Subunit Communications Crucial for the Functional Integrity of the Yeast RNA Polymerase II Elongator ( (cid:1) -Toxin Target (TOT)) Complex*

In response to the Kluyveromyces lactis zymocin, the (cid:1) -toxin target (TOT) function of the Saccharomyces cerevisiae RNA polymerase II (pol II) Elongator complex prevents sensitive strains from cell cycle progression. Studying Elongator subunit communications, Tot1p (Elp1p), the yeast homologue of human IKK-associated protein, was found to be essentially involved in main-taining the structural integrity of Elongator. Thus, the ability of Tot2p (Elp2p) to interact with the HAT subunit Tot3p (Elp3p) of Elongator and with subunit Tot5p (Elp5p) is dependent on Tot1p (Elp1p). Also, the association of core-Elongator (Tot1–3p/Elp1–3p) with HAP (Elp4–6p/Tot5–7p), the second three-subunit subcom-plex of Elongator, was found to be sensitive to loss of TOT1 ( ELP1 ) gene function. Structural integrity of the HAP complex itself requires the ELP4 / TOT7 , ELP5 / TOT5 , and ELP6 / TOT6 genes, and elp6 (cid:2) / tot6 (cid:2) as well as elp4 (cid:2) / tot7 (cid:2) cells can no longer promote interaction

In response to the Kluyveromyces lactis zymocin, the ␥-toxin target (TOT) function of the Saccharomyces cerevisiae RNA polymerase II (pol II) Elongator complex prevents sensitive strains from cell cycle progression. Studying Elongator subunit communications, Tot1p (Elp1p), the yeast homologue of human IKK-associated protein, was found to be essentially involved in maintaining the structural integrity of Elongator. Thus, the ability of Tot2p (Elp2p) to interact with the HAT subunit Tot3p (Elp3p) of Elongator and with subunit Tot5p (Elp5p) is dependent on Tot1p (Elp1p). Also, the association of core-Elongator (Tot1-3p/Elp1-3p) with HAP (Elp4 -6p/Tot5-7p), the second three-subunit subcomplex of Elongator, was found to be sensitive to loss of TOT1 (ELP1) gene function. Structural integrity of the HAP complex itself requires the ELP4/TOT7, ELP5/ TOT5, and ELP6/TOT6 genes, and elp6⌬/tot6⌬ as well as elp4⌬/tot7⌬ cells can no longer promote interaction between Tot5p (Elp5p) and Tot2p (Elp2p). The association between Elongator and Tot4p (Kti12p), a factor that may modulate the TOT activity of Elongator, requires Tot1-3p (Elp1-3p) and Tot5p (Elp5p), indicating that this contact requires a preassembled holo-Elongator complex. Tot4p also binds pol II hyperphosphorylated at its C-terminal domain Ser 5 raising the possibility that Tot4p bridges the contact between Elongator and pol II.
Microbial rivalry between Kluyveromyces lactis killer strains and sensitive Saccharomyces cerevisiae cells relies on secretion of zymocin, a heterotrimeric (␣␤␥) protein toxin complex that acts as a cell cycle blocker in G 1 (1)(2)(3). Zymocin docking involves interaction of its ␣-subunit, an exo-chitinase, with S. cerevisiae cell wall chitin (4,5), and anti-zymotic activity resides within the ␥-subunit, also termed ␥-toxin (6,7). In an effort to identify the intracellular ␥-toxin target (TOT) 1 process, seven TOT genes were found to abrogate toxicity when mutated (8,9). TOT1-3 and TOT5-7 are identical with ELP1-6 coding for Elongator, an RNA polymerase II (pol II)-associated histone acetyltransferase (HAT) complex (8 -15). In addition, loss of KTI11, KTI13, and SIT4 results in tot phenotypes characteristic for TOT mutants, implying that these genes may also be linked to Elongator function (16 -18). Tot4p (Kti12p) can be found promoter-associated, and it contacts Elongator and pol II (8,19,20). Since both removal and overproduction of Tot4p induce zymocin resistance, Tot4p is likely to influence Elongator by modulating its TOT activity (16,20). Holo-Elongator contains the core complex, Elp1-3p (Tot1-3p), and HAP, a second heterotrimer composed of Elp4 -6p (HapI-3p/Tot5-7p) (8,9,(13)(14)(15). Elp1p is the largest subunit within core-Elongator and homologous to human IKK-associated protein, an IB kinase scaffold protein and a member of a five-subunit protein complex (21,22). Elp2p is a WD40 protein homologous to murine STAT3-interacting protein StIP1 (12,23,24), and Elp3p is the HAT subunit (11,25) whose activity requires the HAP complex (26). Consistent with