Functional Characterization of the Drosophila Hmt4-20/Suv4-20 Histone Methyltransferase

Abstract

Di- and trimethylation of histone H4 lysine20 (H4K20) are thought to play an important role in controlling gene expression in vertebrates and in Drosophila. By inducing a null mutation in Drosophila Suv4-20, we show that it encodes the histone H4 lysine20 di- and trimethyltransferase. In Suv4-20 mutants, the H4K20 di- and trimethyl marks are strongly reduced or absent, and the monomethyl mark is significantly increased. We find that even with this biochemical function, Suv4-20 is not required for survival and does not control position-effect variegation (PEV).

CHROMATIN organization is regulated in part by post-translational modification of the four histone proteins H2A, H2B, H3, and H4, which control packaging of DNA into nucleosomes. The histone tails consisting of 20–35 mostly basic amino acids are the main targets for these modifications. Histone tail methylations have been thought to control transcription primarily by packaging chromatin into open domains, accessible to the transcription machinery, and into inaccessible closed domains (Strahl and Allis 2000; Rice and Allis 2001; Sims et al. 2006). Recently, histone tail methylations have been shown to also function in the response to DNA damage in yeast and progression through the cell cycle in Drosophila and vertebrates (Sanders et al. 2004; Jorgensen et al. 2007; Sakaguchi and Steward 2007; Tardat et al. 2007).

Histone lysines can be mono-, di-, or trimethylated (me1, me2, and me3). The only methylated amino acid in the tail of histone H4 is lysine 20 (H4K20). H4K20 me1, me2, and me3 are relatively stable in the cell cycle and development (Rice et al. 2002; Karachentsev et al. 2005). Me1 of histone H4 is controlled in vertebrates and Drosophila by the PR-Set7 (or -Set 8) methyltransferase (Fang et al. 2002; Nishioka et al. 2002).

The PR-Set7 monomethyl transferase is essential in both Drosophila and vertebrates. Drosophila lacking this enzyme die at the larval-to-pupal transition. In diploid brain cells of mutant larvae, H4K20 me1 is strongly reduced or absent, the mitotic progression is abnormal, and the DNA damage checkpoint is activated. Further, these cells show a strong defect in chromosome condensation. H4K20 me1 appears to be essential for the normal progression through the cell cycle and not for transcription (Sakaguchi and Steward 2007).

The histone H4-K20 trimethyltransferase, Suv4-20, was isolated and characterized by Schotta et al. (2004). Two Suv4-20 genes, Suv4-20h1 and Suv4-20h2, exist in vertebrates, but there is only one Drosophila ortholog. In vertebrates, Suv4-20-dependent histone H4-K20 trimethylation is strongly associated with constitutive heterochromatin at the chromocenter. The me3 mark appears to be organ-specifically distributed and to change during the life of an animal (Karachentsev et al. 2007).

Many results in vertebrates documenting changes in histone H4K20 me3 in cancer cells and aging suggest that this modification also has a fundamental role in gene silencing (Sarg et al. 2002; Wang et al. 2007). To investigate the importance of this modification in Drosophila and to determine the difference in requirement between H4K20 me1 and me3, we induced a complete loss-of-function allele of Suv4-20.

Suv4-20 represents the Drosophila H4K20 di- and trimethyltransferase. In larvae lacking the enzyme, H4K20 me1 is significantly increased. We found no requirement of Suv4-20 for survival under normal and stress conditions (DNA damage and heat shock) in flies reared in the laboratory. Unexpectedly, Suv4-20 did not suppress position-effect variegation (PEV) (Schotta et al. 2004).

MATERIALS AND METHODS

Fly stocks:

All fly stocks were obtained from the Bloomington Stock Center except P[EP1216a], sent from the Szeged stock center, and the grapes^fsA4 line from W. Theurkauf. The w¹¹⁸ stock was used as wild-type stock in all experiments.

RT–PCR and Western blots:

For RT–PCR, total RNA was extracted from wild-type and mutant larvae. This RNA was not treated with DNAse and served as template for combined reverse transcription and PCR using the QIAGEN (Valencia, CA) OneStep RT–PCR kit. DNA contaminating the RNA preparation can also be amplified by this method (DNA in Figure 2A). The primers used for amplification are indicated in Figure 1 (Suv4-20 F, 5′-ctcttaaggtggacacgcaacaactgaa-3′; Suv4-20 R, 5′-cctctgtcagctcggcgatgcag-3′; skpA F, 5′-tgggtgagcatttctggaaagtg-3′; skpA R, 5′-gcattctcatcgtcctccatgc-3′). Western blot analysis was performed as described (Whalen and Steward 1993). Non-green (male) larvae were collected from stocks balanced with FM7-GFP, homogenized in 2–3 vol of standard SDS–PAGE protein loading buffer. Antibodies were used in the following dilutions: anti-H4-K20 mono-, di-, and trimethyl and histone H4 rabbit polyclonal antibody (Upstate) 1:1000.

Figure 2.—

Suv4-20 is the di- and trimethyltransferase of histone H4K20. (A) RT–PCR experiment of RNA isolated from hemizygous Df(1)174 and Df(1)109 larvae. Only the skpA mRNA is absent in Df(1)174. (B) Western blots of wild-type and Df(1)109; P{w⁺, skpA} double-homozygous larvae. The monomethyl mark is increased in the deficiency animals, while the H4K20 me2 and me3 marks are strongly reduced or absent.

Figure 1.—

The Suv4-20-skpA genomic region. Three P-element insertions in the Suv4-20 gene; the two newly isolated deletions and the primers (arrows) used in the RT–PCR experiments are indicated.

Salivary gland chromosome preparations and antibody staining:

Salivary glands of third instar larvae were dissected, and polytene chromosomes were isolated and mounted on slides as described (Lis et al. 2000). Polytene chromosomes were stained with anti-H4-K20 di- and trimethyl rabbit polyclonal (Upstate) 1:100.

DNA damage assay:

Sensitivity to hydroxyurea (HU) and methyl methanesulfonate (MMS) was quantified as follows: 10 females and 10 males were crossed in one vial. After 24 hr, the parents were transferred to a new vial and 24 hr later the drug or distilled H₂O was added to the food (more than five vials per assay). All classes of adult progeny in the grp cross or females in the other crosses were scored. The homozygous Df(1)174 and Df(1)109 were homo- or heterozygous for P{w⁺, skpA}. The percentage of homozygous flies is indicated in Table 1.

TABLE 1.

Sensitivity of Suv4-20 to the DNA damage drugs HU and MMS

Genotype	H2O	20 mm HU	0.05% MMS
grp fsA4/+ × grp fsA4/+	29.7 (546)	0 (119)	2.22 (316)
109/FM7; P{w+, skpA}/+ × 109/Y; P{w+, skpA}/+	53.0 (234)	48.3 (331)	50.0 (286)
174/FM7; P{w+, skpA}/+ × 174/Y; P{w+, skpA}/+	42.6 (359)	44.8 (399)	47.4 (388)

The percentage of homozygous flies is indicated. The total number of flies scored is indicated in parentheses.

Response to γ-irradiation:

Third instar larvae were irradiated with 2000 R. Brains were dissected, fixed, and squashed 30 min and 2 hr after irradiation, and the number of cells in mitosis was counted as described (Sakaguchi and Steward 2007). All mitotic indexes (MI; mitotic cells/total cells × 100) were normalized relative to their respective t = 0 control index (2.16 for wild type, 3.10 for mei-41, and 2.22 for Df(1)174 and Df(1)109).

