High throughput analysis of antibody binding and specificity

Abstract

The present disclosure provides a high throughput analysis system for studying binding between two proteins, such as a target protein and a target binding protein (TBP). The analysis system allows for parallel screening of large quantities of proteins and their binding targets by combining a microfluidic system for single cell analysis with an overlap extension reverse transcription PCR technique.

Description

1. Cross-references to related applications

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 480,918, filed January 20, 2023, and U.S. Provisional Patent Application No. 63 / 612,234, filed December 19, 2023, the disclosure of which is incorporated herein by reference in its entirety for all purposes.

[0002] 2. Sequence List This application contains a sequence list that has been electronically submitted in XML format, the entire contents of which are incorporated herein by reference. The XML copy was created in XX year, named 28152-53884-SEQLIST.xml, and is XX bytes in size. 3. Background Technology Antibody therapy is increasingly being used to treat intractable diseases such as cancer. However, the development of antibody drugs is both expensive and complex. This process involves the identification of antigens, the isolation of antibodies, and the characterization and screening of antibodies with the expected activity against the antigens. The workload for antibody development and characterization is enormous, yet the chance of obtaining a functional and effective antibody is extremely slim.

[0004] As highlighted by the 2019 coronavirus disease (COVID-19) pandemic, monoclonal antibodies can be a key tool for preventing and controlling infectious diseases in high-risk populations. The superior specificity of antibodies, while providing safety and efficacy, also means that diverse and evolving targets may escape treatment. For example, the early antibody bamlanivimab developed by Eli Lilly almost completely lost its efficacy against SARS-CoV-2 strains with the E484K mutation in the spike protein, a mutation that occurred less than a year after the start of the pandemic. Conversely, dual-antibody cocktails, serum-derived polyclonal antibodies, or recombinant polyclonal antibody libraries exhibit longer binding lifetimes and protection against evolving pathogens.

[0005] Innate immune responses are highly polyclonal and provide an ideal resource for mining diverse bindings. Methods for capturing and sequencing natural antibody pairings in these diverse libraries have been developed, but a simple, high-throughput method for assessing the overall binding specificity of individual antibodies remains lacking. One approach, called LIBRA-seq, uses DNA barcoded antigens to stain primary B cells, followed by single-cell sequencing to obtain computer-simulated antibody and antigen sequences. This method has been improved and successfully used to identify neutralizing CoV-2 antibodies, but a significant challenge is the need for separately expressed, purified, and barcoded antigens. Another strategy utilizes existing biological systems that selectively fuse two components. For example, Alpha-seq encodes two interacting libraries on opposite yeast mating types, allowing them to mate, followed by sequencing of diploid cells. ENTER-seq and RAPTR display libraries on lentiviral particles and mammalian cells, respectively, and sequence transduced cells to identify interacting pairings. While these methods are highly selective and sensitive, each fusion provides information on only a single pairing event, requiring the analysis of a large number of cells to map more complex interaction networks.

[0006] Ribosome display is an alternative display technique that enables genotype-phenotype linkage of soluble proteins. In its simplest form, the protein is translated from mRNA lacking a stop codon, thus preventing the dissociation of the protein-ribosome-mRNA complex. Over time, this method has been further improved by using additional stop sequences and recombinant expression reagents. Human open reading frame (ORF) libraries are generated in ribosome display platforms, paired with covalently linked DNA barcodes, and used to determine the antigen specificity of autoreactive antibodies in patient samples. This technique allows screening antigen libraries against antibody libraries but cannot recover antibody sequences, thus limiting its use to analytical methods. SMI-Seq uses gel-immobilized ribosomes to display cDNA barcode protein libraries and incubates them with another barcode library. Fluorescent sequencing is used to identify colocalization barcodes in interacting proteins; although this method is innovative, it is quite complex and has limited scalability.

[0007] Therefore, there is a continued need for improved methods for the generation and characterization of recombinant antibodies. Specifically, high-throughput detection of large numbers of antibodies and antigens is required to accelerate the development of therapeutic antibodies. 4. Summary of the Invention This disclosure provides a high-throughput library-by-library interaction platform that utilizes antigens displayed on cell surfaces and antibodies displayed on soluble ribosomes. This polyclonal mapping system (also referred to herein as “PolyMap”) can be used to study binding patterns between different classes of proteins, such as target proteins (e.g., antigens, ligands, or receptors) and their homologous binding partners (e.g., antibodies, receptors, or ligands). For example, this can be achieved by combining a microfluidic system for single-cell analysis with overlap-extension reverse transcription PCR, allowing for the parallel screening of large numbers of proteins and their binding targets. Such methods can be used for high-throughput analysis of target-binding proteins, such as characterizing individual antibodies in polyclonal mixtures without isolation and purification, and identifying binding partners between antibody libraries and antigen libraries (e.g., libraries of naturally occurring antigen variants). Unlike previous work, the antigen libraries are expressed on the surface of mammalian cells, supporting the native structures of a large number of human and viral proteins, and eliminating the need for purification. Antibody libraries are expressed as soluble single-stranded variable fragments (scFvs) via ribosome display, barcoded by their complementarity-determining region triplet sequence (CDR3H), and used for bulk staining of antigen-presenting cells. Each single cell may bind to thousands of antibody-ribosome-mRNA (ARM) complexes, encapsulated with RNA capture beads with unique barcodes, and the barcoded cDNA is sequenced to reveal antibody-antigen pairings. PolyMap is compatible with synthetic library generation methods, supporting not only analysis but also protein engineering applications. For example, PolyMap can be used to identify CoV-2-targeting antibodies with unique antigen-binding specificity.

[0009] Specifically, this disclosure provides a method for high-throughput analysis of target-binding proteins (“TBPs”), comprising: providing a library of target-modified cells, wherein each of the target-modified cells presents a target on a membrane; contacting the library of target-modified cells with a plurality of target-binding protein (“TBP”)-ribosome-mRNA (TRM) complexes to induce binding between the target-modified cells and the TRM complexes; generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the target-modified cells, one or more TRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; capturing RNA released by the single cell on a solid surface or within a semi-permeable shell; and generating a hybrid polynucleotide library comprising sequences from transcripts recognizing targets expressed by the single cells and / or sequences from mRNAs of the TRM complexes bound to the cells. In some embodiments, the method includes generating a library of hybrid polynucleotides comprising sequences from transcripts recognizing a target expressed by the single cell and sequences from mRNA of the TRM complex bound to the cell. In some embodiments, the method includes generating a library of hybrid polynucleotides comprising: a) sequences from transcripts recognizing a target expressed by the single cell and oligonucleotide sequences (e.g., barcode sequences) immobilized on particles (e.g., beads) associated with the single cell; or b) sequences from mRNA of the TRM complex bound to the cell and oligonucleotide sequences (e.g., barcode sequences) immobilized on particles (e.g., beads) associated with the single cell. In some embodiments, the method includes generating a library of hybrid polynucleotides comprising a first set of hybrid polynucleotides and a second set of hybrid polynucleotides, wherein a) the hybrid polynucleotides of the first set of hybrid polynucleotides comprise sequences from transcripts recognizing a target expressed by the single cell, and oligonucleotide sequences (e.g., barcode sequences) immobilized on particles (e.g., beads) associated with the single cell; and b) the hybrid polynucleotides of the second set of hybrid polynucleotides comprise sequences from mRNA of the TRM complex bound to the cell, and oligonucleotide sequences (e.g., barcode sequences) immobilized on particles (e.g., beads) associated with the single cell. In some embodiments, the method further includes sequencing the library of hybrid polynucleotides. In some embodiments, the method further includes the step of identifying target-TBP pairs based on sequencing of the library of hybrid polynucleotides. In some embodiments, the method further includes the step of identifying target-binding proteins that are specific to the target. In some embodiments, the method further includes the step of identifying the binding affinity or specificity of target-binding proteins that are specific to the target.

[0010] In some embodiments, the method further includes the step of determining the readout distribution of a plurality of TBPs in the plurality of TRM complexes on two or more target sites. In some embodiments, the method further includes the step of normalizing the readout distribution based on the input distribution of the plurality of TBPs.

[0011] In some embodiments, at least one target includes, or is conjugated to, a domain capable of inducing the expression of an activation marker in the target-modified cells when the target binds to TBP. In some embodiments, generating a plurality of monodisperse or polydisperse emulsion droplets includes the step of sorting the target-modified cells based on the presence or absence of the activation marker.

[0012] In some embodiments, the target-modified cells express a fusion protein comprising a target and a transmembrane domain. In some embodiments, the target-modified cells comprise a construct encoding said fusion protein comprising the target and a transmembrane domain. In some embodiments, the construct further comprises a barcode sequence. In some embodiments, the construct further comprises a sequence encoding a fluorescent protein. In some embodiments, the construct further comprises a sequence encoding a surface marker or a detectable tag.

[0013] In some embodiments, the method further includes separating the plurality of monodisperse or multidisperse emulsion droplets containing cells modified with the target by detecting the expression of the fluorescent protein.

[0014] In some embodiments, the target-modified cell library contains one cell clone that presents a target. In some embodiments, the target-modified cell library contains two, three, or four cell clones, wherein each cell clone presents a unique target that is different from the other cell clones. In some embodiments, the target-modified cell library contains at least five cell clones, wherein each cell clone presents a unique target that is different from the other cell clones. In some embodiments, the target-modified cell library contains at least ten cell clones, wherein each cell clone presents a unique target that is different from the other cell clones. In some embodiments, the target-modified cell library contains at least 100 cell clones, wherein each cell clone presents a unique target that is different from the other cell clones. In some embodiments, the target-modified cell library contains at least 1000 cell clones, wherein each cell clone presents a unique target that is different from the other cell clones.

[0015] In some embodiments, the transcript from the isolated single cell contains a coding sequence for the target. In some embodiments, the transcript from the isolated single cell contains a barcode sequence that uniquely identifies the target expressed by the cell.

[0016] In some embodiments, each of the TRM complexes contains a target-binding protein. In some embodiments, each of the TRM complexes contains scFv (e.g., each of the TRM complexes contains TBP containing scFv). In some embodiments, each TRM complex contains a heavy chain variable region (e.g., each of the TRM complexes contains TBP containing a heavy chain variable region). In some embodiments, each TRM complex contains a light chain variable region (e.g., each of the TRM complexes contains TBP containing a light chain variable region).

[0017] In some embodiments, the plurality of TRM complexes comprises 1 to 5 unique TRM complexes, wherein each unique TRM complex contains a unique target-binding protein that is different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprises 6 to 10 unique TRM complexes, wherein each unique TRM complex contains a unique target-binding protein that is different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprises at least 10 unique TRM complexes, wherein each unique TRM complex contains a unique target-binding protein that is different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprises at least 1000 unique TRM complexes, wherein each unique TRM complex contains a unique target-binding protein that is different from the other unique TRMs.

[0018] In some embodiments, each of the TRM complexes comprises a target-binding protein and mRNA encoding the target-binding protein. In some embodiments, the mRNA comprises a coding sequence for the complementarity-determining region (CDR) of the target-binding protein. In some embodiments, the mRNA comprises a coding sequence for CDR3 of the target-binding protein. In some embodiments, the mRNA comprises a barcode sequence that uniquely identifies the TBP or target-binding protein encoded by the mRNA.

[0019] In some embodiments, RNA capture is performed using oligonucleotides immobilized in beads. In some embodiments, RNA capture is performed using oligonucleotides immobilized in beads, wherein each bead has a diameter greater than 10 μm, 0.5-10 μm, less than 1 μm, or about 1 μm. In some embodiments, each bead is a solid bead or a porous bead.

[0020] In some embodiments, according to any of the methods described above for using bead-immobilized oligonucleotides, the bead-immobilized oligonucleotides contain a barcode sequence. In some embodiments, the barcode sequence in the bead-immobilized oligonucleotide is unique for a given bead. For example, in some embodiments, 1) a first bead contains an oligonucleotide immobilized on the first bead, wherein the oligonucleotide immobilized on the first bead contains a first barcode sequence; and 2) a second bead contains an oligonucleotide immobilized on the second bead, wherein the oligonucleotide immobilized on the second bead contains a second barcode sequence, and wherein the first barcode sequence and the second barcode sequence are different. This can be extended to any number of beads. In some embodiments, the hybrid polynucleotide also contains a barcode sequence from the bead-immobilized oligonucleotide.

[0021] In some embodiments, the hybrid polynucleotide is generated by overlap extension polymerase chain reaction (OE-PCR). In some embodiments, the hybrid polynucleotide is generated prior to the synthesis of the first-strand cDNA.

[0022] In some embodiments, the method further includes the step of generating a second set of monodisperse or polydisperse emulsion droplets, prior to the step of generating the hybrid polynucleotide library, the second set of monodisperse or polydisperse emulsion droplets containing bead-captured RNA released from the single cell. In some embodiments, the hybrid polynucleotide library is generated in the second set of monodisperse or polydisperse emulsion droplets.

[0023] In some implementations, the contact between the library of target-modified cells and multiple TBP-ribosome-mRNA (TRM) complexes is in the presence of 25 to 100 mM Mg 2+ The procedure is performed in a buffer solution. In some embodiments, the buffer solution contains 50 mMmg. 2+ In some embodiments, the buffer contains 50 mM MgCl2. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, and BSA. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, polysorbate 20, heparin, and BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer also contains an RNase inhibitor.

[0024] In one aspect, this disclosure provides a kit for high-throughput analysis of target-binding proteins, comprising: multiple constructs, each construct encoding a fusion protein containing a unique target and a transmembrane domain; multiple TBP-ribosome-mRNA (TRM) complexes, each TRM complex containing a unique target-binding protein; and a buffer.

[0025] In some implementations, the kit also includes a host cell. In some implementations, the kit includes a single construct encoding a target.

[0026] In some embodiments, the kit contains 2 to 10 unique constructs, each of which encodes a unique target. In some embodiments, the kit contains at least 10 unique constructs, each of which encodes a unique target. In some embodiments, the kit contains at least 100 unique constructs. In some embodiments, the kit contains at least 1000 unique constructs.

[0027] In some embodiments, the kit contains a unique TRM complex, wherein the unique TRM complex contains a target-binding protein. In some embodiments, the kit contains at least 10 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein. In some embodiments, the kit contains at least 100 unique TRM complexes. In some embodiments, the kit contains at least 1000 unique TRM complexes.

[0028] In some implementations, the kit also includes reagents for overlap extension polymerase chain reaction (OE-PCR).

[0029] In some implementations, the buffer solution contains 50 mM Mg 2+ In some embodiments, the buffer contains 50 mM MgCl2. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, and BSA. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, polysorbate 20, heparin, and BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer also contains an RNase inhibitor.

[0030] In another aspect, this disclosure provides a library of hybridized polynucleotides generated by the methods disclosed herein.

[0031] In another aspect, this disclosure provides a method for high-throughput analysis of receptors, comprising: providing a library of ligand-modified cells, wherein each of the ligand-modified cells presents a target ligand on a membrane; contacting the library of the ligand-modified cells with a plurality of receptor-ribosome-mRNA (RRM) complexes to induce binding between the ligand-modified cells and the RRM complexes; generating a plurality of monodisperse or multidisperse emulsion droplets, wherein each droplet contains a single cell from the ligand-modified cells, one or more RRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; capturing RNA released from the single cell on a solid surface or within a semi-permeable shell; and generating a library of hybridized polynucleotides containing sequences of transcripts from the single cell and / or sequences of mRNA from the RRM complexes.

[0032] This disclosure also provides a method for high-throughput analysis of receptors, comprising: providing a library of receptor-modified cells, wherein each of the receptor-modified cells presents a target receptor on a membrane; contacting the library of receptor-modified cells with a plurality of ligand-ribosome-mRNA (LRM) complexes to induce binding between the receptor-modified cells and the LRM complexes; generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the receptor-modified cells, one or more LRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; capturing RNA released from the single cell on a solid surface or within a semi-permeable shell; and generating a hybrid polynucleotide library containing sequences of transcripts from the single cell and / or sequences of mRNA from the LRM complexes.

[0033] This disclosure also provides a method for high-throughput analysis of receptors, comprising: providing a library of antigen-modified cells, wherein each of the antigen-modified cells presents a target antigen on a membrane; contacting the library of antigen-modified cells with a plurality of antibody-ribosome-mRNA (ARM) complexes to induce binding between the antigen-modified cells and the ARM complexes; generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the antigen-modified cells, one or more ARM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; capturing RNA released from the single cell on a solid surface or within a semi-permeable shell; and generating a hybrid polynucleotide library containing sequences of transcripts from the single cell and / or sequences of mRNA from the ARM complexes. 5. Description of the attached drawings These and other features, aspects, and advantages of the invention will become more readily understood with reference to the following description and the accompanying drawings, wherein: Figure 1 A high-throughput method for the analysis of target-binding proteins (antibodies) is outlined. (1a) A library of target (antigen)-modified cells is prepared by expressing the target library on cells; (1b) Multiple TBP-ribosome-mRNA (TRM) complexes are prepared; (2) The target-modified cell library and multiple TRM complexes are batch-mixed to induce binding between the target-modified cells and the TRM complexes; (3) Cells are encapsulated in droplets containing lysis buffer and RNA capture beads; (4) The RNA capture beads are separated and recapsulated in a second droplet in which the target barcode and antibody CDR3 sequence can be ligated and amplified; (5) The amplified DNA is isolated, batch-prepared for sequencing, and sequenced using methods such as Illumina and NGS. Bioinformatics analysis is used to normalize the data and plot TBP:target interaction maps.

