......... A comman- lfiTEREACED RAPID SCAN ' * STOPPED-FLOW 3mm FOR THE ~ * STUDY OF TRANSEENTS EN 7 " ' ENZYME REACTiGNS Dissertation for the Degreerof ”1.20; ” MECHIGAN STATE UNIVERSITY, 4 ‘ ' RICHARD 3‘. COOLER ' W 1974 r L [B R 1'1 2V1; Micltig :7.n St: to University This is to certify that the thesis entitled A COMPUTER INTERFACED RAPID SCAN STOPPED-FLOW SYSTEM FOR THE STUDY OF TRANSIENTS IN ENZYME REACTIONS presented by Richard B. Coolen has been accepted towards fulfillment of the requirements for Ph ° D0 degree in Chem'i Stry lam: Jam. 0 Major profess” Date June 20, 1974 0-7 639 ##— ’ Amuomo av. ~ ' 1 sons 7 BB?! BINDERY INC LIBHA m e NF-LRS mmmnr MICHIGAN ABSTRACT A COMPUTER INTERFACED RAPID SCAN STOPPED-FLOW SYSTEM FOR THE STUDY OF TRANSIENTS IN ENZYME REACTIONS By Richard B. Coolen A rapid scan stopped-flow apparatus was interfaced to a remote PDPBI computer. The data transmission system permits high frequency (less than lOMHz) parallel digital data transfer over several thousand feet. A frequency multiplication system allows sampling to occur at a constant maximum rate which is independent of scan speed and limited essentially by the computer. The data acquisition software establishes a sampling bandpass which varies with time to produce significant real time S/N enhancement in two dimensions while simultaneously re- ducing absolute storage requirements. A versatile dis- play system facilitiates the analysis of data. An existing rapid scan stopped-flow apparatus was modified to permit the study of changes in the spectrum of substrates, transient intermediates, and products of enzyme catalyzed reactions. Significant improvements in the present stopped-flow system include: (1) variable temperature capability; (2) increased reliability and resolution; (3) a multiple push capability for improved reproducibility; (h) a special flag system which permits \0 [6:2, Richard B. Coolen 6) ‘0 easy calculation of the flow velocity and the location of time zero for the reaction; and finally (5) an improved all quartz double mixer with dual pathlength capability. The overall performance of the instrument was evaluated by using a variety of standard reactions. Mixing is complete within the dead time of the instrument. By studying absorbance changes at 600 nm which accompany the formation of peroxychromic acid, at 600 nm, dead times for the short (1.99 mm) and long (1.85 cm) path lengths were found to be 2.5 and 5.5 ms respectively. Three enzyme systems were examined in order to test the feasibility of doing rapid scanning enzyme kinetics studies with this system, and to illustrate the utility of the teChnique. Rapid scanning of the absorption spectrum during reaction allowed direct observation of the spectra of the Tu and Rh complexes which are formed by the interaction of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and NAD+. The Th complex formed rapidly followed by a first-order growth of the Ru complex. The rate constant .1 for the Tu—9R isomerization was 1.37 sec which is u nearly a factor of four larger than previously reported. No effect on either the Th or R absorption bands or the u rate of R)4 formation was observed when less than satur- ating levels of NAD+ were used. The rate constant for the formation of the Ru complex is independent of wave- length . Richard B. Coolen The enzyme L-threonine dehydrase (TDH) is known to exhibit a peak at u15 nm due to Schiff base formation between pyridoxal phosphate and a lysine residue of the protein. The shift of the absorption to give a new band at u60 nm (composite at ABC nm) upon the addition of L- threonine and AMP has been previously observed and presumably reflects the formation of a Schiff base be- tween the enzyme-bound pyridoxal phosphate and aminocro- tonate, a dehydrated intermediate derived from L- threonine. Toward the end of reaction.when.product form- ation is nearly complete, the transient absorption of the enzyme-bound complex decays back to that of the enzyme alone. This system was studied with the rapid scan system in an effort to detect other possible intermediate species which have been postulated for the overall mech- anism. Other than the h30 nm band, no additional absorp- tions were detected on the short time scale (less than 10 sec.). Product growth is first order, and in.the presence of 25 mM L-threonine the steady-state concentra- tion of the enzyme-substrate complex was achieved in less than 100 ms. An isosbestic point occurs at about u05 nm. A third test system involved the enzyme equine liver alcdhol dehydrogenase (LADH) and its catalysis of the reduction of the substrate analog p-nitroso-N, N-dimethyl- aniline (NDMA) by reduced nicotinamide adenine dinucleo- tide (NADH). The disappearance of the NDMA band at th nm Richard B. Coolen followed Michaelis-Menten kinetics. By following the entire decay scheme and fitting the data with a general non-linear curve-fitting program, values of Km (effective) and Vmax could be obtained from single decay curves. During the decay of the NDMA absorption, the absorption ‘band at 3&0 nm shifted to 335 nm.with a slight increase at the maximum. The 335 nm band then decayed by a first order process with a rate constant of 0.70 1 0.0& sec-1. Coin- cident with the decay of the NDMA band is the growth of tin absorption at 5&0 nm with an isosbestic at about 510 nm. The decay of the 5&0 nm band is first order with a rate constant of O.&&7 1 0.01 sec-1. Since the 5&0 nm decay is slower than that at 335 nm, the two absorptions do not represent the same species. The rapid scanning stopped- flow technique is clearly a powerful tool for the study of transients in enzyme catalyzed reactions. A COMPUTER INTERFACED RAPID SCAN STOPPED-FLOW SYSTEM FOR THE STUDY OF TRANSIENTS IN ENZYME REACTIONS By . rt Richard B.y_ Methods Flow methods were developed in order to permit the study of faster reactions. Initially they were used pri- marily to extend the range of the manual mixing technique. As sensitivity and time response improved, more rapid processes were studied. More importantly however, is the fact that comensurate with instrumental improvements came the discovery of many new phenomena. Frequently, the appearance of transient intermediates or induction phases can be observed only with the most sensitive equipment and only when the reaction system is perturbed far from equilibrium. The development of flow techniques and their application to the study of enzyme reactions will be considered in more detail in a later section. 10 2.1.3--Relaxation Methods Although flow methods have permitted measurements on a millisecond time scale, a large number of important reactions still cannot be approached directly. The classic work of Eigen and his collaborators (17) has provided values for the rate constants of many ionic reactions which are important in the study of biological phenomena. These investigators used the so-called relaxation tech- nique. ‘With this technique, a system at equilibrium is perturbed by changing some intensive property very rapidly (typically in a few microseconds). Relaxation of the system to a new equilibrium position occurs'by first order processes provided the perturbation is small. Most of these fundamental data obtained by Eigen utilized the temperature-jump method, electric field jump technique or sound abosrption measurements. Bi— molecular reactions with rate constants as high as lollliflsecm1 have been measured by these methods. The temperature jump apparatus designed by de Maeyer (18) which used joule heating has‘become the most widely used device for the perturbation of equilibria of biological interest. Although there are descriptions of other tempera- ture jump devices in the literature (19, 20) and in commercial production, all of the improvements of the method which.have been successful and practical came from the work of Eigen and de Maeyer (21). Accurate resolution of very small changes in light transmission and the measure- ll ment of fluorescence changes on the microsecond time scale during chemical relaxation processes has permited the observation of multiple relaxation processes. Although the so-called relaxation spectrum of an enzyme system contains a great deal of information, its interpretation can be very difficult. The great advantage of the relaxation method arises because neither mixing nor transport of the reaction system are required to initiate observation of the changes. Thus the time-limiting features of the flow methods are avoided. A number of important applications of relaxation techniques to the study of enzyme-catalyzed reactions have been made, most notably by Hammes (22, 23). Individual mechanistic processes such as protein conformational changes, the presence of multiple equilibria and the determination of binding steps have all been elucidated with this technique. Gutfreund (21) and Gibson (2&) have reviewed some of the recent successes of both relaxation and flow methods, and give several examples of their appli- cation to the study of transient processes. An excellent example of the complimentary use of flow and relaxation methods is the work by Kirschner‘gt.§l on yeast glyceral- dehyde-3-phosphate dehydrogenase (25, 26). The combined flow—perturbation method (27) can be used in either the continuous or stopped-flow mode. If the turnover is very rapid and steady states can be main- tained only for very short periods, the continuous-flow 12 mode is preferable. However, the heated solution (T-jump) remains in the observation chamber for a very short period so that only very fast relaxations can be measured. The stopped-flow'mode overcomes diffi- culties of this kind, but if the utilization of substrate is not slow compared with.the relaxation, large changes in the steady state concentration of intermediates can make the smaller changes due to relaxation uninterpretable. Despite these difficulties, the stopped-flow temperature jump technique has yielded new information about several systems. The mechanism of ribonuclease is an example (28). Theoretical and practical discussions on the com- bined techniques are reviewed by Human and Hammes (27). Relaxation techniques are generally useful but are somewhat limited in their applicability to the studies of enzyme reactions. Many enzyme-catalyzed processes have very small or very large equilibrium constants, and it is often not possible to produce significant perturbations. In T-jump experiments, processes must be faster than a few hundred milliseconds because of cooling effects. Since the time scale of most relaxation processes is microseconds, only spectrophotometric detection at a single wavelength is feasible. If perturbations are large enough, complex relaxation spectra can result which are difficult to interpret. Indeed, if the phenomenon is wavelength dependent the situation might become hopeless without additional information. 13 2.2--The Stopped-Flow Method In the last ten years several comprehensive surveys of rapid reaction techniqu0s have been published (29-31). Both for historical reasons and because of its wide appli- cation to the study of enzyme reactions, the stopped-flow spectrophotometer has received special attention. For a detailed discussion of the development of the stopped-flow method itself the reader is referred to the comprehensive review in the Ph.D. thesis of Papadakis (32). Although :many minor‘improvements on the simple design of Gibson (33) have been introduced (30, 3h), no significant improvement in overall performance has been achieved in recent years. Berger eg_gl,(35) have described a stopped-flow device which.provides a time resolution of about 250.0s. However, this instrument has so far provided very limited data. The only flow method which has been used success- fully for the study of reactions with half-times of less than 1 ms is the continuous flow device of Chance 3}; §_l_. (36). The range of physical parameters and the sensitivity of their detection provide the widest scope for the hm- provement of flow methods. Split-beam and dual wavelength systems similar to the one designed by Hess (37) for'the measurement of the spectral characteristics of transients in the stepped-flow mode have been described. Systems capable of rapidly scanning the absorption spectrum during 1h reaction have been reported (38); however, none of these systems have been applied to the study of enzymic processes. Computers and transient recorders have proven useful for improving the overall reproducibility, sample throughput, and data analysis. More importantly, they permit auto- matic collection and processing of data obtained during the observation of transients. The resolution of second-order processes involved in enzyme-substrate and other ligand-macromolecule interactions ‘which.characteristically approactheMmlsec“1 depends on the detection of the reaction in very dilute solutions. Fluorescence changes have proved especially valuable for this purpose. A number of designs similar to those of Gibson (39) and Fernley and Bisaz (&0) have been used with.only minimal modification of the transmission optics of the stopped-flow apparatus. Second-order rate constants approaching 109M":lsec'1 have been measured on the milli- second time scale (39). There is no doubt that a stopped- flow spectrofluorimeter built for the analysis of fluore- scence spectra of transients will provide more information than merely increased resolution of concentration changes. Many other physical sensing devices have been used for following rapid reactions and characterizing transient intermediates. However, each.has only been applied to one or two systems (39). The use of thermocouples for recording reactions in the stopped-flow mode has been 15 reported (&l); however, it has not been applied to bio- chemical systems. Rapid quenching techniques have been developed (A2); however, the necessity for analyzing a large number of samples individually has limited its application. The use of automated analysis may make this method feashble for the study and isolation of stable transients. Sirs (&3) and Prince (&&) used stopped-flow measure- ments of electrical conductance and pH respectively for the study of relatively slow enzymic processes; and Chance (&5) and Clark (&6) used platinum microelectrodes for following changes in oxygen concentration in a stopped- flow apparatus. Although flow systems have been developed for use with an ESR apparatus (&7) no design has as yet appeared for carrying out stopped-flow work with aqueous solutions. The relatively long response times and dead volumes of most thermal and electrochemical detection devices has limited their application to very rapid processes; con- sequently, spectrophotometry remains the most versatile and most popular detection system for stopped-flow studies. l6 2.3-11he_§eanning Stepped-Flow System The rapid scan stopped-flow system developed in this laboratory has been applied successfully to the study of reactions of the solvated electron and of aromatic radical ions in non-aqueous media (&8-50). A detailed discussion of the development of this system was recently reviewed (32), therefore only a general description will be pre- sented here along with some of the major shortcomings. Additional information may be found in a later chapter of this thesis which is devoted exclusively to the modifi- cation of the previous apparatus. The system has undergone more or less continuous development to accomodate a variety of problems associated with either the acquisition, storage, and processing of scanning data, or to the sensitivity of reactive compounds to air and moisture (51). The reactive nature of solvated electrons and aromatic ions towards air, moisture, and stopcock lubricants necessitated the development (51) of a vacuum-tight, greaseless stopped-flow system in which the solution comes into contact with only glass, quartz, and Teflon. Construction of the mixing chamber and obser- vation cell was based upon the extensive developments in this field by Chance (52), Gibson (53, 5&), and others (30). Vacuum-line techniques of solution preparation were required so that dilutions could be made in a closed system. After 17 