a role in transcription, Elongator facilitates pol II activity through chromatin and stably associates with hyperphosphorylated pol II (II0) (10,27). Its HAT activity is essential for Elongator to function as TOT and HATminus scenarios yield zymocin resistance (8,14). Together with the finding that the TOT function can be dissociated from Elongator by mutagenesis of its HAT gene without affecting other Elongator properties (9), the HAT function of Elongator plays a key role in mediating zymocicity. As judged from the observations that (i) pol II-driven transcription is down-regulated in zymocin-treated cells (8,9), (ii) pol II underassembly and general pol II defects elicit zymocin-hypersensitivity (9, 28), (iii) interfering with pol II C-terminal domain (CTD) modification alters the response of a cell to zymocin (9), and (iv) the phospho-states of pol II are imbalanced in zymocin-treated cells (29), zymocin may work by hijacking the TOT function of Elongator to convert it into a global pol II inhibitor. Genetic analyses have shown that deletion of any one of the ELP/TOT genes phenocopies the full range of elp/tot phenotypes induced by elp3 point mutations that drastically reduce the HAT activity of Elongator (8,12,25). Thus, it has been speculated that the functional integrity of Elongator is compromised in these deletants, leading to non-productive Elongator HAT scenarios (12,25).
The yeast Elongator subunit Tot1p (Elp1p) has been shown to be related to IKAP, a scaffold protein that associates with IB kinase and that is the largest subunit of the human Elongator complex (21,24). To assess its role in mediating the structural integrity of Elongator, we required data concerning individual Elongator subunit interactions in the presence and absence of TOT1 (ELP1). Thus, the ability of Elongator subunit Tot2p (Elp2p) to co-ip the HAT subunit Tot3p (Elp3p) of Elongator was fully dependent on the presence of a functional TOT1 (ELP1) gene ( Fig. 2A). Similarly, Elongator-associated factor Tot4p required Tot1p (Elp1p) to interact with the HAP component Tot5p (Elp5p) (Fig. 2B). Together with our previous findings that the interactions between Tot1p (Elp1p) and Tot3p (Elp3p) as well as Tot5p (Elp5p) and Tot3p (Elp3p) are insen- In A, C-terminal deletions are shown on the horizontal axis, and each plasmid-encoded allele is indicated (e.g. pTOT2⌬1). Numbers denote amino acid residues being deleted (⌬). Black boxes indicate the presence of eight (numbered) WD40 motifs. B, killer eclipse assay of wild-type TOT2 (pTOT2), empty vector (YCplac33), and the episomal TOT2 deletions (pTOT2⌬1-2) in a zymocin-resistant tot2⌬ background. Inability of the TOT2 truncations to complement zymocin resistance is indicated by a Ϫ symbol, whereas complementation by wild-type pTOT2 is denoted with a ϩ symbol. As shown in C, co-ip of Tot3p and Tot4p requires TOT2 function. Protein extracts obtained from the indicated strains were subjected to co-ip using the anti-c-Myc antibody 9E10. The immunoprecipitates were probed with the anti-HA antibody to detect Tot4p and with the anti-c-Myc antibody to detect Tot3p. Their positions are indicated by arrows. D, synthesis of the C-terminal Tot2p truncations. Protein extracts or immune precipitate (ip) of the indicated strains were subjected to Western analysis using the anti-HA antibody. The positions of wild-type Tot2p and the two truncations are shown by arrows. MSM denotes molecular size markers (see the legend for Fig. 1). sitive to Tot2p (Elp2p) (19), our observations suggest that Tot1p (Elp1p) may play a role as a scaffold to which Tot3p (Elp3p) and Tot2p (Elp2p) assemble onto to form core-Elongator. Moreover, Tot1p (Elp1p) can be considered to mediate the contact between core-Elongator and the HAP complex since this inter-subcomplex communication is lost in the absence of Tot1p (Elp1p). As for the capability of Tot3p (Elp3p) to co-ip Tot4p, both Tot1p (Elp1p) and Tot2p (Elp2p) are required (Fig. 3). Together with the surprising observation that this also holds true for a functional TOT5 (ELP5) gene (Fig. 3), these data strongly indicate that the association of Tot4p with Elongator requires a completely assembled six-subunit holo-complex.