[0035] Figure 2 The image shows FACS detection of cells expressing the spike protein (CoV1-S or CoV2-S) from one of six test constructs (V1, V2, V3, V4, V5, and V6) using mAbs targeting the spike protein.

[0036] Figure 3 The expression of the CoV2-S antigen in Expi293 or CHOZN cell lines after stable transfection with CoV2-S coding sequences (Expi293+CoV2-S, CHOZN+CoV2-S, and CHOZN+CoV2-S(HP+F)) is shown, as described in Example 1. CoV2-S expression was detected using bamranizumab (MFI).

[0037] Figure 4 An exemplary construct framework (p2G-FRT-GS) for generating target-modified cells is provided.

[0038] Figure 5 The structure of an exemplary construct (T7-based expression plasmid) for generating the TBP-ribosome-mRNA complex (TRM complex) is provided.

[0039] Figure 6This study demonstrates the detection of binding between the CoV1 or CoV2 spike protein WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) and the TRM complex of an scFv containing one of four mAbs (bamranizumab, camrelizumab, edevimab, or ipilimumab) on target-modified cells. TRM binding to the CoV2-S variant was detected by strep tag staining.

[0040] Figure 7 This study demonstrates the detection of binding between CoV1 or CoV2 spike protein WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) and a TRM complex containing one of four mAbs (bamranizumab, camrelizumab, edevimab, or ipilimumab) on target-modified cells. TRMs binding to CoV2-S variants were isolated and detected by TaqMan RT-qPCR and antibody-specific probes (TBP).

[0041] Figure 8 The specificity of binding between the CoV2-S antigen and the TRM complex was demonstrated by measuring the ratio between specific staining (RNA isolated from the binding of the CoV2-S antigen and the TRM complex containing camrevimumab) and non-specific staining (RNA isolated from the binding of the CoV2-S antigen and ipilimumab scFv). The TRM complex was generated under different cell-free translation (TL) conditions with varying template concentrations and reaction times.

[0042] Figure 9 This study demonstrates the detection of binding between CoV1 or CoV2 spike protein WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) and a mixture of five TRM complexes on target-modified cells, where each TRM complex contains a scFv of one of four mAbs (bamranibumab, camrelizumab, edevimab, or ipilimumab). TRMs binding to the CoV2-S variant were isolated and detected by TaqMan RT-qPCR and antibody-specific probes (TBP).

[0043] Figure 10Polymap scores representing the binding between five target-modified cell lines and five TRM complexes are provided, where each target-modified cell line expresses the CoV1 or CoV2 spike protein WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K), and each TRM complex contains an scFv of one of five mAbs (bamranibumab, carpreximab, edevimab, ipilimumab, or pembrolizumab). The polymap scores are calculated based on average proportions and scaled down using the average readout of the target cells as described in Example 6.

[0044] Figure 11 This demonstrates an overlap extension reverse transcription PCR of scFv RNA in the TRM complex with a barcode sequence in target-modified cells. OE-RT-PCR produces a result containing V H The product of zone sequence and barcode sequence.

[0045] Figure 12 The workflow of the PolyMap platform is outlined as follows: First, an antigen library expressed on the surface of mammalian cells is constructed, and then incubated with a library of soluble antibody scFv in ribosome-displayed form. Stained single cells are encapsulated and lysed using RNA capture beads with unique barcodes (represented by stars, triangles, and rhombuses). The beads are isolated and used to generate cDNA, which is then further amplified using gene-specific primers for sequencing. Analysis of cell barcodes, antigen barcodes, and antibody CDRH3 is used to generate an antibody-antigen interaction map.

[0046] Figure 13 The expression of the CoV-2 spike protein in different cell lines was shown. Results are presented as flow cytometry data, showing the surface expression levels of CoV-2 S on stable Expi293™ and CHOZN® cell lines. Variants marked as “stable” contain six proline mutations and removal of the furin protease site.

[0047] Figures 14A-14D Results for various antigen expression systems are shown. Figure 14A An antigen expression construct was described, comprising a CMV promoter with a translation enhancement element (2G), a signal peptide (SP), a transmembrane (TM) region for surface display, and a unique barcode (BC) located in the 3' untranslated region (UTR). The FRT site allows for the integration and selective translation of the glutamine synthase gene (GS). Figure 14B The surface expression of different spike protein mutants stably expressed in CHOZN cells is shown by monoclonal antibody and flow cytometry. Figure 14CA representative DNA fragment encoding the antibody scFv ribosome display library was depicted, containing the T7 promoter, Strep tag II, post-protein spacer sequence (TolA), and secM ribosome halting sequence. The 40-mer poly-A tail is encoded by the bovine growth hormone (BGH) polyadenylation signal at the 3'UTR. Figure 14D RNA recovered from spike protein variant cells stained with clinical antibodies as ARM complexes is shown, and as measured by RT-qPCR. The table shows the expected results based on literature and affinity studies.

[0048] Figure 15 Optimization of ARM complex production is shown, as illustrated by RT-qPCR of RNA recovered from CoV-2 SWT cells stained with mixtures of ARM complexes produced under different conditions. The recovery rates were quantified using unique specific probes targeting “positive” (carrevimumab) and “non-specific” RNA, and the recovery rates were calculated. The STD iv TXTL kit contains all the components required for transcription, translation, and termination reactions, while the ΔRF kit omits factors that assist in the release of ribosomes from mRNA. The ARM complexes were translated (TL) for 10 or 30 minutes prior to staining.

[0049] Figure 16 Biological layer interferometry (BLI) of clinical antibodies against CoV-2 spike protein variants is presented. The figure shows binding and dissociation curves at three concentrations, as well as globally calculated curve fit values.

[0050] Figures 17A-17E The screening results for five monoclonal antibodies and four SARS-CoV-2 spike protein variants are shown. For all experiments, ARM complexes were generated from equimolar mixtures of the five mAbs. Figure 17A Cell lines expressing the specified spike protein variants (columns) were stained with an ARM complex mixture, and then the RNA was recovered and amplified for Illumina sequencing. The data shown in the heatmap are the readout percentages for each antibody sequence. Figure 17B Single-cell sorting. Spike protein variant cell lines were pooled in equal volumes, stained with an ARM complex mixture, and then single-cell sorted into the wells of a 96-well plate. Antibody weight chain variable regions and antigen barcodes were amplified and sequenced, and the identity of the spike protein variant cell line in each well was determined using the antigen barcode. Data are presented as a heatmap of the average readout percentage of each antibody sequence associated with each antigen cell line. Figure 17C : Figure 17B The data is from single cells, where each cell is represented by five data points, and each data point represents one antibody. The data shown is the percentage of each antibody associated with each cell. Figure 17DDrop-seq binding profile. Equal volumes of spike protein variant cell lines were pooled and stained with an ARM complex mixture. RNA from single cells was captured onto beads, and the antibody heavy chain variable region and antigen barcode were amplified. These were then sequenced along with the linked Drop-seq cell barcodes. Antibody binding analysis was performed on 91 cells after combining the antigen and antibody cell barcode lists. The data are presented as a heatmap of the average readout percentage of the Drop-seq data. Figure 17E : Figure 17D The data is from single cells, where each cell is represented by five data points, and each data point represents one antibody. The data shown is the percentage of each antibody associated with each cell.

[0051] Figure 18 RNA enrichment was shown under two different staining conditions. Based on sequencing of CoV-2 SWT cell lines before and after staining at two different ARM complex concentrations, fold enrichment of a mixture of 14 antibody clones was achieved.

[0052] Figure 19 The image shows the sequence alignment of selected CoV-2 spike protein variants. The amino acids corresponding to the receptor-binding domain (RBD) are marked in the figure.

[0053] Figure 20A and 20B The binding spectra of a library of anti-SARS-CoV-2 antibodies against spike protein variants are shown. Figure 20A The antibody library for generating the ARM complex was identified by sequencing the heavy chain fragments in the RNA mixture used as input for the in vitro translation reaction. The figure plots the top 100 clones, along with the individual and cumulative percentages of each clone in the library. Figure 20B Drop-seq. Cell lines expressing the specified spike protein variants (columns) were merged, stained with an ARM complex mixture, encapsulated with barcoded beads, and subjected to a drop-seq workflow. The normalized readout percentage for each antibody across all cell lines was plotted. The top 40 antibodies are shown in the figure (based on total antibody readouts).

[0054] Figure 21 The distribution of antigen cells in the Drop-seq library experiment is shown. After ARM complex staining, barcode Drop-seq bead encapsulation (ChemGenesCorporation), and antigen amplification, the identity of the recovered antigens was determined by sequencing the relevant antigen barcodes.

[0055] Figure 22A and 22B This demonstrates the functional validation of the antibody binding patterns identified by PolyMap. Figure 22ACell lines expressing the specified spike mutant (x-axis) were stained with an anti-CoV-2 library ARM complex mixture, and then the RNA was recovered and amplified for Illumina sequencing. The data shown in the heatmap are the log2 fold change in enrichment of each antibody sequence associated with each cell line relative to the input. Antibodies with a log2 fold change value lower than any of the three negative control antibodies are marked in white. Figure 22B : A subset of single antibody clones validated by flow cytometry and compared with previous data. For each antibody, CDR3H sequences and heatmaps from three datasets are shown: top, Drop-seq (from Figure 20B ); in the middle, single antigen staining (from Figure 22A (); Bottom; Flow cytometry. For flow cytometry, full-length monoclonal antibodies were expressed in CHO cells, and the supernatant was used to stain single spike protein variant cell lines. The heatmap shows the MFI of the secondary antibody signal.

[0056] Figure 23 RNA enrichment of clones with different affinities is shown. Five clones in the mixture were enriched based on sequencing results before and after staining of CoV-2 S wild-type cell lines. Affinity was determined by BLI.

[0057] Figure 24 This diagram shows the distribution of antigen sequences from batch-transfected cells. gDNA was isolated from the transfected cells and sequenced to determine the relative frequency of each unique antigen barcode.

[0058] 6. Specific Implementation Plan 6.1. Definition As used herein, the term "target-binding protein" or TBP refers to a protein containing one or more target-binding domains that specifically bind to a target. In some embodiments, the target-binding domain binds to a target or a fragment thereof with similar specificity and affinity to a natural antibody. In some embodiments, the TBP comprises an antibody. In some embodiments, the TBP is composed of an antibody. In some embodiments, the TBP is substantially composed of an antibody. In some embodiments, the TBP comprises an alternative scaffold. In some embodiments, the TBP is composed of an alternative scaffold. In some embodiments, the TBP is substantially composed of an alternative scaffold. In some embodiments, the TBP comprises an antibody fragment. In some embodiments, the TBP is composed of an antibody fragment. In some embodiments, the TBP is substantially composed of an antibody fragment. In some embodiments, the TBP comprises a receptor or a fragment thereof that binds to a target. In some embodiments, the TBP is a receptor or a fragment thereof that binds to a target. In some embodiments, the TBP comprises a ligand or a fragment thereof. In some embodiments, the TBP is a ligand or a fragment thereof.

[0059] The term "antibody" is used in its broadest sense herein to include certain types of immunoglobulin molecules that contain one or more antigen-binding domains that specifically bind to antigens or epitopes. Antibodies specifically include intact antibodies (e.g., intact immunoglobulins), antibody fragments, and multispecific antibodies. An example of an antigen-binding domain is a V... H -V L Antigen-binding domains formed by dimers. Antibodies are a type of ABP.

[0060] The term "alternative scaffold" refers to a molecule whose one or more regions can be diversified to generate one or more antigen-binding domains that specifically bind to antigens or epitopes. In some embodiments, the antigen-binding domains bind to a target or fragment thereof with similar specificity and affinity to natural antibodies. Exemplary alternative scaffolds include those derived from fibronectin (e.g., adnectins). TM ), β-sandwich proteins (e.g., iMab), lipid transport proteins (e.g., anticalins) ® ), EETI-II / AGRP, BPTI / LACI-D1 / ITI-D2 (e.g., Kunitz domain), thioredoxin peptide aptamers, protein A (e.g., Affibody) ® Ankylosing skeletal repeats (e.g., DARPins), γ-B-crystal proteins / ubiquitins (e.g., Affilins), CTLD3 (e.g., tetratranectins), Fynomers, and those (LDLR-A modules) (e.g., Avimers). More information on alternative scaffolds is available at Binz. et al. , Nat. Biotechnol. , 2005 23:1257-1268; Skerra, Current Opinion in Biotech. , 2007 18:295-304 and Silacci et al. , J. Biol. Chem. , 2014, 289:14392-14398; the entire contents of which are incorporated herein by reference. An alternative stent is a type of TBP.

[0061] The term "target-binding domain" refers to the portion of the TBP that can specifically bind to a target.

[0062] An "antibody fragment" contains a portion of a complete antibody, such as the antigen-binding region or variable region of the complete antibody. Antibody fragments include, for example, Fv fragments, Fab fragments, F(ab')2 fragments, Fab' fragments, scFv(sFv) fragments, and scFv-Fc fragments.

[0063] A single-chain Fv, sFv, or scFv antibody fragment contains a V in a single polypeptide chain. H A structural domain and a V L Domain. V H and V L Typically, the linker is connected via a peptide linker. See Plückthun A. (1994). In some embodiments, the linker is (GGGGS). n In some implementations, n = 1, 2, 3, 4, 5, or 6. See Antibodies from Escherichia coli .In Rosenberg M.&Moore GP (Eds.), The Phtrmacology of Monoclonal Antibodies vol. 113 (pp. 269-315). Springer-Verlag, New York, which is incorporated into the full text by reference.

[0064] “Affinity” refers to the strength of the sum of non-covalent interactions between a single binding site of a molecule (e.g., ABP) and its binding partner (e.g., an antigen or epitope). Unless otherwise stated, “affinity” as used herein refers to intrinsic binding affinity, reflecting a 1:1 interaction between members of a binding pair (e.g., ABP and antigen or epitope). The affinity of molecule X for its partner Y can be expressed using the dissociation equilibrium constant (K0). D The kinetic components affecting the dissociation equilibrium constant are described in more detail below. Affinity can be measured by methods commonly known in the art, including those described herein. For example, surface plasmon resonance (SPR) techniques (e.g., BIACORE) can be used. ® ) or biological layer interference (e.g., FORTEBIO) ® To determine affinity.

[0065] Regarding the binding of TBP to target molecules, the terms "binding," "specific binding," "specifically bound to," "specifically to," "selectively binding," and "selectively targeting" refer to binding to a specific antigen (e.g., a peptide target) or an epitope on a specific antigen that differs significantly from nonspecific or nonselective interactions (e.g., with non-target molecules). Specific binding can be measured, for example, by measuring the binding to a target molecule and comparing it to the binding to a non-target molecule. Specific binding can also be determined by competition with a control molecule that simulates an epitope recognized on a target molecule. In this case, if the binding of TBP to the target molecule is competitively inhibited by the control molecule, specific binding is indicated.

[0066] As used herein, the term "TBP-ribosome-mRNA complex" or "TRM complex" refers to a complex comprising a target-binding protein, a ribosome, and mRNA. TRM complexes can be generated by any method known in the art. For example, TRM complexes can be generated using an in vitro cell-free system. In vitro ribosome display technology can generate stable protein (target-binding protein)-ribosome-mRNA (TRM) complexes by linking a single target-binding protein to its corresponding mRNA. TRM complexes can be formed by deleting a stop codon from the mRNA, which causes translational ribosome arrest at the mRNA terminus and prevents the release of nascent polypeptides. Protein-mRNA linking allows for the simultaneous separation of mRNA and the desired protein (target-binding protein) by affinity for the immobilized ligand. The protein-mRNA complex, tightly bound to the ligand, can be subjected to in situ reverse transcription PCR (RT-PCR) to recover the DNA sequence encoding the protein and amplified in the PCR reaction to generate a template for further manipulation and protein processing.

[0067] As used herein, the term "droplet" refers to a small amount of liquid. Droplets are typically spherical, but can also consist of cylindrical droplets spanning the entire diameter of a microfluidic channel. Droplets can form in air, oil, or aqueous solutions, depending on their composition and formation method. Droplets can exist in both monodisperse and polydisperse populations.

[0068] As used herein, the term "monodisperse" refers to the property of components having uniform or nearly uniform sizes. For example, monodisperse droplets typically require >90% of the droplets in the mixture to have a size dispersion of <5%. In many cases, monodisperse droplet swarms are more stable than non-monodisperse droplet swarms (i.e., polydisperse droplet swarms). In some embodiments, the generation of monodisperse droplets requires some form of controlled microfluidic device.

[0069] As used in this article, the term "cell clone" refers to at least two similar species or classes of cells.

[0070] As used herein, the term "lysis" refers to the process of disrupting the cell membrane of one or more cells through physical or chemical methods. Lysis can be achieved by chemical surfactants (such as Triton X-100), alkaline lysis buffers, heating, electric current, or physical disruption.

[0071] 6.2. Other Interpretive Conventions The ranges described herein should be understood as abbreviations of all values ​​within that range, including the endpoints. For example, the range 1 to 50 should be understood as including any number, combination of numbers, or subrange of the group consisting of 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, and 50.

[0072] 6.3. Methods for high-throughput analysis of antibodies This disclosure provides a method for high-throughput antibody analysis. Specifically, the method may include the following steps: (i) providing a library of target-modified cells, wherein each target-modified cell presents a target on a membrane; (ii) contacting the library of target-modified cells with a plurality of TBP-ribosome-mRNA (TRM) complexes to induce binding between the target-modified cells and the TRM complexes; (iii) generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the target-modified cells, one or more TRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; (iv) capturing RNA released by the single cell on a solid surface or within a semi-permeable shell; and (v) generating a hybrid polynucleotide library containing sequences of transcripts from the single cell and / or sequences of mRNA from the TRM complexes. This disclosure also provides variations or modifications of the antibody analysis method.

