* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Protein Molecules in Solution
Genetic code wikipedia , lookup
Paracrine signalling wikipedia , lookup
Drug design wikipedia , lookup
Biosynthesis wikipedia , lookup
Gene expression wikipedia , lookup
Point mutation wikipedia , lookup
Photosynthetic reaction centre wikipedia , lookup
Ancestral sequence reconstruction wikipedia , lookup
Expression vector wikipedia , lookup
Multi-state modeling of biomolecules wikipedia , lookup
G protein–coupled receptor wikipedia , lookup
Magnesium transporter wikipedia , lookup
Signal transduction wikipedia , lookup
Evolution of metal ions in biological systems wikipedia , lookup
Bimolecular fluorescence complementation wikipedia , lookup
Size-exclusion chromatography wikipedia , lookup
Protein structure prediction wikipedia , lookup
Interactome wikipedia , lookup
Protein purification wikipedia , lookup
Western blot wikipedia , lookup
Biochemistry wikipedia , lookup
Metalloprotein wikipedia , lookup
Two-hybrid screening wikipedia , lookup
Protein Molecules in Solution By IRVING M. KLOTZ, PH.D. A protein imolecule in solution miay undergo a host of interactions: with solvent, with small molecules and ions, with itself, with other macromolecules. In a physiologic milieu, all of these possibilities are pre-sent. A variety of interactions will occur. On the basis of the behavior of these substances in simiiple svstemiis, reasonable guesses caan often be mlade as to which interactions will predomiiinate in a biologic environmilent. Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 IN THIS paper, we shall examinie soime of the properties, and the behavior, of proteins in solution. Our view will be a maolecular one and we shall interpret the behavior of proteins in terms of the structure of these milacromolecules anid their interactions with their enviroiinment. It is appropriate, therefore, to summarize briefly at the outset, some of the salient features of the structure of proteini molecules. + H HN H C-OH O Brief Review of Protein Structure 3.4A CH-CH -- H~~ H~~ H% / Proteinis are supermolecules made up of small amiino acid building blocks. The chains of amino acids are held together by peptide linkages, each residue contributing about 3.5 A to the length of the chain (fig. 1). As has long been recognized, specificity of proteins must reside ultimately in the particular sequenee of amlino acids in the chains of the supermnolecule. As a result of the magnificent skills of several individuals, sequences (sometimes called the primary structutre) are Inow known completely for 2 proteins, insulin1 and ribonuclease,' 3 and almost completely for several others. The arrangement in insulin is shown in figure 2. Overwhelmiing evidenlee exists that withini a protein molecule the polypeptide chain does not occur as a stretched-out string of residues. While the basic configurations assumed by the chains are not yet established unequivocally, and certainly must differ among different classes of protein, the best model so far is the a-helix coil suggested by Pauling, Corey, and Branson4 (fig. 3). With such a coiling, it takes approximately 7 amino acids to traverse 10 A From the Departinent of CH=CH 1 HC-CH H 0-1 0O Figure 1 Peptide bond (in broken-line circle) bet ween ,9 amino acids, tyrosine (top) (and alaiiine (bottom). Scale in aniigstromi tunits is shoucn at left. in the direction of the axis of the helix. In a helix, furthermore, a given amino acid has as neighbors n-ot onlly the 2 amino acids to which it is attached bv peptide linkages, but also those residues 1 turn in the helix above or below. If these steric relationships were knowl, and as yet they are completely obscure, we would say that the secondary structure of the protein miolecule had been established. If the basic configuration of the polypeptide is a helical coil, then it is clear from the known dimensions of globular proteins that these helices must be folded over in various configurations so as to be contained, in imost cases, in a volume which is very roughly spherical in shape. Sueh conformations have usually Chemistry, -Northwestern University, Evanston, Ill. 828 Circulation, Volume XXI, May 1960 PROTEIN MOLECULES IN SOLUTION 829 Ala Lys Pro *Thr Tyr Phe - Figure 2 Sequence of amino acids in insulin. There are 2 peptide chains linked together by disulfide bonds represented by heavily blackened parts in the diagram. Gly = glycine; Phe = phenylalanine; etc. (Republished by permission of Naturwissenschaften.37) Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 been called the tertiary structure of the protein molecule. Despite heroic efforts over an extended period, information concerning 3dimensional configuration is available for only 1 protein, myoglobin,5 (fig. 4) which shows a very complicated pattern of foldings, the biologic significance of which is not comprehended at all. Nevert-heless, even in the absence of detailed structural models, we do have extensive knowledge of the over-all sizes and gross shapes of mnany protein macromolecules. A few of these are shown in figure 5. It is only * I from