The Evolution of Membranes - Semantic Scholar€¦ · lipid bilayer core of biological membranes is ﬂuid and leads to their characteristic, ultra-softmechanical properties [2]

CHAPTER 2

The Evolution of Membranes

M. BLOOM

Canadian Institute for Advanced Research &Department of Physics, University of British Columbia,

6224 Agricultural Road,Vancouver, B.C., Canada V6T 1Z1

O.G. MOURITSEN

Canadian Institute for Advanced Research &Department of Physical Chemistry,

The Technical University of Denmark,DK-2800, Lyngby, Denmark

1995 Elsevier Science B.V. Handbook of Biological PhysicsAll rights reserved Volume 1, edited by R. Lipowsky and E. Sackmann

65

Contents

1. Introduction: optimization of physical properties via evolutionary processes . . . . . . . . . . . . . 67

2. Magnetic crystals in magnetotactic bacteria: an example of optimization . . . . . . . . . . . . . . . . 68

3. Pre-biotic membranes: did they exist? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4. Molecular biology, geology and the classification of cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4.1. Phylogenetic tree of procaryotes (eubacteria and archaebacteria) and eucaryotes from

ribosomal RNA polynucleotide sequencing (‘molecular phylogeny’) . . . . . . . . . . . . . . . . 72

4.2. Geological information on cellular evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5. Archaebacterial membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

6. Proteins in membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

6.1. Integral membrane proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

6.2. Laboratory experiments on model systems: an illustrative example . . . . . . . . . . . . . . . . . 80

7. The biosynthetic pathway for sterols and the plasma membranes of eucaryotic cells . . . . . . . 81

7.1. The insights of Konrad Bloch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

7.2. Influence of sterols on the physical properties of membranes . . . . . . . . . . . . . . . . . . . . . . 82

8. Lipid diversity and the physical properties of membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

8.1. Lipid polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

8.2. Skin lipids – modification of lipids in situ to provide a water-resistant surface . . . . . . . 89

8.3. Essential fatty acids and brain lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

66

1. Introduction: optimization of physical properties via evolutionary processes

Membranes play a crucial physical role in cells, defining as they do the boundarybetween the inside and outside of cells or organelles of cells. For this reason itis obviously appropriate to examine the evolution of membranes in relation to theevolution of cells. Still, although the study of cellular evolution is currently an ex-tremely active and rapidly moving field, very little consideration has been given tothe evolution of membranes. As an example, the word evolution does not appearin the index of a recent, comprehensive textbook on biomembranes [1]. In a sym-metrical manner, the word membrane is not be found in an authoritative treatmentof evolution in a textbook on biophysics [1a]. This is not really surprising in viewof the fact that the systematic treatment of the evolution of physical properties ofbiological materials is a non-existent field.

Modern biology, and especially evolutionary biology, is dominated by ‘molecularbiology’, the study at the molecular level of the transfer of genetic information.Aside from structural proteins and relevant physical properties of proteins in general,the physical properties of biological materials are not programmed directly by thegenes. The genetic code controls the synthesis of proteins, some of which act asenzymes for the synthesis or modification of other molecules. The physical propertiesare governed by the structures formed when these molecules aggregate in a mannerdictated by the normal laws of physics and physical chemistry. Although the feedbackleading to genetic control can be one or more stages removed from protein synthesisthat is controlled directly by nucleic acid sequences, we assume that evolutionaryprocesses still lead to optimization of the physical properties of biological materialsbecause of the almost unimaginably long time scale available. This assumption isthe motivation for our examination of membrane evolution in this chapter.

The assumption of optimization of physical properties via evolutionary processescannot be proven for cell membranes at this time. A schematic illustration of theessential features of the plasma membranes of eucaryotic cells is shown in fig. 1. Thelipid bilayer core of biological membranes is fluid and leads to their characteristic,ultra-soft mechanical properties [2]. Because physical scientists have had little directexperience with such soft materials, we are still at an early stage in the characteri-zation of their physical properties at the present time. We propose that informationfrom evolutionary considerations be used to improve our perception and understand-ing of ‘Nature’s design’ in order to focus on potentially useful experimental andtheoretical questions.

In view of our relative lack of experience in dealing with the physical propertiesof soft, natural materials, it is useful to provide some clear-cut examples of opti-mization of physical properties of other types of biological materials via evolutionary

67

68 M. Bloom and O.G. Mouritsen

Fig. 1. Schematic illustration of a biological membrane showing the geometrical relationship betweenthe lipid bilayer and other membrane features (courtesy of Mr. Broo Sørensen of the Technical Universityof Denmark). The membrane depicted here is representative of the plasma membranes of eucaryotic cells.The rubber-like ‘cytoskeleton’ attached to the cytoplasmic (inner) surface of the bilayer exerts a stronginfluence on the mechanical properties of such a composite membrane. The glycocalyx carbohydratenetwork on the outside surface is generally believed to be responsible for cell–cell recognition andadhesion to other cells. The black objects extending from one side of the lipid bilayer to the other

represent integral membrane proteins.

processes. This is possible because biological systems also make hard, crystallinematerials whose physical properties are extremely well understood and character-ized using conventional condensed matter methods. Whenever such materials, madeby cells, are carefully examined, they invariably prove to have been optimized viaevolutionary processes. An illustrative example is given in section 2.

2. Magnetic crystals in magnetotactic bacteria: an example of optimization

The study of bacterial swimming [3], chemotaxis [4–6] and magnetotaxis [7] providewonderful examples of the optimization of physical properties of biological systemsby evolutionary processes. In each case one is struck by the manner in whichnature has found elegant solutions to formidable physical problems by exploitingfundamental physics in a subtle manner. Deformable bodies can swim under lowReynolds number conditions only by coupling translational motions to certain non-trivial cyclic changes in shapes [3, 8]. Chemotaxis leads to a variety of problemsranging from how to attain nearly perfect detection of molecules striking a surface inwater with minimum coverage (≈ 0.1% for bacteria) of the surface by detectors [4]to the design of tiny motors [6]. However, we choose magnetotaxis as our examplebecause, in solving the problem of directed bacterial swimming along the earth’s

The evolution of membranes 69

magnetic field, nature has produced an optimal design of magnetic materials viaevolutionary processes.

Magnetotactic behaviour of bacteria was first discovered accidentally in 1975 byRichard Blakemore while a graduate student. Bacteria in water were observed (inthe northern hemisphere) to swim persistently in a northerly direction. If observed ina droplet, they accumulated at its northern boundary. In the laboratory, the bacteriaswam in the direction of an applied laboratory magnetic field. Observations of suchbacteria in a transmission electron microscope always showed a series of opaqueparticles in a row. In fig. 2, a bacterium is shown with about 20 particles havingan average diameter of about 50 nm. Each of these particles is a single crystalof magnetite (Fe3O4), whose magnetic properties are well known, surrounded by alipid bilayer membrane. These particles, called ‘magnetosomes’, have by now beenthoroughly investigated [9, 10].

Fig. 2. Magnetotactic Bacterium is seen in a transmission electron microscope which reveals a strikingfeature of its internal structure: a chain of particles that are opaque to the electron beam, aligned more orless parallel to the long axis of the cell. The particles are called magnetosomes and consist of magnetite(Fe3O4) single crystals of size ≈ 50 nm and surrounded by a lipid membrane. The chain functions asa compass, orienting the bacterium in the same direction as the lines of force of the earth’s magnetic

field. The species shown here has a flagellum (a filamentous propeller) at each end (from ref. [7]).


Blakemore and Frankel [7] quickly realized that the biological reason for theevolution of magnetosomes was that it provided a mechanism for the bacteria todistinguish between up and down via the vertical component of the earth’s magneticfield! Consistent with this, it was found that Fe3O4 crystals in magnetotactic bacteriain the southern hemisphere are magnetized in the opposite direction to those in thenorthern hemisphere.

There are two biological advantages for bacteria to swim downward. One is toswim away from the higher concentrations of O2 (poisonous to anaerobic bacteria)near the surface of the water and the other is to swim towards the higher concentra-tions of food in the ooze at the bottom of the lake, river or ocean. Procaryotic cells,having a size L ≈ 1 µm, cannot distinguish between up and down via gravity withreliability as may be seen from the inequality satisfied by the ‘gravitational orderparameter’ Sg = bmgL/kT � 1, where b ≈ 0.1 is a buoyancy factor and mg is theweight of a bacterium. Sg is the ratio of the change of gravitational potential energyof a cell via shifts of its mass distribution over distances of order its size (i.e. asin shape changes while keeping the vertical position of one end of the cell fixed),to the fluctuations in the energy of the cell, kT , associated with equilibrium at atemperature T . By contrast, the magnetic order parameter Sm = MBe/kT ≈ 10 forthe bacterium shown in fig. 2 having a magnetic moment M in the earth’s magneticfield Be ≈ 0.5 gauss. This result is obtained providing that the magnetite in thebacterium of fig. 2 has a value close to its saturation (maximum) magnetization percm3, i.e. M/V ≈ 500 erg/(gauss-cm3), leading to M ≈ 6.5×10−13 erg/gauss [7, 10].

