High-throughput analysis of antibody binding and specificity

Abstract

The present disclosure provides a high-throughput assay system for studying binding between two proteins, e.g., a target protein and a target binding protein (TBP). The assay system allows parallel screening of a large number of proteins and their binding targets by combining a microfluidic system for single cell analysis and overlap extension reverse transcription PCR.

Description

HIGH-THROUGHPUT ANALYSIS OF ANTIBODY BINDING AND SPECIFICITY1. CROSS REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims the benefit of U.S. Provisional Patent Application No. 63 / 480,918, filed on January' 20, 2023, and U.S. Provisional Patent Application No.63 / 612,234, filed on December 19, 2023, the disclosures of which are hereby incorporated by reference in their entirety for all purposes.2. SEQUENCE LISTING

[0002] This application contains a Sequence Listing which has been submitted electronically in XML format and is hereby incorporated by reference in its entirety. Said XML copy created on XX, is named 28152-53884-SEQLIST.xml and is XX bytes in size.3. BACKGROUND OF THE INVENTION

[0003] Antibody therapeutics are increasingly used to treat intractable diseases such as cancer. However, the process of antibody drug discovery is expensive and tedious. The process involves identification of an antigen, isolation of antibodies and characterization and selection of antibodies with desired activities against the antigen. The workload for antibody development and characterization can be enormous, but the chance to obtain functional and effective antibodies is slim.

[0004] As highlighted by the coronavirus disease of 2019 (COVID-19) pandemic, monoclonal antibodies can be a crucial tool for preventing and controlling infectious disease in high-risk individuals. The exceptional specificity of antibodies, which provides safety and efficacy, also means that a diverse and evolving target can escape the therapeutic. For example, an early antibody developed by Eli Lilly, bamlanivimab, lost almost all potency against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) strains with the E484K mutation in the spike protein, which occurred less than a year after the beginning of the pandemic. Instead, binary antibody cocktails, serum-derived polyclonal antibodies, or recombinant polyclonal libraries show improved longevity of binding and protection against an evolving pathogen.

[0005] The natural immune response is highly polyclonal and is an ideal reservoir to mine for diverse binders. Methods have been developed for capturing and sequencing the native antibody pairing of these diverse libraries, but there is no simple, high-throughput method to assess binding specificity of the individual antibodies en masse. One approach, called LIBRA-seq, uses DNA-barcoded antigens to stain primary B cells, which are then takenthrough a single-cell sequencing workflow to obtain in silico linked antibody and antigen sequences. This method was adapted and used successfully to identify neutralizing CoV-2 antibodies, but a significant challenge is the need for individually expressed, purified, and barcoded antigens. A different strategy is to exploit existing biological systems in which two components selectively fuse. For example, Alpha-seq involves encoding two interacting libraries on opposite yeast mating types, allowing them to mate, and then sequencing the diploid cells. ENTER-seq and RAPTR display separate libraries on lentivirus particles and mammalian cells, with transduced cells sequenced to identify interacting pairs. Though these methods have high selectivity and sensitivity, each fusion only provides information on a single pairing event, requiring the analysis of a large number of cells for mapping more complex interaction networks.

[0006] Ribosome display is an alternative display technology that enables genotypephenotype linkage of a soluble protein. In its simplest form, the protein is translated from an mRNA lacking a stop codon, preventing the dissociation of the protein-ribosome-mRNA complex. The method has been further improved over time through the use of additional stall sequences and recombinantly expressed reagents. A human open reading frame (ORF) library was generated in the ribosome display platform, paired with covalently-linked DNA barcodes and used to determine the antigen specificity of autoreactive antibodies from patient samples. This technology allows screening of an antigen library against an antibody library, but the antibody sequences are not recovered, limiting this to a profiling method. SMI-Seq uses a gel immobilized ribosome displayed-cDNA barcoded protein library, which is incubated with a second barcoded library. Fluorescent sequencing is used to identify co-localized barcodes from interacting proteins; though the methodology is innovative, it is quite complex and of limited scalability.

[0007] Therefore, there is a continuing need for improved methods for generation and characterization of recombinant antibodies. In particular, high-throughput assays of a large number of antibodies and antigens are needed to expedite the discovery process for development of therapeutic antibodies.4. SUMMARY OF THE INVENTION

[0008] The present disclosure provides a high-throughput library-by-library interaction platform that utilizes cell surface-displayed antigens and soluble ribosome-displayed antibodies. Such a polyclonal mapping assay system, also referred to herein as “PolyMap,” allows for studying binding patterns between various classes of proteins, such as target proteins (e.g. , an antigen, a ligand, or a receptor) and their cognate binding partners (e.g. , anantibody, a receptor, or a ligand). This can be achieved, for example, by combining a microfluidic system for single cell analysis and overlap extension reverse transcription PCR to allow for parallel screening of a large number of proteins and their binding targets. Such methods can be used for high-throughput analysis of target binding proteins, for example for characterization of individual antibodies in a polyclonal mixture, without requiring isolation and purification, and for identification of binding partners between a library of antibodies and a library of antigens (e.g., a library of naturally occurring antigen variants). In contrast to previous work, the antigen library is expressed on the surface of mammalian cells, which supports the native structure of a wide range of human and viral proteins without the need for purification. The antibody library is expressed by ribosome display as soluble single chain variable fragments (scFvs), barcoded through their complementarity-determining region 3 heavy chain sequence (CDR3H), and used to stain the antigen presenting cells in bulk. Single cells, each potentially bound by thousands of antibody-ribosome-mRNA (ARM) complexes, are encapsulated with uniquely-barcoded RNA capture beads, and the barcoded cDNA is sequenced to reveal antibody-antigen pairings. PolyMap is compatible with synthetic library generation methods, enabling not only profiling, but also protein engineering applications. PolyMap can be used, for example, to identify CoV-2 targeting antibodies with unique antigen binding specificities.

[0009] Specifically, the present disclosure provides a method for high-throughput analysis of target binding proteins (“TBPs”), comprising: providing a library of target-decorated cells, wherein each of the target-decorated cells presents a target of interest on the membrane; contacting the library of target-decorated cells with a plurality of target binding protein (“TBP”)-ribosome-mRNA (TRM) complexes, thereby inducing binding between the targetdecorated cells and the TRM complexes; generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the target-decorated cells, one or more TRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript identifying the expressed target of the single cell and / or a sequence from the mRNA of the TRM complex bound to that cell. In some embodiments, the method comprises generating a library of hybrid polynucleic acids that comprise a sequence from a transcript identifying the expressed target of the single cell and a sequence from the mRNA of the TRM complex bound to that cell. In some embodiments, the method comprises generating a library of hybrid polynucleic acids that comprise a) a sequence from a transcriptidentifying the expressed target of the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell; or b) a sequence from the mRNA of the TRM complex bound to the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell. In some embodiments, the method comprises generating a library of hybrid polynucleic acids comprising a first set of hybrid polynucleic acids and a second set of hybrid polynucleic acids, wherein a) the hybrid polynucleic acids of the first set of hybrid polynucleic acids comprise a sequence from a transcript identifying the expressed target of the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell; and b) the hybrid polynucleic acids of the second set of hybrid polynucleic acids comprise a sequence from the mRNA of the TRM complex bound to the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell. In some embodiments, the method further comprises sequencing the library of hybrid polynucleic acids. In some embodiments, the method further comprises the step of identifying a target- TBP pair based on the sequencing of the library of hybrid polynucleic acids. In some embodiments, the method further comprises the step of identifying a target binding protein specific to the target of interest. In some embodiments, the method further comprises the step of identifying binding affinity or specificity of a target binding protein specific to the target of interest.

[0010] In some embodiments, the method further comprises the step of determining the read distributions of a plurality of TBPs in the plurality of TRM complexes across two or more targets of interest. In some embodiments, the method further comprises the step of normalizing the read distributions based on the input distribution of the plurality of TBPs.

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

[0012] In some embodiments, the target-decorated cells express a fusion protein comprising a target of interest and a transmembrane domain. In some embodiments, the target-decorated cells comprise a construct encoding the fusion protein comprising a target of interest and a transmembrane domain. In some embodiments, the construct further comprises a barcode sequence. In some embodiments, the construct further comprises a sequence encoding afluorescent 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 comprises isolating the plurality of the emulsion microdroplets containing the target-decorated cells by detecting expression of the fluorescent protein.

[0014] In some embodiments, the library of target-decorated cells comprises one cell clone presenting one target of interest. In some embodiments, the library of target-decorated cells comprises 2, 3, or 4 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of targetdecorated cells comprises at least 5 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises at least 10 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises at least 100 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises at least 1000 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones.

[0015] In some embodiments, the transcript of the isolated single cells comprises a coding sequence of the target of interest. In some embodiments, the transcript of the isolated single cells comprises a barcode sequence which uniquely identifies the target expressed by the cell.

[0016] In some embodiments, each of the TRM complexes comprises a target binding protein. In some embodiments, each of the TRM complexes comprises an scFv (e.g., each of the TRM complexes comprises a TPB comprising an scFv). In some embodiments, each of the TRM complexes comprises a heavy chain variable region (e.g., each of the TRM complexes comprises a TPB comprising a heavy chain variable region). In some embodiments, each of the TRM complexes comprises a light chain variable region (e.g., each of the TRM complexes comprises a TPB comprising a light chain variable region).

[0017] In some embodiments, 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 different from the rest of the unique TRMS. In some embodiments, 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 different from the rest of the unique TRM complexes. In some embodiments, the plurality of TRM complexes comprisesat least 10 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target binding protein different from the other unique TRM complexes. In some embodiments, 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 different from the other unique TRM complexes. In some embodiments, the plurality of TRM complexes comprises at least 1000 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target binding protein different from the other unique TRM complexes.

[0018] In some embodiments, each of the TRM complexes comprises a target binding protein and an mRNA encoding the target binding protein. In some embodiments, the mRNA comprises a coding sequence of a complementarity determining region (CDR) of the target binding protein. In some embodiments, the mRNA comprises a coding sequence of a CDR H3 of the target binding protein. In some embodiments, the mRNA comprises a barcode sequence which uniquely identifies the TBP or target binding protein encoded by the mRNA.

[0019] In some embodiments, RNA capturing is performed using oligonucleotides affixed to beads. In some embodiments, RNA capturing is performed using oligonucleotides affixed to beads, wherein each bead has a diameter larger than 10pm, 0.5-10 pm, smaller than 1 pm or about 1pm. 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 employing oligonucleotides affixed to beads, the oligonucleotides affixed to beads comprise a barcode sequence. In some embodiments, the barcode sequences in the oligonucleotides affixed to beads are unique for a given bead. For example, in some embodiments, 1) a first bead comprises oligonucleotides affixed to the first bead, wherein the oligonucleotides affixed to the first bead comprise a first barcode sequence, and 2) a second bead comprises oligonucleotides affixed to the second bead, wherein the oligonucleotides affixed to the second bead comprise 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 polynucleic acids further comprise a barcode sequence from the oligonucleotides affixed to beads.

[0021] In some embodiments, the hybrid polynucleic acids are generated by overlap extension polymerase chain reaction (OE-PCR). In some embodiments, generation of the hybrid polynucleic acids is preceded by first strand cDNA synthesis.

[0022] In some embodiments, the method further comprises the step of generating a second set of monodisperse or polydisperse emulsion microdroplets comprising the bead capturedRNA released from the single cell, prior to the step of generating the library of hybrid polynucleic acids. In some embodiments, the library of hybrid polynucleic acids is generated in the second set of monodisperse or polydisperse emulsion microdroplets.

[0023] In some embodiments, contacting the library of target-decorated cells with the plurality of TBP-ribosome-mRNA (TRM) complexes is performed in a buffer comprising 25 to 100 mM Mg2+. In some embodiments, the buffer comprises 50 mM Mg2+. In some embodiments, the buffer comprises 50 mM MgCh. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, and BSA. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, polysorbate 20, heparin, and BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer further comprises an RNase inhibitor.

[0024] In one aspect, the present disclosure provides a kit for high-throughput analysis of target binding proteins, comprising: a plurality of constructs, wherein each construct encodes a fusion protein comprising a unique target of interest and a transmembrane domain; a plurality of TBP-ribosome-mRNA (TRM) complexes, wherein each TRM complex comprises a unique target binding protein; and a buffer.

[0025] In some embodiments, the kit further comprises host cells. In some embodiments, the kit comprises a single construct, encoding one target of interest.

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

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

[0028] In some embodiments, the kit further comprises a reagent for overlap extension polymerase chain reaction (OE-PCR).

[0029] In some embodiments, the buffer comprises 50 mM Mg2+. In some embodiments, the buffer comprises 50 mM MgCh. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgC12, and BSA. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgC12, polysorbate 20, heparin, and BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer further comprises an RNase inhibitor.

[0030] In another aspect, the present disclosure provides a library of hybrid polynucleic acids generated by the method disclosed herein.

[0031] In yet another aspect, the present disclosure provides a method for high-throughput analysis of receptors, comprising: providing a library of ligand-decorated cells, wherein each of the ligand-decorated cells presents a ligand of interest on the membrane; contacting the library of ligand-decorated cells with a plurality of receptor-ribosome-mRNA (RRM) complexes, thereby inducing binding between the ligand-decorated cells and the RRM complexes; generating a plurality of monodisperse or polydisperse emulsion microdroplets, wherein each microdroplet contains a single cell out of the ligand-decorated cells, one or more RRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the RRM complex.

[0032] The present disclosure also provides a method for high-throughput analysis of receptors, comprising: providing a library of receptor-decorated cells, wherein each of the receptor-decorated cells presents a receptor of interest on the membrane; contacting the library of receptor-decorated cells with a plurality of ligand-ribosome-mRNA (LRM) complexes, thereby inducing binding between the receptor-decorated cells and the LRM complexes; generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the receptor-decorated cells, one or more LRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the LRM complex.

[0033] The present disclosure also provides a method for high-throughput analysis of antibodies, comprising: providing a library of antigen-decorated cells, wherein each of the antigen-decorated cells presents an antigen of interest on the membrane; contacting the library of antigen-decorated cells with a plurality of antibody-ribosome-mRNA (ARM) complexes, thereby inducing binding between the antigen-decorated cells and the ARM complexes; generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the antigen-decorated cells, one or more ARM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the ARM complex.5. BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0034] These and other features, aspects, and advantages of the present invention will become better understood with regard to the following description, and accompanying drawings, where:

[0035] FIG. 1 outlines the high-throughput method for analysis of target binding proteins (antibodies), (la) a library of target (antigen)-decorated cells are prepared by expressing a target library on cells; (lb) a plurality of TBP-ribosome-mRNA (TRM) complexes are prepared; (2) the library of target-decorated cells and the plurality of TRM complexes are mixed in bulk to induce binding between the target-decorated cells and the TRM complexes;(3) cells are individually encapsulated in droplets with lysis buffer and RNA capture beads;(4) RNA capture beads are isolated, and re-encapsulated in a second droplet where the target barcode and antibody CDR3 sequence could be linked and amplified; and (5) amplified DNA is isolated, prepared in bulk for sequencing, and sequenced using e.g. Illumina NGS methods. Bioinformatic analysis is used to normalize the data and map TBP:target interactions.

[0036] FIG. 2 shows FACS detection of cells expressing a spike protein (CoVl-S or CoV2- S) from one of six tested constructs (VI, V2, V3, V4, V5 and V6), using mAb against the spike protein.

[0037] FIG. 3 shows expression of 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)) as described in Example 1. The CoV2-S expression was detected with bamlamivimab (MFI).

[0038] FIG. 4 provides the structure of an exemplary construct backbone (p2G-FRT-GS) for generation of target-decorated cells.

[0039] FIG. 5 provides the structure of an exemplary construct (T7-based expression plasmid) for generation of TBP-ribosome-mRNA complex (TRM complex).

[0040] FIG. 6 shows detection of binding between CoVl or CoV2 spike WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) on target-decorated cells and an TRM complex containing an scFv of one of four mAbs (Bamianivimab, Casirivimab, Imdevimab, or Ipilimumab). TRM bound to CoV2-S variants were detected by strep-tag staining.

[0041] FIG. 7 shows detection of binding between CoVl or CoV2 spike WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) on target-decorated cells and an TRM complex containing an scFv of one of four mAbs (Bamianivimab, Casirivimab, Imdevimab, or Ipilimumab). TRM bound to CoV2-S variants were isolated and detected by taqman RT- qPCR with antibody (TBP)-specific probes.

