Mechanism of lactose-proton cotransport in Escherichia coli. Kinetic results in terms of the site exposure model

Rigorous kinetic derivations are presented for the Site Exposure mechanism of lactose-proton cotransport in E. coli [J. Theor. Biol. (1978) 75, 35-50]. Proton translocation inwards is solely associated with the external exposure of the galactoside binding site. A symmetric dimer configuration of the transporter is proposed, resulting in two forms corresponding to the cis and the trans orientation of the binding sites. The cis to trans orientation is inherently unfavorable, induced only by transmembrane substrate gradients. Recently reported extensive kinetic data are straightforwardly predicted by this mechanism, including the complicated effects on the apparent affinity and maximal velocity of uptake exhibited by changes in the magnitude of the proton electrochemical gradient.     

Mitochondrial protein import. Requirement of presequence elements and tom components for precursor binding to the TOM complex

Protein translocation across the outer mitochondrial membrane is mediated by the translocator called the TOM (translocase of the outer mitochondrial membrane) complex. The TOM complex possesses two presequence binding sites on the cytosolic side (the cis site) and on the intermembrane space side (the trans site). Here we analyzed the requirement of presequence elements and subunits of the TOM complex for presequence binding to the cis and trans sites of the TOM complex. The N-terminal 14 residues of the presequence of subunit 9 of F(0)-ATPase are required for binding to the trans site. The interaction between the presequence and the cis site is not sufficient to anchor the precursor protein to the TOM complex. Tom7 constitutes or is close to the trans site and has overlapping functions with the C-terminal intermembrane space domain of Tom22 in the mitochondrial protein import.     

RCA I-binding patterns of the Golgi apparatus

The distribution in the Golgi apparatus of binding sites for the galactose-specific Ricinus communis I lectin (RCA I) was studied in differently specialized cells, including goblet cells and absorptive enterocytes of the rat small intestine as well as acinar cells of the rat embryonic pancreas and submandibular gland. For the purpose of localizing the binding reactions, a pre-embedment method using horseradish peroxidase for electron microscopic visualization, and a post-embedding technique making use of the colloidal gold system were employed. The reactions obtained, localizing cell constituents which contain saccharides with terminal or internal beta-D-galactosyl residues, labeled diverse Golgi subcompartments. The goblet cells showed intense RCA I staining of the cisternae of the trans side of the Golgi stacks. The reaction was weak in the medial cisternae and the cis side of the stacks mostly was devoid of label. In the absorptive cells, in addition to the RCA I reaction of trans Golgi elements, binding sites for this lectin were concentrated in the stacks' medial section. In the embryonic acinar cells, accessible galactosyl residues were either confined to the trans and/or medial cisternae, or distributed across elements of all the stacked saccules. In the latter stacks, the reactions mostly were weak in the cis cisternae and increased in intensity towards the trans side. As regards the respective labeling patterns, similar percentages were calculated for the early and late stages of development: they were approximately 62% for the pattern which showed RCA I label limited to trans/medial cisternae and approximately 38% for the "cis-to-trans"-distributed RCA I reaction.(ABSTRACT TRUNCATED AT 250 WORDS)     

A proton motive force transducer and its role in proton pumps, proton engines, tobacco mosaic virus assembly and hemoglobin allosterism

