Steroids contain four rings of carbon atoms and are used by vertebrates as
hormones. Androgens, for example, promote the growth of facial hair and muscle
in men. Plants also have hormones, but those studied until now have been simpler
molecules.

Joanne Chory and her team at the Salk Institute for Biological Studies in La
Jolla stumbled upon the hormonal role of plant steroids while studying a mutant
of thale cress, Arabidopsis thaliana. The mutants remain stunted and
their leaves cannot photosynthesise properly. The researchers pinpointed the
defective gene, called det2, and found that it is similar to a human
gene that codes for an enzyme used to make androgens.

This suggested that the mutant plants were unable to make a steroid required
for growth. The most widespread plant steroids are the brassinosteroids, and the
pathway used to make one of these steroids, brassinolide, involves an enzyme
similar to that encoded by the det2 gene. When the researchers added
brassinolide to det2 mutant seedlings, the plants grew normally (
Science, vol 272, p 398). “Now we have the genetic evidence that if you
block synthesis of the steroid, the plant cannot develop,” says Jianming Li, one
of the researchers.

In the latest issue of Cell (vol 85, p 171), a team led by
scientists at the Max Planck Institute for Plant Breeding Research in Cologne
describes other Arabidopsis mutants which underline the hormonal role
of plant steroids. The mutants are stunted and have defects in other genes
involved in the manufacture of brassinolide.

Peter Glazer and his colleagues at Yale, and at the biotech company OncorPharm in Gaithersburg, Maryland, stumbled across the effect by chance. They were testing ways of directing mutagenic molecules to specific targets on DNA. The aim was to introduce mutations to inactivate a gene – a technique that might be useful in gene therapy. The mutagen was tethered to a short DNA sequence, or oligonucleotide, which is designed to bind to a particular sequence of double-stranded DNA, creating a triplex.

To their surprise, the researchers found that they didn’t need the mutagen to trigger mutations in the target DNA. In the experiment controls, Glazer’s team introduced bare oligonucleotides into cultured monkey cells containing viral DNA. Each oligonucleotide could bind to certain sequences on the viral DNA helix and form a triplex. The target DNA then mutated at about 13 times the rate of background mutations.

“Glazer’s exciting result is that he didn’t need to add a tethered mutagen [in order to] induce repair and mutation,” says chemist Peter Dervan at the California Institute of Technology in Pasadena.

The researchers suggest that enzymes in the cell were cutting out the unusual stretch of triplex DNA and then repairing the double helix by filling the sequence in again. Because the repair is sometimes inaccurate, mutations appear in the target sequence. They confirmed this picture by showing that in special cells that cannot perform certain types of DNA repair, oligonucleotides don’t trigger an increase in the mutation rate. The researchers also showed that this repair process happens in human cell extracts (Science, vol 271, p 802).

Because oligonucleotides and triplexes naturally appear in cells from time to time, Glazer says that they may cause the apparently random mutations in the genome. He also suggests that increased rates of mutation found at hot spots in yeast and bacteria may arise where a sequence is more conducive to binding to these naturally occurring oligonucleotides. Laboratory experiments so far suggest that oligonucleotides are very choosy about where they bind.

Glazer’s team has a warning for scientists using oligonucleotides as therapeutic agents – for example, using “antisense” DNA molecules to bind to specific molecules of RNA and inactivate them. If the DNA strands happen to bind instead to relevant DNA sequences on a chromosome, they might introduce unexpected and unwanted mutations.

Denis Henshaw and his colleagues at the University of Bristol have shown that electric fields attract the radioactive decay products of radon and make them vibrate. This would make these carcinogens more likely to stick to the airways and the skin, the physicists say.

“If [the theory] proves right, people who live in high radon areas and under high electromagnetic fields would be likely to be at greater risk,” says Barry Michael of the Cancer Research Campaign’s Gray Laboratory in Northwood, Middlesex.

Epidemiological studies into the effects of EMFs have produced mixed results. Some have suggested that living near overhead cables and other sources of strong EMFs poses little danger. But a growing number have found a weak link between EMFs and the incidence of childhood and adult leukaemias and a range of other childhood cancers.

