In the past decade, genome-sequencing projects have yielded a great amount of information on DNA sequences in several organisms. The release of the Drosophila melanogaster heterochromatin sequence by the Drosophila Heterochromatin Genome Project (DHGP) has greatly facilitated studies of mapping, molecular organization, and function of genes located in pericentromeric heterochromatin. Surprisingly, genome annotation has predicted at least 450 heterochromatic gene models, a figure 10-fold above that defined by genetic analysis. To gain further insight into the locations and functions of D. melanogaster heterochromatic genes and genome organization, we have FISH mapped 41 gene models relative to the stained bands of mitotic chromosomes and the proximal divisions of polytene chromosomes. These genes are contained in eight large scaffolds, which together account for ∼1.4 Mb of heterochromatic DNA sequence. Moreover, developmental Northern analysis showed that the expression of 15 heterochromatic gene models tested is similar to that of the vital heterochromatic gene Nipped-A, in that it is not limited to specific stages, but is present throughout all development, despite its location in a supposedly “silent” region of the genome. This result is consistent with the idea that genes resident in heterochromatin can encode essential functions.
HETEROCHROMATIN was originally defined cytologically as chromosomal regions that stained strongly at prophase and maintained a compact organization throughout all stages of the mitotic cell cycle (Heitz 1928). In a wide variety of eukaryotes, large chromosomal segments, or even entire chromosomes, can be composed of heterochromatin. Constitutive heterochromatin is a ubiquitous and abundant component of chromosomes of higher organisms, forming ∼5% of the genome in Arabidopsis thaliana, 30% in Drosophila, and 30% in humans, and up to 80% in certain nematodes and plants such as tomato (Moritz and Roth 1976; Peterson et al. 1998; Dimitri et al. 2005b). Despite these fluctuations in abundance, similar unusual structural properties characterize heterochromatin in virtually all animal and plant species, which together led to the view of heterochromatin as a “desert” of genetic functions (reviewed by John 1988). In the past 2 decades, however, genetic and molecular studies in Drosophila melanogaster have implicated constitutive heterochromatin in important cellular functions such as chromosome organization and inheritance and have shown it to contain genes required for viability and fertility (Gatti and Pimpinelli 1992; Williams and Robbins 1992; Weiler and Wakimoto 1995; Dernburg et al. 1996; Elgin 1996; Karpen et al. 1996; Eissenberg and Hilliker 2000; Henikoff et al. 2001; Coulthard et al. 2003; Dimitri et al. 2005b; Fitzpatrick et al. 2005). These genes were identified by recessive lethal mutations genetically linked to the heterochromatin of chromosomes 2 and 3 (Hilliker 1976; Marchant and Holm 1988) and were mapped to a specific region of mitotic heterochromatin (Dimitri 1991; Koryakov et al. 2002). Thus far, at least 32 essential genes have been mapped to the mitotic heterochromatin of chromosomes 2 and 3, but only a few of them—RpL5, light, concertina, rolled, RpL38, Nipped-B, Nipped-A, Parp, and RpL15—were satisfactorily characterized at the molecular level (Hilliker 1976; Devlin et al. 1990; Parks and Wieshous 1991; Biggs et al. 1994; Rollins et al. 1999; Tulin et al. 2002; Myster et al. 2004; Marygold et al. 2005; Schulze et al. et al. 2005). Notably, the presence of genes in heterochromatin, far from being a peculiarity of Drosophila, appears to be a conserved trait in the evolution of eukaryotic genomes. Heterochromatic genes have been recently identified in Saccharomyces cerevisiae, Schizosaccharomyces pombe, rice, A. thaliana, and humans (Kuln et al. 1991; Arabidopsis Genome Initiative 2000; Horvath et al. 2000; Brun et al. 2003; Nagaki et al. 2004).
