Sugar transport by the bacterial phosphotransferase system. Fluorescence studies of subunit interactions of enzyme I.

Enzyme I of the bacterial phosphoenolpyruvate:glycose phosphotransferase system (PTS) exhibits a temperature-dependent monomer/dimer equilibrium. The accompanying paper (Han, M. K., Roseman, S., and Brand, L. (1990) J. Biol. Chem. 265, 1985-1995) shows that the C-terminal -SH residue (Cys-575) can be modified specifically with fluorescent probes such as pyrene maleimide. The derivative retains full enzyme activity, and is capable of forming dimers at room temperature. In the present studies, Enzyme I labeled in this way is found to exhibit a temperature-, concentration-, and pH-dependent monomer/dimer association. The kinetics of dimer formation of Enzyme I is measured in the following way. A derivatized Enzyme I sample is prepared with a pyrene moiety irreversibly attached to the C-terminal -SH residue and 5,5'-dithiobis-2-nitrobenzoic acid reversibly attached to the other 3 -SH residues. This modified enzyme does not form dimers at room temperature. Addition of dithiothreitol results in total release of the thionitrobenzoate anion within 2 min. After the three -SH groups are unblocked, steady-state and nanosecond time-resolved emission anisotropy measurements indicate the dimer is formed over a period of 30 min. In a similar experiment, little dimer formation is observed at 3 degrees C, at temperature at which the native enzyme also does not form dimers. Tryptophan fluorescence is also examined during the release of the thionitrobenzoate. After the completion of thionitrobenzoate release, additional slow steady-state tryptophan fluorescence changes are observed. These results suggest that dimer formation may be preceded by a conformational change following thionitrobenzoate release.

The derivative retains full enzyme activity, and is capable of forming dimers at room temperature.
In the present studies, Enzyme I labeled in this way is found to exhibit a temperature-, concentration-, and pH-dependent monomer/dimer association. The kinetics of dimer formation of Enzyme I is measured in the following way. A derivatized Enzyme I sample is prepared with a pyrene moiety irreversibly attached to the C-terminal -SH residue and 5,5'-dithiobis-2-nitrobenzoic acid reversibly attached to the other 3 -SH residues.
This modified enzyme does not form dimers at room temperature.
Addition of dithiothreitol results in total release of the thionitrobenzoate anion within 2 min. After the three -SH groups are unblocked, steady-state and nanosecond time-resolved emission anisotropy measurements indicate the dimer is formed over a period of 30 min. In a similar experiment, little dimer formation is observed at 3 'C, at temperature at which the native enzyme also does not form dimers.
Tryptophan fluorescence is also examined during the release of the thionitrobenzoate.
After the completion of thionitrobenzoate release, additional slow steady-state tryptophan fluorescence changes are observed.
These results suggest that dimer formation may be preceded by a conformational change following thionitrobenzoate release.
Enzyme I of the bacterial phosphoenolpyruvate:glycose phosphotransferase system (PTS) 1 is responsible for the phosphorylation and concomitant transport of its sugar substrates across bacterial plasma membranes (l-3). The system must not only be regulated stringently, but regulation must be rapid, before the internal sugar-P concentration rises to toxic levels (4). It has been speculated that the most important form of regulation is related to the unusual associative properties of Enzyme I (5-7), the first protein in the phosphorylation sequence. This protein undergoes a striking temperature-dependent monomer/dimer transition (5), and this transition has been proposed as a model for the action of the enzyme and for regulation of the PTS and, thereby, of sugar uptake by the cell. The model consists of a cycle; monomer associates to dimer in the rate-limiting step, the dimer accepts up to two phosphoryl groups from P-enolpyruvate in the presence of Mg'+, the phospho-dimer dissociates to phosphomonomer, and histidine-containing phosphocarrier protein of the phosphotransferase system accepts the phosphoryl group from monomer to complete the cycle. This model is based on the results obtained with Salmonella typhimurium Enzyme I (5-8). Thus, a study of the monomer/dimer associative behavior becomes important in understanding PTS regulation. Emission anisotropy studies of the tryptophan fluorescence of Enzyme I have been used to measure the monomer/dimer equilibrium (9). However, tryptophan fluorescence is not a desirable probe for fluorescence anisotropy studies of large molecules. (For example, the average lifetime of tryptophan in Enzyme I is 5-6 ns and the rotational correlation time of the enzyme is 80-160 ns.) The utilization of extrinsic covalent probes is an alternative choice for fluorescence emission anisotropy studies.
