(5)(6). Removal of residues 127 to 141 from an atomic model of nuclease A causes no change in the coordinates of the remaining residues and does not yield a pocket, groove, or crevice which would allow extensive disruption of hydrophobic interactions (4)(5)(6). Residues 139 and 140 fold back toward residues 136 and 137, and residue 141 protrudes outside the structure at the bottom left * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' On the basis of the discovery of nuclease B, which contains the 19 extra amino acid residues at the NH, terminus and is a possible precursor of staphylococcal nulcease (3), the original nuclease (1,2) is redesignated as nuclease A. front comer of the model (4)(5)(6). This situation also contributes to limitation of the number of buried nonpolar residues (Val 111, Leu 108, and Val 104) which are exposed upon removal of residues 127 to 141 (4)(5)(6). Thus, the bulk of hydrophobic interactions of native nuclease A would be preserved for the native conformation of the fragment containing residues 1 to 126, n~clease-(l-126)~ (7,8).
We have further investigated the conformational properties of nuclease- , particularly in relation to intrinsic enzymatic activity. Sachs et al. (9), using antinuclease antibody directed to residues 127 to 149, have demonstrated that a low level of enzymatic activity of nuclease-( 1-126) (7,8) is intrinsic (9). Since all the residues of nuclease A involved in binding of Ca2+ (required for enzymatic activity (10)) and pdTp3 (a potent competitive inhibitor (13)) are contained in nuclease-(1-126) (4-6),4 if the active site of nuclease-(l-l26) corresponds to that of nuclease A, then these ligands would bind with nuclease-(1-126) in a manner similar to their binding with nuclease A (4-6). Our observations described below have indicated that pdTp binds with nuclease-(l-126) in the presence of Ca2+ with an apparent association constant approxi-mateIy three orders of magnitude smaller than that with nuclease A and that the conformation of liganded nuclease-(1-126) exhibits partial constraint as well as high flexibility.
Thus, although most interatomic interactions such as hydrophobic interactions and dispersion forces operating for native nuclease A, perhaps except for those requiring complete folding of residues 1 to 141 for their operation (14)(15)(16), would also be operative for nuclease-(1-126), stabilization of the active site of nuclease-(l-l26) by binding with ligands does not lead to cooperative folding of the entire polypeptide chain. The result is interpreted as indicating that specific cooperative interactions, which are not operational for nuclease-(l-126), may be operating for cooperative folding of native nuclease A and that disruption of them cannot be compensated for by binding with ligands. MATERIALS  The abbreviations used are: pdTp, deoxythymidine 3',5"diphosphate; nitrophenyl-pdTp, deoxythymidine 3"phosphate 5"p-nitrophenyl phosphate (11); Sepharose-pdTp, Sepharose 4-B (agarose) phate (12).

4557
elsewhere (3,15). Nuclease-(l-126) and other fragments were prepared from purified nuclease A by established methods as outlined in miniprint immediately following this paper.5 Nuclease-( 1-126) thus obtained has been passed through a Sepharose-pdTp column (12) in order to remove contaminating intact nuclease A (see the miniprint supplement). This nuclease-(l-126) sample was further purified to reduce the amount of contaminating nuclease-(l-105) (enzymatically inactive) to less than 3% (see the miniprint supplement). The quantities of nucleases A and B, the fragments, and the fragment complexes were determined by absorbance at 280 nm (3,16,17) unless otherwise indicated.
Details of the analytical methods, including assay of enzymatic activity, equilibrium dialysis, and measurements of intrinsic viscosity, Stokes radius, and fluorescence are also described in the miniprint supplement.

RESULTS
Homogeneity of the Nuclease-(l-l26) Preparation-The purified nuclease-(1-126) was homogeneous as judged by a symmetrical peak obtained in the elution profile by gel fitration (see the legend to to indicate an increase in the hydrodynamic volume and not aggregation of the molecules, since the value from the plot is still significantly smaller than if a dimer is formed. There has been no extensive zonal spreading or deformation of the symmetrical peak in the elution profile (indicating a monomer-dimer (or polymer) equilibrium) in the concentration range from 1 to 6 mg per ml of nuclease-(1-126), and the absorbance at 280 nm obeys Lambert-Beer law for solutions containing from 0.1 to 1.0 mg per ml of nuclease-(1-126) (see the miniprint supplement). Note that nuclease B, the structure of which consists of an ordered portion corresponding to nuclease A and 19 flexible residues attached to the NHZ terminus of the nuclease A portion (3), also shows a similar deviation in the Andrews plot ( Fig. 1). On the basis of these considerations, we assume that nuclease-(1-126) is a monomer under the conditions employed in the present studies.