The release of the sequence of D. melanogaster heterochromatin by the Berkeley Drosophila Genome Project has greatly facilitated the study of the molecular organization and function of heterochromatic genes. Initially, 3.8 Mb of ∼120 Mb of the D. melanogaster euchromatic genome sequence from release1 was found to correspond to heterochromatic sequences (Adams et al. 2000). Later, an improved whole-genome shotgun assembly (heterochomatic WGS3; Hoskins et al. 2002) was produced by the Drosophila Heterochromatin Genome Project (DHGP), which includes 20.7 Mb of draft-quality heterochromatic sequence. About 450 gene models (computational genes [CGs]) have been identified by the annotation of this sequence (Hoskins et al. 2002). In the recent release 5 of the D. melanogaster genome sequence, the number of genes predicted in heterochromatin was estimated to be even larger (DHGP; G. Karpen, personal communication). Several studies have concentrated on an effort to map gene models to the mitotic heterochromatin of D. melanogaster using BACs and P elements (Hoskins et al. 2002; Corradini et al. 2003; Yasuhara et al. 2003). However, the location of several heterochromatic gene models in the heterochromatin genome of D. melanogaster is still unclear and there is a need for further mapping studies.
To expand our current knowledge of genetic functions located in pericentromeric heterochromatin, we have performed detailed fluorescence in situ hybridization (FISH) mapping of 10 cDNA clones and two P-element insertions to mitotic and polytene chromosomes. The results of this analysis led to the cytological location in heterochromatin of eight large scaffolds, which together account for ∼1.4 Mb of sequence that contains 41 gene models. The results of this analysis also extend our understanding of the correspondence between mitotic and polytene chromosome heterochromatin. Finally, Northern analysis demonstrated that, similarly to known essential genes, gene models located in heterochromatin can be expressed throughout all development, despite their location in a supposedly “silent” region of the genome.
MATERIALS AND METHODS
Genetic markers, mutations, and balancer chromosomes are described in Lindsley and Zimm (1992) and at FlyBase (http://www.flybase.org/). Cultures were maintained at 25° on standard cornmeal–sucrose–yeast–agar medium.
Mitotic and polytene chromosome preparations and FISH procedures were performed according to Dimitri (2004). The probes were labeled using a digoxigenin nick-translation kit (Boehringher Mannheim, Indianapolis); digoxigenin was revealed using an antidigoxigenin rhodamine-conjugated IgG as described in Dimitri (2004); plasmids carrying the cDNA inserts were used as probes.
cDNAs clones, gene models, and P-element insertions:
The cDNAs clones and heterochromatic gene models studied in this work are listed in Table 1. We chose 10 cDNAs from scaffolds containing at least two or more gene models whose cytological locations in heterochromatin were not yet determined or were still rough at the time that we started the experiments reported here. The complete sequences of the scaffolds are available at http://www.ncbi.nlm.nih.gov/. We mapped two P-element insertions: l(2)309 and EY09264. l(2)309 (Myster et al. 2004) are derived from KG01086 yellow+ and white+ SUPor-P element (Roseman et al. 1995). EY09264 is a EPgy2 P element carrying both the mini-white transgene, an intronless yellow transgene and the GAGA/Gal4-UAS enhancer. Both insertions are available at the Bloomington Stock Center.
Inverse PCR, genomic DNA, and cDNA preparations:
Extraction of plasmids containing the cDNA inserts was performed as described in the protocol for midi and maxi plasmid kits (QIAGEN, Chatsworth, CA). Genomic DNA extraction and inverse PCR were performed according to the protocol of the Berkeley Drosophila Genome Project (http://www.fruitfly.org/about/methods/index.html). Primers were designed using the Primer3 program (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3.cgi). The following primers were used to amplify by inverse PCR the flanking sequences of the 19.74.3 PLacW insertion: FOR 5′-TCACTCGCACTT ATTGCA-3′ and REV 5′-TAACCCTTAGCATGTCCGTG-3′. PCR products were cleaned up using either the QIAquick PCR purification kit or the QIAquick Gel extraction kit (QIAGEN). Sequencing reactions were performed by CRIBI using an ABI 3730 DNA analyzer. Sequences were analyzed using Sequencher (Gene Codes, Ann Arbor, MI).