The optimization of the magnetic material goes beyond just getting the ‘right an-swer’ for Sm. The membranes surrounding the magnetite particles are of such asize that the magnetosome diameters always fall within the window of sizes cor-responding to mono-domain, ferrimagnetism, i.e. between 40 nm, below which theFe3O4 particles are ‘super-paramagnetic’ rather than ferrimagnetic, and 120 nm,above which they are multi-domained.The largest possible magnetic moment peratom is obtained within this window. This is a high class engineering solution to amaterials problem and nature has found the solution using evolutionary processes.Space does not permit us to explore other subtle aspects of the magnetosome opti-mization [9, 10]. The intent of this brief description has been to illustrate that, instudying the physical properties of materials such as biological membranes in whichphysical scientists have had only limited experience, a consideration of how naturehas optimized some important properties may provide a guide to useful, system-atic laboratory experiments. We will review some such attempts for membranes insections 6.2 and 7.2.

3. Pre-biotic membranes: did they exist?

The origin of life is an unsolved problem and we confine ourselves to a few remarkson whether the type of membranes found in presently existing cells could haveplayed a role in the very early stages of life. Readers interested in learning aboutmore general aspects of pre-biotic aspects of evolution must look elsewhere (see,e.g., [10a]).


As discussed by Deamer [11], the first cellular systems to appear on the Earth, werepresumably assembled from three molecular species: information-storing moleculescapable of replication, enzyme-like catalysts structurally encoded by that informa-tion and able to enhance replication rates, and boundary-forming molecules whichcould encapsulate the system represented by the first two species. To these shouldbe added molecules capable of storing energy and using this energy to convert com-mon molecules into organized assemblies of biologically active molecules [12]. Itis now generally agreed that under the conditions of volcanic and electrical activitythat probably existed on pre-biotic Earth > 4 eons ago (1 eon = 109 years), the nu-cleotides and amino acids leading to polynucleotides and polypeptides, respectively,would likely have been generated spontaneously [13]. Since it is not clear whethermolecules leading to encapsulation would also have been produced in the Earth’spre-biotic atmosphere [14], an investigation was carried out on the physicochemi-cal properties of amphiphilic compounds in carbonaceous meteorites believed to besimilar to those bombarding the atmosphere of pre-biotic Earth [11, 14, 15]. Theresults of this study demonstrate that organic molecules capable of self-assemblyinto membranes could have been available on the primitive Earth through meteoricinfall.

The specific manner in which life on Earth originated is not known [16]. Eventhe ‘Central Dogma of molecular biology’, which asserts that replication informationin living systems flows from polynucleotide to polynucleotide or polypeptide, butnever from polypeptide, and which is taken as a starting point by almost all personsthinking about the origin of life, is really still open to question in this domain; see,e.g., the proposal by Dyson [17] for an evolutionary sequence from cells to enzymes(polypeptides) to genes (polynucleotides) and compare this with the arguments froma broad biological perspective [18] that “...the transition to vesicles led to the directedchemistry or vectorial chemistry necessary for biogenesis...”

One reason for including this brief section in a chapter on the evolution of mem-branes is to point out that an extension of the studies initiated by Deamer [11] onextremely primitive, naturally occurring membranes to include their physical prop-erties could ultimately contribute towards transferring some speculations concerningthe origins of life into the experimental domain. For example, in our review of theevolution of cells in the next section, we will encounter the proposal that procary-otic and eucaryotic cells parted company from each other in pre-cellular times, eachevolving independently from their common non-cellular, or ‘protocell’ progenote an-cestor. Woese [19] has remarked that the nature of the common organismal ancestoris “...probably the most important, and definitely the least recognized, major questionin biology today”. It may be fruitful to ask whether this proposal is consistent withthe physical properties of the type of pre-biotic material brought to the Earth bymeteorites.

4. Molecular biology, geology and the classification of cells


Fig. 3. Universal phylogenetic tree determined from (r) RNA sequence comparisons, i.e. ‘molecularphylogeny’ [19]. Proposals have been made to root the tree in time on the basis of subsequent workusing molecular phylogenic [24, 25] and geological techniques. Some of the proposed dates for important

evolutionary developments are indicated in fig. 4 (from ref. [7]).

4.1. Phylogenetic tree of procaryotes (eubacteria and archaebacteria) andeucaryotes from ribosomal RNA polynucleotide sequencing (‘molecularphylogeny’)

Though the distinction between procaryotic and eucaryotic cells has been recognizedfor more than a century, and codified more than 50 years ago, the phylogeny of cellsentered a new era with the development and subsequent improvement of molecularbiology techniques for the sequencing of polynucleotides. In particular, the sequenc-ing of fragments of ribosomal (r) RNA has proven to be extremely informative [19].Ribosomes are responsible for the production of proteins and are found in all cells.The quantitative study of correlations between nucleotide sequences of ribosomal (r)RNA for different cells can be used to define an empirical ‘evolutionary distance’between the cells. It has led to a definitive classification of cells into three distinctkingdoms, the eucaryotes and two types of procaryotes, namely eubacteria and ar-chaebacteria. A great triumph of the use of (r) RNA has been the identificationof archaebacteria as a distinct kingdom. In the ‘phylogenetic tree’ shown in fig. 3,constructed from (r) RNA nucleotide sequences [19], there is a much larger ‘evolu-tionary distance’ between any two of the three kingdoms than between cells withineach kingdom. There is broad agreement concerning the modern classification ofcells, but some lively disagreements exist concerning the ancient evolutionary historyof different types of cells [13, 19–25].


It is well established that the fossil record of procaryotes goes back at least 3.5eons while that of eucaryotes goes back no further than 1.9 eons [26]. The infor-mation available from the geological record will be discussed further in section 4.2.The conventional wisdom before the introduction of nucleotide sequencing of (r)RNA was that procaryotic cells were more primitive than eucaryotes and evolvedinto them after the ingestion of one procaryote into another that had lost the exter-nal peptidoglycan cell wall common to virtually all eubacteria – the endosymbionthypothesis [13, 27]. There seems to be general agreement that the ancestors of themitochondrial organelles found in most eucaryotes were, indeed, procaryotic cells asstated in the endosymbiont hypothesis. However, Woese and his collaborators haveargued on the basis of (r) RNA sequencing that there is no reason to believe thateither procaryotes or eucaryotes are more primitive or ancient than the other andthat the three kingdoms of cells could well have parted company from their commonancestor very early in the evolutionary history of cells [19, 28], even as far back as3.6–4.7 eons ago [29]. In their most recent work, Woese and his collaborators haveargued that (r) RNA sequencing favours a ‘rooting’ of the phylogenetic tree shownin fig. 3 such that eubacteria and archaebacteria parted company from the putativecommon ancestor of all known cells, and that eucaryotes and archaebacteria divergedsome time later in evolutionary history [24, 25].

A scenario with a different perspective has been developed by Cavalier-Smith [20,21, 30]. He suggests that, sometime between approximately 1.0 and 1.5 eons ago,archaebacteria and eucaryotes evolved from eubacteria that had lost their ability tomake the muramic acid required for the synthesis of murein, an essential ingredient inthe eubacterial peptidoglycan cell wall. This cell wall, which is common to virtuallyall eubacteria, acts as an external armour plate that maintains the shape of the cell,protects it from osmotic shock, and so forth.

Though Cavalier-Smith makes use of the same background data and observationsas Woese, he places a greater emphasis on physical and morphological differencesbetween the different types of cells. We do not attempt to judge here which ofthese two scenarios will ultimately prove to be closer to the truth. In fact they arenot necessarily inconsistent with each other. However, because the arguments ofCavalier-Smith couple physical features of the different cell kingdoms explicitly totheir molecular biology, they allow us to relate measurements of physical propertiesof cells and, in particular, of cell membranes to questions of evolutionary biology.

A few of the physical and morphological differences between the different typesof cells are listed in table 1. We emphasize that our brief list is, necessarily, a highlypersonal and intuitive selection. This is indicated by a perusal of table 6 of the paperby Cavalier-Smith [21], which lists forty-eight unique features of eucaryotes neverfound in procaryotes!

4.2. Geological information on cellular evolution

The fossil record is one of the most important sources of information on evolutionaryhistory. It must, however, be treated with caution. As emphasized by Knoll [26]in his recent review of the geological record in relation to the early evolution of


Table 1

Selected properties and functions of cells of one type essentially not present in the others.

EUBACTERIA: Peptidoglycan cell wall.

ARCHAEBACTERIA: Non-peptidoglycan, non-cellulosic cell wall.Distinctive lipids – e.g., isopranyl glycerol ether lipids

with isoprenoid and hydroisopranoid hydrocarbons.

EUCARYOTES: Compartmentalization – organelles.Cytoskeleton.Endo- and exo-cytosis.Sterols such as cholesterol for animals and ergosterol for plants.

eucaryotes: “Fossil preservation is facilitated by mineralized hard parts and structurescomposed of organic materials that resist degradation. Many organisms have neitherand so have little chance of fossilization. Those that do may only live in certainenvironments or become fossilized under specific sedimentary conditions. Thus,fossil presence and absence cannot be treated symmetrically.”