these very gross models that we can begin our considerations of the properties of protein molecules. Interactions of Protein Molecules with Solvent In any discussion of the behavior of proteins in solution, i.e., of the interactions of these macromolecules with their environment, we recognize that the surfaces shown in figure 5 cannot actually be as smooth as the drawings might imply. If the figure were expanded sufficiently we would be able to visualize the specific side-chains of the individual amino acid residues. A few polar side-chains are shown schematically in figure 6. Such polar and nonpolar substituents play a vital role in the inlteraction of the macromolecule with its environment. That the environment of the protein surface is very different from that in the bulk of the solution is evident from the marked differences in the chemical behavior of functional Circulation, Volume XXI, May 1 960 Figure 3 Schematic diagram z *n of helical configuration of a polypeptide chain (solid line) with hydrogen bonds represented by broken lines. groups which are attached to protein sidechains. For example, sulfhydryl groups in simple compounds, such as glutathione, react rapidly and com-pletely with m-aniy substances, particularly heavy metals: R -SH +Ag = R -SAg±+H+. (1) The same -SH group attached to a protein (for example, f3-lactoglobulin) acts as if it were completely hiddeni fromi any attacking reag,ent, such as Ag+ (fig. 7). The diffusion current of Ag+ added to protein in buffer rises just as it would if protein- were absent. Only if the protein is denat-ured, for example with KLOTZ 830 Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 voA4 Figure 4 Configuration of mYoglobin. The gray disk is the hemne group, the little bla1ck ball attached to it, an atrificially introduced group requiredl for the x-ray analysis of this macrotm-olecule. Thl e white parts are te polypeptide confliguration. ([eUpublished b?y permission of Endeavour. 38) Table 1 Some l7minsn(l P2 's in Pr?oteins PK Protein Group Expected Fotund TRibonuelease OH 10 9.9 (3 groups) >11.5 (.3 groiips) Ovalbuiyiii OH 10 >1- I{eiiioglobixi HN NH+ 6.5 < 5 Circutation, Volume XXI. May 1960 PROTEIN MOLECULES IN SOLUTION Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 urea, does the -SH group become visible to the Ag+. Denaturation is said to "unmask" this mereaptan in the protein molecule. Similarly, if we examine the titration characteristics of acid-base groups of a protein, we often find peculiarities (table 1). For example, the phenolic -OH groups of tyrosine residues often will not dissociate H+ ions at pH values much higher than are characteristic for this substituent.i 7 Also, basic residues, such as histidine, may not accept a hydrogen ion until the pH is lowered much below that at which an imidazolium ion is normally formed.8 A more detailed description of such unusual acid-base behavior is shown in figure 8 for an -N(CH3)2 group which has been artificially introduced into a protein; hydrogen ions are taken up only at much stronger acidities than are necessary when the same group is attached to a simple amino acid.9 There is thus no question that a protein environment may modify the properties of a substituent. However, the molecular basis of the changed behavior is less certain. I should like to present primarily our views'0 on this subject. As we have suggested recently, many aspects of the behavior of proteins in solutioni can be interpreted advantageously if we assign some special properties to the water in the immediate neighborhood of the macromolecule. There are good reasons, aside from the properties of proteins as such, for assuming a special ice-like character of regions of the hydrate water of these macromolecules. As was mentioned previously, proteins contain a large number and variety of nonpolar sidechains (table 2); in simple molecules these side-chains possess the remarkable property of forming crystalline hydrates consisting predominantly of molecules of water in an ice-like lattice with occasional holes into which a nonpolar molecule may fit.1" 12 Such nonpolar hydrates are stable at temperatures well above the melting point of ordinary pure ice; it is the preseniee of the nonpolar molecules in the holes of the polyhedrons (figs. 9 and 10) formed by the H20 molecules which stabilizes the ice-like lattice. Circulation, Volume XXI, May 1960 831 SIZES AND SHAPES OF SOME PROTEINS IOOA H20 e CrGlucos. No+ Cl- Glucose N Chymotrypsinogen Myoglobin ,B-Lactoglobulin Hemoglobin SFibrinogen Hemocyanin Turnip Yellow Virus Figure 5 Sizes and shapes of some protein molecules. For comparison, some small molecules and a scale are shown at the top. In a protein molecule many nonpolar sidechains exist in juxtaposition. Thus, there could be a "coupling" of the ice-like lattices which are formed around each residue with a consequent strengthening of the resultant hydrate structure. The structure of the lattice which is formed by the coupling of water of hydration of protein side-chains may not be 832 KLOTZ >. :-..Xu .::PRO T EIN. -:S . :(~~~~'4H ± H' Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 + + Figure 6 Schematic diagram of some ionic side-cha-ins in protein molecules. Table 2 :TA-LACTOGLOBULIN CComparison of Amino Acid Side-Cha ins with IMolecutles TWhich Form Hydrates 0/ Some molecules which form crystalline hydrates Some amino acid side-chains / / (Ala) -CH3 GH s (Val) --CHz--CH<XCHs (Leu) -C CR4 CHs CH3 2CH< Buffer Ur / / A cf _ AgNO3 _, Figure 7 Detection of -SH groups by amperometric Ag+ titration of protein P-lactoglobulin. In buffer, current due to free Aig+ goes up immediately upon addition of AgNO3. ,i.e., -SH grounps on protein ±SH urea,scurt sotays are not detectable. In buffer ait zero for substantiia addition of of due to removal of {A+ by -SH --.1 L-'y Ag,crra behantor protein. X- -1'11 -CHR--SH (Cys) -CHz--CHR--S CIC3 (Met) -CR2- (Phe) / CHI C CR3CH XCTI,CH CHSH CH3 S-CR3 /9 the same as that of nonpolar hydrates of small molecules, such as those listed in table 2, but the possibilities for cooperative interaction in the macromolecule should be eveu greater, since the residues are even initially relatively fixed in position by the framiework of the protein molecule. A very schematic and exaggerated representationa of such an ice-like lattice is shown in figure 11. The preseniee of patches or ribbonis of icelike coveriugs around various regions of a protein nmoleeule would affect both the kinetics of Circulation, Volume XXI, May 1960 833 PROTEIN MOLECULES IN SOLUTION N (CH5)2 TITRATION OF s02- R Fraction Acid Form Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 pH Figure 8 Uptake of H+ as a function of pH by dimethylamino group (of molecule shown) when attached to glycine and bovine serum albumin (BSA), respectively. A much lower pH. i.e., stronger acidity, is required to place H+ on conjugate to the protein. its reactions and its state of equilibrium. Rates of diffusion of essentially all substances to a reactive site under an ice layer would be greatly reduced. With Li', for example, it is known'3 that the rate of diffusion in iee is at least 10,000 times slower than in liquid water. On this basis, one can readily understand why Ag' titrates very slowly with -SH residues in many proteins, e.g., fl-lactoglobulin (fig. 7). Judging from the observation that mereaptans form erystalline hydrates, we might conclude reasonably that -SH groups in proteins could participate cooperatively with neighboring side-chains in the formation of an ice patch. Diffusion of Ag+ through this ice patch would be greatly hindered as compared to the movement of this ion in ordinary water. Collsequently, the protein mereaptan acts as if it were masked. Similarly, the shift in the titration curve of protein-bound groups such as -N(CH3)2 toward a lower pH (fig. 8) can be readily inCirculation, Volume XXI, May 1960 terpreted. Such a shift means, in essence, that the environmient is more favorable to (CH3 ) 21\N than to (CH3) 2NH'. Sueh a bias seems very reasonable, for the creation of a charged group in place of the uncharged one would require some breakdown of the ice lattice of the protein. Interactions of Protein Molecules with Small Molecules and Ions In addition to interacting with solvent, proteins may also interact with all types of small ions and molecules in the solution. The stability of the complexes which are formed depends largely upon the character of the macromolecule and the small ligand, but it is also affected by the nature of the environment in the solution. An exhaustive survey of these factors would be inappropriate for this symposium, but a few representative examples of some dominant factors in these interactions will be described. 834 KLOTZ Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 Figure 9 JLIodel of arrangement of water mnolecules in crystalline hydrates of nonipola,r taial(cules. Each baill represents 1 water molecule. The molecules are arranged in ptntagoinal planes, 12 of wbhich enclose a dodeaahedral space. The hole insidle this polyhedron is about 3 A in diameter. (Republishedc by permission of Zcitschritff fir Elelctroclh'n?ic.1) Metal Ions and Hydrogen Ionis Figure 6 was iutl oduced inito this paper largely to slhow sonic J)olar side-chains wlhicl occur ini practically all proteinis and which are capable of forminog complexes witlh maimv metal ions. Practically all metal ions forml complexes wvith some protein side-chainis. Neverthcl4ss, there are differenlees in specificity amiong the rm-etals; soIIme examples of these differeinees are summarized in table 3. In (gencral, tlhe alkali iietlals (e.g., Nat) do not form coimiplexes, althi.oug,h with special proteinis (e.g., myosin14) bind(iing is observed, probably due to a clustering of appropriate resid-ues of the macromolecule. The alkalilneeartlh m1-etals (e.g., Cat+) formn complexes largely with car boxylate or' phiosphate sidechainis. HeaIvier metals, particularly of the transition series, are bound to practically all groups of the protein which are capable of Circulation, Volutme XXI, May 1960 835 PROTEIN MOLECULES IN SOLUTION Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 Figure 10 JPolyhedlrous of type of figure 9 arranged to form a large lattice, as is found in crystalline nonpolar hydrates. Compartments of 5, 6 and 7 A are available for enclosure of nonipolar molecules wlithin icater structure. (Republished by permission of Oxford University Press.39) Table 3 Bindiny of Metal lons by tSome Sbubstitnents inl ProteinsHG=CHO-P-O Na (."a+ + Fe+++ Zni Cu++ C / t0 I HN N W \C/ 0 0- H 0 + 0 + + + 0 0 + ± -+ + A + HNH S- 0 0 0 0 + + + +[+ 0 0 + + + + +±+ Hg++ ++ +±+ + + *TEor furtiher qiuantitative estinates of affinities, see (Trd an-id Wileox.13a + + + + + + interacting with II+. Nevertheless, there are differeciees ii:i specificity. Beyond the very qualitative iniformatioln of table 3, one can point to the very high affilnity of Fe-much grieater thani that of Cu-for pheinolic oxygens Circulation, Volumte XXI, May 1960 Figure 11 Schematic diagramii of ice-like character of hydration sheath airounzd nonpolar groups in protein molecules. Small circ-les represent waiter molecules. KLOTZ 836 MOLES BOUND METAL MOLES TOTAL PROTEIN 1O0 5 pH Zn++ pH Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 8 .8 +~~~+ 6.80 6.08 - 7.6 -I 0.0 I -3.0 LOG -2.0 [Me+ ] FREE Figure 12 Uptake of Zn++, and of Ca+, respectively, by serum albumin at different pH's. Ordinate represents number of Zn++ (or Ca±±) ions bound by 1 molecule of protein; abscissa gives the logarithm of the concentration of free metal ion (in qvater) in equilibrium with the bound metal ion on the protein. in compounds such as tyrosine. On the other hand, Cu is bound more strongly to amine nitrogens or to mereaptan groups. An interesting case not shown in table 3 is Pb ion which, like Zn or Cu, is bound strongly to mereaptan groups. On the other hand, Pb shows a much greater affinity for carboxyl than for amine groups; the opposite preference is characteristic of Zn and Cu. There are also specificities among protein molecules, although these are not too well understood in molecular terms, except perhaps for the heme pigments in which a definite prosthetic group can be isolated. Thus, conalbumin or transferrin shows a remarkable affinity for Fe, the interaction occurring with tyrosine residues;15 yet, many other proteins have a substantial complement of tyrosines, but no unusual avidity for Fe. The ligand groups shown in table 3 are all capable of binding H+ ions. It is not surpris- ing, therefore, that H+ ions should compete with metal ions and, hence, that the binding of metals is dependent on the pH. For example, Fe(III)-conalbumin is stable above pH 7 but loses Fe(III) below pH 7 (if citrate is present'5). More detailed results for albumiin complexes with Zn'6 and with Ca'1 are illustrated in figure 12. In all cases, lower pH, i.e., increasing acidity, lowers the amount of bound metal ion. The extent of binding of metal ions by proteins may also be controlled in part by "metal buffers, i.e., complexing and chelating agents with a strong affinity for the cation. These agents thus compete with the protein for the ietal. For example, the binding of zinc ions by serum albumin at pH 6.4 can be reduced by addition of glycine,'8 which forms complexes with the metal. Similarly, acetate competes with protein for lead ions, iodide for mereury,'8 citrate for copper.'9 These "metal Circulation, Volume XXI, May 1960 837 PROTEIN MOLECULES IN SOLUTION Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 -4 log [Am. Acifree Figure 13 Comparison of the extent of binding of L-tryptophan with that of D-tryptophan by serum albumin. Ordinate gives number of bound amino acid molecules on each protein molecule, and abscissa shows concentration of free tryptophan in equilibrium ivith that bound to protein. buffers" can be used to control the extent of binding of cations and, hence, the physical or biologic consequences of such binding. Organic Molecules Proteins also form complexes with organic ions and molecules. For purposes of illustration, we shall consider just a few complexes of serum albumin. In a sense, this protein is not representative because it binds a great variety of molecules and ions whereas one usually emphasizes the high specificity of the interactions of proteins, e.g., enzymes or antibodies. We shall not discuss the molecular interpretations of such specificity. It should be pointed out, however, that serum albumin also shows remarkable discriminatory properties. A most striking illustration occurs with tryptophan (fig. 13), the L-isomer being Circulation, Volume XXI, May 1960 bound substantially, the D-isomer markedly less so.20 Preferences also are shown for posi- tional isomers. For example, albumin distinguishes between o-, m-, and p-methyl red (fig. 14), not only with respect to affinity but also, as judged from spectroscopic manifestations, with respect to the nature of the interaction. As with metal ions, the binding of organic ions by proteins may be decreased by competition of other anions. Thus, the addition of chloride, acetate or phthalate ions, reduces the binding of methyl orange by serum albumin (fig. 15). Obviously, the type of buffer solution used may influence substantially the interaction between