Lithium Chloride Sensitivity Connects the Activity of PEX11 and RIM20 to PGM2 translation

Ashkan Golshani (  Ashkan.Golshani@carleton.ca ) Carleton University Faculty of Science https://orcid.org/0000-0003-3977-5367 Sasi Kumar Jagadeesan Carleton University Faculty of Science Mustafa Algafari Carleton University Faculty of Science Maryam Hajikarimlou Carleton University Sarah Takallou Carleton University Houman Moteshareie Carleton University Azam Tayabali Carleton University Faculty of Science Bahram Samanfar Carleton University Faculty of Science Myron Smith Carleton University Faculty of Science

In the current study, we report that deletion of PEX11 and RIM20 increase cell sensitivity to LiCl when the gene mutant strains are grown in galactose. The primary reported function of PEX11 is to mediate fatty acid oxidation and induce peroxisome proliferation [17]. RIM20 is shown to be actively involved in pH response and multivesicular body formation [18,19]. Our subsequent genetic studies indicate that these genes are involved in regulating the translation of PGM2 mRNA.

Strains and Plasmids
The MAT "α" mating strain, Y7092 can1∆::STE2pr-Sp_his5 lyp1∆ his3∆1 leu21∆0 ura3∆0 met15∆0 and MAT "a" mating strain, Y7092 Y4741 orf∆::KanMAX4 his3∆1 leu2∆0 met15∆0 ura3∆0are utilized in this study. To construct double deletion mutant strains, non-essential gene deletion mutant strains were collected from the Yeast knockout library [20] and used for Synthetic Genetic Analysis (SGA). SGA gene knockouts in MATα mating strains were accomplished using PCR and homologous recombination after the Lithium acetate process, as followed by previous PCR studies [21,22]. The yeast overexpression plasmid library was used to generate overexpression plasmids for the gene candidates [23]. Thermo sher® Yeast gene-GFP fusion library is purchased and used to generate PGM2-GFP fusion strain for qRT-PCR and western blot analysis as shown before [24][25][26]. Chemical sensitivity and PCR analyses are used to validate the overexpression plasmid. A modi ed SGA approach coupled with random spore analysis is used to establish deleted genes reinserted into their respective deletion mutant strains [27,28]. Crossing MATα with nourseothricin sulphate (clonNAT) and MATa with a G418 resistance gene in lieu of the gene deletion of interest results in modi ed SGA. For MATa haploid progeny selection, spores germinated on selective media containing thialysine and canavanine after sporulation. G418 sensitivity was examined using replication plating after the cells were transferred to clonNAT media. PCR analysis was conducted to validate the target gene inside the chromosome in colonies immune to clonNAT but sensitive to G418. The p416 plasmid carried a non-complex construct, used as the control plasmid and carried a LacZ expression cassette under gal promoter control. HIV1 (TAR), RTN4IP1 (RTN), BCL-2 (Bcell) and 2-hair hairpin constructs were designed and cloned into the p416 expression vector upstream of LacZ mRNA using XbaI restriction site between the β-galactosidase reporter gene (LacZ expression cassette) and gal promoter.

Media Inoculation
Yeast cultures were grown with either YP media (1% Yeast extract and 2% Peptone) or Synthetic Complete media with selective amino acids (0.67% Yeast nitrogen base w/o amino acids and 0.2% amino acid dropout mixture). The carbon source in yeast media was either 2% glucose or 2% galactose. Agar at 2% was used to make solid media. LB media (Lysogeny Broth) was used for E.Coli cultures. GeneJET plasmid miniprep kit (Thermo sher®, Ottawa, ON, Canada and Bio-Basics®, Toronto, ON, Canada) was used for E. coli plasmid extraction, and yeast plasmid miniprep kit (Omega Biotek®, Norcross, GA, USA) was used for yeast plasmid extraction; all kits were used as directed by the manufacturer.

