Synthetic dosage lethal (SDL) interaction data of Hmt1 arginine methyltransferase

The introduction of methyl groups on arginine residues is catalysed by Protein Arginine Methyltransferases (PRMTs). However, the regulatory mechanisms that dictate the levels of protein arginine methylation within cells are still not completely understood. We employed Synthetic Dosage Lethality (SDL) screening in Saccharomyces cerevisiae, for the identification of putative regulators of arginine methylation mediated by Hmt1 (HnRNP methyltransferase 1), ortholog of human PRMT1. We developed an SDL array of 4548 yeast strains in which each strain contained a single non-essential gene deletion, in combination with a galactose-inducible construct overexpressing wild-type (WT) Hmt1-HZ tagged protein. We identified 129 consistent SDL interactions for WT Hmt1-HZ which represented genes whose deletion displayed significant growth reduction when combined with WT Hmt1 overexpression. To identify among the SDL interactions those that were dependent on the methyltransferase activity of Hmt1, SDL screens were repeated using an array overexpressing a catalytically inactive Hmt1(G68R)-HZ protein. Furthermore, an additional SDL control screen was performed using an array overexpressing only the protein tag HZ (His6—HA-ZZ) to eliminate false-positive SDL interactions. This analysis has led to a dataset of 50 high-confidence SDL interactions of WT Hmt1 which enrich eight Gene Ontology biological process terms. This dataset can be further exploited in biochemical and functional studies to illuminate which of the SDL interactors of Hmt1 correspond to factors implicated in the regulation of Hmt1-mediated arginine methylation and reveal the underlying molecular mechanisms.


a b s t r a c t
The introduction of methyl groups on arginine residues is catalysed by Protein Arginine Methyltransferases (PRMTs). However, the regulatory mechanisms that dictate the levels of protein arginine methylation within cells are still not completely understood. We employed Synthetic Dosage Lethality (SDL) screening in Saccharomyces cerevisiae , for the identification of putative regulators of arginine methylation mediated by Hmt1 (HnRNP methyltransferase 1), ortholog of human PRMT1. We developed an SDL array of 4548 yeast strains in which each strain contained a single non-essential gene deletion, in combination with a galactose-inducible construct overexpressing wild-type (WT) Hmt1-HZ tagged protein. We identified 129 consistent SDL interactions for WT Hmt1-HZ which represented genes whose deletion displayed significant growth reduction when combined with WT Hmt1 overexpression. To identify among the SDL interactions those that were dependent on the methyltransferase activity of Hmt1, SDL screens were repeated using an array overexpressing a catalytically inactive Hmt1(G68R)-HZ protein. Furthermore, an additional SDL control screen was performed using an array overexpressing only the protein tag HZ (His 6 -HA-ZZ) to eliminate false-positive SDL interactions. This analysis has led to a dataset of 50 high-confidence SDL interactions of WT Hmt1 which enrich eight Gene Ontology biological process terms. This dataset can be further exploited in biochemical and functional studies to illuminate which of the SDL inter-actors of Hmt1 correspond to factors implicated in the regulation of Hmt1-mediated arginine methylation and reveal the underlying molecular mechanisms.

Value of the Data
• The data reported will be useful for identifying putative novel regulators of protein arginine methylation, substrates of arginine methyltransferases and biological processes that are affected by arginine methylation. • The generated dataset is helpful for advancing the work of researchers who are investigating protein function and specifically the biological role of protein post-translational modifications. • These data will direct future studies on the biochemical and molecular mechanism of arginine methylation. • The strategy for generating the SDL interaction data of Hmt1 can be extended to investigate other post-translational modifications and their associated enzymes. • These data describe new collections of mutant yeast strains that could be useful to researchers using S. cerevisiae as a model system.

Data
Raw SDL data from eight replicate screens for WT Hmt1-HZ and eight replicate screens for Hmt1(G68R)-HZ and two replicate screens for HZ-tag are shown in Table S1. For each yeast strain, which was represented by quadruplicate colonies on the arrays, we generated the average growth and standard deviation (StdDev) in galactose (inducing) and glucose (non-inducing) conditions, as well as the calculated SDL score (growth in galactose/glucose) (Table S1). In total, 632 SDL interactions were generated from the eight replicate screens of WT Hmt1-HZ (Table  S2), 622 SDL interactions from the eight replicate screens of Hmt1(G68R)-HZ (Table S3) and 53 SDL interactions from the two replicate screens of HZ-tag (Table S4). To eliminate random SDL interactions, we then selected those that appeared in at least two screen replicates resulting in 129 SDL interactions for WT Hmt1-HZ, 101 for Hmt1(G68R)-HZ and 15 for the HZ-tag ( Table 1 ). Next, we filtered the SDL interactions of WT Hmt1-HZ by removing all interactions also found for Hmt1(G68R)-HZ and HZ-tag, in order to eliminate those that are independent of the Hmt1 methyltransferase activity and false-positive interactions respectively ( Fig. 1 ). This filtering analysis resulted in 50 high-confidence SDL interactions of WT Hmt1-HZ that represent putative arginine methylation regulators and are listed along with their Gene Ontology (GO) Annotated Term in Table 2 . GO enrichment analysis of the 50 filtered SDL interactions (Table S5) identified eight significantly enriched biological process terms, as shown in Table 3 . Notably, Hmt1 (or its human ortholog PRMT1) has already been associated with some of the enriched biological processes like cytosolic transport, metabolism, and RNA processing [5][6][7][8][9] .

