DEVELOPMENTAL BIOLOGY:
Combing Over the Polycomb Group Proteins

The proteins produced by Polycomb and other genes with similar developmental effects--they're called the Polycomb group proteins--for the most part turn off other developmental control genes that establish the fates of specific cells in the developing embryo. Often this suppression--which occurs once the developmental control genes have done their work--is permanent and heritable, passed down to all those cells' daughters throughout the life of the organism. "How can you keep something off for the lifetime of the organism?" asks biochemist Robert Kingston of Harvard's Massachusetts General Hospital in Boston. "That's been fascinating to a lot of us for years."

The new work shows that the Polycomb group proteins, working in various combinations with one another, accomplish this feat by altering chromatin, the complex of DNA and associated histone proteins that together comprise a cell's chromosomes. One set of the proteins first marks the genes to be silenced by attaching methyl groups to a specific histone called H3. A second set then comes in to block transcription of the marked genes into messenger RNA, although there is controversy about how they actually do this.

Biologists are now finding that Polycomb group proteins affect other important developmental events besides cell fate determination. They are apparently needed to maintain the stem cells that form and replenish the body's tissues. They also help inactivate one of the two X chromosomes carried by female cells, which is needed to prevent an overdose of X gene expression. In addition, mirroring findings on other key developmental control genes, researchers have recently linked abnormal expression of one of the Polycomb group genes to the development of prostate, breast, and other cancers.

Gene turnoff. The methyl groups added to histone 3 of chromatin by the Polycomb group complex PRC2 attract PRC1, which then shuts down nearby gene activity. CREDIT: P. HUEY/SCIENCE

The development of the flies had apparently been altered so that their more posterior segments were producing structures ordinarily found on more anterior segments. In the work that would eventually win the Nobel, Ed Lewis went on to discover a series of developmental mutations that disrupted the fly's normal segmentation pattern, often causing anterior structures to shift toward the rear.

Mutational studies suggested that several of the genes responsible for these shifts in cell fate determination were linked together in the genome, forming what became known as the bithorax complex. The genetics also suggested that the Polycomb protein normally represses bithorax gene expression, keeping the genes off in body segments where their products don't belong. This prevents structures such as sex combs or wings from forming in the wrong body segments.

Indeed, this is how the fly permanently shuts down these developmental genes. Polycomb can maintain gene repression for the life of the fly, says Jeffrey Simon of the University of Minnesota, Twin Cities. And the original Polycomb is not alone in this gene repressive activity. Over the years, fruit fly geneticists identified several more genes that can, when mutated, produce similar shifts in segmental structures, indicating that they, too, suppress bithorax and other gene activities.

Today, the Polycomb group of genes has some 15 members. The others were also discovered on the basis of their mutational effects on flies, and for the most part they are not structurally related to one another. They are widely distributed in nature, however. Polycomb group genes "are found in organisms from flies to humans," Simon says. "Nearly every one is conserved."

Tied up in knots. When not condensed, chromatin exists in a "beads on a string" conformation (left). But when treated with PRC1, the beads clump together (middle, right). CREDITS: ; N. J. FRANCIS ET AL., SCIENCE 306 (2004)

Analysis of the gene sequence provided the first clue to the Polycomb protein's modus operandi. The gene encodes a protein with a stretch of 37 amino acids that is similar to a known chromatin-binding domain in a protein called HP1 (for heterochromatin- associated protein 1). That suggested that Polycomb interferes with gene activity by attaching to chromatin in some fashion.

Shortly thereafter, researchers, including Simon, Paro, who is now at the University of Heidelberg, Germany, and Vincenzo Pirrotta, who recently moved from the University of Geneva, Switzerland, to Rutgers University in Piscataway, New Jersey, identified DNA sequences called Polycomb responsive elements (PREs). These are base sequences that are necessary for the repression of nearby genes by Polycomb group proteins. The assumption is that the sequences help attract the proteins to the right genes. Although uncertainties remain, researchers have recently built a picture of how that happens.

