Spectral Differences between Rhodanese Catalytic Intermediates Unrelated to Enzyme Conformation*

Circular dichroism (CD) spectra and UV absorption spectra of two obligatory intermediates in rhodanese catalysis were compared. A broad CD band between 250 and 287 nm increased in a manner stoichiometri- cally related to the content of enzyme-bound persulfide. Titration of a sample of sulfur-substituted rho- danese (ES) with either cyanide or sulfite gave a stoichiometry that is consistent with one persulfide/mole-cule of rhodanese (M, = 33,000). This result agrees with that determined by x-ray crystallography and a method based on quenching of intrinsic fluorescence. Cyanolysis of the persulfide in ES is accompanied by a decrease of UV absorption in the region between 250 and 300 nm. Cyanide titrations followed by the change in absorbance at 263, 272, and 292 nm gave the ex- pected stoichiometry. The magnitude of the difference between the far UV-CD spectra of E and ES found here is smaller than reported previously. This variability suggests that the differences in the secondary structure of these intermediates may not be obligatorily related to the cyanolysis of the persulfide. This view is com- patible with recent evidence which suggested that E and ES may be made different by structural relaxation events that occur outside of the catalytic cycle. Fur-thermore, the methods developed here will be useful in studies on the stability of the catalytic persulfide that has been suggested to be central in the mechanism of several enzymes important in sulfur metabolism.

Spectral Differences between Rhodanese Catalytic Intermediates Unrelated to Enzyme Conformation* (Received for publication, February 26, 1985) Shiu Fung Chow and Paul M. Horowitz

From the University of Texas Health Science Center, Department of Biochemistry, San Antonio, Texas 78284
Circular dichroism (CD) spectra and UV absorption spectra of two obligatory intermediates in rhodanese catalysis were compared. A broad CD band between 250 and 287 nm increased in a manner stoichiometrically related to the content of enzyme-bound persulfide. Titration of a sample of sulfur-substituted rhodanese (ES) with either cyanide or sulfite gave a stoichiometry that is consistent with one persulfide/molecule of rhodanese (M, = 33,000). This result agrees with that determined by x-ray crystallography and a method based on quenching of intrinsic fluorescence. Cyanolysis of the persulfide in ES is accompanied by a decrease of UV absorption in the region between 250 and 300 nm. Cyanide titrations followed by the change in absorbance at 263, 272, and 292 nm gave the expected stoichiometry. The magnitude of the difference between the far UV-CD spectra of E and ES found here is smaller than reported previously. This variability suggests that the differences in the secondary structure of these intermediates may not be obligatorily related to the cyanolysis of the persulfide. This view is compatible with recent evidence which suggested that E and ES may be made different by structural relaxation events that occur outside of the catalytic cycle. Furthermore, the methods developed here will be useful in studies on the stability of the catalytic persulfide that has been suggested to be central in the mechanism of several enzymes important in sulfur metabolism.
Rhodanese, thought to be an important regulator of sulfur flux through a cell's sulfane pool (1,2), can catalyze the transfer of sulfur from thiosulfate to cyanide by a double displacement mechanism in vitro (Reactions 1 and 2). The single polypeptide of 293 amino acids can be isolated as either of two catalytic intermediates: the sulfur-free enzyme (E') and the sulfur-substituted form (ES). The transferable sulfur atom in ES is bound to the active site sulfhydryl in a persulfide bond (3)(4)(5).
Much effort has been focused on deducing an expanded catalytic mechanism that includes putative enzyme conformational changes. Based on the pioneering work of Volini * This work was supported by Research Grant GM 25177 from the National Institutes of Health and Welch Grant AQ-723 (to PMH). 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.
The abbreviations used are: E, sulfur-free rhodanese; ES, sulfursubstituted rhodanese. and Wang (6,7), it has been proposed that E and ES are structurally different in solution; and subsequent studies that have compared the physical properties of these intermediates gave results that are compatible with that proposal (8)(9)(10). Other kinetic studies suggested that conformational maleability of this enzyme may be the basis for the anomalous kinetic behavior observed (11,12). The solution results seemed to indicate that the enzyme can cycle between the two conformers. In contrast to the solution studies, x-ray crystallographic studies have failed to see significant conformational differences between the two intermediates in a crystalline environment (13,14). This observation is incompatible with the idea that enzyme structural changes are catalytically coupled. There is, as yet, little supporting data for explanations of the observed discrepancies between the crystallographic and solution studies (15).
It is generally accepted that E and ES are conformationally distinct in solution; however, the reported magnitude of the difference as indicated by CD and UV absorption spectral differences between the two forms has varied. A comparison of the data from the report that first documented spectral differences between E and ES (6) with the data of a more recent study (16) shows intriguing differences. Understanding these discrepancies is important because spectral differences between E and E S are taken to be reflections of conformational differences between these catalytic intermediates. Furthermore, the magnitude of the structural differences places constraints on the appropriateness of the models used to understand the functional behavior of rhodanese. In this report, we have examined the difference UV absorption spectra and CD spectra of E and ES. The results suggest that the magnitude of the conformational differences between the two catalytic intermediates may be more subtle than previously suggested and those differences may not be obligatorily coupled to catalysis. This view is consistent with the recent evidence suggesting that differences between E and ES may develop in processes that occur outside of the catalytic cycle (12).

