Are Biological Cells Turing-Compatible Quantum Computers?

Active Generation and Stabilisation of Convective Phospholipid Flows Solves the Decoherence Problem and Enables Topological Quantum Computation and Error Correction by Biological Cell Systems

Corresponding Author: Dr. J.T.Lofthouse. MIRth Encryption plc., 21 Lytton Road, Springbourne, Bournemouth, Dorset. BH1 4SH. UK.
Email: DrLofthouse@lycos.co.uk

Abstract:

Keywords: Biomembranes/Topological Quantum Computation/ Decoherence/Error Correction/ Convection/Magnetohydrodynamics

Introduction

Biologist's somewhat limited notions of cells as 'computers' have been almost entirely classical-level, binary paradigms (eg Conread, 1990; Bray, 1995). Focussing exclusively on the cells proteins as informational processors or 'bits', signalling molecules such as hormones, ions or proteins are all assumed to operate by changing their three-dimensional shapes (conformational states), thereby altering their ability to bind others. Whilst Fluid Dynamicists are aware that vast amounts of information can be stored in viscous fluid flows, Cell Biologists prefer to dismiss the highly complex motions of biological cells fluids - the membrane and cytoplasm in which their proteins are continuously embedded - as 'random'. There is clear evidence that certain classes of lipid, particularly phospholipids, are crucial to cell signalling processes (eg inositol lipids), however, the prevailing rubric restricts their role of to that of a 'molecular key' that binds to a protein 'bit', changes its conformation, and renders it able/not able to bind another protein and change its shape etc. This notion that biological cells process information solely by sequential changes in the shape of proteins presents cellular information processing as a highly inefficient 'semaphore' system: given the speed and sheer volume of information they parallel process, the mechanisms used must be significantly more sophisticated than this. Since proteins are undoubtedly 'conductive' In Vitro, theoretical conjectures that they act as classical level conduit for electrical signals in the living cell (Pullmann and Pullman, 1962: Tuszynski et al, 1998) have been posited as alternative computational paradigms, but none of the mechanism suggested to enable them to function as conduit In Vivo have appeared convincing from a Cell Biologists' perspective. At the sub-classical level, there have been more recent speculations that cellular protein molecules might perform 'Quantum Computations', but these have met with almost universal derision, largely because of what is termed 'the Decoherence problem'.

'Quantum computation' is an overall description of several computational models, each based on a theoretical ability to manufacture, manipulate and measure a quantum state. Decoherence arises when the components of a quantum mechanical state lose phase coherence. Entanglement of a system with its environment (as arises for example, in the process of attempting to measure it) delocalises quantum coherence to the combined system, which therefore shows the same properties as an ensemble of states without any coherence between them, and changes the behaviour of the system from quantum mechanical to classical. Like their classical level forebears, all biological QC paradigms posited to date have focussed exclusively on protein molecules holding the 'coherent state', conformationally functioning as 'qubits'. Frohlich (1968, 1979, 1975) was the first to conjecture that sets of protein dipoles might undergo coherent conformational excitations if thermal energy was supplied from the surrounding heat bath, but provided no feasible means of preventing its Decoherence in their aqueous ionic environment. In more recent years, one particular protein, Tubulin, whose subunits assemble to form highly ordered cylinders ('microtubules') in most cells, has received particular attention as a protein 'qubit'. Hameroff et al, (1996) suggested that the range of conformational states these protein subunits adopted indicated they functioned as 'qubits', and further (Jibu, Hameroff et al. 1994) suggested that once assembled into tubules, they could function as waveguides. Biologists dismissed Hameroffs' paradigm chiefly because the suggested energy source for pumping Tubulin subunits into coherent states (hydrolysis of GTP molecules bound to the subunits) is known to actually fragment microtubules (!), (beta-Tubulin subunits in microtubules have bound GDP that cannot be exchanged, alpha Tubulin subunits do not hydrolyse the GTP attached to them In Vivo, and if the beta-Tubulin GTP-cap is hydrolysed, the microtubule actually falls apart: for additional criticisms in Grusch & Churchland, 1999;. Sieffe, 2000, Scott, 1996: Georgiev, 2i003). However, it was the absence of viable shielding mechanisms to isolate coherent states from the aqueous and ionic cytoplasm in which they were embedded that led most Physicists to dismiss 'Quantum Computation' by this or any other cell protein outright (eg Tegmark, 1999). Whilst several hypothetical means of shielding from entanglement were subsequently suggested (ordered water sheaths around proteins, cyclical 'gel-sol' alterations to the cytoplasm which embeds them) no evidence these arise, let alone were capable of remaining stable in the metabolically active cell has ever been presented.

The Decoherence problem then, appears to exclude biological quantum computation altogether, however, more advanced understandings of the theory of Quantum Computation have shown there are two viable strategies to facilitate it that do not demand total isolation of coherent states. The first option is to enable Quantum Error Correction (QEC) to take place. In classical systems, noise, and the system's tendency to depart from the desired state can be suppressed using detection and feedback (eg electronic servo devices based on a stable voltage reference, high-performance amplifiers and low-noise resistors). Passive stabilisation can only be applied to a quantum regime by making the perturbations small rather than by making the system large, and is therefore ruled out on practical grounds. Active stabilisation is based on amplification and dissipation, and since in general terms a) an unknown quantum state cannot be amplified, ('no cloning theorem') and b) dissipation generally prevents unitary evolution, actively stabilisation of a quantum state appears impossible. However, the concept of quantum error correction (QEC) has demonstrated it is possible to restore a quantum state which has been affected by noise (Steane et al, 1997) in a 'fault tolerant' manner, which is itself sufficiently noise insensitive during the corrective process to allow overall stabilisation. Steane (op cit) has demonstrated that the task of applying QEC in the lab essentially rests upon how the quantum state is synthesised. My reading of this work is that QC by biological cells would be feasible they posessed two interacting systems of qubits, with ancillae reversibly mediating between them, performing error correction, and possessed a dissipative process that could generate a coherent state, but none has been posited prior to the model presented in this paper. The second way of circumventing the Decoherence problem would involve cells using TCQ of the type envisaged by Witten et al, 1989; by Kitaev (1997) and by Freedman (2002). Wittens idea for a topological quantum computer is one in which particles moving on a surface trace out paths in space-time, which can intertwine to form braids. If the particles are "Anyons", one can recapture information about a braid by measuring physical properties of the Anyons after they stop moving. The idea of a physical system of knots or other loops performing a computation sounds strange, but examples of such systems abound at the classical level. In an electrical transformer, for instance, two separate loops of wire are coiled around an iron core. An electric current passing through one wire can generate a voltage in the other that is proportional to the number of times it twists around the core. Thus, without seeing the first wire, the number of twists is calculable by measuring a voltage output. Likewise, the Quantum Hall Effect can be exploited to measure either magnetic fields or currents. By encoding information in braids instead of single particles, a small disruption from the environment might disturb the threads of a knot slightly but is unlikely to change its overall knottedness- thus, the topological quantum computer has a built-in defence against Decoherence. TQC would again circumvent the Decoherence problem altogether in 'warm wet and windy' cellular environments, but no cellular mechanism with the necessary attributes to perform it have been posited to date.

This paper uses a new Fluid Dynamics-based model of biological cells to suggest they do indeed possess components capable of performing both TCQ and QEC. Whilst most Physicists are familiar only with low temperature coherent states ( Bose-Einstein condensates, Quantum Hall fluids), and consequently identify 'heat' as an enemy that causes decoherence, high-temperature coherent states do exist, but are not widely known outside the Fluid Dynamics community. Two phenomena in particular, - Benard-Marangoni and Raleigh-Benard convection - use heat constructively to generate coherent fluid states in thin viscous layers when a critical temperature gradient is established across them. Above a critical gradient, viscous fluids such as silanes and nematics begin to flow coherently in a limited range of tesselated toruses, forming hexagonal, triangular, roll and square patterns (Mpegs available at here for those who have never seen them: comprehensive reviews of this field are available in Koschmieder, 1974, 1993). Shear-driven Couette-Taylor flows are also 'coherent ' fluid states, generated mechanically by the application of shear to thin viscous fluid layers, but no links between these phenomena and biological systems have previously been made. Spatially-patterned flows of electrically conductive fluid produce defined magnetic fields by a dynamo action, and it is these that I suggest significantly extend the cells information processing capacity.

In a series of papers (Lofthouse, 1998 a-i) I have shown that both types of coherent fluid states are manifest by the viscous fluids of 'Far from Equilibrium', metabolically active biological cell systems, and are evidenced merely by existing dynamic data. I demonstrated that cell membranes cannot have a 'Bilayer' structure in the metabolically active state, falsified this structure by considering the cell surface as a Topological manifold, and presented detailed evidence that far from being 'random', both membrane and cytoplasmic flows are actually spatially organised by forced convection and shear. This qualitative model has been shown capable of predicting all three major shape changes of the red blood cell in response to perturbations including the application of shear, a decline in metabolic rate, increased calcium levels, exposure to microgravitation and the application of electric fields. In this paper, I discuss how the model significantly extends the 'computational ' capacity of biological cells. The model suggests cells are able to use 'waste heat ' to force millions of membrane phospholipids (whose headgroup dipoles can function as 'qubits) into a coherent state capable of performing Topological QC. Furthermore, by taking into account the continuous low affinity interactions between these fluids and the cell's protein architectures, I show that in a manner identical to Steanes' schemes for QEC, coherent lipid states are actively generated and stabilised by ATP/GTPases, and that protein kinases, phosphorylases, and lipases act as ancillae, mediating interactions with the protein cytoskeleton. Finally, I demonstrate how the this protein/lipid feedback enables changes in the Topology of coherent fluid flows (induced by diverse intracellular and environmental factors) to convert spatially patterned flows into classical level output (current) through protein conduit without decoherence, whilst simultaneously shielding conduit from the aqueous, ionic environment. As the model suggests that cells 'computations' are determined by a 'computable in theory' fluid system, it also represents the first 'Turing compatible' biological QC conjecture.

The model draws together four disparate scientific fields (Fluid Dynamics, Cell Biology, Topology and Computing): this heavily abridged qualitative version is presented for those with prior knowledge of Topology and QC, who may be unfamiliar with the finer points of Cell Biology and Fluid Dynamics. If any existing system harnesses the power of QC, biological cells are strong candidates: hopefully reverse engineering a successful system could inspire construction of a workable 'high temperature' technological device with fewer limitations than currently envisaged low-temperature designs.

1. Why membranes cannot adopt a 'Bilayer' structure in the metabolically-active cell; the universally-observed phenomenon of aminophospholipid 'Flip-Flop' renders them self-intersecting surfaces, homeomorphic to an N-fold Torus. The new 'Dynamic Template' model.

