Colony PCR is a nice way of screening colonies if you have to do a lot of them. It only really works with Taq. You wouldn’t do this unless you needed to screen many colonies, so you’ll want to do a mastermix of the water, buffer, oligos, and polymerase and split it into multiple 20uL reactions. The temperature program is a normal one, except you have a longer pre-denaturation period to kill nucleases present in the thermally-lysed bacteria that will chew up your oligos. You should modify the extension time for the size of your predicted PCR product, assume 45 sec per kb.
There are two variations on the experiment—using oligos that anneal within your part, or oligos that anneal within your vector. If you use even one oligo that anneals within your part, you are looking for a yes/no signal from the PCR. If the part was present, you’ll get a signal. If it isn’t there, you’ll get no signal. That is both good and bad. It is good because it tells you both that your part specifically is there, and you learn its size and can see if it is right. Unfortunately, PCR can pick up on even tiny amounts of template in a reaction and you sometimes get false positives due to amplification of the remnant ligation reaction that you threw onto the plate the day before. The other way to do it is with external oligos. Usually you use your sequencing oligos, ca998 and G00101. Now, you’ll get a PCR product no matter what is inside your plasmid. You’ll determine whether it is your product or the parent vector based on the size of the product. So, for this to work, your part must be significantly different in size from your parent vector’s part.
To set this up, make and distribute your mastermix, then using a tip or toothpick stab a colony. Swoosh it around in the PCR reaction and then either streak an X on a plate, or drop it into a culture. If you don’t make a replica of the culture, your cells are gone! (it’s a common newbie error to forget to do that.)