LiCl sensitivity analysis
Speci c colonies of yeast are cultivated for two days at 30 o C in liquid YPgal (YP + 2% galactose) media before growth saturation. Liquid cultures are standardized to an OD 0.1 at 600nm after incubation, serially diluted (10 −1 to 10 −4 ) and spot platted (10ul per spot) onto agar YPgal media with 10 mM LiCl treatment. Sensitivity to LiCl could be assessed by comparing gene deletion strains with wild type (WT). Overexpression constructs pPEX11 and pRIM20 are transformed into our gene deletion strains to validate observed sensitivities due to the deletion of our candidate genes. For quanti cation analysis, colony counting was performed using 100 µL of diluted 10 −4 concentration cell cultures then spread onto YPgal plates with and without 10 mM LiCl. The colonies were counted after a 48h incubation at 30°C. Each experiment was conducted in triplicates. One way ANOVA analysis (P-value ≤ 0.05) was used to determine signi cant statistical differences.

mRNA quanti cation analysis
PGM2-GFP yeast strain grown in liquid YPgal media with 10 mM LiCl treatment was used to determine PGM2 mRNA concentrations in different deletion strains. Max RNA is collected using the Qiagen® RNeasy Mini Pack. iScript The Bio-Rad® cDNA Synthesis Kit (Mississauga, ON, Canada) was used to create complementary DNA and Bio-Rad® iQ SYBR Green Supermix along with CFX connect real-time system (Bio-Rad®) are used to perform qRT-PCR. The housekeeping gene PGK1 is used as an internal control. Experimentation and data processing is carried out in compliance with MIQE guidelines [29]. Experiments are replicated with three technical and biological replicates. One way ANOVA analysis (P-value ≤ 0.05) was used to determine signi cant statistical differences. Primers PGK1 Forward: ATGTCTTTATCTTCAAAGTT; Reverse: TTATTTCTTTTCGGATAAGA; PGM2 Forward: GGTGACT CCGTCGCAATTAT; R: CGTCGAACAAAGCACAGAAA 2.5 PGM2 Protein content measurement Quantitative western blotting is used to determine the protein content of the PGM2p-GFP fusion protein. Mutant strains were cultured in liquid YPgal media with 10 mM LiCl. The protein extraction process is carried out as mentioned previously [30]. Bradford Protein Assay (BSA) was used to determine the protein concentration. 50ug of total extracted protein were loaded onto a 10% SDS-PAGE gel and ran through the Mini-PROTEAN Tetra cell electrophoresis apparatus method (Bio-Rad®). Trans-Blot Semi-Dry Transfer (Bio-Rad®) was used to transfer protein through a 0.45 µm nitrocellulose membrane. PGM2p-GFP protein levels are detected using a mouse monoclonal anti-GFP antibody (Santa Cruz®), and PGK1 protein levels are measured using a mouse monoclonal anti-PGK1 antibody (internal control). Immunoblot visualization and densitometry research are carried out using chemiluminescent substrates (Bio-Rad®), the Vilber Lourmat gel doc Fusion FX5-XT (Vilber®), and the FUSION FX app (Vilber®). Experiments are replicated with three technical and biological replicates.
One way ANOVA analysis (P-value ≤ 0.05) was used to determine signi cant statistical differences.

Quantitative β-galactosidase Assay
Translation initiation activity of our gene deletion strains is investigated using, LacZ reporter constructs (as previously described). LacZ activity was assessed using the assay as a quantitative tool to quantify translational activity in mutant yeast strains. The assay was carried out using O-nitrophenyl-dgalactopyranoside [31,32]. Experiments are repeated three times with technological and biological replicates. Meaningful variations are evaluated using a one-way ANOVA study (P-value 0.05).

Genetic Interaction Analysis
Genetic interactions analysis for our candidate genes (GI) were investigated using a 384 formatted SGA. The yeast double deletion mutant library is generated as previously shown [24, 25,28]. A query strain (MATα, Y7092) carrying a gene deletion of interest and a clonNAT marker is mated with our three non-essential gene deletion sets, comprising ~1000 MATa mutants. Following rounds of selection, haploid double gene deletion mutants are selected.
Successful gene knockout transformations were identi ed using PCR analysis with pAG25 plasmid as the template. The clonNAT resistance gene marker was ampli ed, with NAT and Kanamycin used as selection markers. Conditional SGA is carried out by growing the generated double mutants in a chemical stress condition [33] on sub-inhibitory, 3 mM LiCl agar media. This facilitates the investigation of genes that only express under LiCl stress conditions [25,34]. Phenotypic Suppression Array (PSA) is performed by mating an empty plasmid, MATa single deletion library, and an overexpression plasmid of a query gene carried by MATα yeast strain, as mentioned previously for the SGA mating protocol. Strains with and without the overexpression plasmid are compared and grown in media with LiCl treatment. Overexpression of query genes was used to examine whether there is some phenotypic shift with inhibitory LiCl treatment. A proposed functional relation between these two genes could then be established [22,33].

