Sho Asakura

Reconstitution of bacterial flagella in vitro

The process by which bacterial flagellar filaments are reconstituted from flagellin molecules in vitro is examined by physical measurements (flow birefringence, viscosity, sedimentation) and by electron microscopic analysis, using flagella isolated from a strain of Salmonella. It is shown to be essentially reversible and characteristic of crystallization. In the presence or absence of salt, the flagella are depolymerized into monomers by heat treatment. At neutral pH, the repoly-merization takes place only in the presence of salts. In monomer solutions, however, spontaneous nucleation rarely happens and the solutions remain in a state of super-saturation. In order to polymerize the monomers in such solutions, it is necessary to add fragmented flagella. Then, the ends of added fragments act as nuclei, resulting in rapid formation of long flagellar filaments. In this process the number of filaments contained in each solution remains unchanged, and a one-to-one correspondence holds between the added fragments and the fully grown filaments. The rate of growth of flagellar filaments in vitro is determined under various salt conditions and found to be comparable to that observed in vivo.

The interaction between G-actin and ATP

Abstract By a mild treatment using an ion-exchange resin, the excess nucleotide present in G-actin solutions can be removed without appreciable direct loss of the polymerizability of actin. G-Actin gradually becomes inactive in the absence of excess adenosine triphosphate (ATP). Kinetic studies were made under various conditions to study this inactivation. The experimental data can be explained quantitatively based on a simple reaction formula involving a dissociation equilibrium between G-actin and ATP. Finally, the nature of the ATP-binding of G-actin is discussed from a thermodynamic point of view.

The cooperative nature of G-F transformation of actin

Abstract Experimental results are presented which suggest that the G-F transformation of actin is a cooperative phenomenon. One of these results is concerned with the relation between the initial rate of transformation and the G-actin concentration. The initial rate increases in proportion to the third to fourth power of G-actin concentration. The dephosphorylation of ATP on G-actin requires the cooperation of two or more G-actin molecules; and the formation of the nucleus of F-actin by three or four G-actins promotes the further transformation of G-actin. The addition of F-actin accelerates the transformation of G-actin and this is further evidence that the G-F transformation of actin is a cooperative phenomenon. Under various salt conditions the transformation rate increases remarkably when increasing amounts of F-actin are added. Moreover, immediately after the addition of F-actin, explosive dephosphorylation of ATP and transformation of G-actin occur.

Helical transformations of Salmonella flagella in vitro.

Abstract Helical transformations of reconstituted Salmonella flagella were visualized by dark-field light microscopy. Flagella from SJ670 strain were lefthanded helices with a pitch of 2.3 μm at neutral pH. When, however, the pH of the solution was lowered to 4.7, they were discontinouously transformed into close-coils with a pitch of 0.5 μm and a diameter of 1.2 μm, and a further lowering of the pH converted these coiled flagella into so-called curly ones, righthanded helices with a pitch of 1.1 μm. The transformation was rapid and reversible. Two other kinds of flagella (SJ25 and SJ30) also underwent such polymorphic conversions. Thus pH is an important factor in the control of flagellar transformation. As a result of the transformation, the degree of flow birefringence of a flagellar solution depends strongly on pH. Measurements of this parameter were useful in the study of the effects on the transformation of salt concentration and temperature.

Salmonella flagella: in vitro reconstruction and over-all shapes of flagellar filaments

In vitro reconstitution of flagellar filaments from monomeric flagellins was carried out by the method used previously ( Asakura, Eguchi & Iino, 1964 ) using flagella isolated from Salmonella strains SJ670, SJ25, SL23 and SJ30: the former three strains possess flagella of the normal type and carry different H-antigen types, and the last one is a curly flagellar mutant obtained from SL23. Monomeric flagellins and short fragments of flagella prepared from these strains were mixed in various combinations to re-form not only homogeneous filaments but also heterogeneous ones. Kinetic studies of these processes by viscosity measurement and by electron microscope observation showed: (1) that for initiating polymerization of any kind of monomer, it is necessary in practice to add short fragments or “seeds”, which need not necessarily be homologous with the monomer; (2) that the over-all rate of polymerization is largely dependent on the nature of monomer, the order of rapidity being SJ670 > SJ25 > SJ30 > SL23; (3) that polymerization of flagellin takes place at a rate less than first order with respect to the concentration of monomer; and (4) that copolymerization of monomers of SJ670 and SJ30 proceeds at a rate markedly less than the sum of polymerization rates of the constituent monomers. These results were consistently explained by assuming trans-conformation of flagellin on polymerization. Over-all shapes of reconstituted filaments were examined by a low-magnification electron microscope. Not only homogeneous filaments but also heterogeneous ones and copolymers could be classified into either one of the two major types, namely the normal and curly types. When a small amount of short fragment was used, the shape of reconstituted filament was determined by the nature of monomer. When, however, a moderately long fragment of SJ30 flagellum (curly) was used for polymerizing the comparable concentration of SJ25 monomer (normal), the grown part of each filament was found to be curly. Copolymerization of comparable concentrations of SJ670, SJ25 or SL23 monomer with SJ30 monomer produced filament of the curlytype, irrespective of the nature of the added fragment. These experimental results show a dimorphic nature of flagellins. The normal type filament was often transformed into the curly type when it was incubated for a long time. The transformed filaments were reversed into the normal type by treatment with low concentrations of pyrophosphate or ATP. From these results it is concluded that the normal type filament possesses an intrinsic ability to transform between the normal and curly types, depending on external conditions, and that this transformation is akin to dimorphic transition in crystals.

