Evidence for Conformers of Rabbit Muscle Adenylate Kinase*

of the apparent M, values and the circular dichroism patterns suggest the existence of three rel- atively stable conformers of rabbit muscle adenylate kinase (RMAK). M,

These studies derive from observations that crystalline porcine skeletal muscle AK' exists in more than one pHdependent conformation (1,2), our observation (3) that a single AK isozyme from cat heart showed multiple M, values by gel filtration (4), and that multiple M, values of AK were reported in the urine of patients with myocardial infarctions (5). The reported variations of conformations from chemical modifications of bovine serum albumin (6), detected as altered M, values, and Stokes radii, determined by gel filtration techniques, serve as a basis for these studies. The skeletal muscle AK isozymes are bilobe molecules consisting of two globular units joined by an unordered region forming a cleft (1,2,(7)(8)(9)(10).
We studied the effect of pH, substrates, and DTT on the M in conformations observed in AK crystals due to pH are an inherent property (2).
CD studies also revealed alterations in the conformations of AK with changes of the pH and in the presence of DTT. In the absence of DTT and with changes in pH, there are very small changes in the a-helix and the ,&pleated sheet regions while the unordered region shows no variations.

RESULTS
The  formation. Table I gives the estimates of the Stokes radii of RMAK at different pH values based on gel filtration data. The minimum Stokes radius of 1.81 nm obtains at pH 5.0 in the absence or presence of DTT, and the maximum Stokes radius of 2.14 nm obtains at pH 9.0 in the presence of DTT. These observations and reports of conformational changes associated with substrate binding (1,8,22) Table II gives the Stokes radii calculated from the gel filtration data shown in Fig. 2.
Some generalizations about the effect of the substates can be derived from the elution patterns shown in  Fig. 3.
The Effect of ATP on Apparent M, Values at pH g--The elution profiles in the presence of ATP are more complex than in the presence of AMP, as shown in Fig. 3. At concentrations of 0.1, 0.5, and 1.0 mM ATP (Fig. 2), three peaks appear with apparent M, values near 22,000, 26,000, and 29,000. As with AMP, there is a shift toward an apparent M, value of 29,000 as the ATP concentration increases and a single peak obtains at 4.0 mM ATP.
The Effect of Both ATP. Mg on Apparent M, Values at pH g--The pattern in Fig. 3 shows two peaks at ATP. Mg con-   peak (not shown) obtains for these conditions. The effect of substrates and DTT on the apparent molecular weight and Stokes radius of RMAK Data are based on the (V, -V,)/Vi (6) determined from the data in Fig. 2. Other details are the same as those given in Table I.
Circular Dichroism Studies-In order to better understand the changes of the apparent M, values, we studied the effects of pH and DTT on the circular dichroic patterns. Fig. 4A shows typical CD patterns when pH is varied, and Fig. 4B shows an example of the differences in the patterns in the presence and absence of DTT at pH 8. Using circular dichroic data from pH 5 to 9 and in the presence and absence of DTT, Fig. 5 shows the percent the secondary structures determined by using the following equation (23)  where X is the total of the secondary structures and fH, fO,fR, and ft are the fraction or percent of a-helix, P-pleated sheet, unordered, and B-turns, respectively, and Xn,Xs,Xr, and Xt are estimated values for the corresponding secondary protein structures. The CD spectra were deconvoluted by the method of Compton and Johnson (24,25) with a program made available to the authors by Dr. Steven Koerber, Biosym, San Diego, CA. The results shown in Fig. 5 indicate little changes in the secondary structure between pH 5 and 9 in the absence of DTT, the a-helix ranging between 10 and 15%, the 8pleated sheet decreasing from 55 to 48%, and the unordered and p-turns being near 30 and 5%, respectively. These results are not in agreement with the values obtained from the crystalline structure of RMAK (2,6). In the presence of DTT over the same pH range, the percent of the a-helix fraction increases from 9 to about 30%; the P-pleated sheet fraction decreases from 55 to about 30%; and the unordered and the P-turn fractions show less than a 10% change.
Correlation of DTT Activation with Molar Elliptic&y-We had determined previously that the enhancement of activity by the presence of DTT was measurable with time (12) . Fig.  6, A and B, shows a preliminary study of the changes of RMAK activity at pH 8.0 and the changes of the molar ellipticity with time under similar conditions. Though 222 nm is generally associated with the a-helix region and 206 nm with the P-pleated regions, these measurements at only two wavelengths with time do not permit an estimation of the changes in the secondary structures of RMAK.

