Insights into Transcriptional Regulation by Individual H3K4 Methylation Marks in S. Cerevisiae

Set1 is a lysine methyltransferase in S. cerevisiae that catalyzes the mono, di and tri methylation of the fourth lysine on the amino terminal tail of histone H3 (H3K4). Set1-like methyltransferases are evolutionarily conserved, and research has linked their function to developmental gene regulation and several cancers in higher eukaryotes. Set1 is a member of the multiprotein COMPASS complex in S. cerevisiae. The H3K4 methylation activity of COMPASS regulates gene expression and chromosome segregation in vivo. The three distinct methyl marks on histone H3K4 act in discrete ways to regulate transcription. Trimethylation of H3K4 is usually associated with active transcription whereas dimethylation of H3K4 is associated with gene repression. In this study, amino acid substitution mutants of SET1 that encode partial function Set1 proteins capable of H3K4me1, H3K4me1 and H3K4me2, or H3K4me1and H3K4me3 were analyzed to learn more about the roles of individual H3K4 methyl marks in transcription. The ndings reveal a previously unappreciated role for H3K4me1 in activation of transcription of the HIS3 gene in S. cerevisiae cultures grown under histidine-starvation conditions. Surprisingly, induction of the HIS3 gene in cultures grown under histidine starvation is not accompanied by signicant changes in the proles of H3K4-methylated nucleosomes at the HIS3 gene in SET1 wild-type strains and set1 partial-function mutants. The data show that H3K4me1 supports induction of HIS3 mRNA to wild-type levels under histidine-starvation conditions and that higher-order H3K4 methylation (H3K4me2 and H3K4me3) is not required.


Introduction
The structural organization of eukaryotic DNA into chromatin regulates transcription by RNA polymerases (Han and Grunstein 1988, Izban and Luse 1992, Kornberg and Thonmas 1974, Wasylyk and Chambon 1979, Workman and Kingston 1998. Nucleosomes, the fundamental structural subunits of eukaryotic chromatin, consist of 147 bp of double-stranded DNA wound around a histone core octamer (Lawrence, et al. 2016, Luger, et al. 1997, Venkatesh and Workman 2015. The histone core octamer has two copies of each of the histones, H2A, H2B, H3 and H4. To overcome the physical barrier imposed by chromatin, RNA polymerases rely on the action of trans-acting proteins and protein complexes, including chromatin remodelers, transcription factors, co-activators and histone-modifying enzymes make chromatin accessible to RNA polymerase II to access DNA sequences (Castillo, et al. 2017, Chatterjee, et al. 2011, Côté, et al. 1998, Lee, et al. 1993, Lee, et al. 2007, Santos-Rosa, et al. 2003).
Post-translationally modi ed histones regulate chromatin characteristics and functions, including accessibility, compaction, nucleosome dynamics, replication, and transcription (Chatterjee, et al. 2011, Lawrence, et al. 2016, Santos-Rosa, et al. 2003, Soriano, et al. 2014, Zhang, et al. 2015. Histones are targets for many covalent modi cations including methylation, acetylation, phosphorylation and ubiquitylation (Allfrey, et al. 1964, Zhang, et al. 2015. Covalently modi ed histones in chromatin may recruit or facilitate the interaction of effectors, such as chromatin-remodeling complexes and transcription factors, with chromatin (Castillo, et al. 2017, Lawrence, et al. 2016. For example, acetylated lysines are recognized by bromodomain-containing proteins ( (Dhalluin, et al. 1999); reviewed in (Filippakopoulos and Knapp 2014)), like the Swi2/Snf2 subunit in the SWI/SNF chromatin-remodeling complex, which aids in the generation of open and accessible chromatin ( (Tamkun, et al. 1992), reviewed in (Swygert and Peterson 2014)). The dynamic exchange of acetyl groups onto and off of histones acts as a switch between active and repressive chromatin, respectively (reviewed in (Eberharter andBecker 2002, Kuo and). In contrast to histone acetylation, the outcome of histone methylation is context-dependent; methylated histones have the capacity to activate or repress transcription (Hyun, et al. 2017).