SDL rationale for this study
The rationale supporting the performed SDL screens is based on the observation that the phenotype resulting from an overexpressing protein is exacerbated by deleting a second gene  Subunit of the membrane-associated retromer complex; essential for endosome-to-Golgi retrograde protein transport; peripheral membrane protein that assembles onto the membrane with Vps5p to promote vesicle formation; required for recruiting the retromer complex to the endosome membranes     [10] . Particularly, overexpressing Hmt1 alone does not lead to lethality or growth reduction compared to endogenous basal expression ( Fig. 2 A) because arginine methylation regulators (i.e. demethylase, PRTM-regulators, modification crosstalk) are still intact to limit the levels of arginine methylation on Hmt1 substrates ( Fig. 2 B). In contrast, overexpression of Hmt1 in combination with the deletion of an arginine methylation regulator will result in toxic hypermethylation of substrates, which cannot be tolerated by cells and therefore causes a growth defect ( Fig. 2 C). Hence, the arising objective was to screen the entire yeast genome for genes whose deletion leads to severe growth defect when the predominant arginine methyltransferase Hmt1 is overexpressed.

Yeast strains and plasmid construction
For the construction of inducible vectors, a multicopy MORF (moveable ORF) plasmid was used, which contained a URA3 selectable marker, a HIS 6 -HA-ZZ (HZ) tag and PGAL, a galactose inducible promoter of the galactokinase gene GAL1, that overexpressed either HMT1 gene with the complete wild type sequence (Hmt1-WT) or hmt1 gene with the catalytic mutant sequence Fig. 1. Comparison of SDL interactions of wild-type Hmt1 and catalytically inactive Hmt1(G68R). Venn diagram indicating the total SDL interactions generated from the WT Hmt1-HZ and Hmt1(G68R)-HZ screens that appeared in at least two screen replicates ( Table 1 ). This has led to 50 SDL interactions that are dependent on the methyltransferase activity of Hmt1. The Venn diagram was generated using the online tool Venny 2.1 ( https://bioinfogp.cnb.csic.es/tools/venny/ ).

Table 3
Significantly enriched Gene Ontology (GO) biological process terms represented within the Hmt1 SDL interactions. . The deleted genes were replaced with the antibiotic marker KanMX4, which confers resistance to Geneticin (G418).

SDL screen data collection
An outline of the SDL procedure performed during this study is indicated in Fig. 4 . The initial DMA library, which consists of 14 array plates was replicated by the BM3-BC automated pinning robot on solid (2% agar) rich medium plates with 2% glucose (YPAD) and the addition of G418. The plates were incubated at 30 °C for 2 days. Four lawn plates of the control strain and four of each query strain were prepared. Each strain was incubated overnight in 5 ml Synthetic Complete liquid medium without uracil (SC-URA) and with 2% glucose at 30 °C. The next day, 800 μl of culture were spread on SC-URA lawn plates and the plates were incubated at room temperature (RT) for 3 days. For the mating process, the query strain was pinned from the lawn plates onto fresh YPAD mating plates, followed by pinning of the DMA library on top of the query cells ( Fig. 4 ). The plates were incubated for 3 days in RT. Selection of diploid cells first involved pinning from the mating plates onto the diploid plates (SC-URA), followed by 2 days incubation at 30 °C. Then, cells from diploid plates were pinned onto SC plates with the addition of G418 and incubated for 2 days at 30 °C. After diploid selection, cells were pinned on enriched solid sporulation medium (reduced carbon and nitrogen levels) and incubated for 7 days in RT for the induction of sporulation and the formation of haploid meiotic spore progeny ( Fig. 4 ). The formed spores were next pinned onto solid Synthetic Defined medium without histidine, arginine, lysine and uracil and with the addition of canavanine and thialysine (SD -His/Arg/Lys/URA + canavanine/thialysine), for the selection of the MATa haploid meiotic progeny. Plates were incubated for 2 days at 30 °C, and the haploid selection step was repeated twice. As a prefinal step, cells were pinned on SD -His/Arg/Lys/URA + canavanine/thialysine plates with the addition of G418, for selection of the meiotic progeny that carries the gene deletion mutation ( genex ::kanMX) and incubated for 2 days at 30 °C. SDL interactions were selected by pinning the constructed cells onto SC plates, either containing 2% glucose or raffinose as non-inductive conditions, or 2% galactose as inductive conditions ( Fig. 4 ). The SDL interactions haploid selection step was repeated 8 times with the query strain carrying the plasmid overexpressing WT Hmt1-HZ and 8 times with the query strain carrying the plasmid overexpressing Hmt1(G68R)-HZ ( Table 1 ). It was additionally assessed 2 times with the query strain carrying the empty vector expressing the HZ tag only, as a control ( Table 1 ). Each plate was digitally photographed and the relative growth of individual colonies was processed with RobosoftPro application, that