macromolecule and small ion. Nonionic organic molecules may also compete with ionic. Thus, decanol reduces the binding of tryptophan,20 and alcohols, from 838 KI,OTZ 3\N N=N w3 C Go O2 z 0 a. PROTEIN Li 0 z -1I D 0 Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 H3C\ e0 O co J 0 0 /=N N H *- 0 12-13A / CO- H3 PROTE IN PROTEIN Figure 14 Structural formulas of some positional isomers between which serum albumin can discriminate either in extent of binding or in effect of protein on color of small molecule. Also shown are distances between groups on protein inrolved in binding of these small molecules, for comparison with distances between (CHS)2N- and -COO- substituents on each small molecule. Protein sidechains are too far apart (12 to 13 A) to link to both (CH,3)2N- and -COO- of top molecule; hence linkage is to -COO- only, and binding and spectroscopic effects differ in this case from results with lower 2 molecules. -5.0 LOG I DY EIFREE Figure 15 Difference in extent of binding of methyl orange by serum albumin in presence of different buffTers: *, acetate buffer, 0, phthalate buffer. Lower binding curve in presence of latter buffer showvs that phthalate ion interferes more thacn does acetate. methanol to decanol, reduce the binding of methyl orange by serum albumin. It is also possible to increase the uptake of organic molecules and ions by a variety of stratagems, 2 of which will be mentioned here. A molecule such as pyridine-2-azo-p-dimethylaniline (I) is not, by itself, bound appreciably by proteins. However, if very small amounts of metal, e.g., Zn++ are added to the solution, very substantial binding of (I) follows (fig. 16). As we now kniow, the metal N-N-=N - -N(CH3)2 (I) ion acts as a bridge between protein and organic molecule (fig. 17). The net result is a strong cooperative interaction. Zinc ion and (I) have a small affinity for each other (k 0.2 x 103) even in the absence of protein; attached to a protein, the affinity of Zn++ for (I) rises almost 50-fold (k = 9 X 103). Equally striking increases in the binding of anions by protein molecules are observed in the presence of glycine (fig. 18), but here Circulation, Volume XXI, May 1960 PROTEIN MOLECULIES IN SOLUTION 839 2 z WZn+i li C 0 a z cn D 00 ED .1 Uf) Uf) LJ 0 0 2 ri Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 No Zn 0 fv I *- -5 LOG -4 [DYE] FREE Figure 16 Increase in degree of binding of a dye (pyridine2-azo-p-dimethylaniline) when Zn++ is added to the protein solution. Each curve shows number of bound dye molecules on each protein molecule, as a function of free dye concentration in the solution. PROTEIN .Zn ON = N o (C H3 Figure 17 a bridge Schematic diagram of Zn++ ion acting between protein and small molecule. This mediating effect of Zn++ increases number of dye molecules bound to protein molecule. as the mechanism of the effect must be markedly different, for glycine itself is not bound by serum albumin. Furthermore, the concentrations of this amino acid which are required to increase the binding of protein22 are very high (0.5 to 2 M) compared to those at which metals manifest their effects. At such high Circulation, Volume XXI, May 1960 log (A) Figure 18 Increased binding of methyl orange by serum albumin when glycine is added to the solution: 0, no glycine; 0, 2.5 M glycine. Each curve shows number of bound dye molecules per protein molecule as a function of the concentration (A) of free dye in solution. concentrations of amino acid, the dielectric constant of the solution is altered appreciably and, hence, interactions between charged species could be changed. Nevertheless, as we have shown, this electrostatic effect is not the dominant one since even the binding of uncharged molecules by proteins is increased. Instead, it is our feeling, for reasons described elsewhere,22 that the influence of glycine is to be understood in terms of what this amino acid can do to the solvent rather than to the protein. Zwitterionic glycine perturbs the water in its neighborhood and, hence, also the hydration of the organic molecule in the aqueous solution. Consequently, the organic moleeule turns more readily to the more hospitable environment of the protein envelope where, with the assistance of nonpolar side- 840 KLOTZ As has been mentioned earlier, mercury combines with -SH groups of proteins. An analogous reaction occurs with organic mercurials, e.g., between (II), (CH8)2N-Q- p Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 Figure 19 Top: protein molecule, P, with some water of hydration (small circles) around nonpolar regions. Bottom: protein complex with an anionic hydrocarbon molecule, A, showing increased hydration in the complex due to cooperative effect betwveen P and A. chains of the macromolecule, a cooperative stabilization of the water lattice may ensue (fig. 19). On this basis, one can see that glycine should affect the binding of uncharged organic molecules as well as of organic ions. So far, we have considered examples of protein complexes which are essentially reversible (in the thermodynamic sense), that is, which can be dissociated by lowering