The effect of H3K4 methylation on chromatin accessibility depends on the chromatin context, consistent with data showing that H3K4 mono-, di-, and trimethylation may have different effects on gene transcription (Kusch 2012, Pokholok, et al. 2005. Chromatin immunoprecipitation (ChIP) and ChIP-seq experiments revealed that the distribution of K4-monomethylated (H3K4me1), K4-dimethylated (H3K4me2) and K4-trimethylated (H3K4me3) histone H3 across open reading frames (ORFs) is not identical (Bernstein, et al. 2005, Soares, et al. 2017. For the most highly expressed genes in S. cerevisiae, nucleosomes with H3K4me3 peak at the promoter and up to ~200 bp beyond the transcription start site, nucleosomes with H3K4me2 are enriched in the middle of the ORF, and nucleosomes with H3K4me1 are found predominantly at the 3' end of an ORF (Berger 2007, Pokholok, et al. 2005. Methylation of histones is not known to change the structure of chromatin on its own, instead methylated histones act by recruiting effector proteins to chromatin (Musselman, et al. 2012, Pray-Grant, et al. 2005, Taverna, et al. 2006).
H3K4-methylated histones may be recognized by chromodomain-containing proteins (Eissenberg 2012) and plant homeodomain (PHD) nger domains, including the COMPASS member Spp1 (He, et al. 2019). H3K4me3 is usually associated with active transcription (Kusch 2012, Ng, et al. 2003, Schneider, et al. 2005. H3K4me3 is recognized by the chromodomain-containing protein, Chd1, a member of the SAGA transcription coactivator complex (Pray-Grant, et al. 2005). The transcription factor TAF3 also interacts with trimethylated H3K4 to recruit TF II D to gene promoters (Vermeulen, et al. 2007). In yeast, the chromatin remodeler Isw1 interacts with methylated H3K4 to generate open, accessible chromatin at the 5' end of the MET16 gene (Santos-Rosa, et al. 2003). H3K4me3 is also recognized by other protein complexes, some of which are negative effectors of transcription (Musselman, et al. 2012, Taverna, et al. 2006). In response to DNA damage, the PHD domain of ING2 (INhibitor of Growth 2), a subunit of the mSin3a-HDAC1 histone deacetylase complex, binds to H3K4me3, stabilizing mSin3a-HDAC1 at the promoter of cyclin D1 gene and other proliferation genes, leading to repression of transcription (Shi, et al. 2006). H3K4me2 is associated with repression of transcription and has been shown to interact with HDACs that reduce acetylated histones at the 5′ ends of some highly expressed genes (Kim and Buratowski 2009, Pinskaya and Morillon 2009).
Progress is being made toward understanding the roles H3K4me1 plays in gene regulation. In yeast, H3K4me1 regulates chromatin remodeling at osmostress-induced genes (Nadal-Ribelles, et al.

2015)
. In mammalian cells, there is compelling evidence that H3K4me1 promotes interactions between enhancers and promoters by facilitating the binding of chromatin remodelers , Yan, et al. 2018). However, a catalytically defective H3K4 HMTase also facilitates enhancer-promoter interactions meaning that an additional mechanism that doesn't require H3K4me1 is likely to promote enhancer-mediated effects as well (Dorighi, et al. 2017, Rickels, et al. 2017. Recent work has shown that the patterns of H3K4me1 with H3K4me3 and H3K27me3 at promoters in human and mouse germ cells and ESCs may predict the transcriptional state of a promoter (Bae and Lesch 2020).
Set1 is the only H3K4 HMTase in S. cerevisiae making it an excellent system to study the effect of the three H3K4 methyl marks on transcription. In a previous study, mutants of SET1 were made that encode proteins with amino acid substitutions in the SET domain of Set1 ( Figure 1) (Williamson, et al. 2013). Wild-type Set1 generates H3K4me1, H3K4me2 and H3K4me3. Amino acid substitution mutants were constructed to alter residues near and in the active site of Set1 to generate mutants with different H3K4 methylation capabilities. The set1-Y967A alleles is null mutant with no detectable H3K4 methylation that is indistinguishable from a set1D mutant. The set1-G951A mutant produces H3K4me1 and very low levels of H3K4me2. Although G951 is not in the active site, it is highly conserved in Set1 homologs and has been shown to be important for Set1 function (Dillon, et al. 2005, Nislow, et al. 1997, Sollier, et al. 2004). Two additional partial function mutants were generated: set1-Y967F that generates H3K4me1 and low levels of H3K4me3 and set1-R1013H that generates H3K4me1 and H3K4me2 (Figure 1).
In this study, the set1 amino-acid substitution mutants and the HIS3 gene were studied to learn about the roles of individual H3K4 methyl marks in RNA polymerase II transcription in S. cerevisiae. HIS3 codes for imidazole glycerol phosphate dehydratase (Fink 1964), the enzyme that catalyzes the sixth step in the biosynthesis of histidine in S. cerevisiae. The herbicide 3-Amino-1,2,4-triazole (3AT) is a competitive inhibitor of the HIS3 gene product and is used to induce histidine-starvation leading to the production of the general amino acid control regulator, Gcn4 (Brennan andStruhl 1980, Hope andStruhl 1985). Gcn4 activates transcription of the HIS3 gene, increasing the level of HIS3 mRNA and biosynthesis of histidine (Hill, et al. 1986).
