Pol5 is required for the recycling of SSU-processome components

Several known AFs engaged in the “middle” steps of LSU maturation and required for C2 cleavage are supposed to have an additional effect in cleavages at A0, A1, and A2 (Gregory et al., 2019; Hong et al., 1997; Saveanu et al., 2003; Talkish et al., 2012; Zanchin et al., 1997). Regarding Pol5, our data provide evidence for its role as “B factor” and for an

additional role in SSU biogenesis. Besides the contact sites of Pol5 within the ITS2 and the domain III of 25S rRNA, an additional crosslinking region within the 5’ETS was observed (see Figure 34A). Mapping these crosslinks in available SSU-processome structures shows that the contact sites of Pol5 are surrounded by the tUTP components (see Figure 34B) (Braun et al., 2020; Chaker-Margot et al., 2017). This observation is also supported by the description of the Pol5-containing UTP-A complex done by the Greenblatt-group, which mostly co-purified 5’ETS fragments (Krogan et al., 2004). A recent publication, comparing the co-purification of Utp8 and Pol5 on short 5’ETS transcripts showed much higher levels of 5’ETS co-purified with Pol5 than with Utp8 (Gallagher, 2019). Nevertheless, we did not look for such interaction in our analyses, since our probes for the 5’ETS were located at the 3’ end.

In the absence of Pol5, our data showed accumulation of purified tUtps with the 5’ETS fragment and several pre-rRNAs, either containing or not the 5’ETS (see Figure 44). This result is consistent with the accumulation of 5’ETS in the absence of Pol5 observed in the data of the Greenblatt-group (Krogan et al., 2004). Altogether, these findings indicate a role of Pol5 in the release of the tUtps from the 5’ETS during disassembly of the 5’ETS particle. Trapping the tUtps and other SSU factors within the 5’ETS particle and the SSU processome would lead to the strong reduction of available SSU AFs, which would induce a defect in the assembly of the SSU. In agreement, our data show an increase of 35S transcripts not associated with tUtps (see Figure 44). Interestingly, the association of tUtps with 5’ETS fragments is not detected when pre-rRNA is synthesized by RNA Pol II (Gallagher, 2019). We think it can be due to a slower production of pre-rRNAs, which are more efficiently processed and, therefore, the transcribed spacers will be more efficiently degraded.

The AIM domain in Pol5 seems to be required for the recycling of the tUtps. Since the exosome activity is involved in the degradation of the 5’ETS (Delan-Forino et al., 2017) and the AIM domain participates in the recruitment of the exosome (Thoms et al., 2015), it is very likely that release of tUtps from 5’ETS requires the action of the nuclear exosome. Consistent with this hypothesis, the truncated NTD, still containing the intact AIM domain, and the complete CTD still support recycling of tUtps although not cell growth (see Figure 52). In contrast, the only presence of the CTD is not enough to support recycling (see Figure 53). The higher stability of the CTD indicates formation of a stable folding, which might be required together with the AIM domain to induce the release of the AFs from the 5’ETS. Therefore, it is now important to determine the function of these Pol5 mutants in the assembly of the LSU. In summary, our model would support the existence of a transient Pol5-containing UTP-A complex associated to the 5’ETS. In this model, formation of UTP-A might be required for Pol5 to recruit the exosome in proximity of the 5’ETS, and most possibly the efficient degradation of the 5’ETS by the exosome recycles the tUtps. Therefore, it will be interesting to investigate if Pol5 is involved in the recruitment of exosome to 5’ETS-containing particles. Moreover, the same approach might be also relevant to look if Pol5 participates in the recruitment of the exosome to degrade the ITS2 region. As an alternative, other proteins as Utp18 and Nop53 might target the exosome to pre-ribosomal particles (Thoms et al., 2015).

tUtps, the primary binders of the pre-rRNA transcripts, stabilize the nascent transcripts by blocking the accessibility of exonucleases (Gallagher et al., 2004; Jakob et al., 2012; Pérez-Fernández et al., 2007; Wery et al., 2009). The depletion of Pol5 led to an accumulation of the 35S pre-rRNA, but the relative 35S levels associated with tUtps were lower (see Figure 44). It is likely that several 35S transcripts are stabilized even if they are not associated with tUtps. This striking result contrasts with the protective role suggested for tUtps (Wery et al., 2009) and it might indicate toward a functional role during RNA Pol I driven transcription. Interestingly, upon prolonged depletion of Pol5, we observed a decrease of RNA Pol I occupancy at the rDNA promoter region (see Figure 46). This effect might be caused by the absence of Pol5 or the reduced pool of tUtps, which are required for efficient rDNA transcription by RNA Pol I (Gallagher, 2019; Gallagher et al., 2004; Schmid, Master thesis, 2019).

Moreover, the absence of Pol5 at the 5’ETS might also influence the cleavage at site D, since a processing intermediate (22S’) could be detected reaching from 5’ETS to site D or an alternative site close to D (see Figure 44). Usually, cleavage at site D is performed by Nob1 in the cytoplasm as one of the last steps of SSU maturation (Fatica et al., 2003, 2004; Turowski et al., 2014). On one hand, this could mean that an immature pre-rRNA has been exported to the cytoplasm, but it involves the bypassing of several quality control steps rendering this scenario very unlikely (Lebaron et al., 2012; Parker et al., 2019; Strunk et al., 2012). On the other hand, the pre-rRNA might have been cleaved in the nucleus by the already associated endonuclease Nob1. Although Nob1 should be kept inactive until required in the cytoplasm (Lamanna and Karbstein, 2009, 2011; Lebaron et al., 2012; Strunk et al., 2012), studies on the endonuclease Utp24 indicated alternative cleavage events when pre-ribosomal particles are accumulated (Choque et al., 2018). This scenario would imply a leaky inhibition of Nob1 in the nucleolus when the 5’ETS particle and maybe also other SSU-processome components are not released from the pre-40S particle. In agreement with our results, the absence of Rps27 also correlates with an anomalous cleavage around D site when A2 cleavage is delayed (Ferreira-Cerca et al., 2005).

Altogether, our data suggest a dual role of Pol5 in ribosome assembly. Pol5 is required for the assembly of domain III and the PET within the large ribosomal subunit and for the recycling of the tUtps from the SSU processome. Moreover, the AIM domain of Pol5 might play a role in recruiting the exosome to the 5’ETS region. We propose that Pol5 might act as a sensor for the communication between both assembly processes. Since 18S rRNA will be synthesized independently of 25S rRNA, processivity of RNA Pol I below 100% efficiency might cause an overaccumulation of 40S subunits, which would disrupt protein homeostasis. Therefore, this control mechanism might be raised to fine tune processivity defects of RNA Pol I to assure the stoichiometric production of both ribosomal subunits.