FASEB J. Mp Biomedicals
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published as doi: 10.1096/fj.08-111484.
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.08-111484v1
22/11/3795    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fan, C.-Y.
Right arrow Articles by Chou, M.-Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fan, C.-Y.
Right arrow Articles by Chou, M.-Y.
(The FASEB Journal. 2008;22:3795-3804.)
© 2008 FASEB

Production of multivalent protein binders using a self-trimerizing collagen-like peptide scaffold

Chia-Yu Fan1,*, Chuan-Chuan Huang1,*, Wei-Chun Chiu*, Chun-Chieh Lai{dagger}, Gunn-Guang Liou{dagger}, Hsiu-Chuan Li* and Min-Yuan Chou2,*

* Biomedical Engineering Research Laboratories, Industrial Technology Research Institute, Taiwan; and

{dagger} Division of Molecular and Genomic Medicine, National Health Research Institutes, Taiwan

2Correspondence: Biomedical Engineering Research Laboratories, Industrial Technology Research Institute, Bldg. 53, No 195, Sec. 4, Chung Hsing Rd., Chutung, Hsinchu 310, Taiwan, Republic of China. E-mail: minyuanc{at}itri.org.tw


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A class of multivalent protein binders was designed to overcome the limitations of low-affinity therapeutic antibodies. These binders, termed "collabodies," use a triplex-forming collagen-like peptide to drive the trimerization of a heterologous target-binding domain. Different forms of collabody, consisting of the human single-chain variable fragment (scFv) fused to either the N or C terminus of the collagen-like peptide scaffold (Gly-Pro-Pro)10, were stably expressed as soluble secretory proteins in mammalian cells. The collabody consisting of scFv fused to the N terminus of collagen scaffold is present as a homotrimer, whereas it exhibited a mixture of trimer and interchain disulfide-bonded hexamer when cysteine residues were introduced and flanked the scaffold. The collagenous motif in collabody is prolyl-hydroxylated, with remarkable thermal and serum stabilities. The collabody erb_scFv-Col bound to the extracellular domain of epidermal growth factor receptor with a binding strength ~20- and 1000-fold stronger than the bivalent and monovalent counterparts, respectively. The trimeric collagen scaffold does not compromise the functionality of the binding moieties of parental immunoglobulin G (IgG); therefore, it could be applied to fuse other protein molecules to acquire significantly improved targeting-binding strengths.—Fan, C.-Y., Huang, C.-C., Chiu, W.-C., Lai, C.-C., Liou, G.-G., Li, H.-C., and Chou, M.-Y. Production of multivalent protein binders using a self-trimerizing collagen-like peptide scaffold.


Key Words: phage-display • single-chain antibody • EGFR • CD3


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
ANTIBODY AFFINITY, MEASURING the strength of an antibody binding to its antigen epitope, is regarded as critical to the success of an antibody as a therapeutic agent. An antibody with high affinity can compete effectively with a natural ligand for the targeted receptor, to reduce dosing, toxicity, and cost of therapy. Affinity can also influence pharmacokinetics, including the distribution and excretion of the designated antibody in the targeted tissue as well as host circulation. A requirement for an effective therapeutic antibody in vivo is strong affinity for the target antigen without nonspecific binding to unrelated proteins. Unfortunately, current technologies in humanization of murine monoclonal antibody (mAb) by complementarity-determining region (CDR) -grafting with or without structure-based design to transform CDR residues from murine to human origin often result in a reduction or loss of binding affinity (1 , 2) . Affinity maturation by chain shuffling using phage-display single-chain variable fragment (scFv) library screening for high-affinity binders is a tedious process, yet the outcome of improving binding affinity is uncertain. Therefore, many therapeutic antibodies may be hampered by low affinity for the target antigen after humanization. Functional affinity (avidity) is a measure of the overall strength of binding of an antigen with many antigenic determinants and multivalent antibodies. Polymerization of antigen-binding partners greatly increases their availability (or valency) for binding to a group of specific identical ligands in extreme proximity to a target cell. Different approaches have been proposed to obtain multivalent molecules for improving avidity. The use of non-IgG protein scaffold fragments, such as core-streptavidin for the formation of a tetrameric complex (3) , the human tetramerization domain of transcription factor p53 (4) , the "ankyrin repeats" (5) , the "A-domain avimers" (6) , and the C-type lectin-like domain of tetranectin (7) have recently been adopted to increase target binding strength, thermal stability, and sensitivity. However, some of these molecules are either nonhuman fragments or not natural components of plasma, and are associated with the risk of an immune response that could severely limit potential therapeutic applications.

