THE Jcmm.a OF BIOLOGICAL CHEMISTRY Vol. 247, No. 17, Issue of September 10, pp. 5580-5586, 1972

Printed in U.S.A.

Intermolecular Cross-linking of Monomeric Proteins and

Cross-linking of Oligomeric Proteins as a

Probe of Quaternary Structure

APPLICATION TO LEUCINE AMINOPEPTIDASE (BOVINE LENS)*

(Received for publication, December 20, 1971)

FREDERICK H. CARPENTER AND KATHRYN TINKER HARRINGTON

From the Department of Biochemistry, University of California, Berkeley, California S/,720

SUMMARY

Reaction of leucine aminopeptidase with dimethyl suber- imidate at pH 8.5 and room temperature, followed by incuba- tion with sodium dodecyl sulfate and mercaptoethanol and then electrophoresis in 3.5% acrylamide gels resulted in six protein bands with molecular weights as integer values of 56,000. Under reaction conditions where appreciable cross- linking took place, the species corresponding to dimers, tetramers, and hexamers were much more prevalent than those corresponding to trimers and pentamers, the latter being barely detectable. Providing that the ability of the protomers to form cross-links reflects their spatial arrange- ment in the oligomer, the results suggest that leucine amino- peptidase is composed of six identical protomers arranged as a trimer of dimers. Of the possible arrangements for a hexamer, the planar hexamer with all heterologous interac- tions is incompatible with these data. Thus, cross-linking reactions may yield information on the spatial arrangement of protomers in oligomers as well as on the number and kinds of protomers in the oligomer. Individual treatment of a number of monomeric proteins with dimethyl suber- imidate in the frozen state at - 10’ in the presence of sodium dodecyl sulfate and mercaptoethanol yielded intermolecular cross-linked species which could be separated into a large number of components on gel electrophoresis in sodium dodecyl sulfate. For any one cross-linked protein the plot of logarithm of molecular weight against RF fell on a straight line for RF values between 0.2 and 0.8. The lines described by different proteins, although generally having similar slopes, were not always superimposable. This pro- cedure for cross-linking monomeric proteins has potential use in determining molecular weights by gel electrophoresis in sodium dodecyl sulfate.

In the past few years increasing attention has been paid to the theoretical and experimental problems associated with the

* This work was supported in part by Grant GB 27573 of the National Science Foundation and Grants AM 00608 and EY 00813 of the National Institutes of Health.

quaternary structure of oligomeric proteins (l-3). Information on the number and kinds of protomers in an oligomeric protein can be obtained by several physical and chemical approaches (2) of which’the cross-linking procedure of Davies and Stark (4) is especially appealing because of its simplicity in terms of equip- ment. Information on the arrangement in space of the proto- mers has been available through the physical techniques of electron microscopy or x-ray crystallography (3, 5). The results presented in this report indicate that, in some instances, the cross-linking reaction of Davies and Stark (4) may yield infor- mation on the arrangement of the protomers in space as well as on the kind and number of protomers in the oligomer.

Leucine aminopeptidase (EC 3.4.1. l), isolated in crystalline form from the bovine lens (6), has a molecular weight of about 320,000 (7, 8). In view of this large size it is not surprising to find that the enzyme has a number of subunits. Kretschmer (9, 10) and Kretschmer and Hanson (II), using a variety of techniques including gel chromatography in sodium dodecyl sulfate (9), electron microscopy (lo), and sedimentation velocity and diffusion studies in sodium dodecyl sulfate and urea (II), have concluded that the molecule is made up of 10 subunits with molecular weights of 32,600. On the other hand, Weber and Osborn (12), using gel electrophoresis in sodium dodecyl sulfate, and Melbye and Carpenter (8), by a variety of techniques in- cluding gel electrophoresis in sodium dodecyl sulfate and sedi- mentation and equilibrium studies in a number of denaturing solvents of both the crystalline enzyme and the reduced and carboxamidomethylated derivative, arrived at a subunit size of about 54,000, indicating a hexameric structure for the enzyme. Recent studies of Vahl and Carpenter (13) and Vahl (14) on the stoichiometry of the zinc content of the enzyme and on the replacement of zinc by magnesium and manganese are in accord with the 54,000 subunit.

