DNA Pol can copy DNA text, but it can’t copy an epigenetic “highlight” – and this can be the source of biochemists’ despair or delight! If you want to steal bacterial machinery like DNA scissors (restriction enzymes) from different bacteria, at least make sure the parts are compatible – if their methylation’s restricting your restriction enzyme, you might need to modify your experiment!
In biochemistry and molecular biology we use a lot of protein machinery, like reaction mediator/speed-uppers called enzymes, that come from bacteria that are super useful – like restriction endonuclease (REases) (aka restriction enzymes) that are like DNA scissors that recognize and cut specific sequences. Different bacteria make different useful enzymes we can exploit, but, believe it or not, those bacteria don’t make that stuff just so that we can take it out of them – they make it for themselves, we just steal it. So sometimes problems arise if we try to mix and match machinery from different bacteria, so it’s important to know what you’re using, why, and where it’s coming from.
Bacteria use something called a restriction-modification (R-M) system to protect themselves from foreign DNA (from things like bacteria-infecting viruses called bacteriophages (phages). They use restriction enzymes (REases) to recognize and cut specific DNA “code words.” To keep from cutting their own DNA accidentally, the bacteria use methyltransferases (MTases) to “hide” the code words wherever they occur in their own DNA by methylating them. It’s kinda like, if safety scissors won’t cut what you want, protect your fingers instead.
They also use MTases for more general purpose uses, like differentiating old and new strands during DNA replication (when a cell copied all its DNA before splitting so each daughter cell gets a full genetic blueprint). In the lab, methylation can be a blessing or a curse – a blessing because it can help you selectively degrade DNA (like the template plasmid in a PCR mutagenesis reaction (will be explained…)) but a curse because it can hide cut sites you want to use.
And people often use then a lot. We utilize that precise specificity to cut precisely where we want in a piece of DNA. This is great for a variety of different tasks including molecular cloning – we can cut a gene out of one place and stick it somewhere else (like in a circular piece of DNA called a vector plasmid we can stick into bacteria) – and analytical restriction digestion (to see if we cloned correctly, we can use an REase we know should cut a sequence we put in and see if it gets cut (a kind of yes/no) – or cut something into pieces and compare the sizes to what we’d expect in different cases (e.g. if the insert got put in or not).
These techniques rely on that specific “code word” being recognized and cut – if that code word’s hidden, it won’t get cut and (in the cloning case) you won’t be able to cut out your gene or (in the analytical digest case) you’ll get “false” results. So let’s take a closer look at when/why/how this happens and what you can do about it
At heart, the issue is that The same universality of the genetic code that makes it so useful also comes with challenges….
Genomes are like cookbooks for recreating an organism. they’re written in DNA, with stretches of DNA (genes) acting as recipes providing instructions on making products (proteins or functional RNAs). All organisms have these genomes, everything from bacteria to humans (and aliens?) and we even use the same language, so we can stick human genes into cells from other organisms (like bacteria or insect cells) and those cells will know what proteins those genes “spell” & make them for us. This is what we do in recombinant protein expression.
The DNA language is the same, but we need a way to “give it to them” – a VECTOR (vehicle) – so we 1st use MOLECULAR CLONING to take a gene from one place (such as by cutting it out with REases) and stick it somewhere else like a PLASMID (a circular piece of DNA that bacteria can host) – kinda like taking a recipe from one cookbook, sticking it into a binder, then sticking that binder into the bacteria.
In one sense, it’s great that all DNA “looks the same” (just long chains of 4 nucleotide “letters” (A, T, C, & G) in various orders) so these other cells can read it, but on the other hand, because all DNA looks the same, how does cellular machinery tell different things apart?
One way is because it’s only the “naked” DNA looks the same & cells have ways to mark it up. We need something above & beyond just the genetic code – something you might call EPIGENETIC… (“epi” meaning over/outside of)
There are different types of such epigenetic modifications & they can serve different functions, but one of the main types is methylation – the addition of a methyl (-CH₃) group. This is like highlighting a letter and the highlighters are enzymes called methyltransferases (MTases)
In bacteria, Many MTases act in combination with restriction enzymes (REases) as part of a “Restriction-Modification” (R-M) system that protects them from foreign DNA by cutting it up – REases recognize and cut specific DNA sequences, and they use this as a way to destroy foreign DNA or RNA coming from foreigners like bacteria-invading viruses called bacteriophages (“phages”).