Our work shows that in the absence of H3K4me2 and H3K4me3, H3K4me1 promotes transcription of the HIS3 gene under histidine-starvation conditions in S. cerevisiae. A role of H3K4me1 in activation of transcription has not been reported previously. In addition, we also demonstrate that one or more of the genes required for biosynthesis of isoleucine and valine  is activated by H3K4me1 when H3K4me3 and H3K4me2 are absent. The results indicate a previously unrecognized role for H3K4me1 in the activation of transcription of the HIS3 gene and one or more of the ILV genes by RNA polymerase II.

Media
Standard media preparation protocols were used (Rose, et al. 1990). YPADTU is YPD media supplemented with 40 mg/L adenine hemisulfate, 80 mg/L L-tryptophan and 20 mg/L uracil. Where indicated, 1 M 3-Amino-1,2,4-triazole (3AT) made in sterile H 2 O was added to media to a nal concentration of 10 mM. Sulfometuron Methyl (SMM), made in dimethyl sulfoxide (DMSO), was added to media to a nal concentration of 1 μg/mL. SC Complete is de ned synthetic medium containing all nutrients required for yeast cell growth. SC-His, SC-Ile Val and SC-Trp are synthetic complete yeast growth media lacking histidine, isoleucine and valine, or tryptophan, respectively.

Yeast Strains
Strains of S. cerevisiae used in this study are listed in Supplementary Table 1. Yeast strains were made by standard genetic crosses and genetic transformation. The generation and initial characterization of mutant alleles of SET1 were described previously (Williamson, et al. 2013). Plasmids containing HHT2-HHF2 and hht2-K4R-HHF2 have been described previously (Briggs, et al. 2001). For this study, the genes encoding HHT2 and HHF2 or hht2-K4R-HHF2 were cloned into pRS414 plasmids (Sikorski and Hieter 1989) that carry a TRP1 selectable marker. The endogenous S. cerevisiae HHT1-HHF1 and HHT2-HHF2 genes encoding the histones H3 and H4 were deleted and replaced with selectable marker genes. At least one copy of the genes encoding H3 and H4 is required for viability. Therefore, the cells expressed either the wild-type HHT2-HHF2 genes or a mutant version of HHT2 (hht2-H3K4R) with a wild-type HHF2 gene from a plasmid. The mutant hht2-K4R gene encodes histone H3-K4R, an amino acid substitution variant of histone H3 that cannot be methylated by Set1 because the lysine at position 4 is replaced with an arginine residue. Plasmids were transformed into cells from 1 ml saturated culture, as previously described except that 40 µg of denatured salmon sperm carrier DNA was used (Chen, et al. 1992). Cells were plated on SC-Trp solid agar and incubated at 30 o C to obtain transformants.

Growth Assays
Cultures were grown to saturation at 30 o C in SC-His or YPADTU liquid medium. Eight, ve-fold serial dilutions were made in sterile water. The last six dilutions were plated (5 µl) on each of four types of solid agar plates: SC-His, SC-His + 10 mM 3AT, SC-complete, and SC-complete + 10 mM 3AT. Plates were imaged after 24 and 42-44 hours of incubation at 30 o C. For experiments with yeast strains expressing histone H3K4 or H3K4R from plasmids, six 10-fold serial dilutions were made in sterile water and dilutions were plated, as described above. Plates were incubated at 30 o C and imaged after 42 hours. For Ile Val starvation growth assays, eight, ve-fold serial dilutions were made in sterile water using overnight cultures grown in YPADTU. The last six dilutions were plated (5 µl) on solid agar plates: SC-Ile Val + DMSO, SC-Ile Val + 1 μg/mL SMM, SC-complete + DMSO, and SC-complete + 1 μg/mL SMM. Plates were imaged after 72-96 hours of incubation at 30 o C.

RNA Isolation and Northern Blotting
Saturated cultures of yeast strains grown in SC-His liquid medium were diluted into 50 mL of fresh SC-His medium at ~4x10 6 cells/mL and grown to a density of 1-2x10 7 cells/mL at 30 o C in a shaker incubator. After 4 hours, 500 μL of 1 M 3AT ( nal concentration 10 mM) or sterile distilled water were added to the cultures, which were incubated at 30 o C in a shaker incubator for one hour. Total RNA was extracted from yeast cultures, as previously described (Schmitt, et al. 1990) with the following modi cations. Cells were resuspended in 450 µl AE buffer (50 mM C 2 H 3 NaO 2 pH 5.3, 10 mM EDTA), and then transferred to a microfuge tube to which 50 µl 10% SDS was added. Extraction with an equal volume of chloroform:isoamyl alcohol (24:1) was performed before the addition of 50 µl 3M sodium acetate pH 5.3 and 2.5 volumes 100% ethanol. After precipitation, RNA was resuspended in sterile milliQ H 2 O and stored at -70 o C. RNA (15 µg) was analyzed by Northern blotting, as described previously (Swanson, et al. 1991).
