Multiple cadherin extracellular repeats mediate homophilic binding and adhesion.

The extracellular homophilic-binding domain of the cadherins consists of 5 cadherin repeats (EC1-EC5). Studies on cadherin specificity have implicated the NH(2)-terminal EC1 domain in the homophilic binding interaction, but the roles of the other extracellular cadherin (EC) domains have not been evaluated. We have undertaken a systematic analysis of the binding properties of the entire cadherin extracellular domain and the contributions of the other EC domains to homophilic binding. Lateral (cis) dimerization of the extracellular domain is thought to be required for adhesive function. Sedimentation analysis of the soluble extracellular segment of C-cadherin revealed that it exists in a monomer-dimer equilibrium with an affinity constant of approximately 64 microm. No higher order oligomers were detected, indicating that homophilic binding between cis-dimers is of significantly lower affinity. The homophilic binding properties of a series of deletion constructs, lacking successive or individual EC domains fused at the COOH terminus to an Fc domain, were analyzed using a bead aggregation assay and a cell attachment-based adhesion assay. A protein with only the first two NH(2)-terminal EC domains (CEC1-2Fc) exhibited very low activity compared with the entire extracellular domain (CEC1-5Fc), demonstrating that EC1 alone is not sufficient for effective homophilic binding. CEC1-3Fc exhibited high activity, but not as much as CEC1-4Fc or CEC1-5Fc. EC3 is not required for homophilic binding, however, since CEC1-2-4Fc and CEC1-2-4-5Fc exhibited high activity in both assays. These and experiments using additional EC combinations show that many, if not all, the EC domains contribute to the formation of the cadherin homophilic bond, and specific one-to-one interaction between particular EC domains may not be required. These conclusions are consistent with a previous study on direct molecular force measurements between cadherin ectodomains demonstrating multiple adhesive interactions (Sivasankar, S., W. Brieher, N. Lavrik, B. Gumbiner, and D. Leckband. 1999. PROC: Natl. Acad. Sci. USA. 96:11820-11824; Sivasankar, S., B. Gumbiner, and D. Leckband. 2001. Biophys J. 80:1758-68). We propose new models for how the cadherin extracellular repeats may contribute to adhesive specificity and function.

Neutrophil activation by monomeric interleukin-8.

Interleukin-8 (IL-8), a pro-inflammatory protein, has been shown by nuclear magnetic resonance (NMR) and x-ray techniques to exist as a homodimer. An IL-8 analog was chemically synthesized, with the amide nitrogen of leucine-25 methylated to selectivity block formation of hydrogen bonds between monomers and thereby prevent dimerization. This analog was shown to be a monomer, as assessed by analytical ultracentrifugation and NMR. Nevertheless, it was equivalent to IL-8 in assays of neutrophil activation, which indicates that the monomer is a functional form of IL-8.

Iron-mineral accretion from acid mine drainage and its application in passive treatment.

This study demonstrates substantial removal of iron (Fe) from acid mine drainage (pH ≈3) in a passive vertical flow reactor (VFR) with an equivalent footprint of 154 m(2) per L/s mine water and residence times of >23 h. Average Fe removal rate was 67% with a high of 85% over the 10-month trial. The fraction of Fe passing a 0.22 µm filter (referred to here as Fe-filt) was seen to be removed in the VFR even when Fe(II) was absent, indicating that the contribution of microbial Fe(II) oxidation and precipitation was not the dominant removal mechanism in the VFR. Removal rates of Fe-filt in the VFR were up to 70% in residence times as low as 8 h compared with laboratory experiments where much smaller changes in Fe-filt were observed over 60 h. Centrifugation indicated that 80-90% of the influent Fe had particle sizes <35 nm. Together with analyses and geochemical modelling, this suggests that the Fe-filt fraction exists as either truly aqueous (but oversaturated) Fe(III) or nanoparticulate Fe(III) and that this metastability persists. When the water was contacted with VFR sludge, the Fe-filt fraction was destabilized, leading to an appreciably higher removal of this fraction. Heterogeneous precipitation and/or aggregation of nanoparticulate Fe(III) precipitates are considered predominant removal mechanisms. Microbial analyses of the mine water revealed the abundance of extracellular polymeric substance-generating Fe-oxidizing bacterium 'Ferrovum myxofaciens', which may aid the removal of iron and explain the unusual appearance and physical properties of the sludge.

