Mechanism of Conversion of the Salmonella O Antigen by Bacteriophage ε34

The structural determinants for antigen 34 in the E group salmonella are glucosyl substituents on the galactosyl units of the O antigen which has a mannosylrhamnosylgalactose repeating sequence. The temperate bacteriophage epsilon(34) brings about the production of antigen 34. It has been shown here that glucose is transferred from uridine diphosphate glucose to the O antigen via a glucosyl-lipid intermediate in a two-step reaction. Glucose is linked through carbon 1 to the lipid by a phosphodiester bridge, the glucosyl bond having the beta-anomeric configuration. The lipid is a C(55)-polyisoprenoid alcohol, each isoprene unit having one double bond. It is the same lipid which is involved in the synthesis of the O antigen repeating sequence.     

Assembly of DNA polymerase delta and epsilon holoenzymes depends on the geometry of the DNA template.

To study in details the assembly of DNA polymerases delta and epsilon holoenzymes a circular double-stranded DNA template containing a gap of 45 nucleotides was constructed. Both replication factor C and proliferating cell nuclear antigen were absolutely required and sufficient for assembly of DNA polymerase delta holoenzyme complex on DNA. On such a circular DNA substrate replication protein A (or E. coli single-strand DNA binding protein) was neither required for assembly of DNA polymerase delta holoenzyme complex nor for the gap-filling reaction. A circular structure of the DNA substrate was found to be absolutely critical for the ability of auxiliary proteins to interact with DNA polymerases. The linearization of the circular DNA template resulted in three dramatic effects: (i) DNA synthesis by DNA polymerase delta holoenzyme was abolished, (ii) the inhibition effect of replication factor C and proliferating cell nuclear antigen on DNA polymerase alpha was relieved and (iii) DNA polymerase epsilon could not form any longer a holoenzyme with replication factor C and proliferating cell nuclear antigen. The comparison of the effect of replication factor C and proliferating cell nuclear antigen on DNA polymerases alpha, delta and epsilon indicated that the auxiliary proteins appear to form a mobile clamp, which can easily slide along double-stranded DNA.     

DNA polymerase epsilon: the latest member in the family of mammalian DNA polymerases

DNA polymerase epsilon is a mammalian polymerase that has a tightly associated 3'----5' exonuclease activity. Because of this readily detectable exonuclease activity, the enzyme has been regarded as a form of DNA polymerase delta, an enzyme which, together with DNA polymerase alpha, is in all probability required for the replication of chromosomal DNA. Recently, it was discovered that DNA polymerase epsilon is both catalytically and structurally distinct from DNA polymerase delta. The most striking difference between the two DNA polymerases is that processive DNA synthesis by DNA polymerase delta is dependent on proliferating cell nuclear antigen (PCNA), a replication factor, while DNA polymerase epsilon is inherently processive. DNA polymerase epsilon is required at least for the repair synthesis of UV-damaged DNA. DNA polymerases are highly conserved in eukaryotic cells. Mammalian DNA polymerases alpha, delta and epsilon are counterparts of yeast DNA polymerases I, III and II, respectively. Like DNA polymerases I and III, DNA polymerase II is also essential for the viability of cells, which suggests that DNA polymerase II (and epsilon) may play a role in DNA replication.     

Isolation and partial characterization of membrane vesicles carrying markers of the membrane adhesion sites.

At areas of adhesion between outer membrane (OM) and inner membrane (IM) in gram-negative bacteria, newly synthesized membrane constituents are inserted, and bacteriophage infection occurs. We describe here the isolation of these sites from cell membrane fractions of Salmonella anatum. Sucrose density gradients yielded membrane vesicles of the OM and IM; their mutual cross-contamination was low, as measured by 2-keto-3-deoxyoctonate and beta-NADH-oxidase activities. To mark the areas of lipopolysaccharide synthesis in the envelope (the adhesion sites), we infected S. anatum with phage epsilon 15, which causes a rapid change (conversion) in the cell's O-antigenic composition from serogroup E1 to E2; lipopolysaccharide of type E2 also serves as receptor for phage epsilon 34. We found that the fractions of intermediate density (Int. M) from briefly converted cells bound both phage epsilon 34 and E2-specific antibody. In the electron microscope, epsilon 34 was seen to have absorbed with a high degree of significance to the Int. M fraction of briefly converted cells, but not to the Int. M fraction of unconverted cells. Furthermore, the Int. M fractions of briefly converted cells coagglutinated anti-E2-coated Staphylococcus aureus, whereas the OM and IM fractions showed comparatively little agglutination. In addition, Int. M material exhibited elevated phospholipase A1 and A2 activities comparable to those of the OM fraction; the IM was essentially phospholipase free. Our data indicate that this membrane fractionation allows one to isolate from Int. M regions a variety of activities associated with adhesion sites.     

In vitro synthesis and assembly of picornaviral capsid intermediate structures.

Cell-free translation of encephalomyocarditis RNA in extracts of rabbit reticulocytes results in the synthesis of viral proteins indistinguishable from those produced during virus infection of cells. The viral capsid proteins are produced in an active form capable of assembly into viral capsid intermediate structures. Protomers (5S), pentamers (14S), and shell-like structures (75 to 85S) can be detected after prolonged incubation in the extracts. Proteolytic cleavage of capsid precursor proteins appears to be a prerequisite for assembly, in apparent contrast to cell-associated assembly. Assembly of pentamers is also preceded by conversion of protein epsilon 1 to epsilon in a step which may reflect an amino-terminal blocking reaction.     

The reaction between ferrous polyaminocarboxylate complexes and hydrogen peroxide: an investigation of the reaction intermediates by stopped flow spectrophotometry

The reactions of Fe(II)EDTA, Fe(II)DTPA, and Fe(II)HEDTA with hydrogen peroxide near neutral pH have been investigated. All these reactions have been assumed to proceed through an active intermediate, I1, (Formula: see text) where pac is one of the three polyaminocarboxylates mentioned above. I1, whether .OH radical or an iron complex, reacts with ethanol, formate, and other scavengers at rates relative to k2 that, with the exception of t-butanol and benzoate, are similar, but not identical, to those expected for the.OH radical. In contrast, at pH 3, in the absence of ligands the reaction of I1 with Fe2+ was inhibited by ethanol and t-butanol and the reactivity of I1 towards these two scavengers relative to ferrous ion is identical to that exhibited by the hydroxyl radical. When pac = HEDTA, the intermediate of the first reaction reacts with formate ion to form the ferrous HEDTA ligand radical complex, which is characterized by absorption maxima at 295 nm (epsilon = 2,640 M-1 cm-1) and 420 nm (epsilon = 620 M-1 cm-1). For the reaction of Fe(II)HEDTA with H2O2, the following mechanism is proposed: (Formula: see text) where k17 = 4.2 X 10(4) M-1 sec-1 and k19 = 5 +/- 0.2 sec-1.     

Probing electrostatic interactions along the reaction pathway of a glycoside hydrolase: histidine characterization by NMR spectroscopy

We have characterized by NMR spectroscopy the three active site (His80, His85, and His205) and two non-active site (His107 and His114) histidines in the 34 kDa catalytic domain of Cellulomonas fimi xylanase Cex in its apo, noncovalently aza-sugar-inhibited, and trapped glycosyl-enzyme intermediate states. Due to protection from hydrogen exchange, the level of which increased upon inhibition, the labile 1Hdelta1 and 1H epsilon1 atoms of four histidines (t1/2 approximately 0.1-300 s at 30 degrees C and pH approximately 7), as well as the nitrogen-bonded protons in the xylobio-imidazole and -isofagomine inhibitors, could be observed with chemical shifts between 10.2 and 17.6 ppm. The histidine pKa values and neutral tautomeric forms were determined from their pH-dependent 13C epsilon1-1H epsilon1 chemical shifts, combined with multiple-bond 1H delta2/epsilon1-15N delta1/epsilon2 scalar coupling patterns. Remarkably, these pKa values span more than 8 log units such that at the pH optimum of approximately 6 for Cex activity, His107 and His205 are positively charged (pKa > 10.4), His85 is neutral (pKa < 2.8), and both His80 (pKa = 7.9) and His114 (pKa = 8.1) are titrating between charged and neutral states. Furthermore, upon formation of the glycosyl-enzyme intermediate, the pKa value of His80 drops from 7.9 to <2.8, becoming neutral and accepting a hydrogen bond from an exocyclic oxygen of the bound sugar moiety. Changes in the pH-dependent activity of Cex due to mutation of His80 to an alanine confirm the importance of this interaction. The diverse ionization behaviors of the histidine residues are discussed in terms of their structural and functional roles in this model glycoside hydrolase.     

Gene Conversion, Linkage, and the Evolution of Multigene Families

The evolution of the probabilities of genetic identity within and between the loci of a multigene family is investigated. Unbiased gene conversion, equal crossing over, random genetic drift, and mutation to new alleles are incorporated. Generations are discrete and nonoverlapping; the diploid, monoecious population mates at random. The linkage map is arbitrary, and the location dependence of the probabilities of identity is formulated exactly. The greatest of the rates of gene conversion, random drift, and mutation is epsilon much less than 1. For interchromosomal conversion, the equilibrium probabilities of identity are within order epsilon [i.e., O(epsilon)] of those in a simple model that has no location dependence and, at equilibrium, no linkage disequilibrium. At equilibrium, the linkage disequilibria are of O(epsilon); they are evaluated explicitly with an error of O(epsilon 2); they may be negative if symmetric heteroduplexes occur. The ultimate rate and pattern of convergence to equilibrium are within O(epsilon 2) and O(epsilon), respectively, of that of the same simple model. If linkage is loose (i.e., all the crossover rates greatly exceed epsilon, though they may still be much less than 1/2), the linkage disequilibria are reduced to O(epsilon) in a time of O(-ln epsilon). If intrachromosomal conversion is incorporated, the same results hold for loose linkage, except that, if the crossover rates are much less than 1/2, then the linkage disequilibria generally exceed those for pure interchromosomal conversion.     

Iron-histidine stretching vibration in the deoxy state of insect hemoglobins with different O2 affinities and Bohr effects

Resonance Raman spectroscopy has been employed to detect the iron-proximal histidine stretching mode in deoxyhemoglobins from insect larvae of Chironomus thummi thummi (CTT). With the excitation of 413.1 nm, we observe a sharp and intense line in the 220-224 cm-1 region. The assignment of this line to the Fe-N epsilon (His) stretching mode was made on the basis of a 3-cm-1 shift upon 57Fe/54Fe isotope substitution. The Fe-N epsilon (His) vibration is used to monitor the possible changes in the Fe-N epsilon (His) bond strength (hence bone length) in the deoxy state of the monomeric (CTT I, III, and IV) and dimeric (CTT II) insect hemoglobins. As these hemoglobins differ in O2 affinity, off-rate and on-rate constants, and in the Bohr effect, they are excellent model systems for investigating the mechanism of protein control of the heme reactivity. Some of these hemoglobins (CTT III, IV, and II) are allosteric, exhibiting two interconvertible conformational states with high and low O2 affinity at high and low pH, respectively. The Fe-N epsilon (His) stretching frequency does not correlate with the O2 affinity, the on-rate and the off-rate constants for different hemoglobins, for different conformational states, and for modified hemoglobins with different heme peripheral groups. This vibrational mode is insensitive to deuteration of the heme vinyl groups. It is important to note that the Fe-N epsilon (His) bonds in the high pH (high-affinity) and the low pH (low-affinity) states are identical. This implies that the O2 molecule, prior to binding, "sees" identical binding sites. Thus, the difference in free energy changes upon O2 binding is manifested only in the oxy form.     

Observation of a novel transient ferryl complex with reduced CuB in cytochrome c oxidase

The reaction between mixed-valence (MV) cytochrome c oxidase from beef heart with H2O2 was investigated using the flow-flash technique with a high concentration of H2O2 (1 M) to ensure a fast bimolecular interaction with the enzyme. Under anaerobic conditions the reaction exhibits 3 apparent phases. The first phase (tau congruent with 25 micros) results from the binding of one molecule of H2O2 to reduced heme a3 and the formation of an intermediate which is heme a3 oxoferryl (Fe4+=O2-) with reduced CuB (plus water). During the second phase (tau congruent with 90 micros), the electron transfer from CuB+ to the heme oxoferryl takes place, yielding the oxidized form of cytochrome oxidase (heme a3 Fe3+ and CuB2+, plus hydroxide). During the third phase (tau congruent with 4 ms), an additional molecule of H2O2 binds to the oxidized form of the enzyme and forms compound P, similar to the product observed upon the reaction of the mixed-valence (i.e., two-electron reduced) form of the enzyme with dioxygen. Thus, within about 30 ms the reaction of the mixed-valence form of the enzyme with H2O2 yields the same compound P as does the reaction with dioxygen, as indicated by the final absorbance at 436 nm, which is the same in both cases. This experimental approach allows the investigation of the form of cytochrome c oxidase which has the heme a3 oxoferryl intermediate but with reduced CuB. This state of the enzyme cannot be obtained from the reaction with dioxygen and is potentially useful to address questions concerning the role of the redox state in CuB in the proton pumping mechanism.     

Spectroscopic and kinetic studies on the oxygen-centered radical formed during the four-electron reduction process of dioxygen by Rhus vernicifera laccase.

The oxygen-centered radical bound to the trinuclear copper center was detected as an intermediate during the reoxidation process of the reduced Rhus vernicifera laccase with dioxygen and characterized by using absorption, stopped-flow, and electron paramagnetic resonance (EPR) spectroscopies and by super conducting quantum interface devices measurement. The intermediate bands appeared at 370 nm (epsilon approximately 1000), 420 nm (sh), and 670 nm (weak) within 15 ms, and were observable for approximately 2 min at pH 7.4 but for less than 5 s at pH 4.2. The first-order rate constant for the decay of the intermediate has been determined by stopped-flow spectroscopy, showing the isotope effect, k(H)/k(D) of 1.4 in D(2)O. The intermediate was found to decay mainly from the protonated form by analyzing pH dependences. The enthalpy and entropy of activation suggested that a considerable structure change takes place around the active site during the decay of the intermediate. The EPR spectra at cryogenic temperatures (<27 K) showed two broad signals with g approximately 1.8 and 1.6 depending on pH. We propose an oxygen-centered radical in magnetic interaction with the oxidized type III copper ions as the structure of the three-electron reduced form of dioxygen.     

Characterization of an O-antigen bacteriophage from Aeromonas hydrophila.

A unique bacteriophage of Aeromonas hydrophila serotype O:34 was isolated, purified, and characterized. The bacterial surface receptor was shown to be the O-antigen polysaccharide component of lipopolysaccharide specific to serotype O:34, which was chemically characterized. The high molecular weight lipopolysaccharide fraction (a fraction enriched in O antigen) was fully able to inactivate bacteriophage PM1. Phage-resistant mutants of A. hydrophila O:34 were isolated and found to be specifically devoid of lipopolysaccharide O antigen. No other cell-surface molecules were involved in phage binding. The host range of bacteriophage PM1 was found to be very narrow, producing plaques only on A. hydrophila strains from serotype O:34.     

Structural Events in a Bacterial Uniporter Leading to Translocation of Glucose to the Cytosol

Among classes of sugar transporters, there exists a comparatively new family of transporters named SWEET transporters (semi-SWEETS in bacteria) that are uniport transmembrane proteins. It is hypothesized that sugar is transported from the extracellular side (via outward-open state) to intracellular side (inward-open state) through intermediate occluded state (both extracellular and intracellular gates closed). In our study, extensive unbiased all-atom molecular dynamics simulations were carried out with the outward-open and inward-open conformations to study this transition mechanism. We find that after 100?ns, the outward-open structure without sugar bound starts changing to the occluded form leading to closure of extracellular gates stabilized by electrostatic and hydrophobic interactions. Further simulations (up to 7?μs) have led to a transition toward the inward-open form and suggest that there exists more than one intermediate occluded conformation. We have also performed 5-μs simulations on the glucose-docked structure to identify the putative substrate-bound translocation pathway. Glucose binds to semi-SWEET with strong hydrogen bonds to Asn66 and Trp50. Comparative simulations of substrate bound, and unbound forms suggested that glucose, the putative substrate, facilitates relatively rapid conformational transitions. For the first time, we captured the release of glucose to the cytosol, in this family of transporters. We find that prior to release of glucose, the glucose forms interactions with polar residues near the intracellular gate which may facilitate its release. The distance between the residues Asn31 and Gly34 of the other protomer was found to play a decisive role in the transport of glucose to the cytoplasmic side.     

