There are 2 strains you use for the same vector: cloning strain (K strain e.g. DH5alpha, XL1Blue) is the one with mutant dnases and recombinases and expression strain (B strain e.g BL21, Rosseta etc) is the one with mutant proteases.
In the K strain your ligated plasmid would be nicely transformed since there are minimum dnases and recombinases to affect it. These strains grow slow but have great transformation efficiency. They are also very easy to make competent using the Hanahan method which was actually developed for K strains. These strains are used mainly for plasmid production for mini/midi preping.
In the B strain, your protein will be nicely expressed and protected from the proteases, so it can accumulate up to the harvest time. Minimum preoteases to chop your protein up. Grow faster but have terrible transformation efficiency.
The story behind the original B-strains and K-strains is long. Initially, they were doing this old fashion UV-mutation/selection for the nice bacterial strains that people wanted to have. So they obtained strains to suit them but with lots of mutations. Hence, there are quite a few differences on the genome level between the strains. For more details, I recommend:
http://www.ncbi.nlm....pubmed/19765592 (as I said, Studier writes all the nice articles

)
The normal workflow from gene to protein in bacteria: PCR your insert with overhangs for the restriction enzymes you want to use; restrict digest the vector and the insert with these enzymes; ligate vector and insert; transform in K strain and plate; check the ontained colonies for the right insert by colony PCR; miniprep a few of them and send for sequencing; once sequence confirmed, transform the plasmid in B-strain and do the expression of your protein in the B-strain.
Now, this is the classical way (details omitted for simplicity; do not take it as a protocol; there are some steps in between that are essential such as PCR clean up...). If you for example want to transform an epPCR library or any kind of library of variants and select based on expression of your protein, then you cannot wait to obtain the plasmid from K strain, you have to go directly to the B-strain. So, in this case, due to the fact that nowadays we have a bit of more knowledge of how to manipulate bacteria, beyond UV, people have modified the classical BL21(DE3) to get BL21-Gold(DE3) (and similar strains) which has the advantage that it is also a mutant of endonucleas A (as K strains are, but this one is done nicely with modern molecular biology methods not UV

) Now, you will be tempted to say, hack, why should I use the long way and not transform the ligation directly to BL21(DE3)-Gold? Well, you could do that, but I personally do not like strains with too many mutations (endonuclease A is needed in the bacteria) because they are not too healthy and stressed bacteria do not produce tons of protein. But this is just me.
Another question that might arise would be: why do you go through K-strain until the expression strain:
1- easy to transform ligation reaction into; if you use BL21(DE3) for transforming you might end up with no clones, but you might also be lucky. The question is: do you want to play the lotery in the lab?

2 - easy to mini/midiprep your plasmid to have enough for sequencing or other things for which you need ug amounts. Again, nobody says you cannot miniprep from BL21, but you need to be good at. I minipreped successfully from B strains, but I am a mixture of luck and good hands:) I have seen people not being able to miniprep from B-strains. So again: how lucky do you feel to do that? You might end up repeating the entire cloning procedure because you were too lazy to go through the K-strain. (we have a saying in my language: the lazy one, runs more; I thought appropriate to insert some wisdom from the elders here

)
This was the story about the strains = living organisms that multiplicate on their own. Not to be confused with plasmids: round pieces of DNA that is not able to multiply on its own. For simplicity, I will refer from now own only to bacterial plasmids, even though the parts are common for many other plasmids.
In a plasmid you generally have:
-origin of replication - needed by the bacteria to recognize the plasmid for replication; these origins are that thing that give the copy number; you can have high copy number plasmids (100 copies per cell) or low copy number plasmids (1-10); this is regulated by the origin of replication. In general, for expressing proteins, you want a low copy number plasmid. But for cloning in bacteria a plasmid to be used for mammalian cell culture imaging or for yeast protein expression, basically to be used in another organism, then you use a high copy number plasmid.
-antibiotic resistance - with its own constitutive promoter and terminator; produces protein to clear the antibiotic; makes the bacterium survive in the certain antibiotic used for selection
-promoter - used for transcription of mRNA from your insert
-terminator -stops mRNA transcription
-rbs - ribosome binding site - ribosome recognition site for translation to happen
-MCS - multiple cloning site with various restriction sites for you to put your insert into
These are the minimum components you have on an expression vector in bacteria. On top, you have some other regulatory components. But this is advanced molecular biology I will let you discover on your own

Andreea