Because it has become the most widely published visual meme in Cell Biology, we frequently forget that the 'Bilayer' membrane is merely assumed an accurate representation of the structure of the cell surface when it is metabolically active. Undoubtedly, two non-self intersecting surfaces compatible with this structure are seen in Electron micrographs of cell sections , but the preparative techniques used to obtain this visual information involve halting metabolic activity by freezing the cell, treating with glutaraldehyde which cross links proteins to lipids, embedding the cell in Araldite (!), cutting sections and finally coating these with a layer of carbon or metal atoms. It is fair to state that such processing probably does little to preserve or mimic the 'Far from Equilibrium' state in which the metabolically active, living cell operates! 'Bi-Layers' then, are only seen in highly perturbed cell systems that have been reduced to an 'Equilibrium' state, and it is only an assumption that they maintain this structure when the cell is metabolically active. In a 'BiLayer' structure, phospholipid molecules are assumed to traject entirely randomly on the surfaces of two concentric spheres, oriented such that their hydrophobic acyl chains are sandwiched between two layers of hydrophilic headgroups. Topologically, both of the 'leaflets' of a Bilayer membrane are homeomorphic to spheres - they do not self-intersect Sadly for text book illustrators, the assumption the membrane retains this structure in the metabolically active cell is soundly and completely contradicted by observations made using dynamic techniques (NMR and ESR).

When metabolically active, and continuously 'heated from within' by the enzymatic hydrolysis of polyphosphates (ATP, GTP) and exothermic reactions, it is universally accepted by the Cell Biology community that the aminophospholipids Phosphatidyl-serine (PS) and -ethanolamine (PE) not only to translate laterally across the membrane surface, but simultaneously undergo a motion- termed 'Flip-Flop' - through it (Lofthouse, 1998a). These motions render the membranes of metabolically active cells self-intersecting surfaces. With N points of Flip and Flop, a rudimentary Topological classification of the membrane as a manifold shows it is homeomorphic to an N-fold Torus, not two non-self intersecting concentric surfaces that are homeomorphic to spheres. This diffeomorphism proves categorically that the membrane lipid cannot form a 'Bilayer' structure when the cell is metabolically active (Figs. 1A and B). So what 'structure' does it have?

Fig. 1A : Cross-section of the cell membrane depicted as a 'Bilayer' - two non-intersecting closed surfaces, each homeomorphic to spheres. Hydrophilic lipid headgroups are in continuous contact with the inner cytoplasm and outer environment, and are free to translate laterally and to rotate. Fatty acid chains attached to them are shielded from the aqueous, ionic cell environment. 1B: With aminophospholipid 'Flip-Flop', lipid trajectories cause the cell membrane to self-intersect, falsifying the 'Bilayer' model, rendering the cell surface homeomorphic to an N-fold torus.

2. Basic Fluid Dynamics: features of phospholipid dynamics suggest they are actually undergoing forced convective and shear- driven organisation when the cell is metabolically active.

Figure 1C: Thin viscous fluid layer manifesting a counter-rotating roll pattern. Wavelength is given by the distance between two fluid upwellings, and is ~ twice the depth of the fluid layer (d). Shear induced Taylor vortice flow of viscous fluids also involves a pattern of counter-rotating rolls. In convective rolls, layers of fluid move at different velocities on nested cylinders within each roll, but move along rolls in advancing spirals when shear driven (Velarde and Normande, 1980: Koschmieder, 1974). Fluid flowing on the surface of each roll is therefore moving on a torus. If the membrane flows were spatially organised, fields would be generated by the motion of phospholipid headgroup dipoles over the surface of these tori. The pi electrons of unsaturated acyl chains attached to phospholipid headgroups would be subjected to the field inside each toroidal structures, shielded from the aqueous cytoplasm and external environment.

Figure 1D: In 'Natural' buoyancy-driven hexagonal convection , fluid flows up at the centre, and down at the perimeter of each convection cell and arises (for a limited range of finite amplitudes) at Rayleigh numbers well below the critical value predicted by Linear theory. Fluid flows on each hexagon move on the surface of one torus.

Figure 1E: In surface-tension driven 'Natural' convection, hexagonal flow pattern is somewhat different: each hexagonal convection cell has 6 'nodes' of the vertical velocity component which extend from the cell centre to each of the corners (as shown above: Mpegs of fluid flows taken from beneath a convecting fluid layer are available at http://members.lycos.co.uk/convection/Frontpage.html ). Fluid molecules thus move on one of six toruses per convection cell, with net flow upwards out of the cell at the centre of each hexagon (arrowed red at X,Y and Z), and down into the cell in the centre of each triangle. In a living cell, this fluid layer runs against the cytoskeletal protein network whose corrugated surfaces would mechanically 'force' the flow pattern from these idealised streamlines. As a consequence of this forcing, linear equations describing the onset of 'Natural' convection are not quantitatively predictive,and cannot be used to falsify my model. Since the transmembrane sections of proteins running through the membrane at X,Y and Z (Glycophorin C/D) and those in the centre of each edge (Anion Transport channel AE1 protein) are known to adopt helical structures, lipid flows would undoubtedly spiral at the centres and corners of each hexagon. Helical flows of electrically conductive fluids can generate a dynamo action, such that these proteins also experienced fields (see later).

This range of nanoscale patterns is identical to the range of patterns in which the membrane-associated proteins are assembled in the cortical networks of all plant and animal cells. Could the lateral and flip/flop motions of membrane phospholipids actually indicate a spatially organised, coherent lipid state exists in cell membranes? Furthermore, could it be induced by these mechanisms? As discussed at length in (Lofthouse 1998a-c), extant linear equations suggest an unfeasibly high temperature gradient would be needed to establish convective flows, but these describe only 'Natural' (ie unforced) convection, and cannot be quantitatively applied in a biological context for a host of reasons. They even fail to predict the empirical behaviour of nematics undergoing 'Natural' convection (Dubois-Violette, 1978)! The contention is qualitatively feasible, is consistent with all empirical data, and importantly, would give the membrane the correct Topology as a manifold. It would also explain why exogenous labelled phospholipids appear to diffuse in metabolically active cell membranes between 10-100 times more slowly than they do in an artificial lipid bilayers, and are constrained to move in patches of distinct sizes on the cell surface, a discrepancy that has puzzled cell biologists for three decades ( Fujiwara et al., 2002: video of the motions of headgroup- labelled lipid on cell surface at http://www.jcb.org/cgi/content/full/jcb.200202050/DC1). Since the headgroups are labelled, reporter lipids cannot enter the coherent state and be transported through the membrane, and are confined by the contours of the convection cells)

(NB: Whilst patterned flows of electrically conductive fluids (akin to Williams roll domains) can also be induced by electrical fields, and would also give the fluid surface a Topology compatible with 'FlipFlop' motions, empirical evidence indicates electric fields cannot be the primary generators of coherent aminophospholipid motions in cell membranes. If generated by charge differentials across a 'Bilayer' membrane, perturbing the charge across the membrane would undoubtedly have a discernible effect on phospholipid flip flop rate. Neither removing all the ferromagnetic haemoglobin from inside red blood cells, nor stripping over 90% of the surface negative charges from cell surfaces has any significant influence. I have demonstrated that red cell shape can be used to indicate membrane phospholipid flow pattern (Lofthouse, 1998a-c): the application of electric fields induces defined shape changes (Goncharenko and Katov, 1985), but each of these can be mimicked exactly by other perturbations that only affect convective flow organisation. Red cell shape recovery after removal of electric fields is temperature and ATP-dependent (Henszen et al., 1997) compatible with a convectively organised membrane. When small amounts of exogenous uncharged lipids (eg cholesterol) are added to normal red cell membranes , much higher electric field strengths are needed to induce the same shape changes : it appears therefore that the mechanisms driving phospholipid flows are primarily sensitive to the convective parameters of the membrane than the charge across it, and for this reason, I suggest the primary coherent phospholipid state is generated by heat. Electrohydrodynamic instabilities in lipid flows can be induced in my new membrane model, but do not generate them de novo (the role of induced electrohydrodynamic instabilities in the functioning of cells of innervated tissues (eg voluntary muscle, neurone) is discussed in Lofthouse, 1998j.

Phospholipid molecules have strong molecular similarities to the chain-like silicon oils and nematics that are used to study convective- and shear-driven organisation (Dubois-Violette et al., 1978). They comprise two fatty acid (acyl) chains (designated SN1 & SN2), which vary in length, and can be saturated or unsaturated: in phospholipids however, these nematics are attached to a negatively charged phosphate group and a positively charged amide moiety forming the headgroup (Fig. 2A).

Fig. 2A: Phospholipid Structure Green shaded area represents the phospholipid headgroup, which contains a circa 2 Debye dipole between the phosphate group and the phospholipid amide group X (where X = Phosphatidyl-ethanolamine (PE), -serine (PS) or -choline (PC)). Two acyl chains (blue shaded: designated SN1 & SN2) are attached, which can either be saturated or unsaturated, and which vary in length from 2-26 Carbons. Sites of cleavage by the enzymes Phospholipase A1,, A2,, C and D are arrowed: note that In Situ the first two lipases attack the acyl chain moieties of lipid molecules, altering the microviscosity of the membrane, whilst the latter two attack headgroups.

According to my model, loss of the coherent lipid state would lead to headgroups being held in the 'up' position at either the cytoplasm/membrane or environment/membrane interfaces. Molecules that bind to phospholipid headgroups, or to specific sites on acyl chains should thus act as probes to indicate whether phospholipid headgroups were in the down (coherent) or up (decoherent: spinning) state. For example, the enzyme Phospholipase A2, is known to bind to a specific site on the SN2 acyl chains of phospholipids. This site lies immediately under the headgroup (Fig. 2A), and would only be freely accessible on molecules that had left the coherent phase. Similarly the enzyme Protein Kinase C, binds to free headgroups of aminophospholipids Phosphatidylserine (PS) and Phosphatidylethanolamine (PE), and if membrane lipid were undergoing surface-tension driven convection, it should not be able to bind the membrane at either the intracellular or extracellular interface (Fig. 2B).

Fig. 2B: Proton transfer complex formation by phospholipid headgroups renders cleavage sites for enzymes inaccessible.