Magnetic fields are thought to be harmless, and electric fields flow around, rather than through the human body, so biologists have struggled to understand how EMFs could cause cancer. Henshaw believes the link with radon, outlined this week in the International Journal of Radiation Biology, may be the answer.

Radon seeps naturally from uranium-containing rocks. In Britain, background levels are highest in Cornwall, Devon, Derbyshire and Northamptonshire. Even in the regions of high background radon, however, there is a wide variation in exposure between one house and the next. Across the country, 100 000 homes are thought to contain radon which gives off more than 200 becquerels of radioactivity per cubic metre, the level considered hazardous for health. To predict who is at the highest risk of developing cancer, says Henshaw, researchers must go from house to house making individual measurements of the concentration of radon decay products and the strength of the EMFs.

Henshaw used plastic film to record the tiny tracks made by high-energy alpha particles released by polonium-214 and polonium-218, the decay products of radon. From the pattern of tracks left by the particles, the researchers calculated how many polonium nuclei had hit the film and how many were in the air nearby. Seventy pieces of the plastic film were suspended from the ceiling of a cellar in a house in Bristol with a radon concentration of more than 200 becquerels per cubic metre. The researchers applied a voltage across two metal plates placed on the cellar floor, creating an electric field of 5 kilovolts per metre, typical of fields measured 1 metre above the ground, outdoors, underneath a National Grid power line.

When the voltage was applied, about twice as many polonium nuclei hit the pieces of plastic film, and there were up to 50 per cent more radon decay products in the air within a metre of the voltage plates.

Henshaw believes that the radon decay products are attached to tiny water droplets that form an aerosol. Aerosols vibrate in an electric field, and drift towards its source. The two effects make the particles more likely to hit the source. People exposed to strong EMFs “are going to inhale more radon decay products”, says Henshaw. “If these aerosols are vibrating, they are going to stick to your tissues more.”

The government’s radiation watchdog, the National Radiological Protection Board (NRPB), however, is unconvinced that EMFs and radon combined are as dangerous as Henshaw’s work suggests. “It’s an interesting theory … The physical measurements we would not argue with,” says John Stather, the NRPB’s senior assistant director. But he argues that EMFs, by increasing the tendency of aerosol particles to stick to other surfaces in a room, will reduce the amount inhaled.

However, the NRPB is now working with the Medical Research Council and the major cancer research charities on the UK Childhood Cancer Study (UKCCS). The study will measure radon levels, and magnetic and electric fields in households where a child has been diagnosed with cancer and the parents agree to participate. Up to 20 000 of these homes will be studied along with a similar number of control homes in which no children have cancer. When the study is completed in two years’ time, says Stather, it should quantify the true risks. “A comprehensive study such as the UKCCS is the most positive way forward,” he says.

But Henshaw wants further action. He argues for a study that looks at the decay products that are affected by EMFs and which actually provide the radiation dose.

Ray Cartwright, an epidemiologist at the Leukaemia Research Fund Centre at the University of Leeds, says that quantifying the risks will be extremely difficult. National Grid figures show that 23 500 British homes lie within 50 metres of power lines carrying 275 000 volts or more. But given variation in the number of electrical appliances and the configuration of the wiring from house to house, he says, domestic EMFs may not be that closely related to the presence of nearby power lines.

Many biologists argue that evolution is really about the reproduction and survival of genes, rather than individual organisms. This means that natural selection can favour DNA sequences which promote their own survival, even if this is at the expense of other genes in the same genome.

The new genes belong to a family of genes that produce enzymes called homing endonucleases. These enzymes recognise specific sequences of DNA and cut them in two. The genes then hijack a cell’s DNA – repair enzymes to repair the damage, and in the process insert a copy of themselves into the gap.