[0073] Analysis or sequencing of a hybrid nucleic acid library provides information related to the binding between the target and the TRM complex. For example, based on the sequence, the binding affinity or specificity of TBP to the target can be studied. Therefore, the method may also include a step of identifying target-TBP pairs based on sequencing of the hybrid polynucleotide library. In some embodiments, multiple target-TBP pairs are identified by sequencing the hybrid polynucleotide library. In some embodiments, more than two, three, four, five, six, seven, eight, nine, ten, twenty, or more target-TBP pairs are identified. In some embodiments, the method further includes a step of identifying a target-binding protein specific to the target. In some embodiments, the method further includes a step of identifying the binding affinity or specificity of a target-binding protein specific to the target. 6.3.1. Target-modified cells The methods disclosed herein use target-modified cells expressing a fusion protein containing a target and a transmembrane domain. In some cases, the target-modified cells express a transmembrane protein containing the target, rather than a fusion protein. The target-modified cells present the target to a surface. In some embodiments, the fusion protein also includes a signaling domain that facilitates the delivery of the target to the surface.

[0074] In some embodiments, the target is an antigen. In some embodiments, the target is a ligand of the receptor or a modification thereof.

[0075] In some embodiments, the target-modified cells contain the coding sequence of the target. In some embodiments, a polynucleotide construct encoding the target is transiently transfected into the target-modified cells. In some embodiments, the polynucleotide construct encoding the target is stably transfected into the target-modified cells. In some embodiments, the polynucleotide construct is a lentiviral vector. In some embodiments, the polynucleotide construct contains an FRT site, which allows the coding sequence of the target to be stably integrated into the landing site. In some embodiments, the polynucleotide construct containing the FRT site is transfected with a recombinase to achieve stable integration.

[0076] In some embodiments, the target is expressed in target-modified cells by exogenous polynucleotides. In some such embodiments, the target-modified cells contain a construct encoding a fusion protein or transmembrane protein containing the target. In some embodiments, the construct contains a coding sequence for a fusion protein or transmembrane protein operatively linked to a regulatory sequence (e.g., a promoter). In some embodiments, the construct also contains a barcode sequence. In some embodiments, the barcode sequence indicates expression or presentation of the target on the target-modified cells. In some embodiments, the construct also contains a sequence encoding a reporter protein. In some embodiments, the reporter protein is a fluorescent protein. In some embodiments, the reporter protein is a surface marker or a detectable tag. In some such embodiments, the reporter protein (e.g., a fluorescent protein) can be used to identify or separate target-modified cells containing said construct or emulsion droplets containing such target-modified cells. The emulsion droplets can be monodisperse or polydisperse.

[0077] In some embodiments, the construct includes an FRT site or other sites for introducing a coding sequence for a target. In some embodiments, the construct includes a self-cleavage site, such as P2A or T2A. In some embodiments, the construct includes a gene (e.g., glutamine synthase (GS)) that allows selection of cells with stable integration of the construct.

[0078] In some embodiments, the target is expressed by the genome of a target-modified cell. In some embodiments, the genome has been genetically modified to include coding sequences for fusion proteins or transmembrane proteins. In some embodiments, the target is expressed by an endogenous sequence in the genome.

[0079] A target-modified cell library may contain a single cell clone presenting a common target. In some embodiments, the target-modified cell library contains more than one cell clone, each expressing a unique target. In some embodiments, the target-modified cell library contains more than one cell clone, each expressing a unique variant of a specific target. In some embodiments, the target targets expressed in the target-modified cell library are different from each other. In some embodiments, the target targets expressed in the target-modified cell library are similar to each other. In some embodiments, the target targets expressed in the target-modified cell library have similar protein sequences. In some embodiments, the target targets expressed in the target-modified cell library are targets of the same or related proteins.

[0080] In some embodiments, the target-modified cell library contains 2, 3, 4, or more cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some embodiments, the target-modified cell library contains at least 5 cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some embodiments, the target-modified cell library contains at least 10 cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some embodiments, the target-modified cell library contains at least 100 cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some embodiments, the target-modified cell library contains at least 100 to 1000 cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some embodiments, the target-modified cell library contains more than 1000 cell clones, wherein each cell clone presents a unique target different from the other cell clones. In some implementations, the library of target-modified cells contains 30 to 1000 cell clones, each of which presents a unique target that differs from the other cell clones.

[0081] The methods described herein can use a variety of target molecules. In some embodiments, targets specific to diseases (e.g., cancer or immune diseases) are used. In some embodiments, targets specific to pathogens (e.g., bacteria or viruses) are used. In some embodiments, the target is a ligand of a receptor. In some embodiments, the target is a receptor, such as a multitransmembrane protein.

[0082] In the methods described herein, various cell types can be used as target cells for modification. Cells can be prokaryotic cells (e.g., bacterial cells) or eukaryotic cells (e.g., mammalian or plant cells, fungi). Cells can be primary cells or cell lines. In some embodiments, the cells are cancer cells. In some embodiments, the cells are derived from multicellular organisms, including, for example, birds, plants, and mammals such as humans, cattle, sheep, apes, monkeys, pigs, rats, mice, dogs, and cats. In some embodiments, the cells are derived from single-celled organisms, including, for example, bacteria and yeast. Cells can be prokaryotes, such as *Escherichia coli*; or eukaryotes, such as single-celled eukaryotes (e.g., yeast or other fungi), plant cells (e.g., tobacco or tomato plant cells), animal cells (e.g., human cells, monkey cells, hamster cells, rat cells, mouse cells, or insect cells), or hybridomas. Examples of cells include CS-9 cells, the COS-7 cell line of monkey kidney cells (ATCC CRL 1651) (see Gluzman et al., 1981, Cell 23:175), L cells, C127 cells, 3T3 cells (ATCC CCL 163), Chinese hamster ovary (CHO) cells or their derivatives, such as Veggie CHO, and related cell lines grown in serum-free medium (see Rasmussen et al., 1998, Cytotechnology 28:31), HeLa cells, the BHK (ATCC CRL 10) cell line, the CV1 / EBNA cell line derived from the African green monkey kidney cell line CV1 (ATCC CCL 70) (see McMahan et al., 1991, EMBO J. 10:2821), and human embryonic kidney cells, such as 293, 293 EBNA, or MSR. 293, human epidermal A431 cells, human Colo205 cells, other transformed primate cell lines, normal diploid cells, in vitro cultured cell lines derived from primary tissues, primary explants, HL-60, U937, HaK, or Jurkat cells. In some embodiments, the cells are Expi293 cells or CHOZN cells. Typically, the cells are cultured cells that have been transformed or transfected with nucleic acids encoding a polypeptide, which is then expressed in the cells.

[0083] 6.3.2. TBP-ribosome-mRNA complex (TRM complex) The method disclosed herein uses multiple TBP-ribosome-mRNA (TRM) complexes comprising a target-binding protein. In some embodiments, each TRM complex contains a TBP containing an scFv. In some embodiments, each TRM complex contains a TBP containing a heavy chain variable region. In some embodiments, each TRM complex contains a TBP containing a light chain variable region. In some embodiments, each TRM complex contains a TBP containing a single-stranded binding fragment (e.g., VH / κ).

[0084] In some embodiments, the plurality of TRM complexes comprise unique TRM complexes containing target-binding proteins. In some embodiments, the plurality of TRM complexes comprise 1 to 5 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprise 6 to 10 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprise at least 10 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprise at least 1000 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein different from the other unique TRMs. In some embodiments, the plurality of TRM complexes comprise at least 10,000 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein that is different from the other unique TRMs.

[0085] In some embodiments, the multiple TRM complexes comprise one or more target-binding proteins with known targets. In some embodiments, the multiple TRM complexes comprise one or more target-binding proteins with unknown targets.

[0086] In some embodiments, the multiple TRM complexes contain more than one unique target-binding protein, each binding to a unique epitope on the same target.

[0087] In some embodiments, each of the TRM complexes comprises a target-binding protein and mRNA encoding the target-binding protein. In some embodiments, the mRNA comprises a coding sequence for the complementarity-determining region (CDR) of the target-binding protein. In some embodiments, the mRNA comprises a coding sequence for CDR3 of the target-binding protein. In some embodiments, the mRNA comprises the target-binding protein V H The region is encoded by a barcode sequence. In some embodiments, the mRNA contains a barcode sequence. In some embodiments, the barcode sequence identifies a target-binding protein encoded by the mRNA.

[0088] 6.3.3. Analysis Process 6.3.3.1 Contact the target-modified cells with the TBP-ribosome-mRNA (TRM) complex. The method disclosed herein involves a step of contacting target-modified cells with a TRM complex. This step is performed under conditions where the target-modified cells and the TRM complex can bind. These conditions can be adjusted or optimized to allow the target-binding protein of the TRM complex to specifically bind to its target. For example, these conditions can be adjusted or optimized by changing the buffer composition.

[0089] In some implementations, the contact step is in the presence of Mg 2+ The process is carried out in a buffer solution. In some embodiments, the buffer solution contains 25 mM to 100 mM Mg. 2+ In some implementations, the buffer solution contains 50 mM Mg. 2+ In some implementations, the buffer solution contains 25 mM Mg. 2+ In some implementations, the buffer solution contains 75 mM Mg. 2+ In some implementations, the buffer solution contains 100 mM Mg. 2+ .

[0090] In some embodiments, the contact step is performed in a buffer containing MgCl2. In some embodiments, the buffer contains 25 mM to 100 mM MgCl2. In some embodiments, the buffer contains 50 mM MgCl2. In some embodiments, the buffer contains 25 mM MgCl2. In some embodiments, the buffer contains 75 mM MgCl2. In some embodiments, the buffer contains 100 mM MgCl2.

[0091] In some embodiments, the buffer also contains HEPES. In some embodiments, the buffer also contains salt. In some embodiments, the buffer also contains NaCl. In some embodiments, the buffer also contains polysorbate 20. In some embodiments, the buffer also contains heparin. In some embodiments, the buffer also contains BSA. In some embodiments, the buffer also contains an RNase inhibitor.

[0092] In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, and BSA. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, polysorbate 20, heparin, and BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer also contains an RNase inhibitor.

[0093] 6.3.3.2 Droplet Generation In the methods provided herein, a sample containing target-modified cells and a TRM complex obtained in the contact step can be separated into microdroplets. Monodisperse emulsions can be formed on a microfluidic chip, or polydisperse emulsions can be formed using a machine (e.g., the IKAUtra-Turrax Tube Drive system). In some embodiments, a microfluidic system with three pressure pumps (e.g., the Dolomite microfluidic system) is used. Methods for generating microdroplets are known in the art, for example, those disclosed in WO2016200577A1, which are incorporated herein by reference in their entirety.

[0094] In some implementations, a microfluidic device is used to generate single-cell emulsion droplets. This microfluidic device sprays single cells from an aqueous reaction buffer into a hydrophobic oil mixture. The device can generate thousands of emulsion droplets per second. Once the emulsion droplets have formed, the device sprays the emulsion mixture into a tank. The mixture can then be pipetted to or collected in a standard reaction tube for subsequent processing, such as thermal cycling.

[0095] Custom microfluidic devices for single-cell analysis are typically fabricated in academic and commercial laboratories (Kintses et al.(2010 Current Opinion in Chemical Biology 14:548-555). For example, the chip can be made of polydimethylsiloxane (PDMS), plastic, glass, or quartz. In some embodiments, fluid is moved within the chip by pressure or the action of a syringe pump. Even single cells can be manipulated on programmable microfluidic chips using custom-designed dielectrophoresis devices (Hunt). et al. (2008 Lab Chip 8:81-87). In one implementation, a pressure-based PDMS chip is used, which consists of a flow-focusing geometry fabricated using soft lithography (e.g., Dolomite Microfluidics (Royston, UK)). (Anna) et al. (Johnston, 2003 Applied Physics Letters 82:364-366). The reservoir design typically produces 10,000 monodisperse water-in-oil droplets per second, with diameters ranging from 10 to 150 μm. In some embodiments, the hydrophobic phase consists of a fluorinated oil containing a carboxyl-containing perfluoropolyether ammonium salt, which ensures optimal conditions for molecular biology and reduces the likelihood of droplet coalescence (Johnston). et al. (1996 Science 271:624-626). To measure the periodicity of cell and droplet flow, images were recorded at 50,000 frames per second using standard techniques such as a Phantom V7 camera or Fastec InLine (Abate). et al. , 2009 Lab Chip 9:2628-31).

[0096] Microfluidic systems can optimize microdroplet size, input cell density, chip design, and cell loading parameters to ensure that over 98% of droplets contain single cells. Three common methods for achieving this statistical outcome are: (i) extremely diluted cell solutions; (ii) fluorescent screening of droplets containing single cells; and (iii) extremely diluted cells to control multiple hit rates, and fluorescent cell sorting to reduce negative rates. In some embodiments, the target-modified cells contain the coding sequence of a reporter protein (e.g., a fluorescent protein). In some such embodiments, the reporter protein can be used to detect and identify droplets containing target-modified cells. In a specific embodiment, the fluorescent protein is used to separate droplets containing a single target-modified cell.

[0097] In some implementations, the input cell flow is synchronized with the droplet formation cycle, such that over 98% of the droplets contain single cells (Edd). et al. , 2008 Lab Chip 8: 1262-1264; Abate et al.(Lab Chip 9:2628-31, 2009). In these microfluidic devices, high-density cell suspensions are forced through high aspect ratio channels, causing the cell diameter to occupy a large portion of the channel width. Various input channel widths and flow rates can be tested to find the optimal solution.

[0098] In some embodiments of the invention, the microfluidic chip is used to isolate 10, 100, 1000, 10,000, 100,000, 1 million, or 1 billion single cells from a heterogeneous pool of target-modified cells. In some embodiments, the method of the invention uses single cells in a reaction vessel instead of emulsion droplets. Examples of such reaction vessels include 96-well plates, 0.2 mL tubes, 0.5 mL tubes, 1.5 mL tubes, 384-well plates, 1536-well plates, etc.

[0099] In some embodiments, the isolated target-modified cells in the droplets bind to the TRM complex. In other embodiments, the isolated target-modified cells in the droplets do not bind to the TRM complex.

[0100] In some embodiments, the isolated target-modified cells are encapsulated with a lysis reagent. In some embodiments, the isolated target-modified cells are placed under conditions where the lysis reagent can induce lysis of the target-modified cells.

[0101] 6.3.3.3 Capturing RNA released by single cells The method described herein also includes the step of capturing RNA released from target-modified cells. In some such embodiments, beads containing polynucleotide probes targeting RNA molecules can be used to capture RNA molecules. In one embodiment, the beads are spherical beads comprising agarose, glass, chemical polymers, or magnetic materials. In some embodiments, the beads are made of materials such as latex, glass, or silica, and have a size ranging from 0.1 micrometers to 1 mm. In some embodiments, the diameter of the beads is greater than 10 μm. In some embodiments, the diameter of the beads is greater than 100 μm. In some embodiments, the diameter of the beads is less than 1 mm. In some embodiments, the diameter of the beads is 0.5-10 μm, less than 1 μm, or about 1 μm. In one embodiment, the probe comprises biotin, and the bead comprises streptavidin attached to the surface of the bead. In some embodiments, the beads are solid beads or porous beads.

[0102] In some embodiments of the invention, target-modified cells are encapsulated in droplets containing beads, the beads comprising bound polynucleotide probes whose sequences are complementary to the target polynucleotide target in a single cell. In some embodiments, the probes are 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130 or more nucleotides in length. The probes can be RNA, DNA, locked nucleic acid (LNA), peptide nucleic acid (PNA), glycol nucleic acid (GNA), or any nucleic acid analogue. The polynucleotide target can be DNA or RNA. The bead-containing reagent mixtures present in these embodiments are specifically designed to both lyse cells and promote polynucleotide hybridization, thereby enabling the beads to capture the target DNA or RNA target.

[0103] In some embodiments of the invention, the polynucleotide probe targets the constant region of an immunoglobulin. In some embodiments, the polynucleotide probe targets the constant region of IgK or IgG. In some embodiments, the polynucleotide probe targets the mRNA sequence in the TRM complex. In some embodiments, the polynucleotide probe comprises a sequence complementary to the mRNA sequence or a fragment thereof in the TRM complex.

[0104] In some embodiments of the present invention, the polynucleotide probe targets and modifies transcripts in cells. In some embodiments, the polynucleotide probe targets and modifies transcripts of a target protein in cells. In some embodiments, the polynucleotide probe targets transcripts of fusion proteins comprising a target protein and a transmembrane domain. In some embodiments, the polynucleotide probe targets transcripts of fusion proteins containing a transmembrane domain. In some embodiments, the polynucleotide probe comprises oligo(dT) polynucleotides capable of hybridizing to the poly(A) tail of mRNA.

[0105] In some implementations, the multiple nucleic acid probe targets a barcode sequence. In some implementations, the multiple nucleic acid probe targets both the TRM complex (e.g., mRNA within the TRM complex) and the fusion protein or transcript of the target.

[0106] In some implementations, a mixture of several polynucleotide probes is used. In some such implementations, a first set of polynucleotide probes targeting the TRM complex and a second set of polynucleotide probes targeting the fusion protein or the transcript of the target can be combined.

