Article pubs.acs.org/ac
Construction of a Gene Screening System Using Giant Unilamellar Liposomes and a Fluorescence-Activated Cell Sorter Takehiro Nishikawa,† Takeshi Sunami,†,‡ Tomoaki Matsuura,†,§ Norikazu Ichihashi,†,‡ and Tetsuya Yomo*,†,‡,∥ †
Yomo Dynamical Micro-scale Reaction Environment Project, ERATO, Japan Science and Technology, 1-5 Yamadaoka, Suita, Osaka, 565-0871, Japan ‡ Graduate School of Information Science and Technology, Osaka University, 1-5 Yamadaoka, Suita, Osaka, 565-0871, Japan § Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka, 565-0871, Japan ∥ Graduate School of Frontier Biosciences, Osaka University, 1-5 Yamadaoka, Suita, Osaka, 565-0871, Japan S Supporting Information *
ABSTRACT: We have constructed a gene screening system composed of an in vitro transcription−translation system encapsulated within giant unilamellar liposomes and a fluorescence-activated cell sorter (FACS), which allows highthroughput screening of genes encoding proteins of interest. A mock gene library of β-glucuronidase (GUS) was compartmentalized into liposomes at the single-molecule level, and liposomes exhibiting green fluorescence derived from hydrolysis of the fluorogenic substrate by the synthesized enzyme were sorted using FACS. More than 10-fold enrichment of GUS gene with higher catalytic activity was obtained when a single copy of the GUS gene was encapsulated in each liposome. Quantitative analysis of the enrichment factors and their liposome size dependencies showed that experimentally obtained and theoretical values were in agreement. Using this method, genes encoding active GUS were then enriched from a gene library of randomly mutated GUS genes. Only three rounds of screening were required, which was also consistent with our theoretical estimation.
I
catalytic activities.12−16 FACS-based IVC is therefore a valuable technique for high-throughput screening of enzymes with desired catalytic activity. In previous studies, FACS has been used for IVC by encapsulating a fluorogenic substrate into water-in-oil-in-water (w/o/w) emulsion droplets, and enzymes exhibiting β-galactosidase activity were screened from a gene library using FACS4 as droplets with genes encoding higher levels of β-galactosidase activity show stronger fluorescent signals. Although double emulsions were successfully applied to the directed evolution of β-galactosidase activity, the environment where the protein is synthesized is different from that of the living cell, which limits the type of proteins that can be used for IVC; for example, this technology is not applicable to membrane-associated proteins. In addition, because of the presence of the oil phase, components that are highly soluble in oil cannot be included. Therefore, the present study was performed to develop a FACS-based IVC that is, in principle, more applicable to a wider range of proteins and reactions. Here, we propose application of giant unilamellar liposomes, also known as giant unilamellar vesicles (GUV), as microcompartments for FACS-based IVC. Our previous attempt to
n vitro compartmentalization (IVC) is a methodology for the screening of genes, which has been applied to the directed evolution of enzymes.1−4 In this method, a gene library is first compartmentalized into water-in-oil (w/o) emulsion droplets at nearly the single-molecule level together with an in vitro transcription−translation system, which enables expression of the encoded enzymes. A screen is then performed for droplets expressing enzymes with the desired properties, e.g., with higher catalytic efficiency. For example, screening of a library of 3.4 × 107 mutated phosphotriesterase genes yielded a phosphotriesterase with kcat 63-fold higher than that of the wild-type enzyme.5 IVC utilizes an in vitro transcription and translation system that enables the expression and screening of proteins that cannot be expressed in vivo (e.g., toxic proteins). In addition, this system allows rapid screening of large gene libraries due to the absence of transformation steps and the time required for growth of living cells. Although other in vitro selection systems, including CIS display, mRNA display, and ribosome display, share these favorable characteristics,6 IVC is the only method that allows direct screening for catalytic turnover reaction entirely in vitro.7 Fluorescence-activated cell sorting (FACS) is a valuable technology for high-throughput screening, which enables the analysis and sorting of >107 droplets/h. FACS has been used successfully to select proteins with high binding affinities8−11 or © 2012 American Chemical Society
