Advanced Cell Membrane Biophysics

Born Self-Energy

$$\Delta G_{\text{Born}} = \frac{z^2 e^2}{8\pi\varepsilon_0 R}\left(\frac{1}{\varepsilon_m} - \frac{1}{\varepsilon_w}\right)$$

For a monovalent ion ($z=1$) with $R = 0.2$ nm:

$$\Delta G_{\text{Born}} \approx \frac{(1.6 \times 10^{-19})^2}{8\pi(8.85\times10^{-12})(0.2\times10^{-9})}\left(\frac{1}{2} - \frac{1}{80}\right) \approx 70\, k_BT$$

Three Contributions to TM Helix Insertion

1. Hydrophobic transfer: Burying nonpolar surface area into the bilayer interior. $\Delta G_{\text{hydrophobic}} = -\gamma \times A_{\text{np}}$, where $\gamma \approx 25$ cal/(mol·Å$^2$) is the atomic solvation parameter and $A_{\text{np}}$ is the nonpolar accessible surface area. For a typical 20-residue TM helix:$$\Delta G_{\text{hydrophobic}} \approx -40\, k_BT$$
2. Backbone hydrogen bond cost: Unsatisfied H-bonds in the low-dielectric environment cost $\sim 4\text{-}5\, k_BT$ per H-bond. The alpha-helix satisfies all backbone H-bonds internally, reducing this penalty.
3. Conformational entropy loss: Restricting the chain to a helix costs $\sim 0.5\text{-}1\, k_BT$ per residue relative to the unfolded state in water.

$$\Delta G_{\text{hydrophobic}} \approx -T\Delta S_{\text{water}} + \Delta H_{\text{vdW}} \quad\text{(entropic dominance at 298 K)}$$

Near room temperature, $|T\Delta S| \gg |\Delta H|$, so the hydrophobic effect is entropy-driven. At higher temperatures ($T > 100\,^\circ$C), the enthalpic contribution from van der Waals interactions becomes more important.

$$d = d_{\text{protein}} - d_{\text{bilayer}}$$

Elastic Energy Density

$$\mathcal{F}[u] = \frac{\kappa}{2}(\nabla^2 u)^2 + \frac{K_A}{2}\left(\frac{u}{d_0/2}\right)^2$$

The first term penalizes curvature of the midplane; the second penalizes thickness changes. Minimizing the total energy with boundary condition $u(R_p) = d/2$ yields a modified Bessel equation.

$$\kappa \nabla^4 u + K_A u = 0 \quad\Rightarrow\quad \kappa\left(\frac{d^2}{dr^2} + \frac{1}{r}\frac{d}{dr}\right)^2 u + K_A u = 0$$

The solution that decays at infinity is:$$u(r) = d \cdot \frac{K_0(r/\xi)}{K_0(R_p/\xi)}$$

where $K_0$ is the modified Bessel function of the second kind and $\xi$ is the decay length.

Characteristic Decay Length

$$\xi = \left(\frac{\kappa}{K_A}\right)^{1/4} \approx 1.5\text{-}2.5 \text{ nm}$$

With $\kappa \sim 20\, k_BT$ and $K_A \sim 250$ mN/m,$\xi \approx 2$ nm. This means the membrane relaxes back to its equilibrium thickness within 2-3 nm of the protein surface—only one lipid shell.

Mismatch Free Energy

$$\Delta F_{\text{mm}} \approx \pi K_A\, d^2\, \xi + \pi\kappa\, \frac{d^2}{\xi}$$

The first term comes from membrane compression/stretching; the second from bending. For $d = 1$ nm, this gives $\Delta F_{\text{mm}} \sim 3\text{-}8\, k_BT$, significant enough to influence protein sorting and membrane domain formation.

Protein-Protein Interaction via Mismatch

$$V(D) \sim \pi K_A\, d^2\, \xi\; K_0\!\left(\frac{D}{\xi}\right)$$

This interaction is attractive at short range for proteins with the same sign of mismatch, with amplitude $\sim 1\text{-}10\, k_BT$. The $K_0$Bessel function decays exponentially at large distances with decay length $\xi$, so the interaction is short-ranged.

Annular Lipid Exchange

$$\Delta\Delta G_{\text{annular}} \approx 0.5\text{-}2\, k_BT$$

This modest preference is sufficient to enrich certain lipid species at the protein boundary. With $\sim 20\text{-}30$ annular sites per protein, the total selectivity can be substantial.

$$Z = 1 + \frac{[\text{chol}]}{K_d}, \qquad \langle n_{\text{bound}} \rangle = \frac{[\text{chol}]/K_d}{1 + [\text{chol}]/K_d}$$

Molecular dynamics simulations yield $K_d \sim 0.1\text{-}1$ mol% for CRAC/CARC motifs, corresponding to $\Delta G_{\text{bind}} \sim -3\text{-}-5\, k_BT$.

Curvature Sensing Free Energy

$$\Delta G_{\text{curv}} = -\kappa_m A_p (C_0^{(p)} - C_0^{(m)}) H + \frac{\kappa_m A_p}{2}(C_0^{(p)} - C_0^{(m)})^2$$

Here $H$ is the local mean curvature, $\kappa_m$ is the monolayer bending modulus, $A_p$ is the protein's footprint area, and $C_0^{(p)}, C_0^{(m)}$ are the spontaneous curvatures of the protein-containing and bare monolayers. The linear term drives curvature sensing (proteins preferentially bind to curved regions), while the quadratic term represents the self-energy cost.

Coupled Hamiltonian

$$\mathcal{H} = \int\left[\frac{\kappa}{2}(\nabla^2 u)^2 + \frac{\sigma}{2}(\nabla u)^2\right] dA + \mathcal{H}_{\text{couple}}$$

where the coupling energy confines membrane fluctuations around the cortex position $h_0$:

$$\mathcal{H}_{\text{couple}} = \frac{\gamma}{2}\int (u - h_0)^2\, dA$$

The coupling constant $\gamma$ (in units of pN/nm$^3$) parameterizes the stiffness of the attachment between bilayer and cortex.

Suppressed Fluctuation Spectrum

$$\langle |u_q|^2 \rangle = \frac{k_BT}{\kappa q^4 + \sigma q^2 + \gamma}$$

Without coupling ($\gamma = 0$), the standard Helfrich spectrum$\langle|u_q|^2\rangle \sim k_BT/(\kappa q^4)$ diverges as$q \to 0$. The coupling constant $\gamma$ acts as a confining potential that suppresses long-wavelength (small $q$) fluctuations, effectively "pinning" the membrane to the cytoskeleton.

Coupling Correlation Length

$$\xi_c = \left(\frac{\kappa}{\gamma}\right)^{1/4}$$

Fluctuations on length scales larger than $\xi_c$ are strongly suppressed by the cytoskeletal coupling. For RBC with $\gamma \sim 10^{-6}$ pN/nm$^3$,$\xi_c \sim 100$ nm.

WLC Force-Extension (Marko-Siggia)

$$F(x) = \frac{k_BT}{L_p}\left[\frac{1}{4(1 - x/L_c)^2} - \frac{1}{4} + \frac{x}{L_c}\right]$$

At low extension ($x \ll L_c$), the chain behaves as a linear spring. Near full extension ($x \to L_c$), the force diverges as $(1 - x/L_c)^{-2}$.

Spectrin Spring Constant

$$k_{\text{sp}} \approx \frac{3k_BT}{2L_p L_c (1 - \bar{x}^2)}$$

where $\bar{x} = x_0/L_c$ is the fractional extension at rest. For spectrin in the RBC at physiological conditions ($\bar{x} \approx 0.4\text{-}0.6$):$k_{\text{sp}} \sim 5\text{-}20\, \mu$N/m.

2D Shear Modulus

$$\mu = \frac{\sqrt{3}}{4}\frac{k_{\text{sp}}}{s_0}$$

For RBC spectrin: $s_0 \approx 80$ nm (junction spacing), giving $\mu \approx 4\text{-}10\,\mu$N/m, in excellent agreement with micropipette aspiration measurements.

Non-Equilibrium Fluctuation Spectrum

$$\langle |u_q|^2 \rangle_{\text{active}} = \frac{k_BT}{\kappa q^4 + \sigma q^2 + \gamma} + \frac{f_{\text{active}}^2 \tau}{(\kappa q^4 + \sigma q^2 + \gamma)^2}$$

The second term represents excess power from ATP-driven processes.$f_{\text{active}}$ is the active force density and $\tau$ is the correlation time of the active kicks. ATP depletion eliminates the excess and restores the thermal spectrum.

Mixing Free Energy Density

$$f(\phi) = \frac{k_BT}{a^2}\left[\phi\ln\phi + (1-\phi)\ln(1-\phi) + \chi\,\phi(1-\phi)\right]$$

The first two terms are the ideal entropy of mixing; $\chi$ is the Flory-Huggins interaction parameter quantifying the energetic penalty of unlike neighbors.$a$ is the lipid spacing.

Spinodal and Critical Point

$$\frac{\partial^2 f}{\partial \phi^2} = 0 \quad\Rightarrow\quad \frac{1}{\phi} + \frac{1}{1-\phi} - 2\chi = 0$$

At the critical point ($\phi_c = 1/2$):$$\chi_c = 2$$

For $\chi > 2$, the mixture is unstable at $\phi = 1/2$ and undergoes spinodal decomposition into $L_o$ and $L_d$ phases.

