Reconstitution of the Escherichia coli pyruvate dehydrogenase complex.

The binding of pyruvate dehydrogenase and dihydrolipoyl dehydrogenase (flavoprotein) to dihydrolipoyl transacetylase, the core enzyme of the E. coli pyruvate dehydrogenase complex [EC 1.2.4.1:pyruvate:lipoate oxidoreductase (decaryboxylating and acceptor-acetylating)], has been studied using sedimentation equilibrium analysis and radioactive enzymes in conjunction with gel filtration chromatography. The results show that the transacetylase, which consists of 24 apparently identical polypeptide chains organized into a cube-like structure, has the potential to bind 24 pyruvate dehydrogenase dimers in the absence of flavoprotein and 24 flavoprotein dimers in the absence of pyruvate dehydrogenase. The results of reconstitution experiments, utilizing binding and activity measurements, indicate that the transacetylase can accommodate a total of only about 12 pyruvate dehydrogenase dimers and six flavoprotein dimers and that this stoichiometry, which is the same as that of the native pyruvate dehydrogenase complex, produces maximum activity. It appears that steric hindrance between the relatively bulky pyruvate dehydrogenase and flavoprotein molecules prevents the transacetylase from binding 24 molecules of each ligand. A structural model for the native and reconstituted pyruvate dehydrogenase complexes is proposed in which the 12 pyruvate dehydrogenase dimers are distributed symmetrically on the 12 edges of the transacetylase cube and the six flavoprotein dimers are distributed in the six faces of the cube.     

Acyl group and electron pair relay system: A network of interacting lipoyl moieties in the pyruvate and α-ketoglutarate dehydrogenase complexes from Escherichia coli

The dihydrolipoyl transacetylase component of the Escherichia coli pyruvate dehydrogenase complex [pyruvate:lipoate oxidoreductase (decarboxylating and acceptor-acetylating), EC 1.2.4.1] bears two sites on each of its 24 polypeptide chains that undergo reductive acetylation by [2-(14)C]pyruvate and thiamin pyrophosphate, acetylation by [1-(14)C]acetyl-CoA in the presence of DPNH, and reaction with N-ethyl[2,3-(14)C]maleimide in the presence of pyruvate and thiamin pyrophosphate. The data strongly imply that these sites are covalently bound lipoyl moieties. The results of similar experiments with the E. coli alpha-ketoglutarate dehydrogenase complex [2-oxoglutarate:lipoate oxidoreductase (decarboxylating and acceptor-succinylating), EC 1.2.4.2] indicate that its dihydrolipoyl transsuccinylase component bears only one lipoyl moiety on each of its 24 chains. Charging of the 48 acetyl acceptor sites on the transacetylase or the 24 succinyl acceptor sites on the transsuccinylase by pyruvate or alpha-ketoglutarate, respectively, and thiamin pyrophosphate was observed in the presence of only a few functionally active pyruvate dehydrogenase or alpha-ketoglutarate dehydrogenase chains. Extensive crosslinking of the transacetylase chains was observed when the pyruvate dehydrogenase complex was treated with pyruvate and thiamin pyrophosphate or with DPNH in the presence of N,N'-o- or N,N'-p-phenylenedimaleimide, respectively. When the alpha-ketoglutarate dehydrogenase complex was treated with DPNH in the presence of N,N'-p-phenylenedimaleimide, only transsuccinylase monomers and crosslinked transsuccinylase dimers were detected. It appears that the 48 lipoyl moieties in the transacetylase and the 24 lipoyl moieties in the transsuccinylase comprise an interacting network that functions as an acyl group and electron pair relay system through thiol-disulfide and acyl-transfer reactions among all of the lipoyl moieties.     

Mechanism of action of the pyruvate dehydrogenase multienzyme complex from Escherichia coli

The extent of cooperativity among the polypeptide chain components in the overall reaction catalyzed by the pyruvate dehydrogenase multienzyme complex from Escherichia coli has been studied. Selective inactivation of the pyruvate dehydrogenase component with thiamin thiazolone pyrophosphate demonstrates that no cooperativity between this component and the overall catalytic reaction occurs: the amount of overall complex activity is directly proportional to the fraction of active pyruvate dehydrogenase component. The transacetylase component has two lipoic acid residues on each of its polypeptide chains that can be modified by N-[(3)H]ethylmaleimide in the presence of pyruvate and thiamin pyrophosphate. The kinetics of the loss of overall complex activity due to modification of the lipoyl residues on the transacetylase component by maleimide reagents shows that not all lipoic acids are coupled into the overall catalytic reaction and that acyl-group and electron pair transfer involving two or more lipoic acids per catalytic cycle must occur. Finally, full complex activity is found when only half the normal flavin content is present. The results indicate that extensive communication among lipoic acids in acyl-group and electron pair transfer must exist in the normal catalytic mechanism. These results are consistent with the average distances between catalytic sites measured by energy transfer experiments.     