The present study investigates the nature of the oligomer by the cross-linking technique of Davies and Stark (4) in which dimethyl suberimidate is used to cross link the protomers while in the oligomeric form, followed by dissociation and gel electro- phoresis of the cross-linked products in sodium dodecyl sulfate. The results not, only indicate the hexameric nature of the en- zyme but also suggest that the protomers are arranged as a trimer of dimers.

The present report also describes procedures by which mono-

5580

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

merit proteins as well as oligomeric proteins can be cross-linked to give a variety of molecular species which in some cases var) from monomer to decamer. The cross-linked proteins yield distinctive patterns on gel electrophoresis in sodium dodecyl sulfate and have potential use as standards for determining molecular weights not only of monomeric but also of oligomeric proteins by the electrophoresis procedure. The procedure for obtaining a large number of cross-linked species consists of per- forming the cross-linking reaction in the frozen state in the presence of sodium dodecyl sulfate and mercaptoethanol.

EXPERIMENTAL PROCEDURE

Materials

Proteins-The proteins, their molecular weight and source were as follows: horse transferrin (74,000), Nutritional Bio- chemicals (lot 4696) ; bovine serum albumin (65,000) (15), Sigma Chemical Co. (lot 99B-0450) ; ovalbumin (45,000) (16), Pentex (lot 10) ; rabbit muscle aldolase (40,000, monomer) (16), Cal- biochem (lot 901485) ; horse apoferritin (19,000, monomer) (17), Mann Research Laboratories (lot T3577). Leucine aminopepti- dase was isolated in crystalline form from bovine lenses by the procedure of Hanson et al. (6).

Chemicals-Dimethyl suberimidate was synthesized from suberonitrile (Aldrich Chemical Co.) according to the directions of Davies and Stark (4), m.p. 212-214”. Triethanolamine, acrylamide, N, N’-methylenebisacrylamide, mercaptoethanol, and ammonium persulfate were products of Eastman Organic Chemicals. The source of N , N , N’, N’tetramethylethylene- diamine was Matheson, Coleman, Bell; of Coomassie Brilliant Blue R250 was Colab Laboratories, Inc., and of sodium dodecyl sulfate was Sigma Chemical Co.

Methods

Cross-linking of Monomeric Proteins-Proteins (0.25 to 0.5 mg) were dissolved in 50 ~1 of a 0.2 M triethanolamine solution at pH 8.5 containing 0.3 mg of dimethyl suberimidate dihydrochloride. To the protein solution was added 50 ~1 of 2% sodium dodecyl sulfate-2% mercaptoethanol in 0.2 M triethanolamine at pH 8.5. The resulting mixture was frozen in Dry Ice-isopropanol and stored at -10” for 16 hours. The solution was then incubated at 37” for 2 hours and subjected to electrophoresis.

Cross-linking of Leucine Aminopeptidase-The reaction with dimethyl suberimidate was performed essentially as described by Davies and Stark (4). Differences are noted on the figure legends.

Sodium Dodecyl Sulfate-Gel Electrophoresis-Most of the runs were performed essentially by the Davies and Stark modification (4) of the Weber and Osborn procedure (12), on gels containing 3.5% acrylamide, 0.135% N, N’-methylenebisacrylamide polym- erized with 0.075% ammonium persulfate and 0.033% N, N, N’, N’-tetramethylethylenediamine. Samples containing approxi- mately 50 pg of protein were introduced on the gels in a solution containing 50% glycerol and 0.02% bromophenol blue. The running buffer was 0.1 M borate, 0.1 M sodium acetate, and 0.1 y0 sodium dodecyl sulfate at pH 8.5. Electrophoresis was per- formed at 6 ma per tube until the dye front approached the bottom of the tube. Gels were removed, cut at the dye front, stained with Coomassie blue, and destained electrophoretically.

RESULTS

Cross-linking Variety of Proteins in Frozen State-A number of proteins were treated with dimethyl suberimidate at -10” in

5581

the presence of sodium dodecyl sulfate and mercaptoethanol, incubated at 37”, and then electrophoresed on 3.5% gels. With the exception of aldolase and apoferritin, these proteins are pre- sumed to be monomeric during reaction with the reagent. In electrophoresis, each gave at least 5 bands of decreasing intensity with increasing molecular weight. As many as 10 bands, cor- responding to the production of decamers, could be discerned in some cases. Aldolase exists normally as a tetramer (4). In agreement with this, the band representing the tetrameric form was the strongest, indicating that intramolecular cross-linking took place more readily than the intermolecular reaction. How- ever, some pentamers, hexamers, and heptamers were also formed. Their appearance indicates that some cross-linking between oligomers also took place under these conditions.

