Active Segregation Dynamics in the Living Cell

Ajay Bansal, Amit Das† , and Madan Rao∗

Simons Centre for the Study of Living Machines,
National Centre for Biological Sciences (TIFR), Bellary Road, Bangalore 560065, India
†
Present address: Department of Physics, Northeastern University, MA 02115, USA

In this paper, we bring together our efforts in identifying and understanding nonequilibrium phase
segregation driven by active processes in the living cell, with special focus on the segregation of cell
membrane components driven by active contractile stresses arising from cortical actomyosin. This
also has implications for active segregation dynamics in membraneless regions within the cytoplasm
and nucleus (3d). We formulate an active version of the Flory-Huggins theory that incorporates
a contribution from fluctuating active stresses. Apart from knitting together some of our past
arXiv:2111.11539v1 [cond-mat.soft] 22 Nov 2021

theoretical work in a comprehensive narrative, we highlight some new results, and establish a cor-
respondence with recent studies on Active Model B/B+. We point to the many unusual aspects of
the dynamics of active phase segregation, such as (i) anomalous growth dynamics, (ii) coarsening
accompanied by propulsion and coalescence of domains that exhibit nonreciprocal effects, (iii) seg-
regation into mesoscale domains, (iv) emergence of a nonequilibrium phase segregated steady state
characterised by strong macroscopic fluctuations (fluctuation dominated phase ordering (FDPO)),
and (v) mesoscale segregation even above the equilibrium Tc . Apart from its implications for actively
driven segregation of binary fluids, these ideas are at the heart of an Active Emulsion description of
the lateral organisation of molecules on the plasma membrane of living cells, whose full molecular
elaboration appears elsewhere.

∗ madan@ncbs.res.in
 2

I. INTRODUCTION

The cell is an organised collection of interacting biomolecules and ions in a highly viscous aqueous medium that
is maintained out-of-equilibrium [1] by active stresses and fluxes [2]. The active forces that drive this organisation,
maintain the cellular system in a nonequilibrium steady state. This molecular organisation and segregation at different
scales is crucial for cellular function and to all processes of life. Here, we will discuss specific nonequilibrium mechanisms
governing the dynamics of segregation of molecular composition within the cell.
A striking example is the molecular organisation at the cell membrane at physiological temperatures [3, 4]. The
components of the cell membrane are subject to fluctuating active contractile stresses from the thin cortical layer of
actomyosin that adjoins it [5, 6] (Fig. 1). We have proposed that these active stresses drive contractile flows that give rise
to a lateral (2d) organisation of multiple species of lipids and proteins over different scales [6–8]. This nonequilibrium
organisation finds experimental support from a variety of high resolution studies such as FRET [6, 7], super-resolution
microscopy [9] and EM [10]. More recently we have investigated how such nonequilibrium active stresses can give rise to
a novel kind of mesoscopic phase segregation on the cell membrane at physiological temperatures, which is larger than
the putative equilibrium liquid-order phase transition temperature [11, 12]. In addition to such motor-cytoskeleton
complexes, localised chemical reactions at the cell surface described by reaction-diffusion-advection equations, can
also generate mechanical or chemical stresses (via a nonequilibrium chemical potential) that may drive mesoscale
segregation [13, 14].

FIG. 1: Schematic of a typical metazoan cell highlighting candidate regions that exhibit nonequilibrium phase segregation
of composition at mesoscales - (A) The plasma membrane, juxtaposed with a layer of cortical actomyosin, is subject to active
stresses which drive lipid and protein mesoscale segregation on the 2d membrane. (B-C) Nonequilibrium chemical processes
have also been suggested as playing a role in liquid-liquid phase segregation in the 3d cell interior, such as in (B) P-bodies
within the nucleus and (C) stress granules within the cytoplasm.

Similarly, stress granules [15, 16] and P-bodies [17], belong to a growing class of membraneless organizations of
specific biopolymers and proteins in the form of liquid-liquid condensates in the cell interior (Fig. 1). While a variety
of equilibrium mechanisms have been proposed for these assemblies (see [18] for a recent review), there have been
recent suggestions that some of these condensates could be 3d realisations of active segregation [19].
In this paper, we combine some of our earlier studies on the active organisation and segregation of cell membrane
composition [6, 13, 20, 21], with newer insights from our [11, 12] and other studies [26, 27], to contribute to the general
understanding of the dynamics of nonequilibrium phase segregation driven by activity. In doing so, we will highlight
those aspects of the active segregation dynamics that are fundamentally different from equilibrium phase segregation.
For specificity, we will focus on the segregation of lipids on the plasma membrane of living cells, but our formalism
has a more general scope.
Giant unilamellar vesicles (GUV) with multi-component lipids, e.g., DOPC+SM+Cholesterol, and giant plasma
membrane vesicles (GPMV) undergo an equilibrium liquid order - liquid disorder phase transition when quenched
 3