If phospholipids were leaving the coherent state, changes in PS/PE 'Flip-Flop' rate and/or a change in PS/PE headgroup orientation would arise, and would be evidenced by a change in the ability of probes PA2 and PKC to bind the membrane. If the coherent lipid state were, I I suggest, caused by forced Benard Marangoni convection, it follows that cellular analogues of factors that directly perturb surface-tension driven convection (changes in the temperature gradient across the layer, changes in the viscosity of the membrane layer, factors that perturb headgroup interactions), should cause more PA2 and PKC to bind to the membrane fraction of whole cells. Existing data supports this contention, since it indicates intact cells membranes are able to bind PKC and PA2 when cells are subjected to :

1. ATP depletion, changes in extra-cellular temperature (Cribier, 1993 ) This is analogous to altering the temperature gradient across a thin fluid layer. Coherent convective flows would induce a 'headgroups down' orientation, explaining why PS binding probes are not be able to bind the metabolically active cell surface.
2. Treatment with Anaesthetics : these bind PS headgroups as well as numerous proteins (Mondal et al, 2002) and would therefore perturb their ability to enter the coherent state. Anaesthetics inhibit the rate of aminophospholipid Flip-Flop (Badger, 1982: Basse, 1992) and increase the amount of PKC bound to the membrane ( Do, 2002: Gomez, 2002: Nirvathi, 1996;Simons, 1998:Kurokawa, 1992: Zhong et al, 2002). In some cases anaesthetics binds directly to PKC and prevents it binding aminphospholipid (Hemmings, 1998: Slater et al, 1997:
3. Loading cells with Ca 2+: (Seelig, 1993) this binds to proteins and to PS headgroups
4. Changing the gravity vector - this perturbs buoyancy-driven convection only, which I have suggested may be manifest by the cytoplasm. Since cytoplasm is mechanically coupled to the membrane, it can exert shear on it, and there is little doubt that shear driven instabilities in systems of two immiscible fluids (eg membrane and cytoplasm) are G-sensitive. Alterations to the Gravity vector could under my model indirectly induce perturbations in coherent membrane lipid flows even if they were predominantly surface tension driven. Significantly more Protein Kinase C is found bound to the membrane of biological cells subjected to a micro-gravitational environment(Hatton et al, 1993).

3. Could Convective and Shear-Driven Membrane Phospholipid Flows Perform Topological Quantum Computations?

The notion of topological quantum computing remains a mathematical theory, and is attributable to Michael Freedman and Kitaev. Though still a mathematical theory, this new approach holds the potential of significantly changing the landscape of quantum computation. The major challenge in stabilising a computer is to control errors. To prevent a crash for a classical computer, we only need to distinguish between ON and OFF. In topological quantum computation however, information is encoded in some topological property of a quantum system, and because these topological properties are global, it is harder for local errors to destroy the information. Freedman's computational model based on topological quantum field theories of anyons is polynomially equivalent to quantum computation. Kitaev proposed to control errors in quantum computing physically using anyons. Kitaev suggested that a quantum computer based on topological theory could be produced using an electron liquid, and a two-dimensional surface chilled to an extremely low temperature subjected to a magnetic field. When the electron liquid is under extreme cold conditions, composite particles called anyons form. These move on the two-dimensional surface, forming a braid. Vaughan Jones discovered a method of mapping a braid to a matrix, or a grid of numbers. Each matrix has an invariant property, a property that remains constant regardless of the shape or form the braids and knots take. The braids created by the trajectories of anyons share this property, and can be used to encode information and protect it from errors. The operation of a topological quantum computer would not depend on the exact details of a path, only on the way the paths (or braids) of the anyons are linked. Since the braids and their matrices are used to perform quantum computations, the result would be a robust and stable system. To have a usable computer would not demand matrices for all possible computations, only for a certain set of them from which other computational functions could be built. (Think of how we use the right combination of simple arithmetic, like adding, to do more complicated arithmetic, like multiplying or squaring numbers.). Although quantum Hall fluids are so far the only systems identified as containing anyons, many other anyonic systems await discovery. Freedman is exploring the possibility of creating non-abelian anyons in systems consisting of many thin layers of magnets. Kitaev, is investigating building memory for a quantum computer from a system in which electrons spinning on the corners of a hexagonal grid give rise to anyons. I suggest that spatially-organised flow topology of phospholipids under the new membrane model yields the behaviour needed to provide a suitable set of matrices 'Far from Equilibrium' to facilitate TQC by the metabolically active biological cells.

Simply speaking, flows of electrically conductive fluids generate magnetic fields. Dynamos are systems that can convert mechanical into electromagnetic energy: electricity is generated by dynamos in which conductive and ferromagnetic materials rotate relative to each other. Likewise, the convective motions of electrically conductive fluids also generates magnetic fields by a dynamo action (Childress and Soward, 1974: Soward, 1974: Feudel et al., 2003). Whilst in some 'Natural' convective flow patterns, gross external fields are zero, and the dynamo is therefore not measurable directly, it is accepted that where flows posess helicity, net non-zero fields are generated inside individual toroids (Ponomarenko, 1974). If we view the membrane as a thin layer of viscous fluid undergoing convective motion, lipid acyl chains and dipoles will follow trajectories on the surfaces of toruses when the cell is dissipating heat (Fig. 2C) . Headgroups effectively flow helically on a series of solenoids and will generate axial fields, experienced by the pi-electrons of unsaturated acyl chains held inside them. In 'Natural' surface-tension driven convection, a hexagonal pattern is composed of 6 toruses, and pattern wavelength is a function of distance from threshold. Within limits, wavelength increases with temperature gradient above a critical value. Altering the wavelength of a hexagonal pattern would also alter the paths of individual lipid molecules on each torus, and would also alter the field experienced by the pi-electrons inside. If the overall flow pattern changed eg from a hexagon to roll ('bifurcations'), lipid coherence would be maintained, but the Topology of flow would alter, hence the fields pi-electrons experienced would differ.

Fig. 2C: Flow Pattern Topology (Planiform) Determines the Fields Generated By Phospholipids in the Membrane. (i) Convective motions of conductive fluids generate magnetic fields via dynamo action Childless and Soward, 1972: Soward, 1974). Numerical modelling of 'Natural' convection (eg Matthews, 2001) indicates that only some patterns generate net non-zero fields external to the fluid: this implies that field strength running horizontally between the individual tori of a hexagonal pattern ( along X-X shown in blue) or vertically through them (along b-b= red) would be zero. This would almost certainly not apply in a cellular context, since here sufficient helicity to generate a dynamo action would be 'forced' by flowing over the helical surfaces of transmembrane proteins that run through it. (ii) In any case, flows of phospholipid headgroups on each torus resemble the helical flows of fluid on Ponomarenko cylinders (Ponomarenko, 1973) , and will therefore possess an internal non-zero field along their axis (green), which will be experienced by shielded pi-electrons on unsaturated phospholipid acyl chains when the cell is metabolically-active.

The new membrane model thus appears to confer on phospholipid molecules the behaviour needed to facilitateTQC, provided the cell remains metabolically active. The trajectories of individual lipid molecules (headgroups and acyl chains) are braids determined by flow pattern ('planiform') and wavelength. The dipolar headgroups thus move on tori, and produce magnetic fields the p-electrons of their unsaturated acyl chains are held inside, isolated from the environment. Changes in the cell or its environment that alter either the wavelength or flow pattern alter the velocity and trajectories of lipid molecules, and consequently the trajectories of their acyl chains and p-electrons, and the strength of the fields they are held within, but the coherent state is maintained The convective organisation of membrane lipids would thus place in the region of 20 million aminophospholipid headgroups in a red blood cell alone into a coherent state with a defined Topology, providing the cell remained metabolically active.

The model thus suggests phospholipids might well have a previously obscured function in biological cells. Currently, Cell Biology in general is 'protein focussed', and attribute the differences between types of cells, and evolutionary development exclusively to alterations in protein sequences. Examination of total cellular composition (membrane lipid components as well as proteins) reveals that whilst point mutations do arise in individual protein sequence, by far the greatest differences between even the same cell types in different species are in the content and distribution of the acyl chain moieties attached to their phospholipid: why such extensive phospholipid polymorphism exists is entirely unknown ("Lipid Polymorphism" is a vast research field in itself eg Dowhan, 1997). If the membrane served only as a 'barrier' between the interior of the cell and its environment, the function could be performed with one type of lipid species alone - there are over 120 different acyl chain combinations on the phosphatidylserine fraction of human red blood cell membranes alone! Interpreted against my model, acyl chains determine the kinematic viscosity of the membrane, and hence the thermal gradient at which they begin to undergo coherent patterned motions. If membrane phospholipids were functioning as a TQC system, the number and distribution of acyl chains attached to headgroups in an individual cell would also be a key factor in determining the range of 'computations' an individual cell could perform.

How could the cell 'output' the results of any TCQ performed in such a manner by its phospholipid flows ? Feedback provides an answer. Membranes do not exist in isolation: in every biological cell they are in continuous contact with a network of proteins that runs against them - their cortical cytoskeleton. The proteins of these nanoscale networks are well known to undergo continuous low affinity interactions with aminophospholipids in all cells (Rybicki et al., 1988; Klausner et al, 1984:Deme et al., 2000; Sikorski et al, 2000: Le Rumeur et al., 2003), and in every cell, are arranged in the same range of nanoscale patterns (hexagons, triangles, squares or parallel lines of constant spacing) as realisable convection- and shear-driven fluid flow patterns. Moreover, each cortical protein is 'acylated' post-translationally, and these lipidic fatty acid chains, once inserted into the membrane, would be subjected to fields generated by phospholipid flows. In the following section, I show how this continuous feedback onto the cells protein subsystem would enable spatially organised membrane flows to generate classical level output- in the form of current flow- via basic electromagnetic induction.

4. Cytoskeletal Protein Networks as Nanowires. Foerster energy transfer routes through proteins. How spatially organised phospholipid flows convert the membrane into a 'Dynamic Template' for the cytoskeleton, which shields specific electron flow routes into and out of the cell. Alterations to membrane flow patterns change protected current input/output routes, and agents that induce them.

Whilst membrane lipid composition shows enormous variations from cell to cell, the highly flexible cortical 'cytoskeletons' - geometric, nanoscale meshwork that runs against the membrane of all plant and animal cells - form a 'minimum inventory-maximum diversity' architectural system. All cytoskeletons contain analogues of the same protein components, assumed to have originated from a common ancestral gene. For example all cells contain a protein that is a member of the Spectrin Gene Superfamily of proteins, one that is a member of the Actin superfamily, a Tubulin analogue etc, which differ in amino acid sequence across species, and in different cell types. Most membrane-bound cytoskeletal proteins are 'acylated' after being produced ie they have acyl (fatty acid) chains of variable length added along their length post-translationally (eg Wolff et al, 2000: Harzsch et al., 2001: Chen & Manning, 2001; Maretzki et al, 2003; Westermann & Weber, 2003) and undergo low -affinity electrostatic interactions with phospholipids (Rybicki et al, 1988; Klausner et al, 1984: Le Rumeur et al.,2003: Deme et al, 2000). According to prevailing Cell Biology paradigms, the sole function of both phenomena is merely to ensure proteins are anchored to the membrane.

In all cells, membrane associated cytoskeletal proteins form networks in a limited range of nanoscale patterns - triangles, hexagons, parallel lines of constant spacing or squares - the same range of patterns that form spontaneously in thin viscous fluids undergoing 'Natural' convective and shear-driven organisation. For example, Fig. 3A shows the arrangement of the ubiquitously expressed protein Spectrin in normal human red blood cells.

Fig. 3. The membrane associated cytoskeleton of the normal human red blood cell. Electron micrograph of the normal red blood cell cytoskeleton - a predominantly nanoscale hexagonal network running against the inner cell membrane. A multi-subunit protein called 'Spectrin' forms the edges of these networks, and is annealed together by other proteins into a continuous network. Some of these other proteins 9JPC and AE1A) run trans-membrane at the mid-point of edges and the vertices, In some abnormally-shaped 'Elliptocytic' red cells with altered membrane phospholipid compositions, Spectrin forms as a rectangular lattice. A fully triangulated Spectrin network also lays underneath the synaptic end of neurones.