Received: March 8, 2012 Accepted: April 22, 2012 Published: April 23, 2012 5017
dx.doi.org/10.1021/ac300678w | Anal. Chem. 2012, 84, 5017−5024
Analytical Chemistry
Article
(pET-gusA-native or pET-gusA-snap, 1 ng/μL), amplification primers (0.4 μM for each primer), T7F (5′-TAATACGATCTACTATATAGGG-3′) and T7R (5′-GCTAGTTATTGCTCAGCGG-3′), 0.2 mM dNTPs, and 0.05 units/μL of Pyrobest DNA polymerase (TaKaRa Bio). The PCR products were purified using a QIAquick PCR purification kit (Qiagen, Hilden, Germany). Library Construction. To construct a gene library for the screening experiment, random mutations were introduced into gus_wt by error-prone PCR using Mutazyme II DNA polymerase (GeneMorph II random mutagenesis kit; Stratagene, La Jolla, CA). The reaction mixture contained the template DNA (gus_wt, 1 ng/μL), amplification primers [0.5 μM for each primer, T7-bGal-Ant 02 (5′-CCCGCGAAATTAATACGACTCACTATAGGG-3′) and T7-bGal-Sen 02 (5′CTCCTTTCAGCAAAAAACCCCTCAAGACCC-3′)], 0.8 mM dNTPs, and 0.05 units/μL of Mutazyme II DNA polymerase. The DNA elongation process was repeated 25 times in error-prone PCR. The PCR products were purified using a QIAquick PCR purification kit (Qiagen) and agarose gel electrophoresis on E-Gel CloneWell 0.8% (Invitrogen). After purification, the mutated GUS gene was amplified by PCR using KOD FX DNA polymerase (Toyobo, Osaka, Japan). Biosynthesis of β-Glucuronidase. β-Glucuronidase (GUS) was synthesized by in vitro transcription−translation using the PURE system,23 prepared in our laboratory by the protocol modified according as described in our previous reports.24−26 Various types of GUS (GUS_wt, GUS_SNAP, mutated GUS, and active variants of GUS) were synthesized from the genes encoding each type of GUS [gus_wt, gus_snap, mt_gus, and Rn_gus (n = 1, 2, and 3) of enriched genes of active GUS variants]. The reaction buffer solutions (20 μL/reaction) for protein synthesis were prepared by mixing the template DNA (PCR product, 0.1 nM) for GUS, PURE system, RNase inhibitor (20 units), T7 RNA polymerase (2.5 units), TA647 (1 μM) as an internal reference dye, PFB-FDGlcU (50 μM) as a fluorogenic substrate, and sucrose (330 mM). The reaction buffer was compartmentalized into giant unilamellar liposomes and incubated at 37 °C for 120 min. GUS synthesis was monitored by FACS (FACSAria; Becton Dickinson, Franklin Lakes, NJ) or real-time PCR system (Mx3005P; Agilent Technologies, Santa Clara, CA). Liposome Preparation. Giant unilamellar liposomes were prepared by the inverted emulsion method,20,27 which involved centrifuging a w/o emulsion layered on an aqueous buffer solution in a test tube. The w/o emulsion was prepared as follows: POPC and cholesterol were dissolved in chloroform (100 mg/mL, each). Liquid paraffin was added to the chloroform solutions of POPC and cholesterol to prepare 5 mg/mL solutions. The lipid solutions were heated at 80 °C for 20 min to remove chloroform. Each of the lipid solutions was mixed at 9:1 by weight to prepare the mixed solution of POPC and cholesterol. The residual chloroform was removed by heating the mixed solution at 80 °C for 20 min. The reaction buffer for GUS synthesis (20 μL) was added to the lipid mixture (400 μL) and stirred vigorously using a vortex mixer for 30 s to prepare the w/o emulsion. Then, 400 μL of the w/o emulsion was layered on 150 μL of aqueous dilution buffer (bottom aqueous phase) in a test tube. The dilution buffer consisted of 100 mM HEPES, 280 mM potassium glutamate, 20 mM Mg(OAc)2, NTPs (3.75 mM ATP, 2.5 mM GTP, 1.35 mM CTP, and 1.35 mM UTP), amino acids (0.357 mM 18 AA,