In this paper I am proposing a new, conformationally dependent basic site in proteins. The initial formulation of this proposal was based on the following: (1) bacteriorhodopsin is a light-driven proton pump and as such is a prototype for understanding proton-mediated energy transduction in biological systems; (2) current evidence suggests about 2 protons are pumped for each photon absorbed; (3) given the usual role of prolines as helix breakers, it is surprising to find about 2 prolines deeply embedded in the membrane-spanning, probably α-helical, portion of the bacteriorhodopsin molecule; (4) another presumptive proton translocator, the F₀ proteolipid, is also helical and has a critical proline in its structure; (5) workers interested in protein folding have explained the existence of fast and slow folding subgroups of the same protein molecule as being due to cis : trans isomerization about the proline imide group; (6) the cis : trans isomerization is acid catalyzed; (7) simple chemical considerations predict that the proton affinity of the proline nitrogen should increase dramatically as the imide group is distorted away from planarity and should be a maximum midway between the cis and trans forms; thus, stabilization of the intermediate by protonation accounts for the acid catalysis of the proline cis : trans isomerization.Linking these observations together suggests that proline-containing α-helices may play a role in proton motive energy transduction. Due to the absence of a proton on the proline nitrogen, a proline-containing helix has a “proton hole” between the proline nitrogen and the carbonyl oxygen four residues earlier in the sequence. Here I propose a model in which the paramount feature is the change in pK_a associated with a change in geometry of the “proton hole.” Order of magnitude calculations suggest that the proton hole should change its pKa by about 8 units, corresponding to a 10⁸ change in proton affinity, for every 10 kcal of distortion energy, V. Calculations also show that it is energetically feasible to modulate the pK_a of this site over the dynamic range of pK_a = 2–14. Such a large value for $\text{ΔpK}{}_{\mathrm{a}}\text{ΔV}$ and such a dynamic range makes this site an ideal basis for an “integral proton injector,” an abstract model for proton pumping suggested on purely theoretical grounds by Nagle &; Mille (J. chem. Phys.74, 1367–1372, 1981).Finally, two well studied proteins, the α-chain of hemoglobin and tobacco mossaic virus coat protein, both show features in their X-ray determined structures suggesting the possibility of protonation and deprotonation of the proton hole in α-helices containing proline. For TMV coat protein, there is a proline-containing α-helix that is located precisely in the region of the protein which undergoes an acid-induced conformational rearrangement. Structural changes at this locus have been singled out in comparisons of the X-ray structures of the TMV protein in its two conformations. For the α-chain of horse hemoglobin, there are two concurrent sites that are likely protonated and one contrary site that likely becomes deprotonated as hemoglobin converts from the liganded to the deoxy form. The contrary proline is proposed to help maintain co-operative oxygen binding over a wide pH range. The absence of one of the concurrent proline site in marsupial hemoglobin accounts for the small Bohr effect exhibited by these hemoglobins. The absence of the contrary proline site in carp hemoglobin accounts in a very logical way for the large Bohr effect and the lack of cooperative oxygen binding at both low and high pH by this hemoglobin.     

Subfractionation of chick embryo epiphyseal cartilage Golgi. Localization of enzymes involved in the synthesis of the polysaccharide portion of proteochondroitin sulfate.

Membranes from chick embryo epiphyseal cartilage were fractionated by equilibrium sucrose density gradient centrifugation and assayed for galactosyl xylose transferase, chondroitin polymerization and sulfation as well as the marker enzymes glucose-6-phosphatase, NADH cytochrome c reductase, galactosyl ovalbumin transferase, and sialyltransferase. The order of distribution of chondroitin sulfate synthesis from dense to light membranes correlated with the established sequence of events for its synthesis. The linkage region enzyme, viz. galactosyl xylose transferase, distributed with NADH cytochrome c reductase in an earlier and heavier cis compartment. Chondroitin polymerization and sulfation had a dual distribution similar to the galactosyl ovalbumin transferase and sialyltransferase in separate later and lighter medial and trans compartments, or in an extended medial or trans compartment. The galactosyl xylose transferase had a distribution distinctly different from that of the galactosyl ovalbumin transferase indicating that these distinct enzymes showed no cross-reactivity with their respective acceptor substrates. The dual distribution of chondroitin sulfate synthesis was consistent with our previous demonstration of the two nascent proteochondroitin populations produced by microsomal preparations from the same source. The results indicated separate subcellular locations for synthesis of the two forms.     

Conformational analysis of cyclic hexapeptides designed as constrained ligands for the SH2 domain of the p85 subunit of phosphatidylinositol-3-OH kinase

