Immunoreactivity of subunits of the (Na+ + K+)-ATPase Cross-reactivity of the α,α+ and β forms in different organs and species

The immunologic cross-reactivity of the α and α+ forms of the large subunit and the β subunit of the ($\text{Na}{}^{\mathrm{+}}\text{+}\text{K}{}^{\mathrm{+}}\text{)-}\text{ATPase}$ from brain and kidney preparations was examined using rabbit antiserum prepared against the purified holo lamb kidney enzyme. As previously reported by Sweadner ((1979) J. Biol. Chem. 254, 6060–6067) phosphorylation of the large subunit of the ($\text{Na}{}^{\mathrm{+}}\text{+}\text{K}{}^{\mathrm{+}}\text{)-}\text{ATPase}$ in the presence of Na⁺, Mg²⁺, and [$\text{γ-}{}^{32}\text{P}\text{]}\text{ATP}$ revealed that dog and, very likely, rat brain contain two forms of the large subunit (designated α and α+) while dog, rat, and lamb kidney contain only one form (α). The cross-reactivity of the α and α+ forms in these preparations was investigated by resolving the subunits by SDS-polyacrylamide gel electrophoresis. The separated polypeptides were transferred to unmodified nitrocellulose paper, and reacted with rabbit anti-lamb kidney serum, followed by detection of the antigen-antibody complex with ¹²⁵I-labeled protein A and autoradiography. By this method, the α and α+ forms of rat and dog brain, as well as the α form found in kidney, were shown to cross-react. In addition, membranes from human cerebral cortex were shown to contain two immunoreactive bands corresponding to the α and α+ forms of dog brain. In contrast, the brain of the insect Manduca sexta contains only one immunoreactive polypeptide with a molecular weight intermediate to the α and α+ forms of dog brain. The β subunit from lamb, dog and rat kidney and from dog and rat brain cross-reacts with anti-lamb kidney ($\text{Na}{}^{\mathrm{+}}\text{+}\text{K}{}^{\mathrm{+}}\text{)-}\text{ATPase}$ serum. The mobility of the β subunit from dog and rat brain on SDS-polyacrylamide electrophoresis gels is greater than the mobility of the β subunit from lamb, rat or dog kidney.     

The isolation and analysis of the luminal plasma membrane of calf urinary bladder epithelium

The luminal plasma membrane of calf urinary bladder epithelium (urothelium) has been isolated by a method designed to preserve enzymic activity as well as structural integrity. The yield was about 80 μg per calf bladder. Low levels of 5′ nucleotidase, Mg²⁺-ATPase and ($\text{Na}{}^{\mathrm{+}}\text{+}\text{K}{}^{\mathrm{+}}\text{)-}\text{ATPase}$ activities were found in the luminal membrane fraction. Cerebroside was the major lipid present and dodecyl sulphate gel electrophoresis revealed a complex protein and glycoprotein composition in the whole membrane. A membrane fraction consisting of only the plaque areas was shown to have a simpler protein composition with major polypeptides of apparent $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 12 000 and 22 000. These may associate to form a 30 000 apparent $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ complex which could represent the individual ‘particles’ of the dodecameric subunits seen by electron microscopy in the plaque regions.     

Transport kinetics of dipicrylamine through lipid bilayer membranes. Effects of membrane structure