Fig. 3B: In voluntary muscle cells, several proteins are related to Spectrin, alpha/beta Spectrin is distributed against the membrane in layers of parallel lines of constant spacing (Porter et al., 1992), whilst another related protein (Dystrophin: shown here by red fluorescent stain) forms an untriangulated hexagonal lattice over the M- and Z-lines of an individual cell. In some abnormal ('Muscular Dystrophy') cell lines whose membranes have abnormal phospholipid compositions, these network patterns are disrupted (Williams et al., 1999: Sato et al., 1992).

In previous papers, I showed qualitatively and quantitatively why Reaction Diffusion mechanisms cannot be the mechanism responsible for establishing these cytoskeletal patterns, and demonstrated that these networks can be assembled against convectively-organised membrane surfaces in a manner that allows form to emerge spontaneously from the growth process (Lofthouse 2004 d, g & h). This mechanism was suggested to operate by virtue of the constant low affinity interaction between phospholipids and protein molecules, which ensured that proteins arriving at the membrane surface would be physically dragged towards and into lipid flow sinks, enabling them to use the membrane surface as a 'Dynamic Template'. In a cell where the membrane manifest a roll flow pattern, for example, cytoskeletal proteins would be dragged into a hexagonal pattern, were it manifesting a roll pattern, they would be dragged into parallel lines with a spacing approximately equal to twice membrane depth (Fig. 4A).

Fig. 4A. The 'Dynamic Template' Membrane. Low-affinity interactions between any cytoskeletal protein and phospholipids ensure that the spatial arrangement of any protein arriving at the membrane surface would be dictated by their flow pattern. Flow pattern in thin layers of similar viscous fluids to lipid acyl chains (eg nematics) cannot be predicted by any extant theory, but are a complex function of chain length and various parameters including the temperature gradient across the layer. By far the greatest difference between any two cell types is the acyl chain composition of the membrane lipid and metabolic rate: different cells could thus manifest differing patterns, the symmetry of which could be continuously broken across the entire developmental trajectory. The Dynamic Template mechanism allows cell form to emerge spontaneously from the growth process itself. This is the first biological pattern forming mechanism in which the symmetry breaking changes necessary can be realistically generated during assembly, and its ability to construct realistic erythrocyte, muscle- and plant cell cytoskeletons is demonstrated elsewhere (Lofthouse 1998 d,, 1998 g and 1998h).

Fig. 4B: Microtubules as Conduit for current flow. (i) In all biological cells, alternating subunits of alpha (white) and beta (shaded) Tubulin protein form linear protofilaments (ii), varying numbers of which align laterally to form 'Microtubules' (iii). Each dimer is acylated, and undergoes low affinity interactions with phospholipids. When presented to a 'Dynamic Template' membrane therefore (iv), one protofilament will be dragged into a flow sink at the inner membrane surface, and depending on the trajectories of phospholipid headgroups past them, both it and the acyl chains attached to it experience magnetic fields produced by their motions. The pi electrons on unsaturated acyl chains form a 'sea', and are subjected to non-zero magnetic fields inside each toroid. Topological quantum computation results could therefore be output by induction into current flows through cytoskeletal proteins without Decoherence of the phospholipid state. In vivo, microtubules are associated with transmembrane proteins, which are also shielded from the aqueous environment by up- and downwelling lipid flows over their helical surfaces, and can therefore conduct current into and out of adjacent cells.

This suggestion is compatible with feasible electron routes already been identified in assembled microtubules (Porter, 2000). The amino acid sequences of Tubulin monomers from different cells are known in many cases, and those most likely to participate in electron transport (aromatic residues with electron resonant ring structures eg Tryptophan or Histidine : Porter et al., 2000) are distributed in patterns with remarkable complementarity to the patterns of phospholipid flows suggested by my model Whilst the 'FÖrster distance' in an ionic solution at constant temperature is circa 1 nm, electron (or proton) transfer through particular amino acids has been demonstrated In Vitro over a distance of ~2 nm in some proteins. By building models of microtubules , it is apparent that in the A lattice formation (ie microtubules containing 13 protofilaments), suitably spaced amino acids with electron resonant ring structures are distributed in distinct winding patterns - Tryptophan residues laying close enough for transfer forming distinct vertical pathways along individual protofilaments in microtubules, whilst Histidine residues with a compatible spacing are distributed in either 3-start or 5-start spiralling paths along its length (Fig. 4C).

Fig. 4C: Electron transfer routes through assembled microtubules (after Tuszynski, Porter et al.,). (i) Tryptophan residues laying within the minimum distance for Foerster energy transfer form continuous routes along each individual protofilament (shaded) , whilst (ii) Histidine residues laying within this distance form two helical paths of different pitch Small differences in amino acid sequences in the Tubulin subunits of different cell types could increase or decrease these distances, generating variation in conductive routes. Deamidation of amino acids induces kinks in the polypeptide backbone, which would also affect the distance between these residues as the cell senesces, which would enable conduction routes to be turned 'on' or 'off' at specific times in the individual developmental trajectory of a cell.

With either convective or shear-driven roll patterns in thin layers of viscous fluids, the depth of the fluid crudely dictates the wavelength of a roll pattern flow: the distance between two upwelling flows is equal to the width of two rolls, and is circa twice the depth of the membrane (d)(Koschmieder, 1973: 1994). Whilst the depth of biological membranes is invariably quoted as 10 nm, recorded literature measurements actually vary from 5 to 18 nm: in good agreement with the observed spacing of Microtubule Associated Proteins (MAPs: range of 10 to 36 nm). MAP proteins connect microtubules to proteins that run through the membrane, connecting it to the external environment (ie connecting it to the shielded transmembrane proteins of the abutting cell if packed into a tissue. If laying against a convectively-organised membrane in which phospholipids flowed in a roll pattern, therefore, at least one vertical Tryptophan or Histidine pathway through the microtubule would lie between counter-rotating rolls of PS/PE, shielded from the aqueous ionic environment by the magnetic field, connected to transmembrane proteins that were themselves shielded from their environment. As shown in Fig. 4D, a re-orientation of phospholipid rolls relative to the underlying cytoskeleton could thus re-route a current flow from a Tryptophan route to a Histidine route in an individual protofilament and thence out of the cell via a distinct set of transmembrane proteins. Changes in the Topology of the coherent lipid state could thus be transduced into classical level current flow through a distinct set of proteins without its decoherence.

Fig. 4D: Electromagnetic induction of current through shielded protein conduit effected by a spatially organised membrane flows. If laying against a 'dynamic Template' membrane, conductive Tryptophan routes through one protofilament of a microtubule would be held between two rolls: a change in the relative orientation of these would protect alternative routes through Histidine residues which appear to lie within the Foerster distance.

Similar electron routes appear to exist in other membrane-associated cytoskeletal proteins. All members of the Superfamily of Spectrin proteins (including Dystrophin, Utrophin and alpha-Actinin in homo sapiens) for example, contain a high sequence homology 'Spectrin repeat' sequence, giving them a molecular structure resembling a string of sausages, separated by flexible linker regions (McGough, 1999: Fig. 4E)

Fig. 4E: (i) Detail of the multiple helical barrel structure of red blood cell protein Spectrin which is known to contain many Tryptophan (blue) and hydrophobic residues (particularly Histidines) that are highly conserved across species (Leto et al., 1988). When lying against the membrane of a metabolically active cell, the metric length of the tetramer (which constitutes one edge of a continuous, triangulated network) is known to be contracted to 1/2.6th is maximum length in vitro, which pushes the subunits together. My feedback mechanism is the first to provide a viable mechanism for effecting this contraction in situ (Lofthouse, 1998c). As shown in (ii), crumpling Spectrin subunits together would produce a continuous electron route, linking transmembrane protein complexes. Electron routes through the membrane bound cytoskeleton could therefore be developmentally regulated by factors that crumple subunits togther: alterations to convective membrane flow wavelength being one of them, and the cellular 'computations' performed by a 'hard-wired' cytoskeleton would be adaptable (see Fig. 4F).

As Spectrin molecules are also acylated post-translationally (ie have fatty acid chains added along their length), and undergo low affinity interactions with aminophospholipids,, they would also be dragged into flow sinks when laid against a membrane. Whilst the length of the helical repeats varies between 99 - 110 residues in different tissues, and the degree of sequence identity between repeats is low, sequence alignment studies reveal these contain highly conserved Tryptophan residues across species: the reason is obscure. It is possible that when the subunits are crumpled together, these and other residues lie within the maximum distance for Foerster energy transfer, and also form protected routes for current flows between transmembrane protein elements, and that convectively organised membrane flows permit them to function as conduit.

Uniquely among models of cellular computation, my feedback mechanism would enable cell signalling routes to be turned on or off by alterations to cellular fluid parameters alone. For example, during red cell development, cell radius is observed to shrink, which necessitates a contraction in the metric length of Spectrin molecules by a factor of 2.6. I demonstrated how this could be achieved by shrinkage in the wavelength of a hexagonal phospholipid flow pattern through the membrane above the cytoskeleton, and showed this was compatible with an identified decline in the metabolic rate of the developing cell, and/or changes in phospholipid composition. Such shrinkage could also ‘crumple’ the subunits of each Spectrin edge together, effectively joining up linked Tryptophan routes, until at a critical point, they became active as shielded conduit, and signalling routes through them that were discontinuous in immature cells became activated in the mature, circulating red cell. Shrinkage beyond a certain point would cause the edges to ‘buckle’, electron routes would no longer be adequately shielded or continuous, and conduction routes would be lost. This would explain the loss of sensitivity to certain signalling molecules during development, and during what is termed ‘programmed cell death’ or Apoptosis (Lofthouse, 1998a, 1998b & 1998c).

Fig. 4F: Signalling routes through a Spectrin cytoskeleton could be turned off (i) or on (ii) by altering the distance between individual subunits in the cytoskeletal network edges, and hence adjusting the distance between participating amino acid residues. The wavelength of a stable hexagonal convection cell is a function of distance from threshold, and within limits expands on increasing the temperature gradient across the fluid layer. In (i), subunits are separated by flexible linker regions, and Trp electron routes are discontinuous. A decrease in membrane flow wavelength would push subunits together (Fig. 4E), producing a continuous shielded electron routes between the transmembrane complexes that Spectrins link together. As shown by Cribier et al, ( 1993), in the case of the human red cell, metabolic rate declines during maturation, concomitant with a decrease in the thermal gradient across the cell. Compatible with my mechanism, the contour length of network edges is known to decrease by a factor of circa 2.6 during the final stages of maturation(Lofthouse, 1998a and 1998b). Routes through Spectrin via Tryptophan pathways could thus be turned on in mature cells that are off in the maturing cell, which might explain how cell responsivity to certain molecular signals (eg hormones), changes during Stem Cell development.