DISCUSSION
These studies suggest RMAK can exist in solution as three relatively stable conformers that are resolvable by gel liltration techniques (3,4,26). The occurrence of several RMAK conformers in solution at different pH values supports the view that the pH-dependent conformations observed in porcine skeletal muscle AK crystals are an inherent characteristic (1, 2) even though such changes may not apply to skeletal muscle AK from all sources (27). The M, values calculated from these gel filtration studies are near 21,500, 26,000, and 29,000, equivalent to estimated Stokes radii (16) of 1.81, 2.02, and 2.14 nm, respectively. We interpret the variations of the Stokes radii as conformational changes because of alterations of intramolecular charge distributions induced by substrates, pH, and DTT. The structural features of mammalian skeletal muscle AK (6,21,22) permit alterations of the effective molecular radius by opening and closing of a cleft formed by two globular domains connected by an unordered region. The addition of substrates, AMP and ATP. MG, and ATP show pronounced complex effects on the gel filtration elution patterns of RMAK. The multiple peaks obtained probably relate to the multiple species of RMAK-adenosine nucleotide complexes evident from kinetic studies that show ATP or AMP can bind at both active sites, as a substrate and to the other site as an inhibitor (2,8,(28)(29)(30). The shifts toward a higher apparent M, value and a single peak at the higher AMP and ATP. Mg concentrations are consistent with a shift toward a single conformer as active sites approach saturation. At the concentrations of the substrate ATPeMg studied, a single peak obtains above 0.5 mM. In kinetics experiments not shown in this study, ATP. Mg cannot be shown to act as an inhibitor at high concentrations and presumably cannot bind to both substrate sites as can AMP and ATP (19, 26, 31-34). The Stokes radius of 2.00 nm estimated for the single peak in the presence of ATP. Mg is intermediate to the extremes determined in this study.
The equilibrium mixture of substrates also gave rise to a single peak but with a Stokes radius of 1.84, equivalent to a M, value of 21,500, indicating a very closed conformation. Such a closed conformation would not seem to permit the placement of the active site of either of the substrates in a cleft and suggests that the active sites may be more superficial than those that have been proposed (2,31,(35)(36)(37)50). The identities of the active sites of the AKs are still controversial, and it has been pointed out that despite very similar enzyme kinetics, there appears to be dramatic changes of amino acid composition from various sources (50) around what is considered the active sites of AKs. The isolation and identification of two peptides that bind either ATP. Mg or AMP seem to have established the major amino acid sequences constituting the active sites of calf muscle and RMAK (19). Recent studies with a synthetic peptide of positions l-45 of RMAK that binds ATP . Mg showed secondary structures calculated from CD data (34) similar to what we found for RMAK in solution.
We are presently attempting to study the effect of substrates on the secondary structures of RMAK.
The opening and closing of the cleft induced by the presence of the substrates is consistent with the kinetic mechanism. The RMAK is a random Bi Bi system (28,38). Holding one of the substrates constant and at relatively high concentrations while varying the second results in a family of rectilinear lines that pass through a coordinate above the reciprocal substrate concentrations axis and to the left of the reciprocal initial velocities axis (28). Such a family of plots is interpreted as showing that the binding of the first substrate increases the affinity of the enzyme for the second substrate (39). The mechanism we propose for RMAK is that the initial addition of either AMP or ATP . Mg to RMAK induces the formation of a more open conformation that allows an easier binding of the second substrate; when the second substrate adds, a very closed conformation then obtains. The enhancement of FIG. 6. An estimation of the internal structural composition of RMAK from CD data as described in the text.
The conditions are 100 mM potassium phosphate, pH 8.0, in the absence (m) and presence (Cl) of 1.0 mM DTT.
The fm, fp, fR, and ft are the percent of the a-helix, the P-pleated sheet, the unordered, and p-turns, respectively. other sulfhydryl reagents appears to be related to the formation of more stable or more enzymatically active conformers rather than a protection from sulfhydryl group oxidations or the prevention of disulfide bonds (38).
radii, but we recognize also that changes in the unordered regions are not likely to be detected by CD under our conditions. In the presence of DTT, the changes of the secondary structures observed by CD do appear to reflect changes in the apparent M, values. Up to pH 7, the presence of DTT appears to have little or no effect on the o-helix region; above pH 7 the percent of a-helix structure increases. The P-pleated sheet region decreases with increasing pH from 5 to 9. Over the pH range tested the apparent M, values in the presence of DTT suggest that as the pH increases a more open conformation results, characterized by an increase in the a-helix region and a concomitant decrease in the @-pleated sheet region.
Our preliminary studies showed that the rates of the enhancement of activity by DTT within the first minutes parallel increases of the [0] at 222 nm while no changes are occurring at 206 nm. At present, it is not possible to associate the kinetics of [0] changes at specific wavelengths with specific secondary structural changes. The kinetics of secondary structure changes as measured by CD techniques are the focus of present studies. Under the conditions of these experiments, the estimations of the secondary structural composition from the CD data do not agree with the results obtained from crystallographic studies (1, 2, 6) as do the CD studies of the AK from Escherichia coli (49). The estimations, at pH 8 in the absence of DTT, are about 15% a-helix fraction, 50% P-pleated sheet, 25% unordered region, and 6% p-turns; as shown in Fig. 5, the reverse of the ordered secondary compositions derived from crystallographic studies. Though not given, when CD data from these studies were applied to equations developed earlier (51, 52) for estimating the percent of o-helix, these results agreed with estimates presented here within a few percent. The differences from the crystallographic data can derive from any of several possibilities. Among those worthy of study are that the structure of RMAK in solution is indeed different from that in the crystal or that the CD spectrum of RMAK is not compatible with the assumptions of the equations (23)(24)(25) developed for the estimates of secondary structures in proteins.