experhmenting with Plexiglass mixing cells, a method was devised for the construction of an all-quartz precision- bore mixing and observation chamber. To completely exclude air from the system, the syringes used for push- ing and stopping were specially constructed to permit back-pumping between two O-rings. A Perkin-Elmer Model 108 Scanning Monochromator is used to scan the spectrum during reaction. This permits one to scan the complete spectrum (limited only by the light source and the detector) at a rate of up to 150 spectra per second. or course, the signal-to-noise ratio (S/N) is less favorable under scanning conditions because of the increased signal bandpass which is required and because of vibrations of the monochromator system. How- ever, the ability to determine changes in the shape of the absorption spectrum during reaction has proved most valuable. Once the general features of the spectrum have been studied, it is possible to increase the S/N by studying the absorb- ance at a fixed wavelength. Reactions with half-lives less than about 15 msec should also be studied at fixed wave- length. The introduction of a scanning capability has not been without its problems. One of these is the wide dynamic range of the output voltage from the photomulti- pliers during a scan. The variations of both the lamp output and the photomultiplier sensitivity with wavelength 18 make it not uncommon for the signal to change by a factor of 100 or more during a scan. To overcome this problem and also to provide a.more convenient display, double-beam techniques together with a logarithmic operational amplifier have been used, so that intensity is converted to absorbance . Because of the rapidity with which data are collected, it is necessary to store data during a run. The previous scanning stopped-flow system used an Ampex SP-3OO FM tape recorder to initially store the data. The analog signal stored on tape was then converted to digital form for computer analysis by using an off-line Varian computer of averaged transients (CAT) coupled to a card-punch. The instrument response as a function of wavelength was calibrated during each run with didymium or holmium oxide glass filters as well as neutral-density filters. The entire system was calibrated from time-to-time by using standard solutions and standard reactions. Path lengths of 1.0 and 2.0 mm have been used and the absorbance follows Beer's law up to an absorbance of at least two (87). Computer programs have been written to correct the raw data for variations of the instrument response as a function of both absorbance and wavelength. Although the FM tape recorder was far more versatile and convenient than the earlier oscilloscope - camera combination, there remained several disadvantages. Chief 19 among these was the noise which was introduced by the tape recorder. This made it difficult to obtain good data, especially near the end of the reaction. Another major disadvantage of the previous system was the inability to rapidly analyze data during a run. With a scanning system it would be an advantage to be able to test for the pre- sence of intermediates or side reactions and to alter conditions to make use of this information. Finally, the information collected during a run of one or two days' duration required several weeks to edit and transfer to punched cards. For all of these reasons, the data collection system was modified to permit use of a PDP 8-1 digital computer for data collection and handling. The stopped-flow method has proven to be an effective tool for the study of a variety of enzymic reactions. At present, only spectrophotometric detection is capable of covering the broad range of reaction rates and signal levels that are encountered in kinetics studies. The ability to either scan the spectrum during reaction or to study absorbance changes at fixed wavelength has proved extremely valuable and should be equally valuable in the study of enzymic reactions. 20 2.h--Data Anal sis Techni ues for the 3tudy of EEzyme §inetics 2.h.lvglgitial Rate Method Ewen the simplest enzyme catalyzed reaction mechanism gives rise to complex rate expressions which are coupled non-linear differential equations that have not yet been solved in analytical form. Most studies in enzyme kinetics avoid this problem by using only "initial rate" data to- gether with three simplifying assumptions; (1) steady-state is maintained for all enzyme-substrate complexes during the initial rate period, (2) inhibition due to product formation is insignificant at early times, and (3) only a small fraction of the substrate reacts so that the substrate concentration may be assumed equal to its initial concentration (80). with these assumptions it is possible to derive initial velocity expressions that are functions of only kinetic constants (Km, Ymax’ etc.) and initial concentrations (30, 30). These functions of kinetic constants become more complicated as the reaction mechanisms become more complex; however, the equations relating the initial velocity (Yo) to initial substrate concentration still have a simple form. Values for the kinetic constants (Km, Vhax, etc.) are most often obtained graphically (55) by using simple plots of l/vi versus l/SO. Although the initial rate method is simple, it has several disadvantages: (1) many runs are needed to construct a single l/Vi versus l/So plot; (2) "initial rate" analysis provides no infor- 21 mation about progress curves for the enzyme-substrate complex. (3) the steady-state assumption may be invalid, and the size of this type of error is often not evaluated (56, 57) (h) in order to differentiate between.mechanisms both forward and reverse reactions must be analyzed (58, 59). In general, for a given reaction the "initial rate" method cannot determine as many rate constants as the method of full-time-course integration (56). Indeed, whenever Etc/So is not small, or the enzyme mechanism involves regulatory processes, the "initial rate" method cannot be applied. Approximate solutions to enzyme kinetic equations have been obtained for some mechanisms (60, 61) using primarily the three techniques which are briefly outlined below. 2.h.2--Approximate Solutions 2.h.2.l--Algebraic Solutions After Simplifying Assumptions It is possible to obtain explicit algebraic solutions by assuming (1) steady-state on all intermediates, (2) rapid equilibrium preceding further steps, and (3) the pseudo-first-order approximation. These assumptions linearize the differential equations and render them solvable. However, such assumptions are not always Justifiable, and the errors which they introduce are often not specified. 22 2.h.2.2--Pertubation Schemes Explicit algebraic solutions can also be obtained by using Self Consistent Well-Ordered Perturbation Schemes. In this technique a perturbation parameter is defined, which, as it approaches zero, linearizes the equations and makes them solvable (for an example see Heineken (56)). The concentration variables are expressed as Taylor's series. The series are then substituted into the differential equations and the series coefficients are determined. As long as the series converges, an exact solution is possible and errors can be estimated. Accuracy can be improved by including higher order coefficients. The perturbation scheme can be used to determine the full- time-course of the reaction for all species. Although the technique is quite general in scope, it has not found widespread application in enzyme kinetics, probably because of the complexity of the equations obtained. 2.h.2.3--Numerical Solutions - (b1 Newton, Rungg,Kutta, n e- rlo or analog methodsIf By far the most general method for solving complex differential equations is by numerical integration with the aid of a computer. This method will always work and it is simple to use no matter how non-linear the differ- ential equations are. Although it yields no explicit algebraic solution, and computing costs can be considerable 23 for some cases, the numerical integration technique pro- vides all of the information available from full-time-course integration, without confining the solution to any set of artificial experimental or mathematical assumptions. 2.h.2.h--Present Study Since it is not the goal of the present work to develop analytical solutions to differential equations, all experimental data are fitted over their entire time course by using appropriate analytical functions or by numerical integration of selected differential equations. Data analysis is accomplished with the help of a general non-linear weighted least squares computer program KINFIT (62). Appropriate statistical information permits esti- mation of the errors, and "goodness of fit." Other data treatment schemes are available to calculate absolute absorbance, time, variance etc. and are described elsewhere (63). No general procedures are available to accomodate the presence of several time-dependent phenomena at a single wavelength. It is not difficult to discern the presence of wavelength dependent rate phenomena; that is, check for the presence or absence of "clean" stoi- chiometry (6h). It is another matter however to resolve complexity when data which do not show "clean" stoichiometry are fitted to selected rate expressions. 2h 2.5--Computer Systems The fundamental advantages of computer interactive instrumentation are manifold and well established. Frazer has presented a review on the general use of small computers in the laboratory (65). Others have presented discussions of terminology and practice in the design and use of digital systems (66-68). In fact, the digital computer has become a common instrument in many laboratories and it is no longer necessary to describe its operating characteristics or the fundamental interfacing procedures in detail. It is not generally true however, that existing on-line systems are currently being used to fully exploit the capabilities of computerized instrumentation. In the present work, a computer interactive system for acquiring and processing data from the rapid scan stopped-flow apparatus is described. A number of inter- ) active on-line stopped-flow systems have been reported (69-71) which were designed primarily for analytical appli- cations at a fixed wavelength. Pardue (72) and a commercial concern (73) have described computerized rapid scanning systems which.utilize silicon-vidicon detectors. The performance of these systems, however, has not been tested on enzyme systems which demonstrate spectral intermediates in the ultraviolet region of the spectrum. The previous data processing system for the scanning stopped-flow apparatus has already been described (87), 25 and some of its major disadvantages were discussed earlier. Since the computer-interfaced system is presented independ- ently in a later chapter, the reader is referred to that section for more details on its operating characteristics. Although the computer system was designed and con- structed specifically for this project, it utilizes typical hardware components and general interfacing and software techniques wherever possible. The hardware and software systems are described in detail in two laboratory manuals (63. 7h). 2.6o-Transient-State Enzyme Studies Three enzyme systems were chosen in order to test the rapid-scan stopped-flow system. Although it was necessary to work at wavelengths above 300 nm, this restriction will soon be removed. The systems chosen and the reasons for their choice are as follows: 1. Complex formation between the co-enzyme nicotin- amide-adenine dinucleotide (NAD+) and the enzyme yeast glyceraldehyde 3-phosphate dehydrogenase (GAPDB) has been extensively studied by Kirschner (25, 26) who showed that a conformational change occurs upon binding accompanied by the growth of an absorption band at 350 nm. We chose to study this system because of its well-known spectral changes in an easily ascessible region. The system provides a good test of sensitivity since a maximum absorbance 26 change of only about 0.08 absorbance units occurs on a rather large background. Also, the necessity to work at uo'b provided a good test of the thermostat system. 2. A large number of enzymes show complex absorption spectra in the visible region because of Schiff-base formation between the enzyme and a co-enzyme. The enzyme L-threonine dehydrase (TDH) of E. coli has an absorption band at AIS nm because of Schiff-base formation between pyridoxal phosphate and lysine residue of the enzyme (5,75). The addition of L-threonine in the presence of AMP causes a transient shift of the absorbance to a new band at u50 nm during the formation of product. Because of the very low turnover number even at high enzyme con- centrations, it is difficult to follow both the rapid for- mation of the complex and the growth of product. However, the averaging scheme available with our system is ideally suited for this purpose. This cooperative study with w. A. Wood of the Biochemistry Department also gave us the opportunity to find out how much information could be obtained with a small amount (6 ml) of solution and less than 10 mg of enzyme. The wealth of information available from a single scanning push in this preliminary study indicates the power of this new instrument for transient-state studies. 3. A third test system for the scanning stopped-flow instrument which proved to be most interesting involved the enzyme equine liver alcohol dehyrogenase (LADH) and 27 its catalysis of the reduction of the substrate analog p-nitroso-H,N~dimethylaniline (NDMA) by reduced nico- tinamide adenine dinucleotide (NADH). This system was selected for several reasons: a) At least one transient species is formed which has an absorption spectrum distinct from those of reactants and products and which appears to form in excess of the active site concentration. The similarity of its spectrum to that of NADH has led to its assignment as a complex between NADH and the substrate analog (109). b) The time course of disappearance of the substrate analog and the intermediate are different which affords an opportunity to use the scanning stopped-flow system to study the kinetics at a number of wavelengths for a single push. The absorption bands occur in a convenient wavelength region. c) The enzyme LADH is well characterized and commercially available in satisfactory purity. Additional background information for these systems will be discussed in the Results section. III. EXPERIMENTAL The following experimental details refer only to the runs with.enzyme. Expertmental details for the other systems which were used for performance tests are given elsewhere(32). 3.1--Glassware All glassware, including the flow system, was soaked overnight with 2N HCl, and subsequently ex- tensively rinsed with quartz distilled conductance waters The glassware was then rinsed with the appropriate buffer solution in preparation for the run. Stock solutions were made up in Pyrex bottles which were equipped with Teflon valves (Kontes Co., Vineland, N. J.) and the appropriate joints (Fisher Porter’S mm Solv-Seal) for connecting the bottles to the input ports of the flow system. Stock solutions were degassed under vacuum and subsequently re-pressurized with purified heliUI. This process helps to reduce foaming which.can lead to denaturation of the enzyme solutions. 