( High-level expression of difficul...)Codon Bias-Adjusted BL21 Derivatives for Protein Expression

High-level expression of difficult heterologous proteins

Carsten-Peter Carstens Anna Waesche
Stratagene

Stratagene has created variants of the Epicurian Coli BL21-Gold cells * called the BL21-CodonPlus -RIL series** of strains, which contain extra copies of the E. coli argU, ileY, and leuW tRNA genes. This modification allows for a high-level expression of proteins that are difficult to express in conventional E. coli hosts due to the codon bias of the gene of interest. The BL21-CodonPlus-RIL strains can eliminate the need to move the gene of interest into eukaryotic expression systems.

BL21-Gold(DE3) cells are the strain of choice for high-level production of heterologous proteins expressed from the T7-promoter. These strains have been engineered to contain an inducible T7-RNA polymerase. They are also EndA- to allow high-quality miniprep DNA preparation.¹ BL21 competent cells are derivatives of E. coliB and lack the Lon protease and the ompT protease genes that may cause instability and degradation of the recombinant protein.

Table 1
Rarest Codons (from E. coli) Preferred by Selected Organisms

Codon frequencies are expressed as codons used per 1000 codons encountered. A complete compilation of codon usage of the sequences placed in the gene bank database can be found at http://www.dna.affrc.go.jp/~nakamura/codon.html Codon frequencies of more than 15 codons/1000 codons are shown in bold to help identify a codon bias that may cause problems for high-level expression in E. coli. The arginine codons AGG and AGA are recognized by the same tRNA (product of the argU gene) and should therefore be combined. However, regardless of the origin of the coding region of interest, each gene should be assessed individually. Even E. coli contains genes with poor codon usage for efficient translation.

AGG arginine

AGA arginine

CGA arginine

CUA leucine

AUA isoleucine

CCC proline

Escherichia coli

1.4

2.1

3.1

3.2

4.1

4.3

Homo sapiens

11.0

11.3

6.1

6.5

6.9

20.3

Drosophila melanogaster

4.7

5.7

7.6

7.2

8.3

18.6

Caenorhabditis elegans

3.8

15.6

11.5

7.9

9.8

4.3

Saccharomyces cerevisiae

9.3

21.3

3.0

13.4

17.8

6.8

Plasmodium falciparium

4.1

20.2

0.5

15.2

33.2

8.5

Clostridium pasteurianum

2.4

32.8

0.8

6.0

52.5

1.0

Pyrococcus horikoshii

30.3

20.4

1.0

18.0

44.9

10.1

Thermus aquaticus

13.7

1.4

1.4

3.2

2.0

43.0

Arabidopsis thaliana

10.9

18.4

6.0

9.8

12.6

5.2

Although strains of the BL21-Gold series are the most convenient and effective bacterial hosts for protein production presently available, expression of heterologous recombinant genes in E. coli is frequently limited when the codon usage in the recombinant gene differs from host cells (Table 1). Forced high-level expression of a gene, with codons that are rarely used by E. coli, results in depletion of the internal tRNA pools.² As a consequence, translation of the recombinant RNA species is delayed, leading to its enhanced degradation. The result of rare codon usage is most frequently seen as ineffective to nondetectable production of the target protein and the formation of aborted translation products that have the appearance of proteolytic degradation. In addition, rare codons may also affect the quality of the produced proteins by inducing translation frameshifts, codon skipping, and frequent misincorporation of amino acids. Misincorporation rates of up to 40% have been reported at specific rare arginine codons.³

The deleterious effects of rarely used codons have been most thoroughly documented for the arginine codons AGA and AGG, which are the rarest codons in E. coli. However, codons for isoleucine (AUA), leucine (CUA), and proline (CCC) have also been demonstrated to affect protein production and quality in E. coli hosts. The expression of affected recombinant genes can be salvaged either by synonymous replacement of the rare codons with more frequently used codons using site-directed mutagenesis or by supplementing the host with extra copies of the cognate tRNA genes.