[0042] FIG. 8 shows specificity of binding between CoV2-S antigen and TRM complexes by measuring the ratio between specific staining (RNA isolated from binding between CoV2-S antigen and TRM complex containing Casirivimab scFv) and non-specific staining (RNA isolated from binding between CoV2-S antigen and Ipilimumab scFv). The TRM complexes were generated under different cell-free-translation (TL) conditions with different template concentrations and reaction times.

[0043] FIG. 9 shows detection of binding between CoVl or CoV2 spike WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) on target-decorated cells and a mixture of five TRM complexes, where each TRM complex contains an scFv of one of four mAbs (Bamianivimab, Casirivimab, Imdevimab, or Ipilimumab). TRM bound to CoV2-S variants were isolated and detected by taqman RT-qPCR with antibody (TBP)-specific probes.

[0044] FIG. 10 provides Polymap scores representing binding between five target-decorated cell lines, wherein each target-decorated cell line expressing CoVl or CoV2 spike WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K), and five TRM complexes, where each TRM complex contains an scFv of one of five mAbs (Bamianivimab, Casirivimab, Imdevimab, Ipilimumab, or Pembrolizumab). The Polymap scores were calculated from average proportion, scaled by average reads from cells of that target as described in Example 6.

[0045] FIG. 11 illustrates the overlap extension reverse transcription PCR of scFv RNA in the TRM complex and the barcode sequence in the target-decorated cells. The OE-RT-PCR generates a product containing the VH region sequence and the barcode sequence.

[0046] FIG. 12 outlines the PolyMap platform workflow, beginning with a library of antigens expressed on the surface of mammalian cells, which are then incubated with asoluble library of antibody scFvs in ribosome display format. Stained single cells are encapsulated with uniquely-barcoded RNA capture beads (barcodes are represented by star, triangle, and diamond shapes) and lysed. Beads are isolated and used to generate cDNA, which is then further amplified with gene-specific primers for sequencing. Analysis of the cell barcode, antigen barcode, and antibody CDRH3 is used to generate maps of antibodyantigen interactions.

[0047] FIG. 13 shows CoV-2 spike expression in various cell lines. Results are expressed as flow cytometry data showing the surface expression level of CoV-2 S on stable Expi293™ and CHOZN® cell lines. The variant labeled “stabilized” includes 6 proline mutations and removal of the furin site.

[0048] FIGS. 14A-14D show results for various antigen expression systems. FIG. 14A depicts an antigen expression construct including a CMV promoter with translationenhancing element (2G), signal peptide (SP), and transmembrane (TM) region for surface display, and a unique barcode (BC) in the 3’ untranslated region (UTR). An FRT site allows integration and translation of a glutamine synthetase gene (GS) for selection. FIG. 14B shows surface expression of different spike mutants expressed stably in CHOZN cells, as measured by a monoclonal antibody and flow cytometry. FIG. 14C depicts a representative DNA fragment encoding the antibody scFv ribosome display library, which includes a T7 promoter, strep-tag II, spacer sequence after the protein (TolA) and secM ribosomal stall sequence. A 40-mer poly-A tail is internally encoded by a bovine growth hormone (BGH) polyadenylation signal at the 3’ UTR. FIG. 14D shows RNA recovered from Spike variant cells stained with clinical antibodies as ARM complexes, as measured by RT-qPCR. The table indicates expected results based on literature and affinity studies.

[0049] FIG. 15 shows optimization of ARM complex generation, as indicated by RT-qPCR of RNA recovered from CoV-2 S WT cells stained with an ARM complex mix generated at various conditions. Unique probes for the “positive” (Casirivimab) and “non-specific” RNA were used to quantify recovery of each, and the ratio calculated here. The STD ivTXTL kit includes all components for transcription, translation, and termination of the reaction, while the ARF kit omits factors assisting in ribosome release from the mRNA. The ARM complexes were translated (TL) for 10 or 30 minutes before staining.

[0050] FIG. 16 shows bio-layer interferometry (BLI) of clinical antibodies to CoV-2 spike variants. Association and dissociation curves at three concentrations are shown, along with the globally-calculated curve fits.

[0051] FIGS. 17A-17E show results for screening of 5 monoclonal antibodies and 4 SARS- CoV-2 spike variants. For all experiments, ARM complexes were generated from an equimolar mixture of the 5 mAbs. FIG 17A: Cell lines expressing the indicated spike variant (columns) were stained separately with the ARM complex mixture, then RNA was recovered and amplified for Illumina sequencing. Data shown in the heatmap is the percent of reads for each antibody sequence. FIG. 17B: Single-cell sorting. Spike variant cell lines were pooled in equal amounts, stained with the ARM complex mixture, then single cells were sorted into wells of 96-well plates. Antibody variable heavy chain regions and antigen barcodes were amplified and sequenced, and the identity of the spike variant cell line in each well was determined using the antigen barcode. The data is shown as a heatmap for the mean percent of reads of each antibody sequence associated with each antigen cell line. FIG. 17C: Data for individual cells from FIG. 17B where each cell is represented as five data points, one for each antibody. The data shown is the percentage of each antibody associated with each cell. FIG. 17D: Drop-seq binding profile. After the spike variant cell lines were pooled in equal amounts and stained with the ARM complex mixture, RNA from individual cells was captured onto beads, antibody variable heavy chain regions and antigen barcodes were amplified, then sequenced along with linked Drop-seq cell barcodes. After merging lists of antigen and antibody cell barcodes, 91 cells were analyzed for antibody binding. The data is shown as a heatmap for the mean percent of reads for the Drop-seq data. FIG. 17E: Data for individual cells from FIG. 17D where each cell is represented as five data points, one for each antibody. The data shown is the percentage of each antibody associated with each cell.

[0052] FIG. 18 shows RNA enrichment at two different staining conditions. Fold enrichment of a mixture of 14 antibody clones based on sequencing before and after staining the CoV-2 S WT cell line at two different ARM complex concentrations.

[0053] FIG. 19 shows alignment of selected CoV-2 spike variant sequences. The amino acids corresponding to the receptor binding domain (RBD) are indicated.

[0054] FIGS. 20A and 20B show binding profiles of a library of anti-SARS-CoV-2 antibodies against a library of spike variants. FIG 20A: Antibody repertoire used to generate ARM complexes was determined by sequencing heavy chain fragments from the RNA mixture serving as input for the in vitro translation reaction. The top 100 clones are plotted, along with the individual and cumulative percentages of each clone in the library. FIG. 20B: Drop-seq. Cell lines expressing the indicated spike variant (columns) were pooled and stained with the ARM complex mixture, then encapsulated with barcoded beads and taken throughthe Drop-seq workflow. The percentage of normalized reads for each antibody across all cell lines is plotted. The top 40 antibodies (based on total number of antibody reads) are shown.

[0055] FIG. 21 shows antigen cell distribution from Drop-seq library experiment. After ARM complex staining, encapsulation with barcoded Drop-seq beads (ChemGenes Corporation), and antigen amplification workflow, the identity of recovered antigen was determined by sequencing the associated antigen barcode.

[0056] FIGS. 22A and 22B show functional validation of antibody binding patterns identified by PolyMap. FIG. 22A: Cell lines expressing the indicated spike variant (x-axis) were stained separately with the anti-CoV-2 library ARM complex mixture, then RNA was recovered and amplified for Illumina sequencing. Data shown in the heatmap is the log2 foldchange of the enrichment over input for each antibody sequence associated with each cell line. Antibodies with log2 fold-change values below any of the three negative control antibodies are colored white. FIG. 22B: A subset of individual antibody clones validated by flow cytometry and compared to the previous data. For each antibody, CDR3H sequences and heat maps for three data sets are shown: top, Drop-seq (from FIG. 20B); middle, singleantigen staining (from FIG. 22A); bottom; flow cytometry. For flow cytometry, full-length monoclonal antibodies were expressed in CHO cells, and the supernatant was used to stain individual spike variant cell lines. Heat maps show MFI of the secondary antibody signal.

[0057] FIG. 23 shows RNA enrichment for clones with different affinities. Enrichment of 5 clones in a mixture based on sequencing before and after staining the CoV-2 S WT cell line. Affinities were determined with BLI.

[0058] FIG. 24 shows the distribution of antigen sequences from bulk-transfected cells. gDNA was isolated from transfected cells and subjected to sequencing analysis to determine the relative number of times each unique antigen barcode appeared.6. DETAILED DESCRIPTION OF THE INVENTION6.1. Definitions

[0059] The term “target binding protein” or TBP as used herein refers to a protein comprising one or more target-binding domains that specifically bind to a target. In some embodiments, the target-binding domain binds the target or its fragment with specificity and affinity similar to that of naturally occurring antibodies. In some embodiments, the TBP comprises an antibody. In some embodiments, the TBP consists of an antibody. In some embodiments, the TBP consists essentially of an antibody. In some embodiments, the TBP comprises an alternative scaffold. In some embodiments, the TBP consists of an alternative scaffold. In some embodiments, the TBP consists essentially of an alternative scaffold. In someembodiments, the TBP comprises an antibody fragment. In some embodiments, the TBP consists of an antibody fragment. In some embodiments, the TBP consists essentially 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.

[0060] The term “antibody” is used herein in its broadest sense and includes certain types of immunoglobulin molecules comprising one or more antigen-binding domains that specifically bind to an antigen or epitope. An antibody specifically includes intact antibodies (e.g., intact immunoglobulins), antibody fragments, and multi-specific antibodies. One example of an antigen-binding domain is an antigen-binding domain formed by a VH -VL dimer. An antibody is one type of ABP.

[0061] The term “alternative scaffold” refers to a molecule in which one or more regions may be diversified to produce one or more antigen-binding domains that specifically bind to an antigen or epitope. In some embodiments, the antigen-binding domain binds the antigen or epitope with specificity and affinity similar to that of naturally occurring antibodies.Exemplary alternative scaffolds include those derived from fibronectin (e.g., Adnectins™), the P-sandwich (e.g., iMab), lipocalin (e.g., Anticalins®), EETI-IVAGRP, BPTI / LACI- D1 / ITI-D2 (e.g., Kunitz domains), thioredoxin peptide aptamers, protein A (e.g., Affibody®), ankyrin repeats (e.g., DARPins), gamma-B-crystallin / ubiquitin (e.g, Affilins), CTLD3 (e.g., Tetranectins), Fynomers, and (LDLR-A module) (e.g, Avimers). Additional information on alternative scaffolds is provided in Binz et al., Nat. BiotechnoL, 2005 23: 1257-1268; Skerra, Current Opin. in Biotech., 2007 18:295-304; and Silacci et al., J. Biol. Chem., 2014, 289: 14392-14398; each of which is incorporated by reference in its entirety. An alternative scaffold is one type of TBP.

[0062] The term “target-binding domain” means the portion of a TBP that is capable of specifically binding to the target.

[0063] An “antibody fragment” comprises a portion of an intact antibody, such as the antigen-binding or variable region of an intact antibody. Antibody fragments include, for example, Fv fragments, Fab fragments, F(ab’)2 fragments, Fab’ fragments, scFv (sFv) fragments, and scFv-Fc fragments.

[0064] “Single-chain Fv” or “sFv” or “scFv” antibody fragments comprise a VH domain and a VL domain in a single polypeptide chain. The VH and VL are generally linked by a peptide linker. See Pliickthun A. (1994). In some embodiments, the linker is a (GGGGS)n. In someembodiments, n = 1, 2, 3, 4, 5, or 6. See Antibodies from Escherichia coli. In Rosenberg M. & Moore G.P. (Eds.), The Phtrmacology of Monoclonal Antibodies vol. 113 (pp. 269-315). Springer-Verlag, New York, incorporated by reference in its entirety.

[0065] “Affinity” refers to the strength of the sum total of non-covalent interactions between a single binding site of a molecule (e.g., an ABP) and its binding partner (e.g., an antigen or epitope). Unless indicated otherwise, as used herein, “affinity” refers to intrinsic binding affinity, which reflects a 1 : 1 interaction between members of a binding pair (e.g., ABP and antigen or epitope). The affinity of a molecule X for its partner Y can be represented by the dissociation equilibrium constant (KD). The kinetic components that contribute to the dissociation equilibrium constant are described in more detail below. Affinity can be measured by common methods known in the art, including those described herein. Affinity can be determined, for example, using surface plasmon resonance (SPR) technology (e.g., BIACORE®) or biolayer interferometry (c.g, FORTEBIO®).

[0066] With regard to the binding of a TBP to a target molecule, the terms “bind,” “specific binding,” “specifically binds to,” “specific for,” “selectively binds,” and “selective for” a particular antigen (e.g., a polypeptide target) or an epitope on a particular antigen mean binding that is measurably different from a non-specific or non-selective interaction (e.g., with a non-target molecule). Specific binding can be measured, for example, by measuring binding to a target molecule and comparing it to binding to a non-target molecule. Specific binding can also be determined by competition with a control molecule that mimics the epitope recognized on the target molecule. In that case, specific binding is indicated if the binding of the TBP to the target molecule is competitively inhibited by the control molecule.

[0067] The term “TBP-ribosome-mRNA complex” or “TRM complex” as used herein refers to a complex comprising a target binding protein, ribosome and mRNA. The TRM complex can be generated by any of the methods known in the art. For example, the TRM complex can be generated using in vitro, cell-free system. The in vitro ribosome display technology can produce stable protein (target binding protein)-ribosome-mRNA (TRM) complexes by linking individual target binding proteins to their corresponding mRNA. The TRM complexes can be formed through the deletion of the terminal stop codon from the mRNA, which causes stalling of the translating ribosome at the end of mRNA with the nascent polypeptide not released. The protein-mRNA linkage allows the simultaneous isolation of the mRNA and desirable proteins (target binding proteins) through an affinity for an immobilized ligand. The protein-mRNA complex that binds tightly to the ligand can be subjected to in situreverse transcription-PCR (RT-PCR) to recover the protein encoding DNA sequence and amplified in a PCR reaction to generate a template for further manipulation and protein.

[0068] As used herein, the term “droplet” refers to a small quantity of liquid. Droplets are typically spherical, but may be comprised of cylindrical slugs that span the full diameter of a microfluidic channel. Droplets may form in air, oil, or aqueous solutions, depending on their composition of matter and the method of formation. Droplets occur in both monodisperse and polydisperse populations.

[0069] As used herein, the term "monodisperse" refers to a property of components characterized by uniform or nearly uniform size. For example, monodisperse droplets typically require size dispersity <5% for >90% of the droplets in a mixture. In many cases, monodisperse droplet populations are more stable than droplet populations that are not monodisperse, z.e., polydisperse droplet populations. In some embodiments, generation of monodisperse droplets requires some kind of controlled microfluidic device.

[0070] As used herein, the term “cell clone” refers to at least two cells of similar kind or classification.

[0071] As used herein, the term “lysis” refers to the process of breaking the cell membrane of a cell or cells through physical or chemical means. Lysis may be achieved through a chemical surfactant such as Triton X-100, an alkaline lysis buffer, heat, electrical currents, or physical disruption.6.2. Other interpretational conventions

[0001] Ranges recited herein are understood to be shorthand for all of the values within the range, inclusive of the recited endpoints. For example, a range of 1 to 50 is understood to include any number, combination of numbers, or sub-range from 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 50.6.3. Methods for high-throughput analysis of antibodies

[0072] The present disclosure provides a method for high-throughput analysis of antibodies. Specifically, the method can comprise the steps of (i) providing a library of target-decorated cells, wherein each of the target-decorated cells presents a target of interest on the membrane; (ii) contacting the library of target-decorated cells with a plurality of TBP-ribosome-mRNA (TRM) complexes, thereby inducing binding between the target-decorated cells and the TRM complexes; (iii) generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the target-decorated cells, one or more TRM complexes bound to the single cell, and a lysis reagent inducing lysis of the single cell; (iv) capturing RNAreleased from the single cell on a solid surface or within a semi-permeable shell; and (v) generating a library of hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the TRM complex. The present disclosure further provides variation or modification of the method for analysis of the antibodies.

[0073] The library of hybrid polynucleic acids can be analyzed or sequenced to provide information related to binding between the target of interest and the TRM complex. For example, based on the sequence, a TBP that binds to the target of interest and / or their binding affinity or specificity can be studied. Accordingly, the method can further comprise the step of identifying a target- TBP pair based on the sequencing of the library of hybrid polynucleic acids. In some embodiments, a plurality of target- TBP pairs are identified by sequencing the library of hybrid polynucleic acids. In some embodiments, more than two, three, four, five, six, seven, eight, nine, ten, twenty or more pairs of target- TBP pairs are identified. In some embodiments, the method further comprises the step of identifying a target binding protein specific to the target of interest. In some embodiments, the method further comprises the step of identifying binding affinity or specificity of a target binding protein specific to the target of interest.6.3.1. Target-decorated cells

[0074] The method disclosed herein uses target-decorated cells expressing a fusion protein comprising a target of interest and a transmembrane domain. In some cases, the targetdecorated cells express a transmembrane protein comprising a target of interest, instead of a fusion protein. The target-decorated cells present the target of interest on the surface. In some embodiments, the fusion protein further comprises a signaling domain that facilitates delivery of the target of interest to the surface.

[0075] In some embodiments, the target is an antigen. In some embodiments, the target is a ligand of a receptor or a modification thereof. In some embodiments, the target is a ligand or a modification thereof.

[0076] In some embodiments, the target-decorated cells comprise a coding sequence of the target of interest. In some embodiments, a polynucleotide construct encoding the target of interest is transiently transfected into the target-decorated cells. In some embodiments, a polynucleotide construct encoding the target of interest is stably transfected into the targetdecorated cells. In some embodiments, the polynucleotide construct is a lentiviral vector. In some embodiments, the polynucleotide construct includes FRT sites that allow stable integration of the coding sequence of the target of interest into a landing pad. In someembodiments, the polynucleotide construct including FRT sites is transfected with a recombinase to allow for the stable integration.

[0077] In some embodiments, the target of interest is expressed from an exogenous polynucleotide in the target-decorated cells. In some such embodiments, the target-decorated cells comprise a construct encoding a fusion protein or a transmembrane protein comprising the target of interest. In some embodiments, the construct comprises a coding sequence of the fusion protein or the transmembrane protein operably linked to a regulatory sequence, such as a promoter. In some embodiments, the construct further comprises a barcode sequence. In some embodiments, the barcode sequence indicates a target of interest expressed or presented on the target-decorated cells. In some embodiments, the construct further comprises 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., the fluorescent protein) can be used to identify or isolate target-decorated cells comprising the construct or emulsion microdroplets containing such target-decorated cells. The emulsion microdroplets can be monodisperse or polydisperse.

[0078] In some embodiments, the construct comprises an FRT site or other site for introducing a coding sequence of a target of interest. In some embodiments, the construct comprises a self-cleavage site such as P2A or T2A. In some embodiments, the construct comprises a gene (e.g., glutamine synthetase (GS)) which allows selection of cells having stable integration of the construct.

[0079] In some embodiments, the target of interest is expressed from the genome of the target-decorated cells. In some embodiments, the genome has been genetically modified to include a coding sequence of the fusion protein or the transmembrane protein. In some embodiments, the target of interest is expressed from an endogenous sequence in the genome.

[0080] The library of target-decorated cells can comprise a single cell clone presenting one common target of interest. In some embodiments, the library of target-decorated cells comprises more than one cell clone, each cell clone expressing a unique target of interest. In some embodiments, the library of target-decorated cells comprises more than one cell clone, each cell clone expressing a unique variant of a specific target. In some embodiments, targets of interest expressed on the library of target-decorated cells are distinct from each other. In some embodiments, targets of interest expressed on the library of target-decorated cells are similar to each other. In some embodiments, targets of interest expressed on the library of target-decorated cells have similar protein sequences. In some embodiments, targets ofinterest expressed on the library of target-decorated cells are targets of the same or related protein.

[0081] In some embodiments, the library of target-decorated cells comprises 2, 3, 4, or more cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises at least 5 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises at least 10 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of targetdecorated cells comprises at least 100 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises 100 to 1000 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises more than 1000 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones. In some embodiments, the library of target-decorated cells comprises 30 to 1000 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones.

[0082] Various targets of interest can be used in the methods described herein. In some embodiments, a target specific to a disease (e.g., cancer or immune disease) is used. In some embodiments, a target specific to a pathogen (e.g., bacteria or virus) is used. In some embodiments, a target is a ligand of a receptor. In some embodiments, the target is a receptor, such as a multi-spanning membrane protein.

[0083] Various cell types can be used as the target-decorated cells in the methods described herein. The cell can be prokaryotic (e.g., bacterial cell) or eukaryotic (e.g., mammalian or plant cell, fungal). The cell can be a primary cell or a cell line. In some embodiments, the cell is a cancer cell. In some embodiments, the cell is from a multicellular organism, including, e.g., birds, plants and mammals such as humans, cows, sheep, apes, monkeys, swine, rats, mice, dogs, and cats. In some embodiments, the cell is from a single-celled organism, including, e.g., bacteria and yeast. The cell can be a prokaryote, for example, E. coli, or it can be a eukaryote, for example, a single-celled eukaryote (e.g., a yeast or other fungus), a plant cell (e.g., a tobacco or tomato plant cell), an animal cell (e.g., a human cell, a monkey cell, a hamster cell, a rat cell, a mouse cell, or an insect cell) or a hybridoma. Examples of the cells include CS-9 cells, the COS-7 line of monkey kidney cells (ATCCCRL 1651) (see Gluzman etal., 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 which grow in serum -free media (see Rasmussen et al. , 1998, Cytotechnology 28:31), HeLa cells, BHK (ATCC CRL 10) cell lines, 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), 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, cell strains derived from in vitro culture of primary tissue, primary explants, HL-60, U937, HaK or Jurkat cells. In some embodiments, the cell is an Expi293 cell or a CHOZN cell. Typically, the cell is a cultured cell that can be transformed or transfected with a polypeptide-encoding nucleic acid, which can then be expressed in the cell.6.3.2. TBP-ribosome-mRNA complexes (TRM complexes)

[0084] The methods of the present disclosure use a plurality of TBP-ribosome-mRNA (TRM) complexes comprising a target binding protein. In some embodiments, each of the TRM complexes comprises a TBP comprising an scFv. In some embodiments, each of the TRM complexes comprises a TBP comprising a heavy chain variable region. In some embodiments, each of the TRM complexes comprises a TBP comprising a light chain variable region. In some embodiments, each of the TRM complexes comprises a TBP comprising a single chain binding fragment (e.g., VH / K).

[0085] In some embodiments, the plurality of TRM complexes comprises one unique TRM complex comprising a target binding protein. In some embodiments, 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 different from the rest of the unique TRM complexes. In some embodiments, 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 different from the rest of the unique TRM complexes. In some embodiments, 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 different from the other unique TRM complexes. In some embodiments, 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 different from the other unique TRM complexes. In some embodiments, the plurality of TRM complexes comprises at least 1000 unique TRM complexes, wherein each of the unique TRM complexes comprises aunique target binding protein different from the other unique TRM complexes. In some embodiments, the plurality of TRM complexes comprises at least 10,000 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target binding protein different from the other unique TRM complexes.

[0086] In some embodiments, the plurality of TRM complexes comprises one or more target binding proteins with a known target. In some embodiments, the plurality of TRM complexes comprises one or more target binding proteins with an unknown target.

[0087] In some embodiments, the plurality of TRM complexes comprises more than one unique target binding proteins, each binding to a unique epitope on the same target. In some embodiments, the plurality of TRM complexes comprises more than one unique target binding proteins, each binding to the same epitope.

[0088] In some embodiments, each of the TRM complexes comprises a target binding protein and an mRNA encoding the target binding protein. In some embodiments, the mRNA comprises a coding sequence of a complementarity determining region (CDR) of the target binding protein. In some embodiments, the mRNA comprises a coding sequence of a CDR3 of the target binding protein. In some embodiments, the mRNA comprises a coding sequence of a VH region of the target binding protein. In some embodiments, the mRNA comprises a barcode sequence. In some embodiments, the barcode sequence identifies the target binding protein encoded by the mRNA.6.3.3. Analysis processes6.3.3.1 Contacting target-decorated cells and TBP-ribosome- mRNA (TRM) complexes

[0089] The methods disclosed herein involves the step of contacting target-decorated cells and TRM complexes. The step is performed under a condition where the target-decorated cells and the TRM complexes can bind. The conditions can be adjusted or optimized to allow the binding of the target binding proteins of the TRM complexes specifically to their targets. For example, the conditions can be adjusted or optimized by changing the buffer composition.

[0090] In some embodiments, the contacting step is performed in a buffer comprising Mg2+. In some embodiments, the buffer comprises 25 mM to 100 mM Mg2+. In some embodiments, the buffer comprises 50 mM Mg2+. In some embodiments, the buffer comprises 25 mM Mg2+. In some embodiments, the buffer comprises 75 mM Mg2+. In some embodiments, the buffer comprises 100 mM Mg2+.

[0091] In some embodiments, the contacting step is performed in a buffer comprising MgCh. In some embodiments, the buffer comprises 25 mM to 100 mM MgCh. In some embodiments, the buffer comprises 50 mM MgCh. In some embodiments, the buffer comprises 25 mM MgCh. In some embodiments, the buffer comprises 75 mM MgCh. In some embodiments, the buffer comprises 100 mM MgCh.

[0092] In some embodiments, the buffer further comprises HEPES. In some embodiments, the buffer further comprises salt. In some embodiments, the buffer further comprises NaCl. In some embodiments, the buffer further comprises polysorbate 20. In some embodiments, the buffer further comprises heparin. In some embodiments, the buffer further comprises BSA. In some embodiments, the buffer further comprises an RNase inhibitor.

[0093] In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, and BSA. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, polysorbate 20, heparin, and BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer further comprises an RNase inhibitor.6.3.3.1 Generation of microdroplets

[0094] In the methods provided herein, the samples containing the target-decorated cells and TRM complexes obtained from the contacting step can be isolated into microdroplets. Monodisperse emulsions can be formed on a microfluidic chip, or polydisperse emulsions can be formed using a machine such as the IKA Utra-Turrax Tube Drive system. In some embodiments, a microfluidic system with three pressure pumps (e.g., Dolomite microfluidics) is used. Methods of generating microdroplets are known in the art, for example, as disclosed in W02016200577A1, which is incorporated by reference in its entirety.

[0095] In some embodiments, a microfluidic device is used to generate single cell emulsion droplets. The microfluidic device ejects single cells in aqueous reaction buffer into a hydrophobic oil mixture. The device can create thousands of emulsion microdroplets per second. After the emulsion microdroplets are created, the device ejects the emulsion mixture into a trough. The mixture can be pipetted or collected into a standard reaction tube for downstream processing, such as thermocycling.

[0096] Custom microfluidics devices for single-cell analysis are routinely manufactured in academic and commercial laboratories (Kintses et al. , 2010 Current Opinion in Chemical Biology 14:548-555). For example, chips may be fabricated from polydimethylsiloxane (PDMS), plastic, glass, or quartz. In some embodiments, fluid moves through the chips through the action of a pressure or syringe pump. Single cells can even be manipulated on programmable microfluidic chips using a custom dielectrophoresis device (Hunt et al., 2008 Lab Chip 8:81-87). In one embodiment, a pressure-based PDMS chip comprised of flow focusing geometry manufactured with soft lithographic technology is used (e.g., Dolomite Microfluidics (Royston, UK)) (Anna et al., 2003 Applied Physics Letters 82:364-366). The stock design can typically generate 10,000 aqueous-in-oil monodisperse microdroplets per second at size ranges from 10-150 pm in diameter. In some embodiments, the hydrophobic phase will consist of fluorinated oil containing an ammonium salt of carboxy-perfluoropoly ether, which ensures optimal conditions for molecular biology and decreases the probability of droplet coalescence (Johnston etal., 1996 Science 271 :624-626). To measure periodicity of cell and droplet flow, images are 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).

[0097] The microfluidic system can optimize microdroplet size, input cell density, chip design, and cell loading parameters such that greater than 98% of droplets contain a single cell. There are three common methods for achieving such statistics: (i) extreme dilution of thecell solution; (ii) fluorescent selection of droplets containing single cells; and (iii) extreme cell dilution to control the multi-hit rate and fluorescent cell sorting to reduce the negative rate. In some embodiments, target-decorated cells comprise a coding sequence of a reporter protein, such as a fluorescent protein. In some such embodiments, the reporter protein can be used to detect and identify droplets containing the target-decorated cells. In particular embodiments, the fluorescent protein is used to isolate droplets containing a single targetdecorated cell.

[0098] In some embodiments, input cell flow is aligned with droplet formation periodicity, such that greater than 98% of droplets contain a single cell (Edd et al., 2008 Lab Chip 8: 1262-1264; Abate et al., 2009 Lab Chip 9:2628-31). In these microfluidic devices, a high- density suspension of cells is forced through a high aspect-ratio channel, such that the cell diameter is a large fraction of the channel's width. A number of input channel widths and flow rates can be tested to arrive at an optimal solution.

[0099] In certain embodiments of the invention, microfluidic chips are used to isolate 10, 100, 1000, 10,000, 100,000, 1 million, or 1 billion single cells from a heterogeneous pool of target-decorated cells. In some embodiments, the methods of the invention use single cells in reaction containers, rather than emulsion droplets. Examples of such reaction containers include 96 well plates, 0.2 mL tubes, 0.5 mL tubes, 1.5 mL tubes, 384-well plates, 1536-well plates, etc.

[0100] In some embodiments, the isolated target-decorated cells in the microdroplets are bound to the TRM complexes. In some embodiments, the isolated target-decorated cells in the microdroplets are not bound to the TRM complexes.

[0101] In some embodiments, the isolated target-decorated cells are encapsulated with a lysis reagent. In some embodiments, the isolated target-decorated cells are in a condition where the lysis reagent can induce lysis of the target-decorated cells.6.3.3.3 Capturing RNA released from a single cell

[0102] The methods described herein further comprises the step of capturing RNA released from the target-decorated cells. In some such embodiments, the RNA molecules can be captured using beads comprising polynucleic acid probes targeting the RNA molecules. In an embodiment, the bead is a spherical bead comprising agarose, glass, chemical polymers, or magnetic materials. In some embodiments, the beads are made of materials such as latex, glass, or silica, ranging in size from 0.1 micron to 1 mm. In some embodiments, the beads have a diameter larger than 10 pm. In some embodiments, the beads have a diameter larger than 100 pm. In some embodiments, the beads have a diameter less than 1 mm. In some embodiments, the beads have a diameter of 0.5-10 pm, smaller than 1 pm or about 1 m. In an embodiment, the probes comprise biotin and the bead comprises streptavidin attached to the surface of the bead. In some embodiments, the beads are solid beads or porous beads.

[0103] In some embodiments of the invention, the target-decorated cells are encapsulated into droplets with beads comprising bound polynucleic acid probes with sequences that are complementary to polynucleic acid targets of interest in the single cells. In certain embodiments, the probes are 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130 or more nucleotides in length. The probes are RNA, DNA, locked nucleic acid (LNA), peptide nucleic acid (PNA), glycol nucleic acid (GNA), or any nucleic acid analogue. The polynucleic acid targets are either DNA or RNA. The reagent mixture present in these embodiments comprising beads is specifically designed to both lyse the cells and encourage polynucleic acid hybridization, thus allowing the beads to capture DNA or RNA targets of interest.

[0104] In some embodiments of the invention, the polynucleic acid probes target the constant region of immunoglobulin. In some embodiments, the polynucleic acid probes target the constant region of IgK or IgG. In some embodiments, the polynucleic acid probes target the mRNA sequence in the TRM complex. In some embodiments, the polynucleic acid probes comprises a sequence complementary to the mRNA sequence in the TRM complex or a fragment thereof.

[0105] In some embodiments of the invention, the polynucleic acid probes target transcripts in the target-decorated cells. In some embodiments, the polynucleic acid probes target the transcripts of targets of interest in the target-decorated cells. In some embodiments, the polynucleic acid probes target the transcripts of fusion proteins comprising a target of interest and a transmembrane domain. In some embodiments, the polynucleic acid probes target the transcripts of the transmembrane domains of the fusion proteins. In some embodiments, thepolynucleic acid probes comprise oligo(dT) polynucleotides capable of hybridizing to the poly (A) tail of mRNA.

[0106] In some embodiments, the polynucleic acid probes target barcode sequences. In some embodiments, the polynucleic acid probes target both the TRM complexes (e.g., mRNA within the TRM complexes) and the transcript of a fusion protein or a target of interest.

