Dear Geya

to my knwoledge there are 2 different phylosophy:

codon optimization or codon armonization;

codon optimization tend to replace all the rare codons identified with codons used in the organism used for the expression

so for example for the arginine in E.coli the 4 more rare codons (AGG,AGA,CGG and CGT are ever replaced with CGT or CGC)

Arginine codon frequency in E.coli

Arginine AGG codon frequency in E.coli --> 0,12% (% of ARG codon: 2,1)

Arginine AGA codon frequency in E.coli -> 0,20% (% of ARg codon 3,6%)

Arginine CGG codon frequency in E.coli --> 0,50% (% of ARg codon 9,1 %)

Arginine CGT codon frequency in E.coli -> 0,35% (% of ARg codon 6,4%)

Arginine CGT codon frequency in E..col--> 2,10% (% of ARg codon 38,3 %)

Arginine CGC codon frequency in E.coli --> 2,20% (% of ARg codon 40,2%)

extrapolated from:

https://openwetware.org/wiki/Escherichia_coli/Codon_usage

codon armonization tend to reduce the number of rare codon and remove the clusters expecially for more critical codons (eg arginines codon in E.coli) but tends to armonize the number of rare codon with the one normal codon usage present in the expression organisms (so for example in the specific case of arginine, for long proteins contaning many arginines some of the less frequent codon are also used (eg if the protein contain 30 arginines, to represent the real codon usage of the expression host, at least 2 CGG codon and 1 AGA codon will be used too)

Of course while the clone desin gby first approach is simple while the second is more complex in absence of an algoritm that facilitate it.

good luck

Manuele

Another strategy to deal with rare codon usage is to use the Rosetta expression strain of E.coli, which expresses genes for several rare tRNAs.

Basically, you translate your open reading frame into a protein sequence and then translate back into cDNA using high frequency codons. There are automated tools for this on the web, where you may define constraints like the restriction sites you need for cloning and manipulating your sequence, adding tags and stuff like Shine-Delgarno resp. Kozak Consensus elements, etc. . Then have your new cDNA synthesized. It never has been this easy.

geneart.com used to have such a nice tool on their website.