28 29 3.2-1glyperaldehydeeggphosphate dehydrogenase Crystalline yeast glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was prepared essentially as described by Kirschner and Voigt (76) with only a few minor modifications. These were associated with the application of the charcoal treatment‘gftgg the first fractionation.with ammonium sulphate and before the protamine sulphate step (as suggested by Kirschner (25)) to remove bound nucleotides. The preparation attained a value of E280/E26O = 2.05 i .07 compared to Kirschner's value of 2.15 1 .05. The enzyme was assayed in the direction of formation of 3-phosphoglyce- rate using essentially the assay technique of Stancel (77). An average specific activity of 133 Units/mg was obtained which is substantially larger than the value of 80'Units/mg reported by Stancel. The concentration of GAPDH for kinetics studies was determined spectrophotometrically using both Eggo = 0.89h m1 mg"1 cm"1 for the pure enzyme and the Biuret method for total protein. The enzyme was electrophoretically homogeneous; however, the kinetics results indicate that the fraction which was used contained components with different kinetic properties. GAPDH from the DEAE Sephadex column was dialyzed against the buffer just prior to use. The buffer (pH 8.5) for all solutions was 0.05M sodium pyrophosphate, 5 mM EDTA, 5 mM sodium arsenate, and 3O 0.2mM dithiothreitol. Nicotinamide adenine dinucleotide (p- NAD+, grade III, Sigma Co.) was used for the kinetics studies. D,L-3-phosphoglyceraldehyde (GAP) was prepared from the barium salt of the diethylacetate (Sigma - Dowex pre- paration), and assayed using NADH production at 3h0nm to determine the concentration of the D enantiomer for kinetics experiments. Concentrations of GAP refer to the D enantiomer only. In general, concentrations refer to the concentration of the species upon reaction (i.e. after mixing). Solutions for the GAPDH runs were equilibrated to 39.9 I .1°C in the bath prior to reaction. Additional information is given in TABLE I. 3.3--Threonine Dehydrase Threonine dehydrase (TDH) from E. coli was a gift from Professor Wood (78). TDH concentrations were deter- 95 .. mined spectrophotometrically using Eicm,h66 — 1.75 (5). A specific activity of 395 Units/mg was determined from 1 1 the production of a-ketobutyrate at 310nm, 6 = 20.1... M- cm- (5) in an assay solution containing 1 ml of 50mM.L-threo- nine in buffer, pH 8.0, 28 °C. Stock enzyme solutions were made up in 0.1M sodium phos- phate buffer pH 8.0 containing lmM adenosine monophosphate and lmM dithiothreitol. In a typical push, solutions of TDH (O.7h.mg/ml final) and L-threonine (25mM) equilibrated 31 to 2h.5‘h were rapidly mixed in the stopped-flow. Sampling parameters for the TDH run are shown in TABLE I. 3.h--Liver alcohol dehydrogenase The commercially available compounds acetonitrile (Baker Chemical Co., reagent grade), p-nitroso-N,N-dimethy1- aniline (NDMA, Aldrich Chemical Co.), and nicotinamide adenine dinucleotide (NADH, Sigma 00., grade V) were used without further purification. The concentrations of NDMA, NADH and LADH stock solutions were determined spectrophoto- metrically using extinction coefficients of 6M0 = 3.51;. x 10"6 M'1 cm-l, 6 3&0 = 6.22 x 103 14-1 cm'l, and £280 = O-MSS m1 mg-1 cm"1 respectively. Solutions were made up in 0.05M sodium pyrophosphate buffer, pH 8.75. NDMA stock solution was prepared by dilution with buffer of an aliquot of 1 mM stock solution of NDMA in acetonitrile. Enuine liver alcohol dehydrogenase (LADH, Sigma Co.) was further purified by overnight dialysis of a concentrat- ed LADH solution against several liters of buffer. A specific activity of 22.9 Units/mg was determined at th nm using an assay mixture of 3.18 X lO'SN NDMA, 8.18 X lO'SM NADH, and h.9 X 10'3 mg/ml LADH in buffer. Rapid mixing experiments were carried out at 23 i .l‘%. Appropriately buffered solutions of NADH (h.09 X 10"5 N after mixing) and NDMA (1.59 X lO'SN) were premixed (just prior to reaction) and subsequently reacted with LADH (3.62 x 10'6n). Refer to TABLE I for the 32 sampling parameters. 3.5--Sampling Parameters The RCA 6903 photmultiplier tubes were used for all three enzyme runs. Glass filters and copper sulphate solutions were used to eliminate unwanted UV and IR radiation. A path length of 1.85 cm and a scan speed of 75 spectra per second were used for all three enzyme runs. The monochromator gives both a forward and a reversed spectrum for each scan. Since only foward scans were collected, the actual scan rate was 37.5 spectra per second (is. the duration of each spectrum was 13.33 ms and the time between spectra was 13.33ms). A slit of 0.5mm was used for the GAPDH run, and 0.6mm was used for the TDH and LADH runs. The remaining sampling parameters for each of the enzyme systems are given in TABLE I. 3.6--Data Analysis Results from the enzyme runs, stored digitally on magnetic tape, were first examined extensively with the display scope. Subsequently, the time development at several selected wavelengths was output for each push to punched cards. The raw data were submitted to the CDC 6500 along with program ABSTIM (63) for additional processing. The ABSTIM routine converts the raw data to absolute absorbance using the values for the 33 TABLE I Sampling parameters for the GAPDH, TDH and LADH Runs. Sampling frequency=20.4KHz Sampling GAPDH-NAD+ GAPDH Parameter Complex Reaction with TDH LADH Formation Substrate Number of Groups 6 5 7 6 Samples/Point 4 4 6 5 Spectra/Group 8 10 12 12 Points/Spectrum 64 64 4O 48 Grouping Factor 2 2 3 2 3h calibrated neutral-density filters in the same way as program PUNDAT (79). Each push has associated with it all the sampling parameters (SAVPAGE). Thus, the time for each point can be determined, and appropriate adjustments for flow and stopping times can be made. Variances are computed from estimates of the standard deviation, taking into account the degree of averaging within each group of points. Finally, data are re- punched in a format suitable for direct submission with the KINFIT program. IV. THE COMPUTER SYSTEM: DESIGN CONSIDERATIONS AND CONSTRAINTS The following chapter is essentially in the form of a manuscript which.will be submitted for publication. The stopped-flow technique involves the rapid mixing of solutions from two syringes in a specially designed mixing and observation cell followed by rapid stopping of the mixed reactions by a third syringe located downstream. In our case the extent of reaction is monitored spectrophotometrically either at a single wavelength or by rapidly scanning the absorption spectrum. Rapid scanning stopped-flow kinetics presents special problems to the experimenter which are not encountered in fixed wavelength work. Since the duration of the experi- ment ranges from a few seconds to several minutes and the monochromator can scan at speeds up to 150 complete spectra per second, an enormous amount of three dimension- al data must be rapidly processed and stored. Actual stopped-flow runs require only a small fraction of the total experimental time, so that a dedicated computer is not necessary. Therefore, although the stopped-flow apparatus was located in the basement, it was interfaced to an existing computer system located on the fourth floor. 35 36 Thus, needless duplication of expensive equipment, parti- cularly the peripheral Input/Output devices was avoided. A remote parallel digital data transmission system of this type should be generally useful where the availability of small computers is limited and high frequency data trans- mission over relatively long distances is required. h.l--Signal Enhancement in Real Time In typical rate studies, the signal varies rapidly at first and then progressively more slowly as the reaction approaches completion (or equilibrium). It is therefore desirable to use a wide bandpass at the beginning of a reaction but this necessity decreases as time increases. Most systems are designed for a fixed bandpass; that is, filters are selected prior to the experiment and then not changed during the experiment. Ryan in these cases, one must be careful in selection of the filters since analog filtering can distort the signal shape. In contrast to this, digital filtering (appropriate averaging or smoothing) can provide a bandpass which varies with time as needed. Narrow slitwidths resulting in decreased light in- tensities, coupled with the mechanical vibrations of the monochromator are the primary causes for the inherently poorer S/N associated with scanning experiments. This is particularly troublesome in the case of enzyme reactions where absorbance changes are generally small and spectral perturbations are often subtle. 37 Optimum on-line use of the mini-computer is achieved by sampling the data signal at a maximum constant rate which is limited by the computer software and not by the expertment (80). Data storage requirements are reduced and simultaneous real time S/N enhancement in both.the wavelength and time domains is achieved by establishing a variable bandpass with the system software. h.2--Determination of the Maximum Sampligg' 3233 In the rapid scanning stopped-flow experiment full scale signal deflections occur on the millisecond time scale as a result of changing wavelength and extent of reaction. With traditional fixed wavelength.systems some kind of clock oscillator provides a convenient sampling frequency which rmmains constant during the entire reaction. More sophisticated instruments have utilized programmable clocks to decrease the sampling rate as the reaction proceeds, thus reducing the quantity of stored data. These approaches were rejected for several reasons: (1) asynchronous sampling would preclude the ultimate effectiveness of spectral averaging, (2) the wavelength variable continues to change on the milli- second time scale even at equilibrium; hence, reduced sampling rates would lead to poorly defined spectra, and (3) decreased sampling rates near the end of reaction ignore some of the most vital data for fitting the kinetic record to complex rate expressions. In the 38 present work the guiding principle was to use the maximum sampling frequency and to perform appropriate digital averaging. This results in real time S/N enhancement and simultaneously reduces absolute storage requirements. It was estimated that approximately h0,us would be required by the computer to complete the data acquisition, averag- ing and storage cycle. Thus the maximum sampling frequency was approximately 25 KHz. This sampling frequency is well within the capabilities of a moderately priced medium speed A/ D converter. We also wanted to synchronize sampling with the nutating mirror of the monochromator. Wavelength.is scanned by the change in angle (relative to the light beam) of this mirror as it rotates. The gear attached to the mirror shaft has 136 teeth which interrupt a light beam and produce a signal (OT) whose frequency varies with the scan speed up to a maximum of 10.2 KHz at 150 spectra per second (see TABLE II). Once during each revolution the light beam is reflected from a polished tooth to produce the beginning of scan (BS) pulse. The GT signal is input to a frequency multiplication system which ultimately generates the trigger used for sampling. In this way, a sampling frequency of 20.h.KHz is readily ob- tainable at any scan speed and is compatible with the time requirements of both the computer and the ADC. 39 comm oom moa mom m.mmm m.a o.m ovum omH m.nma mmm >.©mm mnm.a mn.m oovm ooa mom mow n.moa o.m 0.0 omnm om mmm oam m.mmH mn.m m.n coma cm can omoa hm.mm m.n ma 0mm om ONOH oeom mm.mm ma om OVm 0H ovom omov hm.wa om om mum m ommm ooam mm.ma m.nm me one m omov omam mm.m om oma mma e ooam oomoa no.0 mm oma Esupoomm Hem Azi mmHmEmm uouomm moan mfiam um no t a .uaszkwm BU .vonm Bo Tsuuowmm\omme oem\>mm omm\mcmom .me v.0m u mocmsvmum mcHHQEew usmuoo .muoumEeuem mooq poxooqlomenm one uoumeonnoocoz mCHGCmUm HH Mdmdfi no h.3--Averaging_Scheme Stopped-flow data must be stored initially in computer core because of the time scale of the experiment and the relative slowness of peripheral computer storage (magnetic tape). For a given push, data storage is limited to whatever space remains after resident system software has been loaded. This fact was of primary importance in the development of data acquisition software for the scanning system. At a sampling fre- quency of 20.h KHz one would rapidly exhaust the storage capacity of the 8K computer. While a bandpass which simply increases with time in a pre-determined way would be satisfactory for fixed-wavelength data, it is not applicable to scanning data since the signal changes with time in a way which is determined by the spectrum and the scan rate and not merely by the rate of reaction. We have devised a scheme which operates equally well in either mode. For simplicity its operation with fixed- wavelength data will be considered first. Consider an absorbance peak which decays as a function of time (see Figure l for example). Initially the signal is changing very rapidly with time and excessive averaging would distort the decay curve. At long times, when the signal is slowly changing with time we would like to average a number of adjacent samples (not simply hl decrease the sampling rate) in order to increase the S/N in the very important "tail" of the reaction. Thus, what is needed is a variable frequency bandpass as the reaction proceeds. The averaging process chosen is best represent- ed by an example. Consider a reaction in which the absorbance decays in a second order fashion with a half-life of 50 ms. Suppose we have an absorbance with an initial S/N of 100 (based upon four samples per point). As the reaction proceeds through six half-lives, the absorbance decreases to 1.56 per cent of its initial value with.a concomitant decrease in S/N expected if the number of samples per point were not changed. In the averaging scheme selected for this example, the total time is arbitrarily divided into 10 groups of 16 data points each (see Figure 1). In an actual experiment the number of groups and the number of points would be pro-set. These parameters are chosen according to the rate of reaction and the number of half-lives to be measured. A grouping factor is selected which determines the modulus by which the averaging band- pass is increased as the reaction proceeds. Each of the stored points in Group 1 is actually the average of four h9,Us samples (20.h KHz) and the first group extends in time to 3.1h ms of reaction. The second group also con- tains 16 points, each of which is the average of eight h9 ,us samples.) Group 2 extends in time to 9.h.ms. In this 42 .emmu camcoam>o3 omme e new oEoSUm moammum>e m5» m0 meonm .H muomfim Tommy NEE. 06. A!» 04“ hum n; o; mAv _ )I‘ 7 1 m - _ _ q 00.. I/ \\ma\\ xx/ 36 III“\.\ l/(\ . “mng o. 3N0. Va 0 N. mu on < o\e $.va I / Emma A¢v_ AMM- 00. #3 way each group contains twice as many samples per point as the preceding group and the total elapsed time doubles. For an arbitrary grouping factor, g, group number, n, elapsed time at the end of the group, tn, and samples per point an we have - n-1 8n - g 31 (la) _ n tn - (g ~1)t1 (1b) After six half-lives only 160 points are stored; however, 65,h72 samples are actually measured. One obvious advantage of the technique is that the total number of stored locations is reduced without ignoring useful in- formation. Secondly, we can expect significant S/N enhancement. For a random, constant noise level we have for the signal in group n (Sn) n-l (s/N)n = 37 . (s/N)l . 3‘11 (2) For example, at the ends of the first (Group 5) and sixth (Group 10) half-lives we can expect S/N values of 200 and 35 respectively. Because of this S/N enhance- ment, small changes in the signal at long times can still be detected. Obviously, for very fast processes little can be done to improve the S/N in real time since (7' ’30} 328 the an excessive averaging would lead to signal distortion. The ability to scan wavelength during reaction adds another dimension to the averaging problem. In this case the number of points per spectrum (P/S) is selected by averaging an appropriate number of adjacent samples as the spectrum is scanned. Once this selection is done the averaging bandpass for the wavelength variable re- mains fixed (for that reaction) while averaging in tipg_ proceeds according to the scheme outlined above for fixed wavelength. For moderately fast reactions involving broad featureless spectra, the S/N is enhanced by averaging more adjacent samples. As wider wavelength regions are scanned or as the complexity of the spectrum increases, the number of points per spectrum required for adequate resolution also increases. As the number of points per spectrum increases however, the total number of spectra which can be stored for a given push decreases. The averag- ing bandpass for the wavelength variable is, therefore, determined by the reaction conditions, wavelength resolution desired and the amount of storage available. Averaging in.