Monoubiquitination Is Sufficient To Signal Internalization of the Maltose Transporter in Saccharomyces cerevisiae

Abstract

Monoubiquitination of the 12-transmembrane segment (12-TMS) Saccharomyces cerevisiae maltose transporter promoted the maximal internalization rate of this protein. This modification is similar to that of the 7-TMS α-factor receptor but different from that of the 12-TMS uracil and general amino acid permeases. This result shows that binding of ubiquitin-Lys63 chains is not required for maximal internalization of all 12-TMS-containing proteins.

Binding of ubiquitin (Ub) acts as a signal for at least two different processes in Saccharomyces cerevisiae: for the degradation of cytosolic proteins by the proteasome (3) and for the internalization, for subsequent degradation, of plasma membrane proteins in the vacuole (8). Ub binds through its C terminus to a lysine residue found within target proteins by the action of a cascade of enzymes: Ub-activating enzymes (E1), Ub-conjugating enzymes (E2), and Ub-protein ligase enzymes (E3) (10). Since Ub itself contains seven lysine residues within its sequence, multi-Ub chains bound to proteins can be formed by linking the C terminus of one Ub to a lysine within another Ub. It has been proposed that the differences between the Ub chains bound to cytosolic and plasma membrane proteins could serve for recognition by their respective degradation systems (7, 22, 23).

In yeast cells, Ub chains linked through Lys29, -48, and -63 are present in vivo (1, 6, 11, 21), and it is well established that Ub-Lys48 chains are responsible for the recognition of cytosolic proteins by the proteasome (6). In the case of plasma membrane proteins, it has been shown that Ub-Lys63 chains play a role in the internalization of the uracil (7) and of the general amino acid permeases (22), two 12-transmembrane segment (12-TMS) proteins. However, monoubiquitination is sufficient for internalization of the α-factor receptor (23), a 7-TMS protein. It has been postulated that the different ubiquitination requirements for internalization of these two types of plasma membrane proteins might be related to the differences in their size and the TMS number (23).

To test this hypothesis, we have examined the type of ubiquitination required for internalization of another 12-TMS protein, the maltose transporter. This transporter is internalized and degraded in the vacuole during nitrogen starvation when a fermentable carbon source is present (12, 14, 18). This process requires the binding of Ub and the action of both Ub ligase Npi1/Rsp5 and Ub-protein hydrolase Doa4/Npi2 (13, 15). Free Ub is present at low levels in cells lacking Doa4p (16). For this reason, internalization of plasma membrane proteins is substantially reduced in Δdoa4 mutant cells (7, 13, 15, 22, 23). Based on the fact that this phenotype can be complemented with an overproduction of Ub (7, 22, 23), we investigated the effect of overexpressing mutant Ubs carrying Lys→Arg mutations, which prevent the formation of various kinds of Ub chains in Δdoa4 cells.

The following strains and plasmids were used: MHY501 (MATα DOA4 his3-Δ200 leu2-3,112 ura3-52 lys2-801 trp1-1) and its isogenic doa4::LEU2 derivative, MHY623 (16); RH268-1C (MATa end4 ura3 leu2 his4 bar1-1) (17); pRM1-1, a multicopy plasmid, which carries the MAL1 locus (19); YEp96, which contains a synthetic yeast Ub gene under the control of the copper-inducible CUP1 promoter (5); plasmids derived from YEp96 that encode mutant Ubs, in which distinct lysines (Lys29 [pUbK29R] [5], Lys48 [pUbK48R] [9], and Lys63 [pUbK63R] [5]), all three lysines (Lys29, -48, and -63 [pUbRRR] [5]), and all seven lysines (Lys6, -11, -27, -29, -33, -48, and -63 [pUb-no-Lys] [23]) have been replaced by arginine; and pLP2, also derived from YEp96, encoding a c-myc epitope attached to the amino terminus of Ub (9).