construct a FACS-based IVC system was based on giant multilamellar liposomes,17 which are known to have a complex internal substructure.18 Protein synthesis inside giant multilamellar liposomes was inhibited relative to that in bulk,17 and therefore, this is not an ideal reaction environment. On the other hand, our recent study indicated that GFP synthesis reaction inside giant unilamellar liposomes19 prepared by the inverted emulsion method20 proceeded similar to that in the test tube irrespective of the liposome size or phospholipid composition tested.31 Giant unilamellar liposomes provide a more biologically relevant environment in comparison with emulsion droplets or giant multilamellar liposomes and can be used for screening of proteins other than enzymes, e.g., membrane-associated proteins. As shown in our previous study,21 quantitative analysis of protein synthesis inside giant unilamellar liposomes, as well as that of their structural properties, such as size, lamellarity, and internal substructures, are feasible by FACS. With these techniques, FACS-based IVC using giant unilamellar liposomes can be evaluated quantitatively, i.e., whether the enrichment factors obtained experimentally are reasonable or not can be evaluated. In this paper, we describe the construction of a gene screening system composed of an in vitro transcription− translation system encapsulated in giant unilamellar liposomes and FACS for high-throughput screening of genes encoding proteins of interest. Initially, test gene screening was performed using a mock gene library consisting of a mixture of two different β-glucuronidase (GUS) genes exhibiting contrasting catalytic activities, where the activity of GUS was detected by the presence of a fluorescent substrate in GUS. Liposomes exhibiting stronger fluorescence signals were sorted according to their size using FACS, and the enrichment factors of the gene with higher GUS activity were estimated. Our results indicated that the enrichment factors increased with smaller liposome size, consistent with the theoretical estimates. The genes encoding active GUS mutants were then enriched using our screening method from a gene library of randomly mutated GUS genes. We successfully obtained a gene pool of active GUS mutants through only three rounds of screening, which was also consistent with our theoretical estimates. Finally, we discuss the possible application of the method developed in this study.
■
EXPERIMENTAL SECTION Materials. 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) was purchased from Avanti Polar Lipids (Alabaster, AL). Cholesterol was purchased from Nacalai Tesque (Kyoto, Japan). Transferrin from human serum, Alexa Fluor 647 Conjugate (TA647), was purchased from Molecular Probes (Eugene, OR). 5-(Pentafluorobenzoylamino)fluorescein di-β-D-glucuronide (PFB-FDGlcU) was purchased from Invitrogen (Carlsbad, CA). The in vitro transcription−translation system (PURE system) was a lab-made reagent and supplemented with RNase inhibitor (Promega, Madison, WI), T7 RNA polymerase (TaKaRa Bio, Otsu, Japan), TA647. All other chemicals were purchased from Wako Pure Chemicals (Osaka, Japan) unless otherwise indicated. DNA Templates for GUS Synthesis. Genes encoding wild-type GUS (gus_wt) or SNAP-tagged GUS (gus_snap) were prepared by polymerase chain reaction (PCR) amplification from the plasmids pET-gusA-native (Supporting Information, Figure S1) and pET-gusA-snap, respectively.22 The reaction mixtures for PCR contained the template plasmid 5018
dx.doi.org/10.1021/ac300678w | Anal. Chem. 2012, 84, 5017−5024
Analytical Chemistry
Article
Figure 1. In-liposome GUS synthesis. (a) Schematic representation of GUS synthesis using the PURE system and detection of GUS activity using PFB-FDGlcU. All of the reactants were encapsulated in liposomes. (b) Fluorescence microscopy image of liposomes (with red fluorescent dyelabeled lipid) encapsulating gus_wt, the PURE system, and PFB-FDGlcU after 2 h of incubation at 37 °C. Scale bar: 5 μm. (c) Two-dimensional dot plot of in-liposome GUS synthesis obtained by FACS measurement. Black and red dots indicate liposomes loaded with gus_wt (0.1 nM) before and after 2 h of incubation at 37 °C, respectively. Blue dots indicate liposomes loaded with gus_snap (0.1 nM) after 2 h of incubation at 37 °C.