£ipp_is accomplished by averaging appropriate numbers of consecutive spectra. Analogous to the fixed wavelength case, once the number of samples per point and points per spectrum are determined, the time course of the reaction is divided into groups. 145 Each group contains the same pre-determined number of spectra. Each of the "stored" spectra within.a parti- cular group actually represents the average of c con- secutive spectra where c is determined by the group number n and grouping factor g according to (3) n-l on = g (3) The total number of samples averaged into each point of the stored spectrum is given, as before, by equation 1a. The grouping factor has a marked effect upon the total time span. For example, choosing a grouping factor of 2 with 10 spectra per group and 5 groups yields a total of only 50 stored spectra out of 310 measured spectra. For a grouping factor of 3, the 50 stored spectra would result from 2h30 collected spectra. Obviously when the pimg_development at a particular wave- length is extracted from the scanning record, times appropriate to the averaging scheme must be used. A software-generated variable bandpass averaging scheme is extremely versatile, limited essentially by the experimental conditions and/or data storage capacity. There is a trade-off between resolution and S/N enhance- ment on the one hand and storage limitations on the other. The relationship among the sampling parameters for the available storage in our system is given by h6 (P/S) (S/G) (no) 5 315610 (u) for P/S points per spectrum, S/G spectra per group and NG groups. One of the primary advantages of the variable bandpass technique is that both rapid and slow processes can be recorded in the same experiment. Early groups contain the rapidly changing information while later groups continue to follow slow processes. Even reactions with induction periods can be effectively studied since any number of spectra or samples can be skipped before data collection begins. h.h--§ystem Control and Timing On line experiments require certain instrument generated triggers or timing pulses in order to synchronize sample acquisition and storage with the various components of the system. In our scanning stopped-flow system, trigger signals are provided which mark the flow start and stop,the beginning of each spectral scan and the wavelength. These signals are necessary because the three main components of the system (monochromator, stopped-flow apparatus and computer) operate asynchronously. A common problem in stopped-flow measurements is the determination of the exact zero of time for the reaction. This problem had been particularly troublesome in our scanning experiments when fast reactions were studied #7 because the time at which flow stops had no relation to the wavelength of the spectral scan. One must have an accurate knowledge of the zero of time in order to establish initial absorbance values at all wavelengths and to avoid using data collected during the period of deceleration. In order to directly compare replicate pushes in the scanning mode therefore one must deter- mine precisely when flow stopped with respect to any particular wavelength. This is accomplished by using trigger signals from the stopped-flow apparatus. Two metal flags attached to the stopping plunger break light beams and cause light sensitive transistors to give the start and stop pulses shown in Figure 2. The start pulse is used to start both the Time Shift Clock and the Flow Velocity Clock (Heathkit 00., Universal Digital Instru- ment). The Time Shift Clock utilizes the computer real time clock (81) to measure the time between the start flag and the next beginning of scan (BS) trigger. ‘When the stopping plunger is at a measured small distance from the stopping plate, the stop pulse occurs and stops the Flow Velocity Clock. Since the separation of the two phototransistors is known, the flow velocity can be calculated from the start-to-stop time interval. Measurement of the flow velocity profile showed that the velocity is constant between the start and stop pulses. From Figure 2 it can be seen that knowledge of the flow 4s ucmEHummxo mQHCCmUm o How Emnmmao madame .m musmam H _ aOPm _ _ r us: _ 1 “.05 .525 It" _ _ fiQ .w — _ Hmdkm _ Jul... _ mm We) _ d c. a _ _ . _ _ _ _ _ toot; _ w; 33... szfimzou (I1 do SEE 304m 2.0mm «58% \ oz gown Ix aOh m 30..“— 443.54 1L9 velocity and the timing information permits computation of the zero of time and the corresponding wavelength. Data acquisition begins with the first sample pulse after the Time Shift Clock is stopped, always at the beginning of a new spectral scan. For fixed wave- length pushes, a Wavetek model 116 Signal Generator is used to provide the 20.h KHz sampling trigger and the stop flag is used to initialize data acquisition. Once within the sampling mode the system software takes con- trol until sufficient data are collected as determined 'by the variable bandpass averaging scheme described above. h.5--System Hardware Descriptipp A block diagram of the overall system is shown in Figure 3. The interface system was constructed from Heathkit modules and printed circuit blanks. Most of the individual circuits such as data latches, octal decoders etc. were constructed on Heathkit printed circuit boards using generally available IC components. The sampling system utilizes an Analogic MP 250 fast acquisition Sample and Hold Amplifier and an MP 2212 12 bit ADC. The display and plotting system uses Analogic MT 1810 high.speed 10 bit DAC's. A Digital Equipment Corporation PDP 8-I computer with.8K of core was used. Computer Input/Output peri- pherals include Teletypes, high speed paper tape punch/ 50 soawnowgaowm maficumum oeommaovcw popsaeoo esp hzwawm.w0w¢ m¢w>.zn 2+ Boos 3m” sexezoo . e. M l ,< Be: one wee 03:. on eagmeaexs N mzu>_¢o V-n 32c: 15mm- =3 8.53 pin QED _ oo— 2 3| 1 act. e 10 O a u no to. 2.3 3:20.05 .920 U . to. 500:» dictatotoa .._ A: IOZDQ O¢xoaoQ mo coHuoanm on» now coauera ecu unease pouuoaaou ouuooom .m ouomam .65 59.2063 Nan wmv mun own — a _ q eee so as m e eeeeee a a4 e e eouoqmsqv 66 .mocausou headmap may no moaufiafinoaou one mcasonm xoaoeoo +mzumom may mo COaDmEHOM one you muscloefie .m musmwm m2_b _ o e e e eeeeeeeee ee eee'ee to Dmo. e e e eeeee as e e e e e e e e e e 000. $e M V e e e e eeee a V e e e eee ~ m e ee 0 e e as w Mu e @¢._ n W m. u. m case M 3 e e e e e e 0 LF \e N. e a. _ e e e e e e e e e e e e e e eoeoeeeeoeeeeeeeeoleeelete’\\ 1Q— 67 expansion of both.the time and absorbance scales. This scale expansion procedure is useful when examining flow velocity and stopping characteristics of the system, fast and slow processes, large and small absorbance changes etc. Similar flexible display capabilities are utilized for both wavelength and time displays of data obtained in the scanning mode. The entire system has been used to study the kin- etics of a number of complex reactions involving such diverse fields as transient-state enzyme kinetics and solvated electron reaction rates. The performance characteristics have fully lived up to the expectations developed during the calibration runs. As an example, the spectra obtained during the equine liver alcohol dehydrogenase (LADH) catalyzed reduction of the sub- strate analog N,N - dimethylnitrosoaniline with.NADH, are shown in Figure 10. Three wavelength-dependent time developments have been extracted and analyzed from data obtained in the scanning mode. Details about the construction and performance characteristics of the stopped-flow system are given in the next chapter. 68 .moaz an «2oz mo c3663 poumaouoo mama as» ea mooseno Houuoomm Haouo>o 6&5 £mc0_o>e>> 000 one can own 0N0 q _ .e.‘ 11""; \ .OH musmam N eoueQJosqv v. THE STOPPED-FLOW SYSTEM: DESIGN, EDNSTRUUTIGNTKEEFFETFfifiI The design of the scanning monochromator-stopped- flow system which had been previously used for studies of reactions of solvated electrons and of aromatic ions in non-aqueous media was used as a basis for the present system (51). However, it was modified to permit the study of changes in the absorption spectrum of enzymic systems and to take advantage of the computer system. 'The project was undertaken to make whatever modifications were necessary for the study of enzyme kinetics with a scanning system. Additional technical improvements were effected to enhance the overall utility and re- solution of the apparatus, or as a direct result of the on-line computerization of the flow system. In order that the system remain adaptable to many uses it was constructed of inert materials. An improved quartz flow cell featuring a double mixer and a choice of two path lengths was constructed. The entire flow system, including storage reservoirs and syringes was enclosed in a Plexiglas bath for thermostatting and variable temperature operation. Special seals and 69 70 syringes were designed and constructed to allow lower temperature operation, reduced volume requirements, improved push-to-push reproducibility, and to provide a multiple-push capability. Typical of transient enzyme kinetics experiments is the necessity to study small absorbance changes in the UV region of the spectrum. As a partial solution to this problem a more intense light source was en- ployed in conjunction with an improved light throughput system. However, the greatest enhancement in S/N ratios as well as in overall performance came as a result of the signal averaging capabilities of the on-line computer system. Thus, the new stopped-flow system incorporates many of the features of its predecessor as well as several additional innovations. The introduction of a variable temperature capability to the anaerobic- inert system imposed severe restrictions on the ultimate configuration and construction materials of the flow and optical systems. The nature and availability of enzyme solutions also imposed certain demands on the solution handling system as well as the system's ability to operate in the UV region of the spectrum. A detailed description of the stopped-flow apparatus is presented below, followed by a presentation and discussion of the performance tests and system characteristics. 71 5.1--Flow System Design The flow system was constructed using greaseless vacuum tight seals so that the reacting solutions come into contact only with.quartz, Pyrex and Teflon. Since it is a closed system, a specially designed system for solution makeup and handling was required. Teflon seals, joints and valves are used throughout the flow system to insure an inert environment for the reacting solutions and to facilitate system repairs and alter- ations. A reference cell provides for double-beam operation, and its location beside the reaction cell in the bath insures the same temperature for both cells. 5.2--Flow and Reference Cells The flow cell was constructed entirely from quartz capillary tubing of known diameter. Quartz is required for UV applications and since the remainder of the flow system is Pyrex, one piece construction of the cell is required. Optical windows for the short path length were provided by first having the capillary tubing ground and polished flat on two opposite sides (Precision Glass Products Co., Oreland, Penn.). Holes were drilled in the quartz by using an "airbrasive" unit (8h) which utilizes a fine jet of gas driven alundum powder. A detailed description of the drilling and the glassblowing procedures involved in the actual construction of the cell are given elsewhere (32). A diagram of the flow 72 cell is shown in Figure 11 (32). Earlier designs consisted of a single four-jets mixer with only a short pathlength. The latest flow cell divides the mixed stream from the first four jets into four additional streams which then rejoin to effect a second four-jets mixing of the reactants. This secondary mixing process resulted in an increase in volume between the mixing and observation point of less than lonl. (less than 1 ms increase in dead time). Because of the small changes in absorbance involv- ed in transient enzyme studies it was necessary to pro- vide for a longer path length. This was accomplished by sealing a length of capillary at right angles to both the entrance and exit tubes (see Figure 11). Sections of very thin quartz microscope slides were then fused to the ends of the capillary, and subsequently ground and polished to provide for a 1.85 cm effective path length. The one piece system design permits flow through the cell without any volume expansion or contract- ion. Therefore a common source of cavitation artifacts is eliminated. Construction of the reference cell was somewhat sim- pler since no mixing portion was required. Of course care was taken to insure that no dead spaces were present which might interfere with rinsing or introduce bubbles. .aoxwa o no coavoomnmmoao a o .80 m.oe I m .muchn Nunodo as m aouaomunoamwm u < .Haoo cowum>aomoo one wcfixwa may no anwowo owpeaonom .Ha madman £9... ea .36 face. 58 9.3 MT! 73 714 Both cells were rigidly held in place by clamping them in specially machined aluminum holders, and securing the holders to the flow system frame. The holders also serve as mountings for the fiber optic light pipes and provide for’masking of unwanted stray light. Liquid from the surrounding bath is kept out of the light pipes by sealing the exterior of the cell holders with silicone rubber cement. 5.3-sggshing and Stopping System 5.3.l--§yringe and Plunger Desigg The construction of leak-free greaseless syringes for variable temperature operation presented special problems. Greased syringes were unacceptable because of the tendency for the grease to "creep" to other parts of the system glassware, and because of the solubilization or deterioration of greases in the presence of many solvents. Commercially available syringes with Teflon plungers overcome the problems of greased syringes. However, Teflon has a tendency to "cold flow" over a period of time, and will contract at lower temperatures; both of these phenomena lead to leaky syringes. Recently, syringes with adjustable Teflon seals have been made available commercially; however, generally speaking these 'were not designed for stopped-flow applications. 75 Over the years, this laboratory has attempted to utilize commercially available syringes in anaerobic and/or variable temperature applications with little success. The most satisfactory solution to the problem has been for the laboratory to design and manufacture its own syringe bodies and plungers. The syringe bodies are constructed of heavy wall precision bore glass tubing (diameter 0.553 inches, purchased from ACE Glass Co. on special order in 3 foot lengths). The syringe bodies are made by the glass shop as shown in Figure 12. The base of the syringe is flared and ground flat to provide stability, to permit easy insertion of plungers, and to facilitate a liquid seal with the base of the bath. The top of the syringe body is formed by fusing a flat plate across its diameter. This procedure minimizes distortions and dead spaces between the plunger and the glass body. To completely exclude air from the system, the syringes are constructed to permit back-pumping between two O-rings (Teflon is permeable to oxygen). This is accomplished by attaching a sidearm to the syringe body. These sidearms can also be used as exit ports for the purpose of rinsing the system prior to re-filling, an operation which wastes solution in many stopped-flow systems. Before the sidearm is attached, a small hole (less than 0.5 mm) is drilled through the syringe body using the "Airbrasive" technique. The "Airbrasive" drilling prevents distortion 76 Teflon Wiper 3: Screw /Syrinqe (\ no" / \ /‘ o———Steel Plunger as m Sideorm Pinhole “ ages Emma“ \ p . LII/J j ( / ziReverse Threaded LOCk- fl 0? Rod Figure 12. Syringe and adjustable plunger design. 