Charge-pulse current-relaxation studies have been performed with lipid bilayer membranes in the presence of the hydrophobic ion dipicrylamine. From the analysis of the relaxation times and amplitudes the translocation rate constant $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ of dipicrylamine as well as the partition coefficient β between membrane surface and water could be evaluated. In a first series of experiments membranes made from monoolein or dioleoylphosphatidylcholine in a number of different $\text{n-}\text{alkane}$ solvents were studied, as well as virtually solvent-free bilayer membranes made from monolayers. The thickness $\text{d}$ of the hydrocarbon layer of these membranes varied between 5.0 and 2.5 nm. While β was almost insensitive to variations in $\text{d}$, a strong decrease of $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ with increasing membrane thickness was found; the observed dependence of $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ on $\text{d}$ approximately agreed with the theoretically expected influence of membrane thickness on the height of the dielectric barrier. No specific differences between Mueller-Rudin films and solvent-free (Montal-Mueller) membranes other than differences in thickness were found. In a further series of experiments the chemical structure of the lipid was systematically varied (number and position of double bonds in the hydrocarbon chain, nature of the polar head group). The translocation rate constant $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ was much larger in phosphatidylethanolamine membranes than in phosphatidylcholine membranes. A strong increase of $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ was found when the number of double bonds in the hydrocarbon chain was increased from one to three. These changes were discussed in terms of membrane fluidity and dielectric barrier height. Much higher values of $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ were observed in lipids with ester linkage between hydrocarbon chain and glycerol backbone, as compared with the corresponding ether analogs. This finding is qualitatively consistent with determinations of dipolar potentials in monolayers of ester and ether lipids. When cholesterol is added to phosphatidylcholine membranes, the translocation rate constant $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ increases up to five-fold, while the partition coefficient β remains virtually constant. The variation of $\text{k}{}_{\mathrm{<mtext>i</mtext>}}$ in this case can be largely accounted for by a decrease in membrane thickness and a concomitant reduction in dielectric barrier height. In membranes made from the negatively charged lipid phosphatidylserine the partition coefficient of dipicrylamine strongly increased with ionic strength, as expected from the Gouy-Chapman theory of the surface potential.     

Determination of the distribution of protein and nucleic acid in the 70 S ribosomes of Escherichia coli and their 30 S subunits by neutron scattering.

Neutron small-angle scattering of the 70 S Escherichia coli ribosomes and of its smaller 30 S subunit has been measured in $\text{H}{}_2\text{O}{}^2\text{H}{}_2\text{O}$ mixtures. A linear dependence of the square of the radius of gyration on the reciprocal of the contrast is found, which is qualitatively similar to the results from contrast variation with the larger 50 S subunit. The slope α in this plot is a measure of radial segregation of RNA and proteins. It is most pronounced with the 50 S subunit. The 30 S particle appears to be more homogeneous, whereas the 70 S ribosome assumes an intermediate value of α. Neither the 30 S and 50 S subunits nor the 70 S ribosome show a significant separation of the centres of mass of their RNA part and proteins. A quantitative comparison of the parameters obtained suggest that the interaction between the two subunits and the 70 S ribosomes does not involve any major change in the latter.     

Analysis of the acid polysaccharides from squid cranial cartilage and examination of a novel polysaccharide

The polysaccharides of cranical cartilge were isolated by ethanol precipitation after papain digestion and β-elimination procedures and were fractionated chromatographically on CPC-cellulose. In addition to the previously described, heavily oversulphated chondroitin sulphate, the tissue contained small amounts of hyaluronic acid, which, however, co-eluted with the chondroitin sulphate from the CPC-cellulose. Approx. 20% of the isolated polysaccharides consisted of an acidic polysaccharide which to our knowledge is not previously described. This polysaccharide consists mainly of glucuronic acid, galactose and mannose in a molar ratio of 1:2:1. Gel chromatography of the preparation indicated a polydisperse molecule with an apparent average molecular weight of 39200 on weight basis $\text{(}\text{M}{}_{\mathrm{<mtext>w</mtext>}}\text{)}$ and 31400 on number basis $\text{(}\text{M}{}_{\mathrm{<mtext>n</mtext>}}\text{)}$.     

Major intrinsic polypeptide of lens membrane: Biochemical and immunological characterization of the major cyanogen bromide fragment

A protein of $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 has been shown to be the major component of eye-lens junctions, which are similar but not identical to the gap junctions of liver and other tissues. Cyanogen bromide cleavage of the $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 polypeptide from bovine lenses yields a major fragment of $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 15 000 (fragment 1). However, if the junctions are first treated with trypsin or carboxypeptidase Y, cyanogen bromide treatment yields a fragment of reduced molecular weight. Since protease treatment has been shown to cleave residues almost exclusively from the carboxy-terminal end of the $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 polypeptide, it follows that fragment 1 represents the carboxy-terminal half of this molecule, part of which is exposed to proteolytic attack outside the membrane. This latter result is corroborated by the fact that antisera which recognize both the $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 polypeptide and fragment 1 fail to do so after preadsorption with intact membranes. In addition, comparative amino acid and partial sequence analyses of the $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 polypeptide and fragment 1 indicate that fragment 1 is more hydrophilic in character, suggesting that much of the amino-terminal half of the $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ 26 000 polypeptide is buried within the lipid bilayer.     

