Many pleiotropic roles have been ascribed to small abundant HMG–Box (HMGB) proteins in higher eukaryotes but their precise function has remained enigmatic. To investigate their function genetically we have generated a defined deficiency uncovering the functionally redundant genes encoding HMGD and HMGZ, the Drosophila counterparts of HMGB1–3 in mammals. The resulting mutant is a strong hypomorphic allele of HmgD/Z. Surprisingly this allele is viable and exhibits only minor morphological defects even when homozygous. However, this allele interacts strongly with mutants of the Brahma chromatin remodeling complex, while no interaction was observed with mutant alleles of other remodeling complexes. We also observe genetic interactions between the HmgD/Z deficiency and some, but not all, known Brahma targets. These include the homeotic genes Sex combs reduced and Antennapedia, as well as the gene encoding the cell-signaling protein Rhomboid. In contrast to more general structural roles previously suggested for these proteins, we infer that a major function of the abundant HMGB proteins in Drosophila is to participate in Brahma-dependent chromatin remodeling at a specific subset of Brahma-dependent promoters.

---

CHROMATIN structure has important consequences for all nuclear processes, be they transcription regulation, replication, or DNA repair. In addition to histones, chromatin contains DNA-binding nonhistone proteins. One of the most abundant groups of chromatin-binding proteins are the high mobility group (HMG) proteins. These proteins make up several unrelated classes including the HMGA, HMGB, and HMGN proteins. HMGA are structurally flexible proteins that bind to A/T-rich DNA sequences and promote the binding of other transcription factors (REEVES 2003). HMGN proteins bind nucleosomes and promote chromatin decondensation (BUSTIN 2001).

The HMG–Box (HMGB) class of proteins is characterized by a DNA-binding domain or box, which binds DNA nonspecifically and induces sharp bends and consequently has a preference for distorted or prebent structures. This domain is utilized in many contexts, e.g., sequence-specific transcription factors and subunits of chromatin remodeling complexes (THOMAS and TRAVERS 2001; TRAVERS 2003). However, a conserved group of small, highly abundant HMGB proteins remains, consisting of one or two HMGB domains followed by a stretch of basic residues and an acidic tail. The function of this group is poorly defined.

In vitro studies have examined the role of HMGB proteins in stimulating transcription factor binding to their target DNA sequences (THOMAS and TRAVERS 2001). Increased rates of binding were observed, especially in the case of factors requiring prebent DNA, such as TATA-binding protein (TBP). These observations lead to the proposition that HMGB proteins act primarily as architectural chromatin factors.

The most abundant HMGB proteins in yeast are encoded by the functionally redundant genes Nhp6a and Nhp6b. Mutation of both genes results in a temperature-sensitive growth phenotype and misexpression of a small percentage of genes (MOREIRA and HOLMBERG 2000). Genetic interactions with TBP and the SAGA complex, required for activation of the HO gene, have been observed (YU et al. 2000, 2003; BISWAS et al. 2004). In addition, Nhp6a/b forms part of the yeast FACT complex that facilitates transcription elongation (FORMOSA et al. 2001).

Studies of HMGB proteins in mammals have been complicated by issues of redundancy. In mice three highly related HMGB proteins (HMGB1–3) exist, all ubiquitously expressed during embryogenesis, but with differing spatial expression patterns in adults. Single mutants in Hmgb1 and Hmgb2 show no embryonic defects, but Hmgb1 mutants die 24 hr after birth from hypoglycemia and Hmgb2 knockout mice exhibit reduced male fertility (CALOGERO et al. 1999; RONFANI et al. 2001). Hmgb3 mutants are viable but exhibit erythrocythemia, with a 10% increase in the number of red blood cells (NEMETH et al. 2003, 2005). Thus far, to our knowledge, no double or triple mutants of HMGB1-3 have been generated and therefore it has not been possible to address the issue of redundancy in mice.

In Drosophila small, abundant, non-sequence-specific HMGB proteins include Dorsal switch protein 1 (Dsp1) (MOSRIN-HUAMAN et al. 1998; DECOVILLE et al. 2001), HMGD and HMGZ (WAGNER et al. 1992; NER et al. 1993), and the products of two small, uncharacterized genes termed CG7045 and CG7046. Null mutants of Dsp1 exhibit homeotic transformations relating to downregulation of Sex combs reduced (Scr) expression (DECOVILLE et al. 2001). Biochemical studies have observed competition between HMGD and histone H1 in binding to chromatin, implying that there may be a global role for small abundant HMGB proteins in the regulation of chromatin structure (NER and TRAVERS 1994; NER et al. 2001).

Using rearrangement screen (RS) P-elements available from the DrosDel consortium, we have produced a defined deficiency of HmgD and HmgZ in flies, resulting in a strong hypomorphic allele of these genes. In conflict with the hypothesized role in chromatin architecture, this HmgD/Z mutant is viable and displays no defects in global chromatin organization. Instead we observe strong genetic interactions with members of the Brahma remodeling complex, providing evidence for a role in activation of two Hox genes, Antennapedia (Antp) and Sex combs reduced (Scr), and also of rhomboid (rho), which is involved in EGFR signaling.

Fly stocks and crosses:

All stocks were maintained on standard medium at 25°. All alleles and stocks used are as described in FlyBase (http://flybase.bio.indiana.edu), with the following exceptions: Df(2R)ED3921 and Df(2R)ED50000 (this work) and Df(2R)ED3923 (DrosDel consortium, http://www.drosdel.org.uk). Stocks containing RS elements 5-HA-1624, 5-SZ-3389, CB-5352-3, and CB-0090-3 were obtained from Szeged stock center; Dsp1¹ was kindly provided by D. Locker, Acf1 alleles by J. Kadonaga, and Nurf301 [otherwise known as E(bx)] and Iswi alleles by P. Badenhorst.

Mounting fly wings:

Wings from adult flies were dissected in ethanol and mounted in a 6:5 lactic acid:ethanol mixture.

