Exploring cellular biochemistry with nanobodies

Abstract

Reagents that bind tightly and specifically to biomolecules of interest remain essential in the exploration of biology and in their ultimate application to medicine. Besides ligands for receptors of known specificity, agents commonly used for this purpose are monoclonal antibodies derived from mice, rabbits, and other animals. However, such antibodies can be expensive to produce, challenging to engineer, and are not necessarily stable in the context of the cellular cytoplasm, a reducing environment. Heavy chain–only antibodies, discovered in camelids, have been truncated to yield single-domain antibody fragments (VHHs or nanobodies) that overcome many of these shortcomings. Whereas they are known as crystallization chaperones for membrane proteins or as simple alternatives to conventional antibodies, nanobodies have been applied in settings where the use of standard antibodies or their derivatives would be impractical or impossible. We review recent examples in which the unique properties of nanobodies have been combined with complementary methods, such as chemical functionalization, to provide tools with unique and useful properties.

Tools to detect, visualize, and modulate the properties of proteins are essential to understand the function of the targets recognized and the biology that follows. Introduction of exogenous expression vectors and CRISPR/Cas gene-editing tools provide an unprecedented ability to introduce, alter, or eliminate proteins of choice in cells or intact organisms. These approaches are designed to modify biological processes of interest. Introduction of expression vectors allows production of proteins of choice, WT or mutant, including versions fused with fluorescent proteins or other tags for visualization. Expression of proteins from nonnative loci, as in exogenous expression vectors, or as fusion proteins with tags often alters expression levels, subcellular localization, and biological function. The development of antibody fragments that can interact with and perturb endogenous proteins in cells and organisms without the need for genomic modification would be useful. Nanobodies have unique qualities that make them well-suited for this goal.

Nanobodies, like full-size conventional antibodies, show the affinity and antigen specificity required for specific targeting of molecules of interest, even though they comprise only a single variable region. Nanobodies have several useful features not regularly found in conventional antibodies. These include their small size, the capacity to bind and stabilize specific receptor conformations, and their availability in high yield from bacterial expression systems. Nanobodies have been widely used to target soluble protein antigens or those found at the surface of cells (e.g. for structural studies and imaging applications (for reviews see Refs. 1 and 2). Similar to full-sized antibodies, nanobodies are suitable for flow cytometry, immunoprecipitation, affinity purification, and microscopy (3–7). Although nanobodies are often applied in settings that could just as well use standard, full-size antibodies, we emphasize scenarios where the use of a nanobody provides advantages. In this review, we cover topics including methods for the identification of target-specific nanobodies, functionalization of nanobodies using chemical and enzymatic methods, and the use of nanobodies that engage targets inside or at the surface of the cell as well as viral targets. We cover the development of nanobody-epitope tag pairs and the use of nanobodies in synthetic biology. This review may serve as an accessible resource for scientists looking to identify nanobodies useful for their system of interest. We focus on areas such as nanobody functionalization and synthetic biology, in which methods and use of nanobodies are rapidly evolving.

Screening platforms

Conventional antibodies (Igs) consist of two identical heavy (H) and light (L) chains that pair to form a stably folded protein, with an antigen-binding site to which the two variable (V) domains, V_H and V_L, contribute. Both interchain and intrachain disulfides and N-linked glycosylation are needed for effective assembly of Igs. These requirements preclude the proper assembly of full-size antibodies in the reducing environment of the cytoplasm. Single-chain variable fragments (scFvs) consist of the variable domains from the heavy and light chains, connected by a linker. Although some scFvs can function in the cytoplasm, many scFvs require intrachain disulfides to afford stability and appropriate heavy-light chain pairing. Heavy chain–only antibodies from camelids fold and function in the absence of light chains. These camelid immunoglobulin heavy chains can be shrunk to just their variable domains (Fig. 1) to yield VHHs or nanobodies, which can retain antigen binding in the absence of disulfide bond formation. They can thus be used in the cytosol of live cells, as discussed below. This feature of nanobodies is one of the signature advantages of their application, relative to more conventional alternatives, as discussed below.

Figure 1.

Structures of human and camelid Igs and fragments. Conventional human Igs (i.e. IgG) have been truncated to provide functional fragments (Fab and scFv) that contain variable regions from the light and heavy chains. In the case of the scFv, a linker is required to facilitate appropriate pairing of heavy- and light-chain variable regions. A subset of antibodies from camelids consists of only the heavy chains. Expression of the isolated variable region from heavy chain–only antibodies provides functional single-domain antibodies (VHHs/nanobodies).

Methods to identify nanobodies that bind to targets of interest are essential for their effective deployment (8). Target-binding nanobody clones are usually isolated from screening highly diverse pools of nanobodies. Such pools must be sufficiently large to contain appropriately specific nanobodies, a suitable screening method must be at hand to identify specific binders, and such binders should retain their properties in the relevant contexts, as in the case of cytoplasmic expression or when dealing with membrane proteins. Screening methods that provide nanobodies with desirable functional properties (receptor antagonism, agonism), selectivity for specific target conformations (structural studies, biosensors), and functionality in different subcellular localization (cytoplasmic, cell surface) are in short supply and constitute an area of emphasis for future exploration. Both immunization and screening strategies ought to be designed with the final application(s) of the resulting nanobodies in mind. For example, immunization with unfolded, denatured proteins is more likely to yield reagents that are useful in immunoblotting or immunohistochemistry on fixed samples.

For library construction, B cells from naive or immunized camelids can serve as the point of departure, as can cultured camelid B cells exposed to antigens of interest (9). Purified proteins (10), cells or cell lysates containing antigens of interest (11, 12), or recombinant DNA to induce antigen expression in the host (13, 14) can serve as immunogens. DNA-based immunization has been particularly valuable for the generation of nanobodies against properly folded membrane proteins (15, 16). The diversity of nanobody sequences available in a given pool can be further expanded through mutagenesis. Both natural diversity mutagenesis, in which residues at positions in a nanobody with high diversity in naturally occurring collections of nanobodies are varied (17), and virus-mediated directed evolution (18) can increase diversity and identify novel nanobodies. Important features in the screening approach include the source of the nanobody pool (synthetic versus naive versus immunized library) (8, 19), the mechanism by which nanobody proteins are produced and displayed (phage display versus yeast display versus bacterial display versus ribosome display versus DNA/RNA display) (20), and the method by which antigens of interest are presented for selection (peptide or protein immobilization on solid support versus display of antigens on the cell surface versus labeled soluble antigen) (21). Given the importance of identifying nanobodies that bind to membrane proteins, a variety of approaches have yielded nanobodies that bind to intact, properly folded membrane targets (22–24) and that either block or induce activation (25).

The defining feature of a display method is the mechanism by which the biochemical properties of the nanobodies are linked to the genetic information encoding the nanobodies. The type of display method used also dictates the diversity of the library of nanobody sequences used. Phage display, in which nanobodies are fused in frame with viral proteins for display on the surface of phages—typically an M13 derivative—that encapsulate the relevant DNA sequence, is commonly used to pan for nanobodies (26, 27). Phage display libraries with a diversity of 10⁷ to 10⁸ clones are common. Display-based approaches can also be applied using model single-cell organisms, such as Escherichia coli (20), Staphylococcus sp. (28, 29), and yeast (30, 31).

