Back to mutations. Can there be mutations in DNA stretches constituting promoters? Yes, and more often than in genes themselves. In the 1970s Allan Wilson and Mary-Claire King at Berkeley correctly theorized that the evolution of genes is less important than the evolution of regulatory sequences upstream of genes (and thus how the environment regulates genes). Reflecting that, a disproportionate share of genetic differences between chimps and humans are in genes for TFs.
Time for more complexity. Suppose you have genes 1–10, and transcription factors A, B, and C. TF-A induces the transcription of genes 1, 3, 5, 7, and 9. TF-B induces genes 1, 2, 5, and 6. TF-C induces 1, 5, and 10. Thus, upstream of gene 1 are separate promoters responding to TFs A, B, and C—thus genes can be regulated by multiple TFs. Conversely, each TF usually activates more than one gene, meaning that multiple genes are typically activated in networks (for example, cell injury causes a TF called NF-κB to activate a network of inflammation genes). Suppose the promoter upstream of gene 3 that responds to promoter TF-A has a mutation making it responsive to TF-B. Result? Gene 3 is now activated as part of a different network. Same networkwide outcome if there is a mutation in a gene for a TF, producing a protein that binds to a different promoter type.4
Consider this: the human genome codes for about 1,500 different TFs, contains 4,000,000 TF-binding sites, and the average cell uses about 200,000 such sites to generate its distinctive gene-expression profile.5 This is boggling.
Epigenetics
The last chapter introduced the phenomenon of environmental influences freezing genetic on/off in one position. Such “epigenetic” changes* were relevant to events, particularly in childhood, causing persistent effects on the brain and behavior. For example, recall those pair-bonding prairie voles; when females and males first mate, there are epigenetic changes in regulation of oxytocin and vasopressin receptor genes in the nucleus accumbens, that target of mesolimbic dopamine projection.6
Let’s translate the last chapter’s imagery of “freezing on/off switches” into molecular biology.7 What mechanisms underlie epigenetic changes in gene regulation? An environmental input results in a chemical being attached tightly to a promoter, or to some nearby structural proteins surrounding DNA. The result of either is that TFs can no longer access or properly bind to the promoter, thus silencing the gene.
As emphasized in the last chapter, epigenetic changes can be multigenerational.8 Dogma was that all the epigenetic marks (i.e., changes in the DNA or surrounding proteins) were erased in eggs and sperm. But it turns out that epigenetic marks can be passed on by both (e.g., make male mice diabetic, and they pass the trait to their offspring via epigenetic changes in sperm).
Recall one of the great punching bags of science history, the eighteenth-century French biologist Jean-Baptiste Lamarck.9 All anybody knows about the guy now is that he was wrong about heredity. Suppose a giraffe habitually stretches her neck to reach leaves high in a tree; this lengthens her neck. According to Lamarck, when she has babies, they will have longer necks because of “acquired inheritance.”* Lunatic! Buffoon! Epigenetically mediated mechanisms of inheritance—now often called “neo-Lamarckian inheritance”—prove Lamarck right in this narrow domain. Centuries late, the guy’s getting some acclaim.
Thus, not only does environment regulate genes, but it can do so with effects that last days to lifetimes.
The Modular Construction of Genes: Exons and Introns
Time to do in another dogma about DNA. It turns out that most genes are not coded for by a continuous stretch of DNA. Instead there might be a stretch of noncoding DNA in the middle. In that case, the two separate stretches of coding DNA are called “exons,” separated by an “intron.” Many genes are broken into numerous exons (with, logically, one less intron than the number of exons).
How do you produce a protein from an “exonic” gene? The RNA photocopy of the gene initially contains the exons and introns; an enzyme removes the intronic parts and splices together the exons. Clunky, but with big implications.