[0107] In some implementations, 5'-amino-modified polynucleotide probes are bound to carboxylic acid beads (Kojima) using 2-(N-morpholino)ethanesulfonic acid (MES) buffer. et al.(2005, Nucleic Acids Research 33:el50). In other embodiments, a biotinylated polynucleotide probe is bound to streptavidin-coated beads.

[0108] In some embodiments, the method of the present invention uses single cells in a reaction vessel instead of emulsion droplets. Examples of such reaction vessels include 96-well plates, 0.2 mL tubes, 0.5 mL tubes, 1.5 mL tubes, 384-well plates, 1536-well plates, etc. Various other designs of microfluidic chips can also be used to isolate single cells (Marcus). et al. , 2006, Anal Chem 78:3084-3089).

[0109] In some implementations, cells or cell subpopulations are added to a reaction vessel along with beads containing bound polynucleotide probes and lysis buffer. The lysis buffer lyses the cells to allow the polynucleotide probes to bind to the target polynucleotide from one or more cells. Beads hybridized to the polynucleotide target are isolated from the lysis buffer and added to the reaction vessel as individual beads or subpopulations of beads, and contacted with the PCR mixture to allow for the amplification and / or fusion of the polynucleotide target.

[0110] In some embodiments of the invention, a solvent (e.g., diethyl ether) is used to recover the aqueous phase of the droplet emulsion containing the beads and their binding targets. The beads are separated into emulsions using a PCR mixture, such that, on average, a single bead is separated into a single emulsion microdroplet (DeKosky). et al. (2015, Nat Med 21:86-91). Monodisperse emulsions can be formed on microfluidic chips, or polydisperse emulsions can be formed using a machine (e.g., the IKA Ultra-Turrax Tube Drive system). PCR reagents can amplify multiple target DNA or RNA targets. In some embodiments of the invention, the PCR product is a fusion protein and a ligation transcript of the target target in target-modified cells, as well as the mRNA of the TRM complex.

[0111] In some implementations, the step of capturing RNA released by the target-modified cells is omitted. In some cases, the target-modified cells are encapsulated in a semi-permeable shell, which eliminates the need for capturing RNA released by the target-modified cells in the workflow. In some such implementations, subsequent steps to generate a library of hybridized polynucleotides are performed in microdroplets containing RNA released by the target-modified cells. Amplification reactions can be performed in microdroplets loaded with reagents for reverse transcription and PCR (e.g., OE-RT-PCR).

[0112] 6.3.3.4 Generation of hybrid polynucleotide libraries The method described herein also includes the step of generating a library of hybrid polynucleotides containing sequences of transcripts from the single cell and / or sequences of mRNA from the TRM complex. In some aspects, the hybrid polynucleotide contains sequences of transcripts from the single cell and sequences of mRNA from the TRM complex. In other aspects, the hybrid polynucleotide contains sequences of transcripts from the single cell or sequences of mRNA from the TRM complex. The sequences of transcripts from the single cell can recognize targets expressed by the single cell. The sequences from the mRNA, along with the TRM complex that can bind from the single cell, can therefore be used to recognize the TRM complex. In some embodiments, the hybrid polynucleotides described herein also contain sequences associated with the single cell. For example, in some embodiments, the hybrid polynucleotides also contain barcode sequences from oligonucleotides attached to particles (e.g., beads) associated with the single cell. In some such embodiments, the hybrid polynucleotide comprises: a) a sequence from a transcript that recognizes a target expressed by the single cell, and an oligonucleotide sequence (e.g., a barcode sequence) immobilized on a particle (e.g., a bead) associated with the single cell; or b) a sequence of mRNA of the TRM complex bound to the cell, and an oligonucleotide sequence (e.g., a barcode sequence) immobilized on a particle (e.g., a bead) associated with the single cell. In some embodiments, the hybrid polynucleotide library comprises a first set of hybrid polynucleotides and a second set of hybrid polynucleotides, wherein a) the hybrid polynucleotides of the first set of hybrid polynucleotides comprise a sequence from a transcript that recognizes a target expressed by the single cell, and an oligonucleotide sequence (e.g., a barcode sequence) immobilized on a particle (e.g., a bead) associated with the single cell; and b) the hybrid polynucleotides of the second set of hybrid polynucleotides comprise a sequence from mRNA of the TRM complex bound to the cell, and an oligonucleotide sequence (e.g., a barcode sequence) immobilized on a particle (e.g., a bead) associated with the single cell. In some such implementations, the hybrid polynucleotide library contains sequences associated with multiple cells, and for each of the multiple cells, the barcode sequence of the oligonucleotide fixed on the particle associated with a given cell is unique compared to the barcode sequences of oligonucleotides fixed on the particle associated with other cells. This allows hybrid polynucleotides containing sequences associated with a given cell to share a common barcode that is different from the barcode shared by other hybrid polynucleotides containing sequences associated with other cells.

[0113] Hybridized polynucleotides can be generated from RNA isolated from microdroplets. Specifically, hybridized polynucleotides can be generated from RNA released from target-modified cells and a TRM complex isolated into the same microdroplet. In some such embodiments, the target protein in the target-modified cell and the target-binding protein of the TRM complex can have high affinity and specificity. In some embodiments, beads containing polynucleotide probes targeting RNA molecules are used to capture the RNA. In some embodiments, the beads contain polynucleotide probes with barcoded sequences. In some embodiments, the barcoded sequences of the polynucleotide probes are unique for each microdroplet.

[0114] In some embodiments, PCR is used to generate hybrid polynucleotides. In some embodiments, hybrid polynucleotides are generated via high-throughput amplification in a reaction vessel. As used herein, the term "reaction vessel" refers to any entity capable of physically separating the reaction into independent compartments. A reaction vessel can be a plastic compartment, a microfluidic chamber, or a droplet, such as a droplet of an aqueous reaction solution. When droplets are used, the methods described herein also include the step of generating droplets containing a reaction mixture with bead-separated RNA.

[0115] In some embodiments, the hybrid polynucleotide is generated prior to the synthesis of the first-strand cDNA. In some embodiments, oligo(dT), random primers, or combinations thereof are used to initiate the reverse transcription reaction. In some embodiments, oligo(dT) is used to initiate synthesis, preferably at the 3' end of the RNA fragment.

[0116] In some embodiments, overlap extension PCR is used to generate fusion amplicon products of the transcript and TRM complex sequence of the target-modified cell in a single reaction tube or droplet. In some embodiments, the fusion amplicon product also comprises an oligonucleotide sequence (e.g., a barcode sequence) immobilized on a particle (e.g., a bead) associated with the target-modified cell. The amplification methods and ligation amplicon are disclosed in: Johnson et al. , 2005 Genome Research 15:1315-24; U.S. Patent 7,749,697; PCT Publication No. WO2012 / 083225; and PCT Publication No. WO2013 / 096643, each of which is incorporated herein by reference in its entirety.

[0117] In some embodiments, at least two nucleic acid target sequences (e.g., a first target sequence in a target-modified cell transcript and a second target sequence in a TRM complex) are selected and designated as target sequences. Forward and reverse primers are designed for each of the two nucleic acid target sequences, and these primers are used to amplify the target sequence. "Minor" amplicons are generated by amplifying the two nucleic acid target sequences separately, and then a fusion amplicon, also known as a "major" amplicon, is generated by amplifying the fusion between multiple target sequences. In one embodiment, the "minor" amplicon is a nucleic acid sequence amplified from the first target sequence, and the "major" amplicon is a fusion complex, such as a recombinant fusion polynucleotide, generated from sequences amplified between multiple target sequences. In some embodiments, the target sequence also includes a third target sequence in an oligonucleotide immobilized on particles (e.g., beads) associated with the target-modified cell.

[0118] The method described in this paper uses "internal" primers (i.e., a reverse primer for the first target sequence and a forward primer for the second target sequence) containing a domain that hybridizes with the minor amplicon and a domain that hybridizes with the second major amplicon. The "internal" primer is a restriction agent that, during the exponential phase of PCR, is depleted, driving the annealing of overlapping domains in the minor amplicon and generating the major amplicon.

[0119] Design PCR primers against the target using standard parameters, such as a melting temperature (Tm) of approximately 55-65°C and a length of 20-50 nucleotides. Use primers under standard PCR conditions, such as 1 mM Tris-HCl pH 8.3, 5 mM potassium chloride, 0.15 mM magnesium chloride, 0.2-2 μM primers, 200 μM dNTPs, and a thermostable DNA polymerase. Many commercial kits are available for PCR, such as Platinum Taq (Life Technologies), Amplitaq Gold (Life Technologies), Titanium Taq (Clontech), husion polymerase (Finnzymes), and HotStartTaq Plus (Qiagen). Any standard thermostable DNA polymerase can be used for this step, such as Taq polymerase or Stoffel fragments.

[0120] In one embodiment, a set of nucleic acid probes (or primers) comprising a first probe, a second probe, a third probe, and a fourth probe are used to amplify a first target nucleic acid sequence and a second target nucleic acid sequence to form a fusion complex. The first probe contains a sequence complementary to the first target nucleic acid sequence (e.g., the 5' end of the first target nucleic acid sequence). The second probe contains a sequence complementary to the first target nucleic acid sequence (e.g., the 3' end of the first target nucleic acid sequence) and a second sequence complementary to the exogenous sequence. In some embodiments, the exogenous sequence is a non-human nucleic acid sequence and is not complementary to either target nucleic acid sequence. For example, the exogenous sequence may be a polynucleotide sequence encoding a serine- and glycine-rich polypeptide sequence that links the heavy and light chain variable regions in the scFv (see, for example, PCT / US1992 / 001478). The first and second probes are the forward and reverse primers of the first target nucleic acid sequence, respectively.

[0121] The third probe contains a sequence partially complementary to the second probe, which is complementary to the exogenous sequence, and a sequence complementary to the second target nucleic acid sequence (e.g., the 5' end of the second target nucleic acid sequence). The fourth probe contains a sequence complementary to the second target nucleic acid sequence (e.g., the 3' end of the second target nucleic acid sequence). The third and fourth probes are the forward and reverse primers for the second target nucleic acid sequence, respectively.

[0122] The second and third probes, also known as the "inner" primers of the reaction (i.e., the reverse primer for the first locus and the forward primer for the second locus), are at limited concentrations (e.g., the inner primer concentration is 0.01 μM, and all other primers are at 0.1 μM). This preferentially drives the amplification of the primary amplicon rather than the secondary amplicon. The first and fourth probes are called the "outer" primers.

[0123] The first and second nucleic acid sequences are amplified independently. The first nucleic acid sequence is amplified using a first probe and a second probe, and the second nucleic acid sequence is amplified using a third probe and a fourth probe. Next, complementary sequence regions of the amplified first and second nucleic acid sequences are hybridized, and the hybridized sequences are amplified using the first and fourth probes to generate fusion complexes. This is called overlap extension PCR amplification.

[0124] In overlap extension PCR amplification, complementary sequence regions of the amplified first and second nucleic acid sequences serve as primers, extending the double strand in both directions by DNA polymerase molecules. In subsequent PCR cycles, the outer primer guides the complete fusion sequence, allowing the fusion complex to replicate under the action of DNA polymerase. This method can produce multiple fusion complexes.

[0125] In some embodiments, the fusion complex is cloned into an expression vector. In some embodiments, the expression vector is a plasmid or phage particle. In one embodiment, a regulatory sequence (e.g., a promoter) is inserted into the expression vector by a method selected from Gibson assembly, site-specific digestion and ligation, and targeted recombination.

[0126] In some embodiments, the coding sequence of the target-binding protein in the fusion complex is cloned separately into an expression vector. The expression vector may contain the coding sequence of the target-binding protein operatively linked to a regulatory sequence that induces the expression of the target-binding protein.

[0127] In one aspect, this disclosure provides libraries of hybridized polynucleotides generated by the methods described herein.

[0128] In some embodiments, the methods described herein do not include the step of generating a library of hybrid polynucleotides comprising a sequence from a transcript recognizing a target expressed by the single cell and an mRNA sequence of a TRM complex binding to the cell. The methods may not generate such hybrid polynucleotides, but instead include the steps of generating a library of a first polynucleotide and a library of a second polynucleotide, the first polynucleotide comprising a sequence from a transcript of a single cell and a first barcode, and the second polynucleotide comprising a sequence from mRNA of a TRM complex and a second barcode. The first and second barcodes can be used to identify a target expressed by a single target-modified cell and the TRM complex binding to the single target-modified cell. For example, the first and second barcodes from a single droplet are identical or related. The methods may also include the step of analyzing the libraries of the first and second polynucleotides to identify the target target and target-binding proteins that can bind to the target target.

[0129] 6.3.3.5 Analysis of hybrid polynucleotides The method described herein also includes the step of analyzing the hybrid polynucleotide library. In some embodiments, the analysis involves sequencing. DNA sequencing is used to validate the construct or construct library. In some embodiments, DNA sequencing is performed using a massively parallel method from a vendor (e.g., Illumina) or using a monoclonal sequencing method (e.g., Sanger sequencing from a vendor such as Applied Biosystems). The sequence of the hybrid polynucleotide may contain the sequence of the target and / or the sequence of the TRM complex or a portion thereof. Thus, the sequence can be used to identify the target expressed on the surface of the target-modified cell and the TRM complex bound to the target. The sequence of the hybrid polynucleotide also contains the sequence of an oligonucleotide or a portion thereof immobilized on particles (e.g., beads) associated with the target-modified cell. Thus, the target and the bound TRM complex pair from a given cell can be identified.

[0130] The analysis may involve identifying the target and TRM complex isolated from the droplet. The TRM complex bound to the target may be isolated into the same droplet along with target-modified cells expressing the target. Therefore, in some embodiments, the analysis involves identifying the target and the target-binding protein that binds to it. In some cases, the analysis involves identifying the target and the target-binding protein that can bind to it. In some embodiments, the analysis may also involve identifying multiple TRM complexes bound to the target in the same droplet, for example when the droplet contains particles (e.g., beads) immobilized with oligonucleotides containing barcode sequences.

[0131] In some embodiments, the target-modified cell can bind to more than one TRM complex. In some such embodiments, multiple TRM complexes are isolated together with the target-modified cell into the same microdroplet. In some embodiments, the RNA captured from the microdroplet can contain sequences from more than one TRM complex. Therefore, the hybrid polynucleotide generated from a single microdroplet can contain one or more pairs of target targets and target-binding proteins. Therefore, in some embodiments, multiple target-binding proteins binding to the same target can be identified. In some embodiments, the hybrid polynucleotide generated from a single microdroplet can contain barcode sequences of oligonucleotides immobilized on particles (e.g., beads) within the microdroplet.

[0132] In some implementations, the methods described herein can be performed under conditions of varying stringency. Target-TRM complex pairs identified under different conditions can be analyzed to test their binding affinity.

[0133] In some embodiments, the methods described herein are used to identify a subset of target-binding proteins with high binding affinity to a target. In some embodiments, the methods are used to identify a subset of target-binding proteins with a desired binding affinity to a target.

[0134] In some implementations, the analysis involves quantitative analysis of specific pairs of the target and the target-binding protein. The quantitative data can be used to measure the affinity and specificity of the target-binding protein for the target.

[0135] In some embodiments, the analysis involves identifying all target-binding proteins that bind to a specific target. In some embodiments, the analysis involves quantifying target-binding proteins that bind to the target. Quantitative data can be used to identify target-binding proteins that have the desired affinity and / or specificity to the target. In some embodiments, quantification can be used to determine the binding affinity and / or specificity of a target-binding protein to the target by comparing data with known target-TBP pairs. In some embodiments, the process can be used to mature the antibody's affinity by testing the binding of various modifications to the target. This can be selected based on the binding activity (e.g., affinity and specificity) of the antibody with the specific modification to the target.

[0136] In some embodiments, the analysis involves identifying all target targets that bind to a specific target-binding protein. In some embodiments, the analysis involves quantifying the target targets that bind to a specific target-binding protein. In some embodiments, the methods described herein are used to characterize individual antibodies in a polyclonal mixture without the need for separation and purification. In some embodiments, the methods are used for high-throughput screening of binding partners between diverse antibody libraries and diverse target libraries (e.g., libraries of naturally occurring target variants). Quantification can be used to identify targets for target-binding proteins.

[0137] In some embodiments, the methods of this disclosure are used to analyze the binding of polyclonal antibodies to a target. In some embodiments, the methods of this disclosure are used to analyze the binding of polyclonal antibodies to a target in samples from donors to select donors. In some embodiments, the methods are used to analyze and monitor the quality or yield and consistency of polyclonal antibodies against a target. In some embodiments, the methods are used to characterize polyclonal antibodies. In some embodiments, the methods are used for QC monitoring of antibody libraries over time, assessing library diversity, epitope localization, or evaluating the antimutagenic function of antibodies.

[0138] In some embodiments, the methods described herein are used to identify epitopes of antibodies by identifying binding between an antibody and a target library, wherein the target is a variant of a target protein with one or more mutations at different sites. Epitopes can be identified by analyzing which mutations result in the loss of binding between the antibody and the target protein. In some embodiments, the methods can be used to identify epitopes of polyclonal antibodies. In some such embodiments, the diversity of polyclonal antibodies can be determined based on epitope mapping.

