Sabregirl's Rules for cloning

These are my rules of thumb for cloning. I haven't done many of the more difficult cloning techniques such as blunt-blunt and blunting using fill in but I have had enough trouble from the so-called easy ones to write up this guide. Many of these points come from actual problems I've had in the lab.

The first thing you should always do when planning to make a new construct is make a map.
Sounds silly and maybe you don't have good software for this but even if it's just a hand drawn diagram, it will save you a lot of headaches later on. Be sure to list locations of restriction sites you may be using and distances between them. The ultra basic (and free) Plasmid Processor is often good enough for your own maps.
Make sure you check (by restriction/sequencing) whatever plasmid/clone etc. someone hands you.
It's happened to me and many other people I know that the construct you are given may be wrong, or the map may not be accurate. No matter how much you THINK you can trust who is giving it to you, check it anyway. A few simple restriction digests can save you days, weeks and sometimes even years of problems. You should ESPECIALLY check restriction sites that you intend to use for cloning. Just run a quick test digest of sites you want to use before designing primers that require them. Sometimes silly things like maps being wrong or people failing to mention things like DNA methylation can trip you up.
When you're modifying/ designing new sequence, ALWAYS run it through a restriction site analysis program-no matter how small.
Sometimes you'll accidentally generate annoying things like unwanted restriction sites that can cause problems later.
Avoid unecessary steps when cloning
Most cloning protocols I find seem to say that you should ALWAYS dephosphorylate and/or gel extract your vector. There are times when these steps are vital to successfully cloning something but this is not always the case.
The easiest type of cloning is typcially subcloning.
By subcloning I mean taking a piece from one bacterial plasmid and inserting it into another. This is easiest when you are using two different and non-compatible overhang generating restriction sites. This means you don't have to dephosphorylate your vector. If your vector contains a restriction site that is not in your insert but is between the two restriction sites you are using to clone your insert (as in a multiple cloning site) then you can avoid gel extracting your vector as well. You can use this third restriciton enzyme to cut religated vector molecules after the ligation reaction and prevent them from being transformed and compied by your competent cells. Subcloning without gel extracting the vector and cutting the reaction after ligation is so easy that it is usually hard to find plasmids that don't have the insert. Most of the time I will only pick 3 or 4 colonies with 3 out of 4 or 4 out of 5 will be positive. Because of the effectiveness of this method, I sometimes do several subcloning steps simply to avoid more difficult cloning reactions such as blunts. Sometimes you can get good results from a ligation even if you don't have a third restriction site to use but it may be easier to gel extract the vector in this case.
When you get too much backgroud from non-gel extracted vectors OR you are trying to clone a fragment generated through PCR, gel extract the vector.
Gel extraction makes DNA more difficult to transform. You will get much, much lower transformation efficiencies from gel extracted vectors. Sometimes, though it is necessary to use gel extracted vectors. When doing this you should use as much vector DNA as possible but 100ng is good to shoot for. You NEED to get extract the vector when cloning from PCR. PCR products are non-methylated always gel extracted and just plain hard to clone many times. If you use a non-gel extracted vector, most times you will get far too much background to find your clone. If you do not have to use high fidelity taq (which typically leaves blunt ended fragments) then you can use a TA cloning kit such as the one from invitrogen. These kits are usually MUCH easier than cloning via restriction directly through PCR. If however you HAVE to use high fidelity taq to avoid massive numbers of mutations, digesting the taq polymerase with proteinase K, followed by phenol chloroform extraction can increase your restriction efficiency and increase the number of clones you can get.
Always use fresh ligase buffer
Ligase depends on ATP for function. ATP is usually included in the ligase buffer but it can go off after repeated freeze-thaw cycles. When I buy a new batch of ligase I aliquot out the ligase in small 500uL tubes with about 20uL in each and store them frozen. These can be re-thawed up to about 2 or 3 times then tossed.
Always do a positive and negative control ligation
For a long time I would do ligations and get nothing back out of them. I was unfortunately listeing to an unskilled post-doc who thought controls were "a waste of time". Especially when doing ligations without gel extracting the vector you should make up a ligation without an insert (checks if the DNA prep has an inhibitory component or the ligase is bad) and make up a ligation that doesn't have any ligase but just has vector (this checks the background of your uncut vector which is always present when not gel extracting your vector). When the vector is gel extracted it is less necessary to do a negative and positive control ligation often times I do not do them but it never hurts to check ligase function and that your vector is actually cut the way you think it is.
If it doesn't work one way, try another
There are typically many different ways of achieving a particular construct. Sometimes one way of cloning will simply not work for one reason or another. Screening hundreds of colonies is typically not a good solution to your problem. Usually if you didn't find it in 10 or 20 you're probably not going to find it at all. It's always good to have a backup plan. Often times you may learn later that there was a real reason why it wouldn't work that way.
Translate the sequences you get from a sequencing reaction
While you should be able to find mutations from DNA:DNA alignments translating the sequence is an additional step that should help you find any errors.
Frequently asked Questions about Ligations and cloning

What Ratio should I use?
Remember that the ratio is a ratio of number of molecules and not a ratio of DNA mass. To do this I typically simply take the mass or concentration of DNA I am working with and divide it by the kilobases of the fragment. I call this the "factor" Many protocols have you actually calculating with the average molecular weight of DNA but this is unecessary for setting up ligations as your are simply using a constant in the calculation. As an example if you have a DNA concentration of 5ng/uL and the DNA fragment is 1600bp then the factor is 3.125.
I almost always use a 3 to 1 ratio (insert to vector) when working with inserts that are smaller than vectors and especially when not gel extracting the vector. You can, however use all sorts of ratios including 1 to 3, 9 to 1 etc. Sometimes getting the right ratio is the key to success.
What kind of DNA should I use?
This may seem silly but for a while I was using very "dirty" DNA for my ligations. This stopped them from working almost completely. The DNA you use should be clean meaning either use a kit or a couple of phenol chloroform extractions to remove contaminating protein.
How do I determine how much DNA I have?
I always use gel intensity quantification. We have software (Quantity One) associated with our bio-rad gel documentation system that will allow accurate determination of DNA quantity on a gel. This is then related to the quantity of DNA found in a particular band of the marker. Some people say they can quantify using a spectrophotometer but our spec does not seem to be reliable for gel extracted DNA. You could probably use a flourometer instead.
How much should I transform with?
For me typically the answer is the whole ligation. This is probably excessive especially in the case of a non-gel extracted vector but is usually necessary for a gel extracted vector.