Affinity transfer by CDR grafting on a nonimmunoglobulin scaffold

Authors

  • Magali Nicaise,

    1. Laboratoire de Modélisation et d'Ingénierie des Protéines, UMR8619, Université de Paris-Sud, Bât 430, F-91405 Orsay Cedex, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Marielle Valerio-Lepiniec,

    1. Laboratoire de Modélisation et d'Ingénierie des Protéines, UMR8619, Université de Paris-Sud, Bât 430, F-91405 Orsay Cedex, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Philippe Minard,

    1. Laboratoire de Modélisation et d'Ingénierie des Protéines, UMR8619, Université de Paris-Sud, Bât 430, F-91405 Orsay Cedex, France
    Search for more papers by this author
  • Michel Desmadril

    Corresponding author
    1. Laboratoire de Modélisation et d'Ingénierie des Protéines, UMR8619, Université de Paris-Sud, Bât 430, F-91405 Orsay Cedex, France
    • Laboratoire de Modélisation et d'Ingénierie des Protéines, UMR8619, Université de Paris-Sud, Bât 430, F-91405 Orsay Cedex, France; fax: 33-1-69-85-37-15.
    Search for more papers by this author

Abstract

Neocarzinostatin (NCS) is a small “all β” protein displaying the same overall fold as immunoglobulins. This protein possesses a well-defined hydrophobic core and two loops structurally equivalent to the CDR1 and CDR3 of immunoglobulins. NCS is the most studied member of the enediynechromoprotein family, and is clinically used as an antitumoral agent. NCS has promise as a drug delivery vehicle if new binding specificities could be conferred on its protein scaffold. Previous studies have shown that the binding specificity of the crevasse can be extended to compounds completely unrelated to the natural enediyne chromophore family. We show here that it is possible to introduce new interaction capacities to obtain a protein useful for drug targeting by modifying the immunoglobulin CDR-like loops. We transferred the CDR3 of the VHH chain of camel antilysozyme immunoglobulin to the equivalent site in the corresponding loop of neocarzinostatin. We then evaluated the stability of the resulting structure and its affinity for lysozyme. The engineered NCS-CDR3 presents a structure similar to that of the wild-type NCS, and is stable and efficiently produced. ELISA, ITC, and SPR measurements demonstrated that the new NCS-CDR3 specifically bound lysozyme.

The development of soluble proteins that recognize given target molecules (ranging from small chemical compounds to macromolecular structures) is of increasing importance in therapy and biotechnology. In the past decades work in this direction was centered mainly on the engineering of antibodies as they display specific and tight binding to a huge variety of molecular compounds. However, antibodies and their functional fragments present several disadvantages from a practical point of view. First, they are large molecules: Even the smallest antigen-binding fragment, Fv, consists of approximately 250 amino acids. Second, they are composed of two different polypeptide chains, necessitating complicated cloning steps for the pair of genes, sometimes resulting in unstable domain association in the case of Fv (Hoogenboom 1997). A single-chain engineered fragment of immunoglobulin called ScFv consists of the VH and VL subunits joined by a synthetic peptide (Bird et al. 1988). It is the most widely used fragment of immunoglobulin in biotechnology because it can be produced in various hosts and because it is relatively stable (Harris 1999; Hudson 1999; Kousparou et al. 2002). However, in many cases, ScFv can only be produced in small quantities.

To overcome the drawbacks of the immunoglobulins, other protein scaffolds have been developed to create new types of binding protein. The proteins used include protease inhibitors, DNA-binding protein, Escherichia coli cytochrome b562, helix-bundle proteins, disulfide-bridged peptides, lipocalins, and their derived anticalins (Nygren and Uhlen 1997; Skerra 2000).

Several approaches have been used to obtain new ligand-binding functions drawing on either rational design experiments in conjunction with site-directed mutagenesis (Riechman et al. 1988; Essen and Skerra 1994; Schiweck and Skerra 1995; Ellis et al. 1996; van den Beucken et al. 2001), or combinatorial molecular biology methods (Smith 1991; Wells and Lowman 1992; Hoess 1993). Here, with the same goal, we have explored a new scaffold, the neocarzinostatin, as a potential support for new interaction capacities.