77 of the inside diameter of the barrel and provides good control over the size and positioning of the hole. Thus the hole is positioned so as to lie below the Teflon wiper when the plunger is in the "fill" position. This allows the space between the O-rings to be evacu- ated. The diameter of the hole relative to the width of the Teflon wiper is such that as the plunger is lowered further to the rinse position, a continuous seal is maintained with the inside of the syringe barrel. In this position, old solution may be evacuated to a waste vessel and the syringe efficiently rinsed with solvent or solution from a reservoir. Once rinsed, the plungers are returned to the "fill" position and back-pumping may continue. Care should be taken so that materials which attack the O-rings do not come into contact with thmm. The glass sidearms were attached by the glass-shop using a "stick-seal" technique so as not to distort the interior diameter of the syringe body. This process is difficult and is not recommended for amateurs. The thermal properties of Teflon are such that a seal made at one temperature develops a leak at a lower temperature. To overcome this problem (and the problem of cold-flow), we have designed and tested plungers in which the Teflon seals are forced against the glass wall by Viton O-rings which are held under compression (see Figure 12). A machined Teflon cap is attached to a 78 threaded brass rod which in turn is reverse threaded into a stainless steel shaft. If leakage develops, the lock-nut is loosened and the rod is turned. This action pulls the Teflon cap down onto the O-ring causing it to compress and force the wiper against the glass wall, thus effecting a better seal. The adjust- able wiper system has been tested outside of the flow system at -30°C and was able to hold a vacuum of 10'5 mm Hg. The actual low temperature limit under stopped-flow conditions has not been determined because of the presence of other Teflon materials (valves) which would begin to leak severly at around 5°C. A Teflon valve similar in principle to the adjustable plunger caps is currently under development. The adjustable plungers are still undergoing test- ing. Syringe leakage during a run can be halted effectively by tightening the Teflon cap; however, certain precautions are necessary. The relationship between the O.D. of the compression O-ring and the I.D. of the Teflon wiper is critical, and because Teflon does not hold a thread very well, periodic replacement of the cap is required. 79 5.3.2--Mechanical and Framework The structural backbone of the flow system itself consists of four threaded brass rods which are fastened vertically at one end to an angle iron frame which in turn is bolted to the floor. Aluminum plates are fast- ened with large nuts to the brass rods to serve in holding the syringes in place and for added structural support. The top plate acts as the stopping plate and as a mount for the waste system pneumatic piston. The remaining framework consists mostly of lattice rods which serve to support the glassware of the solution handling and delivery system and the vacuum lines. The vertical mounting of the flow system facilitates the movement of solutions into the system and provides an easy means of sweeping out trapped bubbles. The pushing plungers are mounted in an aluminum block which is in turn connected by a threaded rod to a large pneumatic piston (2" diam. bore, Cathy Co. (Schrader), Lansing, Michigan). Movement of the pushing plungers may be controlled manually with a three position valve which controls the flow of the pressurizing gas. The hand operated valve affords good piston control while the flow system is being filled, rinsed or used to sweep out bubbles. To improve flow velocity reproducibility and to allow for automatic operation, a bypass air system was constructed so that the air pressure to the pneumatic 80 piston could be controlled by a solenoid valve (Cathy Co. (Schrader) Lansing, Michigan). A simple switching circuit thus permits electrical control of the fill, push and hold modes of the piston. The switching action can also be controlled automatically by using TTL pulses. The BS triggers are ignored until the Begin Push switch is activated. Subsequently, the very next BS trigger will cause the solenoid valve to initiate the pushing action. The overall reproducibility in time of the solenoid valve and pneumatic piston operation is sufficient to provide reproducible flow velocities. Data collection is begun with the next BS trigger after constant flow velocity is achieved. Since the time bet- ween BS triggers is constant, and since the period of acceleration is reproducible to within one spectral scan, the scanning monochromator can be roughly "synchronized" with the operation of the stopped-flow system and the collection of data by the computer. 5.3.3--Thermostat System The flow system, including storage reserviors and syringes is enclosed in a Plexiglas thermostat bath similar to that developed by Dewald and Brooks (85). This is important, not only for the study of the tempera- ture dependence of reaction rates, but also to insure that there are no temperature gradients within the system. 81 Such temperature gradients can cause optical artifacts even in the absence of absorbance changes (39). A liquid bath has the advantage of a large heat capacity which tends to diminish temperature differences between different parts of the system. The bath.was constructed from 3/8" Plexiglas plates which.were bonded together by using methylene chloride. One of the sides and a large section of the back wall are removable to give easy access to the flow system. Wherever holes in the bath were required, silicone rubber sealer (Dow Silicone Rubber Cement or General Electric RTV) was used to prevent leakage of the bath fluid. The transparent Plexiglas allows easy visability of the flow system and acts as a reasonably good thermal insulator. A variety of bath fluids could be employed. However water or ethylene glycol-water mixtures are generally satisfactory. The present bath circulation system actually consists of three separate baths. Liquid from a large (25 gal) reservoir is circulated through copper coils which are submerged in an intermediate bath located near the stopped-flow bath. Liquid is rapidly exchanged.by an impeller pump between the stopped—flow bath and the intermediate bath. Circulation to the jacketed burettes is also possible and is intended for use either to pre-equilibrate the solutions or as a means of slowing down decomposition of labile reactants. 82 The large capacity of the total bath system affords temperature regulation to better than i.l°C. However, considerable time is required to establish thermal equilibrium. Thus for variable temperature studies smaller auxiliary baths might be employed. 5.3.h--Solution Mixers Reagents and solvents are delivered by burettes to two storage reservoirs located on either side of the bath for solution makeup. Mixing is accomplished by a Teflon coated magnetic stirring bar which has been sealed in the storage vessel. The length of the stirring bar and the diameter of its container are such that the bar cannot lie flat. Vortex type mixing is effected by air-driven turbine type immersible magnetic stirrers which have been mounted vertically behind the storage vessels. Control of the stirring process is important to avoid denaturation of the protein by excessive foaming. 5.3.5--Stoppipg Syringe System The stopping syringe is similar to the pushing syringes except that the sidearm is used for back-pumping only. The separation between the stopping plate and the base of the stopping plunger determines several important instrumental parameters. First, the total volume required per push is directly related to the length of the stroke for a given cross-sectional area of the top syringe. 83 Second, sufficient time (distance) must be provided to allow the flow velocity to become constant. Finally, as the scanning speed is decreased, longer periods of constant flow velocity are required in order to obtain at least one complete zero-time spectrum (during flow). Given the above considerations, the system currently requires approximately 0.7 ml of each reagent per push. The period of constant flow is approximately 80 ms. (dependent on the air pressure) and the stopping syringe diameter is 0.553 inches. Because the decomposition rate of solvated electron solutions is increased with a large surface-to-volume ratio, fairly wide diameter syringes were employed. A new micro-system designed specifically for enzyme studies is currently under development. The pushing syringes contain a total volume of 2.5 ml each when filled. Hence three pushes can be delivered for each filling of the syringes. A small pneumatic piston and valve are used to return the stopping plunger to the ready position while simultaneously expelling wastes. In order to provide more versatile dual-beam operation, the waste line may be connected to the reference cell. Thus the reference cell can be filled with a variety of reference solutions which can be delivered by the flow system. The appropriate solution is simply push- ed into the stopping syringe and then expelled to the reference cell through the waste system. Once filled, the 8h reference cell is isolated from the waste line and the normal flow of actual waste materials is continued. 5.3.5.l--Start and Stop Flags A versatile and reliable flag system was construct- ed to indicate not only the stopping of flow but also to monitor the flow velocity and to provide control triggers for the data acquisition system. Motion of the stopping plunger is monitored by two metal flags which are mounted on the plunger barrel. As the plunger moves upward, the flags interrupt two light beams which normally illuminate two light-sensitive phototransistors. This interruption causes pulses to be generated. In this way both Start and Stop pulses are provided which can be used to initialize data acquisition, calculate flow velo- cities and determine the zero time of reaction. 5.h--Solution Handling and Delivegy System The entire flow system, including the burettes and solution entry ports can be evacuated routinely to less than lO-h'mm Hg. For applications not requiring high vacuum, a roughing pump or water aspirator is utilized either for drying the system or as an aid in moving solutions through the system. There are four solution entry ports located on each side of the flow system. Each port is connected to the high-vacuum line by a Teflon 85 valve and to the delivery system through 10 ml burettes. For anaerobic work, solutions are first pressurized with an inert gas in glass bottles which have been fitted with Teflon valves and Fisher-Porter type joints. In non- anerobic studies, solutions are introduced into the system through glass funnels which have been fitted with Fisher-Porter joints. After the bottles have been mount- ed, and the remaining spaces evacuated, the high-vacuum valves are closed and the valves on each bottle are open- ed to fill their respective burettes. Each burette is jacketed so that liquid from the flow system bath can be circulated around each solution. The four burette tips for each side have been brought together in a solution receiving area which is designed to permit solution delivery from any burette into a closed system and also to provide for good drainage and rinsing characteristics. From the receiving area, solutions travel to a no ml storage reservoir located within the bath. Here solution makeup is completed with mixing accomplished as described earlier. At this point the solutions are allowed to come to thermal equilibrium with the bath. The 10 ml burettes permit good accuracy and precision while the hO ml storage vessel provides enough.volume for>multiple push applications. The five valves which are used to control the direct- ion of flow of solutions are rigidly mounted to the sides of the Plexiglas bath (see Figure 13). Since these valves FigurelB. 86 Schematic diagram of the thermostated stopped-flow apparatus. A - Joints for rinsing solutions, B — Joints for reactants, C - thermostated burettes, D - reactant reservoirs, E - mixing and observation cell, P - reference cell, G - pushing syringes, H — pneumatic pistons, I - stopping syringe, J - quartz light fibers, K - thermostat bath, L - to vacuum, M — to vacuum and "waste", 1, 2, 3, u, S-flow valves. 87 are operated frequently during a run, their stability is important so that applied torques are not translated into motion of the glassware which can ultimately cause breakage. The pushing syringes are filled by closing the inner two valves and opening the two outer valves while the pushing plungers are lowered to the "fill" position. Reactions are carried out by closing the outer valves and opening the inner valves. The pneumatic piston forces the plungers upward causing the reactants to flow through a short length of spiral tubing included for strain relief and into the mixing chamber. The quartz mixing cell is connected by Fisher-Porter joints to the remainder of the flow system which is constructed from 2 mm ID Pyrex capillary tubing. The moving solution causes the stopping plunger to strike the stopping plate thus halting the flow of solutions as well as the advance of the bottom plungers. With the inner valves closed and the pneumatic piston in the hold position, the exhaust valve is opened to allow the small pneumatic piston on top to expel the spent solutions while returning the top plunger to the "ready" position. With.the exhaust valve closed, and the inner valves reopened, the pneumatic piston is again activated resulting in a second kinetic push. In this way three records can be obtained for each filling of the bottom 88 syringes. S.5--thica1 System Design Flexible quartz fiber optic light pipes with water- tight sheathing have been used to bring light through the liquid bath to the cells and from the cells to the detector (86). Attachment of the fibers directly to the flow cell minimize problems caused by cell movement and provides considerable flexibility in positioning the remaining optical components in addition to reducing the problems due to stray light. Because a double-beam system is used, lamp and photo- multiplier supply fluctuations are largely cancelled out. Two additional "noise" sources are shot noise and ruechanical noise. The S/N resulting from the former is determined largely by the photocathode current and the frequency bandpass of the total detector system. Mech- anical noise results mainly from vibrations of the optical components in the monochromator and decreases when the system is operated at a fixed wavelength. With a scanning system it is necessary to operate with a broader frequency bandpass than is usual at fixed wavelength; that is, there is a wide dynamic range of the output voltage from the photomultiplier tubes during each scan. Variations of both the lamp output and photomultiplier tube sensitivity with wavelength make it not uncommon for the signal to change by a factor of 100 or more during a scan. The double-beam technique together with an operational 89 amplifier logarithmic converter overcome this problem and also provide a more convenient display. The use of double detectors and the ability to provide a variety of refer- ence absorptions, affords the advantages of baseline subtraction and scale expansion techniques. The gear system of the scanning monochromator is used to provide the basic timing signals for the data acquisition system. Light sensitive transistors are used in the scanning mode to provide 68 wavelength.mark- ing pulses for each spectrum, as well as a single pulse each time a new forward scan is begun. The frequency multiplication system is used to provide a constant 20.h KHz sampling frequency which is independent of scan speed. The BS pulses are used by the automatic start system to synchronize the monochromator to the flow system, and by the computer to synchronize the monochromator to the data acquisition and storage process. S.S.l--Light Sources and Detectors Some of the S/N ratio disadvantages of scanning can be overcome by using an intense light source. With this in mind, a 1000 Watt Xenon lamp (Oriel Co.) was installed to provide sufficient light intensity in the UV - VIS region of the spectrum and to overcome lower light levels caused by narrower slits and the use of fiber optic light pipes. Tungsten and deuterium light sources are also available. 