Northern blot analysis and RT–PCR:

Northern blot analysis was carried out using standard procedures. Total RNA was extracted using Trizol and mRNA was purified using the Micro-FastTrack 2.0 mRNA isolation kit (Invitrogen, San Diego), according to manufacturers' instructions. Blots were hybridized with a radiolabeled full-length HMGZ cDNA clone, in ULTRAhyb hybridization buffer (Ambion, Austin, TX), according to manufacturer's instructions, and bands were visualized using a phosphorimager. For RT–PCR, exon-flanking primers were designed for regions common to splice variants of HmgD and HmgZ, and reactions were performed using the SuperScript III One-Step RT–PCR System (Invitrogen), according to manufacturer's instructions. RT–PCR primer sequences are as follows: Act5C 5'-GACACCAAACCGAAAGACTT, Act5C 3'-ACCCACGTACGAGTCCTTCT, HMGD 5'-CCCGCGAGTCGATCAAGC, HMGD 3'-CGGACTCCTCCTTCTTGCTCT, HMGZ 5'-AACGAGACCCGCGAACAG, HMGZ 3'-CTTCTTGGCTGGCTTGGC, CG30403 5'-TACCACCAGGAGCTAAAGAAGA, and CG30403 3'-AAATAAGGAAGAGGGTGTAGCAT.

Glycerol gradients:

Nuclear extract was prepared from an overnight collection of embryos as previously described (KAMAKAKA et al. 1991). Nuclear extract (200 µl) was fractionated using 30–5% glycerol gradients, by centrifugation at 48,000 rpm for 16 hr in a Beckman ultracentrifuge using a SW-60 rotor. Fractions (100 µl) were collected and separated by SDS–PAGE. HMGD and HMGZ were detected by an antibody raised against the HMGB domain of HMGD (rat anti-HMGD-100, 1:250). Brahma was detected with a rabbit polyclonal antibody (gift from C. P. Verrijzer) described elsewhere (KAL et al. 2000) (rabbit anti-Brm 1:2000).

Generation of HmgD and HmgZ deficiency:

The HmgD and HmgZ genes are located on chromosome 2R, at 57F9 and 57F10, 20 kbp apart. Both HmgD and HmgZ produce alternatively spliced mRNA products (WAGNER et al. 1992; NER et al. 1993). Between the two genes lies an uncharacterized ORF, CG30403. DrosDel RS elements within the HmgZ and HmgD genes were selected to produce two deficiencies, as shown in Figure 1A. The first, Df(2R)ED3921, removes the first exon of HmgZ, the CG30403 gene, and the first exon of HmgD. This was generated using RS elements 5-HA-1624 and CB-5352-3.

View larger version (41K): In this window In a new window Download PPT slide   FIGURE 1.— (A) Schematic overview of HmgZ and HmgD gene region. Positions of RS elements are marked, with arrows signifying the orientation of the FRT sequence. Annotated transcription start sites are shown for HmgZ, CG30403, and HmgD. The solid boxes represent exons. The figure is not drawn to scale. 5-HA-1624 and CB-5352-3, and 5-SZ-3389 and CB-0090-3, were used to generate Df(2R)ED3921 and Df(2R)ED50000, respectively. (B) Western blot of Df(2R)ED50000 and Df(2R)ED3921 embryos. Increasing amounts of wild type (wt) and of either of the deficiency (Df) embryos were probed with an antibody against the HMGB domain of HMGD/Z (rat anti-HMGD-100, 1:250). One microliter equals one embryo. (C) Qualititative RT–PCR of HmgD and HmgZ transcripts in Df(2R)ED3921 flies compared to wild type. Z, HmgZ; D, HmgD; CG, CG30403; act, Act5C. Lanes 1–4: wild-type flies. Lanes 5–8: Df(2R)ED50000. Lanes 9–12: Df(2R)ED3921. M, 1-kb marker (Roche, Indianapolis). RT–PCR primers were designed to common spliced exons to generate a single product, and RT–PCR was performed at high cycle number (30 cycles). Df(2R)ED50000 and Df(2R)ED3921 flies show loss of HmgD and CG30403 transcript but retain HmgZ transcript. (D) Northern blot analysis of HmgZ transcripts in Df(2R)ED3921 larvae and adult flies. Blots were probed with a radiolabeled cDNA clone of HMGZ. Hybridization signals of both alternatively spliced HmgZ transcripts (HMGZ RA and HMGZ RB) are 60 and 30% in Df(2R)ED3921 larvae and flies, respectively, compared with wild-type signal intensity.

 
The second deficiency, Df(2R)ED50000, removes CG30403 and the first exon of HmgD. This was made using RS elements 5-SZ-3389 and CB-0090-3 and was used as a control for all further experiments to rule out an effect of CG30403 in genetic interactions.

Flies homozygous for Df(2R)ED50000 are viable and display no observable phenotype. No interaction with Df(2R)ED3921 or a with larger deficiency spanning HmgZ and HmgD [Df(2R)ED3923, breakpoints 57F6: 57F10] was seen. Moreover, no genetic interactions between Df(2R)ED50000 and other alleles were observed in any subsequent crosses.

Even though Df(2R)ED3921 removes the promoter and first exon of both HmgD and HmgZ the HmgD/Z deficiency is viable and exhibits no apparent growth impairment at either high or low temperatures. This phenotype would not be expected if HMGD/Z were involved in the organization of chromatin structure, since gross changes in chromatin would be expected to cause lethality. However, we observed that flies homozygous for Df(2R)ED3921 have a distinctive phenotype, exhibiting loss of the distal portion of the L5 longitudinal vein (Figure 2B), with a penetrance of 80% in females and 50% in males. Full penetrance was seen in flies crossed to the larger deficiency [Df(2R)ED3923]. These observations imply that the deficiency is not a null allele of HmgD/Z, but do verify that the phenotype is due to altered levels of HMGD/Z protein and not the CG30403 gene product.

