Large scale synthetic site saturation GPCR libraries reveal novel

Subscriber access provided by University of South Dakota

Technical Note

Large scale synthetic site saturation GPCR libraries reveal novel mutations that alter glucose signalling David Öling, Lina Lawenius, William Shaw, Sonya Clark, Ross Kettleborough, Tom Ellis, Niklas Larsson, and Mark Wigglesworth ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.8b00118 • Publication Date (Web): 16 Aug 2018 Downloaded from http://pubs.acs.org on August 18, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Synthetic Biology

Large scale synthetic site saturation GPCR libraries reveal novel mutations that alter glucose signalling

Öling, D, 1Lawenius, L, 1Shaw, W3,4, Clark, S5, Kettleborough, R5, Ellis, T,

3,4

Larsson, N1 and

Wigglesworth, M2 1. Discovery Biology, Discovery Sciences, Innovative Medicines and Early Development Biotech Unit, AstraZeneca R&D, Gothenburg, Sweden 2. Hit Identification, Discovery Sciences, Innovative Medicines and Early Development Biotech Unit, AstraZeneca R&D, Macclesfield, Cheshire, UK 3. Department of Bioengineering, Imperial College London, London SW7 2AZ, U.K 4. Centre for Synthetic Biology, Imperial College London, London SW7 2AZ, U.K. 5. Twist Bioscience, San Francisco, CA 94158

Abstract Site saturation mutagenesis (SSM) is a powerful mutagenesis strategy for protein engineering and directed evolution experiments. However, limiting factors using this method are either biased representation of variants, or limiting library size. To overcome these hurdles, we generated large scale targeted synthetic SSM libraries using massively parallel oligonucleotide synthesis and benchmarked this against an error-prone (epPCR) library. The yeast glucose activated GPCR – Gpr1 was chosen as a prototype to evolve novel glucose sensors. We demonstrate superior variant representation and several unique hits in the synthetic library compared to the PCR generated library. Application of this mutational approach further build the possibilities of synthetic biology in tuning of a response to known ligands and in generating biosensors for novel ligands.

ACS Paragon Plus Environment

ACS Synthetic Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 11

Directed evolution or protein engineering experiments rely heavily on random or targeted mutagenesis such as site saturation mutagenesis (SSM). The latter is particularly useful for exploring a small number of amino acid (a.a) positions or domains, but library size quickly becomes a limiting factor as well as requirement of previous structural or biochemical knowledge of the protein1, 2. Recently, technology developments in synthetic DNA manufacturing using low cost, microarraybased oligo synthesis have enabled the use of large scale SSM libraries in a way that has not been economically feasible previously3-5. Similarly, SSM by using degenerate oligos encoding a mutant codon (NNK/NNS) have been optimized to build libraries with high variant representation (N = A/C/G/T, K = G/T and S = C/G)6, 7. Pfunkel or nicking mutagenesis can produce ~92% amino acid variant representation whereas NNK/NNS and massively parallel oligo synthesis based protocols can reach up to 95% and 99% respectively5,

8-10

. However, current oligo based methods generate

unwanted stop codons and considerable error-, codon- or amplification –bias due to PCR amplification steps, annealing and oligo synthesis errors (1:200 nt:s or better)3,

11

. Now, these

drawbacks have been overcome by the development of a silicon nano-well architecture, solid-phase oligo synthesis platform which relies on synthesis miniaturization and robust DNA quality control by Next Generation Sequencing (NGS) to ensure bias and error- free synthesis. We sought to capitalize on this novel DNA synthesis platform to explore the sequence space of a Gprotein coupled receptor (GPCR). This family of receptors belong to a family of proteins that are widely exploited in medicine as they control many disease relevant pathways. Single mutations in transmembrane (TM) domains or extra cellular loops (ECL) of these receptors can dramatically influence GPCR signalling by altering agonist affinity, efficacy or receptor selectivity12-17. Using the yeast glucose receptor Gpr1 as a prototypical GPCR, we generated rationally designed, large scale, synthetic site saturation libraries to search for Gpr1 variants with altered activation and ligand binding. The synthetic library was complemented and benchmarked against a random mutagenesis library generated by error prone PCR (epPCR). Capitalizing on the endogenous glucose signalling cascade of yeast, we adopted a directed evolution platform to screen these libraries and isolated several Gpr1 variants with altered response to glucose. Using deep sequencing, we demonstrate superior variant representation in the synthetic library compared to the PCR based library. Importantly, we isolated several unique gain-of-function variants from the synthetic library that were not found in the epPCR library.