Neocarzinostatin belongs to the family of bacterial chromoproteins. The known members of this family are neocarzinostatin, secreted by Streptomyces neocarzinostaticus; macromomycin, secreted by Streptomyces macromomyceticus (Van Roey and Beerman 1989); C-1027 (Xu et al. 1994) and actinoxanthin (Pletnev and Kuzin 1982; Sakata et al. 1993), secreted by Streptomyces globisporus; and kedarcidin (Constantine et al. 1994), secreted by an unidentified species of actinomycetes. NCS contains a chromophore tightly bound in a cavity. The antibiotic activity of this protein is provided by this enediyne chromophore, which binds to DNA with high affinity and produces damage by radical reaction (Ishida et al. 1965; Kappen et al. 1980). The role of apo-NCS is to carry the chromophore and to protect it. The antitumoral activity of the chromoprotein has elicited considerable interest from chemical, biological, and medical perspectives, and is currently the focus of intense research activity (Sudhahar et al. 2000; Schauss et al. 2001; Urbaniak et al. 2002). The demonstrated ability of NCS to act as a transporter and its very low immunogenicity suggest that this protein could be useful as a general transporter in antitumoral therapy. A form of this product substituted with styrene maleic anhydride copolymer (SMA-NCS) is currently used in Japan to treat hepatic tumors (Maeda 2001).

Although the chromophore without its carrier protein displays some activity in vitro, all clinical trials and current applications of neocarzinostatin are based on the use of neocarzinostatin/neocarzinostatin chromophore complex as the active component. Wild-type (WT) NCS has been demonstrated to accommodate substituted naphthoate compounds related to the neocarzinostatin chromophore within its binding cleft (Urbaniak et al. 2002). Moreover, we have shown that the amino acids in the natural chromophore binding site can be randomly substituted to extend binding specificity to compounds (such as testosterone) completely unrelated to the natural enediyne chromophore (Heyd et al. 2003). This protein shows the same overall folding pattern as immunoglobulins, the topology of the β-strands being identical (Adjadj et al. 1992b). The structural similarity between immunoglobulin and NCS (Adjadj et al. 1992a) suggests that a second region of this scaffold, the NCS loops equivalent to immunoglobulin CDR1 and CDR3, may also be modified to confer new specific interaction capacities for a target protein, thereby mimicking the antibody/antigen complex. This may be particularly relevant for the development of drug targeting. The anchorage points of these loops being located at almost identical relative positions on NCS and immunoglobulins, we checked whether the substitution of the NCS-loop by the CDR-Ig loop can confer on NCS the capacity to interact with the antigen of origin.

Typically, the antigen-binding site of antibodies from vertebrates is formed by combining the variable domains of a heavy chain (VH) and a light chain (VL). However, some antibodies from the Tylopoda (camels and llamas) are an exception, because there are formed only from heavy chains (Hamers-Casterman et al. 1993; Muyldermans et al. 1994). Consequently, in the antigen recognition domain, referred to as VHH, there are only three hypervariable regions, instead of the six antigen-binding loops present in the classical antibodies formed by VH and VL chains. Several VHH/antigen complexes have been crystallized (Desmyter et al. 1996; Spinelli et al. 1996). In the camel VHH/HEWL complex, contact is essentially made via a long CDR3 loop of 27 residues (Desmyter et al. 1996). The N-terminal part of this loop (10 residues) penetrates deeply into the active-site cleft of the lysozyme, providing approximately 70% of the antigen contact (Transue et al. 1998). We transferred this sequence to the equivalent site in the corresponding CDR3 loop of neocarzinostatin. We studied the stability of the resulting structure and its affinity for hen egg white lysozyme.

Results

Construction of the NCS-CDR3

NCS has a similar overall fold with the VHH camel immunoglobulin domains. This similarity is clearly demonstrated by examination of the topological diagram in Figure 1, which shows the global fold of apo-NCS and the variable VHH domain of camel antilysozyme Ig. The structure of NCS is such that the 99–107 loop located at the surface of the protein could be directly interchanged with the equivalent sequences corresponding to the CDR3 (residues 99 to 126) of the VHH domain of camel immunoglobulin. The anchorage points of these loops are located at almost identical relative positions on the two structures. Using an adaptation of the QuickChange site-directed mutagenesis procedure (see Materials and Methods) we replaced the 99–107 loop of NCS by the VHH-CDR3 sequence corresponding to residues 99–126 of the antilysozyme antibodies. Growth conditions based upon those previously developed (Heyd et al. 2000; Valerio-Lepiniec et al. 2002) were optimized so as to overproduce soluble protein. Overproduction of this protein in the BLR21 strain led to secretion of the recombinant protein into the culture medium in a soluble folded form, with a production efficiency of 30–35 mg per liter of culture.