to sufficiently low temperatures, Tc = 24.6◦ C for DOPC+SM+Cholesterol [22] and Tc is typically below 25◦ C for
GPMV [23]. Reference [23] emphasizes that at 37◦ C, the GPMV membranes are almost entirely in the mixed, ho-
mogeneous phase. Studies on the dynamics of coarsening of the lipid domains at T < Tc , subsequent to the initial
linearly unstable growth regime, confirm the standard Cahn-Hilliard growth R(t) ∼ t1/3 driven by chemical potential
gradients, with a prefactor that depends on the interfacial tension. Our understanding of the physics of coarsening in
these few-component “artificial” lipid systems is quite mature and summarized in the classic review of [24].
On the other hand, the physics of segregation in the cell membrane differs fundamentally from that in artificial
GUVs and GPMVs, because the cell membrane is additionally subject to a variety of nonequilibrium forces, primarily
fluctuating active contractile stresses arising from a dynamic actomyosin cortex [8, 25]. This has been the subject of
discussion in several papers written over the years [6, 7, 12, 13, 20, 21] and revisited in the current paper. We will see
that activity drives the system to a nonequilibrium steady state with distinct features both in the dynamical approach
to the steady state and the nature of fluctuations in the steady state. These features depend on how activity enters
into the dynamics of coarsening; specifically, on whether the segregating components act as passive scalars (that do
not affect the agencies of activity, i.e., actomyosin) or active scalars (that act back on the agencies of activity). This
classification was introduced in [6, 8, 13] and the analysis of active scalars initiated in [21]. More fundamentally, as
has been highlighted in [26, 27], we will see that it is the breaking of time reversal symmetry (TRS) at the microscopic
level that is at the root of the profound differences between active and equilibrium phase segregation.
In earlier studies, we have investigated the effects of fluctuating active stresses on the dynamics of clustering using
the equations of active hydrodynamics [6, 20, 21] and agent-based brownian dynamics simulations [13, 28]. In [11, 12],
we have studied the active segregation using a description based on a kinetic Master equation, which we solve by
Monte Carlo simulations. Finally, in [21], we have looked at the clustering of active advective scalars using an active
hydrodynamics approach. In this paper we formulate a new approach based on a time dependent Landau-Ginzburg
dynamics starting from an active version of the Flory-Huggins theory [29], to study a hierarchy of active segregation
models for a binary fluid (lipids) subject to active contractile stresses. We also report some new results on the relevance
of active noise in the dynamics of coarsening using the kinetic Master equation approach.

II. ACTIVE FLORY-HUGGINS THEORY

Consider a 2D surface (cell membrane) comprising two molecular components A (lo-lipids) and B (ld-lipids), whose
areal densities are denoted by ρA and ρB . A simple free energy functional describing its equilibrium configurations is
given by,
Z X Z X
ρα (r)vαβ (r − r0 )ρβ (r0 ) dr dr0 + T
 
F [ρA , ρB ] = ρα (r) ln ρα (r) − ρα (r) dr (1)
α,β∈A,B α∈A,B
0
where vαβ (r − r ) is the non-local 2-body interaction. For short-range interactions, one can expand it locally to obtain,
Z   Z X
1  
F [ρA , ρB ] = vαβ ρα ρβ + cαβ ∇ρα ∇ρβ dr + T ρα (r) ln ρα (r) − ρα (r) dr + . . . (2)
2
α∈A,B

It is convenient to express this as a Flory-Huggins free energy functional [29], written in terms of the volume fraction
of component A, φ = ρA /ρ, together with ρ = ρA + ρB , which we take to be incompressible,
Z  
F [φ] = T − χ(T, ρ) φ2 + κ(T, ρ) (∇φ)2 + µ(T, ρ) φ + φ ln φ + (1 − φ) ln(1 − φ) + . . . dr (3)

where χ(T, ρ) = (2vAB −v2TAA −vBB )

, µ(T, ρ) = (vAB −v
T
BB )
and κ(T, ρ) = cAA +cBB
2T
−2cAB
.
We now describe the dynamics of segregation in terms of the hydrodynamic variables – the volume fraction φ and
the individual velocities vα , where the barycentric hydrodynamic velocity,
v = φvA + (1 − φ)vB (4)
is incompressible, ∇ · v = 0. The volume fraction φ of component A, obeys the usual continuity equation,
∂φ
= −∇ · (φvA ) = ∇ · ((1 − φ)vB ) (5)
∂t
The dynamics of the velocities vα is driven by exchange between the species and local stresses acting on the individual
species [30]; assuming stationary flows,
1
γ(vA − vB ) = − ∇ · ΠA + ∇ · ΣA + φ∇p (6)
2
1
−γ(vA − vB ) = − ∇ · ΠB + ∇ · ΣB + (1 − φ) ∇p (7)
2
 4

where the osmotic pressure is partitioned between the two components,

δF
∇ · ΠA = φ∇ , (8)
δφ
(similarly ΠB ) derivable from the Flory-Huggins free energy Eq. 3 and the total stress tensor Σ is partitioned between
the two components Σα [30], a combination of dissipative and active stresses. The hydrodynamic pressure p can be
eliminated by using the incompressibility condition on the hydrodynamic velocity v.
The stress tensor Σα for the two component fluid membrane subject to active contractile stresses arising from a
coupling of cortical actomyosin to component A alone, is given by,
(diss)
Σα = Σα + Σ(act) δα,A (9)
Σ(diss) ηα(s) T
ηα(b)