As described in the accompanying paper (lo), Enzyme I contains four sulfhydryl groups which are classified into two groups: one (C-terminal Cys-575) reacts 10 times faster than the other three -SH groups with appropriate reagents such as DTNB.
Chemical modification of the C-terminal cysteine residue does not change the activity of the enzyme, whereas the modification of the three other -SH groups results in an inactive and monomeric enzyme.
Modification of the enzyme with a long-lived fluorescent probe at a unique (nonessential) site makes it possible to obtain specific structural insights over a time scale of seconds. Some of the information potentially available from the study of fluorescent-labeled enzymes are the following: accurate hydrodynamic properties of the functional enzyme, changes in conformation, subunit-subunit association, protein-protein interactions, and structure-function correlation.
In this paper, we present characteristics of pyrene maleimide-labeled Enzyme I and studies of the influence of concentration, temperature, and pH on the relative quantities of monomer and dimer of Enzyme I labeled at Cys-575. Data were analyzed by the procedure described by Badea and Brand (15) and also by the global analysis procedures described by Knutson et el. (16) and Beechem et al. (17).   (9) of 3.7 and 7.3 ns, and a small fraction of a short-lived species of 0.6 ns ( Table I). The aim of the experiment reported here was to attempt to determine the origin of the two long decay components with respect to the 2 tryptophan residues. One possible explanation for the lifetimes, 3.7 and 7.3 ns, is that each is associated with an individual tryptophan residue. Therefore, the quenching pattern of these lifetimes by pyrene conjugated to the protein should depend on the distance between these tryptophans and Cys-575 (since the efficiency of energy transfer is inversely proportional to the sixth power of the distance between a pair of fluorophores).
Two types of experiments were performed. In one, Enzyme I was reacted with 0.5-1.2 mol of pyrene maleimide/monomer. The reactions were allowed to go to completion and tryptophan fluorescence decay was measured. In a second set of experiments, approximately 1 mol of pyrene maleimide was reacted per mol of Enzyme I monomer, and a series of decay curves were obtained as a function of reaction time.
The results obtained from the first set of experiments are Sugar Transport by the Bacterial Phosphotransferase System summarized in Table I. The data were analyzed simultaneously by a global procedure with the decay constants (pi to TJ "linked" (16,17), and the lifetimes recovered from this analysis are 0.7, 1.9, 3.8, and 7.1 ns.
The new value is 1.9 ns, since the fluorescence decay of native Enzyme I (Table I) can be resolved into a triple exponential with two major components of 3.7 ns (41%, preexponential) and 7.3 ns (43%), and a minor component of 0.6 ns (16%). We believe that the 1.9-ns decay component is most likely a quenched lifetime due to a resonance energy transfer from tryptophan to pyrene. The four lifetimes obtained with pyrene-conjugated Enzyme I are very similar to the lifetime values obtained from the TNB-bound enzyme w.
Tracing the origins of the 1.9-ns component may provide valuable information for distinguishing the origins of the two native major decay components (i.e. whether they arise from the same or both tryptophan residues). The "map" of the extent of reaction (Table I) shows that the amplitude of the 7.1-ns decay (01~) decreases as it is "exchanged" with the amplitude of the 1.9-ns decay (o(J. On the other hand, the amplitude of the 3.8-ns component is apparently unchanged. Thus, these results suggest that the 1.9-ns component is a quenched decay arising "solely" from the 7.3-ns species of native Enzyme I, and therefore the 3.7-and 7.3-ns decay components originate independently from the two different tryptophan residues. The second type of experiment (time-resolved reaction kinetics of the interaction of pyrene with Enzyme I), was performed to test this conclusion.