Characterization of NCS-CDR3

SDS-PAGE showed one well-resolved band with no visible contaminants on an overloaded gel. Mass spectrometry indicated the presence of a single species with the expected molecular weight, indicating that the recombinant protein was homogeneous, with no proteolysis on the N- or C-terminal end (data not shown).

The recombinant protein was further characterized to verify that replacement of the loop 99–107 by the antilysozyme VHH-CDR3 induced no significant change in the protein structure.

CD spectra for WT and mutant protein were recorded under identical conditions at 25°C (Fig. 2). The spectra have the characteristic of an all β-protein, with a maximum at 195 nm and a minimum at 210 nm. However, unlike immunoglobulins, NCS spectra present a positive contribution around 223 nm previously reported as representing “no typical” secondary structure (Heyd et al. 2000; Sudhahar et al. 2000; Valerio-Lepiniec et al. 2002). Although qualitatively identical to that for the WT protein, the shape of the NCS-CDR3 spectrum is not strictly identical: It displays a shift in the positive contribution around 223 nm, suggesting that the structure of the protein is locally modified.

The fluorescence emission spectrum of the NCS-CDR3 displays a maximum emission wavelength at 343 nm, as for the WT protein (data not shown), suggesting that the insertion does not significantly modify the accessibility of W39, the main contributor to fluorescence (Edo et al. 1991); therefore, the mutation induces no significant structural change in the distal loop 38–45 located at the bottom of the cleft.

Finally, the integrity of the cleft was checked by measuring EtBr binding. This compound binds apo-NCS stoichiometrically in the natural chromophore cleft (Mohanty et al. 1994), and is therefore a convenient tool for monitoring the functional properties of apo-NCS and its variants. The Kd value obtained for the NCS-CDR3 mutant (10 μM) is slightly higher to that for the WT protein (2 μM; Heyd et al. 2000).

Stability of NCS-CDR3 compared to WT-NCS

The stability of the NCS-CDR3 protein was evaluated by analyzing thermally and chemical denaturation-induced unfolding transitions by spectroscopy and calorimetry.

We used DSC (Fig. 3) to compare the thermal stability of the modified NCS with that of the WT-NCS. Unfolding of the WT protein led to a transition peak centered at 67.7°C. The transition peak was analyzed with a non-two-state single transition model, allowing independent determination of ΔHvH and ΔHcal. This analysis gave a ΔHvHHcal ratio of 0.97, with a calorimetric enthalpy (ΔHcal) of 121 ± 5 kcal•mole−1, suggesting a two-state model. The CDR3 insertion has a significant effect on the overall stability of the protein. A decrease in the melting temperature is observed, the shift being of about 10.7°C, accompanied by a decrease in denaturation enthalpy to 82 ± 5 kcal•mole−1, giving a ΔHvHHcal ratio of 1.1. Moreover, whereas the thermal denaturation of WT-NCS was fully reversible (Heyd et al. 2000), that of NCS-CDR3 was irreversible, a second scan showing no endothermic transition.

We also studied the denaturation of WT and NCS-CDR3 induced by GdmCl. Figure 4 shows the normalized variation of the maximum emission fluorescence wavelength of NCS and mutant proteins as a function of denaturant concentration. For the WT protein, the transition occurred between 2 and 3.5 M GdmCl, with a midpoint transition, Cm, at 3 M. The calculated Gibbs free energy of unfolding ΔG0 and the proportionality constant m are 8.8 ± 0.2 kcal mole−1 and 3.0 ± 0.2 kcal mole−1 M−1, respectively. The transition of NCS-CDR3 is shifted towards lower denaturant concentrations (1.7 M) than that of WT-NCS, resulting in a decrease of the ΔG0 value (3.3 ± 0.2 kcal M1) and m value (1.9 ± 0.1 kcal mole−1 M−1). For both proteins, the chemical denaturation was fully reversible (data not shown).