Genetic Interaction data Analysis
The size of the colony determines cell tness [28,35]. Colony size comparisons between the control (WT) and single deletion mutants are performed to determine if any phenotypic changes (synthetic illness or synthetic lethality) exist. Normalized colony sizes were then compared. The tness threshold for yeast cells is set to a minimum of 30% size reduction in contrast to the control. SGA Software was used to measure colony sizes and similarities [36]. Only size reductions of 50-95% are taken into consideration. This experiment was repeated three times, each time with a separate technical replication. Only hits of at least two replicates are taken into consideration. Finally, identi ed hits are then classi ed according to their biological processes and molecular function and then enriched with Gene ontology terminology using tools including the Genemania database. http://genemania.org (accessed on January 18, 2021) and complex functionality pro ling http://webclu.bio.wzw.tum.de/profcom/start.php (Referenced on January 23, 2021).

Results And Discussion:
3.1 Yeast tolerance to Lithium Chloride is eroded by gene deletions of PEX11 and RIM20 Chemical-genetic approaches have the ability to allow the detailed annotation of the functional properties of various chemicals and bioactive compounds.
They can shed light on a compound's primary mode of action as well as its secondary sites of interactions within a cell [24]. The sensitivity of mutant strains to a target compound is a powerful approach for determining the target pathways and the impact a compound has on a cell at the molecular level. In the current study, while investigating yeast gene deletion mutants that are sensitive to LiCl, we discovered two deletion mutants for PEX11 and RIM20, that exhibited increased sensitivity to LiCl ( Figure 1A, B) compared to a control strain ( Figure 1A). Using a spot test analysis, we demonstrate that when cells are grown in media containing galactose as the primary carbon source, deletion of PEX11 and RIM20 resulted in a substantial reduction in growth in the presence of 10 mM LiCl, indicating increased sensitivity. Deletion of TIF2 (eIF4A) was used as a positive control. Further we demonstrated that incorporating the deleted genes back into the corresponding gene deletion mutants reversed the observed growth reduction establishing a correlation between the observed phenotypes and the deleted gene. Colony count measurement analysis further validated these results by offering a quantitative viewpoint ( Figure   1B). The formation of colonies observed in the presence of LiCl is compared and normalized to the number of colonies observed in the control strain. Lower number of colonies formed for gene deletion mutants indicates increased sensitivity to LiCl compared to the control strain. As shown in Figure 1B, deletion of TIF2, PEX11 and RIM20 under the in uence of LiCl results in a signi cant reduction in colony formation. As before reintroduction of the deleted genes back into the corresponding mutant cells increased the number of colonies formed to levels comparable to that of the control strain. Additionally, we examined yeast strains' exposure to LiCl when glucose was used as the carbon source. As expected, no increased sensitivity for the mutant strains was observed when glucose was used (S1. Figure).
Previously, it was stated that overexpression of TIF2 reversed the toxicity of susceptible strains to LiCl. We inserted overexpression plasmids for our candidate genes into corresponding deletion strains spotted on media containing 10 mM LiCl to see whether they would similarly restore LiCl sensitivity.
When the plasmids were integrated into the mutant strains, tness was restored, suggesting that they may function similarly to eIF4A in the cell.
When galactose is used as the carbon source, LiCl inhibits PGM2 expression, resulting in the accumulation of galactose intermediate metabolites that are toxic to yeast cells [11]. GAL1 is the galactokinase in yeast that phosphorylates -D-galactose to -D-galactose-1-phosphate in the initial stages of galactose metabolism. By converting glucose-1-phosphate to glucose-6-phosphate, PGM2 promotes galactose entry into glycolysis. We investigated the impact of PEX11 and RIM20 on LiCl toxicity when galactose metabolism was disabled. For this purpose, we developed double gene deletions for PEX11 and RIM20 with the GAL1 gene. Not surprisingly, we observed GAL1 double mutant cells were no longer hypersensitive to LiCl treatment. PEX11 and RIM20 have not been previously linked to LiCl sensitivity or the underlying molecular mechanisms associated with it, rendering them intriguing gene candidates to investigate.