Mechano-chemical behaviour of F-actin

Even at very low ionic strengths, actin assumes a polymerized form of moderate size under sonic vibration. This was deduced from the following facts: (a) rapid formation of long F-actin filaments takes place after sonication for a short period; (b) the actin shows moderately high viscosity at the time of stopping the vibration; and (c) actin placed in a sonic field is not denatured by EDTA. The short actin polymer formed by the vibration splits added ATP enzymically. After stopping the vibration, i.e. when the transformation of the short actin polymers into long F-actin filaments takes place, a large amount of ATP is split. The initial velocity of this ATP splitting is proportional to the actin concentration. This finding was explained by assuming that each short actin polymer contains an appreciable proportion of partially interrupted structures of the F-form and their reformation is coupled with the ATP splitting. From the re-examination of the facts obtained from the sonic experiments, the following conclusions were drawn. (1) The interruption-reformation process is essentially reversible; (2) interruption is accelerated by mechanical forces; (3) reformation is accelerated by the dephosphorylation of added ATP. Taking into account these results, various dynamic properties of an F-actin filament which theoretically follow from the helical polymer model are presented.

Studies on co-operative properties of tropomyosin-actin and tropomyosin-troponin-actin complexes by the use of N-ethylmaleimide-treated and untreated species of myosin subfragment 1

When subfragment-1 of rabbit skeletal myosin was extensively modified with N-ethylmaleimide, the protein became strongly associable to actin in the presence of MgATP at low ionic strength, while the ATPase ceased to be activated by actin. Various concentrations of the modified protein were mixed with 10 μmol of pure actin or actin complexed with tropomyosin, and the fraction β of actin saturated with the modified protein in each mixture was determined by an ultracentrifugal method. We then added 0.3 μmol of unmodified subfragment-1 to the same sets of mixtures as used in the above experiments and determined the rate of ATP hydrolysis V by unmodified subfragment-1 as a function of β. A biphasic V-β relation was obtained for the tropomyosin-actin complex: when β was increased continuously from zero, the rate first increased substantially, had a maximum value more than tenfold larger than the initial at β ∼- 0.3, and finally decreased to zero. In contrast, the V-β profile for pure actin deviated downwards from a linear relation, showing that there was a weak repulsive interaction between the modified and unmodified subfragment-1 species bound to the actin filament. The occurrence of such a repulsion was interpreted in terms of a steric hinderance model. Assuming that the same kind of repulsion underlay the biphasic V-β relation for the tropomyosin-actin complex, we calculated the relation of V′-β in an ideal case where it was absent. The result was also biphasic. We studied regulated actin in the presence and absence of Ca2+ by the same method and obtained biphasic V′-β relations in both cases. The experimental results were analyzed by a two-state model based on the proposal of Bremel & Weber (1972) that, within tropomyosin-actin or the regulated actin complex, n actin monomers undergo “off”/“on” transitions as a unit. Interactions between units were ignored in order to estimate the apparent size n, as well as the equilibrium constant L for the transition in the absence of myosin heads. Within the framework of allosteric theory (Monod et al., 1965), we derived formulae fit for data analysis, found a satisfactory agreement of the experimental and theoretical results, and obtained values of n = 11, and L = 37 for the tropomyosin-actin complex, and n = 16, L = 9 for regulated actin in the presence of Ca2+. The parameters in its absence could not be determined separately from the V−β relation which, however, was well-approximated with a combination of n = 16 and L = 10,000. It was also shown that tropomyosin-actin complex in the “on” state activated subfragment-1 ATPase eightfold more strongly than pure actin, and 2.2 to 2.6-fold more strongly than regulated actin in the “on” state. The results are compared with those provided by Greene & Eisenberg (1980), Hill et al. (1980) and Trybus & Taylor (1980) and discussed in conjunction with the double helical structure of tropomyosin-actin and regulated actin filaments. A simple allosteric calculation is presented in the Appendix to explain the well-known biphasic dependence on substrate concentration of the rate of regulated actin-subfragment-1 MgATPase (Bremel et al., 1972; Weber & Murray, 1973), with a reference to Deshcherevsky (1977).

