2.1. Deletion of NAM7, PUS2, and RPL27B Increases Yeast Sensitivity to Lithium Chloride
Chemical sensitivity of mutant strains under treatment of a target compound is an investigation tool that allows us to determine the targeted pathways of the compound and allows observation into the effects on the cell at a molecular level [
26,
27,
28]. While investigating yeast gene deletion mutants that are sensitive to LiCl, we identified three gene deletion mutants:
NAM7,
PUS2, and
RPL27B which showed increased sensitivity to LiCl (
Figure 1B), compared to control media without LiCl treatment (
Figure 1A). In the spot test assay, we demonstrate that deletion of
TIF2 (eIF4A) causes increased sensitivity to 10 mM LiCl when grown in media with galactose as the carbon source. Deletion of
NAM7,
PUS2, and
RPL27B showed dramatic growth reduction in the same manner (
Figure 1B), suggesting sensitivity to LiCl. Reintroduction of the deleted genes back into the chromosome of the corresponding gene deletion mutants reversed this growth reduction, further connecting the observed phenotypes to the deleted gene (
Supplementary Figure S1). We also tested to see the sensitivity of yeast strains to LiCl using glucose as the carbon source; however, no significant toxicity was observed (data is not shown).
It has been previously shown that over-expression of
TIF2 reverts the toxicity of sensitive strains to LiCl [
14]. To see if our candidate genes would revert LiCl sensitivity in the same manner, we transferred over-expression plasmid of our candidate genes into corresponding deletion strains spotted on media containing 10 mM LiCl. When the plasmids were introduced to the mutants, the fitness of the strains was recovered, proposing they may have similar function as eIF4A in the cell (
Figure 1C). The molecular activity of
NAM7,
PUS2, and
RPL27B has never been connected to LiCl sensitivity and the molecular pathways related to that, making them interesting targets for further investigation.
These findings were confirmed with colony count measurement analysis, providing a quantitative approach (
Figure 2). In this experiment, the number of colonies seen in the presence of LiCl in the media is compared and normalized to the number of colonies seen in control media and wild type (WT). Using this data, we were able to calculate the decreased percentage of colonies seen in deletion strains. As seen in
Figure 2, deletion of
NAM7,
PUS2,
RPL27B, or
TIF2 lead to decreased colony formation.
LiCl targets
PGM2 expression which leads to an accumulation of galactose intermediate metabolites, which becomes toxic for yeast cells including galactose-1-p [
14,
15,
29]. In yeast,
GAL1 is the galactokinase that phosphorylates α-
d-galactose to α-
d-galactose-1-phosphate in the first step of galactose catabolism. The
PGM2 enzyme facilitates the entry of galactose into glycolysis and converts glucose-1-phosphate to glucose-6-phosphate. To investigate the influence of
NAM7,
PUS2, and
RPL27B on LiCl toxicity through galactose metabolism, we generated double gene deletions for
NAM7,
PUS2, and
RPL27B with the
GAL1 gene. We observed that when
GAL1 was deleted, double mutant cells did not show sensitivity to LiCl treatment (
Figure 1B).
2.3. Translation of β-Galactosidase Reporter mRNA with a Hairpin Structure Is Altered by Deletion of NAM7, PUS2, and RPL27B
The 5′-UTR of
PGM2 mRNA has a highly structured region [
30] (
Figure S2).
PGM2 expression is severely reduced in the absence of translation initiation helicase
TIF2, a protein responsible for unwinding mRNA structures during translation [
14]. Since we observed that
NAM7,
PUS2, and
RPL27B are likely to function in the translation pathway, we decided to examine if these genes are possibly impacting translation of highly structured mRNAs. For this experiment, we inserted 5′-UTR of
PGM2 in front of a
LacZ expression cassette of p416 plasmid (pPGM2). From here, pPGM2 was then transformed into deletion mutant strains of our candidate genes and WT.
β-galactosidase activity was measured as a reference for translation activity (
Figure 4A). It was shown that translation activity significantly decreased in
Nam7Δ,
Pus2Δ, and
Rpl27bΔ carrying plasmid with hairpin when compared to translation activity with control plasmid (
Figure 4B).