NCS-CDR3/HEWL interaction

ELISA

The lysozyme-binding properties of NCS-CDR3 were investigated by ELISA. A microtiter plate was coated with HEWL, incubated with NCS-CDR3, and bound protein was detected by incubation with peroxidase conjugated anti-His-tag antibodies (Fig. 5). NCS-CDR3 displayed significant binding, with a dose-dependent response. Saturation was not reached until a concentration of about 100 μM of NCS-CDR3. No signal was observed in the absence of immobilized HEWL or NCS-CDR3. Very little binding activity was detected with the control, in which WT-NCS was added instead of NCS-CDR3. Soluble HEWL (100 μM) competed with immobilized HEWL for binding to NCS-CDR3 (30 μM; data not shown). Fifty percent of the signal was lost on addition of a mixture of NCS-CDR3 (30 μM) and HEWL (100 μM).

SPR measurements

HEWL was immobilized via classic amine coupling on a biosensor chip. The surface thus prepared was then exposed to NCS-CDR3 or WT-NCS in a Biacore apparatus. Figure 6A presents a sensorgram comparing the binding of 25 μM of WT-NCS and NCS-CDR3. The on- and off-rates observed are very rapid, precluding any analysis of the kinetic parameters. Only qualitative data can be obtained. The binding activity of NCS-CDR3 is correlated with protein concentration (Fig. 6B) and the signal obtained for WT-NCS was consistently weaker than that for NCS-CDR3.

ITC studies

Figure 7 shows HEWL titration of the native purified NCS-CDR3. Raw data obtained from the calorimetric titration at 25°C (Fig. 7A) of NCS-CDR3 with lysozyme displayed a very small amplitude response, generating a low signal-to-noise ratio for the integrated heat per injection (Fig. 7B). As a result, the curve fitted to the integrated binding isotherm gives only a crude estimation of the dissociation affinity constant, 33 ± 1 μM. No binding activity was detected in control experiments in which WT-NCS/ HEWL interaction was tested. Consistently, experiments performed at 25°C led to aggregation process. To overcome this process, ITC experiments were also performed at 15°C. In these conditions, raw data displayed a large amplitude response (Fig. 7C) generating a good signal-to-noise ratio. The curve fitted to the integrated binding isotherm gives a dissociation affinity constant of 0.5 ± 0.3 μM (Fig. 7D). In these conditions, a very small binding activity was detected between WT-NCS and HEWL; this nonspecific binding was totally abolished in presence of 75 mM NaCl (data not shown).

Discussion

Effect of the mutation on the structure and stability of the protein

The insertion of the camel-immunoglobulin CDR3 into NCS induced only local changes in the overall structure of the protein as indicated by fluorescence and CD measurements; the integrity of cleft is preserved as indicated by the Kd value for EtBr, which is close to that of the WT-NCS. The only change observed in CD spectrum is a shift of the peak around 223 nm. This peak has been previously reported as representing “no typical” secondary structure (Heyd et al. 2000; Sudhahar et al. 2000; Valerio-Lepiniec et al. 2002). The “no typical” secondary structure could be related to the loops of the protein. In this case, the slight modification of the spectrum around 223nm only represents the contribution of the large loop inserted in the NCS-CDR3 variant.

This loop replacement reduces the stability of NCS as shown by microcalorimetry and Gdm-Cl denaturation monitored by fluorescence. DSC measurements indicated that the Tm of the NCS-CDR3 mutant is about 11°C lower than that of the WT protein; this shift in Tm is accompanied by a decrease of 39 kcal•mole−1 in ΔHcal. Thermodynamic analysis of the transition obtained with Gdm-Cl–induced denaturation also demonstrated that insertion of the camel VHH-CDR3 loop destabilizes NCS. However, although less stable than the WT protein, the engineered protein has a stability similar to that of an immunoglobulin domain (Ewert et al. 2002, 2003).