View larger version (130K): In this window In a new window Download PPT slide   FIGURE 2.— Wing phenotypes of genetic interaction of HmgD/Z with the Brahma complex. (A) Wild-type wing showing longitudinal veins (L2, L3, L4, L5, and L6), anterior cross-vein (acv), and posterior cross-vein (pcv). (B) Homozygous Df(2R)ED3921 wings show loss of the distal portion of L5. (C) brm2/+ flies display normal wing vein patterning. (D–E) Enhanced loss of veins is seen in Df(2R)ED3921 flies crossed to brm and osa alleles. Severity of the phenotype is increased with the strong osa2 allele (E) and by reducing HMGD/Z levels further in Df(2R)ED3921/Df(2R)ED3923 flies heterozygous for brm or weak osa mutations (data not shown). (F) Df(2R)ED3921. (G) Flies homozygous for Df(2R)ED3921 and heterozygous for brm2 show a held out wings phenotype, which is also observed in flies trans-heterozygous for Df(2R)ED3921 and osa alleles (see Table 1).

 

View this table: In this window In a new window   TABLE 1 Interactions with the Brahma remodeling complex

 
Deficiencies were confirmed using PCR and DNA sequencing (GOLIC and GOLIC 1996; RYDER et al. 2004). To verify the loss of CG30403 in Df(2R)ED50000 we performed qualitative RT–PCR on mRNA isolated from these flies. As shown in Figure 1C we observed no product for CG30403 and HMGD, yet a band for HMGZ. To examine the level of HMGZ expression in Df(2R)ED50000 we probed Western blots of mutant embryos with an antibody raised against the HMGB domain of HMGD, which also recognizes HMGZ. Figure 1B illustrates that HMGZ levels in Df(2R)ED50000 are hardly reduced compared to wild-type expression, suggesting that HMGZ expression is upregulated to compensate for the loss of HmgD.

To verify that Df(2R)ED3921 resulted in mutation of the HmgD and HmgZ loci Western blots of embryo extract were probed with the HMGD antibody. As shown in Figure 1B, very low levels of HMGD/Z were detected, accounting for <5% of the ECL signal in wild-type embryos. Given that the antibody may recognize other uncharacterized HMGB proteins, we performed RT–PCR on adult flies. It should be stressed that this technique provides only qualitative evidence for the absence or presence of HMGD/Z transcripts. In wild-type flies HMGD and HMGZ were detected, as illustrated in Figure 1C. In Df(2R)ED3921 flies HMGD was not detected, yet HMGZ transcript was present. To quantify the amount of HMGZ mRNA remaining in these flies, Northern blot analyses of larvae and adults were performed. Figure 1D shows that flies homozygous for Df(2R)ED3921 retain low levels of HMGZ transcripts (60% of the hybridization signal in larvae and 30% in adults compared with wild type). This agrees with the results obtained from the Western blot, and we conclude that Df(2R)ED3921 can be considered a strong hypomorph for HmgD/Z.

HMGB redundancy:

The lack of a more severe phenotype in the HmgD/Z double mutants could perhaps be due to redundancy with other HMGB proteins. Other non-sequence-specific HMGB proteins in Drosophila are Dsp1 and an uncharacterized gene pair, CG7045 and CG7046. These proteins share high sequence similarity with HMGD/Z in the HMGB domain, but this is not extended to residues flanking this domain. Dsp1 contains two HMGB domains and a C-terminal acidic tail, while both CG7045 and CG7046 contain one HMGB domain, yet the C termini are not enriched for acidic residues.

Mutation of Dsp1 results in homeotic transformation of the first leg to the third at a low frequency (DECOVILLE et al. 2001). Flies homozygous for Dsp1¹ (a null allele) and Df(2R)ED3921 showed no enhancement of homeotic transformations of first leg to third or of an effect on the wing phenotype observed in Df(2R)ED3921. Null mutants do not exist for CG7045/6; therefore, Df(2R)ED3921 flies were crossed to Df(3R)Exel6191 (breakpoints 94A9; 94B2), a deficiency encompassing CG7045/6. No genetic interaction was observed in Df(2R)ED3921; Df(3R)Exel6191/+ flies. This was also true for the triple mutant dsp1¹; Df(2R)ED3921; Df(3R)Exel6191/+. It was not feasible to generate a triple null for all four genes and therefore the possibility of redundancy cannot be excluded. Yet, if the hypothesis of a role for HMGB proteins in determining global chromatin architecture were valid, depletion of a substantial proportion of HMGB proteins would be expected to result in a more pronounced phenotype than that observed.

HMGD/Z interact specifically with the Brahma complex:

Small abundant HMGB proteins have been implicated in chromatin remodeling in vitro with the ISWI-containing CHRAC complex (BONALDI et al. 2002). To test the biological relevance of this observation, Df(2R)ED3921 flies were crossed to available mutants of chromatin remodeling factors. We observed a strong interaction with the brahma amorphic allele, brm², but not with alleles of domino (ATPase subunit of the NuA4 complex), Iswi (ATPase subunit of NURF, CHRAC, and ACF complexes), Nurf301 (member of the NURF complex), Acf1 (member of the ACF complex), dMi-2 (ATPase subunit of the NURD complex), kismet (ATPase subunit, uncharacterized complex), or lodestar (ATPase subunit, uncharacterized complex). Flies of genotype Df(2R)ED3921; brm²/ + have held out wings and loss of vein L5 up to the posterior cross-vein with 100% penetrance (Figure 2, D and G).

Since held out wings have also been observed in flies heterozygous for partially complementing brm alleles and in brm/osa trans-heterozygotes (TAMKUN et al. 1992; VAZQUEZ et al. 1999), we tested the interaction of Df(2R)ED3921 with other members of the Brahma remodeling complex, summarized in Table 1. The same phenotypes were seen with alleles of osa. Notably flies trans-heterozyous for the HmgD/Z deficiency and the strong osa² allele displayed held out wings at 65% penetrance, and a weaker allele, osa⁰⁰⁰⁹⁰, at 35%. This provides evidence of a function for HmgD/Z in the same pathway as the Brahma complex. No significant frequency of held out wings was observed for selected alleles of Brahma complex members moira, Snr1, and deficiencies covering dalao (encodes Bap111) and Bap60.