Yeast display platforms have succeeded in the identification of nanobodies that bind to specific conformations of cell surface proteins, such as G protein–coupled receptors (GPCRs) (30, 31). Bacterial and yeast display platforms of a complexity comparable with that of phage libraries have the advantage that antigen-binding clones can be detected and enriched by flow cytometry (20).

Ribosome display relies on a covalent bond between the nanobody and the encoding RNA chain. Both the translated nanobody sequence and the RNA that encodes it remain tethered to the ribosome when the mRNA lacks a stop codon. Nanobodies that bind to cell membrane proteins in specific conformations were thus obtained (32–34). An approach called RNA display or cDNA display relies on the antibiotic puromycin applied in vitro to enter the ribosomal active site and form a cross-linking covalent bond with the nascent nanobody polypeptide and the encoding RNA sequence to enable selection (35, 36). Library complexity (up to 10¹² unique clones) used in in vitro selection techniques can exceed by far those used in phage, bacterial, and yeast display, but screening then requires multiple rounds of selection to arrive at individual high-affinity binders (33).

Alternative methods of screening have been developed to identify nanobodies that function in their intended environment. In one such method, nanobody-coding sequences (minus the signal peptide) were inserted into lentiviral vectors for expression in the cytoplasm of mammalian cells. Nanobodies that protected cells from a lytic infection with influenza A virus or vesicular stomatitis virus were then identified through enrichment of surviving cells and recovery by PCR of the protective nanobody sequences (37). A similar approach was used to identify a nanobody that protected against porcine reproductive and respiratory syndrome virus (38). The use of a functional readout, like cell survival, ensured that the nanobodies were functional in the cytosol. The size of the library tested in the lentivirus-based approach is similar to that used in phage display (∼10⁷ clones). Another method to identify nanobodies that are functional in the cytosol involves a yeast two-hybrid system, in which propagation of the yeast is contingent on the interaction of a nanobody clone with a target antigen of interest (39). This approach yielded nanobodies that bind HIV VPR and capsid proteins and the hemagglutinin-neuraminidase protein of the Newcastle disease virus (40, 41).

Chemical and enzymatic functionalization

Conventional recombinant expression in bacteria produces nanobodies in high yields, providing ample material for chemical functionalization. Conjugation of nanobodies with fluorescent dyes, small-molecule drugs, oligonucleotides, and other moieties allows complex yet controlled functionalization of nanobodies to extend their application to a wide range of areas, including imaging, therapeutics, and detection, and as delivery agents. Early examples of nanobody functionalization mostly relied on reactivity of cysteine and lysine residues using maleimide (42) and N-hydroxysuccinimide ester–based chemistry (43, 44). Nanobodies typically require the introduction of an unpaired Cys and disulfide reduction prior to labeling. N-Hydroxysuccinimide ester–based labeling lacks selectivity, resulting in heterogeneous mixtures of labeled proteins. Excessive labeling of nanobodies can cause loss of antigen recognition and specificity and result in altered pharmacokinetic properties (45–47). Chemoenzymatic labeling methods, incorporation of unnatural amino acids, and expressed protein ligation are therefore attractive alternatives for the bioconjugation of nanobodies, as will be summarized below (Fig. 2). These methods enable the conjugation of nanobodies with a virtually unlimited selection of chemical cargoes. Even with these advances, it remains difficult to use nanobody conjugates prepared in vitro to address biology inside of live cells because of their membrane impermeability. The development of robust methods for delivery of nanobodies across the cell membrane (48–50) or labeling in cells with minimal background will empower new and powerful applications with conjugates.

Figure 2.

Recent examples of nanobody bioconjugation. A, enzymatic approaches, including OaAEP1 (a), formylglycine-generating enzyme (b), tubulin Butelase and tubulin tyrosine ligase need to be switched in the figure legend (Butelase is panel c, Tubulin tyrosine ligase is panel d) and sortase A (e). B, incorporation of unnatural amino acids by stop codon suppression. C, expressed protein ligation. See the Chemical and Enzymatic Functionalization section for a discussion of strengths and drawbacks of these approaches and associated references.

Oldenlandia affinis asparaginyl endopeptidase (OaAEP1)

Asparaginyl endopeptidases (AEPs) are an increasingly attractive class of enzymes for protein modification. AEPs are nominally Cys proteases that recognize a tripeptide motif, Asn/Asp-Xaa-Yaa, and generate a thioester intermediate C-terminally of the Asn or Asp residue. The thioester can be then attacked by a suitable nucleophile: a dipeptide Gly/Ala-Zaa, where Zaa is a hydrophobic amino acid residue. AEPs of plant origin, such as OaAEP1 from O. affinis, catalyze head-to-tail cyclization of peptides and have been used to prepare cyclic peptides (51–53). The slow kinetics of OaAEP1 limited its application for protein labeling. Substitution of cysteine residue 247 by an alanine residue enhances OaAEP1's catalytic efficiency, making it an efficient tool for protein modification (54). The mutant OaAEP1 has been applied to the modification of nanobodies. By screening different nucleophiles, a Gly-Val dipeptide was identified that readily served as a nucleophile in the ligation reaction, but the product of that reaction (Asn-Gly-Val) was poorly recognized by the enzyme. This yields a ligation product resistant to the reverse reaction, a common shortcoming of enzymatic labeling methods such as sortagging (see “Sortase A”) (55). The use of OaAEP1 with Asn-Gly-Leu–based modified tripeptides allowed efficient modification of the N terminus of a nanobody with the Gly-Val sequence at the N terminus. The use of OaAEP1 thus enabled conjugation of nanobodies with a broad range of different molecules: dyes, lipids, biotin, tetrazine, azide, cyclooctene, small-molecule drugs, PEG oligomers, d-amino acids, and β amino acids (55).

The resistance of the Asn-Gly-Val sequence—the ligation product—to attack by OaAEP1 allowed efficient site-specific modification of a nanobody both at the C and N terminus using the same enzyme, making possible the preparation of doubly functionalized nanobodies (55), as follows. A C-terminal NGL sequence is enzymatically modified first with a Gly-Val modified peptide, yielding a cleavage-resistant product. The N-terminal modification required transient protection of the future N-terminal Gly-Val nucleophile by a TEV protease recognition sequence. Its removal freed up the N terminus for a second OaAEP reaction. Obviously, this method is not limited to modification of nanobodies and can be applied to other proteins of interest, and it might be particularly useful for single-molecule studies. It is also worth noting that, although not observed during nanobody modification, unwanted cleavage within the protein of interest is a side reaction that can occur while using enzyme from the AEP family.

Tub-tag

Tubulin tyrosine ligase (TTL) modifies the C terminus of a protein through conjugation of an unnatural tyrosine residue. This modified tyrosine can be equipped with a wide range of chemical substituents and thus serves to introduce into the newly modified protein important functionalities, such as azides, aldehydes, iodides, alkynes, and dyes. The TTL enzyme recognizes a glutamic acid–rich 14-amino acid sequence, also called Tub-tag (VDSVEGEGEEEGEE), which must be placed at the C terminus of the protein of interest. Labeling with TTL and Tub-tag has been used to modify nanobodies (56) with fluorescent coumarin and biotin derivatives. Recombinantly produced nanobodies equipped with a Tub-tag sequence were likewise used to prepare nanobody-based immunoprecipitation tools and superresolution probes (56, 57).

