Culturing Cells
Ideal method for culturing cells: Start an overnight culture from a frozen stock and use that overnight culture for work. Always start from the frozen stock.
Less ideal method but still acceptable in a pinch: Start an overnight culture from a frozen stock, store that overnight culture at 4C, and use that overnight culture for further subdilutions and work. Usually, I use cultures left in the 4C for starting new cultures for about 3-5 days at most. Use the same overnight stock for your work. This is less ideal than the method above because the cells age and die with time.
Very bad method: Subdilute a stock to create a new working stock, and then use that new working stock to subdilute into another stock, and so on. This is a bad method and should be avoided, as it can lead to accumulations of mutations and unknown changes that will cause fluctuations.
Maintaining Plasmids
In addition, although wild-type E. coli (such as MG1655 or EMG2) may seem to yield higher efficiencies for transformation, you should try to avoid maintaining plasmids in wild-type strains. Instead, use suitable cloning cells, which can have useful properties such as being endA deficient (leading to higher quality plasmid preps), defective restriction/modification systems, etc.
Tables of suitable cloning cells can be found online, such as http://www.neb.com/nebecomm/tech_reference/competent_cells/strainProperties.asp and http://openwetware.org/wiki/E._coli_genotypes