Enzymatic Kznetics-The enzymatic kinetic parameters (V,,,,, K,,,, and K, (with pdTp)) of nuclease-(1-126) were determined at p H 8.8 at 24 f 1 "C in the presence of lo-' M Ca2+ with heat-denatured DNA as a substrate (13) and are presented in Table I  The values for K,,, and K I with nuclease-(l-126) are both considerably greater than those with nuclease A (Table I). As is the case with nuclease A (10,13), no enzymatic activity was detected with nuclease-(1-126) in the absence of Ca2+ (with M EDTA present). The enzymatic activity of nuclease-(1-126) did not change upon addition of nuclease-(127-149), confuming the previous report (8).

TABLE I
The enzymatic kinetic parameters of nuclease-(l-126), nuclease A, nuclease-T, and type ZZ complex with heat-denatured DNA as the substrate and pdTp as a n inhibitor in the presence of lo-' M eaZ+ a t pH 8.8 and 24 f 1 "C Enzymatic activity was measured at pH 8.8 at 24 f 1 "C according to the method of Cuatrecasas et al. (13) as described in the legend to  Enzymatic activity of nuclease-(l-126) and nuclease A (as a control) was measured as a function of temperature from 5-50 "C. Arrhenius plots of these data are shown in Fig. 3. Apparently nuclease-(l-126) exhibits a transition above 30 "C. Below 30 "C there is no apparent transition as the plots tend to lie on a straight line in a manner similar to nuclease A ( Fig. 3). At 30 "C and above, nuclease-(1-126) exhibited biphasic kinetics of enzymatic activity, that is, the velocity for enzymatic activity (the slope) increases within 1 min after starting the assay in contrast to the velocity below 30 "C which is constant. The reason for the biphasic kinetics observed above 30 "C is not clear. This second phase of enzymatic activity appeared to follow a transition in parallel with the f i t phase of the activity (Fig. 3).
Nuclease-(l-126) also exhibited enzymatic activity with the synthetic substrate nitrophenyl-pdTp (11). The results are compared with those of the control samples in Table 11. The relationship of V,,, and K, for nitrophenyl-pdTp between nuclease-(1-126) and nuclease A is simiiar to that with heatdenatured DNA (Tables I and 11 Nuclease-(l-126) 6.8 X 0.05 , Z The possible equilibrium of the active and the inactive species is ignored (see footnote of Table I).
the total concentration of pdTp; see the miniprint supplement for the details). The value obtained was 1.76 0.23 X lo3 M-'. The value with nuclease A, determined in the absence of Ca2+6 at 20 "C at pH 8.0 by the same method as a control, was 6.66 & 0.10 X lo3 M-'. This value is close to that obtained by tyrosine fluorescence titration at 25 "C according to the method of Cuatrecasas et al. (23) as described in the miniprint supplement. But it is smaller than that reported for nuclease A (2.3 X lo4 M-' at pH 8.8 at ambient temperature) (24). The redSon for this discrepancy is unknown.
It has been reported that tyrosine fluorescence of nuclease A changes upon binding with pdTp only in the presence of Ca2+ (23). However, when the concentration of pdTp was increased up to 1.4 X M, tyrosine fluorescence of nuclease A was found to decrease even in the absence of Ca" (see  Table III. 7 It is interesting to note that the intrinsic viscosity of nuclease B falls between nuclease A and nuclease-T-(50-149), the former sample representing a globular protein and the latter a disordered fragment. This may be related to the fact that the structure of nuclease B consists of an ordered portion and a flexible portion (see above). Addition of ligands did not change the intrinsic viscosity with all these control samples. On the other hand, the intrinsic viscosity of nuclease-(1-126), which fell between nuclease B and nuclease-T-  in the absence of ligands, decreased upon increasing the It is known that binding of pdTp to nuclease A (21) and nuclease-T (15) is weaker in the absence than in the presence of Ca'+. It is also shown that nuclease A is not adsorbed to Sepharose-pdTp in the absence of Ca*+ (22).
