BIOLOGICAL OVERVIEW

How do cells maintain or change gene expression during differentiation? Two genetically complex systems are involved: the Polycomb group (Pc-G) of proteins and the trithorax group (TRX-G). Each forms a large multiprotein complex to fulfill the functions of modifying gene expression. Molecular analysis of the components of these multiprotein complexes has resulted in an understanding of the role of chromatin structure in gene regulation. Furthering this understanding has been a comparison of homologies across species, from yeast to humans. Pc-G proteins cooperate to induce changes in chromatin that inactivate genes, while TRX-G proteins cooperate to modify chromatin structure to activate transcription.

Current understanding of the TRX-G stems from the intersection of classical genetics with the biochemistry of chromosomes. trithorax mutants mimic the homeotic transformations caused by mutations of homeotic genes such as Antennapedia and Ultrabithorax. trithorax mutants exhibit a transformation of posterior abdominal structures to more anterior structures. This interaction between trx and homeotic genes has lead to observations that trx is necessary to sustain homeotic gene expression past the gastrulation phase (Breen, 1991).

Chromatin is the constituent of nuclei and chromosomes that interacts with colorful dyes, when viewed with light microscopy. The main constituents of chromatin are DNA and histones, the proteins that organize the long DNA double helix into chromosomes. For many years, histologists have distinguished between two types of chromatin: the genetically active euchromatin and the genetically inactive heterochromatin. It is now clear that the Pc-G and TRX-G proteins are involved in the transition of DNA between activity and inactivity, and that this transition involves a change in the molecular makeup of chromatin. Various yeast proteins, part of a large protein complex, activate silenced yeast genes, in particular mating type genes.

Homology of structure implies homology of function. Discoveries of homology between organisms as diverse as yeast and man can be directly applied to understanding of function across all species. Many of the yeast proteins have both human and Drosophila counterparts. Just as the fly's TRX-G proteins activate transcription in Drosophila, the yeast and human proteins have homologous functions in their respective organisms. One protein of a large gene activating protein complex in yeast, the SWI/SNF complex (pronounced switch-sniff), is homologous to Drosophila brahma gene. The Brahma protein contains a bromo domain and a DNA-dependent ATPase/helicase domain. A second Drosophila protein, SNR1, has been found to be homologous to another subunit of the yeast SW1/SNF complex (Dingwall, 1995). Yet a third Drosophila protein (ISWI) is homologous to another subunit of SWI/SNF and known human member of the SWI/SNF family (Tsukiyama, 1995).

Nuclesomes constitute the principle histone multiprotein complex that make up the chromatin structure that organizes DNA into chromosomes. The biochemical action of the SWI/SNF complex (the disruption of nucleosomal structures) makes promoters available for activation by transcription factors. This chain of events has provided insight into the mechanisms by which TRX-G proteins activate transcription.

The Trithorax protein contains a zinc finger domain, homologous to many transcriptional activators, including a human homolog involved in oncogenic transformation in cases of acute leukemia. Polytene chromosomes are especially thick chromosomes in the fly's salivary glands in which the DNA has been duplicated a thousand times over. TRX protein can be immunolocalized on polytene chromosomes at 16 binding sites that overlap those occupied by Polycomb group proteins. This overlap suggests that Trithorax is involved in a multiprotein complex made up of both gene silencing (Polycomb) and activating (Trithorax) subunits. In the absence of certain members of the Pc-G (Enhancer of zeste or Posterior sex combs proteins), TRX protein becomes disassociated from its chromosomal binding site, providing further evidence of the likelihood that a large protein complex is involved in activating and repressing genes (Kuzin, 1994).

In these investigations, classical genetics and cell morphology have teamed with molecular genetics, cooperatively amassing information in an ongoing clarification of the roles of TRX-G genes. It has been established that there are two large protein complexes involving TRX-G genes, both of which activate genes, one activating inactive genes, and the other sustaining the activity of already activated genes. It is now also clear that transcription factors are not mere lumps of protein, sitting on promoters and activating transcripition by RNA polymerase, but rather are elements in a large array of proteins possessing enzymatic activity, able to use the energy of the cell to open chromatin structures or actively silence genes (Simon, 1995 and Orlando, 1995).

GENE STRUCTURE

The trithorax locus of Drosophila affects segment determination primarily in the thoracic region. Mutant flies show transformations of the third and, to a lesser extent, first thoracic segment toward the second thoracic segment; abdominal transformations also occur. Genetic evidence suggests that these effects are based on interactions between trx and genes of the bithorax complex and Antennapedia complex. Further, interactions between the maternal effect locus female sterile homeotic (fsh: coding for a combined pfkB family of carbohydrate kinase and Bromodomain protein) and trx have been observed. To aid in a molecular analysis of trx function, the locus has been cloned by a P-element transposon tagging approach. Five insertion mutations have been mapped within a region of about 10 kilobases; one of these mutations reverted coincident with the loss of the insertion. Transcription mapping suggests that two RNAs of about 12 and 15 kilobases are the major transcripts of the trx locus and that the transcription unit comprises a region of about 25 kilobases. Transcripts from the trx locus are distributed uniformly in early embryos, but at 14-16 hr after fertilization the ventral nerve cord contains a higher concentration of trx RNA than other regions of the embryo (Mozer, 1989).