Yeast cells were grown in YNB (yeast nitrogen base) minimal medium with 2% maltose as previously described (13). Growth was followed by measuring the optical density at 640 nm. To trigger endocytosis, cells were harvested during early exponential growth (about 0.5 mg [dry weight] per ml), washed, and suspended in an ammonium-free medium containing 2% glucose as previously described (18). Endocytosis of the transporter was determined at different times of incubation by monitoring two steps, internalization and degradation. Internalization was determined by following the decrease in the rate of transport activity with radioactive maltose (18), and degradation was determined by immunoblotting cellular extracts with antitransporter polyclonal antibodies (18). For anti-Ub immunoblot analysis, samples were suspended in Laemmli buffer, boiled for 5 min, and resolved by using sodium dodecyl sulfate-15% polyacrylamide gels in a Tricine system (20) before transfer to an Immobilon-P membrane.

Overexpression of Ub partially restored endocytosis of the maltose transporter in Doa4p-deficient cells.

Overexpression of Ub was achieved by transforming with a multicopy plasmid bearing the Ub gene under the control of the inducible CUP1 promoter (5) and growing the cells in the presence of 0.1 mM CuSO₄. We found that overexpression of Ub had no effect on the internalization (Fig. 1A) or the degradation of the maltose transporter in wild-type cells (Fig. 1B). However, as previously reported (15), in Δdoa4 mutant cells overexpression of Ub substantially increased the rate of both processes (Fig. 1C and D). Control experiments performed in parallel (Fig. 2) demonstrated overproduction of Ub in wild-type and mutant cells as well as the availability of free Ub in the cells during the 5 h of the endocytosis experiments (Fig. 1). The inability of Ub overexpression to completely restore endocytosis is explained by the multiple abnormalities observed in Doa4p-deficient cells which are not related to the lack of free Ub (16).

FIG. 1.

Overexpression of wild-type Ub and of Ub mutants with mutations in the lysine residues partially restored endocytosis of the maltose transporter in Doa4p-deficient cells. MHY501 (wild type) (A and B) and MHY623 (Δdoa4) cells (C and D) transformed with the plasmid pRM1-1 carrying the MAL1 locus (●) or transformed with both pRM1-1 and YEp96 carrying the wild-type Ub gene (○) or with pUbK29R (X), pUbK48R (▵), pUbK63R (▿), pUbRRR (□), or pUb-no-Lys (◊) were harvested during early exponential growth, washed, and suspended in the endocytosis medium. After incubation at 30°C for the indicated times, the cells were harvested and maltose transport activity was determined (A and C). Data are mean values of two experiments. The results of the two experiments differed by less than 10%. The maltose transporter was detected by immunoblotting aliquots containing 30 μg of protein of cellular extracts obtained at the indicated times (B and D).

FIG. 2.

Overexpression of Ub in wild-type and Doa4p-deficient cells. MHY501 (wild type) and MHY623 (Δdoa4) cells transformed, when indicated, with the plasmid YEp96 containing the yeast Ub gene were harvested during exponential growth, washed, and suspended in the endocytosis medium. Ub was detected by immunoblotting aliquots containing 30 μg of protein of cellular extracts obtained at the indicated times.

Overexpression of mutant Ubs with mutations in Lys29, -48, and -63 restored endocytosis in Doa4p-deficient cells.

To determine if any of the three Ub chains detected in vivo (1, 6, 11, 21) is involved in endocytosis of the transporter, the cells were transformed with plasmids encoding Ub mutants carrying Lys→Arg mutations in Lys29, -48, and -63. Overproduction of these Ub mutants restored maltose transporter internalization (Fig. 1C) and degradation (data not shown) in Δdoa4 mutant cells as efficiently as overproduction of wild-type Ub (Fig. 1C and D). No such effect was observed in DOA4 wild-type cells (Fig. 1A). These results suggest that Ub chains linked through Lys29, -48, and -63 are not involved in the internalization of the maltose transporter. This conclusion is supported by the fact that internalization (Fig. 1C) and degradation of the transporter (Fig. 1D) were also restored by overproduction of a triple Ub mutant carrying mutations in all three lysine residues.

Overexpression of a mutant Ub lacking all of its seven lysine residues also restored endocytosis in Doa4p-deficient cells.