α = ∇vα + (∇vα ) + (∇ · vα ) 1 (10)
(act)
Σ = ζ1 (ρm , c) nn + ζ2 (ρm , c) 1 (11)
(diss)
where, the dissipative stress Σα is written in terms of the shear and bulk viscous stresses of the individual compo-
nents, and the active stress Σ(act) , with ζ1 and ζ2 being negative for contractile stresses, is expressed in terms of the
density of myosin ρm and the density c and polarization n of actin filaments in the actomyosin cortex, for which we
need to provide the dynamical equations [6, 21]. Since the dense actomyosin cortex is a momentum sink, one must in
principle include a TRS violating contribution proportional to ρm n, to the right hand side of Eq. 6. These, together
with Eqs. 4-11, will form a closed set of equations that describe the segregation of a 2-component fluid membrane
subject to active stresses.
One could imagine that the inter-particle interaction between the A-components could be modulated by the presence
of a third component which is driven by the active contractile stresses. For instance, as shown in [11], the lower leaflet
PS is such a component, but it could as well be the myosin density ρm . This would lead to a change in the local
interaction parameter χ(r) = χ0 (T ) − χ1 (T )ρm (r), which would have the effect of locally enhancing the tendency for
the A-component to segregate, if myosin (ρm ) or PS were to accumulate.
In what follows, we will ignore the hydrodynamic velocity v in the overdamped dynamics of the 2d membrane
embedded in 3d; this allows us to eliminate the pressure p to get,
 
(1 − φ) φ
vA − vB = −∇ · Π − ∇ · ΣB + ∇ · ΣA + Γ ρm n . (12)
γ 1−φ
It is convenient to go from the Active Flory-Huggins theory outlined above to an active Landau-Ginzburg formalism,
written in terms of the segregation order parameter Φ ≡ 2φ − 1, leading to the dynamical equation,
  
∂Φ δF M (Φ)
= ∇ · M (Φ)∇ − ∇ · ΣA − Γ ρ m n (13)
∂t δΦ (1 + Φ)
where we expect the relative exchange mobility M between the two components to be proportional to (1 − Φ) (1 + Φ).
This reduces to the standard Model B dynamics [31] in the absence of active stresses.
When the dynamics of Φ depends on the agencies of the active stress, but not vice versa, these set of equations
describe the segregation of a passive scalar driven by fluctuating contractile active stresses. On the other hand, when
the agencies of the active stress in turn depend on Φ, the resulting set of equations describe the segregation of an
active scalar driven by fluctuating contractile active stresses. In Section III, we will take up the active segregation of
passive scalars, while in Section IV, we will briefly discuss active segregation of active scalars, and show its relation to
recent work on Active Models B and B+ [26, 27].

III. ACTIVE LANDAU-GINZBURG MODELS FOR SEGREGATION OF PASSIVE SCALARS

The Active Flory-Huggins theory outlined above offers a systematic framework to arrive at a hierarchy of active
Landau-Ginzburg models, each of which highlights a different aspect of the influence of activity on segregation dy-
namics. Starting with the full dynamical theory in Φ and ρm , c, n (actomyosin variables), we systematically prune
variables by declaring them to be “fast” – this implies that the theories higher up in the hierarchy contain all the
features of those lower down.

A. Active Landau-Ginzburg Model 1

The dynamical equations for actin and myosin aggregates in the 2d cortex follow from [6, 21] and many other sources,
e.g., see [2] for the early references. These equations include contributions from active translational and rotational
 5

currents which in turn arise from active forces and torques. The concentration of actin concentration obeys a continuity
equation, ∂t c = −∇ · Jc , with an actin current given by
Jc = −Dc ∇c + c γ ∇ · Σ(act) + v0 (ρm )cn (14)
where the first term is the diffusive flux with diffusion matrix Dc , the second term is the active flux derived from
active stress with active translational mobility γ, and the last term is the flux from the active “body force” (a TRS
violating term that is present because the cortex dissipates momentum via friction).
The dynamics of the density of myosin aggregates has similar contributions to the flux, and in addition, includes a
turnover from binding (kb ) - unbinding (ku ),
 
∂t ρm = −∇ · −Dm ∇ρm + L ∇ · Σ(act) + v0 (ρm )ρm n + kb (c) − ku (ρm ) (15)

