Chemicals and enzymes
Standard chemicals were purchased from Sigma Aldrich and sodium acetate was purchased from Scharlau, imidazole was purchased from PanReac AppliChem and IPTG was purchased from Fischer Bioreagents. Enzymes for standard cloning procedures were purchased from Thermo Fisher Scientific and New England Biolabs, respectively.
Construction of a fluorescence-based protein folding reporter
For construction of a protein folding sensor that reports on the formation of IB, the IbpA promoter (Genbank: LQ302077.1) from E. coli MG1655 was fused to either a stable (GFP-mut3; GenBank: LQ302079.111) or a destabilized version of GFP (GFP-ASV; GenBank: LQ302078.111). The GFP-ASV and GFP-mut3 were amplified by PCR using primer pairs and templates as indicated in Supplementary Table 1 and Supplementary Table 2. PCR products were cloned into pSEVA441 (GenBank: JX560339.1) using the XbaI and SpeI restriction sites, resulting in either pSEVA441-GFP-ASV or pSEVA441-GFP-mut3. The E. coli lbpA promoter was amplified by PCR (Supplementary Table 1) and cloned via the PacI and XbaI restriction sites into pSEVA441-lbpAp-GFP-ASV and pSEVA441-lbpAp-GFP-mut3, respectively. To generate pSEVA631(Sp)-lbpAp-GFP-ASV or pSEVA631(Sp)-lbpAp-GFP-mut3, the lbpAp-GFP reporter gene was subcloned via PacI and SpeI into the pSEVA631 (GenBank: JX560348.1). Finally, the gentamicin cassette of pSEVA631 was replaced by the spectinomycin cassette of pSEVA441 using the SpeI and PshAI restriction sites. All constructs were verified by Sanger sequencing.
Fusion of proteins with a fluorescent translation-sensor
A set of proteins were fused to the translation coupling cassette30 (GenBank: LQ302080.1) followed by mCherry (GenBank: LQ302081.1). The BRCT-domain of human Poly [ADP-ribose] polymerase 1 (PARP1-BRCT, GenBank: LQ302082.1), a truncated version of BRCT-domain of human breast cancer 1, early onset (BRCA1-BRCT, GenBank: LQ302085.1 the human cyclin-dependent kinase 4 inhibitor D (p19, GenBank: LQ302086.1), and protein E6 from human papillomavirus type 16 (GenBank: LQ302087.1) were amplified by PCR using the primers and templates as indicated in Supplementary Table 1. In addition, mCherry was amplified by PCR according to Supplementary Table 1. Each protein encoding DNA fragment was assembled with the mCherry-PCR fragment and NdeI and HindIII digested pET22b vector (Novagen), using a Gibson assembly reaction (New England Biolabs). The resulting expression vectors pET22b-XXX-trans-mCherry (XXX stands for the respective protein; see also Supplementary Table 2) comprise the coding sequence of the different proteins being linked via a C-terminal translation coupling cassette30 to the ORF of mCherry. All cloned constructs were confirmed by Sanger sequencing.
Cloning of NusA and SUMO fusion proteins
For analyzing the impact of NusA and SUMO protein-tags on expression and translation levels of either PARP1-BRCT, BRCA1-BRCT, p19, or E6, proteins were N-terminally fused to NusA (GenBank: LQ302088.1) and SUMO (GenBank: LQ302089.1), respectively59,60. Thereby, NusA and SUMO were amplified by PCR using the primers indicated in Supplementary Table 1 and inserted into pET22-XXX-trans-mCherry via the NdeI restriction site. The final protein expression reporter plasmids named pET22b-NusA-XXX-trans-mCherry and pET22b-SUMO-XXX-trans-mCherry (XXX stands for the respective protein; see also Supplementary Table 2), respectively, were all verified by sequencing.