Collagens are extracellular matrix proteins that contain collagenous domains with the repeating triplet sequence Gly-X-Y. The presence of such triplets allows 3 collagen polypeptide chains ({alpha}-chains) to fold into a triple-helical conformation. In the Y position, Pro is generally post-translationally modified to 4-hydroxyproline (O or Hyp) by prolyl 4-hydroxylase (P4H) to stabilize the triple-helical structure of collagen. In the absence of proline hydroxylation, the essential triple helical conformation of collagen is thermally unstable at below physiological temperatures (8 , 9) . Many collagen-like proteins with collagenous domains are present in human serum and serve as an innate immune system in protection from infectious organisms. These include the complement protein C1q, the collectin family of proteins—mannose binding lectin (MBL), ficolins, and surfactant proteins A and D. A common structural feature among these "defense collagen" molecules is that all are in multitrimeric protein units with a target-binding domain at the C terminus. Consequently, multimerization significantly increases the avidity of the binding domain of these defense collagen molecules. Therefore, it seems to be a feasible approach to use a collagen-like domain as scaffold to create a range of designer protein drugs that have improved therapeutic properties.

We attempted to design a novel antibody format that can use multivalent antigen-binding fragments to increase antibody avidity. As well an improving binding avidity, the designed structure should also satisfy the criteria for biotherapeutics, such as solubility, serum stability, reduced immunogenicity/mitogenicity, and pharmacokinetic standards. The sequence Gly-Pro-Hyp is the most stabilizing and most common triplet in collagen, and the peptide (Gly-Pro-Hyp)10 can self-associate into a highly stable triple helical structure (10 , 11) . Therefore, it is possible to use a self-assembled collagen-like domain to drive the trimerization of a heterologous target-binding domain, resulting in an increase in binding valence compared to the bivalent IgG format. In this study, a short collagen-like peptide, (Gly-Pro-Pro)10, was adopted as a fusion partner with an antigen-binding variable domain, scFv, to facilitate the formation of a stable triple helical structure under physiological conditions by the expression of a fusion construct in a P4H-containing mammalian system. This chimeric protein is designated as a collabody. We demonstrated that the (Gly-Pro-Pro)10 peptide scaffold by itself can drive the trimerization of an N- or C-terminal fusion partner of scFv. The collagen-like motif in collabody was mostly prolyl-hydroxylated, forming a thermally stable triple helical structure. The trimeric collabody format was demonstrated to improve target-binding affinities and it retains the functionality of parental IgG, making it an ideal platform for therapeutic purposes.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Selection of phage library against human epidermal growth factor receptor-extracellular domain (EGFR-ECD)
The erb phagemid containing the EGFR-ECD-binding variable fragment scFv was isolated by screening from a human single-fold scFv phage display library (Tomlinson I+J), kindly provided by I. M. Tomlinson and G. Winter (MRC Laboratory of Molecular Biology, Cambridge, UK). Selections were performed using immunotubes (Maxisorp; Nunc, Roskilde, Denmark) coated with 10 µg of purified recombinant ECD of EGFR (Research Diagnostics, Inc., Flanders, NJ, USA). Blocking, panning, washing, elution, and reamplification of eluted phage were carried out according to published protocols (12) .

