High-throughput screening of soluble recombinant proteins

Abstract

The aims of high-throughput (HTP) protein production systems are to obtain well-expressed and highly soluble proteins, which are preferred candidates for use in structure-function studies. Here, we describe the development of an efficient and inexpensive method for parallel cloning, induction, and cell lysis to produce multiple fusion proteins in Escherichia coli using a 96-well format. Molecular cloning procedures, used in this HTP system, require no restriction digestion of the PCR products. All target genes can be directionally cloned into eight different fusion protein expression vectors using two universal restriction sites and with high efficiency (>95%). To screen for well-expressed soluble fusion protein, total cell lysates of bacteria culture ( approximately 1.5 mL) were subjected to high-speed centrifugation in a 96-tube format and analyzed by multiwell denaturing SDS-PAGE. Our results thus far show that 80% of the genes screened show high levels of expression of soluble products in at least one of the eight fusion protein constructs. The method is well suited for automation and is applicable for the production of large numbers of proteins for genome-wide analysis.

Fig. 1.

Moleclular cloning strategy. Four PCR primers and reactions were used in two separate tubes. An equal amount of the two PCR products were mixed, and then the 5` ends were phosphorylated with T4 polynucleotide kinase. After denaturing (95°C for 5 min) and renaturing (65°C for 10 min), ∼25% of the final products carry EcoRI (5`) and XhoI (3`) cohesive ends and are ready for ligation with the vectors.

Fig. 2.

Recombinant DNA plasmids purified from JM109(DE3). Eight different fusion protein expression vectors are indicated above. Two independent clones from each construct were isolated for characterization (lanes A and B). Plasmid DNAs were purified in a 96-well format using Millpore's Motage plasmid mini-prep kit; 3–5 μL mini-prep DNA was restriction digested with EcoRI and XhoI and separated in 0.8% agarose gel. A 1-kb DNA ladder (from MBI Fermentas, USA) was used as marker (M) and shown in the far left lane. The expected sizes (in base pair) of the desirable restriction fragments of two different target genes are indicated on the right of the figure.

Fig. 3.

Analysis of budding yeast Csm2 (YIL132C) fusion proteins. Samples of total proteins and soluble protein fractions were separated on a 10% SDS-PAGE under reducing conditions and stained with Coomassie Blue. Lane 1, whole cell lysates of induced cells; lane 2, whole cell lysates of uninduced cells; lane 3, soluble proteins with induction. Eight different fusion proteins are indicated above. The molecular weight standards are shown in the center and labeled on the left (×1,000). NusA and MBP fusion proteins show high solubility (indicated by arrows below the lanes of soluble protein fractions); on the other hand, GST and Trx fusion proteins are well induced but not soluble (indicated by open triangles).

Fig. 4.

(A) Statistical analysis of soluble protein ratio obtained in at least one of the eight expression constructs. (B) Eight different gene fusions and their effects were also compared. A total of 40 different genes were tested in this study. Well-induced and highly soluble fusion proteins were identified visually by comparing the relative density of protein bands in SDS-PAGE as shown in Figure 3 ▶.

Fig. 5.

SDS-PAGE analysis of purified proteins. The gel shows typical yields (5–20 mg per liter of Luria-Bertani [LB] culture) and purity (∼90%) obtained from two steps of affinity purification. The database accession or open reading frame number of the expressed proteins and their fusion tags are indicated. The molecular weight standards are labeled on the left.