Isolation of bovine cytochrome c1 as a single non-denatured subunit using gel filtration or high pressure liquid chromatography in deoxycholate

The non-denatured cytochrome $\text{c}{}_1$ subunit of bovine ubiquinone-cytochrome $\text{c}$ reductase was isolated using either gel filtration or high pressure liquid chromatography in 1% deoxycholate. The preparation was a single band on polyacrylamide gel electrophoresis in dodecyl sulfate, had a heme content of 31 nmol heme/mg protein, had an absorbance ratio $\text{A}{}_{417}\text{A}{}_{278}\text{= 2.65}$, a visible spectrum with maxima at 553, 530, 523.5, 417, 317, and 277 nm for the reduced protein, and an amino acid analysis identical to that previously reported for the isolated denatured protein. The Stokes' radius of this non-denatured deoxycholate solubilized protein was 34Å, indicating that the protein either is a dimer in deoxycholate, is asymmetric, or binds large amounts of detergent.     

Elasnin,a new human granulocyte elastase inhibitor produced by a strain of Streptomyces

Elasnin, a new human granulocyte elastase inhibitor, has been isolated from $\text{Streptomyces}\text{noboritoensis}$ KM-2753. Elasnin is a neutral, lipophilic colorless and viscous oil ($\text{n}{}_{\mathrm{D}}{}^{17}\text{=1.4983}$, $\text{[α]}{}_{\mathrm{D}}{}^{18}\text{?0.9°}$, λ_max^EtOH 291 nm (ε, 7760)). The molecular formula was C₂₄H₄₀O₄ (M.W.: 392) as determined by its elemental analysis and mass spectrum. Elasnin inhibits markedly human granulocyte elastase, but is almost ineffective for pancreatic elastase, trypsin, chymotrypsin, thermolysin and papain.     

Crystallographic studies of immunoglobulins: crystallization of the Fc fragment of rabbit IgG with and without cleavage of the inter-chain disulphide bridge

The Fc fragment was prepared from rabbit immunoglobulin G by digestion with papain, both with the inter-chain disulphide bond intact, and after reduction and alkylation. These two types of Fc crystallized in different, yet related forms, each with one dimer in the asymmetric unit. The covalently linked dimer crystallized in space group $\text{P2}{}_1\text{; a = 68.85 ± 0.05}\text{A}\text{?}\text{, b = 72.50 ± 0.05}\text{A}\text{?}\text{, c = 60.40 ± 0.05}\text{A}\text{?}$ and β = 105.1 ± 0.2 °. The reduced, non-covalently linked dimer also crystallized in space group $\text{P2}{}_1\text{; a = 81.55 ± 0.05}\text{A}\text{?}\text{, b = 55.65 ± 0.05}\text{A}\text{?}\text{, c = 68.85 ± 0.05}\text{A}\text{?}$ and β = 1051 ± 0.2 °. A non-crystallographic 2-fold axis relating the two identical polypeptide chains is clearly visible in the h0l projection of the second crystal form.     