5. Changing inputs/outputs by altering membrane flow patterns and wavelengths alone

A problem with all previous processing conjectures based on protein cytoskeletons has been that the cytoskeleton is ‘hard-wired’: by implication this means that if it were to act as conduit, cell inputs and outputs would be fixed. By definition, a 'computation' demands that an input signal alters cellular outputs, and individual cells must respond differently to many signals. In sharp contradistinction from its’ predecessors, my model enables labile response. With a membrane through which phospholipids flowed in a stable pattern with a stable wavelength, a distinct sub-set of network protein routes lie shielded from the environment, consequently a given sub-set of transmembrane proteins co-incide spatially with upwelling and downwellings as input/output from the cell. However, alterations to the classical level routes into and out of the cell can be effected merely by altering membrane lipid flow pattern

My cellular computation scheme would enable inputs/outputs and signalling pathways to be without altering the layout of the cytoskeletal lattice, by subtle alteration to the pattern of membrane flows against which it rests. Phospholipid acyl chains would in each case be moving on a different set of toruses through the membrane above the cytoskeleton (ie performing a different TQC), generating current and input/output through different transmembrane elements. As shown in Figs. 5A - 5F, only some flow pattern changes alter the shape or size of the cell by performing affine deformations of the cytoskeletal lattice, the rest would be indiscernible visually.

The fully mature human erythrocyte has a predominantly hexagonal cytoskeleton network whose edges comprise four strands of the multi-subunit protein Spectrin. This network is ligated to two types of transmembrane protein complexes - Junction Point Complexes (JPCs) that form the vertices of the network, and Ankyrin/AE1 proteins (A/AE1) that join two Spectrin dimers together in an anti-parallel arrangement to form each network edge (Liu et al., 1988). (Fig. 5A). We assume here these complexes have identical protein compositions.

Fig. 5A: Schematic cytoskeletal network of the human red blood cell (i) Network vertices (A - X) are composed of transmembrane (Junction Point Complex) proteins Glycophorin C & D, ligated to intracellular proteins including ERM protein 4.1, a short Actin protofilament, tumour suppressor protein p55. Network edges comprise two dimers of highly flexible Spectrin protein. One end of each dimer is ligated to Ankyrin protein, which ligates two dimers to the transmembrane protein Band 3 (AE1/A: numbered 1-52) at the midpoint of each edge. The other end of each Spectrin dimer is indirectly bound to the transmembrane JPC. When a cell abuts another (in a tissue), their membranes are coupled, and join to the transmembrane components of the next cell in a continuous, shielded route. NB: From the known relative abundances of protein components in the red cell, it is evident that JPCs vary in the number of copies of Glycophorin C/D present: theoretically, JPCs can have 0-6 copies of Glycophorin C, or 0-6 copies of D, or combinations of C/D in various rotational positions. Hexagonal JPCs can also vary in their content of tumour suppressor protein p55, which can theoretically be present in 0-6 copies. The possibility that protein content determines gate functions has been considered elsewhere (Lofthouse, 1998).,/P> First consider a small section of this hexagonal network lying against a membrane manifesting hexagonal surface tension driven convection. All network edges are dragged into the six upwelling flow nodes that run from the centre of each hexagonal cell to the perimeter. Some transmembrane protein complexes co-incide in space with phospholipids flowing upwards at the centre of convection cells, whilst others will be co-incident with downwelling flows at the perimeter of a convecting unit All will therefore be pinned spatially and have phospholipid headgroup dipoles moving in helical trajectories across their surfaces. A specific sub-set of AE1/Ankyrin and JPC proteins can thus act as shielded ‘inputs’ and ‘outputs’ from the cell (shown connected in red in Fig. 5B), and would also lie in the magnetic fields generated by dynamo action of the membrane

Fig. 5B: Shielded electron routes through the cortical cytoskeleton if laying under a membrane manifesting a hexagonal flow planiform. Spectrin connecting the JPC vertices A and B will lie in one of the six ‘nodes’ of the hexagonal cell, and will therefore be pinned in space and shielded from the aqueous cytoplasm beneath it. Spectrin connecting vertices A and E is also shielded. Note that all the AE1/A complex lie in a shielded upwelling of phospholipid with this pattern

If the same Spectrin lattice laid against a membrane manifesting a roll pattern, a different sub-set of edges coincide with flow sources/sinks, and these would elongate to a metric length determined by the length of the flow sink, parallel members separated by a distance approximately twice the depth of the membrane (Fig. 5C).

Fig. 5C: Shielded electron routes through the cortical cytoskeleton if phospholipid flows through the membrane manifest a counter-rotating roll pattern. Note that only a sub-set of network edges (red) lie in an upwelling flow, shielded from the cytoplasm. Transmembrane proteins running through the membrane and connecting to vertices A and B are also shielded, whilst signalling routes between A and E are not. Some AE1/A complexes lie between two upwelling flows, others (eg 5 & 6) between two downwellings. A change in the flow pattern in the membrane from hexagons to rolls thus changes the cells current inputs and outputs, performing a ‘computation’.

Phospholipid flow pattern thus determines which sub-set of transmembrane proteins lie in flow up and downwellings - the cells inputs and outputs, and a simple bifurcation in flow pattern (ie a change in aminophospholipid Topology from hexagons to rolls) or a change in hexagonal flow wavelength can alters the cells’ inputs and outputs. Thus a topological quantum computation by the cells membrane lipids can be transduced into an electrical signal through the cytoskeleton into or out of the cell.

What factors determine and alter the pattern which a thin layer of viscous fluid manifests convectively organised or shear-driven flows? Even for ‘Natural’ (unforced) convection, no available theory allows prediction of the pattern a given fluid will select (Koschmieder, 1993), though empirical observation indicates this is a complex function of ‘distance from threshold’ (ie distance from onset) and fluid composition. Where convection is mechanically forced however (for example, by introducing grids of corrugated surfaces into the fluids) pattern and wavelength are determined to the greatest extent by the pattern and surface contours of bodies they run against. Patterns can also be thermally forced by spatially organising the sites of heat injection into thin fluid layers. In a biological cell therefore, since the membrane of the fully assembled cell runs against a corrugated grid of proteins, cells inputs and outputs could be altered by:

1. Mechanical Forcing via protein conformational changes: factors that altered the surface topography of proteins (eg deamidation, addition of phosphate groups, binding of Ca 2+ ions) and/or hydrolysis of ATP/GTP at specific sites on the protein lattice (eg at network vertices) could induce subtle changes in membrane flow patterns above them, altering protected routes, hence altering cell inputs and outputs. A
2. Thermal forcing by Rac/Rho ATP/GTPases - current models assert that the entire intracellular signalling apparatus operates on a protein Meccano structure, inducing only protein conformational changes: signal molecule A binding substrate B, hydrolysing ATP or GTP, the energy released being used to change conformation, such that a protein ‘lock’ fits or does not fit a ‘key’. My model suggests that these signalling proteins do not operate not exclusively in this fashion, but contribute to cellular dynamics by determining the sites of heat injection into the membrane fluid, and effectively determining which pattern the phospholipid flows are ‘forced’ into. This new membrane model is therefore entirely compatible with extant ‘Molecular computation’ models, but provides a mechanism that enables them to alter cell computations in a non-binary manner. Also, as G -proteins are acylated (Harzsch et al., 2001: Chen & Manning, 2001; Maretzki et al, 2003; Westermann & Weber, 2003), they and their acyl chain moieties would also be subjected to magnetic fields generated by a dynamo action.
3. Anaesthetics: These could alter the coherent lipid state either directly by binding phospholipid headgroups or indirectly by binding regulatory molecules like PKC. This clearly has implications for models of Consciousness (vide supra and Appendix I).
4. 'Morphogens:' are a range of substances that alter plant cell shape, eg Auxins and Colchicine. Under the new model, these alter cell shape not by binding to proteins and changing thei r conformations, but primarily by exerting effects on membrane phospholipid flow patterns. Their binding to lipid molecules, and effects on lipid phase transitions are often marginalised, and in plant cells, there is a substantial body of evidence that falsifies the notion that it acts exclusively by binding and changing the conformation of protein molecules (detailed in 53). Colchicine changes the shape of human red blood cells (which contain no microtubules whatsoever),
5. Changes in velocity and direction of cytoplasmic streaming: These can also induce changes in the cells inputs and outputs indirectly, because the membrane and cytoplasm are mechanically coupled immiscible viscous fluids. Streaming cytoplasm can thus exert shear on the inner membrane surface, and induce flow bifurcations. These are known to occur for example, in plant cells that are treated with several morphogens, and in a range of cells on exposure to a microgravitational environment.
6. Exposure to microgravitational environment I have presented evidence that suggests that cytoplasmic streaming is buoyancy driven with mechanical and thermal forcing: since this is a function of the gravity vector, my new model provides a mechanism that enables the cell is able to ‘compute’ gravitational information, and has been shown to predict the shape change of erythrocytes in microgravity (Lofthouse, 1998 a - f), and to provide a feasible mechanism for gravitropic plant cell growth (Lofthouse, 1998d).
7. Deamidation This ubiquitously observed post-translational amino acid modification induces 180 o kinks in the backbone of proteins and would undoubtedly induce deviations in the trajectories of fluid particles moving across their surfaces. In general, 'wobble' can accumulate in viscous fluid flows so that beyond a certain point, its flow pattern changes completely. Deamidation involves the loss of amide groups from Glutamine and Asparagine amino acids, which converts them into Glutamic and Aspartic Acid respectively. Since these are negatively charged at physiological pH, deviations in the trajectories of negatively charged PS headgroups as they flowed across the protein surface would also be induced by electrostatic repulsion. The rate at which a particular Glutamine or Asparagine residue deamidates is determined by the sequence of a protein - specifically by the nature of the amino acid on its' C-terminal neighbour, and hence ultimately DNA sequences. As shown diagramatically in Fig. 6 , if we assert that identical network edge proteins contained one Asparagine residue that deamidated at approximately the same time, a pattern of hexagonal PS/PE convection cells could be ‘forced’ into a square cell pattern, and cellular ‘inputs’ and ‘outputs’ for current flow would be altered at a certain time in its’ Developmental trajectory. Slight alterations in protein sequence would alter the timing of this event between two cells. Whilst the prevailing metaphor for deamidation is ‘protein fatigue’, I have identified it as a form of PT (and hence CPT) violation - this is discussed in depth in Lofthouse, 1998 f.
8. Protein Phosphorylation The addition of a bulky, negatively charged phosphate group to a protein by kinase enzymes that bind to specific amino acid residues (Serine and Threonine in particular) both changes the target proteins conformation and adds negative charges: to it both these factors would deflect stable PS/PE trajectories through the membrane above the cytoskeletal network, could induce 'wobble' in them, and could therefore alter the topology of the coherent lipid state. Phosphorylation by some enzymes involves the hydrolysis of ATP or GTP , which would cause localised 'heat'injection into the membrane in a specific pattern, which could also force a change in flow pattern (Topology).
9. Electric fields are also known to induce shape changes: in many cell types (Jaffe & Nuccitelli, 1977: Harris et al, 1990: White et al, 1990) which are accompanied by alterations in the alignment of their cytoskeletal proteins. Whilst these realignments have previously been interpreted as due solely to the effect of the applied field on protein dipoles, I have demonstrated (Lofthouse 1998 a - d) that the primary cause is an induced bifurcation in phospholipid flow patterns through the membrane which drag the proteins into new arrangements. In the case of human red blood cells, my Dynamic Template mechanism has been shown to predict precisely the shape changes observed in both normal and abnormally shaped cells.
10. Alterations to membrane lipid composition Roughly speaking, the weight percentage of phospholipids in the membranes is broadly similar in the same cells of different mammalian species, ie they have about the same percentage of phosphatidylserine, phosphatidylethanolamine, phosphatidylcholine etc, but their acyl chain moieties show characteristic variations in length, degree of saturation/unsaturation, even in the same species. Whilst this has long been known to have a significant impact on the microviscosity of the membrane, its functional significance has been unclear. This new model suggests that viscosity alterations (eg the peroxidative damage that has a particular effect on unsaturated acyl chains) affect membrane lipid flow patterns, and thus alter inputs and outputs in response to different signals, and thus plays a key role in the ‘computations’ a particular cell performs in response to a given signal. Empirically, the actual pattern which a given thin layer of Silicon oil adopts when heated from below is a complex function of several parameters, two of which is the distribution of chain lengths/impurities present and distance from threshold (1): the biological analogues of these factors would be the length/degree of saturation of the acyl chains attached to phospholipids,and the gross metabolic rate. This computational mechanism therefore offers an explanation of the long observed phenomenon of ‘Homeoviscous Adaptation’ (HA), in which feedback loops of both plant and animal cells attempt to maintain the effective viscosity of the membrane in response to shifts in environmental temperature, largely by making alterations to the lipid acyl chains (excellent review in Cossins, 1994). My model suggests that HA renders cells capable of altering cellular inputs/outputs without altering their shape, and offers a viable explanation of dormancy/changes in cell function that arise eg during mammalian hibernation - a decrease in environmental temperature cues an alteration in membrane composition and therefore flow patterns, in such a way that certain signalling routes through the cytoskeleton were effectively turned off . Responses to oxidative assault (eg by light) also affects predominantly the acyl chains of membrane lipids. Anaesthetics too operate in this manner.