The dynamics of the polarization of the active actin filaments is given by

Dt n = K1 ∇2 n + K2 ∇(∇ · n) + ζ ∇ · Σ(act) + a(c − c̄) − b|n|2 n


(16)
where the differential operator Dt includes the active self-advection [2, 32], K1 and K2 are the Frank coefficients
associated with splay and bend distortions, respectively, which in principle can depend on ρm , and the last term in the
bracket ensures that the magnitude of |n| ≈ 1, with c̄, being the Onsager concentration of the polar filaments above
which they become orientationally ordered [33].
The cortical fluid adjacent to the PM does not display global orientational order (dilute regime [29]), even so, the
active contractile stress fluctuations can drive large concentration fluctuations, such that locally c(r, t) > c̄, leading to
patterning instabilities which gets positively reinforced due to contractility. Following [21], we highlight 3 cases –
• When the active forces are larger than the active torques, the contractile stresses spontaneously create compact
polar patterns that are motile (Fig. 2(a)). The size of the polar cluster is set by the interplay between filament
diffusion and the contractile active force. The advection term v0 distorts the shape of the cluster from circularity.
It also sets the motility speed of the apolar cluster, which also depends on the asymmetric profile of actin and
myosin in √the cluster. The typical distance between these polar asters is given by the fastest growing mode
−1
kdist ∼ 1/ ζ. These compact contractile clusters of actin and myosin moving with a fixed speed correspond to
travelling pulse solutions [21].
• When the active torques are larger than the active forces, the contractile splay instabilities give rise to compact
apolar clusters that are stationary (Fig. 2(b)). The size of the apolar cluster is again set by the interplay between
filament diffusion and the contractile active force.
• When the active force is comparable to the active torque, there is a spontaneous instability to the formation of
spiral asters that rotate with a local angular frequency proportional to v0 nθ (r) (Fig. 2(c)).
This patterning of concentration and orientation of actomyosin happens over a time scale 1/ζ, with a fast build up of
localised contractile stress that draws in other filaments as well as other neighbouring clusters leading to coalescence.
What effect does this spontaneous formation of contractile clusters have on the segregation dynamics of membrane
components that interact with it? From Eq. 13, it is easy to see that this fast buildup of the active stress and the
concomitant travelling pulse contributes to a driving force for the segregation and growth of the A component. This
can be seen even at T > Tc , from a simple stability analysis about the homogeneous phase.
We perform a preliminary numerical analysis of Eqs. 14-16 and Eq. 13 on a 2d grid, with parameters that lead to
either polar or apolar clusters. Figures 2(d) and 2(e) represent snapshots of the segregation order parameter Φ profile
for a symmetric mixture A-B, starting from the homogeneous phase, when T > Tc and T < Tc , respectively. Our
numerical study confirms the instability towards segregation even when T > Tc . Unlike the equilibrium segregation of
symmetric mixtures, where one expects to see an evolving bicontinuous domain configuration, here we see the evolution
of mescoscale domains, a consequence of the active terms in Eq. 13 that break the Φ → −Φ symmetry.
Figure 2(f) shows snapshots of the order parameter Φ profile, following a quench into the polar cluster phase.
This shows the propulsive movement of domains of component A, and the phoretic reorientation of the direction of
propulsion of one of the domains prior to coalescence. This nonreciprocal long-range sensing and rapid coalescence
of domains is a unique feature of the TRS violating active segregation dynamics [21, 34–38]. The TRS violation can
be traced to the polarization n, which is not a gradient of a scalar. Such nonreciprocal effects of domain propulsion
and growth would eventually lead to anomalous growth of domains. The details of this coarsening dynamics and the
influence of nonreciprocity on growth laws will be taken up later.
The segregation observed here bears some resemblance to the physics of motility-induced phase segregation
(MIPS) [39]. The segregation of the passive scalars is driven by contractile flows into the core of the apolar clusters,
where the density of the A-component gets larger. The mobility that multiplies the active part of the current in Eq. 13,
is Φ dependent and gets smaller as Φ increases (i.e., as the concentration of the A-component increases).
 6

FIG. 2: (a-c) Contractile clusters with definite filament orientation (arrows) and concentration profiles (heat map) emerge
as an instability of the homogeneous, isotropic state by numerically solving Eqs. 14- 16 on a 2d grid for different values of the
Frank coefficients K and active parameter ζ : (a) moving Polar cluster (K = 3, ζ = 30), (b) stationary Apolar cluster (K = 1,
ζ = 30) and (c) rotating Spiral cluster (K = 1, ζ = 10). (d-e) Coarsening dynamics driven by active contractile stresses at
(d) T > Tc and (e) T < Tc . (d) Spatial map of segregation parameter Φ (colour bar) shows coarsening of a symmetric mixture
of A and B at T /Tc = 1.12, starting from a homogeneous state at t = 0 leading to the formation of mesoscale domains of
the A-component at late times (t = 300). (e) Spatial map of Φ at T /Tc = 0.84, shows coarsening into mesoscale domains of
the A-component at late times (t = 300). (f) Phase segregated domains at T > Tc showing non-reciprocal features in the dy-
namics of propulsion and growth, in the regime which seeds polar clusters. Nonreciprocity is apparent during the propulsive
movement of domains of component A (arrows), which shows a phoretic reorientation of the direction of propulsion of one
of the domains prior to coalescence. The length and time scales are set by the filament translational and rotational diffusion
coefficients [21].
 7

B. Active Landau-Ginzburg Model 2

So far we have taken both the actomyosin dynamics and the dynamics of segregation to be deterministic. We
now look at the effects of noise, both thermal and active, which are inevitably present. Thermal noise that appears
additively in Eq. 13 is known to be irrelevant in the coarsening dynamics, save for a finite renormalization of the
interfacial tension [24]. What about active contributions to the noise, which arises as a consequence of stochastic
turnover of actomyosin?
To address this, we note that since the dynamics of the contractile agents is fast in the regime of high activity, it is
useful to construct coarse-grained quantities, the scalar density and polarity of the compact contractile regions Ω,
Z Z
ψ=− cρm ∇ · n p= cn (17)
Ω Ω

