Conversion of interleukin-13 into a high affinity agonist by a single amino acid substitution.

We created a novel mutated form of human interleukin-13 (IL-13) in which a positively charged arginine (R) at position 112 was substituted to a negatively charged aspartic acid (D). This mutant, termed IL-13R112D, was expressed in Escherichia coli and purified to near homogeneity. IL-13R112D was found to be a potent IL-13 agonist with 5-10-fold improved binding affinity to IL-13 receptors compared with wild-type IL-13 (wtIL-13). The conclusion of IL-13 agonist activity was drawn on the basis of approximately 10-fold improved activity over wtIL-13 in several assays: (a) inhibition of CD14 expression in primary monocytes; (b) proliferation of TF-1 and B9 cell lines; and (c) activation of STAT6 in Epstein-Barr virus-immortalized B cells, primary monocytes, and THP-1 monocytic cell line. Furthermore, mutant IL-13R112D neutralized the cytotoxic activity of a chimeric fusion protein composed of wtIL-13 and a Pseudomonas exotoxin A (IL-13-PE38) approximately 10 times better than wtIL-13. Based on these results, it was concluded that IL-13R112D interacts with much stronger affinity than wtIL-13 on all cell types tested and that Arg-112 plays an important role in the interaction with its receptors (IL-13R). Thus, these results suggest that IL-13R112D may be a useful ligand for the study of IL-13 interaction with its receptors or, alternatively, in designing specific targeted agents for IL-13R-positive malignancies.

IL-13 1 is a pleiotropic cytokine that plays a major role in immune response and inflammation (1)(2)(3). It can inhibit production of proinflammatory cytokines IL-1, IL-6, and tumor necrosis factor-␣ and down-regulate the expression of CD14 (a lipopolysaccharide receptor) on monocytes (4). It can also cause the generation of antigen-presenting dendritic cells in combination with granulocyte-macrophage colony-stimulating factor (5). IL-13 also plays a major role in B cells. It can up-regulate CD23, CD72, major histocompatibility complex class II and surface IgM on B cells, drive IgE class switch, and induce production of immunoglobulins by B cells (6).
We have previously reported that a variety of human solid tumor cells express elevated levels of IL-13Rs (7, 10 -12). To target these receptors, we produced a chimeric protein (IL-13-PE38QQR) composed of wtIL-13 and a mutated form of Pseudomonas exotoxin (PE38QQR). This cytotoxin is highly cytotoxic to IL-13R-expressing cells (10 -12). However, the binding affinity of IL-13-PE38QQR was 10 times lower than wtIL-13 (10). To improve the binding affinity of IL-13, we proposed to generate IL-13 muteins by site-directed mutagenesis in which single amino acid substitutions were introduced in the critical region of IL-13 molecule. The selection of single amino acid substitutions was based on structural similarities and reported mutations in a similar cytokine, IL-4. We created and purified human IL-13R112D (in which arginine 112 was changed to aspartic acid) and analyzed its various activities on a variety of cell types. We also tested the capability of IL-13 mutant to inhibit cytotoxicity mediated by IL-13 cytotoxin, IL-13-PE38. We conclude on the basis of these results that IL-13R112D is approximately 5-10 times superior to wtIL-13 in all respects.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and DNA ligase were obtained from New England Biolabs (Beverly, MA), Life Technologies, Inc., Panvera (Madison, WI), and Roche Molecular Biochemicals. Fast protein liquid chromatographic columns and media were purchased from Amersham Pharmacia Biotech. Sequence specific oligonucleotide primers were synthesized at Bioserve Biotechnologies (Laurel, MD). Advantage-HF polymerase chain reaction (PCR) kit was from CLONTECH (Palo Alto, CA).
The pET based expression vector with ampR gene was used for construction of mutein clone. Plasmids were amplified in Escherichia coli (DH5␣ high efficiency transformation) (Life Technologies, Inc.), and DNA was extracted using Qiagen kits (Chatsworth, CA). TF-1 human erythroleukemia cell line was obtained from ATCC (Manassas, VA) and were grown in human granuclocyte-macrophage colony-stimulating factor. B9 mouse plasmacytoma cell line was a kind gift of Giovana Fosato (Center for Biologies Evaluation and Research, Food and Drug Administration, Bethesda, MD) and were grown in human IL-6. PM-RCC renal cell carcinoma cell line was established in our laboratory (19).