Equilibrium Raft Size

$$R^* = a \cdot \exp\!\left(\frac{2\pi\lambda a}{k_BT\, \Delta f}\right)$$

where $\lambda$ is the line tension between $L_o$ and$L_d$ phases ($\lambda \sim 0.1\text{-}5$ pN), and$\Delta f$ is the free energy difference between mixed and demixed states. Near the critical point where $\lambda$ is small, rafts remain nanoscale ($R^* \sim 10\text{-}200$ nm).

Ginzburg-Landau Free Energy

$$F[\phi] = \int\left[f(\phi) + \frac{\varepsilon^2}{2}|\nabla\phi|^2\right] dA$$

The bulk term $f(\phi)$ is the Bragg-Williams mixing free energy (double-well for $\chi > \chi_c$). The gradient term penalizes sharp interfaces and sets the interface width $w \sim \varepsilon/\sqrt{|\chi - \chi_c|}$. The line tension between coexisting phases is$\lambda = \varepsilon\int_{-\infty}^{\infty}(\phi')^2\, dx \propto \varepsilon(\chi - \chi_c)^{3/2}$.

Cahn-Hilliard Equation

$$\frac{\partial \phi}{\partial t} = M\nabla^2\left(\frac{\delta F}{\delta \phi}\right) = M\nabla^2\left[f'(\phi) - \varepsilon^2 \nabla^2\phi\right]$$

$M$ is the mobility (related to lipid lateral diffusion coefficient). The Laplacian ensures conservation: the chemical potential$\mu = \delta F/\delta\phi$ drives diffusive flux$\mathbf{J} = -M\nabla\mu$.

Growth Rate of Perturbations

$$\omega(q) = -Mq^2\left(f''(\phi_0) + \varepsilon^2 q^2\right)$$

When $f''(\phi_0) < 0$ (inside the spinodal), modes with$q < q_c = \sqrt{-f''/\varepsilon^2}$ are unstable ($\omega > 0$).

Dominant Wavelength

Maximizing $\omega(q)$ with respect to $q$:$$q_{\max} = \sqrt{\frac{-f''(\phi_0)}{2\varepsilon^2}}, \qquad \lambda_{\max} = 2\pi\sqrt{\frac{2\varepsilon^2}{-f''(\phi_0)}}$$

This sets the initial domain spacing during spinodal decomposition. At later times, domain coarsening follows a power law: $R(t) \sim t^{1/3}$ (Lifshitz-Slyozov-Wagner growth law for conserved dynamics).

2D Diffusion-Limited Rate

$$k_{2D} = \frac{2\pi D_{2D}}{\ln(L/a)}$$

where $L$ is the system size and $a$ is the target radius. This weak (logarithmic) dependence on system size means that 2D diffusion to a target is only mildly sensitive to confinement, but it also means that the 2D rate depends on the effective membrane domain size.

The average cluster size follows:$$\langle n \rangle \approx 6\text{-}12 \text{ molecules per cluster}$$

Cluster lifetime is $\sim 0.1\text{-}1$ s, set by the balance between Ras-Ras attractive interactions ($\sim 1\text{-}2\, k_BT$ per pair) and entropic cost of confinement. Nanoclusters amplify signaling by creating local high-concentration platforms for effector recruitment.

Electrostatic Association

$$K_{\text{assoc}} \propto \exp\!\left(\frac{n_{\text{basic}} \cdot z_{\text{PIP}_2} \cdot e^2}{4\pi\varepsilon_0\varepsilon_w k_BT\, \ell_D}\right)$$

The association constant depends exponentially on the number of basic residues$n_{\text{basic}}$ in the binding domain. This creates an ultrasensitive electrostatic switch: a protein with 5 basic residues has $\sim 10^3$-fold stronger binding than one with 3, enabling sharp discrimination.

Partition Function for Coupled Array

$$Z = (1 + [L]/K_{\text{off}})^N + L_0 \cdot (1 + [L]/K_{\text{on}})^N$$

The probability of the active state is:$$P_{\text{on}} = \frac{L_0(1 + [L]/K_{\text{on}})^N}{(1 + [L]/K_{\text{off}})^N + L_0(1 + [L]/K_{\text{on}})^N}$$

Here $[L]$ is the ligand concentration, $K_{\text{on}}, K_{\text{off}}$are the dissociation constants in the on/off states, $L_0$ is the allosteric constant, and $N$ is the number of coupled receptors.

Hill Coefficient

$$n_H \approx N \quad\text{(for tightly coupled arrays)}$$

For $E.\, coli$ chemotaxis receptor arrays with $N \sim 10\text{-}20$coupled receptors, the sensitivity gain is $\sim 10\text{-}20$-fold, enabling detection of sub-percent changes in attractant concentration across a vast dynamic range.