The intron/exon structure of the trx gene and the large alternatively spliced TRX mRNAs are capable of encoding only two protein isoforms. TRX proteins differ only in a long Ser- and Gly-rich N-terminal region, encoded by exon III, present only in the larger trx isoform. Trithorax has a novel variant of the highly conserved nuclear receptor type of DNA binding domain (Stassen, 1995). Five different transcripts can be identified. Each one uses the same first exon. Transcript M has no second or third exon, and a short 3'UTR. Transcript ME also has no second or third exon, and a long 3'UTR. These two transcripts code for the low molecular weight protein isoform. Transcripts E1, E2 and L differ in their use of second and third exons as follows: E1 uses the second and not the third; E2 uses the third and not the second, and L uses both. E1and L, both using the third exon, code for the high molecular weight protein isoform, while E2 codes for the low (Sedkov, 1995 and Stassen, 1995).

Exons - eight, with alternative splicing of multiple 5' exons.

Bases in 3' UTR - There are two alternative 3'UTRs, one of 154 nucleotides and the second of 1578 nucleotides (Sedkov, 1994).

PROTEIN STRUCTURE

Amino Acids - 3358 and 3726 for the two isoforms (Stassen, 1995)

Structural Domains

The trithorax gene functions in segment determination in Drosophila through interaction with genes of the bithorax complex and Antennapedia complex. Genetic evidence suggests that trx may be considered a positive regulator of homeotic genes. Sequencing of cDNAs corresponding to the entire trx transcription unit reveal the existence of an unusually long open reading frame encoding 3759 amino acids. The main features of the predicted trx protein are several cysteine-rich regions that can be folded into zinc finger-like domains. Cysteine-rich portions expressed from trx cDNAs in Escherichia coli are capable of zinc binding in vitro, suggesting a possible function for the trx product as a metal-dependent DNA-binding protein. Analysis of trx mutant embryos with antibody to the Ultrabithorax gene product shows decreased staining in parasegment 6 of the ventral nerve cord of late embryos. However, expression of Ubx is not affected in embryos carrying the lethal mutation trxE3, in which one of the putative zinc finger-like domains of the Trx protein is deleted. This differential effect of the E3 mutation suggests that trx exhibits other function(s) besides those involved in the regulation of Ubx expression in the ventral nerve cord of the embryo (Mazo, 1990).

The smaller isoform of TRX lacks an N-terminal serine and glycine rich region of 368 residues. These residues are potential phosphorylation targets. There are four PEST residues found in proteins with short half lives. The DNA binding domain (DBD) found in residues 762-860 is a variant of that found in the nuclear receptor superfamily, and has high homology to Drosophila proteins Knirps, Knirps-like, Ecdysone receptor, Ultraspiracle, FtzF1 and Tailless, as well as retinoid-X receptors. The D box or second finger is involved in DNA binding-induced DBD dimerization The more N-terminal P Box (or first finger) contains three residues responsible for sequence specificity in DNA binding (Stassen, 1995).

The Trithorax protein has a C4HC3 zinc finger motif.

This C4HC3 motif has been identified in a variety of proteins including the Drosophila trithorax and its human homolog ALL-1 involved in oncogenic translocations in acute leukemias. This domain, whose proposed name is TTC (for Trithorax consensus), is more frequently referred to as PHD fingers (for plant homeodomain fingers). This domain is found in many transcriptional regulators or DNA-binding proteins. Interestingly, TTC/PHD was found in several bromo domain containing transcriptional adaptors including the E1A-binding p300 and the CREB-binding CBP proteins. In CBP, this domain does not appear to be involved in DNA, CREB or TFIIB binding. (Koken, 1995).

The cys-rich zinc finger central region is also present in Polycomb-like and RBP2, a retinoblastoma binding protein. TRX proteins terminate with another novel conserved domain. This tromo domain is shared with Drosophila Enhancer of zeste, Sup(var)3-9, and human and mouse ALL-1 proteins (Stassen, 1995).

The C-terminal tromodomain portion of the Trithorax protein (is part of a larger region previously shown to share extensive homology with a human protein (ALL-1/Hrx), implicated in acute leukemias. Molecular analysis of a Polycomb group protein, Enhancer of zeste, predicts a 760-amino-acid protein. A region of 116 amino acids near the E(z) carboxy terminus is 41.2% identical (68.4% similar) with a carboxy-terminal region of the Trithorax protein. Over this same 116 amino acids, E(z) and ALL-1/Hrx are 43.9% identical (68.4% similar). Otherwise, E(z) is not significantly similar to any previously described proteins. It is interesting that this region of sequence similarity is shared by two proteins with antagonistic functions; it may comprise a domain that interacts with a common target, either nucleic acid or protein. It is a region moderately enriched in both acidic and basic residues (Jones, 1993 and Stassen, 1995).

The C-terminal domain of Trx has been termed a tromodomain; the more generally accepted term for this is a SET domain (for Su(var)39, E(z) and trx, the three proteins in which domain was first identified). Trx has an atypical bromodomain between the third and fourth PHD Zn fingers; these Zn fingers are highly conserved domains in the middle of the protein. With regard to the antagonistic functions between Trx and E(z), some E(z) mutants enhance the trx phenotypes rather than repressing them (Manuel Diaz, personal communication to the editor of The Interactive Fly).

The D. virilis trithorax gene has been isolated and sequenced. It produces a set of transcripts similar to that of D. melanogaster, and it encodes a protein that shows sequence similarity in several domains which are also conserved in the human homolog, ALL-1/HRX (Tillib, 1995).

date revised: 5 September 98

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.