For kinetic studies, the rapid collection of decay curves during the reaction is essential and can be achieved with a single photon-counting pulse fluorometer with a synchronously pumped, mode-locked, cavity-dumped dye laser as the excitation source. Approximately one pyrene maleimide was reacted with Enzyme I (to label specifically the fast reacting -SH group, Cys-575) and seven consecutive decay curves were collected (10 s per curve) during the reaction. The best fit of the data was obtained by fixing two long "native lifetimes" with a linkage of four decay components. Any contamination by scattered excitation light was corrected by fixing a O.Ol-ns decay term, and allowing its amplitude to vary. Fig. 2 shows plots of individual intensity (LyiTi) obtained from this systemwide analysis of the reaction. The intensity of the 7.3-ns component (01~7~) decreased substantially, while an additional decay component of 1.9-ns ((~~7~) appeared (initially absent). On the other hand, the 3.7-ns component (LYNCH) changed only slightly. We emphasize that only 10-s collection times are available for each decay curve and, therefore, that high data quality cannot be expected in the standard instrumental configuration.
The kinetic data shown in Fig. 2 independently support the suggestion that only a 7.3-ns component is quenched upon modification of the "fast reacting" -SH group. Characterization of pyrene fluorescence in Enzyme I- Fig.  3 shows emission spectra of pyrene maleimide conjugated mainly to the fast (A) and also mainly to the slow reacting (B) -SH groups of Enzyme I at pH 7.5. Emission spectra were recorded at 24-h intervals. When Enzyme I is labeled with less than 1 mol of PM/monomer at pH 7.5, preferential labeling takes place at the C-terminal cysteine residue (10). The spectra for this case (Fig. 3A) show that the emission peak at 376 nm is higher than the peak at 396 nm. In contrast, when slow -SH groups are labeled at pH 7.5 (with the fast -SH group protected by reaction with DTNB), the ratio of these two emission peak heights is quite different (Fig. 3B). Seven decay curves were collected at the rate 10 s/curve during the reaction of pyrene maleimide with Enzyme I. The experiments were performed in 100 mM HEPES, pH 7.5, and 1 mM EDTA with an enzyme concentration of 0.33 mg/ml. The measurements were begun as 6 pM pyrene maleimide was added. The peak counts were less than 1000 in each curve. The seven decay curves were analyzed simultaneously, with the aid of a global procedure. Two native components, 3.7 (TJ and 7.3 ns (TV), were fixed, since the PM-labeled Enzyme I exhibits two major native components. Two other components were linked (not fixed), and the lifetimes recovered were 0.5 (TV) and 1.9 ns (TV). The global reduced x2 was 1.15. Under certain conditions, the initial adduct formed by coupling pyrene maleimide to protein -SH groups may be unstable. The imide ring can open and rearrangement can occur (22)(23)(24)(25). It is evident that the succinimido ring on the slow -SH groups open much more rapidly than maleimide adduct of the fast -SH group. Moreover, the derivatized "slow" -SH groups exhibit a small but not insignificant pyrene emission peak at 470 nm, indicating excimer formation. Such a peak at 470 nm has never been detected with Enzyme I selectively labeled at the fast -SH group. While the protein labeled only on the "fast" forms a dimer at room temperature, no dimer is formed from the enzyme when the slow -SH groups are derivatized. Therefore, excimer formation is not the result of intermolecular interaction of the subunits, but it arises from intramolecular events. These results indicate that the ring-opening process of pyrene maleimide on the fast -SH group of Enzyme I does not interfere with fluorescence studies of the derivatized enzyme.
Anisotropy Studies of Pyrene Maleimide-labeled Enzyme I-The accompanying paper (10) indicates that the Cys-575 residue of Enzyme I is the fast-reacting -SH group and thus provides a unique labeling site. The C-terminal cysteine residue was therefore labeled with pyrene maleimide, and the monomer/dimer association of the labeled enzyme was characterized by fluorescence emission anisotropy experiments.