We use REases so often in the lab that it’s easy to forget that bacteria aren’t making them just for us to exploit – instead they’re making them for themselves. So one of their key concerns is making sure they don’t cut their own DNA in their quest to protect themselves. One way to do this is by having those restriction enzymes recognize really specific sequences – instead of cutting every “the,” cut every “kayak” – choose “rarer” words and keep those words out of your own DNA. note: I chose “kayak” because REase recognition/cleavage sites are often palindromic so that there’s a cut site on each strand.
But what if the bacteria really needs to write a story about kayaks and racecars? Even if such sagas aren’t in a bacteria’s plot, invaders can be sneaky – what if they don’t have those specific sequences? The phages have small genomes, so the chances of that sequence just randomly being there are small. Bacteria “know” this, so have evolved to target sequences that many phages need. But the phages can evolve to have different sequences.
So the bacteria needs more “generic,” snip-happy scissors as backup. But, *Dam,* how will the bacteria protect its own DNA? Other enzymes called METHYLTRANSFERASES come to the rescue. These guys add methyl (-CH₃) “tags” on the bacteria’s DNA. This methylation acts as a kinda uniform to tell the cutter, hey we’re on the same team! So methylation “hides” the cut sites of some restriction enzymes. It’s like, instead of using safety scissors to protect your fingers, you put on cut-proof gloves.
There also more “generic” methyltransferases which have functions outside of that R-M system, including Dam (DNA Adenine Methyltransferase) which is pretty highlight-happy. It methylates the A everywhere it sees GATC.
Why? One of its main functions is in replication -> Every time a cell replicates (grows by splitting in 2), it has to make a copy of its entire genome (the whole instruction manual) so that each “daughter cell” gets a copy.
This sounds like a huge task (and it is) but it’s made doable by the fact that these genomes are double-stranded and one strand is the “inverse complement” of the other. This sounds all technical but it just means that if you know the sequence of one of the strand you know the sequence of the other because A is always across from T and C across from G. And the inverse part’s because the 2 strands are antiparallel (running in opposite directions) -> so if 1 strand’s GATACA, the other strand’s TGTAGC
So if you “unzip” the strands you can make a copy of the missing side of that zipper based on the sequence of the side you have – an enzyme called DNA Polymerase (DNA Pol) does this letter-laying-down-ing and it does a really great job, but it still makes some mistakes.
So cells have to have ways to recognize these mistakes and fix them. Mismatch repair machinery can “white out” the errors and fix them, but they need to know which is the copy and which is the original so that they change the copy to match the original instead of changing the original to match the copy.
so Dam highlights up the original and as you might have found out the hard way, highlights don’t photocopy. DNA Pol’s great at using one strand of the DNA to recreate the 2nd strand to give you a nice new copy – of the “text” – BUT all that markup, the epigenetic info is NOT copied (at least at first) because the copier can only copy the text.
So the epigenetic modifications have to be re-added by other enzymes, like Dam methyltransferase, which methylates the A in GATC – they tag along after the copier, highlighting that letter by transferring a methyl group from S-adenosylmethionine (SAM) to the N6 position of the adenine (SAM I am until I met Dam!)
A couple useful things about this “teamwork” is that DNA Pol doesn’t get slowed down by having to add things on and the delay time between the letter copying and the modification adding gives you a brief window in which the DNA is “hemimethylated” – only the original, template strand has the methylation. So, if there’s a “typo,” the “editors” (mismatch repair machinery) will know to erase and fix the non-methylated strand (to match the original) before the new strand has time to be methylated. That way, the text of the recipe doesn’t change. In addition to that strand differentiation stuff, bacteria use methylation for various other things including helping regulate when replication occurs; helping regulate when genes are made into protein and, as can be inconveniently important in our cases, protecting the cell’s DNA from their own DNA-cutting machinery. But the first stuff’s good.