the concentrations of the reacting species. It should be pointed out that covalently bound complexes may also be formed at physiologic temperatures and pH's. Among the most interesting of these from a physiologic viewpoint are those involving the mercaptan or disulfide substituents of proteins. N = N-D- Hg-OOCCH8 (II) and almost any protein with mereaptan groups: P - SHl+RHgX--.P-- S--HgR +HX. (2) The mereaptide bonds formed are very stable; although the rate of the reaction is not always high, the stoichiometry23 is very sharp (fig. 20). Similarly, sharp stoichiometry is observed in the reaction of disulfide compounds with protein mereaptan groups :24 P-SH + R-S-S-R --- P-S-S-R. (3) Intramacromolecular Interactions of Proteins It has been known for some time that for many proteins, the kinetic-molecular unit which we see at physiologic pH's and temperatures is not necessarily the smallest molecular species. Under other conditions, the macromolecule may dissociate (or associate) reversibly into smaller (or larger) units. Some of the earliest work in this field is that of Eriksson-Quensel and Svedberg,25 who demonstrated a number of disaggregation reactions in hemocyanins. A more recent study with ft-lactoglobulin,2628 is shown in figure 21. The normal molecular weight for this protein is around 35,000. From the shape shown in figure 5, which is based on x-ray investigations, one might not be too surprised to see that this molecule dissociates into 2 equal fragments at low pH. Furthermore, as figure 21 shows, aggregation into units which have molecular weights of 70,000 may also occur at low temperatures, although this is not as complete as the reaction of disaggregation. In general, disaggregation occurs at pH's far from the isoelectric point, i.e., where the protein acquires a substantial charge. Sueh behavior is characteristic of many other proteins, for example, chymotrypsin29 and insulin,30 as well as of hemocyanin and B&-lactoglobulin. WVith insulin the electrostatic Circulation, Volume XXI, May 90an 841 PROTEIN MOLECULES IN SOLUTION 1.0k- r A 0-a o o0 Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 0.5 r) 0 8 ° C 0 0 0 l p~~~~ O P-S-Hg-R -SH+R-Hg-X 1- 15 30 45 [dye]free aqueous M 60 X Figure 20 Extent of binding of an azomercurial by serum albumin. The binding rises very sharply even at exceedingly small concentrations of mercurial, and levels off very sharply when combination with all available -SH groups has occurred. basis of the dissociation has been indicated by further experiments.30 In acid solution the unit of molecular weight near 36,000 dissociates into particles which have a molecular weight of 12,000 when the charge reaches about +20 to +30 per trimer. In the presence of SCN- ions, known to be bound to insulin in acid solution, lower pH's must be attained to dissociate the trimer. Obviously, the bound anions reduce the internal electrostatic repulsion in the insulin molecule and hence stabilize the trimer. These observations indicate that many protein molecules are composed of smaller units linked together by relatively weak bonds. The nature of these bonds has still not been established unequivocally. Intermacromolecular Interactions In addition to interacting with themselves, protein molecules of a given kind may form complexes with other types of proteins or with other types of macromolecules, including polyCirculation, Volume XXI, May 1960 saccharides and nueleic acids. One can distinguish in these combinations 2 general categories of complexes: (1) nonspecific, in which electrostatic factors usually predominate; (2) specific, in which distinctive structural factors are of major significance. As an example of each category, we shall cite protein-nucleic acid interactions. Interactions of proteins with nucleic acids are excellent examples of cases in which opposite electrostatic charge is the primary condition for reaction. Precipitates are formed, for example, between nucleic acid and lysozyme, ovalbumin, serum albumin, fibrinogen, casein, and aetin, respectively ;31-33 in all cases the protein is cationic, the nucleic acid, anionic. As we would expect for a complex held together primarily by electrostatic forces, small quantities of salt usually redissolve and dissociate the combination. Soluble complexes may also be formed.32 34 In all these cases, the KLOTZ 842 Molecular Weight - 60,000) 0 ~~~~~~~~~20C. IO---0 0' 'IO 0 50,000o 1 I~~~~~~~~~~~~~~~~~~~~~~ O // 0 40,0ooo 0~ -0-0-0- 30,0001 25° C. 20o,0o0 Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 6-v 10,000 2.0 3.0 4.0 5.0 pH Figure 21 Variations in molecular weight of /3-lactoglobulini. Solid line correlates results at 25 C., broken line at 2 C. (Method of presentation taken from Tanford.28) The "normal" moleculaxr weight of /3-lactoglobulin at its isoelectric point (pH 5) and higher is 35,000. At lowtt temperature and pH's of 4 to 5, the molecules aggregate to dimers. At very low pH, the molecule breaks up into 2 subutitts; compare with diagram in figure 5 which is based on x-ray data. wide spectrum of proteius with which nucleic acid complexes are formed shows