Not4 is important for ubiquitin homeostasis in the cell.
Two deubiquitination enzymes, Ubp6 and Doa4, were identified in a synthetic lethal screen with the not4Δ
). We confirmed that deletion of either UBP6
was lethal in the absence of NOT4
(). Since both Ubp6 and Doa4 contribute to the maintenance of wild-type levels of free ubiquitin in the cell, we questioned the possible importance of Not4 in ubiquitin homeostasis. We observed reduced levels of free ubiquitin and the accumulation of polyubiquitinated proteins in the not4Δ
mutant (), and the synthesis of ubiquitin was increased in the not4Δ
mutant (see Fig. S2A in the supplemental material).
Fig. 1. Not4 is important for ubiquitin homeostasis in the cell. (A) Serial dilutions of the indicated mutants, containing a NOT4-URA3 plasmid and either a vector alone or a LEU2 plasmid expressing Not4, were spotted on plates with 5-fluoroorotic acid (FOA) and (more ...)
The proteasome releases polyubiquitin chains from substrates before their degradation. Hence, the phenotype of the not4Δ
mutant is consistent with a defect in proteasome function. Proteasome mutants generally display a synthetic growth phenotype when combined with a deletion of RPN4
, which is a transcription factor for proteasome genes (52
). The rpn4Δ
mutant also displayed synthetic slow growth when combined with the not4Δ
deletion (). The same was true for a deletion of the proteasome subunit Rpn10, a polyubiquitin receptor that plays a role in maintaining the structural integrity of the RP (15
) (). Synthetic mutant phenotypes were additionally observed when the not4Δ
mutant was combined with other proteasome mutants (see Fig. S2B in the supplemental material).
Taken together, these findings are compatible with a defect in proteasome function in not4Δ cells.
Several domains of Not4 contribute to ubiquitin homeostasis in the cell.
Several domains in the Not4 structure have been described (7
) (). A zinc finger domain, specific for RING E3 ligases, is present between amino acids 33 and 78 (21
). The function of Not4 as an E3 ligase was confirmed in vitro
for the human ortholog (1
) and both in vitro
and in vivo
for yeast Not4 (43
). A coiled-coil domain (amino acids 94 to 128) and an RNA recognition motif (amino acids 137 to 228) have been predicted, but the functional importance of these domains is unknown.
Fig. 2. Structure-function analysis of Not4. (A) Schematic representation of the Not4 protein, with its ring finger domain (RING), putative coiled-coil domain, and RNA recognition motif (RRM). (B) ubp6Δ not4Δ cells carrying a NOT4-URA3 plasmid (more ...)
To perform a structure-function analysis of Not4, we created plasmids expressing N-terminally Myc-tagged Not4 derivatives and mutants with N-terminal and C-terminal truncations, as well as a Not4 point mutant in which isoleucine at position 64 was replaced with alanine (Not4 I64A). This mutation disrupts the interaction of Not4 with its partner E2 enzymes, Ubc4 and Ubc5 (41
), and results in the reduced ubiquitination of its substrate, Egd2 (see Fig. S3A in the supplemental material). All of the derivatives were stable (Fig. S3C) and expressed at levels comparable to that of endogenous Not4 (data not shown) but led to different growth rates (Fig. S3B and Table S1). Interestingly, for many Not4 derivatives, we noticed an additional slower-migrating form (Fig. S3C). This form was absent in the I64A mutant and in the mutants lacking the RING domain, suggesting that it could be autoubiquitinated Not4.
We determined which domains of Not4 were essential for the viability of ubp6Δ cells (). The mutants with a growth rate comparable to that of the not4Δ mutant (from Not41–78 [in which amino acids 1 to 78 are present] to Not41–232 and derivatives with a deletion of the N terminus) had a synthetic phenotype comparable to that of the not4Δ mutant. Those with an intermediate growth rate (Not41–330, Not41–430) had an intermediate phenotype. Mutants that grew like a wild type (Not41–480, Not41–530) did not display any synthetic phenotype.
