Novel assays for quantitatively detecting membrane protein binding
The NanoBiT® split enzyme assay, enhanced with HiBit peptide and digitonin, addresses the challenges of low dynamic range and expression variability in membrane protein assays, achieving improved signal-to-noise and accuracy in quantifying membrane protein interactions.
Patent Information
- Application Number
- PCT/US2026/026787
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-06-03
- Filing Date
- 2026-05-05
- Publication Date
- 2026-11-12
AI Technical Summary
Existing assays for measuring membrane protein-protein interactions in mammalian cells suffer from low dynamic range, high background interference, and variability in protein expression levels, making it difficult to discern true positive interactions.
A novel assay system using the NanoBiT® split enzyme, combined with a HiBit peptide and a non-lytic permeabilizing agent like digitonin, allows for orthogonal normalization of protein expression levels by competitively displacing SmBit from LgBit, thereby improving signal-to-noise ratio and accuracy in quantifying membrane protein interactions.
The system achieves a 3-fold increase in signal for natural membrane protein interactions and over 20-fold for designed interactions, significantly enhancing the dynamic range and reproducibility of membrane protein-protein interaction measurements.
Smart Images

Figure IMGF000037_0001_TABLE 
Figure IMGF000037_0002_TABLE 
Figure IMGF000038_0001_TABLE
Abstract
Description
PATENT Attorney Docket No.: 2292.1PC NOVEL ASSAYS FOR QUANTITATIVELY DETECTING MEMBRANE PROTEIN BINDINGCROSS-REFERENCE TO RELATED APPLICATIONS
[0001] The subject patent application claims the benefit of priority to U. S. Provisional Patent Application Numbers 63 / 816,642 (filed June 3, 2025; now pending) and 63 / 799,795 (filed May 5, 2025; now pending). The full disclosures of the priority applications are incorporated herein by reference in their entirety and for all purposes.SEQUENCE LISTING
[0002] This application includes by incorporation of reference a sequence listing in the XML format, 2292 1PC Sequence Listing, which was created on May 1, 2026 and contains 32 KB in content.BACKGROUND OF THE INVENTION
[0003] Membrane proteins assemble into functionally important complexes vital for diverse biological processes and regulation. Interactions between proteins within lipid bilayers, whether transient or constitutive, can also be chemically targeted towards therapeutic modulation1. Several assay systems exist for measuring membrane-embedded protein-protein interactions (PPI), varying drastically in their dynamic range and adaptability2,3. Here, we assess an emerging assay system, NanoBiT4. for the high-throughput analysis of transmembrane (TM) PPI in mammalian cells focused on complexes mediated principally by TM regions.
[0004] Split protein and 2 -hybrid systems fused to TM domains have been extensively explored to characterize homo- or hetero-oligomeric complexes of both natural and engineered assemblies2. E. co / Lbased systems like ToxR use TM helices' ability to self¬ interact and induce homo-dimerization of a fused transcription factor to initiate genetic reporters for mutually parallel single- and multi-pass membrane proteins3 8. Similar colorimetric reporter systems have been modified or developed to measure hetero¬ dimerization of distinct TM spans, i.e. dominant-negative (dn) ToxR9, dnARAC10, BACTH11, or the LexA-LexA* GALLEX12systems. By instead incorporating an antibiotic resistance reporter, the analogous ToxCAT’3assay enables genetic selections of TM span self-assembly in pooled library-based screens14,15. As well, a split beta-lactamase (BlaTM) strategy has beenused to extend measurement beyond mutually parallel TM interactions to antiparallel complexes while also restricting interaction of consequence to those with periplasmic localized enzyme fragments in the intended insertion topology16- 17. This suite of assays has been highly effective for studying TM domain interactions, particularly for single-span domains, leading to foundational discoveries concerning the sequence determinants directing interactions in membranes5’11’1’-16’18’19. However, utilizing E. coli as host drastically restricts investigations into most multi-spanning membrane proteins involved in human health and disease.[0005 j To test mammalian proteins in eukaryotic cell lines, split protein complementation using enzymes (DHFR20, P-galactosidase21, firefly luciferase22, etc.), fluorescence proteins2’’24, and others25-27have been used extensively28; however, many have drawbacks which are further exacerbated when examining membrane proteins. The decisive factor and typical shortcoming of most split proteins are their level of complementation independent of fusion partner PPI due to inherent self-interaction propensity29and / or irreversibility30that establish the basal signal level and usable dynamic range. Often the signal readout (fluorescence, enzyme activity, antibiotic resistance, metabolite dependence, etc.) over background is not substantial and, when combined with variability in protein expression levels, leads to difficulty in discerning true positive PPIs. The crowded cellular membrane environment can further reduce the signal-to-noise between background collisions and true PPIs, making discerning transmembrane PPIs in mammalian cells using biomolecular fluorescence complementation (BiFC) very difficult31,32. Modern assays strive for selectivity by requiring multiple molecular events in series as added layers of control: BRET33,34, TANGO35, SPARK36, mammalian-membrane two-hybrid split ubiquitin complementation (MaMTH) system37. This class of multi-step assays are reliably employed for GPCR recruitment of beta-arrestins or G-proteins and inter-cell surface interactions38, but are not widely employed for measuring PPIs between two transmembrane proteins. These assay¬ formats suffer from high background when both protein components are membrane-spanning. Likewise, strategies fusing multiple domains to large membrane protein constructs often exhibit shortcomings in trafficking and lack versatility for measurements in alternative contexts (e.g., distinct sub-cellular localizations, insertion topologies, engineered sequences).
[0006] There is a need in the art for more dynamic and accurate means for assaying protein interactions, especially membrane-protein-protein interactions. The present disclosure is directed to this and other unmet needs.SUMMARY OF THE INVENTION
[0007] In one aspect, the present disclosure provides assay systems and implementing methods for quantitatively determining membrane protein-protein interaction between a target protein and a binder protein in a mammalian cell. The methods entail ( 1 ) transfecting a target vector and a binder vector into the mammalian cell, wherein the target vector and the binder vector respectively encode the target protein and the binder protein that are respectively fused to the first and the second complementary fragments of the a split enzyme, (2) measuring a first enzyme activity in the cell after culturing the transfected cell under appropriate conditions to form a reconstituted split enzyme, (3) introducing into the cell a variant of the first complementary fragment that can outcompete the first complementary fragment for binding to the second complementary fragment and displace the first complementary enzyme fragment from the reconstituted split enzyme, (4) quantifying binder protein level in the cell by measuring a second enzyme activity from reconstituted split enzyme formed by the variant peptide with the second complementary enzyme fragment the in the cell, and (5) normalizing the first enzyme activity with the second enzyme activity or the quantified binder protein level; thereby quantitatively determining membrane protein¬ protein interaction between the target protein and the binder protein in the mammalian cell.
[0008] Some assay systems of the disclosure for measuring membrane PPIs are based on the NanoLuc® split enzyme. In these assay systems and related methods, the first and second measured enzyme activities are luminescence signals. In some embodiments, the first and the second complementary enzyme portions of the split enzyme are respectively SmBit and LgBit of the NanoBiT® split enzyme. In some of these embodiments, the first complementary enzyme fragment contains the amino acid sequence shown in SEQ ID NO:2 or a conservatively modified variant thereof, and the second complementary enzyme fragment contains an amino acid sequence shown in SEQ ID NO:1 or a conservatively modified variant thereof. In some of these methods, the variant peptide contains an amino acid sequence shown in SEQ ID NO:3 or a conservatively modified variant thereof.
[0009] In some embodiments, the variant peptide is introduced into the cell in the presence of, or after treating the cell with, a non-lytic cell permeabilizing agent. In some of these embodiments, the cell permeabilizing agent is digitonin. In some embodiments, the target protein is fused to the first complementary enzyme portion as shown in SEQ ID NO: 2, the binder protein is fused to the second complementary enzyme portion as shown in SEQ IDNO:1, and the variant peptide contains the amino acid sequence set forth in SEQ ID NO:3. In some embodiments, the employed target protein is a GPCR, and the employed binder protein is an engineered or design protein. In some embodiments, the target vector and the binder vector each further encode one or more peptide tags that are fused to the -terminus or the C-tenninus of the protein and / or the complementary enzyme portion encoded by each of the two vectors. In some embodiments, the novel PPI assays of the invention additionally include a step of normalizing the first enzyme activity' with level of the target protein in the cell. For example, level of the target protein in the cell can be quantified via co-transcription of a reporter molecule (e.g., rFluc) from the target vector.
[0010] In a related aspect, the disclosure provides test kits for quantitatively measuring membrane protein-protein interaction between a target protein and a binder protein. The kits contain (1) a first vector for inserting a polynucleotide sequence encoding the target protein to express a first fusion protein containing the target protein and the SmBit fragment of the NanoBiT® split enzyme, (2) a second vector for inserting a polynucleotide sequence encoding the binder protein to express a second fusion protein containing the target protein and the LgBit fragment of the NanoBiT® split enzyme, and (3) an SmBit variant peptide that (i) has a stronger affinity than SmBit for binding to LgBit and (ii) can displace SmBit from LgBit that is non-covalently bound by SmBit. In some of these embodiments, the SmBit fragment contains the amino acid sequence shown in SEQ ID NO:2 or a conservatively modified variant thereof, the LgBit fragment contains the amino acid sequence shown in SEQ ID NO:1 or a conservatively modified variant thereof, and the SmBit variant peptide contains the amino acid sequence shown in SEQ ID NO:3 or a conservatively modified variant thereof. Some kits of the disclosure can additionally contain a non-lytic cell permeabilizing agent. In some of these embodiments, the included cell permeabilizing agent is digitonin.
[0011] A further understanding of the nature and advantages of the present invention may be realized by reference to the remaining portions of the specification and claims.BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Figure 1. Normalization of PPI signal by relative protein expression using LgBit- HiBit. (a) Schematic of HiBit peptide (SEQ ID NO:3) added to cells with non-lytic permeabilization allows for a secondary event of luminescence indicating HiBit displacement of SmBit (SEQ ID NO:2) and quantification of the total pool of cellular LgBit-fused protein, (b) Immunofluorescence anti-FLAG staining of different LgBit -tagged “binders” RAMP1,CD8a, and InnTag5-eVgL at varying plasmid DNA transfection levels (5, 20, 40, and 80ng), co-expressed with CLR-SmBit (80ng) in HEK293T cells on 96-well glass-bottom plates. Scale bars, 50pm. (c) Raw luminescence quantifying protein expression level upon addition of 1 pM HiBit with 0.001% digitonin for semi-permeabilization, complementing and measuring LgBit-tagged “binders” RAMP1, CD8a, CD4, and eVgL upon DNA titration (1, 5, 10, and 20ng) in co-transfection with CLR-SmBit (20 ng). Mean and replicate luminescence data points are from n=2 biologically independent experiments. Error bars, SEM. Statistical significance from unpaired Welch’s t-test (asterisks, P <0.001). (d) Ratio of NanoBiT and HiBit-LgBit luminescence signals, resulting in a relative measure of PPI strength normalized by ‘"binder” expression level. CLR-SmBit (20 ng) co-transfected with LgBit-tagged “binders” (RAMP1, CD8, CD4, and lnntag5-eVgL) at varying plasmid DNA levels (1, 5, 10, and 20 ng). Mean and replicate luminescence data points are from n=2 biologically independent experiments, annotated with fold-change between positive and negative control PPI pairs. Error bars, SEM. Statistical significance from unpaired Welch’s t-test (asterisks, P <0.001). (e) Normalization of NanoBiT PPI luminescence by “binder” protein levels quantified either directly by NanoLuc complementation (LgBit+HiBit) as in panels (c-d) or indirect by a co-transcriptional rFluc reporter expressed downstream of the binder (T2A-rFluc in plasmid) for 8 SmBit / LgBit-fused protein pairs. Mean values and replicate data points are from n=3 biologically independent experiments. (1) Correlation of mean NanoBiT signal normalized by direct HiBit-LgBit luminescence or indirect T2A-rFluc quantification of samples in panel (e). Bottom line, linear regression line with R2=0.74 and Pearson correlation coefficient of 0.86; Top line, x=y regression line representing an ideal correlation. EpoR- RAMP1 pair is an outlier, caused by higher rFluc luminescence relative to HiBit detection.