90 The monochromator has a maximum scan rate of 150 spectra per second. It is a four-pass Littrow system with continuously variable slits. The scan width is determined by the angle of nutation of a rotating mirror and is independent of either scan speed or center of scan. Unwanted radiation and/or heat are removed by liquid and glass filters. The lamp is fitted with a 2 inch collimating condenser system for increased light input. A specially constructed holder brings the fiber optic light pipes up to the exit slit of the mono- chromator for improved light collection. An attempt was made to combine the individual fiber bundles at one end into a slit shape, to serve as a flexible fiber optic beam-splitter which could make optimum use of the light output of the monochromator. By shaping the fibers to match the shape of the exit slit, and by randomly splitting the fiber bundles, more uniform illumination of the cells could be achieved. Unfortunately the cladding on the fibers is quite fragile and the effort was aborted after a large investment of time. Glass or plastic beam splitters with an assortment of area transfer designs are readily available and reasonably inexpensive. UV trans- mitting quartz fiber beam-splitters are available from Schott Optical Co. on special order but are quite expensive. Future plans in this laboratory include the purchase of such a beam splitter as part of an overall upgrading of the optical system since matching of the two beams is presently 91 one of the limiting contributors to the noise. At present two 50 cm lengths of flexible fibers with round 2 mm diameter ends are brought up to the exit slit of the monochromator for each of the path lengths. The other ends of the fiber are secured to the cell windows. Light from the cells is brought to the detectors by similar round ended fibers. Currently two sets of photomultiplier detectors are used; (1) RCA 6903 for UV-VIS and (2) EMI 968MB for VIS-IR applications. The EMI tubes utilize special magnetic lenses and a cooling box which has been described elsewhere (32) for improved dark current characteristics. The reference and sample output photocurrents are converted to absorbance and amplified by using an operational amplifier circuit. A selection of analog filters is also provided in the circuit. The absorbance signal is biased to allow for full scale utilization of the data acquisition system. The absorbance conversion is linear over the 3 range of 10- to 10-9 amps. of photocurrent. The system response as a function of wavelength is calibrated during each run with rare earth oxide glass filters as well as neutral density filters. Computer programs correct the raw data for variations of the instrument response as a function of both absorbance and wavelength. 92 S.6--Performance Results The overall performance of the system.was evaluated by using a variety of tests including the study of standard reactions. S.6.l—-thical Calibration Data The sensitivity of the system depends essentially upon the optical components. Both light intensity and phototube response decrease substantially at shorter wavelengths; and without averaging of adjacent spectra the scanning mode results in significantly poorer S/N. However, the most severe loss in sensitivity arises from fluctuations in the output signal that are caused by "wander" of the Xenon arc. Since reference and sample fibers are intercepting different regions of the arc, these fluctuations are uncompensated and result in vertical de- flections in the spectrum. Since the net result of all these effects is a complex function of wavelength and instrument settings, it is not possible to state the overall system sensitivity in general. However, by measuring the deflections in the spectrum of holmuim oxide glass on an analogue scope we estimate the r.m.s. noise to range from 0.002 to 0.01 absorbance units in the 500-300 nm region. The wavelength resolution of the instrument also depends upon the instrumental settings. A typical cali- bration spectrum of holmuim oxide glass was shown earlier (see the discussion of the scanning monochromator). The 93 effective path lengths of the cell were determined with Beer's law tests. For the long path length, aqueous solutions of 2,h- dinitrophenolate (360 nm) were used. The solutions were calibrated with a Cary model 15 spectrophotometer using 1.000 f 0.001 cm Beckman cells. In Figure la the results obtained with the stopped-flow system at 360 nm are plotted against the corresponding results from the Cary. Beer's law is obeyed up to an absorbance of approximately 2.0, and the effective path length calculated by using a least squares fit is 1.850 i .005 cm. A similar procedure was used for the short path length. In this case, freshly prepared KMnOh solutions were used (52h nm) and gave a path length of 1.99 t 0.02 mm (32). Stray light which bypasses the cell was therefore shown to be negligible at absorbances less than two. In general, scattered light problems were minimized by positioning appropriate cut-off filters between the lamp and the monochromator entrance slit. S.6.2--Flow Calibration Since it is possible to measure the flow velocity accurately with the phototransistor flag circuit, we are able to determine the time at which the flow velocity reaches a constant value (for a given air pressure on the pneumatic piston). The reproducibility of the absorbance traces during flow depends of course on the reproducibility 94 .msuoumamo BOHMIomQQoum oz» :0 numcma Sumo mcoH mnu mcams oouammwe ouchHOmnm on» umcfimmm omuuoHQ ma Ammo a co consumes mama LUHLB Ascommv mumaoconmouuaclm mo mcoHDDHOm no oUCmAMOmnm 0:9 .vH musmfim roléeaaofi :3: 3:33.34 aim “Nu AYN m; .w; .v; N; o; c. o. e. N. fi _ q _ — _ d _ q _ q p m. Moo use” eauoqlosqv Q N; 95 of the flow velocity. Data are collected only during the period of constant flow. It is possible to obtain initial absorbance data during a given run which are reproducible to approximately 5 per cent as determined by studies of 3 + - the reaction of Fe with SCN . S.6.3--Flow Cell Performance, Fixed Wavelength S.6.3.l--Mixing Efficiency The mixing efficiency of a flow system can be defined as the degree to which the reactants are mixed upon their arrival at the point of observation. In order to test the efficiency of the double mixer cell we studied the reduct- ion of p-nitrophenolate by acid. Under the conditions of the test the reaction was diffusion controlled and the final absorbance was non-zero. Any change in the remaining absorbance after stopping would have indicated incomplete mixing of reactants. Since no such changes were observed on the analogue scope for a series of replicate pushes, it is assumed that mixing is completed within the dead time of the instrument. This test did not provide a quantitative measure of mixing efficiency, however the results indicate that the cell is at least as efficient as any reported to date. In an effort to measure the mixing efficiency quantitatively, additional tests with suitable indicator reactions and more accurate computer analysis are currently underway. 96 When identical solutions or solvents were mixed no change in signal was observed. This indicates that mixing artifacts due to cavitation or thermal gradients are absent. The stopping time is approximately 0.5 ms as mea- sured by the "rounding" of the absorbance trace in the reaction of Fe3+ with SCN". 5.6.3.2--Dead Time The time required to transfer the solutions from the mixing chamber to the observation point and bring them to a complete stop is called the dead time. The dead time, therefore, depends directly on the flow velocity obtained for a particular experiment as well as the mix- ing efficiency and the stopping time. The dead time for the short path length was calculat- ed to be 2.7 ms from the flow velocity and volume displace- ment for the short path length (32). Similarly the dead time for the long path length was calculated to be 6.7 ms. Using data obtained by studying the formation of peroxy- chromic acid at 600 nm and extrapolating to the true time zero of the reaction, dead times for the short and long path lengths under these conditions were computed to be 2.5 ms and 5.5 ms respectively (32). 97 5.6.hs-Qpantitative Measurements The formation of blue peroxychromic acid at 30°C was investigated in detail by Papadakis in order to evaluate the performance of the system in an actual kinetics study (32). Measurement of both a pseudo-first order and a third order rate constant for this reaction at several wavelengths, under a variety of concentration conditions, and with both the scanning and fixed wave- length modes gave reproducible values which are in agree- ment with those reported in the literature. The calculated rate constant for the base hydrolysis of 2,u-dinitrophenylacetate was unaffected by the choice of sampling parameters which.were varied over a wide range. Obviously if an unreasonable averaging bandpass is select- ed the rate curves will be distorted just as is the case with analog filters. Good quantitative performance under actual operating conditions has also been demonstrated with several enzyme reactions. The ability of the stopped-flow system to give reproducible rate constants, enhanced S/N, and to detect subtle changes in the spectrum during reaction (including isosbestics) under conditions of relatively poorer S/N is clearly illustrated with the LADH study discussed in the next chapter. VI TRANSIENT STATE ENZYME RESULTS 6.l--Gl ceraldeh de- - hos hate DEETDEOGENASE (GAPDH) GAPDH catalyzes the conversion of glyceraldehyde-3-phos- phate (GAP) to 3-phosphoglycerate (PGA) in the presence of arsenate and NAD+. In the absence of NAD+ the enzyme exists as an equilibrium mixture of inactive (To, 98.h%) and active (R0, 1.6%) forms (26). Stopped-flow mixing of the apo-enzyme with saturating levels of NAD+ leads to the rapid formation (less than 10 ms) of a catalytically inactive complex containing four NAD molecules (Th). The Th complex subsequently undergoes a slow isomerization to yield an active Rh complex with an absorption band in the same region but with a larger absorbance. The equil- ibria among the different enzyme forms are strongly temp- erature dependent and at uo‘b the conversion to Rh is essentially complete. In solution the substrate (GAP) exists primarily as the acetal. Hence, after a small (3%) initial burst period, the reaction proceeds at a slower rate correspond- ing to the rate of conversion of the acetal to the cat- alytically useful aldehyde. This fact was used to advantage in testing the scanning system since detectable levels of the Ru complex could be produced even in the presence of 98 99 NADH production. The rationale for studying the GAPDH system was to test the scanning system with a well defined enzyme system (GAPDH has been studied extensively by Kirschner (25, 26) and others (88)), and to illustrate the facility with which subtle differences in enzyme forms can be detected with such an instrument. Since the enzyme was prepared in quantity in our own laboratory, a process which required several weeks of work, it was also hoped that additional information on the GAPDH system could be found. 6.l.l--Complex formation between NAD and GAPDH Kirschner determined the spectrum of the Tu and Rh complexes by using a series of fixed wavelength experi- ments (26). This point-by-point mapping technique resulted in rather ill-defined spectral shapes and left open the question of whether both complexes have the same band shape. By using the rapid scan system it was possible to determine both the Th and R,4 spectra directly. The spectrum at various times during the growth of the RM complex minus that of the starting NAD+ absorbance is shown in Figure 15. There is some hint of a wavelength shift as the isomerization proceeds indicating that a difference in the spectrum of the Th and Rh complexes may indeed exist. Because of the tail of the NAD+ 100 absorption and the small enzyme absorbance at early times in this region however, it is difficult to obtain a true band shape by the available subtraction routines. It should be noted that much of the noise shown in Figure 15 is caused by uncompensated spatial fluctuations of the Xenon arc. These fluctuations could be reduced by a better beam-splitter. No significant differences in the spectrum of either complex was observed at lower concentrations of HAD. A time out from Figure 15 at 366 nm shows the growth of the Rh complex (see Figure 16). Note that the initial absorbance is not zero because of the rapid fermation of TM which.has a significant absorption at this wavelength. The Rh growth.was fitted to a simple first order process with rate constant kh' As Figure 17 shows however, there is a systematic deviation of the data from the calculated curve. This is especially true at long times indicating the presence of an additional but slower process. Kirschner has shown that with a GAPDH preparation which showed two bands on electrophoresis, double exponential growth curves for Rh resulted. Holland and Westhead (113) have found however, that even the electrophoretically pure band from the column step can be further broken down into isoenzyme components. Based upon the assumption that our GAPDH preparation was actually a mixture of two enzyme forms which have different rates of isomerization, the Rh growth data were fitted to a double exponential 101 .o. m. on no season moeaouoaz on» no coauooom codumuaumeog ‘31:. one. madman oocaouno euuoomm oucouommao oouomaom .ma ousmam 35.; 59.2263 o¢¢ ON¢ 00¢ can con 8» on» _ m _ _ 4, q q _ a a a a :00. e / ee 0 IVO. ' ee 00 so _ eee / .00. as e e z .o. / . .. / e o o o N 9 saueqmsqv V 102 .mH ousmam scum Edmmm um usoumE«B .mH muzmfim AoomVoE; eoueqmsuv 103 .Aeeoomv madam no xooeeoo em on» no nukoum 0:» mo paw Hoaucmcomxm mamcfim .na musmfim Too: 2:: Qu- od. 0.» 06 0:9. 0 — 1 W — E I x o m gm 0 x .0 108. o 0 ul 0 u O on x o o 12...... x 0 on 0 OX x ox x o x x" 100”. x O x u x x 8 x o x o x x x O o o o o o O o u u o o o x x x x x x (um 99;) eauoqaosqv 10a expression. The results are shown in Figure 18 and Table III. The percentage of each form was allowed to "float" as a parameter along with the rate constants for the iso- merization of both the main (kn) and sub (kh) components. The double exponential expression yielded a better fit based on the non systematic distribution of the experi- mental points about the calculated curve. The average amount of the minor sub-form was 25% compared to Kirschner's value of 20%. Comparison of the two rate constants (kn and kL) with Kirschner's results (Table III) shows that our values are somewhat larger. It is difficult to rationa- lize this discrepancy. Experimental conditions for the run were practically identical with those of Kirschner. The specific activity, E280/260 ratio, and the electrophoresis results indicate that the enzyme preparations were at least comparable. The total absorbance change for the growth corresponds precisely with that expected from the concentration of GAPDH used. The fact that the isomerization process is first order in enzyme concentration, and that I h those determined by Kirschner does not detract from the our values for'both kh and k are several times larger than validity of our results. The relatively large standard ! deviations that are given for kh and kn indicate the diff- iculty in resolving two relatively similar exponential processes from a small absorbance change ( .08) on top of a large background absorbance. Kirschner provides only 105 .Aecommv $93.0 no meQEOU vm mg» 00 guzoum oz» mo yam Hmaucocomxo cannon .mH ousmfim fee: 06:. rim. 0.0. nfi. 0.0 QN 0 a _ A a a u m m 1 00m. 0 0 xx 0 0 x x 0 e L 2h. x 0 x 0 0 x um 100% x 0 no x x u x 0 u 0 O u u o I o 0 u u u x x (um 99;) esuoqaosqv 106 3.0 mooduammd Sioueoedoeooumomd emodflmomd Inconven— m.o 2.93m; «mouse; Soummq 8.3.34 omooumemd A7833. 