PGM2
Expression is Controlled at the Translational Level by PEX11 and RIM20.
PGM2 plays a key role in the sensitivity of yeast cells to LiCl [12], we investigated whether PEX11 and RIM20 will in uence PGM2 expression. For this PGM2p was GFP-tagged, and western blot analysis was performed using anti-GFP antibodies to quantify protein content (Figure 2A). The deletion of PEX11 and RIM20 had no substantial effect on PGM2p protein levels in the absence of LiCl treatment. Interestingly, when cells were treated with LiCl, PEX11 and RIM20 gene deletion resulted in a signi cant reduction of PGM2p protein levels compared to the WT. To analyze the effect of PEX11 and RIM20 gene deletion on PGM2 transcription, the mRNA content of PGM2 was measured using qRT-PCR. Figure 2B indicates that deletion of PEX11 and RIM20 had similar levels of PGM2 mRNA levels as the WT in control media. However, PGM2 mRNA content increased in cells treated with LiCl, but there was no statistically relevant difference between mutant strains and WT. Thus, the deletion of PEX11 and RIM20 seems to have little in uence on PGM2 mRNA content. As a result, it appears that PEX11 and RIM20 in uence PGM2p at the level of protein synthesis. These ndings are similar to the observations by Hajikarimlou  3.3 Deletion of PEX11 and RIM20 In uences Translation of β-Galactosidase Reporter mRNAs with a Hairpin Structure, but not Those Without Several genes that affected PGM2 expression at the translation level, used the 5'-UTR of PGM2 mRNA to exert their activities. The 5'-UTR of PGM2 mRNA is predicted to have a structured region (S2. Figure) and in the absence of TIF2, a protein that unwinds mRNA structures during translation, PGM2 expression is reduced [14]. Next we intended to investigate whether PEX11 and RIM20 affect translation using the structured 5'-UTR of PGM2 mRNA. For this experiment, we utilized pPGM2 plasmid where the PGM2 5'-UTR is placed in front of a LacZ expression cassette in the p416 expression plasmid. The pPGM2 plasmid, as well as the parental p416 plasmid lacking a structural region in front of the LacZ gene were transformed into the deletion mutant strains for our candidate genes as well as the WT strain. β -galactosidase activity was quanti ed as a measure of translation ( Figure 3A). No signi cant difference was observed for β -galactosidase activity derived from mRNAs lacking a 5'-UTR structure (p416) in different strains. β -galactosidase activity was signi cantly reduced however in pex11∆ and rim20∆ when the mRNA carried PGM2 5'-UTR suggesting a connection between the activity of PEX11 and RIM20, and the translation of structured mRNAs.
Next, we examined the effect of PEX11 and RIM20 on additional structured mRNAs. For this we utilized four additional constructs each carrying different structures at their 5'-UTR. pTAR carries a structure derived from HIV1 mRNA containing a ΔG= -57.9kcal/mol and pRTN carries the structured 5'-UTR of RTN4IP1 mRNA with ΔG= -29.8kcal/mol. Similarly, pBcell carries a structure derived from BCL-2 mRNA and a ΔG= -20kcal/mol. The last construct, p2hair contains a synthetic structure designed to have a high degree of complexity ΔG= -33kcal/mol. In our analysis. we observed that deletion of PEX11 and RIM20 signi cantly reduced the expression of all four highly structured mRNAs, comparing the WT suggesting a general role for these two genes in the translation of structured mRNAs (Figure 4).