Binding of NCS-CDR3 to HEWL

The affinity of NCS-CDR3 for HEWL was analyzed by ELISA, ITC, and SPR approaches. A specific interaction was found to occur between NCS-CDR3 and HEWL. At 25°C, this interaction is of low affinity as we obtained no saturation signal in the ELISA and SPR experiments. With ITC, even with noisy data, we obtained an estimated dissociation constant of around 30 μM. Although this crude estimate, the value of the dissociation constant is consistent with the 50% competition observed for 100 μM-soluble HEWL in an ELISA test; the measured dissociation constant would predict a 66% competition.

Difficulties in obtaining a precise dissociation constant were due to the experimental conditions. The low affinity of HEWL binding made it necessary to work with high protein concentrations, and some of the experimental noise can be attributed to the initiation of protein aggregation following complex formation. To prevent this aggregation we worked at a lower temperature (15°C). In these conditions, we observed a clear signal corresponding to a dissociation constant of 0.5 μM. A nonspecific interaction between HEWL and WT-NCS was observed: It is directly related to the pHi values of HEWL (pHi ≈ 9.2) and WT-NCS (pHi ≈ 5) and was fully screened by working in 75mM NaCl. This strong temperature effect on the dissociation constant and the salt effect upon aggregation process suggest that the specific interaction between CDR3 and HEWL is mainly driven by electrostatic interaction.

The dissociation constant obtained at 15°C is higher than those obtained for the complex between the camel immunoglobulin (CabLys3) and HEWL (Kd = 20 nM; Desmyter et al. 1996). The observed difference in Kd values may be due to key structural interactions not present in the hybrid protein. In the original immunoglobulin, the CDR3 loop is stabilized by a disulfide bridge with CDR1. It has been proposed that this disulfide bridge imposes conformational constraints, optimizing the loop for interaction with HEWL (Spinelli et al. 1996; Transue et al. 1998). Such a structural constraint does not exist in NCS-CDR3: This may make this loop much more flexible and decrease the efficiency of interaction with HEWL.

The type of approach taken here has mainly been used with antibodies, in so-called humanization or CDR grafting. A set of six CDRs has been transplanted from a mouse monoclonal antibody onto a human protein framework, resulting in a hybrid Fv moiety retaining its original specificity (Riechman et al. 1988). The human VH domain has been camelized by randomization of CDR3, and clones displaying specific binding to lysozyme have been selected on the basis of Kds lying between 1 and 10 μM (Davies and Riechmann 1996), close to that observed in this study. Other attempts to graft one or more CDR region by a rational approach have been unsuccessful (Schiweck and Skerra 1995; Ellis et al. 1996; van den Beucken et al. 2001). Thus, even though this simple operation can be relatively easily performed, in most cases affinity is lost and the design of humanized antibodies requires correction of the framework (Caldas et al. 2003). The rather low affinity we have obtained is consistent with these results. The originality of our work lies in the fact that we have successfully grafted a CDR loop onto a protein scaffold that is not an immunoglobulin. Other protein scaffolds, including lipocalins and anticalins, may also provide attractive alternatives to recombinant antibody fragments, combining, like NCS, the advantage of much smaller size with that of a single polypeptide chain (Nygren and Uhlen 1997; Skerra 2000). However, NCS presents a major advantage in that two regions of this protein can be modified: the NCS loop equivalent to immunoglobulin CDR and the natural chromophore binding cleft. This could be particularly useful for the development of drug targeting as the NCS/chromophore complex is a powerful antitumoral agent used in tumor targeting. We have shown here that binding of macromolecular antigen should be possible with this scaffold. It is now important to use combinatorial molecular biology to build a library of CDR-loops for NCS (in combination with or independent of cleft mutations) to generate a large diversity for molecular recognition.

Materials and methods

Mutagenesis and protein purification

The cloning of neocarzinostatin has been described elsewhere. The camel VHH-CDR3 sequence (DSTIYASYYECGHGLSTGGYG YDSWGQG) involved in HEWL recognition was inserted in place of the residues 99–107 of NCS (DAAGNGPEG) using a method based on the Stratagene QuickChange site-directed mutagenesis method. We used the same enzymatic reaction as the QuickChange kit, but with different primers. Each primer contained 18 nucleotides (underlined sequence in the primers below), which hybridized upstream and downstream from the sequence of nucleotides corresponding to the 99–107 loop of NCS. Each primer also contained half the desired insertion corresponding to the antilysozyme VHH-CDR3 (bold sequence in the primer below). Primer 1 was 5′-ACCGTGACCGCATTCGTAGTAGGAAGCGTAGATGGT GGAATCACTTAAGCCCACCTGGCA-3′, and primer 2 was 5′-CTGTCCACCGGTGGT TACGGT TACGAT TCCT GGGGT CAGGGTGTGGCAATAAGCTTCAAC-3′. The complete NCS-CDR3 DNA sequence was determined. Mutant plasmids were used to transform E. coli BLR21 (NOVAGEN). Growth conditions were based upon those previously developed (Heyd et al. 2000; Valerio-Lepiniec et al. 2002). Cells freshly transformed with the expression vector were grown for 3 days on 2YT medium containing ampicillin, tetracycline, and chloramphenicol at 30°C, without induction. The culture medium was separated from the bacteria, and soluble proteins directly secreted into the culture medium were precipitated with 650 g of ammonium sulfate per liter. The proteins were collected by centrifugation for 20 min at 17,000g. The precipitate was dialyzed first against double-distilled water and then against 50 mM phosphate buffer, pH 8 containing 300 mM NaCl. The protein solution was then applied onto a Ni-NTA column (Qiagen). After washing with a phosphate buffer (pH 8.0), 50 mM containing 300 mM NaCl and 20 mM imidazol, the protein was eluted with the same buffer containing 250 mM imidazol instead of 20 mM imidazol.

Physicochemical properties

The molecular weight and purity of the variant protein were analyzed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry on a PerSeptive Voyager-DE STR using standard methods. Protein concentrations were determined by spectrometry using extinction coefficients of 13,800 and 24,750 M−1 cm−1 at 280 nm, for WT-NCS and NCS-CDR3, respectively, calculated with the ProtParam program on the expasy server (www.expasy.ch/tools/protparam.html) from their sequences according to Gill and von Hippel (1989; Pace et al. 1995).

Circular dichroism (CD) spectra were recorded from 185 to 250 nm on a Mark VI dichrograph (Jobin-Yvon) equipped with a thermostatically controlled cell holder and connected to a computer for data acquisition. Data were acquired from 13-μM sample solutions in 20 mM phosphate buffer (pH 7.4) in quartz cells with a path length of 1 mm.

Ethidium bromide binding to NCS-CDR3 was studied in 20 mM phosphate buffer (pH 7.4) by fluorimetry with a Cary Eclipse fluorimeter by monitoring the intrinsic fluorescence of a EtBr solution (5 μM final concentration, λexc = 479 nm, λem = 620 nm, bandwidth of 2 nm) at various NCS-CDR3 concentrations. Saturation curve data were analyzed by using the following equation:

equation image((1))

where λFmax = FmaxF0; ΔF = FF0, where F and F0 are the fluorescence intensity measured in the presence and absence of the protein, respectively; P0 is the total protein concentration; B0 the total ethidium bromide concentration; and Kd the dissociation constant. Experimental data were fitted according to equation 1 by using a simplex procedure based on the Nelder and Mead algorithm.

Stability of the mutant protein

The thermal stability of the mutant protein was studied by scanning calorimetry on a DSC apparatus (Microcal Corp). DSC measurements were made with a 1 mg•mL−1 aponeocarzinostatin solution dialyzed overnight against 20 mM phosphate buffer (pH 7.4). Buffer solution from the dialysis bath was used as a reference. All solutions were degassed just before loading into the calorimeter. Scanning was performed at 1 K•min−1. The percentage recovery of native protein after heat denaturation was evaluated by rescanning after cooling a denatured sample.

Thermodynamic parameters, calorimetric enthalpy (ΔHcal) and van't Hoff enthalpy (ΔHvh), were determined as previously described (Valerio-Lepiniec et al. 2002) The heat capacity of the solvent alone was subtracted from that of the protein sample. These corrected data were analyzed using a cubic spline as a baseline in the transition. Thermodynamic parameters ΔHcal and ΔHvh were determined by fitting the following equation to the data:

equation image((2))

where Kd is the equilibrium constant for a two-state process, ΔHcal is the measured enthalpy, corresponding to

equation image((3))

and ΔHvh is the enthalpy calculated on the basis of a two-state process. Fitting was performed using ORIGIN software (Microcal).