Example: how red blood cells ‘compute’ a mechanical shear field

The application of mechanical shear to any thin viscous fluid layer undergoing stable hexagonal convection is known to induce a distinct change (Graham, 1933). At low shear rates, hexagonal fluid flows first change into rolls running transverse to the direction of application, then as shear rate is increased, into a square cell intermediate, and eventually into rolls that are parallel to applied shear rate (Fig. 6A: 48) .

Fig. 6A: The effect of mechanical shear (applied horizontally left to right) on c onvective planiform (see text for details: from Graham, 1933).

Because under my model, the metric length of spectrin (inter-vertice network distance) is determined by the length of the membrane flow sink it lays between, if a red cell membrane were manifesting a hexagonal flow pattern my model predicts that the application of shear should induce a bifurcation into rolls, and the cytoskeleton should undergo a defined deformation. As detailed in (Lofthouse, 1998b), when red blood cells are subjected to shear in a Couette Taylor apparatus, they undergo a Symmetry-breaking shape change from biconcave disc to a prolate ellipsoid whose dimensions can be precisely predicted by my model (49). (Figs. 6B, C and D).

Fig. 6B: The deformation of red cells in shear fields can be predicted precisely by assuming a pitchfork ‘hexagon to roll’ bifurcation takes place in membrane phospholipid flows (Lofthouse, 1998a and b). Because of the subunit structure of the Spectrin edges of the cytoskeletal network, the distance between linked vertices is determined by the length of the membrane phospholipid flow sink in the membrane above them. With a hexagonal pattern, all network edges are therefore equal in length, but an affine deformation of the cytoskeleton can be effected by a bifurcation into a roll pattern, which induces a change in cell shape from a biconcave disc into an ellipsoid.

Fig. 6C and D: Effect of mechanical shear flows on red blood cell shape

On comparing Fig. 6B with Figs 5B and C it is evident that a membrane flow bifurcation indicates deformation in a shear field is accompanied by a change in the cells shielded protein inputs and outputs. Since the rate of shear determines membrane flow pattern, it will also determine which transmembrane proteins and network edges are shielded. A dynamic fluid system is therefore used by the cell to ‘compute’ the strength of the shear field it is experiencing. The same shape change can also be induced by the application of electric fields of appropriate frequencies: I have suggested that the fine details of this indicates that the fields induce bifurcations in membrane flows - detailed in.Lofthouse, 1998 a -d)

8. Error Correction: active stabilisation of the coherent phospholipid state using thermal and mechanical forcing by Protein Kinase C.

Consider the events that arise when red cells are loaded with Ca 2+ using an ionophore. Ca 2+ ions bind PS headgroups and shifts the 'gel-sol' phase transition temperature of membranes. They also bind Spectrin protein, which forms the edges of the cytoskeletal network (42). Increased intracellular Ca 2+ levels are known to increase in the amount of PS that can be cleaved by PA2 at the exterior surface of the cell (43), giving a strong indication that binding disrupts dipole-dipole interactions between headgroups, which would perturb stable PS/PE convection. Once the coherent state is lost, PS headgroups would stick up from the plane of the intracellular membrane, and PKC could bind to them. Thus activated, PKC would attach phosphate groups (- charges) to proteins whilst simultaneously hydrolysing ATP specifically at the vertices of the lattice under the membrane. This localised heating, at the vertices of the cytoskeletal network, where fluid is flowing upwards anyway, would inject sufficient energy back into the lipids to effectively ‘force’ their convective motions back into previously stable trajectories. Phospholipid molecules would dissociate from PKC as they re-enter associations with other lipid headgroups, and de-activated, PKC would be released back into the cytoplasm. At the same time, the increased intracellular calcium activates an ATP-dependent Ca 2+ -pump, which itself hydrolyses ATP as it expels cations, and adds to the generally increased rate of internal heat dissipation. The entire coherent phospholipid state could therefore be actively stabilised providing the cell remained metabolically active. Inhibition of PKC activity should significantly alter either flip flop rate and the amount of PS cleavable by lipases at the extra and intracellular surfaces: this is reportedly the case (Julien et al., 1997).

The model presented in this paper demonstrates that lipids play a role in cellular computation that is far more subtle and central than current models indicate.. Many cell signals are derived from the inositol phospholipids, but under prevailing paradigms, they are seen to act merely as 'molecular keys' that alter the conformation of the proteins they bind to For example, the hormone-stimulated hydrolysis of inositol phospholipids is known to produce the Ca2+ mobilising second messenger, inositol trisphosphate. This reaction generates diacylglycerol, which activates protein kinase C when added to cells. Phospholipase D catalyzes the hydrolysis of phosphatidylcholine to form phosphatidic acid and released choline headgroup. The phosphatidic acid may itself act as a signal molecule (e.g., by activating a PA-activated kinase), or can be hydrolyzed to form diacylglycerol by the enzyme PA phosphohydrolase. . My model suggests they function in a much more sophisticated manner, inositol lipid metabolism altering membrane fluidity (hence viscosity), and (from a Topological perspective) altering the flow trajectories of aminophospholipid molecules through the membrane, and therefore the TQC performed, and helping to determine inputs/outputs from the cell.

Implications for Consciousness models: anaesthetics known effects on the activity of PKC suggest they perturb coherent phospholipid states altering a cells inputs, outputs and shielded routes:

All types of anaesthetics are known to interact directly with phospholipids and alter the structures they form In Vitro (Smith et al, 1991: Cullis and Verklai, 1983, Hornby and Cullis, 1981) In intact cells, they are all known to perturb aminophospholipid ‘Flip-Flop’ rates, and either bind PKC directly , or cause it to become active when added to intact cells (Do et al., 2002; Slater et al., 1997; Kurokawa et al., 2002; Simons eta l., 1998; Zhong et al, 2002; Mondal et al., 2002; Basse, 1992, Badger et al., 1982; Zager et al., 1999; Nirvathi et al., 1996; Hemmings , 1998; Gomez et al., 2002). Interpreted against my model, these data imply that conciousness is linked directly to the coherent phospholipid state, and that anaesthetics medicate their action either directly by binding phospholipid headgroups and interfering with their ability to remain in the coherent membrane state, or indirectly , by inhibiting the ability of PKC to correct perturbations and pump the membrane back into its stable coherent state. (NB: the reagent Colchicine also binds phospholipid headgroups). Interestingly, lipid-soluble psychotropic drugs (eg Neuroleptics) also bind PKC (Kumar et al, 1997), indicating they exert their action on cell signalling either directly by perturbing the coherent lipid state, or indirectly by preventing PKC binding to membranes, incapacitating the regulatory mechanism that maintains it and performs error correction on it. This suggestion has important implications for models of consciousness (discussed in Appendix I).

SUMMARY

My model represents a radical departure from previous models of biological computation, , particularly those of Frolich, since:- a) the fundamental coherent state is held by cell lipids, not proteins, b) their coherent state is generated by 'waste heat' (not destroyed by it), c) the coherent state is not generated by electric/magnetic field gradients, instead, these fields are themselves generated and varied by the primary convective or shear driven mechanism d) the cell system utilises TQC and QEC e) classical level output is achieved by induction, and involves only a change in the Topology of the coherent lipid state, not decoherence The model offers a rationale for the enormous range and polymorphism of lipids that are known to exist in biological membranes (Gunstone, 1994). Taking into account variations in headgroup types, differing types of linkages between them and alterations in the length and degree of saturation of the fatty acyl chains to which they are attached, several thousand have been identified across species. In the human erythrocyte alone, in the region of 500 distinct molecular species of have been identified. Note that Nature only appears to require 4 nucleic acids and 24 (5) amino acid units. The reason for this level of phospholipid diversity is obscure: if membranes functioned merely to provide a barrier between the intra- and extracellular milieu as cell models assume, this could be achieved using only a single unsaturated species. It is firmly established that the headgroups of certain lipids are substrates for particular enzymes in vivo, but no function for the diversity of acyl chain moieties within particular sub-classes of lipid species has previously been suggested. My Dynamic Template model suggests that acyl chain composition, through its’ effect on the kinematic parameters of the membrane as a fluid, determines its flow dynamics, and hence the pattern and Topology of the coherent state that can be generated by each cell.

