THE ANATOMICAL RECORD (NEW ANAT.) 261:198–216, 2000

FEATURE ARTICLE

The New Architectonics: An Invitation To Structural Biology

CLARENCE E. SCHUTT* AND UNO LINDBERG

The philosophy of art might offer an epistemological basis for talking about the complexity of biological molecules in a meaningful way. The analysis of artistic compositions requires the resolution of intrinsic tensions between disparate sensory categories—color, line and form—not unlike those encountered in looking at the surfaces of protein molecules, where charge, polarity, hydrophobicity, and shape compete for our attentions. Complex living systems exhibit behaviors such as contraction waves moving along muscle fibers, or shivers passing through the growth cones of migrating neurons, that are easy to describe with common words, but difficult to explain in terms of the language of chemistry. The problem follows from a lack of everyday experience with processes that move towards equilibrium by switching between crystalline order and chain-like disorder, a commonplace occurrence in the submicroscopic world of proteins. Since most of what is understood about protein function comes from studies of isolated macromolecules in solution, a serious gap exists between what we know and what we would like to know about organized biological systems. Closing this gap can be achieved by recognizing that protein molecules reside in gradients of Gibbs free energy, where local forces and movements can be large compared with Brownian motion. Architectonics, a term borrowed from the philosophical literature, symbolizes the eventual union of the structure of theories—how our minds construct the world—with the theory of structures—or how stability is maintained in the chaotic world of microsystems. Anat Rec (New Anat) 261:198–216, 2000. © 2000 Wiley-Liss, Inc.

KEY WORDS: architectonics; cytoskeleton; functional microanatomy; molecular motor; structural biology; proteins; hydrophobic effect; Gibb’s free energy; thermodynamics; proteomics

REDUCTIONISM IN THE plausible case could be made that they ment—will soon be freely available to POST-GENOMIC ERA didn’t exist at all, and even if they did, everyone to probe, exploit, or to study it wasn’t clear how mere aggregations as new form of knowledge about our- We are coming to grips with our mo- could give rise to distinct qualities selves. While it is not at all obvious lecular destiny. Just one hundred (Duhem, 1914). The application of x- how we will read this map, it is a years ago, when it appeared impossi- ray analysis to crystals by Max von certainty that a sound understanding ble that atoms could ever be seen, a Laue, William Bragg, Sr. and Jr., of the structure of proteins will be changed all of that in short order. needed, because all of life’s in- Dr. Lindberg is Professor of Zoological Their legacy, delivered by subsequent termeshed processes, including the Biology at the Wenner-Gren Institute of Stockholm University. He is First Vice- generations of x-ray crystallogra- natural decoding of the message, act President and Foreign Secretary of The phers, is that structures of complex through them. Royal Swedish Academy of Sciences, and Past Chair of the Stockholm branch of the biological macromolecules, whose We are now accustomed to hearing Swedish Folk University. Dr. Schutt is presence was barely suspected a cen- about “the post-genomics era” and Professor of Chemistry at Princeton Uni- versity. He is a Member of the Program in tury ago, can not only be seen, but words like “proteomics” and “emer- Neuroscience, an Associated Faculty obey the ordinary laws of physical gomics” are meant to alert us to where Member in the Molecular Biology Depart- ment, and former Director of the Molecu- chemistry. Genes themselves, a math- the real intellectual or technological lar Biophysics Program. He is Chairman ematical abstraction needed to ex- (not to say investment) opportunities of the Board of the National Alliance for plain regularities in plant heredity, lie. The “proteome” is the set of all Autism Research and Annual Co-Chair and Founder of the Princeton-Eden Lec- were revealed as elaborate three-di- proteins actually present in a given ture Series on Autism. mensional structures containing nu- cell, as revealed by electrophoresis *Correspondence to: Clarence E. Schutt, The Henry H. Hoyt Laboratory, Wash- cleic acids wrapped around proteins. gels capable of resolving thousands of ington Road, Princeton University, The human genome—nothing less distinct spots, each reporting the pres- Princeton, NJ, 08544, USA. E-mail: [email protected] than a highly encoded map of our ma- ence of a single protein. Similarly, terial being, its history and develop- patterns of gene expression, both spa-

© 2000 Wiley-Liss, Inc. FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 199 tially and temporally, can be found by to the complexity of speakable ideas. teins (see Figure 2) . These surfaces expanding the repertoire of messen- An idea that contains too many are somewhat easier to describe be- ger RNA present in individual cells minute yet closely related parts, too cause they are shaped to conform to and identifying them with antibodies many relations within relations, can- one another. Even though each is or anti-sense oligomers of DNA. The not be ‘projected’ onto discursive complicated in itself, the pleasing amount of information will be stag- forms; it is too subtle for speech.“ “clasped hands” nature of their mu- gering and its implications for under- Langer (1967) is claiming that the tual interactions, allows us to grasp standing human diseases and affect- boundary conditions set by the linear the meaning of the functional associ- ing cures can be hardly exaggerated. structure of sentences limits what we ation. As one becomes familiar with All of this appears to represent the can communicate about the arts different classes of interactions and triumph of reductionism; all compo- through language alone. The great motifs by studying many crystal struc- nents and sub-components of living works of art must speak for them- tures, a kind of visual gallery develops, systems can be described, in princi- selves directly through the multiple which can form the basis for a lan- ple, by atomic structures. Atoms com- portals of our senses to some con- guage having its own syntax and se- bine to form molecules, some of them sciousness beyond words. mantic rules. To participate in the dis- very large indeed, such as proteins, In many ways, the surfaces of pro- courses of structural biology, one which require for their synthesis min- teins share this discursive aspect with must learn how to express what one iature machine-like ribosomes. Pro- the visual arts (see Figures 1 and 2). It means to say about biological func- teins are synthesized as long chains, is difficult to find words to describe tion in this language. annealing into compact structures in the complexity of form one sees. One It is this aspect of structural biology tiny oven-like chambers called chap- that invites comparison with formal eronins. Recently, the structures of criticism in the arts or in literature. some of the chaperonins (Sigler et al., Architectonics pertains When an art critic assesses a new 1998), and even the ribosome itself work, he or she assumes that the (Wimberly et al., 2000; Ban et al., to both “the theory of reader knows something of the history 2000), have yielded their secrets to structure,” as in the of art. To see the new demands knowl- crystallographers, now armed with edge of the old. But there is more to it. powerful synchrotron x-ray sources, architectural literature, References to earlier works carry with and computers with awesome capa- and “the structure of them contexts of discovery, break- bilities for data reduction and visual- through and failure, drawing from the ization. These structures show us the theory,” used by reader’s consciousness a broad spec- arrangement of the working parts and philosophers to denote trum of memories and feelings. It is set the stage for understanding how the intention of the critic to set the these movements are coordinated to the processes by which stage for these evoked responses. The carry out the functions of these com- ideas become fixed in mind then becomes receptive to the plex particles. our consciousness. critic’s message. In structural biology, certain structures are “classic” in the same way that the paintings of Ce- WHAT IS ARCHITECTONICS? zanne, Van Gogh, Picasso, Rothko, or Architectonics is a venerable term in is clearly in the presence of a compo- Stella define unique compositional the philosophical literature. It per- sition, but even the canonical forms of styles, immediately recognizable, tains to both “the theory of structure”, sculpture, including Alexander Calder’s evocative. We will refer to these clas- as in the architectural literature, and mobiles, are inadequate to the task of sic structures as “structural para- “the structure of theory”, used by phi- describing all of the events that take digms.” losophers in a variety of ways to de- place during protein folding or during But, architectonics is not the same note the processes by which ideas be- the encounter of two proteins. Even if as architecture; “tonics” refers to ten- come fixed in our consciousness the surfaces are color-coded for sions inherent in a structure. Return- (Kant, 1787). Suzanne K. Langer charge, hydrophobicity, and polarity, ing to Langer’s analysis of artistic (1967) in her magisterial three-vol- the presence of ridges, clefts, knobs, composition: “The most fundamental ume work on the philosophy of art and holes adds so much texture and elements seem to be tensions; and uses the word “architectonics” to ex- landscape to the chemical complexity upon closer inspection, tensions show pound on the discursive, non-sequen- that one is at a loss for where to begin. some peculiarly interesting traits. By tial, quality of visual art forms: “Their In recent years, crystallographers their very occurrence they immedi- complexity, consequently, is not lim- have determined structures for a large ately engender a structure. They act ited, as the complexity of discourse is number of functional protein com- on each other in a great variety of limited, by what the mind can retain plexes. What these structures reveal is ways—they can be handled so as to from the beginning of an apperceptive a high degree of complementarity, intersect without losing their identity, act to the end of it. Of course, such a both chemical and topological, be- or contrariwise, so that they fuse and restriction on discourse sets bounds tween the surfaces of interacting pro- compose entirely new elements. They 200 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE

to form tertiary, or fully-folded, struc- learned from these classic structures, tures. Proteins interact with each passively absorbed, are unconsciously other to form higher-order structures at play in the formulation of molecu- like filaments, tubules, and viral shells. lar mechanisms for all biological pro- These “first elements” of biological cesses. structure are not rocks or bricks, but As an exercise in thinking without dynamic units with internal tensions structural paradigms, try to define en- between the competing tendencies for zyme specificity without recourse to order and disorder that lie at the heart specific molecules. A good start might of life’s rhythmic processes. The struc- be to invoke the notion of “lock and tures of DNA and RNA are the pri- key”. Fair enough, this idea served en- mary elements of storage and retrieval zymologists rather well during the de- of genetic information. Their twists cades before the structures of the first and turns, and sequence variations, enzymes were obtained. But, exactly define a world of sufficient template how is the lock fashioned from bits of complexity to encode life’s design. In polypeptide chain? How unique is the the cytoplasm, actin is the first ele- lock, how secure from tampering ment of a system of meshed and cross- (e.g., by oxidants)? Once the key is linked filaments that enables cells to inserted (a process signposted by a move, to pull, and to change shape. decreasing Gibbs free energy, de- These molecules are “bobbing” in a scribed below), what makes it turn? torrent of Gibbs free energy, a quan- And how does it get out? Are mole- tity of immense importance for under- cules of a certain softness or pliability standing the emergence and mainte- required to meet a particular func- nance of biological complexity (see tional need? Are these requirements Box I). The flow of free energy starts in line with the properties of bonded in the nuclear transformations in the carbon, nitrogen, oxygen, and hydro- sun’s core, which produce a vast gen? The question boils down to this: Figure 1. Hierarchical structure of proteins. stream of ultraviolet photons that are given complete knowledge of the laws Proteins are covalently-bonded polymers of harnessed by the photosynthetic ma- of physical chemistry, can you pro- the twenty naturally-occurring amino acids. Polypeptide chains fold into three-dimen- chinery of plants. These little factories vide a plausible basis for the diversity sional structures, unique for each sequence produce glucose molecules, rich in of biological reactions carried out by of amino acids. The actin molecule is shown chemical bond energy, that are con- macromolecules (gene regulation, as an example. Higher order structures con- verted by mitochondria in plant and force generation, signal transduc- sist of subunits of proteins held together by extensive non-covalent interactions. The animal cells into ATP, adenosine tion), without referring to results ob- helical actin filament is shown in light relief in trisphosphate, a universal fuel whose tained by x-ray crystallography? the background. Figure originally appeared breakdown into ADP and inorganic For example, compare the struc- in Kreatsoulas et al. (1999). phosphate by muscles and other kinds tures of trypsin, chymotrypsin, and of transport machinery makes life elastase (Stryer, 1988). Notice how possible. naturally the description of specificity can be intensified or muted, resolved follows, how easily the common cata- either by being spent or by being lytic triad is distinguished from the STRUCTURAL PARADIGMS counterbalanced, modified by a specificity pocket—no fuzzy thinking, touch, and all the while they make for It is sometimes claimed by those mo- no dubious assumptions. While its structure. This appears to be true in lecular biologists who can perform function could conceivably be de- all the great orders of art; in every one the neat teleological trick of decanting duced from its structure (it might be of them, a general range of tensions is “function” from “structure” that struc- easy to spot an icosahedral virus, or a set up by the first element —line, ges- tural biology is nothing more than a globin e.g.), the reverse operation can- ture, or tone—which the artist estab- tool that provides scaffolds upon not yet be performed without relying lishes” (Langer, 1967). which to hang notions of information on family resemblances within an es- Proteins can be described as hierar- flow. This view presupposes “structur- tablished paradigm. This is the aspect chical structures using similar lan- al paradigms,” structures whose inter- of biology that makes it so different guage (Figure 1). The lowest level of pretations have led to new principles from physics, where First Principles the hierarchy is the polypeptide chain and ways of looking at biological pro- generally can be specialized to the itself. Under the proper conditions, cesses. The discovery of new struc- case at hand. the chain can fold back upon itself tural paradigms is akin to discoveries The helix-loop-helix DNA-binding forming elements of secondary struc- in the world of art because the micro- motif is another classic piece of pro- ture, such as helices and sheets, which scopic world never looks the same tein architecture. In looking at a new subsequently pack against each other once they have appeared. The lessons DNA-binding protein, one looks first FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 201 to see if this element is present, and 2000). Indeed, the reverse side of the chy. The analogy to boundary condi- considers other precedents in trying coin, how one set of proteins (the hi- tions, a concept taken from mathe- to classify the new in terms of the old. stones) can package up all possible matical physics, is apt. When a taut But in doing so, what should echo in DNA sequences into chromatin, has wire is strung between two points and the mind are the struggles of the early been revealed in an elegant crystal plucked, the possible waveforms on band of protein-DNA crystallogra- structure analysis by Tim Richmond the wire are restricted to those which phers. Carl Pabo, Tom Steitz, Brian and colleagues (Luger et al., 1997). are integral fractions of the distance Matthews, Steve Harrison and others Chris Calladine, a structural engi- between the endpoints; the imposition were trying to find a structural code in neer working a stone’s throw away of a boundary condition, in this case the patterns of charge and polarity on from the Cavendish Laboratory in the pinned ends of the wire, results in the surfaces of these proteins that Cambridge where it all started, devel- the selection of a particular subset of might be used to predict how the oped a set of simple rules that made it possible harmonic motions. edges of specific base pairs could be possible to describe deformations of Similarly, the solutions to Schrod- recognized in the major groove. Im- DNA in terms of base twists and tilts inger’s wave equation, the distinc- plicit in these attempts was the belief and other operations on the molecule tively shaped atomic orbitals (s,p,d,f) that the phosphodiester sugar back- (for recent review see Lu and Olson, familiar from freshman chemistry, bone of DNA was an invariant, se- 1999). While these rules have a me- arise from the restriction that nega- quence-independent, structural fea- chanical flavor easy to comprehend, a tively charged electrons stay in the po- ture. bridge has been crossed where, even tential well of the positively charged The famous Watson-Crick B-form though the machine-like character of nucleus. When electrons are forced to of DNA was originally deduced from accommodate the restrictions of x-ray fiber patterns. Fiber diffraction, multi-centered positive charges, as though clearly of value in defining he- To participate in the when six atoms of carbon come to- lical parameters such as helical pitch gether to form benzene rings, new and repeat distances, utilizes speci- discourses of structural electronic forms emerge, imposing a mens in which the fibrous molecules biology, one must learn planar hexagonal structure on the nu- are randomly oriented about the fiber how to express what clear centers themselves. The benzene axis. Many details are averaged out molecule is frequently used in the sci- when fiber patterns are used to de- one means to say about ence of “complexity” to illustrate duce structural models. The impor- biological function in “emergent properties”, or the idea tance of these details did not become that the whole is greater than the sum clear until Richard Dickerson and this language. of its parts, in this case the aromatic- Horace Drew solved the crystal struc- ity of the de-localized electrons is the tures of several short DNA stretches emergent property. by x-ray crystallography. They found intermolecular interactions involved Polyani saw “DNA structure”, the sequence-dependent variation in the in “recognition” is self-evident, every- double helical pattern of hydrogen- structural parameters for DNA, sug- day words to communicate the se- bonded bases projecting inwards gesting that more subtle geometric as- quential steps in these processes are from a common phosphodiester back- pects of the structure might be in- becoming increasingly more difficult bone, as organizing atoms (P,O,N,C,H) volved in base pair recognition than to find (Kono and Sarai, 1999). from a lower, information-free, level just the matching of rigid surfaces of the hierarchy into a readable blue- (Dickerson and Drew, 1981). MICHAEL POLYANI AND LIFE’S print. Any other constellation of the As crystal structures for more and same atoms would be meaningless IRREDUCIBLE STRUCTURE more protein-DNA complexes were when presented to the transcription determined, it was clear that base pair We will argue here that the Gibbs free apparatus, a complex ensemble of recognition was based, not just on the energy concept, when applied to pro- protein molecules. Polyani stressed geometric possibilities of the helix- tein molecules, is the First Principle of the irreducibility of the information loop-helix motif, but on other motifs structural biology. However, to apply coded in the sequence of bases, the as well, and especially on the potential thermodynamic concepts, such as free fact that knowing everything about for DNA to undergo changes in struc- energy, to complex hierarchical sys- the quantum mechanical laws govern- ture (Anderson et al., 1987; Davies et tems, it is necessary to draw bound- ing atomic structure tells us nothing al., 2000). When the energy of defor- aries (physical or conceptual) be- about the content of the coded mes- mation enters into the balance of fac- tween levels of the hierarchy. Over sage or its origin. He wrote that: “mor- tors determining whether two sur- thirty years ago, just as the revolution phogenesis, the process by which the faces can bind, the search for a simple in structural biology was getting un- structure of living beings develops, structural recognition code (e.g., argi- derway, Michael Polyani (Polyani, can then be likened to the shaping of a nine side-chains recognize GC base 1969) wrote about the machinery of machine which will act as a boundary pairs on a rigid DNA molecule) had to life as having to satisfy “boundary for the laws of inanimate nature”. For be abandoned (Pabo and Neludova, conditions” in a multi-leveled hierar- Polyani, all working parts of a living 202 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE

Figure 2. Surface complexity. Proteins reflect a degree of complexity similar to the paintings of Kandinsky. The profilin:actin complex is shown on the left as a solid object color-coded for electric charge (blue is positive, from lysine and arginine amino acid sidechains, and red for the acidic carboxylate groups of aspartic acid, glutamic acid, and the carboxy termini of actin and profilin). The corrugated surface, pattern of electrostatic charges, hydrophobicity, and hydrogen-bonds of proteins comprise a unique interaction site for other proteins that have complementary surfaces. Kandinsky explored how the diverse elements of painting (color, line, and form) worked harmoniously to convey artistic meaning. Shown on the right is “Small Dream in Red, 1925), which appeared in the first edition of Kandinsky’s Point and Line to Plane (from Kandinsky 1866-1944 The Journey to Abstraction by Ulrike Becks-Malorny, Ko¨ ln: Benedikt Taschen Verlag GmbH (1994), p. 152. ISBN 3-8228-9045-6 Permission to publish obtained from the Kunst museum, Bern, Germany.).

system are under “dual control”, self. Yet, Polyani’s notion of boundary formation across cell membranes to whereby higher principles work on conditions in functional hierarchies is initiate intracellular signaling cas- principles at a lower, more fundamen- a fertile one, and represents the Sec- cades. tal level such as the laws of physics ond Principle of structural biology. and chemistry. Using written lan- Protein molecules can be regarded as guage as an example of a hierarchic PROTEINS AS “BOUNDARY system (letters-to-words-to-sentences- CONDITIONS” to-meaning), he argued that boundary conditions establish a semantic rela- The discovery of new The eukaryotic cell is the basic orga- tion between lower and higher levels. structural paradigms is nizing unit of higher life forms. Thou- The concept of “higher principles” akin to discoveries in sands of proteins on the surfaces of may sound old-fashioned to the mod- cells continuously sample their envi- ern mind, since we can now “see” (ow- the world of art because ronments, searching for molecular ing to the power of electron micros- the microscopic world cues on which directions might be fa- copy and crystallography) genes vorable for finding nutrients, or the packaged into chromatin, the basic never looks the same correct turn to take during develop- unit of structure of chromosomes, once they have ment. These results are passed and how they are transcribed and through the plasma membrane to spe- translated into proteins. Also, Polyani appeared. cialized protein structures, called “sig- was speaking about holistic concepts nal transduction complexes,” that such as “epigenetic landscapes” and shape the message that is transmitted “morphological fields”, which have to the nucleus. The concept of limiting been recast by modern developmental key elements of “Polyani machines” surfaces formed from protein mole- biologists in strictly molecular terms shaped to apply boundary conditions cules allows us to see signal transduc- (Edelman, 1988). For example, a sys- in the functional hierarchy of life. For tion as an “assembly process”, in tem of “semaphorin” molecules estab- example, cell-cell interactions are me- which the signal on the outside, per- lishes signposts for migrating neurons diated by protein-protein interactions haps a growth factor, induces the for- as the developing brain connects it- at the cell surface, which transfer in- mation of structures on the inside of FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 203