[0139] 6.3.4. Analysis of the interaction between ligand and receptor The high-throughput analysis method disclosed herein can be used to study the interaction between ligands and receptors. In some such embodiments, the target may be a receptor or a fragment thereof, and the ABP may be a ligand or a fragment thereof. Alternatively, the target may be a ligand or a fragment thereof, and the ABP may be a receptor or a fragment thereof. In some embodiments, the method includes the steps of: (i) providing a library of ligand-modified cells, wherein each of the ligand-modified cells presents a target ligand on a membrane; (ii) contacting the library of the ligand-modified cells with a plurality of receptor-ribosome-mRNA (RRM) complexes to induce binding between the ligand-modified cells and the RRM complexes; (iii) generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the ligand-modified cells, one or more RRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; (iv) capturing RNA released from the single cell on a solid surface or within a semi-permeable shell; and (v) generating a library of hybridized polynucleotides containing sequences of transcripts from the single cell and / or sequences of mRNA from the RRM complexes. In some embodiments, the method includes the following steps: (i) providing a library of receptor-modified cells, wherein each of the receptor-modified cells presents a target receptor on a membrane; (ii) contacting the library of receptor-modified cells with a plurality of ligand-ribosome-mRNA (LRM) complexes to induce binding between the receptor-modified cells and the LRM complexes; (iii) generating a plurality of emulsion droplets, wherein each droplet contains a single cell from the receptor-modified cells, one or more LRM complexes bound to the single cell, and a lysis reagent for inducing lysis of the single cell; (iv) capturing RNA released from the single cell on a solid surface or within a semi-permeable shell; and (v) generating a library of hybridized polynucleotides containing sequences of transcripts from the single cell and / or sequences of mRNA from the LRM complexes.

[0140] Hybridized polynucleotide libraries can be analyzed or sequenced to provide information related to the binding of a target ligand to an RRM complex or a target receptor to an LRM complex. For example, based on the sequence, the binding affinity or specificity of the receptor and / or its binding to the target ligand can be studied. Therefore, the method may also include a step of identifying ligand-receptor pairs based on sequencing of the hybridized polynucleotide library. In some embodiments, the method further includes a step of identifying ligand-binding proteins that are specific to the target ligand. In some embodiments, the method further includes a step of identifying the binding affinity or specificity of ligand-binding proteins that are specific to the target ligand.

[0141] 6.4. Kits for high-throughput antibody analysis This disclosure also provides a kit for high-throughput analysis of antibodies.

[0142] The kit may contain multiple polynucleotide constructs, each encoding a fusion protein containing a unique target and a transmembrane domain. In some embodiments, the kit contains multiple polynucleotide constructs, each containing a coding sequence for a transmembrane domain and an insertion site for adding a coding sequence for the target.

[0143] In some embodiments, the kit further comprises multiple TBP-ribosome-mRNA (TRM) complexes, wherein each TRM complex contains a unique target-binding protein. In some embodiments, the multiple TRM complexes contain 1, 2, 3, 4, 5, or more unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 5 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 10 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 50 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 1000 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 500 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 1000 unique target-binding proteins. In some embodiments, the multiple TRM complexes contain more than 10,000 unique target-binding proteins.

[0144] In some embodiments, the kit further includes reagents for PCR. In some embodiments, the kit further includes reagents for overlap extension polymerase chain reaction (OE-PCR). In some embodiments, the kit further includes reagents for first-strand cDNA synthesis. In some embodiments, the kit further includes polymerase.

[0145] In some embodiments, the kit also includes a buffer that can be used in the step of contacting the target-modified cells with the TRM complex. In some embodiments, the buffer contains Mg 2+ In some implementations, the buffer solution contains 25 mM to 100 mM Mg. 2+ In some implementations, the buffer solution contains 50 mM Mg. 2+ In some implementations, the buffer solution contains 25 mM Mg. 2+ In some implementations, the buffer solution contains 75 mM Mg. 2+ In some implementations, the buffer solution contains 100 mM Mg. 2+ .

[0146] In some embodiments, the buffer solution contains MgCl2. In some embodiments, the buffer solution contains 25 mM to 100 mM MgCl2. In some embodiments, the buffer solution contains 50 mM MgCl2. In some embodiments, the buffer solution contains 25 mM MgCl2. In some embodiments, the buffer solution contains 75 mM MgCl2. In some embodiments, the buffer solution contains 100 mM MgCl2.

[0147] In some embodiments, the buffer also contains HEPES. In some embodiments, the buffer also contains salt. In some embodiments, the buffer also contains NaCl. In some embodiments, the buffer also contains polysorbate 20. In some embodiments, the buffer also contains heparin. In some embodiments, the buffer also contains BSA.

[0148] In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, and BSA. In some embodiments, the buffer contains HEPES, NaCl, 50 mM MgCl2, polysorbate 20, heparin, and BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer contains 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer also contains an RNase inhibitor.

[0149] 7. Example 7.1. Example 1: Development of target-modified cell lines and constructs for target display Six different constructs (V1-V6) were tested to generate target-modified cells. These constructs had different leader sequences (mouse IgM, native CoV1-S or CoV2-S or tPA) and transmembrane domains (cmyc-PDGFR, native CoV1-S or CoV2-S, or native sequences with 19 or 37 amino acid C-terminal truncations), as shown in Table 1.

[0150] The construct was transiently transfected into Expi293 cells, and transgene expression was assessed by flow cytometry 24 hours later using an anti-spike protein mAb. Figure 2This study shows that constructs with a natural leader sequence, especially those with a natural leader sequence and a 19-amino acid truncated transmembrane domain (natural ∆19), perform very well.

[0151] Further modifications were made to generate three stable cell lines expressing the CoV2-S antigen. First, the V1 construct was packaged into lentivirus and used in Expi293 cells to generate stable cell lines. Cells transfected with V1-lentivirus (Expi293-CoV2-S) were isolated on FACS by detection of positive c-myc signal. The second and third constructs were modified to make them compatible with CHOZN cells. Specifically, the second construct was modified to include components of V5, a native leader sequence, and an FRT site containing the P2A and glutamine synthase (GS) genes, which allowed for screening of stable integrons in glutamine-deficient media. This construct also utilized the 2G-UNIC translation enhancement element and the CMV promoter (CHOZN-CoV2-S). The third construct was modified to contain the same components as the second construct, along with six stable proline mutations (Hsieh et al, Science 369, 2020) and a furin protease site mutation (Peacock et al, Nature Microbiology 6, 2021) to prevent lysis (CHOZN-CoV2-S(HP+F)). The second and third constructs were transfected into CHOZN cells, resulting in stable cell lines containing a single landing site at the locus and exhibiting high and consistent gene expression. In this step, a single copy of the construct with the FRT site was inserted into the landing site using a recombinase.

[0152] All three cell lines and their corresponding negative controls were stained with different concentrations of bamranizumab (Jones et al., BioRxiv 9OCT20), a human IgG1 with high affinity for CoV2-S. Figure 3 As shown, the surface expression of CoV2-S in CHOZN cells (CHOZN+CoV2-S and CHOZN+CoV2-S(HP+F)) was significantly better than that in Expi293 cells (Expi293+CoV2-S). Furthermore, stable mutations (HexaPro, HP) and furin mutations (F) enhanced surface display.

[0153] Based on the research, a construct scaffold (p2G-FRT-GS) was selected to generate target-modified cells. For example... Figure 4As shown, the construct backbone comprises: (i) a CMV promoter with a 2G translation enhancement element; (ii) a blue fluorescent protein filling sequence (for identifying library background during cloning); (iii) a unique 20-nucleotide barcode (e.g., HHSWNNHHCTGGNNHHSWHH (SEQ ID NO:114), NNNWSHHHNNHHHNNWSNNN (SEQ ID NO:115) or a similar sequence); (iv) an FRT site with 2A-GS (for screening integrons entering landing site cell lines); and (v) an ampicillin-resistant and bacterial source that can be grown in E. coli.

[0154] Target libraries can be introduced into the construct backbone in several ways. For small libraries, each insert can be synthesized with a unique barcode and cloned via Gibson assembly or insertion into the HindIII / NheI site. For larger libraries, a backbone library with unique barcodes must first be generated by inserting the barcode library (which may contain gene blocks or oligonucleotides containing degenerate nucleotides) into the MluI / NheI site. For scan mutagenesis, a library of single-point mutants can be synthesized and then cloned into the library barcode backbone via HindIII / MluI site or Gibson assembly. For other mutagenesis methods (e.g., random mutagenesis), a library of inserts can be generated using primers with degenerate nucleotides, such as error-prone PCR or OE-PCR, and then cloned into the barcode backbone using HindIII / MluI. Note that the barcode library needs to be at least 10 times larger than the insert library to ensure that each insert has a unique barcode.

[0155] 7.2. Example 2: Target-modified cell lines expressing CoV2 antigen All available CoV2 sequences as of May 2022 were downloaded from GISAID and analyzed. The analysis revealed 25 unique clades with 20 unique spike protein sequences. Further refinement of the list, removing sequences differing by only a single amino acid, resulted in 17 spike protein sequences. These 17 sequences, along with the CoV1 spike protein, were synthesized and cloned into the 2G-FRT-GS backbone, each possessing a unique barcode. The spike protein sequences contain mutations in the hexapeptide and furin protease to improve surface display, but do not include mutations in the leader sequence.

[0156] DNA from all library members was mixed in equimolar proportions and transfected into the platform CHOZN cell line, followed by selection over several weeks in glutamine-deficient medium. Alternatively, clones were transfected individually, selected, and then merged. RNA was isolated from the clones and sequenced to confirm library distribution.

[0157] 7.3. Example 3: TBP-ribosome-mRNA complex (TRM complex) Five mAbs were cloned into a T7-based expression plasmid, the structure of which is as follows: Figure 5 As shown. This T7-based expression plasmid contains: (i) a T7 promoter and ribosome binding site; (ii) an ampicillin resistance and origin of replication for proliferation in *E. coli*; (iii) a start codon followed by a light chain variable region (without a signal peptide); (iv) a flexible linker, (G4S)4 or similar; (v) a heavy chain variable region; (vi) a strep-tag; (vii) unstructured spacer sequences, such as unstructured spacer sequences derived from *E. coli* tolA protein or other proteins, of varying lengths; (viii) a translation pause sequence derived from *E. coli* secM protein; (ix) a deletion of a stop codon; (x) a polyA region encoding 40x A nucleotides internally; and (xi) a T7 transcription terminator, which may contain multiple tandem sequences or mutants.

[0158] Transcription or coupled transcription-translation was performed using intact plasmids, linearized plasmids, or PCR fragments of plasmids (from promoter to terminator) as templates. Reactions were performed using the NEB PURExpress kit (with or without release factor), the NEB T7 Rapid High-Yield RNA Kit, and the Promega TnT Rabbit Reticulocyte Kit.

[0159] 7.4. Example 4: Binding between the target and the TBP-ribosome-mRNA complex (TRM complex) Cells expressing spike protein variants (CoV2-WT, CoV2-K444T, CoV2-E484K, CoV2-F486K, and CoV1-WT) were stained with TRM complexes at different dilutions, each TRM complex containing a scFv of one of four mAbs (bamranibumab, camrevimab, edevimab, or ipilimumab). TRM binding to the CoV2-S variant was detected by Strep tag staining, and the results are as follows: Figure 6 As shown.

[0160] In another set of experiments, cells expressing spike protein variants were stained with different dilutions of the TRM complex, and RNA was isolated from the stained cells and quantified using Taqman RT-qPCR with antibody (TBP)-specific probes. Results are as follows: Figure 7 As shown. RNA was recovered from the stained cells as expected, and binding between TRM and CoV2-S was observed. CoV variants previously known not to bind to a given antibody showed no binding, as shown. Figure 6 As shown (red line).

[0161] In subsequent experiments, TRM complexes generated under different cell-free translation (TL) conditions were tested. Specifically, TRM complexes were generated using different concentrations of RNA template at different TL reaction times. Figure 8 Data showed that, within a shorter reaction time (0.5 hours), the TRM complex generated using a 150 nM RNA template significantly reduced nonspecific background, providing the largest ratio between RNA detected by the TRM complex containing camrevimumab scFv (i.e., nonspecific staining) and RNA detected by the TRM complex containing ipilimumab scFv (i.e., nonspecific staining).

[0162] 7.5. Example 5: Binding between the target and the merged TBP-ribosome-mRNA complex (TRM complex) Cell lines expressing the CoV2-S antigen were stained with a mixture of TRM complexes. The TRM complex mixture was generated from a library of RNA templates (150 nM template) via in vitro translation (reaction time 0.5 h). RNA was isolated from cells stained with the TRM complex mixture and quantified using Taqman RT-qPCR with antibody-specific probes (TBP). Figure 9 The results provided show specific binding between the target and the TRM complex pair, low background staining, and no signal for variants known not to bind to a given antibody (represented by red text). Specifically, bamranib (“Bam”) deletion is known to bind to the E484K mutant, camrevimib (“Casi”) deletion to the F486K mutant, and edevimab (“Imdevi”) deletion to the K444T mutant. Ipilimumab (“Ipi”) is not expected to bind to any spike protein variant (it binds to CTLA4), and all antibodies are not expected to bind to the CoV1 spike protein. These known binding specificities have been confirmed by the target and TRM complex pair, as shown in [the original text]. Figure 9 As shown.

[0163] 7.6. Example 6: Single-cell isolation and RNA amplification Five target-modified cell lines were pooled and stained with a mixture of five TRM complexes. Single cells were sorted and isolated into 96-well plates. Samples in each well were amplified and barcoded for sequencing. The first RT-PCR reaction added the ligand to the mRNA transcript and amplified the Vt of the antibody mRNA in the TRM complex. HThe regions and cell lines span the target barcode area. A second PCR reaction adds amplicons to the index barcode, which is unique for each column and row in the 96-well plate. After pooling the samples from all wells, sequencing is performed using mi-seq.

[0164] The binding distribution of single-cell data was analyzed. First, the target cells in each well were identified. Empty wells and wells containing more than one cell were excluded by applying a threshold. For each well, the proportion of readouts corresponding to each antibody was calculated. The polymap score was calculated based on the average proportion of target cells and narrowed down by the average readouts of that target cell. Polymap scores for CoV1 or CoV2 spike protein WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) were determined and fed to a TRM complex containing one of four mAbs (bamranib, carprednisone, edevimab, ipilimumab, or pembrolizumab). Polymap scores were significantly lower on cell lines with targets that were not expected to bind to a given TRM complex (shown as red outlines). The proportion of negative antibodies (ipilimumab or pembrolizumab, indicated in red) was low on all targets.

[0165] 7.7. Example 7: Single-cell analysis using a microfluidic system Five target-modified cell lines were pooled and stained with a mixture of five TRM complexes. After staining with the TRM complexes, single cells were encapsulated in water-in-oil droplets for antibody library analysis. The beads were released from the emulsion and washed to remove unbound RNA, and then the RNA-bound beads were isolated from the single cells. Since each cell expresses a single target (i.e., a SARS-CoV-2 spike protein variant), it was expected that each bead would bind to the RNA of that target as well as the RNA of each TRM complex bound to that specific cell.

[0166] Individual RNA-binding beads were then encapsulated into a second emulsion for antibody library analysis. These second emulsions also contained PCR buffer, reagents, and primers for overlap extension reverse transcription PCR (OE-RT-PCR). In the reaction mixture, RNA was converted to cDNA, which was then used to barcode the target and antibody V. H CDR3 is amplified using primers that are physically linked together. The reverse primer for the antibody fragment contains a sequence complementary to the forward primer sequence of the target fragment. In subsequent rounds of PCR reactions, these complementary fragments anneal, and the resulting product contains both the antibody and target sequences. Figure 11 This demonstrates overlapping extension reverse transcription PCR and simultaneous inclusion of antibody (V) H The generation of products of target (barcode) specific sequences.

[0167] Following OE-RT-PCR, primers containing Illumina linkers and an index for the standard Illumina sequencing process were used to further amplify the fused TBP target sequence. In the experiment, the following CDR3 and barcode sequences linked to the CoV2-S antigen were successfully amplified and sequenced.

[0168] Table 2

[0169] 7.8. Example 8: High-throughput specific analysis of antibody clones using ribosome display and microfluidics. High-throughput methods for mapping pairwise interactions of protein-protein interactions allow for deeper understanding of biological interactions, such as natural antibody responses. However, existing techniques are limited by the number of interactions that can be screened at a time. To overcome this limitation, a polyclonal mapping system (also referred to herein as “PolyMap”) has been developed. This system is based on the batch binding of ribosome-display antibody libraries to libraries expressing antigens on cell surfaces, combined with single-cell analysis using droplet microfluidics. In this embodiment, mapping of thousands of antigen-antibody interactions was confirmed for a library of clinically relevant SARS-CoV-2 surface antigens and a variety of naturally occurring anti-SARS-CoV-2 antibodies, and antibodies with selective binding to antigens across a wide evolutionary clade were identified. Although developed using antibody-antigen libraries, PolyMap is well-suited for screening other types of protein-protein interactions involving soluble ligands and their membrane-bound targets, such as cytokines, immunomodulatory factors, and their receptors.

[0170] PolyMap workflow design The PolyMap platform allows for one-pot interaction screening of antibodies and antigen libraries. Figure 12 Antibodies are expressed in a ribosome-displayed form, utilizing bound mRNA to provide genotype-phenotype linkages for soluble proteins. Antigens are highly expressed on the surface of mammalian cells and can then bind to soluble ARM complexes. Ribosome-displayed antibody libraries are incubated with libraries from antigen-expressing cells, allowing sampling of all possible antibody-antigen combinations in a single batch step. To elucidate antibody-antigen interactions, cells stained with ARM complexes are washed and encapsulated individually in microdroplets containing lysis reagents and RNA-capturing beads with unique barcodes. The isolated beads undergo a series of molecular biology steps to generate cDNAs with barcoded antibody and antigen sequences linked together. The barcoded transcripts are then read out and quantified by deep sequencing, and bioinformatics analyses are used to map antibody-antigen pairwise binding specificity.