Viable technological quantum computers, appear theoretically more likely to resemble analogue computers. These are electrical circuits which received input by varying voltages, and used the physical properties of circuit elements like LCR circuits and operational amplifiers, to integrate and differentiate continuous signals and solve differential equations. Analogue systems then work with continuous signals that are processed according to the physical layout of a machine, rather than the digital system manner, which orders a series of ones and zeros according to the dictates of a logic-based program. This gives analogue machines an advantage in terms of size, efficiency and raw computing power. The disadvantage is that analogue systems lack flexibility- when armed with the right software (and assuming its processor is powerful enough) a digital computer can solve any problem that can be expressed as a logical algorithm. But, because analogue computers have their problem-solving abilities built directly into their electronics, if you want to perform a different sort of calculation, you have to design and build a new circuit to do it. At a classical level of interpretation, my new membrane model provides a means of using cellular proteins as conduit at the classical level by providing viable shielding mechanism, and enabling the cells inputs and outputs to be altered during lifespan and in response to shifts in environment without altering the ‘hard wired ‘ cytoskeleton. The feedback mechanism demonstrates that the pattern of cytoskeletal proteins determines the cells circuit, but the planiform of lipid structures which it lies against determine which routes through it are shielded as inputs and outputs. This allows different cells to ‘compute’ different outputs for the same signal, explains why blood cells responds to signals that skin cells do not, and begins to explain why certain areas of the brain are associated with specific functions- because cells in each region have a distinct cytoskeletal architecture, and distinct lipid membrane compositions.

Endnote: DNA has been shown to be conductive In Vitro at constant temperature, but the ‘decoherence problem’ has been responsible for skepticism about a host of quantum conjectures posited for DNA transcription. I point out that DNA , like cytoskeletal proteins, also undergoes low affinity interactions with phospholipids. Although there is little empirical evidence to date that nuclear membrane phospholipids undergo ‘Flip-Flop’ (simply because no-one has addressed the question in a laboratory), the feedback mechanism presented in this paper would also be capable of generating current through nucleic acids during transcription whilst simultaneously shielding them from their aqueous environment. Granted, DNA can be transcripted in vitro without a membrane being present, but virtually nothing is known about how transcription is timed or co-ordinated in the cell. It is clear that many co-enzymes involved in transcription are phospholipid dependent, and that Proteins called ‘Lamins’ form an orthogonal cytoskeleton inside the nuclear membrane of Eukaryotic cells which must somehow be assembled after cell division. Just like the cytoskeletal proteins that run against the plasmalemma, nuclear membrane Lamins also bind phospholipids with a low affinity. It is known that the dissolution of the nuclear membrane during Cell Division involves phosphorylation of lamin proteins by the action of kinase enzymes. The action of some of these kinases, like PKC, are PS-dependent, and would therefore require that PS-PS headgroup interactions be disrupted before they become active ie their activity would be prompted by a change in the coherent state of nuclear membrane phospholipids. According to the prevailing digital rubric for transcription, kinases are suggested to deliver regulatory signals to enzymes involved in phospholipid synthesis by invoking conformational changes in them: the new model would enable them to feed into the systems’ dynamics by changing surface topography of lamins running against the nuclear membrane, causiong ‘wobble’ in stable membrane lipid flows. This would in turn play against DNA , fluid flows exerting forces that could assist in its winding/unwinding. It is clear that phospholipid synthesis is co-ordinated with regulation of the Cell Cycle (Jackowski, 1994; 1996), the new model suggests that a similar DNA/phospholipid Dynamic Template mechanism is deeply significant to the computation of DNA replication. This may begin to explain why DNA intron sequences reflect flicker frequencies identical to those found in the dynamics of viscous fluids, which is explored in depth in my papers.

References:

1. Aswad DW. (1995). Deamidation and iso-Asp formation in peptides and proteins. Boca Raton, FL: CRC Press.
2. Badger CR, Helmkamp GM Jr.(1982). Effects of the tertiary amine local anaesthetics (dibucaine, lidocaine, tetracaine) on aminophospholipid Flip-Flop Biochim Biophys Acta 692(1):33-40 ).
3. Bangham, A.D. In "Cell Membranes: Biochemistry, Cell Biology and Pathology". (G. Weissmann and R. Clairborne, (Hsrg)). HP Publishing, New York. S. 24-34.
4. Basse F, Sainte-Marie J, Maurin L, Bienvenue A.(1992). Effect of benzyl alcohol on phospholipid transverse mobility in human erythrocyte membrane. Eur J Biochem 205(1):155-62
5. Blowers, R., Clarkson, E., and Maizels, M. (1951). Flicker phenomenon in human erythrocytes. J. Physiol. 113, pp. 228-239.
6. Bray,D; 1995. Nature 376 p307
7. Brochard, F., and Lennon, J.F.(1975). Frequency Spectrum of the flicker phenomenon in erythrocytes. Journal Physique(Paris) 36., pp. 1035-1047.
8. Browicz. (1890). Zbl.Med.Wiss. 28, pp. 625 et seq
9. Browning, J. (1981a). Headgroup dynamics of phospholipids at the membrane surface I . Biochemistry. 20 pp 7144-7151.
10. Browning, J. (1981b). Headgroup dynamics of phospholipids at the membrane surface II. Biochemistry. 20pp7133 - 7144
11. Chen CA, Manning DR. Regulation of G proteins by covalent modification. Oncogene. 2001 Mar 26;20(13):1643-52.
12. Childress, S., and Soward, A.M. 1972. Convection -driven hydrodynamic dynamo. Phys. Rev. Lett. 29 pp 837-839.
13. Conread, H. (1990). “Molecular Computing” pp 235-324. In Marchalli, C and Yorits (Eds). Advances in Computers 31. Academic Press, London.
14. Cossins, A.R. ( Editor) (1994). “Temperature Adaptation of Biological Membranes”. Portland Press Proceedings.
15. Cribier, S., Sainte Marie, J., and Devaux, P.F. (1993). Quantitative comparison of ‘aminophospholipid translocase’ in human erythrocytes and in K562 cell lines. Biochim. Biophys. Acta. 1148: 1. pp 85-90.
16. Cullis, P. and Verkleij, A.J. 1983. Modulation of membrane structure by calcium and dibucaine as detected by 31P nmr. Biochim Biophys Aca 552 pp 546-551.
17. Deme, B, Hess, D., Tristl, et al., M (2000) Eur.J.Physics E2 pp.;125-136. Binding of actin filaments to charged lipid monolayers:film balance studies combined with neutronreflectivity
18. Do SH, Fang HY, Ham BM, Zuo Z. (2002). The effects of lidocaine on the activity of glutamate transporter EAAT3: the role of protein kinase C and phosphatidylinositol 3-kinase. Anesth Analg. 95(5):1263-8
19. Dowhan, W. 1997. Molecular Basis for Membrane Phospholipid Diversity:> Why Are There So Many Lipids? Ann. Rev. Biochem. 66 pp 199-232
20. Dubois-Violette, E., Durand, G., Guyon, E., Manneville, P. and Pieranski, P. (1978). "Instabilities in nematic liquid crystals" in "Liquid Crystals" (Liebert, L. Ed.). Solid State Physics Supplement 14, 147-193.
21. Feudel, F., Gellert, M., Rudiger, S, Witt, A and Seehafer, N. Dynamo effect in a driven helical flow. Phys. Rev. E 68, 046302 (2003)
22. Freedman,M, Kitaev, A and Wang, Z. (2002). Simulation of topological field theories by quantum computers, Comm. Math. Phys. 227 (2002), no. 3, pp587-603.
23. Freedman, S. and Kitaev, A. >
24. Frohlich H (1968) Long-range coherence and energy storage in biological systems. Int. J. Quantum Chem. 2:641-9.
25. Frohlich H (1970) Long range coherence and the actions of enzymes. Nature 228:1093.
26. Frohlich H (1975) The extraordinary dielectric properties of biological materials and the action of enzymes. Proc. Natl. Acad. Sci., U.S.A. 72:4211-4215.
27. Fujiwara, T., Ritchie, K, Murakoshi, H., Ken Jacobson3 and Akihiro Kusumi, (2002) A. Phospholipids undergo hop diffusion in compartmentalized cell membrane Journal of Cell Biology, 157, Number 6, June pp1071-1082
28. Georgiev, D. (2003).Revisiting the microtubule based quantum models of mind: tubulin bound GTP cannot pump microtubule coherence or provide energy for alpha/beta computation in stable microtubules. CogPrints ID 00003251
29. Gomez RS, Barbosa J Jr, Guatimosim C, Massensini AR, Gomez MV, Prado MA.2002. Translocation of protein kinase C by halothane in cholinergic cells. Brain Res Bull May;58(1):55-9
30. Goncharenko, M., and Katov, II. (1985). Effect of cholesterol on the stability of human red cells to electric breakdown. Biofizika 30:3, pp. 441-445.
31. Graham, A. (1933). The effects of shear on convective planiform. Phil. Trans. Roy. Soc.Lond. Series A 232, pp. 285-296.
32. Grush & Churchland, 1999 )
33. Hameroff SR, and Penrose R (1996) Orchestrated reduction of quantum coherence in brain microtubules: A model for consciousness. In: Toward a Science of Consciousness - The First Tucson Discussions and Debates, SR 72. Hameroff, A Kaszniak and AC Scott (eds.), MIT Press, Cambridge, MA. pp 507-540. (also published in) Mathematics and Computers in Simulation (1996) 40:453-480.
34. Harris, A.K., Pryer, N.K. and Paydarfar, D. 1990. J. Exp. Zool. 253 p 163
35. Harzsch S, Anger K, Dawirs RR.Immunocytochemical detection of acetylated alpha-tubulin and Drosophila synapsin in the embryonic crustacean nervous system. Int J Dev Biol. 1997 Jun;41(3):477-84.
36. Hatton J.P., Gaubert, F., Lewis, M.L., Darsel, Y., Ohlmann, P., Cazenave, J.P. and Schmitt, D. (1999). The kinetics of translocation and cellular quantity of protein kinase C in human leukocytes are modified during spaceflight FASEB J. 13 Supp.l:S23-33
37. Hemmings HC Jr. (1998). General anesthetic effects on protein kinase C. Toxicol Lett 1998 Nov 23;100-101:89-95
38. Henszen, M.M., Weske, M., Schwarz, S., Haest, C.W., and Deuticke, B. (1997) . Electric field pulses induce reversible shape transformation of human erythrocytes. Mol. Memb. Biol. 14:4, pp. 195-204.
39. Hornby, A. and Culllis P. Influence of local and neural anaesthetics on the polymorphic phase preferences of egg yolk PE. Biochim Biophys Acta 647 pp 285-292. 1981.
40. Jackowski, S. (1994). Co-ordination of membrane phospholipid synthesis with the Cell Cycle, J.Biol. Chem. 269, 3858-3867.
41. Jackowski, S. (1996). Cell Cycle regulation of membrane phospholipid metabolism, J. Biol. Chem. 271, 20219-20222.
42. Jaffe, L., and Nuccitelli, R. 1977. A. Rev. Biophys. Bioeng. 6 p 445
43. Jibu M, Hagan S, Hameroff SR, Pribram KH, and Yasue K (1994) Quantum optical coherence in cytoskeletal microtubules: implications for brain function. BioSystems 32:195-209.
44. Johnson, F.H., and Flagler, E.A. Science 112 pp 91-92
45. Julien, M., Millot, C., Tocanne, J.F. and Tournier , J.F. (1997). TPA inhibits aminophospholipid translocation. Exp. Cell Res. 234:1, pp.125-31
46. Kitaev, A. (1997). Fault tolerant quantum computation by anyons. quant-ph/9707021.
47. Klausner, R.D., Kumar, N., Weinstein, J., Blumenthal, R. and Flavin, M. (1984). Interaction of Tubulin with phospholipid vesicles. J. Biol. Chem. 256:11 pp 5879-85.
48. Koschmeider, E.L. (1974). Benard convection. Adv. Chem. Phys. 26 pp177-212.
49. Koschmeider, E.L. (1993). Benard cells and Taylor Vortices. (Eds. G.K. Batchelor and Freund, L.B.) Cambridge Monographs on Mechanics and Applied Mathematics
50. Kumar R, Holian O, Cook B, Roshani P.(1997). Inhibition of rat brain PKC by lipid soluble psychotropics.Neurochem Res. Jan;22(1):1-10
51. Kurokawa H, Murray PA, Damron DS.(2002). Propofol attenuates beta-adrenoreceptor-mediated signal transduction via a protein kinase C-dependent pathway in cardiomyocytes. Anesthesiology. Mar;96(3):688-98
52. Le Rumeur E, Fichou Y, Pottier S, Gaboriau F, Rondeau-Mouro C, Vincent M, Gallay J, Bondon A. Interaction of dystrophin rod domain with membrane phospholipids. Evidence of a close proximity between tryptophan residues and lipids. J Biol Chem. 2003 Feb 21;278(8):5993-6001. Dec 11.
53. Leto TL, Fortugno-Erikson D, Barton D, Yang-Feng TL, Francke U, Harris AS, Morrow JS, Marchesi VT, Benz EJ Jr. (1988). Comparison of nonerythroid alpha-spectrin genes reveals strict homology among diverse species. Mol Cell Biol. 1988 Jan;8(1):1-9.
54. Ling E; Danilov YN; Cohen CM (1988). Modulation of red cell band 4.1 function by cAMP-dependent kinase and protein kinase C phosphorylation. J Biol Chem, Feb, 263:5, 2209-16.
55. Liu, S.C., Derick, L., and Palek, J. (1987). Visualisation of the hexagonal lattice in the erythrocyte membrane skeleton. J. Cell Biol. 104, pp. 527-536.
56. Lofthouse, J.T. (1998a). Spatiotemporal Pattern Formation in BioFluids I: Cell Shape Perturbants As Evidence of Spatially-Organised Membrane Flows. Los Alamos Archive Arxiv preprint server & here
57. Lofthouse, J.T. (1998b), Red Cell Shape Responses to Shear Fields Evidence the Convective Organisation of 'Far From Equilibrium' Cell Membranes . XXXLanl Preprint or here
58. Lofthouse, J.T. (1998c) SpatioTemporal Pattern Formation In BioFluids III. Red Cell Echinocytosis : Turbulent Membrane Flows At The " Edge of Chaos". Los Alamos Preprint or here.
59. Lofthouse, J.T. (2004d). Fluids as Transducers of Gravitational Information in Biological Cells. How spatially-organised membrane and cytoplasmic flows establish Gravity-sensitive Microtubule/Microfibril parallelism and maintain it during growth and tropic motion. Lanl Preprint and here
60. Lofthouse, J.T. (1998e) The Dynamic Template Membrane. Lanl Preprint.
61. Lofthouse, J.T. (1998f) Biological Clocks and PT violation . Lanl Preprint
62. Lofthouse, J.T. (1998g). Pattern Formation in Biological Fluids VII: Convective and Shear-Induced Instabilities and an Adaptive Pattern-Forming Mechanisms for Cytoskeletal Assembly How Protein/Lipid feedback enables biological form to emerge spontaneously from the growth process across the cells' Developmental trajectory.Preprint
63. Lofthouse, J.T. (2004h). Towards a Universal Constructor: how cellular fluid dynamics construct red cell cytoskeletons. Available at First draft
64. Lofthouse, J.T.(1998i) Evidence for age-related deamidation of tumour supressor protein p55 and glycophorin C and D in the human erythrocyte membrane. Med. Hyp. 51 pp 321-324.
65. Lofthouse, J.T. (1998j): Paper 6 muscle cell contraction
66. Marchant, H.J.(1982) Plant cell shape. (1982). In “The Cytoskeleton in Plant Growth and Development Lloyd, C.W. (Ed). Academic Press. Chapter 12 pp 295-320 Maretzki D, Mariani M, Lutz HU. Fatty acid acylation of membrane skeletal proteins in human erythrocytes. FEBS Lett. 1990 Jan 1;259(2):305-10.
67. Marshall IN (1989) Consciousness and Bose-Einstein condensates. New Ideas in Psychology 7:73-83.
68. McGough, A.M. (1999) How to build a molecular shock absorber. Current Biology V9, R887-R889.
69. Mondal M, Chakrabarti A.(2002). The tertiary amine local anesthetic dibucaine binds to the membrane skeletal protein Spectrin. FEBS Lett. Dec 18;532(3):396-400.). Spinal anesthesia by local anesthetics stimulates the enzyme protein kinase C and induces the expression of an immediate early oncogene, c-Fos.Anesth Analg. Sep;83(3):542-7
70. Nivarthi RN, Grant GJ, Turndorf H, Bansinath M. 91996
71. Papart, A.K., and Hoffmann, J.F. (1956). Red cell flicker. J. Cell. Comp. Physiol. 47 pp. 295.
72. Ponomarenko, Y. B.: 1973, `The theory of the hydromagnetic generator', J. Appl. Mech. Techn. Phys. 14, p. 775-778.
73. Porter, G.A., Dmytrenko, G.M., Winkelmann, J.C. and Bloch, R.J.. (1992). Dystrophin colocalizes with Spectrin in distinct subsarcolemmal domains in mammalian skeletal muscle. J. Cell Biol. 117: 997-1005
74. Porter, M., Tuszynski, J., and Nip., A. 2000. Conduction pathways in microtubules. Lab Pages.
75. Pullman, B and Pullmann A 1962 Nature 196 1137.
76. Riahi, D.N. (1998). Vortex formation and stability analysis for shear flows over combined spatially and temporally structured walls. TAM Report No. 881(June)
77. Rybicki, A.C., Heath, R., Lubin, B., and Schwartz, R. (1988) Human erythrocyte protein 4.1 is a phosphatidylserine binding protein. J. Clin. Invest. 81 pp 255-260..
78. Sato, O., Nonomura, Y., Kimura, S. and Maruyama, K. (1992). Molecular shape of Dystrophin. J. Biochem. 112 pp 631-636.
79. Schwartz, R., Chiu, D. and Lubin, B. (1985). Erythrocyte membrane phospholipid organisation. Current Topics Haematology 5 pp 63-112.
80. Scott, A. (1996). On quantum theories of mind. J. Consc. Studies. 3(5-6) pp 484-49.
81. Seelig, J.(1993). Phospholipid headgroups as sensors of electric charge. In "New developments in lipid-protein interactions and receptor function." (Wirtz, K., Packer, L., Gustafsson, J., Evangelopoulos, A. and Changeux, J. Eds.) .Plenum Press.
82. Seehafer, N., and Demircan, A. 2003. Dynamo action in cellular convection. Magnetohydrodynamics 39, No. 3, 335-344,
83. Sieffe, C. (2000). Cold numbers unmake the quantum mind. Science 287.
84. Sievers, A., Buchen, B., Volkmann, D., and Hejnowicz, Z. (1991). The role of the cytoskeleton in gravity perception. In “The Cytoskeletal Basis of Plant Growth and Form”. (C.W.Lloyd, Editor). pp. 168-181. Academic Press. ISBN 0-12-453770-7.
85. Sikorski, A., Hanus-Lorenz, B., Jezierski, A., and Dluzewski, A. 2000. Interaction of membrane skeletal proteins with membrane lipid domain. Acta Biochimica Polona 47 pp. 565-578.
86. Simons RM, Laughlin KV, Kampherstein JA, Desai DC, Shurina RD, Mullin JM.(1998). Pentobarbital affects transepithelial electrophysiological parameters regulated by protein kinase C in rat distal colon. Dig Dis Sci. Mar;43(3):632-40
87. Slater SJ, Kelly MB, Larkin JD, Ho C, Mazurek A, Taddeo FJ, Yeager MD, Stubbs CD. (1997). Interaction of alcohols and anesthetics with protein kinase C-alpha J Biol Chem. Mar 7;272(10):6167-73)
88. Smith, I.C.P., Auger, M., and Jarrell, H.C., 1991. Molecular details of anaesthetic-lipid interaction, Ann. NY Acad. Sci., 625, 668-684.
89. Soward, A.M. 1974. A convection-driven dynamo. Phil. Trans. R. Soc. Lond A 275 pp 611-630.
90. Steane A. M. (1997). Active stabilisation, quantum computation and quantum state synthesis Phys.Rev.Lett. 78 pp. 2252-2255
91. Tegmark, M. “The Quantum Brain” quant-ph/9907009 July 5th, 1999.
92. Tuszynski J.A., Brown, J.A. (1998) Dielectric polarization, electrical conduction, information processing and quantum computation in microtubules, are they plausible? Phil Trans Royal Society
93. Velarde, M.G. and Normande, C. (1980). Convection, Sci. Am. 243, 89-93.
94. Wallis, C.J., Babitch, J., and Wenegieme, E.F. (1993). Divalent cation binding to erythrocyte Spectrin Biochemistry 32, pp.5045-50
95. Warren, J. Publications
96. Westermann S, Weber K. Post-translational modifications regulate microtubule function. Nat Rev Mol Cell Biol. 2003 Dec;4(12):938-47.
97. White, R. G., Hyde, G. J. and Overall, R.L. 1990. Protoplasma 158 p 73.
98. Witten, E. (1989). Quantum field theory and the Jones polynomial, Comm. Math. Phys., 121 pp. 351-399.
99. Wolff, J., Zambito, A.M., Britto, P.J, and Knipling,L. (2000) Autopalmitoylation of Tubulin. Protein Sci Jul;9(7), pp. 1357-64
100. Williams, M.W. and Bloch, R.J. (1999). Extensive but co-ordinated re-organization of the membrane skeleton in myofibers of dystrophic (MDX) mice. J Cell Biol. 144:6, pp 1259-70
101. Zabielski, L and Mestel, A.J. 2004. Dynamo Action in Steady Helical Pipe Flow. XXI ICTAM, 15-21 August 2004, Warsaw, Poland
102. Zager RA, Burkhart KM, Conrad DS.(1999) Isoflurane alters proximal tubular cell susceptibility to toxic and hypoxic forms of attack.Kidney Int Jan;55(1):148-59
103. Zhong L, Su JY.(2002). Isoflurane activates PKC and Ca(2+) -calmodulin-dependent protein kinase II via MAP kinase signaling in cultured vascular smooth muscle cells. Anesthesiology Jan;96(1):148-54.

Copyright 2000 MIRth plc: permission needed to use all or part of this paper with suitable accreditation to the author