- Email: [email protected]
Can Cytoplasm Exist without Undergoing Phase Separation?
Can Cytoplasm Exist without Undergoing Phase Separation? D. E. Brooks
Department of Pathology and Laboratory Medicine and Department of Chemistry, University of British Columbia, Vancouver, Canada V6T 2B5
~~~
~
Studies on such systems as the lens of the eye and theoretical considerations suggest that phase separation may well occur in cytoplasm. In this chapter, several issues relevant to this question are raised. It is suggested that while the interaction between water and the macromolecules in a mixture is proving crucial to their phase separation behavior, the abnormal water that is widely observed in cytoplasm and concentrated protein solutions is unlikely to constitute a thermodynamic phase in the sense of phase separation studies. The role of fixed structures in the cytoplasm, the likelihood that the volume of separated phases would be small and subject to spreading over the fixed structures and the expectation that much of the phase volume could be occupied and dominated by properties of the interface are also discussed. Finally, some experimental approaches to studying the existence of liquid-liquid phases in cytoplasm are proposed. While there is no proof that phase separation exists in cytoplasm, application of some of the techniques outlined might well provide more positive evidence for its presence. KEY WORDS: Phase separation, Bound water, Scanning probe microscopy, Confocal microscopy.
1. Introduction
The question posed by the title to this chapter is one of the central issues we have tried to explore in this book, the other being the possible consequences of such phase separation, which is addressed in the final chapter. As we discussed in the chapter by Johansson et al., the high average concentration of macromolecules in cytoplasm suggests that phase separation could well occur. But is it inevitable? This is less clear. International Review of Cytology, Val. 192
0074-76W00$30.00
321
Copyright 0 2oM) by Academic Press. AU rights of reproduction in any form reserved.
322
D. E. BROOKS
II. Cytoplasm of the Lens of the Eye In the chapter by Clark and Clark and in the published work from the Boston group (Liu et al., 1995,1996), the very interesting behavior of lens cytoplasm and some of its components is examined. Here is a case in which high protein concentrations are present but evolutionary pressure has acted to produce a system in which light is refracted and transmitted with virtually no losses due to light scattering. That is, fibrils, particulates, and regions of phase separation are absent under normal conditions. However, the referenced work shows that a relatively small decrease in temperature or a somewhat abnormal composition can result in phase separation with an attendant loss of optical clarity. Hence, under physiological conditions, the cytoplasm is a stable single phase but it exists quite close to the binodial on the phase diagram. Because of the requirement for the lens to focus incoming light, the refractive index, hence, the total protein concentration, must be high. Certainly, macromolecular crowding must be occurring in these circumstances, yet phase separation does not occur. Therefore, crowding by itself is insufficient to produce phase separation under all conditions. In fact, it is not apparent that pure crowding effects will necessarily potentiate all kinds of phase separation in macromolecular solutions. Phase separation occurs when the unfavorable entropy changes associated with the formation of phases (which are less “mixed” than the hypothetical single parent solution containing the combined components and volumes of the separated phases) are compensated for by the energetics of interactions among the components. These include paired interactions among all the distinguishable kinds of components in the system, particularly the solvent. Phase separation can result from repulsion (i.e., positive interaction energy) between two macromolecular components (as occurs, for instance, in dextradpoly(ethy1ene glyco1)lwater systems), from poor solubility (positive interaction energy between water and a macromolecule), or from attraction between two macromolecular species (compiex coacervation). Macromolecular crowding would certainly be expected to potentiate complex coacervation, since association reactions, in general, are found to be enhanced under crowding conditions (Zimmerman and Minton, 1993). In the chapter by Johansson et al., it is shown that a crowding agent that has high water solubility (i.e., an athermal interaction with solvent) lowers the critical concentrations for incompatible macromolecules if they do not mutually repel too strongly and if they differ in water solubility, as modeled in our Flory-Huggins (FH) calculations. It is also noted, however, that if the incompatibility between the macromolecules is stronger (more positive) than incompatibility of a macromolecule and the solvent, then phase separation is not affected by crowding by a soluble macromolecule.
CMOPLASM AND PHASE SEPARATION
323
While the FH calculation describes a somewhat different picture of the cytoplasm than does the hard particle exclusion model developed in most of the crowding literature, it has the advantage of being able to take into account the qualitative effects of repulsion or attraction between pairs of the components. A mean field model, which likewise takes into account the energetics of interactions among the three components of mixtures of two lens proteins and water, has recently been developed (Liu et al., 1995, 1996). The model successfully describes the phase separation observed as a function of temperature and composition. It demonstrates that weak attraction between the proteins and a difference in protein-water interactions are required to reproduce the phase separation observed in ternary mixtures in vitro. Hence, as in the work described in the chapter by Johansson et al., protein-water interactions are a major determinant of phase behavior. It would be of considerable interest to examine thermodynamic models of other cytoplasmic mixtures by any of the approaches capable of describing aqueous protein mixtures. Scaled particle theory, mean field calculations, and Monte Carlo computations (Lomakin et al., 1996) could all be examined. The theoretical approach may be at least as productive as an experimental examination of the same issues, due to the difficulties in demonstrating the presence of phases in cytoplasm, as discussed below.
111. Role of Bound Water
The evidence supporting the idea tha water in concentrated protein solutions and in cytoplasm in particular exists in at least two distinct states is compelling, as outlined in the chapter by Garlid. There is clearly a sizeable fraction of the mass of water present that is not osmotically active, in all likelihood because it is associated in some sense with the macromolecules and surfaces present; it will be referred to here as “bound.” As discussed above, water interactions are critical in predicting whether or not a macromolecular mixture will phase separate under crowded conditions. Does water bound to intracellular components, or its conjugate “free” water, constitute a thermodynamic phase in the sense of the chapter by Johansson er al. in this volume? There is no proof either way, but it seems to me that this is, in general, unlikely. As Cabezas points out in his chapter, a phase is a region in which the physical properties are continuous, at most slowly varying functions of position, bounded by an interface which is characterized by having an interfacial tension. For bound water to be inactive osmotically, it is presumably strongly hydrogen bonded either to itself or to the surface of a macro-
324
D. E. BROOKS
molecule or structure. The layer of bound water would then be expected to follow the contours of the surface very closely. Since most proteins are irregular in their surface topography and exhibit a variety of charged (negative and positive), neutral, and hydrophobic groups to the aqueous environment, it is very unlikely, in general, that the properties of the bound water would vary continuously and smoothly over the surface. One would expect instead islands of bound water with particular average orientations that would vary with the nature of the surface groups on the macromolecule. The average amount of bound water could be quite large, but probably not continuous. Hence, it would not satisfy the above definition of a phase. Another argument against the thermodynamic phase interpretation is the requirement for a bounding interface with a finite interfacial tension. The interface would have to delineate the boundary between the free and bound water. There would likely be a concentration gradient present in the vicinity of the interface since the bound water would not be expected to act equally well as a solvent for all small molecular species present (see chapter by Cabezas). The concentration differences on either side of the interface would result in a change in refractive index and a finite interfacial tension. The change in refractive index would produce observable light scattering if the effective diameter of the area bounded by the continuous interface was a significant fraction of the wavelength of light. In concentrated protein solutions, the interfaces would have to interact and, if they had a significant surface tension, the interfaces would tend to fuse and try to minimize their surface area in order to minimize the free energy of the interface. Hence, in the crowded conditions of the cytoplasm, one might expect reasonably large regions of bound water to accumulate by aggregating the macromolecules to which the water was bound. These accumulations, in turn, ought to scatter significant amounts of light. At least in the lens of the eye, however, such scattering is not observed. Since the phenomenon of bound water occurs so generally in protein solutions, it seems highly unlikely that it would be uniformly abnormal in lens cytoplasm. More likely, bound water does not tend to minimize the area of its interface with free water, because it lacks a significant interfacial tension. In this case, it would not be considered a separate phase in the thermodynamic sense. Assuming the lens components are not abnormal in water binding, the conditions that lead to phase separation in vitro when components of this system are examined presumably produce phases in which the proteins separate with their bound water associated with them. It is likely the hydrated species that undergo phase separation at the appropriate temperature and composition, as in any other phase separated mixture of macromolecules. In the cases in which agreement between observation and theory requires that the protein-solvent interaction energy by considered unfavorable (positive), the effective repulsion could result either from exposed
CYTOPLASM AND PHASE SEPARATION
325
hydrophobic amino acids or from bound water interacting unfavorably with free solvent, as discussed in the chapter by Garlid. If bound water does not form a thermodynamic phase in the above sense, this need not eliminate the possibility of phase separation occurring under crowded conditions in the cytoplasm.
IV. Role of Insoluble Structures As is pointed out throughout this book, the cytoplasm contains numerous organelles and fibrillar systems which appear to be insoluble and exist as nondiffusing, solidlike structures. The total exposed area of such structures could be very high (see chapter by Luby-Phelps). This high surface area would be likely to have a significant effect on the distribution of any thermodynamic phases present in the cytoplasm because of the likelihood that one phase would exhibit a higher affinity for a given structure than the second phase of a locally phase separated region. That is, one of the phases would tend to wet or spread over a particular solidlike structure in preference to the other as a further manifestation of the difference in properties of the macromolecular components of each phase. This would occur in such a way as to minimize the free energy of the phases in contact with the structure, perhaps producing highly asymmetrical geometries if the solidlike surfaces were themselves highly asymmetric (e.g., fibrils). In the absence of such structures, multiphase systems are known to form concentric sequences of spherical phases (to minimize the area of the interface) within one another (Bungenberg de Jong, 1949). The geometric features of this type that are visible within the cytoplasm are due to the presence of membrane-bound organelles or nuclei lying within the region bounded by the plasma membrane. Liquid-liquid structures exhibiting these forms have not been reported in cytoplasm to my knowledge. They may be absent because their geometry would be determined by their wetting behavior in contact with the fixed surfaces of the cytoplasm, the geometry of the latter determining the shape of any associated liquid phases.
