Independent Biological and Biochemical Functions for Individual Structural Domains of Drosophila Linker Histone H1

Abstract

Linker histone H1 is among the most abundant components of chromatin. H1 has profound effects on chromosome architecture. H1 also helps to tether DNA- and histone-modifying enzymes to chromatin. Metazoan linker histones have a conserved tripartite structure comprising N-terminal, globular, and long, unstructured C-terminal domains. Here we utilize truncated Drosophila H1 polypeptides in vitro and H1 mutant transgenes in vivo to interrogate the roles of these domains in multiple biochemical and biological activities of H1. We demonstrate that the globular domain and the proximal part of the C-terminal domain are essential for H1 deposition into chromosomes and for the stability of H1-chromatin binding. The two domains are also essential for fly viability and the establishment of a normal polytene chromosome structure. Additionally, through interaction with the heterochromatin-specific histone H3 Lys-9 methyltransferase Su(var)3-9, the H1 C-terminal domain makes important contributions to formation and H3K9 methylation of heterochromatin as well as silencing of transposons in heterochromatin. Surprisingly, the N-terminal domain does not appear to be required for any of these functions. However, it is involved in the formation of a single chromocenter in polytene chromosomes. In summary, we have discovered that linker histone H1, similar to core histones, exerts its multiple biological functions through independent, biochemically separable activities of its individual structural domains.

Introduction

DNA in the eukaryotic nucleus is packaged into a compact nucleoprotein complex called chromatin (1, 2). The histones constitute a family of basic proteins that play a vital role in organizing chromatin. There are five major classes of histones: the core histones H2A, H2B, H3, and H4 and the linker histone H1. The nucleosome core particle, the fundamental unit of chromatin, consists of an octamer of two copies of each of the core histones around which about 146 bp of DNA is wrapped. H1 binds to the nucleosome core particle and the linker DNA between adjacent core particles. The stoichiometry of H1 to nucleosome core particles in cells of complex eukaryotes ranges from about 0.5 to approximately equimolar (3). Given this abundance, the linker histone H1 is expected to play important roles in chromatin structure. Indeed, numerous in vitro studies, as well as experiments in vivo, have shown that H1 has a profound effect on chromatin structure (reviewed in Refs. 4 and 5). For example, incorporation of H1 into chromatin leads to an increase in the spacing between nucleosome core particles. H1 also facilitates the folding of chromatin into more compact structures. In addition to these architectural roles, there are now an increasing number of reports describing interactions between H1 linker histones and a variety of other nuclear proteins (6), including functional interactions between H1, histone-modifying enzymes, and DNA methyltransferases that lead to modulation of epigenetic marking of chromatin (7, 8). In this context, it is notable that the binding of linker histones to chromatin is highly dynamic, with residence times much shorter than that of the core histones (9).

In higher eukaryotes, linker histones are typically ∼200 amino acid residues in length. They comprise a tripartite structure consisting of a relatively short (∼40 residues) unstructured N-terminal domain (NTD),⁴ an evolutionarily conserved central globular domain (GD; ∼80 residues), and a highly basic, disordered C-terminal domain (CTD; >100 residues). Very little is known about the functions of the H1 NTD. The structures of several linker histone GDs have been examined by x-ray crystallography (10) and NMR (11, 12) and have been found to resemble a 3-helix “winged helix” fold. The H1 GD makes specific contacts with the nucleosome core particle and the nucleosomal DNA near the dyad axis (13). Fluorescence recovery after photobleaching (FRAP) experiments have demonstrated that the GD plays a dominant role in H1 binding to chromatin in cultured cells (14). FRAP experiments also indicate that the CTD plays an important role in the residence time of H1 in chromatin (15). In addition, experiments with oligonucleosome arrays assembled in vitro showed that the H1 CTD is required to condense the arrays into higher order structures (16–18).

Despite these advances, there is a paucity of information about the roles of the individual H1 structural domains in the many different biological functions of H1 in vivo. Drosophila melanogaster offers an attractive system for such studies because, in contrast to mammals, it expresses a single linker histone protein during much of its life cycle. H1 is essential for proper development of flies, and it plays key roles in several aspects of chromosome structure and gene regulation (7, 19). Here we investigate the functional roles of the three H1 structural domains, the H1 NTD, GD, and CTD, in the roles of H1 in Drosophila development, chromosome structure, and epigenetic regulation.

Results

The C-terminal Domain of Linker Histone H1, but Not Its N-terminal Domain, Is Required for Localization to Chromatin in Vivo

To develop an in vivo approach for assessing the contributions of the histone H1 structural domains to the various biological functions of H1, we utilized a system that we described previously for depleting H1 in Drosophila by RNAi (19). Flies carrying a GAL4-responsive, UAS-driven transgene (pINT-1-H1) encoding H1-specific hairpin RNA are mated to flies bearing the ubiquitous tubulin-GAL4 driver (see “Experimental Procedures”). Larvae carrying both transgenes express reduced levels of H1 protein (19). The extent of reduction depends upon the integration site of the pINT-1-H1 transgene and the temperature, which is due to higher transcriptional activity of GAL4 at elevated temperatures (20). The pINT-1-H1 transgenes used in this study cause a depletion of H1 protein levels to ∼10–30% of normal. This reduction leads to marked changes in larval chromosome structure and gene expression and sharply reduces the rates of eclosion of adult flies (19).

To utilize this system for H1 structure-function analyses, we generated an additional set of alleles harboring UAS-driven transgenes encoding full-length wild-type, mutant, or truncated H1 polypeptides. The variant proteins included H1 molecules lacking the N-terminal domain (ΔN); the C-terminal domain (ΔC100%); or 25, 50, or 75% of the CTD (Δ25%CTD, Δ50%CTD, and Δ75%CTD, respectively) (Fig. 1A). The H1 proteins were tagged with a FLAG epitope at their N termini. We also constructed a transgene encoding a triply mutated H1 protein with alanine substitutions at three lysine residues (K58A/K91A/K95A) that were found to be important for H1 binding to the nucleosome core particle in vitro (12). Finally, we constructed, as a negative control, a nonsense mutant containing two in-frame stop codons in the H1-coding sequence (substituting Ser-2, which immediately flanks the N-terminal FLAG, and Met-52 within the globular domain, the only remaining methionine codon in the native H1 sequence).

FIGURE 1.

Ectopic expression of H1 transgenes in vivo. A, schematic representations of the H1 variant proteins encoded by the transgenes. The three major structural domains of Drosophila H1 are represented by an unfilled rectangle (NTD), a filled oval (GD), and a shaded rectangle (CTD). Each transgene-encoded protein contains a FLAG epitope on its N terminus. The abbreviated name for each mutant protein is shown on the right along with the residues included in the protein and mutated positions in parentheses. FL (2 … 255), full-length H1 protein. Solid line, sequence corresponding to the H1 RNAi transgenes (19). B, expression levels of H1 transgenic proteins in third instar larvae. L3 larval extracts were analyzed by immunoblotting with antibodies against FLAG (left) and Drosophila H1 (right). The position of endogenous H1 in the gel is indicated by an arrow on the right. The proteins encoded by G3×MUT and FL transgenes migrate slightly more slowly than endogenous H1 due to the FLAG epitope, resulting in two closely spaced bands detected with the anti-H1 antibody. The protein encoded by ΔC25% transgene migrates slightly more rapidly than endogenous H1, due to the combination of a truncation and the added FLAG epitope. Positions of molecular mass marker bands (kDa) are shown on the left.