Back to each particular gene coding for a particular protein.10 Introns and exons destroy this simplicity. Imagine a gene consisting of exons 1, 2, and 3, separated by introns A and B. In one part of the body a splicing enzyme exists that splices out the introns and also trashes exon 3, producing a protein coded for by exons 1 and 2. Meanwhile, elsewhere in the body, a different splicing enzyme jettisons exon 2 along with the introns, producing a protein derived from exons 1 and 3. In another cell type a protein is made solely from exon 1. . . . Thus “alternative splicing” can generate multiple unique proteins from a single stretch of DNA; so much for “one gene specifies one protein”—this gene specifies seven (A, B, C, A-B, A-C, B-C, and A-B-C). Remarkably, 90 percent of human genes with exons are alternatively spliced. Moreover, when a gene is regulated by multiple TFs, each can direct the transcription of a different combination of exons. Oh, and splicing enzymes are proteins, meaning that each is coded for by a gene. Loops and loops.
Transposable Genetic Elements, the Stability of the Genome, and Neurogenesis
Time to unmoor another cherished idea, namely that genes inherited from your parents (i.e., what you started with as a fertilized egg) are immutable. This calls up a great chapter of science history. In the 1940s an accomplished plant geneticist named Barbara McClintock observed something impossible. She was studying the inheritance of kernel color in maize (a frequent tool of geneticists) and found patterns of mutations unexplained by any known mechanism. The only possibility, she concluded, was that stretches of DNA had been copied, with the copy then randomly inserted into another stretch of DNA.
Yeah, right.
Clearly McClintock, with her (derisively named) “jumping genes,” had gone mad, and so she was ignored (not exactly true, but this detracts from the drama). She soldiered on in epic isolation. And finally, with the molecular revolution of the 1970s, she was vindicated about her (now termed) transposable genetic elements, or transposons. She was lionized, canonized, Nobel Prized (and was wonderfully inspirational, as disinterested in acclaim as in her ostracism, working until her nineties).
Transpositional events rarely produce great outcomes. Consider a hypothetical stretch of DNA coding for “The fertilized egg is implanted in the uterus.”
There has been a transpositional event, where the underlined stretch of message was copied and randomly plunked down elsewhere: “The fertilized eggterus is implanted in the uterus.”
Gibberish.
But sometimes “The fertilized egg is implanted in the uterus” becomes “The fertilized eggplant is implanted in the uterus.”
Now, that’s not an everyday occurrence.
—
Plants utilize transposons. Suppose there is a drought; plants can’t move to wetter pastures like animals can. Plant “stress” such as drought induces transpositions in particular cells, where the plant metaphorically shuffles its DNA deck, hoping to generate some novel savior of a protein.
Mammals have fewer transposons than plants. The immune system is one transposon hot spot, in the enormous stretches of DNA coding for antibodies. A novel virus invades; shuffling the DNA increases the odds of coming up with an antibody that will target the invader.*
The main point here is that transposons occur in the brain.11 In humans transpositional events occur in stem cells in the brain when they are becoming neurons, making the brain a mosaic of neurons with different DNA sequences. In other words, when you make neurons, that boring DNA sequence you inherited isn’t good enough. Remarkably, transpositional events occur in neurons that form memories in fruit flies. Even flies evolved such that their neurons are freed from the strict genetic marching orders they inherit.
Chance
Finally, chance lessens genes as the Code of Codes. Chance, driven by Brownian motion—the random movement of particles in a fluid—has big effects on tiny things like molecules floating in cells, including molecule
s regulating gene transcription.12 This influences how quickly an activated TF reaches the DNA, splicing enzymes bump into target stretches of RNA, and an enzyme synthesizing something grabs the two precursor molecules needed for the synthesis. I’ll stop here; otherwise, I’ll go on for hours.
Some Key Points, Completing This Part of the Chapter
Genes are not autonomous agents commanding biological events.
Instead, genes are regulated by the environment, with “environment” consisting of everything from events inside the cell to the universe.
Much of your DNA turns environmental influences into gene transcription, rather than coding for genes themselves; moreover, evolution is heavily about changing regulation of gene transcription, rather than genes themselves.