The unfolding induced by GdmCl was monitored at 25°C by fluorescence spectroscopy on 5 μM protein solutions in 20 mM phosphate buffer (pH 7.4). Fluorescence measurements were performed with a Cary Eclipse fluorimeter, after 12 h of incubation in GdmCl solutions of various concentrations. Ultrapure GdmCl was obtained from Pierce; denaturant concentrations were checked by refractometry, using the relationship provided by Nozaki (1972). Transition curves were constructed by plotting the position of maximum fluorescence emission (λexc = 290 nm, bandwidth = 2 nm) as a function of denaturant concentration.

The model of linear dependency of ΔGx upon denaturant concentration, x, according to Pace (1986) was used for thermodynamic analysis:

equation image((4))

Assuming that the linear dependence of the free energy change on denaturant concentration observed in the transition region can be extrapolated to zero denaturant concentration, ΔG0 represents the standard variation of free energy in the absence of denaturant and m a constant proportional to the increase in the accessible surface area of the protein to the solvent on denaturation. An equation derived from equation 4, taking into account the baselines and the transition region, was used to analyze the data:

equation image((5))

where yx is the experimental signal in the presence of x molar GdmCl, yn the signal of the native form, sn and sd are the solvent effects on the native and denatured protein signal, respectively, and A the amplitude of the transition. Experimental data were fitted according to equation 5 by using a simplex procedure based on the Nelder and Mead algorithm.

Detection of NCS-CDR3 binding activity in an ELISA

ELISA was carried out at 25°C in 96-wells microtitre plates. Each well was coated by incubation for 1 h with 10 μg•mL−1 of HEWL (Sigma) in Tris-buffered saline (TBS). The wells were washed three times with TBS/Tween, blocked by incubation in TBS/ Tween/BSA (1 mg•mL−1) for 1 h and washed three times with TBS/Tween. NCS-CDR3 was applied in a dilution series in TBS/ Tween and the plates were incubated for 1 hour. The wells were washed three times with TBS/Tween and then incubated with 100 μL of a 1/600 dilution of anti-His-tag antibodies-peroxidase (Roche) conjugated for 1 h. Plates were washed three times with TBS/Tween and once with TBS, and the signals were developed by adding the BM blue peroxidase substrate (Roche), which serves as a chromogenic substrate for peroxidase-mediated color development in enzyme immunoassays. The change in absorption at 390 nm was measured in a Victor2 1420 Multilabel Counter (Wallac).

Isothermal titration microcalorimetry

ITC experiments were performed with a VP-ITC isothermal titration calorimeter (Microcal). Two set of experiments were performed: one at 25°C, and the other at 15°C. For each, injections of 10 μL of HEWL (1.3 mM in phosphate buffer 20mM at pH 7.5) were added from a computer-controlled 300 μL microsyringe at intervals of 1 min into the NCS-CDR3 solution (180 μM, cell volume = 1.430 mL) dissolved in the same buffer as the lysozyme, while stirring at 310 rpm. Reference experiments were performed by injecting lysozyme into a cell containing buffer with no protein. Reference values were subtracted from the sample values (lysozyme injected into NCS-CDR3 solution). The control was performed by injecting lysozyme into a cell containing the same concentration of WT-NCS in the same buffer.

A theoretical titration curve was fitted to the experimental data using the ORIGIN software (Microcal). This software uses the relationship between the heat generated by each injection and ΔH° (enthalpy change in kcal mole−1), KA (the association binding constant in M−1), n (number of binding sites per monomer), total protein concentration, and free and total ligand concentrations.