RNA preparations and Northern analysis:
Poly(A+) mRNA was partially purified from 200 mg material of different Drosophila developmental stages using the Quickprep mRNA purification kit (Amersham, Buckinghamshire, UK). The RNA was sized electrophoretically through a 1.2% agarose–formaldehyde gel, blotted to nylon membrane, and hybridized at 65° with 32P-labeled probes. The same filter was used with different probes after removing the hybridization signal by boiling in 0.1× SSC and checking for residual labeling. Autoradiographs were recorded with the Phosphorimager Storm 820 (Molecular Dynamics, Eugene, OR).
Chromosome preparations were analyzed using a computer-controlled Zeiss Axioplan epifluorescence microscope equipped with a cooled CCD camera (Photometrics, Tucson, AZ). Fluorescence was visualized using the Pinkel no. 1 filter set combination (Chroma Technology, Brattleboro, VT). The fluorescent signals were recorded by IP Spectrum Lab Software and edited with Adobe PhotoShop 7.
FISH mapping to mitotic chromosomes:
FISH experiments with cDNA probes were performed to prometaphase chromosomes of third instar larvae neuroblasts from the y; cn bw sp strain used to generate the genome sequence (Adams et al. 2000). The results of this analysis are shown in Figure 1 and Figure 2A. We initially undertook a detailed mapping of 10 cDNAs and two P-element insertions located in eight scaffolds whose cytological locations were not yet determined or were still approximate (Table 1; see supplemental Table 1 in Hoskins et al. 2002; http://www.dhgp.org). Together, according to the recent release 5 (http://www.dhgp.org/; DHGP, unpublished results), the scaffolds defined by cDNAs carried 41 annotated gene models and covered ∼1.4 Mb of heterochromatic DNA (Table 1).
2Lh mitotic heterochromatin:
RE24541 and SD19278 cDNA clones from CG12567 and CG40042 gene models, respectively, were mapped to the mitotic heterochromatin of the left arm of chromosome 2 (2Lh). CG12567, CG40042, CG40040, and CG40041 are located in scaffold AABU01002756, which is ∼199 kb in length. The RE24541 cDNA produces a single FISH signal in heterochromatic bands h35 distal–h36 proximal (Figure 1A and Figure 2A). A similar result was obtained with SD19278. Thus, the FISH data indicate that AABU01002756 is located in region h35–h36.
2Rh mitotic heterochromatin:
Six cDNA clones were mapped to the mitotic heterochromatin of the right arm of chromosome 2 (2Rh). These cDNAs identify five different scaffolds: AABU01002199, AABU01002549, AABU01002750, AABU01002711, and AABU01002748, which together contain 29 gene models.
AABU01002199 (160 kb) carries five gene models: CG40211, CG41264, CG40218 (also called elica or Yeti), CG41265, and CG40220. This scaffold was previously localized to regions h40–h43 on the basis of cytological and molecular analysis of P-element insertion sites (Konev et al. 2003; see the list from the lab of G. Karpen at http://taputea.lbl.gov/research/het/HetPs/data/index.html and http://www.dhgp.org). To define the location of this scaffold more accurately, we performed FISH mapping using the RE36623 cDNA corresponding to CG4218. As shown in Figure 1B and Figure 2A, the RE36623 probe produced a weak, yet clear, fluorescence signal that maps to region h41 of 2Rh. This result is corroborated by our molecular analysis of the P-element insertion 19.74.3, which maps cytologically to h41 (Corradini et al. 2003). Inverse PCR and BLAST-n analysis of flanking sequences showed that the 19.74.3 insert is located in AABU01002199, ∼4 kb downstream of CG40218 (see materials and methods). Together, these data enabled us to restrict the localization of AABU01002199 with its five gene models to region h41.
AABU01002549 (120 kb) contains six gene models: CG41063, CG40190, CG41066, CG40192, CG41065, and CG40191. This scaffold was associated with the heterochromatin of the second chromosome (Drosophila Heterochromatin Genome Project 2004), yet no precise localization was defined. The LD12604 (CG40191) cDNA produced a single FISH signal in h41 distal–h42 proximal, thus mapping AABU01002549 in this position in 2Rh (Figure 1C and Figure 2A).
AABU01002750 (143 kb) contains five gene models: CG17691, CG40068, CG40388, CG40390, and CG40067. As shown in Figure 1D and Figure 2A, the RE25729 (CG17691) cDNA clone produced a sharp hybridization signal located in the proximal portion of h44, placing AABU01002750 in this region.
AABU01002711 (342 kb) contains eight gene models: CG40130, CG40129, CG17665, CG40131, CG40127, CG40128, CG40133, and CG17683. AABU01002711 was localized to h45 distal–h46 proximal (Figure 2A) using a mixture of both RE37350 (CG17683) and GM16138 (CG40127) cDNA clones as the probe. This result was confirmed by FISH mapping of the EY09264 P insertion (assigned to AABU01002711 by the Gene Disruption Project) to h45 distal–h46 proximal (data not shown).
AABU01002748 (134 kb) contains five gene models: CG40084, CG40081, CG40080, CG40085, and CG41098. Using the cDNA clone (LD07633) from CG40080, a single FISH signal was mapped to h45 distal–h46 proximal (Figure 2A), indicating that AABU01002748, like AABU0100271, spans this segment. This observation is in agreement with the finding that AABU01002748 and AABU01002711 are physically linked by the BAC map (DHGP; R. Hoskins and G. Karpen, personal communication).
Chromosome 3 mitotic heterochromatin:
Two scaffolds, AABU01002740 and AABU01002752, were mapped to the heterochromatin of chromosome 3. The RE54950 cDNA probe (CG10837) localized AABU01002740 (252 kb) with its five gene models (CG17514, CG40103, CG40163, CG10837, and CG40105) to region h48 of 3L (data not shown). The RE06111 cDNA clone (CG17419) mapped to region h58 of 3R heterochromatin (Figure 1E) and allowed us to localize AABU01002752 (122 kb) and its three genes (CG17419, CG40065, and CG41101) into this region.
FISH mapping of cDNAs to polytene chromosomes:
FISH experiments using the cDNAs tested on mitotic chromosomes were also performed with salivary gland polytene chromosomes of the y; cn bw sp isogenic strain. The aim of these experiments was to confirm the chromosomal location of the clones that were mapped to mitotic heterochromatin and to learn more about the relationship between mitotic and polytene chromosome heterochromatin. The results are shown in Figure 1, F, G, and H; Figure 2B). The SD19278 (CG12567) cDNA (scaffold AABU01002756; h35 distal–h36 proximal) produced a diffuse hybridization signal that maps to division 40D-F of 2Lh (Figure 1F and Figure 2B). A similar result was obtained with RE24541 (CG12567).
Six cDNAs were assigned to the proximal regions of the right arm of chromosome 2. The hybridization signal of RE36623 (CG40218) cDNA (scaffold AABU01002199; h41 of 2Rh) is formed by multiple dots that map to division 41A and span the inner portions of the chromocenter (Figure 1G and Figure 2B). Similarly, the signal of LD12604 (AABU01002549; h42 proximal) appears as a cluster of dots and maps to 41A (Figure 2B). The RE25729 (CG40191) cDNA (scaffold AABU01002750; h44 proximal) gave a clear fluorescence signal mapping to the distal portion of 41A (Figure 2B). This result apparently conflicts with a previous finding suggesting that the mitotic regions h43–h44 do not undergo polytenization in salivary glands (Koryakov et al. 1996). The probe containing a mixture of both GM16138 (CG40127) and RE37350 (CG17683) cDNAs (scaffold AABU01002711; h45 distal–h46 proximal) maps to 41B-C (Figure 1H and Figure 2B), consistent with previous mapping (supplemental Table 1 in Hoskins et al. 2002). These results extend and confim our previous study (Corradini et al. 2003), indicating that division 41C of salivary gland chromosomes originates from polytenization of DNA sequences located in h46, the most distal region of 2R mitotic heterochromatin. Finally, on chromosome 3, the FISH signal of RE54950 cDNA mapped to the inner portions of the 3L chromocenter; FISH with RE06111 (CG17419) cDNA (scaffold AABU01002752; h58 of 3Rh) produced a hybridization signal mapping to 81F.