[0171] Establishing an antigen expression system This section addresses the development of a robust antigen surface display platform. Specifically, the platform should possess robust surface expression capability for a single antigen / cell, a unique barcode for antigen recognition, and a simple and efficient cloning and cell line generation process suitable for use with libraries. A range of mammalian expression vectors were evaluated to determine the impact of different signal peptides and transmembrane regions on the surface expression of a selected target SARS-CoV-2 surface antigen (CoV-2 S; data not shown). Native signal peptides and transmembrane regions were retained because they performed at least as well as their substitutes. A 19-amino acid C-terminal truncation was implemented to remove the endogenous endoplasmic reticulum retention signal of CoV-2 S.

[0172] Stable cell lines expressing CoV-2 S were generated using two different methods: (1) lentiviral transduction into Expi293™ cells; and (2) Flp recombinase-mediated integration into CHOZN® cells, which were engineered with Flp recognition target (FRT) landing sites for unit site integration. After sorting or screening the cell lines, staining for CoV-2 S surface expression revealed significantly higher relative signal intensity in the CHOZN cell lines. Figure 13 Furthermore, a modified construct was evaluated and found to further enhance the cell surface expression of the antigen. Figure 13 The construct contains six stable proline mutations and one furin protease site mutation to eliminate the fusion activity of the spike protein.

[0173] The final antigen display vector uses a CMV promoter with translation-enhancing sequence elements ((2G UNic®, ProteoNic)) driving efficient antigen expression, a signal peptide, a transmembrane domain for surface display, the aforementioned proline modifications, and a mutated furin protease site. Figure 14A The translation stop codon is followed by a 20-mer nucleic acid barcode (unique to each antigen), followed by a common flanking sequence to allow PCR amplification and antigen recognition via short-read sequencing. The vector also contains restriction enzyme sites for vector linearization and convenient ligation of the antigen library.

[0174] To generate a library for antigen-displaying cells, an antigen-encoding plasmid library, along with a plasmid expressing Flp recombinase, was transfected into a pre-engineered CHOZN cell line whose FRT sites were integrated into highly expressed genomic sites. This ensured single-copy integration and normalized expression levels for each antigen. Successful plasmid integration led to the expression of glutamine synthase, allowing for screening in glutamine-free medium.

[0175] For proof-of-concept studies, ancestral (referred to as wild-type in this study; WT) CoV-2 spike protein and three individual point mutants (K444T, E484K, F486K) were selected, known to significantly interfere with the binding of specific monoclonal antibodies in clinical development. The four spike protein coding sequences (including the native leader sequence, stable mutation, native transmembrane region, and deleted cytoplasmic tail) were cloned into a platform vector with unique barcodes. In this embodiment, the plasmids were transfected into platform CHOZN cells, and these cells were then screened and stained with cross-reactive antibodies to detect spike protein surface expression. Flow cytometry analysis showed high and consistent expression among the four CoV-2-S variants. Figure 14B ).

[0176] Establish a ribosome display system The ribosome display construct includes previously reported basic features, as well as some recent enhancements and features specific to this application. Figure 14C The T7 promoter drives the expression of the scFv gene, followed by the Strep tag II sequence. The absence of a stop codon and the addition of a 17-amino acid SecM stall sequence prevent ribosome separation from the mRNA, physically linking the newly synthesized protein to its coding sequence. The encoded poly-A tail allows for subsequent capture onto magnetic beads for cDNA synthesis, barcoding, and sequencing. The CDR3H sequence provides a unique identifier for the antibody and, like the antigen plasmid, contains commonly used restriction enzyme sites and flanking regions for library cloning, amplification, and sequencing. The entire expression cassette is preserved in a pUC-based plasmid.

[0177] To generate the ARM complex, a DNA fragment from the T7 promoter to the transcription terminator region needs to be amplified by PCR and used as a template to generate a large amount of RNA. While circular plasmids can be used as transcription templates, it was found that PCR fragments with well-defined ends are crucial for RNA consistency; otherwise, severe terminator readthrough occurs even when using constructs with improved terminator designs (data not shown). It has been determined that using a recombinant-based in vitro expression system and starting with a fixed purified RNA input produces more consistent ARM complex yields compared to coupled transcription-translation processes, likely due to differences in transcriptional kinetics between variants. Furthermore, it was determined that omitting the ribosome release factor and using a short translation time of 10 minutes in the in vitro translation reaction further improves the stability and specificity of the ARM complex. Figure 15 ).

[0178] For preliminary testing, clinical SARS-CoV-2 monoclonal antibodies were selected: bamranizumab, camrevimab, and edevimab, known to be sensitive to the spike protein point mutations E484K, F486K, and K444T, respectively. Biological layer interferometry (BLI) was used to confirm the binding affinity (and loss of binding to specific point mutations) of these antibodies (presented as scFv) to each spike protein mutant expressed as a soluble receptor-binding domain (RBD)-Fc fusion variant (Table 3). Figure 16 The scFv was then cloned into a ribosome display construct to prepare an ARM complex, which was then incubated with cells expressing the antigen. Total RNA was isolated from the stained cells and analyzed by RT-qPCR to quantify the recovery rate of antibody mRNA. For each antibody, complete or partial RNA loss was observed in the expected spike protein variant, consistent with the measured binding affinity. Figure 14D Although there are some differences in RNA recovery rates among different antibodies, which may be due to differences in translation efficiency, the concentrations reported in this paper represent RNA recovery rates of 1-4%, which is consistent with previous reports.

[0179] Table 3

[0180] The binding affinity of clinical antibodies to soluble spike protein mutants was measured using BLI. The dissociation rate is below the detection limit.

[0181] Proof-of-concept test using clinical-stage anti-SARS-CoV-2 antibodies Preliminary testing was performed using a “mini-library” of three clinical-stage anti-SARS-CoV-2 antibodies and two negative control antibodies (anti-CTLA-4 ipilimumab and anti-PD-1 pembrolizumab). In the simplest experiment, ARM complexes were generated using an equimolar mixture of RNA, then diluted to a concentration equivalent to 15 nM RNA and incubated separately with each antigen-expressing cell line. Total RNA from stained cells was amplified and sequenced, and each antibody was identified using the CDR3H sequence. The percentage of total readouts assigned to each antibody was tabulated for each antigen-expressing cell line. As expected, the anticipated dropouts were very small, and the readouts for both negative control antibodies were low in all samples. Figure 17A Surprisingly, a large proportion of the readouts came from camrelizumab, although this is consistent with the high RNA recovery rate in the initial RT-qPCR results for this antibody. Figure 14D ).

[0182] The experiment was repeated by mixing four antigen cell lines together and performing batch staining with a miniature antibody library. To pair antigen and antibody data, single cells were sorted into 96-well plates and lysed. Antigen barcodes and antibody heavy chains were amplified using an array of primers containing column and row barcodes to allow for unique identification of the amplicon in each well. After identifying the antigen for each cell, the distribution of antibodies binding to those cells was calculated. As observed previously with individual cell lines, the anticipated strong CoV-2 mutant loss, low background binding, and reproducibility between single cells were observed. Figure 17B , 17C ).

[0183] While single-cell sorting and array PCR are suitable for small sample sets, single-cell droplet-based methods combined with high-throughput sequencing will allow for the analysis of larger groups of antibodies and / or antigens. For a proof-of-concept test, the previously disclosed microfluidic droplet technology Drop-seq was employed, where single cells are co-encapsulated with barcode beads (ChemGenes Corporation) in nanoliter droplets using a microfluidic chip and infusion pump device. The same antibody-RNA mixture and four cell lines were used, but the staining strategy was slightly modified, diluting the ARM complex to a concentration equivalent to 4.5 nM RNA (this concentration was not expected to affect the results, as binding assays of the ARM complex with CoV-2WT at two different concentrations showed similar enrichment levels for many antibody clones). Figure 18 The Drop-seq workflow was performed via cDNA amplification as previously described (https: / / mccarrolllab.org / dropseq / ), and antigen barcodes and antibody heavy chain sequences were amplified using gene-specific primers with Illumina linkers for sequencing. Individual cells were first identified via Drop-Seq barcoding, and then antigen identity was determined via antigen barcoding; a total of 91 cells were identified for mapping antibody-antigen interactions. Figure 17D , 17E Drop-seq results compared to individual cell line staining ( Figure 17A ) and single-cell sorting ( Figure 17B , 17C The results are consistent.

[0184] Testing anti-SARS-CoV-2 immune repertoire against a larger set of spike protein variants PolyMap was then used for larger libraries of antibodies and antigens. All publicly available CoV-2 S sequences (as of August 2023) were downloaded from GISAID.org, from which 16 unique sequences were selected to represent different phylogenetic clades (nextstrain.org). Figure 19(SEQ ID NO: 126-138, 140, 142, and 143). These sequences and control antigens (CoV-1S, CTLA-4, PD-1, and blue fluorescent protein [BFP]) were cloned into an antigen backbone using unique barcodes and then transfected into CHOZN landing site cell lines, respectively. Next, libraries of various naturally paired scFvs were cloned into ribosome display vectors. The libraries were previously isolated from recovered COVID-19 donors and enriched with antibodies that bind to the ancestral spike protein RBD. Sequencing of the libraries showed a wide variant distribution, with 99% of the sequences derived from 300 unique clones, of which the top clone accounted for ~6.9% of the library ( SEQ ID NO: 126-138, 140, 142, and 143). Figure 20A ).

[0185] For drop-seq microfluidic analysis of larger libraries, cells expressing each CoV-2 S sequence variant were pooled together and then stained with the ARM complex of the anti-CoV-2 antibody library described above. After identifying antigen and antibody sequences using drop-seq cell barcoding, a pooled list of each antibody and antigen associated with each cell barcode was generated, ultimately analyzing 2,783 cells for 20 antigens. Figure 21 Since each antibody has an equal chance of binding to each antigen, the total antibody readouts for each clone across all cell lines were counted, and the readout percentage for each antibody was then normalized to the relevant cell number. When the top 40 antibodies were examined based on the total antibody readouts, differences in binding patterns were observed between different antibody sequences. Figure 20B Many antibody clones showed extremely low enrichment for the Omicron spike mutants (21K, 22A / D, 23B), which was expected since the library was generated from a donor recovered from the original ancestor's WT strain in early 2020. Reduced binding to 20H(β), 20J(γ), CoV-1, and negative control cell lines (PD-1, CTLA-4, and BFP) was also observed.

[0186] In this study, when constructing the CoV-2 antigen cell line, the antigen sequences were transfected into the cells individually and then combined for PolyMap operation. Alternatively, multiple antigen sequences can be transfected into the cells in batches. To test the feasibility of the method, 22 antigen sequences were transfected together, including 18 unique CoV-2S sequences representing different phylogenetic clades (16 sequences previously used, such as...). Figure 19As shown, 21M Omicron (SEQ ID NO: 139) and 22COmicron (SEQ ID NO: 141) and four control antigens (CoV-1S, CTLA-4, PD-1, and BFP) were cloned into the antigen backbone along with their unique barcodes and then batch-transfected into CHOZN® landing site cell lines. Genomic DNA (gDNA) sequencing of the batch-transfected cell lines showed that all variants were expressed with minimal frequency variation. Figure 24 This batch transfection method can efficiently generate and test large antigen libraries.

[0187] PolyMap Result Verification After obtaining single-cell lines corresponding to each antigen, each cell line was stained with an ARM complex library. RNA was isolated and sequenced as before, and the readout ratio for each antibody was calculated. Because the antibody library is unevenly distributed, it is important to focus on the enrichment level compared to the input sample, rather than the original antibody readout ratio. The log2 fold change of this enrichment value was used to better distinguish between enriched and de-enriched sequences, showing the top 40 clones with the highest enrichment levels. Figure 22A Similarly, many clones were observed to be de-enriched by binding to 20H(β), 20J(γ), and Omicron variants.

[0188] Eleven antibody clones (designated M1-M11) found in the first 100 sequences and exhibiting different antigen-binding modes were selected for validation. Some of these clones had previously been isolated into individual CHOZN cell lines and produced on a small scale. For the remaining antibodies, a second sequencing run on the input library was used to determine the Vo. L Sequence and CDR3H, then combined with the previous V H Data were paired and assembled into a full-length sequence for expression. These variable regions were cloned into separate expression vectors and transiently expressed in HEK293 cells.

[0189] The binding of these antibodies was validated by staining each spike mutant cell line with antibody-containing supernatant diluted to a concentration expected within the linear sensitivity range. Antibody binding was detected by secondary antibodies specific to human Fc and then analyzed by flow cytometry. The median fluorescence intensity (MFI) of the data was compared with that of Drop-seq. Figure 20B ) and single antigen staining ( Figure 22A The data was compared with PolyMap data. Overall, the unique binding patterns of each clone in Drop-seq were compared with those in single-antigen and flow cytometry data. Figure 22BThe antibodies are closely matched. It is noteworthy that, since single-antigen staining is normalized within a single cell line, enrichment values ​​can provide some insight into the relative binding affinity between antibodies. Figure 23 For example, clone M1 showed good alignment across all datasets. Figure 22B However, due to its low enrichment in the competitive binding scenario of WT antigens, its affinity may be low. This may also explain some false positive results of single antigen staining for Omicron variants, as there are fewer competing sequences. Clone M10 is likely a sequence with very high translational or PCR efficiency, leading to strong enrichment using single antigen staining methods.

[0190] Finally, the sequence diversity and binding modes of the recovered antibodies were further explored. As expected, few clones were found in these early patient samples that were able to maintain binding to the highly mutated Omicron spike protein. Several repetitive loss patterns were observed, which were associated with common mutations in the RBD. For example, the K417 mutation was found in antibodies against β, γ, and Omicron(21K) and beyond. Figure 19 The binding sites are likely epitopes of M1, M2, M3, M6, M7, M9, and M10. Variants of β, γ, κ, Eta, Iota, and Mu, and later, contain the E484 mutation, which may be an epitope of clones such as M5. Binding of clones such as M8 and M11 is limited to the earliest clones and may recognize epitopes containing the aforementioned amino acids as well as other amino acids (e.g., L452 or N501). Finally, clone M4 shows broad binding and some reactivity to Omicron and may bind to more conserved epitopes.

[0191] discuss To seek a method for pairwise interaction mapping of antigen and antibody libraries that is not limited by existing technologies, PolyMap, a platform for probing mammalian cell surface protein libraries using soluble antibody libraries, was developed. The method was developed and validated using a small matrix of clinical SARS-CoV-2 antibodies with known spike protein mutation susceptibility. Subsequently, PolyMap was applied to an uncharacterized immune library from convalescent COVID patients, revealing antibodies with multiple different binding modalities in popular SARS-CoV-2 spike protein variants.

[0192] A significant advantage of the PolyMap platform is its flexible antigen expression system tailored for membrane protein expression. Unlike other published multiplexing methods that require purification, in vitro translation, or chimeras for yeast display, PolyMap uses full-length antigens displayed on the surface of mammalian cells. Robust quality control mechanisms within the cell ensure the native structure and glycosylation of complex membrane-binding proteins, crucial for many therapeutic targets such as GPCRs or oligoviral antigens. A key component of this antigen expression system is the parental CHOZN cell line, characterized by engineered FRT sites at highly expressed sites in recombinase-mediated integration. This platform has been widely used for one-pot transfection and the preparation of thousands of different antibody libraries. While other methods for generating large antigen or antibody libraries in mammalian cells are also applicable to PolyMap, uniform clonal distribution and stable expression levels are highly advantageous.

[0193] The antibody library was expressed as scFv in a ribosome-displayed form, a previously validated format with several advantages. First, the soluble antibody library allows for bulk staining of suspension cells, a process previously thought to be limited to adherent cells. Second, although the ARM complex may carry multiple translational ribosomes, a short 19-amino acid C-terminal linker is used to facilitate the distribution of a single full-length antibody on each ribosome. This functional monovalent form is crucial for preventing cell aggregation and allows for the handling of stained cells via microfluidic systems. Monovalent binding is also essential for eliminating affinity effects and allowing for the differentiation of subtle variations in binding affinity. Other advantages of ribosome-displayed presentation include the ease of library generation, compatibility with very large libraries, and availability of commercially available reagents.

[0194] PolyMap's screening process is simple and scalable; antigen and antibody libraries can be reused an unlimited number of times once generated. PolyMap's unique feature is that each cell supports hundreds to thousands of interactions, meaning that available screening capacity depends primarily on the number of antigen-expressing cells that can be processed individually. If only a small number of antigen variants need to be detected, individual cell lines can be generated, stained, and then directly sequenced using CDR3H without single-cell manipulation. Figure 17A and 22A For studies involving dozens, hundreds, or more antigens, bulk transfection of plasmid libraries into cells is highly convenient and offers options that are not available with some alternative technologies. The Drop-seq used in this paper typically processes up to 10,000 single cells per hour and is suitable for studies involving dozens to hundreds of unique antigens. Other single-cell and microfluidic sequencing technologies are also compatible with PolyMap and can be explored in future iterations.

[0195] While the data processing methods are relatively simple, there are still several important considerations to understand when interpreting the data. For drop-seq analysis, the readout distribution of individual antibodies across all antigens is evaluated. In this method, it's easy to see binding patterns within each antibody, but difficult to understand binding patterns between different antibodies. In single-antigen staining methods, the readouts for each antigen are normalized, allowing for clear identification of the top binders for each variant. It's important to consider that if antibody distribution is biased, normalization must be performed based on the input distribution, and as the total number of antibodies binding to a particular antigen decreases, the proportion of readouts attributed to the remaining antibodies will be higher. In such cases, rare binders to that antigen may not have particularly high affinity, but they could be good candidates for further engineering.