The remarkable structural and functional organization of the eukaryotic pyruvate dehydrogenase complexes

The three-dimensional reconstruction of the bovine kidney pyruvate dehydrogenase complex (M(r) approximately 7.8 x 10(6)) comprising about 22 molecules of pyruvate dehydrogenase (E(1)) and about 6 molecules of dihydrolipoamide dehydrogenase (E(3)) with its binding protein associated with the 60-subunit dihydrolipoamide acetyltransferase (E(2)) core provides considerable insight into the structural and functional organization of the largest multienzyme complex known. The structure shows that potentially 60 centers for acetyl-CoA synthesis are organized in sets of three at each of the 20 vertices of the pentagonal dodecahedral core. These centers consist of three E(1) molecules bound to one E(2) trimer adjacent to an E(3) molecule in each of 12 pentagonal openings. The E(1) components are anchored to the E(1)-binding domain of the E(2) subunits through an approximately 50-A-long linker. Three of these linkers emanate from the outside edges of the triangular base of the E(2) trimer and form a cage around its base that may shelter the lipoyl domains and the E(1) and E(2) active sites. The docking of the atomic structures of E(1) and the E(1) binding and lipoyl domains of E(2) in the electron microscopy map gives a good fit and indicates that the E(1) active site is approximately 95 A above the base of the trimer. We propose that the lipoyl domains and its tether (swinging arm) rotate about the E(1)-binding domain of E(2,) which is centrally located 45-50 A from the E(1), E(2), and E(3) active sites, and that the highly flexible breathing core augments the transfer of intermediates between active sites.     

Subunit stoichiometry and molecular weight of the pyruvate dehydrogenase multienzyme complex from Escherichia coli.

The molar ratio of the component enzymes of the pyruvate dehydrogenase multienzyme complex from Escherichia coli was found to be 1.8:1.7:1[pyruvate decarboxylase (E1):dihydrolipoyl transacetylase (E2):dihydrolipoyl dehydrogenase (E3)]. This ratio was determined by measuring the Coomassie blue staining of the constituent enzymes after sodium dodecyl sulfate/polyacrylamide slab gel electrophoresis. The above ratio is the average of four separate experiments with two different enzyme preparations. The average molecular weights of the individual enzymes were found to be 96,000, 76,000, and 55,000 for E1, E2, and E3, respectively, by sodium dodecyl sulfate and sodium dodecyl sulfate/8 M urea polyacrylamide gel electrophoresis and by column chromatography in 6 M guanidine . HCl. The molecular weight of E2 was reduced to 33,000-36,000 after extensive reduction and alkylation with iodoacetamide. The molecular weights of the complex, E1, and E3 were found to be 4,800,000, 182,000, and 104,000, respectively, with low-angle laser light scattering. Both E1 and E3 are dimeric under the conditions employed. If octahedral symmetry is assumed for the E2 core, a polypeptide chain ratio of 24:24:12 (E1:E2:E3) is in good agreement with the measured molar ratio of component enzymes and the molecular weight of the pyruvate dehydrogenase complex.     

A computer model analysis of the active-site coupling mechanism in the pyruvate dehydrogenase multienzyme complex of Escherichia coli.