The transgenes were integrated at the unique 86Fa attP landing site on right arm of the third chromosome (3R) by PhiC31 integrase-mediated transgenesis (21). Expression of the H1 proteins encoded by the transgenes was examined by immunoblotting with anti-FLAG and anti-H1 antibodies (Fig. 1B). Quantitation of the Western blotting signals showed that the exogenous H1 proteins accumulated to levels comparable with or greater than that of endogenous H1 (Table 1). In addition, RT-PCRs specific to FLAG-tagged transgenic H1 transcripts revealed that all transgenes, including the nonsense 2×STOP mutant containing two in-frame stop codons, were expressed at the level of mRNA in comparable amounts (data not shown). However, as expected, the 2×STOP transgene failed to express a FLAG-tagged protein (Fig. 1B).

TABLE 1.

Summary of expression and chromatin binding properties of transgenic variant H1 polypeptides

All analyses were carried out with L3 larvae reared at 29 °C throughout their life cycles. All transgenic proteins harbor the N-terminal FLAG tag. Homozygous males carrying UAS-driven mutant H1 polypeptide transgenes were crossed to female Tub-GAL4/TM6, Tb counterparts. Tb⁺ progeny were used for all analyses. The calculation of expression levels of transgenic FLAG-H1 relative to that of endogenous H1 was performed by immunoblotting (Fig. 1B) in two independent experiments. Images were collected using the LI-COR Odyssey infrared imaging system, and the intensities of bands (mean values) were measured using ImageJ software. The ratios of intensity values for endogenous and transgenic H1 (H1 and FLAG signals, rectangular grid tool) were calculated; the average values ± S.D. are shown (column 2). Nuclear localization was determined as in Fig. 2B by comparing the anti-FLAG IF with the DAPI staining patterns (column 3). Loading to chromosomes and co-localizations with endogenous H1 (columns 4 and 5, respectively) were determined from anti-FLAG and anti-H1 staining as in Fig. 2, A and C. Stable association with chromatin in vitro (column 6) was assayed by sucrose gradient sedimentation (Fig. 3C). Significant results are highlighted by boldface type. Y, yes; N, no; NA, not applicable; ND, not determined.

H1 mutant transgene	Expression relative to endogenous H1	Nuclear localization	Loading into chromosomes	Co-localization with endogenous H1	Stability in vitro
2×STOP (S2*, M52*)	0.00	NA	NA	NA	NA
G3×MUT (K58A/K91A/K95A)	0.30 ± 0.04	Y	N	N	ND
FL (2 … 255)	0.40 ± 0.02	Y	Y	Y	Y
ΔN (44 … 255)	1.52 ± 0.08	Y	Y	Y	ND
ΔC25% (2 … 224)	0.75 ± 0.07	Y	Y	Y	Y
ΔC50% (2 … 183)	0.91 ± 0.04	Y	Y	Y	Y
ΔC75% (2 … 149)	0.72 ± 0.05	Y	Y	Y	N
ΔC100% (2 … 124)	0.40 ± 0.01	Y	N	N	N

Next, we determined whether the mutant exogenous H1 proteins are incorporated into chromatin in vivo. We employed indirect immunofluorescence (IF) staining for H1 with anti-FLAG and anti-H1 antibodies on spreads of polytene chromosomes and on whole mounts of salivary gland cells (Fig. 2, A and B). IF staining of FLAG-tagged H1 shows that the exogenous H1 is distributed similarly to endogenous H1 (Fig. 2A). High magnification split images of polytene chromosome arms co-stained with anti-FLAG and anti-H1 antibodies strongly suggest that the distributions of exogenous and endogenous H1 polypeptides are essentially identical (Fig. 2C). In contrast, the mutant containing three substitutions in the globular domain failed to bind to polytene chromosomes, consistent with the results from in vitro studies (12). Interestingly, mutant H1 proteins lacking the NTD or up to 75% of the CTD are readily detectable in polytene chromosomes with distributions indistinguishable from that of endogenous H1 (Fig. 2, A and C). Thus, the NTD and a large portion of the CTD are not essential for H1 incorporation into chromatin in vivo. In contrast, complete deletion of the CTD and mutations of three key residues in the globular domain render these mutants unable to localize to chromatin. Whole mount staining experiments nonetheless indicate that these mutants accumulate in the nucleus (Fig. 2B). Thus, their failure to localize to chromatin is not due to a defect in nuclear import. These results implicate the GD and CTD as crucial determinates of Drosophila H1 deposition into chromatin in vivo.

FIGURE 2.

The globular and C-terminal domains but not the N-terminal domain of Drosophila H1 are essential for its localization to chromatin in vivo. A, localization of truncated H1 polypeptides in polytene chromosomes in vivo. Salivary glands from L3 larvae of transgenic Drosophila lines expressing FL H1 or the indicated mutant H1 proteins were squashed, and polytene spreads were stained with DAPI and antibodies against H1 and FLAG. Note the absence of the FLAG-H1 transgenic polypeptides in salivary gland chromatin of the 2×STOP (a non-expressing negative control), G3×MUT, and ΔC100% transgenic alleles. B, subcellular localization of truncated H1 polypeptides in larval salivary gland cells. Whole mount salivary glands from transgenic L3 larvae as in A were fixed and stained with DAPI (to label the nuclei) and antibodies against H1 and FLAG. A comparison of the DAPI fluorescence and anti-FLAG immunofluorescence patterns for the ΔC100% and G3×MUT transgenic lines in A and B shows that these two mutant H1 proteins localize to the nuclei but are not bound to chromatin. An immunofluorescence signal for the anti-FLAG antibody is undetectable in the 2×STOP mutant (a non-expressing negative control). C, precise co-localization of endogenous and transgenic H1 proteins in the polytene chromosome spreads. The images represent higher magnification views of parts of images presented in A. Merged split images illustrate that the indicated transgenic H1 proteins exhibit localization patterns nearly identical to that of endogenous H1 in polytene bands.

The Proximal Part of the C-terminal Domain Regulates the Binding of H1 to Chromatin in Vitro

Although the mutants lacking 25, 50, and 75% of the CTD could be detected in polytene chromosomes, it was important to determine whether these truncations quantitatively affect the stable association of the mutant proteins with chromatin, their proper positioning with respect to the nucleosome core particle, and their ability to protect the linker DNA. Therefore, we studied the properties of chromatin reconstituted in vitro with recombinant full-length Drosophila H1 and the four truncated mutants with progressively larger deletions of the CTD. We prepared bacterial expression constructs that encode the full-length and C-terminally truncated H1 polypeptides corresponding to the truncated transgenic proteins used in vivo (Fig. 1A). The recombinant proteins did not include the N-terminal FLAG but, rather, were purified to apparent homogeneity (Fig. 3A) without a tag by using the intein-chitin binding domain system (see “Experimental Procedures”) and conventional chromatography.