In addition to supplying important information on fossils in rocks of independentlyassessed age, the geological record also provides vital information on chemical com-position of the atmosphere and hydrosphere. Thus, the proliferation of diversity ineucaryotes and in animals can be reliably correlated with increased concentrationsof O2 in the earth’s atmosphere [26, 31]. A crude plot of the variation of the earth’satmospheric oxygen pressure with time [26] is shown in fig. 4 along with suggestedcorrelations with important evolutionary developments as given by the fossil recordand molecular phylogeny.

Prior to the evolution of cyanobacterial photosynthesis (2.8 to 2.4 eons ago) theatmospheric oxygen pressure (pO2) was less than 10−10 atmospheres. Then, pO2increased to about 1 to 2% PAL (present-day oxygen atmospheric level), an increaseby a factor of 108! It is believed that pO2 remained at this level, controlled byhigh rates of oxygen consumption, until about 1.9 eons ago, when paleoweatheringsurfaces indicate a rise to > 15% PAL. There are indications of further dramaticincreases of atmospheric oxygen concentration in the period preceding the appearanceof animals, probably about 0.58 eons ago according to the fossil record [26, 31].

Molecular phylogeny suggests that the ancestor of mitochondria dates back to thefirst increase in pO2 or earlier. Knoll [26, 31] also mentions some evidence for theidentification of certain fossils dated as far back as 1.9 eons ago with early eucaryoticcells. However, this identification is not yet definitive and the first pronounced burstof eucaryotic cell activity may have begun only 1.2 to 1.0 eons ago when the mostobvious diversification in the paleontological record began.

The review of evidence from molecular biology and geology on cellular evolutionin this section reminds us that evolution should be considered as a historical sciencethat can only be tested in a manner different from the usual ‘scientific method’, i.e.we can’t repeat the whole experiment. Here, we have paraphrased some commentsby Stephen J. Gould [32] in which he makes ‘a plea for the high status of naturalhistory’.


Fig. 4. A simplified plot of the atmospheric pressure of O2 on earth (pO2) as a function of time basedon geological evidence [26] is shown on the left hand side of the diagram. Shown on the right hand sideare suggested dates or time periods of some important developments in the evolution of cells. Prior tothe evolution of cyanobacterial photosynthesis, pO2 was, perhaps as low as 10−10 atmos. It increasedto concentrations capable of supporting aerobic respiration ≈ 2.8 to 2.4 eons ago. The increase in pO2about 2 eons ago coincides with the first fossil records of eucaryotic cells and the increase about 1 eon(109 years) ago to the present atmospheric level (PAL) coincided with the proliferation of eucaryoticdiversity according to the fossil record [26] and signaled the subsequent increase in animal phyla about0.5 eons later [31]. It should be emphasized that the dates associated with the progenitor of moderneucaryotes (3.5 eons ago) and the first fossils of eucaryotes are not considered by the experts to be

definitive at the time of writing.


Workers in the field of molecular phylogeny are very cautious about assigningspecific dates to the branches of the phylogenetic tree in fig. 3. For this reason,any connections made between studies of physical properties of membranes, cellularevolution using molecular phylogeny and the geological fossil record could be ofimportance. For example, if it can be shown whether or not the cytoskeleton cancouple to the plasma membrane in the absence of sterols (As indicated in table 1 ofsection 4.1, all eucaryotic cells have both sterols and cytoskeletons!), then the firstappearance of the cytoskeleton could be linked to that of sterols. These could, in turn,be linked to the geological evidence for oxygen concentration since the biosyntheticpathway for sterols requires O2 at many stages as shown by Konrad Bloch and tobe discussed in section 7.

5. Archaebacterial membranes

Archaebacteria were only identified as a separate kingdom of cells a relatively shorttime ago and the physical properties of their membranes have been studied lesscompletely than those of eubacteria and eucaryotes.

In the Cavalier-Smith [21] scenario for the evolution of archaebacteria, the in-tegrity and form of the cell was maintained after the loss of the cell wall by there-evolution of diverse types of cell walls [33], and the evolution of a unique classof archaebacterial lipids [34, 35] in response to the extreme conditions of salt con-centration, temperature and pressure (e.g., as in deep sea hot springs) under whicharchaebacteria occupy their special niche.

Unlike the straight-chain fatty acids and fatty acid ester-linked glycerol lipidsthat are characteristic of eubacteria and eucaryotes, archaebacterial lipids are distin-guished by being isoprenoid and hydroisoprenoid hydrocarbon and isopranyl glycerolether-linked lipids. Glycerol ether lipids are found in other types of cells in minoramounts, but only in archaebacteria have they been evolved as the fundamentalglycerolipid component. While some ether lipids are structurally analagous to theircounterparts in eubacteria and eucaryotes, in particular di-ether lipids, others such asthe tetra-ether lipids are not. Some of these are about forty carbons long and havetwo polar ends. It would seem that such lipids would quite naturally span the lipidbilayer structure and this may well be their configuration in biological membranes.However, surprisingly little is known about the structure of archaebacterial mem-branes, partly, it seems, because of the small quantities of their pure lipids availablefor studies of physical properties [36].

One of the best structural studies using X-ray diffraction [37, 38] has led to theintriguing suggestion that the plasma membrane of the archaebacterium Sulfolobussolfataricus may be composed of two interlocking cubic phases [39]; see section 8.1.It appears that the study of the physical properties of archaebacterial lipids is still inits early stages and capable of yielding surprises.

6. Proteins in membranes


Fig. 4a. Illustration of some representative lipid sterol molecules found in eucaryotic and procaryotic(eubacterial and archaebacterial) cells. Four molecules are portrayed schematically from top to bottom.The molecules shown in (i), (ii) and (iii) are found in a single monolayer of bilayer membranes, whilethat shown in (iv) is believed to extend across the entire membrane. (i) The cholesterol molecule,representative of a class of sterol molecules found universally in the plasma (outer) membranes ofeucaryotic cells, but not synthesized by procaryotic cells. (ii) The 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphocholine (POPC) molecule illustrating the acyl chains (oleoyl above and palmitoyl below) andthe headgroup (phospha-tidylcholine) coupled to the glycerol backbone by ester linkages. Such lipidsare found in eubacteria and in eucaryotic cells in a variety of headgroups and a wide range of acylchains of different lengths and degrees of saturation. (iii) A representative di-ether lipid, 2,3-Di-O-phytanyl-sn-glycerol, a basic lipid component of the membranes of all halophilic archaebacteria [35].(iv) A representative tetra-ether lipid, Di-biphytanyl-diglycerol-tetraether, an isoprenoid ether, backbone

of complex lipids of mathanogenic archaebacteria [35].

6.1. Integral membrane proteins

As depicted in fig. 1 and described elsewhere [2], the central core of any biologicalmembrane is the lipid bilayer. There is an interaction between the lipids and proteinsand this interaction must be controlled by the physical constraints imposed by thelipid bilayer medium. Such constraints include the fluidity and the hydrophobicityof the bilayer interior, which we will discuss in turn below.

The fact that the lipid bilayers of almost all biological membranes are fluid under


physiological conditions is almost certainly dictated by the requirement that most ofthe integral membrane proteins spanning the bilayer must undergo conformationaltransitions in order to function. Fluid membranes have the compliance to accom-modate the modifications in protein geometry that must inevitably accompany suchconformational changes. Membrane fluidity, in turn, implies zero shear restoringforce. As a consequence, the shapes of cells are governed by the membrane super-and sub-structure. Most procaryotic cells have exoskeletal cell walls to control andmaintain their shapes while the shapes of eucaryotic cells are usually controlled bythe subtle, and as yet imperfectly understood and characterized, interactions betweenthe lipid bilayer of the plasma membrane and its associated cytoskeletal structure[40]. It may well be that, when the nature of the cytoskeleton-lipid bilayer aspectof lipid-protein interactions is more fully understood, it’s evolutionary developmentwill turn out to be correlated with the proliferation of eucaryotic diversity depictedin fig. 4 as having probably occurred about 1 eon ago.

In this section we confine ourselves to some remarks about the interaction be-tween lipids and integral membrane proteins which are found in both eucaryotic andprocaryotic cells and are shown schematically in fig. 1 as entities spanning the en-tire lipid bilayer. There is great biophysical interest in the determination of proteinstructure and in relating structure to the biological function of the protein. At thistime, much many more structures have been worked out for cytoplasmic, i.e. watersoluble, proteins than for integral membrane proteins. The reason for this is that theamphiphilic nature of integral membrane proteins has made it difficult to prepare thehigh quality single crystals required for X-ray diffraction measurements while theanisotropic nature of the membrane fluidity has made it impossible to use modernNuclear Magnetic Resonance (NMR) high resolution methods of determining proteinstructure [2].