The structures of the cyclic hexapeptide cyclo(-Gly-Tyr-Val-Pro-Met-Leu-) ( 1 ) and its phosphotyrosyl (pTyr) derivative cyclo[-Gly-Tyr(PO₃H₂)-Val-Pro-Met-Leu-] ( 2 ), designed as constrained models of a sequence that interacts with the src homology 2 (SH2) region of the p85 subunit of phosphatidylinositol-3-OH kinase (PI-3 kinase), were studied in methanol/water solutions by 500 MHz nmr spectroscopy. Compound 1 was found to exist as a 2:1 mixture of isomers about the Val-Pro bond (trans and cis prolyl) between 292–330 K in 75% CD₃O (D,H)/(D,H)₂O solutions. A third species of undetermined structure (ca. 5%) was also observed. Compound 2, a model of phosphorylated peptide ligand that binds to the PI-3 kinase SH2 domain, exhibited similar conformational isomerism. When either compound was dissolved in pure solvent [i.e., 100% CD₃O(H,D) or (H,D)₂O] the ratio of cis to trans isomers was ca 1:1. A battery of one- and two-dimensional nmr experiments at different temperatures and solvent compositions allowed a complete assignment of both the cis and trans forms of 1 and indicated the trans compound to be the major isomer. The spectral properties of the phosphorylated derivative 2 paralleled those of 1 , indicating like conformations for the two compounds. Analysis of rotating frame Overhauser spectroscopy data, coupling constants, amide proton temperature dependence, and amide proton exchange rates generated a set of constraints that were employed in energy minimization and molecular dynamics calculations using the CHARMM force field. The trans isomer exists with the tyrosine and C-terminal Tyr(+3) (Met) residues at opposite corners of the 18-membered ring separated by a distance of 16–18 Å, in contrast with the cis isomer where the side chains of these residues are much closer in space (7–14 Å). It was previously shown that the pTyr and the third amino acid C-terminal to this residue are the critical recognition elements for pTyr-peptide binding to the PI-3 kinase SH2 domain. Such cyclic structures may offer appropriate scaffolding for positioning important amino acid side chains of pTyr-containing peptides as a means of increasing their binding affinities to SH2 domains, and in turn provide a conceptual approach toward the design of SH2 domain directed peptidomimetics. © 1995 John Wiley & Sons, Inc.     

Catalytic site forms and controls in ATP synthase catalysis

A suggested minimal scheme for substrate binding by and interconversion of three forms of the catalytic sites of the ATP synthase is presented. Each binding change, that drives simultaneous interchange of the three catalytic site forms, requires a 120 degrees rotation of the gamma with respect to the beta subunits. The binding of substrate(s) at two catalytic sites is regarded as sufficing for near maximal catalytic rates to be attained. Although three sites do not need to be filled for rapid catalysis, during rapid bisite catalysis some enzyme may be transiently present with three sites filled. Forms with preferential binding for ADP and P(i) or for ATP are considered to arise from the transition state and participate in other steps of the catalysis. Intermediate forms and steps that may be involved are evaluated. Experimental evidence for energy-dependent steps and for control of coupling to proton translocation and transition state forms are reviewed. Impact of relevant past data on present understanding of catalytic events is considered. In synthesis a key step is suggested in which proton translocation begins to deform an open site so as to increase the affinity for ADP and P(i), that then bind and pass through the transition state, and yield tightly bound ATP in one binding change. ADP binding appears to be a key parameter controlling rotation during synthesis. In hydrolysis ATP binding to a loose site likely precedes any proton translocation, with proton movement occurring as the tight site form develops. Aspects needing further study are noted. Characteristics of the related MgADP inhibition of the F(1) ATPases that have undermined many observations are summarized, and relations of three-site filling to catalysis are assessed.     

Alternating translocation of protein substrates from both ends of ClpXP protease

In ClpXP protease complexes, hexameric rings of the ATP-dependent ClpX chaperone stack on one or both faces of the double-heptameric rings of ClpP. We used electron microscopy to record the initial binding of protein substrates to ClpXP and their accumulation inside proteolytically inactive ClpP. Proteins with N- or C-terminal recognition motifs bound to complexes at the distal surface of ClpX and, upon addition of ATP, were translocated to ClpP. With a partially translocated substrate, the non-translocated portion remained on the surface of ClpX, aligned with the central axis of the complex, confirming that translocation proceeds through the axial channel of ClpXP. Starting with substrate bound on both ends, most complexes translocated substrate from only one end, and rarely (<5%) from both ends. We propose that translocation from one side is favored for two reasons: initiation of translocation is infrequent, making the probability of simultaneous initiation low; and, further, the presence of protein within the cis side translocation channel or within ClpP generates an inhibitory signal blocking translocation from the trans side.     

The intermembrane space domain of mitochondrial Tom22 functions as a trans binding site for preproteins with N-terminal targeting sequences.

Mitochondrial protein import is thought to involve the sequential interaction of preproteins with binding sites on cis and trans sides of the membranes. For translocation across the outer membrane, preproteins first interact with the cytosolic domains of import receptors (cis) and then are translocated through a general import pore, in a process proposed to involve binding to a trans site on the intermembrane space (IMS) side. Controversial results have been reported for the role of the IMS domain of the essential outer membrane protein Tom22 in formation of the trans site. We show with different mutant mitochondria that a lack of the IMS domain only moderately reduces the direct import of preproteins with N-terminal targeting sequences. The dependence of import on the IMS domain of Tom22 is significantly enhanced by removing the cytosolic domains of import receptors or by performing import in two steps, i.e., accumulation of a preprotein at the outer membrane in the absence of a membrane potential (delta psi) and subsequent import after reestablishment of a delta psi. After the removal of cytosolic receptor domains, two-step import of a cleavable preprotein strictly requires the IMS domain. In contrast, preproteins with internal targeting information do not depend on the IMS domain of Tom22. We conclude that the negatively charged IMS domain of Tom22 functions as a trans binding site for preproteins with N-terminal targeting sequences, in agreement with the acid chain hypothesis of mitochondrial protein import.     

Different proton-sugar stoichiometries for the uptake of glucose analogues by Chlorella vulgaris: Evidence for sugar-dependent proton uptake without concomitant sugar uptake by the proton-sugar symport system

The uptake of hexoses by Chlorella vulgaris is accompanied by the uptake of protons. For 6-deoxyglucose a stoichiometry of one proton taken up per sugar molecule has been measured, whereas for 1-deoxyglucose approximately two protons are taken up per sugar molecule.It was found that in the presence of 1-deoxyglucose a considerable proportion of “carrier” catalyzes the transport of protons without the concomitant transport of sugar. Presumably the binding of sugar initiates the translocation of the carrier-proton-sugar complex, but whereas 1-deoxyglucose can still dissociate from the complex at the external side of the cytoplasmic membrane, the translocation of the carrier-proton complex continues.This conclusion was reached since (a) the composition of the translocated carrier-proton-sugar complex is the same for both sugar. Its formation is a first order reaction with respect to protons.(b) When 6-deoxyglucose, present inside cells, is exchanged for external sugar, the exchange ratio is two to one when the external sugar is 1-deoxyglucose, two molecules of 6-deoxyglucose are lost for each molecule of 1-deoxyglucose entering. This result indicates that during uptake of 1-deoxyglucose statistically only each second carrier molecule appearing at the internal side of the cytoplasmic membrane is carrying sugar.     

On the logic of Fisherian sexual selection*

In Fisher's model of sexual selection, a female preference for a male trait spreads together with the trait because their genetic bases become correlated. This can be interpreted as a “greenbeard” system: a preference gene, by inducing a female to mate with a trait-bearing male, favors itself because the male is disproportionately likely also to carry the preference gene. Here, we use this logic to argue that Fisherian sexual selection in diploids proceeds via two channels: (i) trait-bearing males are disproportionately the product of matings between preference-bearing mothers and trait-bearing fathers, and thus trait and preference genes are correlated “in trans”; (ii) trait and preference genes come into gametic phase disequilibrium, and thus are correlated “in cis.” Gametic phase disequilibrium is generated by three distinct mechanisms that we identify. The trans channel does not operate when sexual selection is restricted to the haploid phase, and therefore represents a fundamental difference between haploid and diploid models of sexual selection. We show that the cis and trans channels contribute equally to the spread of the preference when recombination between the preference and trait loci is free, but that the trans channel is substantially more important when linkage is tight.     

The Golgi apparatus in the acinar cells of the developing embryonic pancreas: II. Localization of lectin-binding sites

The reaction patterns of the Golgi apparatus following staining with the lectins concanavalin A (ConA), Ricinus communis I agglutinin (RCA I), and Helix pomatia lectin (HPA) were studied in the pancreas acinar cells of rat embryos in the course of cell differentiation from day 13 through day 20 of gestation. The binding reactions were localized by means of pre-embedment incubation of 10-microns-thick cryosections of pancreas tissue, prefixed in a mixture of 4% formaldehyde/0.5% glutaraldehyde, using horseradish peroxidase for electron microscope visualization. ConA, which preferentially binds to alpha-D-mannosyl residues, consistently stained the cisternae of the cis Golgi side. The majority of the stacks also showed ConA staining of medial cisternae. The reaction of the trans side was variable; in each stage of development, the cisternae of the trans Golgi side either were devoid of labeling or appeared intensely stained. The reactions obtained with RCA I, which recognizes terminal beta-D-galactosyl residues, changed in the course of cell differentiation; in the protodifferentiated and early differentiated states, the system of "rigid lamellae," located at the trans side of the Golgi stacks, was intensely labeled, but became unreactive after production of secretion granules had started, the reaction then being restricted to the stacked saccules. In regard to the Golgi stacks in each of the developmental stages, RCA I binding sites either were confined to the trans cisternae, or, in addition, were found distributed across elements of the medial and cis compartments.(ABSTRACT TRUNCATED AT 250 WORDS)     

Dissection of the Golgi complex. II. Density separation of specific Golgi functions in virally infected cells treated with monensin

In the accompanying paper (Griffiths, G., P. Quinn, and G. Warren, 1983, J. Cell Biol., 96:835-850), we suggested that the Golgi stack could be divided into functionally distinct cis, medial, and trans compartments, each comprising one or two adjacent cisternae. These compartments were identified using Baby hamster kidney (BHK) cells infected with Semliki Forest virus (SFV) and treated with monensin. This drug blocked intracellular transport but not synthesis of the viral membrane proteins that were shown to accumulate in the medial cisternae. In consequence, these cisternae bound nucleocapsids. Here we show that this binding markedly increased the density of the medial cisternae and allowed us to separate them from cis and trans Golgi cisternae. A number of criteria were used to show that the intracellular capsid-binding membranes (ICBMs) observed in vivo were the same as those membranes sedimenting to a higher density in sucrose gradients in vitro, and this separation of cisternae was then used to investigate the distribution, within the Golgi stack, of some specific Golgi functions. After labeling for 2.5 min with [3H]palmitate, most of the fatty acid attached to viral membrane proteins was found in the ICBM fraction. Because the viral membrane proteins appear to move from cis to trans, this suggests that fatty acylation occurs in the cis or medial Golgi cisternae. In contrast, the distribution of alpha 1-2- mannosidase, an enzyme involved in trimming high-mannose oligosaccharides, and of galactosyl transferase, which is involved in the construction of complex oligosaccharides, was not affected by monensin treatment. Together with data in the accompanying paper, this would restrict these two Golgi functions to the trans cisternae. Our data strongly support the view that Golgi functions have specific and discrete locations within the Golgi stack.     

Binding of proflavin to T2L bacteriophage and its DNA.

We have measured the binding equilibria of proflavin to T2L bacteriophage, in both “slow” and “fast” sedimenting forms, and to free T2L DNA. Measurements were carried out by difference spectroscopy at 430 nm at temperatures from 13 to 43°C and at pH 5.6 and 7.6. We found no significant difference in the binding parameters of the two phage forms. Also, the fraction of nucleotides available as binding sites for proflavin was the same for both free and intraphage DNA. However, the binding constant is about an order of magnitude lower for encapsulated than for free T2L DNA, due to the decreased exothermicity of the binding reaction within the phage head.     

Coupling of homotropic and heterotropic interactions in Escherichia coli aspartate transcarbamylase

Several types of conditions allow the disconnection of homotropic and heterotropic interactions in Escherichia coli aspartate transcarbamylase. A model that includes a concerted gross conformational change corresponding to the homotropic cooperative interactions between the catalytic sites and local “site by site” effects promoted by the effectors accounts for this disconnection as well as for the other known properties of the enzyme. However, the substrate concentration influences the extent of stimulation and feedback inhibition of the catalytic activity by the effectors. This result is explained by assuming that these effectors promote a “primary effect”, which is exerted locally “site by site”, and a “secondary effect”, which is mediated by the substrate. As predicted by the model, relaxed (R) forms of the enzyme show only the primary effect. In addition 2-ThioU-aspartate transcarbamylase, a modified form of the enzyme in which the homotropic cooperative interactions between the catalytic sites are selectively abolished, shows the same heterogeneity in CTP binding sites as normal aspartate transcarbamylase.     

Squeezing at Entrance of Proton Transport Pathway in Proton-translocating Pyrophosphatase upon Substrate Binding

Homodimeric proton-translocating pyrophosphatase (H⁺-PPase; EC 3.6.1.1) is indispensable for many organisms in maintaining organellar pH homeostasis. This unique proton pump couples the hydrolysis of PP_i to proton translocation across the membrane. H⁺-PPase consists of 14–16 relatively hydrophobic transmembrane domains presumably for proton translocation and hydrophilic loops primarily embedding a catalytic site. Several highly conserved polar residues located at or near the entrance of the transport pathway in H⁺-PPase are essential for proton pumping activity. In this investigation single molecule FRET was employed to dissect the action at the pathway entrance in homodimeric Clostridium tetani H⁺-PPase upon ligand binding. The presence of the substrate analog, imidodiphosphate mediated two sites at the pathway entrance moving toward each other. Moreover, single molecule FRET analyses after the mutation at the first proton-carrying residue (Arg-169) demonstrated that conformational changes at the entrance are conceivably essential for the initial step of H⁺-PPase proton translocation. A working model is accordingly proposed to illustrate the squeeze at the entrance of the transport pathway in H⁺-PPase upon substrate binding.     