[0107] In some embodiments, a mixture of several polynucleic acid probes is used. In some such embodiments, a first set of polynucleic acid probes targeting the TRM complexes and a second set of polynucleic acid probes targeting the transcript of a fusion protein or a target of interest can be combined.

[0108] In some embodiments, 5'-amino-modified polynucleic acid probes are bound to carboxylic acid beads using 2-(N-morpholino) ethane sulfonic acid (MES) buffer (Kojima et al., 2005, Nucleic Acids Research 33:el50). In other embodiments, biotinylated polynucleic acid probes are bound to streptavidin-coated beads.

[0109] In some embodiments, the methods of the invention use single cells in reaction containers, rather than emulsion droplets. Examples of such reaction containers include 96- well plates, 0.2 mL tubes, 0.5 mL tubes, 1.5 mL tubes, 384-well plates, 1536-well plates, etc. A variety of other designs of microfluidic chips can also be used to isolate single cells (Marcus et al., 2006, Anal Chem 78:3084- 3089).

[0110] In some embodiments, the cell or subpopulation of cells is added to a reaction container along with beads comprising bound polynucleic acid probes and a lysis buffer. The lysis buffer lyses the cells to allow the polynucleic acid probes to bind to the polynucleic acid targets of interest from the cell or cells. The beads hybridized to the polynucleic acid targets are isolated from the lysis buffer as single beads or subpopulations of beads into reaction vessels, and are contacted with a PCR mix to allow amplification and or fusion of the polynucleic acid targets.[OHl] In certain embodiments of the invention, the aqueous phase of the droplet emulsions containing beads and their bound targets is recovered using a solvent, such as ethyl ether. The beads are isolated into emulsions with a PCR mix, such that, on average, single beads are isolated into single emulsion microdroplets (DeKosky et al., 2015, Nat Med 21 :86-91). Monodisperse emulsions can be formed on a microfluidic chip, or polydisperse emulsions can be formed using a machine such as the IKA Utra-Turrax Tube Drive system. The PCR reagents amplify a plurality of the DNA or RNA targets of interest. In some embodiments of the invention, the PCR products are linked transcripts of the fusion protein and the target of interest in the target-decorated cells and the mRNA of the TRM complex.

[0112] In some embodiments, the step of capturing RNA released from the target-decorated cells is omitted. In some cases, the target-decorated cells are encapsulated in semi-permeable shells, which allows a workflow without the step of capturing RNA released from the targetdecorated cells. In some such embodiments, the following step of generating a library of hybrid polynucleotide acids is performed in the microdroplet containing the RNA released from the target-decorated cells. The amplification reaction can be performed in a microdroplet loaded with reagents for reverse transcription and PCR (e.g., OE-RT-PCR) reactions.6.3.3.4 Generation of a library of hybrid polynucleic acids

[0113] The methods described herein further comprises the step of generating a library of hybrid polynucleic acids that comprise a sequence from a transcript from a single cell and / or a sequence from an mRNA of a TRM complex. In some aspects, the hybrid polynucleic acids comprise a sequence from a transcript from a single cell and a sequence from an mRNA of a TRM complex. In other aspects, the hybrid polynucleic acids comprise a sequence from a transcript from a single cell or a sequence from an mRNA of a TRM complex. The sequence from a transcript from the single cell can identify a target expressed from the single cell. The sequence from the mRNA can be from the TRM complex bound to the single cell, and therefore can be used to identify the TRM complex. In some embodiments, the hybrid polynucleic acids described herein further comprise a sequence associated with the single cell. For example, in some embodiments, the hybrid polynucleic acids further comprise a barcode sequence from an oligonucleotide attached to a particle, such as a bead, associated with the single cell. In some such embodiments, the hybrid polynucleic acids comprise a) a sequence from a transcript identifying the expressed target of the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell; or b) a sequence from the mRNA of the TRM complex bound to the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell. In some embodiments, the library of hybrid polynucleic acids comprises a first set of hybrid polynucleic acids and a second set of hybrid polynucleic acids, wherein a) the hybrid polynucleic acids of the first set of hybrid polynucleic acids comprise a sequence from a transcript identifying the expressed target of the single cell and a sequence (e.g., a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell; and b) the hybrid polynucleic acids of the second set of hybrid polynucleic acids comprise a sequence from the mRNA of the TRM complex bound to the single cell and a sequence (e.g, a barcode sequence) from an oligonucleotide affixed to a particle (e.g., a bead) associated with the single cell. In some such embodiments, the library of hybrid polynucleic acids comprise sequences associated with a plurality of cells, and for each of the plurality of cells, the barcode sequence of oligonucleotides attached to a particle associated with the given cell is unique as compared to barcode sequences of oligonucleotides attached to particles associated with the other cells. This allows for hybrid polynucleic acids comprising sequences associated with a given cell to share a common barcode, which is different from barcodes shared by other hybrid polynucleic acids comprising sequences associated with other cells.

[0114] The hybrid polynucleic acids can be generated from RNA isolated from a microdroplet. Specifically, the hybrid polynucleic acids can be generated from RNA released from target-decorated cells and TRM complexes isolated into the same microdroplet. In some such embodiments, the target of interest in the target-decorated cell and the target binding protein of the TRM complex can have a high affinity and specificity. In some embodiments, the RNA is captured using beads comprising polynucleic acid probes targeting the RNA molecules. In some embodiments, the beads comprise polynucleic acid probes comprising a barcode sequence. In some embodiments, the barcode sequences of the polynucleic acid probes are unique for each microdroplet.

[0115] In some embodiments, PCR is used to generate the hybrid polynucleic acids. In some embodiments, the hybrid polynucleic acids are generated by high-throughput amplification in reaction vessels. As used herein, the term “reaction vessel” refers to any entity that provides physical separation of a reaction into separate compartments. The reaction vessels can be plastic compartments, microfluidic chambers, or droplets, e.g., of an aqueous reaction solution. When droplets are used, the methods described herein further comprise the step of generating droplets comprising the reaction mixture with the bead-isolated RNA.

[0116] In some embodiments, generation of the hybrid polynucleic acids is preceded by first strand cDNA synthesis. In some embodiments, oligo(dT), random primers, or a combination of them are used to prime a reverse transcription reaction. In some embodiments, oligo(dT) is used to initiate the synthesis preferably at the 3’ end of the RNA fragment.

[0117] In some embodiments, methods for overlap extension PCR are used to create fusion amplicon products of a transcript of the target-decorated cell and a sequence from the TRM complex in a single reaction tube or microdroplet. In some embodiments, the fusion amplicon products further comprise sequences (e.g., barcode sequences) from oligonucleotides affixed to a particle (e.g., bead) associated with the target-decorated cell. Methods to amplify and link amplicons are disclosed in Johnson el al.. 2005 Genome Research 15: 1315-24; U.S. Patent 7,749,697; PCT Publication WO 2012 / 083225; and PCT Publication WO 2013 / 096643, each of which is incorporated herein by reference in its entirety.

[0118] In some embodiments, at least two nucleic acid target sequences (e.g, a first target sequence in a transcript of target-decorated cells and a second target sequence in a TRM complex) are chosen and designated as target sequences. Forward and backward primers are designed for each of the two nucleic acid target sequences, and the primers are used to amplify the target sequences. “Minor” amplicons are generated by amplifying the two nucleic acid target sequences separately, and then fused by amplification to create a fusion amplicon,also known as a “major” amplicon. In one embodiment, a “minor” amplicon is a nucleic acid sequence amplified from a first target sequence, and a “major” amplicon is a fusion complex generated from sequences amplified between multiple target sequences, e.g., a recombinant fusion polynucleotide. In some embodiments, the target sequences further include a third target sequence in an oligonucleotide affixed to particles (e.g., beads) associated with targetdecorated cells.

[0119] The methods described herein use “inner” primers (z.e., the reverse primer for the first target sequence and the forward primer for the second target sequence) comprising one domain that hybridizes with a minor amplicon and a second domain that hybridizes with a second minor amplicon. “Inner” primers are a limiting reagent, such that during the exponential phase of PCR, inner primers are exhausted, driving overlapping domains in the minor amplicons to anneal and create major amplicons.

[0120] PCR primers are designed against targets of interest using standard parameters, e.g., melting temperature (Tm) of approximately 55-65°C, and with a length 20-50 nucleotides. The primers are used with standard PCR conditions, for example, 1 mM Tris-HCl pH 8.3, 5 mM potassium chloride, 0.15 mM magnesium chloride, 0.2-2 pM primers, 200 pM dNTPs, and a thermostable DNA polymerase. Many commercial kits are available to perform PCR, such as Platinum Taq (Life Technologies), Amplitaq Gold (Life Technologies), Titanium Taq (Clontech), Phusion polymerase (Finnzymes), and HotStartTaq Plus (Qiagen). Any standard thermostable DNA polymerase can be used for this step, such as Taq polymerase or the Stoffel fragment.

[0121] 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 includes a sequence that is complementary to a first target nucleic acid sequence (e.g., the 5’ end of the first target nucleic acid sequence). The second probe includes a sequence that is complementary to the first target nucleic acid sequence (e.g., the 3’ end of the first target nucleic acid sequence) and a second sequence that is complementary to an exogenous sequence. In some embodiments, the exogenous sequence is a non-human nucleic acid sequence and is not complementary to either of the target nucleic acid sequences. For example, the exogenous sequence might be a polynucleotide sequence that encodes a polypeptide sequence rich in Ser and Gly amino acids, which links heavy and light chain variable regions in an scFv (see, e.g., PCT / US1992 / 001478). The first and second probes are the forward primer and reverse primer for the first target nucleic acid sequence.

[0122] The third probe includes a sequence that is complementary to the portion of the second probe that is complementary to the exogenous sequence and a sequence that is complementary to the second target nucleic acid sequence (e.g., the 5' end of the second target nucleic acid sequence). The fourth probe includes a sequence that is complementary to the second target nucleic acid sequence (e.g., the 3' end of the second target nucleic acid sequence). The third probe and the fourth probe are the forward and reverse primers for the second target nucleic acid sequence.

[0123] The second and third probes are also called the “inner” primers of the reaction (z.e., the reverse primer for the first locus and the forward primer for the second locus) and are limiting in concentration, (e.g., 0.01 pM for the inner primers and 0.1 pM for all other primers). This will drive amplification of the major amplicon preferentially over the minor amplicons. The first and fourth probes are called the “outer” primers.

[0124] The first and second nucleic acid sequences are amplified independently, such that the first nucleic acid sequence is amplified using the first probe and the second probe, and the second nucleic acid sequence is amplified using the third probe and the fourth probe. Next, a fusion complex is generated by hybridizing the complementary sequence regions of the amplified first and second nucleic acid sequences and amplifying the hybridized sequences using the first and fourth probes. This is called overlap extension PCR amplification.

[0125] During overlap extension PCR amplification, the complementary sequence regions of the amplified first and second nucleic acid sequences act as primers for extension on both strands and in each direction by DNA polymerase molecules. In subsequent PCR cycles, the outer primers prime the full, fused sequence such that the fused complex is duplicated by DNA polymerase. This method produces a plurality of fusion complexes.

[0126] In some embodiments, the fusion complexes are cloned into an expression vector. In some embodiments, the expression vector is a plasmid or a phagemid. In an embodiment, a regulatory sequence (e.g., promoter) is inserted into the expression vector by a method selected from: Gibson assembly, site-specific digestion and ligation, and targeted recombination.

[0127] In some embodiments, the coding sequence of the target binding protein in the fusion complexes is separately cloned into an expression vector. The expression vector can comprise the coding sequence of the target binding protein operably linked to a regulatory sequence that can induce expression of the target binding protein.

[0128] In one aspect, the present disclosure provides a library of hybrid polynucleic acids generated by a method as described herein.

[0129] In some embodiments, the methods described herein do not comprise the step of generating a library of hybrid polynucleic acids that comprise a sequence from a transcript identifying the expressed target of the single cell and a sequence from the mRNA of the TRM complex bound to that cell. Instead of generating such hybrid polynucleic acids, the method can comprise the step of generating a first library of polynucleic acids that comprise a sequence from a transcript from a single cell and a first barcode, and a second library of polynucleic acids that comprise a sequence from the mRNA of the TRM complex and a second barcode. The first barcode and the second barcode can be used to identify a target expressed from a single target-decorated cell, and the TRM complex bound to the single target-decorated cell. For example, the first barcode and the second barcode from a single microdroplet are identical or related. The method can further comprise the step of analyzing the first library of polynucleic acids and the second library of polynucleic acids to identify a target of interest and a target binding protein that can bind to the target of interest.6.3.3.5 Analysis of hybrid polynucleic acids

[0130] The methods described herein can further comprise the step of analyzing the library of hybrid polynucleic acids. In some embodiments, the analysis involves sequencing. DNA sequencing is used to verify the construct or library of constructs. In some embodiments, DNA sequencing is performed using massively parallel methods from vendors such as Illumina or using single-clone sequencing methods such as Sanger sequencing from vendors such as Applied Biosystems. The sequence of the hybridized polynucleic acids can comprise the sequence of a target of interest and / or the sequence of a TRM complex or a portion thereof. Thus, the sequence can be used to identify a target of interest expressed on the surface of target-decorated cells and the TRM complex bound to the target of interest. The sequence of the hybridized polynucleic acids can further comprise the sequence of an oligonucleotide, or a portion thereof, affixed to particles (e.g., beads) associated with targetdecorated cells. Thus, pairs of targets of interest and bound TRM complexes from a given cell can be identified.

[0131] The analysis can involve identification of the target of interest and the TRM complex isolated in the microdroplet. A TRM complex that binds to a target of interest can be isolated in the same microdroplet together with the target-decorated cells expressing the target of interest. Thus, in some embodiments, the analysis involves identification of a target of interest and a target binding protein bound to the target of interest. In some cases, the analysis involves identification of a target of interest and a target binding protein that can bind to the target of interest. In some embodiments, the analysis can further involveidentification of multiple TRM complexes that bind to a target of interest in the same microdroplet, such as where the microdroplet comprises a particle (e.g., a bead) with affixed oligonucleotides comprising a barcode sequence.

[0132] In some embodiments, a target-decorated cell can bind to more than one TRM complex. In some such embodiments, a plurality of TRM complexes are isolated in the same microdroplet with the target-decorated cell. In some embodiments, RNA captured from the microdroplet can comprise sequences from more than one TRM complex. Thus, hybridized polynucleic acids generated from one microdroplet can comprise one or more pairs of a target of interest and a target binding protein. Thus, in some embodiments, multiple target binding proteins that bind to the same target of interest are identified. In some embodiments, hybridized polynucleic acids generated from one microdroplet can comprise a barcode sequence from oligonucleotides affixed to a particle (e.g., a bead) in the microdroplet.

[0133] In some embodiments, the methods described herein can be performed in conditions of various degrees of stringency. Pairs of targets of interest and TRM complexes identified in the various conditions can be analyzed to test their binding affinities.

[0134] In some embodiments, a method as described herein is used to identify a subset of target binding proteins having a high binding affinity to a target of interest. In some embodiments, the method is used to identify a subset of target binding proteins having a desired binding affinity to a target of interest.

[0135] In some embodiments, the analysis involves quantification of a specific pair of a target of interest and a target binding protein. The quantification data can be used to measure affinity and specificity of the target binding protein against the target of interest.

[0136] In some embodiments, the analysis involves identification of all the target binding proteins that bind to a specific target of interest. In some embodiments, the analysis involves quantification of target binding proteins that bind to a specific target of interest. The quantification data can be used to identify a target binding protein having a desired affinity and / or specificity to the target of interest. In some embodiments, the quantification can be used to determine binding affinity and / or specificity of the target binding protein against the target of interest by comparing the data from a known target- TBP pair. In some embodiments, the process can be used for affinity maturation of an antibody by testing binding of various modifications of the antibody to a target of interest. Specifically modified antibodies can be selected based on their binding activity (e.g., affinity and specificity) against the target of interest.

[0137] In some embodiments, the analysis involves identification of all the targets of interest that bind to a specific target binding protein. In some embodiments, the analysis involves quantification of the targets of interest that bind to a specific target binding protein. In some embodiments, a method as described herein is used to characterize individual antibodies in a polyclonal mixture, without requiring isolation and purification. In some embodiments, the method is used for high-throughput screening of binding partners between a diverse library of antibodies and a diverse library of targets (e.g., a library of naturally occurring target variants). The quantification can be used to identify the target of the target binding protein.

[0138] In some embodiments, a method of the present disclosure is used to analyze binding of polyclonal antibodies against a target of interest. In some embodiments, the method of the present disclosure is used to analyze binding of polyclonal antibodies in a sample from a donor against a target of interest to select a donor. In some embodiments, the method is used to analyze and monitor the quality or production yields and consistency of polyclonal antibodies against a target of interest. In some embodiments, the method is used to characterize a polyclonal antibody. In some embodiments, the method is used for QC monitoring of an antibody library over time, assessment of diversity of the library, epitope mapping, or functional assessment of an antibody against mutations.

[0139] In some embodiments, a method as described herein is used to identify an epitope of an antibody by identifying binding between the antibody and a library of targets, wherein the targets are variants of a target protein with one or more mutations at different sites. The epitope can be identified by analyzing which mutations cause loss of binding of the antibody to the target protein. In some embodiments, the method can be used to identify epitopes of polyclonal antibodies. In some such embodiments, based on the epitope mapping, the diversity of the polyclonal antibodies can be determined.6.3.4. Analysis of interaction between ligand and receptor

[0140] The high-throughput analysis methods disclosed herein can be used to study the interaction between a ligand and a receptor. In some such embodiments, the target can be a receptor or a fragment thereof, and the ABP is a ligand or a fragment thereof. Alternatively, the target can be a ligand or a fragment thereof, and the ABP is a receptor or a fragment thereof. In some embodiments, the method comprises the steps of (i) providing a library of ligand-decorated cells, wherein each of the ligand-decorated cells presents a ligand of interest on the cell membrane; (ii) contacting the library of ligand-decorated cells with a plurality of receptor-ribosome-mRNA (RRM) complexes, thereby inducing binding between the liganddecorated cells and the RRM complexes; (iii) generating a plurality of emulsionmicrodroplets, wherein each microdroplet contains a single cell out of the ligand-decorated cells, one or more RRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the RRM complex. In some embodiments, the method comprises the steps of: (i) providing a library of receptor-decorated cells, wherein each of the receptor-decorated cells presents a receptor of interest on the cell membrane; (ii) contacting the library of receptor-decorated cells with a plurality of ligand-ribosome-mRNA (LRM) complexes, thereby inducing binding between the receptor-decorated cells and the LRM complexes; (iii) generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the receptordecorated cells, one or more LRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the LRM complex.

[0141] The library of hybrid polyucleic acids can be analyzed or sequenced to provide information related to binding between the ligand of interest and the RRM complex or the receptor of interest and the LRM complex. For example, based on the sequence, a receptor that binds to the ligand of interest and / or their binding affinity or specificity can be studied. Accordingly, the method can further comprise the step of identifying a ligand-receptor pair based on the sequencing of the library of hybrid polynucleic acids. In some embodiments, the method further comprises the step of identifying a ligand binding protein specific to the ligand of interest. In some embodiments, the method further comprises the step of identifying binding affinity or specificity of a ligand binding protein specific to the ligand of interest.6.4. Kits for high-throughput analysis of antibodies

[0142] The present disclosure also provides a kit for use in the high-throughput analysis of antibodies.

[0143] The kit can comprise a plurality of polynucleotide constructs, wherein each construct encodes a fusion protein comprising a unique target of interest and a transmembrane domain. In some embodiments, the kit comprises a plurality of polynucleotide constructs, wherein each construct comprises a coding sequence of a transmembrane domain and an insertion site for adding a coding sequence of a target of interest.

[0144] In some embodiments, the kit further comprises a plurality of TBP-ribosome-mRNA (TRM) complexes, wherein each TRM complex comprises a unique target binding protein. In some embodiments, the plurality of TRM complexes comprise 1, 2, 3, 4, 5 or more unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 5 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 10 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 50 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 100 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 500 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 1000 unique target binding proteins. In some embodiments, the plurality of TRM complexes comprise more than 10,000 unique target binding proteins.

[0145] In some embodiments, the kit further comprises a reagent for PCR. In some embodiments, the kit further comprises a reagent for overlap extension polymerase chain reaction (OE-PCR). In some embodiments, the kit further comprises a reagent for first strand cDNA synthesis. In some embodiments, the kit further comprises a polymerase.

[0146] In some embodiments, the kit further comprises a buffer that can be used in the step of contacting target-decorated cells and the TRM complexes. In some embodiments, the buffer comprises Mg2+. In some embodiments, the buffer comprises 25 mM to 100 mM Mg2+. In some embodiments, the buffer comprises 50 mM Mg2+. In some embodiments, the buffer comprises 25 mM Mg2+. In some embodiments, the buffer comprises 75 mM Mg2+. In some embodiments, the buffer comprises 100 mM Mg2+.

[0147] In some embodiments, the buffer comprises MgCh. In some embodiments, the buffer comprises 25 mM to 100 mM MgCh. In some embodiments, the buffer comprises 50 mM MgCh. In some embodiments, the buffer comprises 25 mM MgCh. In some embodiments, the buffer comprises 75 mM MgCh. In some embodiments, the buffer comprises 100 mM MgCh.

[0148] In some embodiments, the buffer further comprises HEPES. In some embodiments, the buffer further comprises salt. In some embodiments, the buffer further comprises NaCl. In some embodiments, the buffer further comprises polysorbate 20. In some embodiments, the buffer further comprises heparin. In some embodiments, the buffer further comprises BSA.

[0149] In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, and BSA. In some embodiments, the buffer comprises HEPES, NaCl, 50 mM MgCh, polysorbate 20,heparin, and BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.5% BSA. In some embodiments, the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 0.01% polysorbate 20, 2.5 mg / ml heparin, and 0.05% BSA. In some embodiments, the buffer further comprises an RNase inhibitor.7. EXAMPLES7.1. Example 1: Development of target-decorated cell lines and constructs for target display

[0150] Six different constructs (VI -V6) were tested for generation of target-decorated cells. The constructs had different leader sequences (murine IgM, native CoVl-S or CoV2-S or tPA) and transmembrane domains (cmyc-PDGFR, native CoVl-S or CoV2-S, or native sequences with 19 or 37 amino acid c terminal truncation) as specified in TABLE 1.

[0151] The constructs were transiently transfected into expi293 cells and expression of the transgenes were assessed 24 hours later using mAh against spike protein by flow cytometry (FIG. 2). The study showed that constructs with a native leader, particularly constructs with a native leader and a transmembrane domain with 19 amino acid truncation (Native Al 9) worked well.

[0152] The constructs were further modified for generation of three stable cell lines expressing the CoV2-S antigen. First, VI construct was packaged into lentivirus and used to generate a stable cell line in Expi293 cells. Cells transfected with the Vl-letivirus were isolated by positive c-myc signal on FACS. (Expi293-CoV2-S) The second and the third constructs were modified to be compatible with CHOZN cells. In particular, the second construct was modified to include components of V5, a native leader, and an FRT site with P2A and glutamine synthetase gene (GS), which allowed selection of stable integrants via glutamine-deficient media. This construct also utilizes a 2G-UNIC translation enhancing element along with a CMV promoter. (CHOZN-CoV2-S) The third construct was generated to include the same components as the second construct, as well as six stabilizing proline mutations (Hsieh et al, Science 369, 2020) and a mutation of the furin site (Peacock et al, Nature Microbiology 6, 2021) to prevent cleavage. (CHOZN-CoV2-S (HP+F)) CHOZN cellscontaining a single landing pad in a locus shown to have high and consistent gene expression were transfected with the second and the third construct and stable cell lines were generated. In this step, a recombinase was used to insert a single copy of the construct with an FRT site into the landing pad.

[0153] All three cell lines, and matched negative controls, were stained with different concentrations of bamlamivimab (Jones et al, BioRxiv 9OCT20), a human IgGl with high affinity to CoV2-S. As provided in FIG. 3, CHOZN cells (CHOZN+CoV2-S and CHOZN+CoV2-S (HP+F)) showed significantly better surface expression of CoV2-S than Expi293 cells (Expi293+CoV2-S). Further, the stabilizing mutations (HexaPro, HP) and furin mutation (F) increased surface display.

[0154] Based on the studies, a construct backbone (p2G-FRT-GS) was selected for generating target-decorated cells. As provided in FIG. 4, the construct backbone includes (i) CMV promoter with 2G translation enhancing element, (ii) blue fluorescent protein stuffer (for identifying library background during cloning), (iii) a unique barcode of 20 nucleotides, (e.g., HHSWNNHHCTGGNNHHSWHH (SEQ ID NO: 114), NNNW SHHHNNHHHNNWSNNN (SEQ ID NO: 115) or similar), (iv) an FRT site with 2A- GS (for selecting for in-frame integrants into the landing pad line), and (v) ampicillin resistance and bacterial origin for growth in E. coli.

[0155] Target libraries can be introduced into the construct backbone in several ways. For small libraries, each insert can be synthesized along with unique barcode and cloned via Gibson assembly or insertion into the Hindlll / Nhel sites. For larger libraries, a backbone library with unique barcodes must be generated first by inserting a library of barcodes, possibly with gene blocks or oligos with degenerate nucleotides, into the Mlul / Nhel site. For scanning mutagenesis, a library of single point mutants can be synthesized and then cloned into the library barcode backbone via the Hindlll / Mlul sites or Gibson assembly. For other mutagenesis (e.g., random mutagenesis), a library of inserts can be generated using e.g., Error-prone PCR or OE-PCR with primers with degenerate nucleotides, then cloned into the barcoded backbone using Hindlll / Mlul. Note that the barcode library needs to be at least lOx the insert library to ensure unique barcodes per insert.7.2. Example 2: Target-decorated cell lines expressing CoV2 antigens

[0156] All CoV2 sequences available as of May 2022 were downloaded from GISAID and analyzed. The analysis revealed 25 unique clades, with 20 unique spike protein sequences. This list was further condensed to remove those with only a single amino acid difference, resulting in 17 spike sequences. These 17, along with CoVl -spike, were synthesized and cloned into the 2G-FRT-GS backbone, with unique barcodes. Spike sequences include hexapro and furin mutations for improved surface display and did not include leader sequence mutations.

[0157] DNA of all library members was pooled in equimolar ratios, transfected into the platform CHOZN line, and then selected in glutamine-deficient media for several weeks. Alternatively, clones were individually transfected, selected, and then pooled. RNA from the clones was isolated and then sequenced to confirm library distributions.7.3. Example 3: TBP-ribosome-mRNA complexes (TRM complexes)

[0158] Five mAbs were cloned into a T7-based expression plasmid with the structure provided in FIG. 5. The T7-based expression plasmid includes (i) T7 promoter and ribosome binding site, (ii) Ampicillin resistance and origin of replication for propagation in E. coli, (iii) Start codon followed by the variable light chain region (no signal peptide), (iv) Flexible linker, (G4S)4 or similar, (v) Variable heavy region, (vi) Strep tag, (vii) an unstructured spacer sequence, e.g., derived from the E.coli protein tolA or other, various lengths, (viii) a translational pause sequence derived from the E coli protein secM, (ix) absence of a stop codon, (x) an internally coded polyA region, of 40x A nucleotides, and (xi) a T7 transcription terminator, possibly several in tandem or mutants.

[0159] The whole plasmid, linearized plasmid, or PCR fragment of the plasmid (promoter through terminator) was used as a template for transcription or coupled transcriptiontranslation. NEB PURExpress kit with or without release factors, also NEB T7 quick high- yield RNA kit, and Promega TnT rabbit reticulocyte kit were used for the reaction.7.4. Example 4: Binding between Target and TBP-ribosome-mRNA complex (TRM complex)

[0160] Cells expressing spike variants (CoV2-WT, CoV2-K444T, CoV2-E484K, CoV2- F486K, and CoVl-WT) were stained with different dilutions of the TRM complexes, where each TRM complex contains an scFv of one of four mAbs (Bamlanivimab, Casirivimab, Imdevimab, or Ipilimumab). TRM bound to CoV2-S variants were detected by strep-tag staining and the results are provided in FIG. 6.

[0161] In a separate set of experiment, cells expressing the spike variants were stained with different dilutions of the TRM complexes and RNA was isolated from the stained cells and quantified using taqman RT-qPCR with antibody (TBP)-specific probes. The results are provided in FIG. 7. RNA was recovered from stained cells as expected and binding between TRM and CoV2-S was observed. CoV variants that were previously known not to bind to given antibodies showed no binding to the antibodies as provided in FIG. 6 (red lines).

[0162] In the subsequent experiments, TRM complexes generated under different cell-free translation (TL) conditions were tested. In particular, TRM complexes were generated with different concentrations of RNA templates at different TL reaction times. The data provided in FIG. 8 indicated that TRM complexes produced with a 150nM RNA template in a short reaction time (0.5hr) provide greatly reduced non-specific background, providing the largest ratio between RNA detected by TRM complex containing Casirivimab scFv (ie., specific staining) and RNA detected by TRM complex containing Ipilimumab scFv (z.e., non-specific staining).7.5. Example 5: Binding between Target and Pooled TBP-ribosome-mRNA complexes (TRM complexes)

[0163] CoV2-S antigen-expressing cell lines were individually stained with a mixture of TRM complexes. The TRM complex mixture was generated from a pool of RNA templates (150nM of templates) by in-vitro translation (at 0.5hr reaction time). RNA was isolated from the cells stained with the TRM complex mixture and quantified using taqman RT-qPCR with antibody (TBP)-specific probes. The results provided in FIG. 9 show specific binding between the target and TRM complex pairs, low background staining and no signals for variants which are known not to bind to the given antibodies (indicated in red text). Specifically, it was known that bamlamivimab (“Bam”) loses binding on E484K mutant, Casirivimab (“Casi”) loses binding to F486K mutant, Imdevimab (“Imdevi”) loses binding to K444T mutant. Ipilimumab (“Ipi”) is not expected to bind any of the spike variants (it binds CTLA4), and none of the antibodies are expected to bind the CoVl spike. These knownbinding specificities were confirmed with the target and TRM complex pairs as shown inFIG. 9.7.6. Example 6: Isolation of single cells and amplification of RNA

[0164] Five target-decorated 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 individual wells were amplified and barcoded for sequencing. The first RT-PCR reaction added adapters to the mRNA transcripts and amplified both the VH region of the antibody mRNA in the TRM complex, and a region spanning the target barcode from the cell line. The second PCR reaction added to the amplicons indexing barcodes unique to the column and the row of the 96-well plates. Samples from all the wells were pooled and sequenced on a mi-seq.

[0165] The single-cell data were analyzed for binding distribution. First, the target of the individual cell in each well was determined. Thresholds were applied to exclude empty wells and wells with more than one cell. For each well, the proportion of reads corresponding to each antibody was calculated. The Polymap scores were calculated from average proportion, scaled by average reads from cells of that target. The Polymap scores for the CoVl or CoV2 spike WT or variants (CoV2-K444T, CoV2-E484K, CoV2-F486K) were determined and provided for TRM complexes containing an scFv of one of four mAbs (Bamianivimab, Casirivimab, Imdevimab, Ipilimumab or Pembrolizumab). The Polymap scores were significantly lower on cell lines with a target expected not to bind to a given TRM complex (indicated with red outline). Negative antibodies (Ipilimumab or Pembrolizumab in red text) were at low proportion on all targets.7.7. Example 7: Single cell analysis using microfluidic system

[0166] Five target-decorated cell lines were pooled and stained with a mix of five TRM complexes. Once stained with TRM complexes, single cells were encapsulated in water-in- oil droplets with cell lysis buffer and RNA-capture beads for antibody repertoire analysis. After releasing the beads from the emulsion and washing to remove unbound RNA, the beads bound to RNA were isolated from a single cell. Since each cell expresses a single target (z.e., SARS-CoV-2 spike variant), each bead was expected to bind to RNA for the single target as well as RNA from each TRM complex that was bound to that particular cell.

[0167] Single RNA-bound beads were then encapsulated in a second emulsion for antibody repertoire 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, and then amplified with primers that enabled the target barcode and antibody VH CDR3 to be physically linked together. The reverse primerfor the antibody fragment contained a sequence complementary to a sequence on the forward primer of the target fragment. During subsequent rounds of PCR, these complementary fragments annealed, and the resulting products contained both the antibody and target sequences. FIG. 11 illustrates the overlap extension reverse transcription PCR and production of the products containing both the antibody (VH) and target (barcode) specific sequences.

[0168] Following OE-RT-PCR, the fused TBP-target sequences were further amplified with primers containing Illumina adapters and indexes for standard Illumina sequencing workflows. In the experiment, the following CDR3 and barcode sequences linked to the CoV2-S antigens were successfully amplified and sequenced.Table 27.8. Example 8: High-throughput specificity profiling of antibody clones using ribosome display and microfluidics

[0169] High-throughput methods for pair-wise interaction mapping of protein-protein interactions allows for deeper understanding of biological interactions, such as natural antibody responses. However, current techniques are limited by the number of interactions that can be screened at once. To overcome this, a polyclonal mapping system (also referred to herein as “PolyMap”) was developed. This system is based on bulk binding of a ribosome- displayed antibody library to a library of cell-surface expressed antigens, combined with single cell analysis using droplet microfluidics. In the present example, the mapping of thousands of antigen-antibody interactions was demonstrated for a set of clinically relevant SARS-CoV-2 surface antigens against a diverse library of naturally occurring anti-SARS- CoV-2 antibodies and antibodies with selective binding towards antigens of a broad range of clades were found. While developed using antibody-antigen libraries, PolyMap is well-suited for screening of other types of protein-protein interactions involving soluble ligands and their membrane bound targets, such as cytokines and immune modulatory factors and their receptors.Design of the PolyMap workflow

[0170] The PolyMap platform allows one-pot interaction screening of an antibody and an antigen library (FIG. 12). The antibody is expressed in a ribosome-display format, which uses a tethered mRNA to provide genotype-phenotype linkage of the soluble protein. Antigens are expressed at a high level on the surface of a mammalian cell, which can then be bound by the soluble ARM complex. Incubation of a library of ribosome-displayed antibodies with a library of antigen-expressing cells allows sampling of all possible antibody and antigen combinations, in a single bulk step. To elucidate antibody-antigen interactions, cells stained with ARM complexes are washed and individually encapsulated in microdroplets, which contain lysis reagent and uniquely-barcoded RNA capture beads. A series of molecular biology steps are used on the isolated beads to generate barcoded cDNA of linked antibody and antigen sequences. The barcoded transcripts are then read and quantified by deep sequencing and analyzed using bioinformatics to map antibody-antigen pair-wise binding specificities.Establishing the antigen expression system

[0171] This section relates to the establishment of a robust antigen surface display platform. In particular, the platform should have robust surface expression of a single antigen per cell, a unique barcode for antigen identification, and a simple and efficient cloning and cell line generation process amenable for use with libraries. A series of mammalian expression vectors were evaluated to determine the impact of different signal peptides and transmembrane regions on surface expression of a chosen target, the SARS-CoV-2 surface antigen (CoV-2 S; data not shown). The native signal peptide and transmembrane region were retained as they performed at least as well as the alternatives. A 19 amino acid C-terminal truncation was implemented to remove an endoplasmic reticulum retention signal endogenous to CoV-2 S.

[0172] Stable cell lines expressing CoV-2 S were generated with two different methods, namely (1) lentiviral transduction into Expi293™ cells or (2) flippase (Flp) recombinase- mediated integration into CHOZN® cells engineered with a Flp recognition target (FRT) landing pad for single-site integration. After sorting or selection of the cell lines, respectively, staining for CoV-2 S surface expression revealed a much higher relative signal from the CHOZN cell line (FIG. 13). A modified construct which included six stabilizing proline mutations and a furin site mutation to abrogate fusogenic activity of the spike protein was also evaluated and found to further increase cell surface expression of the antigen (FIG. 13).

[0173] The final antigen display vector uses a CMV promoter with a translation-enhancing sequence element (2G UNic®, ProteoNic) to drive efficient antigen expression, a signal peptide, a transmembrane domain for surface display, the aforementioned prolinemodifications, and a mutated furin site (FIG. 14 A). A 20-mer nucleic acid barcode, unique for each antigen, is included after the translational stop codon, followed by a common flanking sequence to allow PCR amplification and antigen identification with short-read sequencing. Restriction enzyme sites are included for vector linearization and facile ligation of the antigen library.

[0174] To generate the library of antigen-displaying cells, the antigen-encoding plasmid library was transfected along with a plasmid expressing the Flp recombinase into a previously engineered CHOZN cell line with the FRT site integrated into a high-expressing genomic locus. This ensures single-copy integration and normalized expression levels across individual antigens. Successful integration of the plasmid results in expression of glutamine synthetase, allowing selection in glutamine-free media.

[0175] For a proof-of-concept study, the ancestral Wuhan (termed wild-type for this study; WT) CoV-2 spike protein and three individual point mutants (K444T, E484K, F486K) known to significantly disrupt binding of specific monoclonal antibodies in clinical development were selected. The four spike protein coding sequences, including the native leader sequence, stabilizing mutations, native transmembrane region and deleted cytoplasmic tail, were cloned into the platform vector with unique barcodes. In this instance, the plasmids were separately transfected into the platform CHOZN cells, which were then selected and stained for spike surface expression using a cross-reactive antibody. Flow cytometry analysis showed high and consistent expression between the four CoV-2-S variants (FIG. 14B).Establishing the ribosome display system

[0176] The ribosome display construct includes the fundamental features reported earlier by others with some more recent enhancements and features specific to the instant application (FIG. 14C). A T7 promoter drives expression of the scFv gene, which is followed by a strep- tag II sequence. The absence of a stop codon and addition of a 17-amino acid SecM stall sequence prevents the ribosome from dissociating from the mRNA, physically tethering the newly synthesized protein to its coding sequence. An encoded poly-A tail allows subsequent capture onto beads for cDNA synthesis, barcoding, and sequencing. The CDR3H sequence provides a unique identifier for the antibody, and as for the antigen plasmid, common cut sites and flanking regions are included for library cloning, amplification, and sequencing. The whole cassette is maintained within a pUC-based plasmid.

[0177] To generate ARM complexes, a DNA fragment stretching from the T7 promoter through the transcriptional terminator region is amplified with PCR and used as a template to generate large amounts of RNA. While it is possible to use circular plasmid as atranscriptional template, it was found that a PCR fragment with a defined terminus was important for RNA uniformity, as otherwise significant terminator read-through occurred even with constructs including improved terminator designs (data not shown). It was determined that using an in vitro expression system based on recombinant components and starting with a fixed input of purified RNA produced more consistent ARM complex yields than a coupled transcription-translation process, possibly due to varying transcriptional kinetics between variants. It was also determined that omitting the ribosome release factors from the in vitro translation reaction and using a short 10-minute translation time further improved the stability and specificity of the ARM complexes (FIG. 15).

[0178] For initial testing, the clinical SARS-CoV-2 monoclonal antibodies bamlanivimab, casirivimab, and imdevimab were selected, with known susceptibilities to the spike point mutations E484K, F486K, and K444T, respectively. Bio-layer interferometry (BLI) was used to confirm the binding affinities (and the lack of binding to the specific point mutations) of these antibodies presented as scFvs to each spike mutant expressed as a soluble receptor- binding-domain (RBD)-Fc fusion (Table 3, FIG. 16). The scFvs were then cloned into the ribosome display construct, produced as ARM complexes, and individually incubated with antigen-expressing cells. Total RNA was isolated from the stained cells, which was then analyzed with RT-qPCR to quantify the antibody mRNA recovery. For each antibody, there was complete or partial loss of RNA to the expected spike variants consistent with the measured binding affinities (FIG. 14D). Although there was some variation in RNA recovery between the different antibodies, possibly due to differences in translation efficiency, the concentrations reported here represent 1-4% RNA recovery, which is consistent with previous reports.Table 3Measured binding affinities of clinical antibodies to soluble spike mutants by BLI. *, Off-rate below the limit of detection.Proof-of-concept testing using clinical stage anti-SARS-CoV-2 antibodies

[0179] Preliminary testing was performed using a “mini library” of the 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, an equimolar RNA mix was used to generate ARM complexes, which were then diluted to a concentration equivalent to 15 nM RNA and incubated with each antigen expressing cell line separately. Total RNA from the stained cells was amplified and sequenced, with the CDR3H sequence used to identify each antibody. The percentage of the total reads assigned to each antibody was tabulated for each antigen cell line. As expected, there were very few reads for the expected dropouts, and low reads for both negative control antibodies across all samples (FIG. 17A). Surprisingly, an outsized proportion of the reads were from casirivimab, although this is consistent with the higher RNA recovery from the initial RT-qPCR result for this antibody (FIG. 14D).

[0180] This experiment was repeated with the four antigen cell lines mixed together and stained in bulk with the mini antibody library. To pair the antigen and antibody data, single cells were sorted into a 96-well plate and lysed. The antigen barcode and antibody heavy chain were amplified with arrayed primers containing column and row barcodes to allow unique identification of the amplicons from each individual well. After identifying the antigen of each cell, the distribution of antibodies bound to that cell was calculated. As seen previously when staining the individual cell lines, strong expected CoV-2 mutant dropouts, low background binding, and reproducibility between individual cells was observed (FIGS. 17B, 17C).

[0181] Although single-cell sorting and arrayed PCR are feasible for small sample sets, a single-cell droplet-based approach combined with high-throughput sequencing would allow for analysis of larger sets of antibodies and / or antigens. For proof-of-concept testing, the previously published microfluidic droplet technology Drop-seq was utilized, where single cells are co-encapsulated with barcoded beads (ChemGenes Corporation) in nanoliter droplets using a microfluidic chip and syringe pump set-up. The same antibody RNA mixture and four cell lines were used, though with a slightly modified staining strategy with the ARM complexes diluted to a concentration equivalent to 4.5 nM RNA (this concentration is not expected to impact the results, as a test of ARM complex binding to CoV-2 WT at two different concentrations showed the same level of enrichment for many antibody clones; FIG. 18). The Drop-seq workflow through the cDNA amplification step was performed as previously described (https: / / mccarrolllab.org / dropseq / ), and the antigen barcodes and antibody heavy chain sequences were amplified using gene-specific primers with Illumina adapters for sequencing. Individual cells were first identified by the Drop-Seq barcode, then the antigen identity was determined by the antigen barcode; 91 cells were identified for mapping antibody-antigen interactions (FIGS. 17D, 17E). The Drop-seq results wereconsistent with both the individual cell line staining (FIG. 17A) and the single-cell sorting (FIGS. 17B, 17C).Testing an anti-SARS-CoV-2 immune repertoire against a larger set of spike variants

[0182] PolyMap was next applied to a larger library of antibodies and antigens. All publicly available CoV-2 S sequences (as of August 2023) were downloaded from GISAID.org, and 16 unique sequences representing phylogenetically distinct clades (nextstrain.org) were selected (FIG. 19, SEQ ID NOs: 126-138, 140, 142, and 143). These sequences and control antigens (CoV-1 S, CTLA-4, PD-1, and blue fluorescent protein [BFP]) were cloned with unique barcodes into the antigen backbone then individually transfected into the CHOZN landing-pad line. Next, a diverse library of natively paired scFvs were cloned into the ribosome display vector. This library was previously isolated from convalescent COVID-19 donors and enriched for binders to the ancestral spike RBD protein. Sequencing of this library showed a broad variant distribution with 99% of the sequences coming from 300 unique clones, with the top clone making up -6.9% of the library (FIG. 20A).

[0183] For Drop-seq microfluidic analysis of the larger libraries, cells expressing each CoV-2 S sequence variant were pooled together and then stained with ARM complexes of the aforementioned anti-CoV-2 antibody library. After identifying antigen and antibody sequences with Drop-seq cell barcodes, a merged list of every antibody and antigen associated with each cell barcode was generated, ultimately analyzing 2,783 cells across the 20 antigens (FIG. 21). As each antibody has equal opportunity to bind each antigen, the total antibody reads for each clone across all cell lines were counted, then the percentage of reads for each antibody was normalized by the number of associated cells. When examining the top 40 antibodies based on the total number of antibody reads, differential binding patterns were observed across the antibody sequences (FIG. 20B). Many of the antibody clones showed minimal enrichment to the Omicron spike variants (2 IK, 22A / D, 23B), which was expected as this library was generated from donors convalescent from the original ancestral WT strain in early 2020. Reduced binding to 20H (Beta), 20J (Gamma), CoV-1, and the negative control cell lines (PD-1, CTLA-4, and BFP) was also observed.

[0184] In this study, when the CoV-2 antigen cell lines were made, antigen sequences were transfected into cells individually and then pooled for the PolyMap runs. Alternatively, multiple antigen sequences can be transfected into cells in bulk. To examine the feasibility of this approach, 22 antigen sequences were transfected together, including 18 unique CoV-2 S sequences representing phylogenetically distinct clades (16 sequences previously used and shown in FIG. 19, as well as 21M Omicron (SEQ ID NO: 139) and 22C Omicron (SEQ IDNO: 141)) and 4 control antigens (CoV-1 S, CTLA-4, PD-1, and BFP). These antigen sequences were cloned with unique barcodes into the antigen backbone, then transfected en masse into the CHOZN® landing-pad line. Sequencing of the genomic DNA (gDNA) of the bulk-transfected line showed representation of all variants with only modest variation in frequency (FIG. 24). This method of bulk transfection allows for the efficient generation and testing of large antigen libraries.Validation of PolyMap results

[0185] With the individual cell lines for each antigen available, each line was separately stained with the ARM complex library. RNA was isolated and sequenced as before, and the proportion of reads from each antibody was calculated. Due to the skewed distribution of the antibody library, it was important to look at enrichment compared to the input sample rather than the raw antibody read proportions. The log2 fold change of this enrichment value was used to better differentiate enriched versus de-enriched sequences and is shown for the top 40 most enriched clones (FIG. 22A). Again, many clones de-enriched for binding to the 20H (Beta), 20J (Gamma), and the Omicron variants were observed.

[0186] Eleven antibody clones (termed Ml -Ml 1) found within the top 100 sequences and with various antigen binding patterns were selected for validation. Several had been previously isolated as individual CHOZN cell lines, and for those a small production run was performed. For the remaining antibodies, a second sequencing run on the input library was used to determine the VL sequence and CDR3H, which was then paired with the previous VH data to assemble full-length sequences for expression. These variable regions were cloned into separate expression vectors and expressed transiently in HEK293 cells.

[0187] Binding of these antibodies was validated by staining each spike variant cell line with antibody-containing supernatant diluted to a concentration that is expected to be in the linear sensitivity range. Binding of antibody was detected by a secondary antibody specific to a human Fc and then analyzed by flow cytometry. The median fluorescence intensity (MFI) of the data was compared with both the Drop-seq (FIG. 20B) and single-antigen staining (FIG. 22A) PolyMap data. In general, the distinctive binding patterns of each clone from Drop-seq were closely matched by the single-antigen and flow staining data (FIG. 22B). One interesting note is that as the single-antigen stain is normalized within a single cell line, the enrichment values can provide some insight on relative binding affinity between antibodies (FIG. 23). For example, Clone Ml shows good alignment between all data sets (FIG. 22B), but it may have lower affinity due to the lower enrichment in the competitive binding scenario of the WT antigen. This may also explain a few false positives from the single-antigen staining for Omicron variants, as there were fewer competing sequences. Clone MIO may be an example of a sequence with very high translation or PCR efficiency, which led to strong enrichment using the single-antigen staining approach.

[0188] Lastly, the sequence diversity of recovered antibodies and their binding patterns was further explored. As expected, few clones were found that retained binding to the highly- mutated Omicron spike in these early patient samples. Several repeated dropout patterns were observed, which correlate with shared mutations in the RBD. For example, mutations in K417 are seen in Beta, Gamma, and Omicron (2 IK) onward (FIG. 19), and is a likely epitope for Ml, M2, M3, M6, M7, M9, and MIO. The variants Beta, Gamma, Kappa, Eta, Iota, and Mu onward contain an E484 mutation, which is a likely epitope for a clone like M5. Clones like M8 and Ml 1 show binding restricted to only the earliest clones and may recognize epitopes that include the above amino acids and others like L452 or N501. Finally, the clone M4 shows broad binding and some reactivity to Omicron, and may bind to a more conserved epitope.Discussion

[0189] A solution was sought for pair-wise interaction mapping of an antigen library and an antibody library without some of the limitations of existing technologies. For this purpose, PolyMap was developed, a platform which probes a library of mammalian cell-surface expressed proteins with a soluble antibody library. This approach was developed and validated using a small matrix of clinical SARS-CoV-2 antibodies with known spike mutational susceptibilities. PolyMap was then applied to an uncharacterized immune library from convalescent COVID patients, leading to the discovery of antibodies with several distinct binding patterns across circulating SARS-CoV-2 spike variants.

[0190] A significant strength of the PolyMap platform is a flexible antigen expression system tailored for the expression of membrane proteins. Unlike other published multiplexed mapping methods, which require purification, in vitro translation, or chimeras for yeast display. PolyMap uses full-length antigens displayed on the surface of mammalian cells. The robust quality control mechanisms in the cell ensure native structure and glycosylation of complex membrane-bound proteins, which is critical for many therapeutic targets such as GPCRs or oligomeric viral antigens. A key component of the antigen expression system is the parental CHOZN cell line, which features an engineered FRT site at a highly expressed locus for recombinase-mediated integration. This platform line has been used extensively for one-pot transfections and manufacturing of thousands-diverse antibody libraries. Other methods for generating large antigen or antibody libraries in mammalian cells are alsocompatible with PolyMap, though uniform clonal distribution and even expression levels are beneficial.

[0191] The antibody library was expressed as scFvs in a ribosome display format, which has been previously validated and has several advantageous features. First, the soluble antibody library allowed staining of suspension cells in bulk, which is believed to have previously only been described for adherent cells. Second, although the ARM complex can potentially harbor several translating ribosomes, a short 19 amino acid C-terminal linker was used to promote a single full-length antibody on each ribosome. This functionally monovalent format was critical for preventing cell aggregation and allowed processing the stained cells through a microfluidic system. Monovalent binding was also important to eliminate avidity effects and allow for differentiation of smaller changes in binding affinity. Other advantages of ribosome display include ease of library generation, compatibility with extremely large library sizes, and commercially available reagents.

[0192] The PolyMap screening process is simple and scalable, and both antigen and antibody libraries can be reused indefinitely once generated. Unique to PolyMap, each cell supports hundreds to thousands of interactions, meaning that the available screening capacity is mostly defined by the number of antigen-expressing cells that can be individually processed. If there are only a small number of antigen variants to be assayed, individual cell lines can be generated, stained, and directly subjected to CDR3H sequencing without single-cell manipulations (FIGS. 17A and 22A). For studies including tens, hundreds, or more antigens, en masse transfection of a plasmid library into cells is convenient and presents an option not available with some alternative technologies. Drop-seq, as used here, can typically process up to 10,000 single cells per hour and is appropriate for studies involving dozens to potentially hundreds of unique antigens. Other single-cell and microfluidic sequencing technologies are also compatible with PolyMap and could be explored in future iterations.

[0193] Although the data processing method is relatively simple, there are several caveats that are important to understand when interpreting the data. For the Drop-seq analysis, the read distribution of a single antibody across all antigens was evaluated. In this approach it is easy to see binding patterns within each antibody, but harder to understand how different antibodies might bind relative to one another. In the single-antigen staining approach, reads are normalized within each antigen, which can allow for clear identification of top binders to each variant. The consideration here is that if the antibody distribution is skewed, it is critical to normalize to the input distribution, and as the total number of antibodies binding to a specific antigen decreases, the proportion of reads attributed to the remaining antibodies willlook higher. In this case, the rare binders to the antigen may not be particularly high affinity, but they could be good candidates for further engineering.

[0194] PolyMap was applied to an enriched antibody library derived from convalescent COVID donors and a set of known spike variants, resulting in some noteworthy observations. Several distinctive clonal binding patterns were observed, which were shared across several antibodies (with some of these interaction patterns confirmed using full-length antibodies). Though confirmatory experiments were not performed, examination of the spike amino acid sequences (FIG. 19) showed distinct sequence variants associated with the different binding patterns, potentially indicating key amino acids of the antibody epitopes. Interestingly, despite the antibodies being derived from donor samples collected in April 2020, strong binding was found to all later SARS-CoV-2 variants up to and including Lambda, which did not become a variant of interest to the WHO until June of 2021. Several monocl onals with weak binding to the highly mutated Omicron variants were found, which may target conserved epitopes of potential value for future therapeutic development. Finally, as one strategy to provide broad coverage is to target several non-overlapping epitopes, PolyMap data could help select antibodies with possible functional synergies.

[0195] Thoughtful selection of antibody and antigen libraries can enable a broad range of applications for the PolyMap platform. Pairing PolyMap with antibody-repertoire capture techniques can allow profiling of an immune response across donors, across treatments, or across time. Libraries of human surface proteins or other target protein populations can be amplified from cDNA and generated in bulk for antibody specificity profiling. As both the antigen and antibody platforms are compatible with any library generation technique, synthetic libraries can be generated for protein engineering studies. For example, a deep mutational scanning library of an antigen could be generated and screened with an antibody library for a massively parallelized epitope mapping platform. Targeted CDR mutagenesis and a panel of antigen expressing cells could be used to evolve antibodies with either broad or very targeted specificity. An interesting extension of the current method could involve engineered cells with a ligand-induced activatable phenotype. In this instance, ARM stained and activated cells could be sorted based on marker activation and subsequently subjected to PolyMap for identification of the specific binding partners. Furthermore, in principle the technology can be adapted to any library of soluble proteins and their cognate binding partners expressed on the cell surface, such as cytokines and their cell-bound receptors, or immune checkpoint proteins.MethodsAntigen constructs and cloning

[0196] The amino acid sequence for wild-type (WT) SARS-CoV-2 S protein (UniProt P0DTC2) was modified for improved expression: 6 stabilizing proline mutations were added, the furin site was removed (RRAR (SEQ ID NO: 120) ->AGAG (SEQ ID NO: 121)), and 19 amino acids were truncated from the C-terminus. The sequence was codon-optimized, synthesized, and cloned into the antigen expression vector (FIG. 14A) along with an 18-mer barcode sequence. The expression vector includes a CMV promoter with a 2G-UNic® translation enhancing element (ProteoNic) as well as a FRT site followed by a 2A ribosomal skipping motif and glutamine synthetase gene. Point mutations were introduced into the WT spike sequence by Gibson assembly and cloned with unique barcodes into the same expression vector backbone.

[0197] To generate the circulating spike variant library, the protein sequences for all available spike variants were retrieved from GISAID.org as of August 2023. The phylogenetic analysis for 4,217 SARS-CoV-2 genomes from nextstrain.org that were sampled between Dec 2019 and July 2023 were also retrieved. The most common sequence within each phylogenetic clade was selected and filtered for unique sequences. 16 representative spike variants covering a diverse range of mutations were selected (FIG. 19). These sequences, including the proline and furin mutations, and the C-terminal truncation, were synthesized with unique barcodes (HHSWNNHHCTGGNNHHSWHH, SEQ ID NO: 114) and cloned into the antigen expression vector. Wild-type sequences for CoV-1 S (P59594), human CTLA-4 (Pl 6410), human PD-1 (QI 5116) and mTagBFP2 were included as controls.Stable cell line generation and analysis

[0198] Stable cell lines were generated by transfecting each individual antigen construct along with a plasmid encoding the Flp recombinase into CHOZN® cells containing a matched FRT site and landing pad. Cells were electroporated (MaxCyte) using the default CHO protocol with 5 pg of a 4: 1 mixture of antigen and recombinase plasmid per 106cells. After recovery for 4 days in EX-CELL® CD CHO Fusion media (Millipore Sigma, 14365C) supplemented with GlutaMax (ThermoFisher, 35050061), cells were selected in the absence of GlutaMax for 18 days or until viability was >95%. After selection, gDNA was collected and the barcode region amplified using flanking primers containing sequencing adapters (Antigen BC P5 FWD, Antigen BC P7 REV; Table 4 for primer sequences). The DNA samples were quantified using qPCR and normalized to 1.8nM. The library was diluted to a final concentration of 9 pM, supplemented with 5% PhiX DNA and sequenced on an IlluminaMiSeq instrument using custom primers (M13 SEQ, BGH INDEX SEQ, BGHSEQ). Sequencing results were analyzed using Python scripts that calculated the number of times each unique antigen barcode appeared per library and the frequency relative to the total number of barcodes.Flow Cytometry

[0199] To assess spike protein expression, cells were stained with 10 pg / mL anti-COVID-19 & SARS-CoV S glycoprotein antibody (clone CR3022, Absolute Antibody, Ab01680-10.0), followed by PE or APC anti-human IgG (H+L) Antibody (Jackson ImmunoResearch, 709- 606-149). For binding validation of specific antibodies identified by PolyMap, spike variant expressing cell lines were stained with supernatant from CHO cells expressing the indicated antibody clone at 0.5 pg / mL, followed by APC anti-human IgG Fc Antibody (BioLegend, 409306). After secondary antibody staining, cells were stained with 4',6-diamidino-2- phenylindole (DAPI; BioLegend, 422801). The samples were analyzed on a CytoFLEX LX flow cytometer (Beckman Coulter) and the mean fluorescence intensity (MFI) of the APC signal on live cells (DAPI negative) was determined in FlowJo (vlO.6.1, BD Biosciences). Antibody Constructs and Cloning

[0200] Antibody scFv sequences (Vr-linker-Vu) were assembled, codon optimized, and synthesized with overlaps for Gibson assembly into the ribosome display vector. Instead of a stop codon, the vector encodes a strep-tag II, an unstructured sequence derived from tolA (GGQKQAEE, SEQ ID NO: 122), and a secM stall sequence (FSTPVWISQAQGIRAGP, SEQ ID NO: 123). Transcription from the plasmid is controlled by flanking conventional T7 promoter and terminator elements.

[0201] Antibody libraries were amplified from previously generated natively-paired antibody repertoires from convalescent COVID patients using a primer pool (scFv library FWD 1-5, REV 1-3; Table 4) designed to capture all germlines. These sequences were cloned in bulk into the linearized ribosome display plasmid using Gibson assembly. For sequencing, samples were prepared by amplifying the VH region directly from the plasmid library or from cDNA with universal primers binding to the linker and secM regions (scFv linker 1 P5 FWD, scFv P7 REV).ARM complex generation and cell staining

[0202] The ribosome display plasmid was subjected to PCR to generate a fragment spanning the T7 promoter, scFv cassette, and the T7 terminator sequence (primers ARM complex FWD, ARM complex REV; Table 4). The PCR product served as template in an in vitro transcription reaction (HiScribe® T7 Quick High Yield RNA Synthesis Kit, NEB, E2050S)to generate RNA which subsequently was DNase treated and purified. To generate ARM complexes, 150-450 nM of RNA template (based on 1100 bp length) was added to an in vitro transcription-translation kit without release factors (PureExpress® ARF123, NEB, E6850S) and incubated at 37°C for 10 minutes. The reaction mix was immediately cooled on ice and diluted 1 :5 or more with RBTH buffer (20 mM HEPES, 50 mM MgC12, 50 mM NaCl, pH 7.4, 0.05% BSA, 2.5 mg / mL heparin, 0.01% tween-20). l * 106antigen-expressing cells were stained with 50 pL of diluted ARM complex on ice for 30 minutes, then washed 3x with 1 mL of ice-cold RB (20 mM HEPES, 50 mM MgCh, 50 mM NaCl, pH 7.4 + 0.05% BSA) before further analysis.RNA isolation and qPCR

[0203] For qPCR analysis of ARM complex binding after staining antigen-expressing cells (FIG. 14D), the lysis solution and stop solution from the Cells-to-CT™ 1-Step Power SYBR™ Green Kit (ThermoFisher, A25599) was used. In a 96-well plate, 10,000 stained cells per well were lysed in 20 pL lysis solution for 30 min at room temperature, followed by addition of 2 pL stop solution and incubation for 2 min at room temperature. The lysates were probed for secM (using primers ARM qPCR FWD, ARM qPCR REV, ARM qPCR PROBE; Table 4) in a RT-qPCR assay. The TaqMan™ RNA-to-CT™ 1-Step Kit (ThermoFisher, 4392938) was used according to manufacturer’s instructions. 2 pL of each lysate was distributed into 30 pL reaction mixes (15-fold dilution); each reaction was aliquoted in quadruplicates into a 384-well plate, 6 pL per well. A 7-point standard curve using RNA of a known concentration was included, starting from 1 nM to 1 fM (10-fold serial dilutions).Single-antigen staining

[0204] An equimolar mix of ribosome display RNA for bamlanivimab, casirivimab, imdevimab, ipilimumab, and pembrolizumab was used to generate ARM complexes as described above, then diluted to 15 nM input RNA and added to 4 individual antigen lines (WT, K444T, E484K, F486K; FIG. 17B). For the library-by-library experiment (FIG. 20B), ARM complexes were generated using the COVID-19 antibody library and were used to stain cells at 4.5 nM input RNA. After washing, total RNA was collected and antibody VH amplified using universal primers that add sequencing adapters and Illumina barcodes for demultiplexing (scFv linker 1 P5 FWD, scFv linker 2 P5 FWD, scFv P7 REV; Table 4). The resulting libraries were loaded at 1.8 nM onto a 500-cycle Reagent Nano Kit v2 (Illumina, MS-103-1003) and run on a MiSeq with a 255 x 255 read length scheme using custom primers (M13 SEQ, scFV REV SEQ, scFv INDEX SEQ).Single-cell sorting

[0205] An equal mix of 4 antigen lines was stained and washed as above, then single live (DAPI negative) cells were sorted into individual wells of a 96-well plate containing RNAse inhibitor and 2x RT-PCR master mix (SuperScript IV One-Step RT-PCR system, ThermoFisher). After a freeze-thaw cycle, additional reagents were added: master mix, RT, and 2 pairs of primers that amplify the antigen barcode or antibody VH and add a universal PCR adapter (SCS scFv FWD, SCS scFv REV, SCS Antigen FWD, SCS Antigen REV; Table 4). 3 pL of the RT-PCR product is used in a second PCR (NebNext Q5 Ultra II, NEB) with forward primers that bind the adapter and add a plate and one of 8 row barcodes, and antigen- or antibody-specific reverse primers with one of 12 column barcodes (SCS P5 FWD, SCS scFv P7 REV, SCS Antigen P7 REV). The primers are arrayed across the plate so each well gets a unique combination of column and row barcode, allowing single-well identification. The resulting library was loaded at 1.8 nM onto a 500-cycle Reagent Nano Kit v2 (Illumina, MS-103-1003) and run on a MiSeq with a 255 x 255 read length scheme.Drop-Seq

[0206] The Drop-seq encapsulation, cDNA generation, and amplification steps were carried out as previously described (https: / / mccarrolllab.org / dropseq / ), with some modifications. Since the cell barcodes allowed distinguishing droplets containing multiple cells, a higher cell and bead density was used to maximize the chance of co-encapsulation. Cells were resuspended in RB buffer with 12% OptiPrep™ (Sigma-Aldrich, D1556-250ML) at 500 (FIG. 17D) or 400 cells / pL (FIG. 20B), filtered through a 35 pm cell strainer, then a 20 pm cell strainer, and centrifuged at 30xg for 1 min to remove any remaining large clumps of cells. The supernatant was carefully aspirated without disturbing the bottom ~30 pL and loaded into a 1 mL syringe. Drop-seq beads (generated by and purchased from ChemGenes Corporation, Macosko-201 l-10(V+)) were washed according to the manufacturer’s instructions and resuspended in 10 mM Tris pH 8.0, 1 mM EDTA, 0.01% Tween at 120 or 220 bead / pL. Prior to 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 240 beads / pL (FIG. 17D) or 440 beads / pL (FIG. 20B).

[0207] Microfluidic chips for Drop-seq were purchased from FlowJEM. Pico-Surf 1 (Sphere Fluidics, C023) was used for the droplet generation oil. During encapsulation, syringe pumps were operated at flow rates of 900 pL / hr for each aqueous suspension and 1600 pL / hr for the oil. Droplet breakage was carried out as previously described, but 2 mL Pico-Break (Sphere Fluidics, C082) was used to break the droplets.

[0208] Reverse transcription and exonuclease treatment was carried out as previously described, with the exception of a modified template switch oligo (Table 4). After RT and exonuclease treatment, bead concentration was determined using a Fuchs-Rosenthal cell counter (Bulldog Bio, DHC-N01). Aliquots of 8,000 beads were PCR amplified in a volume of 50 pL using lx HiFi HotStart ReadyMix (Kapa Biosystems, KK2602), 0.4 pM each cDNA PCR FWD and cDNA PCR REV primers. For FIG. 17D, three aliquots were amplified; for FIG. 20B, 10 aliquots were amplified.