[0013] Figure 2. Membrane protein construct trafficking leads to different apparent binding, (a) Schematic of construct variants of RAMP 1 with c-terminal LgBit (not shown): full-length WT RAMP1, TM-only truncated RAMP1 (RAMP1TM), TnnTag5-RAMPlTM, and HALO-tag-RAMPITM. Intra: Intracellular; Extra: Extracellular, (b) Immunofluorescence (anti-FLAG) surface staining of non-permeabilized HEK293T cells transfected with 100 ng of RAMP-1 constructs in panel (a) and 150 ng CLR. Scale bars, 20pm. Blue, Hoescht (nucleus), (c) Schematic of construct variants of eVgL with c-terminal LgBit (not shown): TM-only eVgL, InnTag5-eVgL, and HALO-tag-eVgL. Scale bars, 20pm, (d) Immunofluorescence (anti-FLAG) surface staining of non-permeabilized HEK293T cells transfected with 100 ng of eVgL-LgBit constructs in panel (a) and 100 ng eVgL-SmBit. Scalebars. 20pm. Blue, Hoescht (nucleus), (e). Luminescence NanoBiT memPPI measurement of RAMP1 constructs variants (RAMP1, RAMP1TM, InnTag5-RAMPlTM, and HALO-tag- RAMP1TM; 10 ng) co-transfected with CLR-SmBit (20 ng). Left, NanoBiT luminescence measuring interaction strength. Middle, HiBit-LgBit luminescence measuring binder expression level. Right, NanoBiT / HiBit signal ratio indicating expression-normalized relative PPI interaction strength. Means and replicate data points are technical triplicates of N=2 biologically independent trials. NanoBiT and HiBit values of different samples were normalized relative to the signals of CLR-SmBit with WT RAMPl-LgBit (set to 1) from each independent trial. Error bars, SEM; Norm. Luminescence: Normalized Luminescence. Statistical analysis was performed using One-way ANOVA with Dunnett’s comparisons relative to CLR-RAMP1 (asterisks denote p < 0.001). (f). Luminescence measurement of eVgL construct variants (eVgL, InnTag5-eVgL, and HALO-tag-eVgL) co-transfected with eVgL-SmBit (intra-cellular localized), with experimental conditions and display as in panel (e). NanoBiT and HiBit signals are normalized relative to the values of eVgL-e VgL (set to 1) from each independent trial. Means and replicate data points are from technical triplicates ofN=2 biologically independent trials. Error bars, SEM. Statistical analysis as in panel (e), relative to eV L-e VgL: InnTag5-eVgL, p = 0.015; HALO-tag-eVgL, 0.189. (g). Luminescence measurement of “binder” eVgL construct variants (eVgL, lnnTag5-eVgL, and HALO-tag-eVgL) co-transfected with InnTagS-eVgL-SmBit (partially surface-localized), with experimental conditions and display as in panel (e). NanoBiT and HiBit signals are normalized relative to the values of InnTag5-eVgL-SmBit with eVgL-LgBit (set to 1) from each independent trial. Means and replicate data points are from technical triplicates of N=2 biologically independent trials. Error bars, SEM. Statistical analysis as in panel (e), relative to InnTag5-eVgL-SmBit with eVgL-LgBit: InnTag5-eVgL, p < 0.001; HALO-tag-eVgL, p = 0.487.
[0014] Figure 3. Use of RCA DNA with NanoBiT / HiBit assay for fast screening of PPL (a) Schematic of pipeline for fast screening with memPPI. Colonies transformed with candidate “binder” plasmid of interest are resuspended in a 96-deep well plate. 1 pL of resuspension is denatured and subsequently used as input for the RCA reaction. RCA product is then diluted and used as input DNA of the binder for the NanoBiT / HiBit assay. Hits are traced back to the deep well plate. Plasmid DNA of PPI hits is purified and validated further. (b) Comparison of luminescence levels of NanoBiT and HiBit-LgBit obtained from transfection of “binder” DNA as RCA product or purified plasmid (10 ng) alongsidecorresponding SmBit-tagged “target” proteins (10 ng). Mean and data points come from biologically independent experiments: n=3 with technical triplicates for purified plasmid inputs for controls (CHAMP, RAMP1, eVgL); n=3 with technical triplicates for RCA product for controls; n=2 trials with technical duplicates for controls with RCA as “binder” DNA (RAMP-1, eVgL.); n=l for the novel screened binder candidates as purified plasmids or RCA product. Error bars, standard error of the mean (SEM). (c) Comparison of NanoBiT signal normalized by expression via HiBit for experiments using RCA product or plasmid as DNA input of binder shown in panel b. Error bars, standard error of the mean (SEM). (d) Correlation of mean NanoBiT signal normalized by direct HiBit addition for experiments using RCA product or plasmid as DNA input of binder shown in panel c.DETAILED DESCRIPTION OF THE INVENTIONI. Overview
[0015] Membrane protein-protein interactions (memPPIs) underlie critical biological processed including mediating signal transduction, transport, cell communication, and membrane organization. Despite their importance, systematic and quantitative measurement of protein transmembrane domains laterally interacting in mammalian cells remains challenging due to the low’ dynamic range of conventional fusion protein assays, naturally crowded cell membrane environments, and interference from overexpression or self-complementation artifacts.
[0016] The present disclosure is predicated in part on studies undertaken by the inventors to develop a high-throughput and more accurate assay that can quantitatively monitor mammalian transmembrane (TM) protein-protein interactions (PPI) via structural complementation. The inventors implemented, improved, and validated a platform based on the split luciferase complementation system NanoBiT®, establishing its suitability for quantifying membrane protein interaction within human cells. The inventors benchmarked and adjusted the assay platform using a diverse panel of known natural and synthetic membrane protein PPI pairs w’ith various topologies, confidently distinguishing true positive interactions from co-expressed non-interacting pairs. By introducing an additional in situ step to measure and normalize for protein expression levels, the inventors drastically improved the dynamic range and reproducibility of distinguishing PPIs with high interaction propensity across different cellular conditions and for high-throughput screening. To enable versatile memPPI measurements, the inventors also engineered a set of protein expression constructsto improve protein trafficking with defined topology, facilitating membrane protein design. By incorporating topological control and in situ quantification of relative expression levels, the improved memPPI NanoBiT® platform is suitable for high-throughput interaction analysis and screening synthetic TM polypeptides for target human membrane protein binding. Owing to its sensitivity, scalability, and adaptability, this assay platform of the present disclosure will have broad cross-disciplinary utility in protein science, design, and drug discovery.
[0017] As detailed herein, relative to existing PPI assays, the novel and improved system (“memPPI assay” as referred to herein) produced surprisingly advantageous results for examining PPIs. For example, the NanoBiT® system was successfully expanded from soluble proteins to membrane proteins across various topologies, establishing a more comprehensive platform for measuring PPI. In addition, high-throughput capabilities are achieved, which enable efficient screening for designed membrane protein binders. Further, the memPPI assay system achieves a significantly improved signal-to-noise ratio and lowered background interference. It demonstrates a 3-fold increase in signal for natural membrane protein interactions and over 20-fold for designed membrane protein interactions. By producing robust normalization improvement, the methods described herein also represent significant improvement over existing NanoBiT PPI assays without normalization or normalized by methods known in the art (e.g., via cell viability dye, total protein level, or co-reporter). Orthogonally normalizing binder expression level in situ as described herein, as opposed to other means for PPI assay normalization known in the art, enhances robustness by controlling the impact from binder expression level and reducing the need for labor-intensive optimization of binder protein expression levels through DNA titration. As a specific exemplification (see, e.g., Examples 2 and 6 below), a non-lytic permeabilizing agent (e.g., digitonin) in combination with HiBit peptide were used to quantify expressed intracellular LgBit fusion protein for normalization of protein expression level, thereby improving accuracy of the NanoBiT PPI assay.
[0018] Using a permeability-enabled intracellular HiBit complementation readout to normalize NanoBiT PPI as described herein solved a major problem in PPI assays (i.e., expression variability) without lysis and without adding a separate reporter construct. It relies on (i) displacement-based creation of a second signal proportional to LgBit fusion expression, (ii) using digitonin as a non-lytic delivery tool for HiBit peptide in cells, and (iii) using that second signal specifically to normalize NanoBiT PPI accuracy. It must be noted that thisorthogonal and in situ protein expression normalization via by competitive displacement in the NanoBiT PPI assay, exemplified herein with the HiBit peptide and digitonin, were by no means expected. This is because the NanoBiT PPI assay is intended for weak and reversible LgBit / SmBit association while HiBit / LgBit complementation is high-affinity, which would suggest that the normalization readout may not be orthogonal to the PPI readout and could distort PPI kinetics. Also, digitonin permeabilization changes cellular physiology’ and can perturb PPIs, trafficking, and compartmentalization, which would cast serious doubts on whether normalization would be accurate, non-artifactual, and quantitatively reliable in living cells. To summarize, the existing art would not provide any suggestion or motivation for one to introduce a permeabilization agent and an exogenous peptide (for competitive displacement) step into a live-cell PPI assay without distorting PPI kinetics and while achieving quantitative and orthogonal normalization as required by the present invention.
[0019] Unless otherwise stated, the present invention can be performed using standard procedures, as described, for example in Methods in Enzymology, Volume 289: Solid-Phase Peptide Synthesis, J. N. Abelson, M. I. Simon, G. B. Fields (Editors), Academic Press; 1st edition (1997) (ISBN-13: 978-0121821906); U. S. Pat. Nos. 4,965,343, and 5,849,954;Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press. Cold Spring Harbor, N. Y., USA (1982); Sambrook et al., Molecular Cloning: A Laboratory Manual (2nd ed.), Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y., USA (1989); Davis et al., Basic Methods in Molecular Biology, Elsevier Science Publishing, Inc., New York, USA (1986); or Methods in Enzymology: Guide to Molecular Cloning Techniques Vol. 152, S. L. Berger and A. R. Kimmerl Eds., Academic Press Inc., San Diego, USA (1987); Current Protocols in Protein Science (CPPS) (John E. Coligan, et. al., ed., John Wiley and Sons, Inc.), Current Protocols in Cell Biology (CPCB) (Juan S. Bonifacino et. al. ed., John Wiley and Sons, Inc.), and Culture of Animal Cells: A Manual of Basic Technique by R. Ian Freshney, Publisher: Wiley-Liss; 5th edition (2005), Animal Cell Culture Methods (Methods in Cell Biology, Vol. 57, Jennie P. Mather and David Barnes editors, Academic Press, 1st edition, 1998). The following sections provide additional guidance for practicing the compositions and methods of the present invention.IL Definitions
[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which thisinvention pertains. The following references provide one of skill with a general definition of many of the terms used in this invention: Academic Press Dictionary of Science and Technology, Morris (Ed.), Academic Press (1sted., 1992); Oxford Dictionary of Biochemistry and Molecular Biology, Smith et al. (Eds. ), Oxford University Press (revised ed., 2000); Encyclopaedic Dictionary of Chemistry, Kumar (Ed.), Antnol Publications Pvt. Ltd. (2002); Dictionary of Microbiology and Molecular Biology, Singleton et al. (Eds.), John Wiley & Sons (3rded., 2002); Dictionary of Chemistry, Hunt (Ed.), Routledge (1sted., 1999); Dictionary of Pharmaceutical Medicine, Nahler (Ed.), Springer- Verlag Telos (1994);Dictionary of Organic Chemistry, Kumar and Anandand (Eds.), Anmol Publications Pvt. Ltd. (2002); and A Dictionary’ of Biology (Oxford Paperback Reference), Martin and Hine (Eds.), Oxford University Press (4thed., 2000). In addition, the following definitions are provided to assist the reader in the practice of the invention.
[0021] The singular terms "a," "an," and "the" include plural referents unless the context clearly indicates otherwise. Similarly, the word "or" is intended to include "and" unless the context clearly indicates otherwise.
[0022] The term "agent" includes any substance, molecule, element, compound, entity, or a combination thereof. It includes, but is not limited to, e.g., protein, polypeptide, small organic molecule, polysaccharide, polynucleotide, and the like. It can be a natural product, a synthetic compound, or a chemical compound, or a combination of two or more substances. Unless otherwise specified, the terms “agent”, “substance”, and “compound” are used interchangeably herein.
[0023] The term "analog" or “derivative” is used herein to refer to a molecule that structurally resembles a reference molecule but which has been modified in a targeted and controlled manner, by replacing a specific substituent of the reference molecule w’ith an alternate substituent. Compared to the reference molecule, an analog would be expected, by one skilled in the art, to exhibit the same, similar, or improved utility. Synthesis and screening of analogs to identify variants of known compounds having improved traits (such as higher binding affinity for a target molecule) is an approach that is well known in pharmaceutical chemistry.
[0024] As used herein, the term "amino acid" of a peptide refers to naturally occurring and synthetic amino acids, as well as amino acid analogs and amino acid mimetics that function in a manner similar to the naturally occurring amino acids. Naturally occurring amino acids are those encoded by the genetic code, as well as those amino acids that are latermodified, e.g., hydroxyproline, γ-carboxyglutamate, and O-phosphoserine. Amino acid analogs refer to compounds that have the same basic chemical structure as a naturally occurring amino acid, i.e., a carbon that is bound to a hydrogen, a carboxyl group, an amino group, and an R group, e.g., homoserine, norleucine, methionine sulfoxide, methionine methyl sulfonium. Such analogs have modified R groups (e.g., norleucine) or modified peptide backbones, but retain the same basic chemical structure as a naturally occurring amino acid.
[0025] As used herein the term "comprising” or "comprises” is used in reference to compositions, methods, and respective component(s) thereof, that are essential to the invention, yet open to the inclusion of unspecified elements, whether essential or not.
[0026] As used herein the term "consisting essentially of' refers to those elements required for a given embodiment. The term permits the presence of elements that do not materially affect the basic and novel or functional characteristic(s) of that embodiment of the invention.
[0027] The term "consisting of' refers to compositions, methods, and respective components thereof as described herein, which are exclusive of any element not recited in that description of the embodiment.
[0028] The term "conservatively modified variant" or “conservatively substituted variant” applies to both amino acid and nucleic acid sequences. With respect to particular nucleic acid sequences, conservatively modified variants refer to those nucleic acids which encode identical or essentially identical amino acid sequences, or where the nucleic acid does not encode an amino acid sequence, to essentially identical sequences. Because of the degeneracy of the genetic code, a large number of functionally identical nucleic acids encode any given protein. For instance, the codons GCA, GCC, GCG and GCU all encode the amino acid alanine. Thus, at every position where an alanine is specified by a codon, the codon can be altered to any of the corresponding codons described without altering the encoded polypeptide. Such nucleic acid variations are ‘"silent variations,” which are one species of conservatively modified variations. Every nucleic acid sequence herein which encodes a polypeptide also describes every possible silent variation of the nucleic acid. One of skill will recognize that each codon in a nucleic acid (except AUG, which is ordinarily the only codon for methionine, and TGG, which is ordinarily the only codon for tryptophan) can be modified to yield a functionally identical molecule. Accordingly, each silent variation of a nucleic acid that encodes a polypeptide is implicit in each described sequence.
