Available online at www.sciencedirect.com
ScienceDirect Camelid nanobodies: killing two birds with one stone Aline Desmyter1,2, Silvia Spinelli1,2, Alain Roussel1,2 and Christian Cambillau1,2 In recent years, the use of single-domain camelid immunoglobulins, termed vHHs or nanobodies, has seen increasing growth in biotechnology, pharmaceutical applications and structure/function research. The usefulness of nanobodies in structural biology is now firmly established, as they provide access to new epitopes in concave and hinge regions — and stabilize them. These sites are often associated with enzyme inhibition or receptor neutralization, and, at the same time, provide favorable surfaces for crystal packing. Remarkable results have been achieved by using nanobodies with flexible multi-domain proteins, large complexes and, last but not least, membrane proteins. While generating nanobodies is still a rather long and expensive procedure, the advent of naive libraries might be expected to facilitate the whole process. Addresses 1 Aix-Marseille Universite´, Architecture et Fonction des Macromole´cules Biologiques, France 2 Centre National de la Recherche Scientifique, AFMB, UMR 7257, case 932, 13288 Marseille Cedex 09, France Corresponding author: Cambillau, Christian ([email protected])
Current Opinion in Structural Biology 2015, 32:1–8 This review comes from a themed issue on New constructs and expressions of proteins Edited by Imre Berger and Roslyn M Bill
http://dx.doi.org/10.1016/j.sbi.2015.01.001 0959-440X/# 2015 Elsevier Ltd. All rights reserved.
Introduction In the early 1990s, the Hamers group in Brussels discovered that camelids (llama, dromedary, camel) simultaneously possess two different types of immunoglobulin (Ig) [1]: the classical ones, composed of a light chain with two Ig domains and a heavy chain with four such domains, and a new type of Ig with only a heavy chain with three Ig domains (Figure 1a). In practice, the antigen recognition module for classical antibodies involves the assembly of two independently folded domains, the N-terminal Ig domains of the light chain (VL) and the heavy chain (VH). For biotechnological or pharmaceutical uses, the classical Ig recognition module can be formatted as a four www.sciencedirect.com
immunoglobulin domain Fab fragment (Figure 1a) or as a Fv fragment (VL-VH; Figure 1a), sometimes joined by a linker (scFv). These modules require the coexpression of two chains and their independent folding and assembly via hydrophobic interfaces. This difficult process is better suited to expression in eukaryotic cells than in Escherichia coli. In contrast, the recognition module of camelid Ig involves a unique Ig domain (called VHH or nanobody) in which the hydrophobic VH-VL interface is replaced by a hydrophilic surface, thanks to the framework mutations Gly44Glu and Leu45Arg [2–4]. Indeed, this module has been shown in most instances to be expressed in good yields in E. coli. Initial questions were raised, however, concerning the antigen affinity of nanobodies compared to Fabs/Fvs, as they harbor only three complementarity-determining regions (CDRs), the hypervariable loops that bind to antigens, compared to the six CDRs of Fabs/Fvs (Figure 1a, b). Sub-nanomolar/picomolar affinities for nanobodies against a wide range of antigens have been reported, however, indicating that nanobodies could be as efficient binders as Fabs/Fvs. Furthermore, nanobodies revealed unexpected and unique antigen binding characteristics. Studies of the first structure of a nanobody with an antigen, egg white lysozyme, revealed that its CDR3 inserts deeply into the enzyme active site [2]. The nanobody CDR3s are significantly longer than in classical VHs. Nanobody CDR3s also contribute to covering the nanobody surface corresponding to the VL/VH interface, and they often harbor a disulfide bond joining them to the rest of the domain [2,4]. The ‘protruding’ behavior of nanobodies, although not general, has been shown in many instances and is one of their most interesting features: nanobodies most likely bind to concave surfaces, while Fabs/Fvs bind to planar or convex surfaces. Another peculiar behavior of nanobodies is their ability to recognize the antigen (protein or even hapten) laterally, i.e., using the external surface of the CDR3 together with some patches of the framework [5,6]. These unique features have drawn considerable interest from the biotechnology and health/ pharmaceutical fields, as nanobodies are potentially potent enzyme or receptor inhibitors [7]. Finally, the use of nanobodies in structural biology is now well established and is the subject of this review.
