Structural Diversity Problems and the Solving Method for Antibody Light Chains

The structural diversity (heterogeneity) problem of antibodies has become a big subject along with the development of antibody drugs and catalytic antibodies. The detailed studies on the subject have not been conducted because many difficult and complex problems are existed in the phenomena. The heterogeneity problem is observed in a whole antibody as well as a catalytic antibody. The difficulty and complexity of the heterogeneity are in the generation of many isoforms caused by different charges, different molecular sizes, and/or modifications of amino acid residues. We found that the constant region domain of the antibody light chain also plays an important role in the heterogeneity. It is desirable that the antibody and/or the subunits must have a defined structure for practical use. We found interesting phenomena that copper ion can convert the multi-molecular forms of antibodies to mono-molecular forms. The ion contributed greatly to the enrichment of the dimer-form and the homogenation of the differently charged full-length and constant region domain of the light chain. The role of copper ion must be significant for preparing a single, defined, not multiple, isoform structure. Note that the big problem could be solved by using copper ion during the purification process.


Introduction
In recent year, many monoclonal antibody drugs have been developed, and some of them are practically used in therapy [1][2][3]. With respect to catalytic antibodies, they have extensively been developed [4][5][6][7][8][9][10][11][12][13][14] for the last two decades from the viewpoint of both basic research and the application, where it has been proved that there are many catalytic antibodies being effective against anti-rabies virus [15], anti-influenza virus [16], anti-Helicobacter pylori [17], anti-HIV [7,8,10], anti-Alzheimer's disease [14,18], etc. Interestingly, some of them have been advanced to the stage tested in vivo in this decade [15][16][17][18][19]. In the case of catalytic antibodies, some of them play the role as a whole structure of IgG [5,9,11,13], IgA [20], or IgM [21][22][23]. On the other hand, in some cases, their subunits (light chain or heavy chain) exhibit unique functions [1-4, 6-8, 12]. Once the antibody subunits are separated, the structure of the light or heavy chain becomes flexible and has a tendency to possess structural diversity (or molecular heterogeneity). Regarding structural heterogeneity, it was found about 20 years ago that a whole antibody possesses the structural heterogeneity. These studies were extensively studied by Harris et al. [24] and Nebija et al. [25] using the capillary isoelectric focusing and the 2D-gel electrophoresis [26,27].
We have also reported about the molecular heterogeneity caused by different electrical charges and different molecular size in mouse monoclonal antibody [28]. This phenomenon is not good for the preparation efficacy, high reproducibility, and practical application. In addition, the structural diversity leads us to ask what structure plays the most important role in exhibiting the catalytic antibody functions. It also provides us with another subject of how we can best make a significant structure with high reproducibility and productivity.
We have recently found a crucial method to solve the heterogeneity problem by using copper ion, which can convert the multi-molecular forms into mono-molecular forms for many recombinant human antibody light chains. In addition, the constant region domain of the light chain (CL) plays an important role in generating a mono-molecular form.
In this review, we will describe a novel method for preparing a single and defined mono-form structure in detail.