[0029] For polypeptide or peptide sequences, “conservatively modified variants” refer to a variant which has conservative amino acid substitutions, amino acid residues replaced with other amino acid residue having a side chain with a similar charge. Families of amino acid residues having side chains with similar charges have been defined in the art. These families include amino acids with basic side chains (e.g., lysine, arginine, histidine), acidic side chains (e.g., aspartic acid, glutamic acid), uncharged polar side chains (e.g., glycine, asparagine, glutamine, serine, threonine, tyrosine, cysteine), nonpolar side chains (e.g., alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, tryptophan), beta-branched side chains (e.g., threonine, valine, isoleucine) and aromatic side chains (e.g., tyrosine, phenylalanine, tryptophan, histidine).
[0030] A “fusion” protein or polypeptide refers to a polypeptide comprised of at least two polypeptides and a linking sequence or a linkage to operatively link the two polypeptides into one continuous polypeptide. The two polypeptides linked in a fusion polypeptide are typically derived from two independent sources, and therefore a fusion polypeptide comprises two linked polypeptides not normally found linked in nature.
[0031] A "ligand" is a molecule that is recognized by a particular antigen, receptor or target molecule. Examples of ligands that can be employed in the practice of the present invention may include, but are not restricted to, agonists and antagonists for cell membrane receptors, toxins and venoms, viral epitopes, hormones, hormone receptors, polypeptides, peptides, enzymes, enzyme substrates, cofactors, drugs (e.g. opiates, steroids, etc.), lectins, sugars, polynucleotides, nucleic acids, oligosaccharides, proteins, and monoclonal antibodies.
[0032] “Linkage” refers to means of operably or functionally connecting two biomolecules (e.g., polypeptides or polynucleotides encoding two polypeptides), including, wdthout limitation, recombinant fusion, covalent bonding, disulfide bonding, ionic bonding, hydrogen bonding, and electrostatic bonding. " Fused" refers to linkage by covalent bonding. A "linker" or "spacer" refers to a molecule or group of molecules that connects two biomolecules, and serves to place the two molecules in a preferred configuration with minimal steric hindrance.
[0033] The term "operably linked" when referring to a nucleic acid, refers to a linkage of polynucleotide elements in a functional relationship. A nucleic acid is "operably linked" when it is placed into a functional relationship with another nucleic acid sequence. For instance, a promoter or enhancer is operably linked to a coding sequence if it affects the transcription of the coding sequence. Operably linked means that the DNA sequences beinglinked are typically contiguous and, where necessary to join two protein coding regions, contiguous and in reading frame.
[0034] As used herein, the term "variant" refers to a molecule (e.g., a peptide or polypeptide) that contains a sequence that is substantially identical to the sequence of a reference molecule. For example, the reference molecule can be an enzymatic polypeptide disclosed herein or a fusion thereof. The reference molecule can also be a polynucleotide encoding the polypeptide. In some embodiments, the variant can share at least 50%, at least 70%, at least 80%, at least 90%, at least 95% or more sequence identity with the reference molecule. In some other embodiments, the variant differs from the reference molecule by having one or more conservative amino acid substitutions. In some other embodiments, a variant of a reference molecule is a conservatively modified variant, e.g., a variant which has altered amino acid sequences (e.g., with one or more conservative amino acid substitutions) but substantially retains the biological activity of the reference molecule. Conservative amino acid substitutions are well known to one skilled in the art.
[0035] The term “vector” is intended to refer to a polynucleotide molecule capable of transporting another polynucleotide to which it has been linked. One type of vector is a “plasmid”, which refers to a circular double stranded DNA loop into which additional DNA segments may be ligated. Another type of vector is a viral vector, wherein additional DNA segments may be ligated into the viral genome. Certain vectors are capable of autonomous replication in a host cell into which they are introduced (e.g., bacterial vectors having a bacterial origin of replication and episomal mammalian vectors). Other vectors (e.g., non-episomal mammalian vectors) can be integrated into the genome of a host cell upon introduction into the host cell, and thereby are replicated along with the host genome.Moreover, certain vectors are capable of directing the expression of genes to which they are operatively linked. Such vectors are referred to herein as “recombinant expression vectors” (or simply, “expression vectors”).III. memPPI assays with orthogonal protein level normalization
[0036] The present disclosure provides a novel assay system for detecting and quantitatively measure membrane PPIs (memPPI assay), which can be used for high- throughput screening applications in well plates. The assay system represents significant improvements over existing split enzyme complementation systems, the NanoBiT® split enzyme that is routinely used for biotechnology and drug discovery. As detailed herein, theinventors expanded and improved the NanoBiT® system with conditions for optimized signal-to-noise in measuring mixed topology natural and synthetic designed membrane protein complex driven entirely by lipid-embedded TM span interactions. The inventors engineered and validated constructs for topological control of transmembrane insertion and correct trafficking in HEK293T cells tolerant of diverse de novo lipid-spanning sequences. The inventions described herein added the novel ability to orthogonally normalize the NanoBiT® PPI signal. The orthogonal normalization is accomplished by using a non-lytic permeability agent to deliver to the same live cells a HiBit peptide which has a much higher affinity for the LgBit protein and competitively displaces the SmBit peptide. This allows for quantifying expression level of the LgBit-fused component, measuring relative protein binding (i.e., specific activity), which enables accelerated throughput screening procedures. The orthogonal protein level normalization provides a quantitatively linear intracellular measure of LgBit fusion abundance suitable as a normalization denominator in the NanoBiT PPI assay. In various embodiments, the normalization can involve the generation of a ratio of the measured PPI signal to the binder protein level, a linear regression of the two variables, or a nonlinear correction of the PPI signal with the binder protein level.[0037 J Demonstrating NanoBiT ’s adaptability and successful working principles for membrane proteins should facilitate broader investigation of membrane-spanning complexes and increase throughout of protein engineering approaches targeting natural mammalian membrane proteins within the cell, for example via computational design. The novel assay¬ system provided in the disclosure can create utility and advantages for detecting PPIs in protein engineering, especially for targeting membrane proteins. Other than examining natural membrane PPIs, the assay can also be used to screen and identify engineered or computationally designed synthetic proteins that bind a target protein of interest. Any mammalian cells may be used to perform the PPT assays of the invention. Examples include the HEK293T cell line, the CIIO cell line, the HEK293 cell line, as well as any primary cells that are readily available in the art.
[0038] The novel memPPI assay system of the disclosure can be implemented with any split enzyme assay for measuring PPIs, e.g., any assay based on complementation of a split luciferase such as NanoLuc® or a split fluorescent protein such as GFP. Using split NanoLuc® enzyme for illustration, the underlying mechanism of the assay system reads out a protein-protein interaction in the cell based on two protein in question, “protein A” and “protein B”, being fused to complementary partial fragments of a split enzyme (e.g.,NanoLuc®) wherein protein A & B’s formation of a complex results in induced proximity of the otherwise non-interacting enzyme fragments that reconstitutes an active enzyme - glow luciferase signal indicates how much complex of protein A and B is formed in a cell population. In the assay system, “Protein A” is fused to the small fragment of the split luciferase “SmBit” (1 kDa) and “Protein B” is fused to the large fragment of the split luciferase “LgBit” (19 kDa) to be co-expressed together for some time (e.g., 24-48 hrs) for the PPI measurement either in stably transduced cell lines or introduced to cells through transient transfection of host plasmids for those protein construct genes. The major downside of this “NanoBiT” assay’s traditional implementation is that the level of luciferase signal is also highly dependent to the expression levels of fusion proteins A and B. That is, a lOx increase in luciferase signal can occur due to a 1 Ox increase in expression level, rather than due to formation of a stable PPI complex between proteins A and B. In protein engineering applications when exogenously expressing and screening large numbers of distinct protein domains or sequence mutants / variants, the large variance in protein expression levels that is commonly observed can be a major factor detracting from the interpretation of the luciferase signal as a true positive PPI or false positive / negative. The novel and improved assay system described herein directly addresses this major issue of not knowing how underlying protein expression level impacts the resulting PPI-dependent split enzyme activity. As demonstrated, the novel assay system when compared to the existing NanoBiT® assay has increased reliability, enhanced signal-to-noise, and increased resolving power for measuring true positive versus true negative PPIs.
[0039] As detailed in the Examples below, the novel assay system of the invention involves execution of the measurement in a 2 -step format. First, split enzyme luminescence is first read to indicate the level of non-covalent PPI (“the first measured enzyme activity”). Then a subsequent step is added to perform direct quantification of protein copy number within the same assay plate well by obtaining a second measured enzyme activity as detailed below, which is then used to normalize the first measured enzyme activity by protein expression level in the cell. The result is for every assay well, the assay can measure the level of PPI in parallel with the relative protein expression level which leads to that split enzyme PPI assay signal. In some preferred embodiments, the additional assay step for direct protein level quantification is achieved by the delivery’ of a peptide “HiBit” which is a variant of the SmBit peptide (the small enzyme fragment of the split enzyme) which has high affinity for the LgBit fragment. When HiBit is introduced to the cells expressing both SmBit (protein A)and LgBit (Protein B), the HiBit peptide binds with high affinity to the LgBit enzyme fragment to rapidly reconstitute the active NanoLuc® enzyme (out-competing the fused SmBit peptide on “protein A’’ potentially in a non-covalent complex with “protein B”). Thereby, this second step of HiBit delivery to the cells is able to report on the total number of LgBit-fused component “protein B” present in the cells to obtain “the second measured enzyme activity” shortly after the PPI measurement (SmBit-LgBit) for the same cells (i.e., the first measured enzyme activity).
[0040] A key technical detail for successful implementation of quantifying protein expression levels in the assay is the non-lytic delivery' of the HiBit peptide into the cells. The employed non-ly tic permeability agent (e.g., digitonin) enables intracellular access of HiBit so that LgBit fusion protein abundance can be quantified in the same cells as the NanoBiT PPI readout (i.e., orthogonal normalization of protein level), and used as a normalization factor to improve PPI accuracy. In some preferred embodiments, this is achieved by formulation of the HiBit peptide in a mix with a gentle detergent at a low enough concentration where the final dilution in the cell media is below the detergent’s critical micelle concentration. This enables non-lytic cell permeabilization to allow cellular delivery of HiBit peptide for this LgBit copy number quantification in live cell conditions. In addition to the exemplified protocol for delivering the cell permeabilizing agent, any delivery means in a non-lytic and non-perturbative live cell format can also be used in the assay system of the disclosure. These would include delivery with a cell-penetrating peptide, via a liposome / lipid nanoparticle, or using other chemical cross-membrane delivery mode. Other than the exemplified embodiments of quantifying protein level via the enzyme reconstituted by HiBit binding to LgBit. protein level quantification can also be performed with other assay formats. These include, e.g., parallel monitoring of the SmBit component, which can be achieved by orthogonal measure such as fusing it to different enzyme (e.g., luciferase or GFP) or having the SmBit "target" expressed polycistronically with reporter such as T2A GFP / luciferase or IRES-reporter genes. In some embodiments, other than normalization by quantifying the level of LgBit-fused binder protein as noted above, the assay can additionally include a step of normalization by quantifying the level of SmBit-fused target protein. This can be performed, e.g., via a similar co-transcriptional reporter (e.g., rFluc) that is expressed in frame with (e.g., downstream of) the target protein. This indirect normalization of the target protein copy number in the cell can further improve accuracy of the PPI measurement that is normalized by the aforementioned direct normalization of the binder protein level via thedelivery of HiBit with a non-lytic permeabilization agent. Tn addition to the exemplified perrneabilizing agent digitonin, other non-lytic or mild perrneabilizing agents may also be used in the practice of the assays described herein. Examples of such agents include, e.g., Triton X-100, saponin, streptolysin-0 (SLO), CHAPS, Nonidet P-40 (NP-40) (at low concentration), lysolecithin, and recombinant perfringolysin O (e.g., XF PMP reagent).
[0041] Other embodiments of the memPPI assays of the present disclosure include swapping the sides of the fusion partner, e.g., having LgBit instead of SmBit to measure the target. Alternatively, the quantification can be performed by staining for the SmBit "target" protein with suitable antibodies and using FACS to collect a uniformly protein expressing cell population. In some embodiments, protein level quantification can be performed with another protein or peptide that competes with the complementary fragments of the split enzyme, e.g., Rluc which is a bioluminescent enzyme found in Renilla reniformis. In some embodiments, the first measured enzyme activity and the second measured enzyme activity can be obtained with a sequence of adding a substrate (e.g,, furimazine of NanoLuc®) to the cell, reading the first measured enzyme activity, adding the HiBit peptide to the cell, and reading the second measured enzyme activity. In these embodiments, it is not necessary to add substrate again for the reading the second measured enzyme activity. In some other embodiments, the cells can be washed to remove the originally added substrate after reading the first measured enzyme activity. Fresh substrate of the split enzyme (e.g., furimazine) can then be added to the cells for reading the second measured enzyme activity after the HiBit peptide is added.10042 ] The quantification of the LgBit-fused Protein B through cell delivery of the HiBit peptide allows for PPI measurement and protein copy number quantification to occur in the same well, and facilitates the key technical advantage of obtaining a means to normalize the PPI split protein signal by the amount of Protein B (protein cellular copy number aka expression level) together for the same cells in the same assay plate well. The assay steps can be performed in rapid succession (e.g., within 30 minutes) and is amendable to high degrees of variance in the protein copy number (i.e., providing extensive dynamic range for protein expression level). Normalization of a PPI signaling by protein expression level allows for more consistent comparison of the relative binding propensi ty between two proteins and controls for artifact PPI signals resulting from increased signal simply due to increased component expression level rather than stronger PPI or increased co-localization. Likewise, normalization of split protein complementation PPI measurements by direct quantification ofprotein levels allows for in situ determination of the relative binding propensity under a wide range of cellular conditions including comparisons across cells with drastically different protein copy numbers / expression levels, or conditions where the amount of gene delivered to a set of cells / wells is variable or unknown. For these reasons, ability to measure PPI and normalize by direct quantification of protein copy number can enable higher throughput protein engineering screening applications.
