Expression of two green fluorescent protein variants in citrate-buffered media in Pichia pastoris

Abstract

The Aequorea victoria green fluorescent protein (GFP) has become one of the most frequently employed molecular reporters. A variety of mutational methods have created a large number of GFP variants with differing pH sensitivity [1], differing spectra, and optimized expression [2]. Many of the recent mutational approaches have focused on species-specific codon optimization. First established in 1980 [3], the idea that organisms do not display a random pattern of synonymous codon usage has been validated by the large amount of recent genome sequencing projects [4]. A substantial number of studies have examined the effects of varying codon bias in heterologous protein expression systems under the assumption that disparate patterns of codon bias in the transgene and the expression host will have a significant impact on the levels of recombinant protein produced. It has been suggested that the human codon-optimized enhanced green fluorescent protein (EGFP; Clontech, Palo Alto, CA) is unsuitable for expression in yeast host systems due to such preferred codon discrepancy. However, several studies have demonstrated successful expression in yeast, mainly in Saccharomyces cerevisiae [5–7]. Therefore, we compared the expression of two GFP variants, red-shifted GFP (RSGFP from pRSGFP-C1; Clontech), and EGFP, in the methylotrophic yeast Pichia pastoris. Unlike EGFP, RSGFP is not human codon-optimized; however, like EGFP, RSGFP does have an excitation wavelength shifted toward the red end of the spectrum. Both variants also exhibit greater fluorescence intensity than wildtype GFP: EGFP fluoresces 35 times greater and RSGFP fluoresces 4–6 times greater. We chose to examine the expression of both variants in citrate-buffered rather than phosphate-buffered media to determine whether either variant could serve as a fusion partner with proteins reported to exhibit inhibition in the presence of phosphate. Glucocerebrosidase (acid b-glucosidase, EC 3.2.1.45) activity, for example, has been suggested to be inhibited by phosphate ions [8]. Based on the codon bias theory, we expected to see moderate amounts of RSGFP and minimal amounts of EGFP produced. To create the RSGFP and EGFP constructs, the variant cDNAs were PCR-amplified with the following primers: 50-TAGAATTCCCGGTCGCCACCATG-30 and50-TGTTACAGGGCCCGCGGTTCAGTCGAC-30 for RSGFP; 50-AACGGTCGAATTCATGGTGAGCA AGGG-30 and 50-TATGATCTGAATTGCCGGCCGC TTTACTT-30 for EGFP. PCR amplification was performed with an initial denaturation at 94 C for 2.5min, followed by 30 cycles of 94 C for 1min, 60 C (for RSGFP) or 56 C (for EGFP) for 1.5min, and 72 C for 1.5min. Reactions were carried out with 1ll of template (pRSGFP-C1 or pEGFP-N1 vector), 5ll of 10 PCR buffer, 5ll of 25mM MgCl2, 5ll of 2.5mM dNTPs, and 4 ll forward and reverse primers in a total volume of 50 ll. These amplified cDNAs were digested with EcoR1 and ligated into the vector pPICZaA (Invitrogen, San Diego, CA) prior to electroporation into bacterial cells. Transformantswere selected by resistance toZeocin (0.1mg/ml; Invitrogen) on low-salt Luria–Bertani plates (pH 7.5). Clones were confirmed by PCR amplification and sequenced to ensure the fidelity of the PCR and cloning steps. True positive pPICZaA–RSGFP and pPICZaA– EGFP clones were isolated, purified, and linearized at the BstX1 restriction site prior to electroporation into GS115 P. pastoris cells according to the Pichia Expression Manual protocol (Invitrogen). Transformants were Analytical Biochemistry 311 (2002) 193–195