Formylglycine-generating enzyme

Formylglycine-generating enzymes allow the post-translational modification of cysteine or serine residues within distinct consensus motifs ((C/S)XPXR) to produce formylglycine. Such aldehydes are of particular interest as they can selectively react with hydrazides and amino-oxy moieties for site-specific modification of proteins (58, 59). This approach was used to install an aldehyde motif on two nanobodies that recognize different epitopes on human β₂-microglobulin (60). This method allowed the preparation of C-to-C–linked homodimer nanobodies using an unusual method: aqueous solutions of formylglycine-containing nanobodies and bivalent hydrazide or aminoxy linkers were frozen at −20 °C. This reduction in temperature and freezing drastically increased the rate of dimer formation. Using a sequential approach in which one nanobody was reacted with an excess of the bivalent linker, followed by the addition of the second nanobody, yielded heterodimeric bivalent C-to-C–linked conjugates that were superior in antigen binding relative to C-to-N–linked dimers provided by simple genetic fusion.

Butelase 1

Butelase 1 is also a cysteine protease of the AEP family found in the seed pods of Clitoria ternatea. It recognizes a C-terminal Asn/Asp-containing tripeptide motif, Asn/Asp-His-Val, to form an Asn/Asp-Xaa-Yaa peptide bond, where Xaa can be any amino acid and Yaa is a hydrophobic residue. Butelase 1 is more than 10,000 times faster than other known ligases, with catalytic efficiencies of up to 1,340,000 m⁻¹ s⁻¹ (61, 62). This unique characteristic has made butelase 1 a powerful tool for the preparation of cyclic peptides and proteins or for the direct labeling of a protein (61, 63). Our group used this enzyme together with sortase A to prepare homodimeric and heterodimeric nanobody conjugates connected via DNA linkers (64). The use of dsDNA as linker between two nanobodies imparts rigidity on the linkage and is a straightforward method to control the length of spacing. Such control can be precious for biophysical studies (65). Although butelase 1 is an attractive enzyme for protein engineering, a major drawback remains its availability. Despite new approaches to produce it in bacteria (66, 67), the main source of butelase 1 remains its extraction from seed pods of Clitoria ternatea (62).

Sortase A

Sortase A is an enzyme from Staphylococcus aureus that recognizes the amino acid sequence LPXTGG with high specificity (68). This tag can be placed at the C terminus of the protein of interest but also internal to its sequence, as long as the recognition tag remains accessible (69). After recognition, sortase A cleaves between the threonine and glycine residue to form a thioacyl intermediate. An N-terminal polyglycine equipped with a payload of choice can attack this intermediate and form a new peptide bond (70). We have used this approach to prepare diverse nanobody conjugates: nanobody dimers through C-to-C fusion (71), bispecific nanobodies against GFP and mouse class II major histocompatibility complex (71), nanobodies labeled with radionuclides for PET imaging in vivo (72, 73), nanobody-drug conjugates against B-cell lymphoma (74), and many fluorescently labeled versions (73). The orthogonality of sortase A and butelase 1 allowed the preparation, in a one-pot reaction, of C-C fusion nanobody dimers linked together with PEG and oligonucleotide linkers (64). Proteins equipped with a suitably exposed stretch of Gly residues at the N terminus can be labeled with LPXTG-based peptides in a similar way (75).

Native chemical and expressed protein ligation

Native chemical ligation links unprotected polypeptides through an amide bond that relies on the reaction of a C-terminal thioester with an N-terminal Cys. Expressed protein ligation is based on the naturally occurring splicing of proteins, which proceeds via formation of a thioester intermediate (76). The protein of interest is expressed as a fusion with a mutant version of an intein. Activation with a thiol-containing small molecule, such as 2-mercaptoethanol, generates a C-terminal thioester on the protein of interest, such as a nanobody. Using this thioester, native chemical ligation can be used to attach the desired Cys-containing moiety to the nanobody. The expressed protein ligation approach was used to install on nanobodies two distinct arginine-rich cell-penetrating peptides (CPPs) for comparison. These synthetic CPPs contained d-amino acids, are cyclized, and are therefore impossible to introduce by standard genetic means (48). After attachment of the CPPs to an anti-GFP nanobody, these conjugates were delivered to the interior of the cell with an efficiency of up to 95% of cells, in different cell lines, and at relatively low concentrations (10 μm). Nucleolar localization caused by CPP-nucleic acid interactions could be avoided by attaching the CPP via a disulfide linkage.

Unnatural amino acid incorporation using stop codon suppression

Introduction of an unnatural bio-orthogonally functionalized amino acid can be achieved by using the cellular translational machinery and reassignment of a stop codon (77, 78). This new functional group then enables conjugation of a payload of choice in a site-selective manner. For example, a bifunctional unnatural amino acid derivative was introduced into an anti EGFR nanobody. The modified amino acid (AmAzZLys) contains an aryl amine and an azido group, which allows the conjugation of two different probes in orthogonal and selective fashion. Moreover, the azide moiety was used to perform a photoinduced cross-linking reaction to EGFR upon antigen binding (79). Although attractive in principle, codon reassignment remains technically demanding and often suffers from reduced yields compared with conventional production methods.

Targeting membrane proteins

The high stability, propensity to bind and stabilize specific receptor conformations, and ease of production by recombinant expression make nanobodies well-suited for studying membrane proteins. These characteristics have inspired investigators from a variety of disciplines to use nanobodies. Immense efforts have gone into the identification of membrane protein–binding nanobodies. These efforts have positioned nanobodies as useful reagents for structural studies of membrane proteins using cryo-EM (80, 81) in a way analogous to past work with X-ray crystallography (10) (see also Fig. 3).

Figure 3.

Three examples of nanobodies used as crystallization chaperones. A, in the structure of nucleoporin Nup133 from S. cerevisiae, three nanobodies (shades of orange) generated the critical packing interface necessary to build up the crystal lattice (PDB code 6X04). B, in the structure of the nucleoporin complex of Nup107 and Nup133 from H. sapiens two different nanobodies that bind the Nup107 moiety in separate locations were co-crystallized (PDB code 6X03). C, in the TorsinA-LULL1 complex structure, the nanobody recognizes both binding partners and binds neither TorsinA (white) nor LULL1 (gray) individually (PDB codes 5J1S and 5J1T).

Visualization of the trafficking of membrane proteins using monovalent nanobodies avoids cross-linking–induced artifacts that can arise from the use of bivalent antibodies (82, 83). Nanobodies that bind Igs from mice (84), rabbits (84), pigs (85), and humans (86) can facilitate their use in place of conventional secondary antibodies (87). Nanobodies have also proven useful for affinity purification of delicate membrane protein complexes (6, 88), but these are properties they share with conventional antibodies of similar specificity.