The steady-state level of HIS3 transcript was detected by hybridization with a strand-speci c 32 P-labeled riboprobe. To normalize loading of RNA samples, a 32 P-labelled ACT1 DNA probe synthesized by random priming was used to detect the ACT1 mRNA level. For the northern blots analyzing the set1-G951A strain, a 32 P-labelled HIS3 DNA probe synthesized by random priming was used to detect the HIS3 mRNA level. To analyze the Northern blots, the ratio of HIS3/ACT1 mRNA was calculated for each strain grown in SC-His and SC-His + 10 mM 3AT. To calculate the fold change of HIS3 transcript in presence of 3AT, the ratio of HIS3/ACT1 mRNA for each strain grown in SC-His+10mM 3AT was normalized to the corresponding HIS3/ACT1 transcript levels in cultures grown in SC-His.
The optimal induction of HIS3 transcript after addition of 3AT was determined using a time course experiment (Supplementary Figure 2). The steady-state level of HIS3 transcript was detected as described above. To normalize loading of RNA samples, a 32 P-labelled 18S ribosomal RNA riboprobe was used to detect the rRNA level. To analyze the Northern blots, the ratio of HIS3 mRNA/18S rRNA was calculated for each strain at each time point. Then, the ratio of HIS3 mRNA/18S rRNA at 5, 10, 30, 60, and 120 min post addition of 10 mM 3AT was normalized to the corresponding ratio at 0 min (prior to adding 3AT). All blots were imaged using a G.E Typhoon FLA 7000 and quanti ed using G.E Imagequant TL 8.1 software.

Whole Cell Protein Extracts and Western Blotting
Yeast whole cell protein extracts were prepared as described (Mueller, et al. 2006). Proteins from clari ed whole-cell extracts (8, 20 or 40 μg) were separated on 10% SDS-polyacrylamide gels, transferred to PVDF membrane, and probed with α-histone H3 (ab1791, Abcam; 1:1000), α-K4-monomethyl H3 (13- . IPs were processed as described previously (Bryk, et al. 2002), with the following exceptions. Pierce protein A/G agarose beads were used to pull down crosslinked protein-DNA complexes (Pierce Biotechnology, Thermo sher, IL, USA). ChIP eluates were puri ed using the ChIP DNA clean and concentrator kit, as recommended by the manufacturer (Zymo Research Corp, CA, USA). Puri ed extracts eluted in 100 µl elution buffer were stored at -70 o C.
The effects of alterations in the levels of H3K4me1/2/3 on the growth of yeast cultures under histidine starvation conditions were evaluated with saturated yeast cultures expressing either wild-type or aminoacid substitution variants of Set1 on four different solid agar media (Figure 2). After 42 hours of incubation, the six cultures grew similarly on SC Complete medium with the herbicide 3-Amino-1,2,4triazole (3AT), SC Complete medium without 3AT, and synthetic medium lacking histidine, SC-His ( Figure  2A and 2B right panel). In contrast, all cultures grew less well on SC-His+3AT ( Figure 2B, left panel). This result was expected because 3AT is a competitive inhibitor of the enzyme encoded by the HIS3 gene and causes histidine starvation in medium lacking histidine.