A computer modeling system developed to analyze experimental data for inactivation of the Escherichia coli alpha-ketoglutarate dehydrogenase complex (KGDC) accompanying release of lipoyl moieties by lipoamidase and by trypsin [Hackert, M.L., Oliver, R.M. & Reed, L.J. (1983) Proc. Natl. Acad. Sci. USA 80, 2226-2230] was used to analyze analogous data for the E. coli pyruvate dehydrogenase complex (PDC). The model studies indicate that the activity of PDC, as found for KGDC, is influenced by redundancies and random processes, which we describe as a multiple random coupling mechanism. In both complexes more than one lipoyl moiety services each pyruvate dehydrogenase (EC 1.2.4.1) or alpha-ketoglutarate dehydrogenase (EC 1.2.4.2) (E1) subunit, and an extensive lipoyl-lipoyl interaction network for exchange of electrons and possibly acyl groups must also be present. The best fit between computed and experimental data for PDC was obtained with a model that has four lipoyl domains with four or, more probably, eight lipoyl moieties servicing each E1 subunit. The lipoyl-lipoyl interaction network for PDC has lipoyl domain interactions similar to those found for KGDC plus the additional possibility of interaction of a lipoyl moiety and its paired mate on each dihydrolipoamide acetyltransferase (EC 2.3.1.12) (E2) subunit. The two lipoyl moieties on an E2 subunit in PDC appear to be functionally indistinguishable, each servicing the acetyltransferase site of that E2 subunit and a dihydrolipoamide dehydrogenase (EC 1.6.4.3) (E3) subunit if the latter is bound to that particular E2 subunit. The observed difference between inactivation of PDC by lipoamidase and by trypsin appears to be due to dead-end competitive inhibition by lipoyl domains that have been modified by excision of lipoyl moieties by lipoamidase.     

Fatal lactic acidosis due to deficiency of E1 component of the pyruvate dehydrogenase complex

Summary Pyruvate dehydrogenase complex deficiency is thought to be a common cause of lactic acidosis. We report a patient with lactic acidosis and intermittent weakness. The rate of oxidation of pyruvate by intact skeletal muscle and liver mitochondrial fractions was impaired and pyruvate dehydrogenase complex (PDC) activity was low. The amounts of immunoreactive dihydrolipoyl transacetylase and dihydrolipoyl dehydrogenase in liver and skeletal muscle mitochondrial fractions from the patient were normal. However, there were markedly lower concentrations of both the and subunits of the E1 component of PDC.     

Characterization of murine peritoneal macrophage receptors for fibrin(ogen) degradation products

The binding of human fibrinogen degradation fragments D1, E, X, and Y, as well as fibrin fragment D1 dimer, to mouse peritoneal macrophages was examined. A Scatchard plot of fragment D1 binding was biphasic, suggesting two classes of receptors. Fragments D1, D1 dimer, X, and Y in low concentrations bound to macrophages with high affinity (Kd = 23 to 73 X 10(-11) mol/L). Fragment E bound specifically but at a much lower level than the other fragments. Fragment D1 was able to compete for the binding of radiolabeled fragments X and Y but not radiolabeled fragment E. These studies indicate that fragments D and E are recognized by separate receptor systems but that all of the fibrinogen degradation products that contain the D domain are recognized by the same receptor system.     

Characterization of the gamma chain platelet binding site on fibrinogen fragment D.

Glycoprotein (GP) IIb/IIIa on adenosine diphosphate (ADP)-activated human platelets interacts with specific sites on the fibrinogen molecule leading to aggregation. We characterized the platelet-binding site on the gamma chains of fibrinogen using plasmic fragments D gamma A and D gamma'. Fragment D gamma A, which contains the carboxy terminal gamma A400-411 platelet-binding sequence (HHLGGAKQAGDV), was 70-fold more active than the synthetic gamma A400-411 peptide in inhibiting ADP-induced platelet aggregation. Fragment D gamma A inhibited fibrinogen binding and also bound directly to ADP-activated platelets. The Kd values determined for fibrinogen and fragment D gamma A binding were 0.55 mumol/L and 1.2 mumol/L, respectively. In contrast, fragment D gamma', which differs from fragment D gamma A with respect to its gamma chain sequence from position 408 to the COOH-terminus at position 427, did not inhibit platelet aggregation or fibrinogen binding, and did not bind directly to the platelet surface. Denaturation of fragment D gamma A with guanidine-HCl caused a loss of inhibitory activity in platelet aggregation assays. These data indicate that the native conformation of the gamma chain platelet-binding site on fibrinogen is important for optimal binding to GPIIb/IIIa.     

Amino-acid sequence of fragment A, an enzymically active fragment from diphtheria toxin.

The amino-acid sequence of Fragment A from diphtheria toxin is reported. Fragment A (molecular weight, Mr, 21,145) is the major enzymically active fragment produced upon activation of the intact toxin (Mr about 60,000) by limited tryptic digestion and reduction. It, or a similar fragment, is believed responsible for the inhibition of protein synthesis in animal cells exposed to the toxin. Fragment A, which corresponds to the amino terminus of the toxin, is shown here to consist of three major forms (190, 192, and 193 residues) resulting from cleavage by trypsin adjacent to any of three closely spaced arginine residues. All three forms are enzymically active.     