Amplification of Collected Genes in Gene Screening. The genes encoding active GUS variants were screened and enriched from the compartmentalized gene library of mt_gus (randomly mutated GUS gene) using FACS. The details of the gene screening process are described in the main body of this report. The collected genes were amplified by PCR twice to obtain sufficient amounts of the collected genes for transfer to the next round of the gene screening experiment. KOD FX DNA polymerase (Toyobo) was used for PCR amplification because of the high fidelity of the enzyme in the elongation reaction. The reaction mixture for PCR contained the enriched genes of active GUS variants (template DNA), amplification primers [0.3 μM each primer, T7-bGal-Ant 02 (5′CCCGCGAAATTAATACGACTCACTATAGGG-3′) and T7-bGal-Sen 02 (5′-CTCCTTTCAGCAAAAAACCCCTCAAGACCC-3′)], 0.4 mM dNTPs, and 0.02 units/μL of KOD FX DNA polymerase. The DNA elongation process was repeated 30 times in the first amplification and 10 times in the second amplification. Before the second amplification, the PCR products obtained in the first amplification was purified using a spin column (Qiagen) and gel electrophoresis on an E-Gel CloneWell 0.8% (Invitrogen) to remove side products of PCR and remnants of low molecular weight compounds in the reaction mixture. Fluorescence Microscopy. Fluorescence imaging of liposomes was performed on an IX-70 (Olympus, Tokyo, Japan) equipped with a charge-coupled device camera (VB7010; Keyence, Osaka, Japan) and a fluorescence mirror unit, U-MWIBA3 (Olympus) for detection of fluorescein emission and Cy3−4040C (Semrock, Rochester, NY) for detection of rhodamine emission. A xenon short arc lamp (UXL-75XB; Ushio, Tokyo, Japan) was used as a light source for fluorescence microscopy with a U-RX-T power supply (Olympus). Liposomes used for the fluorescence microscope observation were stained with the red fluorescent lipid 2-(4,4difluoro-5-(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-inda-
0.3 mM Tyr, and 0.3 mM Cys), 1.5 mM creatine phosphate, and 330 mM glucose. The two-phase system was centrifuged at 18 000g and 4 °C for 30 min. Liposomes including giant unilamellar liposomes were formed when the emulsion droplets were passed through the oil−water interface and suspended in the bottom aqueous phase. An aliquot of the liposomal suspension was taken out of the bottom aqueous phase. Quantification of DNA Using Real-Time PCR. The amount and species of the gene molecules (gus_wt and gus_snap) were quantified and determined by real-time PCR using SYBR Premix Ex TaqII (TaKaRa Bio) and the primer sets for both of the target genes. The primers used for detection of gus_wt were “sp_wt forward” (5′-CTGCGAACAGGGCCTGCATG-3′) and “sp_wt reverse” (5′-TCAATCGCTTCCGGCTGATG-3′); the primers used for detection of gus_snap were “snap forward” (5′-CGCGTTGGCGGTAACAAGAAAG-3′) and “snap reverse” (5′-GCCGCAAGCTTTTATTGTTTGC3′). The reaction profile of each sample was recorded on a realtime PCR system (Mx3005P; Agilent). FACS Measurement. A fluorescence-activated cell sorter (FACS) (FACSAria: BD) was used for analysis of in-liposome reactions and collection of liposomes encapsulating the genes of interest. The cell sorter was equipped with a 70 μm nozzle, an argon ion laser with a wavelength of 488 nm, and a He−Ne laser with a wavelength of 633 nm. Two band-pass filters of 530 ± 30 and 660 ± 20 nm were installed to detect PFB-F (hydrolyzed product of PFB-FDGlcU, fluorogenic substrate for GUS) and TA647, respectively. Liposome suspensions including giant unilamellar liposomes were diluted 10-fold in 50 mM HEPES−KOH buffer (pH 7.6, 50 mM HEPES, 100 mM potassium glutamate, 13 mM Mg(OAc)2, and 330 mM glucose) and analyzed by FACS. The internal aqueous phase volume (Vliposome) of liposomes was evaluated from the fluorescence intensity (FI) of TA647 (FITA647) encapsulated in liposomes using the following equation: Vliposome(fL) = 0.00252(FITA647).17,21 5019
dx.doi.org/10.1021/ac300678w | Anal. Chem. 2012, 84, 5017−5024
Analytical Chemistry
Article
Figure 2. Schematic representation of our gene screening system. (a) Illustration of the liposome sorting and gene screening system based on the inliposome reaction and liposome size. Liposomes prepared with our method include those of various sizes (left). Liposomes are sorted based on the fluorescence intensity and their sizes using FACS (right). (b) Two-dimensional dot plot of liposome size and fluorescence intensity of reaction products. Gates (P1, P2, P3, and P4) were set in the dot plot for liposome sorting. Black and red dots represent the liposome population before and after incubation for 2 h at 37 °C detected using FACS. Liposomes emitting green fluorescence above the threshold fluorescence intensities of 200 or 1000 were sorted from each fraction defined by the gates. Two thick gray lines, i and ii, indicate the thresholds for sorting liposomes. (c) FACS analysis of the mixture of liposomes encapsulating TA647 (red) and PFB-F (green) before (red and green) and after (orange) incubation.
