Genome-wide SWAp-Tag yeast libraries for proteome exploration

Abstract

Yeast libraries revolutionized the systematic study of cell biology. To extensively increase the number of such libraries, we used our previously devised SWAp-Tag (SWAT) approach to construct a genome-wide library of ~5,500 strains carrying the SWAT NOP1promoter-GFP module at the N terminus of proteins. In addition, we created six diverse libraries that restored the native regulation, created an overexpression library with a Cherry tag, or enabled protein complementation assays from two fragments of an enzyme or fluorophore. We developed methods utilizing these SWAT collections to systematically characterize the yeast proteome for protein abundance, localization, topology, and interactions.

Figure 1. Library generation of genome wide N’ yeast collections by SWAT technology.

Composition of the SWAT full-genome library. Proteins were tagged at the N terminus with an acceptor SWAT module containing the NOP1 constitutive promotor and GFP. Proteins predicted to harbor a mitochondrial targeting signal (MTS) were tagged with a specific acceptor SWAT module containing an MTS upstream of GFP (MTS_Su9). Proteins predicted to harbor a signal peptide (SP) were previously created with a similar acceptor SWAT module containing a SP upstream of GFP (SP_Kar2). Numbers next each tag type name denote how many strains were created/attempted. The SWAT parental library underwent native promoter/regulation swapping: NOP1pr-GFP turned into NATIVEpr-GFP, NOP1pr-MTS_Su9-GFP turned into NATIVEpr-MTS_native-GFP and NOP1pr-SP_Kar2-GFP turned into NATIVEpr-SP_native-GFP. The SWAT parental library was also swapped to create five additional libraries – TEF2pr-mCherry, TEF2pr-VC, CET1pr-VN, NATIVEpr-DHFR F[1,2] and NATIVEpr-DHFR F[3].

Figure 2. Genome wide N’ tagged collections enables investigation of abundance regulation and localization assignment.

(a) Histogram of the expression levels of fluorophore-protein fusions of the TEF2pr-mCherry, NATIVEpr-GFP and NOP1pr-GFP libraries. a.u., arbitrary units. (b) Scatter plots showing the correlation between protein abundance of NATIVEpr-GFP tagged strains versus C’ GFP tagged strains (left) or protein abundance under generic regulation (NOP1pr-GFP) (right). R represents two-sided Spearman correlation test score. a.u., arbitrary units. (c) Spearman correlation scores of the expression levels of fluorophore-protein fusions of the TEF2pr-mCherry, NATIVEpr-GFP and NOP1pr-GFP libraries to mRNA abundance as measured by RNA-seq , protein translation rates as measured by ribosome profiling , mRNA half-lives , protein half-lives and protein abundance as measured by mass spectrometry . (d) Comparison of N’ NOP1pr-GFP library localization assignments to assignments made from the C′-tagged library ,. A complete list of localization assignments is presented in Supplementary table 1. Quantitation of abundance was preformed once. Strains with a final abundance score lower than 1, were excluded from the data analysis and Spearman correlation tests.

Figure 3. Characterization of the mitochondrial proteome.

(a) New mitochondrial proteins not containing an N’ MTS that could only be visualized when N’ tagged with NOP1pr-GFP (Left) or were already visible under the NATIVEpr-GFP (Right). (b) The NATIVEpr-(NATIVE)MTS-GFP library (Not including Su9-MTS) uncovered new mitochondrial proteins containing an N’ MTS. (c) Several proteins can still be seen on mitochondria even after swapping their NOP1pr-Su9MTS-GFP cassette to a TEF2pr-mCherry without an MTS, suggesting alternate targeting information is found in these proteins. (d) MTS and TMD prediction analysis for proteins showing mitochondrial localization with the NOP1pr-GFP, NATIVEpr-GFP, and/or C’ GFP tags. Cx(9)C - cysteine-rich domain , β-barrel domain . All scale bars are 5µm. Imaging of strains was performed a single time. Images represent entire field.

Figure 4. Protein-fragment complementation libraries enable systematic analysis of protein-protein interactions.

(a) Scheme of peroxisome DHFR PCA. Four yeast strain arrays were constructed, each representing 89 peroxisome-related genes tagged with either DHFR F[1,2] or DHFR F[3] fragments of methotrexate-resistant DHFR at either their N’ or C’. These strains were mated to test every protein pair in all permutations. An interaction between two proteins brings the DHFR fragments together resulting in their folding, reconstitution of activity and growth of strains in the presence of methotrexate. (b) The peroxisome interactome. Blue squares represent interactions discovered by the C’ tagged strains alone. Yellow squares represent interactions that required at least one N’ tagged strain for their discovery. The white-to-red spectrum squares correspond to the Z scores for the interaction. Only proteins with at least one interaction are depicted. A complete list of Z scores for the interaction is presented in Supplementary table 4.

Figure 5. Protein-fragment complementation libraries enable systematic analysis of membrane protein topology.

(a) Scheme of split Venus analysis to determine topology for the N’ of proteins. The N’ TEF2pr-VC library was mated with a strain containing a cytosolic VN, and only if the N’ of a membrane protein faces the cytosol will complementation occur and a fluorescent signal appear suggesting the topology of the protein’s N’. A complete list of topology assignments is presented in Supplementary table 6. (b) Correlation graph between the TEF2pr-mCherry library and the TEF2pr-VC library with a cytosolic-VN (only “in” assignments). Correlation score is two-sided Spearman correlation. Quantitation of abundance was preformed once. Strains with a final abundance score lower than 1, were excluded from the data analysis and Spearman correlation tests.