In particular, they've shown that gene inactivation requires the cooperation of two complexes of the various Polycomb group proteins. The first, called PRC1 (for Polycomb repressive complex 1), was isolated from the fruit fly about 5 years ago by Kingston, Nicole Francis, who is also at Harvard, and their colleagues. PRC1 contains four core proteins--Polycomb itself plus PH (polyhomeotic), PSC (posterior sex combs), and dRING1--and binds to chromatin. Once there, it blocks the effects of a known gene-activating protein complex called SWI/SNF. Humans, it turns out, carry structurally similar proteins, which form a complex with similar activity. PRC1 "seems to be the engine of [gene] repression," Kingston says.

The identification of a second complex of Polycomb group proteins, PRC2, provided a major insight into how PRC1 knows which genes to target. In 2002, four groups, those of Kingston, who was working with Simon and Jürg Müller of the Max Planck Institute for Developmental Biology in Tübingen, Germany, Pirrotta, Danny Reinberg of the University of Medicine and Dentistry/Robert Wood Johnson Medical School in Piscataway, and Yi Zhang of the University of North Carolina, Chapel Hill, came across PRC2 more or less simultaneously.

Developmental regulator. The Polycomb protein (cyan), a portion of which is shown here bound to a histone (yellow), helps ensure that structures like sex combs (left, in the micrograph) develop on correct fruit fly body segments. CREDITS (LEFT TO RIGHT): ARTYOM KOPP/UNIVERSITY OF CALIFORNIA, DAVIS; J. MIN ET AL., GENES & DEVELOPMENT (2003)

The researchers found that both complexes target the same chromosomal sites and that PRC2's methylating activity is needed for PRC1 binding. When PRC2 methylates histone 3, it's "like putting a little signpost in the chromatin that says 'PRC1 bind here,' " Simon explains. Although there is still some uncertainty about how PRC2 finds the right chromatin regions to tag, a team including Richard Jones of Southern Methodist University in Dallas, Texas, Judith Kassis of the National Institute of Child Health and Human Development in Bethesda, Maryland, and Zhang have identified proteins that interact both with PREs and with PRC complex proteins that might possibly be involved in such targeting.

Some uncertainties
Another outstanding issue for Polycomb researchers concerns how PRC1 inhibits gene activity. The simplest possibility is that it compacts the chromatin structure so that the transcribing machinery can't get access to the gene. There is some evidence for this. Isolated chromatin looks something like beads on a string; the beads are the so-called nucleosomes, consisting of DNA wound around a cluster of histone proteins, and the string is additional DNA that links the nucleosomes. Last year, Kingston, Francis, and Christopher Woodcock of the University of Massachusetts, Amherst, used electron microscopy to show that PRC1 compacts such nucleosome arrays, apparently causing the beads to clump together to the point at which they can no longer be distinguished. The researchers found that this compaction requires a segment of PSC, one of PRC1's core proteins that is needed for gene repression, a result indicating that the two activities are linked.

But other researchers, such as Pirrotta, aren't so sure that PRC1 works simply by condensing the chromatin and thus blocking out the transcription machinery. Using a standard reporter gene assay for PRC1-mediated silencing, he and his colleagues recently showed that such silencing doesn't prevent binding by RNA polymerase, the enzyme that copies the DNA into messenger RNA. Instead, PRC1 apparently keeps the polymerase from transcribing the gene. "When we looked at the promoter [where the enzyme binds], RNA polymerase is there, but it can't get moving and open the DNA strands" to allow transcription, Pirrotta says. More work will be needed to clarify this issue, but Kingston, for one, suggests that both mechanisms, DNA compaction and inhibition of the transcription machinery, might conceivably come into play.

Although methyl addition to histone 3 by Polycomb group proteins can clearly tag genes for inactivation, the finding doesn't explain what makes the inactivation permanent. "The repressed state remains over many mitotic [cell] divisions. How is it maintained during DNA replication?" Paro asks. Recent results from his lab suggest a possibility.

In work published online on 1 March in Genes and Development, the Heidelberg workers described evidence suggesting that Polycomb inactivation of PRE-associated genes occurs continuously unless something intervenes to prevent it. Thus, the silenced state could be maintained throughout the lifetime of the organism. But obviously, not all of these genes are shut down during development. Some remain "on" to produce the fly's normal segmental structures and perform other cellular functions. The Paro group has evidence that this active state is enabled by ongoing transcription of the PRE sequences, which somehow prevents Polycomb-mediated silencing, possibly because the transcription alters chromatin structure in such as a way as to block Polycomb binding.