Phenotypic characterization of Stubble and white variegation:

To study PEV, w¹¹⁸/w¹¹⁸, Df(1)174/FM7, Df(1)109/FM7, P[BG00814]/P[BG00814], and w;Su(var)3-9⁰⁶, e, ro/TM6 virgins were crossed to T(2;3)Sb^v, In(3R)Mo, Sb¹, and sr¹/TM3 males. In the next generation, the four scutellar bristles of at least 30 trans-heterozygous females were scored as being wild type or Sb. Further, fly stocks bearing w⁺ transgenes inserted into centromeric or telomeric heterochromatin on the second, third, or fourth chromosome (Cryderman et al. 1998, 1999) were crossed with w¹¹⁸/w¹¹⁸, Df(1)174/FM7, or Df(1)109/FM7. The eye colors of F₁ trans-heterozygous females were compared.

RESULTS

Induction of a Suv4-20 null allele:

Three P-element lines with insertions in the Suv4-20 gene are available (Figure 1). P[BG00814] (Schotta et al. 2004), P[EP1216a], and P[EY07422] did not show a significant reduction of H4K20 me3 on Western blots or salivary gland staining and so we decided to produce a null allele to study the function of the gene.

To induce a null mutation, we mobilized P[EY07422] and obtained six larval lethal excision mutants. Molecular analysis and sequencing of the deletion breakpoints indicate that five of our mutant lines were similar to Df(1)174. One breakpoint of this deletion is in the first predicted exon of skpA and the other breakpoint is in the first predicted intron of Suv4-20. We investigated by RT–PCR if either of the two transcripts is present in the deletion and found that a fragment encoded by the second and third exon of Suv4-20 could be amplified from Df(1)174 RNA, suggesting that Suv4-20 is normally expressed in these flies (Figure 1 and Figure 2A). In one additional line, Df(1)109, ∼600 bp at the 5′-end of skpA are missing and the deletion extends into the first protein coding exon of Suv4-20 (Figure 1 and Figure 2A). Consistent with these observations, we detected Suv4-20-dependent H4K20 me3 in extracts from Df(1)174 larvae, but not in extracts from Df(1)109 animals (supplemental Figure 1S).

Suv4-20 is the histone H4 lysine20 di- and trimethyltransferase:

On Western blots of control wild-type and Df(1)174 (not shown) extracts, all three modified histones are detected. In extracts from Df(1)109 larvae, H4-K20 me1 is clearly present, but only weak bands of me2 and me 3 are detected (Figure 2B). These weak me2 and me3 bands are possibly due to cross-reactivity of the antibody with nonmethylated histone H4. At the same time, H4K20 me1 is markedly increased in Df(1)109 extracts in the absence of me2 and me3.

As expected, on salivary gland chromosomes of Df(1)109 larvae, H4-K20 me2 and me3 are strongly reduced or absent, while on wild-type and Df(1)174 chromosomes their levels and distribution are normal (Figure 3). We conclude that Drosophila Suv4-20 encodes a methyltransferase responsible for di- and trimethylation of histone H4K20.

Figure 3.—

The H4K20 di- and trimethyl mark is not detected on Df(1)109/Y salivary gland chromosomes. Wild type (left) and mutant salivary gland chromosomes, stained with anti-dimethylated and anti-trimethylated H4K20 antibody (red), are indicated. DNA was stained with Hoechst (blue).

Functional requirement for Suv4-20:

The genomic skpA transgene P{w⁺, skpA} on the third chromosome rescues skpA null alleles to viability (Murphy 2003). Homozygous Df(1)174 and Df(1)109 flies carrying P{w⁺, skpA} are viable and fertile. In both lines, the hatch rate of embryos is only ∼60% that of wild type.

Because of this partial embryonic lethality, we investigated if the deficiencies affect meiosis. A hallmark of mutations affecting meiosis is chromosome nondisjunction, manifested by the production of XXY females. We did not see any increase in the nondisjunction phenotype when Df(1)174 and Df(1)109 females were crossed to males carrying a marked Y chromosome.

Genes functioning in DNA damage response often are not essential for viability (Song 2005). We determined if Suv4-20 mutants have defects in the DNA damage response by exposing mutant larvae or embryos to hydroxyurea and methyl methanesulfonate (Sibon et al. 1999). Survival of Df(1)174/Df(1)174; P{w⁺, skpA} and Df(1)109/Df(1)109; P{w⁺, skpA} flies was not affected, showing that the response to DNA damage was normal in animals lacking Suv4-20 (Table 1). As a positive control, we exposed grapes^fsA4 (grp^fsA4), the Drosophila Chk1 ortholog, to the same drugs. The survival of homozygous grp^fsA4 flies was, as expected, strongly affected (Sibon et al. 1999).