Although the number of integral membrane proteins for which high resolutionstructures are known is extremely small, some important general structural featureshave been established. The secondary structural elements within the bilayer aremainly α-helices composed of amino acids which have, on average, a relativelyhigh degree of hydrophobicity. The 3D (tertiary) structure involves a clustering ofthese α-helices, stabilized by the ‘hydrophobic interactions’ acting in combinationwith coulomb forces between the small number of charged residues placed strate-gically within the hydrophobic region of the membrane. These features, alreadyestablished with the first low resolution structure of Bacteriorhodopsin [41,42], andconfirmed by subsequent high resolution studies [43, 44], have led to an importantmethod of guessing membrane- embedded domains of integral membrane proteinsvia ‘hydropathy algorithms’ [45–47]. The hydrophobic matching of lipids and inte-gral membrane proteins, established by diffraction methods, has been confirmed bynuclear magnetic resonance (NMR) experiments which show that the orientationalorder of lipid acyl chains in membranes, which is linearly related to hydrophobicthickness, is affected very little by the integral membranes that co-exist with thelipids in biological membranes [2].

As illustrated in fig. 5, an integral membrane protein whose hydrophobic length iseither longer or shorter than the equilibrium hydrophobic thickness of its host lipid


Fig. 5. Schematic diagram illustrating the principles of of the ‘mattress model’ of lipid-protein inter-actions [48]. Cross-section of two lipid bilayers each containing amphiphilic transmembrane proteinsor polypeptides. The lipid molecules are indicated schematically by a circular polar headgroup regionand two flexible acyl chains. The proteins (or peptides) are shown as rod-shaped objects with hy-drophilic ends and intermediate hydrophobic regions (cross-hatched). Two situations with a mismatch

are illustrated; the protein is longer (a) and shorter (b) than the lipid bilayer thickness.

bilayer membrane, would produce a mechanical strain on the bilayer. Consequencesof such mismatch have been developed in the mattress model of lipid-protein in-teractions [48]. Manifestations of the effects and potential effects of interactionsencompassed within the mattress model, which have been reviewed recently [49],include such phenomena as: (i) lateral molecular organization within the bilayer,(ii) membrane heterogeneity and the existence of domains having different lipidcompositions, (iii) lipid selectivity by proteins. It can thus be seen that conforma-tional transitions in proteins can trigger changes in the lipid bilayer medium andvice-versa [50]. This gives a link between membrane function and thermodynamics.

In summary, the now well established matching of the hydrophobic regions ofintegral membrane proteins and their host bilayer lipids demonstrates the role ofevolution in the optimization of a basic, biologically relevant, mechanical propertyof membranes; in this case the introduction of proteins into membranes in a strain-free manner.


Fig. 6. Smoothed acyl chain orientational order profiles from depaked 2H NMR spectra [51]. Plotsa are from bilayers composed of a CHOL/POPC-d31 mixture containing 30 mol% CHOL, called B2 inthe text; plots b are from bilayers of pure POPC-d31, called B1 in the text. Both sets of plots are at atemperature of 25◦C. For each set of plots, the (+) plots are for samples with no peptide added, the (•)plots are for samples with 3.5 and 6.3 mol% of peptide P16 in a and b, respectively, and the (4) plots

are for samples with 3.5 and 4.8 mol% of peptide P24 in a and b, respectively. (From ref. [51].)

6.2. Laboratory experiments on model systems: an illustrative example

We now illustrate how one can explore the effects of the hydrophobic matchingcondition by means of laboratory experiments on simple systems. Ultimately, suchan exploration should lead to an improvement in the understanding of biologicalprocesses. The system considered is a POPC-d31 membrane (B1) in which thethirty one protons on the palmitic acid chain have been replaced by deuterons inorder to carry out deuterium (2H) NMR experiments to determine the orientationalorder parameter S(n) as a function of position on the acyl chain. It may be seenfrom fig. 6 that the introduction of 30 mol% cholesterol into the POPC bilayer (wecall this cholesterol – containing membrane B2) results in a large increase in themagnitude of S(n). The hydrophobic thicknesses of B1 and B2 have been estimatedas d(1) = 2.58 nm and d(2) = 2.99 nm, respectively [51], on the basis of a nowwell-established, semi-empirical relationship between d and 〈S(n)〉 for POPC. Thephysical origin of the increase in d when cholesterol is introduced into the lipidbilayer will be discussed in the next section.

A qualitative test of the mattress model is provided by introducing synthetic am-phiphilic peptides Pn = K2GLnK2A-amide of different hydrophobic lengths, corre-sponding to different values of n into B1 and B2. In the experiments to be described,P16 and P24 were used. Since the polyleucine part of these peptides is known to


form a tight α-helix [52], the anticipated hydrophobic length of Pn is dn ≈ 0.15 n,corresponding to d16 ≈ 2.4 nm and d24 ≈ 3.6 nm. It is clear from fig. 6 that theaddition of the short peptide P16 to the thin membrane (B1) or the long peptide tothe thick membrane (B2) results in little change in d(1) and d(2), respectively, asexpected from the mattress model if d16 is approximately matched to d(1) and d24 tod(2). Addition of the long (or short) peptide to the thin (or thick) membrane gives anincrease (or decrease) in d(1) (or d(2)), again in agreement with the mattress model.

7. The biosynthetic pathway for sterols and the plasma membranes ofeucaryotic cells

7.1. The insights of Konrad Bloch

As discussed in section 4.1 (see table 1), cholesterol or structurally equivalent sterolsare found in virtually all eucaryotic cells; yet, only a few procaryotic cells have beenfound that synthesize sterols. The biosynthetic pathway for cholesterol and other suchsterols was worked out by Konrad Bloch [53–55]. He suggested that the evolutionof sterols was intimately connected with the physical requirements imposed on theplasma membranes of eucaryotic cells, presumably because of their greater size andcomplexity. Since the biosynthesis of sterols involves O2 at several stages, andit is difficult to visualize an anaerobic pathway for sterols, this would imply thatthe evolution of the type of eucaryotic cells of the type that we know today couldnot have taken place before O2 appeared in appreciable quantities in the earth’satmosphere (see fig. 4). It seems to us (see section 2) that this point of view may beat odds with some of the proposals on the relative ages of different kingdoms of cellsthat arise from molecular phylogeny related to (r) RNA nucleic acid sequencing. Itwould be interesting to search for ways of dating some of the important features ofthe plasma membranes of eucaryotic cells in relation to the appearance of O2 in theearth’s atmosphere. Did early eucaryotic cells already exist in the anaerobic earthatmosphere? If ‘yes’, did they have organelles and plasma membranes capable ofendo- and exo- cytosis before sterols evolved?

Bloch was especially impressed by the particular sequence of enzymatically con-trolled reactions that follow the cyclization of squalene to lanosterol. Squalene isthe final molecule in the biosynthesis of sterols that is chemically accessible in ananaerobic environment. He writes [54] that “...selective demethylations occur at thesterol α-face streamlining the sterol structure. Successive methyl group removal im-posed by selection pressures rendered the sterol molecule progressively more and,ultimately, optimally competent for membrane function. The sterol pathway termi-nates with cholesterol, a molecule designed to optimize attractive van der Waal’sinteraction with phospholipid chains in the membrane bilayer.” This raises the ques-tion as to whether “...one can discern in the contemporary sterol pathway and in thetemporal sequence of modifying events a directed evolutionary process operating ona small molecule and, if so, whether each step of the sequence produces a moleculefunctionally superior to its precursor.”

In these remarks, Bloch implicitly identifies sterols along the biosynthetic pathwayto the end product (cholesterol for animal cells and ergosterol for many yeasts and


fungi) as molecular fossils. This concept has also been explored for molecules suchas triterpenes and hopanoids that are believed to play a similar structural role inprocaryotic cells [56–58]. One unusual and instructive example of a procaryotic cellexhibiting ‘sterol fossils’ is the bacterium Methylococcus capsulatis, which synthe-sizes ‘methyl and dimethyl sterols’ of the type found along the biosynthetic pathwayin eucaryotic cells, in the presence of O2 [58a]. Of particular interest in this paper isthe measurement of the change in lipid and sterol composition with oxygen tensions.This may make it possible to correlate the evolution of certain enzymes responsi-ble for sterol synthesis with the earth’s O2 concentration as provided by geologicalevidence and depicted in fig. 4.

The concept of a molecular fossil for molecules along the biosynthetic pathwaycarries with it the possibility of an experimental laboratory program to identify thephysical properties that are relevant to the evolutionary ‘optimization’ (see section 1),if, indeed it is a physical property that is being optimized. A study of the influenceof various sterols along the biosynthetic pathway to cholesterol on membrane ‘mi-croviscosity’ was carried out using steady state fluorescence depolarization by Blochand co-workers [59] and their results, shown in fig. 7, exhibit a systematic increaseof ‘microviscosity’ on progressing from lanosterol to cholesterol. In the light ofthese experiments and the many others carried out since we now review briefly theinfluence of sterols on the physical properties of membranes.

