Secretion of a foreign protein from budding yeasts is enhanced by cotranslational translocation and by suppression of vacuolar targeting

Abstract

Background: Budding yeasts are often used to secrete foreign proteins, but the efficiency is variable. To identify roadblocks in the yeast secretory pathway, we used a monomeric superfolder GFP (msGFP) as a visual tracer in Saccharomyces cerevisiae and Pichia pastoris.

Results: One roadblock for msGFP secretion is translocation into the ER. Foreign proteins are typically fused to the bipartite α-factor secretion signal, which consists of the signal sequence followed by the pro region. The α-factor signal sequence directs posttranslational translocation. For msGFP, posttranslational translocation is inefficient with the α-factor signal sequence alone but is stimulated by the pro region. This requirement for the pro region can be bypassed by using the Ost1 signal sequence, which has been shown to direct cotranslational translocation. A hybrid secretion signal consisting of the Ost1 signal sequence followed by the α-factor pro region drives efficient translocation followed by rapid ER export. A second roadblock for msGFP secretion in S. cerevisiae occurs during exit from the Golgi, when some of the msGFP molecules are diverted to the vacuole. Deletion of the sorting receptor Vps10 prevents vacuolar targeting of msGFP at the expense of missorting vacuolar hydrolases such as carboxypeptidase Y (CPY) to the culture medium. However, a truncation of Vps10 blocks vacuolar targeting of msGFP while permitting CPY to be sorted normally.

Conclusions: With budding yeasts, if the secretion or processing of a foreign protein is poor, we recommend two options. First, use the Ost1 signal sequence to achieve efficient entry into the secretory pathway while avoiding the processing issues associated with the α-factor pro region. Second, truncate Vps10 to suppress diversion to the vacuole. These insights obtained with msGFP highlight the value of applying cell biological methods to study yeast secretion.

Figure 1

Effects on secreted msGFP constructs of including the α-factor pro region after the α-factor signal sequence. (A) S. cerevisiae cells, either expressing wild-type Erv29 (“WT”) or carrying an erv29Δ allele, were engineered to express msGFP fused to either the α-factor signal sequence alone (pre-αf-msGFP) or the complete α-factor secretion signal (pre-pro-αf-msGFP). Samples of the culture medium and cells were analyzed by SDS-PAGE, immunoblotting, and chemiluminescence to detect msGFP. (B) The strains expressing wild-type Erv29 plus the indicated msGFP constructs were imaged by fluorescence microscopy to detect GFP, and by differential interference contrast (DIC) microscopy to detect the cells. Representative cells are shown. Exposure times for the fluorescence images were 100 msec. Scale bar, 2 μm.

Figure 2

Effects on ER-retained msGFP constructs of including the α-factor pro region after the α-factor signal sequence. S. cerevisiae cells carrying erv29Δ and htm1Δ alleles and expressing nuclear-targeted DsRed-Express2 were engineered to express msGFP with a C-terminal HDEL signal, fused to either the α-factor signal sequence alone (pre-αf-msGFP*-HDEL) or the complete α-factor secretion signal (pre-pro-αf-msGFP*-HDEL). Strains were imaged by fluorescence microscopy with 300 msec exposure times to detect GFP and DsRed, and by differential interference contrast (DIC) microscopy to detect the cells. Scale bar, 2 μm.

Figure 3

Effects on ER-retained msGFP constructs of including the α-factor pro region after the Ost1 signal sequence. This experiment was performed in parallel with that of Figure 2 with the same parameters, except that msGFP with a C-terminal HDEL signal was fused to either the Ost1 signal sequence alone (pre-Ost1-msGFP*-HDEL) or the Ost1 signal sequence followed by the α-factor pro region (pre-Ost1-pro-αf-msGFP*-HDEL). Scale bar, 2 μm.

Figure 4

Effects on secreted msGFP constructs of including the α-factor pro region after the Ost1 signal sequence. (A) The analysis was performed as in Figure 1A, except that msGFP was fused to either the complete α-factor secretion signal (pre-pro-αf-msGFP), or the Ost1 signal sequence alone (pre-Ost1-msGFP), or the Ost1 signal sequence followed by the α-factor pro region (pre-Ost1-pro-αf-msGFP). The asterisk marks a band that may represent ER-localized msGFP molecules fused to the pro region. (B) The analysis was performed as in Figure 1B, except with the strains described in (A). Cells were stained with FM 4-64 to visualize the vacuolar membrane. Exposure times for the fluorescence images were 3oo msec. Scale bar, 2 μm. (C) The analysis was performed as in (B), except that vacuoles were not labeled, and the cells carried an erv29Δ allele and expressed nuclear-targeted DsRed-Express2. Exposure times for the fluorescence images were 300 msec. Scale bar, 2 μm.

Figure 5

Effects on secreted msGFP constructs of deleting or mutating Vps10. (A) S. cerevisiae cells, either expressing wild-type Vps10 (“WT”) or carrying a vps10-104 or vps10Δ allele, were engineered to express msGFP fused to the Ost1 signal sequence (pre-Ost1-msGFP). Exposure times for the fluorescence images were 500 msec. Scale bar, 2 μm. (B) Intracellular and extracellular CPY for the indicated strains was analyzed by SDS-PAGE and immunoblotting as described in Methods. (C) msGFP secretion was analyzed as in Figure 1A for the indicated strains, except that immunoblotting was performed with the quantitative procedure described in Methods. The experiment was done twice, with three replicates each time. Immunoblots from one of the experiments are shown. Error bars represent s.e.m.

Figure 6

Effects of the different signal sequences on secretion of msGFP from P. pastoris . Strains of P. pastoris were engineered to express msGFP fused to either the α factor signal sequence alone (pre-αf-msGFP), or the complete α factor secretion signal (pre-pro-αf-msGFP), or the Ost1 signal sequence alone (pre-Ost1-msGFP), or the Ost1 signal sequence followed by the α factor pro region (pre-Ost1-pro-αf-msGFP). Extracellular and intracellular msGFP was analyzed as in Figure 1A.

Figure 7

Effects of the different signal sequences on intracellular accumulation of msGFP in P. pastoris . Strains of P. pastoris expressing the indicated msGFP constructs (see Figure 6) were imaged by fluorescence microscopy to detect GFP, and by differential interference contrast (DIC) microscopy to detect the cells. Representative groups of cells are shown. Exposure times for the fluorescence images were 200 msec. Scale bar, 2 μm.

Figure 8

Predicted structures of domains 1 and 2 of Vps10 compared to the known structure of sortilin. The protein sequences of S. cerevisiae Vps10 domain 1 (residues 22–737) and domain 2 (residues 719–1393) were submitted to the protein homology/analogy recognition engine Phyre (http://www.sbg.bio.ic.ac.uk/phyre/html/) [54], which detected the similarity to sortilin and generated PDB files for the predicted tertiary structures. A file for the experimentally determined structure of the human sortilin lumenal domain (PDB ID 3F6K) [53] was downloaded from the National Center for Biotechnology Information. To generate the images shown, these PDB files were opened with MacPyMOL using the default settings.