Epigenetics can allow environmental effects to be lifelong, or even multigenerational.
And thanks to transposons, neurons contain a mosaic of different genomes.
In other words, genes don’t determine much. This theme continues as we focus on the effects of genes on behavior.
PART 2: GENES FROM THE TOP DOWN—BEHAVIOR GENETICS
Long before anything was known about promoters, exons, or transcription factors, it became clear that you study genetics top down, by observing traits shared by relatives. Early in the last century, this emerged as the science of “behavior genetics.” As we’ll see, the field has often been mired in controversy, typically because of disagreements over the magnitude of genetic effects on things like IQ or sexual orientation.
First Attempts
The field began with the primitive idea that, if everyone in a family does it, it must be genetic. This was confounded by environment running in families as well.
The next approach depended on closer relatives having more genes in common than distant ones. Thus, if a trait runs in a family and is more common among closer relatives, it’s genetic. But obviously, closer relatives share more environment as well—think of a child and parent versus a child and grandparent.
Research grew subtler. Consider someone’s biological aunt (i.e., the sister of a parent), and the uncle who married the aunt. The uncle shares some degree of environment with the individual, while the aunt shares the same, plus genes. Therefore, the extent to which the aunt is more similar to the individual than the uncle reflects the genetic influence. But as we’ll see, this approach has problems.
More sophistication was needed.
Twins, Adoptees, and Adopted Twins
A major advance came with “twin studies.” Initially, examining twins helped rule out the possibility of genetic determination of a behavior. Consider pairs of identical twins, sharing 100 percent of their genes. Suppose one of each pair has schizophrenia; does the twin as well? If there are any cases where the other twin doesn’t (i.e., where the “concordance rate” is less than 100 percent), you’ve shown that the genome and epigenetic profile inherited at birth do not solely determine the incidence of schizophrenia (in fact the concordance rate is about 50 percent).
But then a more elegant twin approach emerged, involving the key distinction between identical (monozygotic, or MZ) twins, who share 100 percent of their genes, and fraternal, nonidentical (dizygotic, or DZ) twins, who, like all other sibling pairs, share 50 percent of their genes. Compare pairs of MZ twins with same-sex DZ twins. Each pair is the same age, was raised in the same environment, and shared a fetal environment; the only difference is the percentage of genes shared. Examine a trait occurring in one member of the twin pair; is it there in the other? The logic ran that, if a trait is shared more among MZ than among DZ twins, that increased degree of sharing reflects the genetic contribution to the trait.
Another major advance came in the 1960s. Identify individuals adopted soon after birth. All they share with their biological parents is genes; all they share with their adoptive parents is environment. Thus, if adopted individuals share a trait more with their biological than with their adoptive parents, you’ve uncovered a genetic influence. This replicates a classic tool in animal studies, namely “cross-fostering”—switching newborn rat pups between two mothers. The approach was pioneered in revealing a strong genetic component to schizophrenia.13
Then came the most wonderful, amazing, like, totally awesome thing ever in behavior genetics, started by Thomas Bouchard of the University of Minnesota. In 1979 Bouchard found a pair of identical twins who were—get this—separated at birth and adopted into different homes, with no knowledge of each other’s existence until being reunited as adults.14 Identical twins separated at birth are so spectacular and rare that behavior geneticists swoon over them, want to collect them all. Bouchard eventually studied more than a hundred such pairs.