Fig. 2. Rationale of synthetic dosage lethality (SDL) for the identification of putative arginine methylation regulators. (A)
The presence of physiological levels of Hmt1 arginine methyltransferase and arginine methylation regulator (R) leads to normal growth. This putative methylation regulator could control the activity of the Hmt1 enzyme or methylation of its substrate(s) or any other step during the methylation/demethylation process. (B) Overexpression of Hmt1 arginine methyltransferase does not lead to a severe phenotype as its activity is antagonised by an intact regulator (demethylase, PRMT-regulator, modification crosstalk). General levels of arginine methylation on substrates (S) are maintained at physiological levels or increased at a tolerable level for the cells. (C) Combination of Hmt1 methyltransferase overexpression with deletion or inactivation of an arginine methylation regulator (R) results in reduced growth due to toxic hypermethylation of substrate(s). At least 30% reduction in growth would indicate an SDL interaction between Hmt1 and the putative regulator. quantifies the pixels of any given colony [3] . Comparison between average growth of the technical quadruplicate colonies representing each strain in non-inducing conditions, with the corresponding four colonies of the strain in inducing conditions was generated as standard deviation (StdDev) by the application (Table S1). An SDL score indicating the overall growth observed was calculated by the equation (WT Hmt1-HZ Galactose/Glucose) (Table S1).

Data analysis and filtering
The raw data of SDL screens were subjected to initial analysis by following three steps: 1) Removal of YOR202W gene values which were located in the outermost two lines of each plate. These colonies represent a border control in order to avoid irregular growth of strains positioned around the edges of the plate and thus serve as an internal quality control for each screen. 2) Removal of strains with < = 60% growth on glucose compared to the average growth of all the other strains on the plate, in order to exclude strains exhibiting severe growth defect even in uninduced conditions. 3) Selection of hits with SDL score (log10 ≤ −0,15) ≤ 70%, meaning that they have at least 30% growth defect (threshold set by [3] ) when grown on galactose compared to glucose. The SDL interactions were further filtered by executing six rigorous selection steps: 1) removal of SDL interactions giving SDL scores > 70% in other screen replicates, 2) removal Overexpression of WT Hmt1-HZ in induced conditions results in increased methylated levels of its known substrate NpI3. Upon overexpression of Hmt1(G68R)-HZ, the methylation levels of NpI3 are unchanged between uninduced and induced conditions. The detected methylation of Npl3 in the Hmt1(G68R)-HZ strains is mediated by the endogenous wild-type Hmt1 enzyme. The levels of overexpressed Hmt1 were detected using an anti-His antibody recognizing the HZ tag and of its substrate Npl3 was detected using an anti-Npl3 methylated antibody. The loading of protein extracts was monitored by an anti-histone H3 antibody.
of SDL interactions generated only once among all screen replicates, 3) removal of all galactoseinduced genes, 4) removal of genes which displayed significant growth changes when comparing all screens under non-inducible (glucose) conditions, 5) removal of 15 SDL interactions generated by the HZ-tag screen, and 6) removal of 102 SDL interactions generated by the Hmt1(G68R)-HZ screen.

Gene ontology (GO) enrichment analysis
The filtered genes ( Table 2 ) corresponding to the high-confidence Hmt1 SDL interactions identified through the screens were clustered into their annotated GO biological processes according to the Saccharomyces Genome Database (SGD) GO Slim Mapper ( https://www. yeastgenome.org/goSlimMapper ) ( Table S5). The Fisher's exact test was used to classify statistically significant enriched GO terms ( p -value ≤ 0,05) ( Table 3 ).

Declaration of Competing Interest
DK was employed by company EFEVRE TECH LTD. All authors declare that they have no known competing financial interests or personal relationships which have, or could be perceived to have, influenced or have been influenced by the work reported in this article.  After several steps of replica-pinning and selection, an array of haploid cells was isolated that contained the overexpressing construct in combination with a specific gene deletion. The array was then grown on un-induced (glucose or raffinose) and induced (galactose) conditions to identify cells that would show reduced growth phenotype when Hmt1 is overexpressed in a specific gene deletion background. ment Fund and the Republic of Cyprus through the Research & Innovation Foundation (Project: EXCELLENCE/1216/0215).

Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.dib.2020.105885 .