EXPERIMENTAL PROCEDURES
All reagents used were analytical grade. Bovine liver rhodanese was prepared as previously described (17) and was stored at -70 "C as a crystalline suspension of E S in 1.8 M ammonium sulfate containing 1 mM sodium thiosulfate. Stock solutions of the ES form were prepared by washing crystalline ES twice with 1.8 M ammonium sulfate to remove the excess thiosulfate. The washed pellet was dissolved to a protein concentration of 40-50 mg/ml with 10 mM sodium phosphate buffer, pH 7.0. Enzyme activity was measured by a colorimetric method based on the quantitation of the product thiocyanate as a ferric thiocyanate complex (18). The protein concentration was determined spectrophotometrically using Eo&%n, , = 1.75 (19). Unless otherwise stated, experiments were performed in buffer containing 50 mM Tris. SO,, pH 8.6. Circular dichroism measurements were performed in a Jasco Model 5-500 spectropolarimeter with data processing accessory (Model DP-500). A small volume of an ES stock solution was added to buffer previously equilibrated to 23 "C. To obtain the CD spectra in the near UV region, cylindrical cuvettes of 1-cm optical path length were used. Titrations were performed by adding to the cuvette microliter quantities of a KCN or a Na2S03 solution made with buffer. The data presented were the average of 32 scans. To obtain the CD spectra in the far UV region, enzyme samples were made in buffer containing 10 mM Tris. SO4, pH 8.6, with or without KCN. A cylindrical cuvette of 0.1-cm optical path length was used.
[O]MRW was calculated using a mean residue weight of 115. Difference absorption spectroscopy was performed in a Cary 219 spectrophotometer using quartz cuvettes having 1-cm optical path lengths. To get the base-line, a sample cuvette containing 1 ml of buffer with ES at 30.9 p~ was scanned against a reference cuvette containing an identical enzyme solution. To titrate the ES sample with KCN, microliter quantities of a KCN solution made in buffer were added to the sample cuvette. The same volume of buffer without KCN was added to the reference cuvette to compensate for dilution of the sample.