Drug resistance of enteric bacteria. V. High frequency of transduction of R factors with bacteriophage epsilon

Kameda, Mitsuo (Gunma University, Maebashi, Japan), Kenji Harada, Mitsue Suzuki, and Susumu Mitsuhashi. Drug resistance of enteric bacteria. V. High frequency of transduction of R factors with bacteriophage epsilon. J. Bacteriol. 90:1174-1181. 1965.-In the transduction of R factors with phage epsilon(15), a lysate capable of transducing the markers for (TC) or (CM.SM.SA) resistance at high frequency was obtained. The transducing agent is a defective element called epsilon(15)dR(23) which lacks certain functions of phage epsilon(15). After lysogenization with normal epsilon(15) phage and ultraviolet (UV) induction, strains carrying the epsilon(15)dR(23) element produce lysates which have a high frequency of transduction (HFT) on group E(1)Salmonella. Lytic lysates prepared on phage epsilon(15) sensitive strain with the epsilon(15)dR(23) element have a low frequency of transduction (LFT). Lytic growth of phage epsilon(34) on an epsilon(15)dR(23) strain or UV induction of an epsilon(34) lysogenic strain containing epsilon(15)dR(23) results in LFT lysates on group E(2)Salmonella. On UV induction, group E(2)Salmonella (epsilon(15) lysogens) with the epsilon(15)dR(23) element give lysates which are HFT on group E(1)Salmonella but are LFT when tested on group E(2)Salmonella. In all instances, the production of drug-resistant transductants requires infection of the cell with only a single epsilon(15)dR(23) element. It appears that the resistance region of the R factor has replaced that portion of phage genome which is essential for vegetative replication and superinfection immunity. The epsilon(15)dR(23) element does not contain the genetic determinants of the R factor responsible for transmissibility, inhibition of F mating, and interference between two R factors.     

Isoform-specific induction of PKC-epsilon by high glucose protects heart-derived H9c2 cells against hypoxic injury

We investigated which PKC isoforms are involved in high glucose-induced protection against hypoxic injury. Treatment for 48 h with high glucose (22 mM) markedly increased the expression of PKC- epsilon in the particulate fraction (213+/-22.1% of the control) but had no effect on other types of PKC isoforms, suggesting that the high glucose-induced increase in PKC expression is isoform-specific. The mRNA level for PKC- epsilon was also substantially increased, reaching its peak after 4h of high glucose treatment. The high glucose increased PKC-epsilon activity in the particulate fraction up to 183+/-32.2% of the control. During hypoxia, the amount of PKC-epsilon in the particulate fraction was remarkably diminished in the low glucose-treated cells, but remained at a higher level in high glucose-treated cells. The treatment with epsilon V1-2 (10 microM), a specific inhibitor of PKC epsilon, abolished the protective effect of high glucose against hypoxia. These results suggest that isoform-specific induction of PKC-epsilon is involved in high glucose-induced protection against hypoxic injury in heart-derived H9c2 cells.     

Analysis of Fc(epsilon)RI-mediated mast cell stimulation by surface-carried antigens

Clustering of the type I receptor for IgE (Fc[epsilon]RI) on mast cells initiates a cascade of biochemical processes that result in secretion of inflammatory mediators. To determine the Fc(epsilon)RI proximity, cluster size, and mobility requirements for initiating the Fc(epsilon)RI cascade, a novel experimental protocol has been developed in which mast cells are reacted with glass surfaces carrying different densities of both antigen and bound IgE, and the cell's secretory response to these stimuli is measured. The results have been analyzed in terms of a model based on the following assumptions: 1) the glass surface antigen distribution and consequently that of the bound IgE are random; 2) Fc(epsilon)RI binding to these surface-bound IgEs immobilizes the former and saturates the latter; 3) the cell surface is formally divided into small elements, which function as a secretory stimulus unit when occupied by two or more immobilized IgE-Fc(epsilon)RI complexes; 4) alternatively, similar stimulatory units can be formed by binding of surface-carried IgE dimers to two Fc(epsilon)RI. This model yielded a satisfactory and self-consistent fitting of all of the different experimental data sets. Hence the present results establish the essential role of Fc(epsilon)RI immobilization for initiating its signaling cascade. Moreover, it provides independent support for the notion that as few as two Fc(epsilon)RIs immobilized at van der Waals contact constitute an "elementary stimulatory unit" leading to mast cell (RBL-2H3 line) secretory response.     

Inactivation of glucose oxidase by diperoxovanadate-derived oxidants.