Interestingly the vein phenotype seen in Df(2R)ED3921 was also enhanced by brm and osa alleles, resulting in loss of vein L5 up to the posterior cross-vein, but this was not observed with other members of the Brahma complex (Figure 2, C–E). When strong alleles of osa were used and/or the levels of HmgD/Z were reduced further by crossing to a larger defined deficiency [Df(2R)ED3923], partial loss of vein L2 was also seen (Figure 2E and data not shown).

HMGD/Z in vein specification:

The phenotype obtained in HmgD/Z and brm double mutants is reminiscent of the loss-of-function rhomboid allele, rho^ve. This allele results in loss of the distal portions of veins L2–L5, with L3 and L4 displaying the greatest variation in the extent of vein shortening (STURTEVANT et al. 1993). Previous studies have examined Brahma complex function in wing vein formation. The complex was found to be required for the correct specification of both vein and nonvein cells. Overexpression of dominant negative Brahma resulted in loss of Rhomboid expression and loss of vein tissue, while mutation of the negatively regulating Snr1 subunit resulted in ectopic vein formation (MARENDA et al. 2003, 2004; ZRALY et al. 2003). Mutations in brm were shown to enhance the loss of vein phenotype of rho^ve mutants, specifically in the L5 region. As shown in Figure 3B, reducing the expression of HMGD/Z also enhances this phenotype and there is also a partial loss of L2. We also observed a mild enhancement of L3 vein loss and variation in the length of L4. It should be noted that rho^ve flies display variation in L4 length and we observed the same range of different L4 lengths in both rho^ve and Df(2R)ED3921 flies, as shown in Figure 3, A and B. We interpret this to mean that Df(2R)ED3921 does not significantly enhance the loss of L4 and L3, but affects L2 and L5. The interaction with rho^ve provides additional evidence of a role for HMGD/Z involvement in gene activation, in specific contexts, as an accessory factor to the Brahma remodeling complex.

View larger version (96K): In this window In a new window Download PPT slide   FIGURE 3.— Interactions with rhomboid. (A) Homozygous rhove flies show incomplete L3, L4, and L5 veins. (B) Reduction in HMGD/Z levels enhances the rhove phenotype in veins L2 and L5 and mildly so in L3. L4 length was variable in both rhove and Df(2R)ED3921; rhove flies. A is representative of the shortest length and B of the longest L4 length observed.

 

Regulation of homeotic genes:

To examine HMGD/Z function in transcription regulation, we have made use of the visible phenotypes associated with Hox gene misregulation. This allowed us to ask whether HMGD/Z act as general transcription factors, since many Hox gene alleles show dominant phenotypes. We reasoned that if HMGD/Z acted as general regulators of transcription, we would see modification of the dominant Hox gene phenotypes since these provide highly sensitized backgrounds. In addition, Brahma was originally isolated in a screen for proteins involved in Hox gene expression and we wanted to examine the effect of loss of HMGD/Z on Brahma-regulated promoters (KENNISON and TAMKUN 1988; TAMKUN et al. 1992).

The held out wings phenotype observed in Df(2R)ED2921; brm²/+ is not uncommon, but notably has been observed in loss-of-function alleles of Antennapedia. Previous studies have linked Brahma complex function to correct expression of Antp (TAMKUN et al. 1992; VAZQUEZ et al. 1999). Vasquez et al. observed that osa and brm alleles interacted with promoter-specific Antp mutants, both proteins being required for activation of Antp from the P2 promoter specifically (VAZQUEZ et al. 1999). We decided to test whether HmgD/Z were needed for this specific activation, by crossing Df(2R)ED3921 to the gain-of-function allele Antp^Ns; this allele contains an internal partial duplication of the Antp gene plus insertion of two transposable elements and additional sequence (TALBERT and GARBER 1994). This results in ectopic activation of Antp in eye-antennal discs, inducing transformation of antennae into leg tissue.

The HmgD/Z deficiency was shown to partially suppress Antp^Ns (see Table 2). The degree of suppression was not as great as that observed with brm and osa mutations [21% for Df(2R)ED3921 vs. 87% for osa²] and underscores the supporting role of HMGD/Z in gene activation.

View this table: In this window In a new window   TABLE 2 Interactions with Antennapedia

 
We have also examined the role of HMGD/Z with other Brahma-regulated homeotic genes. We observed an interaction with Sex combs reduced. Df(2R)ED3921 enhanced loss-of-function alleles of Scr and partially suppressed the gain-of-function allele Scr^S (see Figure 4). This is similar to the interactions observed between mutations in the Bramha complex and Scr, where overexpression of dominant negative Brahma (DN-brm) in leg discs results in a reduced number of sex comb teeth (ELFRING et al. 1998).

View larger version (130K): In this window In a new window Download PPT slide   FIGURE 4.— Interactions with Sex combs reduced. The average number of sex comb teeth per sex comb is displayed for (A) wild type and (B) Scr4/ +. These males show reduction in the number of sex comb teeth on the first leg compared to that in wild type. (C) Df(2R)ED3921; Scr4 / +. Reduction in the levels of HMGD/Z results in enhancement of this phenotype. Reduction in HMGD/Z levels suppresses the gain-of-function allele ScrS. Gain-of-function mutations in Scr result in sex combs on the second and third legs, indicated by arrowheads. The average number of sex comb teeth per sex comb for each leg is indicated for (D) ScrS/ + and (E) Df(2R)ED3921; ScrS/ +. (F) The suppression of extra sex combs by HmgD/Z is dose dependent as Df(2R)ED3921/Df(2R)ED3923 flies show stronger suppression of the ScrS phenotype.