Structural heterogeneity of monoclonal antibodies 2.1. Examples found in natural monoclonal antibodies
We have found an interesting phenomenon in 2D-gel electrophoresis for mouse type monoclonal antibodies (mAbs), which were prepared against the hemagglutinin molecule of influenza virus [29]. In the experiment, these monoclonal antibodies showed many different spots at the same molecular size [28]. Figure 1a shows the results using InfA-9 mAb. In the case, the whole antibody and the subunits, heavy and light chains separated and purified from the parent whole antibody, were analyzed by 2D-gel electrophoresis. In Figure 1a, many spots in the whole antibody of InfA-9 are shown. The clear four spots (pI = 6.0, 6.1, 6.2 and 6.5) in the heavy chain and three spots (pI = 5.9, 6.1 and 6.5) in the light chain were detected. Except these spots, many faint spots were observed in the same molecular size (the spots observed over the heavy chain were unknown). Then, the heavy and light chains were separated from the whole antibody of InfA-9, highly purified, and submitted to 2D-gel electrophoresis. The results exhibited the similar phenomena. In the heavy chain, similar spots are seen, and clear five spots (pI = 6.1, 6.3, 6.5, 6.7 and 6.9) were detected in this case. The pI positions of the spots were a little bit different compared to the whole antibody. For the case of the light chain, clear six spots (pI = 5.7, 5.9, 6.0, 6.1, 6.2 and 6.7) were detected. The number of spots increased compared with those of the whole antibody. These phenomena are not exceptional, but general. The mAb, InfA-15, exhibited similar results. Figure 1b shows 2D-gel electrophoresis using InfA-15 mAb. In the results of the whole antibody and the heavy chain and the light chain, the pattern of spots were a little different from those of InfA-9 mAb, but several different pI spots were observed in all cases. Namely, different pI spots are present at the same molecular size in any mAb. Note that structural diversity (molecular heterogeneity) should be existing even in the monoclonal antibody and the subunits, while they are a single protein. In our case, it is considered that the various electrical charges of the molecule may be one of the causes. SDS-PAGE; Running gel 12.5%. Strip; pH 3-10 nonlinear 7 cm. Sample; whole antibody 3.2 μg, heavy and light chain: 1.6 μg. Staining; deep purple (GE Healthcare). (a) InfA-9 monoclonal antibody. Whole antibody: many spots are seen in the whole antibody of InfA-9. The clear four spots (pI = 6.0, 6.1, 6.2 and 6.5) in the heavy chain and three spots (pI = 5.9, 6.1 and 6.5) in the light chain were detected. Heavy chain (H): clear five spots (pI = 6.1, 6.3, 6.5, 6.7 and 6.9) were detected. The pI positions of the spots were a little bit different compared to the whole antibody. Light chain (L): clear six spots (pI = 5.7, 5.9, 6.0, 6.1, 6.2 and 6.7) were detected. The number of spots increased compared with those of the whole antibody. (b) InfA-15 monoclonal antibody. For the whole antibody and the heavy and light chain, similar results showing many spots at different pI for the same molecular sizes were observed, suggesting that the molecular heterogeneity of antibodies are generally occurring events.

Recombinant monoclonal antibody Herceptin
The structural diversity of antibodies has been found about 20 years ago [26,27], suggesting that an antibody has some different structures (not a mono-form structure) caused by the various electrical charges. Regarding the charge heterogeneity, Harris et al. [24] and Nebija et al. [25,30] have extensively studied this phenomenon with recombinant antibodies using capillary isoelectric focusing and 2D-gel electrophoresis.
The former paper [24] pointed out a deamidation of Asn residues in the protein.
The latter papers showed multiple spots of pI spreading at heavy and light chains, caused by generation of charge-related isoforms. The pI spreading pattern in 2D-gel electrophoresis [30] is similar to our cases. The effects of sugar chains are also taken into account for molecular heterogeneity.
As it is well known that an antibody light chain has no sugar chain, this can be excluded for this subunit. Thus, in our case, it is thought that the structural diversity is mostly due to the heterogeneity of the different electrical charges. In addition, the possibility of deamidation is excluded because it is hardly considered that the addition of a copper ion causes a reverse reaction of the deamidation and the heterogeneity is lost.

Chromatograms of antibody light chain C51
In this section, we will describe the phenomena of structural diversity using full-length light chains of human antibodies. Figure 2a shows the amino acid sequences of human antibody kappa light chains of C51, #4 and #7. The proteins were expressed in Escherichia coli in accordance with the protocol described in the section of Materials and Methods in Refs. [19,31,32]. Methionine was adducted at the N-terminus and confirmed by amino acid sequence analysis after cloning the cDNA of the light chain into the Nco I site of the pET-20b vector. Leu and Glu residues were inserted by employing the Xho I site before a histidine-tag (His × 6) included in the vector for purification.
After the transformed E. coli cells were recovered by centrifugation, they were sonicated. Then, the soluble fraction was subjected to purification using Ni-NTA affinity chromatography. The result of a Ni-NTA affinity chromatography for the C51 light chain is shown in Figure 2b. The C51 light chain was eluted from fraction 28 (Fr.28) to Fr.44. Fr.35 showed the maximum absorbance. Fr.35 was collected and analyzed by SDS-PAGE with CBB staining, where the C51 light chain was mainly the monomer form with a slight contamination of dimer forms.
The eluted fractions from Fr.30 to Fr.40 were collected and subjected to a cation exchange chromatography. Figure 2c shows the chromatogram, where several peaks were observed. The SDS-PAGEs of the peaks were shown in the figure on the right side. The peaks 1, 2, and 3 were the monomer, the mixture of monomer and dimer, and the dimer, respectively.

Effect of copper ions
Twenty μM of CuCl 2 (Cu 2+ ) was added to either the cell suspension after recovery of the cells (cellsuspension) or the eluent of Ni-NTA chromatography (Ni-NTA eluent). By the addition of Cu 2+ to the cell suspension, the chromatogram changed to that shown in Figure 2d, where a main peak (peak 4) was observed at the retention time of 26 min but other peaks (1, 2 and 3) were small. The SDS-PAGE gave a dimer for peak 4. In Figure 2e, 20 μM of Cu 2+ (0.5 equivalent to the light chain) (c) Cation exchange chromatography as the secondstep purification without copper ion. Several peaks were observed from 15 to 27 min. The peaks 1, 2, and 3 were the monomer, the mixture of monomer and dimer, and the dimer, respectively. (d) Cation exchange chromatography as the second-step purification with addition of 0.5 eq. copper ion in cell suspension. A main peak (peak 4) was observed at the retention time of 26 min but other peaks (1-3) were small. The SDS-PAGE gave a dimer for peak 4. (e) Cation exchange chromatography as the second-step purification with addition of 0.5 eq. copper ion in the Ni-NTA eluent. Only a main peak (peak 3) was observed at the retention time of 26 min but other peaks (1 and 2) were scarcely detected. The SDS-PAGE gave a dimer for peak 3. The addition of Cu 2+ led to formation of dimers.
was added to the Ni-NTA eluent. In this case, only a main peak (peak 3) was observed at the retention time of 26 min but other peaks (1 and 2) were scarcely detected. The SDS-PAGE gave a dimer for peak 3. The addition of Cu 2+ led to dimer formation.
Mass spectroscopic (MS) analysis was performed for the main peaks observed above (data not shown). Briefly, a monomeric light chain was detected at 25,000 m/z and a dimer at 49,000 m/z. A small trimer and tetramer were also detected at 74,000 and 98,000 m/z, respectively. By the addition of Cu 2+ to the cell suspension or the Ni-NTA eluent, the signal for the monomer was substantially reduced. Conclusively, the addition of Cu 2+ was effective for the formation of the dimer.
In order to examine the pI of the light chain, 2D-gel electrophoresis was performed with and without the addition of Cu 2+ . The results are shown in Figure 3a and b. In the case without Cu 2+ , many spots at different pIs were observed with the same molecular size of 31 kDa (Figure 3a). The pI spots were widely located from 6.12 to 10.0. The strong spot was observed at pI = 6.45-6.73. In contrast, in the case with the addition of Cu 2+ , the spots were gathered on the strongest spot at pI = 6.57, while two faint spots were detected at around pI = 6.32 and 6.90 (Figure 3b). It is evident that the electrical charges of the molecule became mono-form by the effect of Cu 2+ .

#4 and #7 light chains
It must be invested whether or not the changes from multi-molecular forms to mono-molecular forms by the addition of copper ions is a general phenomenon. The following experiments were carried out.