Fig. 7. Effect of sterol molecules on the microviscosity of vesicles composed of lipids extractedfrom mycoplasma capricolum at 25◦C using steady state fluorescence depolarization (from ref. [59]).

Microviscosity is expressed in units of poise (g cm−1s−1).


7.2. Influence of sterols on the physical properties of membranes

Following Bloch’s suggestions that cholesterol (CHOL) and other structurally equiv-alent sterols provide some sort of optimization of physical properties of membranes,a more precise interpretation of the type of data shown in fig. 7, in conjunction with2H NMR, calorimetric and other measurements, has been developed [2, 60]. It turnsout that the interpretation of steady state fluorescence depolarization simply in termsof ‘microviscosity’ is an oversimplification based on properties of isotropic fluidsbut not valid for anisotropic fluids such as membranes [60]. Nevertheless, the mainproposals put forward by Bloch [53–55] are born out. As described in section 7.1,Bloch suggested that evolutionary pressure, giving rise to systematic enzymaticallycontrolled modification of sterols, led ultimately to a final sterol (e.g., cholesterol orergosterol or some other equivalent sterol) that optimized the physical properties ofmembranes. At this stage it is appropriate to treat this as a hypothesis that must betested in the laboratory. We summarize some of the main features of the influenceof CHOL as determined experimentally and reformulated in terms of our presentunderstanding of membrane fluidity [2, 60].

It has been known for some time that CHOL in large concentrations eliminates thegel – liquid crystalline phase transition. However, as demonstrated [61] in a definitivestudy of DPPC/CHOL mixtures using 2H NMR and calorimetry, and making use ofdata from a wide range of experimental techniques, the observation that CHOLincreases the orientational order at temperatures T > Tc, i.e. above the normal phasetransition temperature Tc for the pure lipid system, and decreases it at T < Tc, doesnot imply that the system becomes more ‘solid-like’ at T > Tc when CHOL isadded. Rather, the effect of CHOL is to produce a new type of fluid phase, onein which the molecules behave as a fluid in two dimensions, i.e. the moleculesdiffuse and reorientate about the bilayer normal as in a normal fluid. Above all, thesystem exhibits no shear restoring force [2], but, at the same time, the acyl chainsare orientationally ordered as in the gel phase of the pure phospholipid.

As we shall explain below, this behavior can be understood in terms of the natureof the interactions between phospholipid and cholesterol molecules [62]. The theoryhas led to the following terminology for the various phases of the PC/CHOL mixtures.The gel phase is denoted by (so), the indices s standing for solid in two dimensionsand o for orientationally ordered chains, the liquid-crystalline phase (ld) by theindices l for two-dimensional liquid and d for disordered chains and, finally, the newliquid ordered phase is denoted by (lo). We believe the (lo) to be the ‘biologicallyrelevant’ phase.

Phase diagrams for PC/CHOL mixtures are shown in fig. 8. The 2H NMR studiesof Vist and Davis [61] have been extended to three mono-unsaturated PC lipids,namely SEPC [63], PPetPC and POPC [63, 64]. This is appropriate since the plasmamembranes of most eucaryotic cells often have high concentrations of unsaturatedlipids. It seems that the phase diagram of PC/CHOL mixtures has a universal form.The 2H NMR data clearly exhibit a ‘demixing’ between the so and lo phases belowTc. The demixing between the ld and lo phases in the DPPC/CHOL system hasbeen inferred from calorimetry and a number of other measurements, but not from


Fig. 8. Partial phase diagrams for several PC/CHOL mixtures. At low temperatures and low cholesterol(CHOL) concentrations, the system is in the gel or ‘solid-ordered’ (so) phase. As the temperature israised, the system undergoes a transition to the liquid-crystalline or ‘liquid-disordered’ (ld) phase. Athigh CHOL concentrations, the system is in the ‘liquid-ordered’ (lo) phase and there is a demixingregion at intermediate CHOL concentrations. The solid lines are theoretical predictions based on amodel of the lipid-lipid and lipid–CHOL interactions [62] with the interaction parameters adjusted forthe DPPC/CHOL system. It should be emphasized that the model parameters were not adjusted foroptimum fit of the phase diagram. We show the theoretical phase diagram to emphasize its structure inrelation to experimental observations. (a) The DPPC-d62/CHOL system. The experiments were carriedout on DPPC-d62 [61] which has a phase transition temperature about 4◦C lower than for undeuteratedDPPC. The experimental data points have been plotted at correspondingly higher temperatures to matchthem with the theoretical fits and to separate them more from the SEPC/CHOL data. (b) The SEPC-d31/CHOL system [64]. (c) The PPetPC-d31/CHOL system [63, 64]. (d) The POPC-d31/CHOL system

[63, 64].


NMR. It cannot be seen in the NMR spectra because of the rapid molecular exchangebetween the two co-existing liquid phases – or so it is believed [61, 63, 64]. Furtherexperimental work is required on the (ld–lo) part of the phase diagram, which mustbe considered as tentative at this point from the experimental standpoint.

In the theoretical interpretation of these phase diagrams, it is assumed that theinteraction between pairs of phospholipid molecules depends on their mutual orien-tation and account is taken of the internal (conformational) energy of the acyl chains.In the pure lipid system, the model leads to a simultaneous discontinuity (so ↔ ld)in:

(i) chain orientational order (o ↔ d, standing for order at low temperatures inthe gel phase to disorder at high temperatures in the fluid phase) and

(ii) two dimensional symmetry (s ↔ l, from the two-dimensional translationalsymmetry of the gel phase to the translationally disordered fluid phase).

As described above, Konrad Bloch [53–55] has emphasized the role of evolutionin producing, in CHOL, a molecule having a rigid ring structure in one face; the‘α-face’ is ultra-smooth. This feature allows the lipid molecules in their vicinity torotate rapidly. The effect of this is modeled theoretically [62] by assuming that thisadditional freedom reduces the orientation dependence of the interaction of a pairof phospholipid molecules that are both close to a CHOL molecule. Intuitively, therotation would partially average out the orientation dependence of the interaction.The consequence of this feature of the lipid–CHOL interaction is to eliminate thegel to liquid crystalline phase transition at high (> 25%) CHOL concentrations andto produce, as a stable phase, a liquid with relatively high orientational order or,equivalently, a relatively thick membrane.

The question still remains as to what biological advantages could be associated withthe stabilization of the lo phase. In section 7.1, we quoted Bloch [54] as suggestingthat the influence of cholesterol would be to optimize the van der Waals interaction.It is not clear whether the words that Bloch uses apply directly to the picture wehave presented, but an important property of the lo phase is the enormous increase inmechanical strength over the ld phase [65], which must be at least partly due to theincreased importance of the van der Waals interaction for the ordered chains in thelipid molecule. This increased strength is established while maintaining membranefluidity and elimating the phase transition. The latter, in combination with increasedbilayer thickness, reduces membrane permeability greatly [2].

8. Lipid diversity and the physical properties of membranes

The modern era of membrane study essentially began with the fluid-mosaic model [1].One class of questions that has recurred frequently since then concerns the largediversity of lipids found in biological membranes. When one takes into account thedifferent polar head groups and acyl chains, a given membrane may often containwell over 100 unique species. Does each lipid have a specific functional role?Or, is the diversity of lipids in many cell types and the non-random differences inlipid composition between different cell types a manifestation of accidental historical


development rather than the importance of specific lipid composition? At the presenttime, “...there is no satisfying explanation for the observed patterns” ([1], page 32).

Some of the systematics are immediately understandable in a qualitative sensein terms of optimization of physical properties. With the appreciation of the im-portance of membrane fluidity, it soon became clear that many organisms actuallyadjust their lipid composition in response to changes in environmental parameterssuch as temperature in order to preserve membrane fluidity ([1], page 173). Forexample, it is well known that lipids with saturated acyl chains have higher meltingtemperatures than those with unsaturated chains. The mono-unsaturated lipid POPChas a transition temperature about 50◦C lower than the saturated lipid DPPC. Manyorganisms systematically increase (or decrease) the concentration of unsaturated rel-ative to saturated chains in their lipids when the growth temperature is decreased(or increased). The ability of organisms to make such compositional adjustments ispresumably due to the enzymes responsible for the synthesis of unsaturated chainshaving higher activation energies, equivalent to higher temperature-derivatives oftheir activation rates, than those of enzymes responsible for the synthesis of satu-rated chains in lipids. This selection of enzymes driven, not simply on the basis ofenzymatic efficiency but also by its temperature-derivative, must have occurred viaevolutionary processes.

In a similar vein, we have proposed in section 7 a scenario for the optimization ofthe physical properties of plasma membranes of eucaryotic cells, these membraneshaving to satisfy special mechanical considerations because of their relatively largesize. In this case, nature seems often to require a preponderance of unsaturated lipids,mono-unsaturated for red blood cells and poly-unsaturated for the central nervoussystem (see section 8.3). Why mono-unsaturated in one case and poly-unsaturated inthe other is not known – it would seem that both satisfy fluidity needs. These plasmamembranes always have large concentrations of cholesterol (≈ 30 to 50 mol%). Wesuggested in section 7 that this is because sterols extend the fluid range, when alloyedwith di-acyl lipids, while eliminating the phase transition completely, reducing thepassive permeability of membrane and greatly increasing its mechanical strength.