BOX I: The Laws of Thermodynamics

All processes in nature are subject to the Electrical of the system increases by a calculable first and second laws of thermodynamics. amount (see Denbigh, 1971 for a rigorous The first law, the equivalence of work and ⌿⌬q discussion of the perfect crystal). Vacan- energy and the conservation of their sum, is (movement of cies increase the heat capacity of the easy to apply once one learns how to quan- charge against an crystal lattice because they allow atoms to tify heat and to calculate the work done on electric potential) more easily hop about. A cornerstone of or by an isolated system (Klotz, 1967). The information theory is a statement that “in- second law, that the entropy of the universe Chemical gradients formation” can be quantified using Bolz- must increase in any spontaneous process, ͚␮ mann’s formula. The basic idea is that a idni is inherently impossible to apply because i system can be completely described in the whole universe is not accessible to mea- (movement of the form of answers to a series of yes or no surement. molecule against a questions expressed as a string of ones Willard Gibbs, in a remarkable scientific concentration and zeroes. Thus, entropy provides a treatise, showed how the concept of free gradient: measure of our ignorance of atomic posi- ␮ ϭ␮0 ϩ energy allows a neat solution to this prob- i i RT ln Ci; tions, quantifiable as an increase in en- lem. The first and second laws can be com- chemical potential tropy. bined into one formula expressed in terms of the ith species) Stretching a rubber band results in its of system variables only (i.e. “changes in the radiating heat in the amount -T⌬S be- rest of the universe” are not explicitly used) The “work” terms include all externally cause stretching reduces the number (Gibbs, 1906; Denbigh, 1971): wrought changes in the structure of a of possible ways that the cross-linked system, measured in terms of volume, polypropylene polymer chains can ar- ,H ؊ T⌬S surface area and length. The fixed at- range themselves is smaller (W ϭ 1⌬ ؍ G⌬ sys sys sys tributes of the system are the pressure, chain ␴ ␶ in the limit). The stretched system no ϩ work done on “system” P, surface tension, , and tension .It longer needs as much random energy to takes work, forces resisted over dis- remain in equilibrium with its surround- (Gibbs’ formula) tances, to bring about these changes. ings (Hill, 1960). Its Gibbs free energy is The re-ordering of matter and charge where, ⌬G ϭ G Ϫ G is the dif- higher because of the work done on it in sys final initial involved in these processes, can be stretching. Slowly releasing the rubber ference in free energy between initial driven by chemical gradients and reac- ⌬ ϭ Ϫ band allows it to lower its free energy as and final states, Hsys Hfinal Hinitial is tions. The free energy of the system will ␶⌬ the difference in internal energy (en- it performs recoverable work (- l) on be changed by the work done on it. This some external object. In the course of thalpy, at constant pressure), T is the will affect the heat capacity of the sys- ⌬ ϭ Ϫ shortening, the polymer chains take up temperature, Ssys Sfinal Sinitial is the tem. In the final analysis, this appears as difference in entropy of the ’system’ heat from their surroundings, increas- changes in bond energy in the molecules ing the total number of allowed con- alone, and the work done on the ’sys- comprising the system or in the intrinsic ϾϾ tem’ can be by a variety of external figurations (Wchain 1). The un- disorder. An internal combustion engine stretched rubber band in thermal agents. is a perfect example of system in which ⌬G Յ 0 expresses the notion that a equilibrium with its surroundings is most sys a complex chemical liquid is converted stable when the polymer chains are least process described by the initial and final to simple gases and heat in a device states of the “system” will not occur spon- structured! producing mechanical work. The total Rubber elasticity illustrates the idea taneously unless the free energy de- free energy available from the bond en- creases. When ⌬ϭ0 the ‘system’ is in that contracting polymers absorb heat at ergy in the molecules, expressed as constant temperature by increasing the equilibrium. Transferring material between chemical potentials, balances the net two states of equal Gibbs free energy, number of degrees of rotational freedom change in heat released and the work about their bonds. The heat capacity of though they may differ in structure, as with produced. two phases in equilibrium (liquid 7 vapor), crystals, on the other hand, has two as- Ludwig Boltzmann proposed a defini- pects: volume changes, which allow will not involve any net differences in en- tion of entropy that enables changes in thalpy or entropy from which useful work more degrees of freedom through va- molecular structure to be used in calculat- cancy creation, and increases in internal can be derived. ing free energy changes The “work” terms most commonly used energy of the molecules in the lattice. in biophysics (Katchalsky and Curran, S ϭ RlnW(Boltzmann’s formula) Proteins continuously interconvert en- 1964; Klotz, 1967; Morowitz, 1970) are thalpy and entropy in performing work in where R is the gas constant (the product cells, where the Gibbs free energy is of Boltzmann’s constant times Avogadro’s kept well above the minimum by external Mechanical number) and W is the number of ways in sources. Without structures such as pro- which the elements of a system can be teins and lipids that can perform the feat P⌬V͑volume change arranged (Hill, 1960. The field of statistical of changing degrees of internal order, in ‘system’ against a thermodynamics is concerned with a life would be impossible at ordinary tem- pressure P careful evaluation of W under all sorts of peratures. In fact, it appears to be gen- ␴⌬A͑change in surface conditions). The entropy of a perfect crys- erally true that already-folded proteins tal (W ϭ 1) of argon atoms, for exam- lose their ability to function at a temper- area at a surface tensions crystal Ά ple, is zero at a temperature of absolute ature below 220 Kelvin (Rasmussen et ␶⌬ ͑ l change in length at zero (we will ignore entropy associated al., 1992). a tension ␶ with the zero point energies of quantum mechanics). By introducing a small num- ber of vacancies or imperfections (by de- creasing the crystal density) the disorder 204 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE the plasma membrane. These protein All machines consume energy. Bio- ores, the core physical process driving complexes initiate and catalyze fur- logical cells are open thermodynamic economic development. ther change within the cell, such as systems, meaning that they can ex- increased protein synthesis, directed change energy and matter across their A PRIMER ON THE HISTORY OF cytoplasmic movements, or cell divi- boundaries. Much like the perennial PROTEIN STRUCTURE sion. waterfall beloved by philosophers of What is remarkable is that the sig- process, like Whitehead and Bergson, When asked what a protein molecule naling machinery comes together as cells can appear unchanging even as looks like, most of us respond with a the process proceeds. The chemically their contents remain continually in vigorous, gesticulating personification and topologically complex surfaces of flux. It was Willard Gibbs’ great dis- of the polypeptide chain coursing interacting proteins select the parts through space, describing coils and tinction to have discovered the laws of that form signaling complexes at the sheets, and generally folding back chemical equilibrium and the subtle cell surface. If a different factor is en- upon itself into a sweaty heap. The interplay of order and thermal energy countered at the surface, some other rigid appearance of helices and sheets combination of parts forms, tailored in chemical systems. The protein ma- gives a misleading picture of the to pass on the requisite signal to the chines operating in biological cells are forces stabilizing proteins and the ex- nucleus. Only one combination of best analyzed in terms of their con- tent of the motions these structures parts is selected from the multitude of sumption of Gibbs free energy, gener- exhibit. For an isolated system at possibilities, enabling a subtle blend- ally in the form of adenosine triphos- equilibrium, the Gibbs free energy is a ing and ordering of cellular responses phosphate (ATP). It is remarkable that minimum. This thermodynamic gem- to an incredible range of external hydrolysis of GTP by “G-proteins” is stone provides the currency for an threats and opportunities. At each integral to the activation of internal economical description of biophysical step in the assembly of these response response networks in response to ex- processes. In particular, it enables a complexes, protein surfaces must ternal signals processed by many precise definition of “structural stabil- match up in a precise way; unused transmembrane signaling complexes. ity” and some clue as to why Nature parts stand in readiness to react to A cost must be paid to send a signal. needs twenty different kinds of amino future or concurrent contingencies. This is a superb example of a “Polyani acid side-chains. The word “machinery ” conjures up machine” acting as a gate-keeper or The preoccupation with protein many images. After Newton formu- constituting a boundary condition, in “folds” is the legacy of the victory of lated his Laws of Motion, it became this case across the physical surface the polymer theory over the colloid fashionable to speak of the universe as separating outside from inside. theory in explaining the mystery of a clockwork in which all that has hap- A Third Principle of structural biol- denaturation, or the ease with which pened, or whatever will happen, is ogy emerges from this analysis, proteins can be induced to give up contained in solutions to differential namely that Gibbs free energy must their selective powers. Owing to the equations. This Laplacean world chal- be spent in sorting cellular proteins to pioneering work of Emil Fischer, lenged the concept of free will and produce differences across bound- Jacques Loeb, Linus Pauling, Fred made determinism a rude guest in aries. The life of a cell can be broken Sanger, Walter Kauzmann, Max Pe- moral dialogues. As a result, Life was down into the “economic” process of rutz, and others, the principles of pro- held to be separate from, and even in separating components, say viral shell tein structure are now mostly under- opposition to, the workings of the proteins and RNA, or the proteins in- stood from the point-of-view of the mechanized world. Living beings volved in passing signals across a chemist. Proteins are polymers with were seen as being built on a separate membrane, until they are later brought any one of 20 naturally occurring plan. Even as the nature of heat and together by transport mechanisms to amino acids spaced regularly along a its role in work-producing chemical “self-assemble” into functional com- backbone spine. The great diversity of reactions became clarified in the 19th plexes. The economist Georgescu- life on Earth is possible because even Century, it was hard to avoid vitalistic Roegen, in a book of astonishing the class of modestly sized polypep- notions of what made cells tick. Curi- intellectual virtuosity (Georgescu- tides having only100 amino acid side- ously, thermodynamics, the physics of Roegen, 1971), tried with some suc- chains contains 20 to the 100th power heat, held up acceptance of Darwin- cess to unify economics with the phys- different protein sequences. One of ism, because the Earth was thought to ical and biological sciences. Central to the mysteries of life is why so few be cooling at too high a rate to allow his thinking was the concept of “sort- protein sequences are found in the enough time for life to have devel- ing” and “recombining” material and biosphere compared with the astro- oped. Lord Kelvin’s famous resistance energy resources as the main driving nomically large number of possibili- to evolutionary ideas was gracefully force in economic life. He quite ties. Is it simply that only a few were dropped when he learned that radio- clearly understood the importance of selected for during evolution, or are active processes in the Earth’s core Gibbs free energy stored over hun- the rules governing how polypeptide could maintain a constant surface dreds of millions of years in petro- chains fold up into three-dimensional temperature for eons (Burchfield, leum and coal, and how humankind structures so restrictive that only a 1975). has used it to separate metals from small number of sequences qualify? FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 205