THP-1 cells were kindly provided by Dr. Ray Donnelly (Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD). Monocytes were isolated from the peripheral blood mononuclear cells of donors seronegative for human immunodeficiency virus and hepatitis viruses after leukapheresis and purified by countercurrent centrifugal elutriation.
Homology Search and Secondary Structure Analysis of IL-13-A computer program, GCG (Genetics Computer Group, Inc., Madison, WI) was used for a homology search, data base search, and prediction of secondary structure of IL-13 on Silicon Graphics Workstation in Human Genome Center, the Institutes of Medical Science, University of Tokyo (Tokyo, Japan) and the Center for Information Technology, National Institutes of Health (Bethesda, MD). The protein sequence of mouse (IL-13_mouse.swissprot), rat (l26913.gbro), Homo sapiens (Caf0433341.gpmain), and Bos taurus (bta1324411.gpmain) were FIG. 1. Homology of IL-13 between species. A, homology of mature IL-13 between species is calculated using pileup of the GCG program. Cysteine residues are shown in yellow. Positively charged groups and negatively charged groups are shown in blue and red, respectively. The numbering shown here may differ from each species due to the inserted gap during the homology pileup. Tyrosine is generally not classified into the charged group but is classified into the aromatic group. However, since tyrosine as well as glutamic acid or aspartic acids make hydrogen ion bonds in acidic conditions, they are shown here as negatively charged groups. Four cysteine residues are completely conserved between the four species. Many charged groups are also conserved. B, amino acid residue Arg-112 of IL-13 was substituted with Asp. obtained from Swiss Prot and GenBank TM . Since l26913.gbro is not a protein sequence but a cDNA sequence, it was used for a homology search after translation into a protein sequence. The IL-13 sequences are shown after deletion of predicted signal sequence of Homo sapiens or the equivalent (see Fig. 1). Hydrophobicity and secondary structure of IL-13 was predicted by the Kyte-Doolittle method and Chou-Fasman method, respectively (24).
Construction of Plasmids Encoding IL-13R112D-The mutagenesis of IL-13 gene was performed with a cDNA as a template using sense primer 5Ј-taa ttt gcc cat atg tcc cca ggc cct gtg cct-3Ј and antisense primer 5Ј-taa ttt gcc cga att cag ttg aag tct ccc tcg cg-3Ј to mutate Arg-112 to Asp (R112D) and incorporate NdeI and EcoRI restriction enzyme sites at 5Ј-and 3Ј-termini, respectively. After subcloning the polymerase chain reaction products into pCR2.1 (Invitrogen®, Carlsbad, CA), the plasmid was digested with NdeI and EcoRI. The fragment was inserted into a prokaryotic pET-based expression vector digested with same restriction enzymes. We confirmed the existence of mutation and restriction sites by sequencing of the plasmid.
Expression and Purification of Recombinant Proteins-Expression and purification of IL-13R112D and wtIL-13 was carried out by essentially similar techniques as previously reported for IL-4 (25). In the present set of experiments, we used BL21(DE3)pLys E. coli that contains T7 RNA polymerase under the lac promoter operator in its genome. The protein expression was induced by adding 1 mM isopropyl-␤-D-thiogalactopyranoside. WtIL-13 and IL-13R112D were produced in inclusion bodies. After washing, inclusion bodies were solubilized, refolded, and purified by fast protein ion-exchange liquid chromatography. The purified protein showed a single band at 13 kDa in Coomassie Blue-stained SDS-polyacrylamide gel (Fig. 2). In this study, we used IL-13-PE38 fusion protein, which was expressed in E. coli and purified as described previously (26). 2 Cell Proliferation Assays-Proliferation assays were performed as described previously (27). Briefly, TF-1 and B9 cells were washed 2-3 times to remove granulocyte-macrophage colony-stimulating factor and IL-6, and then 1 ϫ 10 3 to 5 ϫ 10 3 cells were cultured in 96-well plates in RPMI complete medium containing 10% fetal bovine serum. Varying concentrations of wtIL-13 and IL-13R112D were added to the wells, and the cells were cultured for 1-2 days. Tritiated thymidine (0.5 Ci) was added to each well 6 -9 h before the plates were harvested in a Skatron cell harvester (Skatron, Inc., Sterling, VA). Filter mats were counted in a ␤ plate counter (Wallac, Gaithersburg, MD).