A plot of the logarithms of the molecular weight of each cross- linked protein against the corresponding RF value in the 3.5y0 gels is shown in Fig. 1. The line is the best fit of all of the data by the least squares procedure. For horse transferrin, the mono- mer molecular weight was determined to be 74,000 from its RF value on the least square line and the cross-linked species were plotted as multiples of the 74,000 value. The points for bovine serum albumin all lie slightly above the line while the points for aldolase all lie below the line. Those for ovalbumin closely ap- proximate the line. The average absolute relative deviation of all the points about this line is 6.2%. However, for any one

t 1.01 ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’

0.2 0.4 0.6 0.8 1.0

Rf

FIG. 1. Plot of molecular weight on logarithm scale of cross- linked proteins against RF on 3.5% gels for bovine serum albumin (0)) apoferritin (u), ovalbumin (0)) transferrin (0) and aldolase (A).

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5582

FIG. 2. A, reaction with dimethyl suberimidate before and after treating with sodium dodecyl sulfate and mercaptoethanol. Gel a, leucine aminopeptidase (0.25 mg per ml) was reacted at pH 8.5 in 0.2 M triethanolamine with dimethyl suberimidate (0.5 mg per ml) for 90 min at room temperature, then sodium dodecyl sulfate and mercaptoethanol were added to make 1% solutions and the resulting solutions were incubated at 37” for 2 hours, frozen at -10” overnight, and electrophoresed. Gel b, same as Gel a except that the leucine aminopeptidase was incubated with so- dium dodecyl sulfate and mercaptoethanol before addition of the dimethyl suberimidate. B, effect of concentration of dimethyl suberimidate. Leucine aminopeptidase (8.6 mg per ml) was re- acted with dimethyl suberimidate at (Gel a) 6 mg per ml and (Gel b) at 0.5 mg per ml under conditions described in Gel a of A. C, effect of concentration of leucine aminonentidase. The enzyme at concentrations per ml as noted on the fig&e was treated with dimethyl suberimidate at 0.5 mg per ml under conditions described for Gel CL of A.

particular protein the situation is much better. Thus the line formed by the points for serum albumin has an average absolute relative deviation of 0.5%. In other words, the relationship between logarithm of the molecular weight and RF values holds much better for an individual protein and its cross-linked oligo- mers than for a variety of proteins of differing molecular weights.

Cross-linking of Leucine Aminopepticlase in Aqueous Solution- Gel a of Fig. 2A shows the typical result of reacting leucine aminopeptidase at 0.25 mg per ml with the dimethyl suberimidate at 0.5 mg per ml at room temperature. Similar patterns are to be noted in Fig. 2, B and C. The intensity of the monomer band varies depending on the treatment but in all cases the bands corresponding to dimer, tetramer, and hexamer are much more prominent than those corresponding to trimer and pentamer, the latter being barely detectable. Gel b of Fig. 2A shows the results obtained when the leucine aminopeptidase is first incubated with sodium dodecyl sulfate and mercaptoethanol at 37” for 2 hours, in order to disrupt the oligomeric structure of the enzyme, and then treated with dimethyl suberimidate at -10” in the frozen state. The principal component of the reaction mixture cor- responds to the monomer with a slight amount of dimer also being present. These results show that very little cross-linking of monomeric proteins takes place once they have been fully denatured and enveloped in sodium dodecyl sulfate. Presum- ably the detergent masks the ammo groups so that they can no longer react with the dimethyl suberimidate.

The effect of concentration of dimethyl suberimidate on the degree of cross-linking in the oligomer is shown in Fig. 2B where leucine aminopeptidase at 8.6 mg per ml has been reacted with the suberimidate at 6 mg per ml (Gel a) or at 0.5 mg per ml (Gel b). At the high concentration of reagent, there is very little monomer remaining and the trimer and pentamer bands are scarcely discernible. The principal species is hexamer with tetramer and dimer in reduced amounts in order. However, there is no evidence for a species of molecular weight higher than hexamer which indicates that there was little interoligomeric cross-linking.