Impact of plasmid copy number and GFP stability on protein folding reporter assay sensitivity
The impact of the vector copy number and intracellular turnover rate of GFP, respectively, on the protein folding reporter system was analyzed to optimize the readout sensitivity of the assay. Therefore, pSEVA631(Sp)-lbpAp-GFP-ASV and pSEVA631(Sp)-lbpAp-GFP-mut3 (pBBR1 origin), as well as pSEVA441-lbpAp-GFP-ASV and pSEVA441-lbpAp-GFP-ASV (ColE1 origin) (constructed as described above), were co-transformed with pET22b in E. coli Rosetta2TM(DE3)pLysS (Novagen®). Transformants were selected on LB plates containing 25 µg/mL chloramphenicol, 50 µg/mL spectinomycin, and 100 µg/mL ampicillin. Single clones were inoculated in LB medium supplemented with the corresponding antibiotics and grown at 37 °C and 300 rpm to an OD600 of 0.5. IB formation in E. coli was induced by performing a heat-shock for 10 min at 42 °C. After heat shock, cells were grown for an additional 2.5 h at 37 °C and 300 rpm. Induction of the lbpAp promoter by IBs in single cells was monitored over time by changes of the GFP signal using flow cytometry (Instrument: BD FACS-Aria™SORP cell sorter; Laser 1: 488 nm: >50 mW, Filter: 505LP, 530/30-nm FITC, Laser 2: 561 nm: >50 mW; Filter: 600LP, 610/20-nm PE-Texas Red®). As control, the GFP signal in un-induced cells was monitored for each time point. The GFP (FITC-A, X-mean) values at each time point analyzed using the FlowJo V10 software were normalized to the corresponding background GFP signal.
To further investigate the impact of GFP stability on the sensitivity of the lbpAp-GFP reporter gene assay, pSEVA631(Sp)-lbpAp-GFP-ASV and pSEVA631(Sp)-lbpAp-GFP-mut3, respectively, were co-transformed with either pET22b, pET22-PARP1-BRCT-trans-mCherry or pET22-BRCA1-BRCT-trans-mCherry into E. coli Rosetta2TM(DE3)pLysS (Novagen®). Transformants were selected on LB plates containing 25 µg/mL chloramphenicol, 50 µg/mL spectinomycin and 100 µg/mL ampicillin. Single clones were grown at 37 °C and 300 rpm in LB medium supplemented with the corresponding antibiotics. At OD600 of 0.5–0.7 the expression of the human proteins was induced by addition of 0.5 mM IPTG. Directly after induction, the growth temperature was changed to 30 °C. Induction of the lbpAp-GFP variants by misfolded proteins was analyzed 1 h after induction using flow cytometry as mentioned above. For data analysis the GFP-signal (FITC-A, X-mean) was normalized to the respective GFP-signal of the vector control. To investigate the applicability of the system in other E. coli strains we co-transformed the pSEVA631(Sp)-lbpAp-GFP-ASV together with either pET22-PARP1-BRCT-trans-mCherry or pET22-BRCA1-BRCT-trans-mCherry into E. coli K-12 MG1655 (DE3). Transformants were selected on LB plates containing 50 µg/mL spectinomycin and 100 µg/mL ampicillin. Single clones were grown at 37 °C and 250 rpm in LB medium supplemented with the corresponding antibiotics. At OD600 of 0.5–0.7 the expression of the human proteins was induced by addition of 0.5 mM IPTG. Directly after induction, the growth temperature was changed to 30 °C, and OD600, the mCherry-signal (577,610), and GFP-signal (485,528) was analyzed in a fluorescent plate reader 1 and 3 h after induction.
Determination of protein localization by fractionated cell disruption
Intracellular localization of proteins was further analyzed by fractionated cell disruption. Here, cells (from 1 mL culture) were harvested either 1 h (for immunoblot analysis) or 3 h (for Instant Blue staining) after induction of protein expression. The cell pellet was resuspended in 50 µL resuspension buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl; 10 mM EDTA, 1 × HP-protease inhibitor mix (Serva)) and cells were broken by repeated cycles of freeze and thaw. Afterwards, cells were adjusted to a final OD600 of 5 in resuspension buffer supplemented with benzonase (≥500 units; Sigma Aldrich). After 20 min incubation on ice, cells were spun down for 1 min at 500 × g to remove cell debris. The supernatant containing all soluble and insoluble proteins was transferred to a fresh reaction tube. An aliquot of the supernatant was taken, representing the total protein fraction (total). The remaining cell lysate was spun down for 15 min at 20,000 x g and the supernatant containing all soluble proteins was transferred into a new reaction tube (sol). The isolated fractions were separated on SDS-PAGE (RunBlue 4–20 %, Expedeon; NuPAGE®Bis-Tris gel 4–12%, Invitrogen) and analyzed by Instant Blue staining (Expedeon) and quantitative immunoblotting using an anti-His antibody (Novagen).