9.2 25m 9.2 25 9% 25m oaz :5 9.2 2% 92 2.5 eeooe eeooe semen seoom measeom rezone m.uoc£owu«x Lusouu Hmaucmcomxm cannon Heaucmcomxw onch .muazmmu m.uocnomaax sufiz comauomfioo m can mammo mo xmameoo mm on» no susonm on» you Amvucoumcoo open “mono umnau on» no snow HmcoHuocsm one .numcoao>e3 coHuowucoocoo +0¢z no uoommo one HHH mqm<8 107 an estimate of kg, and in fact used only 80% of the pro- gress curve to determine kh' There is no indication that a correction for the second exponential process was made. Although our results obtained by using the KINFIT program, represent a much more sophisticated treatment, it is still difficult to rationalize such large differences. According to Kirschner the isomerization rate for the forward process T-éR is independent of the number of NAD molecules bound. He accounts for the predominance of the Tuf-vRLL pathway by assuming that the rate for R-’T is smallest when four NAD's are bound, and that the reverse rates increase when fewer than four are bound. To test this assumption GAPDH was rapidly mixed with 1 mM NAD. Since the T and R forms have different saturation char- acteristics, it is possible with this level of NAD to saturate only the R form in a rapid mixing experiment. It was suspected that under such conditions a significant portion of the conformational change might proceed through (say) the T3-§R3 pathway and that this rate might be some- what different from that for the T144 Rh reaction. These results are also shown in Table III. No significant difference for the rate of RLL formation was observed. As a further test of the overall resolution of the instrument, data from Figure 15 at hOO nm were also fitted to a double exponential expression. As expected, no effect at either level of NAD was observed, although the 108 standard deviations at MOO nm are obviously larger because of the smaller absorbance change at this wavelength. 6.1.2--Reaction with Substrate The reaction with substrate (GAP) was carried out under conditions of excess GAPDH and saturating levels of NAD+. Figure 19 shows the development of the spectrum during reaction with only the buffer background subtracted. The Rh complex and NADH have comparable absorptions in this region under these conditions. Note that even when the excess NAD'background is subtracted (see Figure 20), it is still difficult to discern a wavelength dependence in the time development. However, when time-cuts are taken at several wavelengths across the band (Figure 21), the presence of at least two time and wavelength dependent phenomena is quite evident. These of course, correspond to the simultaneous growth of NADH and the RLL complex, followed by continued production of NADH at a steady state level of R . We feel that these experiments clearly illustrate the utility of the scanning experiment and the need for high resolution and versatile data processing capabilities. 6.2--The Threonine Dehydrase System In an effort to further demonstrate the versatility of the rapid scanning system, we report some results on a collaborative project with Professor Wood of the MSU 109 .Aoouoouunsm coon mm: ocsoumxomn uco>H0mv ado one +Q¢z noes m0m¢0 mo coHuomou onu mcauso mmEHu m90aum> um Esnuuomw .ma wusmfim 35: £05.26; 0: one 00¢ can own own on» q _ _ q 1 q d _ q q _ _ oeeeeeeefiufie ...-m..".. :53: «use. ounce mu.. no o u” . ee .0 O O. O O u. . 000 0 see 0 0 see 0 000 e 0000 0 cos. 0 so e. ee 0 e as”... 00.. 0.0000000 O O O 00.000 0.00 e e no ee co. ee 0 O O. 0.0.0.0....OHNOO O 0. eeeeeooonoooeeeeooe 0. eeeeeoeeeeee aauoqmsqv 110 .AooDUmwunsm +Q um Esuuuomm :5: 505.263 91. 0N¢ 00... 00M 00m ovn . d _ .om enemas 0mm 4 fl - u _ a _ a _ 0001‘. so uoqmsq v .coXmu mm: om owsmam LUHLB Bonn moon on» 00 mucoumEHB .Hm ousmfim .630 25... 0.0 0.0 0.? 0.N 0 a _ _ _ O lll eeuoqmsqv 112 Biochemistry Department on an unusual allosteric enzyme, the L-threonine dehydrase of E. Coli. This proved to be an interesting system since the entire time course of the reaction, from the initial growth and wavelength shift of the absorption of the enzyme-bound complex to the growth.of the product absorption, could be followed. The spectral changes are shown in Figure 22 (which does not have the solvent absorption subtracted). The initial single band at #20 nm grew and shifted to h60 nm in about the first 100 ms. During the growth of product the trans- ient absorbance at A60 nm gradually diminished and shifted until at the end of the reaction the band at u20-h30 nm was about the same as at the beginning. This is shown by the time-cuts in Figure 23 at 330 nm and 4&0 nm. The two decay curves are on the same absorbance scale but a con- stant value has been subtracted from each absorbance for display purposes. Toward the end of the reaction when product formation is nearly complete, the transient absorption of the enzyme-bound complex decays back to that of the enzyme alone. An isosbestic point occurs at about u05 nm. This is shown in Figure 2h, which covers the wavelength range from about 370 to 550 nm. The enzyme itself is known to exhibit a peak at h15 nm due to Schiff base formation between pyridoxal phos- phate and a lysine residue (5,75). The shift of the 113 .Aomuoouunsm moaammon ocv $09 an ocacoowculq mo cofiumcfismmo ecu you...m.~.a masouw sown owuommm nouomaom 06: 0 593353 0v» . 00¢ ON? 00» _ .mm museum d A. _ _ «Av VA. 0A0 asuoqlosqv 114 .AEeonm Do: “9:00.: ovv um 53.603 .553» ode Enemm um musonwefiu mos .mm shaman «5538:. m v m N _. 0 . . . . 4 AV 0 00000000000008.8930 N. o: - m. eouquosqv 115 .Emumhm mah on» Ca .Aunmau uev unmanneuu on» no recon one .Aumma Dov auscum uusooum moanso owumoanma so no museumomm< .v~ owsmam A.Ec;uoco_o>m>> 0ND 00¢ 0¢¢ 00¢ 1llgi1llfilflflflmlflflflflflw J. a . d a 4‘ .dl EB uoeom Zoom “0.0000 00 /r k 4/ .0. 0.000600. \ 5:50on one. eoueqlosqv 116 absorption to give a new band at hSO nm (composite at u30 nm) upon the addition of L-threonine and AMP has been previously observed and presumably reflects the formation of a Schiff base between the enzyme-bound pyridoxal phosphate and aminocrotonate, a dehydrated intermediate derived from L-threonine (89, 90). The results are somewhat inconclusive concerning any new mechanistic information; however, a few general statements are possible: (1) by using a faster scan rate or under fixed wavelength conditions, the approach to steady-state of the enzyme-complex should be tractable, (2) product growth is first order, (3) other than the 430 nm band, no additional absorptions were detected on the short time scale (less than 10 sec.), (u) in the pre- sence of 25 mM L-threonine, the steady-state concentration of the enzyme-substrate complex was achieved in less than 100 ms. All of the above information was extracted from a single kinetic record. Although a great deal of informa- tion is made available from a single push it is not possible to draw any conclusions about the mechanism of reaction from only one set of experimental conditions. In addition, since the background absorbance due to AMP and dithio- threitol were large, the results below about 310 nm are unreliable. 117 The amount of information available even from a single push with the scanning system is enormous. For example, in the TDH run, we collected samples at a 20.h.KHz rate as usual and averaged six samples into each point of the spectrum so that during the 13.33 msec required to scan the spectrum in the forward direction, no points were stored. The first 12 spectra of no points each (group 1) were stored as obtained and the intervening back-scans were discarded. The next 36 spectra were stored as 12 averaged spectra (group 2) with each stored spectrum the average of three successive spectra. This process was continued such that by the time group 7 was reached, each spectrum was the average of 729 adjacent spectra with, of course, a substantially improved signal- to-noise ratio. The total elapsed time was 5.83 min., yet the first 12 spectra were collected at the rate of one spectrum every 26.7 msec. Therefore, both fast and slow spectral changes could be observed. With no points in each spectrum, adequate wavelength resolution was avail- able so that spectral shifts could be followed. At each wavelength, a "time-cut" could be made if desired contain- ing 8h points with spacing between points ranging from 26.7 msec in group 1 to 19.39 sec in group 7. The ability to punch selected data directly from the PDP 8-I computer for analysis with Program KINFIT on the University CDC 6500 computer makes the analysis of such large amounts of data feasible. 118 6.3--Alc9hol Dehydrogenase (Horse liver) The third enzyme reaction chosen to test the new instrument was the equine liver alcohol dehydrogenase (LADH) catalyzed reduction of the substrate analog p-nitroso-N,N-dimethylaniline (NDMA) by reduced nicotin- amide adenine dinucleotide (NADH). This thoroughly characterized enzyme has been the subject of numerous steady-state kinetics studies (91-95) and binding studies (96-100). Good evidence exists for two coenzyme binding sites per molecule of enzyme (101-103). The enzyme ex- hibits a rather broad specificity for aldehydes of widely varying structure; however, specificity for coenzyme analogs is somewhat more restrictive (10h). The kinetic sequence for the binding of coenzyme and substrate prior to reaction has been demonstrated to be ordered with com- pulsory binding of coenzyme as the initial step under steady-state conditions (91,9h,95). Wratten and Cleland (9h.95) have demonstrated the presence of ternary complexes among enzyme-coenzyme-substrate (also under steady-state conditions.) More recently, rapid kinetics techniques have made possible the characterization of some individual steps in the overall transformation (105-111). Transient intermediates in the LADH catalyzed reduction of several aromatic aldehydes have recently been observed (109,110). It was this last aspect of the LADH system which seemed most attractive as a means of evaluating the potential of the rapid scan technique. 119 6 . 3 . 1--LADH Rate Measurement a Quantitative rate measurements were made by using data from four replicate scanning pushes in conjunction with the KINFIT program. A variety of rate expressions, derived to reflect specific mechanistic alternatives, were tried in order to determine the relationships among the time developments at several wavelengths. Only those expressions which yielded the best fit of the data will be discussed. The approach was somewhat phenomenological. However the quantity and complexity of the data required an economical and systematic procedure for eliminating alternative mechanisms . 6.3.2--1(_itrietics of NDMA Disappearance (Q0 nm) The disappearance of the NDMA band at 14110 nm followed Michaelis-Menten Kinetics. By following the entire decay scheme and fitting the data with Program KINFIT, values of Km (effective) and Vmax could be obtained from single decay curves. As expected, since saturating levels of NADH were present over much of the reaction, vmax is more reliably and reproducibly measured than is Km. The fit of the data to the Michaelis-Menten expression is shown in Figure 25. Table IV shows a comparison of the results with values which were determined at very low enzyme concentrations, fixed wavelength and with saturating levels of NDMA. .mzoz Mo COauusomu nonmaouou mnuwmno use ooumasoamu .mm mwsmfim 120 0:3 .6: oo— 0. c. c. N. o m u q u q u "A u 04 n w u M _ 4 1+0 x u u 1N. 1 i¢. x 0 x10. 0 causalosqv 121 TABLEIN' Comparison between steady-state and full-time course integration kinetic parameters for substrate dis- appearance in the LADH system. Kinetic Full-Time-Course Steady Parameter Integration State 23:0.190 (Ref.109) 25:1 ’C 4.31:0.46 x1o-5* KgDMA(M) 7.48:1.6 x10—6 _6 5.64:0.13 x10-6 0.7x1o 6.19:0.82 x10-5 Average 5.90:0.74 x10-6 24.7:1.3 ** 33.5:4.2 Vmax 27.713.6 24:2 30.8:2.4 * These values represent an effective Km for NDMA. **moles/l NDMA per sec. per N, where N is the normality of enzyme sites. 122 6.3.3--Kinetics at 235 nm and 5&0 nm During the decay of the NDMA absorption, the absorp- tion band at 3&0 shifted to 335 nm with a slight increase in the absorbance at the maximum. This is shown by the time-cuts in Figure 26 in which the decay of the uuo nm absorption is compared with the growth and decay of the 335 nm absorption. We conclude that the absorption band at 335 nm is the spectrum of an intermediate species as suggested by Dunn (109). Also shown in Figure 26 is the growth and decay of a new absorption at 5&0 nm. (Since these experiments we have learned that this band has also been observed by other workers (112)). The position of the maximum absorbance may be beyond the scan range used. However, we will refer to the absorption as the "5&0 nm band". Essential to the determination of possible overall mechanisms was the question of whether or not the 335 nm and 5&0 nm intermediates were actually two absorption bands of the same species. It was not possible to quantitatively compare both wavelengths over their entire time course of reaction because of the underlying interference of the NDMA and NADH decays with the 335 nm absorption; that is, there were too few points during the growth.period at 335 nm to quantitatively correct for background decays while simultaneously fitting the data to rate expressions. However, the decay portions of the two curves (330 and 5h0) could be directly and quantitatively compared. After 123 .mndz mmouxo ou oso ma EGmNm pm ouconu0mno Hmsowmou one .Eoummm mama on» now onumcoao>o3 mooHuo> um masonoEHB 033 25... 0. 0 .om ousmsa 0 ¢ E:0¢¢. aauoqlosqv l2& approximately one second of reaction, the remaining changes in the absorption spectrum reflect only the changing con- centration of the intermediate species. The decay of the 335 nm band was found to be first order as shown by the fit to first-order kinetics in Figure 27. The spectrum in this region after the decay is that of the excess NADH with the peak position and band shape expected for this concentration of free and enzyme-bound NADH. The average first-order rate constant for decay of the 330 nm absorption (four pushes, see Table V) is 0.70 f .0& sec"1 (95% confidence level). The standard deviation of the rate constant for repreated pushes is about equal to that obtained from a single push. The decay at 5&0 nm is also first order with a rate constant of O.&&7 I 0.01 sec'1 (four pushes, see Table V and Figure 28). Since the ggmg_pushes are used to evalu- ate data at all wavelengths, there can be no doubt that the 5&0 nm decay is slower than at 335 nm. Therefore, the two absorptions do not represent the same species. 'Time-cuts from several wavelengths in the 3&0 nm region were fit to first order kinetics. The results are in agreement at 330, 335 and 3&0 nm (see Table V). There is a systematic increase in the first order constant, however, in going from 3&0 nm to 360 nm. It is worth emphasizing that the values reported horizontally in Table V are from gng_push, and secondly that these values are the result of fitting the data from the period 125 .Eoummm mndq ofiu Ca ucoamcouu ECmNm ofiu mo Mmuoo noUHOIumuHm .nm ousmam To... 25... 0. 0 0 ¢ N 0 u u u o. ... .u oee x4 x _ a _ a a 7 0 x x no 00 u u x n u x x n 00x x0 0 x0 1 x0 xu 0 xx 0 0 .. V no .4 V xx 0. 0x . x o 0 1 ex 0. 0 0x 0 a o 1 I x 0 x o..N 0 x 126 .Eoummm ma mamdfi 128 after NDMA has disappeared. Since the first order constant at 5&0 nm is the smallest, we infer the presence of an additional time dependent absorption band in the 3&0-360 nm region. In addition, the quantity (A - A.) becomes negative in the 325-380 nm region towards the end of the reaction (not shown), indicating that an additional absorbance is present at infinite time and appears with a half-life of approximately 10 sec. In similar experiments, using the chromophore trans -&- N,N - dimethylaminocinnamaldehyde (DACA, 398 nm) Dunn and Hutchison (110) observed the formation of a long wavelength intermediate (&60 nm) analogous to the 5&0 nm species observed with NDMA. In addition, they observed the slow formation of another species which they attribute to be an artifact resulting from the presence of NADH formed.yig the LADH catalyzed oxidation of residual ethanol in the preparation. Since no extreme precautions were taken in the present work, the very slow increase in absorbance around 360 nm at the end of reaction could be attributed to the same artifact. The rate and magnitude of this slowly forming band are too small to account for the variation of the first order constant with wavelength. One of the unique advantages of the rapid scanning stopped-flow experiment is the fact that rate data at several wavelengths can be compared by fitting data from 129 a single push. In this way, variations in conditions from push to push which may affect the rate constant and which often occur with fixed wavelength instruments are eliminated. The agreement (Table V) bath.within each push and among replicate pushes at 330, 335, 3&0 and at 5&0 nm is proof of the excellent reproducibility of the instrument. 