Reduced growth of yeast strains expressing the null mutants (set1D and set1-Y967A) was observed on SC-His+3AT ( Figure 2B, left panel). The extent of growth of yeast strains expressing partially or fully functional Set1 proteins (set1-G951A, set1-Y967F, set1-R1013H, or SET1 + ) was similar, and all grew better than the null mutants. Given that the set1-G951A mutant catalyzes H3K4me1 mainly with very low H3K4me2 and undetectable H3K4me3 ( Figure 1B and C), the results suggest that H3K4me2 and H3K4me3 are not required for wild-type growth under histidine-starvation conditions. A literature search was performed to identify other genes regulated by the GAAC pathway that might be also regulated by H3K4 methylation. The ILV genes, ILV1, ILV2, ILV3, ILV5 and ILV6, encode enzymes that catalyze the biosynthesis of isoleucine and valine in S. cerevisiae. Sulfometuron methyl (SMM) is a competitive inhibitor of the ILV2 gene, acetolactate synthase . SMM was used in plate growth assays to evaluate the role of H3K4 methylation on growth of cultures during starvation for Ile and Val. Saturated cultures of yeast strains expressing wild-type Set1 or set1 aminoacid substitution variants were analyzed for growth on medium with or without SMM. All strains tested grew equally well on SC Complete media + DMSO ( Figure 3A, right). Likewise, differences in growth were not observed on SC Complete containing SMM, indicating that SMM in de ned complete medium does not cause a growth defect ( Figure 3A, left). Equal growth of strains was also observed on synthetic media lacking isoleucine and valine (SC-Ile Val + DMSO, Figure 3B, right). However, the set1 null mutants grew poorly on medium lacking isoleucine and valine in presence of SMM (SC-Ile Val + 1 μg/mL SMM) when compared to strains expressing wild-type SET1 + or the partial function alleles of set1 ( Figure 3B, left). Similar to the results of growth assays performed under histidine-starvation conditions (Figure 2), the partial function mutant set1-G951A, which performs mainly H3K4me1 was su cient to rescue the growth defect exhibited by the set1 null mutants. These results suggest that H3K4me2 and H3K4me3 are not required for wild-type growth under conditions of isoleucine and valine starvation.
In addition to methylation of histone H3, Set1 methylates the Dam1 protein that functions in chromosome segregation in S. cerevisiae (Latham, et al. 2011. To verify that the growth phenotypes observed in Figure 2 were due to changes in methylation of histone H3 and not another target of Set1, SET1 + and set1 mutant alleles were transformed into yeast strains that express either wild-type histone H3 or a mutant version of H3 (hht2-K4R) that cannot be methylated by Set1 due to the substitution of lysine at position 4 with arginine ( Figure S2). The set1 null mutants that express wild-type histone H3 (pH3K4) grow less well than the other yeast strains on SC-His+3AT ( Figure S2C). In contrast, growth of cultures expressing SET1 + or the partial function alleles and the unmethylatable form of histone H3 (pH3K4R) was indistinguishable from that of the null mutants on SC-His+3AT agar ( Figure   S2B). These results indicate that methylation of histone H3 by Set1 is required for robust growth of yeast cultures during histidine starvation. We next explored the role of H3K4 methylation in expression of HIS3 by measuring steady-state levels of HIS3 mRNA and the presence of H3K4me methyl marks at the HIS3 gene in wild-type SET1 strains, set1 null mutants and set1 partial function mutants.
Histone H3K4 methylation by Set1 is required for wild-type expression of the HIS3 gene in cultures grown in nutrient-stress conditions The level of HIS3 mRNA in cultures expressing wild-type SET1 + or the set1 amino-acid substitution variants was measured by Northern hybridization (Figure 4). Histidine starvation was induced by adding 10 mM 3AT to log-phase cultures for 1 hr prior to isolation of RNA. The HIS3:ACT1 level was increased 13.7-fold in SET1 + cultures when histidine starvation was induced. The HIS3:ACT1 levels were increased in the null mutants (set1Δ and set1-Y967A), yet the levels were signi cantly lower than those in the SET1 + cultures. HIS3:ACT1 levels were upregulated in the set1-G951A, set1-Y967F and set1-R1013H mutants and were not signi cantly different from those in the SET1 + cultures. From these results, we conclude that set1-G951A, an allele mainly capable of generating H3K4me1, supports the activation of transcription of the HIS3 gene under histidine-starvation conditions. Therefore, the growth defects observed for the set1 null mutants (set1Δ and set1-Y967A, Figure 2B) may be explained by the fact that these mutants make less HIS3 mRNA than the SET1 + strain and the partial function set1 mutants under histidine-starvation conditions. However, we note that even though the average HIS3:ACT1 level in the set1-R1013H mutant under histidine-starvation conditions was higher than that in the set1 null mutants, the level was not statistically different from the SET1 + or set1 null cultures. This observation will be addressed further in the discussion.
Histone H3K4 methylation patterns at the HIS3 gene do not change signi cantly during histidine starvation To evaluate the presence of H3K4 methyl marks at the HIS3 promoter during histidine starvation, ChIPs were performed to measure H3K4me1, H3K4me2 and H3K4me3 at the HIS3 promoter and two positions in the HIS3 ORF ( Figure 5A and B). H3K4me1/2/3 marks were detected at the HIS3 gene in yeast strains expressing wild-type SET1 and the partial function mutants, set1-G951A, set1-Y967F and set1-R1013H in the absence and presence of 3AT ( Figure 5C and D). The distribution of H3K4 methyl marks was also evaluated at the promoter of the ACT1 gene and an intergenic region on chromosome VIII (Supplemental Figure 3). One surprising result was that for the most part the levels of H3K4me1, H3K4me2 and H3K4me3 associated with the promoter and 5' ORF of the HIS3 gene in wild-type SET1 + cultures did not change in cultures grown in presence of 3AT, conditions that we show in Figure 4 cause an increase in steady-state HIS3 mRNA.