The temperature-dependent subunit interactions were examined, and a Perrin plot of the selectively labeled sample is shown in Fig. 4. Labeled Enzyme I exhibits a transition between 6 and 16 "C. In this range, steady-state anisotropy increased as the temperature increased, in accord with the expected monomer/dimer transition. This transition has also been observed by time-resolved emission anisotropy of tryptophan fluorescence (9) and previously by ultracentrifugation (5) and column chromatography (6). Moreover, the DTNB reaction with -SH groups of Enzyme I indicated changes in reactivity of the slow-reacting -SH groups in this temperature range (10). These results are consistent with previous studies showing that modification of Cys-575 yields both a catalytically active and a dimeric enzyme.
Since the monomer-dimer transition is a function of protein concentration, the emission anisotropy of labeled Enzyme I was studied at various concentrations at 23 "C ( Fig. 5A) and 14 "C ( Fig. 5B). At 23 "C, where native Enzyme I is predominantly a dimer, the pyrene maleimide-labeled enzyme likewise is dimeric down to 0.08 mg/ml; as the concentration is further decreased, the dimer dissociates. By contrast, at 14 "C the pyrene-labeled dimer rapidly shifts toward monomer as the concentration is decreased at 14 "C. It should be noted that the pyrene fluorescence yield is not significantly dependent on the monomer/dimer transition.
In comparison, the steady-state anisotropy values of 16 pg/ml of the enzyme are 0.126 and 0.123 at 23 and 14 "C, respectively. The absolute 6.5' 1 0. 1.

2.
(T/q) anisotropy values are not directly comparable since spherical shape cannot be assumed and the rotational correlation time is temperature-dependent. Therefore, the anisotropy value of 0.126 at 23 "C! represents more than 50% of dimer whereas 0.123 at 14 "C represents predominantly monomer. The shapes of the curves at the two temperatures can be compared and show that the shift of dimer to monomer occurs at higher concentration at 14 *C, in accord with the behavior of native Enzyme I. These data are presented to show that the anisotropy is concentration-dependent and thus reflects the monomer/dimer transition. The monomer/dimer association of the labeled enzyme was examined further as a function of pH, in the range pH 6.0-8.0, in 0.1 M potassium phosphate buffer containing 1 mM EDTA, and the results are shown in Fig. 6. The labeled enzyme remains dimeric between pH 6.8 and 8.0. At lower pH values the anisotropy decreased, indicating dissociation of dimers. Anisotropy values between pH 6.0 and 6.2 were around 0.12, which may indicate that the enzyme is not completely monomeric at these conditions.
The pH-dependent anisotropy experiments were repeated with nanosecond time-resolved emission anisotropy. In order to study the time-dependent decay of the emission anisotropy of the labeled enzyme, two fluorescence decay curves Ivv(t) and I&t) were obtained under dimer conditions (25 "C). The data were analyzed in terms of a single rotational correlation time, which should reflect some average of the monomer and dimer values (Table II). At pH 7.0, the best single rotational correlation time extracted from the data was 105 ns. The values decreased at lower pH (94 ns at pH 6.65 and 74 ns at pH 6.2). Therefore, the effect of pH on the rotational correlation times is consistent with the steady-state emission anisotropy results shown in Fig. 6. In contrast, the rotational correlation time of a dansylated monomer at pH 7.5 and 20.5 "C is 45 ns (10).  can be used to shift the equilibrium, but temperature also changes solvent viscosity and fluorescence properties, thus complicating hydrodynamic measurements.
An alternative method was therefore developed.
The modification of all four -SH groups results in inactivation of Enzyme I with subsequent monomerization (10). When the -SH groups are derivatized with DTNB, the TNB moieties linked to the protein can be removed by reduction with DTT, thereby reactivating the enzyme with concomitant dimerization (10,26). The reversibility of TNB inactivation of Enzyme I was therefore used in steady-state emission anisotropy experiments to measure the kinetics of Enzyme I dimerization. Enzyme I was selectively and irreversibly labeled with pyrene maleimide at the C-terminal cysteine, Cys-575, which does not inactivate the enzyme. The other three -SH groups were reversibly labeled with DTNB, removed with DTT, and the dimerization process was followed by measuring changes in the emission anisotropy.
The time course of dimerization was measured at 3 and 25 "C after addition of dithiothreitol, and the results are shown in Fig. 7 has reasonable spectral overlap with pyrene and is expected to decrease the decay time of pyrene by resonance energy transfer. When the TNB anion is released from the protein, the expected results is an increase in the pyrene decay time or a decrease in the steady-state emission anisotropy (since steady-state anisotropy is inversely related to the lifetime of the probe). The difference in the initial steady-state emission anisotropy (prior to addition of DTT) at 3 "C uersus 25 "C results from temperature effects on anisotropy; higher anisotropy (at time = 0) is expected at lower temperatures mostly because of the increase in the viscosity of water.
Following the initial rapid change (2 min) at 25 "C, the anisotropy slowly increased from about 0.10 to 0.145 with the rate constant of 7.7 x 10-"/s ( tm = 15 min). By contrast, only small anisotropy changes are seen at 3 "C, a condition in which the native enzyme is predominantly monomeric. This confirms the absence of large volume changes upon TNB release. Nevertheless, small changes in anisotropy may indicate either a conformational change without dimer formation or partial dimer formation.
The slow changes in apparent volume during TNB release were also studied by time-resolved techniques. Anisotropy decay was measured before the addition of dithiothreitol and 40 min after. The correlation time prior to DTT addition was 64 ns compared to 90 ns measured 40 min after DTT addition (data not shown). These results, taken together with the steady-state data, suggest that anisotropy changes observed in these experiments represent slow dimer formation. Kinetic Studies of TNB Release from the DTNB-modified Enzyme Z of PTS-Because the absorption spectrum of TNB bound to the protein overlaps the tryptophan emission spectrum, quenching of tryptophan fluorescence was expected and observed upon reaction of Enzyme I with DTNB. In addition, tryptophan fluorescence was sensitive to temperature changes, an effect which can be ascribed to monomer/dimer equilibrium.
As Fig. 8 shows, the intensity of tryptophan fluorescence at 23 "C (as a dimer) was higher than at 1 "C (as a monomer). This intensity change also accompanies a slight spectral shift, suggesting that conformational changes accompany subunit association. Moreover, the difference spectrum is blue-shifted from the total emission spectrum. Neyroz et al. (9) reported that the 2 tryptophan residues of Enzyme I exhibit two distinctive decay-associated emission profiles, one of which (that associated with the shorter decay component) is more sensitive to temperature. This may indicate that each of the two tryptophans exhibit unique decay times. This, in turn, suggests that one of the 2 tryptophan residues is relatively more sensitive to monomer/dimer association than the other tryptophan residue. Thus, derivatization of the sulfhydryl groups of Enzyme I with DTNB may alter tryptophan fluorescence by both resonance energy transfer (tryptophan to TNB) and monomer/ dimer-associated conformational changes. It is important to examine closely the changes in tryptophanyl fluorescence to determine if these two processes can be distinguished.
One way to approach this problem is to study the release of TNB from the modified enzyme. Specifically, one can compare the rate of TNB release to the rate of fluorescence recovery. If only quencher release was involved in tryptophan fluorescence, the rate of fluorescence recovery should be the same as the rate of TNB release. More generally, fluorescence changes may also arise from any conformational changes associated with subunit-subunit interactions and the recovery rate might not be the same.