Dual-reporter system for simultaneous monitoring of protein translation and folding in single E. coli cells
To analyze the combined reporter system, pSEVA631(Sp)-lbpAp-GFP-ASV and the protein expression reporter plasmids (pET22b-XXX-trans-mCherry, pET22b-NusA-XXX-trans-mCherry, pET22b-SUMO-XXX-trans-mCherry) were co-transformed into chemically competent E. coli Rosetta2TM(DE3)pLysS (Novagen®). Transformants were selected on LB plates containing 25 µg/mL chloramphenicol, 50 µg/mL spectinomycin, and 100 µg/mL ampicillin. Single clones were grown in LB medium (supplemented with the corresponding antibiotics) at 37 °C and 300 rpm to an OD600 of 0.5–0.7 and expression of proteins was induced by addition of 0.5 mM IPTG. Directly after induction, the growth temperature was changed to 30 °C. Protein expression and folding was analyzed 1 h after induction using flow cytometry as mentioned above. For data analysis, GFP (FITC-A, X-mean) signal was normalized to the corresponding PARP1-BRCT signal.
To confirm signal of the translation reporter, protein expression levels were further analyzed by instant blue staining and quantitative immunoblotting using an anti-His-Antibody. Cell-disruption was performed by freeze and thaw cycles as described before and the total protein fractions as well as intracellular localization of the proteins were analyzed. Western Blot signal was quantified using the Image J software61.
Identification of PARP1-BRCT mutants with altered folding properties using FACS
To generate a PARP1-BRCT mutant library the PARP1-BRCT domain was randomly mutated, aiming at a mutation rate of 1–3 mutations per construct, using the GeneMorph II random mutagenesis kit (Agilent) according to manufacturer´s instructions. Primers and templates used for the reactions are indicated in Supplementary Table 1. A megawhop reaction was performed with the random mutated PCR product as megaprimer and pET22-PARP1-BRCT-trans-mCherry as template. The resulting linear DNA fragments were transformed into MegaX DH10B™ T1R Electrocomp™ cells (Invitrogen) and transformants were selected on LB plates supplemented with 100 µg/mL ampicillin. The colonies (library size >100,000) were pooled and the plasmids were directly purified without further growth.
The vectors pET22b, pET22-PARP1-BRCT-trans-mCherry, and the created pET22-PARP1-BRCT-trans-mCherry mutant library were transformed into electro-competent Rosetta2(DE3)pLysS cells harbouring the protein folding sensor (pSEVA631(Sp)-IbpAp-GFP-ASV). After recovery, transformants were directly inoculated into 2 mL LB medium containing 20 µg/mL chloramphenicol, 50 µg/mL spectinomycin, 100 µg/mL ampicillin, and grown overnight at 37 °C and 300 rpm. Cells were transferred into fresh medium and grown at 37 °C and 300 rpm to an OD600 of 0.5–0.7. Expression of proteins was induced by addition of 0.5 mM IPTG and the growth temperature of the culture was shifted to 30 °C. 1 h after induction, cells were analyzed by flow cytometry as mentioned above. 150,000 cells expressing a PARP1-BRCT mutant protein at wildtype level based on the translation sensor signal (Fig. 5A, gate 1), and which had an increased GFP signal (Fig. 5A, Gate 2) were sorted into 1 mL LB medium supplemented with antibiotics and grown overnight at 37 °C and 300 rpm. To further enrich the E. coli fraction harbouring proteins with altered folding properties, another round of protein expression and sorting (150,000 events) was carried out as described above.
The following day, the sorted cell population was again analyzed 1 h after induction of protein expression by flow cytometry. Subcellular localization of proteins in the sorted E. coli fraction was analyzed by Immunoblotting using an anti-His antibody as described above.