In the limit of strong contractility ζ, the compact contractile clusters are apolar, and can be described by the coarse-
grained field ψ alone. The apolar cluster is described by a local orientation field that points radially inward n ∝ −r̂
(Fig. 2a) [6, 20, 21]. This leads to an active contractile stress whose explicit form appears in the dynamics of Φ (Eq. 13).
We include the effects of athermal noise in the dynamics of the coarse-grained field ψ – ψ̇ = k+ − k− (Σ(act) ) ψ – where
the (stress dependent) turnover of actomyosin is represented by a Poisson birth-death process [6, 13, 40].
In [11, 12], we study this stochastic dynamics of active segregation using a kinetic Monte Carlo method. In [12]
we describe the active segregation in the simplest 2-component system in 2d driven by fluctuating active contractile
stresses which captures the essential physics. The dynamics of the membrane components, subject to both equilibrium
and active forces, are described in terms of a Master equation for the time evolution of the probability distribution,
P ({Xiα }, {xa }, t), where xa , a = 1, . . . n, denotes the positions of the compact apolar clusters. We solve the Master
equation using a kinetic Monte Carlo approach, where we specify the updates for the positions {Xiα } of the membrane
components and placement {xa } of the active stress events. While the equilibrium exchange transitions obey detailed
balance, the advective moves on the A-component within the contractile regions and the birth-death moves of the
contractile regions do not (details in [11, 12]).
In [11] we have generalised this theoretical framework to describe the active segregation in a 5-component asymmetric
bilayer containing lo components, ld components and inner leaflet PS. We show that it is the combination of actomyosin
derived active contractile stresses and PS-mediated transbilayer coupling that drives the active segregation of the lo
components. These predictions have been verified in high-resolution fluorescence-based experiments [11].
Figure 3 summarizes our results for the nature of the active phase segregation. We first note that as in Sect. III B,
even when T > Tc , an initial homogeneous state of a symmetric mixture of A and B, evolves to a state with mesoscale
segregated domains of A at late times (Fig. 3(a)). The instability of the homogeneous phase is driven by the apolar
contractile clusters whose rapid turnover ensures that the segregation occurs over a large expanse of space. Similarly,
Fig. 3(b) shows snapshots of the coarsening dynamics when T < Tc ; here too, instead of the usual bicontinuous domains
seen during equilibrium segregation, one sees mesoscale domains of the A component. In Fig. 3(c), we find that the
fluctuating active stresses completely destabilise an initial prepared fully segregated configuration at T < Tc .
In contrast to equilibrium segregation dynamics where the dominant fluctuations are restricted to the interface,
here the fluctuating active contractile stresses ensure that the fluctuations in the bulk too are significant. This is
what drives the interface to be diffuse rather than sharp (as expected in equilibrium segregation) and the bulk to
exhibit strong macroscopic fluctuations leading to breakup and re-formation of macroscopic structures. The interface
dynamics is studied by looking at the behaviour of the dynamical structure factor S(k) = hΦ(k, t)Φ(−k, t)i at large
k (small length scales). For sharp interfaces, S(k) ∼ k −3 ; this Porod behaviour characterises the growing domains
during equilibrium segregation [24]. In contrast, active coarsening dynamics shows departures from Porod behavior;
the domains have diffuse interfaces and low interfacial tension.
On the other hand, bulk statistical quantities such as the integrated order parameter and domain size R(t), show
periods of quiescence followed by large changes over very short times (Fig. 3(d)). This leads to intermittency in the
steady state, a feature exhibited by many driven nonequilibrium systems such as turbulence; the fluctuating active
stresses playing the role of stirring. Intermittency, the sudden precipitous drops and rise of R(t) in the active steady
state, is captured by measuring the time dependent kurtosis of R(t) (scaled fourth moment), which shows a divergence
as t → 0 (Fig. 3(e)). As a result, the steady state exhibits a continual breakup and reformation of macroscopically
large structures.
A striking implication of our results is that in contrast to equilibrium coarsening dynamics, where thermal fluctua-
tions are irrelevant in the renormalization group sense [24], in active coarsening dynamics, active stress fluctuations,
parametrised by an active temperature related to the birth-death rates that drive the system to a distinct nonequilib-
rium steady state characterised by large macroscopic fluctuations (FDPO), are relevant. This fluctuation dominated
phase ordering (FDPO), is studied in great detail in [12] where we discuss its relation with other nonequilibrium
models. Activity destroys the very large domains obtained in equilibrium phase segregation and makes them more
dynamic and intermittent.
 8