[0196] PolyMap was applied to an enriched antibody library derived from convalescent COVID donors and a known set of spike protein variants, yielding several noteworthy observations. Several distinct clonal binding patterns were observed, shared among multiple antibodies (some of which have been confirmed using full-length antibodies). Although confirmatory experiments were not performed, examination of the spike protein amino acid sequence ( Figure 19 The data reveals different sequence variants associated with different binding modes, potentially indicating key amino acids of the antibody epitopes. Interestingly, although these antibodies originated from donor samples collected in April 2020, they were found to bind strongly to all subsequent SARS-CoV-2 variants (including the Lambda variant) until June 2021, when they became a variant of concern to the World Health Organization. Several monoclonal antibodies were also found to bind weakly to the highly mutated Omicron variant, which may target conserved epitopes with potential value for future therapeutic development. Finally, since one strategy for providing broad coverage is to target a number of non-overlapping epitopes, PolyMap data can help select antibodies with potential functional synergies.

[0197] Carefully selected antibody and antigen libraries offer broad application prospects for the PolyMap platform. The combination of PolyMap and antibody library capture technology allows for the analysis of immune responses from different donors, treatment regimens, or time points. Libraries of human surface proteins or other target protein populations can be amplified from cDNA and generated in batches for antibody-specific analysis. Because the antigen and antibody platform is compatible with any library generation technology, synthetic libraries can be generated for protein engineering research. For example, deep mutation scanning libraries of antigens can be generated and screened together with antibody libraries for large-scale parallel epitope localization platforms. Targeted CDR mutagenesis and a group of antigen-expressing cells can be used to develop antibodies with broad or high targeting specificity. An interesting extension of the current approach involves engineered cells with ligand-induced activatable phenotypes. In this case, ARM-stained and activated cells can be sorted based on marker activation, and then PolyMap can be used to identify specific binding chaperones. Furthermore, in principle, this technology is applicable to libraries of any soluble protein and its homologous binding chaperones expressed on the cell surface, such as cytokines and their cell-binding receptors, or immune checkpoint proteins.

[0198] method Antigen construction and cloning The amino acid sequence of the wild-type (WT) SARS-CoV-2 S protein (UniProt P0DTC2) was modified to improve expression: six stable proline mutations were added, the furin protease site (RRAR (SEQ ID NO:120)→AGAG (SEQ ID NO:121)) was removed, and the C-terminus was truncated by 19 amino acids. This sequence was codon-optimized, synthesized, and cloned into the antigen expression vector along with an 18-mer barcode sequence. Figure 14A The expression vector contains a CMV promoter with a 2G-UNic® translation enhancement element (ProteoNic) and an FRT site, followed by a 2A ribosomal jumping motif and a glutamine synthase gene. A point mutation was introduced into the WT spike protein sequence via Gibson assembly and cloned into the same expression vector backbone along with a unique barcode.

[0199] To generate a circulating spike protein variant library, protein sequences of all available spike protein variants up to August 2023 were retrieved from GISAID.org, and a phylogenetic analysis of 4,217 SARS-CoV-2 genomes collected from nextstrain.org between December 2019 and July 2023 was also performed. The most common sequences within each phylogenetic clade were selected, and unique sequences were screened. Sixteen representative spike protein variants were identified, covering a variety of mutants (…). Figure 19These sequences (including proline and furin mutations and C-terminal truncation) were synthesized with unique barcodes (HHSWNNHHCTGGNNHHSWHH, SEQ ID NO: 114) and cloned into the antigen expression vector. Wild-type sequences of CoV-1 S (P59594), human CTLA-4 (P16410), human PD-1 (Q15116), and mTagBFP2 were used as controls.

[0200] Generation and analysis of stable cell lines Stable cell lines are generated by transfecting each individual antigen construct, along with a plasmid encoding the Flp recombinase, into CHOZN® cells containing matching FRT and landing sites. Using the default CHO protocol, transfect with 10... 6 Cells were electroporated (MaxCyte) with 5 µg of a 4:1 mixture of antigen and recombinase plasmid. After recovery for 4 days in EX-CELL® CD CHO fusion medium (Millipore Sigma, 14365C) ​​supplemented with GlutaMax (ThermoFisher, 35050061), cell selection was performed under GlutaMax-free conditions for 18 days or until cell viability >95%. After selection, gDNA was collected and barcode regions were amplified using flanking primers containing sequencing ligands (antigen BCP5 FWD, antigen BCP7 REV; primer sequences are shown in Table 4). DNA samples were quantified using qPCR and normalized to 1.8 nM. The library was diluted to a final concentration of 9 pM, supplemented with 5% PhiX DNA, and then sequenced using custom primers (M13 SEQ, BGH INDEX SEQ, BGH SEQ) on an Illumina MiSeq instrument. The sequencing results were analyzed using a Python script that calculated the number of times each unique antigen barcode appeared in each library and its frequency relative to the total number of barcodes.

[0201] Flow cytometry To assess spike protein expression, cells were stained with 10 µg / mL anti-COVID-19 and SARS-CoVS glycoprotein antibody (clone CR3022, Absolute Antibody, Ab01680-10.0), followed by staining with PE or APC anti-human IgG (H+L) antibody (Jackson ImmunoResearch, 709-606-149). For validation of binding to specific antibodies identified by PolyMap, cell lines expressing spike protein variants were stained with 0.5 µg / mL supernatant from CHO cells expressing the specified antibody clone, followed by staining with APC anti-human IgG Fc antibody (BioLegend, 409306). After secondary antibody staining, cells were stained with 4',6-diamidinyl-2-phenylindole (DAPI; BioLegend, 422801). Samples were analyzed on a CytoFLEXLX flow cytometer (Beckman Coulter), and mean fluorescence intensity (MFI) of APC signals on live cells (DAPI negative) was measured using FlowJo (v10.6.1, BDBiosciences).

[0202] Antibody construction and cloning Antibody scFv sequence (V L -Connector-V H After assembly, codon optimization, and synthesis of overlapping sequences, the vector was used for Gibson assembly into a ribosome display vector. This vector does not encode a stop codon, but instead encodes Strep tag II, an unstructured sequence derived from tolA (GGQKQAEE, SEQ ID NO:122), and a secM stop sequence (FSTPVWISQAQGIRAGP, SEQ ID NO:123). Transcription from the plasmid is controlled by flanking conventional T7 promoter and terminator elements.

[0203] Antibody libraries were amplified from previously generated natural paired antibody libraries from convalescent COVID patients using primer pools designed to capture all lineages (scFv libraries FWD 1–5, REV 1–3; Table 4). These sequences were batch cloned into linearized ribosome display plasmids using Gibson assembly technology. For sequencing, V sequences were amplified directly from the plasmid library. H The region, or using universal primers that bind to the adapter and secM region, can be used to amplify V from cDNA. H Samples are prepared using the zone (scFv connector 1 P5 FWD, scFv P7REV).

[0204] ARM complex production and cell staining The ribosome display plasmid was amplified by PCR to generate a fragment spanning the T7 promoter, scFv cassette, and T7 terminator sequences (primers: ARM complex FWD, ARM complex REV; Table 4). The PCR product was used as a template for in vitro transcription (HiScribe® T7 Quick High Yield RNA Synthesis Kit, NEB, E2050S) to produce RNA, which was then treated with DNase and purified. To generate the ARM complex, 150–450 nM RNA template (based on a length of 1100 bp) was added to an in vitro transcription-translation kit (PureExpress® ARF123, NEB, E6850S) without release factors and incubated at 37°C for 10 minutes. Immediately cool the reaction mixture on ice and dilute it with RBTH buffer (20 mM HEPES, 50 mM MgCl2, 50 mM NaCl, pH 7.4, 0.05% BSA, 2.5 mg / mL heparin, 0.01% Tween-20) at a ratio of 1:5 or higher. Dilute 1×10 6 Each antigen-expressing cell was stained on ice for 30 minutes with 50 µL of diluted ARM complex, then washed three times with 1 mL of ice-cold RB (20 mM HEPES, 50 mM MgCl2, 50 mM NaCl, pH 7.4 + 0.05% BSA) before further analysis.

[0205] RNA isolation and qPCR After staining antigen-expressing cells, qPCR analysis of ARM complex binding was performed. Figure 14D The lysis buffer and stop buffer from the Cells-to-CT™ 1-Step Power SYBR™ Green Kit (ThermoFisher, A25599) were used. In a 96-well plate, 10,000 stained cells per well were lysed with 20 µL of lysis buffer for 30 minutes at room temperature, followed by 2 µL of stop buffer and incubation at room temperature for 2 minutes. For RT-qPCR assays, the secM of the lysates was detected using probes (ARM qPCR FWD, ARM qPCR REV, ARM qPCR PROBE; Table 4). The TaqMan™ RNA-to-CT™ 1-Step Kit (ThermoFisher, 4392938) was used according to the manufacturer's instructions. 2 µL of each lysate was added to a 30 µL reaction mixture (15-fold dilution); each reaction system was quadrupled and added to 6 µL of each well in a 384-well plate. A 7-point standard curve was plotted using known concentrations of RNA (ranging from 1 nM to 1 fM, serially diluted 10-fold).

[0206] Single antigen staining As described above, an equimolar mixture of ribosomal display RNA from bamranizumab, carpreximab, edevimab, ipilimumab, and pembrolizumab was used to generate the ARM complex, which was then diluted to 15 nM of input RNA and added to four separate antigen lines (WT, K444T, E484K, and F486K). Figure 17B For library-by-library experiments ( Figure 20B An ARM complex was generated using a COVID-19 antibody library and used to stain cells at 4.5 nM input RNA. After washing, total RNA was collected and amplified using universal primers with added sequencing ligands and Illumina barcodes to amplify antibody V. H Multiplexing was performed (scFv adapter 1 P5 FWD, scFv adapter 2 P5 FWD, scFv P7 REV; Table 4). The resulting library was loaded at a concentration of 1.8 nM onto a Reagent Nano Kit v2 (Illumina, MS-103-1003) for 500 cycles and run on MiSeq with a 255 x 255 readout length scheme using custom primers (M13 SEQ, scFVREV SEQ, scFv INDEX SEQ).

[0207] Single-cell sorting The four antigen lines were mixed in equal volumes, stained and washed as described above, and then single live cells (DAPI-negative) were sorted into the wells of a 96-well plate. Each well contained RNase inhibitor and 2x RT-PCR premix (SuperScript IV One-Step RT-PCR system, ThermoFisher). After freeze-thaw cycles, other reagents were added: premix, RT, and two pairs of primers for amplifying antigen barcodes or antibody V. HUniversal PCR ligands (SCS scFv FWD, SCS scFv REV, SCS antigen FWD, SCS antigen REV; Table 4) were added. A second PCR (NebNext Q5 Ultra II, NEB) was performed using 3 µL of the RT-PCR product, with the forward primers binding the ligands and one of the 8 rows of barcodes added to the plate, and the antigen- or antibody-specific reverse primers with one of the 12 columns of barcodes (SCS P5 FWD, SCS scFv P7 REV, SCS antigen P7 REV). The primers were arranged on the plate so that each well had a unique combination of row and column barcodes, enabling single-well identification. The resulting library was loaded at a concentration of 1.8 nM onto a Reagent Nano Kit v2 (Illumina, MS-103-1003) for 500 cycles and run on MiSeq with a 255 x 255 readout protocol.

[0208] Drop-Seq The drop-seq encapsulation, cDNA generation, and amplification steps were performed according to the previously described method (https: / / mccarrolllab.org / dropseq / ), with some modifications. Because cell barcodes can distinguish droplets containing multiple cells, a higher cell and bead density was used to maximize the possibility of co-encapsulation. Cells were introduced at a rate of 500 cells / µL ( Figure 17D ) or 400 cells / µL ( Figure 20B Resuspend the cells in RB buffer containing 12% OptiPrep™ (Sigma-Aldrich, D1556-250ML), then filter sequentially through a 35µm cell filter and a 20µm cell filter, and centrifuge at 30×g for 1 minute to remove any remaining large cell fragments. Do not agitate the bottom, carefully aspirate approximately 30µL of the supernatant and transfer it to a 1mL syringe. Wash the Drop-seq beads (manufactured and purchased from ChemGenes Corporation, Macosko-2011-10(V+)) according to the manufacturer's instructions, then resuspend them in 10 mM Tris pH 8.0, 1 mM EDTA, and 0.01% Tween solution at a concentration of 120 or 220 beads / µL. Before encapsulation, the beads were resuspended in Drop-seq lysis buffer (DLB, 200 mM Tris pH 7.5, 6% Ficoll PM-400, 0.2% Sarkosyl, 20 mM EDTA, 50 mM DTT) at a concentration of 240 beads / µL. Figure 17D ) or 440 beads / µL ( Figure 20B ).

[0209] The drop-seq microfluidic chip was purchased from FlowJEM. A Pico-Surf 1 (Sphere Fluidics, C023) was used for droplet oil generation. During encapsulation, the syringe pump operated at a flow rate of 900 µL / h for each water suspension and 1600 µL / h for each oil droplet. Droplet breakup was performed as described above, but 2 mL of Pico-Break (Sphere Fluidics, C082) was used to break up the droplets.

[0210] Reverse transcription and exonuclease treatment were performed as previously described, except for the template conversion oligonucleotides (Table 4). Following RT and exonuclease treatment, bead concentrations were determined using a Fuchs-Rosenthal cell counter (Bulldog Bio, DHC-N01). PCR amplification was performed on 8,000 beads in 50 µL volumes using 1x HiFi HotStart ReadyMix (Kapa Biosystems, KK2602), 0.4 µM of each cDNA PCR FWD and cDNA PCR REV primer pair. Figure 17D Three equal-amplified samples were amplified; Figure 20B Ten equal aliquots were amplified.

[0211] To prepare sequencing samples, 1xQ5 high-fidelity premix (NEB, M0494S), 0.5 µM forward primer, and 0.5 µM reverse primer were used in a 50 µL volume to prepare 5 ng ( Figure 17D ) or 20ng ( Figure 20B PCR amplification was performed on a cDNA library containing antibodies and antigens from the cDNA library. The forward and reverse primers were designed to selectively bind and amplify the antibody and antigen sequences in the cDNA library. Primers used for antibody amplification had reverse P5 / P7 Illumina linkers (Table 4): the reverse primer (Drop-seq P5 REV) contained a P5 Illumina linker and a specific Illumina index for demultiplexing; the forward primer (… Figure 17D scFv connector 1 P7 FWD and Figure 20B The scFv adapter 2 (P7 FWD) contains the P7 Illumina linker. Primers used for antigen amplification are... Figure 17D CoV-2 S FWD and Figure 20B The antigen BC FWD; the reverse primer is Drop-seq P7 REV, with different Illumina indexes for demultiplexing.

[0212] Antibody and antigen samples were sequenced separately. Antibody samples were sequenced using the MiSeq Reagent Kit v3 (Illumina, MS-102-3003) with 600 cycles on Illumina MiSeq, and antigen samples were sequenced using the MiSeq Reagent Nano Kit v2 (Illumina, MS-103-1003) with 500 cycles. For antibody sequencing, readout 1 was 26 bp (Cell BC SEQ), which is the minimum requirement for good clustering on MiSeq (bases 1-12 cell barcodes, bases 13-20 UMI); index 1 was 150 bp, used to capture antibody CDR3 (Strep tag II SEQ); and readout 2 was 6 bp, used to read the Illumina barcode (Illumina BC SEQ). For antigen sequencing, readout 1 was as follows: Figure 17D The 270bp (M13 SEQ primer) shown is as follows: Figure 20B The reading is 76 bp; index 1 is 6 bp, used to capture Illumina barcodes (Illumina BC SEQ); readout 2 is 20 bp (bases 1-12 cell barcode, bases 13-20 UMI; Cell BC SEQ).

[0213] Antibody library sequencing analysis Sequencing analysis was performed as previously described. In short, the expected error number (E) was calculated based on the Phred score of each read, and reads with E > 2 were discarded. IMGT immunoglobulin sequences were processed to generate a position-specific sequence matrix (PSSM), and frame / CDR linker sites for each nucleotide sequence were identified. Reads must have a valid predicted CDR3H sequence. Antibody clones were then conserved, and unique sequences were merged if a CDR3H sequence of 5–6 amino acids in length had a 1-amino acid difference, or if a CDR3H sequence of >6 amino acids in length had a 1–2-amino acid difference. Only clones with at least two sequencing reads were included in the analysis.

[0214] Single-cell sorting sequence analysis A custom Perl script was used to identify the plate, rows, and columns of barcodes, as well as the antigen or antibody identity for each sequencing read from the Fastq file. To determine the antigen identity for each well, only wells with a total antigen readout ≥10 were considered. To exclude wells containing multiple cells, only wells with the highest antigen readout accounting for ≥90% of the total antigen readout in a given well were considered. Wells with a total antibody readout ≥600 were further filtered. The percentage of antibody readout for each antigen was visualized as a heatmap / scatter plot using ggplot version 23.4.2 in R.

