Skip to main content

Table 2 Summary of literature references comparing protein expression levels for native and optimized DNAs in bacterial systems

From: Assessing optimal: inequalities in codon optimization algorithms

First author

Optimization algorithma (source)

Target(s)

Number of constructs

Conclusions

Burgess-Brown [40]

Proprietary (Genscript, Sigma, and MediGene)

Various

30

• 26% of targets show higher expression of soluble protein for optimized over native CDS in E. coli

Kudla [27]

CAI

GFP

154

• Fluorescence levels span > 1000-fold across different CDSs

• No correlation between fluorescence levels and CAI

• Modest relationship between mRNA 2° structure and GFP fluorescence

Welch [28]

PLSR (DNA 2.0)

φ29 DNA polymerase

21

• > 100-fold difference in protein yield observed by differently optimized DNAs

Maertens [41]

CAI (GeneArt)

Various

100

• 24% targets showed ≥ 2× yield for optimized CDS

• 20% targets showed lower expression for optimized CDS

Spencer [42]

Undefined

Firefly Luciferase

7

• Optimization increased translation speeds ~ 2× with proportional decrease in functional protein

• 2–2.5× yield and solubility increase when recoded for frequent codons in Drosophila melanogaster

Trösemeier [43]

CAI (GeneArt) COSEM

ova

manA

5

11

• COSEM optimized sequences expressed ≥ 2× the native sequence

• “Ramp” inclusion was necessary for significant boost in protein expression

Konczal [44]

CAI (GeneWiz)

KRas4B

RalA

Rac1

11

11

11

• “Deoptimization” with ≤ 4 rare codons improves solubility ≥ 4× compared to native CDS

  1. CAI Codon Adaptation Index, PLS partial least squares regression, COSEM Codon-Specific Elongation Model
  2. aThe Kazusa database is reportedly used for codon frequency values by most commercial companies. The COSEM algorithm uses codon frequencies defined by Dong et al. for E. coli with a doubling time of 2.5 h−1