An artificial lipid bilayer formed on an agarose-coated glass for simultaneous electrical and optical measurement of single ion channels

The purpose of this study is to develop an apparatus for simultaneous measurement of electrical and spectroscopic parameters of single ion channels. We have combined the single channel recording apparatus with an artificial lipid bilayer and a fluorescence microscope designed to detect single fluorescent molecules. The artificial membranes were formed on an agarose-coated glass and observed with an objective-type total internal reflection fluorescence microscope (TIRFM). The lateral motion of a single lipid molecule (beta-BODIPY 530/550 HPC) was recorded. The lateral diffusion constant of the lipid molecule was calculated from the trajectories of single molecules as D = 8.5 +/- 4.9 x 10(-8) cm(2)/s. Ionic channels were incorporated into the membrane and current fluctuations were recorded at the single-channel level. After incorporation of Cy3-labeled alametithin molecules into the membrane, bright spots were observed moving rather slowly (D = 4.0 +/- 1.6 x 10(-8) cm(2)/s) in the membrane, simultaneously with the alametithin-channel current. These data show the possibility of the present technique for simultaneous measurement of electrical and spectroscopic parameters of single-channel activities.     

Exaggerated Mg2+ inhibition of Kir2.1 as a consequence of reduced PIP2 sensitivity in Andersen syndrome

Andersen syndrome is an autosomal dominant disorder characterized by cardiac arrhythmias, periodic paralysis and dysmorphic features. Many Andersen syndrome cases have been associated with loss-of-function mutations in the inward rectifier K(+) channel Kir2.1 encoded by KCNJ2. Using engineered concatenated tetrameric channels we determined the mechanism for dominant loss-of-function associated with a trafficking-competent missense mutation, Kir2.1-T74A. This mutation alters a conserved threonine residue in an N-terminal domain analogous to the slide helix identified in the structure of a bacterial inward rectifier. Incorporation of a single mutant subunit in channel tetramers was sufficient to cause a selective impairment of whole-cell outward current, but no difference in the level of inward current compared with wild-type (WT) tetramers. The presence of two mutant subunits resulted in greatly reduced outward and impaired inward currents. Experiments using excised inside-out membrane patches revealed that tetramers with one mutant subunit exhibited increased Mg(2+) inhibition. Additional experiments demonstrated that concatenated tetramers containing one T74A subunit had reduced PIP(2) sensitivity, and that outward current carried by mutant tetramers could be restored by addition of PIP(2) in the absence of Mg(2+). Our results are consistent with the involvement of the Kir2.1 N-terminus in PIP(2) modulation of channel activity and support the existence of an inverse relationship between PIP(2) sensitivity and Mg(2+) inhibition of Kir2.1 channels. Our data also indicate that a single mutant subunit is sufficient to explain dominant-negative behavior of Kir2.1-T74A in Andersen syndrome.     

T cell glycolipid-enriched membrane domains are constitutively assembled as membrane patches that translocate to immune synapses

In T cells, glycolipid-enriched membrane (GEM) domains, or lipid rafts, are assembled into immune synapses in response to Ag presentation. However, the properties of T cell GEM domains in the absence of stimulatory signals, such as their size and distribution in the plasma membrane, are less clear. To address this question, we used confocal microscopy to measure GEM domains in unstimulated T cells expressing a GEM-targeted green fluorescent protein molecule. Our experiments showed that the GEM domains were assembled into membrane patches that were micrometers in size, as evidenced by a specific enrichment of GEM-associated molecules and resistance of the patches to extraction by Triton X-100. However, treatment of cells with latrunculin B disrupted the patching of the GEM domains and their resistance to Triton X-100. Similarly, the patches were coenriched with F-actin, and actin occurred in the detergent-resistant GEM fraction of T cells. Live-cell imaging showed that the patches were mobile and underwent translocation in the plasma membrane to immune synapses in stimulated T cells. Targeting of GEM domains to immune synapses was found to be actin-dependent, and required phosphatidylinositol 3-kinase activity and myosin motor proteins. We conclude from our results that T cell GEM domains are constitutively assembled by the actin cytoskeleton into micrometer-sized membrane patches, and that GEM domains and the GEM-enriched patches can function as a vehicle for targeting molecules to immune synapses.     

Density and disposition of Ca2+-ATPase in sarcoplasmic reticulum membrane as determined by shadowing techniques.

