Chromosome 21 Reveals Sparse Gene Content

The unveiling of the DNA sequences of human chromosomes represents a new chapter in the unfolding story of genomics, but one with roots in the half-century-old field of cytogenetics. Chromosome-level looks can reveal the specific genes behind certain traits and disorders while providing information on genome organization. The diminutive chromosome 21--the smallest of the human contingent, despite its number as next to last--is the fourth to be described. Its debut in mid-May attracted attention because 21 is a chromosome with a familiarity factor, in the form of Down syndrome, usually caused by an extra copy of the chromosome. A consortium of 62 scientists, working in Japan, Germany, France, Switzerland, the United States, and the United Kingdom, contributed to the massive effort.1 Chromosome 21's surprisingly sparse gene content has been compared to a desert or even a black hole, calling for a reexamination of total genome size estimates. The sequencing of chromosome 21 provides a powerful new tool for the ongoing investigation of the causes and components of Down syndrome. This is the most commonly seen aberration of a nonsex chromosome, presumably because it involves the fewest genes and is therefore less severe than having extra copies of other autosomes, which so drastically derails development that spontaneous abortion typically results. Down syndrome takes its name from Sir John Langdon Haydon Down, a medical superintendent of a facility for the profoundly mentally retarded. In 1886, he coined the politically incorrect term "mongolism" to denote the 10 percent of patients with distinctive facial features. The chromosome connection came later. In 1914, Theodor Boveri, a professor of zoology at the University of Wurzburg, suggested that abnormal numbers or structures of chromosomes could be associated with specific syndromes. Tying Down syndrome to abnormal chromosome number was suggested as long ago as 1932, although researchers weren't able to clearly visualize an extra tiny chromosome in cells from patients until 1958. Top, triplication of chromosome 21 underlies Down syndrome. Bottom, transmission electron micrograph of chromosome 21. Identification of the extra chromosome in Down syndrome as number 21 came in 1959.2 This was a pivotal time in cytogenetics, the field that matches chromosome variants to traits or disorders. Improved microscopy and staining techniques were enabling researchers to move from a system of lumping chromosomes into a few size-based groups to ordering them individually, with chromosome 1 the largest and 22 the smallest. "During the late 1950s and early 1960s, human chromosomes were assigned sizes, and the two smallest pairs were very close. Then, with the advent of better banding techniques, you could see that chromosome 22 was a tad larger than chromosome 21, but it was too late to switch the nomenclature," relates Katheleen Gardiner, a senior scientist at the Eleanor Roosevelt Institute in Denver, and part of the chromosome 21 sequencing consortium. Much happened between 1960 and publication of the first full chromosome sequence in December 1999. Cytogeneticists developed ever more specific stains and DNA probes to highlight and distinguish the chromosomes, revealing the guideposts that would help anchor linkage data to build the first gene maps. Some of these techniques estimated the gene density that genome sequencing today addresses directly. R banding, for example, reveals CpG islands, the repeats of C and G that herald the start of many a protein-encoding gene. Many R bands means, roughly speaking, many genes. But chromosome 21 is so small that it yields but one not-very-meaningful R band. The rise of positional cloning in the 1980s gradually allowed investigators to fill in chromosome maps, an effort that Science magazine visually tracked in each October's genome issue. But it was the cover of Nature that dramatically heralded the coming of the first human chromosome sequence, number 22. The cover depicts two hands reaching toward each other against a celestial backdrop of a DNA chip.3 The 33.5 million bases, representing 97 percent of the long arm of chromosome 22, with 11 gaps, introduced 545 protein-encoding genes and 134 pseudogenes, which are sequences very similar to those of active genes but unexpressed, like genetic graveyards. "When we talk about chromosomes 21 and 22, we refer mostly to long arms. There is a lack of interest in the short arms because they are so repetitive," explains Gardiner. Variability in short arm length among individuals may have accounted for the initial missizing of 21 and 22, she