Nucleotide sequence selection for increased expression of heterologous genes in Escherichia coli

Fibroblast growth factor 19 increases metabolic rate and reverses dietary and leptin-deficient diabetes

Hormonal control of metabolic rate can be important in regulating the imbalance between energy intake and expenditure that underlies the development of obesity. In mice fed a high-fat diet, human fibroblast growth factor 19 (FGF19) increased metabolic rate [1.53 +/- 0.06 liters O(2)/h.kg(0.75) (vehicle) vs. 1.93 +/- 0.05 liters O(2)/h.kg(0.75) (FGF19); P < 0.001] and decreased respiratory quotient [0.82 +/- 0.01 (vehicle) vs. 0.80 +/- 0.01 (FGF19); P < 0.05]. In contrast to the vehicle-treated mice that gained weight (0.14 +/- 0.05 g/mouse.d), FGF19-treated mice lost weight (-0.13 +/- 0.03 g/mouse.d; P < 0.001) without a significant change in food intake. Furthermore, in addition to a reduction in weight gain, treatment with FGF19 prevented or reversed the diabetes that develops in mice made obese by genetic ablation of brown adipose tissue or genetic absence of leptin. To explore the mechanisms underlying the FGF19-mediated increase in metabolic rate, we profiled the FGF19-induced gene expression changes in the liver and brown fat. In brown adipose tissue, chronic exposure to FGF19 led to a gene expression profile that is consistent with activation of this tissue. We also found that FGF19 acutely increased liver expression of the leptin receptor (1.8-fold; P < 0.05) and decreased the expression of acetyl coenzyme A carboxylase 2 (0.6-fold; P < 0.05). The gene expression changes were consistent with the experimentally determined increase in fat oxidation and decrease in liver triglycerides. Thus, FGF19 is able to increase metabolic rate concurrently with an increase in fatty acid oxidation.

Secretory production of Arthrobacter levan fructotransferase from recombinant Escherichia coli

Levan fructotransferase (LFTase) from Arthrobacter ureafaciens K2032 was expressed with N-terminal fusion of a LacZ-derived secretion motif (TMITNSSSVP) using the lac promoter system in recombinant Escherichia coli JM109 [pUDF-A81]. In flask cultures, recombinant enzyme activity was detected in culture media, and sequence analysis of N-terminal residues showed that about 40% of the extracellular recombinant LFTase had an authentic N-terminus. In a fed-batch bioreactor containing recombinant E. coli at high cell concentrations (OD(600)>200), the extracellular LFTase accumulated to 46000 U ml(-1) (approximately 2.0 g l(-1)) which was almost 40% of total (intra- and extracellular) recombinant LFTase. The synthesized recombinant enzyme was secreted soon after gene expression was induced by IPTG. Prolonged high secretion caused cell lysis and growth inhibition during the production phase in fed-batch cultures. When lactose was added by continuous feed mode, the secretion of recombinant LFTase and hence the cell lysis were significantly delayed in spite of the increased synthesis level. Therefore the induced cell culture of recombinant E. coli could grow up to a much higher cell concentration with continuing recombinant enzyme synthesis. In the case of the controlled feed of lactose, the maximum activities (U ml(-1)) of total and extracellular LFTase were nearly 100% and 70% higher, respectively.

Translational level is a critical factor for the secretion of heterologous proteins in Escherichia coli

A method for enhancing the secretion of heterologous proteins in Escherichia coli by optimizing, as opposed to simply maximizing, the translational level of a given protein is described. Random alteration of the translational initiation region (TIR) of the Heat-Stable Enterotoxin II (STII) signal sequence resulted in a library of vectors with varied translational strengths. Subsequent screening of this library using E. coli alkaline phosphatase as a reporter led to the selection of several TIR variants covering a 10-fold range of translational strength. These TIR variants, in combination with several previously generated variants, are shown to dramatically improve the secretion of a number of heterologous proteins. In fact, the heterologous proteins tested required a narrow translational range for optimal high-level secretion into the periplasm. Interestingly, the secretion of two native E. coli proteins was unaffected by TIR strength when tested over an identical range. The dependence of secretion on a narrow translational level demonstrates its critical role in the secretion of heterologous proteins in E. coli.

Purification and characterization of active and stable recombinant plasminogen-activator inhibitor accumulated at high levels in Escherichia coli

Plasminogen-activator inhibitor type 1 (PAI-1), the primary physiological inhibitor of tissue-type plasminogen activator, is an unusual member of the serine protease inhibitor (serpin) superfamily in that it spontaneously converts to a latent form lacking activity. This latent form can be reactivated by denaturation and refolding, but the activation is usually incomplete and often leads to aggregation of the protein. In this study we have developed a high-level expression system that leads to the accumulation of PAI-1 at 30-50% total microbial protein. We have developed a single-step purification protocol which can be completed in a few hours, yielding approximately 20 mg purified recombinant PAI-1/litre culture. The purified PAI-1 was 80-100% active and was stable upon incubation at 37 degrees C with a half-life of approximately 48 h. At 20 degrees C, PAI-1 activity was stable for a week and at 4 degrees C it retained its activity completely for up to two months. Freezing caused significant loss of activity. The stability of PAI-1 activity was found to be dependent on pH and ionic strength, being most stable at pH 5.6 and at an ionic strength of 1 M salt. We show that by a combination of high-level expression and rapid purification under optimum conditions, it is possible to produce active and stable PAI-1 in high yield.

Solubility and secretability

The solubility and secretability of proteins can often be affected by extremely small changes in their primary structure. Attempts to determine empirical rules for the alteration of protein structure to improve either of these characteristics have met with only partial success. Those (mostly serendipitous) improvements in solubility that have been obtained via mutagenesis cannot be considered to be 'protein engineering'. The most successful examples where directed mutagenesis has been used to alter protein solubility, hemoglobin and insulin, have relied on established crystal structures and a wealth of data about the relationship between sequence and structure of the targeted protein. Currently, optimizing culture growth conditions by trial and error remains the fastest way to improve expression.