Genetic Interaction Analysis
Further Connects the Activity of PEX11 and RIM20 to Protein Biosynthesis Genetic interaction (GI) analysis is based on the principle that parallel pathways allow for exibility and tolerance to random harmful mutations, preserving cells viability and maintaining cell homeostasis [16]. A gene in one pathway might compensate for a gene in another, allowing the cell to survive.
Consequently, when two genes in parallel pathways are deleted, cell tness can be unexpectedly reduced (sickness) or even the cell dies (lethality). Therefore, when deletion of two genes result in a phenotype that is unexpected from the phenotypes of individual gene deletions, it is commonly said that the two genes are forming a genetic interaction. Due to the reduced tness of double mutant this type of interaction is called negative genetic interaction (nGI). Many studies use nGIs to investigate gene function and pathway interactions [15,16,29,37].
The high throughput analysis of GIs in yeast is performed by mating the two yeast mating types: α-mating type (Mat α) and a-mating type (Mat a). Mat "α" carries the target gene deletion and is crossed with an array of single gene deletions of Mat "a," mating type, to produce double gene deletions [28]. Colony size is used to measure the tness of the strains [35,38]. We utilised this method to investigate genetic connections between our query genes, PEX11 and RIM20, with approximately 1000 additional genes that includes approximately 700 genes associated with gene expression pathway and a random collection of 300 genes used as a control. (S1 . Table).
In our analysis, we discovered numerous intriguing nGIs and several common gene hits between PEX11 and RIM20. Functional enrichment analysis of the hits revealed that a signi cant number of them are connected to relevant biological processes that includes translation regulation, ribosome biology,, mRNA catabolic processes, and protein synthesis and protein synthesis associated factors ( Figure 5). For PEX11, we identi ed nGIs including DPH5, PCI8, EAP1, and TIF1 among others. The methyltransferase, DPH5 is engaged in the diphthamide biosynthesis pathway, responsible for assembly of translation elongation factor 2 (Eft1p or Eft2p [39]., PCI8 forms the subunit of translation initiation factor, eIF3b, contributing to the formation of 43S pre-initiation complex during translation initiation. It is a key factor in cellular signalling at all stages of protein synthesis, including elongation, termination, and ribosome recycling [40]. EAP1, is also involved in translation initiation process. It is an associated protein of translation initiation factor eIF4E, which forms a part of eIF4F complex. It could potentially enhance/inhibit the overall translation rate by facilitating mRNA degradation and promoting mRNA decapping [41]. Similarly, RIM20 interacted with several translation machinery associated genes including RPS23A, TOR1, and TMA19 ( Figure 5). RPS23A codes for the ribosomal protein 28 (rp28) and forms a component of the 40S small ribosomal subunit [42]. Translation regulator protein TOR1 is a member of the TOR complex, which is known to regulate global translation rate through signal transduction. Under stress environments, the TOR-related proteins TOR1 and TOR2 control translation initiation and promote early G1 progression in yeast [43]. TMA19 codes for Translation Machinery Associated protein 19 that interacts with ribosomes during translation initiation. [44].
We also observed several common interactors between PEX11 and RIM20 including SLH1, and CAF20. SLH1 is a putative RNA helicase that is directly implicated in the translation inhibition of non-poly(A) mRNAs. [45]. CAF20 is a translational control phosphoprotein found in the mRNA cap-binding complex.