Interaction studies of the 165 000 dalton protein component of the M-line with the S2 subfragment of myosin.

The M-line protein component of molecular weight 165 000 was isolated and purified from rabbit skeletal muscle using ion exchange chromatography. Sodium dodecyl sulphate electrophoresis revealed that the protein was homogeneous. Circular dichroism measurements indicated that the protein interacts with myosin and heavy meromyosin subfragment 2 (S2). There was an increase in negative ellipticity at 221 nm upon interaction, relative to the calculated values assuming no interprotein interaction. The net increaes in negative ellipticity at 221 nm as a result of interaction of M-protein with myosin and subfragment 2 were 600 degrees and 800 degrees respectively. When the protein was mixed with subfragment 2 in a 1 : 1 mol ratio in 0.5 M KCl/25 mM Tris buffer at pH 8.0, low speed sedimentation equilibrium studies gave a molecular weight of 235 000 +/- 10 000 for the complex, indicative of an interaction of the two components. On a Bio gel A 0.5 m column, M-protein and S2 when applied in 1 : 1 mol ratio, were eluted as a single symmetrical peak and a molecular weight of 230 000 was obtained for the complex from the observed elution volume. Both circular dichroism and sedimentation equilibrium studies indicated no interaction of M-line protein with light meromyosin and subfragment 1. Interaction of the 165 000 component with the flexible hinge region of myosin may have special significance in terms of the mechanism accounting for the reversible expansion of the interfilament distance which occurs during contraction.

Physicochemical studies on the creatine kinase M-line protein and its interaction with myosin and myosin fragments.

The M-line protein which is identical to the muscle form of creatine kinase was purified from rabbit skeletal muscle using ion exchange chromatography. Gel electrophoresis in the presence and absence of sodium dodecyl sulfate revealed the protein to be homogeneous. Sodium dodecyl sulfate gel electrophoresis gave 44 000 +/- 2000 as the minimum molecular weight while low speed sedimentation equilibrium experiments yielded a molecular weight of 84 000 +/- 4000, suggesting that the parent molecule is a dimer. Circular dichroism spectra revealed the presence of two negative dichroic bands located at 218 and 208 nm suggesting the presence of some beta-structure. Ellipticity values at these two wavelengths were -8000 +/- 400 and -9000 +/- 400 deg-cm2-dmol-1. Circular dichroism measurements indicated the protein to interact with myosin, heavy meromyosin and heavy meromyosin subfragment 1 (S1). The Ca2+-activated ATPase activities of myosin, heavy meromyosin and subfragment 1 were inhibited by the addition of M-line protein. When the protein was mixed with subfragment 1 in a 1:1 mole ratio in 0.15 M KC1, 50 mM Tris pH 8, low speed sedimentation equilibrium studies gave a molecular weight of 205 000 +/- 10 000 for the complex, indicative of an interaction of the two components. Both circular dichroism and sedimentation equilibrium studies indicated no interaction of M-line protein with light meromyosin.

Isolation and characterization of the 165 000 dalton protein component of the M-line of rabbit skeletal muscle and its interaction with creatine kinase.

The M-line protein component of molecular weight 165 000 was isolated and purified from rabbit skeletal muscle using ion exchange chromatography. Gel electrophoresis, in the presence and absence of sodium dodecyl sulfate, revealed the protein to be homogeneous. Sodium dodecyl sulfate gel electrophoresis and low speed sedimentation equilibrium studies in 0.5 M KCl, 50 mM potassium phosphate gave a molecular weight of 165 000 suggesting the protein to be made up of a single polypeptide chain. Circular dichroism spectra revealed the presence of two negative dichroic bands located at 216 and 208 nm, indicative of the presence of some beta-structure. Ellipticity values at these two wavelengths were --6500 +/- 400 and --7500 +/- 400 deg . cm2 . dmol-1, respectively. Addition of 165 000 component lowered the enzymatic activity of creatine kinase M-line protein and the nature of the inhibition was found to be a competitive one. When the protein was mixed with creatine kinase in a 1 : 1 mole ratio in a medium consisting of 0.2 M KCl, 25 mM Tris, 1 mM dithiothreitol (pH 8.0), low speed sedimentation equilibrium studies gave a molecular weight of 260 000 +/- 10 000 for the complex, indicative of an interaction of the two components of the M-line.