Results and Discussion We hypothesized that glucose is likely to bind to the exofacial and transmembrane domains of Gpr1. To this end, we designed three synthetic SSM libraries to include 19 a.a variants per position with the


Page 3 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


wild type variant excluded. The libraries spanned the three extracellular loops and residues partly down in the TM domains (Fig. 1A). Using a silicon based nano-well oligo synthesis platform, each position was synthesized individually and stored to facilitate downstream validation following a positive hit. Each variant was synthesized with flanking BsaI sites and pooled with respect to TM domains and ECLs where ECL1 contained all variants from 48 positions; ECL2 contained variants from 73 positions and ECL3 contained variants from 40 positions (in total: 3059 Gpr1 variants). In addition, we employed a complementary approach where a mutagenesis library was constructed by epPCR to introduce random mutations along the full length receptor sequence space (Fig. 1B). Deep sequencing (Illumina) confirmed that the majority of a.a variants (>96%) were present in the SSM library (Fig.1C, top left) in a highly homogenous manner (a.a. frequency average = 5,23%±2,2%, CV=41,4%) (Fig.1C, top right). In fact, deeper sequencing of the SSM library confirmed the presence of 3055 of 3059 designed variants (99,9% representation) demonstrating the effectiveness of the technique. In contrast, the epPCR library contained 35% of the theoretical maximum of a.a variants (Fig.1C, bottom left) in a heterogenous distribution (17,7%±27,9%, CV=157,6%) (Fig.1C, bottom right). The libraries were assembled (Golden Gate) and transformed into a S. cerevisiae strain combining two novel glucose activation assays. Briefly, RAS2 was deleted to abolish parallel activation of cAMP and subsequently SUC2 which has previously been linked to Gpr1 signalling (Fig. S1A-D)18, 19. In addition, the SUC2A site was altered to remove suppressor binding sites and boost reporter gene induction (Fig. S1E-I). Thus, the SUC2A* promoter is only regulated by factors binding to the SUC2B site resulting in a more linear response pathway downstream of Gpr1. Roughly, 600 colonies were isolated using the PSuc2LEU reporter followed by measuring PSuc2A*sfGFP / PTEF1mRuby2 ratios for each isolated mutant at basal (no glucose) and induced (0,075%/4.2mM) conditions (Fig. 1E). Mutants above the threshold were validated and the full-length gene was Sanger sequenced.




Page 4 of 11

Page 5 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Design and construction of mutagenesis libraries. A) Schematic illustration of Gpr1 mutagenesis library rational design and variant synthesis. Each position was synthesized as a pool of 19 variants (excluding wild type). Libraries where then created by, massively parallel gene synthesis and pooling of saturation variants covering 161 positions (in black) distributed over ECL1 (48), ECL2 (73) and ECL3 (40). B) Error prone PCR performed on the full-length gene fragment to generate a randomly mutated library. C) Frequency distribution heat map (right) and plot (left) showing codon variants detected at each position by deep sequencing by Illumina MiSeq (wild type a.a excluded by design) in the SSM (top, a.a. frequency mean per position=5,27%±2,2%, CV=41,4%), and epPCR bottom, (17,7%±27,9%, CV=157,6%) libraries. Data is based on read counts of each variant relative to the total read counts (mean ± S.D) for all variants at each site. D) Workflow for library assembly and screening. Successful Gpr1 variant assemblies (white colonies) were harvested and transformed into a gpr1∆ras2∆ Psuc2LEU strain harbouring the pPsuc2A*sfGFP PTEF1mRuby2 reporter plasmid. Roughly 600 colony forming mutants on low glucose (0,001%) concentrations were assayed for fluorescent reporter induction in basal and induced conditions. E) Plot of individual mutants isolated from the SSM and epPCR libraries presented as sfGFP/mRuby2 ratio (X=Basal, Y=Induced). The dotted lines indicate the cut-off which was set to 3 x Wt S.D (0,51/0,78 basal and induced respectively, S.D=+/- 8%).

