Intramolecular ionic interactions of lysine residues and a possible folding domain in fructose diphosphate aldolase

Cytosolic lysine residues enhance anterograde transport and activation of the erythropoietin receptor

Lysine residues are key residues in many cellular processes, in part due to their ability to accept a wide variety of post-translational modifications. In the present study, we identify the EPO-R [EPO (erythropoietin) receptor] cytosolic lysine residues as enhancers of receptor function. EPO-R drives survival, proliferation and differentiation of erythroid progenitor cells via binding of its ligand EPO. We mutated the five EPO-R cytosolic lysine residues to arginine residues (5KR EPO-R), eliminating putative lysine-dependent modifications. Overexpressed 5KR EPO-R displayed impaired ubiquitination and improved stability compared with wt (wild-type) EPO-R. Unexpectedly, fusion proteins consisting of VSVGtsO45 (vesicular stomatitis virus glycoprotein temperature-sensitive folding mutant) with wt or 5KR EPO-R cytosolic domains demonstrated delayed glycan maturation kinetics upon substitution of the lysine residues. Moreover, VSVG-wt EPO-R, but not VSVG-5KR EPO-R, displayed endoplasmic reticulum-associated ubiquitination. Despite similar cell-surface EPO-binding levels of both receptors and the lack of EPO-induced ubiquitination by 5KR EPO-R, the lysine-less mutant produced weaker receptor activation and signalling than the wt receptor. We thus propose that EPO-R cytosolic lysine residues enhance receptor function, most probably through ubiquitination and/or other post-translational modifications.

Fructose-1,6-bisphosphate aldolase from Drosophila melanogaster: primary structure analysis, secondary structure prediction, and comparison with vertebrate aldolases

The amino acid sequence of fructose-1,6-bisphosphate aldolase from Drosophila melanogaster was determined and was compared with those of five vertebrate aldolases on record. The four identical polypeptide chains of the insect enzyme, acetylated at the N-terminus and three residues shorter than the vertebrate chains, contain 360 amino acid residues. Of these 190 (or 53%) are identical in all six enzymes and in addition 33 positions (or 9%) are occupied by homologous residues. Comparison with the muscle-type isoaldolases from man and rabbit and the liver-type isoaldolases from man, rat, and chicken indicates an average sequence identity of 70 and 63%, respectively. Thus, the insect and the vertebrate muscle aldolases are probably coded by orthologous genes. On this basis an average rate of evolution of 3.0 PAM per 10(8) years is calculated, documenting an evolutional divergence slower than that of cytochrome c (4.2 PAM/10(8) years). The rate is also lower than that of the liver isoform (3.6 PAM/10(8) years). Secondary structure prediction analysis for Drosophila aldolase suggests the occurrence of 11-12 helical segments and 8-9 beta-strands. The conspicuous alternation of these structures in all six aldolases, especially in the C-terminal 200 residues, is consistant with the formation of an alpha beta-barrel supersecondary structure as documented for several other glycolytic enzymes.

Characterization of Neurospora crassa catabolic dehydroquinase purified from N. crassa and Escherichia coli

1. Neurospora crassa catabolic dehydroquinase has been purified from N. crassa and Escherichia coli. 2. Protein-sequence and gel-electrophoretic data show that apparently pure, homogeneous native dehydroquinase is a mixture of intact and proteinase-cleaved enzyme monomers. 3. Protein-sequence data and steady-state kinetics show that the catabolic dehydroquinase gene of N. crassa is expressed with fidelity in E. coli.

