American Scientist

The Concept of the Gene in Development and Evolution: Historical and Epistemological Perspectives. Peter Beurton, Raphael Falk and Hans-Jörg Rheinberger (eds.). xvi + 384 pp. Cambridge University Press, 2000. $59.95.

Perhaps 20 years ago I slipped out of an interminable Cold Spring Harbor session and sought out Barbara McClintock. I was troubled by the recurring mantra of molecular biology that "Genes are proteins, proteins genes—that's all ye need know in this world." This recitation results in serious conceptual problems that are simply swept under the rug. Genes can manifest a position effect, changing their nature if they move on the chromosome; organisms that possess mostly the same proteins and associated regulators can vary dramatically; and the evolutionarily early members of a biological order often have many times the active DNA of their later counterparts (the C-value paradox). Most disturbingly, in this view there seems to be no mechanism specifying organism architecture: If genes simply code for proteins, they only specify cellular composition and not the complex structures encountered in Gray's Anatomy. Invoking "regulation" doesn't help much. Biochemists might feel at ease, but this former engineer was mightily bemused.

After offering tea, McClintock answered the question "What is a gene?" with "I really have no idea." Upon reflection, I realized that this was far from whimsy. As to the neo-Darwinist theories that individual mutated genes drifting through populations accomplish evolution, she was of the opinion that "they haven't read the fossil record correctly," which indeed seems the case. No simple answers to my gene questions were on offer.

Today, genes (implicitly, protein coding sequences) are media darlings. They are sold, patented, wagered on and—just maybe—promised for making improved pets with the isolated canid "loyalty gene." The long struggle, begun in the early 1900s, to relate heritable traits to underlying entities termed genes has been thoughtlessly distorted into a logical muddle. Does this make a difference? You bet it does. Visit a grant-review panel, and you'll quickly appreciate the grip of vague ideas firmly held.

So what are genes? The Concept of the Gene in Development and Evolution is a Guide for the Perplexed. The articles contributed illuminate the distance separating popular thought from the difficult and complex questions of the nature and function of "genes." Several require the technical vocabulary of philosophy, which this reviewer lacks. However, most of these essays are eminently clear and present tightly packed, closely reasoned arguments. These reward repeated, careful mining. Most of the prose is blessedly free of the usual logical nebulosities associated with this topic. Probably wisely, the essays are limited to genes as prescribed by current molecular genetics and do not attempt to address the specification and evolution of organismic form (except perhaps implicitly in Michael R. Dietrich's piece on Richard Goldschmidt—who hovers, like Banquo's ghost, over this subject).

A review can only suggest the breadth and depth of the essays. An appreciation of how we arrived at our current views could begin with Jean Gayon's history of heredity as a philosophical and scientific subject. Here are the ideas of the early major players (Sir Francis Galton, Darwin, et al.) leading up to the chromosomal theory of inheritance "consecrated" by Thomas Hunt Morgan in 1915.

Sara Schwartz's account of "The Differential Concept of the Gene: Past and Present" plots the path to our present confused state. Pioneers such as Morgan clearly saw the multivariate relation between genes and traits and struggled to reconcile it with a research method that assumed a one-to-one correspondence between gene and traits. He finally accepted differences at the genotypic and phenotypic level as respective cause and effect. The complexities of traits were eventually forgotten but have reemerged with the surprises from gene knockout experiments. The real charge down the primrose path began with George Wells Beadle and Edward Lawrie Tatum, who, working with a simple organism, found that every enzyme appeared to have a corresponding gene. Unfortunately, this discovery was turned around to imply that every gene made a protein—a profoundly different and logically insupportable proposition. The hypothesis garnered no overt criticism, reflecting the wish to believe that a simple entity underlay the classical gene. The resulting fabulation is now firmly rooted in the laboratory, literature and classroom.

The "genes are proteins" doctrine was sharpened by Seymour Benzer as described in Frederic L. Holmes's excellent brief review. Exhaustively mapping the mutations in a small region of the T4 bacteriophage genome, he connected a historical gene to the DNA molecule. This precision and elegance, rare in biological science, mesmerized subsequent thought: Here indeed must be the gene. Well, yes, sort of—for the special case of viruses that mostly self-assemble and whose architectural information is therefore encoded within each component. Here genes just make proteins that go together only one correct way. In higher organisms, structural proteins behave more like Lego blocks that join in multitudinous ways—hence varied organisms with essentially the same "gene products." The need for the present volume results, in part, from the procrustean fitting of the metazoan genome to a conceptual bed derived from viruses and bacteria.

Language does matter, especially in the intersection between evolution and development. Scott F. Gilbert offers a brief but detailed and trenchant discussion of the very different meanings of gene in population genetics compared with its use in embryology. For example, in population genetics genes are differences (possibly resulting from mutations) affecting the fitness of an individual, whereas in development, genes direct the formation of the embryo. Both Gilbert and Michel Morange consider the strictly developmental genes (hox genes, myoD, etc.) to be fundamentally different from classical genes. They are indeed master switches, triggering vast developmental consequences, and their significance is only now unfolding.

Evelyn Fox Keller examines terms used in the popular but ill-defined concept of a "genetic program of development." The term genetic applied to developmental instructions needlessly emphasizes DNA over all else. What and where is the program and what is "genetic"? With terminology cribbed from 1960s computerese, DNA was suggested as the source of both data and instructions. Actually, the complex extrachromosomal machinery required for DNA to do anything in the egg or later suggests massively parallel computation if a metaphor is needed. Why invoke the computer terms program and network? Keller traces one origin back to François Jacob and Jacques Monod, who saw in the regulatory "operon" of β-galactosidase in Escherichia coli a network switch for building the embryonic differentiation program. She quotes Sydney Brenner on the operon: "[It] does not tell us how to make a mouse but only how to make a switch."

Dietrich reintroduces the gene theories of Richard Goldschmidt, summarizing a literature that's not readily available. A major figure in genetics during the 1930s and 1940s, Goldschmidt was repeatedly excoriated for the unforgivable sin of rejecting the "corpuscular" gene. Rather, he preferred the position effect with chromosome rearrangements to explain mutations; this was correct sometimes but mostly not. Coupled with transgressions in other scientific matters, this elicited an intolerance endemic in the softer sciences. Deitrich makes clear that there is far more to Goldschmidt than gene theory and that he seems to have been ahead of his time. His approach was holistic—certainly not the current fashion. His intriguing take on evolution proposed that the apparent saltation of the fossil record actually reflected saltatory events. "Hopeful monsters," bred of catastrophic genome duplications and rearrangements, were suggested by the dramatic effects of polyploidy on organism form. Although this can neatly account for the C-value paradox, it brought out the neo-Darwinist dogs of war.

Scientists, in general, do not welcome philosophy of science (Richard Feynman is supposed to have described it as being as useful to scientists as ornithology is to birds). It will be lamentable if such proves the case here. There are important public policy issues as well as broad scientific directives that critically depend on the definitions used in genetics. The present volume offers a unique guide to the meanings of the term gene.