The maltose transporter in Saccharomyces cerevisiae is degraded in the vacuole after internalization by endocytosis when protein synthesis is impaired and a fermentable substrate is present. The possible implication of the ubiquitin pathway in this inactivation, known as catabolite inactivation, has been investigated. Using mutants deficient in npi1/rsp5 ubiquitin-protein ligase and npi2/doa4 ubiquitin-protein hydrolase, we have shown that these two enzymes are required for normal endocytosis and degradation of the transporter. These facts indicate that the ubiquitin pathway is involved in catabolite inactivation of the maltose transporter. The results also revealed that both enzymes act in the internalization step of endocytosis.
The maltose transporter is inactivated in yeast when protein synthesis is impaired and a fermentable carbon source is present . This inactivation, known as catabolite inactivation, is due to a proteolysis of the transporter  which does not require the proteasome function and which occurs in the vacuole after internalization by endocytosis . The experiments reported here attempt to establish if the ubiquitin pathway is implicated in the vacuolar degradation of this plasma membrane protein. For many years the most thoroughly characterized function of protein ubiquitination has been as a signal for degradation of cytosolic proteins in the proteasome (for review see ). However, recent findings strongly indicate that ubiquitination may also act as a signal for internalization by endocytosis and subsequent degradation of plasma membrane proteins in the vacuole [5–9].
The ubiquitin pathway consists of two distinct steps, marking of the protein by attachment of ubiquitin and degradation of the tagged protein with recycling of free ubiquitin. In the first step different groups of enzymes are involved: ubiquitin-activating enzymes (E1), which activate ubiquitin by forming a thiol-ester intermediate with the C-terminus of ubiquitin, and ubiquitin-conjugating enzymes (E2), which transfer activated ubiquitin from E1 to the target protein bound to a ubiquitin-protein ligase (E3). In the second step free ubiquitin is released in the terminal stages of protein degradation by the action of ubiquitin-protein hydrolases (E4) . To establish if the ubiquitin pathway is involved in endocytosis and degradation of the maltose transporter we used E3  and E4  deficient mutant strains. In these strains we investigated the internalization step of endocytosis by measuring maltose transport with radioactive sugar and degradation of the transporter with polyclonal antibodies. The results obtained strongly indicate that the ubiquitin pathway is involved in catabolite inactivation of the yeast maltose transporter.
Materials and methods
d-(U-14C)Maltose and ECL chemiluminescence reagents were from Amersham International (Amersham, UK). Goat antibodies anti-rabbit peroxidase conjugate was from Biosource International (Camatillo, CA).
The following strains were used: 23346c (MATa NPI1 ura3), 27038a (MATa npi1 ura3) , MHY501 (MATα DOA4 his3-Δ200 leu2-3,112 ura3-52 lys2-801 trp1-1), MHY623 (MATα doa4::LEU2 his3-Δ200 leu2-3 ura3-52 lys2-901 trp1-1) . These strains were transformed with the multicopy plasmid pRM1-1, which carries the MAL1 locus . The transformed cells, which grew and transported maltose at normal rates, were grown at 30°C in a rotatory shaker (200 rpm) in a YNB minimal medium with 2% maltose and in the presence of 3 ppm antimycin A. Cell growth was monitored by measuring optical densities at 640 nm. To start inactivation, cells were harvested during exponential growth (about 0.7 mg dry weight per ml), washed and suspended in three volumes of an ammonium-free medium as described previously  in the presence of 2% glucose and 250 μg of tetracycline chlorhydrate per ml to avoid bacterial contamination. The suspension was incubated at 30°C in a rotatory shaker (200 rpm). Internalization by endocytosis of the maltose transporter was followed by measuring the decrease in the rate of maltose transport. Maltose transport was determined using radioactive maltose as described previously . Degradation of the transporter was followed with polyclonal antibodies using crude cellular extracts . Samples containing 20 μg protein were resolved by 10% SDS-PAGE and the maltose transporter was visualized using ECL chemiluminescence detection. Antiserum against maltose transporter diluted 1/3000 in blocking buffer and goat anti-rabbit peroxidase conjugate diluted 1/10.000 were used as primary and secondary antibodies respectively. Protein was determined after precipitation with trichloroacetic acid using the method of Lowry et al. .
Inactivation of the maltose transporter in an ubiquitin-protein ligase deficient mutant
Two ubiquitin-protein ligases (E3 enzymes) have so far been identified in yeast cells. One is Ubr1p, which acts through the N-end rule pathway . The other is Npi1p/Rsp5p which is similar to human E6-AP . To investigate if signaling with ubiquitin is involved in endocytosis and degradation of the maltose transporter we used a mutant strain deficient in the Npi1/Rsp5 ubiquitin-protein ligase. This enzyme has been shown to be implicated in endocytosis and degradation in the vacuole of two plasma membrane proteins, the general amino acid permease  and uracil permease . A single and essential gene, gene NPI1/RSP5, codes for Npi1p/Rsp5p S. cerevisiae and we used a viable npi1/rsp5 strain which shows reduced expression of NPI1/RSP5. We found that internalization of the transporter, monitored by measuring the decrease in transport activity, was substantially reduced in the mutant compared with the isogenic wild-type strain (Fig. 1A). While in wild-type cells the calculated half-life was about 0.5 h, the value was about 1.8 h in npi1/rsp5 mutant cells. A similar difference between the mutant and the wild-type strain became apparent when degradation was followed with antibodies (Fig. 1B). Also in this case degradation was substantially reduced in the mutant compared with the wild-type strain.