[0043] The novel assay system of the present disclosure can be readily employed in the specific application of measuring PPI between membrane proteins, where increased assay sensitivity is vital for interpretable measurements given the distinct environment of expressing proteins in 2D vs 3D for lateral interactions in bilayers of cellular membranes versus free in solution. It can be used to identify binder proteins in protein engineering screening system, providing additional protein quantification measurements and improvement in signal to noise ratio and fold-change. This is indispensable for discerning tnie positive and true negative protein-protein interactions. The assay system can also be used to quantify protein express levels in situ from transient transfected cells, allowing molecular biology processing of input genetic material that is more amendable to high-throughput pooled screening approaches. This cannot otherwise be performed or interpreted with the existing split enzyme PPI assay format such as NanoBiT®.
[0044] To more accurately and quantitatively detect membrane protein-to-protein interaction between a target protein (e.g., a GPCR) and a binder protein (e.g., a natural ligand or an engineered / design protein), detailed protocols for carrying out the niemPPI assay of the disclosure are provided herein. The assays can be performed with either transient vs stable protein expression systems. Some exemplifications of the assay systems are set forth in the Examples below. As a specific exemplification, the NanoBiT® PPI assay can be performed with human embryonic kidney 293 T cells. Vectors or plasmids expressing the target protein fusion (SmBit construct or vector) and the binder protein fusion LgBit construct or vector) can be prepared via standard molecular biology methods and the user instructions provided by the NanoBiT® assay manufacturer. The vectors can be isolated using Qiagen or Zymo miniprep kits and diluted to standardized concentrations prior to transfection. The cells can be transfected with an appropriate amount of each of the two constructs (e.g., 20-30 ng) with, e.g., Lipofectamine3000. At the time of co-transfection, the cells can be seeded into, e.g., PDL-coated white opaque, flat bottom 96-well plates. After proper seeding (e.g., 18 to 24 hrs), the cell media can be replaced with, e.g., COz-independent DMEM. Luminescence canthen be measured for each well with, e.g., the Tecan Spark plate reader with appropriate setting (e.g., using a 250 ms integration time and no attenuation after adding Furimazine to each well). After SmBit-LgBit PPI luminescence signal reading, HiBit peptide can be added to the same live cells in conjunction with an appropriate cell-pemieabilizing agent. In some preferred embodiments, the cell-penneabilizing agent is digitonin. Digitonin is a weak nonionic detergent, which at low concentrations selectively renders the plasma membrane permeable, causing it to release cytosolic components but to leave other organelles intact. Thereafter, the cells are subject to a secondary' luminescence measurement to quantify the amount of LgBit expression level. This can be achieved with the same device (e.g., the Tecan Spark reader) and under slightly different setting, e.g., using a 250 ms integration time and OD1 attenuation, with reading acquired every 3 minutes until luminescence values across the plate plateaus and begins to drop. Normalized luminescence can then be calculated by dividing the initial luminescence (RLU) by the highest recorded values for HiBit supplemented luminescence (RLU) of each condition. Multiple biological replicates can be performed for every interaction measured, and technical replicates can be performed in triplicate within each experiment.[0045 J In addition to the described assay format or methods, the disclosure also provides assay kits for carrying out the memPPI assay described herein. The kits can contain a suitable mammalian cell line (e.g., HEK293T cell) and materials for constructing vectors expressing proteins A and B of a split-enzyme system (e.g., SmBit and LgBit of NanoBiT®) that are respectively fused to a target protein and a candidate or test binder protein. An example of the vectors for fusing a target protein to SmBit is shown in Table 2 (first row). Examples of vectors expressing fusions containing LgBit and a candidate binder protein are shown in Tables 3 and 4. The kits additionally contain a variant peptide (e.g., HiBit) of Protein A of the split enzyme (e.g., SmBit of the anoBiT® system) that has stronger affinity for (and can outcompete Protein A for binding to) Protein B of the split enzyme (e.g., LgBit of the NanoBiT® enzyme). Further, the kits include a suitable non-lytic cell permeabilizing agent and other reagents for non-lytic and non-perturbative intracellular delivery of the variant peptide into the mammalian cell line. In some preferred embodiments, the kits contain materials for assaying PPIs via the NanoBiT® split enzyme, the HiBit peptide, and cell permeabilizing agent digitonin. Additional reagents for performing the first PPI measurement and the protein level quantification of the memPPI assay of the present disclosure can also be included in the kits. These additional reagents include, e.g., cell growth media and washingbuffer. The kits can further include packaging material for packaging the reagents and a notification in or on the packaging material, as well as appropriate instructions for use and labels indicating the intended use of the contents of the kits. The instructions can be present on any written material or recorded material supplied on or with the kit or which otherwise accompanies the kit.
[0046] As demonstrated herein, the memPPI assay system of the disclosure can be readily employed to examine pairs of transmembrane proteins under specific working conditions controlling for common artifacts: over-expression, mis-trafficking, confident non-interacting baseline, and normalizing expression level variability across samples. This led to establishing a convincing, robust, fast (transient transfection), and scalable platform in well-plate format tailored for memPPIs. To accommodate high-throughput applications, i.e., screening natural interactomes or engineered TM protein pools, the assay can utilize expression constructs and tagging strategies promoting reliable, robust protein trafficking, which are adaptable to diverse iterations of “Target” or “Binder” proteins. By quantifying and normalizing by protein expression levels in situ in the same wells, the memPPI assay of the disclosure dramatically broadens the assay's practical dynamic range, and reduces technical burden of manually titrating multiple proteins to common expression levels or rigorous quantification of expression in parallel by other means. The resulting measure of relative interaction propensity enables more consistent comparisons even among cell populations varying widely in protein copy numbers. Notably, the confidence to quantify the LgBit-fused component in situ regardless of DNA amount delivered or expression efficacy of individual constructs makes the memPPI assay amendable to expedient approaches screening PPIs, including applications designing TM proteins targeting mammalian GPCRs, chaperones, and ion channels.
[0047] It is to be noted that various further modifications can be made to the embodiments specifically exemplified herein. For example, NanoBiT® experiments can be strengthened with ample verification of expression levels and trafficking patterns, especially for new “Target” constructs, as exemplified herein via immunostaining. In some embodiments, expression levels of a SmBit component both alone and under co-expression conditions with a representative set of LgBil-tagged constructs can be evaluated to account for potential co-variation. In some embodiments, T2A-rFluc quantification or flow cytometry can be used to normalize or collect cell populations to control expression levels of the SmBit component for more precise comparisons. Given possible variation in SmBit orLgBit levels across protein pairs, the memPPI assay can include testing multiple non- interacting negative control TM proteins be assayed in parallel per “Target” in some embodiments of the present disclosure. Obtaining a true positive PPI for that “Target” to assess level of luminescence expected for the high and baseline levels of complementation to compare unknown PPIs pairs is beneficial to interpretation of the assay results. In some other embodiments, water-soluble binding domains like nanobodies can be employed as positive controls if known TM PPIs are lacking. In some embodiments for discovering new PPIs to a “Target” where no true positive control complex exists, the assay system can include a 2-fold expression-normalized luminescence change above the baseline signal of relevant noninteracting pairs with the same “Target” as the conservative minimum indicative of a true complex. Optimized or constitutive high affinity pairs reach into 10’s or 100’s fold-change32.. In some embodiments for examining known complexes, smaller but statistically significant decreases in NanoBiT® signal upon chemical changes may be interpretable as altered interaction relative strength, if expression and localization are equal. Some embodiments of the assay system can use assay conditions trending towards lower total protein expression levels for increased sensitivity to distinguish true PPIs apart from coincident co-localization or over-expression.EXAMPLES
[0048] The following examples are provided to further illustrate the invention but not to limit its scope. Other variants of the invention will be readily apparent to one of ordinary skill in the art and are encompassed by the appended claims.
[0049] We sought to develop an assay to evaluate versatile membrane protein complexes at scale, enabling chemical biology campaigns such as interaction network mapping, drug discovery, or computational protein design. Such measurements create opportunity to connect design principles, atomic-detail rationale, and algorithms for evaluating PPIs in cells - especially for complexes mediated by TM spans. We propose that the split nanoLuciferase (nanoLuc) NanoBiT system’s unique characteristics are uniquely positioned for discerning PPI measurements between mutual membrane-spanning proteins partners of diverse topologies. LgBit (19 kDa) and SmBit (1.3 kDa) components have minimal selfcomplementation affinity (Kd - mM without interacting fusion partners)4, reversible nature, minimal fusion mass, and enzymatic signal amplification - behaviors leading to the observed higher signal-to-noise for most soluble protein than classic PPI assays. Thus far, NanoBiT' sperformance and technical limitations in such TM-focused applications have not yet been thoroughly assessed. Early precedents for monitoring PPIs by NanoBiT in the membrane include interactions between membrane proteins39'41and G-proteins recruitment to GPCRs42. NanoBiT of constitutive GPCR dimers CBjR and 5-HT'2 demonstrates orders of magnitude better fold-change over background compared to analogous BiFC measurements32.
[0050] Here, we evaluated the utility, scope, and conditions for NanoBiT to reliably distinguish true interactions in the same cell between two membrane proteins from non¬ interacting co-expressed protein pairs across distinct protein topologies and architectures. Novel constructs for robust trafficking and TM protein engineering in HEK.293T cells are presented that control topology, hosting diverse synthetic sequences and appropriately orienting TM spans. Normalization of PPI signal by a relative protein’s expression level was also implemented to better discern interaction propensities, aiding high-throughput screening procedures. Establishing successful, reliable working principles for NanoBiT s application to membrane protein-protein interactions (memPPI) broadens the capabilities to investigate transmembrane interactions and facilitates protein engineering targeting mammalian membrane proteins, e.g., by computational design.Example 1. MemPPI can distinguish natural and designed membrane protein interactions of distinct topologies
[0051] We first sought to establish what is the expected signal-to-noise for a range of relevant true positive natural and synthetically optimized membrane protein complexes, as well as true negative constructs to establish the critical non-interacting baseline from non¬ specific complementation. Two key determinants in our design were developing constructs where fused LgBit and SmBit domains are properly trafficked, localized to the same cellular compartment, and quantitatively monitoring apparent protein expression levels - which depend both on the amount of DNA transfected and each protein’s intrinsic expression and trafficking efficacy. A few isolated studies have monitored NanoBiT activity upon transiently transfecting expected TM interaction partners (GPCR dimers32, VEGFR2 / NRP140, HIV VPU / human CD3179, IL-23R dimers41). These cases use a relatively high amount of DNA transfected (40-200 ng for each component) ranging on a case-by-case basis from investigator experience. Likewise, positive control PPI levels upon co-expression were typically compared to that of expressing separate components alone (SmBit or LgBit) or co-expression of the target membrane protein with a non-membrane protein partner (HALO-SmBit orA0SmBit peptide). The NanoBiT assay’s reliability to detect true TM PPI complexes has not yet been thoroughly evaluated, particularly in the most relevant context where signal-to-noise and baseline complementation are compared across conditions co-expressing other non¬ interacting proteins also embedded in cell membranes.
[0052] To establish the benchmark for signal-to-noise fold increase for TM PPIs relative to non-interacting partner pairs, we measured split nanoLuc complementation for several positive control PPI complexes which are all extensively biochemically validated and structurally characterized, including both natural and de novo designed proteins. This set includes 2 natural GPCR complexes with their TM accessory proteins of distinct topologies of known structure: class B calcitonin-like receptor (CLR) and its type I accessory adaptor receptor activity-modifying protein 1 (RAMP1)43; class A sphingosine 1 -phosphate receptor 1 (S1PR1) and its type II adaptor CD6944. The single-pass homodimeric Glycophorin A (GpA) and its drastically weakened G83I mutant were included, given this has served as a standard benchmark model for membrane PPIs in the field. For synthetic complexes engineered for high-stability, we evaluated 2 designed membrane protein assemblies of single-pass subunits: the homo-pentameric “eVgL”45,46assembly of parallel-oriented type I inserted single-pass protein subunits and antiparallel-oriented complex between type I mouse erythropoietin receptor (EpoR) and inhibitory type II designed “CHAMP”47single-pass protein.
[0053] In the assay, one protein is deemed the “Target” (typically the natural membrane protein) and is tagged with the minimal SmBit fragment (usually at its C-terminus) to be minimally perturbative and to avoid modifying N-terminal signal peptides. The interaction partner, referred to as the “Binder”, is LgBit-tagged (typically at its intracellular termini) and FLAG-tagged (typically at the extracellular termini). The latter “binder” construct is envisioned to be made variable during potential protein engineering screens. We chose a series of mock “Binder” TM proteins known to express well and traffic with high fidelity to the plasma membrane surface to serve as non-interacting negative controls which can reveal the baseline level of non-specific split enzyme complementation simply from co-expression with a given “Target”. These include natural proteins CD4 or CD8a and a de novo 3-pass protein (referred to as syn-3TM). Additionally, positive control “Binders” (RAMP1, eVgL) for one “Target” can be re-used as negative controls for a different “Target” when there is no expected interaction.