We have studied the disposition of calcium ATPase in the native sarcoplasmic reticulum (SR) membrane of vertebrate muscles by rotary shadowing of freeze-dried isolated vesicles and of freeze-fractured in situ membranes. The predominant disposition of the ATPase molecules is disorderly, but small oligomers (dimers, tetramers, and occasionally larger aggregates) are seen. In vesicles from white hind legs of rabbits, the density of ATPase over nonjunctional SR is 31-34,000/microns2. ATPase density is always quite high, but small protein-free lipid patches may be interspersed with it.     

Visualization of the hexagonal lattice in the erythrocyte membrane skeleton

The isolated membrane skeleton of human erythrocytes was studied by high resolution negative staining electron microscopy. When the skeletal meshwork is spread onto a thin carbon film, clear images of a primarily hexagonal lattice of junctional F-actin complexes crosslinked by spectrin filaments are obtained. The regularly ordered network extends over the entire membrane skeleton. Some of the junctional complexes are arranged in the form of pentagons and septagons, approximately 3 and 8%, respectively. At least five forms of spectrin crosslinks are detected in the spread skeleton including a single spectrin tetramer linking two junctional complexes, three-armed Y-shaped spectrin molecules linking three junctional complexes, three-armed spectrin molecules connecting two junctional complexes with two arms bound to one complex and the third arm bound to the adjacent complex, double spectrin filaments linking two junctional complexes, and four-armed spectrin molecules linking two junctional complexes. Of these, the crosslinks of single spectrin tetramers and three-armed molecules are the most abundant and represent 84 and 11% of the total crosslinks, respectively. These observations are compatible with the presence of spectrin tetramers and oligomers in the erythrocyte membrane skeleton. Globular structures (9-12 nm in diameter) are attached to the majority of the spectrin tetramers or higher order oligomer-like molecules, approximately 80 nm from the distal ends of the spectrin tetramers. These globular structures are ankyrinor ankyrin/band 3-containing complexes, since they are absent when ankyrin and residual band 3 are extracted from the skeleton under hypertonic conditions.     

Single molecule recognition of protein binding epitopes in brush border membranes by force microscopy

Sidedness and accessibility of protein epitopes in intact brush border membrane vesicles were analyzed by detecting single molecule interaction forces using molecular recognition force microscopy in aqueous physiological solutions. Frequent antibody-antigen recognition events were observed with a force microscopy tip carrying an antibody directed against the periplasmically located gamma-glutamyltrans- peptidase, suggesting a right side out orientation of the vesicles. Phlorizin attached to the tips bound to NA+/D-glucose cotransporter molecules present in the vesicles. The recognition was sodium dependent and inhibited by free phlorizin and D-glucose, and revealed an apparent K(D) of 0.2 microM. Binding events were also observed with an antibody directed against the epitope aa603-aa630 close to the C terminus of the transporter. In the presence of phlorizin the probability of antibody binding was reduced but the most probable unbinding force f(u) = 100 pN remained unchanged. In the presence of D-glucose and sodium, however, both the binding probability and the most probable binding force (f(u) = 50 pN) were lower than in its absence. These studies demonstrate that molecular recognition force microscopy is a versatile tool to probe orientation and conformational changes of epitopes of membrane components during binding and trans-membrane transport.     

Photoactivation and dissociation of agonist molecules at the nicotinic acetylcholine receptor in voltage-clamped rat myoballs.

The photochemical properties of the azobenzene derivative, Bis-Q, were exploited to carry out an agonist concentration jump followed by a molecular rearrangement of bound agonist molecules at acetylcholine (ACh) receptor channels of voltage-clamped rat myoballs. Myoballs were bathed in solutions containing low concentrations of cis-Bis-Q, the inactive isomer. Whole-cell current relaxations were studied following a light flash that produced a concentration jump of agonist, trans-Bis-Q, followed by a second flash that produced net trans----cis photoisomerizations of Bis-Q molecules. The concentration-jump relaxation provided a measure of the mean burst duration for ACh receptor channels occupied by trans-Bis-Q (7.7 ms, 22 degrees C). The second current relaxation was a more rapid conductance decrease (phase 1, tau = 0.8 ms). Phase 1 may represent either the burst duration for receptors initially occupied by a single cis- and a single trans-Bis-Q molecule or that for unliganded receptors. Single-channel current recordings from excised outside-out membrane patches showed that single channels open following an agonist concentration jump comparable to that used in the whole-cell experiments; when many such records were averaged, a synthetic macroscopic relaxation was produced. Individual open channels closed faster following a flash that promoted trans----cis photoisomerizations of the bound ligand, thus confirming the whole-cell observations of phase 1.     

