IN THE GENES CAN A SINGLE GENE INFLUENCE BEHAVIOR?

 When behavior is hard-wired rather than learned, does it mean it's 100 percent genetic? How do genes create behavior? Can a single gene influence behavior? How can we be sure it's not environment that's causing changes in behavior that are deemed genetic?

These are fascinating questions. First, it's important to understand that genes can't 'make' behavior, especially complex behavior, any more than a single engine part 'makes' a car. Having said that, a malfunction of a gene can derail behavior, just as a faulty engine part can cause a car to break down. Behavior is a phenotypic trait - i.e., it is an observable characteristic of an animal. A phenotypic trait - including "hard-wired" or unlearned behavior - is always the end product of an interaction between the animal's genes and the animal's environment. An animal's genotype is the collection of all its genes, whether or not they are expressed ("visible") in the phenotype.

How genes do their part starts with the exact spot on the chromosome, called a locus, where the gene resides. Two alleles - versions of each gene donated from each parent - are at every locus. The relationship between the two alleles may be additive, meaning the phenotype is intermediate between the two. As a hypothetical example, imagine breeding a dog with a high drive for toys to a dog with a low drive and the offspring being moderately drivey for toys.

The relationship between alleles can also be dominant-recessive to varying degrees. Where one allele is expressed in the phenotype, the other - though present and transmissible to the offspring - is nearly or completely silent. This is the Mendelian inheritance taught in many biology classes, with eye colour in humans the classic example.

To answer the question about whether a single gene can influence behavior, the answer is yes with the operative word being "influence." Complex traits, such as behavior, are better understood as the product of the actions and interactions of many, many genes. Such traits are called "polygenic," meaning that each individual gene makes up only a small part of the genetic contribution to the trait. Body size is a well-known polygenic trait. Polygenic traits can be additive, as in body size. They can also be epistatic, where the product is not intermediate between the genes involved, but rather is more analogous to the dominant-recessive relationship that exists at individual loci. Coat-colour heredity in many breeds of dog is a familiar example of a phenotypic trait governed by epistasis.

Remember, it's far from over once a set of genes is turned on or off. Genes interact with the environment to result in the final phenotypic product, the body part or behavior that we can see. Even body size, though largely polygenically determined, is somewhat influenced by environmental factors like nutrition, illness, etc. In the case of behavior, environmental interactions often playa larger role. For example, famous research on mice bred for high aptitude at running mazes (maze-bright mice) and mice bred for low aptitude (maze-dull mice) also found that rearing the maze-dull mice in a more complex and stimulating environment compensated for their maze-dull genes and they performed as well on mazes as maze-bright mice.

Environment not only interacts with genes but can directly affect their activity. It can literally turn them on and off! In fact, about 95 per cent of genes that code at all do not code for proteins on the road to phenotypic traits, but rather regulate the action of other genes. Most of these regulatory genes respond to environmental triggers. This prompts them to turn the protein-coding genes on and off, both in utero when the organism is being built, during post-natal development and right on into adulthood. h1 other words, genes don't provide recipes in stone on how to build bodies and behave to organisms. More accurately; they participate in a give-and-take process with environmental signals throughout life.

Regulator genes may send instructions to genes that are themselves regulator genes. Eventually; further downstream, genes that will build proteins, specifically enzymes, are turned on. Enzymes then affect cellular metabolism, which in turn affects other cells. At the end of this chain reaction comes the tiny chemical increments that make up the genetic part of the genetics-environment cocktail that is behavior.

An interesting result of this string of events, all kicked off by a particular regulator gene, is that a single gene can have two or more effects. This is called pleiotropy and is quite different from polygenic effects, where multiple genes contribute to one effect. A famous example of pleiotropy occurred during the Russian fox domestication experiment (see the August 2003 issue) where strong selection for reduced flight distance also produced floppy ears, curled tails, shorter muzzles and occasionally even white markings on the coat.

There are different flavours of pleiotropy: For instance, one regulator gene can kick off chain reactions in multiple metabolic pathways in different body systems. Pleiotropy can also occur if a gene sits next door to another unrelated gene on a chromosome and then both stick together when cut and pasted into the new generation.

So how do we know it's genes? The way genes influence behavior can be teased out of the gene-environment interaction in a variety of ways. Even before the advent of molecular genetics, animals with identical or near-identical genotypes were studied in different environments and animals with different genotypes were studied in identical or near-identical environments. These measures did a reasonable job of controlling for environmental effects. The serendipitous (for cience) 'experiment' of comparing human identical twins separated at birth yielded revealing data about the effect of genes on behavior.

Direct study of the presence or absence of specific genes and how this impacts behavior is now possible, too. Because it is so unlikely that single genes that have profound behavioral effects will be found, researchers concentrate their search for sets of genes that each make a small contribution to a certain behavior. Also, most behavior traits are distributed in a continuous rather than 'digital' fashion, which strongly implies multiple genes. For example, there isn't an on-off switch for anxiety, with some dogs completely lacking a predisposition and others completely riddled.

Research that attempts to link the presence of gene suites with elevated occurrence of certain traits in families can be confounded by partial dominance effects and the fact that the same traits may be brought about by different genetic routes. Scans of full genomes can narrow the search if there is a marker residing in close proximity to a candidate gene that is suspected to be associated with a certain trait. The spot on the genome of a control group without the trait is then compared to the same spot on the genome of the group that has the trait.

Experiments can really ice the cake when it comes to nailing down which genes do what. For example, once a suspect gene has been identified, it can be 'knocked out' of the genome of mice and replaced with genomes that are otherwise identical to a control group of mice and then the two groups compared in minute detail. Transgenics, where genes from another animal, such as a human, are spliced into an experimental animal is also a fruitful way to directly observe the effects of genes on behavior, body type and body function. When all these methods converge on the same conclusion, the evidence for the effects of certain genes is extremely compelling.

By Canadian Jean Donaldson is the founder of the San Francisco SPCA Academy for Dog Trainers.