Lynn Margulis was
an indefatigable advocate of positive cell action in the
evolutionary process. Lynn focused her work on observing
real-time interactions between cells and advocating the major
role of cell fusions and symbiogenesis in rapid evolutionary
change. Confirmation of the mitochondrion and chloroplast in
eukaryotic cells as descendants of well-defined prokaryotes
was a major turning point away from the gradualist ideology
that dominated evolutionary thinking for most of the 20th
Century. Since then, we have come to appreciate more the major
evolutionary roles of cell-cell interactions and cellular
control of genome structure. The well-established phenomena of
symbiosis, hybridization, horizontal DNA transfers, genome
repair, and natural genetic engineering have revolutionized
our understanding of genome variation. Rather than a series of
accidents randomly changing a ROM (read-only memory) heredity
system, we realize that active cell processes non-randomly
restructure a RW (read-write) genomic storage system at all
biological time scales.
The main goals of this paper are to
explain what molecular biology and genome sequencing have
taught us about the real-time processes of genome variation
and what the DNA sequence record tells us about the history of
evolution. The take-home lessons can be summarized as:
• Genomes are systems under cell
control, not strings of atomistic agents;
• Genomes are RW memory systems;
• Systemic innovation is the key
problem in evolution science;
• Cells are active, cognitive
participants in evolutionary innovation.
For
those readers who wish to see graphic illustrations of some of
the points made below, a .pdf file of the lecture presentation
can be downloaded at http://shapiro.bsd.uchicago.edu/evolution21.shtml. In order to economize on the
reference list, many citations are given as links to online
bibliographies.
1. Lynn Margulis: Looking at
living organisms in action; creating new species and merging
genomes by symbiogenesis
Watching cells in action was always one of the joys of
attending a lecture by Lynn Margulis. Often she showed videos
of eukaryotic microbes and bacteria from an exotic
environment, like the termite intestine, moving with
extraordinary synchrony. Typically, the protist would be
coated with thousands of cells from one or more bacterial
species providing the motive force for the larger eukaryotic
cell.
What I learned from Lynn’s
fascination with symbiogenesis and the intimate associations
between cells from different biological kingdoms was the
largely unacknowledged power of both prokaryotic and
eukaryotic cells to control their activities and to coordinate
with other cells. It is impossible to imagine how a
symbiogenetic relationship could succeed without both
metabolic and genomic integration. How did the merging cells
synchronize their cell cycles so that one did not outgrow and
lose or overgrow and destroy the other? We do not have the
answer to such questions, but merely posing them opens the
door to a realm of active cell control regimes that, I
believe, will revolutionize 21st Century biology.
In this contribution to the ELS
Symposium honoring Lynn, I wish to discuss how much of genome
evolution results from active cell processes. The focus on
genome evolution reflects my personal experience as a
bacterial geneticist, repeatedly surprised by how successfully
E. coli managed to
engineer its own genome (Shapiro 2009). The idea of genome change as a
result of natural genetic engineering is a distinct departure
from the conventional wisdom developed in the middle of the 20th
Century, which posited a stochastic succession of accidental
changes as the ultimate sources of evolutionary variation (Huxley 1942).
If we remember that conventional
evolutionary wisdom was formulated before the discovery of DNA
as a major carrier of inherited information, then we will be
less resistant to thinking that assumptions about the nature
of genome change were destined to change as our knowledge of
and our technology for reading DNA sequences advanced in the
succeeding decades. Lynn, of course, realized that the early
symbiogenetic thinkers were correct in seeing active cell
processes at work in evolution making rapid major changes (Mereschkowsky 1926;
Walin 1927; Kozo-Polyansky 2010). She correctly emphasized two
aspects of the evolutionary process that have been amply
confirmed by DNA sequence evidence:
A large number of symbiotic
associations involve the gonads or other pre-germinal tissues
and so will have a potential evolutionary effect on the genome
passed down to future generations (http://shapiro.bsd.uchicago.edu/Gonadal_Symbiosis.html). Moreover, it is significant to
note that sexual reproduction involves a merger between two
distinct cells. This basic fact means that the evolutionary
history of all sexually reproducing organisms involves
coordinated cell-cell interactions. We generally take these
interactions for granted, but we will see later that sex
between different populations or species lacks the usual
smooth intercellular coordination and often unleashes events
of profound evolutionary importance.
2. DNA and the Irony of Molecular
Biology: Networks, cognition, and agency in all areas of
cell biology.
2.1
Conventional wisdom: Crick’s Central Dogma
The
demonstrations that DNA carries inherited information (Avery 1944;
Hershey and
Chase 1952) and the
elucidation of its self-complementary double helical structure
(Watson
and Crick 1953)
initially offered the possibility of a purely physico-chemical
explanation of heredity (Watson 1953). The
idea of coded information in DNA sequences determining the
functional proteins and thereby the properties of cells found
its sharpest exposition in Francis Crick’s “Central Dogma of
Molecular Biology” (Crick 1958;
Crick 1970).
The
Central Dogma fit well with the idea of the gene as a basic
unit of heredity. But the unitary, elemental notion of the
gene began to dissolve as molecular geneticists probed how
cells utilize DNA information. Starting in the 1950s, the
atomistic gene was deconstructed into complex clusters of
coding sequences and regulatory sites needed for expression
and control (Benzer 1956;
Jacob 1961;
Jacob and
Wollman 1961; Benzer 1962;
Tjian 1995). In the
1970s, many coding sequences were found to be split into
separate exons that had to be spliced together, adding another
layer of control and flexibility (Chambon
1981; Sharp 1994;
Ast 2005).
2.2
Cell and and Genome Networks; Systems Biology
As
molecular genetic analysis dissected various aspects of cell
and developmental biology, the abundance of interacting
execution and regulatory factors came to represent a basic
principle of biological action (http://shapiro.bsd.uchicago.edu/Regulatory_Networks.html ). By
the beginning of this century, the notion of individual cell
or organismal characters encoded by single genes gave way to
the “systems biology” concept of all traits determined by the
coordinated action of molecular networks (http://shapiro.bsd.uchicago.edu/ExtraRefs.SystemsApproachGeneratingFunctionalNovelties.shtml). These
networks contain components both at the genomic level and
outside the genome in proteins and regulatory RNA molecules.
2.3
Bacterial cognition in metabolism
Molecular
studies produced another unanticipated insight into biological
function when they revealed the central role of intracellular
and extracellular sensory processes. These cognitive
functionalities were apparent from the earliest days of
pre-molecular biology. In 1942 Nazi-occupied Paris, Jacques
Monod completed his doctoral research on quantitative
measurements of bacterial nutrition and growth (Monod 1942). (That
was his day job. By night, he was also a leader of the
Resistance.)
By
asking what happened when he presented the bacteria with a
mixture of sugars to consume, Monod made the striking
discovery that they consumed them in two stages. Using the
simple but elegant method of changing the relative proportions
of the two sugars, he demonstrated that the bacteria
completely consumed the preferred sugar before taking a pause
to adjust their metabolism and then consume the other sugar.
This discovery was the start of Monod’s pioneering work with
his Institut Pasteur colleagues on the lac (lactose) operon
of E. coli(Morange
2010; Ullmann 2010)
Together
with other work at the Institut Pasteur on viral control (Lwoff 1954),
Monod’s lac operon
research was to form the basis of our current understanding of
genome regulation. The key concepts of interactive
communication between regulatory proteins, specialized
recognition sites in the DNA, and signal molecules all arose
from this research (Jacob and
Wollman 1961). From
later studies, we learned that bacteria sugar sensing utilizes
molecules which have other tasks as well, such as DNA binding
and sugar transport across the cell membrane (Deutscher
2008; Gorke
and Stulke 2008). From
the molecular evidence on lac and many other
systems, we can conclude that there is no Cartesian dualism in
bacteria. Sensing is built into the basic fabric of the cell.
