Barton: Evolution

a summary of:

Evolution

2007 N.H. Barton et al. Cold Spring Harbor Press

Chapter 11:  Evolution of Developmental Programs

As you probably already know, the homeotic genes are responsible for determining the fate of a particular portion, or segment, of the body.  A homeotic mutation would be, for example, a leg forming in a place where an antenna should be in a fruit fly (Antennapedia).  Lots of such mutants, in fly and mouse, have been created.  In the fly, these genes are divided between two different chromosomes and these patches are called Antp-C and BX-C.  In humans, they are brought together into one patch called Hox.  Humans, then, have the same sorts of genes, arranged in the same way (anterior to posterior like the body parts they control) just like the fly genes.  In humans, however, there are four versions of this Hox patch.  So we can easily see that any mutation in the Hox genes could have tremendous morphological effects – it could totally rearrange the body plan – and the more Hox gene patches you have, the more complex a body plan you can create.  So the Hox genes would seem to be a major engine for the evolution of body plans. 

For example, one Hox gene, Ubx, determines quite neatly where the thoracic appendages will form.  Where Ubx is expressed, that appendage will be of the thoracic, locomotory type, and not a feeding appendage, and this is true across the variations in number of each type across the different crustacean phyla.  So changes in Ubx expression created entirely different body plans. 

But it’s not that easy with Ubx in insects.  Ubx marks the boundary between two thoracic segments: T2 (anterior; not expressed) and T3 (posterior; expressed).  And a defective Ubx gene will lead to a nifty homeotic mutant (see above).  But, unlike in crustaceans, this expression pattern never changes in nature throughout all the insects.  So you can make the gross evolution in the lab, but that’s not how evolution played out in nature.  For example, butterflies have four wings and flies have two, but the Ubx pattern doesn’t change.  But what Ubx does do is to mark the boundary between forewings and hindwings (which are different in butterflies) and between functional wings and halteres (the vestigial hindwings of flies).  So over time, the evolutionary changes in insect wing anatomy (e.g., hindwings evolving to become halteres) occurred because genes downstream of Ubx responded differently to Ubx.   Ubx expression stayed the same.

So the moral here is that one can’t find a phenomenon in the lab and then assume that it will explain the diversity of Life.  It may (as in the crustaceans) but there might be many other genes interacting with it.  The situation may be complex.  But it did finally get figured out.

————————————————-

Sticklebacks are sardine-sized fish with three spines on their backs and two on their bellies at least for marine species.  Over a period of 15,000 years, the fish wound up in inland seas at the end of the last ice age, and these turned into freshwater lakes.  Curiously, these freshwater sticklebacks lost their stickles to varying degrees.  Functionally, it seems that stickles made the fish just too big for their normal predators to swallow, and, if they did, the stickles hurt the soft mouth parts of the predators.  In freshwater, however, the main predators are dragonfly larvae, and the stickles actually help the larvae grab the slippery fish better.  This big of an evolution in morphology is rapid for just 15,000 years.  What changed in the fish genetically? 

When they crossed marine stickled fish with freshwater nonstickle fish, the F1 generation all have stickles.  When they crossed the F1 generation fish with each other, they got a 3:1 ratio of stickle:nonstickle.  So, of course, this points to a single gene with homozygous dominant in the marine fish.  In mice, there are three genes which control hindlimb, but not forelimb, development, which would be orthologous to pelvic development in the sticklebacks.  Through microsatellite mapping, they had identified the position of the gene responsible for the stickles in the crossing experiment above.  An ortholog to one of the three mouse genes, Pix1, was found to map to the same chromosome area.  So they found the single gene whose change in alleles drastically changed the morphology of the sticklebacks in such a short bit of evolutionary time.  Furthermore, the allele difference is not at the level of the amino acid sequence, but a difference in the promoters:  the dominant allele expresses in all the appropriate places in the body, but the recessive allele expresses everywhere except the pelvis.

There are several elements which make this story work well within an evolutionary paradigm.  A simple genetic change, a big morphological change upon which natural selection acts, and a shift in morphology between populations.  But consider also the human gene Ectodysplasin.  People with defective alleles have loss or reduction in teeth hair, sweat glands and eyebrows.  The ortholog to this gene in the sticklebacks controls the amount of body armor the fish have.  This also fits so well into the evolutionary paradigm:  fish and humans use the same basic gene to create very different, but somehow related, body parts.  As the body plan changes radically from fish to mammals and finally specifically into primates, the gene function adapts and gets used in new ways.

