ABSTRACT
Myosin from molluscan adductor muscle is regulated directly by Ca2+ binding. In the absence of Ca2+ the ATPase activity is greatly inhibited. We review the application of transient kinetic methods to this system and show how they can be simple to perform and less ambiguous than steady-state methods.
INTRODUCTION
A characteristic feature of motor proteins is the need for the efficient inhibition of their ATPase activity during periods of inactivity. In turn, this property provides a way of characterising regulatory systems in vitro’, for example, the Ca2+ sensitivity of the steady-state ATPase activity provided the initial clue in the discovery of the diverse mechanisms involved in the regulation of actomyosin-based motors. The degree of control based on steady-state measurement can, however, be ambiguous. Thus a fivefold inhibition of the ATPase on removal of Ca2+ might indicate that all the molecules are subject to limited control, or that 20 % of the molecules are unregulated and display a permanently high ATPase activity while the remaining 80 % are inhibited by an unknown factor of ⪢ 5 fold. This problem can be overcome by investigating a single or limited turnover of the ATPase where the putative regulated and unregulated populations are resolved in time.
Provided the enzyme active site concentration exceeds the of the reaction for ATP, then on addition of stoichiometric amounts of ATP, the binding phase will be nearly complete before the enzyme—nucleotide intermediates begin to decay. The latter process can then be analysed in terms of a multiexponential decay, the amplitudes and rate constants of which reflect the relative concentrations and ATPase activities of each population. Alternatively, a 2-to 20-fold molar excess of ATP may be added to the enzyme, so that the unregulated population contributes a short steady-state phase while the regulated population undergoes a single turnover in the final phase of the reaction (Fig. 1). Variations of this method include chasing a fluorescent ATP analogue with a non-fluorescent one to follow the fate of enzyme-bound nucleotide, or addition of Ca2+ during the final phase to determine the degree of activation directly. All that is required for the application of these methodologies is a means for detecting the enzyme-nucleotide intermediate. In many cases this need not entail specialised equipment and indeed the measurement can be simpler and quicker than performing steady-state kinetic analysis.
In this paper we summarise the methods that we have used to study the regulatory system of the myosin ATPase of molluscan striated adductor muscle. Steady-state ATPase assays were instrumental in demonstrating the role of the myosin regulatory light chain subunits in the suppression of the ATPase activity in the absence of Ca2+ (Szent-Gyorgyi et al. 1973), but transient methods have provided new insight into the coupling between the Ca2+ and ATPase sites on the myosin head and the degree of regulation achieved by this mechanism.
METHODS AND RESULTS
The application of transient-state kinetic methods requires the ability to distinguish between ligand-bound and ligand-free states of the enzyme. While the following methods refer specifically to the myosin ATPase, many of the principles may well apply to other motor proteins (cf. kinesin, Hackney et al. 1989).
Physical separation
The ATPase activity of the inhibited states of the myosin ATPase are sufficiently slow that the kinetics of product release may be followed by physical separation of the protein from the ligands using rapid gel permeation chromatography. For example, column centrifugation was used to follow a single turnover of scallop striated adductor muscle heavy meromyosin (HMM). Stoichiometric [γ-32P]-ATP was added to 20 μM. HMM and after incubation times varying from 1 to 30 min, a 0.1 ml sample was applied to a lml G-50 Sephadex column in a tuberculin syringe. By spinning for lmin in a benchtop centrifuge, the HMM together with bound Pi was eluted, while the free Pi remained in the column. The rate constant for Pi release from the regulated population was determined as 0.002 s-1 in the absence of Ca2+, but >0.02 s-1 in its presence (Wells and Bagshaw, 1985). A comparable time resolution of separation (30 s) was also achieved using a small (5 × 40 mm) Superóse 6 column (Pharmacia) attached to an HPLC pump operating at a flow rate of lmlmin-1. Nevertheless both of these separation methods are limited to processes with rate constants of <0.02 s-1. The optical methods described below provide a much better time resolution but they are indirect.
Turbidity and light scattering assays
A motor protein is almost invariably involved in the transient association with another protein during the course of its mechanochemical cycle. In solution this can lead to a change in light scattering or turbidity of the sample. These signals may therefore be used to follow the kinetics of association or dissociation of the proteins, which may in turn be limited by some other chemical step.
In the absence of ATP, scallop HMM formed a rigor complex with actin which was turbid. Addition of a molar excess of ATP to the complex resulted in its rapid dissociation, and the solution remained clear during the steady-state hydrolysis phase (since the predominant intermediate is the dissociated M.ADP.Pi species). When the ATP was exhausted the actomyosin rigor complex quickly reformed to give a turbidity similar to its initial value. In the absence of Ca2+ the final phase was more complicated. Only 20% of the turbidity was rapidly regained, the remaining 80% recovered slowly with an exponential profile (k=0.002 s-1). It was concluded that the HMM sample comprised a mixture of 20 % unregulated molecules with a permanently high actin-activated ATPase activity (1.2 s-1) and 80% regulated molecules whose ATPase was suppressed some 650-fold in the absence of Ca2+ (Wells and Bagshaw, 1984). In the absence of Ca2+, the steady-state ATPase was dominated by the unregulated fraction which contributed a turnover of 0.24 s-1 (0.2 x 1.2 s-1) while the regulated fraction made a negligible contribution. Thus the degree of Ca2+ activation of the steady-state ATPase rate (5×) reflected the proportion of regulated molecules rather than the true degree of activation of the turnover rate itself. In conjunction with the rapid column separation experiments described above, it was concluded that Ca2+ was controlling the rate of Pi release, rather than the actin association steps which remained practically Ca2+ independent.