Modulation of substrate selectivity in plasma lipid transfer protein reaction over structural variation of lipid particle.

The modulation of substrate selectivity of human plasma LTP reaction is the subject of the present investigation. The moderate selectivity by a factor of 5 to 6 was observed in the LTP-catalyzed transfer of cholesteryl ester over triacylglycerol between plasma lipoproteins. On the other hand, the transfer of cholesteryl ester by LTP was highly selective over the negligible transfer of triacylglycerol, by a factor of 60 to 500, between the microemulsions with LDL size, regardless of the activators such as human and pig apolipoprotein (apo) A-I, human apo C-III and apo E that bound to the surface of the emulsion in equilibrium. The presence of free cholesterol in these microemulsions reduced slightly the rate of cholesteryl ester transfer but had no effect on triacylglycerol transfer. Other surface-active reagents such as cholic acid, Triton X-100 and Tween-20, did not have an effect on the triacylglycerol transfer either. Triacylglycerol transfer by LTP became measurable between such lipid particles as prepared by co-sonication of lipid with pig apo A-I and isolated as the mixed-microemulsions in the density of LDL and HDL. In these conditions, the substrate selectivity for cholesteryl ester over triacylglycerol was a factor of 6 to 16 mimicking the ratio in plasma lipoproteins. The conformation of pig apo A-I estimated by circular dichroism showed that its apparent helical content was further more induced when apo A-I was integrated into the mixed-microemulsion by co-sonication than the lipid-bound apo A-I in equilibrium. Apo A-I, thus integrated into lipid particles, was highly resistant to the denaturation by guanidine hydrochloride while the lipid-bound apo A-I in equilibrium was denatured as readily as the lipid-free protein. Thus, triacylglycerol transfer by LTP was induced by structural modulation of substrate-carrying lipid particles such as higher integration of apolipoproteins.

Circular dichroism of intermediate subviral particles of reovirus. Elucidation of the mechanism underlying the specific monovalent cation effects on uncoating.

1. Circular dichroic (CD) spectra of purified intermediate subviral particles of reovirus were determined in the presence of different monovalent cations. 2. The CD spectra reveal that reo intermediate subviral particles can exist in two conformationally different forms. The two forms are readily distinguished by comparison of their ellipticities in the wavelength regions 210 nm and 220 nm, with a Na+-induced form exhibiting a reduced negative ellipticity relative to a Cs+-induced form. 3. The transition between the Na+- and Cs+-induced forms is reversible by manipulation of the species of monovalent cation present and appears to be temperature independent. 4. Temperature variation studies on dilute suspensions of particles indicate that the Na+-induced form is stable, whereas the Cs+-induced from undergoes a second transition, temperature dependent and irreversible, to become a viral core. 5. A model is presented relating these observations to the known properties of reovirus uncoating and transcriptase activation.

Circular dichroism of lipoxygenase-1 from soybeans.

The circular dichroism spectra of the three forms of lipoxygenase-1 from soybeans show characteristic differences in the region between 300 and 600 nm. Native lipoxygenase-1 only shows a negative dichroic band around 330 nm. Yellow lipoxygenase-1, obtained by addition of an equimolar amount of 13-F-hydroperoxylinoleic acid to the native enzyme, shows a positive Cotton effect at 425 nm, while the negative band band at 330 nm has increased in intensity. The blue enzyme, representing a complex of yellow enzyme with 13-L-hydroperoxylinoleic acid exhibits a negative dichroic band at 580 nm and positive bands at 410 and 391 nm. The near-ultraviolet CD spectra of the three forms of lipoxygenase are very similar, showing several well resolved positive dichroic bands at 0 degrees C. Using the method of Chen et al. (Chen, Y.-H., Yang, J.T. and Martinez, H.M. (1972) Biochemistry 11, 4120--4131) the contents of alpha-helix, beta- and unordered form of native lipoxygenase-1 were estimated to be 34, 27 and 39% respectively.