Structural biologists are well-aware of the ability of nanobodies to facilitate crystallization of otherwise difficult-to-crystallize proteins. Nanobodies frequently contribute to crystal-packing contacts that facilitate structure determination (Fig. 3A). Structures where even two nanobodies bind to a single polypeptide have been produced (Fig. 3B). The binding of nanobodies to discontinuous epitopes that span more than one protein can also facilitate crystallization of the nanobody-bound complex (Fig. 3C). The single-domain nature of nanobodies implies that the universe of epitopes they can sample overlaps with, but is distinct from, that of conventional Igs. These unconventional modes of nanobody-antigen interaction should inspire new modes of application in biological settings, such as nanobody-induced target heterodimerization and target-induced nanobody dimerization. Moreover, such nanobodies can stabilize a receptor in a particular, functionally relevant conformation, as shown for various GPCRs and bacterial proteins. Structural studies have benefited from the use of nanobodies as chaperones, most notably for membrane proteins such as GPCRs (89–91). In most cases, GPCR-binding nanobodies generated for structural studies bind the cytoplasmic face of the protein (89, 92, 93) and have visualized different conformers of the GPCRs to which they bind. These features have been exploited by intracellular expression of GPCR-binding nanobodies, as will be discussed below. Nanobodies that bind to the extracellular domain of GPCRs facilitated crystallization of metabolic glutamate receptor-2 (mGluR2) (94) and of the apelin receptor (95). The latter enabled the first identification of nanobodies that activate a GPCR (25, 95). Bacterial membrane proteins have also been trapped by conformation-specific nanobodies, both for structural characterization (96, 97) and for inhibition of their activity (98). Several accounts and reviews cover the development of nanobodies as reagents for structural studies (10) and their use as chaperones (2, 90). A comprehensive overview of structures containing nanobodies was published (99), and a regularly updated database is available online (100).

Conformation-specific nanobodies that bind the cytoplasmic face of GPCRs have been valuable for structural studies of GPCRs in distinct conformations (101–105). When expressed in cells, they can serve as sensors to report on the localization of active receptors. Several mGluR2-binding nanobodies that bind the receptor ectodomain act as positive allosteric modulators and sensitize the receptors to respond to subthreshold levels of glutamate (94, 106). A nanobody that binds the extracellular face of CXCR4 reports on conformational changes induced by small-molecule allosteric modulators (107). The use of nanobodies that lock β₂-adrenergic receptor (β2AR) into active or inactive receptor conformations allowed identification of small-molecule agonists, antagonists, and inverse agonists using binding assays (108). The high yield of nanobodies from recombinant expression systems and the variety of chemical functionalization methods available have enabled the synthesis of conjugates between nanobodies and other complex molecules. A small-molecule ligand for mGlu2R was tethered to the receptor through a GFP-specific nanobody that recognized GFP grafted onto the receptor(109). The use of a photoactivatable ligand provided reversible photocontrol of receptor activation on a time scale of seconds. In another approach, using a combination of enzymatic labeling and click chemistry, a truncated peptide ligand (PTH_1-11) for the parathyroid hormone receptor (PTHR) was conjugated to a PTHR-specific nanobody (Fig. 4) (110). Conjugation of a suboptimal peptide ligand to the nanobody enhanced the potency of the peptide by >100-fold in some cases and improved selectivity for one PTHR subtype over another. In a separate set of studies, the nanobodies that recognize surface proteins on antigen-presenting cells were conjugated to weakly immunogenic peptides, including those with nonnatural amino acids. These conjugates showed dramatically enhanced immunogenicity relative to free peptides in vitro and in vivo (111, 112). It was thus possible to generate antibodies against the otherwise poorly immunogenic cyclotides, heavily disulfide-bonded circular peptides found in plants (113).

Figure 4.

Nanobody-ligand conjugates to target a G protein–coupled receptor. Synthetic fragments of parathyroid hormone were site-specifically linked to nanobodies to provide conjugates (bottom right) with biological activity (EC₅₀) superior to the free ligand (bottom left). Structures are based on human parathyroid hormone receptor (gray) in complex with PTH (orange) (PDB code 6FJ3) and a generic VHH (blue) with complementarity-determining regions highlighted (red) (PDB code 3K1K). The binding of the nanobody to PTHR1 (bottom right) is shown in two possible orientations as the actual site of binding is unknown.

Targeting extracellular proteins

Many nanobodies were developed to target soluble extracellular proteins. The list of such targets continues to expand at a rapid pace. The only nanobody currently approved for clinical use targets the secreted protein von Willebrand factor to treat a blood-clotting disorder (114). More nanobodies will find application as therapeutics and diagnostics, as evidenced by a wealth of preclinical data, some of which are summarized below. Nanobodies' small size endows them with a short circulatory t_½ and superior tissue penetration. These two features in combination distinguish nanobodies from full-size conventional Igs and will determine the investigative and therapeutic areas to which they can be applied. One ongoing challenge is to identify targets, either extracellular or otherwise, for which nanobody application has advantages over conventional antibodies.

Nanobodies can be used to block a variety of biological processes, such as placental growth factor-induced angiogenesis in cancer (115); the action of inflammatory proteins, such as tumor necrosis factor, interleukin-23, granulocyte colony–stimulating factor, and macrophage migration inhibitory factor (116–120); and the action of various toxins and venoms (121–123). A chimeric heavy chain–only antibody consisting of a proprotein convertase subtilisin/kexin type 9–binding nanobody and a portion of the human immunoglobulin heavy chain lowers low-density lipoprotein levels when administered in transgenic rats (124). Nanobodies against neuronal tau (125), human prion protein (126), and α-synuclein (127) provided insight into structural transitions that lead to amyloid formation (126) and served as sensors to differentiate between fibrils at characteristically different stages (127). Nanobodies raised against β₂-microglobulin, a protein for which mutations frequently lead to amyloidosis, have illuminated structural features of aggregation intermediates in mutant versions of β₂-microglobulin (128, 129). They can prevent amyloid formation (130) and remove β₂-microglobulin from blood to treat dialysis-related amyloidosis (131). Gelsolin, a protein for which mutations lead to aberrant proteolytic processing and the formation of amyloidogenic fragments, has likewise been targeted with nanobodies. Nanobodies that bind gelsolin prevent proteolysis, either extracellularly or in the secretory pathway and reduce amyloidosis (132, 133). In vivo delivery of gelsolin-binding nanobodies using a viral vector reduces the amyloid burden in a mouse model (134). Anti-gelsolin nanobodies have also been used to visualize gelsolin amyloid deposits by SPECT/CT (135).

Early evaluation of nanobodies raised against carbonic anhydrase and amylase demonstrated inhibitory activity for several of them (136), encouraging further experiments to deploy nanobodies to modulate enzyme activity. Nanobodies that bind to and inhibit the protease urokinase-type plasminogen activator (uPA) (137–139), which can contribute to cancer metastasis, may find clinical application. Crystallization of complexes between uPA and nanobodies shows how substrate binds and reveals the conformational equilibria that contribute (137, 139). An inhibitory nanobody against matrix metalloprotease-8, one of ∼25 matrix metalloprotease family members in mammals that contribute to inflammatory responses, provides protection against pathological inflammation induced by lipopolysaccharide (140). Nanobodies that bind β-secretase affect enzyme function, with two nanobodies increasing and one inhibiting activity (141). Injection of the inhibitory nanobody directly into the cerebrospinal fluid decreased deposition of β-amyloid as a result of β-secretase inhibition in a mouse model of Alzheimer's disease (141). γ-Secretase, also relevant for Alzheimer's disease, has likewise been targeted for inhibition by nanobodies (22). Nanobodies raised against plasminogen activator inhibitor-1 induced a profibrinolytic effect through stimulation of protease activity via neutralization of the protease inhibitor (142, 143). Nanobodies that bind thrombin-activatable fibrinolysis inhibitor (procarboxypeptidase U) block protease activation and thereby promote fibrinolysis (144). Combined, these examples demonstrate the versatility of the various nanobody platforms in their application to extracellular space.

