Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

High Throughput Mutagenesis for Identification of Residues Regulating Human Prostacyclin (hIP) Receptor Expression and Function

  • Anke Bill ,

    Contributed equally to this work with: Anke Bill, Elizabeth M. Rosethorne

    Affiliation Developmental and Molecular Pathways, Novartis Institutes for Biomedical Research, Cambridge, Massachusetts, United States of America

  • Elizabeth M. Rosethorne ,

    Contributed equally to this work with: Anke Bill, Elizabeth M. Rosethorne

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Toby C. Kent,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Lindsay Fawcett,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Lynn Burchell,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Michiel T. van Diepen,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Anthony Marelli,

    Affiliation Infectious Diseases, Genomics Institute of the Novartis Research Foundation, San Diego, California, United States of America

  • Sergey Batalov,

    Affiliation Infectious Diseases, Genomics Institute of the Novartis Research Foundation, San Diego, California, United States of America

  • Loren Miraglia,

    Affiliation Infectious Diseases, Genomics Institute of the Novartis Research Foundation, San Diego, California, United States of America

  • Anthony P. Orth,

    Affiliation Infectious Diseases, Genomics Institute of the Novartis Research Foundation, San Diego, California, United States of America

  • Nicole A. Renaud,

    Affiliation Developmental and Molecular Pathways, Novartis Institutes for Biomedical Research, Cambridge, Massachusetts, United States of America

  • Steven J. Charlton,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • Martin Gosling,

    Affiliation Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom

  • L. Alex Gaither,

    Affiliation Developmental and Molecular Pathways, Novartis Institutes for Biomedical Research, Cambridge, Massachusetts, United States of America

  • Paul J. Groot-Kormelink

    paul.groot-kormelink@novartis.com

    Affiliations Respiratory Disease Area, Novartis Institutes for Biomedical Research, Horsham, United Kingdom, Musculoskeletal Disease Area, Novartis Institutes for Biomedical Research, Basel, Switzerland

High Throughput Mutagenesis for Identification of Residues Regulating Human Prostacyclin (hIP) Receptor Expression and Function

  • Anke Bill, 
  • Elizabeth M. Rosethorne, 
  • Toby C. Kent, 
  • Lindsay Fawcett, 
  • Lynn Burchell, 
  • Michiel T. van Diepen, 
  • Anthony Marelli, 
  • Sergey Batalov, 
  • Loren Miraglia, 
  • Anthony P. Orth
PLOS
x

Abstract

The human prostacyclin receptor (hIP receptor) is a seven-transmembrane G protein-coupled receptor (GPCR) that plays a critical role in vascular smooth muscle relaxation and platelet aggregation. hIP receptor dysfunction has been implicated in numerous cardiovascular abnormalities, including myocardial infarction, hypertension, thrombosis and atherosclerosis. Genomic sequencing has discovered several genetic variations in the PTGIR gene coding for hIP receptor, however, its structure-function relationship has not been sufficiently explored. Here we set out to investigate the applicability of high throughput random mutagenesis to study the structure-function relationship of hIP receptor. While chemical mutagenesis was not suitable to generate a mutagenesis library with sufficient coverage, our data demonstrate error-prone PCR (epPCR) mediated mutagenesis as a valuable method for the unbiased screening of residues regulating hIP receptor function and expression. Here we describe the generation and functional characterization of an epPCR derived mutagenesis library compromising >4000 mutants of the hIP receptor. We introduce next generation sequencing as a useful tool to validate the quality of mutagenesis libraries by providing information about the coverage, mutation rate and mutational bias. We identified 18 mutants of the hIP receptor that were expressed at the cell surface, but demonstrated impaired receptor function. A total of 38 non-synonymous mutations were identified within the coding region of the hIP receptor, mapping to 36 distinct residues, including several mutations previously reported to affect the signaling of the hIP receptor. Thus, our data demonstrates epPCR mediated random mutagenesis as a valuable and practical method to study the structure-function relationship of GPCRs.

Introduction

The human prostacyclin (PGI2) receptor (hIP receptor, International Union Pharmacology nomenclature) is a seven transmembrane (TM) G protein-coupled receptor (GPCR) [1]. The 386 amino acid protein is encoded by the human prostacyclin receptor gene (PTGIR) and is composed of a short N-terminal tail, seven TM-spanning alpha-helical domains, three intra- and three extracellular loops and a long c-terminal tail including two palmitoylation sites. The receptor is further stabilized by two disulfide bonds [2], [3]. The hIP receptor is activated by binding of prostacyclin which leads to the activation of membrane-bound adenylyl cyclase, subsequent formation of the second messenger cyclic adenosine monophosphate (cAMP) and activation of various cellular responses [1]. The hIP receptor is ubiquitously expressed in the body, but shows predominant expression in the cardiovascular system, especially on platelets and vascular smooth muscle cells where it plays a key role in vascular smooth muscle relaxation and inhibition of platelet aggregation. Recent studies have revealed a cardioprotective role of the IP receptor and hIP receptor dysfunction has been implicated in numerous cardiovascular abnormalities, including myocardial infarction, hypertension, thrombosis and atherosclerosis [4][9]. Furthermore, genetic variants in the hIP receptor resulting in deficits of hIP receptor function have been correlated with increased disease severity in patients with coronary artery disease [10], [11]. Thus, it is of paramount significance to understand the mechanisms and structural requirements underlying hIP receptor function. Despite significant progress in our understanding of GPCRs in general and multiple studies using site-directed mutagenesis to identify residues critical for agonist binding and hIP receptor activity [9], the details of hIP receptor structure-function remain largely unknown.