Mutants were then validated by single copy GPR1 variants integrated genomically in the URA locus. In the ECL1 library, two mutants in TM2 and TM3, N116V and I113P were isolated that showed increased basal activity but similar maximum efficacy as Gpr1 wild type (Fig 2 A) (Gpr1 topology adopted from20). Similarly, the ECL2 library contained two mutants located in the second loop, W244L and K202G that conferred increased basal activity but similar maximum efficacy (Fig 2A). Strikingly, the ECL3 and the epPCR library both contained an A640V mutation (located in TM6) which conferred significantly enhanced response to glucose. We proceeded by revisiting the stored synthetic library containing all pooled position 640 mutants with the aim to find other residues at this position that alter the glucose response to a varying degree. Cloning and integration of position 640 mutants was greatly facilitated by the uniform variant pool distribution (Fig. 2B). The mutants were classified with respect to gain of function (GOF) or loss of function (LOF) under unregulated or derepressed conditions using both a plate based assay (Fig. 2C) and a fluorescent reporter assay (Fig. 2D). Among the LOF mutants were: A640I, A640M, A640F, A640Y, A640W, A640C, A640G, A640D and A640C. The latter, A640C has been identified previously as a loss of function mutation which was confirmed also in our assays21. Five mutations were classified as GOF: A640V, A640L, A640P and A640E, whereas the remaining mutations remained unaltered compared to wild type. Finally, we asked whether the GOF or LOF phenotypic effects could be explained by altered expression or localization of the receptor. This was not the case since a C-terminal sfGFP tag on A640V (GOF) and A640C (LOF) showed that localization to the membrane and expression levels were normal (Fig. S2A



and B). Taken together, these data indicate that position 640 is crucial for glucose-induced signalling since different mutations at this position have dramatically different outcomes on the response.

Figure 2. Amino acid substitutions at position 640 of Gpr1 alter the glucose response. A) Glucose response (Basal activity=red, induced (0,1%)=black) of integrated mutants from the various libraries generated with the PSuc2A*sfGFP PTEF1mRuby2 reporter (top). Basal activity shown in red and induced (0,1% glucose) shown in black. Data are presented as the mean +/- SEM of three individual experiments. P-values were obtained by a twosided t-test with 95% confidence intervals, *P ≤ 0.05; **P ≤ 0.005; ***P ≤ 0.0005; n.s = P > 0.05. Snake plot showing a topography of positions of isolated single site mutants (Bottom). B) Pooled position 640 mutants were subcloned and integrated individually into the dual glucose reporter strain. C) Spottest of GPR1640 site saturation mutants. Gpr1 mutants integrated at the URA locus in the double mutant gpr1∆ras2∆ Psuc2LEU were pre-cultured in YPD and then ten-fold serially diluted (10-1 to 10-5) on synthetic defined media without leucine with varying glucose concentrations. Plates were incubated for 4-5 days at 30°C. At least three spottests were repeated with similar results. D) Glucose response of Gpr1 mutants from C generated with the PSuc2A*sfGFP


Page 6 of 11

Page 7 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


PTEF1mRuby2 reporter. Basal activity shown in red and induced (0,1% glucose) shown in black. Data are presented as the mean +/- SEM of three individual experiments. P-values were obtained by a two-sided t-test with 95% confidence intervals, *P ≤ 0.05; **P ≤ 0.005; ***P ≤ 0.0005; n.s = P > 0.05.

In this study we generated a large scale, synthetic site saturation GPCR library using a nano-well, silicon based synthesis platform (WO 2015021080). We demonstrated that the synthetic SSM library outperformed the random mutagenesis library both on the DNA and protein level as seen by: i) variant representation and distribution ii) number of unique hits, and iii) timesaving downstream validation and cloning processes. First, the variant representation was high (99,9%) and homogenous in the SSM library compared to the random library (mean frequency/position = 5,27%±2,2%, CV=41,4% vs. 17,7%±27,9%, CV=157,6%). Second, the number of unique validated hits was higher in the SSM library compared to the epPCR library. In fact, of the six isolated mutants, only A640V was also present in the epPCR library (Fig.1C and Fig.1E). Third, once key positions were identified, we could easily access, clone and validate all other variants via the stored synthetic SSM library plates. In summary, the library described herein relies on high quality oligo synthesis (error rate of 1:1000 nt:s; indels 1:1400), contains no unwanted stop codons and lacks exponential PCR amplification steps which ensures consistent, high (>99%) and homogenous variant representation. In addition, while providing novel insight into Gpr1 activation, we have exemplified that the library design used herein can be used as a general approach for elucidating GPCR activation and ligand binding. It may also be applied more widely to any protein evolution project to efficiently select and test desired protein variants increasing the chance of success.

Materials and Methods Yeast media, plasmids and strains All Saccharomyces cerevisiae strains generated in this study are derivatives of BY4741 (MATa his3∆1 leu2∆0 met15∆0 ura3∆0). Strains, primers and plasmids used in this study are listed in Supplementary Table S1, S2 and S3 respectively. Standard culturing conditions were applied. All plasmids were assembled by Golden Gate 22 Assembly and parts were either synthesized or assembled plasmids included in a toolkit (YTK) . Plasmids used in this study can be found in Supplementary Table S3.

Gpr1 mutagenesis library and screening The random mutagenesis library was created by epPCR methodology using Jena Bioscience (JBS) error prone kit #PP-102 according to manufacturer’s instructions (~0,1% error frequency). The synthetic SSM library was created by mutagenic oligos containing the variation for one position (19 codons) in one nano-well of the silicon based platform using a combination of gene synthesis and overlap extension PCR. The library for each position was deposited into one well of a 96 well plate and the % of each codon variant is confirmed by NGS. Cloning of libraries were done using a standard Golden Gate Assembly protocol 22. At least 106 individual E.coli transformants were pooled prior to plasmid extraction whereas the synthetic SSM libraries contained library plasmids from at least 3000 E.coli transformants per library loop/pool. 15 μg DNA was transformed into competent cells (100 ml culture of ODA600nm=1.5 cells). Transformation mixture was plated on selective media (-



Page 8 of 11

Leu, -His, -Ura) containing 0.001% glucose and 600 colonies were picked after 4 days, re-streaked and analysed in downstream assays.

SUC2 Leucine reporter assay -1

-5

Logarithmically growing cells were serially diluted (tenfold) to OD A600=10 to 10 and spotted with a multipipette on plates containing 5% Glycerol and 2% Ethanol with varying concentrations of glucose. Plates were incubated four to five days at 30 ̊C prior to imaging.

Glucose activation assay Logarithmically growing cells were centrifuged, washed and re-suspended in fresh media lacking glucose supplemented with varying glucose concentrations. Cells (90 µl) of OD A600= 0,5 were transferred to black 96 well plates (NUNC) and incubated at 30 ̊ C for four hours on an Eppendorf Thermomixer R with continuous shaking at 850 rpm. Super folder green fluorescent protein and mRuby2 ratios were assayed with a Tecan Sapphire 2 using appropriate wavelengths for excitation/emission.

Deep sequencing For NGS analysis of the libraries, PCR products were fragmented, size selected, end repaired and dA tailed. Illumina adapters were then ligated and the library was amplified with Kapa HiFi polymerase and purified using Ampure XP beads. The resulting Illumina library was run on a 150bp PE run on the MiSeq platform. Raw FASTQ files were trimmed for adaptors and quality using Trim Galore. For each variant position, custom python code was used to search trimmed reads using constant flanking sequence. All observed codons per variant position were summarized as frequency distribution plots. Sequencing depth was calculated based on the total number of variant positions and number of variants at those positions (target average coverage of > 50x).

Associated content Construct design and assay development to gauge Gpr1 activity (Fig. S1). Expression and localization data of Gpr1 mutants (Fig. S2) Oligos, plasmids and strains generated (Table S1, S2 and S3)

Author contributions D.Ö, N.L, T.E and M.W conceived and planned the experiments. D.Ö and L.L carried out experimental work (library generation, assembly, screening and validation). W.S cloned GPR1 and helped with construct design. Library quality control and analysis by NGS was carried out by S.C, D.Ö and R.K. Discussion and interpretation of the results was done by D.Ö, W.S, N.L, T.E and M.W. D.Ö wrote the manuscript. D.Ö, N.L, T.E and M.W edited the manuscript. All authors provided feedback and analysis of the manuscript.

Acknowledgments The authors declare no conflicting interests. David Öling, Niklas Larsson and Mark Wigglesworth are employees of AstraZeneca which also sponsored this work. We are grateful to TWIST Bioscience for providing guidance and Gpr1 synthetic variant libraries.


Page 9 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15.

16.

17.

18. 19.

20.

21.

Smith, M., Nobel lecture. Synthetic DNA and biology. Biosci Rep, 1994. 14(2): p. 51-66. Reidhaar-Olson, J.F. and R.T. Sauer, Combinatorial cassette mutagenesis as a probe of the informational content of protein sequences. Science, 1988. 241(4861): p. 53-7. Kitzman, J.O., et al., Massively parallel single-amino-acid mutagenesis. Nat Methods, 2015. 12(3): p. 203-6, 4 p following 206. LeProust, E.M., et al., Synthesis of high-quality libraries of long (150mer) oligonucleotides by a novel depurination controlled process. Nucleic Acids Res, 2010. 38(8): p. 2522-40. Starita, L.M., et al., Massively Parallel Functional Analysis of BRCA1 RING Domain Variants. Genetics, 2015. 200(2): p. 413-22. Jain, P.C. and R. Varadarajan, A rapid, efficient, and economical inverse polymerase chain reaction-based method for generating a site saturation mutant library. Anal Biochem, 2014. 449: p. 90-8. Acevedo-Rocha, C.G., M.T. Reetz, and Y. Nov, Economical analysis of saturation mutagenesis experiments. Sci Rep, 2015. 5: p. 10654. Firnberg, E. and M. Ostermeier, PFunkel: efficient, expansive, user-defined mutagenesis. PLoS One, 2012. 7(12): p. e52031. Wrenbeck, E.E., et al., Plasmid-based one-pot saturation mutagenesis. Nat Methods, 2016. 13(11): p. 928-930. Majithia, A.R., et al., Prospective functional classification of all possible missense variants in PPARG. Nat Genet, 2016. 48(12): p. 1570-1575. Fowler, D.M., J.J. Stephany, and S. Fields, Measuring the activity of protein variants on a large scale using deep mutational scanning. Nat Protoc, 2014. 9(9): p. 2267-84. Smits, G., et al., Glycoprotein hormone receptors: determinants in leucine-rich repeats responsible for ligand specificity. EMBO J, 2003. 22(11): p. 2692-703. Vassart, G., L. Pardo, and S. Costagliola, A molecular dissection of the glycoprotein hormone receptors. Trends Biochem Sci, 2004. 29(3): p. 119-26. Venkatakrishnan, A.J., et al., Molecular signatures of G-protein-coupled receptors. Nature, 2013. 494(7436): p. 185-94. Weaver, R.E., et al., High affinity binding of the peptide agonist TIP-39 to the parathyroid hormone 2 (PTH2) receptor requires the hydroxyl group of Tyr-318 on transmembrane helix 5. Biochem Pharmacol, 2017. 127: p. 71-81. Weaver, R.E., M.J. Wigglesworth, and D. Donnelly, A salt bridge between Arg-20 on parathyroid hormone (PTH) and Asp-137 on the PTH1 receptor is essential for full affinity. Peptides, 2014. 61: p. 83-7. Miller, P.S., et al., Non-peptidic antagonists of the CGRP receptor, BIBN4096BS and MK-0974, interact with the calcitonin receptor-like receptor via methionine-42 and RAMP1 via tryptophan-74. Biochem Biophys Res Commun, 2010. 391(1): p. 437-42. Belinchon, M.M. and J.M. Gancedo, Different signalling pathways mediate glucose induction of SUC2, HXT1 and pyruvate decarboxylase in yeast. FEMS Yeast Res, 2007. 7(1): p. 40-7. Gancedo, J.M., C.L. Flores, and C. Gancedo, The repressor Rgt1 and the cAMP-dependent protein kinases control the expression of the SUC2 gene in Saccharomyces cerevisiae. Biochim Biophys Acta, 2015. 1850(7): p. 1362-7. Kraakman, L., et al., A Saccharomyces cerevisiae G-protein coupled receptor, Gpr1, is specifically required for glucose activation of the cAMP pathway during the transition to growth on glucose. Mol Microbiol, 1999. 32(5): p. 1002-12. Lemaire, K., et al., Glucose and sucrose act as agonist and mannose as antagonist ligands of the G protein-coupled receptor Gpr1 in the yeast Saccharomyces cerevisiae. Mol Cell, 2004. 16(2): p. 293-9.



22.

Lee, M.E., et al., A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly. ACS Synth Biol, 2015. 4(9): p. 975-86.


Page 10 of 11

Page 11 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


83x36mm (300 x 300 DPI)


Large scale synthetic site saturation GPCR libraries reveal novel

Recommend Documents