The effect of the concentration of leucine aminopeptidase on the cross-linking reaction can be ascertained by comparison of Gel b of Fig. 2B along with the gels of Fig. 2C. The dimethyl suberimidate was kept constant at 0.5 mg per ml while the leucine aminopeptidase concentration was varied from 8.6 mg per ml (Gel b, Fig. 2B) to 0.25 mg per ml (Fig. 2C). The degree and extent of cross-linking was very similar in all cases except for a slightly decreasing diminution of the monomer form as the con- centration of protein was decreased. This is the type of result that is expected of cross-linking reactions that are taking place intraoligomerically in that they should be relatively insensitive to concentration of the protein.

The sizes of the various species present in the cross-linked leucine aminopeptidase were determined by comparing their RF values with those of a cross-linked preparation of ovalbumin as shown in Fig. 3. The line (- ) connects the open circles ( 0) which are plots of the logarithm of molecular weight of various species of cross-linked ovalbumin ranging in size from monomer (45,000) to octamer (360,000) against their corresponding RF values. The cross-linked ovalbumin was prepared by reaction with dimethyl suberimidate in the frozen state. The vertical bars indicate the RF values of the various species of cross-linked leucine aminopeptidase. The molecular weights for each species along with their calculated subunit size (in parentheses) were as follows: monomer, 59,000 (59,000); dimer, 112,000 (56,000) ;

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

a i5 30 -

43

F J 20 -

P 0 4 X

E Qa .-

’

lo=

4-

31 I I I I I I I I I J 0.2 0.4 0.6 0.8

R f FIG. 3. Plot of logarithm of molecular weight against RF for

various species of cross-linked ovalbumin (0) upon electrophore- sis on 3.5% acrvlamide eels at nH 8.5 in 0.10/o sodium dodecvl sulfate. ‘I&e vekical bars indicate the RF values of the vario& species of cross-linked leucine aminopeptidase.

trimer, 168,000 (56,000) ; tetramer, 220,000 (55,000) ; pentamer, 280,000 (56,000) ; and hexamer, 325,000 (54,000). The average subunit size of 56,000 is in good agreement with values previously published from this laboratory (8).

DISCUSSION

Conditions for Cross-linking Monomeric Proteins

There are two interrelated factors involved in the formation of polymeric molecules through intermolecular cross-linking of monomeric proteins with dimethyl suberimidate. These are freezing and the presence of sodium dodecyl sulfate and mercap- toethanol. Freezing alone is not sufficient nor does the presence of sodium dodecyl sulfate and mercaptoethanol in the absence of freezing have an effect. With sodium dodecyl sulfate one can visualize that the freezing process concentrates the protein and the reagent into micelles stabilized by the detergent. In this concentrated form the protein would be in a favorable condition for intermolecular cross-linking. Perhaps mercaptoethanol can operate in a similar fashion, serving as a solvent to concentrate the reagent and protein as the water is frozen out. Other adden- dum might work as well. This point was not closely investigated but it was determined that neither the triethanolamine buffer nor the borate buffer could replace the sodium dodecyl sulfate in the frozen reaction.

The sodium dodecyl sulfate aided in the cross-linking reaction only if the solutions were frozen before incubation in the presence of detergent. I f the proteins were heated in the detergent and then frozen with the dimethyl suberimidate very little cross- linking took place. Evidently denaturation and unfolding of the protein with the concomitant binding of detergent, especially

5583

to the free amino groups, effectively blocks the cross-linking reaction.

The degree of cross-linking increased as the length of time of freezing increased up to the longest time investigated (5 days). However, in most cases there was a high degree of cross-linking after freezing for 16 hours which did not increase dramatically in the next 4 days. So that for most purposes, overnight freezing should be adequate. However, if one desires to prepare cross- linked proteins containing a large variety of aggregates such as for use as standards in gel electrophoresis, it might be well to consider freezing for a prolonged period of time.

Determination of Molecular Weights

Since the introduction by Shapiro et al. (18) of the sodium dodecyl sulfate-gel electrophoresis procedure for estimating molecular weights of proteins, the technique has been applied extensively, especially by Weber and Osborn (12), to a variety of proteins. According to Reynolds and Tanford (19), the electrophoretic mobility of proteins in gels can be a function of molecular weight when (a) the charge per unit mass is constant and (b) the hydrodynamic properties are a function of molecular length only. Plotting the logarithm of the molecular weight for the monomer, dimer, etc. of serum albumin (Fig. 1) against RF gave a very good straight line, especially for RF values between 0.2 and 0.8. Similarly each of the other proteins shown in Fig. 1 yielded excellent straight lines. Thus for any one protein there is good correlation between size and RF. However, the lines for each protein, although generally having similar slopes, do not always coincide. This difference may be due in part to the better control of conditions when cross-linked proteins in one gel are compared rather than when different proteins in different gels are compared. Also the location of the line depends on the value selected for the monomer molecular weight. Using a slightly higher value for aldolase (42,000 rather than 40,000) and a slightly lower value for bovine serum albumin (63,000 rather than 65,000) would make these two proteins closely approximate the least square line.