They can do this because the broken ends of the cut DNA match the genetic sequences found on either side of the endonuclease gene. When DNA is repaired, it must be rebuilt according to a template. So the damaged phage DNA is reconstructed to match the sequence containing the endonuclease gene. By this mechanism, homing endonuclease genes copy themselves from phage to phage like molecular parasites.

As if this behaviour were not selfish enough, Heidi Goodrich-Blair and David Shub of the State University of New York at Albany have now found two genes that are even more self-centred. Rather than locating and splitting vacant insertion sites, the enzymes made by these genes home in on DNA that already contains a homing endonuclease gene and cut the rival gene out (Cell, vol 84, p 211). “Instead of just homing into an empty house, they can displace the current occupant,” notes Russell Doolittle, an expert on DNA evolution at the University of California, San Diego. As before, B. subtilis’s DNA-repair enzymes ensure that a copy of the selfish gene is inserted into the gap.

Goodrich-Blair and Shub’s genes, which recognise and cut slightly different DNA sequences, come from two closely related bacteriophages called SP01 and SP82. The two genes are thought to have evolved from a single common ancestor, presumably a typical homing endonuclease gene.

The SP82 gene is particularly vicious. In a bacterium infected with the two phages, the SP82 gene wins out, and eradicates the SP01 gene from its phage’s genome. “It’s the most self-serving selfish DNA that you could possibly imagine,” says Shub.

“It’s a remarkable report,” agrees Doolittle. The sheer nastiness of the SP82 sequence, in particular, has amazed many molecular biologists. “It’s designed to go in and take out the competition,” says Nancy Maizels of Yale University in New Haven, Connecticut.

The two genes may be the master villains of the genetic world, but even they are not all bad. Goodrich-Blair and Shub have found that, occasionally, they do neighbouring genes an unwitting favour. Sometimes, before B. subtilis’s DNA-repair enzymes get to work, other enzymes nibble away at the loose ends of DNA left behind by the endonucleases. In the course of repairing this more serious damage the neighbouring genes can be copied into the gap, in addition to the selfish sequence.

Neil Brockdorff and his team at the Medical Research Council’s Clinical Sciences Centre at the Royal Postgraduate Medical School in London have made an important breakthrough in understanding X inactivation by showing that a gene called Xist is essential to the process (Nature, vol 379, p 131). “This is the first direct evidence that the Xist gene is involved,” says Peter Goodfellow, a geneticist at the University of Cambridge, who studies sex chromosomes.

Very early in the development of female embryos, one randomly chosen X chromosome is silenced and condensed into a tightly packed form. Thereafter, only a tiny proportion of its genes remain active. Women suffering from incomplete X inactivation may lack part of the X chromosome that contains the human equivalent of the mouse gene studied by Brockdorff’s team – but the missing segment also contains other genes, so scientists had been unable to prove which gene is responsible.

Brockdorff’s team modified a line of cells taken from a mouse embryo. Each cell contained two X chromosomes with subtly different genetic sequences, so that they could be distinguished. One of the chromosomes also lacked about half of the Xist gene. The researchers inserted these cells into early mouse embryos at the 8-cell stage.

Then they looked at what happened to the inserted cells’ X chromosomes some 10 days later. In all cases the copy containing the complete Xist gene was silenced. “The X chromosome with the deleted Xist gene never undergoes inactivation,” says Brockdorff.

The process of X inactivation involves a number of steps. A cell must count its X chromosomes and then “choose” which one to inactivate, before initiating the genetic program that shuts one chromosome down. How this counting and choice is achieved is still unclear. But Mary Lyon of the MRC’s Mammalian Genetics Unit in Didcot, Oxfordshire, who discovered X inactivation in 1961, notes that the chromosome counting step cannot be encoded by the portion of Xist deleted by Brockdorff’s team. Most of the cells still recognised that they needed to inactivate one chromosome, she points out.

One possibility is that different parts of the Xist gene control the counting, choice and shutdown functions. It should be possible to test this by examining the effects of deleting smaller fragments of Xist, says Hunt Willard of Case Western Reserve University in Cleveland, Ohio, whose team discovered the human version of the gene.