Inactivation of the maltose transporter in an ubiquitin-protein hydrolase deficient mutant
Operation of the ubiquitin pathway requires recycling of ubiquitin. This function is carried out by ubiquitin C-terminal hydrolases (E4 enzymes) which may also be implicated in other functions like processing of ubiquitin precursors and disassembling of the polyubiquitin chains linked to the protein substrate during the degradative process . Five E4 hydrolases, at least, have been identified in S. cerevisiae. Npi2p/Doa4p is a member of this family of deubiquitinating enzymes, and elimination of this enzyme from the cells results in a strong inhibition of proteolysis of all tested ubiquitin-dependent substrates . Using a disruption mutant of the NPI2/DOA4 gene  we found that internalization (Fig. 2A) as well as degradation of the maltose transporter (Fig. 2B) in the npi2/doa4 mutant strain were substantially reduced compared with the wild-type isogenic strain. From data of Fig. 2A a half-life value of about 6 h could be calculated in mutant cells, whereas the calculated value was about 0.8 h in wild-type cells.
The results show that two enzymes, at least, of the ubiquitin pathway, Npi1/Rsp5 ubiquitin-protein ligase and Npi2/Doa4 ubiquitin-protein hydrolase, are required for normal catabolite inactivation of the yeast maltose transporter. These facts suggest that ubiquitin is implicated in proteolysis of this transporter. The results also show that these two enzymes act in an early stage of this proteolysis, i.e. in the internalization of the transporter. This conclusion comes from the observation that the maltose transporter remains active in the plasma membrane, being internalized at a low rate, when activity of any of these two enzymes decreased. The observed differences in the effect of these two enzymes could be related with differences in their cellular expression. A reduced expression of the ubiquitin-ligase takes place in the npi1 mutant cells  whereas no expression at all takes place in the case of the ubiquitin-hydrolase . The results are in agreement with the postulated role of ubiquitin-protein ligases facilitating transfer of activated ubiquitin to protein substrates [4, 6, 8]. In addition, the results suggest a role of the Npi2/Doa4 ubiquitin-hydrolase in vacuolar proteolysis not previously reported. Ubiquitin-hydrolases, in general, are involved in recycling of free ubiquitin. In addition, distinct ubiquitin-hydrolases seem to serve different functions at different stages of proteolysis: (i) cleaving of ubiquitin in early stages, during processing of polyubiquitin-chain precursors; (ii) cleaving of ubiquitin from protein conjugates in the middle stages of proteolysis, to reverse the modification of inappropriately targeted proteins; (iii) cleaving of ubiquitin in the late stages, during proteolysis of protein conjugates by the 26S proteasome . In the case of Npi2/Doa4 hydrolase, the contribution to recycling of free ubiquitin seems rather low  whereas this enzyme seems to play an important role in late stages of proteolysis by the proteasome, in disassembling multiubiquitin chains on protein fragments still bound to the 26S proteasome . Our results indicate that, in addition, Npi2p/Doa4p also plays an important role in vacuolar proteolysis of the maltose transporter although, in this case, its function occurs at an early stage, in internalization of the protein. A possible function of Npi2/Doa4 in degradation of this protein in the vacuole is cleaving of ubiquitin during processing of the polyubiquitin chain precursors. This processing occurs in early stages of proteolysis, before signaling the target proteins with ubiquitin. Therefore, a deficient processing of the precursors could affect signaling for internalization not only of the maltose transporter but also of other plasma membrane proteins. This possibility is supported by the observation that Npi2p/Doa4p is also required for internalization of the general amino acid permease , another plasma membrane protein which is degraded in the vacuole .
The results presented in this work support the emerging view [5–9] that ubiquitin binding can function not only as a signal for proteolysis by the proteasome, as accepted for many years, but also as a signal that triggers endocytosis of plasma membrane proteins for degradation in the vacuole.
We are very grateful to A. André for the gift of the npi1/rsp5 strain, to M. Hochstrasser for the gift of the npi2/doa4 strain, to R. Rodicio for the gift of plasmid pRM1-1 and to M. Herweijer for the gift of the polyclonal antibodies. We are also very grateful to J. Pérez and A. Fernandez for help in the preparation of the figures, and to C. Gancedo for critically reading the manuscript. This work was supported by the Spanish Dirección General Científica y Técnica (PB 94-0091-CO201) and by the European Commission (BIO4-CT95-0107).