Yes, you do need to include a properly spaced RBS into your construct, but you have to realize that the frequency of translational initiation, which is what most of us construe as the efficiency of the RBS, actually depends on the sequence and secondary structure of the mRNA around the start codon (including a good part of the coding sequence itself). That's why you see in the literature seemingly suboptimal or poorly spaced RBSs which somehow seem to work like a charm, and why a poorly expressing gene in one expression vector can magically turn nice and express to high levels when subcloned to a different expression vector. There is no 'optimal' RBS: it depends on the 5'UTR of your mRNA (provided by the vector) and the sequence of the gene to be expressed.

The RBS calculator at http://salis.psu.edu/software/ will help you design an RBS for your protein. It has worked wonders for me. Also take a look at the references, etc.

As for the paper you mentioned, my guess is that the authors inadvertently created a cryptic RBS in the intervening sequence they used to separate their RBS from the ATG.

Yes, you do need to include a properly spaced RBS into your construct, but you have to realize that the frequency of translational initiation, which is what most of us construe as the efficiency of the RBS, actually depends on the sequence and secondary structure of the mRNA around the start codon (including a good part of the coding sequence itself). That's why you see in the literature seemingly suboptimal or poorly spaced RBSs which somehow seem to work like a charm, and why a poorly expressing gene in one expression vector can magically turn nice and express to high levels when subcloned to a different expression vector. There is no 'optimal' RBS: it depends on the 5'UTR of your mRNA (provided by the vector) and the sequence of the gene to be expressed.

The RBS calculator at http://salis.psu.edu/software/ will help you design an RBS for your protein. It has worked wonders for me. Also take a look at the references, etc.

As for the paper you mentioned, my guess is that the authors inadvertently created a cryptic RBS in the intervening sequence they used to separate their RBS from the ATG.

Thank you very much!!!

I will definitely try this

Beware though that the codon preference of the species your recombinant protein is from may not be identical to the codon preference of E.coli. This will also affect your expression levels.

I totally agree with the answer of Alejandro Martin. Many years ago, we used different lengts and even diverse nucleotide sequences of the RBS. We observed that there are no rules. You have to try diverse constructions to select the best for good lecture ot the mRNA and efficient expression of the protein. And that has to be done for every messenger. The message above also si quite true especially if you want to express proteins fomr Plasmodium which genome is around 80-85 % AT rich.

Well, if you do not believe in all the "ifs and buts" then the answer is a big NO.

Try to add 10 bp. its quite enough

Try using a RBS seq AGAAGGAGATATACAT just before your initiation codon, It worked fine for me. Thanks

If you are nor using the RBS attached to the primer than, you need to add RBS to the F' Primer spaced between 0.8 or 1.8 gyre spacer.

The thing to recognize is that the answer is relative depending upon the level of expression. Translation can vary by thousand or ten-thousand fold, so at the low end of translation (like lambda cI) you might not need an RBS. Translational restarts within an ORF occur in the absence of RBS. However if you want a high level of translation then you need to have optimization of the upstream region which includes both the RBS sequence, the spacing, and the surrounding mRNA context. That is why expression vectors often work well, because the context is usually optimized.

Without publications I was unable to get a Research Gate account at my previous employer email address; apologies for the multi-year lateness of this response.

I spent some time reading this paper and looking at the primers and restriction enzymes they used, as well as the sequences of the expression plasmids.  In all cases the protein of interest was cloned in-frame and downstream of expressed codons.  These are all fusion proteins with the labeled RBS the typical 7* or so base pairs distant from the actual start codon.  What they called the "start codon" of their genes of interest are actually just in-frame methionines.

What their paper actually demonstrated is that fusion to certain 5' signal sequences or tags can improve expression of difficult to express genes.

Details from what I wrote at the time:

*- Specifically: 7 nucleotides for SS5, SS6, SS7, and SS8 [pET-28A(+)], 8 nucleotides for SS1, SS2, SS3, SS4, SS9, and SS10 [pET-28B(+) and pTYB2]

When their genes of interest were successfully expressed in the "long spacer" versions of each construct, they were expressed as a fusion protein with the signal sequence or tags on their N-terminal end. At least one of these tags is cleaved off the final protein during processing (the pelB leader in pET-26B(+) for I-SS3 and I-SS4), which would result in a protein of the expected size on their gel. Comparison to the relative size of the same proteins expressed in pET-28A(+) (I-SS7 and I-SS8), which have N-term His and thrombin tags that do not get cleaved off, show a slightly higher molecular weight on their gel (at least to my eyes, based on their photo), which is expected from the extra 20 residues still present in these proteins (~2 KDa, I didn't specifically calculate). The extra 10 amino acids from pTYB2 will be almost undetectable on a gel at this resolution (my opinion based on very limited SDS-PAGE experience).