Griffith (20) has recently reported that some cross-linked oligomers formed by reaction of proteins with glutaraldehyde migrated at a faster rate in sodium dodecyl sulfate electrophoresis than would be expected from their calculated molecular weights. Although we did not make a careful study of this possibility, the results shown in Fig. 1 do not reveal a consistent trend of this nature. Presumably both glutaraldehyde and dimethyl suberim- idate form cross links primarily through the epsilon amino groups of lysine. However, they differ in that (a) the charge is retained in the reaction with dimethyl suberimidate while it is largely lost on reaction with the dialdehyde and (b) the length and flexibility of the connection formed with the dimethyl suberimidate is somewhat greater than that formed with the dialdehyde. These differences in the nature of the cross-linking reagent may account for the differences observed on electrophoresis of the two types of branched oligomers. Also in Griffith’s work, the cross-linked oligomers were compared with monomers which did not contain any intramolecular cross links. In our work, the monomeric species could very well contain a number of intramolecular cross links. Such intramolecular cross links might constrain the molecule so as to affect the ratio of length to molecular weight in the denatured molecule.

The above results have some implications on the use of the sodium dodecyl sulfate-gel electrophoresis for determining the molecular weights of unknown proteins. (a) Cross-linking standard proteins makes it possible to cover a very wide range of

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5584

molecular weights. This will be useful especially in studies on oligomeric proteins. (6) Comparable cross-linking of an un- known protein will yield a number of points for comparison rather than just one point. (c) If the proteins selected here are a repre- sentative cross-section of the behavior one can expect to find in nature, then cross-linked ovalbumin would serve as the best single marker protein for proteins above 40,000 daltons in that its regression line closely approximates the least square line for all points. Consequently, cross-linked ovalbumin was used as a standard in the following studies on leucine aminopeptidase.

Leucine Aminopeptidase

All Protomers of Same Size-Several factors indicate that the protomers of leucine aminopeptidase are all of the same size which suggests that they may be identical. Upon disruption in sodium dodecyl sulfate and electrophoresis, predominantly one species was obtained. If the oligomer had contained species of different size they should have been detected as equally promi- nent bands on electrophoresis. Upon disruption of intraoligo- merit cross-linked leucine aminopeptidase and electrophoresis, the sizes of the various species bore a whole number relationship to one another. This would not be possible if the oligomer contained protomers of different size except in the event that, one protomer was exactly one-half of the size of the other protomer.

Number of Protomers Is Six-Disruption of the cross-linked oligomer. followed by electrophoresis, yielded five readily dis- cernible bands and a sixth band, corresponding to pentamer, which was weak but detectable in preparations that had not been too heavily cross-linked. The six bands indicate a hexameric structure for the protein. In preparations which have been heavily cross-linked the hexamer was the most prominent species and there were no bands corresponding in size to anything greater than the hexamer. The molecular weight of the slowest moving band in cross-linked and dissociated leucine aminopeptidase agreed well in size (325,000) with that found by hydrodynamic methods (8) for the native oligomer (326,000). When this is coupled with the fact that the smallest unit detected by a variety of techniques, including sodium dodecyl sulfate-gel electropho- resis of noncross-linked, cross-linked, and carboxamidomethyl- ated leucine aminopeptidase and equilibrium centrifugation in dissociating solvents such as urea and guanidinium chloride (8), has a molecular weight of 54,000 + 4,000, then the weight ratio of oligomer to monomer is 6.

Arrangement of Protomers in Oligomer-The finding that the dimer, tetramer, and hexamer were the predominant species formed in the intraoligomeric cross-linking reaction can be ex- plained in two ways: (a) “spatial reactivity,” meaning that the ability of a reagent to form cross links depends solely on the spatial arrangement of the protomers and their reactive groups; or (b) “induced reactivity,” meaning that the reactive groups on the protomers are so arranged that when a cross link has been formed between any two protomers, the reactivity of the remain- ing groups is changed (in this case, the reactivity is diminished or the groups are masked). Although the experimental evidence does not allow a differentiation between these two explanations, the first is the most attractive because of its implications for use in determination of quaternary structure. When the interfaces between one protomer and any two or more adjacent protomers are not identical, one can expect this difference to be reflected in the ability of the reactive groups to form cross links. In turn this difference in ability to form cross links may be reflected in the distribution of cross-linked species in a partial reaction. For leucine aminopeptidase, the distribution of cross-linked species.