The N-terminal half of the heavy chain of botulinum type A neurotoxin forms channels in planar phospholipid bilayers

The heavy chain of botulinum type A neurotoxin forms channels in planar phospholipid bilayer membranes. Channel activity is confined to the N-terminal half of this chain; the C-terminal half is inactive. Channel activity is stimulated by low pH (4.5-5.5) on the cis side (the side to which protein is added), neutral pH on the opposite (trans) side, and cis positive voltages. These findings are strikingly similar to those previously reported for analogous fragments of diphtheria and tetanus toxins.     

Single streptomyces lividans K(+) channels: functional asymmetries and sidedness of proton activation.

Basic electrophysiological properties of the KcsA K(+) channel were examined in planar lipid bilayer membranes. The channel displays open-state rectification and weakly voltage-dependent gating. Tetraethylammonium blocking affinity depends on the side of the bilayer to which the blocker is added. Addition of Na(+) to the trans chamber causes block of open-channel current, while addition to the cis side has no effect. Most striking is the activation of KcsA by protons; channel activity is observed only when the trans bilayer chamber is at low pH. To ascertain which side of the channel faces which chamber, residues with structurally known locations were mapped to defined sides of the bilayer. Mutation of Y82, an external residue, results in changes in tetraethylammonium affinity exclusively from the cis side. Channels with cysteine residues substituted at externally exposed Y82 or internally exposed Q119 are functionally modified by methanethiosulfonate reagents from the cis or trans chambers, respectively. Block by charybdotoxin, known to bind to the channel's external mouth, is observed only when the toxin is added to the cis side of channels mutated to be toxin sensitive. These results demonstrate unambiguously that the protonation sites linked to gating are on the intracellular portion of the KcsA protein.     

A molecular switch in SecA protein couples ATP hydrolysis to protein translocation

SecA, the dimeric ATPase subunit of bacterial protein translocase, catalyses translocation during ATP-driven membrane cycling at SecYEG. We now show that the SecA protomer comprises two structural modules: the ATPase N-domain, containing the nucleotide binding sites NBD1 and NBD2, and the regulatory C-domain. The C-domain binds to the N-domain in each protomer and to the C-domain of another protomer to form SecA dimers. NBD1 is sufficient for single rounds of SecA ATP hydrolysis. Multiple ATP turnovers at NBD1 require both the NBD2 site acting in cis and a conserved C-domain sequence operating in trans. This intramolecular regulator of ATP hydrolysis (IRA) mediates N-/C-domain binding and acts as a molecular switch: it suppresses ATP hydrolysis in cytoplasmic SecA while it releases hydrolysis in SecY-bound SecA during translocation. We propose that the IRA switch couples ATP binding and hydrolysis to SecA membrane insertion/deinsertion and substrate translocation by controlling nucleotide-regulated relative motions between the N-domain and the C-domain. The IRA switch is a novel essential component of the protein translocation catalytic pathway.     

The MAM (meprin/A5-protein/PTPmu) domain is a homophilic binding site promoting the lateral dimerization of receptor-like protein-tyrosine phosphatase mu

The MAM (meprin/A5-protein/PTPmu) domain is present in numerous proteins with diverse functions. PTPmu belongs to the MAM-containing subclass of protein-tyrosine phosphatases (PTP) able to promote cell-to-cell adhesion. Here we provide experimental evidence that the MAM domain is a homophilic binding site of PTPmu. We demonstrate that the MAM domain forms oligomers in solution and binds to the PTPmu ectodomain at the cell surface. The presence of two disulfide bridges in the MAM molecule was evidenced and their integrity was found to be essential for MAM homophilic interaction. Our data also indicate that PTPmu ectodomain forms oligomers and mediates the cellular adhesion, even in the absence of MAM domain homophilic binding. Reciprocally, MAM is able to interact homophilically in the absence of ectodomain trans binding. The MAM domain therefore contains independent cis and trans interaction sites and we predict that its main role is to promote lateral dimerization of PTPmu at the cell surface. This finding contributes to the understanding of the signal transduction mechanism in MAM-containing PTPs.