For any protein to be expressed in a model system you would need to have a strong promoter, viz., generally from a constitutively expressing protein like actin-1 (actin is highly expressed in all cells without any break).
But, if u wish to express any target protein in HeLa cell line, then make use of constitutive promoter (like ACT-1) just before ur target sequence. Thus on successful integration into the Hela genome, your promoter will fire endlessly expressing high quantity of target protein.
But again not all proteins can be over expressed as folding pattern and stability in the cell could be a problem, thus even if the protein is expressed (measured by mRNA quantification) it could be degraded on translation.
depends on your application. luciferase can be easily expressed to very high levels. also if you then use a retroviral vector to stably integrate the target gene, you can then generate a stable cell line (HeLa, HEK etc), which constitutively expresses your protein of interest to very high levels upon many passages. The advantage of luciferase is that is also provides a very easy way of measuring your protein expression level through bioluminiscence, if that is what your interest is? and as Ravi said above, the promoter for your target gene is also important. make sure that the promoter can maximize your expression. could use RSV for luciferase.