FIG. 3: Coarsening dynamics driven by fluctuating active contractile stresses at T > Tc (T /Tc = 1.06) and T < Tc
(T /Tc = 0.7) obtained using a kinetic Monte Carlo simulation for a symmetric mixture of A and B described in the text. Re-
sults shown here are for the same set of activity parameters - fraction of 2d space under the influence of active stresses is 0.3,
correlation length and time of active stresses is ξ = 4 and τ −1 = 0.13 in units of particle size and particle diffusion rate, re-
spectively. (a) Spatial map of segregation parameter Φ (colour bar) shows coarsening at T /Tc = 1.06, starting from a homoge-
neous state at t = 0 to the formation of mesoscale domains of the A-component (red) at late times (t ∼ 105 MCS). (b) Spatial
map of Φ when T /Tc = 0.7, also shows coarsening into finite domains of the A-component at t = 1000. (c) The fluctuating
active stresses destabilise the fully segregated initial state at T /Tc = 0.7, leading to the formation of mesoscale domains of
the A-component. The nonequilibrium steady state obtained at late times is accompanied by large macroscopic fluctuations
of a variety of statistical quantities. (d) The time series of the domain size R(t) obtained from the first zero of the correlation
function g(r) starting from the homogeneous state. The domain growth during equilibrium segregation at T /Tc = 0.7 is con-
sistent with R(t) ∼ t0.3 [12]. For the active segregation at T > Tc , domain size appears to grow as R ∼ t0.25 before saturating
to a finite value at steady state. The fluctuations of the domain size in the steady state are large and abrupt, and correspond
to large domains suddenly breaking up into smaller ones. (e) The large and abrupt macroscopic fluctuations of the domains is
a hallmark of intermittency, which shows up as a divergence of the time dependent kurtosis κ(t) (4th -moment of the domain
size) as t → 0 (in contrast, κ(t) = 3 during equilibrium segregation). The large macroscopic fluctuations that characterise this
nonequilibrium steady state is an example of fluctuation dominated phase ordering (FDPO) [12].

IV. ACTIVE SCALARS = ACTIVE MODEL B/B+

Cell membrane molecules such as GPI-anchored proteins do not influence the dynamics of cortical actomyosin,
and so it is appropriate to describe their actomyosin-dependent segregation using the active segregation models in
Sect. III [6–9, 11]. However, many signalling membrane proteins such as Integrin [9] and Cadherin, or even myosin as
studied in [21], can locally modulate the active stresses - these behave as active scalars.
Thus taking component A to be an active scalar, we need to propose a dependence of the active contractile stresses
on the segregation parameter Φ. There are three distinct ways to do this:
 9

(i) We take Σ(act) to be purely isotropic and a function of Φ alone; it then contributes to the active renormalization of
the osmotic pressure Π(act) (Φ) [13]. The effect would be indistinguishable from equilibrium segregation but for a shift
in the critical temperature, leading to the formation of segregated domains when T is greater than the equilibrium
transition temperature Tc (as in Sect. III B).

(ii) We take Σ(act) to be isotropic, but now a function of Φ and ∇Φ, i.e. in addition to the above, the active osmotic
pressure has, to lowest order, a contribution from Π(act) (Φ, ∇Φ) ∝ (∇Φ)2 . This would lead to a coarsening dynamics,
 
∂Φ δF 2
= ∇ · M∇ + M λ∇ (∇Φ) (18)
∂t δΦ

which could be interpreted as arising from an active renormalisation of chemical potential,

δF 2
µ= + λ (∇Φ) . (19)
δΦ

This is equivalent to the so-called Active Model B [26] and breaks TRS.

(iii) We now take the active contractile stress to be a general anistropic tensor of the form Σ(act) ∝ −ζ4µ n n [2, 41].
On symmetry considerations, the influence of the segregation order parameter Φ on activity is given by a phoretic
contribution n ∝ ∇Φ (as appears in [21]) corresponding to an contractile stress Σ(act) ∝ ∇Φ ∇Φ. Conversely, a local
patterning of the polarisation n generates irreversible fluxes of the order parameter Φ. This leads to a coarsening
dynamics of the form,
 
∂Φ δF
= ∇ · M (Φ)∇ + L∇ · (∇Φ ∇Φ) (20)
∂t δΦ

This is not a mere change in the chemical potential, but a renormalization of the current vector,

J = M ∇µ + L∇ · (∇Φ ∇Φ) (21)

equivalent to the TRS violating Active Model B+ dynamics [27].

A detailed study of the dynamics of coarsening that shows striking nonreciprocal effects mediated through a long
range wake in Φ, similar to that described in [38], will appear later.

