LONDON ¿ We all inherit one set of genes from our mother and another from our father, and both sets usually take part in making the proteins that these genes encode. Surprisingly, however, researchers have found that some genes fail to follow this pattern: in a few cases, the gene set from the mother stays silent while only the paternal one is involved in making protein, or vice versa.

This phenomenon is called ¿genomic imprinting.¿ Somehow, the genes concerned carry an ¿imprint¿ which allows them to remember from which parent they were inherited, and which dictates whether they should be active or not. With maternally expressed imprinted genes, only the copy inherited from the mother is active, while with paternally expressed imprinted genes, only the copy inherited from the father makes protein. No one knows what the molecular nature of the imprint is, although much research is currently directed at finding out.

One of the teams of scientists which discovered the existence of imprinted genes was led by professor Azim Surani, of the University of Cambridge. Surani and his colleagues, in collaboration with professor Barry Keverne, have now gone on to examine the role of the Peg3 gene, which is a paternally expressed imprinted gene. They report their study in the April 9 edition of Science, in a paper by Li-Lan Li and colleagues, titled ¿Regulation of Maternal Behavior and Offspring Growth by Paternally Expressed Peg3.¿

Samuel Aparicio, a developmental geneticist at the Wellcome/Cancer Research Campaign Institute of Cancer and Developmental Biology, and at the physiological Laboratory at the University of Cambridge, where the study was carried out, told BioWorld International, ¿We found that a mutation in this gene produces a striking impairment in maternal behavior. This is very interesting because it means that there are now two imprinted genes which are implicated in common in a rather complex behavioral function.¿

The same team reported last year in Nature Genetics that the Met gene, which is also a paternally expressed imprinted gene, has a role in regulating growth and maternal behavior.

For the study into the role of Peg3, the researchers first used gene targeting to produce mice which had a mutation in their Peg3 genes. They then crossed males homozygous for this mutation with wild-type females to produce heterozygous offspring which had inherited the mutation from their fathers. Tests showed that, as expected, the maternal allele of Peg3 was not expressed in these animals, while a marker gene inserted in place of the Peg3 gene as part of the gene targeting experiment was.

These animals were designated +/- heterozygous, and bred with each other. It quickly became apparent to the team that the maternal behavior of the +/- mothers was abnormal. Only 8 percent of first litters from +/- mothers survived to weaning age, compared with 83 percent of those litters born to wild-type mothers.

Newborn pups are deaf, blind and immobile, and therefore rely on their mothers to build a nest, retrieve them into it and keep them warm by crouching over them. Experiments showed that the +/- heterozygote mothers took 11 times longer to retrieve the pups and 8 times longer to build a nest than wild-type mothers, and that they never crouched over their pups during the 15 minutes of testing.

Maternal behavior is, of course, affected partly by the hormonal changes and stimuli that occur during pregnancy and giving birth. In mice, maternal behavior is also exhibited by virgin females which are presented with pups, and by females which have had litters in the past but whose pups have been weaned. The researchers found that heterozygote +/- females in these two categories also took significantly longer to show maternal behavior than comparable wild-type females. They concluded that maternal behavior induced solely by exposure to pups was also adversely affected by the mutation.

Other experiments carried out by Li and her colleagues showed that high levels of Peg3 expression were present in several areas of the brain. Several of these areas have already been shown by other researchers, by analyzing the effects of damaging them, to be important for maternal behavior.

Further observations suggested that the heterozygote +/- mothers had abnormalities in lactation. Their pups gained significantly less weight after a given period of time than the pups of wild-type mothers, despite apparently normal suckling .

Tests showed that the mammary glands of the mutant mothers appeared to be histologically normal. The team therefore turned its attention to the neurons in the brain that release oxytocin, a hormone involved in both the birth process and in stimulating milk ejection. They found that the mutant mothers had significantly fewer oxytocin-positive neurons in the hypothalamus than the wild-type animals. Nevertheless, they conclude in the Science paper: ¿This major endocrine dysfunction cannot explain the behavioral phenotype fully, because Peg3 mutant mothers conceived normally, gave birth and had normal development of mammary glands during lactation.¿

Li and colleagues describe how the Peg3 protein could be involved in DNA-binding and protein-protein interactions. It may also play a role in the tumor necrosis factor signaling pathway. Met, by contrast, is thought to be an enzyme, although its true activity is not known.

One of the most surprising features of the findings reported in this paper is how two such very different proteins apparently play similar roles in the same complex behavioral function. As Aparicio concluded, ¿We don¿t know at a very detailed molecular level whether these proteins interact in precisely the same pathway, but they certainly appear to be in the same parts of the brain and, if you mutate them, you see convergent effects on maternal behavior.¿

This is interesting, he continued, because it adds something to the debate about the evolutionary value of imprinting. ¿One hypothesis is that, if parental genomes are competing to ensure their own survival, they might have asymmetric interests,¿ Aparicio said. ¿The paternal interest might be best served by prolonged care and feeding of progeny, whereas the maternal interest might be best served by having more litters.¿

This, Li and her colleagues write in the Science paper, could explain why the paternal genome may have acquired the ability to regulate maternal behavior through imprinted genes such as Met and Peg3. Aparicio believes the study has an important message for those studying the output of the human genome sequencing efforts. ¿There is a huge gap between obtaining the sequence of a gene and knowing what it does,¿ Aparicio said. ¿These two particular genes, Met and Peg3, are nice examples where you could have done any amount of bioinformatics, but it was only by taking a genetic approach in the mouse that we were able to see the rather complex physiological behaviors that these genes are involved in. You can¿t mimic maternal behavior in a Petri dish. I believe that there is a lot of other human physiology whose molecular basis will only become understandable through this type of approach.¿

The team¿s future studies will concentrate on examining the physiological role in the brain of both genes, and the details of the molecular interactions which give rise to the neural circuits which govern maternal behavior. Aparicio said it ¿will also be interesting to look at whether polymorphisms of these genes are associated with any human conditions that might be related to human maternal behavior, for example postnatal depression or difficulty in bonding with babies. We have to be careful not to be too literal in making analogies between murine and human phenotypes, but this work provides a starting point for tackling the molecular basis of complex physiologies, such as the neural basis of behavior.¿ n