The intrinsic viscosity of nuclease-(l-126) obtained in the previous studies is somewfiat greater (9.8 cm:'/g (7)) than that obtained in the present studies, and those of nuclease A (2.4 cm'/g (14)) and nuclease-T-(6-48) (3.0 crn3/g (14)) are smaller in the previous studies than those in the present studies. The reason for this discrepancy is unknown. However, since the intrinsic viscosity (3.44 r 0.26 cm3/g) of RNase A measured as a control in the present studies agrees with those reported (27,28) in contrast with the previous studies in which a smaller value for RNase A was obtained, we assume that the present values are accurate.

Intrinsic viscosity of nucEease-(i-I26), nuclease A, nuclease B, nuclease-T-(M), and nuclease-T-(50-149) in the presence and
absence of ligands, pdTp, and ea2' at 20 "C andpH 8.0 The procedure for measurement of kinematic viscosity is briefly described in the legend to Fig. 4 and in detail in the miniprint section. The concentrations of Ca2+ and pdTp were 0.01 and 0.02 M, respectively, unless indicated otherwise. A small change in the outflow time with change in the concentration of free pdTp was corrected for (see the miniprint section). The intrinsic viscosity was calculated by the method of Tanford (25) using the partial specific volume calculated on the basis of amino acid composition (26). Where the standard error in the mean is indicated, the average of two determinations is presented. seemed to have decreased but it stiU remained greater than that of nuclease A (Table IV) Table IV) resulted in a decrease of accuracy for estimation of zonal spreading. Therefore, precise evaluation of the data (Table IV) is not possible. However, the relationship of zonal spreading of these samples in the presence of ligands may be similar to that in the absence of ligands (Table   IV). Nuclease-T-(6-48) was heterogenous and consisted of a presumably monomeric species and an aggregated species as judged by gel fitration either in the presence or absence of ligands.
Trypsin Digestion-One of the most characteristic properties of native nuclease A (3,19,32) and the noncovalent complexes of the fragments (8,19) is ligand-induced resistance against proteolysis. In order to test whether nuclease-(l-126) bound with ligands exhibits such characteristics, nuclease-( 1-126) was incubated with trypsin (approximately 1% by weight of the substrate) at pH 8 a t 25 "C in the presence and absence of 2 X M pdTp and lo-' M Ca2+. Examination of diquots of the mixture incubated for 5 and 15 min (the digestion was quenched by addition of soybean trypsin inhibitor) by twodimensional peptide mapping (33) indicated complete digestion after 5-min incubation. No difference in the yield of the ninhydrin-positive spots expected for complete digestion was observed by visual inspection in the presence and absence of The degree of zonal spreading in the absence of ligands is somewhat greater for nuclease-T-(50-149) than for nuclease-(l-126) ( Table  IV) . Since it has been shown that nuclease-T-(50-149) is homogenous by ultracentrifugation (19), this greater spreading is considered not to be due to the presence of aggregated species but to a difference in the conformational distribution between the two species.