Targeting intracellular proteins

Many nanobodies require neither glycosylation nor disulfide bond formation to retain their antigen-binding properties. They can thus be expressed as targeting reagents in the reducing environment of the cytosol. Conventional antibodies and their fragments mostly rely on their entry into the endoplasmic reticulum for assembly and glycosylation. The cytosol precludes association of immunoglobulin heavy and light chains, thus compromising their intracellular assembly into a functional unit. Intracellular nanobodies are typically introduced through transfection of DNA. This allows expression of nanobodies in either a constitutive or an inducible manner. Nanobodies can be expressed as monomeric units to modulate the activity of their targets upon binding or as fusions with fluorescent proteins or taggable protein domains for visualization of targets. Nanobodies, their variants, and fusions have also been used to redirect protein localization, induce protein degradation, and serve as biological sensors of protein conformation, abundance, and localization. Nanobodies that target nuclear proteins are often equipped with a nuclear localization sequence, although this is not always required (145). One of the main bottlenecks restricting the deployment of nanobodies in cells is the paucity of intracellular target-specific nanobodies. We provide summary of many relevant examples below without making claims as to completeness of the list provided (Table 1).

Table 1.

Summary of nanobodies used to target intracellular proteins

Note that this table does not include nanobodies used primarily for structural studies or those that target viral proteins and secreted proteins. The use of GFP-targeting nanobodies in this context is discussed in the text.

Intracellular target (function)	Speciesa	Application and biological impact	Reference
Cytoplasmic proteins
STAT3 (transcription factor)	H	Slows breast cancer growth in vitro and in vivo	151
TUT4 (uridytransferase)	H, M	Blocks microRNA uridylation and degradation	152
Calcium-dependent kinase	T	Inhibits kinase activity, crystallization chaperone	153
CapG (actin-capping enzyme)	H	Blocks actin binding, inhibits cancer metastasis	154
HypE (AMPylation)	H	Inhibits or activates enzyme, cellular imaging	155
SpvB (ADP-ribosylation)	S	Inhibits enzyme, blocks cytoskeletal changes	156
ASC (inflammasome adaptor)	H	Interrupts assembly, cellular imaging	160
Roco (GTPase)	B	Destabilizes dimer, enhances GTP hydrolysis	157
RhoA (GTPase)	H	Either inhibits RhoA or tracks localization	19, 161
RhoB (GTPase)	H	Targeted degradation of GTP-bound RhoB	190
Gβ/γ (GTPase subunit)	H	Blockade of signaling following GPCR activation	163
L-plastin (actin-bundling protein)	H	Inhibits enzyme function, defective immune synapse formation	164, 165
p53 (tumor suppressor protein)	H	Relocalization to mitochondria or protection from proteasomal degradation	175, 176
H2A/H2B (histone)	Y, M, H	Directs ubiquitination to induce DNA damage signaling, imaging	189
UBC6e (E2 enzyme)	M, H	Enhances enzyme function in vitro	158
Dynamin (GTPase)	H	Binds GTP-bound enzyme, visualization of localization	162
Intracellular face of plasma membrane protein
VGLUT (Glu transporter)	R	Inhibits glutamate transport, visualization	166
P-glycoprotein (transporter)	M	Inhibits function in vitro	167
β2AR (GPCR)	H	Binds and stabilizes the receptor active or inactive states for structural studies and visualization	194, 195
Muscarinic acetylcholine receptor (GPCR)	H	Binds and stabilizes the receptor active state for structural studies	92
κ- and μ-opioid receptors (GPCR)	H	Binds and stabilizes the receptor active or inactive states for structural studies and visualization	103, 198, 199
CaV1/CaV2 (high voltage–activated calcium channel)	H, M, G	Expressed as E3 fusion to ubiquitinate, redirect localization, inhibit function	188

^aSpecies of the antigen bound by the referenced nanobody. H, human; M, murine; B, bacterial; S, Salmonella; Y, yeast; R, rat; G, guinea pig.

Transfection-based approaches do not allow direct installation on nanobodies of bright organic fluorophores, which mostly requires chemical methods applied to purified nanobodies. Delivery of labeled nanobodies and other antibody derivatives across the plasma membrane (146) has relied on appending cell-penetrating peptides (48, 147, 148), mutagenesis of nanobody surface residues to increase their positive charge (149), complexation with cell-penetrating mesoporous silica nanoparticles (150), and the use of a microfluidics-based cell permeabilization platform (49). Cell-penetrating nanobodies have been used for imaging (48, 49), to inhibit EGFR function from the cytoplasmic side of the membrane in lung cancer cells (147), and to evaluate the importance of subcellular localization in photosensitizer-induced cell killing (148). The design of methods to deliver proteins such as nanobodies across the plasma membrane is an area of active research, but accumulation of most such exogenously added proteins in the endocytic compartment is difficult to avoid and remains a major confounding factor.

Nanobodies that bind intracellular proteins enable characterization and modulation of proteins of interest and can avoid the need for genetic modification of the target protein. A summary of several nanobodies that target intracellular proteins is shown in Table 1. Targets include mammalian and bacterial proteins; soluble cytoplasmic proteins and those embedded in membranes; and proteins in the nucleus, endoplasmic reticulum membrane, and inner leaflet of the plasma membrane. Nanobodies can inhibit intracellular signaling proteins, such as the transcription factor STAT3 (151), the uridyltransferase TUT4 (152), apicomplexan calcium-dependent kinase (153), the actin-capping protein CapG (154), the AMPylation enzyme HypeE (155) and the Salmonella ADP-ribosylation enzyme SpvB (156). A nanobody that targets an allosteric site on a bacterial LRRK2 GTPase homologue modulates oligomerization and enzymatic activity (157). The ER-localized E2 enzyme UBC6e is targeted by a nanobody that augments enzymatic activity without obvious biological consequences (158). A biological sensing platform that relies on the expression of two nanobodies that bind to different epitopes on GFP, each linked to one part of a transcription factor, converts the production of intracellular GFP to a transcriptional output (159).