Surface plasmon resonance

All experiments were carried out on a BIAcore 2000 instrument. The ligand (HEWL) was immobilized on a CM5 chip by standard amine coupling by first activating the carboxydextran layer with a mixture of 50 mM N-hydroxysuccinimide and 200 mM 1-ethyl-3–(3-dimethylaminopropyl)carbodiimide. HEWL (1 mg•mL−1) in 10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA was then coupled via its amino groups at a flow rate of 5 μL•min−1. Surfaces were then inactivated with 1 M ethanolamine. A nonprotein (blank) sensor surface was prepared and used as reference chips. The presented data are corrected from solute effects of the SPR signal on the blank surface. Measurements were carried out at 25°C in TBS (pH 7.4). NCS-CDR3 was dialyzed against TBS (pH 7.4) before injection. For the measurement of NCS-CDR3 binding to HEWL, soluble NCS-CDR3 was injected at a concentration of 20 to 100 μM at a flow rate of 25 μL•min−1. A control was carried out under the same conditions, with injection of WT-NCS.

Figure Figure 1..

(A) Topological diagrams of NCS and the camel VHH domain showing the Greek-key motif of their folds. The β-strand is delimited by the indicated residues numbers. The variable domain CDR1, −2, and −3 are indicated on the diagrams. (B) Comparison of the structure of the VHH domain of a camel antilysozyme immunoglobulin and apo-NCS. The two proteins are folded into an antiparallel β-sheet built up from seven β-strands (represented as large arrows) arranged into β-sheets (one four-stranded and one three-stranded) packed against each other (the images were generated with the Swiss PDB Viewer program). The CDR3 loops of camel VHH (loop 99–126) and the equivalent loop (99–107) of NCS are represented with a wire frame.

Figure Figure 2..

Comparison of the CD spectra of WT-NCS (black line) and NCS-CDR3 (gray line) recorded under the same experimental conditions: 13 μM sample solutions in 20 mM phosphate buffer (pH 7.4) in quartz cells with a path length of 1 mm.

Figure Figure 3..

Baseline-corrected DSC thermograms of WT-NCS (black line) and NCS-CDR3 (gray line) recorded under the same experimental conditions: 1 mg•mL−1 protein solution in 20 mM phosphate buffer (pH 7.4) with a scanning performed at 1 K•min−1.

Figure Figure 4..

Unfolding transition curve assessed by the normalized variation of the maximum fluorescence wavelength at 25°C as a function of GdmCl concentration. Comparison of the WT-NCS (circles) and NCS-CDR3 (triangles) protein recorded under the same experimental conditions: 5 μM protein solutions in 20 mM phosphate buffer (pH 7.4) incubated 12 h in various GdmCl solutions.

Figure Figure 5..

Binding specificity of NCS-CDR3 for immobilized HEWL tested by ELISA. Black columns: Binding activity as a function of protein concentration. (A) Control without immobilized HEWL and NCS-CDR3; (B) control with 100 μM NCS-CDR3 and without immobilized HEWL; (C) control with 100 μM of WT-NCS without immobilized HEWL; (D) control with 100 μM of WT-NCS in the presence of immobilized HEWL. The error bars correspond to the interval of confidence of the mean of three measurements.

Figure Figure 6..

SPR analysis of NCS-CDR3 or WT-NCS binding to immobilized lyzozyme on a CM5 chip. (A) Sensorgram of interaction between purified WT-NCS (25 μM) or NCS-CDR3 (25 μM) with HEWL. (B) Comparison between signals of specific (NCS-CDR3) and nonspecific (WT-NCS) interactions with HEWL as a function of NCS (WT or CDR3) concentration.

Figure Figure 7..

Microcalorimetric titration of NCS-CDR3 with HEWL. Data were obtained at 25°C (A,B) and 15°C (C,D) by an automated sequence of 28 injections of 1.3 mM HEWL from a 300-μL syringe into the reaction cell, which contained 180 μM NCS-CDR3. The volume of each injection was 10 μL, and injections were made at 1-min intervals. (A,C) Raw data from the titration. Each peak corresponds to one injection. (B,D) The peaks in the upper panels were integrated with ORIGIN software and the values are plotted vs. injection number. Each point corresponds to the heat in μcal generated by the reaction upon each injection. The solid line is the curve fit to the data by the Origin program. This fit yields values for Kd.

Acknowledgements

This work was supported by the Centre National de La Recherche Scientifique (CNRS) and by a grant from Institut Curie (“Programme Incitatif et Cooperatif”). We thank Dr. Charles H. Robert for carefully reading the manuscript and Veronique Besson for technical assistance.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.

Ancillary