Chromatograms
For the purpose of confirmation of the observed phenomena on the structural diversity, other antibody light chains such as #4 and #7 were examined. As stated above, the chromatogram in Ni-NTA purification is similar for many light chains. Thus, the following is focused on the results of cation exchange chromatography, which were very different in each light chain used. The effect of copper ions on the diversity issue will be discussed.
The cation chromatograms for #4 light chain are shown in the cases without and with Cu 2+ as presented in Figure 4a and b, respectively. In the case without Cu 2+ , there were several peaks. The results of SDS-PAGE (non-reduced condition) corresponding to three peaks are also shown. The observed peaks were a mixture of monomers and dimers. Namely, several structurally different light chains caused by different electrical charges are coexisting in the solution. However, when 20 μM of Cu 2+ (0.5 equivalent to the light chain) was added to the Ni-NTA eluent, the several peaks in Figure 4a surprisingly became a single peak (Figure 4b), which was mainly the dimer.
In the case of the chromatography for #7 light chain, huge three peaks were observed in the case without Cu 2+ as shown in Figure 5a. The peak of the retention time at 9 min was the monomer. The peak at 13 min was a mixture of monomer and dimer. The peak at 22 min was also a mixture. Note that light chains possess different molecular sizes and electrical charges in solution.
When 15 μM of Cu 2+ (0.38 eq.) was added to the Ni-NTA eluent, two peaks at 12 min and 23 min of retention time were observed in the chromatography. The results of SDS-PAGE analysis (Figure 5b) indicated that both peak A and peak B were dimers. It is interesting that two kinds of dimers with different electrical charges were coexisting. When 40 μM of Cu 2+ (1.0 eq.) was added to the same eluent, only peak C, which corresponds to peak A in Figure 5b, was observed at 12 min and peak B was not detected (Figure 5c). It is thought that peak B moved to peak A in Figure 5b. Conclusively, the dimeric light chains possessing two kinds of electrical charges became one kind of state possessing a unique electrical charge by the addition of 40 μM of Cu 2+ .   Huge three peaks were observed. The peak of the retention time at 9 min was the monomer, peak at 13 min was the mixture of monomers and dimers, and peak at 22 min was a mixture. In this case, the light chains possess different molecular sizes and electrical charges in solution. (b) With copper ion of 0.38 eq. Two peaks at 12 min and 23 min retention time were observed and both peak A and peak B were dimers. (c) With copper ion of 1.0 eq. Only peak C, which corresponds to peak A in (b), was observed at 12 min as the dimer form.
UV/VIS spectroscopy for these peaks was also conducted. The results are shown in Figure 6.
We could see the absorbance around 560 nm, which is assigned to the absorbance of the interaction of cupper with the amino acids for peak A and peak C, but not for peak B, whose spectrum was very similar without copper ion.

AFM analysis
In order to investigate the morphology of antibody light chains, we conducted atomic force microscopy (AFM) analysis. The peaks of A, B, and C were collected and subjected to AFM analysis as shown in Figure 7a-c. Figure 7a shows the AFM image for peak A. The images for peak B and C are shown in Figure 7b and c, respectively. The red circle represents the clear image of the dimeric light chain. The size of the dimer was roughly estimated at an Figure 6. UV/VIS spectroscopy. The absorbance around 560 nm, which is assigned to the absorbance of the interaction of a copper ion with the amino acids, was observed for peak A and peak C, but not for peak B. It is obvious that peak B has no copper ion.  approximate length of 20 nm, the width of 10 nm, and the height of 4 nm. The lateral and height length are comparable with the AFM image of IgG by Querghi et al. [33]. We could not identify the position of the copper ion residing in the light chain from this AFM analysis.

Structural heterogeneity of the constant region domains of light chains (CLs)
In the previous section, we focused on a full-length light chain, which is consisted of the variable and the constant domain. It is noteworthy to study which domain, the former or latter, causes the structural diversity problem. Although there are many studies on the role of the constant domains (especially for a Fc region) of the heavy chain of the antibody, the reports on the role of the constant region domain of the light chain are scarcely seen. From this point of view, we investigated the role of the constant domain as described in the following.

Chromatography and SDS-PAGE analysis
The expression and purification of the kappa type constant domain were similarly conducted as made in the full-length light chain. Ni-NTA chromatography was also performed to purify the recovered constant region domain. The result was also similar with that obtained in the case of full-length light chain except for the molecular size. The SDS-PAGE analysis for the collected fraction in the Ni-NTA chromatography is shown in Figure 9. Under non-reduced condition, a strong band was detected in the monomeric form at 15 kDa as well as a weak band in the dimeric form at 30 kDa. Under reduced condition, only the monomeric form was observed and the purity was over 95%. This sample was applied to cation exchange chromatography. The results are shown in Figure 10 along with the SDS-PAGE analysis under non-reduced condition. Several peaks were observed at retention times from 5 to 25 min while it was a single material of the constant domain. In Figure 10, peak 1 appearing at the retention time of 7.5 min is a monomer, and peaks 2, 3, and 4 appearing at 14-17 min contain mainly monomers. The dimers or/and trimer were detected for peaks 2 and 4. Peaks 5 and 6 appearing at 21-23 min are the dimer. These results mean that differently charged molecules of the constant region domain as well as differently sized molecules coexisted in solution at the same time. It is obvious that a constant region molecule shows molecular heterogeneity (structural diversity) from the viewpoint of both electrical charge and molecular size, which are very similar with those observed in the full-length light chain.

Effect of copper ions
As stated previously, copper ion (Cu 2+ ) hugely influenced the structural diversity of the fulllength light chain. The same experiment was performed with the constant region domain molecule. The results are summarized in Figure 11a-f. In the case of 0.1 eq. addition of Cu 2+ for the constant region domain molecule, we observed a small peak 7 and one main peak  (Figure 11a). Peak 8 was the dimer by the SDS-PAGE analysis under non-reduced condition (peak 7 was not analyzed because of the small peak). In the case of 0.2 eq. addition of copper, mainly two peaks (9 and 10) were obtained (Figure 11b). The elution times of peaks 9 and 10 were identical with those of peaks 7 and 8, respectively. Peak 9 included mainly the dimer with a very slight amount of the monomer. Peak 10 was the dimer. For these two peaks, UV/VIS spectroscopy was performed. The results are presented in Figure 12. Peak 9 showed an absorbance of around 580 nm, which was based on the interaction of Cu 2+ and amino acids of the constant region domain molecule. On the other hand, no absorbance was detected for peak 10. Namely, protein of peak 9 bound to Cu 2+ but peak 10 did not. Though the peaks are dimeric forms of the constant region domain, they were separated by the cation exchange column chromatography whether or not the peak contains Cu 2+ . For the case of 0.3 eq. addition, the main peak was peak 11 observed at the retention time of 16 min, which included the dimer along with a slight monomer and trimer forms (Figure 11c). In the case of 0.4 eq. addition, a clear single peak of the dimer form was detected at the retention time of 16 min (Figure 11d). In Figure 11e and f, peaks 13 and 14 were observed as single peak at the retention time of 16 min. And they were the dimer. It seems that enough content of added copper ion was 0.4 eq. to induce mono-form formation from the multi-forms of the constant region domain molecule.
The amount of Cu 2+ bound to the constant region domain molecule was also quantified using a commercially available copper detection kit (Copper, Low Concentration, Assay Kit, AKJ, CU21M, Metallogenics Co., Ltd., Chiba, Japan). For the peak appearing at the retention time of 16 min and containing Cu 2+ , the ratio of Cu 2+ : constant region domain was around 0.55. This result agreed with those of the UV/VIS spectroscopy, suggesting that two constant region domain molecules bind one copper ion.