adds. Chromosome 22 harbors genes for schizophrenia, chronic myeloid leukemia, cardiovascular disease, and DiGeorge's syndrome. Its high gene density is consistent with long-held estimates that the whole genome has at least 80,000 protein-encoding genes. Chromosome 22 was followed by three others. U.S. Secretary of Energy Bill Richardson announced on April 13 that researchers at the Department of Energy (DOE) Joint Genome Institute in Walnut Creek, Calif., had obtained draft sequences of human chromosomes 5, 16, and 19.4 The three chromosomes represent 11 percent of the total genome and include 10,000 to 15,000 genes (according to current estimates), among them genes for several types of cancer, a form of dwarfism, Crohn's disease, diabetes mellitus predisposition, and DNA repair proteins. Of historical note is that DOE pioneered the Human Genome Project in the second half of the 1980s before becoming overshadowed by the National Institutes of Health's efforts. Anatomy of Chromosome 21 Overall, chromosome 21 has typical features--repeated DNA sequences at the tips and centromere, and protein-encoding genes scattered along its long arm. Sequencing of 33,546,361 base pairs revealed 225 protein-encoding genes, including 127 known genes and 98 genes predicted from computerized (in silico) sequence analysis, plus 59 pseudogenes. Known genes encode signaling proteins, cellular adhesion molecules, transcription factors, and ion channels. Disease-causing genes include those for the beta-amyloid form of Alzheimer's disease; susceptibility to bipolar affective disorder; and forms of epilepsy, hearing loss, leukemia, and hyperlipidemia. Number 21 is also riddled with duplications and other repeats. A 93-base sequence occurs in 10 copies and is also in other chromosomes, and a six-base sequence repeats 17 times. A seven-million-base stretch harbors just one protein-encoding gene, and three one-million-base areas are devoid of genes. Combined, these "black holes" account for a third of the chromosome. Several lines of evidence had led geneticists to expect chromosome 21 to be gene-light, but perhaps not to the extent that the sequence revealed. "When human genome maps started to emerge in the late 1970s, despite a real focus on 21, the map didn't become gene-dense. Then, bits and pieces of the sequence were released as it was generated. These included 6.5 megabases that led to an estimate of 400 to 500 genes," says Roger Reeves, a professor of physiology at the Johns Hopkins University School of Medicine. Teepu Siddique The 225 protein-encoding genes also fell far short of the 371 genes predicted by UniGene, which is the chromosome-by-chromosome listing of expressed sequence tags (ESTs) maintained by GenBank.5,6 Explains Reeves, "The UniGene project took 30,000 ESTs that were unique--that is, all different from each other--and mapped them over the entire genome. Since they were randomly selected, the ESTs were expected to map randomly among the chromosomes. But chromosome 21 turned out to have 80 percent or fewer of the genes expected for its size." Adds Teepu Siddique, director of the neuromuscular disorders program at Northwestern University Medical School, "If the rest of the genome is similar to chromosome 21, then the number of genes will be drastically reduced to maybe less than half the estimated number." But that's a big "if." Is Extrapolation Valid? The sharp contrast of the gene desert that is chromosome 21 to the jungle of chromosome 22 challenges the very validity of extrapolating from any single chromosome to estimate the size of an entire genome. To an extent, this may be an apples to oranges concern, at least for now, Reeves points out, because researchers use different criteria.7 "The chromosome 21 consortium made some assumptions that could increase the number of genes relative to 22, including computer-based predictions. But at the same time, chromosome 21 used stringent criteria for defining a gene, such as evidence of splicing. This approach would miss any gene that has one exon, or genes with very large introns," explains Reeves. Skip Garner, a professor of biochemistry and internal medicine at the University of Texas Southwestern Medical Center in Dallas, who has worked on automation since the early days of the Human Genome Project, questions the predictive power of UniGene. "The way to estimate genes is to compare the sequence of 21 to EST sequences. Our analysis says that the ESTs are contaminated with trash. All EST totals are overestimated by 30 percent due to a difficulty in the database. So chromosomes 21 and 22 may be close to overlappable in size," he says, adding that determining