FIGURE 3.

Properties of chromatin assembled in vitro with C-terminally truncated H1 variant proteins. A, purified recombinant H1 polypeptides. FL and C-terminally truncated H1 polypeptides expressed in E. coli were analyzed by SDS-PAGE and Coomassie staining. Approximately 4 pmol of each protein is loaded per lane. Molecular mass markers (kDa) are shown on the left. B, chromatosome stop assay with recombinant H1 polypeptides. Oligonucleosomes were assembled in vitro with native Drosophila core histones, ACF, and NAP-1, without and with the indicated recombinant H1 polypeptides. Assembled chromatin was subjected to partial micrococcal nuclease digestion, and DNA was analyzed by agarose gel electrophoresis and ethidium bromide staining. Positions of the core particle DNA (CP) and chromatosome DNA (CHR; core particle plus linker region protected from digestion by H1) in chromatin assembled with full-length H1 are indicated by arrows on the left. Note the decreasing size of chromatosome particles formed with truncated H1 polypeptides. DNA fragment sizes in the 20-bp DNA ladder marker are shown on the right. C, sucrose gradient sedimentation of H1-containing chromatin. Free full-length H1 and H1·NAP-1 and H1·DNA complexes as well as chromatin assembled in vitro without and with various H1 polypeptides were subjected to sedimentation on 5–30% sucrose density gradients. ∼1% of the assembly reaction (starting material; SM) and of each gradient fraction (1–10) was subjected to SDS-PAGE and immunoblotting with anti-H1 antibody (left panels). Chromatin assembled with full-length H1 sediments quantitatively to the bottom of the gradient (fraction 10). Control experiments (right panels) show sedimentation profiles of DNA and core histones in chromatin assembly reactions (without H1), full-length H1 alone, and binding reactions containing H1, NAP-1, and ACF or containing H1 and DNA (without core histones), demonstrating that H1 sediments in fraction 10 only if it is associated with fully assembled chromatin. Results similar to those shown in the right panels were obtained in control experiments performed with each of the truncated H1 polypeptides (not shown). Note that substantial amounts of H1 are present in fractions 3–9 of samples derived from chromatin assembled with H1 lacking 75% or more of the CTD, indicating transient dissociation of the truncated H1 polypeptides from chromatin during centrifugation. N, nicked plasmid DNA; SC, supercoiled plasmid DNA.

Oligonucleosomes were assembled in the defined ATP-dependent chromosome assembly system supplemented with recombinant H1 polypeptides (43). The formation of the H1-containing chromatin was examined by the “chromatosome stop” assay that utilizes partial digestion of assembled chromatin by micrococcal nuclease. The presence of stoichiometric amounts of H1 molecules (relative to nucleosomes) modifies the digestion pattern of chromatin by micrococcal nuclease and, in addition to the DNA fragment protected by the nucleosome core particle (∼146 bp), reveals a longer chromatosome DNA fragment (up to 200 bp for the full-length Drosophila H1). Consistent with the in vivo data (Fig. 2), H1 truncations lacking 25, 50, and 75% of the CTD are efficiently assembled into chromatin. However, we observed that the length of the protected linker DNA fragment decreases progressively for the shorter H1 proteins (Fig. 3B). Intriguingly, even the H1 truncation that completely lacks the CTD protects DNA outside of the nucleosome core particle (<20 bp). Thus, in the defined system in vitro (in the absence of additional factors), the CTD appears dispensable for the deposition of H1 into chromatin. However, it should be noted that these in vitro experiments were performed with truncated H1 polypeptides in the absence of competing full-length H1, whereas the in vivo studies showing that the ΔC100% truncation mutant fails to be incorporated into chromosomes (Fig. 2A) occurred in the presence of competing, endogenous full-length H1.

It remains possible, however, that the CTD may contribute to the stability of H1-DNA or H1-nucleosome association. To examine this possibility, we performed sucrose gradient sedimentation of chromatin assembled with the addition of full-length H1 and various truncated polypeptides (Fig. 3C). We discovered that the full-length H1 and the ΔC25% and ΔC50% CTD truncation mutant proteins remained associated quantitatively with chromatin throughout prolonged ultracentrifugation, sedimenting along with the chromatin to the bottom of the gradient. On the other hand, the mutant polypeptides with shorter CTDs exhibited increasing dissociation from chromatin, as evidenced by their presence in several gradient fractions. These results are consistent with FRAP studies of a mammalian H1 showing that the CTD contributes to the stability of H1 association with chromatin (9, 15). These results also further support our observations that the proximal fragments of the CTD contribute to the binding of H1 to chromatin in vivo (Fig. 2).

The C-terminal Domain of H1 Is Essential for Drosophila Development

Depletion of H1 in Drosophila by RNAi leads to lethality at the late larval-pupal stages of development (19). When H1 levels are reduced below ∼50% of normal, no adult offspring are observed. We sought to test the functional roles that H1 structural domains may play in adult viability. To this end, we used the transgenic alleles encoding various truncated and mutant H1 polypeptides to complement a “hypomorphic” RNAi allele that depletes H1 protein to ∼20% of the normal level (pINT-1-H1[2M]) when combined with Tub-GAL4 driver at 29 °C. In this system, the H1-specific dsRNA and rescue transgenes are simultaneously driven by Tub-GAL4.

Expression of the transgene encoding full-length H1 leads to a statistically significant rescue of the adult lethality (Table 2). In contrast, the nonsense transgene containing two in-frame stop codons in the coding sequence does not rescue adult viability, demonstrating that H1 mRNA expression is not sufficient, and translation of transgenic H1 protein is required to complement the H1 RNAi allele. Surprisingly, the transgene encoding the H1 ΔN mutant rescued viability with a similar efficiency as the full-length H1. However, transgenes encoding truncated H1 polypeptides lacking 50% or more of the CTD failed to complement the RNAi allele. Although a few phenotypically normal adult escapers eclosed in crosses with the transgenic allele encoding the 25% CTD truncation mutant, the rescue effect was not statistically significant. These results indicate that the C-terminal domain of H1 is essential for fly viability. However, the N-terminal domain of H1 is largely dispensable for Drosophila development during and after the larval stage.

TABLE 2.

Rescue of adult viability by transgenic variant H1 polypeptides

All crosses were performed at 29 °C. Males homozygous for UAS-driven H1 transgenes were mated to transheterozygous pINT-1-H1[2M]/SM5, Cy; Tub-GAL4/TM3, Sb females. Viability was scored as the number of eclosed Cy+, Sb+ adults relative to the total number of offspring scored (column 2). Percentage viability (column 3), compared with the expected number of Cy+, Sb+ flies (calculated from the Mendelian distribution), is shown in parentheses (column 2). Probability values are calculated by the χ² two-way test (columns 4 and 5). Statistically significant results (p < 0.05) are highlighted in boldface type. NA, not applicable; ND, cannot be determined.