For this experiment, TNB-bound Enzyme I was first prepared by incubating native Enzyme I at room temperature with excess DTNB in 100 mM potassium phosphate, pH 7.5, and 1 mM EDTA. Modified Enzyme I was separated from excess DTNB by chromatography on a Sephadex G-25 column, equilibrated with 100 mM potassium phosphate, pH 7.5, and 1 mM EDTA at 23 "C. The degree of labeling was estimated by treating TNB-bound enzyme with a 20-fold excess of DTT followed by dialysis and Sephadex chromatography, and by measuring the amount of TNB anion released. The enzyme released 3.9 mol of TNB anion/mol, indicating that all four -SH groups per monomer were fully occupied by TNB.
The rate of TNB release can easily be measured by monitoring absorbance at 412 nm (27), and the result obtained at room temperature is shown in Fig. 9A. Fig. 9B shows the concomitant increase in steady-state tryptophan fluorescence. It is interesting to note that the kinetics of fluorescence recovery are biphasic with the rate constants of 8.3 x 10-*/s and 6.6 x 10w4/s. In contrast, the release of TNB (measured by absorbance) is essentially complete within a minute (with the rate constant of 8.45 X 10-'/s).
Thus, the rapid increase in the tryptophan fluorescence intensity is most likely associated with the loss of resonance energy transfer due to the fast release of TNB anion, since these two rates are the same (within experimental error). The slow phase of fluorescence recovery is intriguing because it indicates conformational changes accompanying dimerization.

DISCUSSION
Enzyme I is a key protein in the bacterial PTS since it is the first protein in the phosphotransfer chain and therefore may regulate the PTS and associated systems, such as adenylate cyclase. Previous studies (5-7) based on ultracentrifu- gation and molecular sieve chromatography presented evidence that Enzyme I exhibits a temperature-dependent monomer/dimer transition and that the monomers are identical. It is likely (8, 28) that the dimer form of the enzyme is the active species.
The aim of this study was to use both intrinsic and extrinsic fluorescence probes (a) to characterize the monomer/dimer equilibrium of Enzyme I with fluorescence emission anisotropy, (b) to examine conformational changes of the enzyme associated with the monomer/dimer transition, and (c) to characterize the kinetics of dimer formation from chemically induced monomers. Extrinsic fluorescence probes offer many advantages. For instance, probes with suitable excited state lifetimes, and advantageous excitation and emission spectral characteristics may be selected so that custom tailored experiments can be performed.
Extrinsic probes also have disadvantages. Covalent coupling of the probe at specific and unique sites is often difficult to achieve and the modified protein may have altered functions as compared to the native molecule.
In this study, site-specific labeling was achieved by reacting pyrene maleimide to the fast reacting -SH group, Cys-575, of Enzyme I. The resulting enzyme was fully active and formed a dimer. Confirmation of the site-specific labeling was presented in the accompanying paper (10). Pyrene was chosen as an extrinsic probe because it has a long decay time (29)(30)(31)(32)(33) and is thus useful for time-resolved emission anisotropy studies as well as for simpler steady-state fluorescence ansiotropy studies, which may be extended to stopped-flow speeds. A potential difficulty with the use of pyrene maleimide as a probe is that ring opening can occur after conjugation (22)(23)(24)(25). It has been suggested that S-[N-(l-pyrene)-succinimido] cysteine, a product, of the reaction of the sulfhydryl group of cysteine with the olefinic double bond of the maleimide moiety of N-(1-pyrene)-maleimide, undergoes a slow cleavage of the succinimido ring by either hydrolysis or aminolysis. Aminolysis accompanies cyclization of the succinimido ring to form thiazine derivatives as a result of subsequent nucleophilic attack by the amino group on a carbonyl carbon (22,23).
The stability of N-ethylmaleimide adducts with thiol groups has been described (22)(23)(24). The adduct is reasonably stable at pH 7.0 in potassium phosphate buffer but is much less stable under more alkaline conditions. Smyth and Tuppy (23) concluded that intramolecular cross-linking of amino and thiol groups by maleimide reagents occurs only when the groups are located in sterically favorable positions. They showed further that under optimal conditions, the competing reactions of intermolecular aminolysis and hydrolysis are minor. Wu et al. (25) extended the studies of intramolecular aminolysis with pyrene maleimide (PM). They reported that cleavage of the succinimido ring causes a spectral shift in both the excitation and emission spectra of pyrene and to changes in decay times. The nature of the rearrangements often depends on the environment of the cysteine residue (25), and we find this to be the case with Enzyme I.