Many proteasome mutants are cycloheximide (CHX) sensitive as a result of ubiquitin depletion (20
). Hence, we tested the resistance of cells expressing the different derivatives of Not4 on medium containing CHX. not4Δ
cells and cells expressing all Not4 derivatives except the longest two (Not41–480
), displayed a CHX-sensitive phenotype (). Even Not4 I64A, which grew relatively well, was CHX sensitive. The accumulation of polyubiquitinated proteins mostly correlated with CHX sensitivity ().
Next we determined which domains of Not4 were required for its association with the Ccr4-Not complex. We immunoprecipitated Not5 and analyzed the presence of the different Not4 derivatives in the immunoprecipitates (). Wild-type Not4, all N-terminally truncated derivatives, and Not4 I64A coimmunoprecipitated well with Not5. The coimmunoprecipitation of C-terminally truncated mutants (Not41–232 and Not41–430) was less efficient. The same results were observed after coimmunoprecipitation for other Ccr4-Not subunits (Not3 and Caf40 [data not shown]). These data indicate that the C, but not the N, terminus of Not4 is important for its interaction with the Ccr4-Not complex.
The proteasome is altered in cells lacking Not4.
As mentioned above, the deletion of Not4 results in a phenotype typical for mutants of the UPS. Hence, proteasome function might be affected in the not4Δ mutant. We compared proteasome activities in total extracts from the wild type and the not4Δ mutant. We separated them on a native gel and then incubated the gel with the Suc-LLVY-AMC substrate, without or with SDS treatment to reveal latent activity (). The amount of proteasome subunits (Rpt1) was the same in total extracts from both strains. However, greater levels of proteasome activity were observed in the not4Δ mutant than in the wild type; proteasome activities in the not4Δ mutant were comparable to those of complexes that correspond to double- and single-capped proteasomes (RP2-CP and RP1-CP) ().
Fig. 3. Proteasome integrity is altered in not4Δ cells. (A) Total cellular extracts were prepared from wild-type (MY1) or not4Δ mutant (MY3595) cells collected at an OD600 of 3.0. One-hundred-microgram samples of total protein extracts (TE) were (more ...)
To follow up on this observation, we purified the 26S holoenzyme through Rpn11-ProtA from exponentially growing (OD600 of 3.0) wild-type and mutant cells (). The same amount of Rpt1 was isolated from both strains. The holoenzyme purified from the wild type was detected on a native gel as an active double-capped proteasome. Additionally, a substantial amount of free RP was purified, as seen on the Coomassie blue-stained native gel and by immunoblotting of the native gel with anti-Rpt1 antibodies. Purification at later stages of growth (OD600 of 13.0) resulted in the isolation of a reduced amount of free RP and the appearance of RP1-CP and CP species (see Fig. S1 in the supplemental material).
The spectrum of proteasome complexes purified from the not4Δ mutant was quite distinct from that from the wild type (). The holoenzyme from the not4Δ mutant migrated on a native gel as active double- and single-capped proteasomes. There was an increase of free CP, compared to that in the wild type, but no free RP could be detected, either by Coomassie blue staining or immunoblotting. These profiles of purification were similar for cells at whatever stage of growth they were collected, even as early as at an OD600 of 0.7 (data not shown). These results suggest that the interaction of CP with RP is less stable in the mutant and that either there is less free RP or it is unstable when not associated with CP.
Salt-resistant RP-CP complexes can be purified from the not4Δ mutant.
The proteasome RP-CP interaction is salt sensitive, and free RP and CP can be obtained by incubation of the holoenzyme with 0.5 M NaCl (33
). We used this procedure and cells expressing either tagged RP (Rpn11-ProtA) or CP (Pre1-ProtA) subunits to purify separately RP and CP from the wild type and the not4Δ
mutant. Purified RP and CP were analyzed first by SDS-PAGE and immunoblotting ( and B). There was the same amount of RP (Rpt1) and CP (α1–7) subunits in total extracts and purified material from both strains (). For wild-type cells, as expected, there were no α subunits in the RP purification (, left) or Rpt1 in the CP purification (, right). In contrast, for the not4Δ
mutant, we found α7 in isolated RPs and Rpt1 in isolated CPs. These findings suggest that in the not4Δ
mutant some RP-CP interactions were resistant to salt.
Fig. 4. RP and CP form salt-resistant complexes in not4Δ cells. Total cellular extracts from wild-type or not4Δ cells expressing Rpn11-ProtA (for RP purification) and Pre1-ProtA (for CP purification) were incubated with IgG beads. After the beads (more ...)
Analysis of purified CPs from the wild type or the not4Δ
mutant on a native gel did not reveal any noticeable differences in either migration or activity (see below and ). In contrast, remarkable differences were apparent when Rpn11-ProtA was purified in high salt from both strains (). For wild-type cells, several forms of RP were visible by Coomassie blue staining (, left). They were all inactive and contained Rpt1 but no α subunits (, middle and right). The faster-migrating form corresponds to free or “competent” RP (RPc
), the RP species that can reconstitute a proteasome (reference 29
and see below). The complexes of intermediate migration are “noncompetent” RPs (RPn
s) because they are incapable of CP binding (reference 29
and see below). They have a migration similar to that of double-capped proteasomes. The slowest-migrating forms at the top of the native gel were named “high” RPs (RPh
s). Besides these Rpt1-containing complexes, another complex present at low levels and migrating the fastest, likely to be the lid alone, was visible (, left, and E, bottom). According to the results of mass spectrometry analysis, wild-type RPc
contained all lid and base subunits, Ecm29, and two RP chaperones—Hsm3 and Nas6 (see Fig. S4 and Table S2 in the supplemental material). RPn
complexes contained all lid and base subunits, as well as Caf4, a protein associated with the Ccr4-Not complex (38
). In RPh
s, all lid and base subunits, Ecm29, Ubp6, Nas6, and both subunits of fatty acid synthase (Fas1 and Fas2) were identified.
In contrast, when Rpn11-ProtA was purified in high salt from the not4Δ mutant (the complex purified in this case will be referred to hereinafter as RPnot4Δ for purposes of simplification), a single proteasome complex migrating at the size of a double-capped proteasome was visible by Coomassie blue staining (, left). This complex was active and contained both the Rpt1 and α subunits. Mass spectrometry analysis confirmed the presence of all lid, base, and CP subunits but identified no Ecm29, RP chaperones, or Fas in this complex (see Fig. S4 and Table S2 in the supplemental material). We concluded that it was a salt-resistant RP2-CP form. An additional small amount of RPc that was positive for Rpt1 by immunoblot analysis at long revelation times (not shown) was detectable on the Coomassie blue-stained gel ( and Fig. S4). By mass spectrometry analysis, we identified only the Rpn2, Rpn3, Rpt1, and Rpt3 subunits in this band (see Table S2 in the supplemental material). This may be due to the lack of sufficient material to detect the remaining RP subunits.
We purified Rpn11-ProtA in high salt (0.5 M) from not4Δ cells expressing the different Not4 derivatives. Only the expression of full-length Not4, Not4 I64A, and the longest C-terminally truncated mutant, Not41–480, restored the RP-CP salt sensitivity and normal levels of free RPc ().
RPnot4Δ reconstitutes proteasomes with CP more efficiently than wild-type RP.
To address the question of why the RP and CP interaction is salt resistant in not4Δ cells, we performed in vitro proteasome reconstitution experiments using separately purified, in high salt, Rpn11-ProtA (RP) and Pre1-ProtA (CP) from wild-type or not4Δ cells. Reconstituted proteasomes were analyzed on a native gel for activity or the presence of the base (Rpt1), lid (Rpn5), or CP (α1–7) subunits by immunoblotting ().
Proteasomes reconstituted with wild-type RP and either wild-type or not4Δ CP were indistinguishable (, lanes 2 and 3). Both single- and double-capped proteasomes were created. Reconstituted complexes showed the expected electrophoretic mobilities on a native gel and had hydrolytic activities indicating effective opening of the CP gate. Hence, these reconstituted species represent bona fide 26S proteasomes. We noticed that after reconstitution, free RPc disappeared (, middle panel, compare lanes 2 and 3 to lane 1), indicating that this is indeed the RP species that was used to reconstitute proteasome. In contrast, the levels of RPh and the lid subunit were unaltered, indicating that these species were unable to reconstitute proteasomes.