Inactivation of glucose oxidase occurred in the presence of bromide, vanadate, H(2)O(2), and phosphate (the bromide system), and this was prevented by NADH or phenol red, a bromine acceptor. Glucose oxidase present during the reaction between diperoxovanadate and a reduced form of vanadate, vanadyl (the vanadyl system), but not added after mixing the reactants, was inactivated, and this was accompanied by a loss of binding of the dye, Coomassie blue, to the protein. The transient intermediate of the type OVOOV(O(2)), known to form in these reactions and used in the oxidation of bromide ion and NADH, appears to be responsible for inactivating glucose oxidase. In both systems, the inactivation of the enzyme was prevented by histidine and DTT, known to quench singlet-oxygen. By direct measurement of 1270-nm emission of singlet-oxygen, its generation was demonstrated in the bromide system, and in the reaction of hypohalous acids with diperoxovanadate, but not in the vanadyl system. By themselves both hypohalous acids, HOCl and HOBr inactivated glucose oxidase, and their prior reaction with H(2)O(2) during which singlet-oxygen was released, protected the enzyme. The results provide support for possible oxidative inactivation of glucose oxidase by diperoxovanadate-derived oxidants.     

The effectiveness of different rat IgG subclasses as IgE-blocking antibodies in the rat basophil leukaemia cell model

The degranulation of mast cells in an allergic response is initiated by the aggregation of high-affinity IgE receptors (Fc epsilon RI) by IgE and antigen. Recently it has been shown that such degranulation can be inhibited by cross-linking Fc epsilon RI and low-affinity IgG receptors (Fc gamma RII) which are also expressed by mast cells. The ability of various monoclonal antibodies to block the degranulation of rat basophil leukaemia (RBL) cells sensitized with IgE antidinitrophenyl (DNP) antibodies has been investigated. Sensitized cells were challenged with immune complexes formed using varying concentrations of antigen, and of both high- and low-valency antigen. It is reported here that rat IgG1 antibodies, which are associated in the rat with a Th1-type response, act as highly effective blocking antibodies over a wide concentration range. Rat IgG2a antibodies, which are associated with a Th2-type response, were able only to inhibit degranulation when immune complexes were formed with very low concentrations of high-valency antigen (DNP32-HSA). Under these conditions, some inhibitory activity was seen with high-affinity murine IgA anti-DNP but not with low-affinity rat IgG2b anti-DNP antibody-containing immune complexes. In addition to this inhibitory activity, IgG2a antibodies were shown to be capable of inducing degranulation of cells via unoccupied Fc epsilon RI. These results demonstrate that blocking activity may arise via both inhibitory receptors and by masking of antigen.     

Glucose permease of Escherichia coli. Purification of the IIGlc subunit and functional characterization of its oligomeric forms

The membrane subunit (IIGlc) of the glucose permease has been purified from overproducing Escherichia coli. About 2 mg of pure protein was obtained from 10 g (wet weight) of cells. IIGlc of E. coli and Salmonella typhimurium are functionally indistinguishable. A small difference was revealed, however, by a monoclonal antibody which neutralizes glucose phosphorylation activity of IIGlc from S. typhimurium, but does not cross-react with IIGlc of E. coli. A dimeric form of purified IIGlc can be detected by chemical cross-linking and by zonal sedimentation at 4 degrees C. Upon mild oxidation a disulfide bond is formed between the subunits of the dimer. Oxidized IIGlc is more stable than the reduced form but is inactive because it cannot be phosphorylated by the cytoplasmic subunit (IIIGlc) of the glucose permease. Cys-421 could be identified as the oxidation-sensitive residue, using a novel assay to detect IIIGlc-dependent phosphorylation of nitrocellulose-bound IIGlc that has been purified by gel electrophoresis. No dimeric form of phosphorylated IIGlc could be detected. Because phosphorylated IIGlc is a catalytic intermediate it is concluded that catalytically active IIGlc is a monomer and that the dimeric form is an artefact observed only with purified resting IIGlc. That IIGlc is active as a monomer is further supported by the observation that monomeric IIGlc catalyzes phosphoryl exchange between glucose and glucose 6-phosphate at equilibrium and that an excess of inactive IIGlc with a serine replacing Cys-421 does not interfere with the activity of wild-type IIGlc as would be expected if interaction between the subunits in a dimer were essential for activity.