6.3.&--Rate Law for &&O ang_5&0 nm The relationship between substrate decay and the growth of the two intermediates is central to a determin- ation of the overall mechanism of the system. The fact that a single substrate gives rise to Egg'intermediates and apparently only a single product (109) significantly complicates the problem. Since the substrate follows Michaelis-Menten kinetics, and the intermediates decay by first order processes, it was possible to relate the decay at &&0 nm to the growth and decay at 5&0 nm by using Equation (&). A zero order correction was included (not shown) to account for substrate decay during the time («’2 ms) which elapsed between the two wavelengths. Since the extinction coefficient of the 5&0 nm species is not known, the rate constantkgu0 actually represents the product of a rate constant (k), an extinction ks déX) = 5&0 ' 3o D 1 + KNDMA ' k5h0(x) (h) (NDMA) coefficient (E ), and the path length (1.85 cm). The 5&0 130 overall course of the 5&0 nm intermediate (X) was fitted by a Michaelis-Menten growth term (kg&0)’ and a first order decay (k§&0) which.of course is independent of both substrate binding (KNDMA) and total enzyme concen- tration (E0). The agreement between calculated and ex- perimental points shown in Figure 29 and Table VI are clearly consistent with the postulate that the growth of the 5&0 nm species is proportional to the rate of substrate decay. However, more recent work by 11mm (Private Communication) has shown that as the concentration of enzyme is increased, the decay time of substrate is decreased and its decay is followed by a first-order growth of a 560 nm band. In the present work only one concentration of substrate was used; hence no conclusions can be drawn about this matter. It is possible that our results represent a coincidental case in which the first- order growth at 5&0 nm accidentally matches our particular decay time of substrate. This possibility is supported by the presence of an isosbestic at 510 nm during the growth period. Figure 30 shows an expansion of the region from &00 nm to 5&0 nm during the decay of the &&0 nm band and the growth of the 5&0 nm absorption. Since the growth of the 5&0 nm species coincides with.the decay of the NDMA band, an isosbestic occurs at 510 nm. It should be noted that the maximum absorbance at 5&0 nm is only 0.0& absorb- ance units. In addition, note the agreement between the 131 .soummm mead or» as ouoaooeuoues ucoeoeouu eeovm one no aeooo one rezone .mm mucosa Ado: 2...... - 0. o 0 ¢ u 0 o as e. e an _ 4 a _ q _ A 4 10 x x x a ... o u 0 x x o u x 0 ex :0 o go: o . . u 16 x o o "a u ex ex 0 V on a. 8 o x m x . ex 0 xu u 0 3 o 0 "IX x on x xu o o 0 "inc. .3 x o 0 0 0 x0 0 .33.... 0 0 on .33.." :30 .38 v0. 132 TABLE VI Rate constants for the Michaelis-Menten growth and first order disappearance of the S40nm intermediate (from Equation 4) Push G _ . . D -l D -l k5407k'65401 E0 k540(sec ) k540(sec ) Code (From Table V) J 0.251:0.012 0.41610.014 0.44610.014 K 0.280:0.012 O.433:0.024 0.45610.013 L 0.274:0.016 0.387i0.015 0.449:0.015 M 0.31410.013 0.560:0.103 0.438:0.014 Note: KfiDMA=5o9X10-6M (average from Table IV) was used in fitting data to Equation 4. Total enzyme concentration(Eo)=3.62XlO‘6N 5540 is the extinction coefficient of the 540nm intermediate, and a =1.85cm. 133 .coauQMOmpo aesoamou on» ma v Esuuooam .moooo oueuumnsm mcausu Egoam no oaumoanma oz» mcasoam Eoumxm mama one ca mCOauQHOmno “anaemv ucoamcouu one Aecovvv outnumasm “.65 59.2263 one one oec _ _ 1 ‘ i q 1 1 1 1" H “ I t 1 1 I 1] , N . .om ousmam eauoqaosqv '0 131+ first order decay constant at 5&0 nm as determined by the decay portion only (Table V), and by fitting two wavelengths over the entire time course (Table VI). Although somewhat complicated by additional con- current decay processes, it appears that the band shift and growth at 335 nm (see Figure 26) also coincides with the disappearance of the NDMA band. The subtle changes in spectral shape in this region emphasize the advantages of the scanning system. The overall changes in the spectrum from 290 nm to .Jr . 5&0 nm are shown in Figure 31. Spectrum 1 is that at zero time; spectrum 2 corresponds to the time of maximum 335 nm absorption; spectrum 3 to the decay of the 335 and 5&0 bands after the substrate band has completely decayed; and spectrum & corresponds to the residual NADH absorption after decay of all intermediates. Although a number of kinetic pushes (reactions) were done with.LADH it should be noted thatI§;1_of the figures shown in this section were taken from a single BER- This emphasizes the distinct advantage of the scanning stopped-flow technique over those which utilize fixed wavelength devices. Although it is not possible to rationalize a unique overall mechanism for the LADH-NDMA system from the results of a single experiment, we feel that the present work significantly limits the number of alternatives. 135 .moaz an «ass mo ceauoeeou ooumaeuoo moea or» ea momeeno Houseman aaouoeo 0E: v Ihozw4m><3 can one can own cum oo» (3.41.. _ _ '7 ' .am ousmam BONVBHOSBV VII. SUGGESTIONS FOR FUTURE'WORK Although the performance of the rapid scan stopped- flow system in transient-state enzyme studies is very good, several steps can be taken to make it even better for such research. Specifically, the installation of a UV grade fused silica prism (recently purchased) will allow studies in the spectral region below 300 nm. Because the present sample and reference beams do 223. originate from the same portion of the lamp image, it is impossible to eliminate fluctuations in the output caused by wander of the Xenon arc. Therefore the existing optical system should be replaced with a fiber-bundle beam-splitter (currently being constructed by Schott Optical 00.). If the fibers within the beam-splitter can be randomized, the improvement in S/N should be significant and the set-up time reduced. The limited availability of most enzymes makes it mandatory to use small volumes of sample. Therefore the present flow system should be re-designed to permit the use of smaller samples. At the same time it would be worthwhile to investigate the use of polypropylene tubing in place of glass tubing for construction of the flow 136 137 system in order to increase the flexibility of the system and to minimize its breakage. The computer interfaced rapid scan stopped-flow system described in this work is an ideal tool for simultaneous study of the kinetics of substrates, products and spectral intermediates. Systems which show development of several spectral intermediates during catalysis, such as those involving 36 enzymes (TDH), can now be studied more effectively by observing the relative rates and the sequence of development of intermediate species. Thus the computer interfaced rapid scan stopped-flow system should go a long way towards helping the investigator to correlate complex kinetics data and to help sort out the kinetics and the mechanisms of enzyme reactions. REFERENCES l. M. Dixon and E. C. Webb, Enzymes, Academic Press, New York, 3rd ed., (1967). 2. H. ggtfreund, Eggyme Kinetics, Benjamin, New York, 19 . 3. K. J. Laidler, The Chemical Kinetics of Enz e Actions, Clarendon Press, Oxford,(1958). &. C. J. Gray, Eggyme Catalyzed Reactions Van Nostrand Reinhold, London, (1975). 5. Y. Shizuta, A. Nakazawa, M. Tokushige, and O. Hayaishi, J. Biol. Chem., ggg, 1883 (1969). 6. W. W. Umbriet, R. H. Burris and J. F. Stauffer, Manometric Techni ues Burgess Publishing 00., Minneapolis, (195%). 7. M. Dixon, Manometric Methods, Cambridge University Press, England 1 1 . 8. George G. Guilbault, Enz atic Methods of Analysis, Pergamon Press, New York, p. 31, (1970). 9. H. Pardue, R. Simon, Anal. Biochem. 2, 209 (196&). 10. R. K. Simon, G. D. Christian and W. C. Purdy, Clin. Chem. 1&, &63, (1968). 11. G. G. Guilbault, D. N. Krames and P. L. Cannon, Anal. Chem. 3Q, 8&2 (1962). 12. T. g§)Ridgway and H. B. Mark, Anal. Biochem. 12, 357 19 . 130 go Ge)Gu11bau1t and Fe Re Shu, “1310 Chem. Q, 2161 1972 e l&. S. J. Updike and G. P. Hicks, Nature 21&, 986 (1967). 138 15. 16. 17. 18. 19. 20. 21. 22. 23. 2&. 25. 26. 27. 28. 29. 30. 139 Howard V. Malmstadt, E. A. Cordos and C. J. Delaney, Anal. Chem., g, 26A (1972). C. D. Scott, C. A. Burtis, Anal. Chem. &5, 327A (1973). M. Eigen and L. C. de Maeyer, in Techniques of Organic Chemistgy V01. 8, Investigations of Rates and Mechanisms 0 Reactions, Intersflence, New Rrk, PP- " o M. Eigen in Fast Reactions and Primary Processes in Chemical Kinetics, ed. S. Claesson, Interscience, New YorE, (I957). T. 0. French and G. G. Hammes, Methods Enzymol., 16: 3. (1961;). G. C. Kreshick, E. Hamori, G. Davenport, and H. A. Scheraga J. Amer. Chem. Soc. 88: 2&6, (1966). H. Gutfreund in Annual Rev. Biochem., Esmond E. Snell, ed., Annual Reviews Inc., 132, 31:5 (1971). G. G. Hammes, Advan. Protein Chems., 23: l (1968). G. G. Hammes and E. J. del Rosario, J. Amer. Chem. Soc. 92: 1750 (1970). Q. H. Gibson in Annual Rev. Biochem., 35: 1435 (1966). K. Kirschner, E. Gallego, I. Schuster, and D. Goodall, J. Mol. Biol., 28, 29, (1971). K. Kirschner, J. Mol. Biol., 58, 51 (1971). J. E. Erman and G. G. Hammes, Rev. Sci. Instrum., 31 (6), 71.6 (1966). G. G. Hammes and P. R. Schimmel, J. Amer. Chem. Soc., £31, &665 (see also &6714.) (1965). S. L. Friess, E. S. Lewis, and A. Weissberger, Eds. Techniques of Organic Chemistry. Vol. 8, Investigation of Rates and Mechanisms of Reactions, IntersHence, 119637. B. Chance, Q. H. Gibson, R. E. Eisenhardt, K. K. Lonberg- Hohn, Eds., Ra id Mixi and Sampli Techniques in Biochemistry, feedemIc, New York, I ID. 31. 32- 33. 3&- 35- 36. 37. 38. 39. &0. Al. A2. h3. MS- &6. #7. L180 #9. 1&0 K. Kustin, Ed. Methods Enzygol., 1Q, (l96&). N. Papadakis, Ph.D. thesis, Michigan State University 197 . Q. H. Gibson and L. Milnes, Biochem. J., 191: 161 (196A). H. Gutfreund, An Introduction tgthe Study of Enzymes, Oxford: Blackwell, (1965). Re Lo Berger, Be Balks and He Fe Chapman, Rev. 8010 Instrum., 39: &93 (1966) "“‘““" B. Chance, D. DeVault, V. Legallais, L. Mela, and T. Ybnetani, See Ref. 18, &37. B. Hess, H. Kleinhans and H. Schleuter, z. Physiol. Chem., 351: 515 (1970). M. J. Milano, H. L. Pardue, T. E. Cook, R. E. Santini, 1(3. Margerum, and J.M.T. Raychba, Anal. Chem. &Q, 37& 197 . Q. H. Gibson, Methods Enzymol., 16: 187 (196&). H. N. Fernley and S. Bisaz, Biochem. J., 107: 279 (1968). """"" R. L. Berger and L. C. Stoddart, Rev. Sci. Instrum., 36: 78, 85 (1965). R. W. Lynn and E. W. Taylor, Biochemist , 9: 2975 (1970). J. A. Sirs, Trans. Faradag Soc., 5&, 201 (1958). R. H. Prince, Trans. Faradag Soc., 58, 838 (1958). B. Chance, Biochem. J., &6, 387 (1950). L. 0. Clark, Trans. Am. Soc. Artificial Internal Organs, ll, &1 (1956). R. L. Berger and L. C. Stoddart, in Rapid Mixigg and Samplinngechniques in Biochemis ry, see 6 . 30 L. H. Feldman, R. R. Dewald, and J. L. Dye, "Solvated Eleggron", Advances in Chemistry Series, No. 50, 163 19 . J. Le Dye, Accounts Chemo R630, A, 306 (1968). 50. 51. 52. 53. 5h. 55. 56. 57. 58. S9. 60. 61. 62. 63. 6&. 65. 66. 67. 68. 1&1 E. R. Minnich, L. D. Long, J. M. Ceraso and J. L. Dye, J. Amer. Chem. Soc., 35, 1061 (1973). J. L. Dye and L. H. Feldman, Rev. Sci. Instrum., B. Chance, Discussions Faradag Soc., 3.1, 120 (1955). Q. H. Gibson and L. Milnes, Biochem. J., 21., 161 (l96&). Q. H. Gibson, Methods in Enz olo , Vol. XVI, K. Kustin, Editor, Academic Press, New lEorl'c', 187-228 (1969). If. M.)Plowman, Enzyme Kinetics, McGraw Hill Book Co., 1972 . F. G. Heineken, H. M. Tsuchiya, R. Aris, Mathematical Biosciences, l, 95 (1967). S. I. Rubinow, and Joel L. Lebowitz, J. Amer. Chem. Soc., 23, 3889 (1970). D. S. Bates and Carl Frieden, J. Amer. Chem. Soc., 2&8, 7878 (1973). I. G. Darvey, S. J. Prokhovnik and J. F. Williams, J. Theoret. Biol., 13, 90 (1966). D. Garfinkel, L. Garfinke, M. Pring, S. B. Green, and B. Chance, Ann. Rev. Biochem., 39, &73, (1970). G. Hammes and P. Schimmel in The Enz mes, P. D. Boyer, Ed., Academic Press, New York, 0 . , 67 (1970). J. L. Dye and V. Nicely, J. Chem. Educ., &8, 141:3 (1971). Laboratory Manual 11 (Software), see also Ref. 32. see Ref. 32, Appendix. J. W. Frazer, Chem. Instrum., g (3), 271 (1970). R. E. Dessy and J. A. Titus, Anal. Chem., g5, 2911 A (1973). G. L. Booman, Anal. Chem., 11m. (1966). G. Lauer and R. A. Osteryoung, Anal. Chem., &9, 30A (1968). 69. 70- 71. 72- 73. 7&- 75. 76. 77. 78. 79. 80. 81. 82. 83. 8&. 85. 86. 142 P. M. Beckwith and S. R. Crouch, Anal. Chem., Lg.» 221 (1972). B. G. Willis, J. A. Bittilsofer, H. C. Pardu, and D. W. Margerum, Anal. Chem., &2, l3&O (1970). R. J. Desa, Q. H. Gibson, Comput. Biomed. Rev., g; &9& (1969). Me Jo Milano, H. L. Pardue, T. E. COOK, Re E0 Santini, D. W. Margerwm, J. M. T. Raycheba, Anal. Chem., LLQ’ 37h- (197(4) Tektionix Inc., Beaverton, Oregon. Laboratory Manual I (Hardware). Y. Shizuta and M. Tokushiga, Methods Enzmol” 17B, 575 (1971). K. Kirschner and B. Voigt, Ho e-Se ler's Z. Ph 3101. Chem., (Ger.), 343, 632 (1968)E.'”'_L— """"L""' R. Stancel, Ph.D. thesis, Michigan State University 1971 . We thank Professor W. A. Wood Dept. of Biochemistry, Michigan State University for the threonine dehydrase sample. E. R.)Minnich Ph.D. thesis Michigan State University 1970 . Personal communication to J. L. Dye from L. de Mayer B. K. Hahn and c. G. Enke, Anal. Chem., g5, 651A (1973). Texas Instruments Application Report, Line Driver and Receiver SN 55107 Series. P. Moore, S.F.A. Kettle and R. G. Wilkins, Inorg. Chem., 5(3), &66 (1966). Manufactured by the 8.8. White Dental Mfg. Co., Industrial Division, 10 East &Oth Street, New Yerk 16, New York. R. R. Dewald and J. M. Brooks, Rev. Sci. Instrum., g3, (11), 1612 (1970). Available from Schott Optical Co., Inc., York Avenue, Duryea, Pa.e 143 87. Le Dinh.Long, Ph.D. thesis, Michigan State University 1973 . 88. Be Chance.and Jo He Park, J. B1010 Chem., _2_Q»_2_’ 5093 (1967)- 89. R. A. Niederman, K. W. Rabinowitz, and W. A. Wood, Biochem. Biophys. Res. Comm., 3Q, 951 (1969). 90. M. Tokushige, and A. Nakazawa, J. Biochem. (Tokyo) 12, 713 (1972)- 91. H. Theorell and B. Chance, Acta. Chem. Scand., 5, 1127 (1951). 92. H. Theorell and J. S. McKinley - McKee, Acta. Chem. Scand., 35, 1811 (1961). 93. K. Dalziel, J. Biol Chem., 238, 2850 (1963). 9&. C. C. Wratten and W. W. Cleland, Biochemistry g, 935 (1963). 95. Ibid., g, 2u&2 (1963). 96. H. gh?orell and T. Yenetani, Biochem. Z., 338, 537 19 3 e 97. S. Taniguchi, H. Theorell and A. Akeson, Acta. Chem. Scand., 2;, 1903 (1967). 98. D. s. Sigman, J. Biol. Chem., ggg, 3815 (1967). 99. H. Theorell and K. Tatemoto, Arch. Biochem. Biophys., 100. R. H. Sarma and C. L. Woronuck, Biochemistry, ll, 170 (1972). 101. H. Theorell and R. Bonnichsen, Acta. Chem. Scand., 5, 1105 (1961). 102. H. Th?orell and T. Yenetani, Biochem. Z., 338, 537 1963 . 103. G. Geroci and Q. H. Gibson, J. Biol. Chem., gag, #275 (1967). 10“.. He Smd and He Theorell, EnZEIEGS, l, 26 (1962). 105. J. D. Shore, Biochemistry, 8, 1588 (1969). 106. S. A. Bernhard, M. F. Dunn P. L. Luisi, and R. Shack, Biochemistry, 2, 185 (1970). 107. 108 O 109. 110. 111. 112. 113. 1972, Abstract No. 144 f. C.)Louisi and R. Favilla, Biochemistry, 11, 2303 1972 . J. D.)Shore and H. Gutfreund, Biochemistry, 2, A655 1970 . M. F. Dunn and S. A. Bernhard, Biochemistry, 19, h569 (1971). M. F. Dunn and J. S. Hutchison (Private communication) J. T. McFarland and S. A. Bernhard, Biochemistry, 11, 1h86 (1972). P. Schack and M. F. Dunn, Abstracts of the 8th Meeting of the Federation Euro ean Biochem. Soc., Amsterdam, 503. M. J. Holland and E. W. Westhead, Biochemistry, lg, 2261 (1973). "'TlTliTlflfllLfijMfi!fliflitfliij'ifll'lflfiflfflflmfi'”