Turbidity changes are also associated with the formation of myosin filaments from monomeric myosin. Myosins from muscles which display thick-filament regulation share the characteristic of forming a folded, 10S conformation, which traps nucleotide at the active site and is inert with respect to actin activation and filament assembly (Cross et al. 1988). In order to form filaments, the 10S conformer must first release its nucleotide and straighten to the elongated 6S conformer. Filament formation, as monitored by a turbidity change, therefore provides a measure of the kinetics of product release from the 10S conformer. Using this approach, we have recently shown that scallop striated adductor muscle myosin is capable of forming a 10S conformer. This conformer is destabilised by the addition of Ca2+, the removal of the regulatory light chains or the labelling of the heavy chain reactive thiol (Ankrett et al. 1990). Fig. 2 shows the turbidity profiles of scallop myosin which had been solubilised by the addition of a 20 fold molar excess of ATP in the absence of Ca2+. As the ATP was exhausted there was an initial small increase in turbidity corresponding to filament formation by a population of extended 6S molecules, followed by a much slower rise ( of several hours) as the 10S monomers unfolded and participated in filament assembly. When Ca2+ was added to the assay at any point in time there was a rapid increase in the rate of filament formation owing to the activation of the unfolding transition. The pre-existing unfolded molecules dominated the steady-state ATPase and accounted for the initial phase of the reaction which led to the exhaustion of the ATP. Thus, if additional ATP was added to the assay in the absence of Ca2+, filament assembly was delayed by a time proportional to the added [ATP], but the final slow turnover of ATP by the 10S conformer was unaffected (Fig. 3). In this case it is clear that conventional steadystate assays of the preparation would not yield any information about the kinetics of product release from the 10S conformer. From these studies it is concluded that the 6S-10S transition is too slow to be involved in regulation of contraction, but may be involved in control of filament assembly during growth and development.
The choice of measuring turbidity or light scattering depends on several experimental factors. In both cases a wavelength (typically 320 to 340 nm) is selected just beyond the protein absorption band and preferably in the range where the tungsten lamp can be used. Light scattering can be measured in commercial fluorimeters with good sensitivity, but it is more prone to artifacts from aggregates and arc light sources tend to be unstable. Furthermore the signal is recorded on an arbitrary scale which makes comparison between experiments performed on different instruments more difficult. Turbidity can be recorded using a conventional spectrophotometer, although the reading will depend to a limited extent on the degree to which the optics reject forward low-angle scattering. For the kinds of measurement described here, the amplitudes are interpreted in relative terms and therefore these instrumental characteristics are rarely critical.
Fluorescence assays
Myosin is not particularly specific with regard to its substrate, ATP, and a number of fluorescent analogues have been developed which show large enhancements in the emission signal on binding to the active site. We have used formycin triphosphate (FTP) extensively, which is commercially available from Sigma Chemical Co., although it is considerably cheaper to purchase formycin monophosphate and convert it to the triphosphate enzymically (Jackson and Bagshaw, 1988a). By selecting a λex=313nm, the background contribution from protein fluorescence is reduced while a convenient line may be selected when using a Hg arc lamp. For extension into in vivo measurements, fluorophores responding to longer wavelengths may be required, such as mant-ATP (λex=350nm; Hiratsuka, 1985).
In a typical experiment, a 5-fold excess of FTP was added to scallop HMM and the fluorescence emission was monitored at 350 nm. This ratio is a compromise. When less was used (or a true single turnover measured) the FTP binding phase was not complete before the product decay started (particularly when manual addition was employed) and hence the ratio of the regulated and unregulated HMM fractions could not be determined. When more FTP was added, the background fluorescence from the free formycin nucleotide was higher and hence the change in signal during binding and turnover was reduced. In the absence of Ca2+ a limited FTP turnover assay has a similar profile to that observed using turbidity measurements described above. Following the binding phase there is a short steady-state phase of enhanced fluorescence during which time the unregulated HMM turns over the bulk of the FTP. A partial drop in fluorescence then occurs when the FTP is exhausted, leaving a final exponential phase of fluorescence decay as the products are released from the regulated fraction. The profile, however, is not easy to analyse since an undetermined amount of product FDP remains bound to the active site. For this reason, and also to obtain an estimate for the rate of product release from the unregulated fraction, we normally ‘chase’ the FTP turnover with a large excess of ATP during the brief steady-state phase (Jackson and Bagshaw, 1988a). This method is particularly successful with myosin because the back dissociation rate constant for bound triphosphate is extremely small (Bagshaw and Trentham, 1973) leaving turnover the only route available for the bound FTP. More recently we have used pyrophosphate (PPi) as the non-fluorescent chasing agent since it is not hydrolysed and also causes less problems with solubilisation when assaying filamentous myosin (Walmsley et al. 1990; Ankrett et al. 1990). In the chase assay, the product release steps are revealed as an exponential with at least two phases corresponding to the unregulated and regulated myosin populations. The FTP turnover assays provide a convenient way of following product release steps and confirm the kinetic parameters deduced by the other assays described above. The initial association rate of FTP to myosin, and many other ATP requiring enzymes, is reduced by about 20-fold, but once bound the kinetics appear to mimic those of ATP closely. FTP also becomes trapped at the active site of 10S myosin and provides a sensitive assay for this state (Cross et al. 1988; Citi et al. 1989; Ankrett et al. 1990).