H3K4me3: Wild-type Set1 is able to convert lower-level H3K4 methyl marks to H3K4me3. Consistent with this, H3K4me3 signals at the HIS3 gene and other genomic regions analyzed by ChIP were signi cantly higher in the SET1 + cultures compared to the partial function set1 mutant cultures (Figures 5C, 5D, and S3, blue bars). H3K4me3 was detected above background (set1D mutant) at the HIS3 gene and other genomic regions in the set1-Y967F mutant, although the levels were signi cantly lower than those in the SET1 cultures. Consistent with previously published data (Pokholok, et al. 2005), the average level of H3K4me3 was higher at the HIS3 promoter and 5' ORF region than the HIS3 3' ORF, in the SET1 + and set1-Y967F samples. H3K4me3 was detected at background levels at the three regions of HIS3 gene, the ACT1 promoter or the intergenic region in the set1-G951A and set1-R1013H mutant samples with or without 3AT-treatment.
H3K4me2: The levels of H3K4me2 at the HIS3 gene were low in the SET1, set1-G951A, set1-Y967F and the set1-R1013H cultures when compared to each strain's most prominent H3K4 methyl mark. With the exception of the SET1 and set1-R1013H cultures, the levels of H3K4me2 measured at the HIS3 promoter, 5' ORF and 3' ORF were at or near background levels in the set1-G951A and the set1-Y967F cultures grown without 3AT. A slight but signi cant increase in the level of H3K4me2 relative to background was observed at the HIS3 promoter in the set1-G951A culture grown in the presence of 3AT. At other positions, the level of H3K4me2 was at background levels in the set1-G951A mutant.
The set1-R1013H mutant can generate detectable levels of H3K4me2 ( Figure 1B), and in the ChIP data, the levels of H3K4me2 in the set1-R1013H mutant were similar to those in the SET1 cultures. In contrast, at all positions of the HIS3 gene that were evaluated with or without 3AT treatment, the levels of H3K4me2 in the set1-Y967F mutant remained at background levels. The set1-Y967F mutant is unusual in that it generates H3K4me1 and H3K4me3, but H3K4me2 is not detected above background in whole cell extracts or at positions in HIS3, ACT1 or the intergenic region (Figures 1, 5 and S3). One possibility to explain the phenotype of the set1-Y967F mutant is that the substitution of phenylalanine for tyrosine at position 967 of Set1 alters the active site in a manner that allows the H3 tail to be monomethylated at K4 and possibly released, but when H3K4me1 is the substrate of COMPASS, two additional methylation events occur without the release of H3K4me2 resulting in the production of H3K4me3.
H3K4me1: In the SET1 + cultures, the level of H3K4me1 was lower than that of H3K4me3 at the three positions analyzed in the HIS3 gene in the absence and presence of 3AT ( Figure 5). In contrast, in the set1-G951A strain, H3K4me1 was the predominant form of K4-methylated H3 ( Figure 5). This is not a surprise because the set1-G951A mutant is not pro cient in generating H3K4me2 and H3K4me3. The set1-R1013H mutant is poor at generating H3K4me3. As such, in the set1-R1013H mutant, H3K4me1 and H3K4me2 are more abundant than H3K4me3 at the three positions analyzed in the HIS3 gene in untreated cultures ( Figure 5A). However, the H3K4 methylation pro le for the set1-R1013H mutant is different at the HIS3 promoter and 5' ORF in 3AT-treated cultures, where the levels of the three forms of H3K4me were similar ( Figure 5B). The pro les at the 3' ORF and ACT1 promoter are more like those in the untreated samples (Figures 5 and S3). In the set1-Y967F mutant, we observed slightly higher average levels of H3K4me1 than H3K4me3 at the three positions analyzed in the HIS3 gene in untreated cultures ( Figure 5A). This pro le switched when the set1-Y967F mutant was treated with 3AT such that the average level of H3K4me3 was higher than H3K4me1 at the HIS3 gene ( Figure 5B). A switch in the levels of H3K4me1 and H3K4me3 was not observed at the ACT1 promoter upon treatment with 3AT ( Figure S3).