Even gooder is that, while when the copying’s done in cells, the original & the copy will *eventually* be identical, this methylation only occurs when the copying is done in bacterial cells with Dam. If you’re doing the copying by PCR (Polymerase Chain Reaction, a method to copy specific stretches of DNA) there’s no Dam in those test tubes, so the new strands don’t get methylated. And you can use this to selectively degrade the original strands using an REase called DpnI which *only* cuts *methylated* DNA (yup, it’s a weird one – but useful!). DpnI comes from Streptococcus pneumoniae which has kinda “swapped roles” with phages when it comes to methylation, potentially because some of those phages got tricky and started methylating their DNA. So the bacteria evolved so that DpnI ONLY cuts methylated sites.
This often comes into play when we’re doing site-directed mutagenesis – basically we want to change just a small piece of the gene but keep it in its same plasmid home. We can do this in different ways, but they mostly involve using PCR to make copies of the plasmid with the typo we want. It gives you lots of copies of the new, mutated, plasmid, but these are mixed in with all those “originals” and we need a way to weed out those template plasmids.
When we’re cloning, we usually use a selectable marker like an antibiotic resistance that’s in that plasmid home so that only bacteria with our plasmid can grow if we spike their food with that antibiotic. But antibiotic selection markers won’t help us distinguish between un-mutated “template” plasmid & the mutated plasmid we want because the sequence is in the same plasmid, so they both have the selection marker.
So we need another way to weed out the template plasmid. But they look almost the same right? There’s a key difference we can exploit -> the methylation gets added in bacterial replication, but not during the replication we do in PCR (no methyltransferases in there). So the original templates will be coated in methyl groups but our new copies will be methyl-free.
So, before we put the plasmid we’ve made into bacteria, we add some DpnI and some salts, etc. it likes, get it nice and cozy at 37°C and let it do its thing for a few hours -> the DpnI will cut up the template plasmid but won’t touch your new copies. This way, when you stick the plasmid into the bacteria, only the new copies will be “viable” and only bacteria that take it up will be allowed to live because they have the plasmid selection marker.
So in this case, methylation is helpful to us, but sometimes not so much… The problem comes when you try to mix and match bacterial machinery – use REases from one bacteria to cut DNA that was copied in bacteria whose MTases hide the sites the REases are looking for.
It’s great for replication that Dam methylates those As, but what’s not great is if that A also happens to be part of the “code word” for an REase, that REase won’t recognize it, so won’t cut it. Which can put a real *Dam*per on your DNA cutting plans.
Most of the REases we use in the lab don’t overlap with Dam sites, but ClaI, MboI, and XboI might (note: those “I”s are “ones” not “i’s” — the enzymes get their names from the bacterial strains they come from and the order in which they were found – e.g. ClaI was the 1st REase isolated from Caryophanon latum )
But even if the REases you want to use are safe in terms of Dam compatibility, there are a couple other MTases in the bacteria we commonly use for cloning that you may need to worry about.
Dcm methylates the second C in CCWGG. These sites aren’t as common as Dam sites (~1 per 512bp compared to 1 per 256 bp) but they can interfere with cutting by ApaI, BsaI, and McsI. And, even rarer (~1 site per 8 kb), but potentially problematic, EcoKI methylates As in A(A)CNNNNNNGTGC or GC(A)CNNNNNNGTT, which can hide cut sites for DraI, HpaI, and PmeI – but since these recognition sites are longer and thus rarer (like a stronger password) you don’t have to worry about them as much.
So, you may want to avoid using those enzymes if possible or, if you can’t use different enzymes, you can use different bacteria – there are Dam⁻/Dcm⁻ (the “⁻“ indicating these strains don’t have those enzymes) that you can use to copy the DNA. The DNA won’t get Dam or Dcm methylated, so you don’t have to worry about hidden cut sites, but you do need to worry more about typos since Dam’s not there to help the mismatch machinery know which strand is the right one.