There was no difference in proteasomes reconstituted using RPnot4Δ with either wild-type or not4Δ CPs (, lanes 4 and 6). In both cases, more active double-capped proteasomes were visible after reconstitution, and additional single-capped forms appeared. Interestingly, in both cases, there was a greater incorporation of CPs into single- or double-capped proteasomes than upon reconstitution with wild-type RPs (CP levels are reduced in lanes 4 and 6 compared to in lanes 2 and 3). This indicates that RPnot4Δs incorporate better into CPs during reconstitution.
RP species in the not4Δ mutant that are not incorporated into RP-CP complexes are unstable.
The total amounts of Rpt1 that copurified with Rpn11-ProtA in high salt from wild-type and not4Δ cells were the same (, left). However, after separation of the purified material on a native gel, much less Rpt1 was detectable in the not4Δ mutant purification than in that of the wild type (). Nevertheless, it was possible to reconstitute more RP2-CP and RP1-CP complexes using the proteins purified from the not4Δ mutant than from the wild type, despite the greater levels of the RPc in the wild type (). A possible explanation for these results is that the RPc purified from the not4Δ mutant was unstable and came apart during native gel electrophoresis.
To analyze this in more detail, we prepared extracts from wild-type and not4Δ cells expressing Rpn11-ProtA. Proteasomes were purified from fractions that were adjusted to increasing NaCl concentrations (). For wild-type cells at 0.1 M NaCl, we obtained active RP2-CP as well as RPc. Increasing salt concentration led to the isolation of only RP (at 0.5 M NaCl) or the lid (at 0.75 M and 1.0 M NaCl). In the purification from not4Δ cells, we obtained RP2-CP, RP1-CP, and CP species at 0.1 M NaCl treatment and a salt-resistant RP2-CP species at 0.5 M NaCl, consistently with our results presented above (, , and ). An increase of salt to 0.75 M led to a low level of lid purification. A further increase of NaCl to 1.0 M led to the absence of any detectable proteasome species on a native gel.
Fig. 5. RP species from the not4Δ mutant that are not incorporated into salt-resistant RP-CP complexes are unstable. Total extracts from wild-type or not4Δ cells expressing Rpn11-ProtA were split into four portions. The NaCl concentration in each (more ...)
SDS-PAGE and immunoblotting revealed the same amounts of lid, base, and CP subunits in the total extracts used for purification from both strains (). We isolated more of the α7 subunit at 0.1 M and 0.5 M NaCl from the not4Δ mutant than from the wild type. Less Rpt1 was isolated at high salt concentrations (0.5 to 1.0 M) from the not4Δ mutant than from the wild type. In contrast, the same amount of lid subunit Rpn12 was found in all purifications. In the wild type, Rpn12 was visible in different complexes on a native gel, namely, in the RP-CP, RPc, and lid complexes and then in even smaller complexes at high salt (, upper right). However, in the case of purification from the not4Δ mutant, Rpn12 was visible in RP2-CP complexes isolated in low salt but was hardly detectable at higher salt conditions. Taken together, these results suggest that in the not4Δ mutant, RP species, except the RP-CP complex, are not stable. Interestingly the level of Ecm29, a proteasome-associated protein, was reduced in not4Δ cells compared to the level in the wild type (see below).
Not4 associates with RP species present in holoenzyme preparations but not with purified RPs.
We identified many proteasome subunits in Ccr4-Not purifications, but we did not find Not4 in RP purifications (see Tables S2 and S3 in the supplemental material). Hence, to understand better whether Not4 interacts with the proteasome and, if so, how, we first purified holoenzyme via Rpn11-ProtA from Myc-Not4-expressing cells. Purified proteins were separated by SDS-PAGE and analyzed by immunoblotting. Myc-Not4 was clearly identified in the purification, together with the Rpt1, Rpn8, and α subunits ().
Fig. 6. Not4 associates with RP species present in holoenzyme but not with purified RP. (A) 26S holoenzyme was purified via Rpn11-ProtA from not4Δ cells (MY7820) expressing Myc-Not4. Total extracts (TE) (lane 1) and purified proteins (purif) (lane 2) (more ...)
Next we tried an in vitro binding experiment. We fixed holoenzyme from wild-type cells via Rpn11-ProtA on IgG beads and incubated these beads with separately purified Myc-Not4 (). The IgG beads were washed, and holoenzyme was eluted by TEV cleavage. The presence of Myc-Not4 in the eluate in addition to proteasome subunits was demonstrated by immunoblotting. We also performed the reverse experiment (). We fixed Myc-Not4-ProtA from total extracts on IgG beads and incubated the beads with purified holoenzyme or RP. After being washed, Myc-Not4 was eluted by TEV cleavage. The presence of the Myc-Not4, Rpn8, or α subunit in the eluates was evaluated by immunoblotting. Rpn8 was found in the eluate from the sample incubated with holoenzyme but not RP. In contrast, no α subunits coeluted with Not4 in either case.
Taken together, these results demonstrate that Not4 can associate, directly or indirectly, with RP species that contain at least the Rpn11 and Rpn8 but not the α subunits and that are present in the holoenzyme but not in RP preparations. Furthermore, Not4 cannot associate with purified RP. This explains why we did not find Not4 in the mass spectrometry analyses discussed above, all performed with RP purifications.
Not4 genetically, physically, and functionally interacts with Ecm29.
Our result suggests that, on one hand, Not4 interacts with a proteasome complex or intermediate and that, on the other hand, Not4 is necessary for proteasome integrity. Hence, Not4 might assist proteasome assembly or stability, perhaps through interaction with proteasome assembly or stabilizing factors. One such factor, Ecm29, was identified in affinity-purified Ccr4-Not complexes by mass spectrometry (reference 3
and see Table S3 in the supplemental material). In not4Δ
cells, Ecm29 levels were decreased, and several forms of Ecm29 accumulated (). Less Ecm29 copurified with Rpn11-ProtA from the not4Δ
mutant than from wild-type cells, and Ecm29 was identified in RP species isolated from the wild type but not from the not4Δ
mutant ( and see Figure S4 and Table S2 in the supplemental material).
Ecm29 has been described to enhance the stability of the proteasome (34
). Deletion of Ecm29 results in the same phenotype that we observed for the not4Δ
mutant: it shifts the equilibrium between RP associated with CPs and free components toward RP-CP complexes and leads to accumulation of polyubiquitinated proteins (37
). We hence investigated the relationship between Ecm29 and Not4 in more detail. The mutant with the deletion of NOT4
displayed a synthetic growth phenotype at 37°C or on CHX when it was combined with the deletion of ECM29
. However, deletion of ECM29
suppressed the slow growth of the not4Δ
mutant at 30°C (). A reduced level of Ecm29 in the not4Δ
mutant was confirmed for tagged Ecm29 expressed in wild-type and not4Δ
cells (). N-terminally truncated Not4 derivatives or the I64A mutant could complement this reduction, whereas C-terminal derivatives were not effective (). S1 analysis revealed the same amount of ECM29
mRNA in wild-type and mutant cells (data not shown). Hence, reduction of Ecm29 levels in the not4Δ
mutant occurs posttranscriptionally. Therefore, we investigated the stability of Ecm29 in wild-type and not4Δ
mutant cells. In both the wild type and the not4Δ
mutant, Ecm29 was rapidly turned over, though in the wild type, a small fraction of Ecm29 seemed relatively stable ().
Fig. 7. Not4 genetically, physically, and functionally interacts with Ecm29. (A) The indicated strains were spotted on YPD plates with or without CHX and left to grow at 30 or 37°C. (B) ecm29Δ cells expressing HA-Ecm29 from an episome (the wild (more ...)