Tot2p (Elp2p) contains eight WD40 domains and was hypothesized to serve the structural integrity of Elongator (12). Intriguingly, however, TOT2 (ELP2) deletion does not compromise Tot1-Tot3 (Elp1-Elp3) and Tot3-Tot5 (Elp3-Elp5) proteinprotein interaction, indicating that communication between both of these core-Elongator subunits and a representative HAP component is insensitive to Tot2p (15,19). As for the established role TOT2 (ELP2) plays in the association of Tot4p with Elongator (19), we tested episomal TOT2 (ELP2) deletion alleles coding for C-terminally truncated variants (Fig. 4A, pTOT2⌬1-2) for their abilities to mediate Tot4-Tot3 proteinprotein interaction by co-ip. As illustrated in Fig. 4C, none of these truncations lacking up to one WD40 domain was able to restore Tot4-Tot3 protein-protein interaction in a tot2⌬ (elp2⌬) background, whereas full-length episomal TOT2 (ELP2) mediated association of Tot4p with Tot3p (Elp3p) in a way indistinguishable from chromosomally encoded wild-type TOT2 (ELP2) (Fig. 4C). Anti-HA Western analysis of protein extracts obtained from yeast strains expressing the truncations as HAtagged proteins revealed that the Tot2p (Elp2p) variants were synthesized (Fig. 4D), suggesting that their incapability in mediating co-ip between Tot4p and Tot3p (Elp3p) in tot2⌬ (elp2⌬) cells was not due to protein instability or lack of protein synthesis. Thus, the function of Tot2p in mediating association of Tot4p with Tot3p (and core-Elongator) resides in its extreme C terminus. Consistently, the appropriate truncation alleles conferred zymocin resistance in killer eclipse assays, suggesting that this region plays a role in TOT function and K. lactis zymocicity, too (Fig. 4B). Taken together, these data indicate that Tot2p (Elp2p) mediates at least in part the contact between Tot4p and Elongator. However, for inter-complex communication and interaction between Tot1p (Elp1p) and Tot3p (Elp3p), it appears to be dispensable (15,19). As for the zymocin mode of action, which is abrogated in the absence of the C terminus of Tot2p, our findings indicate that zymocin requires Tot4p to be associated with Elongator, providing further evidence that Tot4p influences the TOT function of Elongator (19,20).
Consistent with copurification of Elongator and pol II0 (10, 24), c-Myc-tagged Tot2p (Elp2p) and, to a lesser extent, c-Myctagged Tot5p (Elp5p), were able to associate with pol II form II0 using co-ip (Fig. 5A). Similarly, c-Myc-tagged Tot4p interacted with pol II form II0 (Fig. 5A). Remarkably, as judged from utilizing different anti-pol II antibodies, the interaction between Tot4p and pol II was solely restricted to form II0 hyperphosphorylated at Ser 5 within the CTD repeat (Fig. 5B). In contrast, neither pol II form IIA (hypophosphorylated on its CTD) nor pol II form II0 (hyperphosphorylated at the Ser 2 of CTD) were co-precipitable with HA-tagged Tot4p (Fig. 5B). Also, lack of interaction could not be overridden by subjecting yeast cells to cross-linking prior to protein extraction and co-ip to amplify weaker interactions (Fig. 5B). Thus, Tot4p interacts with pol II form II0 hyperphosphorylated at the Ser 5 of CTD, a FIG. 5. Elongator and Tot4p associate with pol II hyperphosphorylated within the CTD repeat. A, co-ip between pol II's largest subunit Rpb1p and Tot2p, Tot4p, and Tot5p. Protein extracts obtained from the indicated strains were subjected to co-ip using the anti-c-Myc antibody 9E10. The immunoprecipitates were subjected to 10% SDS-PAGE and immunoprobed with 9E10 to detect Tot2p, Tot4p, and Tot5p and with the anti-CTD-S5-P antibody H14 to detect pol II form II0. The positions of Tot2p, Tot4p, and Tot5p and pol II's largest subunit Rpb1p are indicated by arrows. MSM denotes molecular size markers. wt, wild type. B, co-ip between Tot4p and pol II form II0. Extracts from TOT4-(HA) 6 -expressing cells grown in the absence (Ϫ) or presence (ϩ) of formaldehyde to induce cross-linking (cl) were subjected to co-ip using the anti-HA antibody 3F10. The immunoprecipitates were subjected to 6% SDS-PAGE and immunoprobed with anti-CTD (8WG16), anti-CTD-S5-P (H14), anti-CTD-P (Bentley laboratory), and anti-CTD-S2-P (H5) antibodies to detect both hyperphosphorylated and hypophosphorylated pol II forms. ip, immune precipitate.