Nanobody generation and selection The first step of nanobody generation is generally performed through immunization of a llama (Lama glama) or a dromedary (Camelus dromedarius). Llamas are generally preferred to dromedaries for ease of access, as several companies and public institutions provide immunization Current Opinion in Structural Biology 2015, 32:1–8
2 New constructs and expressions of proteins
Figure 1
(a)
(b)
VH
6 CDR
vHH 3 CDR or nanobody
CH1 VL
Fv
Fab
linker
CL
Fc
(c)
nanobody 1
Fc
diabody Ct GSn repeat
n
C
nanobody 2 : - same antigen, =/ epitope - =/ antigens - HSA (low clearance)
Nt Current Opinion in Structural Biology
Classical and camelid nanobodies. (a) Composite representation of a classical IgG, with heavy (blue) and light (magenta) chains. Two recognition modules, a Fab fragment (VH, VL, CH1, CL) and a Fv fragment (VH, VL), are identified. Six Complementary Determining Regions (CDRs) form the antigen-binding surface. (b) Composite representation of a camelid IgG, with a heavy chain only (blue). The CH1 domain is replaced by a linker. The recognition module is a unique immunoglobulin domain called vHH or nanobody, with three CDRs. (c) Composite representation of a dual nanobody construct (diabody). This construct makes it possible to target two different epitopes on the same antigen, thus tremendously increasing avidity, or to target two different antigens. Diabodies including human serum albumin (HSA) display a longer clearance.
facilities. Nanobody selection follows a well-established protocol [8,9]. Immunization is performed through 5 or 6 injections, one per week, of 0.5 mg protein (Figure 2). Notably, several antigens can be injected simultaneously. After the last immunization, samples of blood are collected, and the lymphocytes are purified. The total RNA from the peripheral blood lymphocytes is extracted and used as a template for first strand cDNA synthesis with an oligo (dT) primer. Using this cDNA, the VHH encoding sequences are amplified by PCR and cloned into the phage-display phagemid vector pHEN4, leading to a nanobody library of approximately 108 independent transformants. Three consecutive rounds of phage display and panning are performed on solid-phase coated protein. The enrichment for antigen-specific phages within the pools at each round of panning is assessed by polyclonal phage ELISA. A clear enrichment is generally seen after the second and third rounds of panning. Tens of individual colonies are randomly selected and analyzed by ELISA for the presence of antigen-specific nanobodies in their periplasmic extracts. The nanobody sequences from the ELISA-positive colonies are subjected to RFLP analysis and nucleotide sequencing. The whole nanobody selection process takes approximately 70 days (2-3 months) from the Current Opinion in Structural Biology 2015, 32:1–8
first injection (Figure 2). Alternatively, immunization can be avoided when using a naive library [10–12]. Naive libraries are generated by collecting the blood of several non-immunized animals. The library is then obtained by the same procedure as described above (Figure 2). The library should contain a larger number of independent transformants (5 108 to 109). Panning selection then isolates pre-existing binders. However, amplification is slower with naive libraries, and 4 rounds of panning might be necessary to obtain a good binder; however, the overall selection time is reduced to 3 weeks. Naive libraries for use in structural biology applications should develop in the future because the need for extremely good binders (
ScienceDirect Camelid nanobodies: killing two birds with one stone Aline Desmyter1,2, Silvia Spinelli1,2, Alain Roussel1,2 and Christian Cambillau1,2 In recent years, the use of single-domain camelid immunoglobulins, termed vHHs or nanobodies, has seen increasing growth in biotechnology, pharmaceutical applications and structure/function research. The usefulness of nanobodies in structural biology is now firmly established, as they provide access to new epitopes in concave and hinge regions — and stabilize them. These sites are often associated with enzyme inhibition or receptor neutralization, and, at the same time, provide favorable surfaces for crystal packing. Remarkable results have been achieved by using nanobodies with flexible multi-domain proteins, large complexes and, last but not least, membrane proteins. While generating nanobodies is still a rather long and expensive procedure, the advent of naive libraries might be expected to facilitate the whole process. Addresses 1 Aix-Marseille Universite´, Architecture et Fonction des Macromole´cules Biologiques, France 2 Centre National de la Recherche Scientifique, AFMB, UMR 7257, case 932, 13288 Marseille Cedex 09, France Corresponding author: Cambillau, Christian ([email protected])
Current Opinion in Structural Biology 2015, 32:1–8 This review comes from a themed issue on New constructs and expressions of proteins Edited by Imre Berger and Roslyn M Bill
http://dx.doi.org/10.1016/j.sbi.2015.01.001 0959-440X/# 2015 Elsevier Ltd. All rights reserved.