Protein Synthesis Inhibition Assay-Protein synthesis inhibition assay was performed as described previously (28). In brief, 1 ϫ 10 3 PM-RCC cells were cultured in leucine-free medium (Biofluids, Rockville, MD) for 4 h to allow adherence to flat-bottomed microtiter plates. 2 Bharat. H. Joshi, and Raj K. Peri, unpublished results. Flow Cytometry-Flow cytometric analysis of monocytes were performed as described elsewhere (4). Primary monocytes were cultured at 1 ϫ 10 7 cells/ml in polypropylene tubes for 72 h with various concentrations of wtIL-13 or IL-13R112D. Cells were washed and incubated at 4°C for 60 min in fluorescence-activated cell sorter staining buffer (Hanks 1 balanced salt solution plus 0.5% fetal bovine serum, 0.1% sodium azide) containing fluorescein isothiocyanate-conjugated anti-human CD14 (Becton Dickinson, San Jose, CA) antibodies as per the manufacturer's recommendations. For controls, cells were either incubated in fluorescence-activated cell sorter staining buffer alone or with isotype control antibody, mouse IgG2a, and then antimouse Ig fluorescein isothiocyanate-conjugated was used as secondary antibody for staining. The cells were subsequently washed, and fluorescence data were collected on a FACScan/C32 equipment (Becton Dickinson). The results were analyzed with a WinList software program, and fluorescence intensity was expressed as mean channel number on 256 channel/ 10 4 log scale.

RESULTS AND DISCUSSION
Selection of Arg-112 for IL-13 Mutation, Protein Expression, and Purification-The crystal structure of IL-13 is not known. According to the conventional prediction algorithm, which predicts approximately 56% accuracy, there may be four major ␣-helices (A, B, C, and D) in the IL-13 molecule. The secondary structure of IL-13 was predicted based on the homology with IL-4 (1) and multiple alignments between human IL-4, human IL-2, granulocyte-macrophage colony-stimulating factor, granulocyte-colony-stimulating factor, and growth hormone (29). Based on these predictions, the locations of ␣-helices A and D were found to be at similar positions, whereas the location of ␣-helices B and C could not be accurately predicted in these models. The sequence homology alignment of IL-13 between species revealed that human IL-13 sequence is prominently conserved between four species (Fig. 1A). For example, the consensus sequence shows that a total of 87 amino acid residues are highly conserved between species. It has been shown that a cluster of amino acid residues in ␣-helix A, B, and C of IL-4 binds to IL-4R␤ chain, whereas a functional cluster that interacts with ␥c is suggested to be located in helix A and D, particularly amino acids Ile-11, Asn-15, Arg-121, Tyr-124, and Ser-125 (30 -32). Since most (approximately 70%) of the conserved and charged residues are located in predicted ␣-helix regions of IL-13, it is hypothesized that these residues may be essential in IL-13 binding to its receptors. Because IL-4R␤ chain and IL-13R␣Ј chains are shared between IL-4R and IL-13R systems and IL-2R␥ chain does not interact with IL-13 (15,23), it was predicted that amino acids in helix D interact with IL-13R␣Ј chain. We further hypothesized that 1) hydrophilic residues (with acidic or basic side chains) are usually exposed at the surface of a protein that might be expected to be involved in receptor binding and 2) highly conserved residues between species might be important for IL-13 to be an IL-13. Based on these hypotheses and the fact that the residue R112 of human IL-13 is hydrophilic and positive charge at this position is conserved between three species, we decided to mutate this amino acid and test the biological activities (Fig. 1B).
WtIL-13 and IL-13R112D were expressed and purified in an identical manner. As shown in Fig. 2 (lanes 1 and 2 in A and B), isopropyl-␤-D-thiogalactopyranoside induced protein expression very efficiently. Purification of proteins from inclusion bodies revealed a major band of 13 kDa (Fig. 2, lane 3 of A and  B). Upon purification on cation exchange chromatography, a highly purified protein (Ͼ95% pure) was obtained. (Fig. 2, lane  4 of A and B). Thus, one-step purification provided highly purified proteins. One liter of each bacterial culture yielded several milligram of each type of pure proteins. To confirm the identity, the protein was shown to react with anti-human IL-13 antibodies on Western blot analysis (results not shown). In some purifications, we also observed a minor 26-kDa protein in SDS-gel that is regarded as dimerized IL-13R112D or wtIL-13 (results not shown).