[0209] To prepare the samples for sequencing, 5 ng (FIG. 17D) or 20 ng (FIG. 20B) of cDNA library was PCR amplified in a volume of 50 pL using lx Q5 High-Fidelity Master Mix (NEB, M0494S), 0.5 pM forward primer, and 0.5 pM reverse primer. The forward and reverse primers were designed to selectively bind and amplify the antibody and antigen sequences in the cDNA library. The primers used for antibody amplification have reversed P5 / P7 Illumina adapters (Table 4): the reverse primers (Drop-seq P5 REV) contain the P5 Illumina adapter and a unique Illumina index for de-multiplexing; the forward primers (scFv linker 1 P7 FWD for FIG. 17D and scFv linker 2 P7 FWD for FIG. 20B) contain the P7 Illumina adapter. The primers used for antigen amplification were CoV-2 S FWD for FIG. 17D and Antigen BC FWD for FIG. 20B; the reverse primers were Drop-seq P7 REV with different Illumina indices for de-multiplexing.

[0210] The antibody and antigen samples were sequenced separately. The antibody samples were sequenced on the Illumina MiSeq using a 600 cycle MiSeq Reagent Kit v3 (Illumina, MS-102-3003), while the antigen samples were sequenced using a 500 cycle MiSeq Reagent Nano Kit v2 (Illumina, MS-103-1003). For antibody sequencing, read 1 was 26 bp (Cell BC SEQ), the minimum requirement for good clustering on the MiSeq (bases 1-12 cell barcode, bases 13-20 UMI); index 1 was 150 bp to capture the antibody CDR3 (Strep-tag II SEQ); read 2 was 6 bp to read the Illumina barcode (Illumina BC SEQ). For antigen sequencing, read 1 was 270 bp (Ml 3 SEQ primer) for FIG. 17D, and 76 bp for FIG. 20B; index 1 was 6 bp to capture the Illumina barcode (Illumina BC SEQ); read 2 was 20 bp (bases 1-12 cell barcode, bases 13-20 UMI; Cell BC SEQ).Antibody repertoire sequencing analysis

[0211] Sequencing analysis was performed as previously described. Briefly, the expected number of errors (E) for a read was calculated from its Phred scores and reads with E >2 were discarded. IMGT immunoglobulin sequences were processed to generate position-specific sequence matrices (PSSMs) and to identify framework / CDR junction for each of the nucleotide sequences. Reads were required to have a valid predicted CDR3H sequence.Antibody clones were then defined conservatively, combining unique sequences if they had 1 amino acid difference for 5-6 amino acid long CDR3H, and 1-2 amino acid differences for >6 amino acid long CDR3H. Only clones with at least two sequencing reads were included in the analysis.Single-cell sorting sequence analysis

[0212] Custom Perl scripts were used to identify the plate, row, and column barcodes, along with the antigen or antibody identity of each sequencing read from Fastq files. To assign the antigen identity of each well, only wells with >10 total antigen reads were considered. To exclude wells with multiple cells, only wells where the antigen with the most reads represented >90% of the total antigen reads in a given well were considered. Further filtering for wells with >600 total antibody reads was carried out. Percent antibody read for each antigen was visualized as heatmaps / scatterplots using ggplot2 version 3.4.2 in R.Single-antigen staining sequencing analysis

[0213] A read count table of antibody clones across all antigen lines was generated. The percent read for each antibody within each antigen cell line was calculated. Differential percent antibody read between the antibody composition was then calculated for each antigen line compared to the input antibody library, followed by calculation of log2 of the foldchange (log2fc). Within each antigen, if the log2fc value for an antibody was less than the log2fc for any of the negative control antibodies, then the log2fc for that antibody was set to NA. For visualization, the top 40 antibodies by sum of log2fc across all antigens were used. The percent antibody reads and the log2fc values relative to the input antibody library were visualized as heatmaps using ggplot2 version 3.4.2. NA values are colored white.Drop-Seq sequencing analysis

[0214] Using custom Python scripts, the Drop-seq cell barcode, unique molecular identifier (UMI), and antigen barcode or antibody CDR3H were identified for every sequencing read. Only reads with unique UMIs were considered. To determine the antigen identity, it was required that each cell contained >10 antigen reads and >90% single antigen to filter out droplets containing multiple cells. The total number of antibody reads was determined for each cell. The percentage of antibody reads for each antigen was then normalized by the number of cells. For the mini library experiment (FIG. 17D), 243 unique antigen cells were identified and 473 cells with associated antibodies were identified. Merging these data sets resulted in 91 overlapping cell barcodes and allowed mapping of antibody-antigen interactions. For the library-by-library experiment (FIG. 20B), 2,783 total cells with antigen identities were identified. The total reads for each antibody clone across all antigen cell lineswere counted, and then the percentage of reads for each antibody was normalized by the number of associated cells. The normalized percent reads were visualized as heatmaps using ggplot2 version 3.4.2.Protein expression

[0215] The receptor-binding domain of spike, with K444T, E484K, or F486K mutations, was cloned in-frame with the hinge and Fc of a murine IgG2a in a mammalian expression construct. 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.

[0216] VH and VL sequences of select monoclonal antibodies were cloned into separate vectors containing the human IgGl or human Kappa constant regions. Plasmids were transiently transfected in HEK293 cells and the antibodies provided as harvested cell culture fluid (Twist Bioscience).

[0217] Some monoclonal antibodies were cloned with human IgGl or kappa constant regions into a stable expression vector containing the FRT and GS gene compatible with the landing pad cell line. These were transfected and selected as above to generate stable antibodyexpressing lines. A production run was performed as previously described and supernatant harvested for binding studies.Affinity measurement

[0218] Bio-layer interferometry (BLI) kinetic affinity measurements were carried out on a GatorPrime Instrument (GatorBio). ScFv was generated using an in vitro transcriptiontranslation kit (PURExpress, NEB, E6800). 5 ug / mL biotinylated strep-tag II capture antibody (GenScript, A01737) was bound at <50% saturation to Gator streptavidin coated probes (GatorBio, 160002) for 30 seconds prior to loading 10-fold diluted scFv reaction mix for 600 seconds. Association and dissociation data were collected for each scFv against each RBD variant for 180 and 300 seconds, respectively. RBD antigen concentrations were titrated to 100, 33, 11, and 0 nM and run on a multicycle kinetic assay and kinetics were analyzed with a double reference strategy, as previously described and fitted to a global 1 : 1 kinetic model with Rmax Unlinked. 30 second baseline measurements proceeded the capture probe, scFv loading, and association steps.Lentiviral transfection and stable cell line generation

[0219] The amino acid sequence for CoV-2 S WT (UniProt P0DTC2) was cloned with a human IgK signal peptide, c-myc tag, and PDGFR transmembrane region into a plasmid with the EFla promoter and 5’ and 3’ lentivirus LTRs. This plasmid was transiently transfectedalong with the Rev, Gag / Pol, and VSV-G envelope plasmid into HEKTa cells for lentiviral packaging. Supernatant containing lentivirus 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 for c-myc expression.Spike surface expression check

[0220] Cells were stained with a dilution series of bamlanivimab and anti-hlgGl-Fc-PE, washed, and analyzed by flow cytometry (CytoFLEX LX flow cytometer, Beckman Coulter). Median fluorescence intensity of the PE signal is reported.In vitro translation optimization

[0221] A 1 :1 mix of RNA encoding a positive (casirivimab) and control (ipilimumab) scFv was added at 450 nM to an in vitro transcription-translation kit with (PURExpress, NEB, E6800) or without (PURExpress ARF123, NEB E6850) ribosome release factors, and incubated at 37°C for 10 or 30 minutes to generate ARM complexes. These mixes were diluted to 90 nM and used to stain CoV-2 S WT CHOZN cells. Total RNA of the washed cells was collected and quantified using RT-qPCR (1-step RNA-to-Ct kit, ThermoFisher). A Taqman probe to the CHO housekeeping gene Fkpbla was used to normalize cell count, while probes specific for the unique CDRH3 sequences of casirivimab or ipilimumab were used along with standard curves to quantify the ARM RNA bound to the cell. The ratio of casirivimab RNA to ipilimumab RNA is reported (FIG. 15).Competitive ARM staining

[0222] An equimolar mix of RNA encoding 14 scFvs in ribosome display format was diluted 1 : 1 with RNA for an irrelevant library and used to generate ARM complexes. The mix was diluted to 45 nM or 2.25 nM positive ARMs and used to stain CoV-2 S WT CHOZN cells. Total RNA was collected, and the VH region amplified and sequenced. Bound ARMs were quantified by identifying reads with exact CDRH3 matches and quantifying the percentage of each clone. The enrichment of each clone is calculated by dividing by the clone frequency in the input stain ample (FIG. 18).Affinity measurement

[0223] Select clones from the competitive staining were expressed as scFvs as previously described. Single cycle BLI kinetic measurements were carried out on a GatorPrime Instrument (GatorBio). ScFvs was generated using an in vitro transcription-translation kit (PURExpress, NEB, E6800) then loaded onto Strep-Tactin XT probes (GatorBio, 160033) for 60 or 120 seconds followed by 30 or 60 second association and 300 second dissociation steps. RBD antigen concentrations were titrated to 100, 33, 11, and 0 nM and analyzed with a singlereference strategy and fitted to a global 1 : 1 kinetic model with Rmax Unlinked. 60 second baseline measurements proceeded the scFv loading and 300 or 700 seconds proceeding the association steps.Table 4List of oligo sequences used in this study. The 6-base Illumina Indexing Barcode is represented by [INDEX], For single-cell sorting (SCS) primers, the 5-base plate, row, and column barcodes are represented by [PLATE], [ROW], and [COLUMN] respectively.8. EQUIVALENTS AND INCORPORATION BY REFERENCE

[0224] While the invention has been particularly shown and described with reference to a preferred embodiment and various alternate embodiments, it will be understood by persons skilled in the relevant art that various changes in form and details can be made therein without departing from the spirit and scope of the invention.

[0225] All references, issued patents and patent applications cited within the body of the instant specification are hereby incorporated by reference in their entirety, for all purposes.Sequences

Claims

WHAT IS CLAIMED IS:

1. A method for high-throughput analysis of target binding proteins (TBPs), comprising: providing a library of target-decorated cells, wherein each of the target-decorated cells presents a target of interest on the membrane; contacting the library of target-decorated cells with a plurality of TBP-ribosome-mRNA (TRM) complexes, thereby inducing binding between the target-decorated cells and the TRM complexes; generating a plurality of monodisperse or polydisperse emulsion microdroplets, wherein each microdroplet contains a single cell out of the target-decorated cells, one or more TRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the TRM complex.

2. The method of claim 1, further comprising sequencing the library of hybrid polynucleic acids.

3. The method of claim 2, further comprising the step of identifying a target-TBP pair based on the sequencing of the library of hybrid polynucleic acids.

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

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

6. The method of any one of claims 2-5, further comprising the step of determining the read distributions of a plurality of TBPs in the plurality of TRM complexes across two or more targets of interest.

7. The method of claim 6, further comprising the step of normalizing the read distributions based on the input distribution of the plurality of TBPs.

8. The method of any one of claims 1-7, wherein at least one target of interest comprises or is conjugated to a domain capable of inducing expression of an activation marker in a target-decorated cell when the target of interest binds to a TBP.85 28152 / 53884 / FW / l 8022236.

49. The method of claim 8, wherein generating a plurality of monodisperse or polydisperse emulsion microdroplets comprises the step of sorting the target-decorated cells based on presence or absence of the activation marker.

10. The method of any one of claims 1-9, wherein the target-decorated cells express a fusion protein comprising a target of interest and a transmembrane domain.

11. The method of claim 10, wherein the target-decorated cells comprise a construct encoding the fusion protein comprising a target of interest 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 isolating the plurality of the monodisperse or polydisperse emulsion microdroplets containing the target-decorated cells by detecting expression of the fluorescent protein.

15. The method of any one of claims 1-14, wherein the library of target-decorated cells comprises one cell clone presenting one target of interest.

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

17. The method of any one of claims 1-14, wherein the library of target-decorated cells comprises at least 5 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones.

18. The method of claim 17, wherein the library of target-decorated cells comprises at least 10 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones.

19. The method of claim 18, wherein the library of target-decorated cells comprises at least 100 cell clones or at least 1000 cell clones, wherein each of the cell clones presents a unique target of interest different from the other cell clones.

20. The method of any one of claims 1-19, wherein the transcript of the isolated single cells comprises a coding sequence of the target of interest.

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

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

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

24. The method of 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 a TBP comprising an scFv.

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

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

28. The method of 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 different from the rest of the unique TRMS.

29. The method of 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 different from the rest of the unique TRM complexes.

30. The method of 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 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 different from the other unique TRM complexes.

32. The method of claim 31, wherein the plurality of TRM complexes comprises at least 1000 unique TRM complexes, wherein each of the unique TRM complexes comprises a unique target binding protein different from the other unique TRM complexes.

33. The method of 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 a coding sequence of a complementarity determining region (CDR) of the target binding protein.

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

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

37. The method of any one of claims 1-36, wherein RNA capturing is performed using oligonucleotides affixed to beads.

38. The method of claim 37, wherein RNA capturing is performed using oligonucleotides affixed to beads, wherein each bead has a diameter larger than 10pm, 0.5-10 pm, smaller than 1 pm or about 1 m.

39. The method of claim 37 or 38, wherein the oligonucleotides affixed to beads comprise a barcode sequence.

40. The method of claim 39, wherein the barcode sequences in the oligonucleotides affixed to beads are unique for a given bead.

41. The method of claim 39 or 40, wherein the hybrid polynucleic acids further comprise a barcode sequence from the oligonucleotides affixed to beads.

42. The method of any one of claims 1-41, wherein the hybrid polynucleic acids are generated by overlap extension polymerase chain reaction (OE-PCR).

43. The method of any one of claims 1-42, wherein generation of the hybrid polynucleic acids is preceded by first strand cDNA synthesis.

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

45. The method of claim 44, wherein the library of hybrid polynucleic acids is generated in the second set of emulsion microdroplets.

46. The method of any one of claims 1-45, wherein contacting the library of targetdecorated cells with the plurality of TBP-ribosome-mRNA (TRM) complexes is performed in a buffer comprising 25 to 100 mM Mg2+.

47. The method of claim 46, wherein the buffer comprises 50 mM Mg2+.

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

49. The method of any one of claims 46-48, wherein the buffer comprises HEPES, NaCl, 50 mM MgCh, polysorbate 20, heparin, and BSA.

50. The method of claim 49, wherein the buffer comprises 20 mM HEPES, 50 mM NaCl, 50 mM MgCh, 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 MgCh, 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:a plurality of constructs, wherein each construct encodes a fusion protein comprising a unique target of interest and a transmembrane domain; a plurality of TBP-ribosome-mRNA (TRM) complexes, wherein each TRM complex comprises a unique target binding protein; and a buffer.

53. The kit of claim 52, further comprising host cells.

54. The kit of claim 52 or 53, wherein the kit comprises a single construct, encoding one target of interest.

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 of interest.

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 of interest.

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 of any one of claims 52-58, comprising one unique TRM complex, wherein the one unique TRM complex comprises one target binding protein.

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

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

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

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

64. The kit of any one of claims 52-63, wherein the buffer comprises 50 mM Mg2+.

65. The kit of claim 64, wherein the buffer comprises 50 mM MgCh.

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

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

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

69. A library of hybrid polynucleic acids generated by the method of any one of claims 1- 51.

70. A method for high-throughput analysis of receptors, comprising:providing a library of ligand-decorated cells, wherein each of the ligand-decorated cells presents a ligand of interest on the membrane; contacting the library of ligand-decorated cells with a plurality of receptor-ribosome- mRNA (RRM) complexes, thereby inducing binding between the liganddecorated cells and the RRM complexes; generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the ligand-decorated cells, one or more RRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the RRM complex.

71. A method for high-throughput analysis of receptors, comprising: providing a library of receptor-decorated cells, wherein each of the receptor-decorated cells presents a receptor of interest on the membrane; contacting the library of receptor-decorated cells with a plurality of ligand-ribosome- mRNA (LRM) complexes, thereby inducing binding between the receptordecorated cells and the LRM complexes; generating a plurality of emulsion microdroplets, wherein each microdroplet contains a single cell out of the receptor-decorated cells, one or more LRM complexes bound to the single cell, and a lysis reagent 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 hybrid polynucleic acids that comprise a sequence from a transcript of the single cell and / or a sequence from the mRNA of the LRM complex.

Images