Binding affinity of copper ions
The UV/VIS spectrum changed as different concentrations of Cu 2+ was added to the Ni-NTA elution after the samples were dialyzed against PBS. The results are presented in Figure 13.
The absorbance of 580 nm became larger along with an increase in the amount of added Cu 2+ , as showing a slight red shift. In Figure 14, the values for the concentration of added Cu 2+ were plotted vs. the maximum absorbance, which is the absorption isotherm curve for Cu 2+ binding to constant region domain molecules. The Langmuir plot is shown in the inset of Figure 14, indicating a good linear relationship. The binding constant was estimated to be 48.0 μM −1 . The binding affinity from several proteins incorporating divalent metal ions was investigated. The values (K) are presented in Table 1. Hemocyanin and metallothionein have very strong affinity to bind Cu 2+ . Carbonic anhydrase-binding Zn 2+ shows a strong affinity. Aminopeptidase III binding Co 2+ possesses a weak affinity. In the case of CL, the value (48.0 μM −1 ) seems to be intermediate among those metalloproteins.
In order to further investigate the molecular heterogeneity of the constant region domain molecule, two-dimensional (2D) electrophoresis was performed using samples with or without Cu 2+ . Figure 15a and b shows the results for the cases without and with the addition of Cu 2+  under non-reduced conditions, respectively. In the former case, many spots were observed: two spots at pI = 6.9 for the dimer and three spots at pI = 6.2, 6.5, and 6.9 for the monomer. In contrast, one strong spot was observed at pI = 6.9 for the dimer in the case with the addition of Cu 2+ , while very faint spots were detected in the positions of the monomer. Note that Cu 2+ can facilitate changes from the multimeric form to the monomeric form as well as from different electrical charges to a single electrical charge.

Other metal ions
About 1.0 eq. of a metal ion such as Ca 2+ , Mg 2+ , Ni 2+ , and Zn 2+ was added in each Ni-NTA elution and incubated overnight. Figure 16 shows the results of the cation exchange chromatography and SDS-PAGE (non-reduced) for peaks P3 and P5. For the cases of Ca 2+ , Mg 2+ , and Ni 2+ (Figure 16a-c, respectively), a large peak P3 was observed at 17 min along with a small peak P5 at 23 min. Interestingly, a large peak P3 was observed and peak P5 became very small in the case of Zn 2+ (Figure 16d). The chromatogram resembled the case of Cu 2+ (Figure 16e). From the results of the SDS-PAGE, the peak P3 was mostly in a monomeric form for all the cases of Ca 2+ , Mg 2+ , Ni 2+ , and Zn 2+ . On the other hand, P3 of Cu 2+ was the dimer. The molecular form (size) of P3 in the case of Ca 2+ , Mg 2+ , Ni 2+ , and Zn 2+ was quite different from that of  Cu 2+ . It must be considered that the Zn 2+ could not accelerate the dimerization of the constant region domain molecule, while the ion decreased the peak P5 and showed a large peak P3. Zn 2+ could have some ability to unify the structural diversity, but the effect is different from that of copper.

Hemocyanin (Cu
Zn 2+ did not accelerate the dimerization of the constant region domain molecule but has some functions that may contribute to solve the heterogeneity problem. Out of the several metals analyzed, Zn 2+ exhibited an interesting behavior, which must be a characteristic feature : two spots at pI = 6.9 for the dimer and three spots at pI = 6.2, 6.5, and 6.9 for the monomer were found. (b) With Cu 2+ : one strong spot was observed at pI = 6.9 for the dimer. It is revealed that copper ion accelerates both dimerization and generation of a mono-molecular form of the light chain.
of Zn 2+ . Although there are several reports on the relationship between metal ions and the enzymatic activity of catalytic antibodies, details of the contributions of metal ions to the molecular structure of catalytic antibodies are unclear at present [20,34]. Paul et al. reported an interesting function regarding Zn 2+ , which was essential for exhibiting the catalytic function of the antibody light chain to cleave beta-amyloid peptides, while the ion will not affect the catalytic site [14].

Consideration about unstable forms and a stable form of CL
For the reason why the addition of copper hugely effects the formation of a mono-form structure of the constant light chain domain, we postulated one of the situations from the viewpoint of potential energy and the wall height as illustrated in Figure 17. It is likely that the energy  (Figure 16a-c, respectively), a large peak, P3, was observed at 17 min along with a small peak, P5, at 23 min. In the case of Zn 2+ (Figure 16d), a large peak P3 was observed and peak P5 became very small. The chromatogram seemed to be like the case of Cu 2+ .
Structural Diversity Problems and the Solving Method for Antibody Light Chains http://dx.doi.org/10.5772/intechopen.72516 potential of each molecular form is at a comparable level after the preparation of the molecule (CL) without Cu 2+ , as shown in Figure 17a. In this case, transfer of the potential well A to B (or C) is easy because the walls of the potential energies of the wells are low (Figure 17b). However, the energy potential is drastically changed when copper ions are added. The multimolecular forms of the constant region domain, which are sitting in each potential well, drop in one deep potential energy level, as shown in Figure 17c, resulting in the formation of a mono-molecular form from the multi-molecular forms. Once the molecule dropped into the deep potential well, the form would be no longer able to transfer to other forms. As a consequence, the monomolecular form of the constant region domain molecule became stable. This situation can be achieved by the presence of copper ion in a ratio of more than 0.5 eq. of Cu 2+ to the constant region domain molecule. Figure 17. Consideration about conversion of unstable forms to a stable form of CL. (a) State A (corresponding to peak 1 in Figure 10), State B (corresponding to peaks 2, 3, and 4 in Figure 10), and State C (corresponding to peaks 5 and 6 in Figure 10) may stay in a chemical equilibrium. (b) Assumed situation in potential energy for the case without Cu 2+ : each potential energy level for the case without Cu 2+ may be comparable in wells of A, B, and C. The walls among the potential energy wells are not high. (c) Assumed situation of potential energy for the case with Cu 2+ : when Cu 2+ is incorporated, a deep potential level can be generated, and all molecules showing a different heterogeneity may drop into the well and exist as a stable form.

Preparation of mutants and their uptake of copper ions
In order to clarify the copper-binding site, two mutants were prepared from the C51 light chain, because the light chain has no histidine residues in the variable region compared to the sequence of the constant region domain comprising 2 His residues (Figure 8). Both histidine and cysteine residues are considered as the most plausible candidates for the binding site. Therefore, the residues of His195, His204, and Cys220 present in the constant region domain of the C51 light chain were mutated to Ala. As the consequence, two mutants were made. One is Cys220Ala (C220A: mono-mutant) and another is His195Ala, His204Ala, and Cys220Ala (H195A/H204A/C220A: triple-mutant; Figure 18a). The locations of the mutated residues are shown in Figure 18b.  These mutants were similarly expressed and purified as stated in the previous experiments. Fifty μM of Cu 2+ (1.25 eq.) was added to both the cell suspension and the Ni-NTA eluent, where all light chains gave a single peak in the cation exchange chromatography. The copper uptake by the wild type and each mutant was chemically analyzed and the results are presented in Table 2