[0054] In designing optimal NanoBiT conditions for signal-to-noise, we posited that only a few copies of each membrane protein would be necessary to detect a protein complex. Conversely, transfecting large amounts of DNA leading to overexpression likely is detrimental to signal-to-noise - from potential mechanism potentially exacerbated from increased expression levels: non-specific collisions, misfolding, accumulation, aggregation, etc. For the diverse PPI pairs described, co-expression of “Target” and “Binder” proteins at total 10-30 ng combined DNA in 96-well plates resulted in strong NanoBiT signal for the expected positive PPI pairs (~106RLU). Relative to the baseline complementation signal set from co-expressing each " Target” with one or more non-interacting TM protein controls, the fold change increase in NanoBiT luminescence spanned 3-10x for known natural complexes and reached up to 60-fold increase for the designed eVgL (p-values < 0.001). By contrast, when large amounts of plasmid are transfected (range of 100-200 ng range per component, typical for lipofectamine and reported for GPCR dimers32), we found a drastic impairment in discrimination of known positive PPIs from non-interacting protein pairs. We observed that co-transfecting large amounts of “Binder” DNA (200 ng) alongside 100 ng CLR “Target” DNA results in the true negative pair CLR-CD8a having a greater assay signal than true positive CLR-RAMP1 pair. At 100 ng “Binder” this is only a 1.2- fold difference, and it becomes a significant 2-fold change at 50 ng “Binder” (p<0.001, nA2). These data and our lab's accumulated experience testing many membrane protein pairings indicate that under standard high-plasmid assay conditions NanoBiT luminescence does not readily discern complexes between membrane proteins without optimization.
[0055] We next further tested the hypothesis that lower TM protein expression level might increase the ability to distinguish true positive from non-specific membrane-spanning PPIs through systematic titration of DNA transfected. Luminescence signals from true positive PPI pairings eVgL -eVgL and CLR RAMP 1 were compared with that of those “Targets” instead co-transfected with non-interacting TM proteins as mock “Binders’-. Fixed amount of SmBit-tagged “Target” DNA (20 ng) was co-transfected with varied amounts of “Binder” DNA. Comparing cells co-transfected with eVgL-SmBit and eVgL-LgBit (true positive) to those with eVgL-SmBit and non-interacting CD8a-LgBit, we observed substantial increases in luminescence from 34 to 80-fold across the DNA mass ratios tested (1:1, 1:5, 1:10, 1:20 ng; target:binder; each n=2, all p<0.001). Thus, detection of eVgL’s selfinteraction is very robust, relative to appropriate controls, suggesting that homo- pentamerization occurs in cells even at low protein copy numbers. For CLR, its NanoBiTsignal when co-expressed with RAMP1 was drastically increased relative to that of pairs with non-interacting partners (CD8a, CD4, eVgL). The fold increase in luminescence between CLR-RAMP1 and CLR-CD8a pairings becomes greater as less “binder’’ DNA is transfected per-well from 20 to 1 ng, from 2.2 ± 0.1 fold-change to 13.0 ± 1.5 fold-change, respectively (p<0.001, calculated from n=2). These fold-increases for CLR-RAMP1 were consistent, trending similarly whether additional empty plasmid was supplemented as vehicle or not, i.e., to reach an equal total amount of DNA transfected (40 ng) across varying “Binder” DNA delivered. For the self-interaction of GpA, NanoBiT luminescence was significantly increased relative to that of the weakened mutant GpA(G83I) between 4-5x fold-change for 5-20ng “Binder” DNA per well and up to 8-fold at Ing). These data further support that keeping membrane protein expression levels low can be beneficial in improving the signal-to-noise observed for TM PPIs.
[0056] Finally, we tested 2 sets of TM protein constructs with one of the most commonly used split protein assays, split-mVenus BiFC, to compare side-by-side and assess discriminating power relative to NanoBiT’s. For the self-interaction propensity of GpA’s and mutant GpA(G83I)’s measured by fusion and co-expression of split-mVenus constructs, we observed only a very modest 1.4- fold increase in cellular fluorescence intensities at best, detectable only at highest amount of plasmid DNA transfected (200 ng). At lower plasmid levels, the differences between GpA WT and mutant fluorescence were reduced (just 10-20% at 50-100 ng DNA) and fluorescence of mutant GpA(G83I) self-pairing was greater than that of WT GpA at the lowest DNA level (20 ng). For the 2ndcomparison between true positive EpoR-CHAMP pairings and true negative EpoR-CD69, split-mVenus fluorescence showed a modest but consistent 20% increase (1.2x fold) when transfected with high plasmid levels (50-200 ng), but fluorescence was indistinguishable at the at lowest level of plasmid added (20 ng). Another limitation for BiFC is the additional maturation time for fluorescent protein, here requiring 48-hours after transfection for mVenus versus NanoBiT’s being ready faster (18-24 hr). The substantially higher fold-change in luminescence observed with NanoBiT confirms its superior sensitivity and dynamic range over BiFC in most cases for studying and screening membrane protein interaction complexes.Example 2. Quantitative normalization by binder expression increases consistency and discriminatory power across scales
[0057] Given the apparent relationship with expression level and dynamic range of NanoBiT between membrane proteins, we devised a strategy for quantification of protein copy number in situ to better distinguish mechanisms of heightened luciferase complementation occurring due to true PPI or simply from higher protein expression levels of relative components. The approach leverages the LgBit domain to independently quantify the “Binder” protein’s copy number in situ. Luminescence is measured following intracellular delivery of the HiBit peptide, which binds LgBit with nanomolar affinity to spontaneously reconstitute NanoLuc®, outcompeting any expressed SmBit. This live-cell method avoids lysis, which can aggregate membrane proteins, and reports the total accessible pool of LgBit- tagged “Binder”. In the two-step protocol (Fig. 1, panel a), NanoBiT luminescence is first measured, assessing PPIs, followed by HiBit addition. Then, luminescence re-measured in the same well, assessing LgBit-HiBit reconstitution and quantifying the total cellular “Binder” levels. This approach is fast, easy to implement, and especially useful for monitoring very low’ protein expression levels (recommended for memPPI) near or below detections limits of conventional microscopy and immunoblotting.
[0058] To benchmark the utility of adding this LgBit-HiBit quantification step to the NanoBiT assay, we first evaluated the variance in expression levels for 3 LgBit-fused “Binder” constructs co-expressed with “Target” CLR (Fig. 1): RAMP1, CD8a, and InnTag5-eVgL (i.e. a construct of the eVgL TM span fused at its N-terminus to the InnTag5 globular domain, improving plasma membrane trafficking). First, we monitored and quantified trends between “Binder” expression level and DNA transfected by immunostaining fluorescent microscopy (Fig. 1, panel b). CD8a expression increases relatively linearly with more DNA, whereas RAMP1 plateaus at high expression early in the DNA titration and InnTag5-eVgL expression remains low until the highest DNA level. Next, we monitored expression level by LgBit luminescence in HiBit-treated cells co-transfected with increasing amounts of DNA under analogous conditions (Fig. 1, panel c). The observed “Binder” expression levels for each construct’s match those DNA-dependent trends observed by microscopy, indicating LgBit-HiBit luminescence can report in situ protein copy number reliably. We then integrated this HiBit-LgBit measurement to normalize the preceding memPPI NanoBiT signal, resulting in an interaction propensity relative to the amount of LgBit-Binder expressed. Traditional NanoBiT comparing CLR-RAMP1 and CLR-CD8a co-expressed pairs yielded differences in luminescence depending heavily upon the nanogram (ng) of DNA delivered, ranging between 2 to 13 -fold. After normalizing by the HiBit-LgBit expression level, the ratio of normalizedPPI signals for CLR-RAMP1 versus negative PPI pairs becomes far more consistent across DNA transfection conditions (Fig. 1, panel d): CLR-RAMP1 versus CLR-CD8a, 2.1 + 0.1, 2.7 ± 0.1, 2.9 ± 0.1, 2.7 ± 0.2 fold-changes for 20, 10, 5, and 1 ng Binder, respectively (all p < 0.001; N=2). Likewise, for GpA WT dimer and known disruptive G83I mutant, the foldchange in expression normalized luminescence between self-pairings (GpA-GpA and G83I- G83I) remains stable across “Binder” DNA titration from 20 to 1 ng, ranging consistently from 1.6x to 1.8x.
[0059] To further evaluate the normalization of NanoBiT signal by protein expression level, we tested the HiBit-LgBit method for 8 additional PPI pairs and compared this luminescence with that of a canonical co-translational reporter approach indirectly tracking an orthogonal rFluc (red firefly luciferase) downstream of the “Binder” (“Binder”-T2 A-rFluc) in transfected plasmids. We tested 2 known positive pairings (CLR-RAMP1, EpoR- CHAMP) and 6 expected negative pairings with “Targets” CLR, mu-opioid receptor (MOR), and EpoR (Fig. 1, panel e). NanoBiT luminescence normalized by either rFluc or HiBit-LgBit show strong agreement for both positive and negative PPIs (Pearson correlation, R=0.86; Fig.1, panel f). Both normalization strategies similarly discriminate positive PPIs and yield consistent relative expression measurements for “Binder” in 7 of 8 pairings. The outlier PPI EpoR-RAMP1, has not been previously detected or otherwise measured to our knowledge. This Target-Binder pair exhibits modest NanoBiT PPI levels, but when normalized by rFluc appears as low (negative PPI) or as high by HiBit-LgBit (positive PPI). It is unclear whether this PPI is a false negative from rFluc detection, false positive from HiBit-LgBit, or a new true PPI, but it is noted that RAMPs are chaperones with broad scopes of interaction, immunoprecipitating with several receptor classes48,49. Nonetheless, the data supports clear utility of 2-step NanoBiT method for normalization by expression level in most cases. The HiBit-LgBit method avoids interference from bicistronic elements and enables direct in situ measurements of protein levels where co-transcriptional or co-translational reporters can fail to reflect endpoint abundance like membrane protein trafficking and stability. This simple method assaying protein copy number in the same cells addresses challenges of interpreting PPIs from comparisons of absolute NanoBiT signals across many different protein pairs that inevitably vary in underlying protein expression levels.Example 3. Engineering membrane protein topological control for relevant PPI measurements
[0060] Important underlying assumptions for relevant interpretation of PPI signal is that the membrane proteins are expressed (detectable by LgBit / HiBit), trafficked correctly, and topologically oriented to correctly position the LgBit domain in the same cellular compartment as the SmBit tag. To facilitate this behavior, membrane protein constructs must be evaluated and often improved by optimal sequences of signal peptides, epitope sites, soluble domains, and other tags. These are difficult facets to control at scale during protein engineering campaigns, e.g., screening many different TM architectures or pools of diverse (possibly de novo) sequences. We report 3 practical case studies emphasizing the need for suitable protein constructs and how working principles of membrane protein trafficking can be incorporated to enable assaying several natural and synthetic transmembrane domains by memPPI.
[0061] For the first case, we sought to develop scaffold protein expression constructs with robust trafficking directed into a Type II protein insertion topology, which can tolerate diverse natural or synthetic TM sequences. This would enable of screening of libraries of engineered TM domain sequences for antiparallel-oriented binding interactions with natural Type I cellular receptors, such as immune or cytokine receptors. We designed and benchmarked several constructs (Table 4) when hosting the de novo designed TM CHAMP protein which binds EpoR in an antiparallel-oriented complex which inhibits receptor signaling47.
[0062] Construct design was divided into three main components meant to encourage Type II plasma membrane topology: soluble domains, N-terminal signal peptide, and TM-proximal linkers. Type II membrane proteins characteristically lack a cleavable N -terminal signal peptide, instead having modest-sized polar leader sequences that facilitate post-translational inversion of their first TM segment in membranes (i.e., N-terminus facing the cytosol). A sufficiently long water-soluble region C-terminal to the TM span is also favorable for this energetic transition50,51. To chose these elements, we mined the UniProt database for type II single-pass membrane proteins (SL-9906) annotated with characteristic features: plasma membrane localization, >3 predicted extracellular glycosylation sites requiring type II topology, monomeric, and those with >110 residue of C-terminal regions that fold into globular domain which lack enzyme activity. The search identified two candidates as C-terminal extracellular regions for our constructs, human C-type lectin domain family 2 member B (HsCLC2B) and natural killer cell receptor E (HsNKG2-E). We also included the monomeric InnTag5 domain, a minimally perturbative domain tag52, and LgBit as candidateextracellular domains. For the intracellular N-terminal leader sequences and domains, we selected one short and one long N-terminal leader sequence from our bioinformatic search to facilitate type II protein topology; “ssPolarl” from HsNKG2-E and “ssPolar2” Oxidized low- density lipoprotein receptor 1 (HsOLRl)33. The construct N-terminus varied between having only the minimal polar leader sequence followed by an ALFA epitope (e.g., Construct-2) or including a folded domain (e.g., Constructs 1, 3, 4), chosen as LgBit for easy assay readout. Inter-domain linkers were designed to be flexible: the intracellular linker following the positive-inside rule and the extracellular linker having net negative charge. Out of 6 designs, 3 constructs (#2-4) were predicted in silico with strong confidence to adopt the desired type II topology54and surface localization33, thus chosen to test in cells. As well, a negative control construct (#1) anticipated to yield type I insertion using the hemagglutinin signal sequence (ssHA) was included. Each should position a FLAG-tag in the expected extracellular terminus, also varying in whether the LgBit module present in intracellular or extracellular spaces (Table 4).
[0063] Localization of the constructs was tested by immunofluorescence upon transfection into HEK293T cells (polycistronic constructs co-expressing mCherry) and staining with or without membrane permeabilization to assess FLAG accessibility’ and thus construct orientation. The ssHA Construct-1 stains minimally when non-permeabilized, but presents a strong signal for anti-FLAG antibody when penneabilized, indicating intracellular localization i.e., Type I plasma membrane insertion - aligned with expectations. By contrast, Constructs-2 and -3 hosting the ssPolarl both show surface staining in non-permeabilize cells, indicating a significant amount of extracellular FLAG epitope at the plasma membrane and Type II oriented TM protein as designed. Construct-4 with ssPolar2 showed similar localization pattern as construct 1: minimal extracellular FLAG presentation. Construct-2 shows superior surface expression and trafficking compared with Construct-3, lacking a folded domain at its N-terminus and having extracellular LgBit. However, Construct-3 was chosen for further studies due to intracellular positioning of LgBit and compatibility with intracellular SmBit of our existing “Target” constructs.
[0064] To assess the viability of Construct-3 hosting CHAMP for memPPI, we measured its compatibility to interact with EpoR. As expected, the EpoR-CHAMP pair resulted in high normalized PPI signal while negative control CLR-CHAMP did not. To show that protein contributing to PPI signal is at the cell membrane and not in intracellular aggregates, we made protein constructs where SmBit and LgBit component are expected in the extracellularspace. Next, we performed a 3-step NanoBiT method in which surface-exposed LgBit is first detected by adding HiBit without digitonin permeabilization, then cells are permeabilized with sub-lytic concentrations of digitonin to detect whole cell LgBit content as before, revealing the ratio correctly oriented protein. As expected, LgBit-HiBit luminescence of ‘‘Binders” with extracellular LgBit (e.g., modified e-CD8a for Type I and CHAMP in Type II-construct-2) showed high signal without digitonin permeabilization and are minimally perturbed by digitonin addition, confirming the topology’ and that most activity comes from membrane surface-localized proteins. Finally, we measured NanoBiT PPI for pairings of “Target” and “Binder constructs where SrnBit and LgBit are mutually oriented in extracellular or intracellular spaces, or oriented on opposite sides of the plasma membrane. Raw NanoBiT luminescence for mEpoR CHAMP with extracellularly-oriented SmBit and LgBit reaches the level of positive control CLR-RAMP1 (intracellular), while mEpoR-CHAMP pairings with mixed intracellular / extracellular SmBit / LgBit topologies show much lower signal, suggesting the former interaction is occurring between correctly oriented membrane surface localized protein complexes.
[0065] These results validate Type II single-pass membrane protein constructs compatible with intracellular or extracellular forms of NanoBiT PPI assays. Having ssPolarl as the N-terminal leader sequence appears to be a key element determining Type II trafficking. However, a polar N-terminus is alone insufficient, given ssPolar2’s inability to direct this topology. These novel constructs should provide topological control over measurements of TM span interactions for natural or synthetic sequences and be applicable across protein or cellular engineering applications.Example 4. Modulating membrane protein expression, trafficking, and sub-cellular PPI specificity
[0066] In the second case study of topological control, we characterized variants of the RAMP1 protein, which interacts with a broad scope of GPCRs48. Structures of the CLR-RAMP1 complex (CGRPR, PDB: 6E3Y) revealed that both their TM and extracellular domains directly interact4’’, whereas the glucagon receptor’s structure with RAMP2 (PDB: 8FU6) lacked the latter’s TM domain density, suggesting extracellular domains dominate interactions56. To test whether RAMP1 TM span alone can interact with CLR, we generated a minimal RAMP1TM construct fused to LgBit (Fig. 2, panel a ). However, from immunofluorescence microscopy and HiBit-LgBit luminescence, we found that this constructwas not expressed enough to appropriately assess this PPI (Fig. 2, panel b). To create a well- expressed construct displayed RAMPl’s isolated TM domain, we hypothesized that a globular domain preceding RAMPl’s TM span, akin to its extracellular domain, may aid its correct production and trafficking. We devised and co-expressed 2 chimeras of RAMPl’s TM domain fused at its N-terminus to InnTag5 and HALO-tag globular domains and LgBit at its C -termini, finding both expressed at levels similar to WT RAMP1 with sufficient plasma membrane localization (Fig. 2, panels a and b). Raw and expression normalized NanoBiT luminescence of ImiTag5-RAMPlTM and HALO-RAMP 1TM with CLR-SmBit exhibited 2- fold lower signal than WT CLR-RAMP1 (Fig. 2, panel e), being only slightly higher than non-specific interaction pairs CLR-CD8a or CLR-CD4 NanoBiT signals (3-fold less than CLR RAMP1, Fig. 1, panel d). Thus, RAMPl’s TM span may contribute to its interaction with CLR, but RAMPl’s extracellular domain appears more critical for its complexation with CLR. This experiment highlights that addition of innocuous stable soluble domains is a useful strategy to facilitate the expression and trafficking of minimal TM domains, allowing NanoBiT measurements that isolate TM domains’ contributions to protein interactions.
[0067] Similar improvement of expression and trafficking is observed for eVgL’s fusion to extracellular domains. In testing the de novo designed eVgL expressed as a direct LgBit fusion lacking a soluble domain, we find that, despite bearing a ssHA signal peptide, it appears to localize primarily to intracellular membranes rather than the cell surface (Fig. 2, panels c and d). Its exceptionally high expression level and presence of intracellular puncta by microscopy cast doubt on whether its high PPI NanoBiT signal arises mainly from homo- pentamerization rather than overexpression or intracellular aggregation artifacts. As for RAMP1, we tested fusion of InnTag5 and HALO-tag to eVgL-LgBit- tagged “Binder” constructs for improved plasma membrane trafficking (Fig. 2, panels c and d). Both domains drastically increased eVgL’s cell surface localized fraction, although overall expression levels were reduced (Fig. 2, panels f and g).
[0068] With untagged eVgL-SmBit as the “Target”, HALO-tag-eVgL and InnTag5-eVgL exhibit similar NanoBiT PPI signal after normalization by protein expression level (HiBit method), indicating the complex is formed within the intracellular compartments (Fig. 2, panel f). Conversely, w'hen InnTag5-eVgL-SmBit is used as the “Target” construct, the substantially surface localized InnTag5- & HALO-tag-eVgL “Binder” constructs have largely higher normalized PPI signal compared to the intracellularly localized untagged eVgL-LgBit, whose normalized luminescence appears to be a true negative non-interacting pair (InnTago-eVgL-SmBit and eVgL LgBit) (Fig. 2, panel g). As observed for Type II CHAMP EpoR, studies with eVgL constructs indicate NanoBiT exhibits some selectivity for co-localization to the same cell membrane compartment to detect a PPI complex. Together, these case studies exemplify folding considerations and protein engineering strategies for obtaining appropriate constructs with suitable trafficking for natural or synthetic TM interactions, including how to host and display TM spans in mammalian cells when engineering de novo protein complexes.
[0069] Finally, we monitored whether the expression level for the ‘“Target” is stable or significantly varied when co-expressed across a panel of “Binder” constructs, for 1 example CLR-SmBit paired with these different RAMP1 constructs (Fig. 2, panel a) measured by immunofluorescence microscopy. While the observed expression level of the binder was observed to vary' over an order of magnitude (> 10-fold) when “Binder” DNA was varied from 5 to 80 ng, the CLR expression under these conditions (fixed 20 ng) varied modestly but was general stable across constructs at each level of “Binder” DNA. The change in “ Target” level due to the “Binder” construct in most cases is low, at most 50%, relative to the drastic many- fold (well over 100%) increase in raw NanoBiT luminescence observed for wildtype RAMP1 and CLR. These data reflect that, for this positive control complex, the PPI causes elevated luminescence and not an increase in “Target” expression (thus complementation) resulting from “Binder” co-expression. However, other “Targets” may behave differently, and users should use orthogonal methods to track the “Target” level and not assume it as constant while the “Binder” protein level is monitored by the HiBit-LgBit method.Example 5. MemPPI can distinguish specificity between close GPCR homolog complexes
[0070] We next tested the applicability of memPPI to detect subtle chemical changes by assaying interaction selectivity between natural protein homologs: the TM adaptor complex of CD69 that favors S1PR1 versus S1PR244. Expression normalized NanoBiT of CD69 was 2-fold greater when paired with S1PR1 than with to S1PR2 (n=3, p<0.001), aligned with previous co-immunoprecipitation experiments finding CD69 interacts more strongly with SI PR I44. W e then assessed whether memPPI could discern the consequences of CD69 mutations on S1PR1 interaction as previously described44. NanoBiT of the V48F-V49F CD69 double mutant paired wuth S1PR1 showed significant 30% decrease in the normalized signal relative to WT (n=3, p<0.05). Finally, SlPRl-like mutations were introduced to S1PR2 to potentially reconstitute a synthetic interface with CD6944. The L157V-L168M variant showed a modest but consistent 15% increase in PPI signal relative to the WT SlPR2’sinteraction < 3. p<0.()01 ). This case study demonstrates that the memPPI assay can detect subtle changes in interactions propensity observed between known PPI and can probe selectivity, illustrating a broad potential scope of use from interface mapping to interactome analysis this assay may be suited for.Example 6. Jn situ expression normalization enables scalable high-throughput protein engineering screens
[0071] Finally, we leveraged the protein normalization step of our 2-step memPPI and evaluated its performance in an accelerated screening pipeline testing a panel of computationally designed TM domains with intended Type II topology as binders for a GPCR of interest, referred to as GPCR-A. From a pool of synthetic TM domains, cloned in the aforementioned Type II Construct-3 in E. coll, individual clones were subject to plasmid amplification by Rolling Circle Amplification (RCA). Unpurified diluted RCA product was used directly to co-transfect library members alongside 20 ng GPCR-A-SmBit to HEK293T cells to read the NanoBiT PPI alongside protein expression levels (Fig. 3, panel a). Thus, the laborious step of plasmid purification was bypassed and resulting protein expression level is determined in situ in lieu of knowing the precise “Binder” genetic material delivered. RCA products can be sequenced individually or from pool after barcoding and next-generation sequencing (NGS). Interestingly, for the included true positive control PPIs EpoR-CHAMP and CER-RAMP1, transfection with the RCA product results in the same normalized memPPI signal as delivering approximately 20ng and lOng, respectively, of “Binder” plasmid as input DNA, despite considerable differences in raw luminescence values (Fig. 3, panels b and c). Across Target-Binder pairs, although EgBit-HiBit quantification shows order of magnitudes of difference in binder expression levels, the normalized PPI signal compared between plasmid- versus RCA-based transfection is highly similar with excellent linear correlation, R?:::0.98 (Fig. 3, panels c and d). These results establish that normalization by “Binder” protein expression yields self-consistent PPI measurements across an extensive dynamic range and that relative protein levels can be determined in situ on the fly in a protein engineering screen. This adaptability facilitates memPPI’ s integration into high-throughput platforms and opens opportunities for protein design and engineering screens of synthetic TM domain complexes.Example 7. Some exemplified materials and methods
[0072] Expression plasmids: Membrane proteins were all expressed from pcDNA3.1 plasmids under CMV promoter. SmBit-tagged “Target” protein genes such as GPCR were expressed tagged at the N-terminus with the Hemagglutinin signal sequence (ssHA) followed by epitope tags, either Myc alone or 2x HA-tag followed by myc-tag. Fusion to SmBit at the C-terminus was mediated by 10 amino acid Gly-Ser linker.
[0073] For constructs of “Binder” LgBit tagged proteins, LgBit was fused to the membrane at its expected intracellular termini and with a FLAG-tag at the expected extra¬ cellular exposed termini. Variations of protein expression constructs were cloned including additional bl-terminal or C-terminal soluble domains such as HALO, InntagS, or those data mined from natural protein.
[0074] Domain structures and sequences of various Target and Binder constructs expressing a target protein or a binder protein respectively fused to SmBit and LgBit are listed in Tables 2-4. Sequences of LgBit and SmBit used in the constructs, as well as the sequence of HiBit peptide, are listed in Table 1.
[0075] Primary NanoBiT: HEK 293T cells (ATCC) were grown at 37°C with 5% CO2in DMEM (Gibco) supplemented with 10% fetal bovine serum (HyClone) and 1% (v / v) Penicillin-Streptomycin (VWR). For luminescence measurements, cells were transfected with Hpofectamine 3000 in Opti-MEM(Invitrogen) and cultured in 24 hours in white-walled, opaque-bottom 96-well plates (Costar) pre-treated overnight with 0.1 mg / m poly-D-lysine (VWR). For the experiments involving S1PR1 and 2, HEK293T cells were progressively adapted to grow in 10% charcoal-stripped fetal bovine serum (Gibco ), and grown in this media thereafter. All other experimental setups were carried out the same as for WT HEK293T cells.
[0076] For luminescence measurement experiments in 96-well plates, cells were retro- cotransfected with two plasmids encoding Target-SmBit and Binder-LgBit, respectively. The DNA amounts used were specified in each figure or caption. For transfection in each well, DNA was first mixed with 5 pL Opti-MEM (Gibco) containing 0.2 pL Lipo P3000 In a separate tube, 0.3 pL Lipofectamine 3000 reagent was added to another 5 pL of Opti-MEM. The two solutions were combined, incubated at room temperature for 15 minutes, and then added to the cell suspension containing 0.05x106 cells m 100 pL DMEM per well. Cells were cultured for 22-24 hours before luminescence signals were measured.
[0077] After 22-24 hours of expression, the culture media was replaced with 100 pL of scrum-frcc CO -independent media (Gibco) containing 10 pM furimazine (ChemShuttle).Luminescence was immediately recorded every minute for 15-20 minutes using a Tecan SPARK plate reader without the lid with 250ms integration time.
[0078] Two-step NanoBiT: Subsequent to the primary NanoBiT assay, a HiBit detection solution was prepared by mixing 10 pL of 1% digitonin (Research Products International Corp) and 1 pl. of 20 mg / mL HiBit peptide (Abclonal) in 1 rnL of CO2 -independent media. For each well, 10 pL of this solution was added, and luminescence was measured every minute for another 15-20 minutes with OD1 attenuation, again without the lid.
[0079] rFluc luminescence measurement: For comparison with HiBit normalization, a 96-well plate was divided into two halves, with each Target-Binder pairing undergoing NanoBiT / HiBit measurement in one half, and NanoBiT / rFluc in the other. NanoBiT measurement of the whole plate was performed as described above. A blank measurement was taken for the rFluc half for signal between 595-700nm with integration time 500ns and no attenuation. 100 pL Steady-Gio reagent (Promega) was added to rFluc half of the plate, followed by addition of digitonin and HiBit peptide to the other half. Luminescence signal for rFluc was measured with same settings as the blank measurement, while HiBit luminescence was measured as described in previous section.[00801 Immunostaining: Cells were seeded at a density of 0.15x106cells per well in 8- well chambered coverslip slides (Ibidi), 0.2x106cells per well in 48-w’dl plates, or 0.05x106cells per well in black-walled, glass-bottom 96-well plates (Cellvis). All plates and slides were pre-coated with 0.1 mg / mL poly-D-lysine (VWR) overnight. Transfection was performed the following day once cells reached 60-70% confluency.
[0081] Cells in the Ibidi U-slide chambers were transfected with 150 ng of Target plasmid and 100 ng of Binder plasmid using 0.5 pL of Lipofectamine P3000 and 0.5 pL of Lipofectamine 3000, each diluted in 12.5 pl, Opti-MEM (Gibco). For cells in 48-wcll plates, 250 ng of type II constructs were transfected with 0.75 L Lipofectamine 3000 and 0.5 L P3000 reagent. For 96-well glass-bottom plates, each well ’as transfected with 50 ng of Target plasmid along with varying amounts of Binder plasmid (200, 100, 50, or 10 ng).
[0082] Transfection complexes were prepared as described previously: DNA w’as first mixed with P3000 reagent in Opti-MEM, followed by addition of diluted Lipofectamine 3000 following the standard commercial protocol. After a 15-minute incubation at room temperature, the complexes were added dropwise to the cells.
[0083] After 22- 24 hours of expression, cells w ere fixed with ice-cold pure methanol for 10 minutes at -20°C or 4% paraformaldehyde for 15 min on ice, washed three times with coldPBS, and blocked with 3% bovine serum albumin (BSA; Fisher) in PBS for 30 minutes at room temperature.
[0084] Primary immunostaining was performed overnight at 4°C using a mouse anti-FLAG M2 antibody (1:1000, Sigma) in 1% BSAZPBS. The next day, cells were washed twice with PBS at room temperature and incubated with an Alexa Fluor 488 or 647-conjugated anti-mouse secondary antibody (1:2000, Life Technologies) in 1% BSA / PBS for 30 minutes at room temperature. For Target and Binder double-staining, cells were further washed three times with PBS, and incubated with anti-HA-rabbit (1:500 Cell Signaling) and AlexaFluor-647 conjugated anti-mouse secondary antibody (1:2000 Life Technologies). After two PBS washes, nuclei were stained with Hoechst (1:10,000, Life Technologies), followed by another PBS wash. Imaging was performed using a Zeiss LSM 780 confocal microscope with 40x and 63x oil immersion objectives or a Nikon T12-E Inverted Microscope with a 20x ELWD objective. All images were processed with Fiji (Image.!).
[0085] RCA-based NanoBiT / HiBit assay: For the RCA reaction, colonies with Binder of interest were resuspended in 50 pL of water each in a 2niL deep 96-well plate. RCA reaction was carried out in a 96-well PCR plate, and final reaction included 1 pL colony resuspension, 2.5 pL 2x reaction buffer (2mM dNTPs, 100 pM exonuclease-resistant random primer, 2x phi29-XT Reaction Buffer), 1.25 pL water, and 0.25 pL phi29-XT polymerase (NEB). All components, minus polymerase, were incubated for 3 minutes at 95°C for lysis of cells. After polymerase addition, reaction was incubated at 42°C for 2.5-3h, followed by 10 min inactivation at 65°C. RCA product was diluted with 35 pL water and stored at -20°C until needed. The rest of the colony resuspension in deep well-plate rvas supplemented with LB media with carbenicillin and incubated overnight at 37°C. Glycerol was added to a final concentration of 25%, and plate was stored at -80“C.[00861 Cotransfection was carried out as detailed above for the transfection for luminescence measurements. Input DNA for each well consisted of 3 pL of water mixed with 2 pl. of diluted RCA product containing the Binder, and 20ng of Target. NanoBiT and HiBit measurements were carried out as explained above. During a screening campaign, identity of wells of interest can be obtained by sequencing the diluted RCA product, growing colonies from the glycerol stock plate, or PCR barcoding the entire plate for Next Generation Sequencing.
[0087] Split-m Venus Assay: WT HEK293T cells were retro-cotransfected with two plasmids encoding Target- Ven usNtenn and Binder-VenusCterm, respectively. The DNAamounts used were specified in each figure or caption. For transfection in each well, DN A was first mixed with 5 pL Opti-MEM (Gibco) containing 0.2 pL Lipo P3000. In a separate tube, 0.3 pL Lipofectamine 3000 reagent was added to another 5 pL of Opti-MEM. The two solutions were combined, incubated at room temperature for 15 minutes, and then added to the cell suspension containing 0.05 l06cells in 100 pL DMEM per well. Cells were cultured for 48 hours before fluorescence signals were measured. The plate reader is set to Bottom Reading with Excitation at 500nm and Emission 530nni.
[0088] Three-step NanoBiT: A HiBit detection solution was prepared by mixing 1 pL of 20 mg / rnL HiBit peptide (Abclonal) in 1 niL of CO2-independent media. A separate solubilization solution was prepared by mixing 10 pL of 1 % digitonin (Research Products International Corp) in 1 mL of CO2-independent media. Following the primary NanoBiT assay, 10 pL of HiBit detection solution was added to each well, and luminescence was measured every minute for an additional 15 minutes with OD1 attenuation, without the lid. Immediately after the last measurement, 10 pL of solubilization solution was added to each well, and luminescence was measured for an additional 15 minutes with the same settings.Table 1. Sequences for LgBit, SmBit, HiBitTable 2. Type 1 “Target” SmBit-tagged constructsGeneral expression constructs and representative examples for SmBit-tagged targets utilizing ssHA, HA / myc epitope tag(s). TM protein, and SrnBit domain. Examples of eVgL with and without InnTag5 soluble domains are shown. Specific examples of CLR, S1PR1, S1PR2, and mEpoR target constructs are listed. Throughout Tables 2-4, “GS” and similar 2 or 3 letter designations (e.g., PSG) in the constructs’ names indicate short GS-rich peptides or spacers that separate the various components.Table 3. “Binder” LgBit-tagged constructsExpression constructs for Typel LgBit-tagged targets utilizing ssHA, FLAG epitope tag, TM protein, and LgBit domain. Examples of eVgL and RAMP1 with and without soluble domains (InnTag5 or HALO-tag) are shown. Specific examples of CD8 and CD4 “binder” constructs are listed.Expression constructs for Type II LgBit-tagged target utilizing ssPolarl, FLAG, ALFA epitope tag, InnTag5 and LgBit domain. Specific example of CD69 transmembrane domain (40-64) is listed.Table 4. Type II constructsExpression constructs for TypeII LgBit-tagged constructs utilizing ssHA / ssPolar1, FLAG and ALFA epitope tag, type II TM protein CHAMP, HsCLC2B extracellular domain, and LgBit domain.
[0089] Some cited references1 Westerfield, J. M. & Barrera, F. N. Membrane receptor activation mechanisms and transmembrane peptide tools to elucidate them. Journal of Biological Chemistry’ 295, 1792-1814 (2020).2 Duart, G., Grau, B., Mingarro, I. & Martinez-Gil, L. Methodological approaches for the analysis of transmembrane domain interactions: A systematic review. Biochim Biophys Acta Biomembr 1863, 183712 (2021). https: / / doi.org / 10.1016 / j.bbamem.2021.183712 3 Zoned, A., Duneau, J. P. & Cascales, E. Bacterial One- and Two-Hybrid Assays to Monitor Transmembrane Helix Interactions. Methods Mol Biol 2715, 259-271 (2024). https: / / doi.org / 10.1007 / 978-1-0716-3445-5_174 Dixon, A. S. et al. NanoLuc® Complementation Reporter Optimized for Accurate Measurement of Protein Interactions in Cells. ACS Chem Biol 11, 400-408 (2016). https: / / doi.org / 10.1021 / acschembio.5b007535 Langosch, D., Brosig, B., Kolmar, H. & Fritz, H. J. Dimerisation of the glycophorin A transmembrane segment in membranes probed with the ToxR transcription activator. J Mol Biol 263, 525-530 (1996). https: / / doi.org / 10.1006 / jmbi.1996.05956 Kolmar, H. et al. Membrane insertion of the bacterial signal transduction protein ToxR and requirements of transcription activation studied by modular replacement of different protein substructures. EMBO J 14, 3895-3904 (1995).https: / / doi.org / 10.1002 / j.1460-2075.1995.tb00061.x7 Armstrong, C. R. & Senes, A. Screening for transmembrane association in divisome proteins using TOXGREEN, a high-throughput variant of the TOXCAT assay. Biochim Biophys Acta 1858, 2573-2583 (2016). https: / / doi.org / 10.1016 / j.bbamem.2016.07.008 8 Joce, C., Wiener, A. A. & Yin, H. Multi-Tox: application of the ToxR- transcriptional reporter assay to the study of multi-pass protein transmembrane domain oligomerization. Biochim Biophys Acta 1808, 2948-2953 (2011).https: / / doi.org / 10.1016 / j.bbamem.2011.07.0089 Lindner, E. & Langosch, D. A ToxR-based dominant-negative system to investigate heterotypic transmembrane domain interactions. Proteins 65, 803-807 (2006).https: / / doi.org / 10.1002 / prot.2122610 Su, P. C. & Berger, B. W. A novel assay for assessing juxtamembrane and transmembrane domain interactions important for receptor heterodimerization. J Mol Biol 425, 4652-4658 (2013). https: / / doi.org / 10.1016 / j.jmb.2013.07.02211 Steindorf, D. & Schneider, D. In vivo selection of heterotypically interacting transmembrane helices: Complementary helix surfaces, rather than conserved interaction motifs, drive formation of transmembrane hetero-dimers. Biochim Biophys Acta Biomembr 1859, 245-256 (2017). https: / / doi.org / 10.1016 / j.bbamem.2016.11.01712 Schneider, D. & Engelman, D. M. GALLEX, a measurement of heterologous association of transmembrane helices in a biological membrane. J Biol Chem 278, 3105-3111 (2003). https: / / doi.org / 10.1074 / jbc. M20628720013 Russ, W. P. & Engelman, D. M. TOXCAT: a measure of transmembrane helix association in a biological membrane. Proc Natl Acad Sci U S A 96, 863-868 (1999).https: / / doi.org / 10.1073 / pnas.96.3.86314 Dawson, J. P., Weinger, J. S. & Engelman, D. M. Motifs of serine and threonine can drive association of transmembrane helices. J Mol Biol 316, 799-805 (2002).https: / / doi.org / 10.1006 / jmbi.2001.535315 Russ, W. P. & Engelman, D. M. The GxxxG motif: a framework for transmembrane helix-helix association. J Mol Biol 296, 911-919 (2000). https: / / doi.org / 10.1006 / jmbi.1999.348916 Schanzenbacli, C., Schmidt, F. C., Breckner, P., Teese, M. G. & Langosch, D. Identifying ionic interactions within a membrane using BLaTM, a genetic tool to measure homo- and heterotypic transmembrane helix-helix interactions. Sci Rep 7, 43476 (2017). https: / / do i. org / 10.1038 / srep4347617 Julius, A., Laur, 1... Schaiizenbach, C. & Langosch, D. BLaTM 2.0, a Genetic Tool Revealing Preferred Antiparallel Interaction of Transmembrane Helix 4 of the DualTopology Protein EmrE, J Mol Biol 429, 1630-1637 (2017).https: / / doi.org / 10.1016 / j.jmb.2017.04.00318 Elazar, A. et al. Mutational scanning reveals the determinants of protein insertion and association energetics in the plasma membrane. Elife 5 (2016).https: / / doi.org / 10.7554 / eLife.1212519 Lis, M. & Blumenthal, K. A modified, dual reporter TOXCAT system for monitoring homodimerization of transmembrane segments of proteins. Biochem Biophys Res Commun 339, 321-324 (2006). https: / / doi.org / 10.1016 / j.bbrc.2005.11.02220 Remy, I. & Michnick, S. W. Clonal selection and in vivo quantitation of protein interactions with protein-fragment complementation assays. Proc Natl Acad Sci U S A 96, 5394-5399 (1999). https: / / doi.org / 10.1073 / pnas.96.10.539421 Rossi, F., Charlton, C. A. & Blau, H. M. Monitoring protein-protein interactions in intact eukaryotic cells by beta-galactosidase complementation. Proc