V. DISCUSSION

In this paper, we have described a framework to study nonequilibrium phase segregation in a binary mixture driven
by active contractile stresses using an active version of the Flory-Huggins formalism. Though our description is more
general, here we focus on compositional segregation in the plasma membrane, where several studies [6–10] have shown
that one of the primary driving forces behind lateral segregation of composition, both lipids and proteins, at the cell
surface at physiological temperatures, is the nonequilibrium active contractile stresses from the actomyosin cortex
adjoining it. In applying our general formalism to this context, we have brought together some of our past theoretical
work on active clustering and segregation of passive and active scalars [6, 12, 13, 20, 21].
In this paper, we have used both a hydrodynamic approach [6, 20, 21] and a kinetic Monte Carlo simulation [12].
The most striking results are (i) instability to segregation even at T > Tc , (ii) segregation into mesoscale domains
for symmetric binary mixtures even when T < Tc , (iii) nonequilibrium segregated state characterised by macroscopic
and abrupt fluctuations, (iv) propulsion and nonreciprocal features enroute to domain coalescence in systems where
frictional dissipation dominates, and (v) fuzzy domain interfaces suggesting low interfacial tension. These studies have
culminated in a recent proposal of an Active emulsion [11] as a description of the mesoscale organisation of lipids on the
cell membrane at physiological temperatures, much larger than the putative equilibrium liquid-order phase transition
temperature, and contingent on both actomyosin activity and a (lower leaflet) PS-mediated transbilayer coupling. The
generality of our theoretical approach suggests that such Active emulsions could be a general description of mesoscale
segregation in actively driven binary fluids and of mesoscale organisation of molecules driven by a combination of
nonequilibrium stresses and thermodynamic forces, in a variety of cellular contexts. Its immediate implications for
local composition control in the cell make this viewpoint compelling.
 10

ACKNOWLEDGEMENTS

We would like to acknowledge the tremendous contribution of many past students and postdocs, in particular, K.
Gowrishankar, K. Husain, A. Chaudhuri, B. Bhattacharya and A. Polley, whose work over the years has culminated
in our current proposal of active emulsions described here and in [11]. We would also to thank our experimental
colleagues, S. Mayor, and S. Saha, with whom we have had years of active collaborations. We thank S. Ramaswamy for
discussions on the manuscript. We acknowledge support from the Department of Atomic Energy (India), under project
no. RTI4006, and the Simons Foundation (Grant No. 287975), and computational facilities at NCBS. AD acknowledges
support from the Centre for Theoretical Biological Physics at Northeastern University and the Discovery Cluster at
Northeastern University. MR acknowledges the award of JC Bose Fellowship from SERB-DST, India.

[1] R. Phillips, J. Kondev, J. Theriot, and H. Garcia, Physical Biology of the Cell, (Garland Science, New York, 2012).
[2] M. C. Marchetti, J.-F. Joanny, S. Ramaswamy, T. B. Liverpool, J. Prost, M. Rao, and R. A. Simha, Hydrodynamics of soft
active matter, Rev. Mod. Phys. 85, 1143 (2013).
[3] M. Edidin, Lipids on the frontier: a century of cell-membrane bilayers, Nat. Rev. Mol. Cell Biol. 4, 414–418 (2003).
[4] D. Lingwood and K. Simons, Lipid rafts as a membrane-organizing principle, Science 327, 46–50 (2010).
[5] K. Jacobson, P. Liu, B. C. Lagerholm, The Lateral Organization and Mobility of Plasma Membrane Components, Cell 177,
806-819 (2019).
[6] K. Gowrishankar, S. Ghosh, S. Saha, C. Rumamol, S.Mayor, and M. Rao, Active remodeling of cortical actin regulates
spatiotemporal organization of cell surface molecules, Cell 149, 1353-1367 (2012).
[7] R. Raghupathy, A. A. Anilkumar, A. Polley, P. P. Singh, M. Yadav, C. Johnson, S. Suryawanshi, V. Saikam, S. D. Sawant,
A. Panda, Z. Guo, R. A Vishwakarma, M. Rao and S. Mayor, Transbilayer lipid interactions mediate nanoclustering of
lipid-anchored proteins, Cell 161, 581-594 (2015).
[8] M. Rao and S. Mayor, Active organization of membrane constituents in living cells, Curr. Opin. Cell Biol. 29, 126-132
(2014).
[9] T. S. van Zanten, A. Cambi, M. Koopman, B. Joosten, C. G. Figdor, M. F. Garcia-Parajo, Hotspots of GPI-anchored
proteins and integrin nanoclusters function as nucleation sites for cell adhesion, Proc. Natl. Acad. Sci. 106, 18557–18562
(2009).
[10] S. J. Plowman, C. Muncke, R. G. Parton, J. F. Hancock, H-ras, K-ras, and inner plasma membrane raft proteins operate
in nanoclusters with differential dependence on the actin cytoskeleton, Proc. Natl. Acad. Sci. 102, 15500–15505 (2005).
[11] S. Saha, A. Das, A.A. Anilkumar, C. Patra, S. Mayor and M. Rao, Active contractile stress and transbilayer coupling drive
cell surface organisation from nano to mesoscales, in preparation.
[12] A. Das, A. Polley, and M. Rao, Phase segregation of passive advective particles in an active medium, Phys. Rev. Lett. 116,
068306 (2016).
[13] A. Chaudhuri, B. Bhattacharya, K. Gowrishankar, S. Mayor, and M. Rao, Spatiotemporal regulation of chemical reactions
by active cytoskeletal remodeling, Proc. Natl. Acad. Sc. 108, 14825-14830 (2011).
[14] S. D. Hansen, W. Y. C. Huang, Y. K. Lee, P. Bieling, S. M. Christensen and J. T. Groves, Stochastic geometry sensing and
polarization in a lipid kinase–phosphatase competitive reaction, Proc. Natl. Acad. Sc. 116, 01744 (2019).
[15] C. P. Brangwynne, Phase transitions and size scaling of membrane-less organelles J Cell Biol. 203, 875-881 (2013).
[16] J. Wheeler, T. Matheny, S. Jain, R. Abrisch and R. Parker, Distinct stages in stress granule assembly and disassembly,
eLife 5, e18413 (2016).
[17] A.A. Hyman, C. A. Weber and F. Jülicher, Liquid-liquid phase separation in biology, Annu. Rev. Cell Dev. Biol. 30, 39-58
(2014).
[18] J. Berry, C. P. Brangwynne and M. Haataja, Physical principles of intracellular organization via active and passive phase
transitions, Rep. Prog. Phys. 81, 046601 (2018).
[19] C. A. Weber, D. Zwicker, F. Jülicher, and C. F. Lee, Physics of active emulsions, Rep. Prog. Phys. 82, 064601 (2019).
[20] K. Gowrishankar and M. Rao Nonequilibrium phase transitions, fluctuations and correlations in an active contractile polar
fluid, Soft matter 12, 2040-2046 (2016).
[21] K. Husain, and M. Rao. Emergent structures in an active polar fluid: Dynamics of shape, scattering, and merger Phys.
Rev. Lett. 118, 078104 (2017).
[22] E. R. Farkas and W. W. Webb, Precise and millidegree stable temperature control for fluorescence imaging: Application to
phase transitions in lipid membranes, Rev. Sci. Instr. 81, 093704 (2010).
[23] T. Baumgart, A. T. Hammond, P. Sengupta, S. T. Hess, D. A. Holowka, B. A. Baird and W. W. Webb, Large-scale
fluid/fluid phase separation of proteins and lipids in giant plasma membrane vesicles 104, 3165-3170 (2007)
[24] A. J. Bray Theory of phase-ordering kinetics, Adv. Phys. 51, 481-587 (2002).
[25] S. Saha, I-H. Lee, A. Polley, J. T. Groves, M. Rao, and S. Mayor, Diffusion of GPI-anchored proteins is influenced by the
activity of dynamic cortical actin, Mol. Biol. Cell 26, 4033-4045 (2015).
[26] R. Wittkowski, A. Tiribocchi, J. Stenhammar, R. J. Allen, D. Marenduzzo, and M. E. Cates, Scalar φ4 field theory for
active-particle phase separation, Nat. Comm. 5, 1-9 (2014).
[27] M. E. Cates, and E. Tjhung, Theories of binary fluid mixtures: from phase-separation kinetics to active emulsions, J. Fluid
Mech. 836 (2018).
 11