Since we previously noted slower-migrating forms of Ecm29 in the not4Δ mutant (), we checked whether Ecm29 might be ubiquitinated. For this, we expressed His ubiquitin in the wild type or in not4Δ cells expressing HA-Ecm29 and purified ubiquitinated proteins on a nickel column. After being immunoblotted, ubiquitinated forms of Ecm29 were detectable in the wild type and, more importantly, in the mutant ().
These results suggest that Not4 has an impact on ubiquitination and steady-state levels of Ecm29. Since we previously detected Ecm29 in affinity-purified Ccr4-Not complexes by mass spectrometry (see Table S3 in the supplemental material), we looked further for an interaction between Ecm29 and Not4. We immunoprecipitated Myc-Not4 from total extracts with anti-Myc antibodies and could identify Ecm29 in the immunoprecipitate (, left). Then we did the reverse experiment and immunoprecipitated Ecm29 from the total extracts of cells expressing Myc-Not4. We clearly identified Myc-Not4 in the immunoprecipitate (, right). These results confirm that Ecm29 and Not4 can interact, either directly or indirectly.
Some of Not4's impact on the proteasome is through Ecm29.
To answer the question of whether some or all of Not4's impact on the proteasome might be related to its interaction with Ecm29, we purified RPs from the wild type, 2 different not4Δ strains, the ecm29Δ mutant, the ecm29Δ not4Δ double mutant, and, finally, the not4Δ ecm29Δ double mutant transformed either with a multicopy plasmid overexpressing Ecm29 or with a centromeric plasmid expressing wild-type levels of HA-Ecm29 (). The same amounts of the Rpt1, Rpn8, and α subunits were found in the total extracts before purification (, left). As expected, less Ecm29 was found in the total extracts from not4Δ cells.
Fig. 8. Some of Not4's impact on the proteasome is through Ecm29. (A) Rpn11-ProtA was purified at 0.5 M NaCl from cells of the wild type (MY5559) (lane 1), 2 different not4Δ mutants (MY7367 [lane 2] and MY7820 [lane 3]), the ecm29Δ mutant (MY7827) (more ...)
First, we analyzed the proteasome activities in total extracts prepared from the strains indicated in by native PAGE (, top). Holoenzyme activity was higher in the not4Δ mutant than in the wild type, as expected (, top, compare lanes 2 and 3 to lane 1). In contrast, activity in the ecm29Δ mutant was indistinguishable from that of the wild type (, top, lanes 1 and 4). Activity in the double mutant was higher than in the not4Δ single mutant, and an additional active complex (marked by an asterisk), migrating more slowly than the RP2-CP complex, was detected (, top, lane 5). Surprisingly, expression of Ecm29 from either the multicopy or the single-copy plasmid in the not4Δ ecm29Δ mutant reduced the holoenzyme activity to levels lower than those in the not4Δ single mutant (, top, compare lanes 6 and 7 to lanes 2 and 3).
Immunoblot analysis of the native gel with antibodies to follow Rpn11-ProtA revealed that all active complexes (RP2-CP and RP1-CP), as well as RPn, RPc, RPh, and free lid, were present in different amounts in the total extracts (, top right). The free lid was absent in the not4Δ mutant but present in the ecm29Δ mutant and in the double mutant. Hence, the deletion of Ecm29 suppressed the instability of CP-less RP complexes in the not4Δ mutant. No free lid was detectable in the double mutant overexpressing Ecm29, suggesting a complementation of the ECM29 deletion. That did not occur when Ecm29 was expressed only from the centromeric plasmid (, top right, compare lanes 6 and 7 to lane 5). In this context, it is interesting to note that purification of the proteasome via Rpn11-ProtA from the ecm29Δ mutant led to purification of stable lid even under the highest salt conditions (see Fig. S5 in the supplemental material).
