Structural determinants that make hexokinase susceptible to ubiquitin-and ATP- dependent proteolysis

Abstract

Intracellular protein degradation is highly selective, although the regulation of the mechanism(s) involved this phenomenon have not been fully elucidated. Protein ubiquitination has been implicated a variety of biological processes including the marking of protein for degradation. Hershko et al. [ l ] found that a free NHz-terminal group is essential for protein ligation to ubiquitin and subsequently others [2,3] reported that also the nature of the NH,-terminal residue has a strong influence. However other structural determinants appear to be important protein recognition by the ubiquitin system [4,5]. We have previously shown that purified rabbit hexokinase (HK) type I, the same HK isozyme present also mammalian brain is conjugated to ubiquitin and then degraded by a rabbit reticulocyte fraction I1 when is soluble 161 but not when is the mitochondrial bound form [7]. In fact we reported that the mitochondrii bound hexokinase is stable for several hours the same proteolytic system both the presence or absence of ATP. El, & and E,, the enzymes of the ubiquitin conjugating system, are able to incorporate 'Ior biotin-labelled ubiquitin, an ATP dependent manner, soluble HK as well as other mitochondria1 proteins and furthermore the mitochondria by themselves have a marked ATPdependent ability to conjugate lzsI-ubiquitin [8]. However the results of Western blotting experiments, using a specific anti-hexokinase type I antibody, confmed that there is a time-dependent decrease the HK immunoreactive material the presence of ATP when the soluble enzyme was tested contrast, the bound hexokinase was absolutely stable both the presence or absence of ATP (Fig. 1). Thus, mitochondrial bound hexokinase is neither ubiquitinated nor degraded. These results suggest that the intracellular distribution of a protein is an important feature which determines its susceptibility to ubiquitin dependent degradation. The different behaviour of the enzyme when is the soluble or the bound form may be explained different ways: one possibility is that the ubiquitin conjugating system is not able to ligate ubiquitin to hexokinase when it is bound to mitochondria due to the masking of the binding sites. The second explanation is that the soluble enzyme can undergo post-translational modifications, including partial NH2-terminal proteolysis producing in vivo truncate forms of hexokinase type I (Fig. 2) [9]. To differentiate between these possibilities and especially to investigate the role of the last phenomenon on ATPand ubiquitin-dependent proteolysis we have cloned and expressed bacterial cells the full-length human hexokinase type I, a truncate form lacking the first 11 amino acids and a mini-hexokinase corresponding to the COOH-terminal half of the enzyme (which contains the catalytic site). The expression E. coli cells of the different forms of recombinant HKs was obtained by the PET expression system (Novagen). Briefly, the cDNAs coding for the full-length HK, the HK lacking the first 11 aa and the COOH-terminal domain were all derived from the hhHEX-15 clone, purchased from ATCC, Maryland, (which includes nucleotides 383598, encoding a complete 917 amino acid protein) and then ligated with a PET expression plasmid. The recombinant PET expression vectors were then transformed into the E. coli cells BL21(DE3) and cultures of the transformants were induced with IPTG to express the recombinant HKs. The bacterial cells after induction were harvested and sonicated as reported (101 and the recombinant catalitycally active hexokinase proteins were purified to homogeneity as described [ 111. All these purified recombinant HKs were fmally tested for their stability a cell-free, ATPand ubiquitindependent pruteolytic system.