Introduction In the early 1990s, the Hamers group in Brussels discovered that camelids (llama, dromedary, camel) simultaneously possess two different types of immunoglobulin (Ig) [1]: the classical ones, composed of a light chain with two Ig domains and a heavy chain with four such domains, and a new type of Ig with only a heavy chain with three Ig domains (Figure 1a). In practice, the antigen recognition module for classical antibodies involves the assembly of two independently folded domains, the N-terminal Ig domains of the light chain (VL) and the heavy chain (VH). For biotechnological or pharmaceutical uses, the classical Ig recognition module can be formatted as a four www.sciencedirect.com
immunoglobulin domain Fab fragment (Figure 1a) or as a Fv fragment (VL-VH; Figure 1a), sometimes joined by a linker (scFv). These modules require the coexpression of two chains and their independent folding and assembly via hydrophobic interfaces. This difficult process is better suited to expression in eukaryotic cells than in Escherichia coli. In contrast, the recognition module of camelid Ig involves a unique Ig domain (called VHH or nanobody) in which the hydrophobic VH-VL interface is replaced by a hydrophilic surface, thanks to the framework mutations Gly44Glu and Leu45Arg [2–4]. Indeed, this module has been shown in most instances to be expressed in good yields in E. coli. Initial questions were raised, however, concerning the antigen affinity of nanobodies compared to Fabs/Fvs, as they harbor only three complementarity-determining regions (CDRs), the hypervariable loops that bind to antigens, compared to the six CDRs of Fabs/Fvs (Figure 1a, b). Sub-nanomolar/picomolar affinities for nanobodies against a wide range of antigens have been reported, however, indicating that nanobodies could be as efficient binders as Fabs/Fvs. Furthermore, nanobodies revealed unexpected and unique antigen binding characteristics. Studies of the first structure of a nanobody with an antigen, egg white lysozyme, revealed that its CDR3 inserts deeply into the enzyme active site [2]. The nanobody CDR3s are significantly longer than in classical VHs. Nanobody CDR3s also contribute to covering the nanobody surface corresponding to the VL/VH interface, and they often harbor a disulfide bond joining them to the rest of the domain [2,4]. The ‘protruding’ behavior of nanobodies, although not general, has been shown in many instances and is one of their most interesting features: nanobodies most likely bind to concave surfaces, while Fabs/Fvs bind to planar or convex surfaces. Another peculiar behavior of nanobodies is their ability to recognize the antigen (protein or even hapten) laterally, i.e., using the external surface of the CDR3 together with some patches of the framework [5,6]. These unique features have drawn considerable interest from the biotechnology and health/ pharmaceutical fields, as nanobodies are potentially potent enzyme or receptor inhibitors [7]. Finally, the use of nanobodies in structural biology is now well established and is the subject of this review.
Nanobody generation and selection The first step of nanobody generation is generally performed through immunization of a llama (Lama glama) or a dromedary (Camelus dromedarius). Llamas are generally preferred to dromedaries for ease of access, as several companies and public institutions provide immunization Current Opinion in Structural Biology 2015, 32:1–8
2 New constructs and expressions of proteins
Figure 1
(a)
(b)
VH
6 CDR
vHH 3 CDR or nanobody
CH1 VL
Fv
Fab
linker
CL
Fc
(c)
nanobody 1
Fc
diabody Ct GSn repeat
n
C
nanobody 2 : - same antigen, =/ epitope - =/ antigens - HSA (low clearance)
Nt Current Opinion in Structural Biology
Classical and camelid nanobodies. (a) Composite representation of a classical IgG, with heavy (blue) and light (magenta) chains. Two recognition modules, a Fab fragment (VH, VL, CH1, CL) and a Fv fragment (VH, VL), are identified. Six Complementary Determining Regions (CDRs) form the antigen-binding surface. (b) Composite representation of a camelid IgG, with a heavy chain only (blue). The CH1 domain is replaced by a linker. The recognition module is a unique immunoglobulin domain called vHH or nanobody, with three CDRs. (c) Composite representation of a dual nanobody construct (diabody). This construct makes it possible to target two different epitopes on the same antigen, thus tremendously increasing avidity, or to target two different antigens. Diabodies including human serum albumin (HSA) display a longer clearance.
facilities. Nanobody selection follows a well-established protocol [8,9]. Immunization is performed through 5 or 6 injections, one per week, of 0.5 mg protein (Figure 2). Notably, several antigens can be injected simultaneously. After the last immunization, samples of blood are collected, and the lymphocytes are purified. The total RNA from the peripheral blood lymphocytes is extracted and used as a template for first strand cDNA synthesis with an oligo (dT) primer. Using this cDNA, the VHH encoding sequences are amplified by PCR and cloned into the phage-display phagemid vector pHEN4, leading to a nanobody library of approximately 108 independent transformants. Three consecutive rounds of phage display and panning are performed on solid-phase coated protein. The enrichment for antigen-specific phages within the pools at each round of panning is assessed by polyclonal phage ELISA. A clear enrichment is generally seen after the second and third rounds of panning. Tens of individual colonies are randomly selected and analyzed by ELISA for the presence of antigen-specific nanobodies in their periplasmic extracts. The nanobody sequences from the ELISA-positive colonies are subjected to RFLP analysis and nucleotide sequencing. The whole nanobody selection process takes approximately 70 days (2-3 months) from the Current Opinion in Structural Biology 2015, 32:1–8
first injection (Figure 2). Alternatively, immunization can be avoided when using a naive library [10–12]. Naive libraries are generated by collecting the blood of several non-immunized animals. The library is then obtained by the same procedure as described above (Figure 2). The library should contain a larger number of independent transformants (5 108 to 109). Panning selection then isolates pre-existing binders. However, amplification is slower with naive libraries, and 4 rounds of panning might be necessary to obtain a good binder; however, the overall selection time is reduced to 3 weeks. Naive libraries for use in structural biology applications should develop in the future because the need for extremely good binders (