Now, Stratagene has generated BL21-CodonPlus-RIL chemically competent cells that carry extra copies of the argU, ileY, and leuW tRNA genes. The tRNAs encoded by these genes recognize the AGA/AGG, AUA, and CUA codons, respectively. To demonstrate their ability to facilitate the expression of heterologous recombinant genes that are only marginally expressed in conventional BL21 strains, BL21-CodonPlus(DE3)-RIL cells were transformed with vectors expressing genes fromhuman, yeast, and archaeal origins using the T7 RNA polymerase responsive promoter. IPTG-induced expression of all recombinant genes greatly increased in the BL21-CodonPlus(DE3)-RIL cells, compared to their expression in BL21-Gold(DE3) (Figure 1).

Fig.1

Expression of Difficult Proteins

To demonstrate the efficiency of BL21-CodonPlus-RIL cells to facilitate protein production in genes affected by poor codon usage, we compared the expression of a codon-adjusted form of the human cardiac troponin-T gene in conventional BL21-GoldDE3 cells (Figure 1,hcTnT-mut) with expression of the wild-type (hcTnT-wt) form in BL21-CodonPlus(DE3)-RIL cells. The expression of hcTnT in E. coli was previously shown to be solely restricted by the presence of two rare arginine codon pairs.⁴ This was demonstrated by synonymous replacement of the rare arginine codon pairs with frequently used arginine codons (hcTnT-mut), which results in efficient production of hcTnT-mut in E. coli. In Figure 1, BL21-CodonPlus(DE3)-RIL cells express hcTnT-wt at the same level observed for the codon-adjusted hcTnT-mut in conventional BL21-Gold(DE3) cells. Hence, using BL21-CodonPlus(DE3)-RIL strains are as efficient as employing the conventional and time-consuming method of synonymous replacement of rare codons to achieve high-level expression of codon-usage restricted recombinant genes.

Fig.2

The simultaneous expression of several rare tRNA genes in BL21-CodonPlus(DE3)-RIL not only provides a generic host for expression of recombinant genes affected by a range of rare codons but is also necessary for some recombinant genes to achieve a high level of expression. In Figure 2, using an archaeal gene as an example, efficient production of this recombinant gene in E. coli hosts requires the simultaneous presence of functional extra copies of both argU and ileY tRNA genes. Variants of the BL21-CodonPlus(DE3)-RIL strains lacking either the argU or ileY gene, while retaining the other tRNA genes, failed to produce the target gene at the same level as observed in the BL21-CodonPlus(DE3)-RIL strain. The functional expression argU and ileY genes in each of thevariant strains were confirmed using expression constructs, which are dependent only on either argU function (hcTnT) or ileY function (3xi) for efficient expression. The artificial construct 3xi is derived from a well-expressed recombinant gene by insertion of three AUA codons (ile) at the N-terminus.

The host cells did not suffer any obvious deleterious effects from the extra copies of tRNA genes. Growth rates and transformation efficiencies were comparable to the parental BL21-Gold cells. When comparing BL21-CodonPlus(DE3)-RIL cells with BL21-Gold(DE3) cells, production of recombinant genes like Jun-N-terminal kinase (JNK) or l-phosphatase, that are well expressed in conventional BL21 cells, were indistinguishable between the two strains (Figure 1).

Conclusions

Stratagenes new BL21-CodonPlus series of competent cells allows high-level expression of recombinant genes that are restricted in their expression in conventional E. coli due to codon usage bias. Many proteins can be expressed in BL21-CodonPlus-RIL cells instead of moving into a different expression system.

Acknowledgements

The authors thank Alan Greener, Connie Hansen, and Xiuyuan Hu.

REFERENCES

1. Jerpseth, M., et al. (1998) Strategies 11: 3-4.

2. Kane, J.F. (1995) Curr. Opin. Biotechnol. 6: 494-500.

3. Forman, M.D., et al. (1998) Protein Sci. 7: 500-503.

4. Hu, X., et al. (1996) Protein Expr. Purif. 7: 289-293.

* Patent pending
** Patents pending