High-yield expression and purification of recombinant proteins in bacteria: a versatile vector for glutathione S-transferase fusion proteins containing two protease cleavage sites.

small standard deviations of 0.11 and 0.39 pg/ml, respectively. Additional studies demonstrated that the recovery of 0.13, 0.35, 0.62, 2.7, 24, and 260 pg/ml VEGF in 10% pooled mouse EDTA plasma was 100– 120%; the average recoveries of 2, 5, and 20 pg/ml VEGF in 10% individual human serum (n 5 4, with 98–130 pg/ml endogenous VEGF) were 83, 94, and 96%, respectively. This immuno-PCR therefore can measure 2 pg/ml VEGF in human or mouse serum. In another study, VEGF concentrations in 40 normal humans were measured by immuno-PCR and fluorometric ELISA using the same capture and detection antibodies (9). Although VEGF concentrations measured by immuno-PCR tend to be higher and further studies are needed to understand the discrepancy, a good correlation was seen between the two methods (Fig. 3). In summary, we have developed an immuno-PCR using real-time PCR for VEGF. Using real-time PCR improves the DNA quantitation and eliminates postPCR manipulations of samples. Although the assay precision needs to be improved further, the sensitivity of this assay (0.20 pg/ml or 0.005 pM) is a fivefold improvement over the sensitive fluorometric ELISA (9). The assay background seen in the blank sample (Fig. 2A) may be lowered by reducing the amount of streptavidin coated on the tubes or using better blocking or washing methods. If the background signal is lowered, the efficiency of DNA–antibody conjugating is improved, and the PCR instrument variation is reduced, it is possible to speculate that the assay sensitivity and precision would also improve.