Purification and characterization of Aspergillus niger exo-1,4-glucosidase.

A specific exo-1,4-glucosidase (1,4-alpha-D-glucan glucohydrooase, EC 3.2.1.3) from Aspergillus niger has been partially purified and subsequently characterized by biochemical, physico-chemical and optical methods. Molecular sieve chromatography yields an enzyme with maximal activity at pH 4.2-4.5 close to its isoelectric point. Reduction and carboxymethylation leads to complete loss of activity and O-acetylation of 3 of the 13 tyrosine residues results in loss of 20 % of the activity. Sodium dodecylsulfate-polyacrylamide gel electrophoresis indicates that the native enzyme consists of two major components of molecular weights 63 000 and 57 500, respectively. Small amounts of dissociated material of molecular weight 28 000 and 16 000 as well as aggregates of the order of 100 000 are also present to the extent of 2-5% of the total potein. Following reduction and carboxymethylation under forcing conditions, the bands around 60 000 diminish and the 28 000-30 000, 16 000 and aggregate bands are dominant...

Large-scale purification and biochemical characterization of crystallization-grade porin protein P from Pseudomonas aeruginosa.

A large-scale purification scheme was developed for lipopolysaccharide-free protein P, the phosphate-starvation-inducible outer-membrane porin from Pseudomonas aeruginosa. This highly purified protein P was used to successfully form hexagonal crystals in the presence of n-octyl-beta-glucopyranoside. Amino-acid analysis indicated that protein P had a similar composition to other bacterial outer membrane proteins, containing a high percentage (50%) of hydrophilic residues. The amino-terminal sequence of this protein, although not homologous to either outer membrane protein, PhoE or OmpF, of Escherichia coli, was found to have an analogous protein-folding pattern. Protein P in the native trimer form was capable of maintaining a stable functional trimer after proteinase cleavage. This suggested the existence of a strongly associated tertiary and quaternary structure. Circular dichroism studies confirmed these results in that a large proportion of the protein structure was determined to be beta-sheet and resistant to acid pH and heating in 0.1% sodium dodecyl sulphate.

Recombinant locust apolipophorin III: characterization and NMR spectroscopy.

Apolipophorin III (apoLp-III) from the locust Locusta migratoria is an exchangeable apolipoprotein that reversibly binds to lipoproteins. During lipid binding the protein has been proposed to undergo a major conformational change. To study the mechanism of lipid binding we have cloned and expressed recombinant protein in bacteria, permitting stable isotope enrichment for heteronuclear NMR spectroscopy and site-directed mutagenesis. The cDNA coding for apoLp-III was subcloned into the pET expression vector and transformed into Escherichia coli cells. Induction of expression resulted in the specific appearance of apoLp-III in the cell culture medium, indicating it escaped the bacteria without lysis. The protein was purified from the cell-free supernatant by reversed-phase HPLC, characterized and compared to the natural protein isolated from locust hemolymph. SDS-PAGE revealed the recombinant protein has a molecular mass of approximately 17 kDa, similar to that of deglycosylated natural apoLp-III. Monoclonal antibodies were used to detect recombinant apoLp-III in the cells as well as in cell-free medium of induced bacterial cultures. Amino acid sequencing and analysis confirmed the identity of the recombinant protein as L. migratoria apoLp-III. Circular dichroism spectroscopy of recombinant and natural apoLp-III showed similar spectra, both displaying high contents of alpha-helical secondary structure. Denaturation studies of lipid-free apoLp-III with guanidine hydrochloride showed that both proteins have similar denaturation midpoints and DeltaG values indicating similar protein stability. The natural and recombinant protein were functional in lipoprotein binding assays. Using recombinant protein, uniformly and specifically labeled with 15N-amino acids, two dimensional 1H-15N heteronuclear single quantum correlation spectra were obtained. The spectra revealed excellent chemical shift dispersion in both the 1H and 15N dimensions with a well defined resonance pattern. Studies with 15N-leucine specifically labeled apoLp-III in the presence and absence of the micelle forming lipid, dodecylphosphocholine, provided evidence for a significant conformational change upon lipid association.

Distribution of surface-exposed and non-accessible amino acid sequences among the two major structural domains of the S-layer protein of Aeromonas salmonicida.

The tetragonally arranged crystalline surface protein array (A-layer) of the fish pathogenic bacterium Aeromonas salmonicida is a virulence factor. Circular dichroism studies in the presence or absence of 0.1% sodium dodecyl sulfate showed that the secondary structure of A-protein, and its 39,439 molecular weight amino-terminal trypsin-resistant peptide, were altered. In both cases alpha-helix was increased significantly at the expense of beta-structure when SDS was added. Western and dot immunoblotting, immuno-microscopy and enzyme-linked immunosorbent assay with monospecific polyclonal antiserum and eight monoclonal antibodies specific for epitopes exposed on the surface of native A-layer showed that the 481 residue A-protein subunit and the surface of A-layer were conserved antigenically. Mimeotope analysis of nonapeptides representing the sequence of A-protein allowed identification of 146 residues in presumed linear epitopes accessible on the surface of A-layer. Inaccessible or non-epitopic residues accounted for 70% of the protein. The majority of inaccessible residues were in the N-terminal 301 residues of A-protein. Dispersed among these were 65 surface-accessible residues in five linear epitope clusters illustrating the complex folding of this major structural domain of A-protein. The C-terminal 180 residues carried fewer linear epitopes but contained the major region of A-layers surface-accessible sequence, including four linear epitopes in predominantly hydrophobic sequence. Four A-layer surface-binding monoclonal antibodies also bound to this minor structural domain, although the epitopes of only two were identified by mimeotope analysis. The epitopes of six A-layer surface-binding monoclonals could not be identified, suggesting that A-layer may also contain conformation dependent surface epitopes.

Orientation, positional, additivity, and oligomerization-state effects of interhelical ion pairs in alpha-helical coiled-coils.

The role of interhelical g-e' ion pairs in the dimerization specificity and stability of alpha-helical coiled-coils is highly controversial. Synthetic 35-residue coiled-coils based on the heptad repeat QgVaGbAcLdQeK f were used to investigate the effect of orientation of interhelical ion pairs between lysine and glutamic acid residues on coiled-coil stability. Stability was estimated from urea denaturation at 20 degreesC, monitoring unfolding with circular-dichroism spectroscopy. Double mutant cycles were employed to estimate the net interaction energy, Delta DeltaGuint, for the two orientations of the ion pair; Ee-Kg and Ke-Eg. Delta DeltaGuint was found to be about 1.4-fold higher for the Ee-Kg orientation in a coiled-coil containing an N-terminal disulfide bridge. The Delta DeltaGuint value was similar whether obtained from the middle heptad or averaged over all five heptads of the coiled-coil, suggesting that ion pairs contribute additively to coiled-coil stability. The effect of uncompensated charges was also illustrated by single substitutions of Gln with either Lys or Glu, resulting in Lys-Gln or Glu-Gln g-e' pairs. These substitutions were found to be twice as destabilizing at position g as at position e, and Lys was about twice as destabilizing as Glu at both positions e and g. In the absence of an interhelical disulfide bridge, Glu and Lys substitutions in the middle heptad were equally destabilizing at positions e and g (Lys continued to be more destabilizing than Glu) and the Delta DeltaGuint value for Lys-Glu ion pairs was not orientation dependent. These and previous results suggest the non-covalently-linked synthetic coiled-coils behave as molten globules, whereas a disulfide-bridge may "lock in" the structural differences between positions of the heptad repeat. Interhelical Lys-Glu ion pairs in either orientation promoted the formation of trimeric coiled-coils (in the absence of a disulfide bridge) while Gln-Gln g-e' interactions led to dimer formation. The results support a role for g-e' ionic attractions in controlling coiled-coil specificity, stability and oligomerization state, possibly through effects on the side-chain packing at the subunit interface.

The role of interhelical ionic interactions in controlling protein folding and stability. De novo designed synthetic two-stranded alpha-helical coiled-coils.