Next, we wanted to identify human homologs of D. melanogaster heterochromatic gene models and study their chromosomal location in humans. To this aim, the putative protein sequence of 98 gene models assigned to heterochromatin (Hoskins et al. 2002; Corradini et al. 2003; Yasuhara et al. 2003) was used as query for a BLAST-p search against the human protein database. The results of this analysis are shown in Table 2. The 14 genes located at the hetero-euchromatic transition region were also included in this group. The results of this analysis indicate that ∼70% of the putative proteins encoded by heterochromatin-linked gene models studied in this work share significant similarity with human proteins. It is worth noting that all conserved hits identify genes that are located in euchromatic regions of the human genome. In particular, among the human genes that have been mapped in detail (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM), 35% were found to be associated with G bands and 65% with R bands. The observed distribution may reflect the average gene distribution found in human chromosomes, in which R bands are more gene enriched compared to G bands.
Northern blotting analysis of heterochromatic gene models in the wild-type Oregon-R strain of D. melanogaster:
To learn more about the function of the genes resident in heterochromatin, we investigated transcription of the vital gene Nipped-A and of the following 15 gene models: CG12567 and CG40042 of 2Lh, CG40218, CG41063, CG17691, CG40080, CG40130, CG40129, CG17665, CG2944, CG2682, and CG30440 of 2Rh, CG17514, CG10837 of 3Lh, and CG17419 of 3Rh (Tables 1 and 2). Notably, the protein products encoded by 13 of these gene models are conserved in humans (see Table 2). Transcription was studied throughout developmental stages in the wild-type Oregon-R strain by Northern blotting (see examples in Figure 3). Using the Nipped-A cDNA probe, a transcript of ∼11 kb was detected in all stages, suggesting that the expression of this essential gene is required throughout all development. A similar picture emerges from the analysis of the gene models tested: for example, homologous transcripts to cDNAs of CG17514, CG10837, CG30440, CG17691, CG40080, and CG2682 are present at different stages of development (Figure 3); the transcript of CG17514 was not seen in embryos. In each case, the molecular weight of bands is fully consistent with the predicted mRNA size.
In previous work we mapped 22 D. melanogaster gene models, defined in release 3 of the D. melanogaster genome, to the distal heterochromatin of chromosome 2 (Corradini et al. 2003). The results of this analysis expand our knowledge of the genetic functions located in the heterochromatin of this chromosome. Here, we have performed a detailed FISH mapping to the mitotic heterochromatin of 41 additional gene models of D. melanogaster, whose location was unclear at the time that we started our experiments. Moreover, transcription of the vital gene Nipped-A and 15 heterochromatic gene models was also investigated by Northern blotting analysis.
Known genes and gene models in the heterochromatin of chromosome 2:
The mapping results provided by this work complement those obtained in previous studies (Dimitri 1991; Hoskins et al. 2002; Corradini et al. 2003; Yasuhara et al. 2003), enabling us to construct a map of the heterochromatic genes located in chromosome 2 (Figure 4 and Table 2). To date, 11 scaffolds, which together account for ∼6 Mb of genomic DNA, have been mapped to the pericentric heterochromatin of chromosome 2. Thus, it emerges that this region contains at least 76 annotated genes, including 17 genes defined as essential by genetic analysis (Hilliker 1976; Dimitri 1991; reviewed by Coulthard et al. 2003 and by Dimitri et al. 2005b). Recently, Myster et al. (2004) have found additional lethals on 2Rh, but it is still unclear whether or not these lethals actually correspond to new vital genes.