Exaggerated Mg2+ Inhibition of KCNJ2 as a Consequence of Reduced PIP2 Sensitivity in Andersen Syndrome

Andersen syndrome is an autosomal dominant disorder characterized by cardiac arrhythmias, periodic paralysis and dysmorphic features. Many Andersen syndrome cases have been associated with loss-of-function mutations in the inward rectifier K+ channel Kir2.1 encoded by KCNJ2. Using engineered concatenated tetrameric channels we determined the mechanism for dominant loss-of-function associated with a trafficking-competent missense mutation, Kir2.1-T74A. This mutation alters a conserved threonine residue in an N-terminal domain analogous to the slide helix identified in the structure of a bacterial inward rectifier. Incorporation of a single mutant subunit in channel tetramers was sufficient to cause a selective impairment of whole-cell outward current, but no difference in the level of inward current compared with wild-type (WT) tetramers. The presence of two mutant subunits resulted in greatly reduced outward and impaired inward currents. Experiments using excised inside-out membrane patches revealed that tetramers with one mutant subunit exhibited increased Mg2+ inhibition. Additional experiments demonstrated that concatenated tetramers containing one T74A subunit had reduced PIP2 sensitivity, and that outward current carried by mutant tetramers could be restored by addition of PIP2 in the absence of Mg2+. Our results are consistent with the involvement of the Kir2.1 N-terminus in PIP2 modulation of channel activity and support the existence of an inverse relationship between PIP2 sensitivity and Mg2+ inhibition of Kir2.1 channels. Our data also indicate that a single mutant subunit is sufficient to explain dominant-negative behavior of Kir2.1-T74A in Andersen syndrome.     

Pretransition and progressive softening of bovine carbonic anhydrase II as probed by single molecule atomic force microscopy

To develop a simple method for probing the physical state of surface adsorbed proteins, we adopted the force curve mode of an atomic force microscope (AFM) to extract information on the mechanical properties of surface immobilized bovine carbonic anhydrase II under native conditions and in the course of guanidinium chloride-induced denaturation. A progressive increase in the population of individually softened molecules was probed under mildly to fully denaturing conditions. The use of the approach regime of force curves gave information regarding the height and rigidity of the molecule under compressive stress, whereas use of the retracting regime of the curves gave information about the tensile characteristics of the protein. The results showed that protein molecules at the beginning of the transition region possessed slightly more flattened and significantly more softened conformations compared with that of native molecules, but were still not fully denatured, in agreement with results based on solution studies. Thus the force curve mode of an AFM was shown to be sensitive enough to provide information concerning the different physical states of single molecules of globular proteins.     

Potential role of NKG2D/MHC class I-related chain A interaction in intrathymic maturation of single-positive CD8 T cells

The nonclassical MHC class I molecule MHC class I-related chain A (MICA) interacts with the NKG2D receptor expressed at the surface of most peripheral CD8 T cells, gammadelta T cells, and NK cells. We investigated the role of MICA-NKG2D interactions in the selection or maturation of the T cell repertoire within the thymus using MICA tetramers and anti-MICA mAbs. MICA tetramers identified a small population of late stage CD8 single-positive, CD45RA(+) CD62L(+) CCR7(+) CD69(-) thymocytes, a phenotype compatible with that of fully mature CD8(+) cells ready to emigrate to the periphery as naive cells. MICA molecules were expressed in the outer layer of Hassal's corpuscles within the medulla of normal thymus. In thymomas, an overexpression of MICA in cortical and medullar epithelial cells was observed. This was associated with a decreased percentage of NKG2D-positive thymocytes, which expressed a less mature phenotype than in normal thymus. These results indicate that CD8(+) thymocytes up-regulate NKG2D as they complete their developmental program before leaving the thymic medulla to seed the periphery, and identify NKG2D as a potential regulator of the developmental processes in T cells that are essential for immune homeostasis.     