Natl Acad Sci U S A 94, 8405-8410 (1997). https: / / doi.org / 10.1073 / pnas.94.16.840522 Paulmurugan, R., Umezawa, Y. & Gambhir, S. S. Noninvasive imaging of protein¬ protein interactions in living subjects by using reporter protein complementation and reconstitution strategies. Proc Natl Acad Sci U S A 99, 15608-15613 (2002). https: / / doi.org / 10.1073 / pnas.24259429923 Kodama, Y. & Hu, C. D. An improved bimolecular fluorescence complementation assay with a high signal-to-noise ratio. Biotechniques 49, 793-805 (2010).https: / / doi.org / 10.2144 / 00011351924 Romei, M. G. & Boxer, S. G. Split Green Fluorescent Proteins: Scope, Limitations, and Outlook. Annu Rev Biophys 48, 19-44 (2019). https: / / doi.org / 10.1146 / annurev-biophys-051013-02284625 Yao, Z. et al. Split Intein-Mediated Protein Ligation for detecting protein -protein interactions and their inhibition. Nat Commun 11, 2440 (2020).https: / / doi.org / 10.1038 / s41467-020-16299-126 Johnsson, N. & Varshavsky, A. Split ubiquitin as a sensor of protein interactions in vivo. Proc Natl Acad Sci U S A 91, 10340-10344 (1994).https: / / doi.org / 10.1073 / pnas.91.22.1034027 Tebo, A. G. & Gautier, A. A split fluorescent reporter with rapid and reversible complementation. Nat Commun 10, 2822 (2019). https: / / doi.org / 10.1038 / s41467-019-10855-028 Shekhawat, S. S. & Ghosh, I. Split-protein systems: beyond binary protein-protein interactions. Curr Opin Chem Biol 15, 789-797 (2011).https: / / doi. org / 10.1016 / j.cbpa.2011.10.01429 Cabantous, S., Terwilliger, T. C. & Waldo, G. S. Protein tagging and detection with engineered self-assembling fragments of green fluorescent protein. Nat Biotechnol 23, 102-107 (2005). https: / / doi.org / 10.1038 / nbt104430 Magliery, T. J. et al. Detecting protein-protein interactions with a green fluorescent protein fragment reassembly trap: scope and mechanism. J Am Chem Soc 127, 146-157 (2005 ). https: / / doi.org / 10.1021 / ja046699g31 Duart, G. et al. Computational design of BclxL inhibitors that target transmembrane domain interactions. Proc Natl Acad Sci U S A 120, e2219648120 (2023).https: / / doi. org / 10.1073 / pnas.221964812032 Botta, J., Bibic, L., Killoran, P., McCormick, P. J. & Howell, L. A. Design and development of stapled transmembrane peptides that disrupt the activity of G-protein -coupled receptor oligomers. J Biol Chem 294, 16587-16603 (2019).https: / / doi.org / 10.1074 / jbc.RA119.00916033 Gales, C. et al. Real-time monitoring of receptor and G-protein interactions in living cells. Nat Methods 2, 177-184 (2005). https: / / doi.org / 10.1038 / nmeth74334 Angers, S. et al. Detection of beta 2-adrenergic receptor dimerization in living cells using bioluminescence resonance energy transfer (BRET). Proc Natl Acad Sci U S A 97, 3684-3689 (2000). https: / / doi.org / 10.1073 / pnas.97.7.368435 Barnea, G. et al. The genetic design of signaling cascades to record receptor activation. Proc Natl Acad Sci U S A 105, 64-69 (2008).https: / / doi.org / 10.1073 / pnas. O71048710536 Kim, M. W. et al. Time-gated detection of protein -protein interactions with transcriptional readout. Elife 6 (2017). https: / / doi.org / 10.7554 / eLife.3023337 Petschnigg, J. et al. The mammalian-membrane two-hybrid assay (MaMTH) for probing membrane-protein interactions in human cells. Nat Methods 11, 585-592 (2014). https: / / doi.org / 10.1038 / nmeth.289538 Xu, Y., Piston, D. W. & Johnson, C. H. A bioluminescence resonance energy transfer (BRET) system: application to interacting circadian clock proteins. Proc Natl Acad Sci U S A 96, 151-156 (1999). https: / / doi.org / 10.1073 / pnas.96.1.15139 Li. B. et al. High-Throughput NanoBiT -Based Screening for Inhibitors of HIV- 1 Vpu and Host BST-2 Protein Interaction. Int J Mol Sci 22 (2021).https: / / doi.org / 10.3390 / ijms2217930840 Peach, C. J., Kilpatrick, L. E., Woolard, J. & Hill, S. J. Use of NanoBiT and NanoBRET to monitor fluorescent VEGF-A binding kinetics to VEGFR2 / NRP1 heteromeric complexes in living cells, Br J Pharmacol 178, 2393-2411 (2021).https: / / doi.org / 10.1111 / bph.1542641 Lay, C. S., Kilpatrick, L. E., Craggs, P. D. & Hill, S. J. Use of NanoBiT and NanoBRET to characterise interleukin-23 receptor dimer formation in living cells. Br J Pharmacol 180, 1444-1459 (2023). https: / / doi.org / 10.1111 / bph.1601842 Inoue, A. et al. Illuminating G-Protein-Coupling Selectivity of GPCRs. Cell 177, 1933-1947 el925 (2019). https: / / doi. Org / 10.1016 / j.cell.2019.04.04443 Liang, Y. L. et al. Cryo-EM structure of the active, G(s)-protein complexed, human CGRP receptor. Nature 561, 492-497 (2018). https: / / doi.org / 10.1038 / s41586-018-0535-y 44 Chen, H., Qin, Y., Chou, M., Cyster, J. G. & Li, X. Transmembrane protein CD69 acts as an S1PR1 agonist. Elife 12 (2023). https: / / doi.org / 10.7554 / eLife.8820445 Mravic, M. et al. Packing of apolar side chains enables accurate design of highly stable membrane proteins. Science 363, 1418-1423 (2019).https: / / doi.org / 10.1126 / science.aav754146 Kratochvil, H. T. et al. Transient water wires mediate selective proton transport in designed channel proteins. Nature Chemistry 15, 1012-1021 (2023).https: / / doi.org / 10.1038 / s41557-023-01210-447 Mravic, M. et al. De novo-designed transmembrane proteins bind and regulate a cytokine receptor. Nature Chemical Biology (2024). https: / / doi.org / 10.1038 / s41589-024-01562-z48 Kotliar, I. B. et al. Multiplexed mapping of the interactome of GPCRs with receptor activity-modifying proteins. Sci Adv 10, eado9959 (2024).https: / / doi.org / 10.1126 / sciadv.ado995949 Huttlin, E. L. et al. Dual proteome-scale networks reveal cell-specific remodeling of the human interactome. Cell 184. 3022-3040 e3028 (2021).https: / / doi.org / 10.1016 / j. cell.2021.04.01150 Devaraneni, P. K. et al. Stepwise insertion and inversion of a type II signal anchor sequence in the ribosome-Sec61 translocon complex. Cell 146, 134-147 (2011). https: / / doi.org / 10.1016 / j.cell.2011.06.00451 Dou, D., da Silva, D. V., Nordholm, J., Wang, H. & Daniels, R. Type II transmembrane domain hydrophobicity dictates the cotranslational dependence for inversion. Mol Biol Cell 25, 3363-3374 (2014). https: / / doi.org / 10.1091 / mbc. E14-04-087452 Georgieva, M. V. et al. Inntags: small self-structured epitopes for innocuous protein tagging. Nat Methods 12, 955-958 (2015). https: / / doi.org / 10.1038 / nmeth.355653 Murase, T. et al. Identification of soluble forms of lectin-like oxidized LDL receptor- 1. Arterioscler Thromb Vase Biol 20, 715-720 (2000).https: / / doi.org / 10.1161 / 01.atv.20.3.71554 Tsirigos, K. D., Peters, C., Shu, N., Kall, L. & Elofsson, A. The TOPCONS web server for consensus prediction of membrane protein topology and signal peptides. Nucleic Acids Res 43, W401-407 (2015), https: / / doi.org / 10.1093 / nar / gkv48555 Thumuluri, V., Almagro Amienteros, J. J., Johansen, A. R., Nielsen, H. & Whither, O. DeepLoc 2.0: multi-label subcellular localization prediction using protein language models. Nucleic Acids Res 50, W228-W234 (2022). https: / / doi.org / 10.1093 / nar / gkac278 56 Krishna Kumar, K. et al. Negative allosteric modulation of the glucagon receptor by RAMP2. Cell 186, 1465-1477 el418 (2023). https: / / doi.org / 10.1016 / j.cell.2023.02.028 57 Levy, S. & Shoham, T. The tetraspanin web modulates immune-signalling complexes. Nat Rev Immunol 5, 136-148 (2005). https: / / doi.org / 10.1038 / nril548sfc
[0090] Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to one of ordinary skill in the art in light of the teachings of this invention that certain changesand modifications may be made thereto without departing from the spirit or scope of the appended claims.
[0091] All publications, databases, GenBank sequences, patents, and patent applications cited in this specification are herein incorporated by reference as if each was specifically and individually indicated to be incorporated by reference.
Claims
1. WHAT IS CLAIMED IS:
1. A method for quantitatively determining membrane protein-protein interaction between a target protein and a binder protein in a mammalian cell, comprising (1) transfecting a target vector and a binder vector into the mammalian cell, wherein the target vector and the binder vector respectively encode the target protein and the binder protein that are respectively fused to the first and the second complementary enzyme portions of a split enzyme, (2) measuring a first enzyme activity in the cell after culturing the transfected cell under appropriate conditions to form a reconstituted enzyme from the first and the second complementary enzyme portions, (3) introducing into the same cell a variant peptide that outcompetes the first complementary enzyme portion for binding to the second complementary enzyme portion and displaces the first complementary enzyme portion from the reconstituted enzyme, (4) quantifying in situ binder protein level in the same cell by measuring a second enzyme activity from a reconstituted enzyme formed from the variant peptide and the second complementary enzyme portion in the cell, and (5) normalizing the first enzyme activity with the quantified binder protein level; thereby quantitatively determining membrane protein-protein interaction between the target protein and the binder protein in the mammalian cell.
2. The method of claim 1, wherein the split enzyme is NanoLuc®, and the first and second measured enzyme activities are luminescence signals.
3. The method of claim 1, wherein the first and the second complementary enzyme portions of the split enzyme are respectively SmBit and LgBit of the NanoBiT® split enzyme.
4. The method of claim 1, wherein the first complementary enzyme portion comprises SEQ ID NO:2 or a conservatively modified variant thereof, and the second complementary' enzyme portion comprises SEQ ID NO:1 or a conservatively modified variant thereof.
5. The method of claim 1, wherein the variant peptide comprises SEQ ID NO:3 (HiBit) or a conservatively modified variant thereof.
6. The method of claim 1, wherein the variant peptide is introduced into the cell in the presence of a non-lytic cell permeabilizing agent.
7. The method of claim 6, wherein the cell permeabilizing agent (i) allows entry of the variant peptide into the cell, and (ii) itself does not disrupt the reconstituted enzyme formed from the first and the second complementary enzyme portions.
8. The method of claim 6, wherein the cell permeabilizing agent is digitonin.
9. The method of claim 1, wherein the target protein is fused to the first complementary enzyme portion as shown in SEQ ID NO:2, the binder protein is fused to the second complementary enzyme portion as shown in SEQ ID NO:1, and the variant peptide comprises the amino acid sequence set forth in SEQ ID NO:3.
10. The method of claim 1, wherein the target protein is a GPCR, and the binder protein is an engineered or design protein.
11. The method of claim 1, wherein the target vector and the binder vector each further encode one or more peptide tags that are fused to the N-terminus or the C- tenninus of the protein and / or the complementary enzyme portion encoded by each of the two vectors.
12. The method of claim 1, wherein the mammalian cell is the HEK293T cell line.
13. The method of claim 1, wherein normalizing the first enzyme activity with the quantified binder protein level comprises dividing the first enzyme activity by the quantified binder protein level.
14. The method of claim 1, further comprising normalizing the first enzyme activity with level of the target protein in the cell.
15. The method of claim 14, wherein level of the targe t protein in the cell is quantified via co-transcription of a reporter molecule from the target vector.
16. The method of claim 15, wherein the reporter molecule is red firefly luciferase (rFluc).
17. A kit for quantitatively measuring membrane protein-protein interaction between a target protein and a binder protein, comprising (1) a first vector for inserting a polynucleotide sequence encoding the target protein to express a first fusion protein comprising the target protein and the SmBit fragment of the NanoBiT® split enzyme, (2) a second vector for inserting a polynucleotide sequence encoding the binder protein to express a second fusion protein comprising the target protein and the LgBit fragment of the NanoBiT® split enzyme, and (3) a SmBit variant peptide that (i) has a stronger affinity than SmBit for binding to LgBit and (ii) can displace SmBit from LgBit that is non-covalently bound by SmBit.
18. The kit of claim 17, wherein the SmBit fragment comprises SEQ ID NO:2 or a conservatively modified variant thereof, the LgBit fragment comprises SEQ ID NO:1 or a conservatively modified variant thereof, and the SmBit variant peptide comprised SEQ ID NO:3 or a conservatively modified variant thereof.
19. The kit of claim 17, further comprising a non-lytic cell permeabilizing agent.
20. The kit of claim 19, wherein the cell permeabilizing agent is digitonin.