 
The Brahma complex has also been shown to be required for activation of Ultrabithorax, as brm, osa, and mor mutations enhance loss-of-function Ubx alleles (TAMKUN et al. 1992). We did not observe an increase in haltere-to-wing transformation when Df(2R)ED3921 was crossed to Ubx alleles (see supplemental Table 1 at http://www.genetics.org/supplemental/). This result suggests that HMGD and HMGZ do not function as ubiquitous general transcriptional activators that promote gene activation and also that the interaction of HmgD/Z with Brahma is required only at certain Brahma-regulated promoters.

Trithorax (trxG) and polycomb (PcG) group genes were originally characterized as opposing effectors of homeotic gene expression (see KENNISON and TAMKUN 1988 and references therein). Subsequent studies showed that many PcG and trxG proteins have roles in chromatin modulation. To further assess the role of HMGD/Z in chromatin regulation and to look for interactions with this group of genes we crossed Df(2R)ED3921 to alleles of Pc (Polycomb), Psc (Posterior sex combs), Sce (Sex combs extra), ph-p (polyhomeotic proximal), and esc (extra sex combs). No interactions were observed with any PcG mutants or with ash1 (absent, small, or homeotic discs 1) and trx (trithorax) of the trithorax group of proteins (see supplemental Table 1 at http://www.genetics.org/supplemental/). This result indicates that HMGD/Z do not act as typical trxG or PcG proteins, but interact specifically with Brahma.

Proteins that modulate chromatin structure have also been isolated in screens for genes that affect position effect variegation (PEV). If the HmgD/Z deficiency had an effect on PEV it could imply a role in global modulation of chromatin structure, as seen, for example, in Su(var)3-9 or Su(var)205 (heterochromatin protein 1) mutations. HMGB proteins have been hypothesized to promote a more open chromatin conformation and would therefore be expected to act as enhancers of PEV. Df(2R)ED3921 did not suppress or enhance the variegated phenotype of rst³ (roughest) and y^3P (yellow) inversions, providing further evidence against the proposed role of HMGD/Z in global chromatin architecture (see supplemental Table 1 at http://www.genetics.org/supplemental/).

HMGD and HMGZ are not integral components of the Brahma complex:

We next wanted to examine whether HMGD and HMGZ are directly associated with the Brahma complex in vivo. Nuclear extract prepared from Drosophila embryos was separated on 15–35% glycerol gradients. Both purified Brahma complexes are 2 MDa in size and sediment in the bottom fractions of the gradient (MOHRMANN et al. 2004); see Figure 5. As shown in Figure 5, there was no overlap between the Brahma-containing fractions and HMGD/Z, which was located in fractions ranging from 1 MDa (estimated from markers) to the top fractions, corresponding to HMGD/Z bound to sheared chromatin.

View larger version (29K): In this window In a new window Download PPT slide   FIGURE 5.— HMGD and HMGZ are not integral components of the Brahma complex. Nuclear extract derived from 0- to 12-hr wild-type (yw) embryos was fractionated on 15–35% glycerol gradients and probed with antibodies against Drosophila Brahma (rabbit anti-Brm, 1:2000) and HMGD/Z (rat anti HMGD-100, 1:250). HMGD/Z is associated with sheared chromatin but does not cofractionate with the Brahma complex. L, input fraction.

 

The HMGB domain is found in many proteins, from sequence-specific transcription factors to members of chromatin remodeling complexes. The DNA-binding and bending properties of HMGB proteins are clearly important, yet defining a function for small, abundant chromatin-associated HMGB proteins has proven difficult.

We have investigated the role of HMGD and HMGZ in Drosophila and have observed a strong genetic interaction with the Brahma remodeling complex, the first genetic link observed in multicellular organisms. We have also provided evidence for a role of HMGD and HMGZ in aiding the Brahma complex in activation of certain target genes. The genetic interaction of HmgD/Z with brm and osa alleles results in held out wings and loss of wing veins. The held out wings phenotype is attributed to decreased levels of Antp, through loss of transcription activation at the P2 promoter. Furthermore, reduction of HmgD/Z levels partially suppresses the Antp^Ns gain-of-function mutation, caused by ectopic activation of the P2 promoter (VAZQUEZ et al. 1999). These results indicate that HmgD and HmgZ cooperate with the Brahma complex in activating transcription of Antp.

Another gene regulated by the Brahma complex is rhomboid, a gene required for vein specification in the Drosophila wing. Loss of HMGD/Z enhances the rho^ve phenotype in a similar way to brm mutants (MARENDA et al. 2004). We observe loss of both L2 and L5 veins in Df(2R)ED3921; rho^ve flies; a similar phenotype is observed for combinations of the HmgD/Z deficiency with brm and osa. This strengthens the concept that HMGD/Z aid the Brahma complex in activation of its targets.

In examining whether HMGD/Z were involved in chromatin remodeling in vivo, we have observed a genetic interaction of these proteins with a SWI/SNF-like remodeler (Brahma) but not with an ISWI type (NURF, CHRAC, and ACF) although in vitro experiments have previously shown that HMGB1 can facilitate remodeling by the ACF complex (BONALDI et al. 2002). The modes of nucleosome remodeling effected by these two types of complex are thought to be different, where SWI/SNF, but not ISWI, remodelers are able to alter DNaseI cleavage periodicity in nucleosomes and release supercoils (FLAUS and OWEN-HUGHES 2004).

It is interesting to note that the Brahma complex contains a protein, Bap111 (encoded by dalao), with an HMGB domain (PAPOULAS et al. 2001). One might therefore expect that loss of HMGD/Z would not affect Brahma complex function, since it already contains an HMGB protein. We did not observe an interaction in flies heterozygous for a deficiency covering dalao. However, in brm mutants, reduced amounts of functional Brahma complex may render certain promoters more dependent on small HMGB proteins such as HMGD/Z, and it is possible that HMGD/Z and Bap111 have different functions in Brahma-mediated transcriptional activation.