Purification and characterization of branched chain α-keto acid dehydrogenase complex of bovine kidney

A branched chain alpha-keto acid dehydrogenase-dihydrolipoyl transacylase complex was purified to apparent homogeneity from bovine kidney mitochondria. As usually isolated, the complex (s(20,w) = 40 S) contained little, if any, dihydrolipoyl dehydrogenase. When saturated with the latter enzyme the complex had a specific activity of about 12 mumol of alpha-ketoisovalerate oxidized per min per mg of protein at 30 degrees with NAD(+) as electron acceptor. In addition to alpha-ketoisovalerate, the complex also oxidized alpha-ketoisocaproate, alpha-keto-beta-methylvalerate, alpha-ketobutyrate, and pyruvate. The ratios of the specific activities were 2.0:1.5:1.0:1.0:0.4, and the apparent K(m) values were 40, 50, 37, 56, and 1000 muM. The complex was separated into its component enzymes. The branched chain alpha-keto acid dehydrogenase (6 S) consists of two different subunits with estimated molecular weights of 46,000 and 35,000. The dihydrolipoyl transacylase (20 S) contains apparently identical subunits of molecular weight about 52,000. In the electron microscope, the transacylase has the appearance of a cube, and the molecules of branched chain alpha-keto acid dehydrogenase appear to be distributed on the surface of the cube. In contrast to the pyruvate dehydrogenase complex of bovine kidney, the branched chain alpha-keto acid dehydrogenase complex apparently is not regulated by phosphorylation-dephosphorylation. Its activity, however, is subject to modulation by end-product inhibition.     

Identification of functional domains on von Willebrand factor by binding of tryptic fragments to collagen and to platelets in the presence of ristocetin

With the use of monoclonal antibodies that inhibit the ristocetin- induced binding of von Willebrand factor (VWF) to platelets and the binding to collagen, we have previously identified two distinct tryptic fragments. To prove that these fragments contain the platelet binding or the collagen binding domain, we investigated the direct binding of tryptic fragments of 125I-VWF to platelets in the presence of ristocetin and to collagen fibrils. During the course of the tryptic digestion, there was a rapid and parallel decrease in binding to platelets and collagen. In the first ten minutes, binding decreased greater than 50%; a further decrease to 19% and 29%, respectively, was noted at 90 minutes, but no further decrease was observed thereafter. The bound fragments were eluted from platelets and collagen and analyzed on polyacrylamide gradient gels. The fragments bound to the platelets appeared to be reduced, probably by endogenous reducing substances from the platelets. This was prevented by addition of N- ethylmaleimide during the incubation. After 24 hours of digestion, platelets predominantly bound fragments of 116 kd and collagen bound a single fragment of 48 kd. These fragments are similar to those previously identified with the monoclonal antibodies.     

Antimitochondrial antibodies in primary biliary cirrhosis recognize both specific peptides and shared epitopes of the M2 family of antigens

Sera from patients with primary biliary cirrhosis exhibit variable autoantibody reactivity against mitochondria, the commonest antigen (designated M2) including three structures of approximate M.W. 70, 50 and 40 kD. The nature of these antigens has only recently been established; the 70 and 50 kD are the transacetylase E2 and component X, respectively, of the pyruvate dehydrogenase complex and are distinct polypeptides. We have demonstrated, by immunoblotting, elution and rebinding of antibodies, unequivocal cross-reactivity between the major bands of the M2 antigen. In addition, cross-reactivity has been shown between antibodies binding to each of the three M2 bands of mitochondria and two major antigenic bands of both Gram-negative and Gram-positive bacteria. Conversely, antibodies eluted from these two bands of Escherichia coli were found to bind all three M2 bands of mitochondria. These results suggest that the antibodies of primary biliary cirrhosis contain both peptide-specific and cross-reacting antibodies, the latter recognizing a common "M2 epitope" that might include nonprotein components of the peptides. However, direct and competitive enzyme-linked immunosorbent assays failed to implicate the coenzyme of the pyruvate dehydrogenase complex, lipoic acid or its amide, as the common antigenic moiety.     