In the present study it has been shown that, while pyrene maleimide conjugated to the slow -SH residues is susceptible to rapid ring opening, the derivative at Cys-575 is relatively stable to the ring opening reactions. Spectral changes taking place over a 48-h period at pH 7.5 with pyrene specifically conjugated to the fast -SH (Cys-575) and also with pyrene reacted only with the slow cysteine residues are shown in Fig.  3. The derivative of the fast -SH is reasonably stable for 7 days at pH 7.5 and even more stable at more acidic pH. The pyrene maleimide derivatives of the three slow -SH groups gave different results. Fig. 3B shows that the imide ring opened much more rapidly, so that at 48 h (pH 7.5), the derivative had lost about 50% of the initial emission. Thus, the fast and slow sulfhydryl residues are distinguishable by: (a) their relative reactivity to -SH reagents, (b) a spectrally observed ring opening reaction which depends on the microenvironment, and (c) excimer formation between closely located pyrene probes on the protein. Because of these results, the remaining experiments were conducted with freshly prepared conjugates of Enzyme I, labeled with pyrene maleimide specifically at Cys-575, in which negligible ring opening had taken place.
Enzyme I labeled specifically at Cys-575 showed a temperature-, concentration-, and pH-dependent monomer/dimer equilibrium.
As indicated by a Perrin plot (Fig. 4), the emission anisotropy showed an inflection at the temperature region of 6-16 "C, conditions where the native enzyme undergoes the monomer/dimer transition (9, 10). These results, together with concentration-dependent anisotropy data, indicate that the labeled Enzyme I behaves similar to the native enzyme. Moreover, emission anisotropy of the labeled enzyme was pH-dependent; optimum ansiotropy was found between pH 7 and 7.5. Interestingly, Waygood and Steeves (34) reported that the activity of Enzyme I was also pH-dependent with maximum activity at pH 7. These results indicate a good correlation between enzyme activity and the amount of dimerit form present, suggesting that the dimer form of the enzyme is the active species.
As indicated by Neyroz et al. (9), tryptophan fluorescence of Enzyme I exhibits two major decay-associated spectra (35): the 3.7-ns component is blue-shifted compared to the 7.3-ns component.
Moreover, the spectrum associated with the 3.7ns component comprises 17% of total fluorescence intensity at 1 "C (monomeric condition), whereas this contributes 44% of total intensity under dimeric conditions, at 23 "C! (9). This is also clearly evident in Fig. 8 which shows that fluorescence intensity at 23 "C is higher than at 1 "C, and that the difference emission spectrum is suggestive of the 3.7-ns decayassociated spectra. These data suggest that the two decay times originate from tryptophan residues in different unique environment.
The results obtained with the tryptophan to pyrene resonance energy transfer experiments provide valuable information for distinguishing the origins of the two major decay times. Enzyme I with pyrene specifically conjugated to the fast -SH (Cys-575) contains two fluorescence energy donors (Trp-355 and Trp-498) and a common acceptor (pyrene). Both titration and kinetic data indicate that the conjugation of pyrene maleimide to the protein result in quenching of the 7.3-ns decay component (but not the 3.7-ns decay time) to the 1.9-ns component. The only tryptophan close to a cysteine in the primary structure is Trp-498, which is close to cysteine-502. Therefore, we would predict strong quenching between Tip-498 and an energy transfer acceptor bound to Cys-502, one of the slow-reacting cysteines. In fact, the 3.7-ns decay component was quenched to 0.4-ns when the slow -SH groups were labeled with DTNB (21).
The behavior of the 3.7-ns decay component, which is sensitive to subunit association as shown by the temperature studies indicated in Fig. 8, is consistent with the kinetic results observed with the slow -SH groups. These data suggest that Trp-498, which is in close proximity to the slow-reacting