TABLE IV
Relative elution position and zonal spreading by gel filtration of nuclease-(l-126), nuclease A, nuclease B, and nuclease-T-  in the presence and absence of ligands at 6-8 "C at p H 8 The procedure for gel filtration is described in the legend to Fig. 1. The elution volume ( V,) as measured with the fraction exhibiting the maximum absorbance was reproducible within 1 ml. For gel fdtration in the presence of ligands, the buffer containing given concentrations of pdTp and 0.01 M Ca2' was used to equilibrate and elute the column. In this case, the elution of the sample was monitored by measuring the protein concentration by the method of Lowry et al. (30) or amino acid analysis after desalting (see 'measurement of viscosity" in the miniprint section) or a combination of these two methods. The spreading is presented as the width (A&) at a half-height of the peak in the elution profile, measured by the difference in Kd between the two plots, respectively, lying on the ascending and the descending slope of the profile at the half-height. Where no error is indicated, only one determination has been made. In this case, experimental error for Kd is assumed to be 2.5%. Where error i s indicated, the average value of two to four determinations is presented for both Kd and A&. In the presence of both pdTp and Ca'+, the values for a AKd are only qualitative.  (8). If this secondary structure corresponds to a completely folded population which is in equilibrium with disordered populations, this folded population (with a theoretical heIical content of 27% (4-6))' would be approximately 19% of the total population. If this folded population binds with pdTp with the same dissociation constant M) as that with nuclease A in the presence of M Ca", addition of ligands would have shifted the equilibrium in favor of the folded population, resulting,in a n increase in the helical content. Being consistent with the previous observations with ORD, in the present study no change in the ellipticity a t 220 nm was observed with 1.4 X lo-' M nuclease-(l-126) upon Nuclease-T-(50-149) contains all the residues involved in the helical structure of nuclease A (residues 55 to 67, 99 to 106, and 122 to 133 (4-6)). However, the helical structure is not detectable for nuclease-T-(50-149) on the basis of ORD at 233 nm measured from 10-60 "C and CD (7,14,19), in contrast with nuclease-(l-l26) which exhibits a low level of helical content. This may suggest that interactions with residues 1 to 48 are involved in the formation of the helical structure for nuclease-(I-l26), though the interactions between residues 6 to 48 and 50-126 is not strong as indicated by the lack of binding of nuclease-T-(6-48) and nuclease-(50-126) at 4  at 25 "C at pH 8.0 (Fig. 5). If  This should have resulted in the increase in the helical content from 5 to 12%, detectable by the present measurement. Thus, these observations exclude the assumption made above. Then, three alternative possibilities remain. One is that all the populations of nuclease-(l-126) exhibit a low helical content; a second is that the species containing the helical structure (at a lower content than that expected for the native conformation) bind to pdTp with a dissociation constant much greater than nuclease A , and a third is that both the helical and nonhelical species exhibit the same dissociation constant with pdTp.

DISCUSSION
The present studies show that a low level of enzymatic activity is clearly associated with nuclease-(l-126) under the conditions of gel filtration which partially separate nuclease-(1-126) and native nuclease A (see Fig. 2 of the miniprint supplement). This c o n f m s t h e earlier conclusion (9). Furthermore, the present studies have shown that the enzymatic activity of nuclease-(l-126) requires Ca2+ and is apparently competitively inhibited by pdTp. A synthetic substrate, nitrophenyl-pdTp ( l l ) , is also found to serve as a substrate for nuclease-(l-l26). These properties are characteristic of the activity of nuclease A (11,13). This indicates that the active site of nuclease-(l-126) resembles that of nuclease A (4-6). However, some differences in enzymatic properties between the two species were observed with respect to optimum pH, the optimum concentration of Ca", V,,,.,, K,, and KI. A low content of helical structure has also been indicated previously for nulcease-(l-126) (see under "Results") (7,8). This helical  (2) were calculated on the basis of average residues weights as described previously (7). For measurement of CD in the presence of ligands, a given volume (10 to 20 ~1 ) of known concentration of CaC12 and pdTp was added to the nuclease-(l-l26) solution structure has exhibited a heat-induced transition above 30 "C with a midpoint around 48 "C (7,8).
Although nuclease-(1-126) exhibits these elements of an ordered structure, the previous physicochemical and immunological studies (7,8) as well as the present measurements of intrinsic viscosity at 20 "C and Stokes radius and the zonal spreading at 6-8 "C have indicated a disordered conformation for nulcease-(l-126). The high flexibility of the conformation has also been inferred by the observations that nuclease-(1-126) readily interacts even at 6 "C with an overlapping fragment nuclease-T-(50-149) (within 1-2 min (35)) to simultaneously form two alternative complementing structures, type I and I1 (8,16)." (Available evidence has indicated that unless the fragment is unfolded, it would not interact with a second complementing fragment to form an ordered complex (15,16).) Thus, nuclease-(1-126) characteristically exhibits two different aspects, ordered elements (constrained) and a highly flexible nature.