Construction of recombinant plasmids
The coding region of scFv-Col collabody is composed of an N-terminal scFv nucleotide sequence and a C-terminal synthetic gene coding for a peptide sequence of EPKSCDKTHTCPPCPRSIP(GPP)10GICDPSLCFSVIARRDPFRKGPNY, which comprises a hinge region of human IgG (underlined), a collagen-like peptide (GPP)10, and the NC1 domain of type XXI collagen (in italics). This synthetic gene was prepared by overlapping polymerase chain reaction (PCR). The PCR product flanked with NotI and XhoI sites was cloned into the expression vector pSecTag2/Hygro (Invitrogen, Carlsbad, CA, USA) at the same sites. The cDNAs coding for the scFvs of erb and the murine IgG2a anti-CD3 mAb OKT3 (Ortho Biotech Inc., Bridgewater, NJ, USA) were PCR amplified from the erb phagemid and the reverse transcription product from OKT3 hybridoma (CRL-8001; American Type Culture Collection, Manassas, VA, USA), respectively. The cDNAs for the VL and VH of the OKT3 IgG were obtained by reverse transcriptase (RT)-PCR based on the published nucleotide sequence (13) . The entire collabody cDNAs for erb_scFv-Col and OKT3_scFv-Col were cloned inframe with the N-terminal Ig{kappa} chain leader sequence and the C-terminal myc epitope/polyhistidine tag of the pSecTag2/Hygro expression vector for secretion, detection, and purification purposes.

The coding region of Col-erb_scFv collabody is composed of an N-terminal synthetic collagen scaffold gene coding for a peptide sequence of TCPPCPRSIP (GPP)10 GICDPSLC, and a C-terminal erb_scFv nucleotide sequence. The cDNA of erb_scFv was cloned inframe to the above collagen scaffold sequence at the C terminus by overlapping PCR, and the PCR product flanking with AgeI and BamHI sites was cloned into the expression vector pSecTag2/Hygro to make the expression construct of Col-erb_scFv.

The coding region of erb_scFv-GPP10 included an N-terminal nucleotide sequence of erb_scFv and a C-terminal synthetic collagen scaffold gene coding for a peptide sequence of GSP(GPP)10GPSSGG. The cDNA of erb_scFv was cloned inframe to the above C-terminal collagen scaffold sequence by overlapping PCR, and the PCR product flanking with AscI and AgeI sites was cloned into the expression vector pSecTag2/Hygro to make the expression construct of erb_scFv-GPP10.

Expression and purification of antibodies
All recombinant antibodies were obtained by stable transfection of expression constructs in mouse myeloma NS0 cells (European Collection of Animal Cell Cultures, Wiltshire, UK) using Effectene (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. After selection with 400 µg/ml Hygromycin B (A. G. Scientific, San Diego, CA, USA) for 4 wk, a stable clone was cultured in a shaker flask at an initial seeding density of 5 x 105 cells/ml in a chemically defined medium HyQCDM4NS0 (Hyclone Laboratories, Logan, UT, USA) containing 2% fetal bovine serum. The culture was maintained at 130 rpm for 5 days at 37°C. Sodium ascorbate (80 µg/ml) was added to the culture media daily for those cells carrying collabody expression constructs. For the purification of collabodies, approximately 2 L each of the filtered culture media was applied to a T-gel column (1.5x8 cm; Pierce Biotechnology, Rockford, IL, USA) equilibrated with 50 mM Tris-HCl buffer containing 0.5 M of KCl, pH 8.0, at a flow rate of 60 ml/h. After washing with the same buffer, the recombinant antibodies were eluted with 50 mM of sodium acetate buffer, pH 4.0. The UV absorbance was monitored at 280 nm; the peak fraction was collected, neutralized with 1.0 M Tris base to pH 8.0, and applied onto a ZnSO4-charged chelating Sepharose HiTrap column (1 ml bed volume; GE Healthcare, Chalfont St. Giles, UK) equilibrated with 50 mM Tris-HCl buffer containing 0.5 M NaCl, pH 8.0, at a flow rate of 60 ml/h. The column was first washed with 20 mM of imidazole, then the bound antibodies were eluted with 0.25 M of imidazole in the same buffer. The final preparation was dialyzed against 50 mM of Hepes buffer (sodium salt), pH 7.0.

