More Articles on Evolution
A Third Way
James A. Shapiro
The recent reviews in your columns of books by Dennett, Dawkins, and Behe
are testimony to the unflagging interest in controversies about evolution.
Although such purists as Dennett and Dawkins repeatedly assert that the scientific
issues surrounding evolution are basically solved by conventional neo-Darwinism,
the ongoing public fascination reveals a deeper wisdom. There are far more
unresolved questions than answers about evolutionary processes, and contemporary
science continues to provide us with new conceptual possibilities.
Unfortunately, readers of Boston Review may remain unaware of this
intellectual ferment because the debate about evolution continues to assume
the quality of an abstract and philosophical "dialogue of the deaf" between
Creationists and Darwinists. Although our knowledge of the molecular details
of biological organization is undergoing a revolutionary expansion, open-minded
discussions of the impact of these discoveries are all too rare. The possibility
of a non-Darwinian, scientific theory of evolution is virtually never considered.
In my comments, then, I propose to sketch some developments in contemporary
life science that suggest shortcomings in orthodox evolutionary theory and
open the door to very different ways of formulating questions about the evolutionary
process. After a discussion of technical advances in our views about genome
organization and the mechanisms of genetic change, I will focus on a growing
convergence between biology and information science which offers the potential
for scientific investigation of possible intelligent cellular action in evolution.
The past five decades of research in genetics and molecular biology have
brought us revolutionary discoveries. Upsetting the oversimplified views of
cellular organization and function held at mid-century, the molecular revolution
has revealed an unanticipated realm of complexity and interaction more consistent
with computer technology than with the mechanical viewpoint which dominated
the field when the neo-Darwinian Modern Synthesis was formulated. The conceptual
changes in biology are comparable in magnitude to the transition from classical
physics to relativistic and quantum physics.
Four categories of molecular discoveries are especially important in opening
up exciting new ways of thinking about the biological processes that underlie
(1) Genome Organization. Our current ideas of genome organization
are completely different from the "beads on a string" view that dominated
genetics in the 1940s and 1950s. At that time genes were "units" which corresponded
to individual organismal traits, and the "one gene-one enzyme" hypothesis
told us that the essential business of each gene was to encode a specific
protein molecule linked to a particular phenotype. We have now deconstructed
each genetic locus into a modular assembly of regulatory and coding motifs.
Most of these motifs are shared among many loci, suggesting that genomes are
assembled Lego-like from a repertoire of more basic sequence elements, many
of which do not encode proteins but determine other important functions (transcription,
translation, RNA processing, DNA replication, chromatin condensation, etc.).
As we analyze genome expression during cellular proliferation and multicellular
development, we have learned that diverse genetic loci are organized hierarchically
into interconnected genome-wide networks which function dynamically. Not confined
to a single pathway, many genetic loci are active at different times, participating
in the expression of more than one phenotypic trait. Comparisons of genomes
in different organisms have revealed unexpected patterns of evolutionary conservation
across large taxonomic distances, while closely-related genomes frequently
differ significantly in the arrangement of repetitive DNA elements which do
not encode proteins.
How all of this modularity, complexity, and integration arose and changed
during the history of life on earth is a central evolutionary question. Localized
random mutation, selection operating "one gene at a time" (John Maynard Smith's
formulation), and gradual modification of individual functions are unable
to provide satisfactory explanations for the molecular data, no matter how
much time for change is assumed. There are simply too many potential degrees
of freedom for random variability and too many interconnections to account
Studies of the molecular sources of genetic variability have taught us two
major lessons about how cells take care of their genomes--one about self-protection,
the other about self-reorganization.
(2) Cellular Repair Capabilities. First, then, all cells from bacteria
to man possess a truly astonishing array of repair systems which serve to
remove accidental and stochastic sources of mutation. Multiple levels of proofreading
mechanisms recognize and remove errors that inevitably occur during DNA replication.
These proofreading systems are capable of distinguishing between newly synthesized
and parental strands of the DNA double helix, so they operate efficiently
to rectify rather than fix the results of accidental misincorporations of
the wrong nucleotide. Other systems scan non-replicating DNA for chemical
changes that could lead to miscoding and remove modified nucleotides, while
additional functions monitor the pools of precursors and remove potentially
mutagenic contaminants. In anticipation of chemical and physical insults to
the genome, such as alkylating agents and ultraviolet radiation, additional
repair systems are encoded in the genome and can be induced to correct damage
when it occurs.
It has been a surprise to learn how thoroughly cells protect themselves against
precisely the kinds of accidental genetic change that, according to conventional
theory, are the sources of evolutionary variability. By virtue of their proofreading
and repair systems, living cells are not passive victims of the random forces
of chemistry and physics. They devote large resources to suppressing random
genetic variation and have the capacity to set the level of background localized
mutability by adjusting the activity of their repair systems.
(3) Mobile Genetic Elements and Natural Genetic Engineering. The second
major lesson of molecular studies into the origins of genetic change is that
all cells possess multiple biochemical agents for natural genetic engineering--processes
that include the cutting and splicing of DNA molecules into new sequence arrangements.
Most frequently, natural genetic engineering capabilities reveal themselves
through the activities of mobile genetic elements--DNA structures found in
all genomes that can move from one position to another. Mobile genetic elements
are the most fluid components of the genome and also the most taxonomically
specific. In human cells, mobile elements include retrotransposons, like the
half-million or more Alu sequences dispersed over all our chromosomes, as
well as the inherited gene fragments which our lymphocytes assemble daily
to form active genetic loci encoding the key antigen recognition molecules
of our immune system. The biochemical agents of DNA restructuring include
the enzymes used in our own genetic engineering for research and biotechnology
(nucleases, ligases, reverse transcriptases and polymerases) as well as other
proteins that combine to form molecular machines capable of mobilizing different
The existence of cellular biochemical activities capable of rearranging DNA
molecules means that genetic change can be specific (these activities can
recognize particular sequence motifs) and need not be limited to one genetic
locus (the same activity can operate at multiple sites in the genome). In
other words, genetic change can be massive and non-random. Some organisms,
such as the ciliated protozooan Oxytricha, completely reorganize their genetic
apparatus within a single cell generation, fragmenting the germ-line chromosomes
into thousands of pieces and then reassembling a particular subset of them
into a distinct kind of functional genome. Furthermore, natural genetic engineering
systems can operate premeiotically during the somatic development of tissues
that will ultimately produce gametes. This means that major chromosome reorganizations
can be present in multiple gametes. Consequently, the appearance of new genome
architectures during evolution is not necessarily limited to isolated individuals.
The discovery that genome reorganization is largely a biological process
traces back to Barbara McClintock's pioneering studies of mutation and chromosome
rearrangement in maize from the 1940s through the 1960s. She linked these
genetic events to changes in the regulation of gene expression programs during
plant development. We can now appreciate her tremendous wisdom and foresight
by seeing how the Lego-like patterns of integrated genome organization mentioned
above could be created by the activity of cellular natural genetic engineering
systems. Because, like all cellular functions, natural genetic engineering
systems are subject to control circuits, they can be held in abeyance for
long periods and then called into action at certain key times. Sometimes these
activations can be regularly programmed, as in the development of our immune
systems, and sometimes activations can occur in response to crisis, as McClintock
documented in maize.
The point of this discussion is that our current knowledge of genetic change
is fundamentally at variance with neo-Darwinist postulates. We have progressed
from the Constant Genome, subject only to random, localized changes at a more
or less constant mutation rate, to the Fluid Genome, subject to episodic,
massive and non-random reorganizations capable of producing new functional
architectures. Inevitably, such a profound advance in awareness of genetic
capabilities will dramatically alter our understanding of the evolutionary
process. Nonetheless, neo-Darwinist writers like Dawkins continue to ignore
or trivialize the new knowledge and insist on gradualism as the only path
for evolutionary change.
(4) Cellular Information Processing. While it is easy to see how advances
in our understanding of genome organization and genetic change will impact
theories of evolutionary processes, another development in contemporary biology
is of less obvious but even more basic relevance. This is the growing realization
that cells have molecular computing networks which process information about
internal operations and about the external environment to make decisions controlling
growth, movement, and differentiation. This realization has come, in large
measure, from detailed genetic analysis of cellular processes and multicellular
development. The inducible repair systems mentioned above provide a relatively
simple, well-studied example. Bacterial and yeast cells have molecules that
monitor the status of the genome and activate cellular responses when damaged
DNA accumulates. The surveillance molecules do this by modifying transcription
factors so that appropriate repair functions are synthesized. These inducible
DNA damage response systems are sophisticated and include so-called "checkpoint"
functions that act to arrest cell division until the repair process has been
completed. When the checkpoints do not function, cell division proceeds before
repair is completed, and the damaged cells die or produce inviable progeny.
One can characterize this surveillance/inducible repair/checkpoint system
as a molecular computation network demonstrating biologically useful properties
of self-awareness and decision-making.
There are many other cellular systems that display comparable information-processing
capabilities. Fro example, it is now common among molecular biologists who
study the cell cycle to speak of various checkpoints (Is DNA replication complete?
Are the chromosomes properly condensed and aligned on the metaphase plate?)
and decision points (e.g., when to initiate chromosome movement and cytokinesis).
A recent special issue of Scientific American1 describes
beautifully how cancer is now seen as a disease of the molecular information
processing routines that ensure orderly cell growth and behavior in the healthy
organism. Aberrant tumor cell growth appears to result from at least two kinds
of malfunction: the loss of checkpoint controls, or the failure of decision-making
routines that dictate programmed cell death (apoptosis) for cells in inappropriate
surroundings. During embryonic development, cells make decisions about differentiation
based on multiple molecular signals picked up from their environment and from
their neighbors by means of surface receptors. These receptors are linked
to intercellular molecular cascades called "signal transduction pathways"
which integrate the inputs from the receptors to generate appropriate patterns
of differential gene expression and morphogenesis of specialized cell structures.
Signal transduction is not limited to multicellular development. We are learning
that virtually every aspect of cellular function is influenced by chemical
messages detected, transmitted, and interpreted by molecular relays. To a
remarkable extent, therefore, contemporary biology has become a science of
sensitivity, inter- and intra-cellular communication, and control. Given the
enormous complexity of living cells and the need to coordinate literally millions
of biochemical events, it would be surprising if powerful cellular capacities
for information processing did not manifest themselves. In an important way,
then, biology has returned to questions debated during the mechanism-vitalism
controversy earlier this century. This time around, however, the discussion
is informed by two new factors. One is that the techniques of molecular and
cell biology allow us to examine the detailed operation of the hardware responsible
for cellular responsiveness and decision-making. The second is the existence
of computers and information networks, physical entities endowed with computational
and decision-making capabilities. Their existence means that discussing the
potential for similar activities by living organisms is neither vague nor
What significance does an emerging interface between biology and information
science hold for thinking about evolution? It opens up the possibility of
addressing scientifically rather than ideologically the central issue so hotly
contested by fundamentalists on both sides of the Creationist-Darwinist debate:
Is there any guiding intelligence at work in the origin of species displaying
exquisite adaptations that range from lambda prophage repression and the Krebs
cycle through the mitotic apparatus and the eye to the immune system, mimicry,
and social organization? Borrowing concepts from information science, new
schools of evolutionists can begin to rephrase virtually intractable global
questions in terms amenable to computer modelling and experimentation. We
can speculate what some of these more manageable questions might be: How can
molecular control circuits be combined to direct the expression of novel traits?
Do genomes display characteristic system architectures that allow us to predict
phenotypic consequences when we rearrange DNA sequence components? Do signal
transduction networks contribute functional information as they regulate the
action of natural genetic engineering hardware?
Questions like those above will certainly prove to be naive because we are
just on the threshold of a new way of thinking about living organisms and
their variations. Nonetheless, these questions serve to illustrate the potential
for addressing the deep issues of evolution from a radically different scientific
perspective. Novel ways of looking at longstanding problems have historically
been the chief motors of scientific progress. However, the potential for new
science is hard to find in the Creationist-Darwinist debate. Both sides appear
to have a common interest in presenting a static view of the scientific enterprise.
This is to be expected from the Creationists, who naturally refuse to recognize
science's remarkable record of making more and more seemingly miraculous aspects
of our world comprehensible to our understanding and accessible to our technology.
But the neo-Darwinian advocates claim to be scientists, and we can legitimately
expect of them a more open spirit of inquiry. Instead, they assume a defensive
posture of outraged orthodoxy and assert an unassailable claim to truth, which
only serves to validate the Creationists' criticism that Darwinism has become
more of a faith than a science.
A sounder perspective on the history of science would be very helpful to
all concerned. For example, a parallel has been drawn by Allen Orr and others
between criticisms of Darwinian orthodoxy and assaults on the Law of Gravity,
presenting them as equally deplorable examples of anti-science obscurantism.
Yet, if truth be told, gravity is far from a settled matter. The relativistic
Law of Gravity at the end of the 20th century is not the same as the classical
Law of Gravity at the end of the 19th century, and discovering how the continuous
descriptions of general relativity can be integrated into a single theory
with the discrete accounts of quantum physics is still an active field of
research. From a scientific point of view, then, the Law of Gravity has quite
properly been under continuous challenge. Dogmas and taboos may be suitable
for religion, but they have no place in science. No theory or viewpoint should
ever become sacrosanct because experience tells us that even the most elegant
Laws of Nature ultimately succumb to the inexorable progress of scientific
thinking and technological innovation. The present debate over Darwinism will
be more productive if it takes place in recognition of the fact that scientific
advances are made not by canonizing our predecessors but by creating intellectual
and technical opportunities for our successors.
1 Robert Weinberg, "How Cancer Arises," Scientific American
275, no. 3 (September 1996), pp. 62-70.