In order to analyze the present data, two alternative models were considered as limiting cases for interaction of nuclease-(1-126) with ligands or substrates. In the first model the 10 In type I complex, like nuclease-T, residues 1 to 48 of nuclease-  populations of nuclease-(1-126) molecules are distributed in a set of conformations which are in equilibrium with each other." Virtually all these populations may interact with pdTp or a substrate (in the presence of Ca2+), resulting in formation of the active site as schematically presented in Fig.  6. Here, the conformations (bl, bp, etc., Fig. 6) assumed by liganded nuclease-(1-126) are not supposed to exist before binding with ligands. The equations in Fig. 6 may be collectively represented by equation 1. (1) where A and B are the set of 81, 82, etc. and of bl, bz, etc., respectively.
In the second model (equation 2), an enzymatically active population ( E ) , in equilibrium with inactive population (D), would bind with ligands I' Such a conformational equilibrium process is assumed to be fast since the reaction of disordered nuclease-T-(50-149) and nuclease-(99-149) with antinuclease-antibody is apparently complete within 5 min (36). The following points are also to be considered. If a polypeptide chain with, say, 1 0 0 residues is disordered and assumes many conformations which are in equilibrium with each other under a set of conditions, the number of conformations in this ensemble must be limited (14). Otherwise, the equilibrium state would never be attained in a period of time realistic for the experiment (37). This limitation of the conformations is presumably established by interatomic interactions of the system (e.g. dispersion forces and hydrophobic interactions). It would follow that if some dynamic events of the system permit a disordered conformation to reach the native conformation in a period of time realistic for the biological system (37)(38)(39)(40), such dynamic events would also be required for transformation from one disordered conformation to another in the equilibrium ensemble. Therefore, it is not clear at the present time whether such dynamic events considered for the earlier phase of the folding process are specific for the conformational transformation to the native structure. Three independent measurements, enzymatic kinetics, equilibrium dialysis, and intrinsic viscosity, give information for K, of nuclease-(1-126) with pdTp in the presence of 0.01 M Ca2+. The two models described in the text result in mathematically different forms or derivations for K,. But (Table 111) (see the miniprint section for details.)

Measurements
Mathematical forms Observed

Model I Model I1
Enzymatic kinetics 4.6 x lo3 M" at 24 f 1 "C, pH 8. The first and the second models lead to different forms of the Lineweaver-Burk equation, as shown in the miniprint supplement. However, the term l/Vmax is the same for both the models. Therefore, the very low value for VmaX indicates that the active site of nuclease-(l-l26) is not exactly the same as that of nuclease A whatever the active species may be.
The three independent measurements (enzymatic kinetics, equilibrium dialysis, and intrinsic viscosity) have provided information on the apparent association constant (K,) of nuclease-(l-126) with pdTp in the presence of Ca'+. In theory the two models give numerically identical values for K , for each of these measurements (Table V). Therefore, these measurements cannot differentiate the two models." Nonetheless, if all these measurements are related to binding of pdTp with the active site, the observed values for K, should be consistent regardless of which model might be the case. Indeed, the three measurements have resulted in qualitatively consistent values for K, as summarized in Table V. On the basis of these results, we conclude that under the conditions employed, the properties of nuclease-(l-126) observed in the presence of 2 X lo-' M pdTp and lo-' M Ca2+ represent liganded nuclease-(l-l26) as follows: 1) the conformational constraint for nuclease-(l-l26) increases upon binding with ligands (a decrease in the intrinsic viscosity and an apparent decrease in the hydrodynamic volume); 2) liganded nuclease-(l-126) exhibits significantly disordered properties The value of KI or K , is, in theory, different between the species A (equation 1) and E (equation 2) (see Table V). However, the results of measurements of CD (see "Results") have made it unlikely that Kl with species E is equal to that with native nuclease A. Therefore, there is no information available for K, with species E which difierentiates the two models. T-(50-149). However, the initial ratio of type I to type I1 complex as well as their yields were within experimental error independent of temperature and the presence or absence of ligands (16). Thus, in the line of reasoning described previously (see footnote 9 of reference 16), taking into account the rate of formation of the complexes (35) and the rate of digestion of the fragments with trypsin, it can be argued that nuclease-(l-l26) bound with ligands directly interacts with nuclease-T-(50-149). Then, it follows that liganded nuclease-(1-126) is flexible in that the probability of formation of type I and type I1 complexes is the same at 23 "C regardless of whether or not nuclease-(l-126) is liganded. A similar argument holds also for the interaction of the two fragments in the presence of ligands at 6 "C (incubation with trypsin was at 23 "C after temperature equilibration at 23 "C for 2 to 3 min) (16).
