Poor or no expression, troncated proteins, amino-acids misincorporation are the most common consequences of codon bias and unbalanced tRNA pool. 

This issue has been extensively studied for E. coli expression (1) and is perfectly examplified by the arginine amino acid. Among the codons that may translate to arginine with the same tRNA: AGG and AGA are hardly ever found in E. coli highly and continuously expressed ORFs during exponential growth.   However, these 2 arginine codons are found with significantly higher frequency in coding DNAs of many other organisms (see table below).

Thus, when considering E. coli  expression system, it is important to replace the latter AGG and AGA by arginine codons used more frequently in your host organism e.g. CGT or CGC.

(1) Hénaut and Danchin, Analysis and Predictions from Escherichia coli sequences.
Escherichia coli and Salmonella, Vol. 2, Ch. 114:2047-2066, 1996, Neidhardt FC ed., ASM press, Washington, D.C. 
(2) Recombinant gene source, all genes frequencies, www.kazusa.or.jp/codon/

Designing a gene that will express in a particular organism a recombinant protein boils down to choosing the most appropriate triplet for each amino acids. With a ratio of 64 codons to 20 aa plus termination there is quite some flexibility to include other constraints in addition to that of codon bias adjustment such as: 

• Gene & protein engineering: Addition or removal of specific motives (e.g. restriction sites), tags for purification, multiple stops, etc. 

• Gene manufacturing: Maintain average GC content, avoid long repeats, palindroms, etc.