Monitoring Protein Dynamics in Protein O-Mannosyltransferase Mutants In Vivo by Tandem Fluorescent Protein Timers

Abstract

For proteins entering the secretory pathway, a major factor contributing to maturation and homeostasis is glycosylation. One relevant type of protein glycosylation is O-mannosylation, which is essential and evolutionarily-conserved in fungi, animals, and humans. Our recent proteome-wide study in the eukaryotic model organism Saccharomyces cerevisiae revealed that more than 26% of all proteins entering the secretory pathway receive O-mannosyl glycans. In a first attempt to understand the impact of O-mannosylation on these proteins, we took advantage of a tandem fluorescent timer (tFT) reporter to monitor different aspects of protein dynamics. We analyzed tFT-reporter fusions of 137 unique O-mannosylated proteins, mainly of the secretory pathway and the plasma membrane, in mutants lacking the major protein O-mannosyltransferases Pmt1, Pmt2, or Pmt4. In these three pmtΔ mutants, a total of 39 individual proteins were clearly affected, and Pmt-specific substrate proteins could be identified. We observed that O-mannosylation may cause both enhanced and diminished protein abundance and/or stability when compromised, and verified our findings on the examples of Axl2-tFT and Kre6-tFT fusion proteins. The identified target proteins are a valuable resource towards unraveling the multiple functions of O-mannosylation at the molecular level.

Keywords: O-mannosyl glycans; PMT1; PMT2; PMT4; Saccharomyces cerevisiae; fluorescent protein timers; glycosylation; mannosyltransferase; protein turnover; secretory pathway; yeast.

Figure 1

Tandem fluorescent protein timer (tFT) screening. (a) Representation of a C-terminal tFT-fusion protein analyzed. The slow maturing mCherry and the fast maturing sfGFP are fused in tandem to the C-terminus of O-mannosylated proteins of interest. (b) Workflow of the screening of the selected fusion proteins. In brief, 137 individual tFT fusions of the tFT library established by Khmelinskii and coworkers [21] were selected based on the presence of O-mannosyl glycans on these proteins [7]. The 137 tFT query strains were crossed with pmt1Δ (MLY201), pmt2Δ (MLY202), or pmt4Δ (MLY204) mutants using synthetic genetic array methodology [24]. Haploid yeast strains carrying both genetic modifications (tFT-fusion and pmt deletion) were selected. mCherry/sfGFP ratio was calculated for each protein and used for comparison between wild-type and mutants.

Figure 2

Identification of proteins affected in pmt1Δ, pmt2Δ and pmt4Δ deletion mutants. (a) Volcano plots illustrating changes in tFT-fusion protein stability in the indicated mutant strains with regard to statistical significance of data as inferred from variance analysis. Plots show Δ-scores mCherry/sfGFP for changes in protein stability on the x-axis and the negative logarithm of p-values on the y-axis. Data were subset for relevance based on thresholds as indicated by red dashed lines (p-value < 0.1 and net Δ-score > 0.2). (b) Correlation of Δ-scores mCherry/sfGFP, as a measure of tFT-fusion protein stability and turnover (x-axis), and Δ-scores sfGFP, as a measure of change in protein abundance (y-axis). (a,b) Data referring to proteins further analyzed or discussed in this study are labeled in red. The example of Vrg4-tFT shows, that even minor effects could be reproducibly detected. (c) Heatmap with hierarchical clustering of Δ-scores mCherry/sfGFP from a subset of data that passes the thresholds of significance in at least one of the indicated mutant strains (p-value < 0.1 and net Δ-score > 0.2). Cytosolic and luminal orientation of the tFT reporter is indicated with C and L, respectively.

Figure 3

Analyses of Axl2-tFT protein. (a) Live fluorescence microscopy of both wild-type (WT; WT Axl2-tFT) and pmt4Δ (EZY107) cells expressing the Axl2-tFT. Prior to imaging, cells were stained with the vacuolar vital dye 7-amino-4-chloromethylcoumarin (CMAC). Scale bar, 5 µm. (b) Membranes (equivalent to 1 OD₆₀₀ units of yeast cells) from wild-type and pmt4Δ cells expressing Axl2 C-terminally tagged with either HA or tFT (strains MGY69, MGY72, WT Axl2-tFT, and EZY107), were resolved on 8% polyacrylamide gels and subjected to Western blot analysis using anti-HA and anti-GFP antibodies, respectively. Full length form of tagged Axl2 (black arrows) is less abundant in pmt4Δ than in corresponding wild-type cells. In the cells lacking Pmt4, C-terminal proteolytic fragments of the protein (white arrows) are detected. Differences in the apparent molecular masses of the full-length forms and proteolytic fragments observed for Axl2-HA and Axl2-tFT, respectively, correspond to the calculated mass difference (−60 kDa) of the tags. Asterisks indicate an unrelated cross-reactive band, present also in the membranes isolated from wild-type cells without tFT-fusion protein (YMaM330). As a loading control, the blot was subsequently incubated with anti-Sec61 antibody. Experiments have been replicated three times; representative results are shown.

Figure 4

Localization of Kre6-tFT protein. Live fluorescence microscopy of wild-type (WT Kre6-tFT) and pmt1Δ (EZY91) cells expressing the Kre6-tFT (a), and of wild-type (WT Mnn11-tFT) and pmt4Δ (EZY109) cells expressing the Mnn11-tFT (b). (a) In wild-type and pmt1Δ cells, the type II transmembrane protein Kre6-tFT (Figure 5a) is present in the ER, but mainly in the vacuole. (b) Localization of another type II transmembrane protein, Mnn11-tFT (Figure 5a), to the Golgi is not affected by the C-terminal tFT reporter. Lower abundance of old Mnn11-tFT (mCherry) in pmt4Δ cells when compared to wild-type strain confirms the destabilization of this protein observed upon decreased O-mannosylation (Figure 2c). Prior to imaging, cells were stained with the vacuolar vital dye CMAC. Scale bar, 5 µm.

Figure 5

Analyses of Kre6-tFT protein. (a) Topology model of Axl2, Kre6 and Mnn11 depicting the orientation of the tFT timer for type I and type II transmembrane proteins. (b,c) Cell lysates from wild-type (WT; WT Kre6-tFT) and pmt1Δ (EZY91) cells expressing Kre6-tFT were resolved on 8% polyacrylamide gels and subjected to Western blot analysis using anti-GFP antibodies. Pgk1 served as a loading control. (b) Steady state levels of Kre6-tFT. Cell lysates equivalent to 0.2 OD₆₀₀ units of yeast cells were analyzed. (c) Cycloheximide chase analysis as detailed in Materials and Methods. Cell lysates equivalent to 0.2 OD₆₀₀ units (1x; pmt1Δ) and 0.4 OD₆₀₀ units (2x; WT) of yeast cells were analyzed, to allow for better comparability. (b,c) Three forms of Kre6-tFT were detected as previously demonstrated for the native protein [33]. The major band is more abundant at steady state (b) and slower degraded (c) in mutant pmt1Δ. Experiments have been replicated at least two times; representative results are shown.