For NGS, plasmids were isolated from the sorted E. coli population. As control, plasmids were isolated from the PARP1-BRCT mutant library, which was used as starting material for sorting. Two 300 bp DNA fragments were amplified from the PARP1-BRCT library using a high-fidelity polymerase (primers as indicated in Supplementary Table 1). The amplified fragments were purified using AMPure XP beads (Beckman Coulter) to remove free primers and primer-dimer species. Both PCR-products were mixed in a one-to-one ratio.
Next, a PCR reaction was performed to attach Illumina sequencing adapters (Nextera XT Index Kit, Illumina) to the DNA fragments. For the reaction a KAPA HiFi HotStart Polymerase (Kapa Biosystems) was used. The resulting PCR products were purified with AMPure XP beads. The product size of the PCR reaction was verified on a Bioanalyzer DNA 1000 chip and the DNA was quantified using a Qubit® 2.0 Fluorometer. DNA fragments were normalized to 10 nM in 10 mM Tris pH8.5, 0.1% Tween 20. In order to reduce the background signal, the sample was spiked with 5% Phi-X control DNA (Illumina). The DNA was loaded onto the flow cell provided in the MiSeq Reagent kit v2, subjected to 300 cycles (Illumina), and sequenced on a MiSeq sequencing system (Illumina).
Enrichment analysis
The analysis was carried out using Enrich2 software43. However, due to issues running Enrich2 directly from raw fastq files, we converted the fastq files into Enrich2 compatible variant counts using python scripts. The scripts for doing this as well as an Enrich2 analysis config file are available at https://doi.org/10.11583/DTU.10265420. The script does the following: Reads were merged using FLASH v.1.2.1162 and mapped to the reference sequence using bowtie2 v.2.3.4.163. The SAM files that bowtie2 outputs are then parsed to create Enrich2 compatible variant count files.
Folding properties of PARP1-BRCT single mutants
To generate PARP1-BRCT single mutants, a two-fragment Gibson assembly reaction was performed. For each single mutant two overlapping DNA fragments were amplified by PCR using pET22b-PARP1-BRCT-trans-mCherry as template. Primer pairs are listed in Supplementary Table 3. Finally, the two DNA fragments were joined using Gibson Assembly® Cloning Kit (New England Biolabs) according to manufacturer’s instructions. The sequence of each single mutant was confirmed by sequencing. Resulting mutant constructs are listed in Supplementary Table 3.
To examine the translation levels and protein stability of PARP1-BRCT single mutants, each mutant construct (Supplementary Table 3) was co-transformed with pSEVA631(Sp)-lbpAp-GFP-ASV into chemically competent E. coli Rosetta2TM(DE3)pLysS (Novagen®). Protein expression was induced by addition of IPTG and protein translation and folding were analyzed by flow cytometry and quantitative immunoblotting as described before. To determine the percentage of soluble protein, the western blot signal was quantified using the Image J software.
Isolation of PARP1-BRCT-I33N single mutants with rescued folding properties using the dual-reporter system
A PARP1-BRCT-I33N library was generated as described before, using pET22b-PARP1-BRCT-I33N-trans-mCherry as template. The plasmids pET22b, pET22-PARP1-BRCT-I33N-trans-mCherry and the created pET22-PARP1-BRCT-I33N-trans-mCherry mutant library were transformed into electro-competent Rosetta2(DE3)pLysS cells harbouring the protein folding sensor (pSEVA631(Sp)-IbpAp-GFP-ASV). Protein expression was induced with IPTG and flow cytometry was performed as described above. 64 single clones that show protein expression (Fig. 7A, Pool 1, Gate 1) in combination with a decreased GFP signal (Fig. 7A, Pool 1; Gate 2) were sorted in 200 µl LB medium supplemented with antibiotics and grown to stationary phase at 37 °C and 300 rpm. To further enrich the E. coli fraction harbouring proteins with rescued folding properties, a pool of 150,000 cells was sorted (identical gating as for single clones) into 1 mL LB medium supplemented with antibiotics and grown again overnight at 37 °C and 300 rpm. Subsequently, a second round of IPTG induction and sorting was performed to gain another 64 single clones (Fig. 7A, Pool 2; Gate 1 and 2). To verify GFP signal, all single clones (Pool 1 and Pool 2) were inoculated into fresh medium, protein expression was induced, and GFP expression was analyzed using a BD LSRFortessa™ cell analyzer in the HTS mode (Laser 1: 488 nm: >50 mW, Filter: 505LP, 530/30-nm FITC). Finally, plasmids were isolated from single clone cultures, which showed no GFP signal after induction, and analyzed by Sanger sequencing.