How did the modern idea of a pro- sitivity to metal ions and their salts thetic group, catalyze a reaction? The tein as a folded polypeptide chain hav- that apparently explanations could answer was provided by John ing a specific sequence of amino acid only be provided in terms of colloidal Northrup whose systematic crystalli- side-chains develop? John Edsall (Ed- behavior. Loeb showed that these ef- zation of enzymes demonstrated un- sall, 1962) has provided the perfect fects were simply consequences of the ambiguously that proteins are true jumping off point in an essay that inability of large charged proteins to molecules (Northrop et al., 1938) with traces the vicissitudes of the idea that cross semipermeable membranes, re- distinct masses and properties. This proteins are very large molecules. He sulting in a redistribution of ions proof rests on appreciating the Gibbs points out that, even though proteins across membranes separating the Phase Rule (Denbigh, 1971). Essen- were crystallized in the last century proteins from bulk solvents. Further- tially, Northrup showed that a satu- and were determined to have large more, he showed that all of the phe- rated protein solution was a true molecular weights, there was still a nomena, discussed in exquisite termi- phase in equilibrium with the solid great reluctance to consider them as nology long since abandoned, could crystalline phase. As more protein is distinct molecules. Indeed, scientific be quantitatively accounted for when added, it cannot dissolve and must philosophy continued to entertain se- the pH was adequately controlled contribute mass to the growing crys- rious doubts as to the ultimate reality since it is the controlling variable for tals. If it does dissolve, then it cannot of atoms (Pais, 1982), as held by the surface charge on the protein mole- be pure, and another chemical species schools of Ernst Mach and Pierre Du- cules. Ironically, it was a colloid must be present. In this way, it was hem (Duhem, 1914), much less of established that protein molecules molecules. alone, rather than bound impurities Nineteenth century biological chem- or groups, could be the sole source of ists considered the central mystery to To master structural the catalytic power found in protein be the process by which enzymes biology requires solutions. could lose their highly specific prop- The first rule of protein structure erties (“denaturation”) by very gentle familiarity with the was formulated by Linus Pauling, treatments. In Edsall’s analysis (Ed- history of its who predicted the existence of regular sall, 1962), the belief that proteins elements of structure (alpha helices, were colloids, self-aggregating masses development—much beta sheets, and gamma turns) from a of heterogeneous peptides, was the like within the world of consideration of the implications of prevalent one because it led to plausi- the planarity of the peptide bond. Dis- ble explanations for many of the prop- art—because the covered by Fischer at the turn of the erties of proteins. For example, the questions posed by the century, the peptide bond was shown variegated surfaces could provide in the 1930’s to have a planar struc- sites for catalysis and safe harbors for early pioneers still linger ture based on Bernal’s determination the rigid organic structures that car- and color the language of the structure of diketopiperazine. ried out the “real” chemistry. Denatur- we use to describe Planarity was explained in terms of ation was simply explained as the col- quantum resonance between the loid losing its grip on the active agent proteins. amide and carbonyl groups contrib- and reaggregating into a non-produc- uted by each amino acid along the tive substance. Organic chemists were chain. Maximal overlap of electronic very reluctant to accept the idea that chemist, The Svedberg who produced orbitals, related to bond strength, oc- long chain-like polymers could pro- the clearest evidence that proteins curs when the NH and CO atoms lie in vide the specificity exhibited by en- were molecules with distinct molecu- a plane. Pauling reasoned that planar zymes. Even the great Emil Fischer, lar weights. The first photographic peptide bonds stiffened the polypep- who discovered the peptide bond and plates from a centrifuge run on a so- tide backbone just enough to enable stable structures to form. Otherwise, established the chirality of the natural lution of hemoglobin, surprisingly, amino acids, felt that short polymers if there were three flexible bonds per showed sharp boundaries for hemo- no longer than twenty amino acids or amino acid, rather than the two globin rather than the broad ’smear’ so could by combination account for present when the peptide bond is that would have the distribution of the diversity of biological molecules fixed, hydrogen bonds between adja- smaller peptides making up the col- (Fruton, 1972). cent chains would be too weak to sta- Jacques Loeb’s direct attack on the loid. bilize a distinctly folded polymer colloid school (Loeb, 1922), resem- Yet, according to Edsall (Edsall, structure in the face of the universal bling in tone the fable of the “Emper- 1962), the majority of chemists tendency of matter to expand into the or’s New Clothes”, established that doubted whether polymers could space of all possibilities (the 2nd Law proteins are polyelectrolytes. The wild achieve the precision displayed by of Thermodynamics), in this case of variation in the colligative properties proteolytic enzymes in attacking their the polypeptide chain to take on all of protein solutions (e.g., osmotic substrates. The question became, rotational angles about its main-chain pressure), exhibited such a strong sen- could a protein alone, without a pros- bonds. 206 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE

The key to Pauling’s analysis was tion, muscle movements, nerve con- proteins capture our intuitive need to that each unit of the repeating back- duction) without local changes in predict or quantify function? Thermo- bone structure contains both a single temperature (Morowitz, 1970). There dynamics is often said to be irrelevant hydrogen bond donor and an accep- exists a theoretical apparatus of de- for analyzing molecular mechanisms tor. Since the most stable structures ceptive simplicity for analyzing and in detail since it involves relationships are those for which the greatest num- describing processes taking place un- between bulk properties. But this ber of bonds are formed, a repeating der isothermal conditions. These tools sidesteps the point that Gibbs free en- pattern stabilized by bonds formed are called irreversible thermodynam- ergy provides a “plausibility index” between different units along the ics (Katchalsky and Curran, 1965; Pri- that is useful in rooting out non-sen- chain naturally provides an acceptor gogine and Stengers, 1984) and the sical or purely mechanical models for for each donor. Thus, two side-by-side centerpiece is the Gibbs free energy. phenomena that manifestly require a polypeptide chains running in oppo- To venture into structural neurobiol- non-equilibrium treatment. Maybe, site directions can bond each other ogy or physiology without under- some day, free energy changes in- along their lengths as part of a beta- standing the idea that life is an adap- volved in storing and retrieving infor- sheet. It is a remarkable fact that these tation of free energy flow through mation from genes and in sorting cel- regular structures are possible inde- macromolecules is to invite confusion lular constituents, will be added to a pendently of the exact sequence and and disappointment (See Box I). free energy “cost of living index” and chemical nature of the amino acid Thermodynamics, the science of the study of life’s processes will have side-chains, which project out and what drives the process of change, an economic flavor. away from the hydrogen-bonded originated in a commonplace observa- An extreme post-modern structural- backbones. This indifference to spe- tion that was difficult to explain in ist might now say, without fear of rid- cific sequence is what allows three- terms of Newton’s laws: heat sponta- icule, that proteins do not ’fold’, a pe- dimensional protein structures having neously flows from a hotter to a cooler destrian view, but rather ’condense’ to quite different amino acid sequences object and not the other way (Denbigh, the maximally dense set of side-chain to be classified according to backbone 1971). The first crude steam engines clusters compatible with the se- chain topology. were constructed from simple ex- quence. As the condensation ensues, panding cylinders. When cooled by diffusion is not in the three-dimen- GIBBS FREE ENERGY: NATURE’S small children throwing buckets of sional space of the solvent, but rather water at it, a collapsing cylinder could in the one-dimensional space along ACCOUNTANT OF ’PROCESS’ be used to lift water (via a lever) from the polypeptide, which continually From the point-of-view of a physical flooded coal mines, providing access writhes upon itself while hydrophobic chemist, the most remarkable aspect to desperately needed stores of chem- editing of side-chains takes place (see of biology is that energy conversions, ical energy. The children, using water Box II). Protein factors, such as the principally the production of work from the mines to cool the cylinders, chaperonins (Pelham, 1986) might (mechanical, osmotic, or electrical) to get the coal, to heat the cylinder, work simply by holding onto the N- from chemical bond breakage, take were the first of the “Maxwell De- and C-terminal ends of the polypep- place under isothermal conditions. As mons.” Although a few deep thinkers, tide while supplying enough free en- argued below, there must be a con- notably the economist Jevons, wor- ergy to shake the self-reptating chain stant interchange of chemical bond ried about the supplies of coal run- enough to keep it out of local minima. energy and entropy within mole- ning out, most persons were happy This might explain why a major do- cules that can undergo order-disor- just to keep warm (Georgescu-Roe- main in chaperonins resemble actin der transitions. For example, the en- gen, 1971). molecules in having a strategically- ergy released when a bond forms Obviously, force production and placed hydrolyzable ATP molecule between a solvated metal ion and a other physiological processes do not that potentially can mediate changes protein can bring about changes in depend upon heat flows through tem- in protein conformation. the flexibility of some part of the perature gradients as in simple “heat Thus, a protein molecule is a unique polypeptide chain at the expense of engines.” Although muscle fibers do object for chemical study—a colloid changes in the ordered structure of lengthen and contract, they do so on a string—a fixed length polymer the solvation sphere. Thus, changes without large changes in volume. Fur- having a specific sequence of surface- in structure alter the heat capacity of thermore, muscle fibers utilize chem- active chemical groups that can aggre- a system, forcing it to radiate or ical energy from the hydrolysis of gate so as to find a minimum Gibbs absorb heat energy from the sur- ATP, the universal biological fuel, to free energy in an aqueous milieu. The roundings to maintain a fixed tem- perform mechanical work at constant unique colloidal collapse of the poly- perature. temperature. Rather than with vol- mer places catalytic residues and rec- Of course, there is a net exudation ume or temperature changes, it is cy- ognition sites at distinct regions on in the form of broken down molecules clic changes in the structure of muscle the surface. The early chemists who (waste) and randomized energy (heat) proteins that enable them to trans- struggled with the baffling questions but, within the organism, steady-state duce chemical free energy into work. posed by biological enzymes caught a processes take place (memory forma- Can a purely physical description of glimpse of the truth with their notion FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 207

Figure 3. Tension resolution. The polypeptide backbones of proteins are semi-flexible. Restricted rotations, termed phi and psi, about the two covalent bonds flanking the planar peptide bond are possible. The course of the polypeptide chain can be completely specified by the set of phi/psi angles for each protein. When these are plotted as points on a two-dimensional “Ramachandran Diagram” (shown on the left, with the values plotted for profilin and actin), it is seen that they cluster predominantly in three regions (highlighted in red). These regions characterize alpha helices (lower left), beta sheets (upper left), and reverse turns (upper right), which are folds that allow the chain to turn without clashes between side chains and backbone atoms. The allowed regions in the Ramachandran diagram thus represent a resolution of the tension between the freedom of the backbone chains to take on all rotational possibilities and the restrictions that the volumes of the side chains place on allowed folds. Mondrian believed that the intersections of lines to form planes, and the coloration of these enclosed areas, allowed only a limited number of aesthetic possibilities. He created paintings which he though best resolved these internal tensions in composition. Shown is “Tableau No. 111; Composition No. 14; Composition with Red, Black, Yellow, Blue, and Gray (1921)” from Piet Mondrian, 1872-1944 by Yve-Alain Bois, Joop Joosten, Angelica Zander Rudenstine and Hans Janssen. Boston: Bullfinch Press, 1994 ISBN 0-8212-2164-7. Permission to publish obtained from the Phillips Collection, Washington, D.C.

of achieving specificity through com- a part of what presents itself to the self”? How can it be described in bination, but they considered poly- mind as the writhing polypeptide terms of the laws of chemistry? The merization to be too uncontrolled, chain or the aperiodically packed function of the immune system is to yielding too heterogeneous a mixture, side-chains in the structural core of detect and take countermeasures to provide dedicated single macro- the protein. And how is the composi- against threats to the health of the molecules. The existence of molecular tion held as a piece? Its Gibbs free organism posed by viruses and bacte- mechanisms to control the ordering of energy is lower than some other ar- ria. How do cells “decide” on the basis amino acids in heteropolymers was rangement of parts. of molecular interactions to clear the apparently too speculative for any body of an antigen encountered in the nineteenth century thinker to imagine THE IMMUNE SYSTEM: “SELF- bloodstream? (Edsall, 1962; Fruton, 1972). The NON-SELF” AS A PROBLEM IN The surfaces of most cells contain “crystoidal” character of proteins, molecules of the cellular immune sys- ARCHITECTONICS arising from definite sequences of tem, such as the major histocompati- amino acids, is an endowment only Polyani’s analysis of biological com- bility complexes (MHCs), which grip understandable when viewed as the plexity leads us to examine the ques- between “Jaws-like” alpha helices product of a genetically encoded in- tion of boundaries between different small lengths of polypeptide chain formation system. levels in a hierarchy. In adopting par- (Stern and Wiley, 1994; Stern et al., But an artistic tension, precisely in ticular folds, protein molecules create 1994; Strominger and Wiley, 1995). the sense meant by Langer, is revealed unique surfaces out of which bound- These peptides are sampled from the in the composition (Figure 3). We can- aries in the living hierarchy are estab- collection arising from the ceaseless not look at a protein structure piece- lished. One of the most fascinating “grinding up” of a large fraction (per- wise, either as bits of structure, or as questions is how an individual animal haps up to 30%) of all newly synthe- resolved components of force. Water can distinguish its own proteins from sized proteins into short peptides by molecules, squeezed out in the pro- those of foreign invaders. Where is the maw-like structures called protea- cess, invisible to the eye, are as much boundary between “self” and “non- somes. MHC molecules, once loaded 208 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE with peptide, are transported to the lem of predicting three-dimensional plying thermodynamic principles (see surface of the cell where they are pre- structures of proteins. One aim of Box II) to validating proposed struc- sented to receptors on circulating T- physical chemistry is to predict all tures. It brings closer the day when a cells. T-killer cells are commissioned states of organization of matter from gene sequence alone can be used to to pass a life-or-death sentence on the a knowledge of the forces between produce a structural model, provided peptide-presenting cells, depending molecules (represented as derivatives that at least one representative of the upon the “meaning” carried by the of potential energies, discovered by structural class is known to atomic presented peptide. The “boundary” extrapolation from simpler systems). resolution. Atomic detail, particularly between an individual animal and its The achievement of these goals re- at domain/domain or subunit/subunit pathogenic invaders is thus not one quires an analytical procedure for cal- interfaces, will certainly require more contiguous surface, but rather the set culating the Gibbs free energy for var- structural data, but essential catalytic of all cellular surfaces presenting ious states of the polypeptide chain residues or recognition sites might MHC complexes. and the solvent molecules surround- be readily located in many instances. A key aspect of this “clonal selection ing it. Stated this way the problem is The main point is that substantial theory of immunology” is that each unlikely to be solved in terms of the progress will be made in our ability T-cell recognizes only one MHC-pep- atomic interactions of all molecules to predict three-dimensional struc- tide surface. T-cells bearing receptors involved (but see Kono et al., 2000). tures by classification well before the that could potentially bind MHC-pep- Why then is there a sense of excite- “protein folding problem” is solved. tide complexes derived from normally ment in the field of structure predic- Whewell may have anticipated our occurring cellular proteins are se- tion, and a feeling that “proteomics” point-of-view when he wrote “Classi- lected against during a prenatal will be able to shoulder its load after fication is the architectonic science, to “learning” phase. Before birth, any T- the genome appears? The reasons are which Crystallography and the Doc- cell that binds to an MHC-peptide several. trine of External Characters are sub- complex, presumably presenting only The first is that a useful rule-of- ordinate (Whewell, 1857 ).” “self” peptides, is killed. After birth, a thumb has gained wide currency and Thus, the reductionist approach is developmental “switch” is thrown that computer programs have been written running the risk of being bypassed by causes the activation of any surviving to apply it (Richmond, 1984; Eisen- taxonomy, a source of biological truth T-cell line that successfully binds an berg and McLachlan, 1986). The idea even before Linnaeus. Furthermore, MHC-peptide complex. Since there is that, since the structure of water is selected-site mutagenesis of proteins, are no “self” recognizing T-cells in cir- extremely sensitive to the properties combined with measurements of culation after the switch is thrown, of surfaces with which it is in contact changes in their heat capacity as they owing to clonal selection before birth, (water on wax), changes in solvent ac- fold is allowing a precise dissection of only “foreign” peptides, originating cessible surface area of aliphatic side- the free energy changes involved in from infecting viral or bacterial pro- chains during protein-folding should the folding-condensation process teins, are recognized in the context of be proportional to the decrease in (Matsumura et al., 1988; Fersht et al., an MHC surface. Notice the essential Gibbs free energy accompanying the 1992; Fersht, 1993) and during pro- role played by molecular “sorting” in formation of the hydrophobic core tein/protein association. Growth hor- establishing the boundary between (Chothia, 1984; See Box II). This sim- mone binding to its receptor provides “self” and “non-self”. Selection and ple relationship provides a means for an especially apt example (Clackson et transport of sampled peptides are cen- estimating Gibbs free energy changes, al., 1998; Bass et al., 1991). The idea is tral mechanisms governing the flow of without having to calculate the actual that the free energy of proteins in so- information from the cell interior to changes in the average positions and lution can be shifted by changes in pH its outer surface, a stunning example velocities of the solvent molecules in or the addition of denaturants, such of Georgescu-Roegen’s architectonics order to arrive at enthalpy and en- as urea and guanidine hydrochloride. of economic process. The 19th Cen- tropy values separately. The resulting changes in protein tury chemists even got part of it right A second path-breaking develop- structure can be monitored spectro- in recognizing that combinations of ment is the introduction of “tertiary scopically as equilibrium is ap- peptide fragments can generate the di- filters,” a set of computer programs proached. When temperature is versity needed to explain Life’s contin- that can be used to determine whether changed instead of chemical poten- gent responses. a given sequence of amino acids qual- tials, changes in heat capacity can be ifies the folded protein as a member of resolved into enthalpy and entropy THE PROTEIN CLASSIFICATION some family of structurally-homolo- changes. gous proteins (Bowie et al., 1991). Finally, our knowledge of the pro- PROBLEM These programs go beyond mere com- tein folding process is being put to the One of the great challenges to protein parisons of location of helices and test by direct design of novel proteins, science, especially now that the hu- sheets in the linear sequence of amino having amino acid sequences that man genome has become available, is acids and incorporate the specific con- have not been refined by natural selec- to apply these general thermodynamic text of each side-chain. This repre- tion (Hecht et al., 1990). A major con- and structural principles to the prob- sents an exciting step forward in ap- clusion from these studies is that it is FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 209

BOX II: “The hydrophobic effect:” looking at proteins with thermodynamic eyes

The concept of Gibbs free energy can be invoked to explain how the tradeoff between disorder, represented in the entropy of the backbone chain and the surrounding sol- vent molecules, and order, apparently con- ferred by hydrogen bonds, salt-bridges, and disulfide links is achieved. Since polar water molecules can interact with the hydrogen bond donors and acceptors along the back- bone of the unfolded chain, there is no net gain in stability in forming helices and sheets via hydrogen bonds as the chain folds up. There must be something else that explains how the polypeptide chain eludes the dic- tates of the 2nd Law of Thermodynamics. That something else is to be found in the chemical differences amongst the 20 amino acid side-chains, which can be charged, neutral, wax-like, polar, or be able to form covalent disulfide bonds. A folded protein molecule is a very im- probable state for a polymer, representing ϭ Figure 6. Close packing of hydrophobic sidechains in the interior of cofilin. Cofilin (also only one set of torsional angles (Wchain known as actin depolymerization factor—ADF) controls the dynamic turnover of actin 1) out of a vast multitude. Substituting subunits on one of the ends of the filament. Shown here are a set of conserved hydro- Boltzmann’s formula into Gibbs equation ⌬ ϭ phobic sidechains that form the core of ADF/cofilin. Thes sidechains are shaded light at equilibrium ( G 0), allows a neat ar- brown. Part of the ADF/cofilin polypeptide chain is shown as a silver wire. The requirement gument (Schulz and Schirmer, 1979) for a closely-packed core dictates the path of the polypeptide chain which must also showing that hydrogen-bonding is only obey the “law of the chain” to remain in allowed regions of the Ramachandran diagram marginally effective at stabilizing rather ␣ (Fig. 3). The particular fold of ADF/cofilin results in the placement of key sidechains on the long -helices against the pull of entropy surface involved in actin binding (shown in green). Atoms are color-coded: carbon- in a vacuum at a temperature of 300 kelvin ⌬⌯⌯Ϫ ϭ ⌬ brown, nitrogen-blue, and oxygen-red (Bowman et al., 2000). ( bonds T Schain). Unless the pep- tide bond is assumed planar, thereby freezing out one of the three torsional modynamics and polypeptide chain flex- vent water molecules. Thus, the origins degrees of freedom along the chain per ibility (Kauzmann, 1959). He showed that of protein stability lie, not necessarily in amino acid, the randomness of the dis- the most stable state of the polypeptide the strengths of internal chemical inter- ordered chain overwhelms any tendency chain in solution arises when the Gibbs actions, as in the structures of everyday to form stable helices or sheets; hydro- free energy of the complete system, in- life such as bridges, towers, and walls, gen bonds simply are not strong enough cluding the contributions arising from the but equally in the freedom of water mol- to overcome the entropy of the freely- entropy and changes in structure of the ecules to take on a larger number of ϾϾ jointed chain (Wchain 1). solvent water molecules, is a minimum. arrangements in the surrounding me- According to the quantum theory of ra- The essential breakthrough was to rec- dium. This is the basis for the second diation, the ’vacuum’ of outer space con- ognize that the proper “system” to study rule of protein structure, the “hydropho- tains radiation with associated degrees of was protein plus solvent. The chain bic effect”, or entropy stabilization. freedom (Dyson, 1954) and the disordered adopts a “fold” that, given the con- The dawn of the “crystallographically energy (“heat”) at any temperature is par- straints of the peptide bond, frees up the correct” synthesis of structure and ther- titioned between the vacuum radiation greatest number of water molecules to modynamics is heralded by the work of and the material structures with which it is roam freely in solution rather than having Frederic Richards at Yale, who asks in equilibrium. At lower temperatures, he- to adopt more restricted ’iceberg-like’ somewhat elliptically why Nature has se- lices could form in a vacuum because environments in the vicinity of aliphatic lected a particular set of hydrophobic Ͼ thermal radiation would predominate over side-chains. Therefore, WH2O O even amino acids. After all, the most austere chain entropy in the equitable distribution though Wchain 1, with the result that application of Kauzmann’s principle only ⌬ ϩ Ͼ of heat. Of course, the ordering of protein SH2O chain O and the folding pro- requires that a certain relative number of chains does not take place in a vacuum cess is spontaneous. fatty side-chains be buried within the and the folding process does not involve The apparent stability of a soluble compact “folded” structure. Richards radiating disordered chain energy into the protein cannot be explained unless the takes the analysis one step further by frozen reaches of space. Some other ve- entropic contribution of water molecules showing that for most proteins the num- hicle must transport heat (“disordered en- is included in Nature’s free energy led- ber of packing defects or voids is rather ergy”) away from the chain as it takes on a ger. Structurally this manifests itself as a small in the hydrophobic core (Lee and more ordered form (Figure 6). partial segregation of hydrophobic, or Richards, 1971). This implies that there Indeed, it was Walter Kauzmann’s wax-like hydrocarbon side-chains, into might be some relatively restricted set of celebrated paper on “the hydrophobic the interior of the protein molecule side-chain packing combinations that effect” that ushered in the modern syn- where they cannot, by virtue of their sur- are to be found in natural protein struc- thesis of the concepts of solution ther- face activity, induce local order in sol- tures (Ponder and Richards, 1987). 210 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE difficult to design sequences that away in the flux of radiation from the The paradigm for chemomechani- achieve the final degree of close-pack- sun. The photoreaction center is the cal process in biology is the “sliding ing in the hydrophobic core charac- “Polyani machine” that extracts Gibbs filament model of muscle contrac- teristic of natural proteins (DeGrado free energy from ultraviolet light. The tion,” in which cross-bridges project- et al., 1991). The protein engineers crystal structure of this protein ma- ing out of the myosin thick filaments have yet to master the intricate ten- chine is known and can be used to bind to actin thin filaments and pull sion between packing distinctively- explain how the absorption of pho- them towards the center of the sarco- shaped hydrophobic sidechains in the tons is converted into a flow of elec- meres, the basic units of contraction interior of the protein and bending of trons used in the synthesis of ATP (De- in muscle fibers (See Figure 4). Actin the polypeptide backbone within al- isenhofer et al., 1985). is generally thought to be an inert rod- lowed limits. like element in this process. The my- THE ORIGIN OF CELLULAR osin cross-bridges bind ATP as they TENSION THE FREE ENERGY BOUNDARY detach from actin and hydrolyze it in Viruses typify the value of the fac- the unattached state. Upon rebinding BETWEEN SUN AND EARTH tory metaphor for understanding cel- actin, the myosin head ’rotates’ The apparently spontaneous appear- lular economics; parts are gathered in through several binding sites on actin ance of order in the living world was staging areas and then brought to- of successively lower energy while for many centuries adduced as evi- gether for assembly processes driven stretching a molecular “spring” that dence for vitalistic forces, and later as by ordinary crystallization. The work then pulls on the actin filament (Hux- support for the argument that living is done in the sorting (see Box III). ley and Simmons, 1971). In this man- systems run counter to the 2nd law of However, something radically new ner, converting bond energy into elas- thermodynamics. This is no longer an happens when protein subunits can tic energy, it is believed that the free issue. To see why, it is useful to con- carry free energy with them into poly- energy of ATP hydrolysis is trans- sider the planet Earth as an “open sys- merizing structures. Tubulin-GTP duced into work. Myosin is often tem (Schro¨dinger, 1946; Penrose, monomers form microtubules. Actin- called a “motor molecule,” because 1989). The earth radiates energy in the ATP monomers form microfilaments. the macroscopic forces generated by form of radiant heat in an amount Both of these filament systems can muscle fibers could be explained as approximately equal to what it ab- generate forces if these subunits can the summed effect of hundreds of my- sorbs from the sun. It is in an ener- change shape in converting NTP to osin heads independently pulling on getic steady state. However, there is a NDP, where N can be adenosine (A) or each actin filament. That situation has difference in entropy between the ab- guanine (G). These forces result from changed very recently. New measure- sorbed light, peaking in the ultraviolet a conversion of Gibbs free energy ments on the extensibility of actin fil- region of the electromagnetic spec- from nucleotide hydrolysis into length aments, and reconsideration of the trum, and the heat radiated in the changes along some direction (say thermodynamics of muscle, (Baker et form of infrared photons. For a fixed along x) since f ϭ -dG/dx. The struc- al, 1999) have cast doubts on the va- amount of energy, each ultraviolet tural basis for length changes in actin lidity of the conventional cross-bridge photon carts a bigger share of the to- is the principle of subdomain rota- theory of contraction. Attention is be- tal than the more numerous infrared tions about hinge and shear points in ing increasingly focused on models photons streaming from the earth into the actin molecule (Page et al, 1998) that take into account the overall spa- the frigid void (Dyson, 1954). The following the structural paradigm of tial and temporal organization in number of ways (W: see Box I) of dis- viral capsid proteins (see Box III). muscle lattices, and the possibility of tributing a fixed amount of energy These forces can be rather large com- cooperativity amongst the myosin into packets is smaller for ultraviolet pared with Brownian forces, 100 pi- motors. photons than for infrared ones. Hence conewtons in the case of actin (Schutt We have analyzed the energetics of the entropy of the incoming light is and Lindberg, 1998). It is curious that muscle contraction in terms of a new, substantially less than the heat radi- microfilaments and microtubules are quite different, model (Schutt and ated by the earth. To a first approxi- considered by most experts in the field Lindberg, 1992; 1993; 1998; Kreatsou- mation, the increase in the entropy of of “motor proteins” to be merely las et al., 1999) in which actin is not a light itself drives all the living pro- “tracks” along which transporting mo- passive rope-like element but rather cesses on earth without a net warming tors such as myosin and kinesin move. an active force generator. Actin not of the planet. This argument is valid Since microtubules display highly un- only possesses a three-dimensional even though other processes such as usual dynamic motions in GTP-con- structure similar to other ATPases, natural radioactivity in the Earth’s taining solutions (Mitchison and such as hexokinase, but actually hy- crust and geothermal heating contrib- Kirschner, 1984; Desai and Mitchison, drolyses ATP during the assembly of ute to maintaining the earth’s temper- 1997) it is natural to wonder what actin filaments. In our model, actin ature. role they could play in more “engine- itself hydrolyzes ATP successively The “boundary” between the sun like” mechanisms, where kinesin subunit by subunit along actin fila- and the biosphere is the set of all pho- and tubulin work cooperatively to ments, each triggering phosphate re- tosynthetic proteins in plants toiling produce forces. lease into solution by its neighbor, as FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 211

BOX III: An example of a structural paradigm: self-assembly of simple viruses

Explaining the existence of symmetric, quasi-spherical) shells through the applica- port mechanisms) raises the free energy lo- highly ordered structures in biology (simple tion of icosahedral point group symmetry. cally, setting the stage for a return to equi- viruses, microtubules, muscle fibers) typifies This is an example of Polyani’s notion librium that appears as “self-assembly”. The the difficulty of applying the laws of thermo- that events at lower levels in a hierarchical driving force for the process, really just a dynamics to the living process. After all, it system, in this case the self-association of flow down a chemical gradient, is set up would seem that crystalline order, more proteins, entail functional “meaning” at with the original work expended in segregat- characteristic of the inorganic chemistry of higher levels that cannot be deduced from ing the viral components. inert materials, must represent an unlikely full knowledge of the lower level. There are a TMV is the paradigm for helically-sym- occurrence in the interior of cells, where number of functional advantages in having a metric particle assembly. Tomato bushy Brownian motion, the incessant random structure constructed from identical sub- stunt virus (TBSV) is the paradigm for spher- movement of all bits of matter at ordinary units (Caspar and Klug, 1962): (1) genetic ical viruses, which are based on icosahedral temperatures, is the dominant phenome- efficiency - a particle of the size and com- point group symmetry. Formally, each sub- non. There are two structural paradigms, position of TMV could never have its protein unit can be in an identical environment in an tobacco mosaic virus (TMV) and tomato shell coded for by the TMV-RNA without icosahedrally-symmetric shell only if it con- bushy stunt virus (TBSV), two unlikely- some element of repetition in the capsid, (2) tains exactly 60 subunits, three arranged sounding characters for the center stage, error editing - an improperly folded subunit, about each of the 20 three-fold axes. The that exemplify self-assembling structures. say one with a misincorporated amino acid, striking aspect of TBSV is that it contains TMV (Bloomer et al., 1978; Namba and not tailored to fit into the niche presented by 180 protein subunits, all with the same Stubbs, 1986) and TBSV (Harrison et al., the growing particle , would be rejected, and amino acid sequences, and therefore the 1978) have been solved at atomic resolution (3) self-assembly - no further protein scaf- subunits in the TBSV shell cannot all have by x-ray diffraction, and there are many les- folds or enzymatic steps are required. identical three-dimensional structures (Cas- sons to be learned from these simple iso- The simple picture of single subunits fit- par and Klug, 1962) and must fall into three metric viruses (Harrison, 1991; Perutz, 1992). ting into the growing helical ramp has been different ’quasi-equivalent’ classes. These Bernal deduced from the x-ray diffraction supplanted by a scheme involving a disk- deviations from strict symmetry have impor- patterns that viruses had multiple copies of like assembly intermediate (Butler and Klug, tant biological implications. The three single protein subunits arranged according 1978). TMV protein subunits first assemble classes of subunit structure found in the to some rule of regularity, not unlike those into disks which have the capacity to nucle- intact TBSV capsid derive from two features found in crystals. The analogy with crystal ate assembly by binding to a specific se- of the protein (Harrison, 1991). First, the N- growth eventually led to the concept of self- quence on TMV-RNA, providing selectivity, terminal 66 residues are highly charged pos- assembly; i.e., the ability of a large structure while at the same time overcoming the en- itively and are involved in binding RNA, while to spontaneously assemble from its compo- tropic barrier in getting the first few subunits the next 35 residues form “arms” that have nents, without the need for templates upon to form a helical nucleus (Namba and different degrees of packing order in the which to grow, or enzymes to direct bond Stubbs, 1986). Lowering the pH of a solution shell depending on the local environment. In formation. The subunits aggregate by non- of TMV subunits in the absence of RNA two of the three “quasi-equivalent” sub- covalent interactions (hydrophobic, for ex- results in the assembly of empty helical units, the arms are disordered. However, ample) by which the system minimizes its shells. Caspar (1963) observed that a pair of every third subunit has an ordered arm that, Gibbs free energy (Janin and Chothia, carboxyl groups exhibited abnormal pKs with its 60 symmetry-related partners, knits 1990). Thus, in the case of simple viruses, (near 7.0) in the intact virus, but behaved the whole particle together. Secondly, there the design for the whole particle is specified normally for the disks (near 4.5). Since the is a hinge between the two major domains by the direction and strength of ’bonds’ be- carboxyl groups are charged at neutral pH of the TBSV subunits that allows them to tween subunits; the design of the whole is in the disk, he postulated that a ’negative take up different positions in the viral capsid contained within the parts, an idea advo- electrostatic switch’ prevents “lockwasher- without relative slippage in the subunit-sub- cated by Buckminster Fuller for the blue- ing” into helices. However, when RNA is unit bonds. This design feature, maintained print-free assembly of geodesic domes added to the protein solution, the binding contacts between variably-linked domains, (Caspar and Klug, 1962). energy of RNA to the disk overcomes the confers flexibility to the viral shell under con- A system composed of identical subunits electrostatic repulsion, forcing the carboxyl ditions mimicking those in the host cell. can most easily minimize its Gibbs free en- groups into environments in the helical virus The expansion of the shell is thought to ergy by forming symmetric structures. This that resemble the abnormally-titrating car- be important during the assembly and un- can be understood by considering the sim- boxylate groups in maleic acid (Namba and packaging of the virus (Harrison, 1991). The ple case of linear polymerization where each Stubbs, 1986). It is clear that the TMV as- design of TBSV, like TMV, incorporates con- subunit associates end-to-end with two oth- sembly process in solution can be fully un- trolled polypeptide flexibility in two different ers. If the directions of the ’bonds’ between derstood in terms of the system flowing ways, variable terminal arms and internal these subunits are changed, the linear down gradients of Gibbs free energy. The hinges, that enable proteins to encapsulate structure can spiral outwards or inwards de- free energy involved in melting stems and RNA as they move down gradients of Gibbs pending on the bond angles. For a particular loops of the free RNA as it becomes pack- free energy. The 2nd law of thermodynamics set of angles, the subunits will fold into a aged in the virus includes the role of water, is not violated when all factors are taken into helix generating another class of stabilizing bound and free, in the process. The sub- account. These include the work required to contacts between turns; i.e., each subunit units in the disk assembly are less tightly- keep the components separated before as- now has four or six neighbors in the helix. packed compared with the helix and have sembly, the heat lost during the initial ATP Any distortion in the bond angles that de- more flexible side-chains exposed to sol- hydrolysis involved in establishing gradi- stroys the equivalence of the subunits, that vent. They are entropy stabilized. When ents, and the change of entropy of the sys- is the helical symmetry, necessarily leads to RNA becomes encapsulated, the formation tem (virusϩsolvent) as the particle anneals. a less stable particle since the solvent ac- of stronger, better-packed, inter-subunit More complicated viruses, such as poliovi- cessible area increases. Therefore, greater bonds in the (more symmetrical) helical rus, use proteolytic cleavage to “trigger” particle stability is a direct consequence of state lowers the free energy of the particle. transitions in viral capsids during the assem- having a subunit capable of forming sym- In vivo, the orchestrated process of separat- bly process, and other elaborate strategies, metric structures. This argument can be ex- ing proteins and nucleic acids into different but are understandable “mechanisms” in tended to the case of closed isometric (or cellular compartments (via vesicular trans- the context of free energy flow. 212 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE

to drive enzymes through their catalytic cycles. In the standard theories of en- zyme kinetics, it is assumed that bind- ing to the substrate helps to lift the free energy of the enzyme into an ac- tivated state and that random Brown- ian motion ’kicks the protein’ with suf- ficient force to bring about the required transitions. Perhaps the free energy of activation (a form of work) can be supplied by attached actin fil- aments undergoing force-producing length changes. In effect, the actin- rich cell cortex is a “tension bath” analogous to a thermodynamic “heat bath,” but capable of directed vecto- rial action that enables the enzyme to

The actin-rich cell cortex is a “tension bath” analogous to a thermodynamic “heat bath,” but capable of directed vectoral action that enables the enzyme to not only respond more quickly, but to be synchronized to other enzymes elsewhere in the cell and drawing free energy from the same source.

