Search

Upon fasting, the ubiquitin-dependent proteolytic system is activated in skeletal muscle in parallel with the increases in rates of proteolysis. Levels of mRNA encoding the 14 kDa ubiquitin-conjugating enzyme (E2(14K)), which can catalyse the first irreversible reaction in this pathway, rise and fall in parallel with the rates of proteolysis [Wing and Banville (1994) Am.J. Physiol. 267, E39-E48], indicating that the conjugation of ubiquitin to proteins is a regulated step. To characterize the mechanisms of this regulation, we have examined the effects of insulin, insulin-like growth factor I (IGF-I) and des(1-3) insulin-like growth factor I (DES-IGF-I), which does not bind IGF-binding proteins, on E2(14K) mRNA levels in L6 myotubes. Insulin suppressed levels of E2(14K) mRNA with an IC50 of 4 x 10(-9) M, but had no effects on mRNAs encoding polyubiquitin and proteasome subunits C2 and C8, which, like E2(14K), also increase in skeletal muscle upon fasting. Reduction of E2(14K) mRNA levels was more sensitive to IGF-I with an IC50 of approx. 5 x 10(-10) M. During the incubation of these cells for 12 h there was significant secretion of IGF-I-binding proteins into the medium. DES-IGF-I, which has markedly reduced affinity for these binding proteins, was found to potently reduce E2(14K) mRNA levels with an IC50 of 3 x 10(-11) M. DES-IGF-I did not alter rates of transcription of the E2(14K) gene, but enhanced the rate of degradation of the 1.2 kb mRNA transcript. The half-life of the 1.2 kb transcript was approximately one-third that of the 1.8 kb transcript and can explain the more marked regulation of this transcript observed previously. This indicates that the additional 3' non-coding sequence in the 1.8 kb transcript confers stability. These observations suggest that IGF-I is an important regulator of E2(14K) expression and demonstrate, for the first time, stimulation of degradation of a specific mRNA transcript by this hormone, while overall RNA accumulates.

Most of the increased protein degradation in muscle atrophy caused by starvation and denervation is due to activation of a non-lysosomal ATP-dependent proteolytic process. To determine whether expression of the ubiquitin-proteasome-dependent pathway is activated in atrophying muscles, we measured the levels of mRNA for ubiquitin (Ub) and proteasome subunits, and Ub content. After rats had been deprived of food for 1 or 2 days, the concentration of the two polyubiquitin (polyUb) transcripts increased 2-4-fold in the pale extensor digitorum longus muscle and 1-2.5-fold in the red soleus, whereas total muscle RNA and total mRNA content fell by 50%. After denervation of the soleus, there was a progressive 2-3-fold increase in polyUb mRNA for 1-3 days, whereas total RNA content fell. On starvation or denervation, Ub concentration in the muscles also rose by 60-90%. During starvation, polyUb mRNA levels also increased in heart, but not in liver, kidney, spleen, fat, brain or testes. Although the polyUb gene is a heat-shock gene that is induced in muscles under certain stressful conditions, the muscles of starving rats or after denervation did not express other heat-shock genes. On starvation or denervation, mRNA for several proteasome subunits (C-1, C-3, C-5, C-8 and C-9) also increased 2-4-fold in the atrophying muscles. When the food-deprived animals were re-fed, levels of Ub and proteasome mRNA in their muscles returned to control values within 1 day. In contrast, no change occurred in the levels of muscle mRNAs encoding cathepsin L, cathepsin D and calpain 1 on denervation or food deprivation. Thus polyUb and proteasome mRNAs increased in atrophying muscles in co-ordination with activation of the ATP-dependent proteolytic process.

The rapid loss of skeletal-muscle protein during starvation and after denervation occurs primarily through increased rates of protein breakdown and activation of a non-lysosomal ATP-dependent proteolytic process. To investigate whether protein flux through the ubiquitin (Ub)-proteasome pathway is enhanced, as was suggested by related studies, we measured, using specific polyclonal antibodies, the levels of Ub-conjugated proteins in normal and atrophying muscles. The content of these critical intermediates had increased 50-250% after food deprivation in the extensor digitorum longus and soleus muscles 2 days after denervation. Like rates of proteolysis, the amount of Ub-protein conjugates and the fraction of Ub conjugated to proteins increased progressively during food deprivation and returned to normal within 1 day of refeeding. During starvation, muscles of adrenalectomized rats failed to increase protein breakdown, and they showed 50% lower levels of Ub-protein conjugates than those of starved control animals. The changes in the pools of Ub-conjugated proteins (the substrates for the 26S proteasome) thus coincided with and can account for the alterations in overall proteolysis. In this pathway, large multiubiquitinated proteins are preferentially degraded, and the Ub-protein conjugates that accumulated in atrophying muscles were of high molecular mass (> 100 kDa). When innervated and denervated gastrocnemius muscles were fractionated, a significant increase in ubiquitinated proteins was found in the myofibrillar fraction, the proteins of which are preferentially degraded on denervation, but not in the soluble fraction. Thus activation of this proteolytic pathway in atrophying muscles probably occurs initially by increasing Ub conjugation to cell proteins. The resulting accumulation of Ub-protein conjugates suggests that their degradation by the 26S proteasome complex subsequently becomes rate-limiting in these catabolic states.

Ubiquitin-conjugating enzymes (E2s) play a key role in ubiquitin-mediated proteolysis by catalysing the conjugation of ubiquitin to protein substrates. We have previously reported the cDNA cloning of a 14 kDa conjugating enzyme [E2(14)k; Wing, Dumas and Banville (1992) J. Biol. Chem. 267, 6495-6501] that efficiently supported ubiquitination and protein degradation in reticulocyte extracts. Surprisingly, the structure of this E2 was markedly more similar to the Saccharomyces cerevisiae DNA repair gene RAD6, than to the S. cerevisiae UBC4/UBC5 genes which are required for the degradation of short-lived proteins and support much of the ubiquitination of yeast proteins. This suggested that mammalian homologues of UBC4/UBC5 remained to be identified. Using oligonucleotides derived from the S. cerevisiae UBC4 sequence as primers in a PCR reaction with rat muscle cDNA as a template, a 390 bp DNA fragment was amplified which predicted an amino acid sequence that was 83% identical to yeast UBC4. Screening a rat testes cDNA library identified a family of cDNAs which predicted two very similar proteins with basic pIs and molecular masses of approx. 16,700 Da. Isoform 2E was expressed in Escherichia coli and purified to homogeneity. It supported ubiquitination to reticulocyte and testis proteins more rapidly in vitro and produced larger conjugates than E2(14)k. Examination of RNA from different tissues indicated that this type of E2 was expressed in a broad spectrum of tissues but at particularly high levels in the testis. Fractionation of a testis extract by anion-exchange chromatography identified several putative ubiquitin protein ligase activities with which this E2 could interact in promoting conjugation of ubiquitin to proteins. One of these activities supported conjugation of ubiquitin to histone H2A, a substrate degraded in the ubiquitin system by a non-N-end rule mechanism. This paper reports the first cloning of a apparent mammalian homologue of S. cerevisiae UBC4/UBC5. Its high expression in testis and ability to efficiently support conjugation to testis proteins suggest that this family of E2s may play a role in the proteolysis that occurs during spermatogenesis.

In response to serum withdrawal, when overall rates of proteolysis increase in cultured fibroblasts, proteins containing peptide regions similar to Lys-Phe-Gln-Arg-Gln (KFERQ) are targeted to lysosomes for degradation, and the intracellular concentrations of these proteins decline [Chiang & Dice (1988) J. Biol. Chem. 263, 6797-6805]. To test whether such proteins are also selectively depleted in mammalian tissues in vivo, we have used affinity-purified polyclonal antibodies to KFERQ to detect proteins containing such sequences in tissues of fed and fasted rats. Immunoreactive cytosolic proteins were partially depleted from liver and heart of fasted rats, but the time course differed for these two tissues. Immunoreactive proteins in liver were lost during days 2 and 3 of fasting, whereas such proteins in heart were depleted within day 1 of fasting. In the same fasted rats, levels of immunoreactive cytosolic proteins did not change in two skeletal muscles, the dark soleus and the pale extensor digitorum longus. Immunoreactive proteins in a myofibrillar fraction were also partially depleted in heart, but not in skeletal muscles, of fasted rats. The most likely explanation for these results is that the protein loss in different tissues upon fasting results from selective activation of different proteolytic pathways. The increased proteolysis in liver and heart of fasted animals includes activation of the KFERQ-selective lysosomal pathway, whereas increased proteolysis in skeletal muscle does not.