Dark-field light microscopic study of the flexibility of F-actin complexes

The flexibility of F-actin complexed with saturating amounts of myosin subfragments has been measured by the use of a dark-field light microscope and a high-sensitivity television camera. When dilute solutions of F-actin complexes were observed in the microscope, single filaments in flexural thermal motion were visible to the eye. Images of the fluctuating filaments were recorded on videotapes using the high-sensitivity camera, and these records were used for the analysis of fluctuation to calculate flexibility in the framework of statistical mechanics of thermal fluctuation in semi-flexible rods. The analysis was carried out by two different methods. In method A, we selected many filaments (the entire length appeared near focus occasionally in the limited period of 10 to 100 seconds), measured the mean square end-to-end distance 〈R2〉 of each filament during the period and also its contour length L, and calculated a parameter λ representing flexibility by the equation given by Landau & Lifshitz (1958): 〈R2〉 = [2λL − 1 + exp(−2λL)]2λ2. Then, we obtained a value for λ = 0.040 ± 0.010 μm−1 for the acto-heavy meromyosin filament at 24.0 °C ± 1.0 deg. C, and λ = 0.027 ± 0.005 μm−1 for the acto-tropomyosin-heavy meromyosin filament at the same temperature. In method B, still photographs were taken of the video screen to collect a great number of filaments or parts of filaments which appeared just in focus over their length, and the contour length L of each filament and the angle θ(L) between the tangents at its two ends were measured, on the basis of the assumption that the whole length of each filament was in a plane perpendicular to the direction of view. The data were treated statistically and the results were approximated with 〈cosθ(L)〉 = exp(−λL), which holds for an ensemble of filaments with flexibility λ but in two-dimensional thermal motion (Landau & Lifshitz, 1958). The λ-values obtained by this method for acto-heavy meromyosin and acto-tropomyosin-heavy meromyosin filaments were both in good agreements with those obtained by method A, confirming the reliability of our measurement. F-actin complexed with a saturating amount of myosin subfragment-1 was examined by method B, and its flexibility was shown to be little different from that of acto-heavy meromyosin filaments.

The G - F equilibrium in actin solutions under various conditions

Abstract The reversible establishment of equilibrium states in actin solutions, has been investigated under various solvent conditions. In the equilibrium state F-actins coexist with G-actins, the concentration of the latter being determined by the solvent condition. The physical feature of the G - F equilibrium is expressed as a kind of condensation phenomenon. The phase diagram of the equilibrium has been obtained over wide ranges of solvent conditions, i.e. , pH, concentrations of monovalent salts, magnesium, calcium, nucleotides, and organic solvents, and temperature. It has been found that the difference in the effects on the equilibrium of magnesium and calcium increases with increasing concentration of monovalent salts, with increasing pH, and with decreasing temperature. A parallelism has been found between the equilibrium level and the G - F transformation rate and, within the range of the solvent condition here examined, the structure of F-actin is little changed by the solvent condition.

Unidirectional growth of Salmonella flagella in vitro

In previous papers (Asakura, Eguchi & Iino, 1964, 1966), in vitro reconstitution of flagellar filaments from monomeric flagellins, using strains of Salmonella , was shown to be similar to crystallization. At physiological ionic strength and pH, reconstitution consists only of growth, that is, the polymerization of a monomer on to the ends of added fragments of flagella or “seeds”. The present study was undertaken to investigate whether growth takes place at two ends of each fragment or only at one of the two ends. For this purpose, we used two kinds of flagella, having i - and 1,2 -antigens, respectively, together with anti-flagella sera, anti- i and anti- 1,2 . Monomers and seeds derived from i - and 1,2 -flagella were cross mixed to produce heterogeneous filaments or “block copolymers”, which, after treatment with either one of the two antisera, were observed in an electron microscope with negative staining. When a product of copolymerization was treated with anti- i , for example, a part of each filment containing i -flagellin was uniformly labelled with the antibody and could be distinguished from another unlabelled part composed only of 1,2 -flagellin. In a large number of the filaments observed, the great majority were of the form of i-1,2 ; filaments of the form of i-1,2-i or 1,2-i-1,2 , were never found. On the other hand, it was shown that the addition of a small amount of seed to a mixture of i - and 1,2 -monomers results in the formation of long flagellar filments, each of which is uniformly labelled with anti- i and anti- 1,2 , respectively. In view of thes observations, it was concluded that during reconstitution flagellar filaments grow in a unidirectional manner. Abram, Koffler & Vatter (Abst. 45, 2nd Int. Cong. Biophys., Vienna , 1966) have reported that when negatively stained fragments of flagella ( Bacillus ) were observed in an electron microscope, only their distal ends appeared to be frayed. It will be shown in this paper that this end corresponds to the end where in vitro growth takes place.