We observed stronger genetic interactions between osa mutants and Df(2R)ED3921 than with other Brahma complex members. Osa is able to bind DNA, and HMGD/Z could possibly facilitate Osa binding to nucleosomes, thereby enhancing the rate of association of the remodeling complex to nucleosomes. A recent study of the interaction between HMGB1 and the glucocorticoid receptor (GR) found that the residence time of GR bound to chromatin was increased in the presence of HMGB1 (AGRESTI et al. 2005).

We have attempted to look for redundancy between Drosophila HMGB proteins, but are limited by the lack of available mutants for CG7045/6. The developing wing is particularly sensitive to perturbation. If there was any functional redundancy between HMGD/Z and CG7045/6 or Dsp1, modification of the phenotype would have most likely been observed in this tissue. However, until CG7045/6 mutants are made the issue of redundancy cannot be completely discarded. In examining HmgD/Z function in transcription regulation, we made use of the visible phenotypes associated with Hox gene misregulation. These phenotypes are known to be sensitive to changes in the activity of chromatin regulators such as Brahma and Polycomb (KENNISON and TAMKUN 1988; TAMKUN et al. 1992). In a manner similar to that of brahma HmgD/Z were found to facilitate activation of Antp and Scr. In contrast, brm mutants enhance loss-of-function Ubx mutations, yet we observe no effect for the HmgD/Z deficiency. We infer that HmgD/Z may be functionally necessary only at some Brahma-regulated promoters.

Interestingly both Dsp1 and HmgD/Z play a role in the activation of Scr. Dsp1 suppresses the extra sex combs phenotype of Pc, but shows no interactions with brm (DECOVILLE et al. 2001). The opposite is true for the HmgD/Z deficiency, suggesting that these proteins act in independent pathways to activate Scr. Dsp1 has recently also been shown to promote the association of PcG group proteins to PREs (PcG response elements) by binding to a G(A) sequence motif, suggesting a role in transcription repression for this HMGB protein (DEJARDIN et al. 2005).

Our observations in Drosophila have interesting parallels to studies in Saccharomyces cerevisiae. Mutation of both nhp6a and nhp6b genes results in a temperature-sensitive growth phenotype. Synthetic lethal phenotypes with SWI/SNF mutants have been observed for snf5 and more recently for swi2 (YU et al. 2000; BISWAS et al. 2004). In addition Nhp6a/b has been shown to be required for repression of the CHA1 locus, and an interaction with the RSC remodeling complex has been inferred (MOREIRA and HOLMBERG 2000). Similarly Nhp6a/b have been connected with the INO80 remodeling complex; overexpression of Nhp6a rescues the phenotypes associated with mutations in arp6 and arp7 (nuclear actin-related protein, Arp), which are components of the INO80 complex (SZERLONG et al. 2003).

HMGB proteins are widely thought of as architectural chromatin proteins. This implies a global chromatin role, and mutation of such proteins should have dramatic effects on chromatin structure if removed.

Previous work by NER and TRAVERS (1994) hypothesized a role for HMGD in chromatin structure in the syncytial blastoderm. HMGD was proposed to act as an early embryonic linker binding protein, producing a more open chromatin conformation to facilitate rapid nuclear divisions in the absence of H1. Embryos containing the HmgD/Z deficiency have normal nuclear morphology and no abnormal nuclear divisions were observed (A. RAGAB, unpublished observations). This argues against an "early embryonic linker" role for HMGD/Z and suggests that other factors, or the lack or histone H1, are responsible for the decondensed chromatin state during these stages. In addition, no alteration either in micrococcal nuclease digestion kinetics of embryonic chromatin or in the morphology of mitotic and polytene chromosomes of these mutants was seen (E. C. THOMPSON and A. RAGAB, unpublished observations).

Recently an additional role for yeast Nhp6a/b and mammalian HMGB1 in genome stability has been proposed (GIAVARA et al. 2005). Increased sensitivity to UV irradiation was seen in hmgb1–/– primary embryonic fibroblasts and in nhp6a/b double-mutant yeast. In addition, sensitivity to hydrogen peroxide and methyl methanesulphonate (MMS) was also observed in the yeast mutants. We have not observed any enhanced sensitivity to UV irradiation, X irradiation, or MMS in the HmgD/Z deficiency (E. C. THOMPSON, unpublished observations). We therefore infer that a major role for the abundant HMGB proteins in Drosophila is cooperation with the Brahma complex.

Previous biochemical studies are able to provide a model for how HMGD/Z assists chromatin remodeling by exposing DNA at the nucleosome surface. HMGB1 has been shown to bind to DNA at the entry site to nucleosomes (AN et al. 1998). We have previously shown that HMGD binding induces an area of DNA accessibility at the entry site and dyad of nucleosomes in vitro, presumably by bending the DNA (RAGAB and TRAVERS 2003). The exposed DNA could promote association of remodeling complexes to nucleosomes and be used by the complex to propagate nucleosome sliding (BONALDI et al. 2002) or promote association of transcription factors (AGRESTI et al. 2005). Significantly the region where DNA accessibility is increased is immediately adjacent to the binding site for the motor polypeptide of the SWI/SNF complex (SAHA et al. 2005). HMGB proteins are known to bind chromatin with high on/off rates, which would result in rapid transient exposure of target sites and perhaps in more readily recognized and remodeled nucleosomes. This may afford organisms an advantage in the control of gene transcription and would explain the conservation of these small HMGB proteins from yeast to man.

We acknowledge D. Locker, P. Badenhorst, and J. Kadonaga for providing fly stocks; C. P. Verrijzer for the Brm antibody; and J. Brennecke, K. Brown, and A. Buscaino for valuable discussions and critical reading of the manuscript. A. Ragab was supported by a Medical Research Council (MRC) studentship. E. C. Thompson is supported by an MRC studentship.