H1 mutant transgene	Eclosed adults	Percentage expected	p relative to FL	p relative to 2×STOP
		%
2×STOP (S2*, M52*)	0/167 (42)	0	0.0075	NA
G3×MUT (K58A/K91A/K95A)	0/129 (32)	0	0.019	ND
FL (2 … 255)	5/119 (30)	17	N/A	0.0075
ΔN (44 … 255)	3/84 (21)	14	0.82	0.014
ΔC25% (2 … 224)	2/171 (43)	5	0.098	0.16
ΔC50% (2 … 183)	0/131 (33)	0	0.018	ND
ΔC75% (2 … 149)	0/157 (39)	0	0.0095	ND
ΔC100% (2 … 124)	0/122 (31)	0	0.022	ND

The C-terminal but Not the N-terminal Domain of H1 Is Required for Normal Polytene Chromosome Architecture

To examine the roles of H1 structural domains in chromosome structure, we analyzed the ability of the truncated H1 polypeptides to support the formation of phenotypically normal salivary gland polytene chromosomes. Strong depletion of H1 leads to marked changes in several features of polytene chromosome architecture and activity (19). For example, polytene chromosomes from H1-depleted larvae rarely exhibit the regular pattern of bands and interbands. The chromosomes are often tangled and appear as unstructured clumps of chromatin. We examined the ability of transgenes encoding full-length and mutant H1 transgenes to reverse these defects of polytene structure. Large numbers (>35) of DAPI-stained polytene chromosome spreads were prepared from progeny of crosses between the RNAi allele that depletes endogenous H1 below 10% of normal (19) and each transgenic allele. The polytene chromosomes were scored for the presence of a normal pattern of alternating bands and interbands (Fig. 4A and Table 3). We observed that expression of full-length H1 restored the regular pattern in ∼80% of salivary gland cells. No significant rescue was observed with the nonsense transgene containing two stop codons. Strikingly, H1 lacking its N-terminal domain exhibited nearly the same ability to rescue the phenotype as the full-length H1. Significant but only partial rescue was seen with H1 missing the distal 25% of its C-terminal domain. In contrast, H1 mutants missing 50% or more of their CTD failed to rescue the phenotype. These results show that the H1 CTD is required for the formation of normal polytene chromosomes. However, the H1 NTD appears to be dispensable for normal polytene chromosome morphology, which parallels our observation that it is not required for adult viability (Table 2).

FIGURE 4.

N- and C-terminal domains of H1 play roles in chromosome organization. A, the N-terminal fragment of the H1 CTD is required for a normal larval polytene chromosome structure. Salivary glands of L3 larvae with H1 depleted by RNAi and expressing the indicated transgenic H1 polypeptides were analyzed for polytene chromosome morphology. DAPI-stained polytene chromosomes exhibit abnormal, diffuse band-interband patterns and dissociated chromocenters upon abrogation of H1 expression. These defects are partially ameliorated in animals that express FL, ΔN, and ΔC25% H1 transgenes but not ΔC50%, ΔC75%, or H1 transgenes that are not deposited into chromatin (ΔC100% and G3×MUT). Representative images are shown; data from large numbers of samples are compiled in Table 3. B, the H1 NTD and the proximal fragment of the CTD are required for the establishment of a single HP1-containing chromocenter in salivary gland cells, whereas the proximal fragment of the CTD is additionally required for deposition of the H3K9me2 mark in pericentric heterochromatin. Whole mount salivary glands from transgenic L3 larvae expressing the indicated H1 proteins were fixed and stained with DAPI and antibodies to HP1 and H3K9me2. White arrowheads, positions of multiple HP1- or H3K9me2-containing foci. H1-depleted (H1-RNAi) and control RNAi (Nau-RNAi) animals served as positive and negative controls, respectively. The negative control and wild-type cells contain a single focus of HP1 staining and an overlapping focus of H3K9me2 staining. The positive control cells contain dispersed HP1 staining (two or more foci) and lack detectable H3K9me2 staining. Representative images are shown; data from large numbers of samples are compiled in Table 3.

TABLE 3.

Rescue of chromosome structure phenotypes by transgenic variant H1 polypeptides

All crosses were performed at 29 °C. Males homozygous for UAS-driven H1 transgenes were mated to transheterozygous pINT-1-H1[5M]; Tub-GAL4/T(2:3), Cy; TM6, Tb females. Tb⁺ L3 larvae were used for the analyses. Band-interband structure (column 2) was analyzed by DAPI staining of polytene chromosome spreads as in Fig. 4A. Single chromocenter (column 4) and H3K9 dimethylation (column 6) were examined by anti-H1 and anti-H3K9me2 IF staining, respectively, as in Fig. 4B. Numbers represent phenotypically wild-type individual polytene chromosomes or nuclei relative to the total number of chromosome spreads examined. Probability values (columns 3, 5, and 7) are calculated by the χ² two-way test. Statistically significant results (p < 0.05) are highlighted in boldface type. Negative control, yw males (no transgene). NA, not applicable.

H1 mutant transgene	Normal band-interband structure	p	Single chromocenter	p	H3K9 dimethylation	p
Negative control	4/71	NA	13/240	NA	5/168	NA
2×STOP (S2*, M52*)	5/63	0.63	37/385	0.060	4/179	0.66
FL (2 … 255)	42/53	4.7 × 10−17	232/303	1.7 × 10−81	147/202	4.8 × 10−42
ΔN (44 … 255)	29/37	6.8 × 10−18	23/283	0.22	159/207	1.3 × 10−46
ΔC25% (2 … 224)	18/52	3.4 × 10−5	117/289	1.3 × 10−36	67/179	1.3 × 10−36
ΔC50% (2 … 183)	5/59	0.53	23/298	0.29	11/192	0.21
ΔC75% (2 … 149)	3/47	0.87	10/162	0.75	7/102	0.13
ΔC100% (2 … 124)	6/49	0.20	18/219	0.23	3/107	0.93

Both the N- and C-terminal Domains of H1 Play Roles in the Organization and Epigenetic Modification of Pericentric Heterochromatin

A prominent feature of salivary gland nuclei is the chromocenter, a single region of underreplicated heterochromatin formed by coalescence of the pericentric portions of all chromosomes. The chromocenter is distinguished by the presence of heterochromatin protein 1 (HP1) and a high density of histone H3 dimethylated on lysine 9 (H3K9me2), which is usually a repressive histone mark (22, 23). Whole mounts of salivary gland cells stained for HP1 and H3K9me2 typically show a single, overlapping focus of nuclear staining for both features (Fig. 4B). These aspects of normal polytene chromosome structure are drastically perturbed in H1-depleted salivary glands. We observed a strongly diminished staining for H3K9me2 in the vast majority of H1-depleted cells (Fig. 4B and Table 3). Although staining for HP1 was still observed in H1-depleted salivary glands, the normal pattern of singular focal staining was disrupted in nearly all cells; instead, two or more HP1-enriched foci were observed.

