Monday, March 19, 2012

Genomic head to head telomeric sequences detection by one primer PCR

Dicentric chromosomes are the signature of recombination molecular events which may imply telomeric motifs. To observe dicentric chromosomes in senescent cells, a cell culture  followed by an enrichment step is performed to get enough metaphasic cells. However metaphasic cells become difficult  to get as the cell culture reaches senescence or if cells differenciate, thus most of  fusions between short telomeres, when they occur, remain hidden in interphasic cells.
Telomeric fusions can generate head to head TTAGGG motifs as:

Head to head telomeric sequences (blue) at a fusion point (several possibilities), flanked by proximal (orange, green) and distal (purple) subtelomeric sequences.

 Such kind of sequence is known in the human genome at the 2q13-q14 locus (Ijdo et al.).

>gi|180516|gb|M73018.1|HUMCHR2A Human ancestral telomeric fusion DNA sequence

The fusion point of this sequence is:
It is the cicatrix of an ancestral chromosomic fusion, these motifs are short and degenerated so they cannot be detected by QFISH. Using Polymerase Chain Reaction and only one TTAGGG primer, it is possible to detect head-to-head telomeric resulting from chromosomal fusion (Jean-Patrick Pommier, unpublished data, 1998-1999).
Detection of a head-to-head telomeric sequence at a genomic locus by PCR amplification with one (TTAGGG)n primer (Dark Blue big arrow)

Material and method:
The  cycle conditions of the first PCR assay were inspired by the paper of Ijdo et al. . As no PCR product was detected with these conditions, the stringency conditions were relaxed by decreasing the annealing temperature, without my lab-book, I cannot bring more precision on the cycling conditions.
Two kinds of assay were performed: G-PCR with (TTAGGG)n primers (n=3 or 5) or C-PCR with (CCCTAA)n primers (n=3 or 5) , both with 100ng of genomic DNA (high molecular weight placenta DNA, primary fibroblasts, SV40 transformed fibroblast pre-crisis/post-crisis, wild type S. cerevisiae genomic DNA was also tested). In a second assay, both G an C primers were used at high (100 nM?) or low (20 nM?) concentration. G3-primer and C3-primer had 5' non telomeric sequence allowing a second PCR round with a non telomeric primer.
In an assay, prior PCR, 1 microgramme of genomic  DNA was digested by Hpa II (methylation sensitive) or Msp I restriction endonuclease (NEB) targeting GGCC or by an other endonuclease targeting a TTAA restriction site. A 100ng aliquot was then submitted to G-PCR.
PCR products were analysed in a 2-3% agarose gel electrophoresis, stained with ethidium bromide, with a molecular weight standard in the range 200 to 2000 bp approximatly (possibly molecular weight marker IV, Roche?).

Genomic DNA from placenta, primary fibroblasts, precrisis SV40 transformed fibroblasts submitted to G5-PCR (wells n°6,8,10) yield a smear like PCR product with possibly discrete bands. The same genomic DNA submitted to C5-PCR yield a smears as PCR product with no bands.
From left to right: Wells n°2, 13: molecular weight markers. Wells n°6, 8, 10: G5-PCR with (TTAGGG)5 primer. Wells n°7, 9, 11: C5-PCR with (CCCTAA)5 primer. Well n°12:?. The wells n°3, 4 may be some control (no primer, no genomic DNA).

More conditions were tested in the following assay:
C5-PCR or C3-PCR yield a smear as PCR product at high concentration. The PCR product becomes undetectable at low primer concentration. I performed C5-PCR or C3-PCR as a control, I didn't expect to detect a PCR product with a (CCCTAA)n primer due to the head to head orientation of the telomeric motifs in a telomeric fusion. Queue-to-queue telomeric motifs may occur, but those PCR products are likely PCR artifacts.
More interestingly, G5-PCR and G3-PCR yield bands-like PCR products with human DNA in all conditions:
  • Bands are observed with human DNA as PCR target, whereas a smear is observed with yeast DNA, indicating that the PCR product detect a specific structure in the human genome which is not present in the yeast genome.
  • With the "eyes of faith" (les yeux de la foi), G5-PCR at hight concentration primer, seems to distinguish postcrisis DNA (well 8, upper) from normal or precrisis DNA (wells 6,7 upper) by an relative increase of the amount of the short band. As the SV40 transformed fibroblasts reach the crisis, the frequency of dicentric chromosomes increase and become stable when an immortal clone emerge from the crisis. It is consistent with an accumulation of head-to-head telomeric motif generated by chromosomal fusion between short telomeres.
  • The shortest band (~600 bp) vanishes from the PCR product as the G5 primer concentration is diminished regardless the source of DNA (primary cells, precrisis or postcrisis cells). There are possibly several ancestral HTH telomeric motifs per cell, whereas there is less than one neo HTH telomeric motif per cell (statistically around the crisis, cells have 0 or 1 dicentric chromosomes, even some cells have more). As neo HTH telomeric motifs are rare events, decreasing G5 primer would decrease PCR detection efficiency. The same for G3 primer due to thermal stability.
In the last assay the presence of palindromic sequences in potential HTH telomeric motifs was tested with G5-PCR only.
Head-to-head telomeric motif generation by chromosomic fusion

Some of them generate palindromic sequences as GGCC cut by the isoschisomeres Hpa II / Msp I restriction endonucleases, or TTAA .There are 36 possible head-to-head telomeric motifs as follow:


If the telomeric fusion point contains a palindromic sequence, submitting the genomic DNA to endonuclease prior G5-PCR may prevent the PCR amplification:
Endonuclease digestion of DNA may prevent G-PCR amplification.
In uncut DNA (primary fibroblasts (line 1), precrisis fibroblasts (line 6), immortalized clone (line 11)), G5-PCR produces two major bands bellow 1200 bp. Those bands vanished as DNA is cut with MspI (GGCC) or with an endonuclease targeting TTAA sequence (from left to right, primary cells, pre or post-crisis fibroblasts:lines 2, 4, 7,9, 12, 13), indicating that palindromic sequences exist between inverse tandem telomeric motifs.
When genomic DNA is cut with HpaII, an endonuclease cutting GGCC motifs free of methyl-cytosine, the G-PCR product cannot be distinguished from the uncut DNA (lines 3, 8, 13), indicating that the cytosines are methylated in/between inverse telomeric motifs.

G5-PCR following endonuclease cut (Agarose gel + EtBr).(well 1: primary fibroblast DNA (PF) uncut, well2: PF+MspI, well3: PF+HpaII, well4: PF+TTAA cut, well5: molecular weight markers (possibly :2176, 1766, 1230, 1033, 653, 517, 453 bp), well6: precrisis transformed (PT) fibroblast uncut, well7: PT+MspI,well8: PT+HapII, well9: PT+TTAA cut, well10: mw marker(?), well11: Immortalized postcrisis fibroblast (IF) uncut, well12: IF+MspI, well13: IF+HpaII, well14: IF+TTAA cut.

If the aim  was to detect somatic telomeric fusions associated to cellular senescence or to cellular transformation, the G5-PCR assays failed to distinguish clearly the different cell states. The G-PCR to detect end-to-end telomeric fusion may be optimized (PCR optimization) by increasing stringency:
  • by increasing annealing T°
  • by decreasing elongation time
  • using hot-start PCR (Taq hot-start or wax)
Prior PCR, genomic DNA could be digested by Bal31 to completely remove terminal telomeric motifs and to keep telomeric motifs belonging to potential fusion points.
Polyacrylamide gel electrophoresis (PAGE) with P32-labelled (TTAGGG)n primer  instead of ethidium-bromide stained agarose gel to visualize PCR product, may detect somatic telomeric fusions in the range 0-600 bp. In the last example, the six last lines correspond to G-PCR products were visualized by PAGE (may be products corresponding to lines 6 to 11 in the second assay).
Real time PCR, in-situ PCR on nuclei could be also considered.
Six G5-PCR products (right side) in a PAGE+Ethidium Bromide staining
I was expecting that detection of somatic telomeric fusions yields short PCR products, bellow 500 bp. The bands in the 1Kb range, are not so short. They may correspond to head-to-head telomeric motifs flanking non telomeric sequences (containing TTAA and methyl-C GGCC motifs).
Two kinds of sequences leading to bands in G-PCR product.
This last hypothesis could be checked by FISH. The G-PCR product could be labelled (after a second round of PCR) with Fluo-dNTP and hybridized on metaphasic chromosomes. Three FISH conditions could be set:
  1. labelled G-PCR product alone: some locci should be detected.
  2. labelled G-PCR product+excess of Cot1 DNA: should compete satellite DNA. Detected locci should contain non satellite sequence, possibly telomeres or unique sequences.
  3. labelled G-PCR product+excess of unlabeled telomeric DNA (produced by PCR, Ijdo & al.): detected locci should not contain telomeric motifs.
Of course, G-PCR product should be insensitive to Bal-31 digestion of the genomic DNA prior PCR amplification.

Thursday, March 1, 2012

Accessing images through a xml object.

I realize that to go further toward an image processing application,  it becomes necessary to  store informations about processed images. Those informations could be stored in a xml object mirroring the tree directories on the hard disk.
Initially I hesitated between an ini file handled with configobj and a xml file. The facilities provided by Lxml convinced me to use that library.
The images are stored as follow: Project / Slide / Field / Fluorochromes/ images. A typical xml object representing a project would be:
  <slide name="A">
    <field name="2" ROI="x0,y0,x1,y1">
      <fluorochrome name="F1">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
      <fluorochrome name="F2">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
    <field name="1" ROI="x0,y0,x1,y1">
      <fluorochrome name="F1">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
      <fluorochrome name="F2">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
  <slide name="B">
    <field name="1" ROI="x0,y0,x1,y1">
      <fluorochrome name="F1">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
      <fluorochrome name="F2">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
      <fluorochrome name="F3">
        <image position="-1,-1" name="1.TIF" exposure="0.0"/>
For example, we have here a project called Cytogenet, two slides: A or B. In the slide A, two microscopic fields where observed, a region of interest was recorded by its corners (x0,y0) and (x1,y1); the first field was observed with two fluorochromes F1 or F2, and one image was taken for each with a given exposure time.