AGRESTI, A., P. SCAFFIDI, A. RIVA, V. R. CAIOLFA and M. E. BIANCHI, 2005 GR and HMGB1 interact only within chromatin and influence each other's residence time. Mol. Cell 18: 109–121.[CrossRef][Medline]

AN, W., K. VAN HOLDE and J. ZLATANOVA, 1998 The non-histone chromatin protein HMG1 protects linker DNA on the side opposite to that protected by linker histones. J. Biol. Chem. 273: 26289–26291.[Abstract/Free Full Text]

BISWAS, D., A. N. IMBALZANO, P. ERIKSSON, Y. YU and D. J. STILLMAN, 2004 Role for Nhp6, Gcn5, and the Swi/Snf complex in stimulating formation of the TATA-binding protein-TFIIA-DNA complex. Mol. Cell. Biol. 24: 8312–8321.[Abstract/Free Full Text]

BONALDI, T., G. LANGST, R. STROHNER, P. B. BECKER and M. E. BIANCHI, 2002 The DNA chaperone HMGB1 facilitates ACF/CHRAC-dependent nucleosome sliding. EMBO J. 21: 6865–6873.[CrossRef][Medline]

BUSTIN, M., 2001 Chromatin unfolding and activation by HMGN chromosomal proteins. Trends Biochem. Sci. 26: 431–437.[CrossRef][Medline]

CALOGERO, S., F. GRASSI, A. AGUZZI, T. VOIGTLANDER, P. FERRIER et al., 1999 The lack of chromosomal protein Hmg1 does not disrupt cell growth but causes lethal hypoglycaemia in newborn mice. Nat. Genet. 22: 276–280.[CrossRef][Medline]

DECOVILLE, M., E. GIACOMELLO, M. LENG and D. LOCKER, 2001 DSP1, an HMG-like protein, is involved in the regulation of homeotic genes. Genetics 157: 237–244.[Abstract/Free Full Text]

DEJARDIN, J., A. RAPPAILLES, O. CUVIER, C. GRIMAUD, M. DECOVILLE et al., 2005 Recruitment of Drosophila Polycomb group proteins to chromatin by DSP1. Nature 434: 533–538.[CrossRef][Medline]

ELFRING, L. K., C. DANIEL, O. PAPOULAS, R. DEURING, M. SARTE et al., 1998 Genetic analysis of brahma: the Drosophila homolog of the yeast chromatin remodeling factor SWI2/SNF2. Genetics 148: 251–265.[Abstract/Free Full Text]

FLAUS, A., and T. OWEN-HUGHES, 2004 Mechanisms for ATP-dependent chromatin remodelling: farewell to the tuna-can octamer? Curr. Opin. Genet. Dev. 14: 165–173.[CrossRef][Medline]

FORMOSA, T., P. ERIKSSON, J. WITTMEYER, J. GINN, Y. YU et al., 2001 Spt16-Pob3 and the HMG protein Nhp6 combine to form the nucleosome-binding factor SPN. EMBO J. 20: 3506–3517.[CrossRef][Medline]

GIAVARA, S., E. KOSMIDOU, M. P. HANDE, M. E. BIANCHI, A. MORGAN et al., 2005 Yeast Nhp6A/B and mammalian Hmgb1 facilitate the maintenance of genome stability. Curr. Biol. 15: 68–72.[CrossRef][Medline]

GOLIC, K. G., and M. M. GOLIC, 1996 Engineering the Drosophila genome: chromosome rearrangements by design. Genetics 144: 1693–1711.[Abstract]

KAL, A. J., T. MAHMOUDI, N. B. ZAK and C. P. VERRIJZER, 2000 The Drosophila Brahma complex is an essential coactivator for the trithorax group protein Zeste. Genes Dev. 14: 1058–1071.[Abstract/Free Full Text]

KAMAKAKA, R. T., C. M. TYREE and J. T. KADONAGA, 1991 Accurate and efficient RNA polymerase II transcription with a soluble nuclear fraction derived from Drosophila embryos. Proc. Natl. Acad. Sci. USA 88: 1024–1028.[Abstract/Free Full Text]

KENNISON, J. A., and J. W. TAMKUN, 1988 Dosage-dependent modifiers of polycomb and antennapedia mutations in Drosophila. Proc. Natl. Acad. Sci. USA 85: 8136–8140.[Abstract/Free Full Text]

MARENDA, D. R., C. B. ZRALY, Y. FENG, S. EGAN and A. K. DINGWALL, 2003 The Drosophila SNR1 (SNF5/INI1) subunit directs essential developmental functions of the Brahma chromatin remodeling complex. Mol. Cell. Biol. 23: 289–305.[Abstract/Free Full Text]

MARENDA, D. R., C. B. ZRALY and A. K. DINGWALL, 2004 The Drosophila Brahma (SWI/SNF) chromatin remodeling complex exhibits cell-type specific activation and repression functions. Dev. Biol. 267: 279–293.[CrossRef][Medline]

MOHRMANN, L., K. LANGENBERG, J. KRIJGSVELD, A. J. KAL, A. J. HECK et al., 2004 Differential targeting of two distinct SWI/SNF-related Drosophila chromatin-remodeling complexes. Mol. Cell. Biol. 24: 3077–3088.[Abstract/Free Full Text]

MOREIRA, J. M., and S. HOLMBERG, 2000 Chromatin-mediated transcriptional regulation by the yeast architectural factors NHP6A and NHP6B. EMBO J. 19: 6804–6813.[CrossRef][Medline]

MOSRIN-HUAMAN, C., L. CANAPLE, D. LOCKER and M. DECOVILLE, 1998 DSP1 gene of Drosophila melanogaster encodes an HMG-domain protein that plays multiple roles in development. Dev. Genet. 23: 324–334.[CrossRef][Medline]

NEMETH, M. J., D. J. CURTIS, M. R. KIRBY, L. J. GARRETT-BEAL, N. E. SEIDEL et al., 2003 Hmgb3: an HMG-box family member expressed in primitive hematopoietic cells that inhibits myeloid and B-cell differentiation. Blood 102: 1298–1306.[Abstract/Free Full Text]