To investigate whether this effect is specific to the 5′-UTR of
PGM2 mRNA, we used a second construct with a strong hairpin structure on its 5′-UTR (p281-4) [
31]. As a control, we used p281, which lacks a secondary structure [
31]. We observed that
β-galactosidase activity was reduced in p281-4 when
NAM7,
PUS2, or
RPL27B was deleted, whereas, for p281, there was no significant difference between the mutants and WT (
Figure 5A,B). These results demonstrate that the tested genes do not appear to affect translation of mRNAs lacking structured regions and exert their activities specifically on structured mRNAs.
Since
NAM7,
PUS2, and
RPL27B impacted translation of structured reporter mRNAs, we wanted to see if they were able to impact translation of other naturally structured mRNAs compared to the previous ones which were designed by software. For this purpose, we designed two
β-galactosidase mRNA reporters with different complex RNA structures: pTAR and pRTN. pTAR has the 5′-UTR of the
HIV-tar1 gene, which has a strong hairpin loop, while pRTN has the 5′-UTR of the
FOAP-
11 gene, which has a highly structured region [
32,
33]. When
NAM7,
PUS2, or
RPL27B were deleted, levels of
β-galactosidase expression were significantly reduced (
Figure 5C,D).
2.4. Genetic Interaction Analysis Further Connects the Activity of NAM7, PUS2, and RPL27B to Protein Biosynthesis
Genetic interaction (GI) analysis assumes that parallel pathways allow for plasticity and tolerance against random deleterious mutations, protecting cells and maintaining cell homeostasis if one gene is deleted or mutated in a pathway [
34]. This means that a gene in one pathway can compensate for the lack of gene activity in a parallel pathway, allowing the cell to survive. Accordingly, when two genes in parallel pathways are deleted, cell fitness decreases (sickness) or the cell dies (lethality). As a result, we assert that the two genes are genetic interactors, or in other words, they are functionally working in parallel pathways. These aggravating interactions are also known as negative genetic interactions (nGIs). nGIs are useful in many studies to understand gene function and pathway crosstalk [
26,
35,
36].
Analysis of GIs in yeast is done by mating two types of yeast: α-mating type (Mat α), and a-mating type (Mat a). Mat “α” carries the target gene deletion, which is crossed with Mat “a”, an array of single gene deletions to produce double gene deletions [
37]. Fitness of double deletions is measured using colony size assessment [
38]. Using this method, we made double deletions for each of our three query genes
NAM7,
PUS2, and
RPL27B with nearly 1000 other genes to screen for possible genetic interactions. This experiment included a random set of 304 genes as controls.
We observed several interesting nGIs with
NAM7, including
PRP22,
TIF2,
GCD11, and
PRT1. PRP22 is a DEAH-box RNA helicase,
TIF2 codes for the translation initiation factor eIF4A,
GCD11 forms part of the small subunit of eIF2, and
PRT1 is the subunit of eIf3. Other interacting genes involved in translation initiation are
YGR054W and
HCR1 (
Figure 6).
PUS2 interacted with
DHH1,
EAP1, and
HCR1 among others.
DHH1 codes for an ATP-dependent RNA helicase,
EAP1 codes for an eIF4E-associated protein, and
HCR1 codes for a subunit of the eIF3 translation initiation factor, which is also important in binding of initiation factors to 40S subunit and AUG recognition along with
HCR1 as an RNA recognition motif [
39]. Many of
PUS2 nGIs were involved with ribosomal structural proteins (
Figure 6).
As expected,
RPL27B interacted with a number of translation genes (
Figure 6). This included
DHH1,
EAP1,
SLH1, and
SKI2, which have RNA helicase activity, as well as
ECM32, which has DNA helicase activity and is also involved in modulating translation termination.
Comparing the nGI profiles for
NAM7,
PUS2, and
RPL27B we noticed some interesting common hits.
HCR1 is a subunit of eIF3 and is known to be important for scanning efficiency specially in cooperation with
DED1 (RNA helicase) on long 5′-UTRs [
40] and binding with
DHX29 in human cells [
41].
PRT1 is also crucial in recognition of the right start codon. During scanning,
PRT1 inhibits leaky scanning by promoting the stability of ribosomes on mRNAs possibly by changing its conformation [
42,
43].
XRN1 is known to be involved in mRNA decay and transcription regulation, but recently it was proposed that it might also play a role in translation pathway by regulating translation of specific mRNAs through binding to eIF4F complex, a translation initiation complex [
44].
DED1 is an ATP-dependent helicase that associates with eIF4A to regulate translation initiation [
45,
46]. Methylation of
DED1 strengthens its binding to eIF4A and to
XRN1 [
47], suggesting its potential effect on translation regulation.
XRN1 was also proposed to interact with another helicase,
DHH1, in yeast to control translation by decapping mRNAs for degradation [
48,
49].
P54 (homologue of
DED1 in humans) was shown to be important in localization and assembly of P-bodies in the cell [
48].
The comparison of this data to the published nGIs in Saccharomyces Genome Database (SGD) shows an expected degree of overlap. For example, for NAM7, except for an observed nGI with YGR054W in the current study, other observed nGIs overlap with the reported interactions. This provides further confidence for the validity of our observations.
Another form of nGIs, called conditional nGI, can be studied under sub-inhibitory concentration of chemicals. It is used to investigate possible candidate gene functions that are activated under a specific condition [
27,
50]. For this experiment, we studied nGIs for
NAM7,
PUS2, and
RPL27B under a mild sub-inhibitory concentration of LiCl (3 mM). Shown in
Figure 7, new nGIs were observed for our candidate genes. The new interactions hinted to new conditional functions for
NAM7,
PUS2, and
RPL27B in regulation of translation. We observed a number of common interactors between the three query genes including
BCK1,
CTK1, and
MCK1.
BCK1 is important in negative regulation of translation under stress conditions. It is involved in deadenylation and decapping of mRNAs to be degraded in connection with
DHH1 [
51,
52,
53,
54]. On the other hand,
CTK1 is a conserved kinase that phosphorylates RPS2p, one of the components of small ribosomal subunit. It affects translation fidelity during elongation, as well as phosphorylation of other translation initiation factors, including eIF4A, eIF5, eIF4G, and eIF3 [
55,
56].
MCK1 is involved in phosphorylation-dependent protein degradation, among other roles.
Phenotypic Suppression Array (PSA) analysis focuses on another form of interaction where over-expression of one gene compensates for the absence of another gene, under a drug treatment [
26,
36,
57,
58]. Here, we treated the arrays of mutant strains with 10 mM LiCl, and, consequently, some showed sensitivity. Then, we were able to revert LiCl sensitivity in a number of deletions by introducing over-expression plasmid of
NAM7,
PUS2, and
RPL27B. Interestingly, over-expression of
NAM7,
PUS2, and
RPL27B compensated for the sensitivity of four mutual gene deletions
ITT1Δ, eIF2A
Δ,
EAP1Δ, and
PSK2Δ (
Figure 8).
Eap1 encodes for an eIF4E-associated protein that accelerates decapping of mRNA and negatively regulates translation [
53].
ITT1 interacts with the translation release factor eRF3 and modulates the efficiency of translation termination. eRF3 is not only important in regulation of translation termination and cell cycle regulation but also is shown to mediate mRNA decay through interaction with UPF family of proteins that include one of our query genes
NAM7 [
59]. Itt1p interacts with Mtt1p, an RNA helicase and poly(A) binding protein Pab1p involved in mRNA circularization and ribosome recycling [
59,
60]. eIF2A deletion also showed sensitivity to 10 mM LiCl and was reverted by over-expression of
NAM7,
PUS2, and
RPL27B in our experiments. eIF2A, a translation initiation factor, is shown to regulate IRES-mediated translation and because of its physical interaction with Ded1p and eIF4A, it has been suggested to play a role in translation regulation of certain mRNAs [
61].
PSK2 is known to be important in regulating both sugar metabolism and translation through phosphorylation of intermediate molecules [
62]. In addition,
PSK2 is shown to function as a trans-acting factor for translational regulation of mRNAs with upper Open Reading Frames (uORFs), including
ROK1 that encodes an RNA helicase [
63].