Possibility of a zinc finger in the constant region domain
It is not well known that there is a zinc finger-like motif in the constant region of the antibody light chain. Interestingly, Radulescu pointed out that the motif is a type of Cys-X 3 -His [35]. The sequence Cys-X 3 -His-X 15 -Cys-X 3 -His is a complete motif of a zinc finger. The aa sequence of the constant region domain used in this article is presented in Figure 8. The sequence from positions 190th-224th of the constant region domain is CEVTHQGLSSPVTKSFNRGECLEHH.
The sequence of LEHH was adducted as a His-tag was introduced. The underlined amino acids agree with those of the zinc finger motif, Cys-X 3 -His-X 15 -Cys-X 3 -His, in which His224 is a part of the His-tag. This motif can uptake a metal ion such as Zn 2+ , which is a divalent metal ion. As Cu 2+ is also a divalent metal ion, it can bound to the motif. Based on the results of the chemical analysis of copper ions in mutants, those histidine and cysteine residues must be responsible to uptake the copper ion. It is plausible that a copper ion is able to bind to the zinc finger motif instead of a zinc ion. As seen in the investigation of divalent metal ions on the structural diversity, zinc ions showed some effect. This maybe ascribed to the presence of a zinc finger motif in the constant domain in the light chain (note that the zinc finger motif is conserved in both human and mouse antibody light chain (kappa type)). In addition, the similar motif composed of the same amino acids exists in the CH1 domain of the heavy chain.
As well known, a zinc finger can function as a transcription factor influencing gene regulation and protein expression. Few studies on the relationship between antibody and zinc finger have been made so far. From the viewpoint that one protein can have dual or multi-functions in case, the chemical and biochemical functions of each domain of an antibody should be investigated in detail.

Assumed binding site of a copper ion
Taking those facts mentioned above and discussed, the binding site of copper is assumed as illustrated in Figure 19. Cu 2+ must be coordinated with histidine and cysteine residues of dimeric CL molecules. In this situation, the copper ion mediates the CL molecule placing the lowest potential energy level.   In the case of the #4 wild type, two kinds of forms were observed. One is the dimer circled with #1 red color (a). This seems to be a cis form, whose molecular conformation corresponds to that of (b). A variable region faces to another variable region. The dimer circled with #2 red color (c) seems to be a trans form, whose molecular conformation corresponds to that of (d). In the case of the #4 wild type, two kinds of form were observed. One is the dimer circled with #1 red color (Figure 20a). This seems to be a cis form, whose molecular conformation corresponds to that of Figure 20b. A variable region faces to another variable region. On the other hand, the dimer circled with #2 red color (Figure 20c) seems trans form, whose molecular conformation corresponds to that of Figure 20d. Figure 20e was another spot, where cis and trans forms were observed. In contrast, a very simple form was observed in the case of #4 mutant C220A, as shown in Figure 20f and g. In any spots, only monomeric forms were observed.
If copper ion is incorporated into the zinc finger motif residing in the constant region domain, the dimeric form observed in the #4 wild type is easily formed. On the other hand, the mutant C220A exists as the monomeric form, which indicates that the cysteine locating at position 220th is a crucial amino acid to bind to a copper ion.

X-ray diffraction analysis
We are trying to determine the detailed steric conformation of a catalytic light chain. At present, the structure of the main chain of the #4 mutant C220A was clarified as a preliminary experiment. Figure 21a shows a single crystal of the #4 mutant C220A formed in the experiment. By using the single crystal, X-ray diffraction analysis was performed. The result is presented in Figure 21b, where a 3.1 Å resolution was attained. Interestingly, there are eight molecules of the #4 mutant in one lattice mediated by hydrophobic interaction.  In the case of the AFM analysis, the #4 mutant C220A was a monomer. This is ascribed that the mica as the supporting material firmly interacts with the #4 mutant molecule. In solution, the #4 mutant may interact with each other with a strong Van der Waals force.

Perspectives and conclusions
As stated in the abstract and introduction, the issue of structural diversity (heterogeneity) of antibodies has become a big subject along with the development of antibody drugs and catalytic antibodies. This subject has not been solved for a long period, because many difficult and complex problems were existing. For this long-period unsolved problem, copper ion showed a drastic effect and gave one of the answers for solving the structural diversity issue. Note that the antibody and/or the subunits must have a defined structure for practical use.
In recent years, many possibilities of the development of antibody drugs and catalytic antibodies have been reported by research groups throughout the world. This article offers huge insights into the development of catalytic antibodies, maybe, as well as antibody drugs.
Because the preparation can be standardized, many scientists and engineers will easily be able to produce the defined structure and the same functional antibody under any circumstances and anywhere in the world.