[0215] Single antigen staining and sequencing analysis A readout count table for all antigen cell lines and antibody clones was generated. The readout percentage for each antibody in each antigen cell line was calculated. Then, the percentage of differential antibody readouts between the antibody composition of each antigen cell line and the input antibody library was calculated, along with the log2 (log2fc) of the fold change. For each antigen, if an antibody's log2fc value was less than that of any negative control antibody, the log2fc of that antibody was set to NA. For ease of visualization, the top 40 antibodies with the highest sum of log2fc across all antigens were used. The antibody readout percentages and their log2fc values ​​relative to the input antibody library were visualized as a heatmap using ggplot version 23.4.2. NA values ​​are represented in white.

[0216] Drop-Seq sequencing analysis Each sequencing read was identified using a custom Python script, drop-seq cell barcoding, unique molecular identifiers (UMIs), and antigen barcoding or antibody CDR3H. Only reads with unique UMIs were considered. To identify antigens, each cell was required to contain ≥10 antigen reads, with single antigens ≥90%, to filter out droplets containing multiple cells. The total number of antibody reads per cell was determined. The percentage of antibody reads for each antigen was then normalized to the cell count. In microlibrary experiments (… Figure 17D In the study, 243 unique antigen cells and 473 cells with associated antibodies were identified. Merging these datasets produced 91 overlapping cell barcodes, allowing for the mapping of antibody-antigen interactions. In library-by-library experiments… Figure 20B In the study, a total of 2,783 cells with antigenic identity were identified. Total readouts for each antibody clone across all antigen cell lines were counted, and the readout percentage for each antibody was then normalized based on the relevant cell count. The normalized readout percentages were visualized as a heatmap using ggplot version 23.4.2.

[0217] Protein expression In mammalian expression constructs, spike protein receptor-binding domains with K444T, E484K, or F486K mutations were cloned in the same frame as the hinge and Fc regions of mouse IgG2a. Each plasmid was transiently expressed using the Expi293™ system (ThermoFisher) and purified using protein A chromatography (PrismA, Cytiva). Size and purity were validated using SDS-PAGE and SE-HPLC.

[0218] Selected monoclonal antibodies V H and V LThe sequences were cloned into vectors containing either human IgG1 or the human κ constant region. The plasmids were transiently transfected into HEK293 cells, and the antibodies were provided as harvested cell culture medium (Twist Bioscience).

[0219] Monoclonal antibodies were cloned with human IgG1 or κ constant regions into a stable expression vector containing the FRT and GS genes compatible with the landing site cell lines. Transfection and selection were performed as described above to generate stable antibody-expressing cell lines. Production runs were performed as previously described, and supernatants were collected for binding studies.

[0220] Affinity measurement Biolayer Interference (BLI) kinetic affinity measurements were performed on a GatorPrime instrument (GatorBio). ScFvs were generated using an in vitro transcription-translation kit (PURExpress, NEB, E6800). 5 μg / mL of biotinylated Strep-tag II capture antibody (GenScript, A01737) was bound to a Gator streptavidin-coated probe (GatorBio, 160002) at ≤50% saturation for 30 seconds, followed by loading of a 10-fold diluted scFv reaction mixture for 600 seconds. Binding and dissociation data for each scFv against each RBD variant were collected at binding and dissociation times of 180 and 300 seconds, respectively. RBD antigen concentrations were titrated to 100, 33, 11, and 0 nM for multi-cycle kinetic assays, and kinetics were analyzed using a dual-reference strategy as previously described, with a global 1:1 kinetic model fitted using Rmax Unlinked. A 30-second baseline measurement was performed for the capture probe, scFv loading, and association steps.

[0221] Lentiviral transfection and stable cell line production The amino acid sequence of CoV-2 SWT (UniProt P0DTC2), along with the human IgK signal peptide, c-myc tag, and PDGFR transmembrane region, was cloned into a plasmid containing the EF1α promoter and 5' and 3' lentiviral LTRs. This plasmid, along with Rev, Gag / Pol, and VSV-G envelope plasmids, was transiently transfected into HEKTa cells for lentiviral packaging. The lentiviral supernatant was collected at 24 and 48 hours, filtered, and added to Expi293™ cells (ThermoFisher) at a final dilution of 1:4. Transduced cells were stained and sorted to detect c-myc expression.

[0222] Spike protein surface expression examination Cells were stained with a series of dilutions of bamranibumab and anti-hIgG1-Fc-PE, washed, and analyzed by flow cytometry (CytoFLEX LX flow cytometer, Beckman Coulter). The median fluorescence intensity of the PE signal is reported.

[0223] In vitro translation optimization A 1:1 mixture of RNA encoding positive (camrevimumab) and control (ipilimumab) scFv was added at 450 nM to an in vitro transcription-translation kit containing (PURExpress, NEB, E6800) or without (PURExpress ΔRF123, NEB E6850) ribosome releasing factor and incubated at 37°C for 10 or 30 min to generate ARM complexes. These mixtures were diluted to 90 nM and used to stain CoV-2 SWT CHOZN cells. Total RNA was collected from washed cells and quantified using RT-qPCR (1-step RNA-to-Ct kit, ThermoFisher). Cell counts were normalized using a TaqMan probe targeting the CHO housekeeping gene Fkpb1a, while cell-bound ARM RNA was quantified using a standard curve with a specific probe targeting the CDRH3 sequence specific to either camrevimumab or ipilimumab. The ratio of camrevimumab RNA to ipilimumab RNA was reported. Figure 15 ).

[0224] Competitive ARM staining An equimolar mixture of RNA encoding 14 scFv ribosome-displayed forms was diluted 1:1 with RNA used for an irrelevant library and used to generate ARM complexes. The mixture was diluted to 45 nM or 2.25 nM positive ARMs and used to stain CoV-2 SWTCHOZN cells. Total RNA was collected, the VH region was amplified, and sequencing was performed. The percentage of bound ARMs was quantified by identifying readouts that perfectly matched the CDRH3 gene and quantifying each clone. The enrichment of each clone was calculated by dividing by the clone frequency in the input stained sample. Figure 18 ).

[0225] Affinity measurement As previously described, clones selected from competitive staining were expressed as scFv. Single-cycle BLI kinetic measurements were performed on a GatorPrime instrument (GatorBio). ScFv was generated using an in vitro transcription-translation kit (PURExpress, NEB, E6800) and then loaded onto a Strep-Tactin XT probe (GatorBio, 160033) for 60 or 120 seconds, followed by a 30 or 60-second binding and 300-second dissociation step. RBD antigen concentrations were titrated to 100, 33, 11, and 0 nM, analyzed using a single-reference strategy, and a global 1:1 kinetic model was fitted using Rmax Unlinked. Baseline measurements were performed for 60 seconds before scFv loading and at 300 or 700 seconds after the binding step.

[0226] Table 4

[0227] List of oligonucleotide sequences used in this study. [INDEX] represents a 6-base Illumina index barcode. [PLATE], [ROW], and [COLUMN] represent single-cell sorting (SCS) primers, 5-base base, row, and column barcodes, respectively.

[0228] 8. Equivalents and References Although the invention has been specifically shown and described with reference to preferred embodiments and various alternative embodiments, those skilled in the art will understand that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.

[0229] All references, published patents and patent applications cited in the text of this specification are incorporated herein by reference in their entirety for all purposes.

[0230] sequence

Claims

1. A method for high-throughput analysis of target-binding proteins (TBPs), comprising: A library of target-modified cells is provided, wherein each of the target-modified cells presents a target on a membrane; The library of the target-modified cells is brought into contact with multiple TBP-ribosome-mRNA (TRM) complexes, thereby inducing the binding between the target-modified cells and the TRM complexes; Multiple monodisperse or multidisperse emulsion droplets are generated, wherein each droplet contains a single cell from the target-modified cells, one or more TRM complexes bound to the single cell, and a lysis reagent that induces the lysis of the single cell. RNA released by the single cell can be captured on a solid surface or within a semi-permeable shell. and A library of hybrid polynucleotides is generated, which contains sequences of transcripts from the single cell and / or sequences of mRNA from the TRM complex.

2. The method according to claim 1 further includes sequencing the hybrid polynucleotide library.

3. The method according to claim 2, further comprising the step of identifying target-TBP pairs based on sequencing of the hybrid polynucleotide library.

4. The method of claim 2, further comprising the step of identifying a target-binding protein that is specific to the target.

5. The method of claim 2, further comprising the step of identifying the binding affinity or specificity of a target-binding protein that is specific to the target.

6. The method according to any one of claims 2-5, further comprising the step of determining the readout distribution of the plurality of TBPs in the plurality of TRM complexes on two or more target targets.

7. The method according to claim 6 further includes the step of normalizing the readout distribution based on the input distribution of the plurality of TBPs.

8. The method according to any one of claims 1-7, wherein at least one target comprises, or is conjugated to, a domain capable of inducing the expression of an activation marker in the target-modified cell when the target binds to TBP.

9. The method of claim 8, wherein generating a plurality of monodisperse or polydisperse emulsion droplets includes the step of sorting the target-modified cells based on the presence or absence of the activation marker.

10. The method according to any one of claims 1-9, wherein the target-modified cell expression comprises a fusion protein containing a target and a transmembrane domain.

11. The method of claim 10, wherein the target-modified cell comprises a construct encoding the fusion protein comprising a target and a transmembrane domain.

12. The method of claim 11, wherein the construct further comprises a barcode sequence.

13. The method of claims 10-12, wherein the construct further comprises a sequence encoding a fluorescent protein.

14. The method of claim 13, further comprising separating the plurality of monodisperse or multidisperse emulsion droplets containing cells modified with the target by detecting the expression of the fluorescent protein.

15. The method according to any one of claims 1-14, wherein the library of target-modified cells comprises a cell clone that presents a target.

16. The method according to any one of claims 1-14, wherein the library of target-modified cells comprises 2, 3 or 4 cell clones, wherein each of the cell clones presents a unique target that is different from the other cell clones.

17. The method according to any one of claims 1-14, wherein the library of target-modified cells comprises at least five cell clones, wherein each of the cell clones presents a unique target that is different from the other cell clones.

18. The method of claim 17, wherein the library of target-modified cells comprises at least 10 cell clones, wherein each cell clone presents a unique target that differs from the other cell clones.

19. The method of claim 18, wherein the library of target-modified cells comprises at least 100 cell clones or at least 1,000 cell clones, wherein each of the cell clones presents a unique target that differs from the other cell clones.

20. The method according to any one of claims 1-19, wherein the transcript of the isolated single cell contains the coding sequence of the target.

21. The method according to any one of claims 1-19, wherein the transcript of the isolated single cell comprises a barcode sequence.

22. The method according to any one of claims 1-21, wherein the target is an antigen, a ligand, a receptor, or a modification thereof.

23. The method according to any one of claims 1-22, wherein the target-binding protein is an antibody, ligand, receptor, or a modification thereof.

24. The method according to any one of claims 1-23, wherein each of the TRM complexes comprises a target-binding protein.

25. The method of claim 24, wherein each of the TRM complexes comprises TBP containing scFv.

26. The method according to any one of claims 1-23, wherein each of the TRM complexes comprises a TBP containing a heavy chain variable region.

27. The method according to any one of claims 1-23, wherein each of the TRM complexes comprises a TBP containing a light chain variable region.

28. The method according to any one of claims 1-27, wherein the plurality of TRM complexes comprises 1 to 5 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target-binding protein that is different from the other unique TRMs.

29. The method according to any one of claims 1-27, wherein the plurality of TRM complexes comprises 6 to 10 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target-binding protein that is different from the other unique TRM complexes.

30. The method according to any one of claims 1-27, wherein the plurality of TRM complexes comprises at least 10 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target-binding protein that is different from the other unique TRM complexes.

31. The method of claim 30, wherein the plurality of TRM complexes comprises at least 100 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target-binding protein that is different from the other unique TRM complexes.

32. The method of claim 31, wherein the plurality of TRM complexes comprises at least 1,000 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target-binding protein that is different from the other unique TRM complexes.

33. The method according to any one of claims 1-32, wherein each of the TRM complexes comprises a target-binding protein and an mRNA encoding the target-binding protein.

34. The method of claim 33, wherein the mRNA comprises the coding sequence of the complementarity-determining region (CDR) of the target-binding protein.

35. The method of claim 34, wherein the mRNA comprises the coding sequence of CDR3 of the target-binding protein.

36. The method according to any one of claims 1-35, wherein the mRNA comprises a barcode sequence.

37. The method according to any one of claims 1-36, wherein RNA capture is performed using oligonucleotides immobilized in beads.

38. The method of claim 37, wherein RNA capture is performed using oligonucleotides immobilized in beads, wherein each bead has a diameter greater than 10 μm, 0.5-10 μm, less than 1 μm, or about 1 μm.

39. The method of claim 37 or 38, wherein the oligonucleotide immobilized in the bead comprises a barcode sequence.

40. The method of claim 39, wherein the barcode sequence in the oligonucleotide immobilized in the bead is unique for a given bead.

41. The method of claim 39 or 40, wherein the hybrid polynucleotide further comprises a barcode sequence from the oligonucleotide immobilized in the bead.

42. The method according to any one of claims 1-41, wherein the hybrid polynucleotide is generated by overlap extension polymerase chain reaction (OE-PCR).

43. The method according to any one of claims 1-42, wherein the hybrid polynucleotide is generated prior to the synthesis of the first-strand cDNA.

44. The method according to any one of claims 37-43, further comprising the step of generating a second set of emulsion droplets prior to the step of generating the library of the hybrid polynucleotides, the second set of emulsion droplets containing bead-captured RNA released from the single cell.

45. The method of claim 44, wherein the hybrid polynucleotide library is generated in the second group of emulsion droplets.

46. ​​The method according to any one of claims 1-45, wherein the library of the target-modified cells is contacted with the plurality of TBP-ribosome-mRNA (TRM) complexes in a solution containing 25 to 100 mM Mg 2+ It is carried out in a buffer solution.

47. The method of claim 46, wherein the buffer solution comprises 50 mM Mg 2+ .

48. The method of claim 47, wherein the buffer solution comprises 50 mM MgCl2.

49. The method according to any one of claims 46-48, wherein the buffer comprises HEPES, NaCl, 50 mmol / L MgCl2, polysorbate 20, heparin and BSA.

50. The method of claim 49, wherein the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin and 0.5% BSA.

51. The method of claim 49, wherein the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin and 0.05% BSA.

52. A kit for high-throughput analysis of target-binding proteins (TBPs), comprising: Multiple constructs, each encoding a fusion protein containing a unique target and transmembrane domain; Multiple TBP-ribosome-mRNA (TRM) complexes, each containing a unique target-binding protein; and Buffer solution.

53. The kit according to claim 52, further comprising host cells.

54. The kit according to claim 52 or 53, wherein the kit comprises a single construct encoding a target.

55. The kit of claim 52 or 53, wherein the kit comprises 2 to 10 unique constructs, wherein each of the unique constructs encodes a unique target.

56. The kit of claim 52 or 53, wherein the kit comprises at least 10 unique constructs, wherein each of the unique constructs encodes a unique target.

57. The kit of claim 54, wherein the kit comprises at least 100 unique constructs.

58. The kit of claim 57, wherein the kit comprises at least 1000 unique constructs.

59. The kit according to any one of claims 52-58, comprising a unique TRM complex, wherein the unique TRM complex comprises a target-binding protein.

60. The kit according to any one of claims 52-58, comprising at least 10 unique TRM complexes, wherein each of the unique TRM complexes contains a unique target-binding protein.

61. The kit according to claim 59, comprising at least 100 unique TRM complexes.

62. The kit according to claim 61, comprising at least 1000 unique TRM complexes.

63. The kit according to any one of claims 52-62, further comprising reagents for overlap extension polymerase chain reaction (OE-PCR).

64. The kit according to any one of claims 52-63, wherein the buffer contains 50 mM Mg 2+ .

65. The kit of claim 64, wherein the buffer contains 50 mM MgCl2.

66. The kit according to any one of claims 52-65, wherein the buffer comprises HEPES, NaCl, 50 mM MgCl2, polysorbate 20, heparin and BSA.

67. The kit according to claim 66, wherein the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin and 0.5% BSA.

68. The kit according to claim 66, wherein the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCl2, 0.01% polysorbate 20, 2.5 mg / ml heparin and 0.05% BSA.

69. A library of hybridized polynucleotides produced by the method of any one of claims 1-51.

70. A method for high-throughput analysis of receptors, comprising: A library of ligand-modified cells is provided, wherein each of the ligand-modified cells presents a target ligand on a membrane; The ligand-modified cell library is brought into contact with multiple receptor-ribosome-mRNA (RRM) complexes, thereby inducing the binding between the ligand-modified cells and the RRM complexes; Multiple emulsion droplets are generated, wherein each droplet contains a single cell in the ligand-modified cell, one or more RRM complexes bound to the single cell, and a lysis reagent that induces the lysis of the single cell; RNA released from the single cell can be captured on a solid surface or within a semi-permeable shell. and A library of hybrid polynucleotides is generated, which contains sequences of transcripts from the single cell and / or sequences of mRNA from the RRM complex.

71. A method for high-throughput analysis of receptors, comprising: A library of receptor-modified cells is provided, wherein each of the receptor-modified cells presents a target receptor on a membrane; The receptor-modified cell library is contacted with multiple ligand-ribosome-mRNA (LRM) complexes, thereby inducing the binding between the receptor-modified cells and the LRM complexes; Multiple emulsion droplets are generated, wherein each droplet contains a single cell from the receptor-modified cell, one or more LRM complexes bound to the single cell, and a lysis reagent that induces the lysis of the single cell; RNA released from the single cell can be captured on a solid surface or within a semi-permeable shell. and to generate a library of hybrid polynucleotides containing sequences of transcripts from the single cell and / or sequences of mRNA from the LRM complex.

Images