It acts as a regulator of cap-dependent translational initiation [46].
Conditional nGIs are those interactions that are formed under a speci c condition that includes the presence of a sub-inhibitory concentration of a bioactive compound, cold shock, heat shock, minimum media etc. They represent gene functional associations that are formed in response to a certain environment [30,47]. For example, in the presence of DNA damage, the functions of certain genes may be altered, and it is the altered functions that are functionally related and hence form the basis of GIs in the presence of DNA damage [48]. We investigated nGIs for PEX11 and RIM20 using a mild subinhibitory concentration of LiCl (3 mM). The observed GIs in this case are those that are formed as a response to the presence of LiCl. As seen in Figure 6, we found new nGIs for our candidate genes PEX11 and RIM20. PEX11 interacted with several translation regulating genes including PBP1, and EBS1 among others. PBP1 is a component of glucose-deprived induced stress granulates that in uences global translation rate through the TORC1 signalling pathway and autophagy. It is also known to have an active role in controlling mRNA polyadenylation by interacting with Pab1P, which in uences initiation of translation during protein biosynthesis [49]. EBS1 codes for a protein involved in translation inhibition by interacting with cap binding proteins CDC33 and NAM7 upon glucose starvation. It also regulates translation initiation factor eIF4E, another cap binding protein that binds to 5' end of the mRNA during the formation of 48S initiation complex in translation initiation step [50]. Interestingly, eIF4E is a component of the eIF4F complex, that comprises of eIF4A, eIF4G and eIF4E [51]. eIF4G is a scaffolding protein and eIF4A is a DEAD-box helicase protein that is actively involved in unwinding structured mRNAs during translation initiation. PEX11 interacting with EBS1 may indirectly in uence helicase activity of eIF4A and potentially affect other helicases in uencing global translation rate.
Similarly, RIM20 interacted with several news genes involved in translation control and regulation of translation. ETT1 is a nuclear protein that regulates overall translation rate and enhances/inhibits protein synthesis by in uencing initiation and termination step during translation [52]. Under extreme stress conditions, PTP2 modulates phosphorylation in the MAPK signalling pathway, affecting mRNA decapping and mRNA stability in uencing overall translation rate [53]. RPS19B is a protein component of the 40s ribosomal subunit that participates in the assembly of 43s pre-initiation complex by binding matured mRNA to small ribosomal subunit with translation initiation factors eIF1A and eIF3 during protein synthesis [54].
Notably, we also observed several common interactors between PEX11 and RIM20 including DHH1, TIF2, and SHE3 that are involved in translation control. DHH1 is a cytoplasmic DEAD-box helicase that regulates translation process in uencing mRNA stability, mRNA degradation and polyadenylation at 5' end of mRNA [55]. As explained above, translation initiation factor, TIF2 is another DEAD box RNA helicase that facilitates the binding and unwinding of mRNA during translation initiation [51]. SHE3 is a RNA binding protein that constitutes a part of the mRNA localization machinery and plays an active role in recruiting Myo4p-She3 complex that is involved in mRNA export / localization into the cytoplasm, in uencing global translation rate[56].
Phenotypic suppression array (PSA) analysis focuses on another kind of interaction in which overexpression of one gene compensates for the absence of another [24,33,57,58]. This is an important type of interaction as it can highlight certain functional overlaps between two genes. To this end we exposed the gene expression mutant arrays to 10 mM LiCl. Some of the investigated strains showed increased sensitivity. Then, by introducing PEX11 and RIM20 overexpression plasmids, we attempted to compensate for the observed sensitive phenotypes. Interestingly, the overexpression of either PEX11 or RIM20 compensated for the sensitivity of two gene deletions for RPS30A and SLF1 ( Figure 6A). RPS30A encodes a cytosolic small ribosomal subunit protein that can also bind to mRNAs and contributes to the selection of the correct tRNA molecules [59]. SLF1 is an RNA binding protein that regulates mRNA translation, especially under stress conditions. [60].
The recovery of the sick phenotype in gene deletions for SLF1, RPS30A by introducing RIM20 and PEX11 overexpression plasmid was con rmed using spot test ( Figure 6B). This was further veri ed by cell quanti cation analysis using colony count ( Figure 6C). Using spot test and colony count analysis we veri ed that the overexpression of RIM20 and PEX11 compensated the sensitivity to LiCl that was seen for deletion strain for RPS30A and SLF1. Both SLF1 and RPS30A have reported genetic interactions with TIF2. Consequently, we also included the deletion strain for TIF2 in the spot test for analysing LiCl sensitivity compensation. Interestingly, when PEX11 and RIM20 are overexpressed in TIF2 deletion, LiCl sensitivity caused by TIF2 deletion was also recovered further connecting the activity of these genes.