FIG. 4. A, diagram of a planar hexsmer with all heterologous interactions. R, diagram of a planar hexamer with all isologous interactions.

can arise from a spatial arrangement consisting of a trimer of dimers where the rates of formation of intradimer cross links are not the same as the rates for interdimer cross links.

Fig. 4 represents two planar arrangements of six protomers to give a hexamer. Cornish-Bowden and Koshland (3) have recently discussed the theory of quaternary structure of proteins composed of identical subunits for oligomers composed of dimers, trimers, and tetramers. In their terms of reference, Fig. 4A represents a planar hexamer composed of all heterologous binding domains (p&. The lower diagram (Fig. 4B) represents a case where the hexamer is formed by all isologous interactions (pp3, 443, + - + - + -). In the theoretical treatment developed by Cornish-Bowden and Koshland (3), all possible interactions were allowed, i.e. pp, QQ, and pq, in the same oligomer. For the planar hexamer this treatment gives rise to a number of asym- metric forms between the extremes of the all heterologous and all isologous (Fig. 4) both of which are symmetrical. For sim- plicity the present discussion will be confined to those cases where all interactions are not allowed, i.e. an isologous interaction (pp) excludes a heterologous interaction (pg) and vice versa. With this limitation for the planar hexamer, only the two forms depicted in Fig. 4 exist. Of these two forms, the upper (Fig. 4A), consisting of all het.erologous interactions, is incompatible with the cross-linking results on leucine aminopeptidase in that it should yield, depending upon the degree of cross-linking, either

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5585

Octahedron

Pq,, r

c

6

FIG. 5. A, diagram of a hexamer in the form of a trigonal prism. B, diagram of a trigonal prism containing six heterologous (pp~) and three isologous (I?~) interactions. C, diagram of a hexamer in the form of an octahedron. D, an octahedral form containing six heterologous (pqc) and two kinds of isologous interactions b.3, ssd. Looking at. each vertex, one proceeds in a clockwise direction from p to p to T to s.

increasing or decreasing amounts of all of the species from monomer to hexamer. On the other hand, the all isologous structure (Fig. 4B) is compatible with the cross-linking results on leucine aminopeptidase. If one assumes that cross-linking reactions take place more readily between pp interfaces than between qq (or vice versa), then in a partially cross-linked product one should obtain a prevalence of dimers, tetramers, and hexa- mers over trimers and pentamers.

Two additional configurations for a hexamer are shown in Fig. 5. In Fig. 5A a trigonal prism is illustrated in which each protomer interacts with three other protomers. Shown in Fig. 5B is the only form of the trigonal prism which is permitted when one allows only one type of interaction for each interface. It contains a mixture of heterologous and isologous interactions (pqs, rra). Forms which are completely isologous or completely heterologous are impossible. Since this form has two different types of interactions (pq and TT), it could give rise to the results found for leucine aminopeptidase in much the same fashion as the all isologous planar hexagon of Fig. 4B. In Fig. 5C an octahedral st.ructure is illustrated where each protomer interacts with four other protomers. The one form of the octahedral structure that

meets the restrictions of allowing only one type of interaction for each interface is illustrated in Fig. 50. This form has a total of six heterologous reactions of one kind (pq6) and of six isologous interactions of two different kinds (TT~ and ss~). This form could also give rise to the distribution of cross-linked products observed for leucine aminopeptidase. For example, if cross-linking at rr interfaces is much more facile than at pq or ss, then dimer, tetra- mer, and hexamer could exceed trimer and pentamer.

All of the situations depicted here are subject to a theoretical treatment similar to that used by Cornish-Bowden and Koshland (3) with the exception that rates of cross-linking between inter- faces are substituted for free energies of binding at interfaces. From such a procedure one could calculate the relative composi- tion in terms of monomer, dimer, etc. of a partially cross-linked oligomer as a function of the degree of reaction. Conceivably there could exist enough differences in behavior between the four structures depicted here to allow their complete resolution by a comparison of experimental data with theoretical yields. The data on leucine aminopeptidase was not obtained in sufficient detail or accuracy to warrant such an attempt. At present the only structure which can be definitely excluded is the planar hexamer with all heterologous interactions. Even this con- clusion is subject to some ambiguity because of the two explana- tions for the cross-linking results given at the start of this section. Nevertheless the cross-linking reactions deserve further consider- ation both theoretically and experimentally as a method for elucidating quaternary structure.