Intracellular proteins that require oligomerization and those trafficked through the endoplasmic reticulum have also been probed using nanobodies. Co-translational delivery of a nanobody to the endoplasmic reticulum prevents aberrant processing and aggregation of a variant of the secreted form of gelsolin (132). Expression of a nanobody that recognizes ASC, an adaptor protein important in inflammasome assembly, enabled visualization of inflammasome assembly in cells and altered the morphology of assembled inflammasomes (160). Nanobodies raised against the active (GTP-bound) forms of the GTPases RhoA and dynamin enabled tracking of active enzymes in living cells (161, 162). Using these tools, active RhoA was detected at the inner plasma membrane upon overt activation; active dynamin was formed in stochastic bursts associated with membrane fission. A nanobody that binds and inhibits the function of β/γ-subunit complex of the heterotrimeric G proteins showed that blockade of β/γ-subunit function has minimal impact on Gα function (163). The expression of nanobodies that inhibit the actin-bundling protein L-plastin uncovered a role for this protein and the T cell integrin LFA-1 in facilitating the formation of the immune synapse (164, 165). Two proteins found in the plasma membrane, VGLUT and P-glycoprotein, are targeted at their cytoplasmic faces by nanobodies that inhibit their function (166, 167).

Intracellular nanobody constructs can be adapted to yield fusion proteins that enable the relocalization, destruction, or enzymatic modification of nanobody-bound targets. The fusion of a nanobody to O-GlcNAc transferase enabled directed glycosylation of proteins targeted by the nanobody fusion (168), a post-translational modification that is otherwise widespread. Nanobodies, expressed as fusions with tags that dictate a particular subcellular localization, can redirect the localization of proteins of interest and serve to control protein diffusion. GFP-binding nanobodies routed to subcellular sites have been used to assess the impact of redirecting GFP-tagged targets in living multicellular organisms (Fig. 5A) (169–171). Forced mislocalization of the drosophila regulatory myosin light chain, tagged with GFP, to either the basolateral or apical membrane, caused an alteration in the shape of wing cells or aberrant sibling cell asymmetry in neural cells (169, 170). Membrane anchoring of a protein essential for the development of cellular polarity in Caenorhabditis elegans revealed the importance of clustering for asymmetry (171). This approach has been used to redirect mRNAs engineered to contain the GFP-binding binding sequence MS2, which showed that forced mRNA relocalization also caused protein relocalization (172). In a variation on this approach, a secreted protein was tethered to the cell surface, with rates of diffusion controlled by the strength of the nanobody-epitope interaction employed, to assess the importance of local and distal action of secreted proteins (173, 174). This approach showed that diffusion of a secreted morphogenic protein in Drosophila is essential for proper wing patterning (173). A p53-binding nanobody with a mitochondrial localization tag showed that mitochondrial mislocalization led to loss of cell viability in some cases (Fig. 5B) (175). A p53-binding nanobody that blocked the degradation of p53 mediated by human papillomavirus E6 protein failed to promote apoptosis (176).

Figure 5.

Nanobodies as redirecting and sensing agents in live cells. A, use of GFP-binding nanobodies to redirect tagged proteins to subcellular locations or for degradation. B, use of a p53-binding nanobody to block HPV E6-mediated ubiquitination and degradation. C, use of orthogonal anti-NP nanobodies as biosensors coupled to a transcriptional output (214). The DNA-binding domain (DBD) and VP64 activation domain are separately fused to anti-NP nanobodies (VHHs). UAS, upstream activator sequence that binds DBD. Transcription of the reporter gene produces GFP. D, use of nanobodies as biosensors to detect active and inactive states of GPCRs.

The targeting function of nanobodies can be exploited to link particular substrates to the degradation machinery. The modular nature of nanobodies thus enabled their use in a variety of fusion proteins for targeted degradation of proteins of interest (177). This approach typically involves genetic fusion of a nanobody with a (fragment of a) ubiquitin ligase (E3), which recruits the endogenous ubiquitination machinery to tag nanobody-targeted proteins for degradation by the proteasome. One widely deployed version of this approach relies on a fusion of a nanobody that binds GFP/YFP with an F-box protein domain from the Skp cullin F-box E3 complex, which can degrade GFP/YFP-tagged proteins (Fig. 5A) (178, 179). An alternative version uses von Hippel–Lindau (VHL)-nanobody fusions to degrade GFP-tagged proteins (180, 181). Targeted degradation using this approach, in combination with expression under tissue-specific promoters, has enabled evaluation of the role of targeted proteins in specific tissues in Drosophila development (182, 183). Tissue-specific degradation of myosin-II showed it is not essential for tracheal elongation or the closure of the dorsal opening during development of Drosophila. Alternative versions of nanobody-mediated target degradation, relying on other E3 fragments, have also been developed for application in zebrafish (184, 185) and C. elegans (186). The use of nanobody-E3 fragment fusions to target cell surface ion channels, either as a YFP fusion or as the WT protein, showed that ubiquitylation can have divergent consequences for the trafficking and function of cell surface proteins (187, 188). An E3 fragment–nanobody fusion that bound directly to the histone H2A-H2B protein dimer enabled targeted ubiquitylation of histones and caused signaling associated with DNA damage (189). A nanobody that selectively bound to the active (GTP-bound) form of RhoB GTPase was applied as an F-box fusion to knock down active RhoB. It showed that the GTP-bound fraction of RhoB mediates its role in cell invasion (190). Control of the properties of the E3 fragment–nanobody produced in cells, such as expression level and degradation rate, allows quantitative control of cellular protein levels (191).

Specialized, targeted approaches have produced nanobodies that bind to the target only when found in a specific conformation, with a particular emphasis on membrane proteins. These approaches rely on screening with antigens locked into the desired conformation: amyloidogenic protein variants at various stages of self-assembly, complexes formed by protein-protein interactions, and receptors bound to ligands constitute some of the targets of conformation-specific nanobodies (30). They have served as biosensors to visualize the distribution of proteins in a specific conformation in living cells (Fig. 5D) (192, 193), with a particular emphasis on GPCRs. A nanobody raised against the β2AR, used as a chaperone to facilitate its crystallization (89), binds to the cytoplasmic face of β2AR and stabilizes its active state, much like a G protein would. This same nanobody, when expressed as a fluorescent fusion protein served as a biosensor to visualize ligand-bound β2AR in its active state (194). Surprisingly, activated β2AR was found both at the plasma membrane and in early endosomes. Further characterization of other β2AR-binding nanobodies identified one that bound to and stabilized the inactive form of β2AR (101, 195). This assembly of nanobodies enabled the classification of several β2AR ligands as agonists, antagonists, or inverse agonists (101, 108). Certain β-adrenergic receptor ligands affect the conformation of receptor molecules found in the Golgi, suggesting that receptors en route to the cell surface can be activated by cell-permeable ligands (196). This possibility is of interest also in view of the exclusive Golgi localization of GPCRs such as GPR107 (197). Conformation-specific nanobodies have been applied for similar applications to the muscarinic acetylcholine receptor (92) and for the κ- and μ-opioid receptors (103, 198, 199). Although these conformation-specific nanobodies are restricted to the indicated receptors, study of a nanobody specific for the κ-opioid receptor in its active state showed that the intracellular loop fragment responsible for nanobody binding could be grafted onto other GPCRs with retention of binding, highlighting the exciting possibility of designing receptor chimeras bound by conformation-specific nanobodies (103). This select set of examples shows that intracellular expression of nanobodies directed against cytoplasmic targets clearly is a feasible approach to modulate cellular functions. However, considerable efforts must be expended to identify nanobodies with the desired properties. As methods for the production of large and completely synthetic nanobody libraries improve, the need for deliberate immunization of animals is reduced. Synthetic libraries may also make it possible to obtain nanobodies against proteins that are not immunogenic in camelids.