Concluding Remarks:
When used as a therapeutic molecule, LiCl has a broad impact on the expression of various genes and pathways among other activities. Its precise mechanism of action however remains unclear. Here, we examined yeast gene deletion mutant strains for LiCl sensitivity and discovered that when two genes, PEX11 and RIM20, were deleted, the corresponding deletion mutant strains were more sensitive to LiCl. TIF2 overexpression, which codes for eIF4A, the helicase enzyme involved in translation initiation, has been shown to restore LiCl sensitivity in yeast cells cultured on galactose media [14]. We determined that overexpression of our candidate genes causes a phenotypic recovery similar to TIF2, suggesting that these genes may have a similar/related function. In the literature the molecular activities for PEX11 and RIM20 have not been linked to LiCl sensitivity making these genes intriguing target for further study. LiCl toxicity in yeast is thought to be caused by the accumulation of galactose intermediate metabolites, particularly galactose-1p [14,15]. It has been shown that LiCl inhibits the conversion of glucose-1-phosphate to glucose-6-phosphate via modifying the expression of the phosphoglucomutase enzyme PGM2 [11]. The current study shows that PEX11 and RIM20 in uence PGM2 expression. We demonstrate that PEX11 and RIM20 regulate the translation of PGM2 mRNA via its 5′-UTR region. Similar observations were made using other mRNAs with various 5'-UTRs that carry different structures, but not for the ones without. We recently reported on the nding of several additional genes involved in the translation of structured mRNAs, [15,16]. The identi cation of new gene functions that in uences the translation of structured mRNAs suggests that translation of mRNAs with structures seems to be more complex than what was originally thought. Furthermore, it shows that the structured mRNAs might be subjected to a more complicated regulatory network during translation than previously hypothesized.
The impact of PEX11 and RIM20 on structured mRNA translation may be explained in a variety of ways. The simplest explanation is that one or more of these components may have mRNA helicase activity, which may directly contribute to the unwinding of the structured mRNAs at 5'UTR. In the absence of recognizable helicase motifs, the possibility of a helicase activity for these proteins is unlikely. It is also possible that these factors might affect the activity of other helicases and consequently, mediate the translation of structured mRNAs. Modifying the activity of ribosome can be another way that that PEX11 and RIM20 might in uence the translation of structured mRNAs. There are no reported physical interactions between these factors and ribosomes making this a less likely scenario. However, in the current study we uncovered several interactions for RIM20 with several ribosomal proteins suggesting a potential link between RIM20 and ribosomes. An alternative explanation is that these factors may exert their activity on the biology of the mRNAs. In this way PEX11 and RIM20 might affect mRNA biogenesis or affect factors that contribute to mRNA biogenesis.
The ndings of the current study further connect the biological activity of LiCl to the translation of structured mRNA in yeast. This knowledge may contribute to our understanding of the biology of BD. LiCl is widely considered to be a neuroprotective drug that prevents neuronal cell death. According to the observations made in the current study, LiCl's mode of action may include altering translation of certain structured mRNAs linked to the process of cell death and apoptosis. In future, it would be interesting to study the effect of LiCl on the translation of such mRNAs. Inhibiting the expression of these mRNAs may be considered for therapeutic purposes. Figure 1 Spot test and colony count analysis reveal increased LiCl sensitivity for tif2∆, pex11∆ and rim20∆, in comparison to WT, with respective overexpressions and GAL1 overexpression causing similar tolerance compared to WT. A) Yeast cells are serially diluted (10 −1 to 10 −4 ) and spotted onto Ypgal media with 10 mM LiCl. Each experiment was repeated at least three times with similar outcomes as shown. WT is used as a control. B) Mutant strains tif2∆, pex11∆ and rim20∆ formed signi cantly reduced number of colonies under LiCl treatment compared to the WT. Re-introduction of the target genes results in a reversal of LiCl sensitivity in gene deletion mutants, with GAL1 overexpression of target genes resulting in similar reversal of sensitivity. Bars represent mean values (± S.D., n ≥ 3), error bars represent standard deviation, and different letters indicate statistically signi cant differences (One-way ANOVA with Post Hoc Tukey Test, a = 0.05).

