Day 2 Evening Lecture Notes
Barton Slatko, New England Biolabs
June 7, 2004
When phage are killed by bacteria they are trying to infect, they were historically said to be "restricted." "Plaques" are "holes in the bacterial lawn" of a petri dish and are comprised of phages that are not killed. Survivor phage can be modified and used to infect another lawn in an iterative process until all phage survive. The phage were historically said to be "modified."
If the modified survivor phage are now exposed to a different strain of E. coli, they will once again be killed. Once again modified phage can be isolated. Bacteria have "restriction enzymes" that kill attacking phages. Restriction endonucleases cut phage DNA internally. The phages that survive are methylated -- that is, a methyl group protects DNA from cutting. In this example the second strain of E. coli has a different restriction enzyme that cuts at another still unmethylated site.
Bacteria have these cleavage enzymes for self-protection. Where does the methylation defense come in? A cell doesn't want to cut its own DNA so it has evolved a "cognate methylase" to protect its own DNA.
Some bacteria have evolved as many as 12 restriction enzymes. The proliferation is a sign of the arms race between bacteria and viruses. Usually methylation is slower than restriction but occasionally phage progeny are still produced. Most progeny are killed but some survive since the methylated ssDNA template created during replication will have a greater tendency to get a methylated second strand, which in turn will favor another methylated strand during the next replication. (This propagating methylation appears to be a unique example of Lamarckian inheritance of an acquired characteristic.)
EcoR1 comes from E. coli, strain R. Other enzymes come from different bacteria. A few hundred restriction enzymes have been identified that fall into four classes. The type II are most commonly used. Type II enzymes can be monomers (like methylase) or dimer (endonucleases). Many recognize like EcoR1 recognize dyad (mirror) sequences. Dimers sit on both strands of DNA simultaneously. Cleavage enzymes stress and distort bonds until they break. The recognition mechanism that enzymes use is heavily studied. A methyl group on the second nt of the EcoR1 site (the first A of GAATTC) prevents EcoR1 from cutting it. Each restriction enzyme has its own methylase. Methyl transferase moves methyl groups from S-adenosyl methionine (SAM) to DNA. SAM is often used in buffer solutions.
In general restriction enzymes can leave overhangs at the 5' or 3' end, or they can be "blunt cutters." Clearly enzymes that leave "blunt ends" have less specificity in religation than those that produce matching overhangs. With EcoR1, two overhangs that ligate together can be cut again with the same enzyme. This is not true with blunt cutters, where a blunt of the recognition site
5' GGGCCC 3'
3' CCCGGG 5'
5' GGG CCC 3'
3' CCC GGG 5'
where the two ends can now bind to another sequence that will not reproduce the GGGCCC recognition site.
Example: HincII cuts
5' GTPyPuAC 3'
3' CAPuPyTG 5'
where "Py" is any pyrimidine and "Pu" any purine. This non-specificity is called "enzyme wobble."
Note that an enzyme that cuts at the center of a recognition site is a blunt cutter while one that cuts nearer an end produces an overhang. Less-specific blunt cutters are harder to use because of more possibilities for undesirable ligation products but convenient recognition sites for enzymes that produce overhangs are not always available in the region of interest. Enzymes that cut near the beginning of a recognition sequence (like EcoR1) generate a 3' overhang while those that cut near the end generate a 5' overhang.
The length of recognition sites varies. To cut DNA into many small fragments use an enzyme with a short recognition sequence, e.g. a "4-cutter." To cut DNA into fewer, longer fragments, use an enzyme with a longer recognition sequence like an "8-cutter."
For a random DNA sequence, the probability of finding the EcoR1 recognition site is 0.254 = 1/4096. However, genomes are not truly even in nucleotide distribution. Thermophile organisms have more of the strongly bonded GC nucleotides while many organisms have more than 50% AT.
Two enzymes that have the same restriction sequence are called "isoschizomers." They may still cut at different places and generate different resulting overhangs. For example HhaI and HinPI both cut GCGC, but one generates a 3' overhang and one a 5' overhang. The New England Biolabs catalog has a chart explaining which enzymes produce compatible overhangs.
A major difficulty in developing a cloning strategy is that much of mammalian and plant DNA is methylated. Different restriction enzymes are blocked by different methylation sites even if they have the recognition sequence. If the sequence of the gene to be cloned is not known, the scientist is advised simply to try a bunch of different restriction enzymes.
Except right after replication, methylation is always symmetric on both strands. This is why methylated phages tend to have methylated progeny.
In choosing an enzyme, be aware of: