Description

The present invention relates to novel anti-IgG nanobodies, particularly nanobodies directed against rabbit or mouse IgG. Further, the invention relates to the use of said nanobodies and methods for producing them.

Mouse and rabbit antibodies are fundamental tools for numerous basic research techniques as well as medical diagnostic assays. The detection or immobilization of these primary antibodies is most often performed indirectly via polyclonal anti-IgG secondary antibodies. Yet, the need for a continuous supply of anti-IgG sera requires keeping, immunizing, bleeding and eventually sacrificing large numbers of goats, sheep, rabbits, or donkeys, which is not only costly but also a major animal welfare and ethical problem (Shen, 2013; Reardon, 2016). Furthermore, every new batch of serum contains another heterogeneous mixture of antibodies, which need to be affinity-purified on IgG columns and then depleted (by pre-adsorption) of nonspecific and crossreacting antibodies. Moreover, the success of this procedure has to be laboriously quality controlled each time. The large size of secondary antibodies (-10-15 nm; 150 kDa) is also a disadvantage, since it limits tissue penetration and introduces a considerable label displacement, reducing the obtainable image resolution by super-resolution fluorescence microscopy methods (Ries et al., 2012; Szymborska et al., 2013; Pleiner et al., 2015). Their non-recombinant nature further precludes genetic engineering i.e. tagging or fusion to reporter enzymes.

Why then, have recombinant anti-IgG detection reagents not yet replaced polyclonal secondary antibodies? The major issue is regarding signal strength. The signal in traditional immunofluorescence, for example, is amplified by: (i) multiple secondary IgG molecules binding to distinct epitopes of a primary antibody; (ii) a large IgG tolerating many labels per molecule; and (iii) by their bivalent binding mode exploiting avidity for high affinity target recognition. In the light of these facts, it appears very challenging to achieve comparable signal levels with a small, monovalent and monoclonal reagent.

Yet, we considered nanobodies, single-domain antibodies derived from camelid heavy-chain antibodies (Hamers-Casterman et al., 1993; Arbabi Ghahroudi et al., 1997; Muyldermans, 2013), as perhaps the best candidates for such reagents. Due to their small size (~3 x 4 nm; 13 kDa), the possibility of their renewable production as recombinant fusion proteins, as well as favorable biophysical properties, nanobodies attracted considerable attention as powerful tools in cell biology (Helma et al., 2015), structural biology (Desmyter et al., 2015) and as future therapeutic agents (Van Bockstaele et al., 2009; Kijanka et al., 2015). They are particularly useful for super- resolution imaging (Ries et al., 2012; Szymborska et al., 2013; Pleiner et al., 2015; Gottfert et al., 2017; Traenkle and Rothbauer, 2017). The resolving power of some of the best microscopes reported to date (e.g. ~6 nm by Balzarotti et al., 2017; -10-20 nm by Huang et al., 2016 or Xu et al., 2012) may be reduced due to the offset between fluorescent label and target introduced by primary and secondary antibodies (20-30 nm). Site-specifically labeled nanobodies represent a promising solution to this problem, since they can place fluorophores closer than 2 nm to their antigen and, despite their small size, even tolerate up to three dyes (Pleiner et al., 2015).

In this study, we describe the generation of a comprehensive toolbox of nanobodies (Ig single variable domains) against all mouse IgG subclasses and rabbit IgG. This work required very extensive optimizations of our routine nanobody selection efforts, such as a time-stretched and thus affinity- enhancing immunization scheme, subsequent affinity maturation including off-rate selections, as well as testing and improving ~200 initial candidates. When labeled site-specifically with fluorophores, the resulting nanobodies performed remarkably well in Western Blotting and immunofluorescence. In contrast to polyclonal secondary antibodies, they even allow a single-step multicolor labeling and co-localization. Moreover, we show that anti-lgG nanobodies can be conjugated to horseradish peroxidase (HRP) or expressed as fusions to ascorbate peroxidase (APEX2) (Lam et al., 2015) and thus used for enhanced chemiluminescence Western blotting or colorimetric ELISAs or immuno-EM detection. These monoclonal recombinant nanobodies are thus perfect substitutes for conventional animal- derived polyclonal secondary antibodies. We envision that they can be engineered to enable a more versatile use of the plethora of existing antibodies and even allow the development of more sophisticated antibody- based diagnostic tests.

A first aspect of the invention relates to a nanobody directed against rabbit IgG comprising

(a) a CDR3 sequence as shown in SEQ ID NO. 1 or 2, or

(b) a CDR3 sequence which has an identity of at least 80%, particularly at least 90% to a CDR3 sequence as shown in SEQ ID NO. 1 or 2. In a particular embodiment, the nanobody comprises

(a) a combination of CDR1 , CDR2 and CDR3 sequences as shown in SEQ ID NO. 1 or 2, or

(b) a combination of CDR1 , CDR2 and CDR3 sequences which has an identity of at least 80%, particularly at least 90% to a combination of CDR1 , CDR2 and CDR3 sequences as shown in SEQ ID NO. 1 or 2.

In a further particular embodiment, the nanobody comprises

a sequence as shown in SEQ ID NO. 1 or 2, or

a sequence which has an identity of at least 70%, particularly at least 80% and more particularly at least 90 % to a sequence as shown in SEQ ID NO. 1 or 2. A further aspect of the invention relates to a nanobody directed against mouse IgG comprising

(a) a CDR3 sequence as shown in any one of SEQ ID NO. 3-34, or

(b) a CDR3 sequence which has an identity of at least 80%, particularly at least 90% to a CDR3 sequence as shown in any one of SEQ ID NO. 3-34.

In a particular embodiment, the nanobody comprises

(a) a combination of CDR1 , CDR2 and CDR3 sequences as shown in any one of SEQ ID NO. 3-34, or

(b) a combination of CDR1 , CDR2 and CDR3 sequences which has an identity of at least 80%, particularly at least 90% to a combination of CDR1 , CDR2 and CDR3 sequences as shown in any one of SEQ ID NO. 3-34.

In a further particular embodiment, the nanobody comprises

a sequence as shown in any one of SEQ ID NO. 3-34 or a sequence which has an identity of at least 70%, particularly at least 80% and more particularly at least 90% to a sequence as shown in any one of SEQ ID NO. 3-34.

Still a further aspect of the present invention is a reagent for detecting, isolating and/or purifying IgG, particularly rabbit or mouse IgG comprising at least one nanobody as described above.

Still a further aspect of the invention is a method for detecting, isolating and/or purifying IgG, particularly rabbit and/or mouse IgG, comprising binding of at least one nanobody as described above to IgG.

In a preferred embodiment, the nanobody is specifically directed to one of the following groups of types of IgG molecules and epitopes: - rabbit IgG Fab fragment

- rabbit IgG Fc fragment

- mouse IgG kappa light chain or mouse IgG lambda light chain - mouse lgG1 , e.g. lgG1 Fc fragment, lgG1 hinge region, or lgG1 Fab fragment

- mouse lgG1/lgG2a, e.g. lgG1/lgG2a Fab fragment

- mouse lgG1/lgG2a/lgG2b, e.g. lgG1/lgG2a/lgG2b Fab fragment - mouse lgG2a, e.g. lgG2a Fc fragment or lgG2a hinge region

- mouse lgG2a/2b, e.g. lgG2a/lgG2b Fc fragment

- mouse lgG2b, e.g. lgG2b Fab fragment

- mouse lgG3, e.g. lgG3 Fc fragment

- mouse lgG2a/3, e.g. lgG3 Fc fragment

Specific embodiments of preferred anti-IgG nanobodies are shown in Tables 1 and 2. Table 1 indicates isotype, epitope and species specificity of particular nanobodies of the present invention.

Table 1

N/A = not applicable; Fab = Fragment antigen-binding; Fc = Fragment crystallisable; K = kappa light chain; λ = lambda light chain.

Table 2 indicates the amino acid sequences as well as the individual framework portions, i.e. framework 1 , 2, 3 and 4, and CDR sequences, i.e. CDR1 , CDR2 and CDR3 sequences of particular nanobodies of the present invention.

In its broadest sense, the nanobody of the invention is defined by the presence of at least one CDR3 loop which, as such, is capable of mediating efficient binding to the target molecule. In many cases, CDR1 and/or CDR2 sequences may be varied e.g. by loop grafting, without detrimentally affecting the target specificity. The CDR3 sequence of a nanobody of the invention is defined as a sequence having an identity on the amino acid level of at least 80%, particularly of at least 90% and more particularly of at least 95% with a CDR3 sequence shown in any one of SEQ ID NOs 1-34. In a special embodiment, the CDR3 sequence of the nanobody is as defined in any one of SEQ ID NOs 1 -34.

In a more specific embodiment, the nanobody of the present invention is defined by a combination of CDR1 , CDR2 and CDR3 sequences as shown in any one of SEQ ID NOs 1-34 or a combination of CDR1 , CDR2 and CDR3 sequences having a sequence identity on the amino acid level of at least 80%, of at least 90% or at least 95% to a combination of CDR1 , CDR2 and CDR3 sequences as shown in any one of SEQ ID NOs 1 -34.

In an even more specific embodiment, the nanobodies of the invention have an amino acid sequence as shown in any one of SEQ ID NOs 1-34 or an amino acid sequence having an identity of at least 70%, at least 75%, at least 80%, at least 85%, at least 90% or at least 95% on the amino acid level to an amino acid sequence as shown in any one of SEQ ID NO. 34. In particular, the nanobodies of the invention have an amino acid sequence as shown in any one of SEQ ID NOs 1-34.

The nanobodies of the present invention are characterized by a high binding specificity to different IgG subclasses, in particular to the four mouse IgG subclasses, e.g. to the mouse lgG1 subclass, mouse lgG2a subclass, mouse lgG2b subclass and/or mouse lgG3 subclass, or to the rabbit IgG subclass. For example, nanobodies of the present invention recognizing mouse lgG1 subclass moelcules may be selected from nanobodies of SEQ ID NOs: 8-24. In particular, anti-mouse nanobodies specifically targeting mouse lgG1 molecules may be selected from nanobodies of SEQ ID NOs: 8-19. Further, nanobodies of the present invention recognizing mouse lgG2a isotype molecules may be selected from the nanobodies of SEQ ID NOs: 20-30 and 34. In particular, anti-mouse nanobodies specifically targeting lgG2a molecules may be selected from nanobodies of SEQ ID NOs: 25-29. Nanobodies of the present invention which recognize mouse lgG2b subclass molecules are selected from nanobodies of SEQ ID NOs: 24 and 30-32, whereby nanobodies of SEQ ID NOs: 31 and 32 specifically bind to mouse lgG2b subclass molecules. The present invention also provides nanobodies which recognize mouse lgG3 subclass molecules, selected from the nanobodies of SEQ ID NOs: 32 and 33, whereby SEQ ID NO: 33 specifically recognizes mouse lgG3 subclass molecules. Nanobodies of the present invention may also be able of binding specifically to the mouse light chain, independently of the heavy chain of the IgG subclass (nanobodies of SEQ ID NOs: 3-7). For example, specific anti-kappa chain nanobodies of the invention may be selected from nanobodies of SEQ ID NOs: 3-6. Preferred embodiments of preferred anti-lgG nanobodies of the invention specific for particular IgG subclasses are shown in Table 3:

Table 3

The nanobodies of the present invention are further characterized by high binding specificity for different binding regions of the IgG molecules, e.g. specifically taregeting epitopes on the mouse kappa and/or lambda light chain or the mouse Fc, hinge or Fab fragment on the IgG molecules. Some of nanobodies may have mixed specificities, e.g. may be mouse Fab-binders, which target an interface between a kappa and/or lambda light chain and an IgG, e.g. lgG1 , lgG2a or lgG2b, heavy chain. For example, nanobodies of the present invention specifically binding to an epitope comprising a portion of the and/or λ light chain of mouse IgG molecules may be selected from nanobodies of SEQ ID NOs: 3-7, 19-24 and 31-32. Anti-IgG nanobodies of the present invention specifically targeting an epitope on the Fc fragment of mouse IgG molecules may be selected from nanobodies of SEQ ID NOs: 8-17, 25-28, 30 and 33-34. In particular, nanobodies of SEQ ID NOs: 8-17 target to a mouse lgG1 Fc fragment, nanobodies of SEQ ID NOs: 25-28, 30 and 34 bind to mouse lgG2a Fc fragment and nanobodies of SEQ ID NOs: 33 and 34 recognize an epitope location on the Fc fragment of lgG3 subclass molecules. Table 4 indicates preferred embodiments of preferred anti-IgG nanobodies targeting specific epitope regions on particular IgG sbtype moleules:

Table 4

The nanobodies of the present invention can be characterized by a high target affinity and/or a very low off-rate. The target affinity may be in the range from low nanomolar to sub-picomolar, preferably below 10 nM, below 1 nM or below 100 pM as measured by quantitative phage display. This is supported by the following experimental observations. When titrating target IgG down to subpicomolar concentration, a stoichiometric phage retrieval during anti-IgG nanobody selections was observed, arguing for a non-affinity limited process. Off-rate selections using high excesses of competitor IgG showed no significant nanobody- IgG complex dissociation over a course of 4 h. Similarly, we observed no spectral intermixing during one-step co-immunolocalizations with differentially labeled, pre- formed IgG-nanobody complexes. Due to their low off-rate, the anti-IgG nanobodies remain bound to their target even after extended washing procedures. Further, the nanobodies of the present invention can be characterized by a weak crossreactivity to IgGs from other species, e.g. from rat, guinea pig or from humans. Accoring to a preferred embodiment of the invention the nanobodies show a crossreaction of 15% or less, preferably of 10% or less, more preferably of 5% or less to IgGs from other species, in particular to IgGs from rats, guinea pigs or humans, expecially to IgGs from humans. In a very preferred embodiment of the invention the nanobodies are exclusively specific to rabbit and/or mouse IgG molecules and show no crossreaction to IgGs from other species, in particular to human IgGs. The crossreactivity of the anti-lgG nanobodies of the present invention is measured by conventional methods, such as fluorescent dot blots or ELISA. A specificity profiling dot blot assay of the anti-lgG nanobodies of the present invention to analyze their crossreaction to IgG from other species is described in Example 1.5 and Figure 1 - supplement 1a. The nanobodies of the invention may be genetically modified, e.g. by incorporating an additional cysteine residue at the N-terminus and/or at the C-terminus and/or other surface-exposed positions within the framework region. This facilitates coupling to heterologous moieties via the SH-side chain of cysteine. Further, the present invention refers to conjugates of the above nanobodies. Such conjugates may be genetic fusions, wherein the nanobody is conjugated via peptide bonds to a heterologous peptide or protein sequence. Examples of heterologous protein sequences are peroxidases such as class I peroxidases, e.g. APEX2, or phosphatases (e.g. lambda phosphatase; alkaline phosphatase), or luciferase.

On the other hand, the conjugate may comprise heterologous moieties, e.g. proteinaceous moieties and/or non-proteinaceous moieties coupled to amino acid side chains, e.g. thiol, amino, guanidino, carboxy or hydroxy groups on amino acid side chains, or other reactive groups such as azide or alkyne groups on modified amino acid side chains, and/or to the N- or C terminus by non-peptidic bonds. Suitable conjugation partners are labelling groups, e.g. enzyme reporter groups as described above or fluorescent labelling groups, solid phase-binding groups such as streptavidin or biotin. In a preferred embodiment, one, two or three labelling groups, e.g. fluorescent labelling groups may be coupled to the side chain of cysteine residues. In a further specific embodiment, one, two or three enzyme reporter groups such as horseradish peroxidase or phosphatases may be coupled to the side chain of cysteine residues.

In a further preferred embodiment, one, two or three labelling groups, e.g. fluorescent labelling groups may be coupled to the side chain of lysine residues. In a further specific embodiment, one, two or three enzyme reporter groups such as horseradish peroxidase or phosphatases may be coupled to the side chain of lysine residues.

A further aspect of the present invention relates to a nucleic acid molecule encoding a nanobody as indicated above. The nucleic acid may be a double stranded or single stranded nucleic acid, e.g. DNA or RNA. The nucleic acid molecule encoding the nanobody may be an operative linkage with an expression control sequence, in particular with an expression control sequence which is heterologous to a native nanobody expression control sequence.

The nucleic acid molecule can be incorporated in a prokaryotic or eukaryotic vector suitable for transfecting or transforming hosts, e.g. host cells, e.g. bacterial cells such as E. coli, or eukaryotic host cells, e.g. yeast cells (e.g. S. cerivisiae or Pichia pastoris), insect cells or mammalian cells, e.g. cultured mammalian cells (e.g. HEK-293, HeLa or CHO cells). For this purpose, the nucleic acid molecule may have a codon-optimized sequence with regard to expression in the desired host. Suitable non-human host organisms include multicellular organisms, e.g. mammals, such as mice, rabbits, rats. Suitable types of vectors include plasmids, phages, phagemids, viruses etc. as known to the skilled person. In a preferred embodiment, the vector is a prokaryotic expression vector.

Still a further aspect of the invention is a recombinant cell or recombinant non- human organism transformed or transfected with a nucleic acid molecule or a vector as indicated above. Preferably, the cell or non-human organism is capable of expressing the nanobody of the invention.

The nanobodies of the present invention are particularly suitable for detection, purification and/or isolation of IgG molecules. Specific embodiments include purification and/or isolation of IgG from biological fluids such as blood, plasma, serum or all culture supernatant, detection, purification and/or isolation of IgG- antigen-complexes, immunofluorescence procedures including indirect one-step- immunofluorescence, e.g. by premixing of anti-lgG nanobodies with primary antibodies, indirect one-step co-localisations, e.g. by using different primary antibodies with differently labelled anti-lgG nanobodies, immunoblots with labelled, e.g. fluorescence-labelled and/or enzyme-labelled anti-lgG nanobodies, immunoassays, e.g. in the ELISA format, with labelled, e.g. enzyme-labelled anti- lgG nanobodies etc. Further, the nanobodies can also be used as intracellular antibodies, i.e. intrabodies.

The nanobodies of the invention may be used alone or as combinations comprising several different nanobodies, e.g. a nanobody directed against an Fc fragment of a specific IgG subtype such as mouse lgG1 may be combined with a nanobody directed against the same IgG subtype such as mouse lgG1 , but against a different epitope, e.g. Fab, hinge etc. Alternatively, combinations of nanobodies directed against different subtypes of IgG, e.g. lgG1 and lgG2a, preferably each carrying a different reporter and/or labelling group, are provided. These combinations of compatible nanobodies provide strong signal amplification and may be used for immunostaining, e.g. in multi-colour immunostaining (co- localizations of multiple targets) or immunoblotting detecting e.g. two or more antigens at the same time.

Thus, the invention encompasses combinations of several nanobodies.

In one embodiment, the combination comprises at least 2 nanobodies, e.g. 2 or 3 nanobodies each recognizing the same type of IgG molecules, e.g. mouse lgG1 , mouse lgG2a or rabbit IgG, wherein individual nanobodies of said combination bind to non-overlapping epitopes on the IgG molecules, e.g. respetively to the kappa and/or lambda light chain or to the Fab, Fc or hinge fragment of the IgG molecules. For example, a combination of nanobodies recognizing mouse lgG1 may be selected from the nanobodies of SEQ ID NO: 5 and 8, SEQ ID NO: 8 and 19, SEQ ID NO: 18 and 19 and SEQ ID NO: 5 and 18 or nanobodies comprising at least the CDR3 sequences thereof. A combination of nanobodies recognizing mouse lgG2a may be selected from the nanobodies of SEQ ID NO: 5 and 25 or nanobodies comprising at least the CDR3 sequences thereof. A combination of nanobodies recognizing rabbit IgG may be selected from the nanobodies of SEQ ID NO: 1 and 2 or nanobodies comprising at least the CDR3 sequences thereof. The simultaneous use of combinations of two or more of such nanobodies results in an adequate signal amplification in applications like immunofluorescence and immunoblotting.

Table 5 indicates preferred embodiments of combinations of amti-IgG nanobodies of the invetion directed to the same IgG subtype molecules but individually binding to non-overlapping epitopes on the IgG molecules:

Table 5

Further, the invention encompasses combinations of nanobodies which recognize IgG molecules of different types without cross-reaction, e.g. a nanobody recognizing mouse lgG1 and/or a nanobody recognizing mouse lgG2a and/or a nanobody recognizing rabbit IgG as described above. When appropriately labeled, such a combination of nanobodies can be used for colocalizations in immunofluorescence or multiplexing in immunoblotting.

The nanobodies of the invention may also be provided as a preformed immune complex with an IgG antibody directed against an antigen of interest. Due to their high target affinity and/or low off-rate, the immunocomplexes are stable and can be preformed before use. The immunocomplexes may be used alone or as combinations comprising several different immunocomplexes, e.g. immunocomplexes comprising different IgG antibodies each complexed with a specific nanobody. The nanobodies may be directed against the same IgG subtype or against different IgG subtypes. Thus, combinations of different preformed complexes, preferably each carrying a different reporter and/or labelling group, are provided. These complexes and combinations may be used for immunostaining, e.g. in multi-colour immunostaining or immunoblotting.

Thus, the invention encompasses combinations comprising differently labeled versions of the same nanobody, each bound to different IgG molecules of the same type, but recognize different antigens. Such combinations of pre-formed immune complexes may be used for colocalizations in immunofluorescence or multiplexing in immunoblotting.

The nanobodies of the present invention can be produced in bacteria such as E.coli. They provide reproducible quality, since they are defined by means of their sequence, thus obviating the use of immune sera, i.e. varying mixtures of polyclonal antibodies.

The nanobodies of the present invention have advantageous properties in view of known anti-lgG nanobodies, e.g. as commercially available from Abeam. These advantages are demonstrated in the present examples.

Further, the invention is explained in more detail by the following figures and examples: Figure 1.

Characterization of the anti-lgG nanobody toolbox.

(a) Overview of selected anti-lgG nanobodies identified according to the invention. The obtained nanobodies were characterized for IgG subclass specificity, epitope location on Fab or Fc fragment and species crossreactivity (Figure 1 - figure supplement 1 ). The protein sequences of all anti-lgG nanobodies can be found in Table 2. Nb = nanobody; CDR III = Complementarity-determining region III; Gp = Guinea pig; Hs = Human; κ = kappa light chain; λ = lambda light chain; Fab = Fragment antigen-binding, Fc = Fragment crystallizable.

(b) IgG subclass reactivity profiling of selected anti-mouse IgG nanobodies representing all identified specificity groups. The indicated IgG species were spotted on nitrocellulose strips and the strips blocked with 4 % (w/v) milk in 1x PBS. Then 300 nM of the indicated tagged nanobodies were added in milk. After washing with 1x PBS, bound nanobodies were detected using a fluorescent scanner. Note that the signal strength on poylclonal IgG depends on the relative abundance of the specific subclass (e.g. lgG2b and lgG3 are low-abundant) or light chain (kappa : lambda ratio = 99:1 ). TP885 and TP926 showed no detectable binding to polyclonal Fab or Fc fragment and might bind to the hinge region. MBP = maltose binding protein; poly = polyclonal.

Figure 1 - figure supplement 1.

Species crossreactivity profiling and native target IgG isolation.

(a) Crossreactivity profiling of anti-lgG nanobodies. Using the same Dot blot assay as described in Figure 1 b, the crossreactivity of anti-lgG nanobodies to polyclonal IgG from the indicated species was determined.

(b) Isolation of polyclonal rabbit IgG from rabbit serum. Anti-rabbit IgG nanobodies TP896 and TP897 carrying an N-terminal Avi-SUMOStar tag were biotinylated and immobilized on magnetic Streptavidin beads. After incubation with crude rabbit serum and washing, nanobody-bound polyclonal rabbit IgG was specifically eluted via SUMOStar protease cleavage in physiological buffer. Empty beads served as negative control.

(c) Isolation of anti-Nup62 mouse lgG1 kappa mAb A225 from hybridoma supernatant with anti-mouse lgG1 nanobodies TP881 and TP885 as described in a. The asterisk indicates the SUMOStar protease used for elution. Figure 2.

Application of peroxidase-linked anti-lgG nanobodies.

(a) A twofold dilution series of Xenopus laevis egg extract was blotted and probed with anti-Nup62 mouse lgG1 mAb A225. It was then decorated with horseradish peroxidase (HRP)-conjugated goat anti-mouse polyclonal IgG (5 nM) or anti- mouse lgG1 Fc nanobody TP1 107 (5 nM) and detected via enhanced chemiluminescence (ECL). Similarly, a rabbit polyclonal antibody targeting Nup54 was decorated with HRP-conjugated goat anti-rabbit polyclonal IgG or anti-rabbit IgG nanobody TP897 (5 nM).

(b) A twofold dilution series of Xenopus egg extract was blotted on nitrocellulose and probed with an anti-Nup62 mouse lgG1 monoclonal antibody (upper panel). It was then detected either via HRP-conjugated anti-mouse lgG1 nanobody ab193651 (Abeam, United Kingdom, used at 1 :3,500 dilution, ~5 nM) or 5 nM anti- mouse lgG1 Fc nanobody TP1 107. For this, TP1 107 was conjugated to maleimide-activated HRP (#31485, Thermo Fisher Scientific, USA) via a C- terminal cysteine by incubating both in equimolar amounts for 1 h at room temperature. The blot was developed using Enhanced chemiluminescence (ECL). (Lower panels) A twofold dilution series of Xenopus egg extract was blotted on nitrocellulose and probed with polyclonal rabbit antibodies against Nup54 or Nup107. They were then detected either via HRP-conjugated anti-rabbit IgG nanobody ab191866 (Abeam, United Kingdom, used at 1 :3,500 dilution, ~5 nM) or 5 nM anti-rabbit IgG nanobody TP897. TP897 was conjugated to HRP as described above and the blot developed by ECL.

(c) Oxidation of the fluorogenic ELISA substrate Amplex Ultra Red. A dilution series of pure HRP or recombinant anti-mouse lgG1 Fc nanobody TP1 107- Ascorbate peroxidase (APEX2) fusion was incubated with Amplex Ultra Red and H2O2. Oxidation leads to formation of the fluorescent compound resorufin. The obtained data were fitted with a four-parameter logistic regression. The inflection points of the curves can be used to compare attainable sensitivity. A.U. = arbitrary units. Figure 2 - figure supplement 1.

Anti-lgG nanobody conjugation to HRP and fusion to APEX2.

(a) Anti-mouse lgG1 Fc nanobody TP1 107 with a C-terminal cysteine was conjugated to maleimide-activated horseradish peroxidase (HRP) by incubation of equimolar amounts for 1 h at room temperature.

(b) Expression of anti-mouse lgG1 Fc nanobody TP1 107-ascorbate peroxidase (APEX2) fusion in E. coli. After binding to nickel beads via the N-terminal Hisu- £>c/NEDD8-tag, untagged fusion protein was eluted by on-column /x/NEDP1 cleavage (Frey and Gorlich, 2014).

Figure 3.

Western blotting with infrared dye labeled anti-lgG nanobodies.

(a) A twofold dilution series of Xenopus laevis egg extract was analyzed by SDS- PAGE and Western Blotting. The indicated rabbit polyclonal antibodies were used to detect nucleoporins (Nups). These primary antibodies were then decorated either via IRDye 800-labeled goat anti-rabbit polyclonal IgG (1 :5,000; LI-COR Biosciences, USA) or anti-rabbit IgG nanobody TP897 (10 nM). Blots were analyzed with an Odyssey Infrared Imaging System (LI-COR Biosciences, USA).

(b) (Left panel) A twofold dilution series of HeLa cell lysate was analyzed by SDS- PAGE and Western Blotting. The indicated mouse lgG1 mAbs were decorated either via IRDye 800-labeled goat anti-mouse polyclonal IgG (1 :1 ,340, 5 nM, LI- COR Biosciences, USA) or anti-mouse lgG1 Fc nanobody TP1 107 (5 nM). (Right panel) A twofold dilution series of Xenopus egg extract was blotted and probed with anti-Nup62 mouse lgG1 mAb A225. It was then detected either via IRDye 800-labeled goat anti-mouse polyclonal IgG (5 nM), anti-mouse lgG1 Fc nanobody TP1 107 (5 nM), anti-mouse lgG1 Fab nanobody TP886 (5 nM), anti- mouse kappa chain nanobody TP1 170 (2.5 nM), a combination of TP1 107 and TP886 or TP1 107 and TP1 170. Blue pixels indicate signal saturation.

(c) A dilution series of filamentous bacteriophages was blotted and probed with an anti-minor coat protein pill mouse lgG2a mAb. It was then decorated either via

IRDye 800-labeled goat anti-mouse polyclonal IgG (2.5 nM) or anti-mouse kappa chain nanobody TP1 170 (2.5 nM). Figure 4.

Imaging with anti-lgG nanobodies.

(a) Immunofluorescence with anti-mouse lgG1 nanobodies. HeLa cells were stained with the indicated mouse lgG1 kappa mAbs. These primary antibodies were then detected with Alexa 488-labeled goat anti-mouse polyclonal antibody, anti-mouse lgG1 Fab nanobody TP886 or anti-mouse lgG1 Fc nanobody TP1107. A combination of TP886 and TP1107 yielded increased staining intensities. Laser intensities used to acquire the anti-lgG nanobody images were normalized to the intensity used to acquire the anti-mouse polyclonal antibody image (RLI = relative laser intensity is used here as a measure of fluorescence signal strength).

(b) Immunofluorescence with anti-mouse lgG2a nanobodies. HeLa cells were stained with the indicated mouse lgG2a mAbs. These primary antibodies were then detected with Alexa 488-labeled goat anti-mouse polyclonal antibody, anti- mouse lgG2a Fc nanobody TP1129 or anti-kappa chain nanobody TP1170. A combination of TP1129 and TP1170 yielded increased staining intensities.

(c) Immunofluorescence with anti-rabbit IgG nanobody TP897. HeLa cells were stained with the indicated rabbit antibodies. These primary antibodies were then detected with Alexa 488-labeled goat anti-rabbit polyclonal antibody or anti-rabbit IgG nanobody TP897.

(d) Multicolor-staining of HeLa cells. HeLa cells were incubated with the indicated mouse lgG1 , mouse lgG2a or rabbit IgG antibodies. These primary antibodies were detected via anti-mouse lgG1 Fc nanobody TP1107, anti-mouse lgG2a Fc nanobody TP1129 or anti-rabbit IgG nanobody TP897, respectively, labeled with the indicated Alexa dyes. The upper two panels show dual and the lower panel shows a triple co-localization.

Figure 4 - figure supplement 1.

Immunofluorescence with anti-mouse IgG nanobodies.

(a-b) Images for a given mAb or polyclonal antibody were acquired under identical settings and pixel intensities are represented via a false-color lookup table, (a) HeLa cells were stained with the indicated mouse lgG1 mAbs. These primary antibodies were then detected with Alexa 488-labeled goat anti-mouse polyclonal antibody or a combination of anti-mouse lgG1 Fab nanobody TP886 and anti- mouse lgG1 Fc nanobody TP1107. (b) HeLa cells were stained with the indicated mouse lgG2a mAbs. These primary antibodies were then detected with Alexa 488- labeled goat anti-mouse polyclonal antibody or a combination of anti-mouse lgG2a Fc nanobody TP1129 and anti-kappa chain nanobody TP1170.

(c) Protein sequence alignment of anti-mouse lgG2a nanobody TP921 and the variant TP1129 obtained after affinity maturation. HeLa cells were stained with a mouse lgG2a mAb targeting Lamin A/C. The mAb was detected via TP921 or TP1129 labeled with a single Alexa 488 dye and the images acquired under identical settings.

(d) Protein sequence alignment of anti-mouse kappa chain nanobody TP974 and the variant TP1170 obtained after DNA shuffling and affinity maturation. HeLa cells were stained with a mouse lgG2a mAb targeting Lamin A/C. The mAb was detected via TP974 or TP1170, both labeled with two Alexa 488 dyes.

(e) HeLa cells were stained with the indicated mouse lgG1 kappa mAbs. These primary antibodies were then detected with Alexa 647-labeled goat anti-mouse polyclonal antibody, anti-mouse lgG1 Fc nanobody TP1107 or anti-mouse kappa chain nanobody TP1170. A combination of TP1107 and TP1170 yielded increased staining intensities, see (f) for identical settings scan. RLI = relative laser intensity (as defined in Figure 4a). Figure 5.

One-step immunostaining of HeLa cells with anti-lgG nanobodies.

(a) The indicated mouse lgG1 mAbs were pre-incubated with an equal amount of Alexa 488-labeled goat anti-mouse secondary antibody or a combination of anti- mouse lgG1 Fab nanobody TP886 and anti-mouse lgG1 Fc nanobody TP1107. Likewise, the anti-LAP2 rabbit polyclonal antibody was pre-incubated either with Alexa 488-labeled goat anti-rabbit secondary antibody or anti-rabbit IgG nanobody TP897. The resulting mixes were then applied to fixed and blocked Hela cells. After washing, the cells were directly mounted for imaging. For every primary antibody, images were acquired under identical settings and pixel intensities are represented via a false-color lookup table.

(b) Multicolor-staining of HeLa cells with mouse lgG1 subclass mAbs. The indicated mouse lgG1 mAbs were separately pre-incubated with Alexa 488, Alexa 568 or Alexa 647-coupled anti-mouse lgG1 Fc nanobody TP1107 and then mixed before staining HeLa cells in a single step. Washed cells were directly mounted for imaging.

Examples

1. Methods

1.1 Alpaca immunization

Two female alpacas, held at the Max Planck Institute for Biophysical Chemistry, were immunized 4 times with 1.0 mg polyclonal mouse or rabbit IgG at 3 week intervals. The anti IgG project turned out to be the so far most challenging nanobody project in the lab, because we aimed at an extremely low off-rate for imaging and blotting applications. We therefore resumed immunizations after a 12 months (rabbit IgG) or an 8 months break (mouse IgG). Nanobodies obtained after these late immunizations still showed very clear phage enrichment (> 1000-fold) even with femtomolar concentrations of the IgG baits. We therefore assume that they have very high affinity.

1.2 Selection of anti-lgG nanobodies

The generation of nanobody immune libraries and the selection of antigen-specific nanobodies by phage display from these libraries were performed as previously described (Pleiner et al., 2015). IgG was biotinylated at accessible lysines by addition of a 4x molar excess of NHS-PEGi ₂ -biotin (Iris Biotech GmbH, Germany) for 2 h at room temperature in 1x PBS. Then the reaction was quenched and the excess of unreacted biotin separated from biotinylated IgG via buffer exchange into 50 mM Tris/HCI pH 7.5, 300 mM NaCI using PD-10 Desalting columns (GE Healthcare, USA).

1.3 Expression and purification of untagged nanobodies

Nanobodies with engineered cysteines were expressed in the cytoplasm of E. coli NEB express F' (New England Biolabs, USA). A 50 ml preculture (2YT medium containing 50 g/ml Kanamycin) was grown overnight at 28°C. The culture was then diluted with fresh medium to 250 ml. After 1 h of growth at 25°C, protein expression was induced for 3-5 h by adding 0.2 mM IPTG. After addition of 1 mM PMSF and 10 mM EDTA to the culture, bacteria were harvested by centrifugation, resuspended in lysis buffer (50 mM Tris/HCI pH 7.5, 300 mM NaCI, 10 mM imidazole, 5 mM DTT) and then lysed by sonication. The lysate was cleared by ultracentrifugation for 1.5 h (T647.5 rotor, Sorvall, 38,000 rpm) at 4°C. Nanobodies with engineered cysteines carried an N-terminal Hisi4-6cffslEDD8-tag and were affinity purified via Ni ²⁺ chelate affinity chromatography. After washing with two column volumes (CV) of lysis buffer and one CV of maleimide-labeling buffer (100 mM potassium phosphate pH 7.5, 150 mM NaCI, 1 mM EDTA, 250 mM Sucrose), untagged nanobodies were eluted by on-column cleavage with 500 nM bdNEDPI protease (Frey and Gorlich, 2014) in maleimide-labeling buffer for 45 min at 4°C and labeled immediately with fluorophores. For longer storage, 10 mM DTT or TCEP were included in the maleimide-labeling buffer to keep cysteines reduced. Purified nanobodies were aliquoted and frozen in liquid nitrogen

1.4 Site-specific fluorescent labeling of nanobodies with engineered cysteines

The fluorescent labeling of nanobodies with maleimide dyes was described in detail before (Pleiner et al., 2015). Briefly, stored nanobodies were thawed and the buffer was exchanged again to Maleimide-labeling buffer to remove the reducing agent, using either illustra NAP-5 or PD-10 desalting columns (GE Healthcare). For a standard labeling reaction, 5-10 nmoles of nanobody were rapidly mixed with 1.2x molar excess of fluorescent dye per cysteine on the nanobody and incubated for 1.5 h on ice. Free dye was separated from labeled nanobody by buffer exchange to Maleimide labeling buffer on illustra NAP-5 or PD-10 desalting columns. Quantitative labeling was quality controlled by calculating the degree of labeling (DOL). Fluorescently labeled nanobodies were always aliquoted, snap- frozen in liquid nitrogen and stored at -80°C until further use.

1.5 Dot blot assay for anti-lgG nanobody specificity profiling

For profiling the binding of anti-lgG nanobodies to different IgG subclasses and to analyze their crossreaction to IgG from other species, a dot blot assay was performed. Nitrocellulose membrane was cut in strips and different IgGs (500 ng for polyclonal total IgG, Fab and Fc fragments; ~250 ng for monoclonal IgG in 1 μΙ) were spotted. Strips were blocked with 4 % milk (w/v) in 1xPBS for 30 min at room temperature. Then, nanobodies were added at -300 nM in 1 ml milk for 30 min. After washing two times with 1x PBS for 10 min each, bound nanobodies were detected at 488 nm in a fluorescence scanner (Starion FLA-9000, Fujifilm, Japan). The following IgGs were used: lgG1 kappa mAb A225 (Cordes et al., 1995); lgG1 lambda (#010-001-331 , Rockland, USA); lgG2a kappa (#02-6200, Thermo Fisher Scientific, USA); lgG2b kappa (#02-6300, Thermo Fisher Scientific, USA); lgG3 kappa (#401302, BioLegend, USA); polyclonal IgG Fab fragments (#010-0105, Rockland, USA); polyclonal IgG Fc fragments (#31205, Thermo Fisher Scientific, USA). Polyclonal IgG of the following species were used: rabbit (self-made, affinity-purified from serum); mouse (#18765); rat (#14131); goat (#15256); sheep (#15131 ); human (#14506, all Sigma-Aldrich, USA) and guinea-pig (#CR4-10, Sino Biological, China).

1.6 Native isolation of IgG with anti-lgG nanobodies

Polyclonal rabbit IgG from serum or mouse mAbs from hybridoma cell culture supernatant were isolated natively with anti-lgG nanobodies. For this, 0.3 nmoles of biotinylated nanobodies carrying a N-terminal His -Biotin acceptor peptide- (GlySer) ₉ -SUMOStar-(GlySer) ₉ -tag were immobilized on 1 mg magnetic Dynabeads MyOne Streptavidin T1 (Thermo Fisher Scientific, USA). Excess biotin binding sites were quenched with biotin-PEG-COOH (#PEG1053, Iris Biotech, Germany). The beads were then incubated with 1 ml pre-cleared (10 min, 16,000g at 4°C) serum or hybridoma supernatant for 30 min at 4°C. After washing two times with wash buffer (50 mM Tris/HCI, 300 mM NaCI), nanobody-bound IgG was eluted by addition of 50 μΙ 0.5 μΜ SUMOStar protease (Liu et al., 2008) in wash buffer for 20 min on ice. An aliquot of the eluate was then analyzed by SDS-PAGE and Coomassie staining.

1.7 Western Blotting

Bacteriophage protein III was detected with a mouse anti-pill lgG2a mAb (#E8033S, New England Biolabs, USA). Mouse mAbs used for detection of human proteins in HeLa cell lysate were the following products: anti-Skpl (clone H-6, #sc- 5281 , Santa Cruz Biotechnology, USA), anti-a-tubulin (clone DM1 A, #T6199, Sigma-Aldrich, USA) and anti-Histone H3 (clone 96C10, #3638, Cell Signaling Technologies, USA). Polyclonal goat anti-mouse IgG coupled to IRDye 800CW (#925-32210; LI-COR Biosciences, USA) was used to detect primary mouse antibodies at a dilution of 1 :1340 (5 nM). Polyclonal rabbit antibodies against Xenopus laevis nucleoporins Nup98, Nup93, Nup54 and Nup88 were prepared in the lab (Hulsmann et al., 2012). Polyclonal goat anti-rabbit IgG coupled to IRDye 800CW (#925-3221 1 ; LI-COR Biosciences, USA) was used to detect primary rabbit antibodies at the lowest suggested dilution of 1 :5,000. Anti-mouse lgG1 Fab nanobody TP886 (5 nM), anti-mouse lgG1 Fc nanobody TP1 107 (5 nM) and anti- rabbit IgG nanobody TP897 (10 nM) were labeled with a single IRDye 800CW maleimide (#929-80020, LI-COR Biosciences, USA) via a C-terminal cysteine and used at the indicated concentrations in 4 % (w/v) milk in 1x PBS. Polyclonal goat anti-mouse-HRP conjugate was from DakoCytomation (Denmark) and used at 1 : 1 ,000 dilution (5 nM). Anti-mouse lgG1 Fc nanobody TP1 107 was conjugated to maleimide-activated HRP (#31485, Thermo Fisher Scientific, USA) via a C- terminal cysteine by mixing both in equimolar amounts and incubation for 1 h at room temperature. The conjugate was used at 5 nM in 4 % (w/v) milk in 1x PBS. The ECL solution was self-made and contained 5 mM Luminol (#A4685, Sigma- Aldrich, USA), 0.81 mM 4-lodophenylboronic acid (#471933, Sigma-Aldrich, USA) and 5 mM freshly added H ₂ 0 ₂ in 0.1 M Tris/HCI pH 8.8.

1.8 Amplex Ultra Red assay

APEX2 was derived from pTRC-APEX2 (Addgene plasmid #72558), which was a gift from Alice Y. Ting (Lam et al., 2015). The anti-mouse lgG1 Fc nanobody TP1 107-APEX2 fusion was expressed from pTP1 135 with an N-terminal Hisu- bc NEDD8-tag in E. coli NEB express F' (New England Biolabs, USA) for 6 h at 25°C in the presence of 1 mM of the heme precursor 5-aminolevulinic acid (#A3785, Sigma-Aldrich, USA). Following lysis, the protein was purified by nickel chelate affinity chromatography and eluted by cleavage with 500 nM £>c/NEDP1 protease (Frey and Gorlich, 2014) in 100 mM potassium phosphate pH 7.5, 150 mM NaCI, 250 mM sucrose. The final assay mix contained 160 μΜ Amplex Ultra Red, 160 μΜ H ₂ 0 ₂ in either 100 mM Citrate pH 6.6, 150 mM NaCI (optimal pH for APEX2) or 100 mM potassium phosphate pH 6.0, 150 mM NaCI (optimal pH for HRP). 50 μΙ of this mix was used per reaction. Anti-mouse lgG1 Fc nanobody TP1 107-APEX2 was titrated from 167 nM to 470 fM in a 1 .8-fold dilution series and 2 μΙ of each dilution added to 50 μΙ reaction mix in triplicates. HRP (#31490, Thermo Scientific, USA) was titrated from 31 nM to 5 fM in a 2.4-fold dilution series and 2 μΙ per dilution added to 50 μΙ reaction mix in triplicates. The 96-well plate containing these reactions was incubated at room temperature for 30 min and then resorufin fluorescence was measured at 590 nm (530 nm excitation) in a Bio-Tek Synergy HT Multi-Detection Microplate Reader (BioTek Instruments Inc., USA).

1.9 Immunofluorescence

HeLa cells grown on glass coverslips were fixed for 10 min at room temperature with 3 % (w/v) paraformaldehyde (PFA) and then washed two times with 1x PBS for 5 min each. Residual PFA was quenched by incubation with 50 mM NH ₄ CI in 1x PBS for 5 min. After two washes with 1x PBS for 5 min each, the cells were permeabilized with 0.3 % (v/v) Triton-X-100 for 3 min. Then the cells were washed three times quickly with 1x PBS and blocked for 30 min with 1 % (w/v) BSA in 1x PBS (blocking buffer). Following blocking, the coverslips were stained with primary antibody, which was diluted in blocking buffer, in a humid chamber for 1 h at room temperature. The coverslips were then washed two times in 1x PBS for 15 min each and added again to a humid chamber for incubation with secondary antibody or anti-lgG nanobody diluted in blocking buffer. Afterwards, the cells were washed two times in 1x PBS for 15 min each and the coverslips mounted with Slow Fade Gold (Thermo Fisher Scientific, USA) for imaging on a Leica TCS SP5 confocal microscope equipped with hybrid detectors (Leica, Germany). For methanol fixation, the cells were incubated with -20°C-cooled methanol for 6 min at room temperature, washed two times in 1x PBS for 5 min each and then blocked in blocking buffer. The staining was performed as described above.

1.10 Antibodies for immunofluorescence

The following rabbit antibodies were used for immunofluorescence on HeLa cells: anti-Lap2 polyclonal antibody (1 :100 dilution, #14651 -1-AP, Proteintech, UK); anti- Ki-67 mAb clone D3B5 (1 :200 dilution, #9129, Cell Signaling Technologies, USA). The following mouse mAbs were used for immunofluorescence on HeLa cells: anti-Vimentin mAb clone V9 (1 :10 dilution of Hybridoma supernatant, kind gift of Mary Osborn); anti-Ki-67 mAb clone B56 (1 :50 dilution, #556003, BD Bioscience, USA); anti-TPR mAb 203-37 (1 :500 dilution, Matritech Inc., USA; (Cordes et al., 1997)); anti-Cytochrome (Cyt) c mAb clone 6H2.B4 (1 :50 dilution, #556432, BD Bioscience, USA); anti-Lamin A/C mAb clone 4C1 1 (1 :50 dilution, #4777T, Cell Signaling Technologies, USA); anti-CD44 mAb clone 156-3C1 1 (1 :200 dilution, #3570T, Cell Signaling Technologies, USA). Polyclonal goat anti-rabbit IgG (#1 1 1 - 545-003) and goat anti-mouse IgG (#1 15-545-003, Jackson ImmunoResearch, USA) coupled to Alexa Fluor 488 were used as secondary antibodies at 1 :150 dilution (~33 nM). Anti-IgG nanobodies were labeled with maleimide Alexa Fluor dyes at engineered surface cysteines (Pleiner et al., 2015) and used at 20 nM. The used nanobodies had the following degree of labeling: TP886-Alexa 488 = 1 .9, TP1 107-Alexa 488 = 2.7, TP1 107-Alexa 647 = 2.2, TP1 129-Alexa 488 = 2.5, TP1 129-Alexa 568 = 2.0, TP1079-Alexa 488 = 2.2, TP897-Alexa 488 = 2.2.

2. Results

2.1 A comprehensive anti-lgG nanobody toolbox

We immunized two alpacas separately with polyclonal mouse or rabbit IgG and used chemically biotinylated mouse monoclonal antibodies (mAbs) of defined subclasses as well as rabbit IgGs for phage display selections of nanobodies from the resulting immune libraries. First results with the initially obtained anti-lgG nanobodies were rather disappointing, i.e. we experienced dim and noisy signals in immunofluorescence as well as in Western blots. We reasoned that an increase in affinity and specificity might yield improved reagents and therefore re- immunized the animals after a one-year pause. For this, we used IgGs pre-bound to multivalent particulate antigens expected to provide strong T-helper cell epitopes. Moreover, we increased the stringency of the subsequent phage display selections by lowering the bait concentration down to the femtomolar range, which should not only select per se for sub-nanomolar binders, but also bring displayed nanobodies in direct competition with each other, because the number of bait molecules was up to 1000-fold lower than the number of displaying phages. Finally, we performed in vitro affinity maturations by random mutagenesis and further rounds of phage display, this time also combined with off-rate selections. In this way, we obtained a large toolkit of anti-rabbit and anti-mouse IgG nanobodies (Figure 1 a).

All nanobodies were extensively characterized for subclass specificity, epitope location on Fab or Fc fragment and crossreactivity to IgGs from other species (Table 1 , Figure 1 b, Figure 1 - figure supplement 1a). Their full protein sequences are listed in Table 2. Notably, we identified nanobodies against all four mouse IgG subclasses and the sole rabbit IgG subclass. Strikingly, many anti-mouse IgG nanobodies target lgG1 , which represents the most abundant subclass of commercially available mouse mAbs (-62-64 %), followed by lgG2a (-22-24 %) and the less frequent lgG2b (-13 %) and lgG3 (-1-2 %). Since the vast majority (-99 %) of mouse mAbs possess a kappa light chain, anti-kappa chain nanobodies promised to be the most broadly useful tools and we therefore actively selected for such binders by swapping the IgG heavy chain subclass during sequential selection rounds. For the identification of binders targeting the rare lambda chain, we had to pre-deplete the nanobody immune library of heavy chain and kappa chain-binders. Some of the identified nanobodies have mixed specificities, e.g. multiple mouse Fab-binders target an interface between kappa light chain and lgG1 or lgG2a heavy chain. Most anti-mouse IgG nanobodies are exclusively mouse-specific, while others additionally crossreact with rat IgG (Figure 1 - figure supplement 1a). The anti-rabbit IgG nanobody TP897 also efficiently recognizes guinea pig IgG. All nanobodies were produced by cytoplasmic expression in E. coli, mostly with an N-terminal His-NEDD8-tag for purification by Ni(ll) chelate affinity capture and proteolytic release (Frey and Gorlich, 2014). They were further equipped with ectopic cysteines for subsequent maleimide labeling reactions (Pleiner et al., 2015). Without further optimization, we typically obtained yields of 15 mg per liter of bacterial culture, which already suffices for a million immunofluorescence stains or 200 liters of Western blotting solution (see below).

We first assessed if the anti-IgG nanobodies were specific and could purify their IgG target from its common source. Anti-rabbit IgG nanobodies TP896 and TP897 isolated polyclonal rabbit IgG from crude rabbit serum with high specificity (Figure 1 - figure supplement 1 b). Likewise, anti-mouse IgG nanobodies TP881 and TP885 could purify an lgG1 mAb from hybridoma cell culture supernatant (Figure 1 - figure supplement 1c). Notably, nanobody-bound IgG was released under physiological conditions using SUMOStar protease cleavage (Pleiner et al., 2015). The main virtue of this approach is perhaps not to purify IgGs from sera, but to perform immune-affinity purifications of antigens or antigen complexes that have been pre-bound to the primary antibodies. In contrast to traditional IPs, this approach allows to release the purified complexes under fully native conditions.

2.2 Western blotting with horseradish peroxidase-conjugated anti-lgG nanobodies

We next tested the performance of anti-lgG nanobodies as detection reagents in Western Blotting, which is a major application for secondary antibodies. A popular mode of signal detection in Western Blotting is enhanced chemiluminescence (ECL) in which antibody-horseradish peroxidase (HRP) conjugates are used. HRP is a heme-containing enzyme that catalyzes the oxidation of luminol in the presence of H2O2 to yield bright chemiluminescence, which is greatly increased by phenol-derived enhancers. We conjugated maleimide-activated HRP to anti- mouse lgG1 Fc nanobody TP1107 via a C-terminal cysteine (Figure 2 - figure supplement 1a) and used the resulting conjugate in ECL Western Blotting. The nanobody-HRP conjugate is functional and outperformed a polyclonal secondary antibody-HRP conjugate from a commercial supplier (Figure 2a). The anti-rabbit IgG nanobody TP897 could also be linked to HRP and the resulting conjugate was functional and specific. 2.3 Comparison with commercially available anti-lgG nanobodies

Commercially available anti-mouse lgG1 or anti-rabbit IgG single-domain nanobodies designated as ab193651 and ab191866 (Abeam, United Kingdom) were compared with anti-mouse lgG1 nanobody TP1107 and anti-rabbit IgG nanobody TP897 as HRP conjugates (Figure 2b). The nanobodies of the invention provide a substantially higher sensitivity when used at equal concentrations.

2.4 Recombinant ascorbate peroxidase fusion to anti-lgG nanobodies

Due to its stability and the breadth of its catalyzed colorimetric or chemiluminescent reactions that allow strong signal amplification, HRP is the most preferred enzyme for conjugation to secondary antibodies. However, it still has to be isolated from horseradish roots as a mixture of different isoforms, cannot be made in a practical scale and with a useful specific activity in E. coli (Krainer and Glieder, 2015), and it fails entirely as a genetic fusion to bacterially expressed nanobodies. As an alternative, we tested the engineered APEX2 ascorbate peroxidase (Martell et al., 2012; Lam et al., 2015) as a fusion partner of the anti-mouse lgG1 Fc nanobody TP1 107. The TP1 107-APEX2 fusion was not only well-expressed and soluble in E. coli (Figure 2 - figure supplement 1 b), but also, it was active and efficiently catalyzed the oxidation of the initially colorless substrate Amplex Ultra Red to the highly fluorescent resorufin (Figure 2b). In line with previous reports (Lam et al., 2015), HRP seemed slightly more efficient than APEX2 in catalyzing this reaction. Nonetheless, low femtomole amounts of TP1 107-APEX2 could be detected, suggesting its applicability e.g. in ELISA assays as well for immunohistochemistry and enzymatic antigen-localization in immunoelectron microscopy applications.

2.5 Western Blotting with infrared fluorophore-linked anti-lgG nanobodies

A convenient alternative to peroxidase conjugation or fusion is the labeling of secondary antibodies with infrared fluorescent dyes. In fact, infrared fluorescent Western blotting has emerged as a superior alternative to classical ECL. It offers high signal-to-noise ratios, allows straightforward quantification due to signal linearity over many orders of magnitude and even enables the simultaneous dual color detection of multiple proteins. We thus labeled anti-lgG nanobodies site- specifically with the infrared fluorophore IRDye 800 at a C-terminal cysteine (Pleiner et al., 2015). The anti-rabbit IgG nanobody TP897 alone performed just as well as a commercial polyclonal anti-rabbit IgG secondary antibody, when it was used with rabbit polyclonal antibodies to detect various nucleoporins (Nups) in a Xenopus egg extract (Figure 3a). Similarly, the anti-mouse lgG1 Fc-specific nanobody TP1 107 gave comparable or even higher signal intensities than a polyclonal anti-mouse IgG secondary antibody in Western Blotting on HeLa cell lysate (Figure 3b). Combinations of TP1 107 with the compatible anti-mouse lgG1 Fab-specific nanobody TP886 or the anti-mouse kappa chain nanobody TP1170 provided a clearly better detection sensitivity than the polyclonal secondary antibody. TP1 170 allows sensitive detection of lgG2a subclass mAbs, as shown here for the detection of the bacteriophage minor coat protein pill (Figure 3c). We routinely found infrared fluorophore-labeled anti-lgG nanobodies to yield higher detection sensitivity than their HRP-conjugated counterparts. When combined with the compatible IRDye 680, dual color blots using e.g. mouse and rabbit primary antibodies are easily possible (not shown). In contrast to polyclonal secondary antibodies, IRDye-labeled anti-IgG nanobodies give also a clean and strong signal when pre-bound to primary antibodies before application. This makes a separate incubation with the secondary antibody dispensable and saves up to 2 h processing time per blot. We explored such a one-step staining strategy in more detail below for immunofluorescence.

2.6 Single and multi-color imaging with anti-IgG nanobodies

We next sought to assess the performance of the anti-IgG nanobodies as detection reagents in conventional indirect immunofluorescence. For this, cells are incubated sequentially with primary and secondary antibodies with intervening washing steps. Fluorophore-linked polyclonal secondary antibodies are routinely used for detection, since they can bind primary antibodies at multiple sites and thus deliver many fluorophores to enable large signal amplification. In contrast, individual anti-IgG nanobodies target only a single epitope per antibody (or two for symmetrical binding sites) and we therefore expected only modest signal amplification. Strikingly however, the anti-lgG1 nanobodies TP886 and TP1107, which specifically target lgG1 Fab and Fc fragment, respectively, not only performed well in Western Blotting, but also were well-behaved imaging reagents. For maximum brightness, we labeled these nanobodies with 2-3 fluorophores each at defined cysteines (Pleiner et al., 2015) and used them individually for the detection of mouse lgG1 mAbs in an indirect HeLa cell immunostaining (Figure 4a). Surprisingly, both were only slightly dimmer than the polyclonal mixture of anti-mouse secondary antibodies. We assume that the excellent nanobody signal is also due to less steric hindrance as compared to the much larger conventional secondary antibody. When both nanobodies were used in combination, we detected increased signal strengths that often were directly comparable to those obtained with the secondary antibody (e.g. for Vimentin or Ki-67) (see also Figure 4 - figure supplement 1a). Importantly, despite a high labeling density with (the always somewhat sticky) fluorophores, we observed no detectable background staining with these anti-IgG nanobodies. This probably relates to the fact that the affinity of our nanobodies is very high, which allows their use at rather low nanomolar concentrations. The poor performance of the first anti-IgG nanobody generation indeed suggests that such excellent signal to noise ratio is not a trivial feature for a monovalent detection reagent.

For the detection of lgG2a subclass mAbs, we used a combination of two nanobodies, TP1129 and TP1170 (Figure 4b, Figure 4 - figure supplement 1 b). The lgG2a-specific nanobody TP1129 targets an epitope on the Fc-fragment and was obtained after affinity maturation of a lower affinity precursor (Figure 4 - figure supplement 1c). Likewise, the kappa chain-specific nanobody TP1170 is an affinity-optimized variant, obtained after error-prone PCR, DNA shuffling and affinity selection (Figure 4 - figure supplement 1d). TP1170 also proved effective in combination with the anti-lgG1 Fc nanobody TP1107 for the detection of lgG1 kappa mAbs (Figure 4 - figure supplement 1e and 1f). The anti-rabbit IgG Fc nanobody TP897 can be used for the detection of polyclonal and monoclonal rabbit IgG (Figure 4c).

The presented nanobodies are specific for their respective IgG subclass, as shown in the specificity profiling dot blot assay (Figure 1 b). We exploited this for multicolor imaging of HeLa cells with different IgG subclasses (Figure 4d). Mouse lgG1 , mouse lgG2a and rabbit IgG-specific nanobodies did not show any crossreaction and consequently allowed for clean co-localization experiments. Even triple co-localizations were readily possible.

2.7 Rapid one-step immunostaining and co-localization

The main reasons for separate incubation steps of primary and secondary IgGs in indirect immunofluorescence and Western blotting are the large size, as well as the bivalent and polyclonal nature of conventional secondary antibodies. If primary and secondary antibodies are pre-incubated, large oligomeric complexes form, which in immunofluorescence cannot easily penetrate into cells to reach their target and thus create background and poor signal (see Figure 5a). In contrast, anti-lgG nanobodies are monovalent and therefore do not crosslink primary antibodies. This allows streamlining the conventional immunostaining procedure to a single step. The primary antibodies are simply pre-incubated with fluorescently labeled anti-lgG nanobodies and then applied to cells together. After washing, the cells can be directly mounted for imaging. In such a workflow, anti-lgG nanobodies perform exceptionally well (Figure 5a). This time-saving protocol is also suitable for co-localization studies combining mouse and rabbit IgGs or combining mouse mAbs of different sub-classes.

If the off-rate of the IgG pre-bound nanobodies were negligible over the staining period, then an exchange between the different pre-formed complexes would also be negligible. This would also make it unnecessary to use different IgG subclasses for multicolor imaging. We thus tested a multicolor staining workflow of HeLa cells relying solely on lgG1 subclass mAbs (Figure 5b). For this, we labeled anti-lgG1 Fc nanobody TP1107 with either Alexa 488, Alexa 568 or Alexa 647 maleimide and pre-incubated it with different lgG1 mAbs. The separately pre-incubated mixes were then combined and applied to HeLa cells for staining in one-step. Strikingly, we obtained clean dual and even triple co-localizations. In order to preclude an intermixing of colors, unlabeled TP1107 can be added in excess to the final mix and cells can be post-fixed after staining and washing.

3. Discussion

Due to the absence of more sustainable alternatives in the past, the great usefulness of polyclonal secondary antibodies in basic research certainly justified their animal-based production. However, in order to guarantee their constant supply to an ever-growing market, the producing companies had to dramatically increase their livestock, aim for very high antibody titres using aggressive hyper- immunization strategies causing strong side effects and increase the frequency and volume of collected bleedings. It is therefore not surprising that the global industrial scale production of antibodies causes severe animal welfare and ethical problems. The magnitude of these problems recently surfaced in the Santa Cruz Biotechnology scandal (Shen, 2013; Reardon, 2016).

Ideally, one should replace all animal immunization by selecting binders from synthetic libraries (Gray et al., 2016; Moutel et al., 2016; McMahon et al., 2017; Zimmermann et al., 2017). Yet, with a purely synthetic approach it is still not straightforward to obtain high-affinity binders. Further, the synthetic strategy is typically also inferior in terms of binder-specificity, because it lacks the stringent selection against self-reactivity that happens in antigen-exposed animals. The requirement for specificity is particularly high for secondary antibodies. We therefore see the here applied approach of using an immune library for binder selection as the best possible compromise. Since it is generally sufficient to obtain a few good nanobodies out of a small blood sample containing -100 million lymphocytes, and since we found ways of further improving the initially found ones in vitro, there was no need for any hyper-immunization aiming at high titers. Importantly, once ideal nanobodies are identified, they are defined by their sequence and they can be renewably produced in E. coli at constant quality and without any further animal involvement. Since polyclonal secondary antibody production accounts for the largest share of immunized animals in the world, the anti-IgG nanobodies described in this study have the potential to make a great step forward towards reducing animal use and further contribute to a future of standardized recombinant antibodies (Marx, 2013; Bradbury and Pluckthun, 2015a; Bradbury and Pluckthun, 2015b).

We expect that our anti-IgG nanobodies will replace polyclonal secondary antibodies in many of their applications, e.g. in Western blotting and immunofluorescence. For both applications, their site-specific and quantitative modification with fluorophores via maleimide chemistry creates superior reagents with predictable label density and position. Furthermore, the precise targeting of primary mouse antibodies at the kappa chain with a specific nanobody could substantially reduce the label displacement in super-resolution microscopy. In the future, we will also explore the direct coupling of anti-IgG nanobodies with engineered cysteines onto colloidal gold particles for electron microscopy, which also suffers from the large linkage error introduced by bulky secondary antibodies.

Due to their monovalent and monoclonal nature, anti-IgG nanobodies do not crosslink primary antibodies and we exploited this for a one-step immunostaining workflow that saves valuable hands-on time and can also be extended to Western blotting. We envision that for routine stainings, preformed complexes of primary antibodies and labeled nanobodies can be prepared as stock solutions or simply bought from commercial suppliers. Due to the high affinity of the described nanobodies, the same strategy also enables multicolor immunostainings based on a single IgG subclass, which could also be relevant for flow cytometry sorting of specific cell types. This would be a cheaper and more flexible alternative to differentially labeled primary antibodies, it does not pose the risk of inactivating an antigen-binding site and it can easily be done if only small amounts of primary antibody are available.