In Schizosaccharomyces pombe, histone H4-K20 methylation is not required for viability but controls the recovery time after DNA damage induced by irradiation (Sanders et al. 2004; Du et al. 2006). In essentially the same experiment as done in yeast, we examined the response of wild-type and mutant flies to γ-irradiation. By determining the MI, we found that the response is normal (Table 2). In wild type, as expected, the MI fell drastically 30 min after irradiation. Df(1)174 and Df(1)109 showed a similar response to wild type. In wild type, the MI remained suppressed for at least 2 hr. Both deficiencies recovered more quickly than did wild type. This phenotype is not dependent on Suv4-20 because both deletions behaved in the same way. We used the fly ATR ortholog mei-41 as a positive control and found that the mei-41 MI showed only 30% reduction of MI 30 min after irradiation (Hari et al. 1995).

TABLE 2.

Suv4-20 and response to γ-irradiation

	Relative MI (±2SD) after γ-irradiation
Genotype	30 min	2 hr
Wild type	0.27 ± 0.14	0.48 ± 0.23
mei-41D3/mei-41D3	0.70 ± 0.07	1.00 ± 0.40
109/109; P{w+, skpA}/P{w+, skpA}	0.33 ± 0.17	1.05 ± 0.29
174/174; P{w+, skpA}/P{w+, skpA}	0.16 ± 0.11	1.38 ± 0.20

We further tested if H4K20 me2 and me3 have any effect on the heat-shock response. We exposed adult homozygous Df(1)174 and Df(1)109 females carrying one copy of the skpA transgene and wild-type females to 36° for 60 min. Western blots of extracts of these females and non-heat-shocked females were performed. The induction of expression of HSP70 and HSP 23 proteins was similar in wild-type and mutant females. In the extracts of all heat-shocked flies, the induction of the two proteins was normal (results not shown).

Suv4-20 and suppression of PEV:

Schotta et al. (2004) showed that a P-element insertion into the 3′-UTR of Suv4-20, P[BG00814] functions as a dominant suppressor of T(2;3)Sb^v. On the T(2;3)Sb^v chromosome, the dominant bristle marker Stubble (Sb) is localized close to the pericentric heterochromatin. The expression of this translocated marker is variegated in a wild-type background. The variegation becomes more pronounced in the P-induced Suv4-20 mutant background.

We repeated these experiments with the same P-element line used in the Schotta et al. (2004) experiment with Df(1)174 and Df(1)109 and, as controls, with white¹¹⁸ and Su(var)3-9 (Table 3). We scored the same four scutellar bristles and found similar numbers in all genetic combinations with the exception of Su(var)3-9 (Table 3). The lowest number of Sb bristles (29%) was present in the T(2;3)Sb^v stock.

TABLE 3.

Phenotypic characterization of Stubble variegation

Genotype	% Sb bristles
Df(1)109/In(3R)Mo, Sb1, sr1	40
Df(1)174/In(3R)Mo, Sb1, sr1	50
P[BG00814]/In(3R)Mo, Sb1, sr1	43
w118/In(3R)Mo, Sb1, sr1	34
Su(var)3-906/In(3R)Mo, Sb1, sr1	100
TM3/In(3R)Mo, Sb1, sr1	29

We further tested the function of Suv4-20 in suppression of PEV by crossing Df(1)174 and Df(1)109 males carrying one copy of the skpA transgene to virgins from 10 PEV tester chromosomes carrying w⁺ insertions in the centromeric and telomeric heterochromatin of all chromosomes (see materials and methods; Cryderman et al. 1998, 1999). In the next generation, the heterozygous females not carrying P{w⁺, skpA} were examined. The eye color of all trans-heterozygous females was indistinguishable from that of siblings that did not carry the deficiencies.