These analyses indicate that formation or stabilization of the active site of nuclease-(l-126) by binding with ligands is not associated with cooperative folding of the entire poly-

Conformation of Liganded Nuclease-(2-126)
peptide chain as observed below 30 "C, outside the thermal transition range exhibited by the ordered elements of nuclease-(1-126) (see above). The Arrhenius plot of enzymatic activity (Fig. 2) also indicates that no further folding of nuclease-(1-126), which would affect enzymatic activity, occurs by decreasing the temperature down to 5 "C. Thus, the flexibility of the conformation seems to be intrinsic to nuclease- (1-126).
The dramatic stabilization of native nuclease A (19,32,41,42) and all the ordered complexes of the fragments (8,15,16,19) by binding with ligands has been shown to be solely due to suppression of unfolding (15,35), which is, in turn, explained by assuming that the interatomic interactions at the ligand binding site are coupled with the cooperative interatomic interactions operative throughout the three-dimensional structure (the global cooperative interaction^)'^ (15). In this context, the flexibility of liganded nuclease-(1-126) may be explained by the assumption that such global cooperative interactions would be disrupted with nuclease-(1-126) so that the ligand binding interactions at the active site would not be linked to the interatomic interactions of the rest of the conformation. Conversely, the observed weak binding of pdTp with nuclease-(1-126) (in the presence of Ca") may be explained, at least in part, by the absence of contribution to the binding force from such global cooperative interactions (15).
Although nuclease-( 1-126) lacks information for cooperative folding, nuclease-(l-126) has an ability to interact and fold cooperatively with a carboxyl-terminal fragment, nuclease-(lll-149) (8) (see also the miniprint). On the other hand, combination of nuclease-( 1-126) and nuclease-( 127-149) does not generate a noncovalent complex even in the presence of pdTp (immobilized) and Ca2+ at 6-8 "C (see the miniprint supplement). This suggests that the mechanism of such global cooperative interactions would be specific in that hydrolytic cleavage of only one peptide bond between residues 126 and 127 of native nuclease A would be sufficient for their disruption, resulting in loss of noncovalent interactions to bind the two derived fragments. In fact, only the two specific sites of nuclease A (residues 48 to 50 and 114 to 124) are permissible for cleavage without disrupting the force holding together the resulting fragments (8,19,34,43) (see also the miniprint supplement for nuclease-( 1-105)).
Specific global cooperative interactions have been proposed to come into existence only after completion of folding of native nuclease A to provide the force for maintaining the ordered structure (14)(15)(16). The formation of noncovalent complexes, consisting of the ordered structures plus the flexible segments protruding from the specific sites," as a thermodynamically stable system may also be explained by the assumption that the segments incorporated into the ordered structure are determined on the basis of the requirement for such global cooperative interactions so that the redundant segments are superfluous to and do not interfere with the ordered structure (3, 16).
Thus, it may be thought that information for such specific global cooperative interactions underlies cooperative folding of native nuclease A as well as formation of ordered complexes from the fragments." In this context, the fact that stabilization of the active site of nuclease-(l-126) by binding with ligands is not followed by cooperative folding of the rest of the polypeptide chain may also be considered as indicating the l 3 Note that the global cooperative interactions are not long range interaction itself. The latter interaction is the interaction between residues close in space but f a r apart in the amino acid sequence (cf. Ref. 39). The global cooperative interactions involve linkage of interactions throughout the three-dimensional structure which is assumed to cause a further decrease in energy (14-16). specific nature of the mechanism of such global cooperative interactions, the disruption of which cannot be compensated for by binding with ligands.

( 7 1
Taking the reciprocal we obrain For rhe second model