Figure 4. Muscle as a hierarchical system. Muscle cells contain hierarchical arrays of fibers. The basic functional unit of muscle is the sarcomere, which consists of interdigitating arrays not only respond more quickly, but to of actin and myosin filaments. When a fiber contracts, each of the thousands of sarcomeres be synchronized to other enzymes arranged end-to-end contract in unison. In this way, the ten micron per second sarcomeric contractions are amplified into meter per second shortenings of the whole fiber. At the elsewhere in the cell and drawing free molecular level, the actin filaments move along the myosin filaments, powered by the energy from the same source. hydrolysis of ATP. The background image of Leonardo Da Vinci reminds us of mankind’s Eukaryotic cells are notable for the centuries-long search for the secret of animal locomotion. Figure originally appeared in broad range of motile activities in Kreatsoulas et al. (1999). which they can engage over the course of their lifetimes. Cell division is cer- tainly the most balletic but on the lo- cal scale membrane ruffling and mi- it moves relative to myosin. The role ling the kinetics of biophysical pro- crospike movements probably qualify of the well-known actin-activated my- cesses, “Polyani machines” operating as the most violent. Neurons and im- osin ATPase is to catalyze the uptake at boundaries in the time domain. The mune cells are especially rich in mo- of ATP on actin and to provide trac- implications of this new model for tile activity (Smith, 1988; Davenport tion points for the crawling actin studying actin-based movements are et al., 1993; Matus, 1999). Neurotrans- filaments. This suggests a new “gate- many, but one idea is that actin sub- mitters, synthesized in the cell soma keeping” role for ATPases in control- units generate forces that can be used of motor neurons, are transported to FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 213

CONCLUSION Structural Biology is a system of thought. It is not a simple system. It is an architectonic science because it is concerned with how living systems are built up. To master it requires fa- miliarity with the history of its devel- opment—much like within the world of art—because the questions posed by the early pioneers still linger and color the language we use to describe proteins. Statistical thermodynamics is essential in order to understand how structure can be sustained in a micro-world where Brownian fluctua- tions are the dominant processes. Ul- timately, if we are to understand how thoughts and beliefs can be fixed in our minds, we must look to proteins as the material substrates for mental Figure 5. Mind as a problem in architectonics. Neurons (shown on the bottom left) connect processes and devise theories that rec- up with each other along their dendrites (shown enlarged on bottom right). The small, ognize their special character as hier- micron-sized, variously-shaped projections from the dendrites are called spines. Spines are the sites of contact (synapses) with the axons of other neurons. Synaptic plasticty or archical structures under tension. changes in synaptic strength are believed to form the basis for learning and memory. At the Change is possible because we live in molecular level, these changes in spine shape are mediated by the polymerization of actin, a flux of Gibbs free energy, generated a universal “architectonic element” (the backbone representation of an actin subunit is in the sun, captured by plants, and shown on the upper right). Actin forms helical filaments (shown at low resolution in the parceled out as fuel to drive the sort- center. Immanuel Kant revolutionized philosophy with his views on how the structure of our minds filters the stream of sensory information in categorizing external reality (his multiple ing and motile activities of cells. images form a tableau for offering a structural biological view of the mind’s hierarchy). Architectonics began with Kant (1787) and his preoccupation with the problem of how our minds devise the- ories about the external world from neuromuscular synapses along the Actin-binding proteins, such as sense-given information; the structure microtubules in axons. The vast ma- members of the formin and dystro- of our minds limits what we can know jority of all excitatory synapses in cor- phin families, link membrane-bound of the world (Figure 5). In Kant’s tical neurons are located at the tips of enzymes and channels with the actin words: “By the term Architectonic I actin-rich structures called “dendritic microfilament system (Lindberg and mean the art of constructing a system. spines” which are densely distributed Schutt, 1999). Actin’s role may ex- Without systematic unity, our knowl- along the dendritic arbors of pyrami- tend beyond merely providing base- edge cannot become science; it will be dal cells (Harris and Kater, 1994) (See ment support. Microfilament tension an aggregate, and not a system. Thus Figure 5, lower right). Spines, which transmitted directly to ion channels Architectonic is the doctrine of the scientific in cognition, and therefore are 1-2 microns in length, change in the plasma membrane could shape on a second by second basis as necessarily forms part of our method- affect the magnitude of the electro- they adjust to change the strength of ology.” chemical resting voltages that gov- their synaptic junctions with neigh- Susanne Langer’s great mentor, ern the propagation of action poten- boring neurons (Fischer et al., 1998). Ernst Cassirer, carried Kant’s pro- This process lies at the heart of learn- tials in the nervous system. Actin gram forward by seeking the universal ing and memory as first outlined by binding proteins are important in metaphorical structures behind lan- E. O. Hebb in 1951. The precise mech- mediating vesicle transport of neu- guage, arts, and mathematics. Al- anisms underlying “synaptic strength” rotransmitters in pre-synaptic ter- though he could never anticipate what have not been established, although mini. Is it really too fanciful to won- kinds of physical structures in the polymerization and cross-linking of der how mechanical, membrane brain could filter sensory data into actin filaments are certainly factors, electrical potential, and chemical categories, Cassirer (1957) may have and many neurotransmitter receptors neurotransmission, might mutually captured the essence of modern think- are physically linked to actin via spe- shift the overall free energy balance ing about how neurons construct the cialized proteins, such as gephyrin, in sheet in distributed networks of neu- mind. He wrote in concluding his the case of GABA receptors (Lindberg rons? Hebbian memory must after magnum opus: “We have so far tried and Schutt, 1999). all be Gibbsian! to show how the individual symbolic 214 THE ANATOMICAL RECORD (NEW ANAT.) FEATURE ARTICLE forms—language, myth, theoretical Figure 5). The boundary is ever- Burchfield JD. 1975. Lord Kelvin and the knowledge—are aspects in the struc- changing as we encounter new expe- Age of the Earth. Univ. of Chicago Press. ture of the intelligent organization of riences, sift through our memories, Chicago, IL. Butler PJG, Klug A. 1978. The assembly of reality. Each of them presented us form new concepts. In the architec- a virus. Sci Am 239:62-69. with an independent, architectonic tonic view, our minds are written on Caspar DLD. 1963. Assembly and stability principle, an ideal ”structure“, or, bet- this surface by microscopic dendritic of the tobacco mosaic virus particle. Adv ter, — since we are here never dealing spines, small actin-rich “Polyani ma- in Protein Chem 18:37-121. Caspar DLD, Klug A. 1962. Physical prin- with describing purely static relation- chines” reaching out to connect with ciples in the construction of regular vi- ships, but rather with exposing dy- axons bearing gifts of precious neuro- ruses. Cold Spring Harbor Symp. Quant namic processes—a characteristic transmitters, signaling changes in the Biol 27:1-27. way of ”structuring“ itself. . .”. Cassir- external world. Actin is thus a pro- Cassirer E. 1957. The Philosophy of Sym- er’s language suits our present pur- totypical architectonic element: struc- bolic Forms. Yale University Press. New Haven. Volume 4. 50p. poses because we are concerned with turing and shaping the synaptic sur- Chothia C. 1984. Principles that determine “structuring” itself, but as it occurs at faces between communicating neu- the structure of proteins. Annu Rev Bio- the surfaces of communicating cells. rons, building new worlds from our chem 53:537-572. Applied to neurons, the dynamic present experiences and past memo- Churchland PM. 1984. Matter and Con- changes at the tips of dendritic spines ries. sciousness. MIT Press. Cambridge, MA. Clackson T, Ultsch MH, Wells JA, de Vos are brought about by changes in the AM. 1998. Structural and functional architecture of the underlying actin analysis of the 1:1 growth hormone: filaments. ACKNOWLEDGMENTS receptor complex reveals the molecular There is a strain of modern philos- basis for receptor affinity. J Mol Biol. This work was supported by grants 277:1111-28. ophy that continues the tradition of from The Nancy Lurie Marks Family Davenport RW, Dou P, Rehder V, Kater SB. Cassirer. It is the “hermeneutical Foundation (to CES), the National In- 1993. A sensory role for neuronal growth school” whose central premise is that stitutes of Health (GM44038 to CES), cone filopodia. Nature 361:721-724. our minds build worlds out of histor- the Swedish Cancer Foundation (to Davies DR, Goryshin IY, Reznikoff, Ray- ically learned knowledge and fresh ment I. 2000. Three-dimensional structure UL), and the Swedish Natural Science of the Tn5 synaptic complex transposition sensory experiences. Rather than the Research Council. Figure Credits: Fig- intermediate. Science 289:77–85. fixed categories of cognition imagined ures 1 and 2 were prepared by Con- DeGrado WF, Raleigh DP, Handel T. 1991. by Kant, these philosophers (Gada- stantine Kreatsoulas and originally De novo protein design: what are we learn- ing? Curr Opin Struct Biol 1:984-993. mer, 1976; Goodman, 1978) posit an appeared in Kreatsoulas et al. (1999). infinite plenitude of frames of refer- Denbigh K. 1971. The principles of chem- Figures 3,4,6 were prepared by Greg- ical equilibrium. Cambridge University ence depending on the conversations ory D. Bowman. Figure 5 was pro- Press, Cambridge. in which one engages. To understand duced by Alex Plakantonakis. Deisenhofer J, Epp O, Miki K, Huber R, Structural Biology requires a close Michel H. 1984. The structure of the study of the theoretical problems that photoreaction center from rhodopseudo- beset its founders, a kind of dialogue monas viridis. Nature 318:618-624. LITERATURE CITED Desai A Mitchison TJ. 1997. Microtubule with the past reflected in terms like polymerization dynamics. Annu Rev Cell “denaturation”, “folding”, “self- Anderson JE, Ptashne M, Harrison SC. 1987. Structure of the repressor-opera- Dev Biol 13:83-117. Review. assembly”, “lock-and-key”, and “sort- tor complex of bacteriophage 434. Na- Dickerson RE, Drew HR. 1981. Kinematic ing.” ture 326:846–852. model for B-DNA. Proc Natl Acad Sci Perhaps the highest level of the Ban N, Nissen P, Hansen J, Moore PB, Steitz USA 78:7318–7322. TA. 2000. The complete structure of the Duhem P. 1914. The aim and structure of functional hierarchy is the brain itself physical theory. Atheneum (Second edi- and the place where theory-building large ribosomal subunit at 2.4 Angstroms resolution. Science 289(5841):905–920. tion reprinted by Princeton University takes place. The modern cellular the- Bass SH, Mulkerrin MG, Wells JA. 1991. A Press, 1954). ory holds that the combinatorial com- systematic mutational analysis of hor- Dyson FJ. 1954. What is heat? Sci Am 58- plexity inherent in networks of bil- mone-binding determinants in the hu- 63. Edelman GM. 1987. Neural Darwinism. lions of neurons connected by man growth hormone receptor. Proc Natl Acad SciUSA.May15;88:4498- Basic Books. New York. thousands of synapses is enough to 4502. Edelman GM. 1988. Topobiology: an intro- sort the received sensory streams Bloomer AC, Champness JN, Bricogne G, duction to molecular embryology. Basic from the outside world into learned Staden R, and Klug A. 1978. Protein disk Books. New York. and remembered categories. Whether of tobacco mosaic virus at 2.8 Å resolu- Edsall JT. 1962. Proteins as macromole- cules: an essay on the development of the one believes that the synapses are tion showing interactions within and be- tween subunits. Nature 276:362-368. macromolecular concept and some of its changeable nodes in a network that Bowie JU, Luthy R, Eisenberg D. 1991. A vicissitudes. Arch Biochem, Suppl. 1:12- encodes concepts (Churchland, 1984), method to identify proteins that fold into 20. or link up neurons into groups, as in a known 3-dimensional structure. Sci- Eisenberg D, McLachlan AD. 1986. Solva- Edelman’s theory of neuronal selec- ence 253:164-170. tion energy in protein folding and bind- Bowman GD, Nodelman IM, Hong Y, Chua ing. Nature 319:199-203. tion (Edelman, 1987), the boundary N-H, Lindberg U, Schutt CE. 2000. A Fersht AR. 1993. Protein folding and sta- between our outer and inner worlds is comparative structural analysis of the bility: the pathway of folding of barnase. the set of all synaptic surfaces (see ADF/cofilin family. Proteins 41:374–384. FEBS Lett 325:1, 2, 5-16. FEATURE ARTICLE THE ANATOMICAL RECORD (NEW ANAT.) 215

Fersht AR, Matouschek A, Serrano L. 1992. the Myb DNA-binding domain utilizing (1969): essays by Michael Polyani. The folding of an enzyme . 1. Theory of free energy calculations. Proteins Grene, M. ed. Routledge & Kegan Paul protein engineering analysis of stability 38:197–209. Ltd. London (1969). and pathway of protein folding. J Mol Kreatsoulas C, Lindberg U, Schutt CE. Ponder JW, Richards FM. 1987. Tertiary Biol 224:771-782. 1999. “Nouveau Regard sur le Muscle“. templates for proteins: use of packing Fischer M, Kaech S, Knutti D, Matus A. Les Cahiers de Science & Vie. 53:78-87. criteria in the enumeration of allowed 1998. Rapid actin-based plasticity in Langer SK. 1967. Mind: an essay on feel- sequences for different structural dendritic spines. Neuron 20:847-854. ing. Volume 1. p. 93 and 155. classes. J Mol Biol 193:775-791. Fruton JS. 1972. Molecules and life: histor- Lee B, Richards FM. 1971. The interpreta- Prigogine I, Stengers I. 1984. Order out of ical essays on the interplay of chemistry tion of protein structures: estimation of chaos: man’s new dialogue with nature. and biology. Wiley-Interscience, New static accessibility. J Mol Biol 55:379-400. Bantam, New York. York. Lindberg U, Schutt CE. 1999. “Actin-bind- Rasmussen BF, Stock AM, Ringe D, Petsko Gadamer HG. 1976. Philosophical Herme- ing Proteins”. Encyclopedia of Molecu- G. 1992. Crystalline ribonuclease A loses neutics. Univ. of California Press. Berke- lar Biology. (T. Creighton, Ed). John function below the dynamical transition ley, CA. Wiley & Sons, Inc. p.32-48 . at 220 k. Nature 357:423-424. Georgescu-Roegen N. 1971. The entropy Loeb J. 1922. Proteins and the Theory of Richmond TJ. 1984. Solvent accessible sur- law and the economic process. Harvard Colloidal Behavior. McGraw-Hill, New face area and excluded volume in proteins: University Press, Cambridge, MA. York. analytical equations for overlapping Gibbs JW. 1906. The Scientific Papers of J. Lu X-J, Olson W. 1999. Resolving the dis- spheres and implications for the hydro- Willard Gibbs (Longmans, Green C.; re- crepancies among nucleic acid conforma- phobic effect. J Mol Biol 178:63-89. printed as Dover paperback, New York, tional analyses. J Mol Biol 285:1563–1575. Schro¨dinger E. 1946. What is life? Re- 1961). Luger K, Mader AW, Richmond RK, Sargent printed by Cambridge University Press, Goodman N. 1978. Ways of Worldmaking. DF, Richmond TJ. 1997. Crystal structure Cambridge 1967. Hacket Publishing Company. Indianap- of the nucleosome core particle at 2.8 A Schulz GE, Schirmer RH. 1979. Principles olis, IN. resolution. Nature 389:251-260 of Protein Structure. Springer-Verlag Harris KM, Kater SB. 1994. Dendritic Matsumura M, Becktel WJ, Matthews BW. (New York Berlin Heidelberg Tokyo). spines: cellular specializations imparting 1988. Hydrophobic stabilization in T4 Schutt CE, Lindberg U. 1991. Scratching the both stability and flexibility to synaptic lysozyme determined directly by multi- back of the cortex. Curr Biol 4:257-258. function. Annu Rev Neurosci 17:341-371. ple substitutions of ile 3. Nature 334: Schutt CE, Lindberg U. 1992. Actin as the Harrison SC. 1991. What do viruses look 406-410. generator of tension during muscle con- like? Harvey Lectures 87:127-152. Matus A. 1999. Postsynaptic actin and neu- traction. Proc Natl Acad Sci USA 89:319- Harrison SC, Olson AJ, Schutt CE, Winkler ronal plasticity. Curr Opin Neurobiol 323. FK, Bricogne G. 1978. Tomato bushy 9:561-565. Schutt CE, Lindberg U. 1993. A new per- stunt virus at 2.9 Å resolution. Nature Mitchison T, Kirschner M. 1984 Dynamic spective on muscle contraction. FEBS 276:368-373. instability of microtubule growth. Na- Lett 325:59-62. Hecht MH, Richardson JS, Richardson ture 312:237-42. Schutt CE, Lindberg U. 1998. Muscle con- DC, Ogden RC. 1990. De novo design, Morowitz HJ. 1970. Entropy for biologists: traction as a Markov process: energetics expression, and characterization of Fe- an introduction to thermodynamics. Ac- of the process. Acta Physiol Scand 163: lix: a four-helix bundle protein of native- ademic Press, New York. 307-323. like sequence. Science 249:884-891. Namba K, Stubbs G. 1986. Structure of Sigler PB, Xu Z, Rye HS, Burston SG, Fen- Hill TL. 1960. An introduction to statistical tobacco mosaic virus at 3.6 Å resolution: ton WA, Horwich AL. 1998. Structure thermodynamics. Addison-Wesley, Read- implications for assembly. Science 231: and function in GroEL-mediated protein ing, Mass. 1401-1406. folding. Annu Rev Biochem 67:581–608. Huxley AF, Simmons R. 1971. Proposed Northrup JH, Kunitz M, Herriott RM. Smith SJ. 1988. Neuronal cytomechanics: mechanism of force generation in stri- 1938. Crystalline Enzymes. Columbia the actin-based motility of growth cones. ated muscle. Nature 233:533-538. University Press, New York. Science 242:708-715. Janin J, Chothia C. 1990. The structure of Pabo CO, Nekludova L. 2000. Geometric Stern LJ, Wiley DC. 1994. Antigenic peptide protein-protein recognition sites. J Biol analysis and comparison of protein-DNA binding by class I and class II histocom- Chem 265:16027-16030. interfaces: why is there no simple code patibility proteins. Structure 2:245-251. Kant I. 1787. Critique of Pure Reason. Ev- for recognition? J Mol Biol 301:597–624. Stern LJ, Brown JH, Jardetzky TS, Gorga eryman’s Library. J.M. Dent & Sons Ltd Page R, Lindberg U, Schutt CE. 1998. Do- JC, Urban RG, Strominger JL, Wiley DC. London (reprinted in 1964, 471p). main motions in actin. J Mol Biol 280: 1994. Crystal structure of the human Katchalsky A, Curran PA. 1965. Nonequi- 463-474. class II MHC protein HLA-DR1 com- librium thermodynamics in biophysics. Pais A. 1982. “Subtle is the Lord”: the sci- plexed with an influenza virus peptide. Harvard University Press, Cambridge, ence and the life of Albert Einstein. Ox- Nature 368:215-221. Mass. ford University Press, Oxford. Strominger JL, Wiley DC. 1995. The class Kauzmann W. 1959. Some factors in the Pauling L, Corey RB, Branson HR. 1951. I and class II proteins of the human interpretation of protein denaturation. The structure of proteins: two hydrogen- major histocompatibility complex. The Protein Chem 14:1-63. bonded helical configurations of the 1995 Albert Lasker Medical Research Kirschner M, Mitchison T. 1986. Beyond polypeptide chain. Proc Natl Acad Sci Award. JAMA 274:1074-6. self-assembly: from microtubules to USA 37:205-211. Stryer L. 1988. Biochemistry, 3rd ed. W. H. morphogenesis. Cell 45:329-342 Pelham HRB. 1986. Speculation on the Freeman, New York. Klotz IM. 1967. Energy changes in bio- functions of the major heat shock pro- Whewell W. 1847. The Philosophy of the chemical reactions. Academic Press Inc., teins and glucose-regulated proteins. Inductive Sciences. 2nd. Ed. Johnson re- New York. Cell 46:959-961. print Corp. New York. Kono H, Sarai A. 1999. Structure-based Perutz MF. 1992. Protein structure: new Wimberly BT, Broderson DE, Clemons Jr, prediction of DNA target sites by regula- approaches to disease and therapy. WM, Morgan-Warren RJ, Carter AP, Von- tory proteins. Proteins 35:114–131. W. H. Freeman & Co., New York. rheim C, Hartsch T, Ramakrishnan V. Kono H, Saito M, Sarai A. 2000. Stability Polyani M. 1968. Life’s Irreducible Struc- 2000. Structure of the 30S ribosomal sub- analysis of the cavity-filling mutations of ture. Chapt. 14 Knowing and Being unit. Nature 407:327–339.