The role of interchain ionic interactions in controlling protein folding and stability has been studied by using de novo designed synthetic two-stranded alpha-helical coiled-coils. The model coiled-coil (denoted as EK) consists of two identical 35-residue polypeptide chains with a heptad repeat KgLaGbAcLdEeKf and a Cys residue at position 2 and an Ala residue at position 16 in each chain. The Lys residues at positions "g" in one chain and Glu residues at positions "e" in the other chain are expected to form interchain ion-pairs in the coiled-coil structure. This peptide forms a stable coiled-coil structure in benign medium (50 mM KCl, 25 mM PO4, pH7) with a [urea]1/2 value of 3.5 M. In contrast, two peptide analogs EE (EgLaGbAcLdEeKf) and KK (KgLaGbAcLdKeEf), which differ from EK in that EE contains only negatively charged Glu residues and KK contains only positively charged Lys residues at both positions e and g, each show a random coil structure in benign buffer. However, peptide EE or KK can form a stable coiled-coil structure if the interchain ionic repulsions are effectively suppressed either by changing pH or by using high salt concentrations. An equimolar mixture of these two peptides displays 100% alpha-helical content under the same conditions. These results demonstrate that although the interhelical ionic attractions are not essential for coiled-coil formation, a large number of these weak interactions can play an important role in the assembly of helices. Though interhelical ionic repulsions destabilize the homo-stranded coiled-coil, electrostatic attractions may stabilize the hetero-stranded coiled-coil. In addition, this study also suggests that the folding process for these synthetic model coiled-coils does not involve a single-stranded alpha-helix as a significantly populated folding intermediate.

Salt effects on protein stability: two-stranded alpha-helical coiled-coils containing inter- or intrahelical ion pairs.

An investigation into the role of surface-accessible ion pairs in protein stability was carried out by determining the effects of added salt (KCl, MgCl2 and LaCl3) at neutral and acidic pH on the stability of de novo designed two-stranded alpha-helical coiled-coils. The effects of salt on the stability of coiled-coils containing interhelical i to i' + 5 or intrahelical i to i + 4 and i to i + 3 Lys-Glu ion pairs were compared to the effects on the stability of a control coiled-coil, which contained no intra- or interhelical ion pair. These studies indicate that ionic interactions contribute to coiled-coil stability. The results show that added salt can have complex effects on protein stability, involving stabilizing and destabilizing contributions with the net effect depending on the nature of the charged residues and ionic interactions present in the protein.

Roles of structural domains in the morphology and surface anchoring of the tetragonal paracrystalline array of Aeromonas hydrophila. Biochemical characterization of the major structural domain.

The tetragonally arranged S-layer of Aeromonas hydrophila contains two morphological domains. The mature S-layer protein of A. hydrophila has a subunit molecular weight of 52,000, and has been reported to contain two structural domains. Here a mutant has been isolated which produces an S-layer of subunit molecular weight 38,650 as determined by sedimentation analysis. This truncated S-protein was exported via the periplasm to the cell surface, but could not self-assemble into a tetragonal array or be anchored to the cell surface. Instead the truncated protein formed cup-like structures which were purified and characterized biochemically. Automated Edman degradation showed that the truncated protein comprised the amino-terminal structural domain of the S-protein. This domain had an increased hydrophobic amino acid content relative to the wild-type protein, and contained approximately 42% beta-sheet, 10% alpha-helix, and 19% beta-turn. Differences in alpha-helix and beta-turn contents between the wild-type and truncated proteins were observed when the effects of pH and SDS were examined, indicating that the carboxy terminus influences the effects of environmental change on the conformation of the S-protein. This lesser carboxy-terminal array also appears to be required for both correct array morphology, and array anchoring, while the greater amino-terminal domain appears to comprise the major morphological core of the surface array.

NOE data demonstrating a compact unfolded state for an SH3 domain under non-denaturing conditions.