Letter: Electron microscopic visualization of the folded chromosome of Escherichia coli

The folded Escherichia coli chromosome has been analyzed in the electron microscope following the Kleinschmidt spreading technique. In all cases, the DNA molecule appears intact and supercoiled. In addition, the membrane-associated form of the chromosome shows DNA fibers attached to one or two membrane patches. The DNA molecule is intact, as evidenced by the absence of free ends. Two types of DNA coiling are generally observed: (1) spirals with about four to seven turns per spiral, and (2) stretched plectonemic supercoils, more abundant on the outside of the complex where the spreading has been more extensive.The released folded chromosome appears to be less compact; it extends over a larger area and frequently breaks with the spreading. The intact molecules which can be found show the DNA concentrated in few nueleation areas, with one or a few interconnecting DNA fibers. The folded DNA has very few or no single strand nicks, as evidenced by its extensive supertwisting at high ethidium bromide concentrations.     

A novel receptor on allograft (H-2d)-induced macrophage (H-2b) toward an allogeneic major histocompatibility complex class I molecule, H-2Dd, in mice

The generation of knockout mice demonstrated that noncytotoxic CD4(+), but not cytotoxic CD8(+), T cells were essential for the rejection of skin or organ allografts. Earlier we reported that allograftinduced macrophages (AIM) in mice lysed allografts with H-2 haplotype specificity, implying screening of grafts by AIM. Here, we isolated a cDNA clone encoding a novel receptor on AIM (H-2D(b)) for an allogeneic major histocompatibility complex (MHC) class I molecule, H-2D(d), by using H-2D(d) tetramer and a monoclonal antibody (mAb; R15) specific for AIM. The cDNA (1,181-bp) encoded a 342-amino acid polypeptide with a calculated molecular mass of 45 kDa and was found to be expressed on AIM, but not on resident macrophages or other cells, infiltrating into the rejection site. HEK293T cells transfected with this cDNA reacted with R15 mAb and H-2D(d), but not H-2L(d), H-2K(d), H-2D(b), H-2K(b), H-2D(k), or H-2K(k), molecules; and the H-2D(d) binding was suppressed by the addition of R15 or anti-H-2D(d) mAb. AIM yielded a specific saturation isotherm in the presence of increasing concentrations of H-2D(d), but not H-2D(b) or H-2D(k), molecules. The dissociation constant of AIM toward H-2D(d) tetramers was 1.9 x 10(-9) M ; and the binding was completely inhibited by the addition of R15 or anti-H-2D(d) mAb. These results reveal that a novel receptor for an allogeneic H-2D(d) molecule was induced on effector macrophages responsible for allograft (H-2(d)) rejection in H-2(b) mice.     

Establishment of an immunoscreening system using recombinant VP1 protein for the isolation of a monoclonal antibody that blocks JC virus infection

Polyomavirus JC (JCV) infection causes the fatal human demyelinating disease, progressive multifocal leukoencephalopathy. Although the initial interaction of JCV with host cells occurs through direct binding of the major viral capsid protein (VP1) with cell-surface molecules possessing sialic acid, these molecules have not yet been identified. In order to isolate monoclonal antibodies which inhibit attachment of JCV, we established an immunoscreening system using virus-like particles consisting of the VP1. Using this system, among monoclonal antibodies against the cell membrane fraction from JCV-permissive human neuroblastoma IMR-32 cells, we isolated a monoclonal antibody designated as 24D2 that specifically inhibited attachment and infection of JCV to IMR-32 cells. The antibody 24D2 recognized a single molecule of around 60 kDa in molecular weight in the IMR-32 membrane fraction. Immunohistochemical staining with 24D2 demonstrated immunoreactivity in the cell membrane of JCV-permissive cell lines and glial cells of the human brain. These results suggested that the molecule recognized by 24D2 plays a role in JCV infection, and that it might participate as a receptor or a co-receptor in JCV attachment and entry into the cells.     

Construction of bioactive chimeric MHC class I tetramer by expression and purification of human-murine chimeric MHC heavy chain and beta(2)m as a fusion protein in Escherichia coli

Major histocompatibility (MHC) class I tetramers are used in the quantitative analysis of epitope peptide-specific CD8+ T-cells. An MHC class I tetramer was composed of 4 MHC class I complexes and a fluorescently labeled streptavidin (SA) molecule. Each MHC class I complex consists of an MHC heavy chain, a beta(2)-microglobulin (beta(2)m) molecule and a synthetic epitope peptide. In most previous studies, an MHC class I complex was formed in the refolding buffer with an expressed MHC heavy chain molecule and beta(2)m, respectively. This procedure inevitably resulted in the disadvantages of forming unwanted multimers and self-refolding products, and the purification of each kind of monomer was time-consuming. In the present study, the genes of a human/murine chimeric MHC heavy chain (HLA-A2 alpha1, HLA-A2 alpha2 and MHC-H2D alpha3) and beta(2)m were tandem-cloned into plasmid pET17b and expressed as a fusion protein. The recombinant fusion protein was refolded with each of the three HLA-A2 restricted peptides (HBc18-27 FLPSDFFPSI, HBx52-60 HLSLRGLPV, and HBx92-100 VLHKRTLGL) and thus three chimeric MHC class I complexes were obtained. Biotinylation was performed, and its level of efficiency was observed via a band-shift assay in non-reducing polyacrylamide gel electrophoresis (PAGE). Such chimeric MHC class I tetramers showed a sensitive binding activity in monitoring HLA/A2 restrictive cytotoxic T lymphocytes (CTLs) in immunized HLA/A*0201 transgenic mice.     

Structural models of the supramolecular organization of AQP0 and connexons in junctional microdomains

Membrane proteins perform many essential cellular functions. Over the last years, substantial advances have been made in our understanding of the structure and function of isolated membrane proteins. However, like soluble proteins, many membrane proteins assemble into supramolecular complexes that perform specific functions in specialized membrane domains. Since supramolecular complexes of membrane proteins are difficult to study by conventional approaches, little is known about their composition, organization and assembly. The high signal-to-noise ratio of the images that can be obtained with an atomic force microscope (AFM) makes this instrument a powerful tool to image membrane protein complexes within native membranes. Recently, we have reported high-resolution topographs of junctional microdomains in native eye lens membranes containing two-dimensional (2D) arrays of aquaporin-0 (AQP0) surrounded by connexons. While both proteins are involved in cell adhesion, AQP0 is a specific water channel whereas connexons form cell–cell communication channels with broad substrate specificity. Here, we have performed a detailed analysis of the supramolecular organization of AQP0 tetramers and connexon hexamers in junctional microdomains in the native lens membrane. We present first structural models of these junctional microdomains, which we generated by docking atomic models of AQP0 and connexons into the AFM topographs. The AQP0 2D arrays in the native membrane show the same molecular packing of tetramers seen in highly ordered double-layered 2D crystals obtained through reconstitution of purified AQP0. In contrast, the connexons that surround the AQP0 arrays are only loosely packed. Based on our AFM observations, we propose a mechanism that may explain the supramolecular organization of AQP0 and connexons in junctional domains in native lens membranes.     

Closure of potassium M-channels by muscarinic acetylcholine-receptor stimulants requires a diffusible messenger.

The M-current (IK(M)) is a slow voltage-gated K+ current which can be inhibited by muscarinic acetylcholine-receptor (mAChR) agonists. In the present experiments we have tested whether this inhibition results from a local (membrane-delimited) interaction between the receptor and adjacent channels, or whether channel closure is mediated by a diffusible messenger. To do this, single KM(+)-channel currents were recorded from membrane patches in dissociated rat superior cervical sympathetic neurons by using cell-attached patch electrodes. Channel activity was inhibited when muscarine was applied to the cell membrane outside the patch but persisted when channels were exposed to muscarine added to the pipette solution. We conclude that a diffusible molecule (or molecules) is (are) required to induce intrapatch channel closure following activation of extra-patch receptors.     

Structural aspects of aldehyde dehydrogenase that influence dimer-tetramer formation

Aldehyde dehydrogenases are isolated as dimers or tetramers but have essentially identical structures. The homotetramer (ALDH1 or ALDH2) is a dimer of dimers (A-B + C-D). In the tetrameric enzyme, Ser500 from subunit "D" interacts with Arg84, a conserved residue, from subunit "A". In the dimeric ALDH3 form, the interaction cannot exist. It has been proposed that the formation of the tetramer is prevented by the presence of a C-terminal tail in ALDH3 that is not present in ALDH1 or 2. To understand the forces that maintain the tetramer, deletion of the tail in ALDH3, addition of different tails in ALDH1, and mutations of different residues located in the dimer-dimer interface were made. Gel filtration of the recombinantly expressed enzymes demonstrated that no change in their oligomerization occurred. Urea denaturation showed a diminution to the stability of the ALDH1 mutants. The K(m) for propionaldehyde was similar to that of the wild-type enzyme, but the K(m) for NAD was altered. A double mutant of D80G and S82A produced an enzyme with the ability to form dimers and tetramers in a protein-concentration-dependent manner. Though stable, this dimeric form was inactive. The tetramer exhibited 10% activity compared with the wild type. Sequence alignment demonstrated that the hydrophobic surface area is increased in the tetrameric enzymes. The hydrophobic surface seems to be the main force that drives the formation of tetramers. The results indicated that residues 80 and 82 are involved in maintaining the tetramer after its assembly but the C-terminal extension contributes to the overall stability of the assembled protein.     

Effect of Lipid Head Groups on Double-Layered Two-Dimensional Crystals Formed by Aquaporin-0

Aquaporin-0 (AQP0) is a lens-specific water channel that also forms membrane junctions. Reconstitution of AQP0 with dimyristoyl phosphatidylcholine (DMPC) and E. coli polar lipids (EPL) yielded well-ordered, double-layered two-dimensional (2D) crystals that allowed electron crystallographic structure determination of the AQP0-mediated membrane junction. The interacting tetramers in the two crystalline layers are exactly in register, resulting in crystals with p422 symmetry. The high-resolution density maps also allowed modeling of the annular lipids surrounding the tetramers. Comparison of the DMPC and EPL bilayers suggested that the lipid head groups do not play an important role in the interaction of annular lipids with AQP0. We now reconstituted AQP0 with the anionic lipid dimyristoyl phosphatidylglycerol (DMPG), which yielded a mixture of 2D crystals with different symmetries. The different crystal symmetries result from shifts between the two crystalline layers, suggesting that the negatively charged PG head group destabilizes the interaction between the extracellular AQP0 surfaces. Reconstitution of AQP0 with dimyristoyl phosphatidylserine (DMPS), another anionic lipid, yielded crystals that had the usual p422 symmetry, but the crystals showed a pH-dependent tendency to stack through their cytoplasmic surfaces. Finally, AQP0 failed to reconstitute into membranes that were composed of more than 40% dimyristoyl phosphatidic acid (DMPA). Hence, although DMPG, DMPS, and DMPA are all negatively charged lipids, they have very different effects on AQP0 2D crystals, illustrating the importance of the specific lipid head group chemistry beyond its mere charge.     

Block of single L-type Ca2+ channels in skeletal muscle fibers by aminoglycoside antibiotics

The activity of single L-type Ca2+ channels was recorded from cell- attached patches on acutely isolated skeletal muscle fibers from the mouse. The experiments were concerned with the mechanism by which aminoglycoside antibiotics inhibit ion flow through the channel. Aminoglycosides produced discrete fluctuations in the single-channel current when added to the external solution. The blocking kinetics could be described as a simple bimolecular reaction between an aminoglycoside molecule and the open channel. The blocking rate was found to be increased when either the membrane potential was made more negative or the concentration of external permeant ion was reduced. Both of these effects are consistent with a blocking site that is located within the channel pore. Other features of block, however, were incompatible with a simple pore blocking mechanism. Hyperpolarization enhanced the rate of unblocking, even though an aminoglycoside molecule must dissociate from its binding site in the channel toward the external solution against the membrane field. Raising the external permeant ion concentration also enhanced the rate of unblocking. This latter finding suggests that aminglycoside affinity is modified by repulsive interactions that arise when the pore is simultaneously occupied by a permeant ion and an aminoglycoside molecule.     

The pigeon heart 5''-nucleotidase responsible for ischaemia-induced adenosine formation.

The pH-induced reversible dissociation of pigeon liver malic enzyme (EC 1.1.1.40) was studied by combined use of chemical cross-linking and SDS/polyacrylamide-gel electrophoresis. The tetrameric enzyme showed a pH-dependent dissociation in an acidic environment. At pH values above 8.0 most molecules existed as tetramers. The enzyme was gradually dissociated at lower pH. When the pH was below 5.0 most of the enzyme was present as the monomeric forms. Reassociation of the subunits was accomplished by adjusting the pH to neutrality. The dissociation and reassociation were almost instantaneous. No trimer was detected. The pigeon liver malic enzyme was thus shown to have a double-dimer quaternary structure with D2 symmetry. In the presence of substrates, the monomer-dimer-tetramer equilibrium favours the direction of dissociation. Tartronate, an L-malate analogue, was found to be more effective than L-malate in this process. When the monomeric forms were immobilized, the enzyme subunits were found to be fully active in catalysis. A possible arrangement of the four identical subunits of the enzyme molecule is proposed to account for the results obtained in this investigation. The origin of the half-of-the-sites reactivity of pigeon liver malic enzyme is also discussed.