Figure 2
Protein and mRNA content analysis for PGM2-GFP with and without 10 mM LiCl treatment via western blot and qRT-PCR analysis. A) PGM2p-GFP protein content is reduced in deletion strains pex11∆ and rim20∆ when cells are treated with LiCl. PGK1p is used as the housekeeping gene/control with obtained values normalized to it. B) No difference was observed for the mRNA content for the deletion strains. Please note that there was an approximately 40% increase in the absolute values for the mRNA content level between LiCl treated and non-treated cells. Each experiment is repeated at least three times. Bars represent mean values (± S.D., n ≥ 3), error bars represent standard deviation and '*' represents statistically signi cant results compared to the WT (One-way ANOVA with Post Hoc Tukey Test, a = 0.05). Data are normalized to those for the WT strain under the same condition (with or without LiCl treatment) Figure 3 β-galactosidase activity analysis of simplistic construct p416, with and without PGM2 hairpin. A) Each strain carries a p416 mRNA construct with no complex structure upstream at the 5' UTR of the LacZ reporter. No signi cant difference in β-galactosidase activity between yeast strains. B) Each strain carries a complex PGM2 mRNA construct at the 5'UTR, upstream from the LacZ reporter. β-galactosidase activity is reduced in tif2∆, pex11∆ and rim20∆ strains in comparison to WT. Bars represent mean values (± S.D., n ≥ 3) and '*' represents statistically signi cant results compared to WT, (One-way ANOVA with Post Hoc Tukey Test, a = 0.05). β-galactosidase activity analysis in yeast strains containing varying complex secondary structures. A) Each strain carries a complex p2-hair mRNA construct at the 5'UTR, upstream from the LacZ reporter. β-galactosidase activity is reduced in tif2∆, pex11∆ and rim20∆ strains in comparison to WT. B) Each strain carries a complex pRTN mRNA construct at the 5'UTR, upstream from the LacZ reporter. β-galactosidase activity is reduced in tif2∆, pex11∆ and rim20∆ strains in comparison to WT. C) Each strain carries a complex pTAR mRNA construct at the 5'UTR, upstream from the LacZ reporter. β-galactosidase activity is reduced in tif2∆, pex11∆ and rim20∆ strains in comparison to WT. D) Each strain carries a complex pBcell mRNA construct at the 5'UTR, upstream from the LacZ reporter. β-galactosidase activity is reduced in tif2∆, pex11∆ and rim20∆ strains in comparison to WT. Each analysis was normalized to each hairpin structure's respective WT. Bars represent mean values (± S.D., n ≥ 3) and '*' represents statistically signi cant results compared to WT, (One-way ANOVA with Post Hoc Tukey Test, a = 0.05).

Figure 5
Negative genetic interactions (nGIs) for PEX11 and RIM20. A cluster of interactors fall in the category of protein synthesis and protein synthesis associated factorsPEX11 (p = 4.49E-06) and RIM20 (p = 6.32E-07). SLH1, SGN1, DPH2 and CAF20, are mutual hits shared between two target genes. Genes are represented via circles (nodes) and lines/dotted lines (edges) represent nGIs identi ed. The red circle within the inset represents an example of a typical negative genetic interaction. Intensity of interactions is shown through color gradients, darker coloration indicating reduced tness. Existing genetic interactions refer to those GIs that have been previously reported in the literature. Conditional nGIs for PEX11 and RIM20 in the presence of 3mM LiCl. In the presence of LiCl, a new set of genetic interactors are revealed that are involved in the regulation of translation for both PEX11 (p=7.83E-06) and RIM20 (p= 5.33E-6). CPA1, SHE3, DHH1 and TIF2 are mutual interactors shared between PEX11 and RIM20. Genes are represented via circles (nodes) and lines/dotted lines (edges) represent nGIs identi ed. The red circle within the inset represents an example of a typical negative genetic interaction. Intensity of interactions is shown through color gradients, darker coloration indicating reduced colony growth. Existing genetic interactions refer to those GIs that have been previously reported in the literature. Overexpression of PEX11 and RIM20 compensate for the sensitivity of SLF1∆, and RPS30A∆ to 10mM LiCl A) RPS30A and SLF1 are both involved in translation control. Overexpression of PEX11 and RIM20 recovers the phenotype caused by LiCl sensitivity for the deletion strains for RPS30A and SLF1. B) Spot test analysis further showed that PEX11 or RIM20 overexpression increased tolerance to 10 mM LiCl in the deletion strains for SLF1, RPS30A and TIF2.
For this experiment, yeast cells are serially diluted from 10 −1 to 10 −4 and spotted onto Ypgal media with 10 mM LiCl. Each experiment was repeated at least three times with similar outcomes as shown. WT is used as a control. C) Colony count analysis further veri ed that PEX11 and RIM20 overexpression compensated for the LiCl sensitivity in SLF1, RPS30A and TIF2 gene deletion strains. Bars represent mean values (± S.D., n ≥ 3), and different letters indicate signi cant differences (One-way ANOVA with Post hoc Tukey Test, α = 0.05).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.