In this connection it should be noted that Kohlhaw and Boat- man (21) observed a preponderance of dimers and tetramers over trimers in cross-linking studies performed on the tetrameric form of isopropylmalate synthetase with dimethyl suberimidate. This result is incompatible with an all heterologous planar struc- ture (pq4) for the tetramer. Either of the other two potential structures for a tetramer, the isologous planar tetramer (pp~ qqs,

+ - + -) or the tetrahedron (pp~, qqz, VP) (3), could yield this cross-linking result. However, the fact that under some con- ditions the enzyme exists as a dimer (21) would favor the isolo- gous planar structure for the tetramer.

Acknowledgments-We are indebted to Charles Sawyer for computer aid in determining the least square lines.

REFERENCES

1. MONOD, J., WYMAN, J., AND CHANGEUX, J.-P. (1965) J. Mel Biol. 12, 88~218

2. KLOTZ, I. M., LANGERMAN, N. R., AND DARNALL, D. W. (1979) Annu. Rev. Biochem. 39, 25

3. CORNISH-BOWDEN, A. J., AND KOSHLAND, D. E., JR. (1971) J. Biol. Chem. 246, 3092-3102

4. D~vrns. G. E., AND STARK. G. R. (1970) Proc. i\iat. Acad. Sci. U. ,%‘A. 66,‘651

5. PERUTZ, M. F., MUIRHEAD, H., Cox, J. M., AND GOAMAN, L. C. G. (1968) Nature 219, 131-139

6. HANSON, g., G~~~ssER, D., AND KIRSCHKE, H. (1965) Hoppe- Seyler’s Z. Physiol. Chem. 340, 107

7. KRETSCHMER, K., .~ND HANSON, H. (1965) Hoppe-Seyler’s Z. Physiol. Chem. 340, 126

8. MELEIYE, S. W., AND CARPENTER, F. H. (1971) J. Biol. Chem. 246, 2459-2463

9. KRETSCHMER, K. (1967) Hoppe-Seyler’s Z. Physiol. Chem. 348, 1723

10. KRETSCHMER, K. (1968) Hoppe-Seyler’s Z. Physiol. Chem. 349, 715

11. KHETSCHMER, K., .~ND HANSON, H. (1968) Hoppe-Seyler’s Z. Physiol. Chem., 349, 831

12. WEBER, K., AND OSUORN, M. (1969) J. Biol. Chem. 244, 4406- 4412

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

5556

13. VAHL, J. M., AND CARPENTER, F. H. (1971) Fed. Proc. 30, 1183 14. VAHL, J. M. (1950) Doctoral dissertation, University of Cali-

18. SHAPIRO, A. L., VIRUELA, E., AND MAIZEL, J. V. JR. (1967)

fornia, Berkeley Biochem. Biophys. Res. Commun. 28, 815

15. SCATCHARD, G., AND PIGLIACAMPI, J. (1962) J. Amer. C&m. 19. REYNOLDS, J. A., AND TANFORD, C. (1970) J. Biol. Chem. 246

51616165 Sot. 84, 127

16. CASTELLINO, F. J., AND BARKER, R. (1968) Biochemistry 7,2207 20. GRIFFITH, I. P. (1972) Biochem. J. 126,553

17. BJORK, I., AND FISH, W. W. (1971) Biochemistry 10, 2844-2848 21. KOHLHAW, G., AND BOATMAN, G. (1971) Biochem. Biophys.

Res. Commun. 43, 741

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Frederick H. Carpenter and Kathryn Tinker HarringtonLEUCINE AMINOPEPTIDASE (BOVINE LENS)

Oligomeric Proteins as a Probe of Quaternary Structure: APPLICATION TO Intermolecular Cross-linking of Monomeric Proteins and Cross-linking of

1972, 247:5580-5586.J. Biol. Chem. 

  http://www.jbc.org/content/247/17/5580Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/247/17/5580.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on October 8, 2020

http://ww

w.jbc.org/

Dow

nloaded from