FIG. 6. Working model on how Tot4p (Kti12p) may regulate the TOT function of Elongator. Holo-Elongator formation requires the association of core-Elongator (Elp1-3p/Tot1-3p) with the HAP complex (Elp4 -6p/Tot5-7p) (1) in a manner largely dependent on the putative scaffold protein, Elp1p/Tot1p. Individual subunit interactions within the HAP complex cannot be derived from the co-ip studies presented, although individual deletions (tot5⌬, tot6⌬, and tot7⌬ cells) remove association with core-Elongator, and two-hybrid interaction has been shown between Elp5p/Tot5p and Elp6p/Tot6p (45). Fully assembled holo-Elongator enables interaction with Tot4p/Kti12p (2) in a fashion that requires the functional integrity of the C terminus of Elp2p/Tot2p. This contact may be necessary for piggy-backing Elongator onto elongation-competent pol II form II0 (3). During this transition, Tot4p/ Kti12p becomes replaced by Elongator and can be recycled (4). For simplicity, Tot4p/Kti12p is shown monomeric. The possibility, however, that Tot4p/Kti12p contacts holo-Elongator and pol II form II0 as part of a protein complex cannot be eliminated for the time being. modification that is transcriptionally characteristic for a postinitiation event and for pol II0 engaged in promoter clearance and/or early transcript elongation (37)(38)(39). DISCUSSION Taken together, our findings can be hypothetically incorporated in a working model (Fig. 6) in which the assembly of holo-Elongator requires the association of preformed core-Elongator with the HAP complex. This inter-complex communication largely relies on Tot1p (Elp1p), which may serve as a scaffold protein. Fully assembled holo-Elongator is capable to contact Tot4p (Fig. 6), an Elongator-associated factor that is not a structural subunit but rather transiently contacts Elongator, presumably to promote its interaction with elongationcompetent pol II0. Consistent with this, Tot4p associates both with core-Elongator subunits and HAP components (8) as well as with pol II form II0 hyperphosphorylated at the Ser 5 of CTD. Since this modification occurs after preinitiation complex formation (37)(38)(39) and since Tot4p is able to occupy the promoter rather than the coding sequence of the ADH1 gene in chromatin immune precipitations (19), Tot4p may be recruited to and communicate with pol II form II0 engaged in promoter escape and/or early transcript elongation (39). One interpretation of such a scenario includes the conclusion that Tot4p mediates the association of Elongator with pol II to yield HAT-productive holo-enzymes ready for efficient transcript elongation. Consistent with such a role for Tot4p as a loading factor, its removal yields tot/elp phenotypes indistinguishable from TOT/Elongator mutants (8,9). Moreover, multicopy TOT4 induces zymocin resistance and intermediate tot phenotypes, indicating that excess Tot4p levels affect Elongator function, too (8,16,20). According to the identification of a putative ATP/GTP binding P-loop motif in the N terminus of Tot4p that is necessary for protein function, Tot4p may be a G-protein (20). Together with the fact that a deletion of KTI13 (ATS1) encoding a putative GTP exchange factor (40) results in zymocin resistance and tot phenotype expression (18) that can be suppressed by multicopy TOT4 (16), excess Tot4p levels may bypass the requirement for this GTP exchange factor under normal conditions (20). Thus, the role of Tot4p in Elongator function is likely to be regulatory. Provided that Tot4p was a key component of a distinct complex that associates both with Elongator and pol II, loss of TOT4 function would impair complex formation, suppress Elongator loading, and ultimately result in a HAT-minus scenario and zymocin resistance. Alternatively, excess Tot4p levels due to multicopy TOT4 might perturb normal complex stoichiometry by separately