The estimated DNA content in heterochromatin of chromosome 2 is ∼20 Mb (Hoskins et al. 2002), 70% of which (14 Mb) is made up by satellite DNA (Lohe and Hilliker 1996). Thus, ∼6 Mb of the remaining heterochromatin should correspond to islands of nonsatellite, complex DNAs enriched in transposons and genes. The observation that the 11 heterochromatic scaffolds mapped to chromosome 2 account for ∼6 Mb of genomic DNA suggests that almost the entire amount of nonsatellite heterochromatic DNA of chromosome 2 has been assembled. Notably, four scaffolds molecularly linked by the BAC map (Hoskins et al. 2002) map to region h46 (Figure 4 and Table 2) and together account for 3.5 Mb, which is reasonably close to the inferred size of this region (∼2.5 Mb).
From Figure 4 it emerges that most gene models are located in distal heterochromatin; together, 54 gene models, carried by seven scaffolds, map to regions h35–h36 on 2Lh and h45 (distal)–h46 on 2Rh, which together represent ∼30% of chromosome 2 heterochromatin. By contrast, only 22 gene models, carried by four scaffolds, are present in the large h37–h45 (proximal) segment that accounts for ∼70% of chromosome 2 heterochromatin. Finally, 14 more genes (scaffolds AE003787 and AE003786) have been included in the map, which are likely to be scattered somewhere at the transition region between heterochromatin and euchromatin.
Among the 76 heterochromatic gene models annotated on chromosome 2, 7 corresponded to the known essential genes defined by genetic analyses: RpL5, light, concertina, rolled, RpL38, Nipped-B, and Nipped-A (Hilliker 1976; Devlin et al. 1990; Parks and Wieshous 1991; Biggs et al. 1994; Rollins et al. 1999; Tulin et al. 2002; Myster et al. 2004; Marygold et al. 2005). Thus, a significant portion of essential genes that mapped to chromosome 2 heterochromatin (Figure 4) still need to be identified molecularly. Notably, among those genes, l(2)41Ad is the only known vital gene mapping to region h44 (Dimitri 1991) and may be implicated in proper morphogenesis of the legs and wings (Hilliker 1976; Dimitri et al. 2005b). Region h44 contains at least five gene models (CG17691, CG40067, CG40068, CG40388, and CG40390) carried by AABU01002750, but, at present, there are no molecular data linking any of them to l(2)41Ad. Notably, CG17691, CG40068, and CG40390 encode polypeptides that share similarity with known human proteins, which are apparently not related to morphogenetic functions (Table 2). Interestingly, l(2)41Ad is highly mutable in I-R hybrid dysgenesis (Dimitri et al. 1997) and most 41Ad lethal alleles are associated with cytologically visible deletions of h43–h44 spanning up to ∼1 Mb of DNA (Dimitri et al. 2005a). These genetic and cytological features suggest that l(2)41Ad is a large locus. The largest gene model found in AABU01002750 is CG40388, which is ∼12 kb, a size similar to that of the light gene, but smaller compared to that of other vital heterochromatic genes of chromosome 2 such as Nipped-B, rolled, and Nipped-A, which are ∼39, 50, and 73 kb long, respectively (http://www.flybase.org). Together, these observations do not support any obvious correspondence between l(2)41Ad and any given gene model mapped to h44. More studies are required to establish the molecular identity and function of l(2)41Ad.