V. Phase and Interface Volumes One of the interesting results, reported in the chapter by Johansson et al., when the effects of a high background concentration of uninvolved macromolecule on phase separation of two macromolecules was investigated, was the prediction of the very small phase volume that resulted.
326
D. E. BROOKS
This could well be a general result since most of the individual components in cytoplasm are present in low total amount. Hence, although crowding effects can be expected in some cases to significantly lower the concentrations required to produce two phases, the total amount of material available may limit the volume of the phase to a very small fraction of the total available volume. Such phases could be difficult to detect, as pointed out in the chapter by Johansson et af. Moreover, when the possible thickness of the interface, as discussed in the chapter by Cabezas, is considered, the properties of these small volume phases could be somewhat different from those expected based on experience with larger volumes of phases in which the interface occupied a small fraction of the phase volume. For instance, Cabezas gives as an estimate an interface thickness of perhaps 10molecular diameters. If a phase occupies a volume of diameter 1 p and the l!rgest macromolecule in the phase has a characteristic dimension of 100 A, the interface occupies about 60% of the phase volume. Hence, the properties of such a phase would be much more dominated by the interfacial properties than would be expected at first glance. Understanding just how such phases would behave in the complex surroundings of the cytoplasm will require much further work.
VI. Experimental Approaches to Detecting Phases in Cytoplasm While the theoretical possibility of phases appearing in cytoplasm seems very real, whether or not such phases actually exist in nature is a moot point. Experimental evidence for such phases would be invaluable. However, great difficulties attend such a demonstration because of the small volumes involved and the presence of organelles and surfaces that can be expected to complicate interpretation of observations and measurements in living cells. Nonetheless, the following avenues might be usefully explored. A. Microelectrode Measurements
Neurophysiologists have, for many years, examined the electrical properties of intact excitable cells by recording the electrical potential difference between the inside of the cell and the external bathing medium (Katz, 1966).While the physical form of the electrodes began as fairly large, crude glass tubes, suitable for recording from the very large axon of the giant squid, more recently, much finer electrodes have been employed. It is now routine for investigators interested in sensory cells in the visual and auditory
CYTOPLASM AND PHASE SEPARATION
327
organs, for instance, to use glass micropipettes drawn down to diameters too small to observe in the optical microscope (Crawford and Fettiplace, 1980). The inside tip diameters are estimated to be a few hundred nanometers in these cases. These very small electrodes are inserted through the cell membrane via oscillation or in small excursions controlled by piezoelectric drivers that allow fine control over the tip movement. The penetration of the membrane is detected electrically by a jump in potential (the membrane potential) or capacitance, not by visual observation. If such electrodes were used to probe the cytoplasm of a variety of cells, it is possible that the tip could sample any phases present. It is well known that the phases in aqueous macromolecular solutions frequently exhibit a potential difference of a few millivolts across the interface (Brooks et al., 1985), a potential which is within the measurement range of the microelectrode systems referred to above. The capacitance of the liquid-liquid interface would probably be extremely low, however, due to its high water content. Hence, some indication of the presence of phases might be obtained if a careful series of investigations was performed using this technology. €3. Scanning Probe Microscopy
Scanning Probe Microscopy (SPM) has provided a relatively new method for examining structures at the molecular level (Wiesendanger, 1994). It has the outstanding advantage, from the biological perspective, of allowing measurements to be made with a resolution of nanometers in aqueous media. Hence, living cells, working enzyme systems, etc., can be examined. This approach, in one mode, allows the force acting on the instrument’s tip to be recorded as a function of distance from a surface, which provides a characterization of the nature of the surface, The liquid-liquid interface of a phase boundary would be expected to show a characteristic force-distance response, depending on the wetting properties of the SPM tip for the phases involved. Hence, in principle, it might prove possible to distinguish liquid-liquid phase boundaries from, for instance, membrane-bound organelles or from cross-linked macromolecular fibers or surfaces on the basis of their response to the SPM tip. Clearly, much work on model systems would have to be done to allow any reasonable interpretation of such data. Also, some reproducible method for allowing the SPM tip access to the cell interior without diluting it so much that any phases present would be dissolved would have to be developed. Perhaps mechanically lysing the cell by sonication in a chamber filled with air equilibrated with an aqueous salt solution of the appropriate concentration (to avoid evaporation or dilution) could provide such access. While this approach is clearly pure speculation
328
D. E. BROOKS
at this stage, the unparalleled sensitivity, resolution, and versatility of SPM suggest that the method deserves consideration with respect to the demonstration of phases in cytoplasm
C. Confocal Microscopy The confocal microscope was designed to provide optical images in the presence of highly scattering backgrounds (Pawley, 1995). As such, it allows good observations to be made in the cytoplasm of cells. The method still suffers from the limitations of alI optical instruments, namely, that structures smaller than about the wavelength of light cannot be resolved. However, the use of fluorescence reporting molecules (see the chapter by LubyPhelps for examples) increases the effective resolution of the method, as signals from even single macromolecules can be detected and located in an image. Hence, if fluorescently labeled molecules could be utilized that either formed or partitioned strongly into cytoplasmic phases, there would be some chance of observing the geometry and behavior of such phases in a living cell. Information obtained perhaps from the in vitro experiments described below could be utilized to guide selection of appropriate molecular species to use as probes.
