Skip Navigation

2nd Principal Investigators Meeting

Identification of m5CpG Alterations Associated with Breast Carcinomas
Graham J. Brock
University of Glasgow

Table of Contents:

Identification of m5CpG Alterations Associated with Breast Carcinomas

Dr. Graham Brock Dr. Graham Brock
University of Glasgow

CpG Islands

CpG Islands
(click images for more detail)

CpG Island methylation

Isolation of methylated CpG Islands

  • Methyl-CpG binding domain (MBD) of MeCP2, fused to a histidine tag (HMBD) in a Ni-agarose matrix
  • Each HMBD binds a single m5CpG, thus genomic DNA retained non-specifically at low salt concentration.
  • However at increased salt conc. retained sequences will contain multiple m5CpG’s

The Methyl-CpG Binding domain column

Testing the MBD columns ability to separate three plasmids

End-labelled human DNA fractionated using an MBD column.

Additional analysis of the ‘bound’ fraction.

DNA eluting at salt concentration of ~0.8M and above was cloned and analysed, the resulting library contained.
  • SINEs (Alu’s), LINEs and other high copy no. repeats
  • rDNA (Non Transcribed Spacer), Mer22 and other low/medium copy no. repeats
  • Heterochromatin from Chromosomes 1,9 & 16
  • Exons/Pseudogenes
  • CGIs from the inactive X

Subtractive Hybridzation.

  • Although Methylated CGIs can be extracted the sample obtained is not pure and contains other methylated sequences
  • To remove these we use extracted DNA from tumor as ‘tester’ and from normal sample as ‘driver’
  • The subtraction will recover CGIs methylated in the tumor but not in the normal sample

Outline of Subtractive Hybridization Procedure

Recovery of plasmid spike

Samples Analysed

Specimens were obtained from patients undergoing mastectomies; tumors used in this study were classified as infiltrating ductal carcinomas. Adjacent normal parenchyma was obtained to serve as normal control

ID No. Age Clinical Stage Histological Grade
91/92 55 II MD
129/130 76 II PD
155/156 43 - PD
157/158 49 II WD

Yan, P. S., et al. (2000). "CpG island arrays: an application toward deciphering epigenetic signatures of breast cancer." Clin Cancer Res 6(4): 1432-8

Subtraction using densely methylated fraction from matched pair 155/156

Subtraction using densely methylated fraction from matched pair 155/156 Lanes 1 & 2
Tester = 155pd
Driver = 156n

Lanes 3 & 4
Tester = 155pd
Driver = 155pd

S = ‘Sigma Taq’
B = ‘Bio-taq’

Sequenced clones %GC vs.CpGObs/Exp

Preliminary results library 155pd-156n

Class of Sequence Number % of clones analysed
Predicted CGI 39 29
No. Significant Homology 14 10
High copy no. repeat 26 19
Low copy no. repeat 19 14
Exon/Pseudogene 14 10
Other 25 18
Total 137  

What is a predicted CGI?

The algorithm to determine CpG Islands is based on the definition of CpG Islands by Gardiner-Garden and Frommer (J. Mol. Bio 196:261-282, 1987)

    Region of ~200bp which has a GC content >50% and a CpGObs/Exp of > 0.6

    Additional Criteria - must be unmethylated in normal tissue

Identification of CpG methylation changes

Sequence comparison

(after bisulfite treatment)

Sequence comparison

Sequence comparison 1.

(after bisulfite treatment)

Sequence comparison 1.

Sequence comparison 2.

(after bisulfite treatment)

Sequence comparison 2.


  • Method extracts cloned inserts with the sequence criteria of a CGI.
  • Some of these predicted CGIs have altered methylation patterns when DNA from tumor and normal tissue is compared.
  • Further analysis is currently underway of the remaining cloned sequences.


For helpful discussion
Members of the Dynamic Mutation Group
the MBSU for sequencing services
both @ University of Glasgow
Tim Hui-Ming Huang @
University of Missouri

Adrian Bird @
University of Edinburgh

Funded by NIH Grant No. 1R21CA86305-01