Identification of "buried" lysine residues in two variants of chloramphenicol acetyltransferase specified by R-factors

Two variants of chloramphenicol acetyltransferase which are specified by genes on plasmids found in Gram-negative bacteria were subjected to amidination with methyl acetimidate to determine the relative reactivity of surface lysine residues and to search for unreactive or "buried" amino groups which might contribute to stabilization of the native tetramers. Representative examples of the type-I and type-III variants of chloramphenicol acetyltransferase were found to have one lysine residue each in the native state which appears to be inaccessible to methyl acetimidate. The uniquely unreactive residue of the type-I protein is lysine-136, whereas the lysine that is "buried" in the type-III enzyme is provisonally assigned to residue 38 of the prototype sequence. It is suggested that the lysine residue in each case participates in the formation of an ion pair at the intersubunit interface and that the two amino groups in question occupy functionally equivalent positions in the quaternary structures of their respective enzyme variants. Lysine-136 of type-I enzyme is also uniquely unavailable for modification by citraconic anhydride, a reagent used to disrupt the quaternary structure of the native enzyme. Contrary to expectation, exhaustive citraconylation fails to dissociate the tetramer, but does destroy catalytic activity. Removal of citraconyl groups from modified chloramphenicol acetyltransferase is accompanied by a full region of catalytic activity. Analysis of the rate of hydrolysis of citraconyl groups from the modified tetramer by amidination of unblocked amino groups with methyl [14C]acetamidate reveals difference in lability for several of the ten modified lysine residues. Although the unique stability of the quaternary structure of chloramphenicol acetyltransferase may be due to strong hydrophobic interactions, it is argued that lysine-136 may contribute to stability via the formation of an ion pair at the subunit interface.

The use of naturally occurring hybrid variants of chloramphenicol acetyltransferase to investigate subunit contacts

1. Hybrids of the tetrameric enzyme chloramphenicol acetyltransferase (EC 2.3.1.28) were formed in vivo in a strain of Escherichia coli which harbours two different plasmids, each of which normally confers chloramphenicol resistance and specifies an easily distinguished enzyme variant (type I or type III) which is composed of identical subunits. Cell-free extracts of the dual-plasmid strain were found to contain five species of active enzyme, two of which were the homomeric enzymes corresponding to the naturally occurring tetramers of the type-I (beta 4) and type-III (alpha 4) enzymes. The other three variants were judged to be the heteromeric hybrid variants (alpha 3 beta, alpha 2 beta 2, alpha beta 3). 2. The alpha 3 beta and alpha 2 beta 2 hybrids of chloramphenicol acetyltransferase were purified to homogeneity by combining the techniques of affinity and ion-exchange chromatography. The alpha beta 3 variant was not recovered and may be unstable in vitro. 3. The unique lysine residues that could not be modified with methyl acetimidate in each of the native homomeric enzymes were also investigated in the heteromeric tetramers. 4. Lysine-136 remains buried in each beta subunit of the parental (type I) enzyme and in each of the hybrid tetramers. Lysine-38 of each alpha subunit is similarly unreactive in the native type-III chloramphenicol acetyltransferase (alpha 4), but in the alpha 2 beta 2 hybird lysine-38 of each alpha subunit is fully exposed to solvent. Another lysine residue, fully reactive in the alpha 4 enzyme, was observed to be inaccessible to modification in the symmetrical hybrid. The results obtained for the alpha 3 beta enzyme suggest that lysine-38 in two subunits and a different lysine group (that identified in the alpha 2 beta 2 enzyme) in the third alpha subunit are buried. 5. A tentative model for the subunit interactions of chloramphenicol acetyltransferase is proposed on the basis of the results described.

Binding of nicotinamide-adenine dinucleotide to rabbit muscle aldolase

A direct interaction of rabbit muscle fructose-1,6-bisphosphate aldolase with NAD+, NADH, and NAD-agarose was demonstrated. The electrostatic forces are primary involved in this interaction. Two specific binding sites for the dinucleotide were observed. One of them is located at the active site of the enzyme, the second is in a region of weak binding site for ATP and fructose 1,6-bisphosphate.