Previous works identified vital genes and gene models in the heterochromatin of chromosome 3 (Marchant and Holm 1988; Hoskins et al. 2002; Koryakov et al. 2002; Corradini et al. 2003; reviewed by Fitzpatrick et al. 2005). Here we mapped eight additional gene models: CG17514, CG40103, CG40163, CG10837, CG40105 to region h48 of 3Lh and CG17419, CG40065, and CG411013 to region h58 of 3Rh. Interestingly, five vital genes were mapped to these regions by Koryakov et al. (2002): l(3)80Fa, l(3)80Fb, and l(3)80Fc, which lie in h48 and l(3)81Fa and l(3)81Fb located in h58. On the basis of the complementation behavior of mutant alleles, l(3)81Fb was suggested as corresponding to the zeppelin (zep) locus (Fitzpatrick et al. 2005). At present it is unclear whether or not the eight gene models that mapped to h48 and h58 are candidates for essential genes in these regions; a possible exception may be CG17419. This gene model encodes a protein carrying ankyryn repeats (Table 2) and may correspond to zep. In fact, the zep product was suggested to be involved in cell–cell adhesion (Fitzpatrick et al. 2005), which is consistent with the finding that ankyrin repeats can be found in cell–cell interaction proteins (Mosavi et al. 2004).
As shown in Table 2, some of the products encoded by gene models share significant identity with known proteins from humans and other organisms. However, gene models and essential genes in D. melanogaster heterochromatin apparently do not have molecular functions that would distinguish them from genes located in euchromatin (Hoskins et al. 2002). Essential genes are on average significantly larger than euchromatic genes, due to the occurrence of long introns enriched in transposable element (TE)-related DNA sequences (Devlin et al. 1990; Hoskins et al. 2002; Tulin et al. 2002; Dimitri et al. 2003, 2005a). Surprisingly, although some of the gene models mapped to heterochromatin are also large, such as CG40130 (54 kb), CG40129 (55 kb), CG17514 (45 kb), and CG40080 (34 kb), this does not seem to be a general rule. For example, with the exception of rolled (CG12559), which is large (∼50 kb), most genes carried by scaffolds AABU01002199 and AABU01001947 (region h41) are <6 kb, and six of them are even very short (<600 bp). A similar situation is found in scaffolds AABU01002549 and AABU01002750.
Several hypotheses could explain these observations. First, during evolution, older genes in heterochromatin may increase their size due to TE insertions in the introns. If that is the case, then it would follow that short gene models in heterochromatin are younger, or, in other words, have recently “moved” to heterochromatin. Second, genes in heterochromatin may be differentially targeted by TEs, with some genes being somehow more refractory than others. Third, there might be a selective pressure to maintain some genes of short size in heterochromatin due to particular functional properties. Highly expressed genes indeed tend to have substantially shorter introns than those expressed at low levels (Castillo-Davis et al. 2002). In this regard, it is worth noting that three highly expressed ribosomal protein genes, i.e., RpL38, RpL5, and RpL15, all of short size, are located in heterochromatin (Marygold et al. 2005; Schulze et al. 2005). However, we did not detect significant differences in the expression of the short and large gene models tested by Northern blotting. Finally, some predictions may represent different portions of the same gene, or may be spurious, leading to an underestimate of the actual size of genes in heterochromatin and to an overestimate of the actual gene number.
Are D. melanogaster heterochromatic genes also found in heterochromatin in other organisms? It recently has been found that genes located in D. melanogaster heterochromatin have euchromatic orthologs in D. pseudoobscura and D. virilis (Yasuhara et al. 2005). Here, we report that 70% of heterochromatic gene models of chromosome 2 encode putative proteins sharing significant similarity with human proteins (Table 2). However, the corresponding human orthologs are invariably located in euchromatin. Although the function and regulation of known heterochromatic genes and their homologs in other eukaryotes is far from being elucidated, these data suggest that a heterochromatic location may not be crucial for the proper function of a given gene.
Developmental expression of heterochromatic gene models:
We also analyzed heterochromatic gene expression throughout the Drosophila life cycle in Northern experiments (Figure 3). RNAs for the vital gene Nipped-A and 15 gene models were found to be transcribed in embryos, larvae, pupae, and adults. Although much remains to be learned about the number, structure, regulation, and evolution of heterochromatic genes, some interesting features are beginning to emerge from the developmental expression of gene models resident in pericentric heterochromatin. In particular, our observations indicate that, similarly to Nipped-A, the expression of the heterochromatic gene models studied here is not limited to specific stages, but is present throughout development. This is also true for vital heterochromatic genes such as light, rolled, RpL5, RpL38, RpL15, Nipped-B, and Parp (Biggs et al. 1994; Rollins et al. 1999; Tulin et al. 2002; Marygold et al. 2005; Schulze et al. 2005). This finding is consistent with the idea that genes resident in a heterochromatic environment can code for essential functions.
Correspondence between mitotic and polytene chromosome heterochromatin:
In D. melanogaster polytene tissues, the heterochromatic regions from all chromosomes aggregate to form a cytological structure called the chromocenter. Two cytological domains of heterochromatin can be distinguished in the chromocenter (Heitz 1934): α-heterochromatin, which corresponds to a small compact region located in the middle of the chromocenter and undergoes little if any replication during polytenization (Gall et al. 1971), and β-heterochromatin, a diffusely banded mesh-like material that lies between euchromatin and α-heterochromatin and is extensively polytenized (Zhang and Spradling 1994; Berghella and Dimitri 1996). The assembly and organization of the chromocenter represents one of the most intriguing aspects in the mitotic-to-polytene chromosome transition. Although the structural relationships between mitotic and polytene chromosomes have been analyzed in several studies, only a few have focused on the correspondence between mitotic and polytene chromosome heterochromatin (Koryakov et al. 1996).
Here, we report the mapping of 10 cDNAs to mitotic and polytene chromosome heterochromatin. All tested probes on polytene chromosomes produced hybridization signals of significant fluorescence intensity, indicating that the corresponding gene models undergo polytenization, at least to some extent. This observation is consistent with previous findings showing that single-copy genes located in deep heterochromatin are extensively polytenized (Devlin et al. 1990; Berghella and Dimitri 1996), in contrast with satellite DNA blocks that fail to undergo polytenization and exemplify the well-established paradigm of heterochromatin underreplication. In addition, several cDNA probes produced heterochromatic FISH signals split into two. A similar pattern was produced by some PZ heterochromatic insertions (Zhang and Spradling 1994) and by the rolled cDNA (Berghella and Dimitri 1996) and suggests that sister chromatids retain their identity within the chromocenter (Zhang and Spradling 1994).
Our cDNA mapping may have an intersting outcome for studying the correspondence between mitotic and polytene chromosome heterochromatin of chromosome 2 (Figure 2 and Figure 5). On 2Lh, polytenization of sequences in region h35 distal and h36 proximal contributes to division 40D-F. This correspondence is consistent with the previous finding by Koryakov et al. (1996) showing that region h35 corresponds to division 40D-F. In the same study (Koryakov et al. 1996), the authors also suggested that h46, the most distal region of 2Rh, corresponds to division 41C. More recently, we have extended this observation and shown that polytenization of DNA sequences located in h46 contribute to form division 41C-D (Corradini et al. 2003). Moreover, we found that, in the very proximal heterochromatin of 2R, sequences from the large h41–44 portion contribute to form polytene division 41A, while sequences located at the h45–h46 transition can contribute to polytene division 41B.
We are grateful to Gary Karpen and Patrizia Lavia for critical reading of the manuscript and suggestions; to the Drosophila Heterochromatin Genome Project (Roger Hoskins, Chris Smith, Joe Carlson, and Gary Karpen) for sharing unpublished data; and to Rachele Alicandro, Katia Gazzetti, Cinzia Rizzo, and Laura Somma for helpful assistence with BLAST analysis and polytene chromosome preparations. This work was supported by grants from Ministero dell'Università e della Ricerca “Progetti di ricerca di interesse nazionale 2004–2006” and from Istituto Pasteur-Fondazione Cenci-Bolognetti.
Communicating editor: K. G. Golic
- Received August 29, 2006.
- Accepted November 1, 2006.
- Copyright © 2007 by the Genetics Society of America