How you can tell if you are a victim of abduction., page 5

Here's more- if I had more room, I could do it all at once!!

As our appreciation of nucleotide sequences and their information content increases, it is becoming increasingly difficult to find any sequences that are absolutely free of selective constraint – don’t try too hard to learn a hard and fast definition of this term.

We know of a very large number of nucleotide sequences that are non-coding yet play vitally important regulatory elements and are therefor under strong selective constraint – most are things you are already familiar with from this course and should not require much additional attention. They include: promoter elements, introns, and 3’ untranslated regions.

[Regarding introns: 1) The mouse and human introns for the T cell receptor complex genes were found to be very similar: over 70% identical. This implied that these introns must do something useful or else they would have been free to accumulate mutations without any constraints. Perhaps they too are important for regulation.

2) Another example of a set of genes that has been unambiguously shown to have introns that play a role in the regulation of gene expression are the alpha globin genes which have numerous Sp1 binding sites within their first introns.

3) Additionally, the introns might be important for exon shuffling. It has been hypothesized that exons correspond to a limited number of useful functional protein domains. Recombining these domains (much like recombining the words in our language) could give rise to proteins with different functions without having to “re-invent” a domain like a DNA binding site (or, using the previous analogy, without having to redefine a given word like “dog” every time you used it in a sentence).

4) Numerous other examples of sequences playing important regulatory roles that don’t fall neatly into any of the categories already described are also becoming known. For instance, Lin4 is a developmental control gene for C. elegans that does not code for a protein but does code for an RNA that binds to the mRNA of lin14 – interestingly, in its intron. XIST is a non-coding locus in the human genome but when its mRNA is made it binds to its parent chromosome and leads ultimately to the inactivation of both its own locus and the entire chromosome it is associated with – when it is not present on an X chromosome, Barr bodies cannot be formed.]

Pseudogenes and gene families

Gene duplication followed by the accumulation of mutations in one of the two copies has been a very important theme in the evolution of eukaryotic genomes. When genes are duplicated (often through a process of unequal crossing over – to be discussed in greater detail in Wednesday and Friday’s lectures) one of the two copies is usually no longer under selective constraint (only one copy of the gene is needed to carry out the job that the organism needs it to do) and is free to accumulate changes. Very often these mutations result in a non-functioning gene that eventually accumulates so many mutations that it is no longer recognizable as being related to the gene that is still under selective constraint. Occasionally however, the new mutations give rise to genes with subtly different and new functions and provide the organisms that possess them with a selective advantage.

Globin genes have been a classic example of just such an evolutionary history. Over a billion years ago the common ancestor of plants and animals all appear to have had a single globin gene. The plant globin went on to become the leghemoglobin used by plants to sequester oxygen today whereas the animal globin has become a complex family of alpha- and beta-like globin genes. Each of the two globin gene families in mammals has numerous relic pseudogenes as well as several different functional versions of globins that are expressed at different times in the course of the development of the organism.

Highly repetitive satellite DNA

Aside from the families of near-duplicate genes (such as the globin genes) and families of tandemly arranged, perfectly copied genes (such as tRNAs, rRNAs and histones), eukaryotic cells contain other regions of DNA that are generally referred to as repeated sequences or as repetitive DNA. Repetitive DNA falls into several different kinds of categories based largely on the manner in which the sequence is multiplied (copied) and the length of the repeating unit itself. Lengths of the repeated sequence can be just a few nucleotides (as in CA repeats) or a few hundred nucleotides (as in Alu repeats) to several thousand nucleotides (as in L1 repeats).

Isolation

The existence (and copious quantity of) repetitious DNA in mammals was first recognized when their denatured DNA was observed to renature very quickly in reassociation experiments. In these experiments a mammal’s total genomic DNA is randomly sheared into fragments approximately 1,000 nucleotides long. These double-stranded fragments are dissociated into single strands and then allowed to base pair again. Strands containing a repeated sequence would find a complementary partner to base pair with more quickly that strands without repeated sequences.

Two factors affect these reassociation kinetics: the initial concentration (C0) and time (t). A convenient term for comparisons of these reassociation rates is the C0t1/2 value – the C0t that allows one half of the DNA to become double stranded.

When these types of experiments were performed on a typical mammalian genome such as our own, about 10 to 15% of the DNA reassociates immediately, another 25 to 40% reassociates at an intermediate rate and the remaining 50 to 60% reassociated at a much slower rate (and is “single copy” DNA – most genes would fall in to this category).

Micro- and mini-satellites

The very first sequences to reassociate in C0t experiments are often found as part of micro- or mini-satellites. The term “micro-satellite” is commonly used to describe sequences with little information content composed of repeating units that are as short as one or two nucleotides in length and as long as seven nucleotides in length. The most abundant, stable and highly polymorphic microsatellites, for instance, are tandem copies of the di-nucleotide (CA)n which are present at between 50,000 and 100,000 copies in the human genome.

In addition to mammals, microsatellites have been found to be widely distributed in the genomes of several other “complex” eukaryotes including frogs (Xenopus), zebrafish and plants. In the past few years a number of microsatellite repeats have been found to be associated with SINEs and these SINEs have been postulated to play a role in the de novo creation of microsatellites (to be discussed shortly – in 1995, Deininger and Batzer reported that the microsatellites actually tend to spring up where the SINEs have inserted). Regions of the genome rich in homopurine or homopyrimidine nucleotides such as those associated with human Alu elements (the predominant SINE in primates) appear to be ideal hotspots for the genesis of microsatellite repeats.

The term “mini-satellite” has been arbitrarily used to describe sequences whose repeat units that are longer than seven base pairs and are typically in the size range of 15 to 30 nucleotides. These sequences are obviously not complex enough to encode any useful proteins and the number of copies of the repeated unit are typically variable from one individual to another (hence their use in DNA profiling settings) though they rarely exceed a couple thousand nucleotides in length. Such repeats are quite common in the sequence of the human genome and occur with a frequency of about once per 10,000 base pairs. It has recently (1994) been speculated that mini-satellites are the evolutionary progeny of micro-satellites (Wright, 1994).

This type of DNA was originally termed “satellite” DNA because some varieties of simple sequence DNA separates from the majority of DNA during equilibrium centrifugation in CsCl.

The high degree of polymorphism in the size and structure of these repeats have made them important tools in population studies, genetic linkage mapping and forensic identity testing. All the mechanisms by which these micro- and mini-satellites become so polymorphic may not have been elucidated yet but strand slippage during DNA replication (also defined as slipped-strand mispairing) and intra-allelic recombination have been shown to be at least partly responsible for the variability in the length and number of these repeats (Levinson and Gutman, 1987; Scholtterer and Tautz, 1992).

SINEs and LINEs

The portions of the genome that reassociate at an intermediate rate correspond to sequences present in multiple copies within the genome but which have longer stretches of sequences associated with the repeat units. These repetitive elements are typically broken in to two categories: short interspersed repetitive elements (SINEs) and long interspersed repetitive elements (LINEs).

Between 10 and 15% of mammalian genomes come from members of the SINE families of repeats and the repeats are usually distinctive of the mammalian order in which they are found. For instance, primates all have about half a million copies of a SINE called an Alu repeat per haploid genome while lagomorphs (rabbits) all have about one million copies of a repeat known as a C repeat. The members of both of these SINE families (and those of other mammalian orders) are about 300 nucleotides in length and can be actively transcribed by RNA polymerase III.

SINEs like Alu repeats can be divided into various subfamilies of distinct genetic ages based on the presence of commonly shared diagnostic mutations. Several groups have independently identified a series of subfamilies on the basis of these diagnotsic mutations. Although the vast majority of Alu repeats have amplified over the past 55 to 65 miliion years, there is evidence which indicates that retroposition of new Alu elements is still an active process. Some Alu insertions are so recent in human evolution that they are not fixed in human populations.

A total of 59 out of 345 (17%) of the (CA)n repeats that Deininger and Batzer studied were found to be in immediate association with Alu elements. Approximately 75% (44) of these Alu elements had (CA)n repeats as a part of their 3’ A-rich tails, approximately 24% (14) had them as a part of their middle A-rich region and a single Alu had the microsatellite repeat at its 5’ end. The position of the (CA)n repeats relative to Alu repeats is not random. Interestingly, the majority of the Alu elements identified in this search belong to the older subfamilies.

About 5% of mammalian genomes comes from the sequence of members of the LINE class of repeats. All mammalian orders have the same LINE, a repeat known as an L1 repeat. Full length L1s are typically about 7,000 nucleotides long and have at least two long open reading frames. They can be transcribed by RNA polymerase II and are likely to be a “captured” or “stranded” retrovirus.

Retrotransposition

Both SINEs and LINEs appear to propagate by a mechanism known as retrotransposition. They are first transcribed into an RNA molecule, then processed, then reverse-transcribed (made from RNA back in to DNA) and then reinserted into the genome at a distant site. Interestingly, the SINE families of repeats all bring with them all the sequences necessary for their own subsequent transcription (again, because RNA polymerase III’s promoters are internal to the transcript itself).

Evolutionary history

None the less, it appears as if there are only a limited number of both L1 and SINE repeats that can give rise to additional copies of themselves.

Alu repeats are mobile genetic elements in primates and provide a potentially good phylogenetic marker as they are highly stable once inserted. Furthermore, independent Alu insertions at the same chromosomal locations are extremely unlikely. Batzer et al. used four human specific Alu insertions to test the ‘out of Africa’ hypothesis of human origin. These Alu repeats have recently retrotransposed and are not yet fixed in humans, allowing screening for the presence or absence of the insertions. Distance-based maxium-likelihood trees divided 16 populations from around the world into four major groups: Africa, Europe, Asia/Americas and Australia/New Guinea. Human-specific Alu insertions provide the advantage that the ancestral state is known to be the absence of the insertion. Therefore a tree based on insertion polymorphism can be rooted without additional information or assumptions. The authors found the root to be placed within the African branch, suggesting an African origin of the polymorphic Alu insertions. (PNAS, 91:12288-12292).

Possible roles

There has been much speculation about the possible roles of these repeats. In many ways they appear to be the epitome of junk DNA in that they do not appear to be under selective constraint and appear to be dispensable. They have even been called the ultimate parasites in that they appear to be best at simply making copies of themselves without conferring any clear advantage to the organisms in which they are found.

That sort of thinking about SINEs at least has been changing in the last decade or so. They have been implicated as occasionally playing important roles in polyadenylation (by providing functioning polyadenylation signals to the genes into which they insert), possible sites that act as origins of replication, break points to homologous recombination, starting points for non-homologous recombination, “tire-patches” involved in the repair of DNA breaks, as transposable protein domains, and even simply by introducing variability to the genome upon which natural selection can act (for instance, some individuals have hemophilia B due to the insertion of a human-specific Alu subfamily member within the coding region of the clotting factor IX gene). Still a large fraction (probably greater than 50%) of SINE sequences are probably disposable.