Regulation of calcium accumulation and efflux from platelet vesicles: Possible role for cyclic-AMP-dependent phosphorylation and calmodulin

Calcium-accumulating vesicles were isolated by differential centrifugation of sonicated platelets. Such vesicles exhibit a $\text{(Ca}{}^{\mathrm{2+}}\text{+ Mg}{}^{\mathrm{2+}}\text{)-ATPase}$ activity of about 10 nmol (min·mg)^?1 and an ATP-dependent Ca²⁺ uptake of about 10 nmol (min·mg)^?1. When incubated in the presence of Mg[γ-³²P]ATP, the pump is phosphorylated and the acyl phosphate bond is sensitive to hydroxylamine. The [³²P]phosphate-labeled Ca²⁺ pump exhibits a subunit molecular weight of 120 000 when analyzed by lithium dodecyl sulfate-polyacrylamide gel electrophoresis. Platelet calcium-accumulating vesicles contain a 23 kDa membrane protein that is phosphorylatable by the catalytic subunit of cAMP-dependent protein kinase but not by protein kinase C. This phosphate acceptor is not phosphorylated when the vesicles are incubated in the presence of either Ca²⁺ or Ca²⁺ plus calmodulin. The latter protein is bound to the vesicles and represents 0.5% of the proteins present in the membrane fraction. Binding of ¹²⁵I-labeled calmodulin to this membrane fraction was of high affinity (16 nM), and the use of an overlay technique revealed four major calmodulin-binding proteins in the platelet cytosol ($\text{M}{}_{\mathrm{r}}\text{= 94 000, 87 000, 60 000 and 43 000}$). Some minor calmodulin-binding proteins were enriched in the membrane fractions ($\text{M}{}_{\mathrm{r}}\text{= 69 000, 57 000, 39 000 and 37 000}$). When the vesicles are phosphorylated in the presence of MgATP and of the catalytic subunit of cAMP-dependent protein kinase, the rate of Ca²⁺ uptake is essentially unaltered, while the Ca²⁺ capacity is diminished as a consequence of a doubling in the rate of Ca²⁺ efflux. Therefore, the inhibitory effect of cAMP on platelet function cannot be explained in such simple terms as an increased rate of Ca²⁺ removal from the cytosol. Calmodulin, on the other hand, was observed to have no effect on the initial rate of calcium efflux when added either in the absence or in the presence of the catalytic subunit of the cyclic AMP-dependent protein kinase, nor did the addition of 0.5 μM calmodulin result in increased levels of vesicle phosphorylation.     

CNDO/2 calculations of the relative stability of the goniomers in poly(dl-alanine)

Goniomers of single-stranded poly(dl-alanine) $\text{π}{}_{\mathrm{<mtext>dl</mtext>}}$ helices were treated by the semiempircal CNDO/2 MO procedure. The α_R and $\text{α}{}_{\mathrm{<mtext>dl</mtext>}}$ forms were also calculated. From the calculated total energies and partitioned energies we discuss the conformational stablity of goniomers of poly(dl-alanines). The conformational stability of the α and π forms is also compared.     

Abnormal neutrophil myeloperoxidase from a patient with chronic myelocytic leukemia

Neutrophil myeloperoxidase from a patient with chronic myelocytic leukemia was isolated under conditions designed to minimize proteolysis. Those methods yielded an α₂β₂ form of myeloperoxidase from normal individuals. Purified enzyme from the patient had electronic absorbances ($\text{A}{}_{430}\text{A}{}_{280}\text{= 0.85}$), enzymatic activity, and electrophoretic and Chromatographic behavior indistinguishable from that of normal myeloperoxidase. Edman degradation and physical studies after reduction and denaturation, however, showed that as compared to normal enzyme, one 55,000-dalton α subunit of the patient's myeloperoxidase was replaced by a 39,000-dalton peptide with a different amino-terminal sequence, a mixture of smaller peptides, and an heme derivative. Myeloperoxidase from the leukemic neutrophils appeared to have been partially degraded in vivo by lysosomal proteases.     

Pulse-label analysis and mapping of the two terminal regions of asynchronous complementary strand replication of mitochondrial DNA in transformed hamster cells

In vivo pulse-chase radioactive labeling studies were performed to localize within the physical map of $\text{C}{}_{13}\text{B}{}_4$ hamster mtDNA² the two terminal regions of heavy and light complementary strand synthesis. These terminal segments have been defined operationally as that region on the H- and L-strand that is synthesized last. mtDNA of monolayer cultures was pulsed with [³H]thymidine for a minimum period of 10 minutes, which is about one-tenth of one round of mtDNA synthesis, followed by chase periods of up to 120 minutes. The properties of the labeled closed circular replicative intermediates E-mtDNA, C-mtDNA and D-mtDNA were analyzed in $\text{CsCl}\text{PrI}{}_2$ gradients and in neutral sucrose velocity and alkaline CsCl gradients. Both terminally labeled α and β daughter molecules were found to pass through the E-mtDNA stage. Sensitivity of $\text{C}{}_{13}\text{B}{}_4$ mtDNA to alkali and ribonuclease A indicated the presence of covalently linked ribonucleotides. The distribution (specific activity) of pulse-chase radioactivity relative to uniform label was followed in electrophoretically separated HpaII + HinIII and HpaI restriction fragments (freed of 7 S initiation sequences) and corrected for thymine content. The strand specificity of the pulse-label was determined by hybridization of restriction fragments with H- and L-strands of mtDNA. The kinetic data agree precisely with electron microscopic determinations of H- and L-strand origins at respective genome positions of 0 and 67 ± 3, which are located on HpaII + HindIII fragments 9 and 6, respectively (Nass, 1980). The two terminal regions are within the predicted genome sector between about 67 and 100/0 map units; the highest terminal pulse-chase radioactivity extends within 5 to 15% of the genome's length behind each origin. The kinetics of early labeling events were found to differ at the two termini. The evidence indicates that the majority of L-strand initiation/termination sites are in the region near map position 67 ± 3, and confirms the highly asynchronous replication mechanism of this DNA.     

Conformational states of enkephalins in solution.

Chicken erythrocyte chromatin was partially digested with micrococcal nuclease and separated into multimeric subunit fractions by gel permeation chromatography. The fractions were characterized by their Svedberg constant, diffusion coefficient, circular dichroism, and electrophoresis pattern of the extracted DNA. The molecular weight dependence of the sedimentation coefficient was found to be S_20,w = .011 × M^.554. The molecular weight dependence of $\text{rmf}\text{f}{}_{\mathrm{o}}$ is best represented in the Kirkwood theory by either a helical superstructure $\text{or}$ a flexible coil $\text{with}\text{attractive}\text{interactions}$ between nucleosome units. The dimer calculations of $\text{f}\text{f}{}_{\mathrm{o}}$ suggest that the core particles are separated by spacer regions which contribute up to ～20% of the frictional properties of the molecule.     

Mechanism of action of thrombin on fibrinogen: Reaction of the N-terminal CNBr fragment from the Bβ chain of bovine fibrinogen with bovine thrombin

The N-terminal cyanogen bromide fragment from the Bβ chain of bovine fibrinogen was isolated, and its molecular weight was estimated to be approximately 14,000–15,500. The ratio of the Michaelis-Menten constants, $\text{k}{}_{\mathrm{<mtext>cat</mtext>}}\text{K}{}_{\mathrm{m}}$, for its hydrolysis by bovine thrombin was found to be 3 × 10^?7 [(NIH unit/liter)s]^?1, indicating that the Bβ fragment is a poor substrate for thrombin compared to the corresponding Aα chain fragment. This value of $\text{k}{}_{\mathrm{<mtext>cat</mtext>}}\text{K}{}_{\mathrm{m}}$ is too small to account for the rate of release of fibrinopeptide B from fibrinogen by thrombin. It is suggested that, while the Aα chain contains all of the amino acid residues necessary to interact with thrombin, the Bβ chain does not; i.e., some of the binding sites that are used in the hydrolysis of the Bβ chain are assumed to be located on either the α or γ chains of fibrinogen. An alternative hypothesis is that, after the Bβ chain fragment is removed from the fibrinogen molecule, it does not have the proper conformation to be hydrolyzed by thrombin.     

Ligand-dependent tryptic inactivation of the ouabain sensitivity of ADP-ATP exchange catalyzed by canine renal Na+, K+-ATPase.