the true number of unique ESTs is the best way to estimate gene density. "Calculating genome size from the density on a few sample chromosomes is not the way to do it. Fragments of several chromosomes might be better," he says. Garner and others also note that the numbers of protein-encoding genes identified in the genomes of Drosophila and yeast fell short of predictions. Déjà Vu The idea that genes are not laid out evenly among the chromosomes is not new, as the high-resolution chromosome banding introduced in the 1970s vividly illustrated. "Cytogeneticists predicted about 25 years ago that genes are not distributed at random, but are rather clustered in R bands. The genome, being about half R bands, has fewer genes than predicted by a 'clothesline model' of evenly spaced genes," says Julie R. Korenberg, the Geri and Richard Brawerman Chair of Molecular Genetics at the University of California, Los Angeles, who is one of those cytogeneticists. The role of whole-chromosome sequences in confirming and extending what was inferred or known from classical cytogenetics brings up a more subtle aspect of the current evolution/revolution in genetics. That is, genome projects and chromosome sequences have hardly sprung from the ether since 1990. Researchers have been scrutinizing specific regions of specific chromosomes for decades. "The chromosome 21 sequence will be fun to wade through, but we had most of the genes before. Anyone in the field would have looked at genome organization themselves," says Korenberg. The legacy of cytogenetics and early molecular genetics can be lost in the excitement of sequencing and the accompanying enthusiasm. For example, upon the revelation of the chromosome 22 sequence, Francis Collins, director of the National Human Genome Research Institute, was widely quoted as comparing the feat to "seeing an ocean liner emerge out of the fog, when all you've ever seen before were rowboats." Likewise, DOE's Richardson commented at the agency's unveiling of chromosomes 5, 16, and 19, "Scientists can already mine this treasure trove of information for the advances it may bring in our basic understanding of life, as well as applications such as diagnosing, treating, and eventually preventing disease." Although the rowboaters of genetics have been clarifying basic metabolism and physiology, work that has yielded said diagnostics and treatments since long before the genomics era, having the sequence of chromosome 21 will boost ongoing research efforts. Consider "Lou Gehrig's disease." "We identified the first familial amyotrophic lateral sclerosis [FALS] locus on chromosome 21 in 1989-1990, discovering mutations in superoxide dismutase 1 [SOD1]," recalls Siddique.8 "But there may be genes or sequences close to the gene responsible for FALS that may influence the effects of mutant forms of SOD1" revealed in the sequence, he adds. Having the sequence will also catalyze Korenberg's ongoing efforts to dissect Down syndrome, to identify which genes, when present in an extra dose, cause or contribute to the many signs and symptoms. "We look at people who have extra pieces of 21, and see what they have in common. Worldwide, 45 such people are known. We are the major group that takes anyone with anything wrong with 21, and ask, which genes are extra? How many cause heart defects? Visual problems? Language problems?" Korenberg has long considered the study of Down syndrome to be a way to shed light on a myriad of medical problems. "Understanding the ways in which genes are responsible for the associated defects may provide windows to leukemia, aging, congenital heart disease, congenital gut disease, and risk for developing a number of solid tumors. That's not bad for a chromosome that also models the structural characteristics of most of the genome." S Ricki Lewis (rickilewis@nasw.org) is a contributing editor for The Scientist. References 1. M. Hattori et al, "The DNA sequence of chromosome 21," Nature, 405:311-9, May 18, 2000. 2. J. Lejeune, "Les chromosomes somatique des enfants mongoliens," Comptes Rendus des Academy des Science Paris, 248:1721, 1959. 3. I. Dunham et al., "The DNA sequence of human chromosome 22," Nature, 402:489-5. 4. www.doe.gov/news/releases00/aprpr/pr00104.htm 5. www.ncbi.nlm.nih.gov/UniGene 6. P. Deloukas et al., "A physical map of 30,000 human genes," Science, 282:744-746, Oct. 23, 1998. 7. R. Reeves, "Recounting a genetic story," Science, 405:283-4, May 18, 2000. 8. T. Siddique, "Molecular genetics of familial amyotrophic lateral sclerosis," Advances in Neurology, 56:227-31, 1991.

Chromosome 21 Reveals Sparse Gene Content

Interested in reading more?

Become a Member of