Proliferation Activity of wtIL-13 and IL-13R112D on Hematopoietic Cell Lines-After purification of wtIL-13 and IL-13R112D, the goal was to compare their biological activities on various cell types that express different types of IL-13R. First we tested their mitogenic activity. wtIL-13 has been shown to induce proliferation of TF-1 human erythroleukemia cell line (26,33). We tested the proliferative activity of IL-13R112D on TF-1 cell line. As shown in Fig. 3A, proliferative activity of IL-13R112D was more than 10 times better than induced by wtIL-13. The concentration of wtIL-13 that produced half-maximal proliferation (ED 50 ) was about 2 ng/ml compared with less than 0.2 ng/ml for IL-13R112D. Similarly, IL-13R112D stimulated mouse plasmacytoma cell line B9 much stronger than wtIL-13. IL-13R112D was 5.7-19-fold better than wtIL-13 in proliferation assays. Thus, proliferation activity of IL-13R112D on hematopoietic cells that express type III IL-13R is about one log greater than wtIL-13.
Down-regulation of CD14 Expression on Monocytes by IL-13R112D-Since IL-13 has been shown to down-regulate CD14 expression on monocytes (4), we investigated whether IL-13R112D mutant has stronger activity compared with wtIL-13. As shown in Fig. 4, IL-13R112D and wtIL-13 suppressed CD14 expression on monocytes in a dose-dependent manner. IL-13R112D was 10 times superior to wtIL-13 in down-regulation of CD14. For example, 1 ng/ml IL-13R112D induced downregulation of CD14, which was similar to that induced by 10 ng/ml wtIL-13.
Inhibition of IL-13 Toxin Mediated Cytotoxicity by IL-13 Mutant-We previously demonstrated that IL-13 toxin (IL-13-PE38QQR) is specifically highly cytotoxic to the PM-RCC cell line (10). To determine the superiority of IL-13R112D over wtIL-13, we compared the activity of wtIL-13 and IL-13R112D as it displaced cytotoxicity mediated by IL-13-PE38 in PM-RCC cells. As shown in Fig. 7A, IL-13R112D appeared to be better than wtIL-13 in blocking the cytotoxicity of IL-13-PE38. IL-13-PE38 was highly cytotoxic to these cells with a concentration that inhibited protein synthesis by 50% (IC 50 ) was less than 0.1 ng/ml. In the presence of 2 g/ml wtIL-13, the IC 50 increased to 60 ng/ml, whereas in the presence of IL-13R112D, the IC50 reached to 105 ng/ml. To carefully determine the extent of superiority of IL-13R112D in blocking cytotoxicity of IL-13-PE38, we used varying concentrations of cytokines in the presence of a fixed concentration of IL-13-PE38 (Fig. 7B). In this assay, IL-13R112D appeared to be approximately 10 times better than wtIL-13 in blocking the cytotoxicity (Fig. 7B).
IL-13 is a central mediator of asthma and may play a major role in cancer biology, because a variety of solid tumor cells express abundant numbers of receptors for IL-13 (3, 7, 10 -12, 17, 18, 21, 36 -40). We hypothesize that elimination of IL-13R-expressing cells may provide therapeutic benefit for disease processes where IL-13 and IL-13R are involved. To achieve that goal, we produced a IL-13 cytotoxin that eliminated IL-13R-expressing cells, but this molecule had 10-times lower binding affinity to IL-13R than wtIL-13 (10). To improve the binding affinity, in the current study we made IL-13R112D that not only bound better than wtIL-13 to IL-13R but turned out to be a IL-13 agonist with improved biological activities on various cell types that express different types of IL-13R. These results suggest that Arg-112 is involved in IL-13 binding to its receptors.
Since IL-13R112D had agonistic activity in cell types that expressed type I, type II, or type III IL-13R, it could not be concluded whether this amino acid residue is responsible for binding to IL-13R␣ or IL-13R␣Ј chains or both. Because IL-13R112D also displaced IL-4 binding more effectively than wtIL-13 and since IL-13R␣Ј chain but not IL-13R␣ is shared with IL-4R system, it is presumed that amino acid residue Arg-112 more likely interacts with IL-13R␣Ј chain. Additional studies are ongoing to determine specific interaction with different receptor subunits.
In conclusion, we produced a novel IL-13 agonist with improved binding affinity to IL-13R. This molecule may be useful in the study of IL-13 function, for example in the activation of potent antigen-presenting dendritic cells, its interaction with IL-13R and its role in inflammatory diseases, and in designing IL-13 cytotoxins with improved binding activity and cytotoxicity to IL-13R-expressing cells.