It is likely that many other examples of lipid diversity can be explained in termsof the optimization, via evolution, of the physical properties of membranes. Weclose this final section with brief accounts of three explicit cases that are presentlybeing studied in many laboratories. Each of these cases focuses on different ways inwhich evolutionary forces have manifested themselves in the diversity of lipids. Lipidpolymorphism brings us into contact with the role of symmetry and the breaking ofsymmetry in membranes as well as the influence of membrane curvature. Skin lipidsgive an example of lipids being modified in situ in order to play slightly differentphysical roles as they move from deep in the skin to the surface. Brain lipids bringus into contact with an important aspect of human evolution, brain development,and how that was made possible through the availability of new types of materials,essential fatty acids. These three topics are large ones and our treatments are briefand not intended as reviews. We hope that they will encourage further work froman evolutionary perspective on the manner in which specific lipids have been usedto optimize physical properties of membranes of a variety of organisms.


8.1. Lipid polymorphism

One of the areas of modern physics research closely coupled to membrane biophysicsis the study of liquid crystals, especially, that of lyotropic liquid crystals, i.e. liq-uid crystals whose properties depend strongly on water concentration as well as ontemperature. The lipid structure of prime importance to cell biology is the fluid bi-layer. Lipids that spontaneously aggregate in a bilayer structure, in combination withproteins under physiological conditions, will usually but not always, form lamellararrays having long range order when mixed with water. Such an aggregate, whenfluid, is called the (lamellar) Lα phase and is comprised of a periodic arrangement oflipid bilayers of thickness dl alternating with water layers of thickness dw to definea one-dimensional liquid crystal of repeat distance equal to dl + dw.

It has been known for many years that lipid-water mixtures can give rise to a varietyof liquid-crystalline structures of different symmetries. Their long range order canbe elucidated using standard methods such as X-ray diffraction. In addition to the Lαphase, two types of hexagonal phases are found, corresponding to two-dimensionalarrays of hexagonally coordinated cylinders in which the lipid acyl chains are orientedinside (HI) or outside (HII) the cylinders. As illustrated in fig. 9, the molecularorigin of the hexagonal phases can be understood qualitatively [39, 66] in terms ofthe shapes of the lipid molecules. Not shown in fig. 9, but discussed extensively inthe paper from which this figure was taken [39], is the cubic liquid crystal phaseconsisting of cubically coordinated cylinders, mentioned previously (see section 5)as being potentially relevant to archaebacterial membranes [38]. Cubic phases arefound, for some lyotropic liquid crystalline systems, in a narrow range of waterconcentrations between the low concentrations at which the Lα phase is stable andthe higher concentrations at which the HII phase is stable [67].

Biological membranes are not liquid crystalline in the sense of having long rangeorder so that one does not expect to see hexagonal phases in natural materials. Still,the study of hexagonal liquid crystalline phases has been of great value in focusing onpotential structural and transitional roles of local cylindrical geometry in biologicalfunction.

As an illustration of a potentially important structural role played by local cylin-drical geometry, consider the tight junction, which is crucial to the function ofepithelial sheets. To quote from a well-known cell biology text [13]: from page 29,“The epithelial sheet has much the same significance for the evolution of complexmulticellular organisms that the membrane has for the evolution of single cells.” andfrom page 793, “Epithelia have at least one important function in common: they giveus selective permeability barriers separating fluids on each side that have differentchemical compositions. Tight junctions play a doubly important role in maintainingthe selective-barrier function of cell sheets”.

It is clear from reading further in the text of Alberts et al. [13] that, though thestructure of the tight junction is not known with any confidence at this time, theauthors favour a protein-based structure for the strands that produce the symmetrybreaking required to couple the two bilayers and maintain the permeability barrierrequired of the tight junction. Nevertheless, it has been proposed, on the basis of


Fig. 10. a: diagram of a cross-section of a tight junction illustrating the offset disposition of a pairof intramembrane cylinders. b: diagram illustrating the paired offset cylinders at a tight junction strandand the different possible planes of fracture through such a junction. c: proposed organization of thephospholipids at a tight junction strand. d: diagram of phospholipids combined with freeze-fracturemicrograph to show how fractures through lipid micelles could produce images characteristic of tight

junctions. (From ref. [68].)

freeze fracture electron microscopy [68], see fig. 10, that a pair of intramembrane

Fig. 9. Structure of different anisotropic liquid crystalline phases that membrane lipids can form. (A)Normal hexagonal (HI) phase in which the acyl chains are in the interior regions of the hexagonallycoordinated cylinders; (B) lamellar (Lα) phase; and (C) reversed hexagonal (HII) phase in which theacyl chains are in the exterior regions of the hexagonally coordinated cylinders. We do not show thecubic liquid crystal phases discussed extensively in the paper [66] from which this figure is taken. (From

ref. [39].)


cylinders composed of lipids, i.e. of the type that form the building blocks of theHII phase, may well be the principal structural elements comprising tight junctions.If this proposal is correct, then the structure of tight junctions represents a solutionvia evolutionary methods to a physical symmetry problem, how to couple lamellarstructures to each other in such a way as to maintain good control of permeabilitybetween them, while the system participates in biological activity. The proposedsolution makes use of the range of possibilities opened up by the variety of molecularshapes of the lipid building blocks of membranes. Note that a protein based solutionto this problem would involve the same type of symmetry considerations.

There have been many candidates [39, 66, 67, 69, 70] proposed as probablebiological manifestations of lipid polymorphism via transitional states of membranes.As in the case of structural examples such as the one described above, most of thesecandidates involve a change in local symmetry, transient in this case, which is madepossible by local cylindrical symmetry of the boundary between the hydrophilic andhydrophobic regions of the lipid-water interface. The fusion of pairs of liposomes,budding of small vesicles from larger liposomes, endo- and exo-cytosis are all relatedto each other and are examples of changes in the topology of the membrane surfaces.In the transition from one vesicle to two or vice-versa, say, there must be a transitionalstate in which some part of the local membrane surface deviates appreciably fromlamellar symmetry. The study of such changes of topology and shapes is currentlyan active and important area of both theoretical and experimental study [67, 69, 70].Though most of the work underway at the present time is on model systems, weanticipate that these studies will have an impact on our understanding of physiologicalprocesses in the near future.

8.2. Skin lipids – modification of lipids in situ to provide a water-resistant surface

The properties of skin lipids give an example of how nature has evolved a structureto provide protection of large scale biological organisms from potential environmen-tal dangers, as in water loss, microbial invasion, ultraviolet irradiation, mechanicaltrauma, etc. We discuss here, very briefly one of these features, water loss, to makethe point that as the cells making up human skin, for example, move from the inner(dermis) to the outer (epidermis) layers (see fig. 11), systematic modifications ofthe lipids take place in such a way as to ensure that, among other things, the outer(stratum corneum) layer of the epidermis is relatively water repellent. This can onlybe done with a significant change in cellular structure as ordinary cells are relativelywater permeable. The detailed structure and thermodynamic phase properties of theepidermis and its associated epidermis has yet to be worked out. Those readerswishing to learn about recent work in this field as seen by the specialists in thefield should consult the literature [71]. We are indebted to Dr. Neil Kitson for hisassistance in providing the following extremely simplified description of a complexsystem.

The epidermis, though only the outer layer of the skin, is itself a composite, but iscomposed mainly of cells derived from one type known as keratinocytes. There areother cells but no blood vessels; a complex vascular network is located immediatelybeneath the epidermis in the layer known as the dermis.


Fig. 11. Schematic diagram of the stratum corneum in relation to the overall geometry of the epidermis.The upper part of the figure shows its location and thickness in relation to the outer layers of theskin, with an expanded sketch indicating the type of network of diffusive paths whereby water isbelieved to pass through the stratum corneum. The lower part of the figure shows the manner inwhich terminally differentiated epidermal cells (corneocytes) in the stratum corneum are surroundedby intercellular membranes described as ‘lamellar lipid sheets’. (This figure was provided courtesy of

Dr. Russell O. Potts.)


The epidermis operates in a manner somewhat analagous to that of blood. In theformation of blood, a population of cells (‘stem cells’) spends its time making newcells. These new cells differentiate into ‘mature’ cells (red blood cells) which arebiologically useful for short periods of time, ≈ 120 days in humans [13], and arethen lost, destroyed or recycled. In the case of the epidermis, the stem cells liveat the bottom of the epidermis, next to the dermis, and the maturing cells move uptowards the skin surface. It takes them about two weeks to move from the bottomto the top of the epidermis where they are discarded as dust. The equivalent of thered blood cell is the ‘corneocyte’, and the equivalent of the circulating blood is thestratum corneum, the very top layer of the epidermis. Corneocytes are filled withkeratins (rather than hemoglobin), are glued together with intercellular connections(called desmosomes), and have between them unusual lipid lamellae that provide thewaterproofing.