Fluorescence spectroscopy also provides a convenient means of following the kinetics of Ca2+ binding to proteins. In practice, many Ca2+-binding proteins are near saturated at contaminant levels of Ca2+ and it is more convenient to follow the kinetics of dissociation. In any event such measurements, in conjunction with the equilibrium constant, indicate that the association reaction would be too fast to measure by rapid mixing methods at the concentrations required to obtain an adequate signal. Intrinsic protein fluorescence (tryptophan or tyrosine) often provides a convenient probe. Fig. 4A shows the release of Ca2+ from the specific, high affinity sites of scallop heavy meromyosin on mixing with EGTA. Better signal-to-noise ratios can be obtained by using the fluorescent indicator, quin-2 as a combined chelator and probe (Fig. 4B). These procedures can potentially monitor different steps in the Ca2+ release mechanism, but in this case they are probably both limited by the same conformational change which occurs at about 40 s.
Instrumental considerations
As mentioned above many of the key features of the regulatory mechanism of the molluscan myosin ATPase activity were evaluated using a standard spectrophotometer and fluorimeter. The recent introduction of ultramicrocuvettes taking 0.1ml samples (Hellma Co.) has increased the scope of these measurements for cases where material is limiting. There are, however, a number of assays where more specialised approaches are warranted. The ATPase rates when activated by Ca2+ are generally too fast to be evaluated by manual mixing methods and require stopped-flow instruments for transient kinetic analysis. Such apparatus is available commercially (e.g. Applied Photophysics Ltd, Hi-Tech Ltd) or can be constructed in the laboratory from commercial components (e.g. Ealing Electro-optics Ltd, Oriel Corporation). In either case the light source and detector are modular and therefore can be arranged to suit the experiment at hand. For example, we have made extensive use of these components in conjunction with a custom-built fluorescence housing which accommodates a standard cuvette (Jackson and Bagshaw 19886). The instrument comprises a light guide input for excitation, two 90° photomultipliers for fluorescence and/or light scattering detection and a photodiode for transmittance detection. The latter has proved of great value for detecting artifacts in the fluorescence signal due to inner filter effects, arc lamp jumps or changes in the assembly state of the myosin. Components may be added to the cuvette through a syringe needle and the contents stirred by an overhead paddle, allowing readings to be obtained within about 2 s.
The spectroscopic signals are digitised and stored using a microcomputer for subsequent analysis. The main problem facing the experimentalist these days is not so much finding an appropriate system for data capture but achieving compatability between different instruments so that data may be moved freely between systems for analysis and presentation. The latter can be achieved by hardwire connections through the serial ports and a communications package for ASCII files or by using emulation software.
Transient kinetic profiles usually comprise one or more exponential phases and non-linear, least squares analysis can be used to extract the amplitudes and rate constants. The efficiency of the control of the scallop myosin ATPase leads to a wide dynamic range in the observed transients so that selecting an appropriate timebase for the record can cause difficulties. In these critical situations we have digitised records using a logarithmic timebase to ensure that the fast and slow phases are defined by a sufficient number of points and are subject to even weighting in the analysis (Walmsley and Bagshaw, 1989; Walmsley et al. 1990).
As in all hypothesis testing, a match between the observed and simulated records does not prove that a mechanism is correct, but an inappropriate mechanism may be ruled out.
These mechanisms only become distinguishable if the rate constant for trapping is comparable or slower than the rate constant for turnover, in which case the pathway will take several turnovers to build up the trapped state (M.*ADP.Pi) to its steady-state value. Thus the extent of trapping, as estimated from the amplitude of the slow phase of product release, will be dependent on the time of chasing the system during a limited turnover experiment.
CONCLUSION
Motor proteins require rigorous control so as not to waste chemical energy during periods of inactivity. In vitro studies of myosin have now revealed a number of preparations where long-lived intermediates exist under relaxing conditions, but which are capable of rapid activation on addition of Ca2+. The true dynamic range of these preparations has only been revealed by transient state methods. However, because the time scale of these transients extends over several minutes, or even hours, specialised rapid reaction equipment is not required for their detection. This methodology should find application in other motor systems.
ACKNOWLEDGEMENTS
We are grateful to the SERC for financial support.