Here we set out to explore the relationship between hIP receptor structure and function using an unbiased, high-throughput, random mutagenesis approach. We compared the applicability of chemical mutagenesis and error-prone PCR mediated mutagenesis to create a library covering the complete coding region of the hIP receptor and analyzed the quality of our library by next generation sequencing (NGS). Furthermore, we report the functional characterization of >4000 hIP receptor mutants and the identification of 18 mutants of the hIP receptor that maintained expression but demonstrated a full or partial reduction in receptor activity. Our results highlight the advantages of random high throughput mutagenesis to gain insights in the structure-function requirements of hIP receptor, as compared to biased approaches using site-directed mutagenesis.

Experimental Procedures

Hydroxyl amine mutagenesis

Hydroxylamine mutagenesis, which reacts with pyrimidine nucleotides to generate cytosine to thymidine transitions, was adapted from a previous published protocol [12]. 5 ug HA-tagged hPTGIR in pcDNA5 was incubated with 250 µl 1 M hydroxylamine hydrochloride (Sigma Aldrich), pH 7 at 70°C for 32 to 128 min. Mutagenised plasmids were then isolated and washed at least 3 times using the QiaQuick PCR purification kit (Qiagen), and finally eluted with 50 µL EB buffer from the kit and diluted with 50 µL sterile water. 1 µL of purified mutated plasmid was transformed into 100 µL Invitrogen Library Efficiency DH5α competent E. coli (Invitrogen). Transformed cells were then plated onto Luria Broth (LB)-agar media supplemented with ampicillin and grown O/N at 37°C. Individual bacterial colonies were picked and grown O/N at 37°C, DNA was isolated, the hPTGIR insert sequenced and the mutation rate per mutant determined (Table 1).

thumbnail
Table 1. Colony number and mutation rate obtained by hydroxyl amine mediated mutagenesis.

https://doi.org/10.1371/journal.pone.0097973.t001

Error-prone PCR mediated mutagenesis

Error-prone PCR (epPCR) mediated mutagenesis was performed using the GeneMorph II Random Mutagenesis kit (Stratagene). Briefly, 500–1000 ng of plasmid DNA (pcDNA5-HA-hPTGIR) was mutagenized by PCR according to the manufacturer's protocol using the following primers: IPR-for: 5′- CGACTCACTATAGGGAGACCCAAGCTGG -3′, IPR-rev: 5′- CAACTAGAAGGCACAGTCGAGGCTGATC -3′. PCR products were separated by gel electrophoresis and purified using a gel-extraction kit (Qiagen). After restriction enzyme digest (HindIII and XhoI), the DNA was purified with a PCR purification kit (Qiagen) and cloned into the parental vector (pcDNA5-HA). Mutagenized DNA was transformed into E.coli DH5α Max Efficiency (Invitrogen). For initial studies to determine the mutation rate, between 100–200 single colonies were handpicked. For the large mutant library prep around 4000 individual colonies were picked with a Mega-Pix robot (Genetix). In both cases, colonies were transferred to 96-deep well blocks containing 1.2 mL of terrific broth (TB)/Carbenicillin/Tylosin, and grown at 37°C O/N in the HiGro incubator (Gene Machines, Inc.). The next day glycerol stocks were made of the whole library by transferring 100 µl of the bacterial culture into 70 µL of 50% glycerol/50% TB solution. The glycerol stocks were stored at −80°C for future amplification of the library. DNA purification of mutant clones from the remaining 96-deep well bacterial cultures was performed by a RoboSeq 4204 prep robot (MWG Biotech) with the NucleoSpin Robot-96 Plasmid Miniprep kit (Macherey-Nagel). Purified plasmids in 96-well microtiter plates were quantified with a PowerWave XS spectrometer (Bio-Tek) and normalized to a final concentration of 40 ng/µL with a TheOnyx (MWG Biotech) liquid-handling robot. The 96-well plates containing normalized mutant DNA were compressed into 384-deep well barcoded bioassay plates (Greiner Bio-One) by transferring 62.5 µL of mutant plasmid with a MiniTrak robot (Perkin Elmer) into assigned 384-well plate coordinates. In total, twelve unique 384-well plates (approx. 4,000 mutants) containing an individual clone within each well were generated. All plates were stored at −20°C.

Next generation sequencing

All library constructs were pooled and the genomic library was prepared and sequenced by Axeq Technologies on an Illumina HiSeq 2000. The 101 bp paired-end reads were aligned to the wild-type-sequence of human PTGIR using BWA (v0.7.4) with the parameters -n 2 -t 12 (pmid 19451168). Alignments were converted to bam (pmid: 19505943) and only high quality nucleotides were counted at each position using the bamutils basecalls function (pmid 23314324; bamutils basecall -ref wildtype.fa -qual 40 epPCR.bam). Mutations were called against the reference sequence and mutation rates calculated with R and the plyr package (R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org; [13]). Non-biased assessment of codon usage was calculated by dividing the total mutations across the coding sequence by the potential mutations. All potential codons were then determined based on the three alterative bases at each location, assuming one mutation per codon. The data has been deposited in the NIH SRA database (BioProject: PRJNA242280).