Second, we purified Rpn11-ProtA in high salt (0.5 M) from all strains and analyzed the purified proteasomes on a native gel for activity (, middle) and by SDS-PAGE (, right). The same amounts of the base (Rpt1) and lid (Rpn8) subunits were purified for all samples (, right). As expected, there was no activity detected for the purification from wild-type cells (, middle, lane 1), and consistently, no CP subunits copurified with Rpn11 (, right, lane 1). For the single mutants, active proteasome forms were observed, mostly RP2-CP for the not4Δ mutant (, middle, lanes 2 and 3) and RP1-CP for the ecm29Δ mutant (, middle, lane 4). The presence of CP subunits in both single mutants was confirmed by immunoblotting (, right, lanes 2 to 4). These results indicate that in the ecm29Δ mutant, as in the not4Δ mutant, salt-resistant active RP-CP complexes were present. In the not4Δ ecm29Δ double mutant, the amount of active RP-CP was increased and the presence of CP subunits was more important, revealing an aggravated phenotype ( and B, lane 5). Expression of Ecm29 complemented the increased amount of salt-resistant RP-CP from the not4Δ ecm29Δ mutant compared to that from the single not4Δ mutant (, middle, and B right, compare lanes 6 and 7 to lanes 2, 3, and 5). Curiously, many faster-migrating RP subcomplexes were detectable by Coomassie blue staining in the purification from the ecm29Δ mutant, which were absent in the purification from the double mutant (, bottom, compare lane 4 to lane 5).
Our results suggest that the impact of Not4 on Ecm29 contributes to the wild type's RP-CP salt sensitivity. Indeed, (i) the salt resistance of the RP-CP interaction was observed in both the not4Δ mutant and the ecm29Δ mutant, (ii) the levels of Ecm29 were lower in the not4Δ mutant than in the wild type, and (iii) the overexpression of Ecm29 in the ecm29Δ not4Δ double mutant could entirely suppress this RP-CP salt resistance. On the other hand, greater instability of CP-less RP complexes specific for the not4Δ mutant was not observed in the ecm29Δ mutant (, right, lane 4) and in fact were partially suppressed by the deletion of ECM29 in the not4Δ mutant (, right, lane 5). Conversely, the accumulation of small RP complexes in the ecm29Δ mutant was suppressed by the not4Δ mutant (, bottom, lanes 4 and 5). Taken together, these results suggest that the functions of Ecm29 and Not4 are partially interconnected and that the phenotype of the not4Δ mutant may in part be due to an inappropriate function of Ecm29 in this mutant.
Ecm29 association with proteasome is deficient in the not4Δ mutant.
To address how Not4 may affect Ecm29 function, we purified Ecm29-ProtA from the wild type and the not4Δ
mutant. The ProtA tag on Ecm29 stabilized the protein such that its levels in the wild type and in the not4Δ
mutant were relatively comparable. We analyzed the Ecm29 purification by mass spectrometry (see Table S4 in the supplemental material) and immunoblotting (). This analysis led to several interesting observations. A number of proteasome subunits, but not all, could be copurified with Ecm29 from the wild type. These are the subunits of the lid, Rpn1 and -2 and Rpt1 to -6 of the base, and the α1 to α4, α6, and α7 as well as the β1, β5, and β6 subunits of the CP. Additionally, Blm10 copurified. This is compatible with the fact that Ecm29 interacts with both RPs and the α ring of CP (34
). In the not4Δ
mutant, neither Rpt1, Rpt3, β1, α3, nor Blm10 was detectable in the purification of Ecm29, and the levels of some other subunits (Rpt2, -4, and -6 and α2, α7, and β5) appeared to be reduced. These results of mass spectrometry were confirmed by immunoblotting (). Indeed, the same amounts of the Rpt1, Rpt2, Rpn8, and α subunits were detected in total extracts before purification, but slightly less Rpt2 and many fewer α subunits were purified from the not4Δ
mutant than from the wild type. No Rpt1 at all was purified from the mutant. These results suggest that Ecm29 might make primary contacts with the lid, whereas Not4 contributes to Ecm29 association with the other proteasome components.
Many other proteins were identified in the Ecm29 purifications. These are proteins related to vesicle sorting and transport, cytoskeleton and motors proteins, proteins involved in lipid biogenesis, chaperones, and ribosomal proteins. This is consistent with a recent study that established a link between Ecm29 and both endosomal compartments and molecular motors (18
) and suggested that Ecm29 has cellular partners other than the proteasome.