During the maturing process that occurs between the stem cells on the bottom ofthe epidermis and the stratum corneum on the top, the keratinocytes make small or-ganelles (‘lamellar bodies’) filled with lipids and lipid metabolizing enzymes. Theseare eventually pushed out of the cell (exocytosis) and fuse with each other to formmyelin-like layers around and between the cells. This intercellular lipid is usuallyreferred to as the ‘stratum corneum lipid’ or the ‘intercellular matrix’ or ‘intercellulardomains’. It is technically difficult to analyze the lipid composition at various levelsof the epidermis. However, there is general agreement in the field now that, fromthe time they are synthesized in the lamellar bodies until they form mature inter-cellular membranes, three major classes of lipid are represented in approximatelyequal concentrations: sphingolipids, phospholipids and sterols. During the matur-ing process, enzymatic modifications take place so that a recent model membranestudy [72] could reasonably be based on the assumption that a better approximationto the lipid composition just before the cells are discarded is a three-componentequimolar mixture of free fatty acid, ceramide and cholesterol.

The study of skin lipids may have a special twist with regard to evolution. Afternature had evolved soft materials, it had to find a way of modifying them so as toprovide the special protective features of the skins of animals while using the softmaterials as a starting point. The precise way in which this has been done is stillnot really known, but should be amenable to study using modern physical methods.

8.3. Essential fatty acids and brain lipids

The role of dietary fatty acids in the evolution of the human brain has been reviewedby Crawford [73] and discussed in relation to a broad evolutionary perspective in abook aimed at the general reader [74]. In this section we present a brief account ofsome of the arguments made in the review. An important feature of this discussionis that brain lipids contain a very large fraction of poly-unsaturated fatty acyl chains,which mammals are unable to synthesize because they lack the enzymes to introducedouble bonds at carbon bonds beyond C-9 in the fatty acid chain [75]. In humansthe poly-unsaturated lipids that seem to be of importance to the central nervoussystem are formed via elongation and desaturation from essential fatty acids (EFA),


where the word essential means that they must be supplied in the diet and cannotbe endogenously synthesized. The EFA’s occur as linoleic (n-6) and α-linoleic (n-3)acids, which are elongated and desaturated from 18-carbon chain lengths with twoor three double bonds to 20- and 22-carbon chain lengths with four and six doublebonds. Since the human brain evolved quite recently on the evolutionary time scale, itwill be interesting to understand the physical role played by these poly-unsaturatedfatty acids in brain lipids, and why nature evolved a brain system dependent onmaterial derived from the local environment rather than being under the control ofthe cellular synthetic apparatus. The book by Crawford and Marsh [74] examinesthe consequences, in evolutionary ‘theory’, of important steps “being dictated bychemistry and physics responding to the prevailing conditions”. Such considerationswould, for example, play a role in the relatively simple case of magnetotactic bacteriadescribed in section 2 of this article, where the size distribution of the bacterialmagnets is undoubtedly a reflection of the amount of iron available in the bacterialdiet.

As expressed by Crawford [73]: “Brain chemistry is characterized by two uniquefeatures. First, the brain maintains a constant flow of ionic and electrical informa-tion without which it dies. Its sophisticated communication network is achievedby transmembrane transfer systems with an outstandingly heavy investment in lipidbiotechnology: some 60% of its structural material is lipid. Brain lipid is composedof cholesterol and phosphoglycerides that are rich in the preformed fatty acids, pri-marily arachidonic and docosahexaenoic (DHA), and not the parent EFAs. Forinstance, in the rat, the 20- and 22-carbon long-chain fatty acids are incorporatedinto the developing brain ten times more efficiently than are the parent EFAs. Mostexperts agree that both n-6 and n-3 families of fatty acids are required in the diet.It is also clear that the dietary supply of EFAs is limiting for brain growth”.

Some of the physical implications of the correlations discussed by Crawford [73],deserve serious study from an evolutionary perspective: e.g., correlations between theuse of DHA and the emergence of important biological features in human evolutionsuch as the development of a photoreceptor, synaptic membranes and the role ofinositol phosphoglyceride in its involvement with calcium transport. Such studiescould lead to a better understanding of past and future human evolution. In addition,the inherent limitations in brain development now known to result from maternalmalnutrition due to inadequate supplies of EFAs in the diet of babies in some thirdworld countries provides a social motivation for serious study of how EFAs influencethe physical aspects of fundamental processes in the central nervous system.

Acknowledgements

We wish to thank Neil Kitson and Jenifer Thewalt for helpful contributions to thisarticle.

References1. Gennis, R.B., 1989, Biomembranes (Springer, New York).

1a. Kuhn, H. and Jurg Waser, 1983, Evolution, in: Biophysics, eds W. Hoppe, W. Lohmann, H. Markland H. Ziegler (Springer, Berlin) p. 830.


2. Bloom, M., E. Evans and O.G. Mouritsen, 1991, Physical properties of the fluid lipid-bilayercomponent of cell membranes: A perspective, Q. Rev. Biophys. 24, 293–297.

3. Purcell, E.M., 1977, Life at low Reynolds number, Am. J. Phys. 45, 3–11.4. Berg, H.C. and E.M. Purcell, 1977, Physics of chemoreception, Biophys. J. 20, 193–219.5. Berg, H.C., 1983, Random Walks in Biology (Princeton Univ. Press, Princeton, NJ).6. Berg, H.C., 1988, A physicist looks at bacterial chemotaxis, in: Cold Harbor Symposia on Quan-

titative Biology LIII, 1–9.7. Blakemore, R.P. and R.B. Frankel, 1981, Magnetic navigation in bacteria, in: Sci. Am., December,

pp. 58–65.8. Shapere, A. and F. Wilczek, 1987, Self propulsion at low Reynolds number, Phys. Rev. Lett. 58,

2051–2054. See also 1989, J. Fluid Mech. 198, 557–585 and 587–599.9. Gorby, Y.A., T.J. Beveridge and R.P. Blakemore, 1988, Characterization of the bacterial magneto-

some membrane, J. Bacteriol. 170, 834–841.10. Mann, S. and R.B. Frankel, 1989, Magnetite biomineralization in unicellular microorganisms,

in: Biomineralization – Chemical and Biochemical Perspectives, eds S. Mann, J. Webb andR.J.P. Williams (VCH Verlagsgesellschaft GmbH, Weinheim, Germany) p. 389.

10a. Wachterhauser, G., 1988, Microbiol. Rev. 52, 452–484.11. Deamer, D.W., 1985, Surface structures are formed by organic components of the Murchison

carbonaceous chondrite, Nature 317, 792–794.12. Morowitz, H.J., B. Heinz and D.W. Deamer, 1988, The chemical logic of a minimum protocell,

Origins Life 18, 281–287.13. Alberts, A., D. Bray, J. Lewis, M. Raff, K. Roberts and J.D. Watson, 1989, Molecular Biology of

the Cell, 2nd edition (Garland, New York).14. Deamer, D.W., 1986, Role of amphiphilic compounds in the evolution of membrane structure on

the early Earth, Origins Life 17, 3–25.15. Deamer, D.W. and R.M. Pashley, 1989, Amphiphilic components of the Murchison carbonaceous

chondrite: Surface properties and membrane formation, Origins Life 19, 21–38.16. Morgan, J., 1991, In the beginning, Sci. Am. 257(2), 116–125.17. Dyson, F.J., 1985, Origins of Life (Cambridge Univ. Press, Cambridge).18. Morowitz, H.J., 1992, Beginnings of Cellular Life (Yale Univ. Press, New Haven, CT).19. Woese, C.R., 1987, Bacterial evolution, Microbiol. Rev. 51, 221–271.20. Cavalier-Smith, T., 1975, The origin of nuclei and of eucaryotic cells, Nature 256, 463–468.21. Cavalier-Smith, T., 1987, The origin of eucaryote and archaebacterial cells, Ann. NY Acad. Sci.

503, 17–54.22. Lake, J.A., 1989, Origin of the Eukaryotic nucleus: eukaryotes and eocytes are genotypically

related, Can. J. Microbiol. 35, 109–118.23. Rivera, M.C. and J.A. Lake, 1992, Evidence that eukaryotes and eocyte prokaryotes are immediate

relatives, Nature 257, 74–76.24. Woese, C.R., O. Kandler and M.L. Wheelis, 1992, Towards a natural system of organisms: Proposal

for the domains Archaea, Bacteria and Eucharya, Proc. Nat. Acad. Sci. USA 87, 4576–4579.25. Wheelis, M.L., O. Kandler and C.R. Woese, 1992, On the nature of global classification, Proc.