SDS-PAGE
SDS-PAGE was carried out using either a 10% NuPAGE Bis-Tris polyacrylamide gel with 3-morpholinopropanesulfonic acid (MOPS) or a 7% SDS/Tris acetate polyacrylamide gel with sodium acetate as running buffer (Invitrogen). Proteins were stained with Coomassie brilliant blue R-250.

Surface plasmon resonance (SPR)
The binding kinetics of erb antibody variants to the EGFR-ECD were measured using a BIAcore X biosensor (Biacore, Inc., Uppsala, Sweden) in the running buffer HBS-EP (10 mM HEPES, pH 7.4; 150 mM NaCl; 3 mM EDTA; 0.005% surfactant P20). Briefly, EGFR-ECD was immobilized onto a C1 sensor chip via amine coupling to a level of 1700 response units (RU), and purified antibodies with different concentrations were injected at a flow rate of 10 µl/min. The surface was regenerated by injection of 5 µl of 10 mM glycine-HCl, pH 2.5. Sensorgrams were obtained at each concentration and were evaluated using the program BIA Evaluation 3.2 (Biacore). Binding data were fitted with a 1:1 Langmuir binding model to calculate the affinity constant KD, which was defined as the ratio of dissociation rate (koff)/association rate (kon).

Stability and pharmacokinetic assays
A quantitative ELISA employing the recombinant EGFR-ECD as capture reagent and anti-c-myc mAb (9E10; Sigma-Aldrich Corp., St. Louis, MO, USA), followed by a horseradish peroxidase (HRP) -conjugated affinity-purified polyclonal goat anti-mouse IgG and chemiluminescent substrates (Pierce Biotechnology) as detecting reagent, was used to evaluate serum concentrations of erb antibody variants carrying the C-terminal myc-tag during the course of the experiment. For serum stability assay, the individual erb antibody variant was diluted in sterile human serum at a final concentration of 5 µg/ml and incubated at 37°C. Aliquots (100 µl) at different incubation time points were transferred and stored at –80°C. The amount of active anti-EGFR remaining in each aliquot was measured by ELISA. For pharmacokinetic assay, 18 7- to 8-week-old male C57BL/6 mice (3 per group) were used to analyze erb_scFv-Col clearance. Following a prebleed, each mouse was injected intravenously with 2 mg/kg of erb_scFv-Col and during the next 72 h, periodic blood samples were collected. The amounts of active anti-EGFR remaining in plasma were quantitated by ELISA using rabbit anti-c-myc antibodies conjugated with HRP (Bethyl Laboratories, Montgomery, TX, USA) as detecting antibody. The pharmacokinetic experiment was performed at the Center of Toxicology and Preclinical Sciences at the Development Center for Biotechnology (Taiwan, Republic of China) by Dr. Charlene Chen. The concentration of erb_scFv-Col in plasma was determined by ELISA for each time point and fitted to a two-compartment elimination model using WinNonlin version 3.0 (Pharsight, Mountain View, CA, USA).

Antibody displacement assays
Human peripheral blood mononuclear cells (PBMCs) were obtained from healthy blood bank (Hsinchui Blood Center, Taiwan, Republic of China) donors by Ficoll-Hypaque density gradient centrifugation of EDTA-anticoagulated whole blood. All of the following procedures were conducted at 4°C. Human CD3(+) T cells were isolated using human T-cell enrichment columns (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions and were suspended in FCM buffer (PBS with 2% fetal bovine serum and 0.1% sodium azide) at a density of 1 x 106 cells/ml. The cells were treated with mouse total IgGs (2 µg/ml, Jackson ImmunoResearch Laboratories, West Grove, PA, USA) for 30 min and were then incubated with a serial dilution of OKT3_scFv-Col or OKT3 IgG for 1 h. A fixed, saturating amount (0.25 µg/ml, determined by flow cytometry) of fluorescein isothiocyanate (FITC) -conjugated OKT3 IgG (eBioscience, San Diego, CA, USA) was added directly. After incubation for 1 h, the cells were washed with FCM buffer and analyzed for immunofluorescence by flow cytometry on a FACScan (Becton Dickinson, San Jose, CA, USA). The data are presented as percentage inhibition of maximal fluorescence intensity, which is defined as the mean fluorescence intensity obtained by staining T cells with OKT3-FITC in the absence of blocking mAbs. The concentration of each mAb required to inhibit half the maximal fluorescence intensity (IC50) was calculated.