————————————————-

Teosinte is a much branched large cane-like grass that grows in Mexico.  It has a small seed head with 5-12 kernels instead of the 500 kernels of maize ears.  The seed coats are hard and indigestible for animals.   It is the ancestor of our modern maize plants, created by Native American breeders thousands of years ago, but it doesn’t look much like maize at all.  Plant geneticists have traced the breeding of this crop via genetic analysis.  It is an interesting example of how far morphological evolution can go with just a few genes changed, in this case by human hands.

Genetic analysis answered questions concerning the evolution of each part of the plants.  The change in ear morphology was determined to be due to approximately five genes by crossing teonsinte and maize and taking it to the F2 generation, where only 1 in 500 progeny produced heads that looked like either parent.  This is close to the expected 1/1024 that one would expect if there were 5 genes responsible (a 3:1 phenotypic ratio (1/4) for each gene, to the fifth power).  So the wild difference in the ears was due to just 5 genes.  Next the branching habit of the two plants is very different – a shrub versus a single tall cane.  The gene responsible, tb 1,  is expressed in the lateral growth nodes and inhibits branching.  The amino acid sequence encoded  is the same for the alleles of both plants; only the promoter sequences differ.  Maize expresses twice as much tb 1.  Next, a single allele of another gene, tga 1, is responsible for the hard fruit case. 

Corn cobs hang around a long time.  I can remember seeing ancient cobs on a long hike in Utah to an ancient Native American encampment.  It was strange being far back in the wild, with no other people, looking at pottery shards and old cobs about the size of hot dogs.  They have found cobs as old as 4400 years.  DNA analysis of these showed the presence of the modern tb 1 allele, but a modern allele for another gene, su 1, which allows for tortilla production, was not present.  There are many other genes with lesser effects that have maize-specific alleles compared to teosinte.  The thought is that their effects were revealed when the important mutations occurred with tb 1, tga 1 and su 1.  Then, the farmers could select the best plants in their eyes and these lesser genes could have their best alleles selected for.

Even though this was human selection, natural selection could operate the same way.  Is there really that much difference?  The alleles appear, they confer benefit or advantage, and the progeny multiply.  And just a few genes can make a big difference, which would allow other, lesser genes to evolve as well.

Chapter 12:  Generation of Mutations

There are many degrees of mutation.  Point mutations involve just a single base change in the DNA (each unit that makes up DNA, the G, A, T or C is a “base”).   Synonymous mutations change the codon to one that still codes for the same amino acid.  Synonymous mutations are always far more numerous than nonsynonymous mutations.  As well, the rate of synonymous mutations is pretty much the same across all genes, but the rate of nonsynonymous mutation is very different; some gene sequences are “conserved” over time and some are variable.  Thus the DNA sequence itself reveals natural selection: freedom to change when change has no effect on the organism, restrictions to change when change is seen at the level of the protein.  For example, a single base nonsynonymous change causes sickle-celled anemia.

Another interesting phenomenon is replication slippage.  If you have a sequence with repeats; e.g., GATGATGAT, then it is possible for the polymerase to replicate this sequence, and then come back and replicate a portion of it again, with part of the newly made strand binding to the first part of the GAT and then adding on two more GAT’s.  In this way, once a small patch of GAT’s gets started, it can easily expand to a much larger patch.  These entrepreneurial patches of repeats are called “microsatellites” and they are very common in the human genome.  That says that, even at the level of DNA, things just kind of go any way they want.  I know they don’t think, but there is a correspondence in terms of independence from God in the chance events of Nature and the choice events of Man.  The equivalent to microsatellites in our macroscopic world would be the inevitability of those Starbuck’s places popping up all over. 

Genes themselves also duplicate through much the same mechanism.  What do you do with an extra gene?   It can suffer through mutations as long as the original copy stays functional.  Over many generations, it now has a chance to evolve into a functional, but different, gene.  These clusters of gene families are very common in the genome.  The book cites fly histone gene clusters, but I can remember a colleague finding a whole cluster of virus resistance genes in a plant that each coded for resistance to a different virus.  Each gene in the cluster had mutated enough to have a new, slightly different role.  This sort of mechanism is an easy way to see how genes could evolve over time with no pressure from natural selection.  It is perhaps the main mechanism for gene evolution.

Chapter 15:  Random Genetic Drift

We always think of natural selection being the only driving force determining whether a particular mutation propagates in a population or is eliminated.  However, sheer randomness seems to play a very important role with many evolutionary events.  Of course, the vast bulk of the nongenic DNA (DNA that is not a gene) will not have any selective force for or against it because its presence makes no difference to the success of the organism.  Furthermore, since there are so many genes, changes in an individual gene very commonly have no significant effect on the fitness of the organism. 

In physics, there are so many molecules in a gram of gas that the random fluctuations cancel out and allow us to construct predictable gas laws.  In populations of organisms, there are far fewer individuals, so random fluctuations do not cancel out and the evolution of the population actually goes somewhere, even if all the mutations are under no selective pressure and just increase or decrease randomly.

Mathematical models incorporating replication within a population of stable size (# deaths = # births) with no selection come up with some surprising conclusions.  Since the chance of death of an individual before reproduction is 50%, and the same for any progeny it manages to produce, the chances for extinction of any mutation that arises in a gene in that individual are very high.  Thus, most alleles of a target gene in the model population are represented by only a few copies and then go extinct.  Tracking alleles in this population would show that one or a few alleles dominate for many generations and then also go extinct.  In this way, dominance can be achieved even without natural selection.  The smaller the population, the more this effect occurs.  When actual populations are used, the patterns predicted from the models hold true.

Another phenomenon predicted mathematically from these model systems is coalescence.  If, in general, populations grow more and more divergent as we move forward in time, then the large number of alleles will coalesce more and more as we move backwards in time, finding common ancestors, which finally coalesce into the ultimate single common ancestor for the entire population.  By using these models, the evolution of a population of HIV strains can be deduced and the putative ancestral sequence can be postulated. 

So once again, chance seems to play an important role in evolution, and models predicated on chance correctly predict the actual data seen in natural populations. 

Neutral Theory

Well, you’re supposed to have a bunch of mutations and then these get sorted out as to whether they go extinct or survive via natural selection.  However, when protein sequences started becoming available during the 1960’s, it just didn’t seem to work out that way.  Regardless of which organism it was in, a particular gene had the same rate of change.  For example, alpha-globin accumulates one amino acid change every 6 million years whether it’s in fish, birds or mammals.  However, there is a tremendous difference between different genes; some of them evolve 1000x faster than alpha-globin, but do this no matter which creature they are in.  When you start looking at non-genes, like introns or the DNA found between genes, then the mutation rate approaches that of random DNA.  To explain these differences, it appears that natural selection is indeed powerfully at work.  Only functional genes survive – most of the mutant versions tend be weeded out.  Non-gene DNA can accumulate lots of mutants however; there is no natural selection against it.

So the neutral theory would posit that, as long as the gene is still functional, mutations tend to accumulate at a steady rate, drifting with no particular direction.  And in a small population, we would expect that most of the variants would be eliminated at random.  As we saw in a previous post, individual mutations have only a fraction of a chance of being passed down in a population that is not increasing; there is only a 1/2 chance of getting the mutation to the next generation (of one individual) and then a 1/2 chance again of surviving to the next generation, and so on.  So the allele that survives is randomly chosen, not necessarily the most fit, but eventually comes to dominate the population.  Also, this means that all members of the current population will be descended from a single ancestor and share a common line of descent from that ancestor.  

From this, one would expect that in large populations, more mutants could hang out without being eliminated by this effect.  As it turns out, large populations do have more diversity, but not nearly as much as you would expect.  I was the first person to study the evolution of plant viruses using large amounts of straight sequence data from viral populations.  I expected to get lots of mutations as I passaged viruses in plants over one and a half years.  However, I found far, far fewer than anyone had expected, so few that I had difficulty working out the statistical analyses for the experiments.  It was at that time that I decided to find an easier subject with more data points!   One explanation for this is populational bottlenecks.  When a population is drastically reduced in numbers, it will behave like a much smaller population, even when it makes a tremendous comeback in numbers.  Viral populations do this all the time, as only a few virus particles move from one cell to colonize another cell, or move into the phloem to colonize the rest of the plant, or from one plant to another plant. 

So there you have it.  Darwin modified.  It’s not just natural selection that can drive evolution.  It can move rather randomly all on its own.

Chapter 16:  Population Structure

I’ll start by describing some terms and concepts.  In subsequent paragraphs, I’ll show the surprising implications of some mathematical models.  Finally, I’ll apply these models to the diversity of ideas and the need for Christian universities.

First, some terms and concepts.  In a mathematically ideal population, every individual member would have an equal chance of mating and interacting with every other member.  This is called a panmictic population.  Populations with barriers to this sort of interaction are said to be structured.  They consist of many demes or subdivisions of the population which have barriers separating themselves from the rest of the population.  These barriers may be complete or partial; in other words, there may be no gene flow or a little gene flow between the demes.  The demes may be sharply delineated (such as communities living on cliff compared to those in a neighboring flat area) or may have a transition zone or cline (such as foothills between a mountain and a valley).  Gene flow depends on how quickly the organism gets around.  Flightless grasshoppers were studied in Europe and it was calculated that it would take over a 100 generations for a gene to move just 200 meters through the metapopulation (a group of demes that interact with each other).  So for such organisms, there are local demes that can look very different from other demes. 

Much of population genetics is based on mathematical models, and these can make very interesting predictions.  As I noted in an earlier post on this book, populations tend to be much less diverse that you would expect just looking at the mutation rate.  When a mutation occurs, it has very little chance, mathematically, of becoming established in the population; it usually goes extinct quickly.  However, some lucky variants tend to take over the population for a certain number of generations.  This happens even in the absence of natural selection and the phenomenon is known as random genetic drift. 

However, if a population is divided into many demes that have little gene flow between them, the diversity of the population is preserved, being locked up in the local demes.  This also results in less drift in the population as a whole.  So increased diversity, but less history (none of the change of the guard as one variant dominates and then recedes, yielding to the next variant).  This mathematical model helps to explain the very low diversity seen in HIV infections, given the very high mutation rate of this virus.  The virus has a population structure comprising demes found in clusters of spleen cells.  However, these clusters tend to die out, destroying the deme structure.  In this way, the genetic drift seen in panmictic populations kicks in, with a succession of dominating variants and less diversity.

Getting philosophical, it does make you think about our society.  Thanks to TV, Top 40 music charts, and all the rest of the entertainment/communications media, our social deme structure has been largely dismantled.  One hundred years ago, there was an enormous difference growing up in the country versus in the city.  Today, I see almost no difference.  Louie Giglio has noted, as he took Passion events on international venues, that the youth culture is now the same the world over.  Population theory would predict that, with greater flow, overall diversity will decrease drastically and a succession of dominant variants will dance across the stage.   Of course, these are different phenomena, ecological vs. cultural, but it is interesting. 

One example is the lack of diversity between universities and the corresponding need to have distinctive universities, especially Christian universities.  The same culture tends to predominate at almost all universities.  It demands that all voices be given the same hearing.  That sounds great, but, of course, it is impossible.  The voices heard are those approved by the surrounding culture for the most part, modified by the influence of “university culture”.  University culture differs from the surrounding culture, but it is heavily influenced by it.  University culture is fairly uniform across all the universities.  The same coffeehouses, the same books, the same ideas.  A truly Christian university has a robust history to base itself on.  It is also based on a truly external, eternal source, and that is the ultimate outlier for increasing intellectual diversity.  Working at a Christian university myself, I think that many folks fear being different.  They fear saying “no” to dominant cultural variants traipsing across the stage at this point in time.  They fear exploring ideas that no one else in championing.  I think, though, that this is why we’re here as intelligent animals.  We are to move past our cluttering culture and successfully identify the things of God, whatever the cultural cost. 

Chapter 17:  Selection on Variation

One way to create a better product is to learn all you can about how it works and then rationally design a better variant.  However, with a very complex structure, such as an enzyme, many times it is faster and easier to make random mutations and then let some sort of selection process find the best of the random mutants.  This, then, mimics the natural selection process seen in Nature.  It is amazing at how efficient and productive this method is.  In this chapter, one of these case studies is presented.

In this case we have a ribozyme that binds a magnesium ion for its activity, but we seek to create one that binds calcium instead.  A ribozyme is an RNA that is able to perform a particular enzymatic function.  It does so not because it codes for a particular enzyme but because it folds to form a certain structure, the same way protein enzymes do.  In this example, the Tetrahymena ribozyme that cuts RNA (an RNase) is the focus.   The experiment starts by mutating the RNA and then taking the mutants through a selection process, in this case, their ability to bind and cut an RNA substrate.  Those that effectively do so stay bound to the substrate in the column.  Those that are not effective are eluted (washed) away.   The best mutants are cloned and mutagenized again, cycling through this through several “generations” until the very best structure is obtained. 

The “evolutionary process” of this experiment was similar to what is seen in Nature with the evolution of alleles in populations.  In this case, the “population” of RNA molecules was very large (10¹³ molecules) and the selection critierion very harsh.  Thus, the evolution was speeded up:  it took fewer than a dozen generations to obtain a ribozyme bound to calcium but had the same enzymatic activity as the wild type ribozyme bound to magnesium.  Along the way, different sequence variants dominated and then receded, and finally the population was dominated by certain variants which had obtained the peak fitness.  So, as from an earlier post, different random variants traipse across the stage, but, in this case, the very stringent selection conditions winnow out the very best.  In Nature, that doesn’t always happen: when natural selection is not powerful, genetic drift takes over and a series of variants continually appear and disappear.