NEMETH, M. J., A. P. CLINE, S. M. ANDERSON, L. J. GARRETT-BEAL and D. M. BODINE, 2005 Hmgb3 deficiency deregulates proliferation and differentiation of common lymphoid and myeloid progenitors. Blood 105: 627–634.[Abstract/Free Full Text]

NER, S. S., and A. A. TRAVERS, 1994 HMG-D, the Drosophila melanogaster homologue of HMG 1 protein, is associated with early embryonic chromatin in the absence of histone H1. EMBO J. 13: 1817–1822.[Medline]

NER, S. S., M. E. A. CHURCHILL, M. A. SEARLES and A. A. TRAVERS, 1993 dHMG-Z, a second HMG-1-related protein in Drosophila melanogaster. Nucleic Acids Res. 21: 4369–4371.[Abstract/Free Full Text]

NER, S. S., T. BLANK, M. L. PEREZ-PARALLE, T. A. GRIGLIATTI, P. B. BECKER et al., 2001 HMG-D and histone H1 interplay during chromatin assembly and early embryogenesis. J. Biol. Chem. 276: 37569–37576.[Abstract/Free Full Text]

PAPOULAS, O., G. DAUBRESSE, J. A. ARMSTRONG, J. JIN, M. P. SCOTT et al., 2001 The HMG-domain protein BAP111 is important for the function of the BRM chromatin-remodeling complex in vivo. Proc. Natl. Acad. Sci. USA 98: 5728–5733.[Abstract/Free Full Text]

RAGAB, A., and A. TRAVERS, 2003 HMG-D and histone H1 alter the local accessibility of nucleosomal DNA. Nucleic Acids Res. 31: 7083–7089.[Abstract/Free Full Text]

REEVES, R., 2003 HMGA proteins: Flexibility finds a nuclear niche? Biochem. Cell Biol. 81: 185–195.[CrossRef][Medline]

RONFANI, L., M. FERRAGUTI, L. CROCI, C. E. OVITT, H. R. SCHOLER et al., 2001 Reduced fertility and spermatogenesis defects in mice lacking chromosomal protein Hmgb2. Development 128: 1265–1273.[Abstract]

RYDER, E., F. BLOWS, M. ASHBURNER, R. BAUTISTA-LLACER, D. COULSON et al., 2004 The DrosDel collection: a set of P-element insertions for generating custom chromosomal aberrations in Drosophila melanogaster. Genetics 167: 797–813.[Abstract/Free Full Text]

SAHA, A., J. WITTMEYER and B. R. CAIRNS, 2005 Chromatin remodeling through directional DNA translocation from an internal nucleosomal site. Nat. Struct. Mol. Biol. 12: 747–755.[CrossRef][Medline]

STURTEVANT, M. A., M. ROARK and E. BIER, 1993 The Drosophila rhomboid gene mediates the localized formation of wing veins and interacts genetically with components of the EGF-R signaling pathway. Genes Dev. 7: 961–973.[Abstract/Free Full Text]

SZERLONG, H., A. SAHA and B. R. CAIRNS, 2003 The nuclear actin-related proteins Arp7 and Arp9: a dimeric module that cooperates with architectural proteins for chromatin remodeling. EMBO J. 22: 3175–3187.[CrossRef][Medline]

TALBERT, P. B., and R. L. GARBER, 1994 The Drosophila homeotic mutation Nasobemia (AntpNs) and its revertants: an analysis of mutational reversion. Genetics 138: 709–720.[Abstract]

TAMKUN, J. W., R. DEURING, M. P. SCOTT, M. KISSINGER, A. M. PATTATUCCI et al., 1992 brahma: a regulator of Drosophila homeotic genes structurally related to the yeast transcriptional activator SNF2/SWI2. Cell 68: 561–572.[CrossRef][Medline]

THOMAS, J. O., and A. A. TRAVERS, 2001 HMG1 and 2, and related ‘architectural’ DNA-binding proteins. Trends Biochem. Sci. 26: 167–174.[CrossRef][Medline]

TRAVERS, A, A., 2003 Priming the nucleosome: A role for HMGB proteins? EMBO Rep. 4: 131–136.[CrossRef][Medline]

VAZQUEZ, M., L. MOORE and J. A. KENNISON, 1999 The trithorax group gene osa encodes an ARID-domain protein that genetically interacts with the brahma chromatin-remodeling factor to regulate transcription. Development 126: 733–742.[Abstract]

WAGNER, C. R., K. HAMANA and S. C. ELGIN, 1992 A high-mobility-group protein and its cDNAs from Drosophila melanogaster. Mol. Cell. Biol. 12: 1915–1923.[Abstract/Free Full Text]

YU, Y., P. ERIKSSON and D. J. STILLMAN, 2000 Architectural transcription factors and the SAGA complex function in parallel pathways to activate transcription. Mol. Cell. Biol. 20: 2350–2357.[Abstract/Free Full Text]

YU, Y., P. ERIKSSON, L. T. BHOITE and D. J. STILLMAN, 2003 Regulation of TATA-binding protein binding by the SAGA complex and the Nhp6 high-mobility group protein. Mol. Cell. Biol. 23: 1910–1921.[Abstract/Free Full Text]

ZRALY, C. B., D. R. MARENDA, R. NANCHAL, G. CAVALLI, C. MUCHARDT et al., 2003 SNR1 is an essential subunit in a subset of Drosophila brm complexes, targeting specific functions during development. Dev. Biol. 253: 291–308.[CrossRef][Medline]

Communicating editor: J. TAMKUN

This article has been cited by other articles:

				K. S. Weiler The Multi-AT-Hook Chromosomal Protein of Drosophila melanogaster, D1, Is Dispensable for Viability Genetics, May 1, 2009; 182(1): 145 - 159. [Abstract] [Full Text] [PDF]

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu