Cancer Genet. 2013 Jun;206(6):222-6.

The importance of codon context for understanding the Ig-like somatic hypermutation strand-biased patterns in TP53 mutations in breast cancer.

Lindley RA.

Melville Analytics Pty Ltd, Jindabyne, Australia. admin@melvilleanalytics.com

 

Abstract

Evidence already exists that the activation-induced deaminase (AID)/APOBEC family constitutes a set of differentially expressed enzymes capable of deaminating cytosines (C to U) in single-stranded DNA (ssDNA) and that they are potentially powerful mutagens. The mutagenic processes involved are believed to be activated in many nonlymphoid tissue types-for example, initiating some cancers and/or leading to further somatic mutagenesis. To investigate the extent that codon context might be important in influencing the likely location of TP53 mutations in breast cancer, the codon-bias patterns resulting from the ssDNA target specificities of cytidine deaminases of the AID/APOBEC family were analyzed. The data indicate that codon context strongly influences the likely location of mutations at motifs for AID/APOBEC1/APOBEC3G, and at WA sites. An unexpected finding is a highly significant preference for transitions of cytosine to occur at the first nucleotide position and for transitions of guanosine to occur at the second nucleotide position in the mutated codon (read 3′ to 5′). Thus, the mechanisms involved appear to be sensitive to codon reading frames and to have an intrinsic ability to differentiate between the cytosines on the nontranscribed strand and those on the transcribed strand in the context of an open “transcription bubble.”

Copyright © 2013 Elsevier Inc.

KEYWORDS: TP53, codon-bias, breast cancer, somatic hypermutation, mutagenesis

PMID: 23880211

 

Supplement:

As the cost of whole genome and exome-wide sequencing continues to decrease, a huge amount of effort is being focused on adding ‘medical meaning’ to the massive amount of sequence data being produced: understanding the mechanisms directing the generation of somatic mutations is fundamental to these efforts.

In this study, it is hypothesised that codon context influences both the location and type of mutations occurring in the TP53 gene in breast cancer tissue. The possibility that the transcription-linked mechanisms involved might be able to read the underlying codon structure of ssDNA in-frame during transcription has been overlooked until now.

To test this hypothesis, 2,514 breast cancer mutations from the International Agency for Research on Cancer (IARC) TP53 Database were analysed (http:www-p53.iarc.fr/). The IARC TP53 breast cancer dataset is selected as representative of the endogenous somatic mutation pattern, as this mutation pattern appears to arise in tissue that is least accessible to exogenous carcinogens such as those in tobacco smoke. Most of the mutations are single point mutations, predominantly focused in the DNA binding region (codons ~130-300) of the TP53 gene.

A key finding is a highly significant preference for transitions of cytosine (C) to occur at the first nucleotide position of the mutated codon (MC-1, read 5-prime to 3-prime), and for transitions of guanine (G) to occur at the second nucleotide position in the mutated codon (MC-2) for mutations occurring at AID/APOBEC3G/APOBEC1 motifs. Thus, the mechanisms involved appear to be sensitive to codon reading frames, and have an intrinsic ability to differentiate between the Cs on the non-transcribed strand, and those on the transcribed strand in the context of an open “transcription bubble”. The results imply the active involvement of a transcription-linked codon sensor operating at the level of single stranded DNA (ssDNA).

The findings reported in this study also reveal that the different members of the AID/APOBEC family of cytosine deaminases display highly significant and differential codon-biased mutation spectra. For the dominant transitions of guanine to adenine (G>A) and cytosine to thymine (C>T) at motifs for AID and APOBEC3G, and for the dominant transition A>G at WA (W = A/T) sites, a highly significant MC-1/MC-2 codon bias is found (p<0.001 level for 1df). In the case of the APOBEC3G data subset, there are 505 mutations occurring off G at the selected motif CG, that result in a G>A, G>C or G>T mutation. Each possible mutation type can occur at an MC-1, MC-2 or MC-3 site within the mutated codon structure.  That is, there are nine (3×3) possible classes of base substitution occurring off an APOBEC3G motif. Of the 505 mutations in this data subset, 358 mutations (71%) are G>A mutations coincident with an MC-2 site. If the mutations are randomly generated, we would expect that around 56 (i.e. 1/9 or 11%) would be G>A at an MC-2 site. The predominance of G>A mutations at CG targets in MC-2 sites suggests that APOBEC3G-mediated deamination is a major contributor to at least some breast cancers, and that the mutations are far more specifically targeted than previously thought.

In comparison, the APOBEC1 dataset reveals a statistically significant codon-bias for C>T transitions, but not for G>A transitions. One possible explanation is that the simple TG/CA sequence does not adequately define the APOBEC1 motif previously defined for its interaction with DNA (or RNA) templates. In addition, in cancer cells there may be stochastic competition among AID/APOBEC family members for the selection of binding sites, as some nucleotide sequences are a match with more than one of the potential candidates in vivo. Further research will be able to address these and some other possibilities.

There is quite a remarkable preference for the dominant transitions of C to target MC-1 sites, and for the dominant transitions of G to target MC-2 sites at AID, APOBEC3G and WA motifs. A key inference is that the mechanisms involved are able to discriminate between Cs on the non-transcribed strand, and those on the transcribed strand of ssDNA in the context of a transcription bubble. As it is known that the AID/APOBEC family of cytosine deaminases are differentially expressed among various tissue types, the results also add to our understanding of how variations in the somatic hypermutation-like strand-bias patterns observed for different cancer types might be generated (1).

This initial study is important in that the differential codon-bias spectra observed for the various AID/APOBEC proteins analysed, provides a rational basis for the development of new genetic tests to identify the likely source of new somatic mutations arising in individuals. Further analyses using the targeted somatic mutation (TSM) approach described might also lead to the design of more specific and effective genomic therapies, and the production of highly specific substrate-based inhibitors for targeting the mutagenic proteins identified in disease-affected tissue.

Addendum

Since the publication of this paper, a similar analysis applied to whole-genome cancer datasets show that the phenomena described here are evident in other genes, and in all of the non-lymphoid and lymph tissue types analysed so far (In Preparation). That is, the phenomena described in this study appear to be a part of a universal set of endogenous TSM mechanisms that may result in somatic mutations associated with disease, or perhaps give rise to more beneficial somatic diversity within an individual during development.

 

References:

  1. Steele EJ, Lindley RA 2010 Somatic mutation patterns in non-lymphoid cancers resemble the strand biased somatic hypermutation spectra of antibody genes. DNA Repair 9: 600-603.
  2. Lindley RA, Steele EJ 2013 Critical analysis of strand-biased somatic mutation signatures in TP53 versus Ig genes, in genome-wide data and the aetiology of cancer. Review Article, ISRN Genomics 2013:1-18.

 

Dr Robyn A LindleyContact

Robyn A Lindley PhD

Melville Analytics Pty Ltd

Unit 14 Nuggets Crossing

Jindabyne, NSW 2627

AUSTRALIA

Tel/Fax: +61 (0) 2 6456-1393

robyn@melvilleanalytics.com,

Multiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier SchönmannMultiselect Ultimate Query Plugin by InoPlugs Web Design Vienna | Webdesign Wien and Juwelier Schönmann