We analyzed the ability of the transgenes encoding full-length and mutant H1 proteins to restore the formation of a single chromocenter. For each RNAi-transgene combination, we examined >100 individual whole mount salivary gland nuclei by DAPI and IF staining for H3K9me2 and HP1. We found that full-length H1 restored H3K9me2 staining in 72% of cells (Table 3). It also restored a single focus of HP1 staining in 77% of cases. No significant rescue was observed with the nonsense transgene. As in our analyses of polytene chromosome banding patterns, H1 proteins missing 50% or more of their CTD residues did not rescue H3K9 dimethylation or formation of a single chromocenter. However, H1 lacking the distal 25% of its CTD exhibited a significant ability to rescue both features. Interestingly, the H1 transgene missing its NTD was unable to restore the single HP1 focus in individual nuclei but exhibited a robust rescue of H3K9 dimethylation (Fig. 4 and Table 3). Notably, the H3K9me2 IF signal co-localized with that for HP1 and appeared in multiple foci, a phenotype never observed in wild-type or H1-depleted cells. (No significant difference in the intensity of H3K9me2 staining was observed with the ΔN transgene as compared with the FL transgene.)

These results indicate that the H1 CTD plays a dominant role in several aspects of normal polytene chromosome structure. On the other hand, the H1 NTD is dispensable for polytene chromosome morphology and for promoting pericentric H3K9 dimethylation, but it appears to be important for organizing HP1-containing chromatin into a single chromocenter. Interestingly, the rescue of pericentric H3K9 methylation by the N-terminally truncated H1 transgenic protein (ΔN) is not sufficient to reinstate the single chromocenter. Furthermore, because the H1 ΔN transgene readily rescues adult fly viability (Table 2) but does not significantly ameliorate the defect in forming a single chromocenter (Table 3), it appears that the formation of a chromocenter in endoreplicating salivary gland cells is not essential in flies.

Physical Interactions with HP1 and Su(var)3-9 Individually Require the Globular and C-terminal Domains of H1, Respectively

Taken together, the foregoing results indicate that the H1 CTD plays a dominant role in many biological functions of H1. In contrast, so far, we have only detected a role for the H1 NTD in coalescence of HP1-containing pericentric heterochromatin into a single chromocenter, a function to which the CTD also contributes. We sought to examine the putative separate roles that these individual domains may play in the molecular functions of H1, such as physical interactions with downstream effector proteins. For instance, we have previously observed that the H1 CTD is sufficient for binding to Drosophila STAT92E, which together with H1 coordinates a unique pathway of heterochromatin organization in flies (42). H1 linker histones have been also shown to interact with HP1 (7, 24, 25). Finally, we reported that Drosophila H1 interacts directly with the heterochromatin-specific histone methyltransferase Su(var)3-9 and cooperates with it to mark heterochromatin with H3K9me2 (7). Therefore, we determined whether the mutant H1 proteins used in this study interact with HP1 and Su(var)3-9 (Fig. 5A).

FIGURE 5.

The H1 C-terminal domain is required and sufficient for binding to Su(var)3-9, and the H1 globular domain is sufficient for binding to HP1. A, GST pull-down analyses of direct physical interactions of HP1 and Su(var)3-9 with truncated H1 polypeptides. Recombinant truncated H1 polypeptides were analyzed by GST pull-downs with GST, GST-HP1, and GST-Su(var)3-9 fusion proteins. Coomassie-stained SDS-PAGE gels are shown to demonstrate equivalent loading and efficient pull-down of GST fusion proteins. Co-purifying H1 polypeptides are analyzed by anti-H1 Western blotting. Loading control (10% of material used for incubation with GST fusions) is shown as a separate panel. Positions of protein molecular mass markers (kDa) are shown beside the panels. The binding of Su(var)3-9 to H1 is not affected by a deletion of the N terminus. However, the deletion of more than 50% of the H1 C terminus abolishes the binding to Su(var)3-9. The binding of HP1 to H1 is not affected by the deletion of either the entire C terminus or N terminus of H1. B, GST pull-down analyses of direct physical interactions of HP1 and Su(var)3-9 with individual structural domains of H1. Purified recombinant HP1 and Su(var)3-9 proteins were analyzed by GST pull-downs with fusions of individual H1 domains and GST. Coomassie-stained SDS-PAGE gel is shown to demonstrate equivalent loading and efficient pull-down of GST fusion proteins. Co-purifying FLAG-HP1 and His₆-Su(var)3-9 are analyzed by anti-HP1 or anti-His₆ Western blotting. Loading control lanes (10 or 5% of material used for incubation with GST fusions) are included on Western blots. Positions of protein molecular mass markers (kDa) are shown beside the panels. FLAG-HP1 and His₆-Su(var)3-9 bands are indicated by arrows. Consistent with the results in A, the globular domain of H1 is sufficient for binding to HP1. Similarly, the C-terminal domain of H1 is sufficient for binding to Su(var)3-9.

Recombinant full-length and truncated mutant H1 proteins were tested for interactions with GST-HP1 and GST-Su(var)3-9 by pull-down assays followed by immunoblotting for H1. Deleting the NTD or the distal 25% of the CTD did not eliminate the binding of H1 to Su(var)3-9. However, truncating the H1 CTD at residue 183 (in the ΔC50% polypeptide) greatly diminished Su(var)3-9 binding. Therefore, residues 184–224 of the H1 CTD are required for interaction with Su(var)3-9. In contrast, the deletion of either the H1 NTD or CTD failed to abolish physical interactions between H1 and HP1 (Fig. 5A). To further delineate the protein binding region(s) of H1, we utilized GST fusions with full-length H1 and its individual globular and N- and C-terminal domains (42) for pull-down experiments with purified recombinant HP1 and Su(var)3-9 proteins (Fig. 5B). Consistent with our analyses of truncated H1 polypeptides, the H1 globular domain alone and the H1 CTD alone exhibited strong physical interactions with HP1 and Su(var)3-9, respectively. These interactions were indistinguishable from those made by the full-length H1 protein. The results of these protein-protein interaction studies provide a molecular explanation for the observed in vivo role for the H1 CTD in H3K9me2 marking of heterochromatin. Apparently, the Su(var)3-9 binding interface of the H1 CTD, including residues within the 184–224 fragment, promotes the recruitment of this histone methyltransferase to heterochromatin and dimethylation of H3K9.