Evolutionary biologists have been probing the processes that create new species ever since Charles Darwin published his theory of evolution by natural selection. More recently, geneticists have debated whether this speciation always involves the gradual acquisition of many mutations, each with a tiny effect, or whether a small number of genetic changes with big effects could do the same job.

Toby Bradshaw and his colleagues at the University of Washington in Seattle now say that for monkey flowers, at least, speciation need not be a painfully gradual process. They studied two monkey flower species which do not interbreed in the wild, but which do produce fertile offspring when artificially crossed.

The flowers – Mimulus lewisii, which is pollinated by bumblebees, and M. cardinalis, which is pollinated by hummingbirds but is thought to have evolved from a bee-pollinated species – seem perfectly adapted to their respective pollinators. M. lewisii has pink flowers with two yellow streaks which guide bees towards a drop of nectar. Bees then brush against the pollen-bearing anthers and the pollen-catching stigma. M. cardinalis is red, which is attractive to birds but invisible to bees. Its petals form a narrow tube with a larger volume of more dilute nectar at its base. As a hummingbird hovers with its beak down this tube, protruding anthers and stigma make contact with its forehead.

Bradshaw and his colleagues measured nine easily quantifiable traits in the two monkey flower species, first-generation hybrids of the two and a second generation produced by self-pollinating these hybrids. The traits included the concentration of red pigment in the flowers’ petals, nectar volume and the length of the stamens bearing the anthers.

The researchers then correlated these traits with the inheritance of a set of 153 short marker DNA sequences which differed between the two parent species, scattered randomly across the plants’ eight chromosomes. If a genetic marker lies close to the gene influencing a particular trait, it should be inherited along with the trait. This technique can be used to examine the genetic differences between any two species which can still successfully interbreed.

For eight of the traits, a gene or genes lying in a single small stretch of DNA seemed to be responsible for more than a quarter of the variation they measured. The researchers conclude that just three of these genetic changes would have been enough to separate the two species, switching M. cardinalis from bee to bird pollination (Nature, vol 377, p 762).

It should now be possible to use further breeding experiments to work out the order in which these changes occurred. “I could see nectar concentration or flower colour being the first [change], with others following,” speculates Mark Macnair of the University of Exeter. He believes that the ability of monkey flowers to self-pollinate was important, as the intermediates between the two species may have attracted neither bees nor birds, or may have had a structure that resulted in inefficient pollination.

Sleep scientists have long been on the trail of one or more “sleep factors”. Such substances are thought to accumulate in the cerebrospinal fluid (CSF) that bathes the brain, until they reach a concentration that tells the brain to “nod off”. The new compounds, described by Richard Lerner and colleagues from the Scripps Research Institute, La Jolla, in the 9 June issue of Science (vol 268, p 1506), belong to a new family of fatty acid amide molecules. The most potent analysed so far, cis-9,10-octadecenoamide, is also present in rat and human CSF.

When the team injected it into rats that were 2 hours into a 12-hour dark cycle (being nocturnal they are wide awake and active at this time), the rats went to sleep within 4 minutes, and stayed that way for between 1 and 2.5 hours, depending on the dose.

Recordings of electrical activity in the rats’ brains showed that the induced sleep followed a normal pattern of cycling between different depths of drowsiness, featuring slow-wave sleep and shorter periods of rapid eye movement (REM) sleep. The trouble with most sleeping pills is that in humans they tend to produce a much deeper sleep than this, with even slower brain waves and less REM sleep, leading to morning hangovers and a rebound insomnia.

The new lipids “provide a whole new approach to looking at sleep and wakefulness”, says Steven Henriksen of the Scripps team. Tests are under way of the long-term effects of the compound in rats, he says. Experiments in humans are some way off.

The scientists still do not know where or how the lipids actually work – for example, whether they interact with unique receptors, and what enzymes might be involved in synthesising and destroying the compounds. Such processes are potential targets for a future sleeping pill that might work by regulating the lipid’s activity. Answering these questions looks set to keep the researchers awake at night for some time to come.