Cancer indicator? Micrograph A shows normal prostate epithelium, B shows a precancerous lesion, and C, full-fledged cancer. As the cancer progresses, EZH2 expression (purple in D, E, and F) increases. CREDIT: A. KUZMICHEV ET AL., PNAS 102, 6:1859 (2005)

The Polycomb group proteins also have roles beyond developmental regulation. By surveying the fruit fly genome for PRE sequences, Paro and his colleagues identified more than 150 genes throughout the genome that could be subject to Polycomb repression. Among these were various genes involved in controlling cell growth and division.

Consistent with that, researchers have recently linked anomalies in Polycomb group gene expression with cancer development and progression. In particular, Arul Chinnaiyan of the University of Michigan Medical School in Ann Arbor, Mark Rubin of the Dana-Farber Cancer Institute in Boston, and their colleagues have looked at the expression of EZH2, the human equivalent of the fruit fly E(z) protein, in prostate and breast cancers. They found that the expression is much higher in cancers that have spread (metastasized) to other tissues than it is in localized tumors or normal tissue. Working with a mouse model of prostate cancer, Reinberg and his colleagues have confirmed that EZH2 production goes up as the cancers progress from localized to metastatic.

Increased EZH2 expression may in fact be a much-needed prognostic indicator for prostate cancer. Although many men develop small, localized prostate tumors as they age, most of these never progress and metastasize. "Most people die with [prostate cancer] rather than of it," Chinnaiyan says. But some of those localized tumors will metastasize, and currently it's impossible to identify the dangerous ones. This means that men may have to undergo therapy unnecessarily, and that can produce unpleasant side effects such as incontinence and impotence.

But in a small study of surgically removed human prostate cancers, published in the 10 October 2002 issue of Nature, the Chinnaiyan team found that increased EZH2 expression in small, localized tumors was associated with a high risk of eventual disease spread. The overexpression "portends aggressiveness and metastasis," Chinnaiyan says. He and his colleagues are now organizing a larger clinical trial to confirm these preliminary findings. In addition, the protein may even provide a target for anticancer drugs. Chinnaiyan and colleagues have found that blocking production of the protein inhibits the proliferation of prostate cancer cells.

How EZH2 overproduction contributes to cancer development remains murky, but one possibility is that it disturbs normal gene control. Because Polycomb group proteins mainly repress genes, a flood of EZH2 may inhibit tumor-suppressor genes or genes that make proteins that keep cells anchored in place so that they can't migrate to new tissues as metastatic cells do.

Another clue comes from Reinberg and his colleagues. They found that EZH2 overproduction leads to formation of a Polycomb protein complex that differs in protein composition from PRC2. This could also lead to changed patterns of gene expression, he suggests.

Intriguingly, EZH2 overexpression and formation of the PRC variant occurs in undifferentiated cells as well as in cancer cells. This is consistent with the views of some researchers that cancer cells behave as if they have regressed to a more primitive developmental state. It is also consistent with recent findings by Jenuwein, Azim Surani of the Wellcome/CRC Institute of Cancer and Developmental Biology in Cambridge, U.K., and others suggesting that histone methylation mediated by EZH2 helps maintain stem cells in their pluripotent developmental state.

The Polycomb group proteins are clearly turning out to be highly versatile players in a wide range of cellular activities. And still more revelations may be in store. Within the past year, researchers including Brockdorff and Zhang have reported that some Polycomb group proteins can add the small protein ubiquitin to histone H2A. Originally discovered as a tag that marks proteins for destruction, ubiquitin has since been shown to have many other roles in the cell, including regulation of gene expression and protein migrations (Science, 13 September 2002, p. 1792).

The Polycomb-mediated histone ubiquitination is involved in gene silencing, but Zhang says its exact role isn't yet known. One thing is clear, however. At 60 years of age, the Polycomb group proteins are still showing plenty of life.

Summary of this Article PDF Version of this Article E-Letters: Submit a response to this article   Download to Citation Manager Alert me when: new articles cite this article   Search for similar articles in:   Science Online   PubMed Search Medline for articles by: Marx, J.   Request permission to use this article   This article appears in the following Subject Collections: Development

Volume 308, Number 5722, Issue of 29 Apr 2005, pp. 624-626.
Copyright © 2005 by The American Association for the Advancement of Science. All rights reserved.