D. In vitro Studies on Isolated Cytoplasmic Components By far the best information on phase separation in cytoplasm has been obtained in the calf lens cytoplasm system described in the chapter by Clark and Clark. The approach has been powerful because it has been possible to isolate large amounts of the lens proteins and to study their phase behavior in vitro. This has allowed phase diagrams to be constructed and detailed theoretical models to be tested. There does not seem to be any reason why such experiments could not be camed out on cytoplasm from other organs. It should be possible, using bacterial, tissue culture, or animal sources, to isolate and fractionate cytoplasm from relatively large numbers of cells. This would, in principle, allow the composition and temperature dependence of phase separation to be examined in the laboratory, perhaps using confocal microscopy to visually examine relatively small volumes of concentrated solutions, to save on material requirements. Fluorescently labeled species isolated from the original mixtures could be examined for their potential as reporter moIecules for phase separation that could be applied in viva Certainly, such experiments would be difficult, expensive, and would require considerable development of purification protocols, but
CYTOPLASM AND PHASE SEPARATION
329
the results would seem to be sufficiently valuable for our understanding of the organization of cytoplasm that they would be worth undertaking.
VII. Conclusions It is interesting with respect to the question of whether or not phase separation occurs in cytoplasm that in the one system for which a strong evolutionary pressure exists to eliminate phase separation (the lens of the eye), a composition has evolved that does phase separate when the conditions are altered by only a small change in temperature or composition. It, therefore, seems legitimate to wonder if, in systems where the evolutionary pressure for transparency is not present, might phase separation not occur? There is presently no proof of its presence or absence in any of these systems. If, in fact, it is shown not to occur, it would suggest that further evolutionary advantage beyond transparency is gained by eliminating phase separation. What might this be? These questions cannot at present be answered. We hope, however, that by examining the issue and accumulating information relevant to it, such questions might be examined more definitively in the future. Acknowledgments I would like to expressmy appreciation to Drs. Tolstoguzov,Kopperschl&ger,Garlid, Cabezas, Johansson, Haynes, and particularly Harry Walter for provoking the interesting discussions and ideas that evolved while editing this book.
References Brooks, D. E., Sharp, K. A., and Fisher, D. (1985). Theoretical aspects of partitioning. In “Partitioning in Aqueous Two-Phase Systems: Theory, Methods, Uses and Applications in Biotechnology” (H. Walter, D. E. Brooks, and D. Fisher, eds.), pp. 11-85. Academic Press, Orlando, FL. Bungenberg de Jong, H. G. (1949). Morphology of coacervates. In “Colloid Science” (H. R. Kruyt, ed.), Vol. 2, pp. 433-480. Elsevier, New York. Crawford, A. C., and Fettiplace, R. (1980). The frequency selectivity of auditory nerve fibres and hair cells in the cochlea of the turtle. J. Physiol. (London)306,79-125. Katz, B. (1966). “Nerve, Muscle and Synapse.” McGraw-Hill, New York. Liu, C., Lomakin, A., Thurston, G. M., Hayden, D., Pande, A., Pande, J., Ogun, O., Asherie, N., and Benedek, G. B. (1995). Phase separation in multicomponent aqueous-protein solutions. J. Phys. Chem. 99,454-461.
330
0.E. BROOKS
Liu, C., Asherie, N., Lomarkin, A., Pande, J., Ogun, 0.. and Benedek, G. B. (1996). Phase separation in aqueous solutions of lens ycrystallins: special role of ‘ys. Proc. Nutl. Acud. Sci. U.S.A. 93, 377-382. Lomakin, A., Asherie, N., and Benedek, G. B. (1996). Monte Carlo study of phase separation in aqueous protein solutions. J. Chem. fhys. 104, 1646-1656. Pawley, J. B. (1995). “Handbook of Biological Confocal Microscopy.” Plenum, New York. Wiesendanger,R. (1994). “Scanning Probe Microscopy and Spectroscopy: Methods and Applications.” Cambridge University Press, New York. Zimmennan, S. B., and Minton, A. P. (1993). Macromolecularcrowding: Biochemical,biophysical and physiological consequences. Annu. Rev. Biophys. Biomol. Struct. 22,27-65.
Department of Pathology and Laboratory Medicine and Department of Chemistry, University of British Columbia, Vancouver, Canada V6T 2B5
~~~
~
Studies on such systems as the lens of the eye and theoretical considerations suggest that phase separation may well occur in cytoplasm. In this chapter, several issues relevant to this question are raised. It is suggested that while the interaction between water and the macromolecules in a mixture is proving crucial to their phase separation behavior, the abnormal water that is widely observed in cytoplasm and concentrated protein solutions is unlikely to constitute a thermodynamic phase in the sense of phase separation studies. The role of fixed structures in the cytoplasm, the likelihood that the volume of separated phases would be small and subject to spreading over the fixed structures and the expectation that much of the phase volume could be occupied and dominated by properties of the interface are also discussed. Finally, some experimental approaches to studying the existence of liquid-liquid phases in cytoplasm are proposed. While there is no proof that phase separation exists in cytoplasm, application of some of the techniques outlined might well provide more positive evidence for its presence. KEY WORDS: Phase separation, Bound water, Scanning probe microscopy, Confocal microscopy.
1. Introduction
The question posed by the title to this chapter is one of the central issues we have tried to explore in this book, the other being the possible consequences of such phase separation, which is addressed in the final chapter. As we discussed in the chapter by Johansson et al., the high average concentration of macromolecules in cytoplasm suggests that phase separation could well occur. But is it inevitable? This is less clear. International Review of Cytology, Val. 192
0074-76W00$30.00
321
Copyright 0 2oM) by Academic Press. AU rights of reproduction in any form reserved.
322
D. E. BROOKS
II. Cytoplasm of the Lens of the Eye In the chapter by Clark and Clark and in the published work from the Boston group (Liu et al., 1995,1996), the very interesting behavior of lens cytoplasm and some of its components is examined. Here is a case in which high protein concentrations are present but evolutionary pressure has acted to produce a system in which light is refracted and transmitted with virtually no losses due to light scattering. That is, fibrils, particulates, and regions of phase separation are absent under normal conditions. However, the referenced work shows that a relatively small decrease in temperature or a somewhat abnormal composition can result in phase separation with an attendant loss of optical clarity. Hence, under physiological conditions, the cytoplasm is a stable single phase but it exists quite close to the binodial on the phase diagram. Because of the requirement for the lens to focus incoming light, the refractive index, hence, the total protein concentration, must be high. Certainly, macromolecular crowding must be occurring in these circumstances, yet phase separation does not occur. Therefore, crowding by itself is insufficient to produce phase separation under all conditions. In fact, it is not apparent that pure crowding effects will necessarily potentiate all kinds of phase separation in macromolecular solutions. Phase separation occurs when the unfavorable entropy changes associated with the formation of phases (which are less “mixed” than the hypothetical single parent solution containing the combined components and volumes of the separated phases) are compensated for by the energetics of interactions among the components. These include paired interactions among all the distinguishable kinds of components in the system, particularly the solvent. Phase separation can result from repulsion (i.e., positive interaction energy) between two macromolecular components (as occurs, for instance, in dextradpoly(ethy1ene glyco1)lwater systems), from poor solubility (positive interaction energy between water and a macromolecule), or from attraction between two macromolecular species (compiex coacervation). Macromolecular crowding would certainly be expected to potentiate complex coacervation, since association reactions, in general, are found to be enhanced under crowding conditions (Zimmerman and Minton, 1993). In the chapter by Johansson et al., it is shown that a crowding agent that has high water solubility (i.e., an athermal interaction with solvent) lowers the critical concentrations for incompatible macromolecules if they do not mutually repel too strongly and if they differ in water solubility, as modeled in our Flory-Huggins (FH) calculations. It is also noted, however, that if the incompatibility between the macromolecules is stronger (more positive) than incompatibility of a macromolecule and the solvent, then phase separation is not affected by crowding by a soluble macromolecule.
CMOPLASM AND PHASE SEPARATION
323
While the FH calculation describes a somewhat different picture of the cytoplasm than does the hard particle exclusion model developed in most of the crowding literature, it has the advantage of being able to take into account the qualitative effects of repulsion or attraction between pairs of the components. A mean field model, which likewise takes into account the energetics of interactions among the three components of mixtures of two lens proteins and water, has recently been developed (Liu et al., 1995, 1996). The model successfully describes the phase separation observed as a function of temperature and composition. It demonstrates that weak attraction between the proteins and a difference in protein-water interactions are required to reproduce the phase separation observed in ternary mixtures in vitro. Hence, as in the work described in the chapter by Johansson et al., protein-water interactions are a major determinant of phase behavior. It would be of considerable interest to examine thermodynamic models of other cytoplasmic mixtures by any of the approaches capable of describing aqueous protein mixtures. Scaled particle theory, mean field calculations, and Monte Carlo computations (Lomakin et al., 1996) could all be examined. The theoretical approach may be at least as productive as an experimental examination of the same issues, due to the difficulties in demonstrating the presence of phases in cytoplasm, as discussed below.
111. Role of Bound Water
The evidence supporting the idea tha water in concentrated protein solutions and in cytoplasm in particular exists in at least two distinct states is compelling, as outlined in the chapter by Garlid. There is clearly a sizeable fraction of the mass of water present that is not osmotically active, in all likelihood because it is associated in some sense with the macromolecules and surfaces present; it will be referred to here as “bound.” As discussed above, water interactions are critical in predicting whether or not a macromolecular mixture will phase separate under crowded conditions. Does water bound to intracellular components, or its conjugate “free” water, constitute a thermodynamic phase in the sense of the chapter by Johansson er al. in this volume? There is no proof either way, but it seems to me that this is, in general, unlikely. As Cabezas points out in his chapter, a phase is a region in which the physical properties are continuous, at most slowly varying functions of position, bounded by an interface which is characterized by having an interfacial tension. For bound water to be inactive osmotically, it is presumably strongly hydrogen bonded either to itself or to the surface of a macro-
324
D. E. BROOKS
molecule or structure. The layer of bound water would then be expected to follow the contours of the surface very closely. Since most proteins are irregular in their surface topography and exhibit a variety of charged (negative and positive), neutral, and hydrophobic groups to the aqueous environment, it is very unlikely, in general, that the properties of the bound water would vary continuously and smoothly over the surface. One would expect instead islands of bound water with particular average orientations that would vary with the nature of the surface groups on the macromolecule. The average amount of bound water could be quite large, but probably not continuous. Hence, it would not satisfy the above definition of a phase. Another argument against the thermodynamic phase interpretation is the requirement for a bounding interface with a finite interfacial tension. The interface would have to delineate the boundary between the free and bound water. There would likely be a concentration gradient present in the vicinity of the interface since the bound water would not be expected to act equally well as a solvent for all small molecular species present (see chapter by Cabezas). The concentration differences on either side of the interface would result in a change in refractive index and a finite interfacial tension. The change in refractive index would produce observable light scattering if the effective diameter of the area bounded by the continuous interface was a significant fraction of the wavelength of light. In concentrated protein solutions, the interfaces would have to interact and, if they had a significant surface tension, the interfaces would tend to fuse and try to minimize their surface area in order to minimize the free energy of the interface. Hence, in the crowded conditions of the cytoplasm, one might expect reasonably large regions of bound water to accumulate by aggregating the macromolecules to which the water was bound. These accumulations, in turn, ought to scatter significant amounts of light. At least in the lens of the eye, however, such scattering is not observed. Since the phenomenon of bound water occurs so generally in protein solutions, it seems highly unlikely that it would be uniformly abnormal in lens cytoplasm. More likely, bound water does not tend to minimize the area of its interface with free water, because it lacks a significant interfacial tension. In this case, it would not be considered a separate phase in the thermodynamic sense. Assuming the lens components are not abnormal in water binding, the conditions that lead to phase separation in vitro when components of this system are examined presumably produce phases in which the proteins separate with their bound water associated with them. It is likely the hydrated species that undergo phase separation at the appropriate temperature and composition, as in any other phase separated mixture of macromolecules. In the cases in which agreement between observation and theory requires that the protein-solvent interaction energy by considered unfavorable (positive), the effective repulsion could result either from exposed
CYTOPLASM AND PHASE SEPARATION
325
hydrophobic amino acids or from bound water interacting unfavorably with free solvent, as discussed in the chapter by Garlid. If bound water does not form a thermodynamic phase in the above sense, this need not eliminate the possibility of phase separation occurring under crowded conditions in the cytoplasm.
IV. Role of Insoluble Structures As is pointed out throughout this book, the cytoplasm contains numerous organelles and fibrillar systems which appear to be insoluble and exist as nondiffusing, solidlike structures. The total exposed area of such structures could be very high (see chapter by Luby-Phelps). This high surface area would be likely to have a significant effect on the distribution of any thermodynamic phases present in the cytoplasm because of the likelihood that one phase would exhibit a higher affinity for a given structure than the second phase of a locally phase separated region. That is, one of the phases would tend to wet or spread over a particular solidlike structure in preference to the other as a further manifestation of the difference in properties of the macromolecular components of each phase. This would occur in such a way as to minimize the free energy of the phases in contact with the structure, perhaps producing highly asymmetrical geometries if the solidlike surfaces were themselves highly asymmetric (e.g., fibrils). In the absence of such structures, multiphase systems are known to form concentric sequences of spherical phases (to minimize the area of the interface) within one another (Bungenberg de Jong, 1949). The geometric features of this type that are visible within the cytoplasm are due to the presence of membrane-bound organelles or nuclei lying within the region bounded by the plasma membrane. Liquid-liquid structures exhibiting these forms have not been reported in cytoplasm to my knowledge. They may be absent because their geometry would be determined by their wetting behavior in contact with the fixed surfaces of the cytoplasm, the geometry of the latter determining the shape of any associated liquid phases.
V. Phase and Interface Volumes One of the interesting results, reported in the chapter by Johansson et al., when the effects of a high background concentration of uninvolved macromolecule on phase separation of two macromolecules was investigated, was the prediction of the very small phase volume that resulted.
326
D. E. BROOKS
This could well be a general result since most of the individual components in cytoplasm are present in low total amount. Hence, although crowding effects can be expected in some cases to significantly lower the concentrations required to produce two phases, the total amount of material available may limit the volume of the phase to a very small fraction of the total available volume. Such phases could be difficult to detect, as pointed out in the chapter by Johansson et af. Moreover, when the possible thickness of the interface, as discussed in the chapter by Cabezas, is considered, the properties of these small volume phases could be somewhat different from those expected based on experience with larger volumes of phases in which the interface occupied a small fraction of the phase volume. For instance, Cabezas gives as an estimate an interface thickness of perhaps 10molecular diameters. If a phase occupies a volume of diameter 1 p and the l!rgest macromolecule in the phase has a characteristic dimension of 100 A, the interface occupies about 60% of the phase volume. Hence, the properties of such a phase would be much more dominated by the interfacial properties than would be expected at first glance. Understanding just how such phases would behave in the complex surroundings of the cytoplasm will require much further work.
VI. Experimental Approaches to Detecting Phases in Cytoplasm While the theoretical possibility of phases appearing in cytoplasm seems very real, whether or not such phases actually exist in nature is a moot point. Experimental evidence for such phases would be invaluable. However, great difficulties attend such a demonstration because of the small volumes involved and the presence of organelles and surfaces that can be expected to complicate interpretation of observations and measurements in living cells. Nonetheless, the following avenues might be usefully explored. A. Microelectrode Measurements
Neurophysiologists have, for many years, examined the electrical properties of intact excitable cells by recording the electrical potential difference between the inside of the cell and the external bathing medium (Katz, 1966).While the physical form of the electrodes began as fairly large, crude glass tubes, suitable for recording from the very large axon of the giant squid, more recently, much finer electrodes have been employed. It is now routine for investigators interested in sensory cells in the visual and auditory
CYTOPLASM AND PHASE SEPARATION
327
organs, for instance, to use glass micropipettes drawn down to diameters too small to observe in the optical microscope (Crawford and Fettiplace, 1980). The inside tip diameters are estimated to be a few hundred nanometers in these cases. These very small electrodes are inserted through the cell membrane via oscillation or in small excursions controlled by piezoelectric drivers that allow fine control over the tip movement. The penetration of the membrane is detected electrically by a jump in potential (the membrane potential) or capacitance, not by visual observation. If such electrodes were used to probe the cytoplasm of a variety of cells, it is possible that the tip could sample any phases present. It is well known that the phases in aqueous macromolecular solutions frequently exhibit a potential difference of a few millivolts across the interface (Brooks et al., 1985), a potential which is within the measurement range of the microelectrode systems referred to above. The capacitance of the liquid-liquid interface would probably be extremely low, however, due to its high water content. Hence, some indication of the presence of phases might be obtained if a careful series of investigations was performed using this technology. €3. Scanning Probe Microscopy
Scanning Probe Microscopy (SPM) has provided a relatively new method for examining structures at the molecular level (Wiesendanger, 1994). It has the outstanding advantage, from the biological perspective, of allowing measurements to be made with a resolution of nanometers in aqueous media. Hence, living cells, working enzyme systems, etc., can be examined. This approach, in one mode, allows the force acting on the instrument’s tip to be recorded as a function of distance from a surface, which provides a characterization of the nature of the surface, The liquid-liquid interface of a phase boundary would be expected to show a characteristic force-distance response, depending on the wetting properties of the SPM tip for the phases involved. Hence, in principle, it might prove possible to distinguish liquid-liquid phase boundaries from, for instance, membrane-bound organelles or from cross-linked macromolecular fibers or surfaces on the basis of their response to the SPM tip. Clearly, much work on model systems would have to be done to allow any reasonable interpretation of such data. Also, some reproducible method for allowing the SPM tip access to the cell interior without diluting it so much that any phases present would be dissolved would have to be developed. Perhaps mechanically lysing the cell by sonication in a chamber filled with air equilibrated with an aqueous salt solution of the appropriate concentration (to avoid evaporation or dilution) could provide such access. While this approach is clearly pure speculation
328
D. E. BROOKS
at this stage, the unparalleled sensitivity, resolution, and versatility of SPM suggest that the method deserves consideration with respect to the demonstration of phases in cytoplasm
C. Confocal Microscopy The confocal microscope was designed to provide optical images in the presence of highly scattering backgrounds (Pawley, 1995). As such, it allows good observations to be made in the cytoplasm of cells. The method still suffers from the limitations of alI optical instruments, namely, that structures smaller than about the wavelength of light cannot be resolved. However, the use of fluorescence reporting molecules (see the chapter by LubyPhelps for examples) increases the effective resolution of the method, as signals from even single macromolecules can be detected and located in an image. Hence, if fluorescently labeled molecules could be utilized that either formed or partitioned strongly into cytoplasmic phases, there would be some chance of observing the geometry and behavior of such phases in a living cell. Information obtained perhaps from the in vitro experiments described below could be utilized to guide selection of appropriate molecular species to use as probes.
D. In vitro Studies on Isolated Cytoplasmic Components By far the best information on phase separation in cytoplasm has been obtained in the calf lens cytoplasm system described in the chapter by Clark and Clark. The approach has been powerful because it has been possible to isolate large amounts of the lens proteins and to study their phase behavior in vitro. This has allowed phase diagrams to be constructed and detailed theoretical models to be tested. There does not seem to be any reason why such experiments could not be camed out on cytoplasm from other organs. It should be possible, using bacterial, tissue culture, or animal sources, to isolate and fractionate cytoplasm from relatively large numbers of cells. This would, in principle, allow the composition and temperature dependence of phase separation to be examined in the laboratory, perhaps using confocal microscopy to visually examine relatively small volumes of concentrated solutions, to save on material requirements. Fluorescently labeled species isolated from the original mixtures could be examined for their potential as reporter moIecules for phase separation that could be applied in viva Certainly, such experiments would be difficult, expensive, and would require considerable development of purification protocols, but
CYTOPLASM AND PHASE SEPARATION
329
the results would seem to be sufficiently valuable for our understanding of the organization of cytoplasm that they would be worth undertaking.
VII. Conclusions It is interesting with respect to the question of whether or not phase separation occurs in cytoplasm that in the one system for which a strong evolutionary pressure exists to eliminate phase separation (the lens of the eye), a composition has evolved that does phase separate when the conditions are altered by only a small change in temperature or composition. It, therefore, seems legitimate to wonder if, in systems where the evolutionary pressure for transparency is not present, might phase separation not occur? There is presently no proof of its presence or absence in any of these systems. If, in fact, it is shown not to occur, it would suggest that further evolutionary advantage beyond transparency is gained by eliminating phase separation. What might this be? These questions cannot at present be answered. We hope, however, that by examining the issue and accumulating information relevant to it, such questions might be examined more definitively in the future. Acknowledgments I would like to expressmy appreciation to Drs. Tolstoguzov,Kopperschl&ger,Garlid, Cabezas, Johansson, Haynes, and particularly Harry Walter for provoking the interesting discussions and ideas that evolved while editing this book.
References Brooks, D. E., Sharp, K. A., and Fisher, D. (1985). Theoretical aspects of partitioning. In “Partitioning in Aqueous Two-Phase Systems: Theory, Methods, Uses and Applications in Biotechnology” (H. Walter, D. E. Brooks, and D. Fisher, eds.), pp. 11-85. Academic Press, Orlando, FL. Bungenberg de Jong, H. G. (1949). Morphology of coacervates. In “Colloid Science” (H. R. Kruyt, ed.), Vol. 2, pp. 433-480. Elsevier, New York. Crawford, A. C., and Fettiplace, R. (1980). The frequency selectivity of auditory nerve fibres and hair cells in the cochlea of the turtle. J. Physiol. (London)306,79-125. Katz, B. (1966). “Nerve, Muscle and Synapse.” McGraw-Hill, New York. Liu, C., Lomakin, A., Thurston, G. M., Hayden, D., Pande, A., Pande, J., Ogun, O., Asherie, N., and Benedek, G. B. (1995). Phase separation in multicomponent aqueous-protein solutions. J. Phys. Chem. 99,454-461.
330
0.E. BROOKS
Liu, C., Asherie, N., Lomarkin, A., Pande, J., Ogun, 0.. and Benedek, G. B. (1996). Phase separation in aqueous solutions of lens ycrystallins: special role of ‘ys. Proc. Nutl. Acud. Sci. U.S.A. 93, 377-382. Lomakin, A., Asherie, N., and Benedek, G. B. (1996). Monte Carlo study of phase separation in aqueous protein solutions. J. Chem. fhys. 104, 1646-1656. Pawley, J. B. (1995). “Handbook of Biological Confocal Microscopy.” Plenum, New York. Wiesendanger,R. (1994). “Scanning Probe Microscopy and Spectroscopy: Methods and Applications.” Cambridge University Press, New York. Zimmennan, S. B., and Minton, A. P. (1993). Macromolecularcrowding: Biochemical,biophysical and physiological consequences. Annu. Rev. Biophys. Biomol. Struct. 22,27-65.