Solubilizing buried domains of proteins: a self-assembling interface domain from glutathione reductase.

In the dimeric glutathione reductase (GR) from Escherichia coli, the interface domain is largely surrounded by the other three domains in each subunit of the protein. Subgenes encoding three forms of the interface domain have been expressed in E. coli and the products purified from inclusion bodies: INT is the excised interface domain, as it is found in native GR; INTN and INTFN are variants carrying exchanges of surface residues in what would have been hydrophobic contact regions with other neighboring domains. The isolated INT domain was found to be a soluble and folded protein, but it was isolated as a mixture of the dimer and at least two species of higher molecular weight. The latter were believed to arise by further association of the dimer via the newly exposed and unsatisfied hydrophobic contact regions. In the variant INTN, three hydrophobic residues normally involved in the contact with the NADPH-binding domain in GR were replaced. This partly suppressed the further aggregation of the dimers. However, continued aggregation at high protein concentrations suggested that at least one further site of unwanted aggregation was still present. After four additional amino acid replacements in the region normally in contact with the FAD-binding domain, the resulting variant INTFN exhibited no unspecific aggregation, even at concentrations as high as 3.2 mg/mL.(ABSTRACT TRUNCATED AT 250 WORDS)

Coenzyme binding of a folding intermediate of aspartate aminotransferase detected by HPLC fluorescence measurements.

Equilibrium dissociation and unfolding of dimeric aspartate aminotransferase from Escherichia coli proceeds via two compact monomeric intermediates which have similar hydrodynamic volumes but different fluorescence properties. We probed binding of the coenzyme pyridoxal 5'-phosphate to these intermediates by coupling fluorescence detection to size-exclusion HPLC. This procedure gave additionally an internal conformational probe of the unfolding transitions of the enzyme. It was shown that the first intermediate, M, is able to bind the coenzyme, whereas the second intermediate, M*, is not. It is likely that M is the correctly folded monomer of the protein.

Collapsed intermediates in the reconstitution of dimeric aspartate aminotransferase from Escherichia coli.

Aspartate aminotransferase from Escherichia coli, which had been denatured by guanidinium chloride, refolded and reassembled to active dimers in two distinct phases. The unfolded monomer U collapsed within 20 s to an intermediate I* that was inactive, fluoresced more strongly than, but had the same peptide CD signal as the native dimer. The formation of crosslinkable dimers, as well as the recovery of enzyme activity, occurred with a biphasic progress curve which was independent of protein concentration. The half-lives of the two phases were 100 s and 2000 s. The data are consistent with a three-step mechanism, in which the overall rate of reassembly is determined by an isomerization of I* to the assembly-competent monomer M. The latter does not accumulate because it dimerizes rapidly to the active enzyme (D). Reassembly of the enzyme from the compact intermediate M*, which is stable at 1.0 M guanidinium chloride, also proceeded in a rapid and a slow phase. Moreover, the formation of M* from the unfolded state was rapid, whereas its refolding to the native dimer was slow. Both the transient intermediate I* and the equilibrium intermediate M* qualify as 'collapsed intermediate' or 'molten globule' states.

Autonomous folding and coenzyme binding of the excised pyridoxal 5'-phosphate binding domain of aspartate aminotransferase from Escherichia coli.

The coenzyme (PLP) binding domain (residues 47-329) of the dimeric aspartate aminotransferase from Escherichia coli was produced separately by recombinant DNA methods. It folded autonomously both in vivo and in vitro, that is, independently of the native N- and C-terminal extensions that combine to form the small domain of eAAT. The PLP-domain had one binding site for PLP of relatively high affinity involving a covalent bond to the protein. It was monomeric, although the major subunit-subunit interface at the 2-fold symmetry axis remained unchanged. This effect appears to be due mainly to the absence of the N-terminal extension that contains hydrophobic residues, which interact with the PLP-domain of the second subunit in the wild-type dimer. Judged by circular dichroism, fluorescence, and HPLC gel filtration at increasing concentrations of guanidinium chloride, the PLP-domain underwent a three-state unfolding transition (M' in equilibrium M'* in equilibrium U') involving a compact intermediate M'*. This behavior parallels the unfolding of the dissociated native monomer of cAAT.

Unfolding of truncated and wild type aspartate aminotransferase studied by size-exclusion chromatography.

The reversible unfolding of globular proteins with increasing concentration of guanidinium chloride (GuCl) can be analysed by size-exclusion chromatography, because the hydrodynamic volume of the proteins increases during unfolding. The dimeric enzyme aspartate aminotransferase (AAT) shows an uncoupled dissociation of the identical subunits followed by the unfolding of the monomers. During the monomer unfolding formation of an intermediate is observed. A monomeric mutant of AAT unfolds with a similar shape of the unfolding transition phase, but is less stable, as shown by a shift of the transition mid-point from 1.7 M GuCl for the wild type to 1.3 M GuCl for the mutant.