Phosphate transporter of bovine heart mitochondria was purified by solubilization of submitochondrial particles with octylglucoside and fractionation of the extract with ammonium sulfate. After reconstitution into liposomes the purified protein catalyzed phosphate transport which was sensitive to mersalyl and other SH reagents. Transport measured either as $\text{P}{}_{\mathrm{i}}\text{OH}$ or $\text{P}{}_{\mathrm{i}}\text{P}{}_{\mathrm{i}}$ exchange was proportional to protein concentration and time. The $\text{P}{}_{\mathrm{i}}\text{OH}$ but not the $\text{P}{}_{\mathrm{i}}\text{P}{}_{\mathrm{i}}$ exchange was stimulated several fold by valinomycin plus nigericin in the presence of K⁺. The reconstituted system provides a suitable assay during purification of the mitochondrial phosphate transporter.     

Subunits of RNA polymerase in function and structure: VI. Sequence of the assembly in vitro of Escherichia coli RNA polymerase

Detailed analysis of the assembly in vitro of Escherichia coli RNA polymerase reveals that core enzyme subunits are assembled in the following sequence: $\text{2 α → α}{}_2\text{→}\text{β}\text{α}{}_2\text{β}\text{→}\text{β′}\text{α}{}_2\text{ββ′}$(premature core) → E (active core). Activation of the premature core enzyme, the rate-limiting step in this sequence, can be achieved in three different ways: self-reactivation, sigma subunit (σ or σ′)-promoted reactivation, and DNA-promoted reactivation.Although there has been disagreement on the enhancement of core enzyme maturation by sigma subunit or DNA, the discrepancy is resolved by the present finding that the premature core alone can be activated in the presence of high concentrations of salt or glycerol, whereas at a salt concentration as low as that in vivo, sigma subunit or DNA is required for maximum activation. However, the question remains unsolved as to which of the three ways operates in the in vivo process of RNA polymerase formation.     

Soluble and enzymatically stable (Na++K+)-ATPase from mammalian kidney consisting predominantly of protomer αβ-units: Preparation,assay and reconstitution of active Na+, K+transport

Soluble $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$ consisting predominantly of αβ-units with $\text{M}{}_{\mathrm{<mtext>r</mtext>}}$ below 170 000 was prepared by incubating pure membrane-bound $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$ (35–48 μmol P_i/min per mg protein) from the outer renal medulla with the non-ionic detergent dodecyloctaethyleneglycol monoether (C₁₂E₈). $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$ and potassium phosphatase remained fully active in the detergent solution at C₁₂E₈/protein ratios of 2.5–3, at which 50–70% of the membrane protein was solubilized. The soluble protomeric $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$ was reconstituted to Na⁺, K⁺ pumps in phospholipid vesicles by the freeze-thaw sonication procedure. Protein solubilization was complete at C₁₂E₈/protein ratios of 5–6, at the expense of partial inactivation, but $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$ and potassium phosphatase could be reactivated after binding of C₁₂E₈ to Bio-Beads SM2. At C₁₂E₈/protein ratios higher than 6 the activities were irreversibly lost. Inactivation could be explained by delipidation. It was not due to subunit dissociation since only small changes in sedimentation velocities were seen when the C₁₂E₈/protein ratio was increased from 2.9 to 46. As determined immediately after solubilization, $\text{S}{}_{\mathrm{<mtext>20,</mtext><mtext>w</mtext>}}$ was 7.4 S for the fully active $\text{(Na}{}^{\mathrm{+}}\text{+K}{}^{\mathrm{+}}\text{)-ATPase}$, 7.3 S for the partially active particle, and 6.5 S for the inactive particle at high C₁₂E₈/protein ratios. The maximum molecular masses determined by analytical ultracentrifugation were 141 000–170 000 dalton for these protein particles. Secondary aggregation occurred during column chromatography, with formation of enzymatically active $\text{(αβ)}{}_2\text{-}\text{dimers}$ or $\text{(αβ)}{}_3\text{-}\text{trimers}$ with $\text{S}{}_{\mathrm{<mtext>20,</mtext><mtext>w</mtext>}}\text{=10–12}$ S and apparent molecular masses in the range 273 000–386 000 daltons. This may reflect non-specific time-dependent aggregation of the detergent micelles.