FACS-based CI2 stability assay
To generate five CI2 mutants with varying stabilities, Site-Directed II Lightning mutagenesis kit (Agilent Technologies) was used with CI2 WT as template. Each mutant was amplified by PCR using primer pairs as indicated in Supplementary Table 1. PCR products were cloned into pET22b-mCherry vector using the NdeI and SpeI restriction sites and joined using Gibson Assembly® Cloning Kit (New England Biolabs) according to manufacturer’s instructions.
The CI2 variants were co-transformed with pSEVA631(Sp)-IbpAp-GFP-ASV into Rosetta2 (DE3) pLysS chemically competent cells and expressed in 50 ml LB media supplemented with 100 µg/µl ampicillin, 25 µg/ml chloramphenicol, and 50 µg/ml spectinomycin at 30 °C, 250 rpm to an OD600 of 0.8. Protein expression was induced with 0.5 mM IPTG. Cells were extracted before and 1 h after induction and kept on ice until FACS analysis. The mCherry and GFP fluorescence was analyzed on a BD FACS-ARIATMSORP cell sorter as mentioned above.
CI2 expression and purification for stability measurements
CI2 variants transformed into Rosetta2 (DE3) pLysS competent cells and expressed in 1 L LB in the presence of 100 µg/µL ampicillin and 25 µg/ml chloramphenicol at 37 °C. Protein expression was induced at OD600~0.5–0.7 with 0.5 mM IPTG and cells were further grown at 30 °C for 4–5 h. Cells were harvested by centrifugation at 5000 × g for 20 min. Cell pellets were resuspended in 20 mL buffer A (20 mM sodium acetate pH 5.3) and frozen at −20 °C. Cell lysis was performed by two rounds of sonication (1 min, 80% amplitude, 0.5 cycles, (Hielscher UP200S)) followed by 30 min incubation on ice in presence of 1 mg DNase. Cell debris and protein aggregates were removed by centrifugation at 20,000 × g, 4 °C for 30 min. The supernatants were loaded onto a 1 mL HisTrap HP column (GE Healthcare) equilibrated with buffer A, and eluted with a gradient of buffer B (20 mM sodium acetate pH 5.3, 1 M imidazole) from 0 to 100 %. Fractions containing CI2 determined from SDS-PAGE analysis were concentrated and loaded onto a superdex75 10/300 GL column (GE Healthcare) equilibrated with 20 mM sodium phosphate pH 7.4, 150 mM NaCl. For buffer exchange the samples were concentrated and loaded onto a superdex75 10/300 GL column equilibrated with 50 mM MES pH 6.25. The purity of the proteins was assessed by SDS-PAGE and the protein concentration was determined using a spectrometer (PerkinElmer lambda40) with an extinction coefficient of 6990 M−1 cm−1.
The CI2 variants were diluted to 10 µM in MES pH 6.25 with or without 6 M guanidium chloride. Using both solutions a dilution series of guanidium chloride ranging from 0 to 6 M guanidium chloride was prepared. Intrinsic tryptophan and tyrosine fluorescence of the CI2 variants was measured in triplicates using nanoDSF technologies on a Prometheus NT.48 instrument (nanoTemper technologies) with a temperature range from 15 to 95 °C with 1 °C/min increments. Global fitting of the temperature and denaturant unfolding was performed using the 330 and 360 nm fluorescence and ∆GU and was obtained as described before64.
Generating and scoring decoy structures
We generated 20,000 decoy structures using Rosetta’s threading protocol45 with PDBID: 2COK (Solution structure of BRCT domain of poly(ADP-ribose) polymerase-1) as a template. As a means to score a given decoy structure, we calculated the spearman’s correlation coefficient ρ between the residue depths of the decoy structure and the mean Enrich2 positional score for the corresponding positions in the sequence.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