The results show that H3K4me3 is present at background levels at the HIS3 gene in the set1-G951A mutant, yet robust growth and induced transcription of the HIS3 gene occurs during treatment of the set1-G951A mutant with 3AT. The phenotypes of the set1D mutant indicate that H3K4 methylation is required for robust growth and induction of HIS3 transcription during histidine starvation. Therefore, we conclude that H3K4me1 supports transcription of the HIS3 gene under histidine-starvation conditions in the absence of H3K4me2 and H3K4me3. The implications of these ndings with respect to transcription and the mechanisms of gene regulation are discussed below.

Discussion
Analysis of partial function variants of Set1 identi ed a role for H3K4 monomethylation in transcription by RNA polymerase II. Much of the existing literature on H3K4 methylation focuses on the role of H3K4me2 and H3K4me3 on transcription (Kusch 2012, Musselman, et al. 2012, Pray-Grant, et al. 2005, Schneider, et al. 2005, Taverna, et al. 2006). The effect of H3K4 monomethylation on transcription is not well characterized.
Previous studies in S. cerevisiae showed that H3K4me1 inhibits RSC-independent chromatin remodeling thereby preventing induction of osmostress genes (Nadal-Ribelles, et al. 2015). H3K4me1 has also been associated with transcriptional silencing in skeletal muscle cells, embryonic broblasts, macrophages, and human embryonic stem cells (Cheng, et al. 2014). In contrast to a repressive role in transcription, the ndings here show that H3K4me1 supports the activation of transcription by Pol II when S. cerevisiae cultures are subjected to nutrient-starvation conditions.
The three states of methylation at lysine 4 of histone H3 play different roles in the regulation of gene expression (Kusch 2012, Pokholok, et al. 2005. Here, mutants of SET1 were used to characterize the effects of different states of H3K4-methylation on expression of the HIS3 gene in S. cerevisiae. Growth assays, Northern blotting experiments and chromatin IPs (Figures 2-5) show that a partial function mutant of Set1, set1-G951A, which catalyzes H3K4me1 mainly, supports increase in the steady-state level of HIS3 mRNA and wild-type growth under histidine-starvation and isoleucine-valine starvation conditions. Wild-type growth under histidine-starvation and isoleucine-valine starvation conditions is also seen with a second partial function allele, set1-R1013H, that is capable of H3K4 mono-and dimethylation but not H3K4me3 (Figure 1). H3K4me3 is not required for induction of HIS3 transcript in set1-R1013H cultures grown under histidine-starvation condition. Both set1-G951A and set1-R1013H are incapable of accumulating H3K4me3 at the HIS3 gene ( Figure 5), therefore activation of transcription cannot be attributed to the accumulation of H3K4me3 at the HIS3 promoter. Also, because H3K4me2 is less than 5% of the wild-type level in the set1-G951A mutant and does not accumulate at the HIS3 gene under histidine-starvation condition, we conclude that H3K4me1, in the absence of H3K4me2 and H3K4me3, supports the increase in steady-state levels of HIS3 transcript under histidine-starvation conditions and promotes wild-type growth.
The growth defects exhibited by yeast strains expressing either of the two set1 null mutants, set1D or set1-Y967A, under histidine-starvation conditions were due to the lack of Set1 histone methylation activity. Comparison of data from null mutants to those from the set1-G951A mutant shows that Set1 monomethylation activity is required for the upregulation of HIS3 mRNA to wild-type levels under histidine-starvation conditions. It is possible that in addition to the HIS3 gene, one or more other genes in the histidine biosynthetic pathway is also upregulated in the presence of H3K4me1. Robust growth under isoleucine-valine starvation also occurred in the set1-G951A mutant but not the null mutants ( Figure 3). Therefore, the data suggest that induction of one or more of the ILV genes that catalyze the biosynthesis of Ile and Val (ILV1, ILV2, ILV3, ILV5, ILV6) occurs the presence of H3K4me1 in cells lacking H3K4me3. The speci c role of H3K4 methylation in upregulation of the ILV genes will be determined in the future.
ChIP analysis showed that the three forms of K4-methylated H3 are present at the HIS3 promoter and the 5' and 3' regions of the HIS3 ORF in wild-type SET1 + cultures. In SET1 + cultures, H3K4me3 was detected across the gene, with the highest signals at the promoter and 5' ORF ( Figure 5). This pro le was expected based on previous studies showing a similar H3K4me3 pro le at many genes in yeast (Berger 2007, Pokholok, et al. 2005. H3K4me2 levels at the positions analyzed in the HIS3 gene were lower than those of H3K4me3 and H3K4me1 ( Figure 5). In general, H3K4me1 levels at these regions were lower than those of H3K4me3 and slightly higher than those of H3K4me2.