titrating the other subunits and thereby reducing the amount of completely assembled Tot4p complex able to interact with Elongator and pol II. In favor of this model, tandem affinity purification-tagged Tot4p has been recently reported to be a constituent of a protein complex that is distinct from Elongator but whose identity awaits further analysis (15). If Tot4p was able to regulate TOT by loading Elongator onto pol II (Fig. 6), one would assume that the underlying contact between Tot4p and Elongator involved interaction with a preassembled complex rather than individual Elongator subunits. Consistently, TOT4 deletion does not affect the structural integrity of Elongator (20), and the capability of Tot4p to interact with Elongator requires TOT1 (ELP1), TOT2 (ELP2), TOT3 (ELP3), and TOT5 (ELP5) gene functions (Figs. 1-4) (Ref. 19). Thus, disruption of TOT1 (ELP1) abolishes the ability to co-ip Tot4p by Tot2p (Elp2p), tot3⌬ (elp3⌬) cells lack Tot2-Tot4 protein-protein interaction, deletion of TOT2 (ELP2) no longer admits Tot4p to associate with Tot3p (Elp3p), and tot5⌬ (elp5⌬) cells do not allow Tot4p to interact with Tot3p (Elp3p). These data suggest that association of Tot4p with Elongator requires not only more than one subunit but a preassembled holo-Elon-gator complex. It will be interesting to determine whether the inability of mutant Elongator to interact with Tot4p prevented it from being associated with pol II0. If this holds true, Tot4p may as well be envisaged to be instrumental in bridging Elongator and pol II, a scenario necessary for Tot4p to act as a putative Elongator loading factor. In contrast to a previous report (13), which considers the HAP complex to be preferentially associated with pol II-free rather than pol II-bound core-Elongator, the HAP component Tot5p (Elp5p) was found to co-ip pol II (II0), albeit to a lesser extent than the core-Elongator subunit Tot2p (Elp2p). Also, our model (Fig. 6) predicts that a functionally productive Elongator HAT complex requires the association of the six-subunit holo-Elongator with elongating pol II. The recent demonstration that all three HAP subunits Elp4 -6p (Tot5-7p) are indeed required for Elongator HAT activity in vivo supports the model and indicates an essential role for Elp4 -6p in either structural organization or substrate recognition of holo-Elongator (26). Intriguingly, both Elp4p and Elp6p were recently suggested to be ATPase homologues, implying putative roles, for instance in ATP-dependent chromatin remodeling (41). Consistent with results showing that the majority of all human core-Elongator subunits are located in the cytoplasm (24,27), yeast core-Elongator has also been reported to localize primarily to the cytoplasm (19,42). Together with the finding that Elongator subunits did not appear to occupy promoters or coding regions of genes in chromatin immune precipitations, whereas other transcriptionally relevant elongation factors clearly did, the likelihood that Elongator is at all recruited to transcribed genes as part of the pol II elongation apparatus has been questioned (42,43). Nevertheless, these data could not eliminate the possibility that Elongator may enter the nucleus to become engaged in transcription. The finding that StIP1, the murine homologue of yeast Elp2p (Tot2p) and interactor of STAT3, is located in the cytoplasm and becomes translocated into the nucleus upon IL-6 treatment (23) supports this notion and indicates that murine Elongator is subject to compartmentalization. Our data showing that the TOT function of Elongator depends on an nuclear localization signal present in Elp1p (Tot1p) provide evidence that yeast Elongator function may require nuclear localization signal-mediated nuclear import (19) and may be regulated by subcellular compartmentalization in a manner similar to other transcription factors (44).