Transposable elements or Controlling elements.
 

    We normally (and correctly) think of a gene as a very stable unit. Mutation rates are low (10^-6 to 10^-7 ) and the position of the gene on the chromosome is constant. That this is not always the case comes from studies of  mobile DNA elements  termed transposable elements (TE), or controlling elements or transposons, that have the ability to move throughout the genome. These elements have been found in virtually all organisms.

     Transposable elements were first found in maize by Barbara McClintock who was later awarded the Nobel Prize for this work.  She identified these elements due to a property that some - but not all - TEs have, namely causing mutability of normally very stable genes.  McClintock defined TEs as follows:  Transposable elements, of apparently sporadic occurrence, which make themselves visible through their effects on gene activities.  The element may move to a new site in which case gene action at the former position maybe  restored.

    Studies of genes and genomes continue to identify the significant impact of transposable elements on  living organisms.  Four fundamental properties of  our genetic material have been impacted by TEs.
 

•     25% of all mutants are caused by transposable elements   (insertions as well as footprints following excision).
•     Non-coding, repetitive  DNA separating coding genes, in many cases,  appears to be of  TE origin.  Moreover, sequences appear to be species specific.
•     TEs may be the origin of  at least some of the introns found in eukaryotic organisms.
•     TEs may duplicate and move genes to new places in the genome.
  

        First, some general comments about transposable elements:
 

  1) Some elements are associated with genetic instability while others are not.

 In the former case, the element at the gene leaves the locus and moves elsewhere (cut and paste).  Some, but not all of the exit events from the gene restore gene function and this is seen as restoration of gene function. These are seen as mutable genes.  The classic plant elements are examples.   In the latter cases mentioned above (copy and paste), a copy of the element at the gene is made and the copy goes to a new site.  The IS elements (insertional sequences) of E. coli are an example of that.

The latter (through RNA) elements are termed retrotransposons or retro elements.  The terms are coined after retroviruses which use a similar mechanism for survival.  The Ty element in yeast is an excellent example of this.  The classic plant elements are an example of the former class.

These can be inverted repeats or direct repeats. Note that both types of repeats can allow for pairing of the ends.  Elements can be identified and placed into families by their terminal sequences.   An exception to this  are elements that move through RNA.  In many cases, one finds an insertion ending in a poly A tail.  Also when one looks at members of the family one finds that the 5' terminus is variable.  Likely this is due to an incomplete RNA copy being made from the DNA template.  A newly-described transposable element, termed Helitron,  also lacks repeats on its ends.  There is no evidence to date that these move through RNA.  The evidence favors movement through DNA.

 Insertion of the TE usually involves a staggered cut in the DNA followed by DNA repair.  This results in a duplication of host sequence flanking the TE.  The size of the duplication is family specific.  For example, the Ac/Ds family in maize, see below,  makes an 8 bp duplication while the Spm family makes a 3 bp duplication..   Usually, but not always, there is no obvious preference for a particular sequence.  The newly-described Helitron does not make a duplication.

Usually some remnant of the duplication (termed a footprint)  is left behind or there is a partial deletion of adjacent host sequences.  Depending on  the position of the insertion in the gene (in the promoter, protein-coding exon, intron or terminator) some footprints allow normal or almost normal gene expression.  Introns, for example, are fairly tolerant of changes, especially in the middle of the intron, so virtually all excisions give rise to wildtype phenotype.  In contrast, there are huge constraints on  excision events in protein coding regions that restore enough gene function to condition "wildtype."  In some cases in which the insertion is in a motif critically important for protein function,  only perfect excision events (those leaving no footprints or any other type of alterations) give wildtype.  In one case, the footprint actually created an advantagous trait in the resulting gene.

 Besides the need for repeated sequences on their termini, DNA- based transposable elements  require an element-encoded protein for transposition.  While the exact role of the protein is not totally known, at least some of these proteins  -- termed transposases  - have been shown to recognize the termini of the element.  Because the protein acts in trans,  a particular element that has lost its ability to produce a transposase because of  mutation  can still transpose by the action of a transposase encoded by another member of the TE family.  Elements encoding a functional transposae are said to be autonomous since their movement is not dependent on another member of the family being present.  Non-autonomous elements refer to those requiring a second member of the family.

Excision events restoring gene function can, with some exceptions, occur at any time during growth.  Hence,  cells located together containing a reversion event most likely arose from a single event and those cells exhibiting the revertant phenotype trace back to a single progenitor cell.  This allows one to follow the path or fate of cell divisions.    For example, the anthocyanin pigments are found in the aleurone or outer layer of the endosperm.  Analysis of mutable alleles of genes involved in the pigment synthesis show that cell divisions are such that circles or spots of cells arise from a single cell.  In the pericarp or seed coat, the same type of analysis reveals stripes of pigmented cells.

   Non autonomous elements, particularly Ds  can become quite divergent, probably because  transposition to nearby sites and from sequences recognized for transposition.  Ds elements differ in size from 405 bp to 20,000 bp and show very little sequence similarity in the internal regions. The inverted terminal repeats and about 20 bp adjacent to one terminus of all Ds elements, however, are identical. One mechanism leading to such divergent Ds elements  -- which takes into consideration their propensity to transpose to nearby sites --   involves the recognition of the two external termini of the two elements in one transposition event.

The question of the true origin of all of the transposable elements is of the same magnitude as the question of the origin of life. Furthermore, the complexity of transposable elements, as eluded to above, raises the distinct possibility of multiple origins.  And, there exists good evidence for what is termed horizontal transmission  – the direct  passage of these elements from one species to another.  Hence, relatedness, as conventionally determined by DNA similarity, barriers to hybridization, characteristics of the organism, etc,  is not adequate for inferences concerning actual times of origins.

Because many of the elements are cloned, it is now fairly simple to ask whether sequences of a particular element are present within a genome.  This type of analysis reveals that there exists much variation within a species for the presence/absence/copy number of a particular element.  For example, a molecular probe for the termini of the Ac/Ds family of elements detects multiple copies of the element in the vast majority of corn lines.  Drosophila P element, on the other hand, appear to have originated in one area of the world and are in the process of spreading throughout the world

Early on, it became clear that some elements required a form of "genomic shock" for activation.  In other words, the elements were present, but inactive and hence not detectable by genetic analysis.  At least three quite different  forms of "genomic shock" have been reported in maize to activate elements.  The first report, from McClintock, involved chromosome breakage.  Here, she established  what is called the breakage - bridge - fusion cycle.  A chromosome undergoing this is broken at each anaphase division until the newly broken ends "heal".   This activated at least two families of transposable elements.

The second form of genomic shock was an atomic bomb blast.  During the early tests of these weapons, plants were situated at variable distances from the bomb blast.  The En system was activated by one such bomb.  Fortunately this experiment will not be repeated.   Finally, it is now known that inactive elements are activated by growth in tissue culture, followed by plant regeneration.

  Ac Activator, genetically defined as the element that activates Ds (Dissociation) to transpose or to break chromosomes.

Ds  Dissociation.  The non-autonomous second element of the Ac-Ds system.

Spm (En)  Suppressor-mutator,  Enhancer.  Discovered independently  by two investigators and hence still carries two names. Receptor element has various names.

Mu  Robertson's Mutator.   Named after Dr. Donald Robertson.  Mu is the preferred element for gene tagging because of its very high forward mutation rate.  Unfortunately, reversion events of the element leaving the locus usually occur very late in development and hence it is very difficult (but not impossible) to get germinal (heritable) reversion events. This is discussed in more detail below.

Dt  dotted .  Much like Ac-Ds  discussed above in that it is a two-element system.  Non-autonomous element does not really have a name.
 

    Rationale: In  modern studies of biological sciences, the relationships between a gene and  its biochemical function and the phenotype following loss of its function are becoming more and more common.   Historically, people have entered the "triangle" from basically two positions.  Some start with a mutant and wish to decipher the underlying biochemical lesion (forward genetics) while others start with a cloned gene and wish to know the phenotype of a knock-out mutant (reverse genetics).  Transposable elements have been used in both approaches.  In fact, because of the wealth of cloned genes now in databases and more refined experimental procedures,  it is now possible to start with a biochemical pathway and isolate genes and mutants in one step. Procedures for each of the three goals is detailed below.

  (a) Have mutant, want gene (forward genetics).
 

   This approach is used when one has a cloned transposable element whose sequences can be used to isolate the newly tagged mutant.  The procedure works best when the mutant is recessive, one has an easily discernible phenotype and one can easily generate lots of progeny. For example, in plants a seed phenotype is better than a plant phenotype, since progeny do not have to be grown to be scored.  It also works best with those organisms easily out-crossed (maize, Drosophila, animals, etc.)  One should also remember the formula:

               n =   log (1-probability of finding mutant)
                                log (1- mutation rate)

 where n is the number of progeny to score.
With a mutation rate of 10^-4  and a probability of 0.99, this turns out to be 46, 051 progeny.  With a mutation rate of 10^-5,  n is  460,510 progeny.  With a probability of .9, and the first mutation rate, it is approximately 23,000 progeny.

 The other, hopefully obvious point is that one wants as high a forward (wildtype to mutant) mutation rate as possible.  In the case of maize, the system, Robertson's Mu is the best.    A second and perhaps more important aspect of Mu for tagging is that it preferentially inserts into single copy DNA.  Since genes fall into the "single-copy DNA" category and since genes represent only a small percentage of the total DNA, this is a very significant advantage to Mu tagging.

                               Cross 1:   TE stock (M1/M1)   X  mutant  (m1-R/m1-R)

Step 2.  Select resulting progeny exhibiting mutant phenotype. Rationale: What one is looking for is an insertion into the gene of interest  abolishing gene function.  Such insertions are seen in the cross above as mutant progeny.  They are rare events. 

Step 3.   Cross selected mutant progeny with a wildtype stock lacking the TE.  Rationale:  From this point on, the goal is to associate, via co-segregation,  a TE-hybridizing DNA fragment with the newly-created mutant.  This step is necessary since there are multiple copies of the element in the genome.  Since the newly-created mutant and the starting mutant likely have the same phenotype, it is best to separate the two as quickly as possible.  A cross to wild type does this.  Each offspring from this cross will be heterozygous with a functional or wild type allele; however, it is impossible at this point to identify which mutant allele is contained in each heterozygote.   There are a number of ways to address this.  For example, if the starting mutant allele were closely linked to another recessive, morphologically-detectable mutant, selfing of the heterozygous seed from cross 2 would identify those families derived from the starting allele.  Closely-linked molecular markers (RFLPs, RAPDs, etc) are now available in many organisms and could be used for this purpose.  Below we will proceed with the worst case scenario, namely that we have no outside markers to aid in allele identification.

                            Cross 2:   selected mutant  X  wildtype lacking TE
                               (m1-R/m1-TE)     X   M1/M1     => M1/m1-R     &    M1/m1-TE

Step 4.  Cross each of five heterozygotes  from Cross 2 with  the starting mutant.

  a)   M1/?  X  m1-R/ m1-R   => M1/m1-R (wildtype phenotype)   &  ?/ m1-R (mutant phenotype)
  b)   M1/?  X  m1-R/ m1-R   => M1/m1-R (wildtype phenotype)   &  ?/ m1-R (mutant phenotype)
  c )   M1/?  X  m1-R/ m1-R   => M1/m1-R (wildtype phenotype)   &  ?/ m1-R (mutant phenotype)
  d )   M1/?  X  m1-R/ m1-R   => M1/m1-R (wildtype phenotype)   &  ?/ m1-R (mutant phenotype)
  e )   M1/?  X  m1-R/ m1-R   => M1/m1-R (wildtype phenotype)   &  ?/ m1-R (mutant phenotype)

    Rationale: As noted above, the heterozygote used in this cross will contain either the starting allele,  m1-R or the allele of interest, m1-TE.   The probability that not one of the  five heterozygotes contains m1-TE  is   (.5) ⁵ or 3.13%.  In other words, we have 97% chance that at least one of the resulting progenies in 4a through 4e contains m1-TE.

Step 5.  Grow representative individuals of each of the ten groups above, and extract DNA from pools of each class.  Rationale: Note that there are ten groups above.  In those progenies segregating for m1-TE, all mutant progeny will contain TE in a new position  (assuming the new mutant is due to insertion of the TE) while none of the wildtype progeny from the same family will contain the new band.  Families containing the m1-R allele will show no association of a new TE insertion with the mutant allele.

 DNA from several (the more, the better) individuals per class is used  (pooled DNA) to cancel out random associations of bands with the new mutant.  For example, if we randomly chose one wildtype progeny and one mutant progeny from a particular family, the chance of finding a genetically unlinked  DNA band in the mutant DNA prep but not in the wildtype prep is 1 in 4.  Pooling cancels out these associations.  Note that one can identify TE-hybridizing DNA fragments  linked but separate from the new mutant.  These fragments would be separable by recombination from the new mutant.  These, then, would be seen as bands of higher intensity in the mutants but present (and of lower intensity) in the wildtype DNA.

Step 6.  The 10 DNA preparations are digested with a restriction enzyme, electrophoresed, blotted and probed with the TE probe.  Band present in the mutant track but not present in the wildtype track from the same family is chosen for cloning.

Step 7.  Clone band.   Several approaches can be used here.  The simplest is to isolate DNA in the size range of the hybridizing fragment, synthesize a genomic library in a conventional lambda vector, probe the library with the TE and purify the hybridizing band.

 Step 8.    Isolate fragment(s)  abutting the TE element.   One should be able to isolate two junction fragments of the TE with the gene; however one will suffice.  Standard molecular techniques are used here.  (If you are not familiar with them, I would be happy to discuss these with you. LCH)

Step 9  Abutting fragments isolated in (8) are used to isolate wildtype sequences.

Step 10.  Prove cloned fragment is really the gene.    Several methods can be used. The best and most direct is to transform the original mutant with the cloned wildtype gene and obtain complementation.  Other approaches include:  probing  other, independently-derived mutants from step 1 and showing that the same fragments are altered or probing heritable revertants of the original mutant and showing that the wildtype pattern seen on Southern blots returns.  (This last approach is problematic with Robertson's Mu since its reversion rate is quite low. )
 

(b) Have gene, want mutant (reverse genetics).

    In lots of situations, an investigator has cloned a gene and wishes to know what happens when its function is lost.  This is sometimes called "reverse genetics".
Fortunately, the procedure here is much less involved than that discussed above.  It can be as simple as a few phone calls. Unfortunately some of these calls are with the university attorneys.

Rationale:  Given an actively mobile  TE element and given a population of sufficient size, every gene should contain at one insertion in the entire population. Such populations are now available in several organisms.  In the case of maize and Robertson's Mu,  investigators first at Pioneer Hi-Bred, later at Norvartis  and now at the University of Florida have generated populations of sufficient size to tag every gene.  Pioneer synthesized 50,000 maize families from various divergent sources but which all contained Mu.   The population has been screened for insertions into about 100 genes and the success rate has been found to be quite high.

    (Note that other forms of insertion can also be used as along as the sequence of the insertion is known.  A very large population of Arabidopsis plants containing T DNA insertions has been generated at several locations.  Several UF laboratories have taken advantage of the Salk collection.)

    Detection of a Mu insertion into a cloned gene turns out to be rather simple. Recall that Mu, like many TEs,  contains repeated sequences on its termini.  Here they are inverted and ~ 200 base pairs in length.  Because they are inverted, one PCR primer complementary to the terminus can synthesize DNA in both directions.

      Assume that you have DNA from a line with Mu inserted into your gene and another line with no such insertion.  If you added to this DNA a primer complementary to your gene and another primer complementary to a Mu end and perform a PCR reaction, a DNA product should be produced from the first line but not from the second line.   This is the basis to the approach.

   Fortunately 50,000 PCR reactions need not be run.  Empirically, it has been found that one can mix 300 samples and detect a single insertion in a single genome.  Hence, instead of running 50,000 samples, one need only perform 50,000/300 or 167 initial reactions.  Given identification of a particular pool containing the insertion, another trick is played to cut the work even further.    Note that  289 samples can be arranged in a grid 17 samples long and 17 samples wide.  If we arrange the samples like this and then make mixtures along both axes, any one sample is represented in two pools.  There are  17 + 17 or 34 pools and hence only 34 PCR reactions are needed to probe 289 samples.   Hence, one can screen a population of 50,000 families and identify the one family with the insertion in your gene of interest in only (167 + 34) or  201 PCR reactions.  This is basically a day or two job.

      There is another advantage to this.  Note that the PCR screen can detect the insertion in your gene of interest when the gene is heterozygous with another allele.  Hence mutants that are lethal in homozygous condition can be detected in the screen.

 The procedure here is simple.  One sends to Pioneer or Norvatis a primer for one's gene.  (Actually, Pioneer likes to do this with two primers, one from each end of the gene.)  The complicated part of this is getting an agreement between the seed company and your institution.  The goal of seed companies is to make money and they wish to tie up patenting/licensing aspects of any technology resulting from your subsequent work with the gene.

 Given agreement, seeds of the appropriate stock(s) are sent to you.

 Interestingly, less than 10% of the insertions into cloned genes so far isolated by Pioneer have clear, discernible phenotypes.  There are two classical explanations for this (1) the gene is so important that homozygotes die at a very early stage in development, or (2) there exists gene or functional redundancy such that removal of one copy is compensated for by other members of the family or by a redundant pathway.  One could also argue that the gene is simply uninportant.  Also quite interesting is the fact that the frequency of  gene insertions with obvious  mutant phenotypes is also low (<10%) in yeast.

(c) Have nothing, want it all, a.k.a. Don McCarty.

       Recent developments in PCR technology, and the availability of micro-arrays of cDNA clones and sequences of known and unknown genes in databases have opened up a new approach to studying function, genes and mutants.  Dr. Don McCarty, who will present the material in this course in developmental genetics is a leader in this area. 

    In this approach,  an inbred with an active transposable element is simply self pollinated and one looks for new mutants.  Mutants can be of any type.  This works well for people interested in a particular phenotype but not necessarily tied to a particular gene.  There was a project here where seed mutants were collected -- any mutant that altered the seed in any way.  In this experiment, approximately 7% of the ears segregated for a seed mutant.  Hence, there are lots of target genes; estimated to be around 300. 

    Each of these mutants then is a candidate for gene isolation, as described below. 

 The first development is termed TAIL-PCR.  Specific to tagging and more specifically to Mu-tagging, PCR reactions are run with a PCR primer complementary to the Mu terminus (as discussed above) and also with what is termed a promiscuous primer.  This primer has two important properties: (1) It is degenerate.  In other words, instead of being one sequence, it is a mixture of several but a finite number sequences.  (2) It has an AT content higher than the Mu primer.

 The PCR machine is programmed such that the first few rounds of amplification are done at higher temperatures of hybridization.  In doing so, only the Mu primer hybridizes and amplifies.  There is what is called linear amplification of Mu sequences.  This increases the number of templates for the second round of PCR.

 The second round of PCR is done at lower temperatures.  This allows the promiscuous primer to hybridize and one get exponential amplification.   The third step of PCR involves replacement of the first Mu primer with a second Mu primer internal or nested to the first one.  Again a round of high-temperature PCR followed by a second round of low-temperature PCR is performed.

    Another approach is to do first strand synthesis using the Mu primer, then place a run of dA's on the 3' end of the first strand.  Then oligo- dT is added to prime off the run of A's. 

 The two approaches each result in several to many amplified sequences termed amplicons.   There are several things one can do:
 

    (1) Sequence the amplicons and compare resulting sequences to sequences from other very closely-related Mu lines in the same inbred that are not exhibiting a mutant phenotype. This comparison quickly  identifies Mu insertions that are common to both lines and hence are not candidates for TE insertion into our gene.  Since the first line has the mutant we are following and the second line does not, we would focus on Mu inserts in the first line but not in the second line. Also, amplicon sequences can be compared to sequences in GenBank or comparable databases.    This is the most direct approach to the gene - clone - phenotype question.  Given all sequences in Genbank and all genes tagged, this should identify all mutants of all genes in maize. 

    The recent development of pryo-sequencing or 454 sequencing has greatly aided in this procedure.  This technique can generate several hundreds of thousands of sequence in one run.  While one sequence is only 100 bp in length, it is enough to identify junction sequences.  And with coding of primers,  massive mixing of PCR products from many mutants can be made and sequenced at one time. 

   (2) Do the TAIL-PCR analysis, starting with a mutant of interest tagged with Mu.  This then  replaces the  Have mutant, want gene  approach described above.  Here one needs the identified tagged mutant and a family segregating for the mutant and wildtype allele.   The amplicon hybridizing to a band in the mutant but not in segregants lacking the mutant allele likely is the gene.

(3) Pool amplicons and probe a micro-array of cDNA or other clones.  Hybridization occurs if Mu has inserted into one of the genes represented on the micro-array.  This approach has many advantages; the major one is that in one probing, one can identify a mutant for a cDNA.  The cDNA clone can be sequenced and it can be used to probe micro-arrays of amplicons (or pooled-by-griding amplicons) to quickly identify the plant with the insertion.  Note that Mu insertions into non-coding DNA   -- really of no interest to anybody  -- fall by the side in this type of analysis.

(4)  Pool amplicons and probe BAC genomic clones from homologous or closely related organism. Probing of BACs from related organism has the advantage that only conserved sequences (i.e. genes) will hybridize.

For your reading pleasure:
 

McClintock, B. (1984) Science 226, 792-801.  (McClintock's Nobel Prize acceptance presentation.  She emphasizes genomic shock as activating the elements in this article.)
 

Walbot, 1992  Annual Review of Plant Physiol & Plant Molecular Biology 43: 49-82.
   Transposons as tags.   ( A good discussion of how-to-clone with the elements.)

 Grant et al 1993. Molecular and General Genetics 241: 153 and references cited therein
     Molecular biology of Spm/En suppression (This and the article below discuss some of the complexities of the Spm/En system we didn't have time to cover.)

Menseen et al  1990.  EMBO Journal 9:  3051
     Interesting Spm/En responsive alleles.

Giroux et al  1994 Proc. Natl Acad Sci., 91:12150-154
       The element Ds acts as a perfect intron.   (A particularly clear example ( to quote W. Gilbert) of the creation of an intron. This is one of my favorites for all kinds of reasons!!)

Fedoroff.  1989.  Cell 56: 181-191 (A good review for the time.).
Gierl and Saedler.  1989. Annual Review of Genetics 23: 71-85. (A good review for the time.).

 McCarty, D. R., Settles, A. M., Suzuki, M., Tan, B. C., Latshaw, S., Porch, T., Robin, K., Baier, J., Avigne, W., Lai, J., Messing, J., Koch, K., and L. Curtis Hannah 2005   Steady-state transposon mutagenesis in inbred maize.  Plant J. 44: 52-61.  ( Gene isolation with maize Mu)

 Till, B.J., Reynolds, S.H., Greene, E.A. et al. (2003b) Large-scale discovery of induced point mutations with high-throughput TILLING. Genome Res. 13, 524–530  (TILLING)