[28] Sk. Raj Hosein, R. Mandal and M. Rao, Stratification, multivalency and turnover of the active cortical machinery are
required for steady active contractile flows at the cell surface, arXiv:2105.11358 [cond-mat.soft].
[29] M. Doi and S.F. Edwards, The Theory of Polymer Dynamics, (Clarendon Press, Oxford, 1986).
[30] H. Tanaka, Viscoelastic phase separation in soft matter and foods, Faraday discussions 158, 371-406 (2012).
[31] P. C. Hohenberg, and B. I. Halperin, Theory of dynamic critical phenomena Rev. Mod. Phys. 49, 435 (1977).
[32] J. Toner and Y. Tu, Flocks, herds, and schools: A quantitative theory of flocking, Phys. Rev. E 58, 4828 (1998).
[33] P-G De Gennes and J. Prost, The physics of liquid crystals, (Oxford University press, vol. 83, 1993).
[34] S. Saha, R. Golestanian, and S. Ramaswamy, Clusters, asters, and collective oscillations in chemotactic colloids, Phys. Rev.
E 89, 062316 (2014).
[35] S. Saha, J. Agudo-Canalejo, and R. Golestanian, Scalar Active Mixtures: The Nonreciprocal Cahn-Hilliard Model, Phys.
Rev. X 10, 041009 (2020).
[36] Z. You, A. Baskaran and M. Cristina Marchetti, Nonreciprocity as a generic route to traveling states, Proc. Natl. Acad. Sc.
117, 19767 (2020).
[37] R. K. Gupta, R. Kant, H. Soni, A. K. Sood, and S. Ramaswamy, Active nonreciprocal attraction between motile particles
in an elastic medium, arXiv:2007.04860 [cond-mat.soft].
[38] J. P. Banerjee, R. Mandal, D. Banerjee, S. Thutupalli and M. Rao Active ploughing through a compressible viscoelastic
fluid: Unjamming and emergent nonreciprocity, arXiv:2109.10438 [cond-mat.soft].
[39] M. E. Cates and J. Tailleur, Motility-induced phase separation, Annu. Rev. Condens. Matter Phys. 6, 219 (2015).
[40] A. Basu, J. F. Joanny, F. Julicher, and J. Prost, Eur. Phys. J. E 27, 149 (2008)
[41] Y. Hatwalne, S. Ramaswamy, M. Rao, and R. A. Simha, Phys. Rev. Lett. 92, 118101 (2004).