Both these experiments are consistent with one another and indicate that Suv4-20 does not function in controlling gene expression. Hence the name of the gene, Suv4-20, is misleading and we propose that the gene should be renamed histone methyltransferase 4-20 (hmt4-20).

DISCUSSION

The two vertebrate HMT4-20 (Suv4-20h1 and Suv4-20h1) proteins have been shown to act as specific nucleosomal H4K20 trimethylating enzymes (Schotta et al. 2004). Recently, Suv4-20h2 was identified as responsible for H4K20 me3 at vertebrate telomeres (Benetti et al. 2007). The enzymatic activity of Suv4-20h1 in vertebrates has not been determined yet.

We found that the Drosophila HMT4-20/Suv4-20 enzyme has di- and trimethylating activity. Concomitant with the absence of the H4K20 me2 and me3 marks, the level of H4K20 me1 is significantly increased (Figure 2B), suggesting that me1 represents a substrate for the HMT4-20 enzyme. This result is consistent with our previous observation that, in the absence of PR-Set7 and strong reduction of H4K20 me1 on salivary gland chromosomes in late third instar larvae, me2 and me3 are also reduced (Karachentsev et al. 2005). This conclusion is also supported by the observation of Pesavento et al. (2008), who found by mass spectrometry that in HeLa cells H4K20 is first monomethylated. H4K20 me1 is then converted into me2 and a fraction of this H4K20 me2 is converted into me3.

Our genetic experiments show that histone H4K20 me2 and me3 does not affect PEV. This result is surprising because Schotta et al. (2004) reported that, in Drosophila, a P-insertion into the third exon (3′-UTR) of hmt4-20/Suv4-20 P[BG00814] suppressed PEV of an Sb reporter gene translocated to heterochromatin. We tested several mutant and control chromosomes for their effect on the expression of the translocated Sb gene and found that in both our deficiencies, only one of them missing hmt4-20/Suv4-20, and in P[BG00814], the number of Sb bristles was similar (40, 50, and 43%, respectively). In combination with w¹¹⁸ and in the stock carrying the translocation, the numbers were lower (34 and 29%). In combination with a well-defined suppressor of PEV, Su(var)3-9, 100% of the bristles showed the Sb phenotype. Our second approach—in which we tested the effect on PEV of w⁺ inserted in centromeric or telomeric heterochromatin by w¹¹⁸ and the two deletion lines—showed no significant change of expression of the w⁺ transgenes. We conclude that hmt4-20/Suv4-20 does not have a significant function in transcriptional regulation.

Histone methyl marks are thought to be required for fundamental aspects of chromatin structure and control of gene expression, DNA replication, or mitosis, processes essential for flies in the laboratory and in the wild. We previously studied the distribution of the three H4K20 methyl marks and showed that they are developmentally regulated and are detected cell-cycle-specifically on chromosomes (Karachentsev et al. 2007). The requirement of the H4K20 monomethyl mark is fundamental in vertebrates and flies. In its absence, the DNA damage checkpoint is activated and chromosome integrity is affected (Jorgensen et al. 2007; Sakaguchi and Steward 2007; Tardat et al. 2007).

Therefore it is all the more surprising that our experiments aimed at determining the function of hmt4-20/Suv4-20 in viability, meiosis, fertility, modification of PEV, DNA damage response, and response to stress showed that me2 and me3 methylation of H4K20 is dispensable for all these processes.

In yeast, the Set9 enzyme controls H4K20 me1, me2, and me3. While the methyl marks are dispensable for viability, they show a weak phenotype in DNA damage response (Sanders et al. 2004; Du et al. 2006). This aspect of H4K20 methylation has been lost in Drosophila.

The vertebrate HMT4-20/Suv4-20 enzymes also have a significantly less important function than the mono-methyltransferase encoded by PR-Set7. Loss of both HMT4-20/Suv4-20 enzymes is apparently viable in mouse embryonic fibroblasts, but results in abnormalities in telomere-length regulation (Benetti et al. 2007).