Genes are found on chromosomes.
Rule: Gene action is usually independent of chromosomal location.
Exception 1: Not always true, there are position effects, where gene regulation is affected by the chromosomal location. Especially true in developmental genes. For example, Hox genes, which are found in clusters in linear orders that correspond to the order of the segments of the body that they control.But, in general, the many genetic engineering successes indicates that there is little effect of chromosomal location, providing the gene and its cis-acting regulatory machinery are transplanted together.
Exception 2: Tight linkage of several genes together may also influence evolution. With tight linkage, epistasis or certain forms of migration, this may lead to linkage disequilibrium, one of the ways in which genes do not act independently.But again, this is a relatively trivial effect, which has little effect in the absence of strong epistasis like that in Papilio. Significant linkage disequilibria, and therefore dependencies between genes, are present over very tiny distances in most eukaryotes, say 1 Mb in humans, and 100 Kb in Drosophila melanogaster. Both of these distances are essentially intragenic, or perhaps rarely may affect adjacent genes.
However, the fact that genes are on chromosomes influences evolution far beyond the minor effects of position effects and linkage disequilibria. Because the genes are arranged on long strings, and because chromosomes themselves act as genetic elements:-
There may be
holistic selective effects that act on 100s to 1000s of genes at a time.
For a basic understanding
of chromosome structure and function, as well as rearrangements and how
they come about, I strongly recommend chapters 11-12 in Hartwell et al.
Although we understand some of the processes involved in chromosomal evolution, we understand by no means all of them.
Chromosome number, for example, varies enormously from organism to organism.
In Drosophila melanogaster, there are only 4 pairs of chromosomes (n = 4, 2n = 8). Of these, one pair is a microchromosome which has hardly any genes on it; one pair is a sex chromosome, which leaves only 2 active pairs of autosomes.What explains these patterns? It is not entirely clear. There is usually about one chiasma (causing a crossover) per chromosome arm; perhaps, therefore, chromosome number (like sex) is an adaptation which affects the general level of recombination in the genome. Many chromosomes means lots of of recombination (50% recombination between chromosomes, plus a lot of chiasmata); few chromosomes means little recombination.
In humans, there are 23 pairs (n = 23, 2n = 46) chromosomes. Mammals in general are highly variable in chromosome number.
Across the whole Lepidoptera, a group of similar age to the mammals, there is quite a bit of variability, but there is a strong modal number which many species in different groups of Lepidoptera actually have, of n = 31.
In any case, it is worth exploring some of the stuff we do know about chromosomes and their evolution.
But first, we have to learn a bit of terminology,
I am afraid!
Instead of genotypes, chromosomes have karyotypes. Typically, the karyotype refers to the number of chromosomes in a set, for instance, "the human karyotype is 2n = 46".
[This topic fits here, but also in the lecture on speciation ...
Polyploidy. A common kind of chromosomal mutation involves a doubling of numbers of copies, or polyploidy, and comes in two flavours, autopolyploidy (doubling of endogenous chromosomes), and allopolyploidy. Sometimes a single chromosome pair adds more copies (aneuploidy), but this is rare, because it leads to unbalanced gene dosages, causing developmental defects and sterility.
Autopolyploidy and allopolyploidy seem particularly popular in flowering plants (an estimated 30% of flowering plant species are of polyploid origin). Perhaps because plants can self; so an autopolyploid branch or a new and rare allopolyploid form can mate with itself, forming offspring with fully balanced gametes. If a polyploid mates with a normal diploid, the F1 hybrid is triploid; this causes massive amounts of aneuploidy in the offspring, leading to almost invariable sterility of hybrid offspring.]
By chromosome breakage. Can occur via radiation, mutagens etc. Repeated sequences, often transposable elements, in the DNA may frequently be involved, i.e. non-homologous recombination
e.g. P- elements are involved in chromosomal mutation in Drosophila.Breakage leads to "sticky ends", ? something to do with the lack of a telomere?
Alu elements probably do in mammals; perhaps in us?
General rule: Heterozygous rearrangements often lead to the production, in meiosis, of UNBALANCED GAMETES; often,
e.g. Paracentric inversions
If no crossing over in inversion: gametes
If crossing over in inversion,
dicentric bridge (breaks at cell division)Because a paracentric inversion heterozygote produces unbalanced gametes, the rearrangements cause:
acentric fragment (lacks centromere, becomes lost)
heterozygote disadvantage -> fixation (overhead).Because the deleterious effects act only if there is a crossover within the inversion, the inversion acts as a crossover suppressor: any crossovers that are produced become involved in unbalanced gametes.
(See drawings of Chironomus - Drosophila is similar)
However, no heterozygous disadvantage in Drosophila and many other Diptera (flies). Here, paracentric inversions are common, even within populations.
In males, there is no crossing over, so no unbalanced gametes.e.g. Pericentric inversions
In females, dicentric fragments are preferentially shunted into polar bodies. Mainly balanced gametes are found in the eggs. Thus, again, no heterozygote disadvantage.
Like paracentric inversions, only worse.e.g. Reciprocal translocations
(See overhead of meiosis in translocation heterozygote)
If no crossing over: approximately 50% (or more) unbalanced gametes due to non-disjunction, or non separation of homologous chromosomesBecause paracentric inversion and translocations heterozygotes produces such unbalanced gametes, the rearrangements cause heterozygote disadvantage.
If crossing over: similar problems
Translocations are a common rearrangement in mammals. Usually, however, populations are fixed for a translocation, as expected from the deleterious effects on gametes. Populations fixed for alternative chromosomal rearrangements are often called chromosomal races. They are common in species such as the european house mouse Mus musculus domesticus. See also, overhead of the Mexican pocket mouse Perognathus goldmani, showing chromosomal races.
However, non-disjunction rates are often
low, 0% - 15%, not usually the approx. 50% that simple theory would lead
one to expect. Mammals, like Drosophila, appear to have some mechanisms
which reduces the production of unbalanced gametes.
As we have mentioned, many Diptera do not suffer from unbalanced gametes when heterozygous for paracentric inversions, leading to a lack of heterozygous disadvantage. As expected, paracentric inversions are a common polymorphism in Diptera.
Strangely, there is even evidence for HETEROZYGOUS ADVANTAGE, which, as we have seen, will maintain polymorphisms. This is found in flies such as Drosophila, and also the malaria carrier, Anopheles mosquitoes.
Why should this be? Dobzhansky in 1930s suggested that inversions trap "coadapted gene complexes" his word for groups of genes that interact epistatically in a positive way, having been built up by selection. The crossover suppression prevents these advantageous genes from being broken up, so that they stay together in linkage disequilibrium. Dobzhansky showed cyclical fluctuations of chromosomal polymorphisms with the seasons, and also that different forms were favoured at different altitudes in the mountains.
Today we are not so sure. For one thing,
this explanation does not explain why there should be heterozygous advantage.
Even in captivity, there is enormously greater death rate of homozygotes
than heterozygotes In some cases (Anopheles quadrimaculatus) chromosomal
homozygotes are virtually lethal. This suggests that there are balanced
deleterious recessive genes on each chromosome. In other words, the
lack of recombination in some way allows a pathological accumulation of
deleterious genes, different on each inversion morph, which cannot be exchanged
with fitter alleles because of crossover suppression. The deleterious equivalent
of Müller's ratchet is operating because of a lack of recombination.
Many chromosomes are morphologically differentiated enough, if stained appropriately, that chromosomes or parts of chromosomes can be clearly identified.
In humans/apes, chromosome banding patterns first showed that chimps are more closely related to humans than gorillas.
There is an example of a phylogeny based on multiple translocations in the overhead for Perognathus goldmani.
For Drosophila, phylogenies based on chromosmal inversions visualised via polytene chromosomes are very commonly used. Here's how:
(picture of ABCDEF, inversion phylogeny of 3 morphs)
However, the DIRECTION of evolution cannot
be inferred from chromosomal phylogeny, meaning the tree cannot be rooted.
For that, we need to find a good estimate of the primitive state. This
is sometimes done by finding an OUTGROUP, which is close to the primitive
pattern; sometimes by assuming that a particular form is primitive based
on distributional or other data.
This is the phenomenon by which similar repeated change seem to happen in many chromosomes at once. The phenomenon is very interesting, but has not been fully explained.
For example, the primitive chromosome number of chromosomes in Mus musculus domesticus, the house mouse, is 2n = 40, all acrocentrics. However, by a series of Robertsonian fusions, there are multiple chromosomal races, some of which have as few as 2n = 22. In many cases, there are races with similar chromosomal numbers, but where the different races have undergone completely different fusions. Nobody knows why.
In mammals in general, reciprocal translocations (including chromosome number changing Robetsonians) are very common. Again, we don't know why.
In Drosophila and other Diptera, we probably DO understand why paracentric inversions are so common.
In the Australian grasshopper called Caledia
captiva, strangely, the situation is reversed from that in Mus.
The primitive karyotype is all metacentrics; but in two chromosmal races,
this has evolved into all acrocentrics.
Factors which may prevent evolution of chromosome rearrangements
The generality of heterozygous disadvantage caused by chromosomal rearrangements suggests that mostpopulations will be fixed for chromosomal rearrangements. In general, this is true.
However, there are frequent polymorphisms. We have already discussed the exception formed by the paracentric inversions of Diptera.Often, non-disjunction rates are surprisingly low; for instance, for Mus, they can be a few percent, when a heterozygote would be expected, on theoretical grounds alone, to produce about 50% non-disjunction. Presumably selective abortion of unfit progeny, or resorption of unbalanced gametes, is to blame forthis. This means that the evolution of rearrangements will be easier than expected.
However, there should still be some heterozygous disadvantage, leading to fixation. This can cause a partial barrier between populations that are fixed for different rearrangements. Species are often characterised by an absence of hybrids, by hybrid inviability, or sterility of hybrids.The barriers between populations are therefore similar to barriers between species, and it may be that these barriers are important in speciation.It is somewhat controversial that chromosomal evolution CAUSES speciation on its own, an idea promoted by MJD White, Guy Bush, and others. But it is clear that chromosomal rearrangements do contribute the the isolation between species, because species often differ chromosomally.
For example, humans have a chromosome number of 2n = 46, whereas chimps have 2n = 48. Humans differ from their closest relatives by 9 pericentric (I think) inversions and one centric fusion. Clearly, even if a hybrid between humans and chimps were possible (I don't know, but it probably is possible, judging from the ability of mammals of about that old to hybridize), the hybrids would almost certainly be very infertile, based on chromosomal problems alone, because they would be heterozygous for so many different rearrangements.Possible advantages to chromosomal mutants
Are there any advantages to chromosomal rearrangements? As we mentioned at the start, there may be position effects - cis-acting effects which change gene regulation.
There may also be advantages due to reductions
or increases of recombination; again we know little about these.
If they are generally at heterozygote disadvantage, and fixed within populations, how on earth do rearrangements evolve?
Genetic drift seems a likely starter to chromosomal evolution.
[ IF TIME (PROBABLY NOT!!): The "shifting balance"; see also later in courseHowever, there is much controversy about this mode of evolution. Many would argue that natural selection is more important as a cause of chromosomal evolution than genetic drift.
MJD White suggested that they might evolve by a process he called "stasipatric speciation". It is now realized that instant speciation via chromosomal evolution is unlikely, except in polyploidy.
However, White's mechanism is an example of a general mode of evolution, called by Sewall Wright the "shifting balance". In which drift interacts with selection to cause the evolution of a new adaptive peak. The "shifting balance" is divided into three phases:
For heterozygous disadvantage it is easy to see what the adaptive peaks represent. Each fixation is stable, and advantageous. Any chromosomal rearrangement will at first (because of the Hardy-Weinberg law) be found mainly in heterozygotes. Since the heterozygotes are at a disadvantage, there will be selection against them, and the fixations will thus be stable.
Phase I: Local genetic drift. However, if selection can't do it, it is possible that drift in small, local populations may sometimes occur to overcome the often weak effects of non-disjunction to lead to a temporary rise in the frequency of a new chromosomal mutation.
Phase II: Local mass selection. Once the frequency reaches the "saddlepoint" or unstable equilibrium given, as we have seen by peq=t/(s+t), selection pushes to the new peak.
Phase III: Interdemic selection. If the local area is fixed for the new morph, and if that morph is a successful experiment, then the area becomes stable to invasion. The new and old forms interact in a cline of frequency of the rearrangement, or a hybrid zone. If the new form is MORE successful than the old form, this implies that reproductive rate or migration rates (or both) from the area outweigh the input from the old form, so that the new form spreads outwards behind a moving cline. This process of population advance is called "interdemic selection" because whole populations (demes) are competing to produce an outcome that is not possible, except due to random factors, within populations.
FUTUYMA, DJ 1998.
Evolutionary Biology. Chapter 10 (pp. 286-294); for shifting balance:
Chapter 14 (pp. 408-409).
For introduction to chromosome structure and function:
L. HARTWELL, L. HOOD, M.L. GOLDBERG, L.M. SILVER, R.C. VERES, A. REYNOLDS (2003) Genetics: From Genes to Genomes. McGraw-Hill, chapters. 12 & 13.
I like this latter text
because it points up the holes in our knowledge so far. They say
things like: ".. this large variety of [non-histone chromosomal]
proteins fulfils many different functions, only a few of which have
been defined to date"; "despite bits and pieces of experimental
evidence, study that directly confirm the radial loop-scaffold model
have not yet been completed"; "[G-]banding patterns are highly
reproducible, but noone knows for sure what they represent". In
other words, this text is highly unusual in pointing out areas we still
do not understand, rather than covering up our ignorance.