The C-terminal Domain of H1 Is Involved in Silencing Transposable Elements and Repetitive Sequences in Heterochromatin

Eukaryotic genomes harbor transposable elements and other types of repetitive DNA sequences, many of which reside in pericentric heterochromatin (26). Transcription of these sequences needs to be silenced for genome integrity. We described previously a role for Drosophila H1 in transcriptional repression of these sequences (7). We found that Drosophila H1 and the heterochromatin-specific histone methyltransferase Su(var)3-9 interact and cooperate to mark heterochromatin with H3K9me2, resulting in silencing of the repetitive elements. Depletion of H1 in Drosophila salivary glands leads to activation of transposable element transcription associated with the loss of H3K9me2 from heterochromatin. We used quantitative RT-PCR assays to assess the ability of transgenes encoding full-length and truncated H1 proteins to restore the repression of these elements (Fig. 6A). Activation of these elements is very strong, providing a large dynamic range for assaying the repression activities of the H1 transgenes. We observed that expression of full-length H1 transgenic protein largely restored repression of each of the repetitive sequences studied. On the other hand, H1 proteins missing 50% or more of the CTD exhibit no repression activity. The H1 polypeptide missing 25% of its CTD exhibited substantial repression activity, although it was not able to fully repress transcription. In contrast, H1 transgenic protein lacking its NTD showed as much repression activity as the full-length H1. We conclude that the H1 CTD plays an essential role in silencing transposable elements in heterochromatin, whereas the H1 NTD is dispensable for this function. These results are fully consistent with the central role of the CTD and the dispensability of the NTD for the physical interaction of H1 with Su(var)3-9 in vitro (Fig. 5) and the marking of heterochromatin with H3K9me2 in vivo (Fig. 4).

FIGURE 6.

The H1 C-terminal but not the N-terminal domain of H1 is required for repression of transposable elements. A, the expression of a subset of transgenic H1 polypeptides partially ameliorates derepression of transposable/repetitive elements observed upon depletion of endogenous H1 in vivo. The transcripts were analyzed by quantitative RT-PCR of RNA isolated from salivary glands of L3 larvae with H1 depleted by RNAi (H1 KD) and expressing the indicated transgenic H1 polypeptides. All numbers were normalized to the values for rp49 RNA. -Fold changes were calculated as a ratio of signals for the indicated samples relative to a control line with depleted control protein, Nautilus. Regulation of the αTub84B gene was used as a control. Error bars from the S.D. of triplicate PCRs of two experiments are too small to be observed on the graph scale. B, schematic of individual contributions of isolated H1 structural domains to its functions in vivo and in vitro. A schematic diagram of the H1 protein with its individual structural domains (NTD, GD, and CTD) of H1 is shown in the center of the figure. The CTD is divided into four equal sized segments representing the regions sequentially removed in the four CTD truncation mutants used in this study. The biochemical properties of H1 and the biological processes affected by H1 that were studied in this report are represented as ovals, except for fly viability, which is represented by a large, rounded rectangle. The regions of H1 required for its roles in these functions are connected to the ovals by arrows. The broken arrow indicates that the 75% CTD truncation affects the stability of H1 binding to chromatin in vitro, although this truncated polypeptide does bind to chromatin in vivo. Adult fly viability requires both the GD and CTD. See “Results” for further details.

Discussion

The past 20 years have seen an enormous increase in our understanding and appreciation of the impact of chromatin structure on gene regulation and development. Much of this progress has focused on the role of the core histones, especially their post-translational modifications that mark chromatin and influence its interactions with many types of regulatory proteins (reviewed in Refs. 27 and 28). As mentioned, linker histones are nearly as abundant as core histones in chromatin. Their binding to chromatin is also much more dynamic than nucleosome core particles, providing them with significant regulatory potential. The tripartite structure (NTD, GD, and CTD) that comprises most H1 linker histones has been conserved in metazoans. Our current knowledge about the roles of these individual structural domains in the many biological functions of H1 in vivo is quite limited. In part, this is because genetic approaches for studying H1 have been difficult to develop. The H1 counterpart in Saccharomyces cerevisiae is not an essential protein (29). On the other hand, the existence of multiple, compensating H1 isoforms in many metazoans has slowed progress in the H1 field (30, 31). The current study was designed to overcome these problems by taking advantage of fly genetics and the fact that D. melanogaster expresses a single linker histone protein during much of its development. Our previous findings showed that H1 is essential for fly development beyond the late larval stage and that it is required for normal chromosome architecture (19). Moreover, we found that H1 plays a particularly prominent role in the structure and genetic activity of heterochromatin, where it participates in epigenetic marking and transcriptional silencing of transposable elements through its interaction with the Su(var)3-9 methyltransferase and STAT92E (7, 19, 42). By generating transgenic fly lines expressing full-length H1 and six different mutant H1 polypeptides, we interrogated its structural domains for their roles in multiple H1-mediated processes in vivo.

We expressed mutant H1 proteins lacking the most C-terminal 25, 50, and 75% of CTD residues as well as a mutant H1 completely lacking the CTD. All of these mutant proteins localized to the nucleus, and the three partial CTD truncation mutants were incorporated into chromatin (Fig. 2). H1 entirely lacking the CTD, however, could not be detected in polytene chromosomes (Fig. 2A), similar to the point mutant harboring mutations of three key GD residues previously characterized as essential for H1 binding to chromatin in vitro. Although three of the four CTD mutants were assembled into chromatin in vivo, none of them were able to complement the lethal effects of depleting H1 in vivo (Table 2). Because the 25% CTD truncation mutant did exhibit partial rescue activity in our other assays, its failure to complement the adult lethality (in a statistically significant manner) is presumably due to the very stringent requirements for H1 functions in the multiplicity of complex processes involved in fly development from larvae to adult, one or more of which must require full H1 functionality. The results of these studies point to particularly important roles for the H1 CTD in these processes (summarized in Fig. 6B).

Importantly, using assays for individual biological and biochemical functions of H1, we observed clear differences in the ability of the 25% CTD truncation, versus the other CTD truncation mutants, to rescue specific molecular phenotypes. H1 lacking the distal 25% of the C terminus can clearly rescue several specific defects associated with H1 depletion, namely disruption of polytene chromosome structure and the organization and marking of pericentric heterochromatin as well as the derepression of transposable elements. In contrast, more extensive truncation of the CTD leads to complete loss of these functions. At the molecular level, the differences in the activity of the 25% CTD truncation mutant versus the other CTD truncation mutants in heterochromatin functions is probably due to their differential binding to Su(var)3-9, the histone methyltransferase responsible for marking pericentric heterochromatin with H3K9me2. Indeed, the H1-Su(var)3-9 protein interaction studies indicate that the region between residues 184 and 224 of the H1 CTD, which is removed in the 50% truncation mutant, is critical for H1 binding to Su(var)3-9 (Figs. 5 and 6B).

On the other hand, differences in the mutants binding to Su(var)3-9 presumably do not account for the differences in their ability to restore disrupted polytene chromosome structure. More likely, differences in chromatin binding properties of the mutant polypeptides explain their different abilities to rescue perturbations in chromosome structure. Consistent with this view, using the in vitro chromatin assembly assay, we observed that the 25% CTD truncation mutant stably associates with chromatin, whereas more extensive truncations of the CTD lead to increasing instability of H1 association with chromatin (Fig. 3B). These in vitro findings are consistent with FRAP experiments demonstrating the influence of the CTD on the chromatin-bound residence time of mammalian H1s in cultured cells (15).

