Restriction of Foreign DNA by E. coli K-12

E. coli has several mechanisms for identifying foreign DNA and destroying it. This can be a significant problem in cloning experiments, resulting in substantially reduced recovery of desired sequences. The problem can be avoided by the use of strains in which these mechanisms are disabled by mutation. A strain completely disabled for restriction will be defective at the hsd, mcrA, mcrBC, and mrr loci (see below).


EcoKI restriction, encoded by the hsdRMS genes, attacks DNA that is not protected by adenine methylation at the appropriate recognition site (AAC[N6]GTGC or GCAC[N6]GTT) (1). McrA, McrBC, and Mrr, encoded by mcrA, mcrBC, and mrr (2-6), are methylation-requiring systems that attack DNA only when it is methylated at specific positions. All three of the latter systems restrict DNA modified by CpG methyltransferase (M.Sss I) to contain methylcytosine in CpG dinucleotides (5). Mrr will also attack DNA with methyladenine in specific sequences (4,6). One or more of these is present in most commonly used strains of E. coli. (7-12).

The methylation-requiring restriction systems are sequence specific; McrA, McrBC, and Mrr do not restrict DNA modified at dcm sites, nor does Mrr restrict DNA modified at dam, EcoKI, or EcoRI sites. In addition to restricting M.Sss I-modified DNA, McrA restricts DNA modified by the HpaII methylase (5´ CmeCGG; 2), while McrBC restricts DNA modified at the sequence 5´ RmeC (2,3,13). McrA and McrBC apparently do not distinguish between 5-methylcytosine and 5-hydroxymethylcytosine; McrBC also restricts DNA containing N4-methylcytosine in appropriate sequences (3). Mrr restricts DNA modified by a variety of adenine methyltransferase and several 5-methylcytosine methyltransferase but no consensus recognition sequence has yet been deduced (4,6).

Almost all laboratory strains of E. coli are derivatives of wild isolates K-12 or B. They do not carry EcoRI or other type II restriction systems, which were identified in other wild isolates.

When to worry

The mcr and mrr loci have been shown to reduce recovery of methylated sequences from mammals (8,9,11,14) and plants (9,15) in cloning experiments. In general, it is wise to use a strain lacking the Mcr and Mrr systems when cloning genomic DNA from an organism with methylcytosine [e.g., NEB 10-beta Competent E. coli (NEB #C3019), NEB Express Competent E. coli (NEB #C2523) or T7 Express Competent E. coli (NEB #C2566)]. This includes all mammals, higher plants and many prokaryotes (16), but not the important experimental organisms Drosophila melanogaster (17) and Saccharomyces cerevisiae (18). In addition, such a strain should be used when using cytosine methyltransferases to generate novel specificities (19) or to protect cDNA from subsequent digestion (20). Since methyladenine is present in many bacteria and lower eukaryotes, Mrr-mediated restriction should also be considered when cloning genomic DNA from these organisms.

When not to worry

Note that the foreign methylation pattern will be lost (and the E. coli methylation pattern acquired) upon replication of the clone in E. coli, unless the clone carries methyltransferase activity. Once successfully introduced, clones can be moved freely among Mcr+ Mrr+ E. coli strains, since the methylation pattern will no longer be foreign. Be sure the DNA is Eco-KI-modified before trying to introduce it into an EcoKI-restricting strain. If necessary, use NEB 5-alpha Competent E. coli (NEB #C2987) or NEB 5-alpha F´Iq Competent E. coli (NEB #C2992) to modify and protect EcoKI sites against EcoKI restriction activity.

Properties of strains

The table here summarizes known McrA, McrBC, and EcoKI strain phenotypes. Mrr phenotype is indicated under "background" when it is known. Only a few strains have been explicitly tested, but it is reasonable to assume that most strains are Mrr+. We have not tested for the prophage-encoded EcoP1 endonuclease; to our knowledge, only JM103 (not in the table) carries this. Strains with a deletion (Δ) of the mcrBC-hsdRMS-mrr cluster were shown to be more permissive hosts than strains with point mutations in hsd and mcrB (9,11,14).


  1. Bickle, T. (1993) in Nucleases eds Linn, S.M., Lloyd, R.S. and Roberts, R.J. (CSH, NY) p. 89-109.
  2. Raleigh, E.A. and Wilson, G. (1986) Proc. Natl. Acad. Sci. USA 83, 9070-9074. PMID: 3024165
  3. Raleigh, E.A. (1992) Mol. Microbiol. 6, 1079-1086. PMID: 1316984
  4. Heitman, J. and Model, P. (1987) J. Bacteriol. 169, 3243-3250. PMID: 3036779
  5. Kelleher, J. and Raleigh, E.A. (1991) J. Bacteriol. 173, 5220-5223. PMID: 1830580
  6. Waite-Rees, P. et al. (1991) J. Bacteriol. 173, 5207-5219. PMID: 1650347
  7. Raleigh, E.A. et al. (1988) Nucl. Acids Res., 16, 1563-1575. PMID: 2831502
  8. Whittaker, P.A. et al. (1988) Nucl. Acids Res., 16, 6725-6736. PMID: 2841642
  9. Woodcock, D.M. et al. (1989) Nucl. Acids Res., 17, 3469-3478. PMID: 2657660
  10. Raleigh, E.A., Trimarchi, R., and Revel, H. (1989) Genetics 122, 279-296. PMID: 2548920
  11. Grant, G.N. et al (1990) Proc. Natl. Acad. Sci. USA 87, 4645-4649. PMID: 2162051
  12. Krüger, T. et al. (1992) Gene 114, 1-12. PMID: 1316864
  13. Sutherland, E. et al. (1992) J. Mol. Biol. 225, 327-348. PMID: 1317461
  14. Doherty, J.P. et al. (1991) Gene 98, 77-82. PMID: 1849497
  15. Woodcock, D. M. etal (1989) Nucleic Acid Res., 17, 3469-3478. PMID: 2657660
  16. Ehrlich, M. and Wang, R.Y. (1981) Science 212, 1350-1357. PMID: 6262918
  17. Urieli-Shoval, Y. et al. (1982) FEBS Lett. 146, 148-152. PMID: 6814955
  18. Proffitt, J.H. et al. (1984) Mol. Cell. Biol. 4, 985-988. PMID: 6374428
  19. McClelland, M., Nelson, M. and Raschke, E. (1994) Nucl. Acids Res. 22, 3640-3659. PMID: 7937074
  20. Meissner, P.S. et al. (1987) Proc. Natl. Acad. Sci. USA 84, 4171-4175. PMID: 2438693