Second, we measured in-liposome gene expression using two GUS genes with contrasting expression patterns of catalytic activities, which were used in the following test gene screening experiment. Wild-type GUS (gus_wt) and GUS fused to the SNAP tag22,31 at the C terminus (gus_snap) genes were used in the test gene screening experiments. Here, we compared the inliposome gene expression of gus_wt with that of gus_snap to investigate how these two enzymes differ when analyzed by FACS. For FACS measurements, a red fluorescent protein (transferrin Alexa Fluor 647 conjugate) was included in the PURE system, the fluorescence signal of which was converted into the internal aqueous phase volume of each liposome. Twodimensional FACS dot plots (Figure 1c) demonstrated the distribution of liposomes in the 2D space represented by fluorescence intensity of PFB-F and internal aqueous phase volume. First, the data showed that liposomes prepared by the method used here generated those with a variety of sizes, ranging between 1 and 100 fL. Second, the number of liposomes exhibiting catalytic activity of GUS (fluorescence intensity of green light >200; Supporting Information Figure S3) was very different between gus_wt and gus_snap (10% or 0.1% of 30 000 liposomes exhibited GUS activity), respectively. On the basis of the contrasting expression patterns of GUS activity, we prepared a mock gene library consisting of gus_wt and gus_snap for the test gene screening experiment. Test Gene Screening on a Mock Library of the GUS Gene. Figure 2a illustrates how the genes of interest are enriched in our screening system. The mixture of two GUS genes, gus_wt and gus_snap (mock gene library), at a DNA concentration of 0.1 nM (0.06 molecules/fL), was compart-
cene-3-pentanoyl)-1-hexadecanoyl-sn-glycero-3-phosphocholine (β-BODIPY 581/591 C5-HPC; Molecular Probes, Eugene, OR) and encapsulated green fluorogenic substrate (PFBFDGlcU) for GUS.
■
RESULTS AND DISCUSSION Detection of In-Liposome GUS Synthesis. Our gene screening system consists of the genes of interest (encoding GUS), an in vitro transcription−translation system (PURE system23), giant unilamellar liposomes (referred to hereafter as “liposomes”), and FACS. Liposomes are identified using FACS by the methods introduced in our recent reports.21,28 A region, “P0” in a two-dimensional (2D) dot plot of side scattering (SSC)−forward scattering (FSC) (Supporting Information Figure S2) indicates where giant unilamellar liposomes are detected in the SSC−FSC dot plot. First, we compartmentalized a reaction system for GUS synthesis into liposomes. Figure 1a shows the synthetic route of GUS from DNA in a liposome. DNA encoding GUS is transcribed by T7 RNA polymerase to mRNA, which is then translated into monomeric GUS. The monomers are assembled into tetrameric GUS capable of expressing catalytic activity.29 The active GUS catalyzes the hydrolysis of the fluorogenic substrate PFB-FDGlcU.30 The fluorescent dye part (PFB-F) released from the substrate emits green light. Figure 1b shows a fluorescence microscopy image of liposomes in which the wild-type GUS gene (gus_wt) was expressed using the PURE system. In the image, green emission can be seen inside the orange circumference of the lipid bilayer membrane. The green fluorescence indicates the synthesis of active GUS inside the liposomes. 5020
dx.doi.org/10.1021/ac300678w | Anal. Chem. 2012, 84, 5017−5024
Analytical Chemistry
Article
mentalized into liposomes ranging in size between 1 and 100 fL. When DNA with such low concentration is partitioned into compartments of