The attraction was obvious—same genes, different environments (and the more different the better); thus, similarities in behavior probably reflect genetic influences. Here’s an imaginary twin pair that would be God’s gift to behavior geneticists—identical twin boys separated at birth. One, Shmuel, is raised as an Orthodox Jew in the Amazon; the other, Wolfie, is raised as a Nazi in the Sahara. Reunite them as adults and see if they do similar quirky things like, say, flushing the toilet before using it. Flabbergastingly, one twin pair came close to that. They were born in 1933 in Trinidad to a German Catholic mother and a Jewish father; when the boys were six months of age, the parents separated; the mother returned to Germany with one son, and the other remained in Trinidad with the father. The latter was raised there and in Israel as Jack Yufe, an observant Jew whose first language was Yiddish. The other, Oskar Stohr, was raised in Germany as a Hitler Youth zealot. Reunited and studied by Bouchard, they warily got to know each other, discovering numerous shared behavioral and personality traits including . . . flushing the toilet before use. (As we’ll see, studies were more systematic than just documenting bathroom quirks. The flushing detail, however, always comes up in accounts of the pair.)
Behavior geneticists, wielding adoption and twin approaches, generated scads of studies, filling specialized journals like Genes, Brain and Behavior and Twin Research and Human Genetics. Collectively, the research consistently showed that genetics plays a major role in a gamut of domains of behavior, including IQ and its subcomponents (i.e., verbal ability, and spatial ability),*15 schizophrenia, depression, bipolar disorder, autism, attention-deficit disorder, compulsive gambling, and alcoholism.
Nearly as strong genetic influences were shown for personality measures related to extroversion, agreeableness, conscientiousness, neuroticism, and openness to experience (known as the “Big Five” personality traits).16 Likewise with genetic influences on degree of religiosity, attitude toward authority, attitude toward homosexuality,* and propensities toward cooperation and risk taking in games.
Other twin studies showed genetic influences on the likelihood of risky sexual behavior and on people’s degree of attraction to secondary sexual characteristics (e.g., musculature in men, breast size in women).17
Meanwhile, some social scientists report genetic influences on the extent of political involvement and sophistication (independent of political orientation); there are behavior genetics papers in the American Journal of Political Science.18
Genes, genes, everywhere. Large genetic contributions have even been uncovered for everything from the frequency with which teenagers text to the occurrence of dental phobias.19
So does this mean there is a gene “for” finding chest hair on guys to be hot, for likelihood of voting, for feelings about dentists? Vanishingly unlikely. Instead, gene and behavior are often connected by tortuous routes.20 Consider the genetic influence on voter participation; the mediating factor between the two turns out to be sense of control and efficacy. People who vote regularly feel that their actions matter, and this central locus of control reflects some genetically influenced personality traits (e.g., high optimism, low neuroticism). Or how about the link between genes and self-confidence? Some studies show that the int
ervening variable is genetic effects on height; taller people are considered more attractive and treated better, boosting their self-confidence, dammit.*
In other words, genetic influences on behavior often work through very indirect routes, something rarely emphasized when news broadcasts toss out behavior genetics sound bites—“Scientists report genetic influence on strategy when playing Candyland.”
The Debates About Twin and Adoption Studies
Many scientists have heavily criticized the assumptions in twin and adoption studies, showing that they generally lead to overestimates of the importance of genes.* Most behavior geneticists recognize these problems but argue that the overestimates are tiny.21 A summary of this technical but important debate:
Criticism #1: Twin studies are premised on MZ and same-sex DZ twin pairs sharing environment equally (while sharing genes to very different extents). This “equal environment assumption” (EEA) is simply wrong; starting with parents, MZ twins are treated more similarly than DZ twins, creating more similar environments for them. If this isn’t recognized, greater similarity between MZs will be misattributed to genes.22
Scientists such as Kenneth Kendler of Virginia Commonwealth University, a dean of the field, have tried to control for this by (a) quantifying just how similar childhoods were for twins (with respect to variables like sharing rooms, clothing, friends, teachers, and adversity); (b) examining cases of “mistaken zygosity,” where parents were wrong about their twins’ MZ/DZ status (thus, for example, raising their DZ twins as if they were MZ); and (c) comparing full-, half-, and step-siblings who were reared together for differing lengths of time. Most of these studies show that controlling for the assumption of MZs sharing more environment than do DZs doesn’t significantly reduce the size of genetic influences.*23 Hold that thought.
Behave: The Biology of Humans at Our Best and Worst Page 24