Further, since the DNA sequences of these anti-IgG nanobodies are essentially synthetic building blocks, they can be genetically appended to the multitude of available tags, fluorescent proteins or enzymes to generate fusion proteins with novel functions for tailored applications in basic research and medical diagnostics, and also become valuable tools for immunology to study Fc or B cell receptors and downstream signaling cascades. Furthermore, anti-IgG nanobodies equipped with protease-cleavable affinity tags (Pleiner et al., 2015) will allow the native isolation of any antibody-target complex e.g. for structural studies by cryo-EM or functional assays. Even though the here presented anti-IgG nanobody toolbox is already highly optimized, we will continue to extend it by identifying new nanobodies that decorate complementary binding sites and thus allow a further signal enhancement, and combine them with additional functional elements. In any case, it will be an open resource for all interested labs.

References

Arbabi Ghahroudi, M., Desmyter, A., Wyns, L, Hamers, R. & Muyldermans, S.

(1997) Selection and identification of single domain antibody fragments from camel heavy-chain antibodies. FEBS Lett, 414, 521-526.

Balzarotti, F., Eilers, Y., Gwosch, K.C., Gynna, A.H., Westphal, V., Stefani, F.D., Elf, J. & Hell, S.W. (2017) Nanometer resolution imaging and tracking of fluorescent molecules with minimal photon fluxes. Science, 355, 606-612. Bates, M., Dempsey, G.T., Chen, K.H. & Zhuang, X. (2012) Multicolor super- resolution fluorescence imaging via multi-parameter fluorophore detection. Chemphyschem, 13, 99-107.

Bates, M., Huang, B., Dempsey, G.T. & Zhuang, X. (2007) Multicolor super- resolution imaging with photo-switchable fluorescent probes. Science, 317, 1749-1753.

Bradbury, A. & Pliickthun, A. (2015a) Reproducibility: Standardize antibodies used in research. Nature, 518, 27-29.

Bradbury, A.R. & Pluckthun, A. (2015b) Getting to reproducible antibodies: the rationale for sequenced recombinant characterized reagents. Protein Eng

Des Sel, 28, 303-305.

Cordes, V.C., Reidenbach, S. & Franke, W.W. (1995) High content of a nuclear pore complex protein in cytoplasmic annulate lamellae of Xenopus oocytes.

Eur J Cell Biol, 68, 240-255.

Cordes, V.C., Reidenbach, S., Rackwitz, H.R. & Franke, W.W. (1997)

Identification of protein p270/Tpr as a constitutive component of the nuclear pore complex-attached intranuclear filaments. J Cell Biol, 136, 515-529.

Desmyter, A., Spinelli, S., Roussel, A. & Cambillau, C. (2015) Camelid nanobodies: killing two birds with one stone. Curr Opin Struct Biol, 32C, 1 -8. Frey, S. & Gorlich, D. (2014) A new set of highly efficient, tag-cleaving proteases for purifying recombinant proteins. J Chromatogr A, 1337, 95-105.

Gottfert, F., Pleiner, T., Heine, J., Westphal, V., Gorlich, D., Sahl, S.J. & Hell, S.W.

(2017) Strong signal increase in STED fluorescence microscopy by imaging regions of subdiffraction extent. Proc Natl Acad Sci U S A, 114, 2125-2130. Gray, A.C., Sidhu, S.S., Chandrasekera, P.C., Hendriksen, C.F. & Borrebaeck,

C.A. (2016) Animal-Friendly Affinity Reagents: Replacing the Needless in the

Haystack. Trends Biotech nol, 34, 960-969.

Hamers-Casterman, C, Atarhouch, T., uyldermans, S., Robinson, G., Hamers,

C, Songa, E.B., Bendahman, N. & Hamers, R. (1993) Naturally occurring antibodies devoid of light chains. Nature, 363, 446-448.

Helma, J., Cardoso, M.C., Muyldermans, S. & Leonhardt, H. (2015) Nanobodies and recombinant binders in cell biology. J Cell Biol, 209, 633-644.

Huang, B., Babcock, H. & Zhuang, X. (2010) Breaking the diffraction barrier: super-resolution imaging of cells. Cell, 143, 1047-1058.

Huang, F., Sirinakis, G., Allgeyer, E.S., Schroeder, L.K., Duim, W.C., Kromann,

E.B., Phan, T., Rivera-Molina, F.E., Myers, J.R., Irnov, I., Lessard, M., Zhang,

Y., Handel, M.A., Jacobs-Wagner, C, Lusk, CP., Rothman, J.E., Toomre, D.,

Booth, M.J. & Bewersdorf, J. (2016) Ultra-High Resolution 3D Imaging of

Whole Cells. Cell, 166, 1028-1040. Hulsmann, B.B., Labokha, A.A. & Gorlich, D. (2012) The permeability of reconstituted nuclear pores provides direct evidence for the selective phase model. Cell, 150, 738-751 .

Kijanka, M., Dorresteijn, B., Oliveira, S. & van Bergen en Henegouwen, P.M.

(2015) Nanobody-based cancer therapy of solid tumors. Nanomedicine

(Lond), 10, 161 -174.

Krainer, F.W. & Glieder, A. (2015) An updated view on horseradish peroxidases: recombinant production and biotechnological applications. Appl Microbiol

Biotechnol, 99, 161 1 -1625.

Lam, S.S., Martell, J.D., Kamer, K.J., Deerinck, T.J., Ellisman, M.H., Mootha, V.K.

& Ting, A.Y. (2015) Directed evolution of APEX2 for electron microscopy and proximity labeling. Nat Methods, 12, 51 -54.

Liu, L., Spurrier, J., Butt, T.R. & Strickler, J.E. (2008) Enhanced protein expression in the baculovirus/insect cell system using engineered SUMO fusions. Protein Expr Purif, 62, 21 -28.

Martell, J.D., Deerinck, T.J., Sancak, Y., Poulos, T.L., Mootha, V.K., Sosinsky,

G.E., Ellisman, M.H. & Ting, A.Y. (2012) Engineered ascorbate peroxidase as a genetically encoded reporter for electron microscopy. Nat Biotechnol, 30,

1 143-1 148.

Marx, V. (2013) Calling the next generation of affinity reagents. Nat Methods, 10,

829-833.

McMahon, C, Baier, A.S., Zheng, S., Pascolutti, R., Ong, J.X., Erlandson, S.C., Hilger, D., Ring, A.M., Manglik, A. & Kruse, A.C. (2017) Platform for rapid nanobody discovery in vitro. bioRxiv, doi: https://doi.org/10.1 101/151043. Moutel, S., Bery, N., Bernard, V., Keller, L., Lemesre, E., de Marco, A., Ligat, L., Rain, J.C., Favre, G., Olichon, A. & Perez, F. (2016) NaLi-H1 : A universal synthetic library of humanized nanobodies providing highly functional antibodies and intrabodies. Elite, 5,

Muyldermans, S. (2013) Nanobodies: natural single-domain antibodies. Annu Rev Biochem, 82, 775-797.

Pleiner, T., Bates, M., Trakhanov, S., Lee, C.T., Schliep, J.E., Chug, H., Bohning, M., Stark, H., Urlaub, H. & Gorlich, D. (2015) Nanobodies: site-specific labeling for super-resolution imaging, rapid epitope-mapping and native protein complex isolation. Elite, 4, e1 1349. Reardon, S. (2016) US government issues historic $3.5-million fine over animal welfare. Nature, doi:10.1038/nature.2016.19958.

Ries, J., Kaplan, C, Platonova, E., Eghlidi, H. & Ewers, H. (2012) A simple, versatile method for GFP-based super-resolution microscopy via nanobodies. Nat Methods, 9, 582-584.

Rust, M.J., Bates, M. & Zhuang, X. (2006) Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat Methods, 3, 793-

795.

Sahl, S.J., Hell, S.W. & Jakobs, S. (2017) Fluorescence nanoscopy in cell biology.

Nat Rev Mol Cell Biol,

Shen, H. (2013) Discovery of goat facility adds to antibody provider's woes.

Nature, doi: 10.1038/nature.2013.12203.

Szymborska, A., de Marco, A., Daigle, N., Cordes, V.C., Briggs, J.A. & Ellenberg,

J. (2013) Nuclear pore scaffold structure analyzed by super-resolution microscopy and particle averaging. Science, 341 , 655-658.

Traenkle, B. & Rothbauer, U. (2017) Under the Microscope: Single-Domain

Antibodies for Live-Cell Imaging and Super-Resolution Microscopy. Front

Immunol, 8, 1030.

Van Bockstaele, F., Holz, J.B. & Revets, H. (2009) The development of nanobodies for therapeutic applications. Curr Opin Investig Drugs, 10, 1212-

1224.

Weber, K., Rathke, P.C. & Osborn, M. (1978) Cytoplasmic microtubular images in glutaraldehyde-fixed tissue culture cells by electron microscopy and by immunofluorescence microscopy. Proc Natl Acad Sci U S A, 75, 1820-1824. Xu, K., Babcock, H.P. & Zhuang, X. (2012) Dual-objective STORM reveals three- dimensional filament organization in the actin cytoskeleton. Nat Methods, 9, 185-188.

Zimmermann, I., Egloff, P., Hutter, C, Stohler, P., Bocquet, N., Hug, M., Siegrist, M., Svacha, L., Gera, J. & Gmuer, S. (2017) Synthetic single domain antibodies for the conformational trapping of membrane proteins. bioRxiv, doi: https://doi.Org/10.1101/168559.