Nat. Acad. Sci. U.S.A. 89, 2930–2934.26. Knoll, A.H., 1992, The early evolution of eukaryotes: A geological perspective, Science 256,

622–627.27. Margulis, L., 1975, Origin of Eucaryotic Cells (Yale Univ. Press, New Haven, CT).28. Woese, C.R. and R.S. Wolfe, 1985, Archaebacteria: The Urkingdom, in: The Bacteria, Vol. 8, eds

C.R. Woese and R.S. Wolfe (Academic Press, Orlando, FL) p. 459.29. Ragan, M.A., 1988, Ribosomal RNA and the major lines of evolution: A perspective, Biosystems

21, 177–188.30. Cavalier-Smith, T., 1981, The origin and early evolution of the eucaryotic cell, in: Molecular

and Cellular Aspects of Microbial Evolution, eds M.J. Carlile, J.F. Collins and B.E.B. Moseley(Cambridge Univ. Press, Cambridge) p. 33.

31. Knoll, A.H., 1991, End of the proterozoic eon, in: Sci. Am., Oct. 1991, pp. 64–73.


32. Gould, S.J., 1989, Wonderful Life: The Burgess Shale and the Nature of History (Norton and Co.,New York).

33. Kandler, O. and H. Konig, 1985, Cell envelopes of Archaebacteria, in: The Bacteria, Vol. 8, edsC.R. Woese and R.S. Wolfe (Academic Press, Orlando, FL) p. 413.

34. Langworthy, T.A., 1985, Lipids of Archaebacteria, in: The Bacteria, Vol. 8, eds C.R. Woese andR.S. Wolfe (Academic Press, Orlando, FL) p. 459.

35. de Rosa, M., A. Gambacorta and A. Gliozzi, 1986, Structure, biosynthesis, and physicochemicalproperties of archaebacterial lipids, Microbiol. Rev. 50, 70–80.

36. Lo, S.-L. and E.L. Chang, 1990, Purification and characterization of a liposomal-forming tetraaetherlipid fraction, Biochem. Biophys. Res. Commun. 167, 228–243.

37. Gulik, A., V. Luzzati, M. de Rosa and A. Gambacorta, 1985, Structure and polymorphism ofbipolar isopranyl ether lipids from archaebacteria, J. Mol. Biol. 182, 131–149.

38. Luzzati, V.I., A. Gulik, M. de Rosa and A. Gambacorta, 1987, Lipids from Sulfolobus solfataricus.Life at high temperatures and the structure of membranes, Chem. Scr. B 27, 211–219.

39. Lindblom, G. and L. Rilfors, 1989, Cubic phases and isotropic structures formed by membranelipids – possible biological relevance, Biochim. Biophys. Acta 988, 221–256.

40. Sackmann, E., 1990, Molecular and global structure and dynamics of membranes and lipid bilayers,Can. J. Phys. 68, 999–1011.

41. Henderson, R. and P.N.T. Unwin, 1975, Three-dimensional model of purple membrane obtainedby electron microscopy, Nature (London) 257, 28–32.

42. Henderson, R., 1981, Membrane protein structure, in: Membranes and Intermolecular Communi-cations, eds R. Balian, M. Chabre and P.F. Devaux (North-Holland, Amsterdam) p. 231.

43. Deisenhofer, J. and H. Michel, 1989, The photosynthetic reaction center from the purple bacteriumRhodopseudomonas viridis, Science 245, 1463–1473.

44. Henderson, R., J.M. Baldwin, T.A. Ceska, F. Zemlin, E. Beckmann and K.H. Downing, 1990,Model for the structure of Bacteriorhodopsin based on high-resolution electron cryo-microscopy,J. Mol. Biol. 213, 899–929.

45. Kyte, J. and R.F. Doolittle, 1982, A simple method for displaying the hydropathic character of aprotein, J. Mol. Biol. 157, 105–132.

46. Guy, H.R., 1985, Amino side chain partition energies and distribution of residues in soluble proteins,Biophys. J. 47, 61–70.

47. Eisenberg, D., 1984, Three-dimensional structure of membrane and surface proteins, Annu. Rev.Biochem. 53, 595–623.

48. Mouritsen, O.G. and M. Bloom, 1984, Mattress model of lipid-protein interactions in membranes,Biophys. J. 46, 141–153.

49. Mouritsen, O.G. and M. Bloom, 1993, Models of lipid–protein interactions in membranes, Annu.Rev. Biophys. Biophys. Chem. 22, 145–171.

50. Sackmann, E., 1984, Physical basis of trigger processes and membrane structures, in: BiologicalMembranes, Vol. 5, ed. D. Chapman (Academic Press, Orlando, FL) p. 105.

51. Nezil, F.A. and M. Bloom, 1992, Combined influence of cholesterol and synthetic amphiphilicpeptides upon bilayer thickness in model membranes, Biophys. J. 61, 1176–1183.

52. Davis, J.H., D.M. Clare, R.S. Hodges and M. Bloom, 1983, Interaction of a synthetic amphiphilicpolypeptide and lipids in a bilayer structure, Biochemistry 22, 5298–5305.

53. Bloch, K., 1965, The biological synthesis of cholesterol, Science 150, 19–28.54. Bloch, K., 1983, Sterol structure and membrane function, CRC Crit. Rev. Biochem. 14, 47–92.55. Bloch, K., 1985, Cholesterol, evolution of structure and function, in: Biochemistry of Lipids and

Membranes, eds D.E. Vance and J.E. Vance (Benjamin Cummings, New York) p. 1.56. Rohmer, M., P. Bouvier and G. Ourisson, 1979, Molecular evolution of biomembranes: Structural

equivalents and phylogenetic precursors to sterols, Proc. Nat. Acad. Sci. USA 76, 847–851.57. Ourisson, G., P. Albrecht and M. Rohmer, 1984, The microbial origin of fossil fuels, Sci. Am.

250(8), 44–51.58. Krajewski-Bertrand, M.-A., A. Milon, Y. Nakatani and G. Ourisson, 1992, The interaction of

various cholesterol ‘ancestors’ with lipid membranes: A 2H-NMR study on oriented bilayers,Biochim. Biophys. Acta 1105, 213–220.


58a. Jahnke, L.L. and P.D. Nichols, 1986, J. Biol. Chem. 167, 238–242.59. Dahl, C.E., J.S. Dahl and K. Bloch, 1980, Effect of alkyl-substituted precursors of cholesterol on

artificial and natural membranes and on the viability of ‘Mycoplasma capricolum’, Biochemistry19, 1462–1467.

60. Bloom, M. and O.G. Mouritsen, 1988, The evolution of membranes, Can. J. Chem. 66, 706–712.61. Vist, M.R. and J.H. Davis, 1990, Phase equilibria of Cholesterol/DPPC mixtures: 2H-NMR and

differential scanning calorimetry, Biochemistry 29, 451–464.62. Ipsen, J.H., G. Karlstrom, O.G. Mouritsen, H.W. Wennerstrom and M.J. Zuckermann, 1987, Phase

equilibria in the phosphatidylcholine-cholesterol system, Biochim. Biophys. Acta 905, 162–172.63. Thewalt, J.L. and M. Bloom, 1993, Phosphatidylcholine: Cholesterol phase diagrams, Biophys. J.

63, 1176–1181.64. Thewalt, J.L., C.E. Hanert, F.M. Linseisen, A.J. Farrall and M. Bloom, 1992, Lipid-sterol interac-

tions and the physical properties of membranes, Acta Pharm. 42, 9–23.65. Needham, D. and R.S. Nunn, 1990, Cohesive properties (elastic deformation and failure) of lipid

bilayer membranes containing cholesterol, Biophys. J. 58, 997–1009.66. Cullis, P.R., M.J. Hope, B. de Kruijff, A.J. Verkleij and C.P.S. Tilcock, 1985, Structural properties

and functional roles of phospholipids in biological membranes, in: Phospholipids and CellularRegulation, Vol. 1, ed. J.F. Kuo (CRC Press, Boca Raton, FL).

67. Gruner, S.M., 1989, Stability of lyotropic phases with curved interfaces, J. Phys. Chem. 93,7562–7570.

68. Kachar, B. and T.S. Reese, 1982, Evidence for the lipidic nature of tight junction strands, Nature296, 464–466.

69. Lipowsky, R., 1991, The conformation of membranes, Nature 349, 475–481.70. Lipowsky, R., 1992, The physics of flexible membranes, Solid State Phys. 32, 19–44.71. Elias, P.M., 1989, The stratum corneum as an organ of protection: old and new concepts, Curr.

Probl. Dermatol. 18, 10–21.72. Thewalt, J., N. Kitson, C. Araujo, A. MacKay and M. Bloom, 1992, Models of stratum corneum

intercellular membranes: The sphingolipid headgroup is a determinant of phase behavior in mixedlipid dispersions, Biochem. Biophys. Res. Commun. 188, 1247–1252.

73. Crawford, M.A., 1992, The role of dietary fatty acid in biology: Their place in the evolution ofthe human brain, Nutr. Rev. 50, 3–11.

74. Crawford, M.A. and D.E. Marsh, 1989, The Driving Force (Heinemann, London).75. Stryer, L., 1981, Biochemistry, 2nd edition (Freeman, San Francisco), p. 403.

Documents

The Evolution of Membranes - Semantic Scholar€¦ · lipid bilayer core of biological membranes is ﬂuid and leads to their characteristic, ultra-softmechanical properties [2]