2.4
Bacterial cognition in proofreading DNA replication
The role
of sensory molecules also emerged from studies of how cells
protect their genomes. The accuracy of E. coli DNA
replication is such there are fewer than one mistake per
billion nucleotides incorporated (Kunkel and
Bebenek 2000). This
are more than five orders of magnitude (100,000-fold) better
precision than the replication polymerase achieves on its own.
All the extra precision comes from two overlapping
proofreading systems that detect and then correct mistakes
after they occur (Radman and
Wagner 1988; Rennie 1991). When
errors are found, the proofreading apparatus can distinguish
the old (presumably correct) strand from the newly synthesized
(presumably mistaken) strand and fix only the latter. In this
way, assuring the accuracy DNA replication has the same
cognitive basis as quality control processes in human
manufacturing.
2.5
Bacterial cognition and action in responding to UV damage
When
genomes suffer damage from radiation and chemicals, the
results include cell death and increased frequency of genetic
changes (mutations). Decades of research have shown that cells
play an active role in repairing both lethal and mutagenic
damage – and also, surprisingly, an active role in producing
mutations in response to damage.
The
inducible nature of the repair and mutagenesis phenomena was
clearly shown in some highly original experiments during the
1950s by the Swiss physicist-turned-molecular geneticist, Jean
Weigle (Weigle 1953;
Weigle and
Bertani 1953). In
order to distinguish between the genetic target of UV
irradiation and the cell environment, Weigle used the
bacterial virus, lambda, as his test system. In the viral
particles, lambda DNA could be treated independently of the
host cell, while the host cell could be treated independently
of the target DNA prior to infection.
Weigle
obtained the following set of results from the four possible
combinations of independently irradiating virus and host cell:
Infected
cells:
Unirradiated
UV-irradiated
Unirradiated virus
No
lethality, low viral mutation frequencies
No
lethality, elevated viral mutation frequencies
UV-irradiated virus
High
lethality, low mutation frequencies
Low
lethality, high viral mutation frequencies
Unirradiated virus infecting unirradiated cells was the
baseline control situation. Greater
lethality
but reduced
frequency of mutations when irradiated virus infected
unirradiated cells showed that radiation of the cells prior to
infection induced both the capacity to repair to lethal damage
(so-called “Weigle repair”) and also stimulated the ability to
produce mutations (“Weigle mutagenesis”).The
unexpected mutation frequency when unirradiated virus infected
irradiated cells confirmed that UV induced mutagenesis
capacity in the cells prior to infection and showed that mutational
activity could occur on untreated DNA.
A couple decades’ research revealed
the bacterial reaction to UV irradiation and chemical
mutagenesis to be an integrated “SOS” response to DNA damage (Witkin 1975; Devoret 1979; Howard-Flanders 1981;
Witkin 1991). A multifunctional sensor protein,
RecA, binds to regions of single-stranded DNA that accumulate
in cells with damaged genomes. This binding activates the
expression of a suite of protein functions that include repair
and recombination proteins, several of which are “mutator” DNA
polymerase activities that actively introduce the incorrect
nucleotides which register as mutations at the end of the
experiment (Sommer, Knezevic et al. 1993;
Goodman 2002).
2.6 Cell cognition in checkpoint
control of the cell cycle
Among the SOS functions is an
intriguing protein that inhibits bacterial cell division until
DNA repair has been completed (Huisman, D'Ari et al. 1984). This constitutes what we now
recognize as a “checkpoint” function that holds up the cell
division cycle until the genome and other necessary components
are in the proper state to proceed (Hartwell 1989; Murray and Kirschner 1991;
Weinberg 1996).
Checkpoints
are ubiquitous in biology and represent an inherently
cognitive set of controls that ensure the reliability of cell
division, a process that involves hundreds of millions of
biochemical and biomechanical events (http://shapiro.bsd.uchicago.edu/ExtraRefs.CellCycleCheckpoints.shtml ). The cell monitors is own
internal status and adjusts the progress of different
subroutines so that none outruns the coordination needed to
produce intact daughter cells with complete genomes. Among the
most impressive is the “spindle checkpoint,” which senses and
guarantees that all the chromosomes are properly aligned along
the cell division spindle apparatus so that each daughter cell
receives one, and only one, copy of each duplicated chromosome
(McIntosh and McDonald 1989;
Musacchio 2011). As with the super-low error
frequencies in DNA replication, the remarkable success of cell
division depends upon cognition rather than mechanical
precision.
2.7 Sensing, communication and
response in the programmed cell death decision
Not only cell survival and
reproduction but also programmed cell death depends on
cognitive processes. All cells, from bacteria to plants and
animals, have systems for a regular cell death process (http://shapiro.bsd.uchicago.edu/Apoptosis.html ). These cell suicide routines are
useful for defense against pathogens, modeling multicellular
development, establishing symbioses, and eliminating in a
predictable manner cells that cannot repair serious damage.
Cell death decisions are typically
contingent upon the receipt of intercellular signals. A good
example of how cognition operates in cell death decisions can
be seen with mammalian cells exposed to DNA damaging agents,
like ionizing radiation. If the cell receptors detect
anti-survival signals, such as tumor necrosis factor (TNF),
they tend to suffer higher levels of programmed cell death,
whereas cells that detect proliferative signals, such as
extracellular matrix proteins or tumor growth factor beta,
undergo more frequent genome repair and experience much less
lethality for the same amount of damage (Meredith, Fazeli et al. 1993;
Ferrer and
Planas 2003; Marini and
Belka 2003).
By the beginning of this century,
the results of molecular cell and developmental biology made
it clear that new basic concepts were necessary to explain how
the properties of living organisms came to be and were passed
on from one generation to the next. The one gene – one
character paradigm (Beadle 1948) and the unidirectional flow of
information in Crick’s Central Dogma (Watson 1953; Crick 1958; Crick 1970) were no longer tenable (Shapiro 2009). Complexity, networks, signaling
and cognition (i.e.,
sensing and decision-making) had emerged as essential features
of living cells.
3. Barbara McClintock and the
discovery of the internal processes of genome repair and
genome restructuring
An independent chain of events leading to our modern
understanding of cell action on the genome dates back to the
1930s, when Barbara McClintock began her studies on the
chromosomal basis of X-ray mutagenesis in maize. She took
pains to recount this history in her Nobel Prize lecture (McClintock 1984).
3.1 Repair of
chromosomes broken by X-rays
"X-ray mutagenesis" was discovered
in the 1920s by Herman J. Muller (Muller 1927). X-rays were the first of many
external agents that induce mutations, and thus hereditary
changes, in living organisms. It is not very well known that
these X-rays triggered cell repair functions and sophisticated
chromosome rearrangements, not simple "gene mutations."
McClintock
set out in 1931 to analyze the X-ray mutants Louis Stadler had
isolated from maize plants at the University of Missouri. She
was the right person to do this. As a Cornell graduate student
in the 1920s, she personally developed the microscopic methods
that allow scientists to visualize the 10 maize chromosomes.
Some X-ray mutants were unstable and "variegated" in their
inherited properties. We can see similar variegating
instabilities in spotted or striped kernels on ears of Indian
corn. Based on her understanding of chromosome behavior,
McClintock was able to explain the "variegating" mutants.
She
proposed that variegating plants carried ring chromosomes.
Such rings formed by the fusion of two broken ends near the
telomeres of a previously linear chromosome. Ring chromosomes
sometimes failed to distribute properly to daughter cells as
the plant grew, thus producing the variegated patterns.
Although colleagues scoffed at McClintock's idea, she went on
to identify the predicted ring chromosomes in the variegating
mutants (McClintock
1932). Other
mutants induced by X-ray treatment also carried chromosome
rearrangements. Sections of chromosomes were deleted,
translocated, inverted, and duplicated. All of these
rearrangements could result from breaking chromosomes at two
sites and then rejoining the broken ends to build brand new
chromosome structures. These were quite different from the
"gene mutations" imagined by Muller and his colleagues.
3.2 The
cognitive nature of broken chromosome repair
McClintock
reasoned that maize cells must have an inherent capacity to
join broken chromosome ends when two of them are present in
the cell. In the later 1930s, she devised experimental methods
to induce new chromosome arrangements. Using experimentally
generated breaks, she demonstrated conclusively that maize
cells can detect, juxtapose, and fuse broken chromosome ends (McClintock
1939; McClintock
1942).
McClintock
realized that the X-rays broke chromosomes wherever they
happened to strike. Breakage alone, however, was insufficient
to generate a mutant chromosome. Broken chromosomes would get
lost. The cell's ability to repair the damage by fusing broken
ends was essential. In other words, X-ray mutagenesis provoked
cell action. McClintock had started down a novel path of
thinking. As she
wrote in her Nobel lecture:
“The conclusion seems inescapable
that cells are able to sense the presence in their nuclei of
ruptured ends of chromosomes and then to activate a mechanism
that will bring together and then unite these ends, one with
another…The ability of a cell to sense these broken ends, to
direct them toward each other, and then to unite them so that
the union of the two DNA strands is correctly oriented, is a
particularly revealing example of the sensitivity of cells to
all that is going on within them…There must be numerous
homeostatic adjustments required of cells. The sensing devices
and the signals that initiate these adjustments are beyond our
present ability to fathom. A goal for the future would be to
determine the extent of knowledge the cell has of itself and
how it utilizes this knowledge in a "thoughtful" manner when
challenged” (McClintock
1984).
By raising
the idea of cells operating in a “thoughtful manner,”
McClintock was self-consciously trying to point biological
research in a revolutionary new direction. This comment led
Dennis Bray to point out in his 2009 book, Wetware: A computer in
every living cell, that she was the first scientist to
ask what a cell knows about itself (Bray 2009).
3.3 Transposable
element activity as a response to “genome shock”
The studies on chromosome breakage
and rejoining led directly to McClintock’s unexpected
discovery of mobile genetic elements (McClintock 1987). They prepared her for analyzing
variegation and chromosome breakage, and they led her to
explore the consequences of introducing broken ends into maize
cells at fertilization. Rather than explain in her Nobel
lecture how she demonstrated genetic mobility, McClintock
chose to relate experiments confirming that the “shock” from
receiving a single broken chromosome end had the extraordinary
effect of awakening previously latent transposable elements in
the genome: “It seemed clear that these elements
must have been present in the genome, and in a silent state
previous to an event that activated one or another of them…It
was concluded that some traumatic event was responsible for
these activations.”
Most of the remainder of McClintock’s lecture provided
“further
examples of
response of genomes to stress.” She wished to emphasize the
generality of her observations across the living world, taking
examples from both plants and animals. With respect to
evolutionary change, she emphasized the “shock” of
interspecific hybridization, a process that leads to whole
genome doubling, widespread genome reorganization, and
formation of novel species (http://shapiro.bsd.uchicago.edu/ExtraRefs.WholeGenomeDoublingCriticalStagesEvolution.shtml). For McClintock, genome change was
not accidental. Change was a response to life history
challenges.
Why did McClintock focus so much attention on cell
sensing and not on research that provided molecular evidence
in support of her previously heretical views (Shapiro 1983)? From the way she ends her lecture,
we can appreciate that McClintock had her perspective directed
towards the years ahead, not those behind:
“In the future attention
undoubtedly will be centered on the genome, and with greater
appreciation of its significance as a highly sensitive organ
of the cell, monitoring genomic activities and correcting
common errors, sensing the unusual and unexpected events, and
responding to them, often by restructuring the genome. We know
about the components of genomes that could be made available
for such restructuring. We know nothing, however, about how
the cell senses danger and instigates responses to it that
often are truly remarkable.” (McClintock 1984)
McClintock wanted to draw special
attention to what she saw as the most challenging problems in
biology: cognition and purposeful action by living cells. As
she knew well from her long experience with 20th
Century genetics and cell biology, whether life has special
“vital” properties that separate it from inorganic matter has
been among the most fiercely disputed topics in the history of
science. In its early days, molecular biology promised to
provide us with an explanation of life in terms of physics and
chemistry. However, as we saw previously, it has succeeded
instead in amazing us with the richness and sophistication of
intra- and inter-cellular control and communication networks.
3.4 McClintock’s preview of
molecular discoveries
In three
important ways, McClintock’s work anticipated the molecular
rediscovery of mobile genetic elements in the 1960s and 1970s
(Bukhari 1977;
Shapiro 1983;
Shapiro 2011):
(i)Demonstrating
that genome repair and restructuring are regulated and
responsive processes the cell executes with a range of
dedicated functions;
(ii)Showing
how mobile fragments of genetic material could alter the
expression regime of virtually any coding region or genetic
locus throughout the genome (McClintock
1953; McClintock
1961);
(iii)Illustrating
the ability of these mobile “controlling elements,” as she
named them, to generate genome networks coordinating the
expression of unlinked loci through insertion of related
elements (McClintock
1956).
She
laid the groundwork for a revolution in our understanding of
the possibilities for genome variation in evolution.
4. The molecular toolbox for
genome restructuring: Natural Genetic Engineering
(Technical)
While McClintock was working out
many details of action involving her “controlling elements,”
the pioneers of molecular genetics were discovering the many
ways that cells acquire, transfer and restructure DNA
molecules (http://shapiro.bsd.uchicago.edu/ExtraRefs.MolecularMechanismsNaturalGeneticEngineering.shtml). Multiple seemingly independent
lines of research supplied abundant evidence of these
capabilities. Although this section will seem rather technical
to some readers, it is essential to mention the details and
thus emphasize the point that DNA change is not a series of
accidents. On the contrary, all DNA changes result from
dedicated cell functions doing their jobs (“the molecular
toolbox”). The sheer variety of cellular DNA restructuring
functions is unknown to all but a small number of specialists
in the field. The existence of these functions needs to be
more widely appreciated because they are such basic components
of cell action in evolution.
4.1 DNA uptake from the
environment
Studies of bacterial transformation,
first identified in 1928 (Griffith 1928), provided the first evidence that
DNA was the “transforming principle” and therefore carried
hereditary information (Avery 1944). More recent studies have
identified “competence” for DNA uptake from the external
environment in many diverse groups of bacteria (http://shapiro.bsd.uchicago.edu/TableIII.1.shtml ).
Analysis of the genetic control of
competence has revealed many different modes of regulation,
including control by stress responses, and intercellular
signaling molecules, and activation by metabolic signals in
the environment. Later studies, mostly for genetic engineering
purposes, using cultured plant and animal cells have shown
that the ability to take up and integrate DNA from the
environment is widespread among living cells. Unfortunately,
little is known about the molecular apparatus which enables
these examples of DNA incorporation by eukaryotic cells.
4.2Viruses, proviruses and
site-specific recombination
By following single dividing
bacteria through hours of patient microscopic examination,
Andre Lwoff, Monod’s doctoral supervisor, convincingly
documented that some bacteria possessed the hereditary
capacity to produce viruses (Lwoff 1953; Lwoff 1954; Lwoff 1966). This property was named
“lysogeny” (= causes lysis) because the liberated viruses
could infect and lyse closely related bacteria.
The study of lysogeny in the
following years provided many important insights into the
relationships of virus and cell genomes, a major source of
evolutionary novelty (Forterre 2006; Koonin, Senkevich et al. 2006;
Koonin 2010). Viruses can infect cells and
establish their own genome as part of the cell genome, called
a “provirus” (Lwoff 1957). In the quiescent provirus state,
parts of the viral DNA can encode functions that alter the
properties of the host. For example, the bacteria that cause
diphtheria do so only when they harbor a provirus that encodes
the diphtheria toxin (Groman 1953).
Viruses can also mobilize DNA
sequences from one cell to another. This DNA “transduction”
capacity was discovered in the 1950s with bacteria (Zinder and Lederberg 1952;
Zinder 1958), but it appears to have few limits.
So-called “metagenomic” analysis (Gilbert
and Dupont 2011) of unpurified environmental
viruses has shown that some contain combinations of sequences
from all three kingdoms of life: bacteria, archaea and eukarya
(http://shapiro.bsd.uchicago.edu/Viral_Composites.html). This means that the progenitors
of these viruses passed through cells from all kingdoms and
thus can carry sequences from one kingdom to another. The
giant DNA viruses that infect amoebae are particularly
noteworthy in this regard (Colson and Raoult 2010). So it appears as though amoebae
and other protists may constitute an evolutionary “melting
pot” for mixing DNA coming from different sources (http://shapiro.bsd.uchicago.edu/Amoebal_Viruses.html ). Some amoebal viruses are also
known to infect other kinds of cell, including plants and
animals.
The study of how provirus genomes
enter and leave host cell chromosomes has revealed a number of
different mechanisms for mobilizing DNA molecules. In
bacteria, one of the most important mechanisms is the process
called “site-specific recombination,” in which specialized
“recombinase” proteins act on particular short sequences to
carry out reciprocal exchanges (http://shapiro.bsd.uchicago.edu/Site-specific_Recombination.html). This form of specialized
recombination has been adapted for many additional purposes in
bacterial cells (Hallet and Sherratt 1997;
Hallet and Sherratt 2010): e.g., turning protein
synthesis on and off, reducing double circles of replicated
DNA to two single circles for equal transmission to daughter
cells, excision of DNA introns during cell differentiation,
and modification of protein structures by inverting segments
within the coding sequence (http://www.huffingtonpost.com/james-a-shapiro/dna-as-poetry-multiple-me_b_1229190.html).
At a low but significant frequency
(about .001%), the provirus exits from the bacterial
chromosome by an aberrant event rather than precisely-targeted
site-specific recombination. In these aberrant events, the
viral genome can acquire an adjacent fragment of the bacterial
chromosome (http://Shapiro.bsd.uchicago.edu/Specialized_Transduction_and_in_vivo_Cloning.html ). The bacterial fragment is
effectively cloned into the virus genome by cell action and
reproduces together with the virus. This kind of in vivo cloning can
be applied to many regions of the bacterial chromosome. It
served as the cloning process used to obtain the specific DNA
sequences for the first isolation of a defined genetic locus (Shapiro, Machattie et al. 1969;
Shapiro 2009).
4.3Antibiotic
resistance, pathogenicity, symbiogenesis and horizontal DNA
transfer in bacteria
The
spread of bacterial antibiotic resistance among bacterial
pathogens starting in the 1950s and 1960s was a real-world
stimulus to research in molecular genetics. The story of how
we came to understand this major evolutionary episode that
took place in real time subject to detailed investigation is a
fascinating and instructive chapter in the history of science.
It illuminates the insight that scientific reality consists of
more than experimentally confirming hypothetical predictions.
In the
early days of molecular biology, bacterial geneticists applied
conventional evolutionary concepts to explain the evolution of
antibiotic resistance. The theory was that mutations could
alter the structure of cell components and either block entry
of the drugs into the bacteria or prevent their action on
cellular targets, such as the enzymes essential to cell wall
synthesis. Even if the initial mutation did not confer a high
degree of resistance, accumulation of several sequential
changes would result in resistance to the antibiotic levels
used in clinical medicine. Indeed, a wide variety of
laboratory experiments confirmed this theory, and bacterial
geneticists isolated the predicted mutant strains (Hayes
1968). In
virtually all cases, the resistant mutants grew less well than
the parental sensitive bacteria, leading to the comforting
conclusion that resistant bacteria would not significantly
accumulate in nature (Gorini
1966). The
degree of confidence was so great that the US Surgeon General
in 1967 declared that
“the war against infectious diseases has been won”(Fauci
2001).
There
were problems both with the science and the new public health
policy based on it. The Surgeon General “misunderestimated”
bacteria, which followed their own evolutionary rules and did
not listen to what the scientists said they should do.
Although experimentally confirmed, the mutation theory of
antibiotic resistance failed to account for most cases in the
real world. Resistance continued to spread among bacteria
isolated in clinics around the globe. Even more ominously,
different strains of pathogenic bacteria increasingly
displayed resistance to more than one antibiotic at a time.
Research
pioneered in Japan found that multiple antibiotic resistances
could be transferred simultaneously from one bacterial species
to another (Watanabe
1967). The DNA
agents responsible for this transfer are circular molecules
that are called multidrug resistance plasmids, which can move
from one cell to another (Clowes
1973; Novick 1980). Multiply
resistant bacteria were not altered in their cellular
structures or inhibited in their growth properties. Rather,
they had acquired new biochemical activities that could
destroy or inactivate the antibiotics, chemically alter their
targets, or remove them from the bacterial cell (Davies
1979; Foster 1983;
Levy 1998).
Multiple
antibiotic resistance clearly represented genome change and
evolution of a type unimagined in the pre-DNA period. DNA
molecules transferred “horizontally” between unrelated cells
rather than descending from ancestral cells. Horizontally
transferred DNA could carry complex sets of genetic
information encoding multiple distinct biochemical activities.
Evolutionary leaps involving several characteristics at once
could occur through horizontal DNA transfer. In the early
1980s, two obscure French-Canadian microbiologists published a
book called A New
Bacteriology, postulating a radically different approach
to thinking about bacterial evolution (Sonea
and Panisset 1983). Sonea
and Paniset argued that bacteria have a huge collective genome
distributed throughout nature in different kinds of cells, in
viruses and latent in the environment. When a new ecological
niche appears, bacteria can assemble the genomic assets they
need to exploit the opportunity.
Subsequent
research has bolstered Sonea and Paniset’s initially
outlandish idea. As we have seen, bacteria have all the
abilities they need to acquire DNA from the environment, from
viruses and from other cells (http://shapiro.bsd.uchicago.edu/TableIII.1.shtml). Viruses
are the most abundant biological entities in the environment (Edwards
and Rohwer 2005) and
carry a multitude of sequences encoding important cellular
capabilities, such as photosynthesis (http://shapiro.bsd.uchicago.edu/Viral_Composites.html).
Detailed
study of many bacterial characteristics, especially
pathogenicity and virulence, indicate that they are encoded by
plasmids or by critical segments of the DNA, so-called
“genomic islands” (Hacker
and Carniel 2001; Juhas,
van der Meer et al. 2009). The sequences
of genomic islands show that they have been acquired from
unrelated organisms and often integrated into the cellular
genome by site-specific recombination.
Among the
DNA sequences most often found on plasmids and in genomic
islands are those encoding molecules needed for virulence in
pathogenic bacteria (http://shapiro.bsd.uchicago.edu/Bacterial_Cell_Biology_In_Pathogenesis_and_Symbiosis.html).
Among
the most important virulence molecules are those that form
intricate complexes in the cell envelope and allow the
bacterial pathogens to inject protein and RNA molecules into
cells of the host organism, whether animal or plant. In so
doing, they subvert host cell regulatory circuits in a way
that meets the invading bacterium’s needs. Similar
macromolecular injection systems are involved in established
symbiotic relationships. Bacteria must be the smartest cell
biologists on the planet; they control events in cells of
higher organisms in a way that mere human scientists can only
dream of imitating.
The
macromolecular transport structures bacteria use to acquire
DNA from the environment and transfer plasmids and other DNA
molecules between cells are very similar (http://shapiro.bsd.uchicago.edu/Competence_for_DNA_Uptake.html
). These DNA mobilization structures, in turn, are closely
related to other envelope-spanning structures involved in the
synthesis of high-energy storage molecules and rotation of
bacterial flagella (literally, “whips”) for swimming through
fluids (http://shapiro.bsd.uchicago.edu/Bacterial_Surface_Structures.html
). They are even used as push-pull apparatus for so-called
“twitching” movement across solid surfaces. Clearly, there has
been wide-ranging use and reuse of these elaborate systems for
multiple functions in the course of bacterial evolution. [1]
4.4 Horizontal transfer outside bacteria
We have
learned through genome sequences that horizontal DNA transfer
is not limited to the prokaryotic kingdoms of bacteria and
archaea (http://shapiro.bsd.uchicago.edu/TableIII.1.shtml ). There
is well-documented DNA transfer from bacteria to plants in
nature and, in the laboratory, to yeast and animal cells (http://shapiro.bsd.uchicago.edu/Interkingdom_and_Eukaryotic_Horizontal_Transfer.html
). The genomes of endosymbiotic bacteria are found in the
nuclear genomes of host invertebrates. There are DNA sequences
that have moved between unrelated plants, animals, fungi, and
protists. All microbial kingdoms have contributed functional
coding regions to multicellular plants and animals, and
genomic systems permitting certain key ecological adaptations
include horizontally acquired sequences.
In
addition to these examples of intercellular horizontal
transfer, it is highly appropriate in a symposium honoring
Lynn Margulis to mention intracellular
horizontal transfer following symbiotic events. Most of the
proteins originally encoded in the genomes of bacterial
ancestors of mitochondria and plastids are now encoded in the
nuclear genomes of eukaryotic cells because of
post-endosymbiosis organelle to nucleus DNA transfer (http://shapiro.bsd.uchicago.edu/Intracellular_Horizontal_Transfer.html
). In cases where algae have formed secondary endosymbioses in
other eukaryotic lineages, there are four or more genome
compartments: nucleus, mitochondrion, nucleomorph (former
algal nucleus) and plastid (Embley
and Martin 2006).
Comparing the genomes of free-living algae with their
symbiogenetic descendants, it is evident that DNA sequences
migrated from the nucleomorph and plastid genomes to the
nuclear chromosomes.
4.4Transposable elements in bacteria
McClintock’s discovery of mobile
elements in the 1940s was extended by bacterial geneticists in
the 1960s and 1970s studying so-called “spontaneous” mutations
(i.e., genetic
changes not induced by mutagenic treatment whose nature and
origins were unknown) (Bukhari 1977; Cohen and Shapiro 1980;
Craig 2002; Shapiro 2009). A number of these mutations had
properties that were not explicable by the recognized types of
DNA changes: nucleotide substitutions, frameshifts and
deletions. They could terminate or activate transcription (http://shapiro.bsd.uchicago.edu/Bacterial_Transposons.html). The anomalous mutations proved to
be the result of insertion of specific DNA sequences that
occurred at many locations in the genome.
The mobile genetic elements
responsible for the insertions came to be called “insertion
sequences” or IS elements. Later, certain viruses were found
to integrate their proviruses into the bacterial chromosome
and replicate their genomes by the same processes that
mobilized a subset of IS elements from one genomic location to
another. Some bacterial species contained dozens or hundreds
of IS elements in their genomes. The majority of the cellular
and viral elements encoded proteins, named “transposases,”
that mediated their movement (transposition) through the
genome (Polard and Chandler 1995).
The discovery of IS elements was
quickly followed by the observation that many antibiotic
resistance determinants are also mobile (Hedges and Jacob 1974). The ability to move DNA to new
locations in the genome had evolved as a powerful means to
combine different antibiotic resistances into a single plasmid
DNA molecule. The mobile elements that encoded a specific
phenotype, like resistance, were called “transposons,” and
that has become the generic term for mobile DNA elements. As
studies of mobile DNA in bacteria continued, naturally
occurring and experimentally derived transposons were shown
able to include any segment of the bacterial genome.
4.6Integrons and the rapid evolution
of multiple antibiotic resistance
In the 1990s, as they analyzed
multiple antibiotic resistance plasmids and transposons, Ruth
Hall and her colleagues recognized an additional mode of DNA
mobility used to construct complex genomic determinants (Hall and Stokes 1993; Hall and Collis 1995). They discovered that many
plasmids and transposons acquired additional individual coding
sequences one at a time by the same kind of site-specific
recombination mechanism that inserted proviral genomes into
bacterial chromosomes. The insertion site was part of a
structure she called an “integron,” which encoded the
recombinase activity necessary to insert the cassettes
downstream of an active transcription signal, where they would
be expressed (http://shapiro.bsd.uchicago.edu/Integrons_Super-integrons.html ).
Where the circular cassette
structures originate remains a mystery. But we now know that
integrons and larger “super-integrons” have acquired a large
number of different kinds of coding sequences, including those
for bacterial virulence, control of transcription, synthesis
of various biochemicals, motility and intercellular signaling.
4.7 “Legitimate” and
“illegitimate” recombination
It is interesting to remark, as we
pass through our review of how cells actively restructure and
assemble their genomes, that it was a commonplace in the
latter 20th Century to call site-specific
recombination, transposition and other kinds of DNA
rearrangements “illegitimate recombination,” and the practice
continues (http://shapiro.bsd.uchicago.edu/Legitimate_and_Illegitimate_Recombination.html ). This exclusionary term was meant
to distinguish these “exceptional” DNA exchange processes from
the homology-dependent recombination that had been used to
construct genetic linkage maps since the early 20th
Century (Hayes 1968; Stahl 1987).
We can understand the initial
surprise that geneticists comfortable with assumptions about
the rules of meiotic crossing over experienced when they
encountered first McClintock’s work and then bacterial
research on DNA transfer and rearrangements involving
plasmids, transposons, integrons and viruses. Clearly, these
genome restructuring events would appear strange to scientists
used to thinking of the Constant Genome that changes only by
accidents or the homologous recombination of allelic
differences. Nonetheless, it is satisfyingly ironic to note
that this century has found the most abundant DNA coding
sequences in the genome databases to be those for transposases
(Aziz, Breitbart et al. 2010). The reason for this abundance is
that transposons and other mobile elements are the most common
sequences in many genomes. For example, the draft human genome
revealed at the start of the 21st Century, our DNA
is less than 1.5% protein coding sequences and over 40%
dispersed repeats consisting of various mobile genetic
elements (Lander, Linton et al. 2001).
In addition to playing an essential
role in meiosis (Ding, Haraguchi et al. 2010;
Szekvolgyi and
Nicolas 2010), homologous (“legitimate”)
recombination has been adapted for specialized targeted
(“illegitimate”) functions in microbes.
Among yeasts and other fungi that
reproduce both as diploids and haploids, growth as diploids is
more stable against inevitable DNA damage. Diploids have two
copies of the genome, and one can serve as a template for
repairing breakage in the other (Goldfarb and Lichten 2010). In order to form diploid cells
following meiosis and sporulation, the haploid progeny of both
budding and fission yeast spores have adapted homologous DNA
breakage repair to undergo self-mediated sex-change operations
(Haber 1998; Klar 2007; Yamada-Inagawa, Klar et al.
2007; Klar 2010). When a haploid cell has changed
its mating type, it can fuse with a cell having the opposite
parental mating type and form a diploid.
By using a sequence-specific DNA
break at the mating type locus, both yeasts direct a process
of homology-dependent recombination that changes genetic
content and cellular mating type. Budding yeast use the HO
endonuclease (Nasmyth 1993) to make the targeted break, while
fission yeast use a modified transposase protein (Rusche and Rine 2010). So this process of directed
recombination must have undergone at least two independent
evolution steps. Once the break has occurred, it is repaired
by recombination with a silent DNA cassette containing
opposite mating type information surrounded by homologous
sequences. The end result is a unidirectional gene conversion
and transfer of new mating-type information to an expression
site, generating the phenotypic sex change (Parvanov, Kohli et al. 2008).
A similar strategy of directed gene
conversion of new information from silent cassettes to
expression sites is used for the process of “antigenic
variation” by numerous prokaryote and eukaryote microbes (http://shapiro.bsd.uchicago.edu/Antigenic_Variation.html; http://shapiro.bsd.uchicago.edu/TableII.5.shtml). The directed gene conversions
change the structures of cell surface proteins and help these
microbes escape the defenses of the host immune system. While
the phenomenon of cassette exchange and the role of homologous
recombination systems are well documented, these antigen
variation events are not as thoroughly understood as the yeast
mating-type switches.
4.8 Retroviruses
Among the major shocks to
conventional reductionist thinking in the 1960s, was Howard
Temin’s assertion about certain RNA viruses. He claimed they
reproduced their genomes by inserting a DNA copy into the host
cell chromosomes and then transcribing the proviral DNA into
more viral RNA (Temin 1972). The connection with Lwoff’s
pioneering work with bacteria is clear. Temin’s arguments were
widely ridiculed until he demonstrated an activity in the
viral particles that “reverse transcribed” RNA into DNA (Temin 1970; Varmus 1987). This activity came to be “reverse
transcriptase” and viruses that employed this strategy
“retroviruses” (Coffin, Hughes et al. 1997). Reverse transcription disrupted
the neat picture of Crick’s Central Dogma. Crick was quick to
add an exception for RNA to his scheme of one-way information
transfer from the genome (Crick 1970), but the genie was out of the
bottle. A major new capacity for genome modification had
appeared on the scene.
Retroviruses have many parallels
with bacterial viruses, including the use of site-specific
recombinases or, more commonly, transposase-like proteins for
proviral integration (http://shapiro.bsd.uchicago.edu/Retrovirus_Integration.html ). We now recognize retroviruses as
the prototype of a family of “retrotransposable elements,” i.e., mobile elements
that move through the genome by virtue of an RNA intermediate
(Varmus 1987). Retroviruses have a complex
reproductive cycle that includes the formation of DNA
proviruses at many different sites in the host cell genome. It
can be thought of as an infectious process of transposition
from one genomic site in the first cell to a different site in
a second cell.
Occasionally, a cellular RNA
sequence can be incorporated into the retroviral genome. This
happened in the case of the first retrovirus to be discovered,
in 1910 (!), the Rous Sarcoma Virus that induces a
transmissible cancer in chickens (Rous 1910). Because of the connection with
cancer, these incorporated cell sequences are often called
“oncogenes” (Weinberg 1996). But we now understand that the
product of an oncogene is simply a cell function that disrupts
normal proliferation when produced in an aberrant manner. Even
though we can reproduce the process in the laboratory (Swain and Coffin 1992), we do not know how retroviruses
incorporate the cellular RNA sequences. But it greatly extends
the genome restructuring capability of this group of mobile
elements. Since retroviruses carry powerful signals for
activating transcription, they can also induce tumors by
inserting in the genome near coding sequences for a cell
function whose aberrant expression leads to oncogenic changes
in cell behavior (Nusse and Varmus 1982; Tsichlis 1987; Tsichlis, Lee et al. 1990). Activating expression by mobile
element insertions had been well documented in both bacteria (Pilacinski, Mosharrafa et al.
1977) and yeast (Errede, Cardillo et al. 1981).
4.9 Endogenous retroviruses (ERVs)
Many genomes, including our own,
harbor endogenous retroviruses (ERVs) (http://shapiro.bsd.uchicago.edu/Endogenous_Retroviruses.html ). ERV proviruses display
characteristic expression patterns during development and
produce viral particles in certain tissues. It is now clear
that ERVs have played several important roles in human and
mammalian evolution. They have contributed to functional human
coding sequences, genome variability, and chromosome
rearrangements in primate evolution. They have provided mobile
expression signals for the generation of genomic networks
active in embryonic development and the cell cycle. Most
notably, ERVs were central to the evolution of the placenta, a
key event leading to the radiation of eutherian mammals (http://shapiro.bsd.uchicago.edu/Retroviral_involvement_in_placenta_evolution.html ): placental tissues have high
levels of ERV expression and contain numerous retroviral
particles, and placenta formation requires proteins
(syncytins) evolved from ERV precursors. In addition, ERVs
play important roles in stress responses and disease.
4.10 Retrovirus-related LTR
retrotransposons
There are many mobile elements that
resemble retroviruses but lack sequences for essential cell
infection steps. These non-infectious elements are called “LTR
retrotransposons” and are found in many, if not all, groups of
eukaryotes (Coffin, Hughes et al. 1997;
McDonald, Matyunina et al.
1997). LTR strand for the “long terminal
repeats” characteristic of provirus structure. LTRs are the
products of the specific way that retroviruses and related
retrotransposons convert their RNA into proviral DNA (Varmus 1987; Coffin, Hughes et al. 1997). LTR retrotransposons are capable
of intracellular retrotransposition and provide a major source
of genome variability in virtually all eukaryotes, from yeast
and protists through higher plants and animals. All LTR
retrotransposons carry powerful signals stimulating genome
expression, and genome sequence data show that many played
important roles in the evolution of genome regulatory circuits
(http://shapiro.bsd.uchicago.edu/LTR_retrotransposons_and_genome_evolution.html ). Amplification of LTR
retrotransposons is also a major contributor to DNA expansion
in both plants and animals with large genomes (http://shapiro.bsd.uchicago.edu/Genome_Size.html ).
4.11 Non-LTR retrotransposons
(LINEs and SINEs) and insertion of reverse-transcribed RNA
into the genome
In addition to the retrotransposable
elements that share the LTR structure with proviruses, there
is another class of “non-LTR retrotransposons” which utilize
an entirely distinct process for reverse transcription and
integration of DNA copies (http://shapiro.bsd.uchicago.edu/Non-LTR_Retrotransposons.html). Genome analysis identified some
of these non-LTR elements as dispersed repeat sequences and
gave them the names SINEs (= short interspersed nucleotide
elements) and LINEs (= long interspersed nucleotide elements).
LINE elements encode their own reverse transcription and DNA
integration functions, while SINE elements are too short and
depend on LINE-encoded activities for their retrotransposition
to new locations in the genome. In some genomes, non-LTR
elements make up the most abundant classes of DNA. In our own,
for example, there are 850,000 LINE elements comprising 21% of
the sequenced genome and about 1,500,000 SINE elements that
account for 13% (Lander, Linton et al. 2001). Intriguingly, SINE elements tend
to be extremely taxon-specific. When Rick Sternberg and I
examined SINEs in mammals, we found that each order had its
own special group of elements (Sternberg and Shapiro 2005). It appears that these elements
originated and expanded rapidly when a new order emerged.
One of the most versatile features
of non-LTR retrotransposition is its lack of sequence
specificity for reverse-transcribing RNA and integrating the
DNA copy. This major difference from the LTR elements (which
use specific LTR sequences) means that LINE-encoded activities
can reverse transcribe and integrate the DNA copy of virtually
any RNA sequence into the genome (Kazazian 2000). This molecular promiscuity adds
greatly to a cell’s genome restructuring capabilities (http://shapiro.bsd.uchicago.edu/RNA_Reverse_Transcription_and_Genome_Insertion.html ). All kinds of processed RNA
molecules can be immortalized. This is the way that
intron-free copies of spliced RNAs enter the genome as
additional copies of a coding sequence (“retrogenes”) or as
“pseudogenes” that may lack full coding capacity but often
play regulatory roles. Sometimes the insertions produce novel
exons or fusion sequences that generate sequences encoding
novel proteins.
Another way that LINEs and SINEs
mediate genome creativity is by picking up adjacent sequences
and moving them to new genome locations. LINEs and some SINEs
frequently include downstream sequences when transcribed, and
these external sequences are incorporated into the genome
together with the non-LTR element (http://shapiro.bsd.uchicago.edu/LINE_SINE_Retrotransduction.html ). There are also SINEs which
similarly mobilize upstream sequences to new locations.
Laboratory studies and genome sequencing have shown that
non-LTR elements can insert either upstream or downstream
exons into a distant coding region. This exon mobilization is
significant in evolution because it mediates protein
innovation by “domain shuffling” (http://shapiro.bsd.uchicago.edu/Exon_Shuffling.html; http://www.huffingtonpost.com/james-a-shapiro/genetic-engineering_b_1541180.html).
DNA transposons have the same innovative exon shuffling
capacity.
4.12 Non-homologous end-joining
(NHEJ) in genome rearrangements
Although haploid cells cannot use
homologous exchanges to restore double-strand DNA breaks in
unreplicated genomes, repair still takes place by a process
known as “non-homologous end-joining” (http://shapiro.bsd.uchicago.edu/NHEJ.html). In NHEJ, the ends of broken DNA
molecules are processed and joined together. There are
multiple ways these coordinated events can occur and multiple
possible outcomes. In some cases, the original DNA structures
can be restored precisely. In other cases, mutations or
deletions occur at the site of joining broken ends of the same
molecule. When the ends of different molecules are joined
together, of course, various types of chromosome
rearrangements occur.[2]
There are two important evolutionary
consequences of chromosome changes. One consequence is the
potential effect of location at different chromosome positions
on the expression of coding regions. This can have a direct
input into physiological or morphological alterations. The
second evolutionary consequence is that chromosome differences
serve as a barrier to sexual reproduction between two recently
diverged species. Most mating events will be sterile without
genome doubling. In this way, chromosome changes separate the
reproductive destinies of species that may still be quite
similar to each other under normal conditions. At the same
time, as we shall discuss below, chromosome restructuring and
accumulated epigenetic changes set up the potential for
unleashing genome variability after interspecific matings
under crisis conditions (http://www.huffingtonpost.com/james-a-shapiro/epigenetics-iii-epigeneti_b_1683713.html).
4.13 Natural Genetic Engineeringas a new evolutionary
process
When we consider all the active cell
processes that alter DNA molecules discussed above (and they
are far from comprehensive), we have to conclude that living
cells possess all the tools they need to restructure their
genomes in any fashion compatible with the basic chemistry of
DNA. I have called this capacity “natural genetic engineering”
(NGE) because it is analogous to the kind of genetic
engineering we carry out in the laboratory or for
biotechnology (Shapiro 1992).
Like its human analogue, NGE is not
a random or accidental process. Each process makes predictable
types of changes in DNA molecules. For example, the movement
of a transposon or retrotransposon to a new location always
introduces the same constellation of coding, regulatory and
DNA restructuring signals. This kind of reproducibility is key
to cells having the ability to construct coordinated networks
rapidly throughout the genome. We know from genome sequence
data that this NGE capacity has been utilized in evolution (http://shapiro.bsd.uchicago.edu/Natural_genetic_engineering_and_multimolecular_networks.html).
Recognizing natural genetic
engineering as a cellular capability makes a radical change in
our understanding of genome variation in evolution. Rather
than a series of random, accidental, independent events, we
have the documented ability of cells to make non-random,
reproducible and concerted changes to the genome. This widens
our ability to think about the sources of biologically useful
novelty in evolution. Rather than selection slowly molding a
gradual process of “slight, successive modifications” (in
Darwin’s words) towards adaptive improvements, we can envisage
functional systems arising by a process analogous to human
engineering, where defined sets of components are assembled in
new combinations to accomplish novel tasks. This is not
deterministic because even the conscious, intelligent example
of human engineering occurs by trial and error, with the
failures typically outnumbering the successes.
In order to consider genome
variation as a serious candidate to be the major innovative
force in evolutionary change, we need to ask two questions:
(1) Is there a known evolutionary
system where novelty clearly results from natural genetic
engineering?
(2) Do cells have the ability to
control natural genetic engineering in ways that would be
necessary for concerted functional innovation?
The
answers to these questions are “Yes” and “Yes.” The questions
will help us define unknown issues that remain to be explored
in evolution science.
5. Evolved evolution by NGE: the
adaptive immune system
The adaptive immune system provides
a positive answer to the first of our two questions about NGE
and evolution. Adaptive immunity is a naturally evolved system
for rapid protein evolution. It solves the seemingly
intractable problem of how to generate a limitless variety of
properly structured antigen recognition molecules while using
only finite DNA coding capacity.
Lymphocytes teach us three important
lessons about evolution that are well outside orthodox
thinking:
(i) They demonstrate the virtually
unlimited creativity of NGE processes. In generating endless
pairs of antibody chains, B lymphocytes show that DNA change (http://shapiro.bsd.uchicago.edu/VDJ_joining.html) can be both targeted and highly
flexible simultaneously. Determinism and randomness are not
the only possible outcomes.
(ii) In switching selected
antibodies from one class to another by targeted exon swapping
(http://shapiro.bsd.uchicago.edu/Ig_Class_isotype_switching.html ), lymphocytes demonstrate that
cells can signal one another and direct the outcome of an NGE
process.
(iii) By undergoing a well-defined
series of steps before and after contact with the invading
antigen, lymphocytes illustrate how evolutionary NGE proceeds
in a stepwise fashion. The first step (VDJ joining) produces a
molecule that works at a basal level. The next step (http://shapiro.bsd.uchicago.edu/Somatic_hypermutation.html ) fine-tunes that initial product.
And the third step (class switching) adapts the fine-tuned
product to a particular application in the body.
If we keep the immune system example
in mind, we can distinguish parallel cases of sequential
multi-step NGE events with varying degrees of precision,
communication and control in a wide range of evolutionary
scenarios. If we think about the evolution of new body parts,
for example, we can trace their initial appearances,
subsequent refinements and further adaptations in the fossil
record (http://shapiro.bsd.uchicago.edu/Fossil_Record.html).
6. Regulating and targeting
natural genetic engineering functions: epigenetic and
ecological connections – a 21st Century
evolutionary scenario
A principle philosophical tenet of the neo-Darwinian
Modern Synthesis is that there can be no biological input to
the process of hereditary variation. Once we realize that this
variation results from dedicated cell functions, as detailed
in sections 3 and
4, our experience
with all aspects of cell biology teaches us that regulation
and control must play a central role in their operations. From
a molecular perspective, the idea of a segregated germ plasm
makes no sense (Weismann 1893). There is no reason to separate the
biochemistry of DNA molecules from other biochemical
functions.
6.1 Activating
and targeting natural genetic engineering
The answer to our second question at the end of section
4.13 is definitely
“Yes.” Cells can control and target NGE. The first way we can
document these assertions is to tabulate the many stimuli that
activate episodes of genome change (http://shapiro.bsd.uchicago.edu/TableII.7.shtml) and the many different ways that
NGE events are targeted within the genome (http://shapiro.bsd.uchicago.edu/TableII.11.shtml). With respect to cell NGE
targeting capabilities, it is important to note that they
utilize the same basic molecular interaction specificities as
other genome control functions: protein recognition of DNA
structure or sequence, DNA-DNA and RNA-DNA sequence pairing,
and protein-protein binding.[3]
6.2 The effects of hybridization
(crosses between different populations and different
species)
From an ecological, population and
evolutionary perspective, among the most important stimuli for
genome restructuring by NGE are mating events between
individuals from different populations or different species (http://shapiro.bsd.uchicago.edu/Hybrid_dysgenesis_interspecific_hybridization.html ). The work of Stebbins on abrupt
speciation by interspecific hybridization was mentioned above
(Stebbins 1951; Anderson 1954). His work is but one example of
many where interspecific hybridization leads to rapid changes
in genome structure and phenotype that generate new species of
plants or animals.
Within-species genome
destabilization goes by the name “hybrid dysgenesis.” Hybrid
dysgenesis has been documented in insects and mammals; it
reflects the way mobile elements establish themselves in the
genome of a species after horizontal transfer to the germline
of an individual, presumably by some kind of viral, symbiotic
or parasitic vector. Hybrid dysgenesis manifests itself when a
sperm introduces chromosomes carrying transposons,
retrotransposons or ERVs into an egg cell that lacks the
epigenetic controls to inhibit their expression. The results
of such an introduction are activation of that element and
high frequencies of transposition and chromosome restructuring
during germline development.
Sometimes the effects of hybrid
dysgenesis on germline development in the progeny are so
severe that the offspring are sterile. But often, dysgenic
germline development successfully produces gametes carrying a
dramatically restructured genome. Since these genome changes
occur before meiosis, hybrid dysgenic progeny can produce
multiple gametes carrying the same restructured chromosomes.
In species that produce multiple progeny from a single mating,
these gametes can lead to the formation of siblings capable of
interbreeding and founding a population with a new genome
architecture (Woodruff and Thompson 2002).
Interspecific hybridization has an analogous
destabilizing effect on the progeny genomes, altering
epigenetic modifications (http://shapiro.bsd.uchicago.edu/TableII.10.shtml), transcription patterns, splicing
patterns, and genome stability (http://shapiro.bsd.uchicago.edu/TableII.8.shtml). Sometimes these destabilization
effects last two or more generations. The precise molecular
basis of this destabilization is not as well defined as in
hybrid dysgenesis, but we can reasonably hypothesize that
species different in their epigenetic control systems (Brown and O'Neill 2010). According to this hypothesis, NGE
operators in chromosomes from the male parent would not be
properly regulated and would initiate an episode of genome
instability.
Interspecific hybridization as a trigger for genome
change is so important in evolution because it is a causal
link between failing species and episodes of genome
restructuring. Individuals will normally mate with members of
their own species. But when populations are depleted because
of some ecological disruption, they will not find normal
partners and consequently will mate with other species. This
feedback process promotes evolutionary innovation because the
hybrid progeny undergo germ-line destabilization for at least
one and often for several generations. In other words, the
ecological conditions that require genome renewal result in
precisely those sexual events that stimulate the needed
natural genetic engineering processes.
7. The genome as a RW memory
system within a cognitive cell
Evolution science, like the rest of biology, will have
to incorporate two fundamental new concepts:
(i)The genome is a RW memory device
subject to modification by the cell at all biological time
scales.
(ii)Cells are cognitive entities
operating on acquired information about external conditions
and internal operations.
The key novelty in thinking about
regulation by transient nucleoprotein complexes and epigenetic
modifications as part of a RW memory system is to realize that
they are more or less temporary inscriptions on the genome
which do not alter DNA sequence information. The genome
operates on three biological time scales:
7.1.1 Within the cell cycle, where
transient nucleoprotein complexes predominate.
As outlined in sections 3 and 4, the study of
changes in genome sequence information have revealed a whole
world of cell functions that restructure DNA molecules at all
levels, from single nucleotide substitutions to massive genome
rearrangements. As whole genome sequencing reveals, each new
taxon has its own characteristic genome features. In other
words, natural genetic engineering constitutes the active
process of genome rewriting in evolution.
7.2 The cognitive cell and the 21st
Century Research Agenda
The area where we have the most to
learn is in what I have called cellular cognition. How do
cells acquire and process information about external and
internal conditions, process that information, and make
decisions about the appropriate biochemical and biomechanical
operations to undertake? We know a lot about the different
kinds of molecules that comprise cell sensory receptors,
signal transduction pathways, and internal networks as well as
their biochemical and biophysical interactions. But we have
virtually no understanding of how these networks operate. What
logical and informatic principles do they use? How are
algorithmic computations performed? In terms of evolutionary
biology, how do these cell networks use the tools available
for regulating and targeting natural genetic engineering to
generate useful adaptations to new conditions?
These and many other questions need
answers in order to take biology forward in this new century.
I wish I had the answers. Nonetheless, as I used to explain to
my students, and as Lynn Margulis would surely agree, you know
you’re truly doing science when you have questions to ask but
can’t anticipate the answers. In that situation, you’re
guaranteed to learn something new.
[1]In terms of contemporary
public controversies, the Intelligent Design (ID)
advocates use the bacterial flagellum as their poster
child for an “irreducibly complex” structure that could
not have evolved by Darwinian evolutionary processes (Behe, M. (1996). Darwin's
Black Box: The Biochemical Challenge to Evolution Free
Press. .)(http://www.wesjones.com/darwin.htm). Both the ID and scientific evolution camps need to
address how the flagellum and related biological
inventions came to be diversified for so many different
uses. Certainly, the ID argument is greatly undermined if
it has to invoke supernatural intervention for the origin
of each modified adaptive structure. At the same time, it
is fair to recognize that evolutionary science faces the
challenge to provide a plausible account for the origin
and diversification of this intricate functional design
for moving large molecules across the bacterial envelope.
[2]Since, as we have seen,
genomes are replete with repetitive elements, it is also
possible to generate many chromosome rearrangements by
homologous recombination between repeats at different
genomic locations (http://shapiro.bsd.uchicago.edu/Chromosome_Rearrangements_by_Repeat_Recombination.html ). Sequencing of the rearranged
sites sometimes makes it possible to distinguish between
the different mechanisms.
[3]Although documented cases
have not yet occurred, I am confident we will eventually
find examples of NGE targeted by RNA-RNA pairing as well.
This appears to be the mechanism behind RNA targeting of
chromatin formatting.