Targeting viral proteins

Nanobodies have been used to target viral proteins. Inhibition of viral entry by nanobodies is well-documented (200, 201). The recent description of camelid-derived nanobodies capable of neutralizing SARS-CoV-2, the coronavirus responsible for the COVID-19 pandemic, is one such example (202, 203). There are fewer examples of nanobodies that target cytoplasmic viral proteins. Because infected cells produce cytoplasmic proteins required for proper virus replication, assembly, and release, nanobody-mediated interference with intracellular viral proteins requires cytoplasmic expression as well. Inhibition of viral polymerases by a cytoplasmically expressed nanobody inhibits influenza propagation (204). A nanobody that blocks multimerization of HIV REV causes defects in viral RNA transport and inhibits virus production (205). HIV NEF is a nonstructural protein that alters protein trafficking in infected cells, and a nanobody directed against NEF blocks these effects (206). Nanobodies that target viral nucleoproteins (NPs) inhibit viral fitness by several mechanisms, including indirect inhibition of polymerase function (207), inhibition of nuclear import of viral ribonucleoproteins (208), and disruption of virus assembly by NP cross-linking with dimeric nanobodies (209). A nanobody that binds grapevine fanleaf virus capsid protein provides resistance to the virus when expressed in plant cells (210, 211) These few examples show that cytoplasmically expressed viral proteins can serve as excellent nanobody targets to explore aspects of viral replication that might be more challenging to study otherwise. For example, the many essential roles of influenza NP have made a dissection of its various functions difficult by standard mutagenesis. The use of NP-specific nanobodies has uncovered novel aspects of nuclear localization (208).

Nanobodies that target viral proteins can be turned into tunable biological tools and sensors. A nanobody-fluorescent protein fusion that recognized the HIV capsid protein p21 was deliberately destabilized by introducing mutations that resulted in rapid degradation of the fusion protein (212). This fusion was stabilized upon binding of p21, thus providing a fluorescent reporter of HIV infection (212). Intracellular expression of influenza NP-specific nanobodies enabled the tracking of viral distribution (213). Use of a pair of orthogonal anti-influenza NP nanobodies linked to different domains of a transcription factor, which must be brought into proximity to function, provided a method in which transcription of a fluorescent protein reporter is controlled by the presence of NP (Fig. 5C) (214). A nanobody that blocks nuclear import of influenza ribonucleoproteins through occlusion of interactions of the viral nuclear localization sequence with its host receptor was repurposed as a way to control viral infectivity: engraftment of the nuclear localization sequence onto the nanobody itself restored infectivity (215).

Epitope tags

Antibodies that target peptide fragments of less than 20 amino acids, commonly called epitope tags, have been widely applied to target tagged proteins of interest in a variety of assays, including microscopy, immunohistochemistry, immunoblotting, and immunoprecipitation experiments (216). Such epitope tags are valuable when high-quality antibodies or nanobodies against the protein portion of the tagged protein are lacking (7) and to avoid cross-reactions with related proteins. Small epitope tags can be preferable to the use of GFP, as the presence of a GFP moiety can impact protein behavior (217). Only a handful of nanobodies that recognize short peptides suitable as epitope tags have been reported thus far. The reason for the scarcity of nanobodies that recognize epitope tags in a variety of settings is unknown but most likely relates to the often convex nature of their antigen-binding portion (paratope) and the need to achieve the necessary buried surface area upon binding of antigen through interaction with a single variable region, instead of the conventional V_H-V_L pair. The reported nanobody-peptide epitope pairs are summarized in Table 2.

Table 2.

Summary of nanobody-epitope tag pairs and applications

Listed affinities were determined using a variety of methods as described in the listed references.

Name	Affinity	Sequence/applications	Notes	References
	nm
EPEA	190	EPEA/affinity purification, flow cytometry	Tag must be at C terminus, cross-reacts with α-synuclein, marketed as C-tag	127, 168, 218
myc	ND	EQKLISEEDL/enzyme-linked immunoassay	Not extensively characterized	219
BC2	2–6	PDRVRAVSHWSS/affinity purification, immunoblotting, (superresolution) microscopy	Cross-reacts with endogenous β-catenin, marketed as Spot-tag	4, 220
6E tag	1	QADQEAKELARQIS/affinity purification, immunoblotting, cell surface tethering of GPCR ligand	Reacts with endogenous UBC6e	110, 158
Alfa	0.02	SRLEEELRRRLTE/affinity purification, immunoblotting, superresolution microscopy, target relocalization in live cells	Tag adopts α-helical structure	6, 221, 223
Moon	30	KNEQELLELDKWASL/live-cell microscopy	Tag derived from HIV gp41	221
PepTag	0.6	AVERYLKDQQLLGIW/immunoprecipitation, live-cell microscopy	Tag derived from HIV gp41, helical, previously named VHH 2E7	239, 240

A nanobody (NbSyn2) raised against α-synuclein binds to the extreme C terminus of the protein, involving the four terminal amino acids (EPEA) (127, 218). Although the tetrapeptide epitope tag is the smallest nanobody tag reported, the NbSyn2-EPEA interaction is of lower affinity (K_d ∼190 nm) than other nanobody-epitope tag pairs and requires that EPEA be placed at the C terminus (127). There is a report of nanobodies that recognize the myc tag (219) derived from the c-Myc protein, but the affinities of these nanobodies have not been reported, and they have not been widely tested. A different nanobody-epitope tag pair resulted from characterization of nanobodies raised against β-catenin (BC2 nanobody) (4, 5). The BC2 nanobody binds to a 12-residue epitope with high affinity (K_d ∼2-6 nm). This tag can be placed at the N or C terminus of the protein of interest with retention of binding (4) and has been used in proteomics and superresolution microscopy studies. Optimal results in the latter application required the use of a homodimeric nanobody construct to increase avidity (4, 220). A nanobody that recognizes a 14-mer peptide with high affinity (K_d ∼1 nm), derived from the endoplasmic reticulum-localized protein UBC6e (6E tag) is nanobody VHH05 (158). Incorporation of the 6E tag into the extracellular portion of a GPCR enabled efficient delivery of ligands linked to VHH05 to a tagged receptor, resulting in potent activation (110). Most nanobody-recognized epitopes are derived from endogenous cellular proteins and can be used in a heterologous setting but may cause high background in applications where the endogenous protein is present. This complication was circumvented by generating a nanobody (NbAlfa) that binds with very high affinity (K_d ∼20 pm) to a synthetic 14-mer helical peptide (Alfa tag) not found in nature (6). NbAlfa was used for affinity purification, high-sensitivity immunoblotting, and superresolution microscopy (6). Whether the introduction of a tag with a strong propensity toward helical structure and book-ended by proline residues disrupts the structure and function of proteins engineered to contain it must be determined on a case-by-case basis.