the absence of any specificity in these combinations. Theoretical calculations35 also lead to the conclusion that electrostatic attraction can account quantitatively for the strength of association. In contrast to these nonspecific complexes, one may cite the highly specific complexes between nucleic acid and protein which are exemplified in the viruses. These combinations lead to special supermolecular structures, with characteristic arrangements of constituent macromolecules. A model for tobacco mosaic virus36' 36a, 36b iS showin in figure 22. This supermolecule, with a nlolecular weight of about 40,000,000, has an over-all shape of a long thin rod, 3,000 A in length and 170 A in width. It is composed of about 5 per cent ribonucleic acid and 95 per cent protein. The association, however, is hardly the random type found in the complexes described in the preceding paragraph. The nucleic acid seemiis to be a single strand winding its wvay through the rod in a helical fashion. In contrast, the protein componeiit seemlls to be present in subunits which have a molecular weight of about 17,000. These subunits envelop the coiled strand of nlueleic acid by winding around it, also in a helical array. There are other rod-shaped viruses with structures similar to that of tobacco mosaic virus. Many small viruses, however, are quite different in structure, being spherical in shape (for example, turnip yellow virus, fig. S). Again, interactioni between niucleic acid and protein leads to a very specific structure, but with the nucleic acid in the core of the spherical supermolecule anil protein subunits arranged symmetrieally in an outer tbick shell. Circulation, Volume XXI, May 1960 -PROTEIN MOLECULES INT SOLUTTION 843 Conclusion see that a protein molecule in solution may undergo( a hlost of intteractionis: witli solvent, withl smnall moleeules land ioiis, wvitlh itself, witlh other mnaeromolecules. In a jthy-So1dogie miliell all of these possibilities resciLt tIheniselv es a.d(1 wvill ronipete amiong thlemselves. Reasonable guesses (as t-o whieb xwill be the doiminaiit inter-et 10118 (1ii often be Ina(le Lfioini lelia-vioi observe(l in somie of' thie simple sy slens iwllie we hiave (lescribed. It seem11S reaW-sonlable to expert tait further studies of s11(c1 int ew-rartions in arntifieial simple vitviroiiinieiits 1iiayf iltiiately provide! a finin basis f1 relaftilln) thle strlneture, of ninae romoleruldes to liciir biologir art ivity. 'Thus,iwe Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 References N 1. lh ri^ A. \ P., S ANOVR, F., SLITII, L A.ANT) :Errf u,N IL.: l)isulphide 1)0(ods of insilin. BioCeieii. .J. 60: 541, 1955. 2. lJiTs;, C 11. \V., STEN, W. If., A \N- tOOR, S.: S1(tuies oil the 11 sittitre of rihoimleclr':ise. 11( Svmposlum (o11 Piroteini Struiceture. Neuberger, A., Eld. Liondon, Methlinll & C1o. Lttl., 195S, pt. 211. ii, A. P., AsN ANFINseN, C. IL: Stlldies oll 3. As thle (lisultide bridIg-es ill rihollnuelease. B3iochimii. e(t bId)hYs. acta 24: 633, 19;57. 4. L'ING, 1.., (CoRY\, . I'., \NND IBRANSON, It. I.: p St ructu e of r-ote ns: Two xv l rogell-bolnded lelic:il eolfiguratiills of tlle polYp)eptile ch"lil. Prloc. NlA-l. Acuol. Sci., U.S. 37: 205, 1951. sw IKv\w 1 C., D1hno, G., 1)NTrZIS, I. AM., P.A1SI 1, II .,NDWCO\FVFi , IJ.: A. three dinieii,8io01,1t modlel of the myoglobin mIIolecule ohtamdni(l iv X royt! nalysis. Na(lture 181: 662, 1958. XXI) \ 6. (nsnsmniu, .1 . A., A.: State of tmr)osinie in egg all)iil 1d1(1 in insilulin as deby spectropliotol3metric titration. Bioterf'ini chemo. J. 37: 302, 1943. 7. T NNFORDJ, C. Ti XSTEiX, .7 . S., -.-ND P DS, P. G.: Phlenloli, lc1droxyl ionizaition in protr IsS. I i-. Hibolclese. .1. Ani. Chewm. Soc. 77: 6409, 1955. S. 1IEYCEIOKR, S., XI) SinsARiIT, 3.: IncreTF(ase in aci(l-dbilnding sites oil denatrlluation of boIse ferrihemnoglobin ait ,O. .1. Am. Clew,l. Soc. 81: 10. nucleic acid hell.. 7'he p)rotneif siibltoits also uoint Kliol'z, 1M.AL, .o) LPu:ss, I-. A.: Ionie eq(uilibria aroutndl n a hclical a-rr'a q. Tile cueter of the r od is h0ollo0w (butt filled with water mooculies) for a distance of aboutl; &0 - arocl)d the aris. (RepvblishedI byJ Permission O(f J<ndea(o.301 im a pIroteill conjug.ate, of a sulfonamide type. Bliochim. et biophys. acta 38: 57, 196(. Protein hydration and(] behavior. Science 128: 81.5, 19-58. 11. ST'WKRELERG, M. V., AND ML:LLEiR, IH. R.: Feste Gashydrate. IT. Strullktur uiil Ralumehemie. Ztschr. Elektrocbeiu. 58: 25, 1954. 5679, 1959. 9. Figure 22 X-4 model of tobat co mosaic virusE. The ithite ellip.soids rej)present hr proteiu Subunits (each, abodt 18',000 molecular weight); tile block coil, the ribo- Circulation, Volume XAI, May 1960 844 Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 12. -, AND MEUTHEN, B.: Feste Gashydrate. VII. Hydrate Wasserlislicher Xther. Ztschr. Elektrochem. 62: 130, 1958. 13. EIGEN, M., AND DEMAEYER, L.: Self-dissociation and protonic charge transport in water and ice. Proe. Roy. Soc. London s.A:247: 505, 1958. 13a. GURD, F. R. N., AND WILCOX, P. E.: Complex formation between metallic cations and proteins, peptides, and amino acids. Advane. Protein Chem. 11: 311, 1956. 