Cell culture and transfection

HEK293 GloSensor-20F cAMP cells (Promega) were maintained in DMEM supplemented with foetal bovine serum (10% v/v) and hygromycin (200 µg/mL) at 37°C, 5% CO2. For experiments, cells were harvested using trypsin/EDTA buffer and seeded on to multiwell plates in CO2-independent medium containing foetal bovine serum (10% v/v). Reverse-transfection of cells with the mutant library was performed in 384-well black, clear bottom plates using FuGENE6 (Promega) according to manufacturer's instructions. Briefly, 1 µL of cDNA library or wild-type hIP receptor (40–100 ng) was seeded into the 384-well assay plate using the Mosquito nanoliter dispenser. FuGENE6 was diluted in OptiMEM to give a 1∶3 DNA:FuGENE ratio (for 40 ng cDNA, 0.12 µl FuGENE6 per well in 4 µl OptiMEM/384 well). 4 µl OptiMEM/FuGENE mixture was then added to cDNA in 384 well plate, and incubated at room temperature for 30 minutes, with gentle shaking. 15 µl of cells (0.67×106 cells/mL; 10,000 cells/well) was then added to the DNA/OptiMEM/FuGENE mixture. Cells were then incubated overnight at 37°C, 5% CO2.

GloSensor cAMP assay

Kinetic analysis of cAMP accumulation in cells expressing hIP receptor was conducted using transiently transfected cells. 10 µL of 12% GloSensor reagent (4% [final]) in CO2-independent medium containing FBS (10% v/v) was added to the cells and the plate was incubated for 2 hour in the dark at room temperature. Luminescence (analogue mode, sensitivity 2) was then read on the FDSS7000 every 1.5 sec (integration time of 1.2 sec), for a total of 1000 samples. 5 µL of 7x MRE-269 [4-[(5,6-Diphenylpyrazin-2-yl)(isopropyl)amino]butoxy] acetic acid (synthesized in house) was transferred to the 384-well cell plate using the FDSS7000, after the first 5 samples. Hits were characterized as mutants that resulted in <25% of the response to MRE-269 in cells transfected with the wild-type hIP receptor.

Receptor expression

The above assay was then repeated with expression analysis. After transfection of selected mutant hIP and wild type receptors, cells were incubated in CO2-independent medium containing FBS (10% v/v); GloSensor reagent (4% v/v); Hoechst (1 µM); Alexa Flour® -647 conjugated anti-HA antibody (1∶500). Receptor expression was analyzed after 30 min on the InCell2000. Nuclei were visualized using DAPI settings (0.05 s exp), HA-tagged receptor using Cy5 settings (1.5 sec exp) and cells were visualized using bright field settings (0.05 s exp). The expression of mutant receptor was analyzed using InCell Developer Toolbox 1.9.1 software. Cells were detected using Object Segmentation (kernel size  = 57, Sensitivity  = 55), followed by Fill Holes and Sieve (Binary) based on size (>10 µm2). Mean and Sum of the fluorescence detected was taken for each well as an estimate of the relative expression levels for each mutant receptor. Cells were then incubated further at room temperature (2 hour in total) and activity assessed as above.

Results and Discussion

Chemical mutagenesis

Mutations in the hIP receptor and impaired receptor function have been correlated with cardiovascular disease [10], [11], however little is known about the structure-function relationship of the receptor. The hIP receptor has been intensively studied by site-directed mutagenesis, leading to the identification of critical residues for receptor activation and ligand binding [9], [11], [14][17]. However, because of the labor-intensive nature of this approach, these studies were focused on a limited number of residues that were selected based on homology with other receptors or molecular modeling approaches and are therefore highly biased. In order to identify residues important for hIP receptor function we set out to generate an unbiased library of hIP receptor mutants, covering the complete coding region of the receptor, for subsequent functional characterization.

In contrast to site-directed mutagenesis, random mutagenesis does not require the a priori selection of particular residues but rather allows for the generation and testing of thousands of amino acid substitutions in an unbiased way. However, depending on the mutagenesis method used, different degrees of bias regarding the type of substitution made can be achieved.

Chemical mutagenesis is the most convenient method to mutagenize a plasmid, since it does not involve any recombinant manipulation or sub-cloning of the parental DNA, i.e. generally the parental plasmid is incubated directly with one of several chemicals and can be used directly for transformation after a simple purification step [18][21]. Chemical mutagenesis can be especially useful if low mutation rates are required or sub-cloning of the targeted region is not possible. An obvious disadvantage is the lack of specificity for a specific region and the limitation of possible base pair changes, resulting in a highly biased library.

We set out to test the applicability of chemical mutagenesis to generate a mutant library for the hIP receptor to gain additional, unbiased insights into the relationship between the structure and function of this receptor. For this, we incubated the parental plasmid coding for the hIP receptor with 1 M hydroxyl amine for 32 min–128 min. Treatment of double stranded DNA with Hydroxyl amine is known to result in C->T and G->A transitions thus allowing for only a limited fraction of possible substitutions [21]. After a purification step, the DNA was transformed in E.coli. As shown in Table 1 the number of resulting colonies was inversely correlated to the time used for incubation with hydroxyl amine. Longer incubation times are expected to result in higher mutation rates, increasing the probability of mutations in regions of the vector coding for the ampicillin resistance used for selection, resulting in a loss of resistance. To determine the mutation rate in the coding region of the hIP receptor, a representative number of colonies were subsequently analyzed by Sanger sequencing. The observed rate of mutations in the coding region of the hIP receptor was very low, with an observed maximum of approximately one mutation every 2.6 kb (Table 1). Notably, the number of colonies obtained under these conditions would not suffice to generate a mutagenesis library of sufficient size to cover the coding region of hIP receptor. These results demonstrate the suitability of hydroxyl amine mediated mutagenesis for the generation of libraries with low mutations rates. For the purpose of our study, however, the observed mutation rate was not sufficient.

Error-prone PCR mediated mutagenesis

Next we explored the feasibility of epPCR mediated mutagenesis to generate a mutant library for the hIP receptor. epPCR is fast and facile, and the most commonly used method to create random mutagenized libraries [22]. It is used to introduce random mutations during PCR by compromising the fidelity of DNA polymerase, resulting in the incorporation of incorrect nucleotides during the PCR reaction and yielding randomly mutated products [23]. The fidelity of the DNA polymerase and therefore the error rate can be reduced by altering the reaction conditions for the PCR, i.e. ion concentrations, amount of starting material, number of cycles, or by using special variants of DNA polymerases with known deficits in 3′-5′ exonuclease activity [22], [24][26]. After the epPCR, the library of mutated DNA sequences is cloned in a suitable plasmid, making an additional purification and ligation step necessary, but on the other hand ensuring that no mutations are present outside the region of interest in the expression plasmid used. A disadvantage of epPCR mutagenesis lies in the mutational preference of the low-fidelity DNA polymerases used to introduce the mutations, favoring substitution of certain nucleotides over others resulting in a biased library [27], [28]. One approach to reduce the bias of the library and to increase diversity is to combine two DNA polymerases with opposite mutational spectra [29]. This approach has been proven to significantly reduce the bias in the library and is used by multiple commercially available mutagenesis kits. We used the commercially available Genemorph II kit which uses a mixture of two polymerases, Taq and Mutazymell, with opposite mutational spectra to reduce mutation bias.

The estimated mutation frequency can be described as a product of the DNA polymerase error rate and the number of amplifications [30]. The more often a target is replicated, the more errors are accumulated. Therefore the initial DNA amount can be used to adjust the mutation frequency. We performed three independent epPCR reactions with varying template DNA concentrations, sub-cloned the PCR fragments in a plasmid including an HA-tag and analyzed the mutation rates by sequencing a small subset of the resulting colonies after transformation in E.coli. As expected, the highest input of starting material (1 ug) resulted in a low mutation rate of approximately 0.5 mutations per kb, the lower DNA input of 750 ng and 500 ng yielded significantly higher mutation rates of approximately 0.8 and 1.5 mutations/kb, respectively (Table 2, Figure 1 A, B).

thumbnail
Figure 1. Error-prone PCR mediated mutagenesis of hIP receptor.

(A, B) Number and distribution of mutations per clone after error-prone PCR mediated mutagenesis of hIP receptor using 750 ng (A) or 500 ng (B) parental PTGIR plasmid as input. Grey bars represent the expected mutational spectrum at a mutation rate of 0.8 mut/kb (750 ng) or 1.5 mut/kb (500 ng), respectively. Black bars represent the experimentally observed number of mutations after sequencing a small subset of clones. (C, D) Expected number and distribution of amino acid changes per clone at a mutation rate of 0.8 mut/kb (750 ng, C) or 1.5 mut/kb (500 ng, D) (E) Expected and observed number and distribution of mutations after pooling both libraries (500+750 ng) depicted as described in (A,B). (F) Expected number and distribution of amino acid changes per clone in our library with an estimated mutation rate of 1.3 mut/kb. (G) Scatter blot showing the mutation rate for each nucleotide position of the coding region of PTGIR as determined by next generation sequencing. The overall mutation rate was 1.48 mutations/kb. (H) Comparison of the observed codon changes in our library (bars) with the expected codon changes (considering an unbiased mutation rate and one nucleotide substitution per codon, red symbol).

https://doi.org/10.1371/journal.pone.0097973.g001

thumbnail
Table 2. Colony number and mutation rate obtained by error-prone PCR mediated mutagenesis.

https://doi.org/10.1371/journal.pone.0097973.t002

Simulation of optimal mutation rate

Low mutation rates lead to libraries with many functional sequences (i.e. wild-type sequences and mutants with only a single amino acid change), but small numbers of total mutations and therefore overall low coverage (i.e. frequency with which a particular amino acid has been mutated in the library). High mutation rates enrich libraries for sequences with high numbers of mutations therefore leading to less functional sequences and a good coverage. However, the accumulation of mutations leads to a decrease in the unique mutation rate (i.e. how often a particular amino acid has been changed without any other mutations occurring in the same plasmid) [31]. Thus, one has to find the optimal mutation rate that balances overall mutation rate (coverage), functional retention and uniqueness. Low mutation rates with maximal unique mutation rate are optimal, but require a larger number of mutants (library size) to guarantee sufficient coverage. The distribution of the number of mutations per construct is expected to be a binomial distribution, and for the ease of downstream analysis of the ensemble of mutants we aimed to maximize the number of single amino acid mutations and minimize multiple and null mutations (i.e. the mode of the distribution would be at 1). The penetrance of DNA mutations into amino acid mutations is incomplete, and by simple in silico simulation for all positions in a representative set of genes, we determined the penetrance correction coefficient to be 0.75 (for the target genes in the study, the coefficient was between 0.72 and 0.77). We have prepared a simple worksheet that allowed us to simulate the expected number and distribution of amino acids in our library as a function of the mutation rate (see Table S1). Using this algorithm, we simulated the expected number and distribution of amino acids in our library as a function of the mutation rate and compared it to the distribution obtained by sequencing a small number of mutated colonies (Figure 1C, D). As can be seen in Table 3, a mutation rate of 1.1 mutations/kb is expected to produce the most single amino acid changes, but at the same time retaining 35% wild-type (non-mutated or silence nucleotide mutations) sequences. The lowest fraction of wild-type sequences in combination with high fractions of single and dual amino acid changes can be obtained with a mutation rate of 1.5 mutations/kb. Therefore, we pooled our two epPCR reactions, resulting in a library of >4000 hIP receptor constructs with a mutation rate of approximately 1.3 mutations/kb (Figure 1E, F).

thumbnail
Table 3. Estimated distribution of amino acid (AA) changes relative to the mutation rate (library size: 4000 plasmids).

https://doi.org/10.1371/journal.pone.0097973.t003

Analysis of mutagenesis libraries by next generation sequencing provides valuable information about coverage, mutation rate and mutational bias

Sanger sequencing is a rapid and cost-effective method to determine the mutation rate of a subset of clones which can be used to predict the expected mutation rate of the library. Due to limitations in throughput, scalability and coverage range, sequencing efforts are typically focused on a subset of clones only. However, sequencing of only a subset of clones does not provide reliable information about the mutational coverage of the targeted region or the expected mutational bias. NGS has made it possible to process multiple sequencing reactions in parallel, enhancing throughput and allowing for better coverage and resolution. To analyze the quality of our mutagenesis library and to obtain more reliable information about coverage and mutation rate, we sequenced our complete mutagenesis library for the hIP receptor using NGS. The pooled library was sequenced on an Illumina HiSeq 2000 (Axeq Technologies); the 101 bp paired-end reads were aligned to the wildtype-sequence of hIP receptor and the mutations were assessed (Table 4). The average coverage for the coding region of hIP receptor was ∼21000 reads. The data revealed a homogenous coverage of the complete coding region of hIP receptor with only a few outliers (Figure 1G). The overall mutation rate was 1.48 mutations/kb, very similar to our predicted mutation rate of 1.3 mutations/kb. We used the NGS data to assess the mutational bias in our library of >4000 hIP receptor clones (Table 5). We found no bias for the mutation of purines and pyrimidines, i.e. A's and T's were mutated at a similar rate as G's and C's (ratio 0.76). Transition mutations were more abundant than transversions (ratio 1.36). Mutations resulting in C's or G's were ∼4 times less likely than mutations resulting in T's and A's, favoring substitution in certain amino acids over others. These results are in accordance with previous reports and highlight the known mutational bias resulting from polymerase mediated mutagenesis [29], [32]. Comparison of the expected codon changes (considering an unbiased mutation rate and one nucleotide substitution per codon) with the observed codon changes in our NGS data showed a good overlap, indicating that the bias in nucleotide substitution only partially translates into a bias in amino acid substitutions (Figure 1H) [33]. Taken together, these data demonstrate the utility of NGS for validating the quality of mutagenesis libraries and for providing valuable information about coverage, mutation rate and potential mutational bias. Further optimization and the development of novel mutagenesis techniques are needed to generate mutagenesis libraries with less bias and increasing biological diversity.

thumbnail
Table 4. NGS results: Number of reads for each base are shown relative to reference base.

https://doi.org/10.1371/journal.pone.0097973.t004

thumbnail
Table 5. Characteristics of the hIP receptor mutant library (4000 plasmids).

https://doi.org/10.1371/journal.pone.0097973.t005

Functional characterization of hIP receptor library identified 32 mutants with reduced activity

In order to identify functional mutations of hIP receptor we tested the >4000 mutants in our library for activity using hIP receptor-mediated production of cAMP as a readout. The GloSensor cAMP assay utilizes a genetically encoded biosensor variant with cAMP binding domains fused to luciferase. Upon binding to cAMP, conformational changes occur that promote large increases in light output upon addition of a chemiluminescent substrate [34], [35]. We transiently transfected HEK293T cells stably expressing the GloSensor cAMP construct with our hIP receptor constructs and stimulated cAMP production with the selective hIP receptor agonist (4-[(5,6-Diphenylpyrazin-2-yl)(isopropyl)amino]butoxy)acetic acid (MRE-269) [36]. Primary hits were defined as mutants that resulted in <25% of the response to MRE-269 in cells transfected with the wild-type hIP receptor (Figure 2A). To eliminate false-negative hits due to insufficient transfection or deficiencies in trafficking of the receptor, we validated cell surface expression of ∼800 hIP receptor mutants in a subsequent step, utilizing the extracellular, N-terminal HA-tag of the protein. High content imaging of non-permeabilized cells stained with a fluorescently-labeled anti HA-antibody identified 34 mutants which demonstrated a loss of hIP receptor function but showed some cell surface expression of the receptor (for representative images see Figure 2B). These mutants were selected for analysis by sequencing, and cell surface expression and functional characterization were confirmed in 3 independent repeats of the assay (Figure 2C). Of these 34 mutants, 18 were expressed at the cell surface and demonstrated full or partial loss of receptor function.

thumbnail
Figure 2. Error-prone PCR mediated mutagenesis of hIP receptor identified 32 mutants with reduced activity.

(A) Scatter blot showing activity data of the hIP receptor library in the primary screen. Cells stably expressing the GloSensor cAMP plasmid were transfected with our hIP receptor constructs and stimulated with 1 µM MRE-269. The luminescent signal was normalized to the signal obtained in hIP receptor wild-type transfected cells. The experiment was performed in duplicates, shown on separate axis. Red box indicates clones with <25% activity as compared to wild-type hIP receptor. (B) Exemplary images of vector or hIP receptor expressing cells, stained with Alexa-647 anti HA-antibody. (C) Bar chart showing the relative expression level (as assessed by staining with Alexa-647 anti-HA-antibody) and the relative activity (as assessed by GloSensor cAMP assay) for 34 selected constructs. Asterisk indicates mutants classified as hits. Data were normalized to hIP receptor wild-type expressing cells and represents the mean ± SEM of three independent experiments.

https://doi.org/10.1371/journal.pone.0097973.g002

Random mutagenesis of hIP receptor identified inactive mutants

The observed distribution of mutations within the 18 clones showed an enrichment in clones with >1 mutations as compared to the distribution of the complete library (Figure 3A, Table 6), reflecting the higher probability of accumulated mutations to result in non-functional sequences [31]. A total of 38 non-synonymous mutations were identified within the coding region of hIP receptor, mapping to 36 distinct residues across the whole length of the protein (Figure 3B, Table 6). As depicted in Table 6 and Figure 3A, most of the mutants contained more than one non-synonymous mutation, making the interpretation challenging. Further work is therefore required to deduce which mutations are causative of loss of function. However, our hits included several previously reported mutations affecting the signaling of hIP receptor (Table 7). As shown in Figure 3B, our data identified three regions of hIP receptor with an accumulation of inactive mutations, indicating regions critical for hIP receptor function: 1) the region around Tyr-75, 2) the region around Cys-165 and 3) the region around Pro-289. Stitham et al. identified the highly conserved residues Arg-279 and Tyr-75 as lying within the PGl2 binding pocket of the hIP receptor and to be critical for ligand binding [14][16]. Cys-165 is described to form a non-conserved disulfide-bridge in human hIP receptor and mutations at this position has been shown to affect receptor expression, dimerization and activation [2], [3]. Pro-289 maps to TMVII and is part of the highly conserved DPXXF motif of the hIP receptor, a region involved in receptor activation [14], [15]. Our list of mutations also included a mutation of Asn-78 to Ile (N78I), a known glycosylation site of IP-receptor and previously implicated in receptor activation and ligand binding [37], Met-99, previously identified to contribute to ligand binding [16] and Ser-106, described to contribute to the process of signal transduction and receptor activation [17]. Notably, we identified the mutations shown in Tables 6 and 7 and Fig. 3B in a single random mutagenesis experiment. Thus, our findings corroborate the advantages of an unbiased random mutagenesis approach over targeted, site-directed mutagenesis of selected residues in multiple experiments. Taken together, these data prove the usefulness of random mutagenesis to gain insights in the structure-function requirements of hIP receptor and to identify mutations that prevent activation of the receptor.

thumbnail
Figure 3. Functional characterization of hIP receptor identified inactive mutants of hIP.

(A) Bar chart depicting the number and distribution of amino acid changes found in the 18 inactive mutants. Black bars represent the total number of amino acid changes; grey bars represent non-synonymous mutations only. (B) Bar chart showing the number and position of all 38 identified non-synonymous mutations in hIP receptor. Black boxes represent the α-helical, membrane-spanning domains of hIP receptor.

https://doi.org/10.1371/journal.pone.0097973.g003

thumbnail
Table 6. Summary of identified non-synonymous mutations in hIP-receptor.

https://doi.org/10.1371/journal.pone.0097973.t006

thumbnail
Table 7. Overview of identified mutations in hIP-receptor previously reported in literature.

https://doi.org/10.1371/journal.pone.0097973.t007

Conclusions

The recent advantages in sequencing techniques have allowed us to study genetic variants of genes at a genome-wide scale. A growing list of studies and clinical trials has started to investigate the role of SNPs in the origin and natural progression of diseases. However, the great number of identified variants and the often low prevalence have complicated the studies and only an unexpectedly small number of novel examples linking genetic variations to cardiovascular disease have been described [38]. Recent genomic sequencing studies have identified numerous SNPs in the PTGIR gene encoding the hIP receptor [14], [39][41]. Comparison of these mutations with the available data of previously biochemically characterized mutants has revealed a correlation between genetic variants in the hIP receptor resulting in deficits of hIP receptor function with increased disease severity in patients with coronary artery disease [10], [11]. Thus, detailed biochemical analysis of genetic variances in functional assays and insights into the structure/function relationship of the protein of interest can support the discovery of functional correlations of naturally occurring genetic variances and disease. In addition, information about the structural requirements underlying receptor function and dysfunction can aid to guide rational drug design and the development of novel therapies for cardiovascular disease.

Here we compared the applicability of chemical mutagenesis and epPCR mediated mutagenesis to generate an unbiased library of hIP receptor mutant plasmids, covering the complete coding region of PTGIR. We found that while chemical mutagenesis is especially useful to generate libraries with low frequencies of mutations it was not suitable for receptor mutation studies like the hIP receptor, where higher mutation rates are needed to guarantee sufficient coverage of the complete coding sequence [32], [42]. We describe a simple algorithm to simulate the optimal number and distribution of mutated amino acid in mutagenesis libraries as a function of the mutation rate. Using this simulation we determined the optimal mutation rate needed to generate a library of >4000 mutants of the hIP receptor by epPCR. We analyzed the quality of our library using NGS and present NGS as a rapid and non-laborious approach to determine the exact mutation rate and the mutational bias in the final library. The functional characterization of our library of >4000 mutants of the hIP receptor generated by epPCR identified 18 mutants of the hIP receptor with partial or complete loss of receptor activity while maintaining cell surface expression, translating into a hit rate less than 0.5%. Given the ∼800 mutants with impaired receptor activity in our primary screen (∼18% hit rate – not controlled for membrane expression), the difference in hit rate highlights the importance to determine the optimal mutation rate needed to avoid the accumulation of multiple mutations in one construct, thereby increasing the risk of preventing proper folding and trafficking to the membrane. The mutations identified in this mutagenesis screen included multiple novel mutations as well as mutations mapping to residues previously reported to affect the signaling of the hIP receptor [9], [11], [14][17], [37]. In these 18 constructs, we identified mutations across the whole length of the protein, mapping to 36 distinct residues of the hIP receptor. Due to the prevalence of multiple mutations within each construct, further work is required to characterize the novel mutations identified in this study. Notably, all these mutants were identified in a single experiment, highlighting the utility of unbiased mutagenesis approaches to investigate the structure-function relationship of GPCRs and other protein classes. Taken together, the results presented in this study demonstrate that epPCR mediated mutagenesis is a valuable method for unbiased screening of residues regulating hIP receptor expression and function.

Supporting Information

Table S1.

Worksheet to simulate the expected number and distribution of amino acids in mutagenesis libraries as a function of the mutation rate. Insert information in the yellow highlighted fields.

https://doi.org/10.1371/journal.pone.0097973.s001

(XLSX)

Acknowledgments

We thank Joseph Loureiro for scientific support and guidance.

Author Contributions

Conceived and designed the experiments: AB ER TCK APO LM SJC MG LAG PJGK. Performed the experiments: AB ER TCK LF LB MTvD AM SB NAR. Analyzed the data: AB ER TCK LF LB MTvD AM SB NAR. Wrote the paper: AB ER TCK APO LM SJC MG LAG PJGK.

References

  1. 1. Narumiya S, Sugimoto Y, Ushikubi F (1999) Prostanoid receptors: structures, properties, and functions. Physiol Rev 79: 1193–1226.
  2. 2. Giguere V, Gallant MA, de Brum-Fernandes AJ, Parent JL (2004) Role of extracellular cysteine residues in dimerization/oligomerization of the human prostacyclin receptor. Eur J Pharmacol 494: 11–22.
  3. 3. Stitham J, Gleim SR, Douville K, Arehart E, Hwa J (2006) Versatility and differential roles of cysteine residues in human prostacyclin receptor structure and function. J Biol Chem 281: 37227–37236.
  4. 4. Egan KM, Lawson JA, Fries S, Koller B, Rader DJ, et al. (2004) COX-2-derived prostacyclin confers atheroprotection on female mice. Science 306: 1954–1957.
  5. 5. Xiao CY, Hara A, Yuhki K, Fujino T, Ma H, et al. (2001) Roles of prostaglandin I(2) and thromboxane A(2) in cardiac ischemia-reperfusion injury: a study using mice lacking their respective receptors. Circulation 104: 2210–2215.
  6. 6. Cheng Y, Austin SC, Rocca B, Koller BH, Coffman TM, et al. (2002) Role of prostacyclin in the cardiovascular response to thromboxane A2. Science 296: 539–541.
  7. 7. Murata T, Ushikubi F, Matsuoka T, Hirata M, Yamasaki A, et al. (1997) Altered pain perception and inflammatory response in mice lacking prostacyclin receptor. Nature 388: 678–682.
  8. 8. Stitham J, Arehart EJ, Gleim SR, Douville KL, Hwa J (2007) Human prostacyclin receptor structure and function from naturally-occurring and synthetic mutations. Prostaglandins Other Lipid Mediat 82: 95–108.
  9. 9. Martin KA, Gleim S, Elderon L, Fetalvero K, Hwa J (2009) The human prostacyclin receptor from structure function to disease. Prog Mol Biol Transl Sci 89: 133–166.
  10. 10. Arehart E, Stitham J, Asselbergs FW, Douville K, MacKenzie T, et al. (2008) Acceleration of cardiovascular disease by a dysfunctional prostacyclin receptor mutation: potential implications for cyclooxygenase-2 inhibition. Circ Res 102: 986–993.
  11. 11. Stitham J, Arehart E, Elderon L, Gleim SR, Douville K, et al. (2011) Comprehensive biochemical analysis of rare prostacyclin receptor variants: study of association of signaling with coronary artery obstruction. J Biol Chem 286: 7060–7069.
  12. 12. Sato TK, Yamada RG, Ukai H, Baggs JE, Miraglia LJ, et al. (2006) Feedback repression is required for mammalian circadian clock function. Nat Genet 38: 312–319.
  13. 13. Wickham H (2011) The Split-Apply-Combine Strategy for Data Analysis. Journal of Statistical Software 40: 1–29.
  14. 14. Stitham J, Stojanovic A, Hwa J (2002) Impaired receptor binding and activation associated with a human prostacyclin receptor polymorphism. J Biol Chem 277: 15439–15444.
  15. 15. Stitham J, Martin KA, Hwa J (2002) The critical role of transmembrane prolines in human prostacyclin receptor activation. Mol Pharmacol 61: 1202–1210.
  16. 16. Stitham J, Stojanovic A, Merenick BL, O'Hara KA, Hwa J (2003) The unique ligand-binding pocket for the human prostacyclin receptor. Site-directed mutagenesis and molecular modeling. J Biol Chem 278: 4250–4257.
  17. 17. Stitham J, Stojanovic A, Ross LA, Blount AC Jr, Hwa J (2004) Clusters of transmembrane residues are critical for human prostacyclin receptor activation. Biochemistry 43: 8974–8986.
  18. 18. Kadonaga JT, Knowles JR (1985) A simple and efficient method for chemical mutagenesis of DNA. Nucleic Acids Res 13: 1733–1745.
  19. 19. Selifonova O, Valle F, Schellenberger V (2001) Rapid evolution of novel traits in microorganisms. Appl Environ Microbiol 67: 3645–3649.
  20. 20. Shortle D, Nathans D (1978) Local mutagenesis: a method for generating viral mutants with base substitutions in preselected regions of the viral genome. Proc Natl Acad Sci U S A 75: 2170–2174.
  21. 21. Stolarski R, Kierdaszuk B, Hagberg CE, Shugar D (1987) Mechanism of hydroxylamine mutagenesis: tautomeric shifts and proton exchange between the promutagen N6-methoxyadenosine and cytidine. Biochemistry 26: 4332–4337.
  22. 22. Labrou NE (2010) Random mutagenesis methods for in vitro directed enzyme evolution. Curr Protein Pept Sci 11: 91–100.
  23. 23. Cadwell RC, Joyce GF (1992) Randomization of genes by PCR mutagenesis. PCR Methods Appl 2: 28–33.
  24. 24. Beckman RA, Mildvan AS, Loeb LA (1985) On the fidelity of DNA replication: manganese mutagenesis in vitro. Biochemistry 24: 5810–5817.
  25. 25. Eckert KA, Kunkel TA (1991) DNA polymerase fidelity and the polymerase chain reaction. PCR Methods Appl 1: 17–24.
  26. 26. Eckert KA, Kunkel TA (1993) Effect of reaction pH on the fidelity and processivity of exonuclease-deficient Klenow polymerase. J Biol Chem 268: 13462–13471.
  27. 27. Cline J, Braman JC, Hogrefe HH (1996) PCR fidelity of pfu DNA polymerase and other thermostable DNA polymerases. Nucleic Acids Res 24: 3546–3551.
  28. 28. Tindall KR, Kunkel TA (1988) Fidelity of DNA synthesis by the Thermus aquaticus DNA polymerase. Biochemistry 27: 6008–6013.
  29. 29. Vanhercke T, Ampe C, Tirry L, Denolf P (2005) Reducing mutational bias in random protein libraries. Anal Biochem 339: 9–14.
  30. 30. Wang D, Zhao C, Cheng R, Sun F (2000) Estimation of the mutation rate during error-prone polymerase chain reaction. J Comput Biol 7: 143–158.
  31. 31. Drummond DA, Iverson BL, Georgiou G, Arnold FH (2005) Why high-error-rate random mutagenesis libraries are enriched in functional and improved proteins. J Mol Biol 350: 806–816.
  32. 32. Wong TS, Zhurina D, Schwaneberg U (2006) The diversity challenge in directed protein evolution. Comb Chem High Throughput Screen 9: 271–288.
  33. 33. Wong TS, Roccatano D, Zacharias M, Schwaneberg U (2006) A statistical analysis of random mutagenesis methods used for directed protein evolution. J Mol Biol 355: 858–871.
  34. 34. Fan F, Binkowski BF, Butler BL, Stecha PF, Lewis MK, et al. (2008) Novel genetically encoded biosensors using firefly luciferase. ACS Chem Biol 3: 346–351.
  35. 35. Binkowski BF, Fan F, Wood KV (2011) Luminescent biosensors for real-time monitoring of intracellular cAMP. Methods Mol Biol 756: 263–271.
  36. 36. Kuwano K, Hashino A, Asaki T, Hamamoto T, Yamada T, et al. (2007) 2-[4-[(5,6-diphenylpyrazin-2-yl)(isopropyl)amino]butoxy]-N-(methylsulfonyl)acetam ide (NS-304), an orally available and long-acting prostacyclin receptor agonist prodrug. J Pharmacol Exp Ther 322: 1181–1188.
  37. 37. Zhang Z, Austin SC, Smyth EM (2001) Glycosylation of the human prostacyclin receptor: role in ligand binding and signal transduction. Mol Pharmacol 60: 480–487.
  38. 38. Koboldt DC, Steinberg KM, Larson DE, Wilson RK, Mardis ER (2013) The next-generation sequencing revolution and its impact on genomics. Cell 155: 27–38.
  39. 39. Saito S, Iida A, Sekine A, Kawauchi S, Higuchi S, et al. (2003) Catalog of 178 variations in the Japanese population among eight human genes encoding G protein-coupled receptors (GPCRs). J Hum Genet 48: 461–468.
  40. 40. Stitham J, Arehart EJ, Gleim S, Douville K, MacKenzie T, et al. (2007) Arginine (CGC) codon targeting in the human prostacyclin receptor gene (PTGIR) and G-protein coupled receptors (GPCR). Gene 396: 180–187.
  41. 41. Stitham J, Arehart E, Gleim SR, Li N, Douville K, et al. (2007) New insights into human prostacyclin receptor structure and function through natural and synthetic mutations of transmembrane charged residues. Br J Pharmacol 152: 513–522.
  42. 42. Rasila TS, Pajunen MI, Savilahti H (2009) Critical evaluation of random mutagenesis by error-prone polymerase chain reaction protocols, Escherichia coli mutator strain, and hydroxylamine treatment. Anal Biochem 388: 71–80.