Histidine-starvation caused a 13.7-fold increase in the level of HIS3 gene transcript (Figure 4) in SET1 + yeast cultures, yet increases in the levels of H3K4me3, H3K4me2 and H3K4me1 were not detected at the HIS3 promoter in cultures grown under histidine-starvation conditions ( Figure 5A and B). In fact, signi cant changes in H3K4me signals in the presence and absence of 3AT were not detected (except for H3K4me1 in the 5' ORF in set1-Y967F and set1-R1013H). These results indicate that while histidine starvation causes an increase in HIS3 mRNA expression (Hill, et al. 1986), it does not affect the H3K4 methylation at the positions of the HIS3 gene that were analyzed in this study. The data show that S. cerevisiae cells require Set1 and H3K4 methylation to elicit a wild-type transcriptional response to histidine starvation. However, higher-level H3K4 methylation is not required; H3K4me1 supports this transcriptional response even when present at levels similar to those observed under non-starvation conditions.
These results suggest that H3K4 methylation alone is unlikely to determine the transcriptional response to histidine starvation at the HIS3 gene. Changes in the levels of other histone marks, such as acetylation, in combination with H3K4 methylation, may be required to recruit transcription effectors responsible for induction of the HIS3 gene under histidine-starvation conditions. The HIS3 gene is regulated by the General Amino Acid Control (GAAC) pathway and requires Gcn4 for initiation of transcription. Growth of yeast cultures in 3AT activates Gcn4, which recruits the histone acetyltransferase (HAT) Gcn5 to many promoters, including the HIS3 promoter, leading to activation of transcription (Kuo, et al. 2000). In the future, we will determine if H3K4me1-mediated activation of transcription is accompanied by an increase in acetylation of histones at the promoters of target genes. Genome-wide studies using the partial function set1 mutants under normal and nutrientstress conditions will provide high-resolution data, potentially identify more genes regulated by H3K4me1, and may reveal interesting patterns not observed in single gene analysis.
A related goal is to identify transcription effectors that read H3K4me1 and contribute to the activation of transcription at HIS3 and other genes. Initiation of transcription requires the assembly of many transcription factors at gene promoters. Gcn4 recruits the SAGA complex (Spt-Ada-Gcn5 acetyltransferase) to gene promoters (Kuo, et al. (2000). Gcn4 also interacts with Gcn5, SWI/SNF, the SRB/Mediator complex, RNA polymerase II, TFIID, and Nua4 (Ginsburg, et al. 2009, Natarajan, et al. 1999, Swanson, et al. 2003. Identi cation of a protein that reads the H3K4me1 will help elucidate an H3K4me1-dependent mechanism of transcription activation. H3K4 methyltransferases are conserved across species from yeast to humans and have important roles in regulation of gene expression (reviewed in Shilatifard (2012)). The involvement of trimethylated H3K4 in gene activation (Pray-Grant, et al. (2005)) and gene repression (Shi, et al. (2006)) provide an illustration of the complexity of Set1-like methyltransferase-mediated regulation. The identi cation and interpretation of H3K4me1-mediated mechanisms will contribute to our understanding of transcriptional regulation and their impact on complex processes, like development and differentiation, in higher organisms. Moreover, evaluating the effects of altered H3K4 methylation patterns in the partial function set1 mutants on other transcription-related processes, such as splicing and elongation, will provide deeper insight into the mechanisms that regulation transcription. Funding: Texas A&M AgriLife Research Con icts of interest/Competing interests: Not applicable Availability of data and material: The authors will make all data and unique research materials and data freely available to other investigators.
Code availability: Not applicable Authors' contributions: SHP, RJ, ND and MB conceived the experiments. Preliminary data were generated by RJ and ND. Data in gures were generated by ND. The manuscript was written by ND and MB. After review and comments, the manuscript was approved by all authors.
Ethics approval: Not applicable Consent to participate: Not applicable Consent for publication: Not applicable  Growth of mutants lacking Set1 methylation activity is sensitive to histidine starvation in the presence of 3AT. Five-fold serial dilutions of yeast cultures expressing wild type SET1+ or set1 mutant alleles were spotted onto SC Complete agar with or without 10 mM 3AT (A) and SC-His agar with or without 10 mM 3AT (B). Allele, relevant genotype, and Activity, H3K4 methylation capacity based on Figure 1B, are indicated on the left of the panels. Null, no detectable H3K4me1/2/3; mono, H3K4me1 with low or no H3K4me2/3; mono/di, detectable H3K4me1 and H3K4me2; mono/tri, detectable H3K4me1 and H3K4me3