Autoantibodies in primary biliary cirrhosis: analysis of reactivity against eukaryotic and prokaryotic 2-oxo acid dehydrogenase complexes

Six components of the mammalian 2-oxo acid dehydrogenase complexes have previously been identified as M2 autoantigens in primary biliary cirrhosis. In this report, we present data showing that both polypeptide-specific and cross-reacting antibodies are present in patients' sera. Antibodies reacting with E2 of the pyruvate dehydrogenase complex cross-react with protein X but not with any other mammalian antigen. The main immunogenic region on protein X has been localized to within its single lipoyl domain. Polypeptide-specific antibodies bind to E1 alpha and E1 beta of the pyruvate dehydrogenase complex. Antibodies reacting with the E2 polypeptides of the 2-oxoglutarate dehydrogenase complex and branched-chain 2-oxo acid dehydrogenase complex show some cross-reactivity but do not recognize any of the antigens of the pyruvate dehydrogenase complex. Antibodies against the E2 component of the mammalian pyruvate dehydrogenase complex cross-react effectively with the corresponding protein from yeast but not with E2 from Escherichia coli. Antibody titer against mammalian antigens is significantly higher than against the bacterial antigens, arguing against a bacterial origin for primary biliary cirrhosis.     

Acrylamide Gel Electrophoresis of Fibrinogen and Fibrin Degradation Products

S ummary . Thin layer two phase acrylamide gel electrophoresis was used for separation of fibrinogen and fibrin degradation products which were subsequently quantitated by scanning. The use of 6 m urea in the spacer gel greatly improved separation. Three main fibrinogen fragments X, Y and D and minor fragments E₁, E₂ and E₃ were identified. During plasmin digestion of fibrinogen, Fragment X gradually decreased and Fragment D increased. The concentration of Fragment Y was rather low as would be expected of an intermediate product in the conversion of X to D. At various times of proteolysis, there was good agreement between the level of Fragment X determined by electrophoresis and the amount of protein precipitated at low concentration of protamine sulphate (PS) from the digest pretreated with thrombin. The study supports the concept of two varieties of Fragment X, one with fibrinopeptides (X^fp)—in fibrinogen degradation products (FDP)—and one devoid of fibrinopeptides (X°)—in fibrin degradation products (fdp). During electrophoresis without urea, Fragment X° undergoes extensive polymerization in the slots and gel because of its dissociation from complexes with Fragments Y and D. In the presence of urea, Fragment X° of the early fibrin digest had slower electrophoretic mobility than Fragment X^fp. The mobility of other fibrinogen fragments was not affected by thrombin.     

Evidence for a multiple random coupling mechanism in the alpha-ketoglutarate dehydrogenase multienzyme complex of Escherichia coli: a computer model analysis.

A computer modeling system was used to analyze experimental data for inactivation of the Escherichia coli alpha-ketoglutarate dehydrogenase complex accompanying release of lipoic acid residues by lipoamidase and by trypsin [Stepp, L. R., Bleile, D. M., McRorie, D. K., Pettit, F. H. & Reed, L. J. (1981) Biochemistry 20, 4555-4560]. The results provide insight into the active-site coupling mechanism in the alpha-ketoglutarate dehydrogenase complex. The model studies indicate that the overall activity of the alpha-ketoglutarate dehydrogenase complex is influenced by redundancies and random processes that we describe as a multiple random coupling mechanism. More than one lipoyl moiety services each E1 subunit (alpha-ketoglutarate dehydrogenase, EC 1.2.4.2), and an extensive lipoyl-lipoyl interaction network for exchange of electrons and possibly acyl groups must also be present. The best fit between computed and experimental data was obtained with a model that has two lipoyl moieties servicing each E1 subunit and a lipoyl-lipoyl interaction network that links all lipoyl moieties on the E2 cube (dihydrolipoamide succinyltransferase, EC 2.3.1.61). The single lipoyl moiety on an E2 subunit is assumed to service the coenzyme A-dependent succinyltransferase site of that E2 subunit as well as an E3 subunit (dihydrolipoamide dehydrogenase, EC 1.6.4.3) if the latter is bound to that particular E2 subunit.     

Binding of fibrin fragments to one-chain and two-chain tissue-type plasminogen activator.