Ub chains linked through Lys6 and -11 have been observed in vitro (2). A role of these chains in vivo might be possible since they are able to bind to subunit 5S of the human 26S proteasome with affinities comparable to chains linked through Lys48 (2). To investigate if these Ub chains or Ub chains linked through Lys27 or -33 are involved in endocytosis of the maltose transporter, a plasmid encoding a mutant Ub lacking all of its seven lysine residues (Ub-no-Lys) was used. Overexpression of this mutant Ub had no effect in DOA4 wild-type cells (Fig. 1A and B), whereas in Δdoa4 mutant cells it restored internalization (Fig. 1C) and degradation of the transporter (Fig. 1D) as efficiently as overexpression of wild-type Ub (Fig. 1C and D). This demonstrated that endocytosis of the transporter did not require binding of any type of Ub chains and suggested that binding of a single Ub molecule to one or more lysine residues within the transporter served as an internalization signal.

In order to detect binding of Ub to the maltose transporter, a temperature-sensitive end4 mutant strain (17) was used. End4p-deficient cells show a slow down in internalization and degradation of the transporter at 35°C (18), and therefore, enhanced levels of putative ubiquitinated species of the maltose transporter could be present in these cells. The mutant strain was transformed with a multicopy plasmid encoding a c-myc-tagged Ub allele (9). As in previous work (7, 9, 15), this tagged Ub allele was used to increase the differences in size between the ubiquitinated species. As in similar experiments performed with End3p-deficient cells (15), a band 7 to 9 kDa greater than the one corresponding to the transporter appeared in the immunoblots of samples collected 1 h after the transfer of the cells to 35°C under endocytosis conditions (data not shown). The appearance of this band, which was not observed in cells growing exponentially at 24°C or in cells harvested immediately after suspension in the endocytosis medium, was consistent with the presence of a monoubiquitinated form of the transporter in end4 cells. It should be noted that this form of the transporter did not appear in Δdoa4 cells (data not shown) in which free Ub was not available.

The results herein indicate that monoubiquitination, i.e., binding of a single Ub molecule to one or more lysine residues of the maltose transporter, signals maximal internalization of this protein and support the view that monoubiquitination is the basic unit for triggering internalization of plasma membrane proteins. In some cases, i.e., the maltose transporter (this work) and the α-factor receptor (23), this basic unit is sufficient for promoting a maximal internalization rate, whereas in other cases, i.e., the uracil permease (7) and the general amino acid permease (22), maximal internalization requires additional binding of Ub-Lys63 chains. It has been speculated that this additional requirement of the permeases could be related to their high TMS number and sizes (12-TMS and about 72 and 66 kDa, respectively). Compared with the α-factor receptor (7-TMS and about 48 kDa), the permeases could, for steric reasons, require longer Ub chains to interact with the endocytic machinery or merely to provide more binding sites for a putative interacting protein (23). However, this possibility is unlikely as the maximal internalization rate of the maltose transporter (a 12-TMS protein of about 68 kDa) did not show this additional requirement.

Formation of specific Ub-Ub linkages is thought to be a property of Ub-protein ligases (E3) in association with Ub-conjugating enzymes (E2) (2, 10, 11). It has been shown that a complex formed by Ufd4p (E3) and Ubc4p and Ubc5p (E2) plays a role in the binding of Ub-Lys29 to certain proteins (11) and also that a complex formed by Ubrp1p (E3) and Rad6p (E2) plays a role in the binding of Ub-Lys48 chains to substrates of the N-end rule (4). In the case of Ub-Lys63 chains, it has been speculated that their binding to plasma membrane proteins could be a result of the action of Npi1p (E3) in association with E2 enzymes not yet identified (7, 22). This was based on the observation that internalization of the uracil and the general amino acid permeases, which is stimulated by binding of Ub-Lys63 chains, requires Npi1p. However, the fact that internalization of the maltose transporter, which is not stimulated by binding of Ub-Lys63 chains, also requires this E3 enzyme (13, 15) seems to rule out this possibility.

In conclusion, the results presented in this paper indicate that monoubiquitination is sufficient to promote a maximal internalization rate of the 12-TMS maltose transporter. This modification is similar to that required for the 7-TMS α-factor but different from that required for the 12-TMS uracil and general amino acid permeases. This indicates that monoubiquitination is not specific for the 7-TMS proteins and that binding of Ub-Lys63 chains is not a general requirement for the maximal internalization rate of the 12-TMS proteins.