In a system designed to evaluate translation kinetics in live cells, efforts to identify nanobody-epitope tag pairs that function in live cells evaluated seven different candidate nanobody-epitope tag pairs. Only a single nanobody-epitope tag pair was found to function in cells (221). This pair (dubbed “moon tag”) consists of a nanobody raised against the HIV envelope protein gp41 and a 15-mer peptide that interact with an affinity of ∼30 nm (222). The moon tag was used along with an orthogonal epitope recognition pair based on an scFv antibody (named “sun tag”) to demonstrate a high degree of stochasticity and variation in translation kinetics in live cells (221). Both the moon tag and the alfa tag system can be used to deliver a nanobody-fluorescent protein fusion to a subcellular site of choice (nucleus, mitochondria, cell membrane) when the nanobodies carry the appropriate targeting signals (223). The orthogonality of the moon tag and alfa tag suggests the potential for multiplexing (223). The VHH05-6E pair can also efficiently redirect subcellular protein localization.⁴ An important consideration for the selection of tag-specific nanobodies is a lack of cross-reactivity with the species of origin in which these nanobodies will be used. Nanobodies against pathogen-specific antigens may provide a possible source of such reagents.

Synthetic biology with nanobodies

The adaptable and modular nature of nanobodies makes it possible to reprogram, or report on, cell function. Two widely deployed examples of this approach are the targeted degradation of cellular proteins tagged with GFP (discussed above) and the expression of nanobody-containing chimeric antigen receptors in T cells (CAR-Ts) (reviewed recently (224)). Many of the applications of nanobodies in synthetic biology described below rely on widely deployed clones that target GFP or viral proteins. Extension of these approaches to epitope tag–targeting nanobodies will expand their adaptability.

One goal of synthetic biology is to allow cells to bind or respond to molecules for which they do not have a natural receptor. This goal can be achieved through expression of cell surface protein fusions that contain nanobodies that bind to a molecule of interest. Adaptation of the NOTCH signaling pathway through introduction of a nanobody-sensing element enabled initiation of transcriptional responses in mammalian cells in the presence of extracellular GFP (Fig. 6A) (225). An analogous platform developed in bacteria used a nanobody fused to a dimerizable DNA-binding domain, which then enabled caffeine-stimulated induction of transcription (Fig. 6B) (226). Modification of a bacterial flagellar protein by insertion of a GFP-binding nanobody yielded bacteria that bound GFP with retention of flagellar structure (Fig. 6C) (227). Other bacterial outer membrane proteins could also be expressed as fusions with a GFP-binding nanobody (228). Modification of the bacterial hemoglobin protease gene yielded membrane vesicles (OMVs) equipped with a recognition tag for SpyCatcher that allowed production of nanobody-modified OMVs (229). The ability to equip bacteria or their outer membrane vesicles with nanobodies sets the stage for delivery to desired biological sites through nanobody targeting. A glycosylphosphatidylinositol-anchored version of a nanobody that binds to HIV gp41 protects susceptible cells from HIV infection, even where the free nanobody itself fails to protect (230). Nanobodies localized at the cell surface can trap secreted proteins near their site of secretion (169, 173). Variation in the strength of the interaction between cell surface nanobody and secreted protein can then be used to control protein diffusivity (174).

Figure 6.

Application of nanobodies in synthetic biology. A, use of nanobodies as recognition elements in programmable synNotch constructs (225). Upon ligand binding, sometimes found on the surface of neighbor cells, synNotch receptors undergo a conformational change that promotes cleavage and release of an intracellularly linked transcription factor. B, nanobodies as recognition elements in bacterial constructs for sensing of extracellular ligands (226). The addition of caffeine causes clustering of extracellular single-domain antibodies and subsequent assembly of split DBDs and a transcriptional output. C, scheme for attachment of nanobodies to bacterial outer membranes or OMVs using SpyCatcher/SpyTag (229). An outer membrane protease, hemoglobin protease (gray), is linked to SpyTag (red). A SpyCatcher/nanobody fusion (black/gold) covalently labels the hemoglobin protease-SpyTag fusion.

Nanobody variants that are unstable in mammalian cells unless bound to their target can report on relative protein abundance (212). Several nanobody constructs whose function or localization can be controlled by light or small molecules have been reported. This approach can achieve improved spatial (localized illumination) and temporal control of cellular function relative to processes that require transcription and translation, which take much longer. The fusion of nanobodies with proteins that engage in light-stimulated interactions with binding partners enables light-directed routing of nanobodies or nanobody-bound targets to large clusters formed by protein hetero-oligomerization (231) or to specific subcellular compartments (232). In these examples, illumination controls only the nanobodies' localization and therefore access to targets with restricted subcellular localization. An alternative strategy hinges on coupling illumination with nanobody function directly. Proteins that heterodimerize upon illumination have been fused with nanobody fragments to enable light-stimulated binding and inhibition of the nanobody targets by light-induced reassembly of nanobodies inside cells (233). Introduction of a photoresponsive protein into the nanobody framework regions also enabled light-induced or light-inhibited nanobody function, although many constructs retained target binding to some extent even in the absence of illumination (234). To enable small-molecule control of nanobodies, a dihydrofolate reductase variant, which folds only upon ligand binding, was inserted into a nanobody complementarity-determining region (235), which allowed small-molecule control of nanobody binding. A different strategy for generating photoresponsive nanobodies relies on stop codon suppression to incorporate a nonnatural photocaged tyrosine residue into a nanobody complementarity-determining region (236, 237). Nanobodies with an appropriately placed caged tyrosine, produced through recombinant expression in bacteria, failed to bind to targets expressed on the cell surface, but upon irradiation, binding was restored.

Concluding remarks

Properties unique to nanobodies make them useful tools, with features that differentiate them from conventional antibodies and synthetic recognition modules. Nanobodies have proven their worth as chaperones for structural studies, particularly in the study of membrane proteins such as GPCRs (10). Expanding efforts to apply nanobodies as intracellular reagents and to equip nanobodies with synthetic moieties using chemical and enzymatic methods are opening up exciting new possibilities.

Nanobodies appear particularly well-suited to report on cellular events that happen on a rapid timescale, where reporters based on transcription and translation may be too sluggish. The specialized properties of nanobodies have allowed insight at high temporal resolution into complex and stochastic events, such as the dynamics of protein translation (221), endomembrane trafficking (162), and receptor conformational changes (103). Designs that turn on or off the function of nanobodies with light (233) or small molecules (235) will allow unprecedented temporal control over processes such as protein degradation (191), subcellular relocalization (175), and even post-translational modifications (168). This strategy relies on the availability of nanobodies that target proteins of interest in the context of the cytoplasm, of which there is a substantial but still limited selection. New screening methods to identify nanobodies against new targets will allow extension to other targets. The emergence of nanobodies that bind epitope tags in the cytoplasm will help to bridge this gap (223).

Chemical and enzymatic labeling strategies applied to nanobodies provide semisynthetic conjugates with properties not achievable without this combination. The use of nanobodies chemically modified with moieties for visualization as in vivo imaging agents is an example (1). The conjugation of nanobodies with synthetic ligands for cell surface receptors provides a path toward conjugates with improved receptor selectivity (110) and light-responsiveness (109). The ability to site-specifically modify nanobodies at more than one position should enable further elaboration (55). The ability to prepare nanobodies through chemical synthesis would permit the ultimate level of control for nanobody chemical functionality (238). The unique properties of nanobodies have firmly secured their place in the biochemist's toolbox, with yet more exciting and original applications to emerge.