Current interest
in the occurrence of conformational changes in rhodanese during catalysis makes it important to resolve the discrepancies between the reports of the CD and UV absorption spectra of two rhodanese catalytic intermediates. Volini Table I). It was postulated that this broad band could be due to the perturbation of aromatic side chains and that the band between 255 and 277 nm could be due to the persulfide group. However, no direct correlation between the intensity of the CD band with the content of enzyme-bound sulfur was shown. If this band indeed reflects the content or the environment of the persulfide, then its intensity should be related stoichio-  The E sample contained 39 p~ enzyme, 10 mM Tris. so,, pH 8.  to a sample of ES is accompanied by an increase of the ellipticity of a broad band between 260 and 287 nm. The peak around 292 nm that has been attributed to tryptophanyl side chains is not significantly changed with KCN additions (Fig.  1). The per cent change in ellipticity at 263 nm is plotted as a function of the ratio of cyanide to enzyme in Fig. 2a. The equivalence point is reached at the cyanide to enzyme ratio of 0.89. This value is consistent with the expected stoichiometry of 1 cyanolyzable persulfide/molecule of rhodanese (Mr = 33,000). Furthermore, as implied by Reaction 2, cyanolysis of the persulfide in ES has been reported to be irreversible and to include no kinetically significant complex (21,22). This is reflected by the sharp break in the titration curve once the expected stoichiometry has been reached.
A similar titration performed using sodium sulfite as the titrant also gives the expected stoichiometry. The equivalence point is reached at a sulfite to enzyme ratio of 1.1. The reversibility of Reaction 1 apparently is reflected in the shape of the titration curve shown in Fig. 2b as contrasted with that shown in Fig. 2a. Likewise, addition of the donor substrate thiosulfate which reforms the persulfide conforms to the equilibrium process (about 75% restored at a ratio of [thiosulfate]/[enzyme] 2 : 30). The data shown in Fig. 2b fit reasonably well with those reported in the intrinsic fluorescence quenching studies (23) as shown by the open circles.
Specific activity measurements indicate that the enzyme is fully active at the end of the titration. Neither the addition of NaCl nor Na2S0, appears to affect the ellipticity in a comparable manner, so nonspecific anion-induced conformational changes can be discounted. The close correlation of the CD change with the addition of substrates that lead to reactions as predicted by Reactions 1 and 2 suggests that the titration of enzyme-bound persulfide is being monitored.
The previous reports further disagreed in the difference UV and 300 nm. The relationship of the CD difference in the aromatic region to the persulfide suggested that an absorption difference in the same region may also be related to the persulfide. Fig. 3a shows a difference spectrum of the two intermediates. Titration of an ES sample with KCN was performed by monitoring the difference absorption change. Changes of intensity a t 263, 272, and 292 nm were plotted as a function of the ratio [cyanide]/[enzyme] (Fig. 36). The plots of the fraction of the difference change monitored at 263 nm and 292 nm are analogous to that shown in Fig. 2a and gave the expected stoichiometry. The plot of the change a t 272 nm showed scatter that could be a reflection of the noise in this region (Fig. 3a). The stoichiometry determined by this method is the same as that determined when the near UV-CD change is monitored.
Several reported CD spectra of E and ES in the peptide backbone region differed in magnitude (see Table 11). Fig. 4 shows the far UV-CD spectra of E and ES. Although the overall shapes of the CD spectra are comparable, the magnitude of the difference between the spectra of E and E S is about one-third of that reported previously. The ellipticity of ES at 220 nm (about -6000 degrees cm2 dmol") agrees with the value reported previously; however, the ellipticity of E at the same wavelength (about -6400 degrees cm2 dmol") is significantly less. Since the magnitude of the difference in the far UV-CD spectra is taken to indicate the extent of secondary structural differences, this result leads one to suspect that the conformational differences between E and ES may be more subtle than previously suggested. Furthermore, the stoichiometric relationship between the magnitude of this CD difference and enzyme-bound sulfur has not been shown; so, the question of whether the presumptive structural differences are obligatory to the interconversion between the two intermediates remains open.
In view of the evidence showing conformational changes in E in the absence of sulfur donors (7) and the more recent evidence suggesting that the conformation of the enzyme is determined by a number of competing processes (12), the observed discrepancies may have bases related to the functional behavior of the enzyme. One factor that could have contributed to the inconsistencies observed is that composition of an enzyme sample thought to be ES may contain some E forms. It has been pointed out that sulfur can dissociate from the enzyme during recrystallization and other purification procedures (24). Different solution methods used to quantitate the bound sulfur have reported a range of stoichiometries (4,25,26). This uncertainty may reflect the inherent instability of the persulfide bond. The ability to quantitate and monitor the enzyme-bound persulfide is particularly important because of its central role in the catalytic mechanism.
The various methods designed to quantitate the persulfide have been unsatisfactory. One based on the occurrence of a weak absorption band ( E = 80 cm" M" at 330 nm (4)) in ES required conditions ([enzyme] 60 mg/ml) too restrictive to be generally useful in studies on the functional relevance of enzyme conformational change; furthermore, no stoichiometric relationship between this band's intensity and sulfur content of the enzyme has been established. A more sensitive method is based on the discovery that the intrinsic fluorescence of ES is about 30% lower than that of E (17). Although this method gives the accepted stoichiometry and equilibrium constant for Reaction 1, the quenching mechanism is unknown and may involve processes that are not directly related to the presence of the persulfide (12).
A more interesting possibility is that the different conditions used may have contributed to the variability observed in the far UV-CD spectra of rhodanese. In the earlier study, the buffer contained a delicate balance of thiosulfate, sulfite, and sulfate (6). As implied by Reaction 1, a mobile equilibrium exists between E, ES, and ( E . S20z-) under these conditions such that the spectrum reflects contributions from each of these species. In the more recent study, the spectrum of E was obtained after an ES sample was treated with a 2-fold molar excess of cyanide. In light of the evidence of E conformational autoconversions, it is questioned whether the far UV-CD spectrum reported for E is in fact an average of all the sulfur-free conformers that existed in solution under these conditions. Variations in the populations of sulfur-free conformers in the sample could account for the variability in the magnitudes reported.
It is important to note that it has not been established that the differences in the near UV-CD and UV absorption reflect conformational differences between the two intermediates.   However, the broad negative band from 250 to 300 nm also overlaps the absorption region of tryptophans, tyrosines, and phenylalanines. It is possible that sulfur removal from the active site perturbs the aromatic side chains, possibly those that make up the hydrophobic cluster (31) implicated in stabilizing the persulfide. This would be consistent with the suggestion that the conversion is associated with conformational change. Since the relative contribution to the observed absorption difference by the persulfide and the perturbation of the aromatic side chains is unknown, the extent of conformational change remains an open question.
In conclusion, the negative CD band between 255 and 270 nm appears stoichiometrically related to the persulfide within the rhodanese active site. These findings provide a basis for a facile procedure which quantitates and monitors the enzyme-bound persulfide. In addition, these methods may be generally useful to study the properties of enzyme-bound persulfides (32-34). More importantly, the assignment of this band to the persulfide explains part of the near UV-CD spectral differences found between E and ES. Subtle structural differences are also indicated. Likewise, the UV absorption difference found between the two intermediates reflects the presence of the persulfide in ES and may also indicate small conformational differences between the two forms. The results reported here are compatible with the recent evidence suggesting that some conformational differences between E and ES depend on processes that take the enzyme outside of the catalytic cycle (12). One might speculate that although E and ES have distinct conformational potentialities, their structures during catalysis may be more similar than previously thought. This view would accommodate the observation that there is no significant conformational change associated with the cyanolysis of the persulfide in the rhodanese crystal.