The prominent role of the CTD in multiple biological functions of Drosophila H1 described here adds to a small but growing list of reports indicating the importance of this domain in processes mediated by mammalian H1 subtypes (reviewed in Ref. 32). For example, a set of studies on the role of the murine H1(0) subtype in chromatin folding in vitro suggests that this function resides in the unique amino acid composition of two subdomains within the CTD (17, 18). The H1(0) CTD was also reported to bind and stimulate the activity of the apoptotic nuclease, DNA fragmentation factor DFF40/caspase-activated DNase (33). Recently, we reported that certain murine H1 subtypes directly interact with DNA methyltransferases DNMT1 and DNMT3B, recruiting them to chromatin to promote DNA methylation and gene silencing (8). A high degree of specificity was observed for these interactions because only certain H1 subtypes interact with these DNA methyltransferases. Strikingly, the specificity among H1 subtypes maps to the CTD. It has also been reported that p53-mediated transcriptional repression is dependent on interaction of p53 with human H1.2, which is mediated by the H1.2 CTD (34).

In contrast to the developing understanding of the H1 CTD functions, virtually nothing is known about the biological roles of the H1 NTD. Important functions for this domain are to be expected, first because it is present in all metazoan H1 subtypes and second because it harbors a large portion of the post-translational modifications described in metazoan (including Drosophila) H1 proteins (35, 36). Our identification of a role for the NTD in promoting the formation of a single chromocenter in endoreplicating salivary gland cells begins to fill this gap. The coalescence of pericentric heterochromatin in a single chromocenter of polytenized chromosomes is probably linked to its severe underreplication (37). We now show that formation of a single chromocenter requires the N terminus of Drosophila H1. Therefore, the H1 NTD may contribute to the recruitment or regulation of additional, possibly yet unknown, effectors of endoreplication. Despite the apparent lack of a requirement for the NTD to rescue other phenotypes analyzed in this study, it remains formally possible that the presence of a small amount of endogenous, wild-type protein remaining in H1 knockdown flies interferes with the results. Thus, future experiments in a true His1 null background will be required to reveal any such roles if the NTD is indeed required for the cognate processes.

It is expected that the globular domain of H1 should play a crucial role in its functions in vivo because it is the region most intimately involved in binding to nucleosome core particles in vitro (12, 13, 16, 38). Consistent with these in vitro studies, when we expressed a mutant H1 polypeptide containing alanine substitutions in three key residues in the globular domain, it was not loaded into chromatin in vivo (Fig. 2A). Furthermore, the globular domain mediates the physical interaction between Drosophila H1 and HP1 (Fig. 5B). Thus, the biochemical separation of these H1 functions (chromatin and HP1 binding) is difficult to achieve. In the future, it will be interesting to identify individual sequences within the H1 globular domain that are required for HP1 binding but are dispensable for H1 incorporation into chromosomes and vice versa. Interestingly, binding of HP1 to human H1.4 was reported to depend on methylation of a specific lysine residue (Lys-26) in the NTD (24), whereas our work provides no evidence for an interaction between the H1 NTD and HP1. However, there is no strong sequence conservation between the NTDs of Drosophila H1 and human H1.4. Nevertheless, the approaches implemented in the current work will allow investigation of the role of post-translational modifications in this and other biological functions of H1 and will provide additional insights into how H1 and HP1 cooperate in the formation of pericentric heterochromatin.

Experimental Procedures

Transgenic Fly Strains

pUASTattB-FLAG-H1 transgenes encoding full-length and mutant H1 variant proteins were constructed by PCR amplification of the corresponding H1 cDNA sequences and cloning into the EcoRI and XhoI sites of pUASTattB (39). Primer sequences are available upon request. A sequence encoding the FLAG epitope was inserted in-frame at the N terminus of the H1-coding sequences by PCR. The constructs were sequenced across the cloned PCR fragments. All pUASTattB-FLAG-H1 transgenes were inserted at the 86Fa attP site on the right arm of the third chromosome by PhiC31 integrase-mediated transgenesis. Embryo injections were performed by Bestgene, Inc. in yw M{eGFP.vas-int.Dm}ZH-2A; M{RFP.attP}ZH-86Fa background.

Fly Genetics

Flies were grown on standard corn meal, sugar, and yeast medium with Tegosept. Stocks were maintained at 18 °C. Crosses were performed in an environmental chamber at 29 °C. Flies carrying the pINT-1-H1[2M] and pINT-1-H1[5M] P-element insertions used to deplete H1 were described previously (19). They encompass a hairpin of two fragments of the first 600 bp of the D. melanogaster H1-coding sequence (encoding the first 200 amino acids of the 256-amino acid, full-length H1 protein) in opposite orientations on both sides of the first intron of the actin 5C gene (Fig. 1A) under the control of the UAS promoter. For studies of viability, polytene chromosome staining, and isolation of RNA, all animals were incubated at 29 °C throughout their life cycles.

For studies of rescue of viability by H1 transgenes, pINT-1-H1[2M]/SM5, Cy; Tub-GAL4/TM3,Sb flies were generated by a series of crosses and maintained at 18 °C. The males were mated at 29 °C to yw females or females homozygous for the pUASTattB-FLAG-H1 transgene insertions. The number of Cy⁺; Sb⁺ adult flies were counted each day for 7 days from the onset of eclosion, and the numbers were compared with the expected number. For studies of H1 transgene rescue of polytene chromosome structure, pericentric heterochromatin properties, and transposable element silencing, pINT-1-H1[5M]; UAST-FLAG-H1/T(2:3) Cy, TM6B, Tb flies were generated by a series of crosses and maintained at 18 °C. The males were mated at 29 °C to Tub-GAL4/TM6b, Tb females. Salivary glands and total RNA from Tb⁺ third instar larvae were isolated and used for the analyses.

Immunohistochemistry

Salivary glands of wandering third instar larvae were dissected in PBS + 0.1% Triton X-100. For whole mount staining, the glands were fixed for 15 min in PBS containing 3.7% formaldehyde, washed in PBS + 0.1% Triton X-100, and permeabilized by treatment for 1 h with PBS + 1% Triton X-100. Alternatively, to prepare polytene chromosomes, the glands were fixed in 3.7% paraformaldehyde for 30 s, squashed in 45% acetic acid + 3.7% formaldehyde, and frozen in liquid nitrogen.

The whole mount glands or polytene chromosome spreads were incubated overnight in PBS + 10% fat-free milk + 0.1% Triton X-100 with primary antibodies at the indicated dilutions: monoclonal mouse anti-Drosophila HP1, C1A9 (1:50; Developmental Studies Hybridoma Bank); monoclonal mouse anti-FLAG (1:1,000; Sigma); affinity-purified rabbit anti-Drosophila H1 (1:5,000 (19); affinity-purified rabbit anti-H3K9me2 (1:100; Abcam). DNA was stained by adding 1.5 μg/ml DAPI (Vectashield) to the mounting medium. The preparations were washed twice in PBS + 400 mm NaCl + 0.2% Nonidet P-40 for 30 min. Appropriate Cy2- and Cy3-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) were used at 1:200. The specificity of staining was verified by appropriate controls, including staining with secondary antibodies alone and staining of polytene chromosome spreads from H1 knockdown animals. Fluorescent images were acquired on a Zeiss Axioplan microscope equipped with a Zeiss digital microscopy camera AxioCam ICC1 and AxioVision digital image processing software. Stereoscopic images were acquired on a Zeiss Stereo Discovery V8 microscope.