The N-terminal SH3 domain of drk (drkN SH3) is unstable, existing in equilibrium between a folded state (Fexch) and an unfolded state (Uexch) under non-denaturing buffer conditions. Using a15N/2H-labeled sample, long range amide NOEs can be observed in the Uexchstate as a result of reduced relaxation, in some cases correlating protons over 40 residues apart. These long range NOEs disappear upon addition of 2 M guanidinium chloride, demonstrating that there are substantial differences between the Uexchand the guanidine denatured states. Calculations using the long range NOEs of the Uexchstate yield highly compact structures having non-native turns and a non-native buried tryptophan residue. These structures agree with experimental stopped-flow fluorescence data and analytical ultracentrifugation results. Since protein stability depends on the structural and dynamic properties of both the folded and unfolded states, this study provides insights into the stability of the drkN SH3 domain. These results provide the first strong NOE-based evidence for compact unfolded states of proteins and suggest that some unfolded states under physiological conditions have specific interactions leading to compact structures.

Insights into the mechanism of heterodimerization from the 1H-NMR solution structure of the c-Myc-Max heterodimeric leucine zipper.

The oncoprotein c-Myc (a member of the helix-loop-helix-leucine zipper (b-HLH-LZ) family of transcription factors) must heterodimerize with the b-HLH-LZ Max protein to bind DNA and activate transcription. It has been shown that the LZ domains of the c-Myc and Max proteins specifically form a heterodimeric LZ at 20 degreesC and neutral pH. This suggests that the LZ domains of the c-Myc and Max proteins are playing an important role in the heterodimerization of the corresponding gene products in vivo. Initially, to gain an insight into the energetics of heterodimerization, we studied the stability of N-terminal disulfide-linked versions of the c-Myc and Max homodimeric LZs and c-Myc-Max heterodimeric LZ by fitting the temperature-induced denaturation curves monitored by circular dichroism spectroscopy. The c-Myc LZ does not homodimerize (as previously reported) and the c-Myc-Max heterodimeric LZ is more stable than the Max homodimeric LZ at 20 degreesC and pH 7.0. In order to determine the critical interhelical interactions responsible for the molecular recognition between the c-Myc and Max LZs, the solution structure of the disulfide-linked c-Myc-Max heterodimeric LZ was solved by two-dimensional 1H-NMR techniques at 25 degreesC and pH 4.7. Both LZs are alpha-helical and the tertiary structure depicts the typical left-handed super-helical twist of a two-stranded parallel alpha-helical coiled-coil. A buried salt bridge involving a histidine on the Max LZ and two glutamate residues on the c-Myc LZ is observed at the interface of the heterodimeric LZ. A buried H-bond between an asparagine side-chain and a backbone carbonyl is also observed. Moreover, evidence for e-g interhelical salt bridges is reported. These specific interactions give insights into the preferential heterodimerization process of the two LZs. The low stabilities of the Max homodimeric LZ and the c-Myc-Max heterodimeric LZ as well as the specific interactions observed are discussed with regard to regulation of transcription in this family of transcription factors.

The entericidin locus of Escherichia coli and its implications for programmed bacterial cell death.

Antidote/toxin gene pairs known as "addiction modules" can maintain plasmids in bacterial populations by means of post-segregational killing. However, several chromosome-encoded addiction modules may provide an entirely distinct function in the programmed cell death of moribund subpopulations under starvation conditions. We now report a novel chromosomal bacteriolytic module of Escherichia coli called the entericidin locus, which is activated in stationary phase under high osmolarity conditions by sigmaS and simultaneously repressed by the osmoregulatory EnvZ/OmpR signal transduction pathway. The entericidin locus encodes tandem paralogous genes (ecnAB) and directs the synthesis of two small cell-envelope lipoproteins. An attenuator precedes ecnA and an ompR-sensitive sigmaS promoter governs expression of ecnB. The entericidin A lipoprotein is an antidote to the bacteriolytic lipoprotein entericidin B. The entericidins are predicted to adopt amphipathic alpha-helical structures and to reciprocally modulate membrane stability. The entericidin locus is not present on any known plasmids, but is conserved in the homologous region of the Citrobacter freundii chromosome. Although the cloned C. freundii entericidin locus is expressed in E. coli independently of ompR, it carries an additional ompR-like gene called ecnR. The organization of the entericidin locus as a chromosomal antidote/toxin gene pair, which is regulated by both positive and negative osmotic signals during starvation, is consistent with an emerging paradigm of programmed bacterial cell death.