14. LEWIS, M. S., AND SAROFF, H. A.: Binding of ions to muscle proteins: Measurements on the binding of potassium and sodium ions to myosin A, myosin B and actin. J. Am. Chem. Soc. 79: 2112, 1957. 15. WARNER, R. C., AND WEBER, I.: Metal combining properties of conalbumin. J. Am. Chem. Soc. 75: 5094, 1953. 16. GURD, F. R. N., AND GOODMAN, D. S.: Preparation and properties of serum and plasma proteins. XXXII. The interaction of human serum albumin with zinc ions. J. Am. Chem. Soc. 74: 670, 1952. 17. KATZ, S., AND KLOTZ, I. M.: Interactions of calcium with serum albumin. Arch. Biochem. 44: 351, 1953. 18. GURD, F. R. N.: Specificity of metal-protein interactions. In Ion Transport Across Membranes. Clarke, H. T., Ed. New York, Academic Press, 1954, p. 264. 19. KLOTZ, I. M., AND FIESS, H. A.: Thermodynamics of metalloprotein combinations: Buffer effects in copper-albumin complexes. J. Phys. and Colloid Chem. 55: 101, 1951. 20. MCCMENAMY, R. H., AND ONCLEY, J. L.: Specific binding of L-tryptophan to serum albumin. J. Biol. Chem. 233: 1436, 1958. 21. KLOTZ, I. M., BURKHARD, R. K., AND URQUHART. J. M.: Structural specificities in the interactions of some organic ions with serum albumin. J. Am. Chem. Soc. 74: 202, 1952. 22. -, AND LUBORSKY, S. W.: Binding of organic ions by proteins: Effects of changes in solTent and their implications as to the nature of the complexes. J. Am. Chem. Soc. 81: 5119, 1959. 23. HOROWITZ, M. G., AND KLOTZ, I. M.: Interactions of an azomercurial with proteins. Arch. Biochem. 63: 77, 1956. 94. KLOTZ, I. M., AYERS, J., Ho, J. Y. C., HOROWITZ, M. G., AND HEINEY, R. E.: Interactionns of proteins with disulfide compounds: Some implications for electron transport in proteins. J. Am. Chem. Soc. 80: 2132, 1958. 25. ERIKSSON-QUENSEL, I. B., AND SVEDBERG, T.: Molecular weights and pH-stability regions of the hemocyanins. Biol. Bull. 71: 498, 1936. 26. TOWNEND, R., AND TIMASHEFF, S. N.: The pH KLOTZ dependence of the association of 8-lactoglobulin. Arch. Biochem. 63: 482, 1956. 27. -: Molecular weight of P-lactoglobulin. J. Am. Chem. Soc. 79: 3613, 1957. 28. TANFORD, C.: Configuration of globular proteins in aqueous solution and its dependence on pH. In Symposiumi on Protein Structure. Neuberger, A., Ed. London, Methuen & Co. Ltd., 1958, p. 61. 29. SCHWERT, G., AND KAUFM AN, S.: Molecular size and shape of the pancreatic proteases. III. a-Chymotrypsin. J. Biol. Chem. 190: 807, 1951. 30. FREDERICQ E., AND NEURATH, H.: Interaction of insulin with thiocyanate and other anions: The minimum niolecular weight of insulin. J. Am. Chem. Soc. 72: 2684, 1950. 31. CHEPINOGA, 0. P., AND GROSBLAT, R. SH.: Characteristics of salt-like compounds of deoxyribonucleic acid with proteins. Ukrain. Biokhim. Zhur. 21: 121, 1949. Chem. Abstr. 48: 4014, 1954. 32. GEIDUSCHEK, E. P., AND DOTY, P.: A light scattering investigatioii of the interaction of sodium desoxyribonucleate with bovine serum albumin. Biochim. et biophys. acta 9: 609, 1952. 33. TONGUR, V. S., DISKINA, B. S., AND SPITKOVSKII, D. M.: Production and properties of artificial complexes of desoxyribonucleic acid with serum albumin. Biokhimiya 22: 879, 1957. 34. HAM,ER, D.: Interaction of acids of high molecular weight with albumin. Biochem. J. 56: 610, 1954. 35. STEINER, R. F.: Reversible association processes of globular proteins. II. Electrostatic complexes of plasma albumin and lyzozyme. Arch. Biochem. 47: 56, 1953. 36. FRANKLIN, R. E., KLUG, A., AND HOLMES, K. C.: X-ray diffraction studies of the structure and morphology of tobacco mosaic virus. In Ciba Foundation Symposium on the Nature of Viruses. Wolstenholme, G. E. W., and Miller, E. C. P., Eds. London, Churchill Ltd., 1957, p. 39. 36a. CRICK, F. H. C., AND WATSON, J. D.: Virus structure: General principles. In Ciba Foundation Symposium on the Nature of Viruses. Wolstenholme, G. E. W., and Miller. E. C. P., Eds. London, Churchill Ltd., 1957, p. 5. 36b. CASPAR, D.: Discussion. In Ciba Foundation Symposium on the Nature of Viruses. Wolstenholme, G. E. W., and Miller, E. C. P., Eds. London, Churchill Ltd., 1957, p. 14. 37. Tuppy, H.: Aminosiure Sequenzen in Proteinen. Naturwissensehaften 46: 35, 1959. 38. PERUTZ, M. F.: Some recent advances in molecular biology. Endeavour 17: 190, 1958. 39. WELLS, A. F.: The Third Dimension in Chemistry. New York, Oxford University Press, 1956, plate X. Circulation, Volume XXI, May 1960 Protein Molecules in Solution IRVING M. KLOTZ Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017 Circulation. 1960;21:828-844 doi: 10.1161/01.CIR.21.5.828 Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231 Copyright © 1960 American Heart Association, Inc. All rights reserved. Print ISSN: 0009-7322. Online ISSN: 1524-4539 The online version of this article, along with updated information and services, is located on the World Wide Web at: http://circ.ahajournals.org/content/21/5/828 Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: http://www.lww.com/reprints Subscriptions: Information about subscribing to Circulation is online at: http://circ.ahajournals.org//subscriptions/