Immunoblot Analyses

5 larvae of each genotype were dissected and homogenized in 200 μl of Laemmli loading buffer. The lysates were boiled for 5 min and centrifuged. 10 μl of lysate was loaded on a 15% SDS-polyacrylamide gel. Equal protein loading was verified by Coomassie staining of equivalently loaded SDS-polyacrylamide gels. Proteins were transferred to PVDF membranes using an electroblot apparatus at 200 mA for 1 h. The membranes were blocked for 1 h in blocking buffer (LI-COR Biosciences). Rabbit anti-H1 (1:5,000) and mouse monoclonal anti-FLAG (1:500) antibodies were incubated with the membranes overnight. Subsequently, the blots were washed in PBS + 0.1% Tween 20 and incubated with infrared dye-labeled secondary antibodies (1:15,000; LI-COR Biosciences).

Measurements of the expression levels of transgenic FLAG-H1 relative to that of endogenous H1 (Table 1) were performed by immunoblotting (Fig. 1B) in two independent experiments. Images were obtained using the LI-COR Odyssey infrared imaging system. The intensities of bands (gray mean values) were measured in ImageJ software. The ratios of intensity values for endogenous and transgenic H1 (H1 and FLAG signals) were calculated using the rectangular grid tool, and the average values ± S.D. were calculated.

Quantitative Real-time RT-PCR

Total RNA from 10 pairs of salivary glands from larvae of each genotype was isolated using TRIzol extraction (Invitrogen) and quantified with a NanoDrop 1000 Spectrophotometer (Thermo Scientific). One μg of total RNA was treated with RNase-free DNase I (Promega), and random-primed cDNA was prepared using the SuperScript II kit (Invitrogen). Real-time quantitative PCR amplifications were carried out in ABI Prism 7700 sequence detection system (Applied Biosystems). One-step RT-PCR was done using a SYBR Green quantitative RT-PCR kit as per the manufacturer's instructions. To quantify the expression levels, cycle threshold values of an endogenous reference gene, rp49, were included. All crosses and RNA extraction procedures were performed twice or more. All RT-PCRs were carried out in triplicate, along with no-template controls.

Recombinant Proteins

Recombinant Drosophila His₆-Su(var)3-9 and FLAG-HP1 proteins were purified as described (40, 41). Full-length Drosophila H1 and truncated H1 polypeptides were cloned into NdeI and SalI sites of pET24-iCBD vector in-frame with intein and chitin binding domain. The vector sequence and map are available upon request. The constructs were transfected into Rosetta 2 cells, and the proteins were induced overnight at 16 °C with 0.42 mm isopropyl 1-thio-β-d-galactopyranoside. The H1 fragments fused to intein and CBD were purified using chitin beads (New England Biolabs) using the manufacturer's instructions (Impact-CN expression system). H1 polypeptides were further purified by FPLC on Source 15S Sepharose (GE Healthcare).

GST fusions of H1; H2A; Su(var)3-9; HP1; full-length H1; and H1 globular (residues 41–119), N-terminal (residues 1–40), and C-terminal (residues 120–256) domains were expressed in E. coli (BL21(DE3)pLys) and purified by glutathione-Sepharose chromatography as described (7, 42). The purified proteins were analyzed by SDS-PAGE, and protein concentrations were determined by Coomassie staining along with BSA protein mass standards (Pierce). All GST-H1 polypeptides (full-length and fragments) were subjected to an additional round of purification by FPLC on a Source 15S column.

Protein-Protein Interaction Studies

In GST pull-down assays, purified recombinant H1 fragments were incubated with GST or GST fusion proteins (HP1 and Su(var)3-9) for 90 min at 27 °C and purified on glutathione-Sepharose as described (42). H1 fragment binding to GST fusion proteins was detected by anti-H1 Western blotting of the pull-down samples. Additionally, the pull-down samples were examined for the presence of GST fusion proteins by SDS-PAGE and Coomassie staining. Binding of GST-H1 fragments to FLAG-HP1 and His₆-Su(var)3-9 was detected by anti-FLAG (1:1,000; Sigma) or anti-His₆ (1:2,500; Sigma) Western blotting of the pull-down samples.

Chromatin Assembly and Chromatosome Stop Assay

In vitro chromatin assembly reactions were performed exactly as described (43). Briefly, 70-μl reactions containing 2.9 nm DNA (pGIE-0 plasmid; ∼3.2 kbp), 64–80 nm each native Drosophila core histone, 0–40 nm H1 polypeptides, 800 nm recombinant Drosophila NAP-1, and 5 nm recombinant Drosophila ACF were incubated in a buffer containing ∼2 ng of recombinant Drosophila topoisomerase I fragment ND423, 3 mm ATP, and ATP regeneration system for 2 h at 27 °C. The reaction products were analyzed by partial micrococcal nuclease digestion, and DNA was run on a 1.25% agarose gel in Tris/borate/EDTA buffer and stained with ethidium bromide. Core histone and H1/DNA mass ratios were optimized in a series of reactions as described (43) to achieve extended arrays of nucleosomes and H1 abundance of ∼1:1 relative to the core particle. For the chromatosome assay, 4% low melting agarose gels were used.

Sucrose Gradient Sedimentation Analysis of Chromatin

Sucrose gradient sedimentation assays were performed as described (44). Briefly, 10 standard 70-μl reactions were loaded on 5–30% sucrose gradients in 25 mm HEPES, pH 7.6, 200 mm KCl, 0.1 mm EDTA, pH 8.0, 1 mm DTT and centrifuged in a Beckman SW-41 rotor for 18 h at 41,000 rpm. The gradients were cut into 10 fractions and analyzed by Western blotting for the presence of H1. The following antibodies were used: rabbit anti-H1 (1:5,000) and secondary HRP-conjugated donkey anti-mouse (Jackson ImmunoResearch; 1:5,000).

As a reference, similar chromatin assembly reactions without H1 were also subjected to sucrose gradient sedimentation. Gradient fractions were deproteinated by treatment with Proteinase K and phenol/chloroform extraction, ethanol-precipitated, and examined for the presence of DNA by agarose gel electrophoresis and staining with ethidium. The fractions were also TCA-precipitated and analyzed for the presence of core histones on Coomassie-stained SDS-PAGE.

Control sucrose gradient sedimentation assays were performed with 1 μg of recombinant full-length H1 protein alone or after a 10-min incubation (at 0 °C) with 40 μg of recombinant Drosophila NAP-1 and 1.3 μg of recombinant Drosophila ACF or 5 μg of plasmid DNA in 0.7 ml of 25 mm HEPES, pH 7.6, 200 mm KCl, 0.1 mm EDTA, pH 8.0, 1 mm DTT. Individual truncated H1 polypeptides were also analyzed by sucrose gradient sedimentation to verify the lack of abnormal protein aggregation.

Author Contributions

H. K. and A. V. E. performed the experiments and analyzed the results; D. V. F. and A. I. S. conceived the experiments, analyzed the results, and wrote the manuscript. All authors reviewed the manuscript.