To explore whether fibrin fragments have binding affinity for the tissue-type plasminogen activator (t-PA) molecule, the interactions were studied of (DD)E complex and fragments DD, E1, and E3 with one-chain and two-chain t-PA. For this purpose, a solid-phase binding assay was developed using microtiter plates with nitrocellulose filters. It was found that (DD)E complex and fragments DD and E3 retained the t-PA binding function of the parent fibrin molecule, thus demonstrating that t-PA binds to both the D and E domains of fibrin. Unexpectedly, fragment E1 did not bind t-PA. Fibrin fragments had different binding properties for one-chain and two-chain t-PA. (DD)E complex had the highest and fragment E3 the lowest affinity for one-chain t-PA, both binding curves being consistent with one class of binding sites. However, binding of the fragments with two-chain t-PA was distinguished by more than one class of binding sites, with fragment E3 having the highest affinity for this form of the activator. epsilon-Aminocaproic acid, even at 50 mmol/L concentration, had only minimal effect on binding of (DD)E complex or fragment DD to either one-chain or two-chain t-PA. The potentiating effect of fibrin fragments on plasminogen activation by t-PA was measured by a chromogenic substrate assay. Fragment DD was the most effective stimulator of plasminogen activation by t-PA. In conclusion, (DD)E complex and fragment DD retained most of the regulatory functions of fibrin, which included t-PA binding and t-PA-mediated acceleration of plasminogen activation to plasmin.     

Activation of calcineurin by limited proteolysis.

Calcineurin, a heterodimer of calcineurin B, a 19,000 Mr Ca2+-binding subunit, and calcineurin A, a 61,000 Mr calmodulin-binding subunit, was previously proposed to be a calmodulin- and Ca2+-regulated protein phosphatase. Like other calmodulin-stimulated enzymes, calcineurin can be activated and rendered calmodulin- and Ca2+-independent by limited proteolysis. By glycerol gradient centrifugation, the native enzyme has a s20,w of 4.5 S in EGTA and 5 S in the presence of Ca2+-calmodulin. Under the same conditions, the s20,w of the trypsin-activated enzyme (4.3 S) is not affected by Ca2+ and calmodulin. The trypsin-treated enzyme is a heterodimer of calcineurin B and a 45,000 Mr fragment of calcineurin A that has lost its ability to interact with calmodulin. Phosphatase activity sediments with calcineurin or its proteolytic fragments, providing further evidence that calcineurin is indeed a protein phosphatase. Calmodulin protects calcineurin against tryptic digestion; proteolysis occurs more slowly, yielding fragments with Mr 57,000, 55,000, and 54,000 that have preserved their ability to interact with calmodulin. After trypsin treatment in the presence of calmodulin, the protein phosphatase activity of calcineurin is still regulated by calmodulin. Prolonged trypsin treatment in the presence of calmodulin produces a 46,000 Mr fragment. Unlike the fragments generated in the absence of calmodulin, this 46,000 Mr fragment still interacts weakly with calmodulin. Thus, calcineurin, like other calmodulin-regulated enzymes, consists of a catalytic domain resistant to proteolysis and a calmodulin-binding regulatory domain susceptible to protease action in the absence of calmodulin but not in its presence. In the absence of calmodulin, the regulatory domain exerts an inhibitory effect on the catalytic domain; the inhibition is relieved upon calmodulin binding to or tryptic degradation of the regulatory domain.     

Principles of quasi-equivalence and Euclidean geometry govern the assembly of cubic and dodecahedral cores of pyruvate dehydrogenase complexes

The pyruvate dehydrogenase multienzyme complex (M_r of 5–10 million) is assembled around a structural core formed of multiple copies of dihydrolipoyl acetyltransferase (E2p), which exhibits the shape of either a cube or a dodecahedron, depending on the source. The crystal structures of the 60-meric dihydrolipoyl acyltransferase cores of Bacillus stearothermophilus and Enterococcus faecalis pyruvate dehydrogenase complexes were determined and revealed a remarkably hollow dodecahedron with an outer diameter of ≈237 Å, 12 large openings of ≈52 Å diameter across the fivefold axes, and an inner cavity with a diameter of ≈118 Å. Comparison of cubic and dodecahedral E2p assemblies shows that combining the principles of quasi-equivalence formulated by Caspar and Klug [Caspar, D. L. & Klug, A. (1962) Cold Spring Harbor Symp. Quant. Biol. 27, 1–4] with strict Euclidean geometric considerations results in predictions of the major features of the E2p dodecahedron matching the observed features almost exactly.     

Sequence specificity of actinomycin D and Netropsin binding to pBR322 DNA analyzed by protection from DNase I.

A direct approach to determining the sequence specificities of equilibrium binding drugs by using the DNase protection technique is described. The method utilizes singly end-labeled restriction fragments and partial digestion of the drug fragment complex with DNase I. Microdensitometry of autoradiograms produced after electrophoretic separation of digestion products allows determination of sequences that are affected by drug binding. The feasibility of the technique for locating small ligands bound to DNA and its eventual use as a quantitative thermodynamic approach to studying ligand binding to heterogeneous DNA as a function of sequence is illustrated by using actinomycin D and Netropsin.