Extended amino acid sequences around the active-site lysine residue of class-I fructose 1,6-bisphosphate aldolases from rabbit muscle, sturgeon muscle, trout muscle and ox liver

1. Amino acid sequences covering the region between residues 173 and 248 [adopting the numbering system proposed by Lai, Nakai & Chang (1974) Science 183, 1204-1206] were derived for trout (Salmo trutta) muscle aldolase and for ox liver aldolase. A comparable sequence was derived for residues 180-248 of sturgeon (Acipenser transmontanus) muscle aldolase. The close homology with the rabbit muscle enzyme was used to align the peptides of the other aldolases from which the sequences were derived. The results also allowed a partial sequence for the N-terminal 39 residues for the ox liver enzyme to be deduced. 2. In the light of the strong homology evinced for these enzymes, a re-investigation of the amino acid sequence of rabbit muscle aldolase between residues 181 and 185 was undertaken. This indicated the presence of a hitherto unsuspected -Ile-Val-sequence between residues 181 and 182 and the need to invert the sequence -Glu-Val- to -Val-Glx- at positions 184 and 185. 3. Comparison of the available amino acid sequences of these enzymes suggested an early evolutionary divergence of the genes for muscle and liver aldolases. It was also consistent with other evidence that the central region of the primary structure of these enzymes (which includes the active-site lysine-227) forms part of a conserved folding domain in the protein subunit. 4. Detailed evidence for the amino acid sequences proposed has been deposited as Suy Lending Division, Boston Spa, Wetherby, West Yorkshire LS23 7BQ, U.K., from whom copies can be obtained on the terms indicated in Biochem. J. (1978) 169, 5.

Folding domains and intramolecular ionic interactions of lysine residues in glyceraldehyde 3-phosphate dehydrogenase

1. Treatment with methyl acetimidate was used to probe the topography of several tetrameric glyceraldehyde 3-phosphate dehydrogenases, in particular the holoenzymes from rabbit muscle and Bacillus stearothermophilus. During the course of the reaction with the rabbit muscle enzyme, the number of amino groups fell rapidly from the starting value of 27 per subunit to a value of approx. five per subunit. This number could be lowered further to values between one and two per subunit by a second treatment with methyl acetimidate. The enzyme remained tetrameric throughout and retained 50% of its initial catalytic activity at the end of the experiment. 2. Use of methyl [1-14C]acetimidate and small-scale methods of protein chemistry showed that only one amino group per subunit, that of lysine-306, was completely unavailable for reaction with imido ester in the native enzyme. This results is consistent with the structure of the highly homologous glyceraldehyde 3-phosphate dehydrogenase of lobster muscle deduced from X-ray-crystallographic analysis, since lysine-306 can be seen to form an intrachain ion-pair with aspartic acid-241 in the hydrophobic environment of a subunit-subunit interface. 3. Several other amino groups in the rabbit muscle enzyme that reacted only slowly with the reagent were also identified chemically. These were found to be located entirely in the C-terminal half of the polypeptides chain, which comprises a folding domain associated with catalytic activity and subunit contact in the three-dimensional structure. Slow reaction of these 'surface' amino groups with methyl acetimidate is attributed to intramolecular ionic interactions of the amino groups with neighbouring side-chain carboxyl groups, a conclusion that is compatible with the reported three-dimensional structure and with the dependence of the reaction of ionic stength. 4. Very similar results were obtained with the enzymes from B. stearothermophilus and from ox muscle and ox liver, supporting the view that the ion-pair involving lysine-306 and aspartic acid-241 will be a common structural feature in glyceraldehyde-3-phosphate dehydrogenases. The B. stearothermophilus enzyme was fully active after modification. 5. No differences could be detected between the enzymes from ox muscle and ox liver, in accord with other evidence that points to the identify of these enzymes. 6. The pattern of slowly reacting amino groups in the enzyme from B. stearothermophilus, although similar to that of the mammalian enzymes, indicated one or two additional intramolecular ionic interactions of lysine residues that might contribute to the thermal stability of this enzyme.