Quantitation of TCR-CD3 modulation/coating
PBMCs from a healthy donor were plated at 2 x 106 cells/well in a 24-well plate (Nunc) and incubated in RPMI 1640 plus 10% FCS with varying amounts of OKT3 and OKT3_scFv-Col. After 24 h incubation, cells were harvested and stained with FITC-conjugated OKT3 or anti-T-cell receptor (TCR){alpha}/β mAb IP26 (eBioscience). The stained cells were counterstained with phycoerythrin-conjugated anti-CD5 mAb UCHT2 (eBioscience) for T cells and analyzed by flow cytometry.

Calculation of CD3 modulation and coating was performed as described elsewhere (14) :

Formula

Formula
where F represents mean fluorescence of stained cells.

Electron microscopy and circular dichroism spectroscopy
Uranyl formate-stained samples were prepared for electron microscopy and images were taken as described elsewhere (15) . Structural figures were prepared with the program PyMol (available at http://pymol.sourceforge.net/). Circular dichroism (CD) spectroscopy was made on an Aviv 62DS spectrometer (Aviv Biomedical, Lakewood, NJ, USA). Wavelength scan was first performed from 200 nm to 260 nm (0.1 cm path length) with purified erb_scFv-Col at a concentration of 0.5 mg/ml in PBS buffer (pH 7.0). The change in ellipticity {theta} at 222 nm was monitored as the protein sample was heated from 20°C to 96°C in 2°C increments at a heating rate of 5°C min–1. The fraction folded was calculated from the CD melting curve as F(T) = [{theta}(T)–{theta}u(T)]/({theta}n(T)–{theta}u(T)], where {theta}n and {theta}u represent ellipticities of the native and unfolded monomeric forms, respectively, at temperature T. The melting temperature Tm was obtained as a midpoint of the transition, that is, F(Tm) = 1/2.

Other assays
Protein concentration was determined by Bradford assay (Coomassie plus reagent, Pierce Biotechnology) using human IgGs as the standard (16) . For amino acid analysis, purified erb_scFv-Col was dialyzed against 50 mM acetic acid, hydrolyzed in 6 N of HCl at 110°C for 24 h and subjected to amino acid analysis in a Waters PicoTag system (Waters, Milford, MA, USA).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Phage-display screening for scFv binding to EGFR and construction of antibody variants
The human synthetic scFv library containing 108 recombinant phages was screened for adsorption to human EGFR-ECD that was immobilized to an immunotube. After the third round of screening, 96% of the phages reacted positively in ELISA. Sequencing 10 scFv cDNAs confirmed that all clones were identical. This scFv clone was designated erb, and its cDNA (Supplemental Fig. 1) was used to construct various collabody formats, including erb_scFv-Col, Col-erb_scFv, and erb_scFv-GPP10. In addition, the scFv derived from OKT3 IgG was fused to the C-terminal collagen scaffold backbone of scFv-Col to make an OKT3_scFv-Col. The molecular structures of the antibody variants used in this study are summarized in Fig. 1 .


Figure 1
View larger version (16K):
[in this window]
[in a new window]

  		Figure 1. Schematic representation of different antibody formats. Top panel, collabodies: erb_scFv-Col or OKT3_scFv-Col contains an amino terminus of either erb_scFv or OKT3_scFv, followed by a hinge region of human IgG1, a collagen-like peptide (GPP)10, and a carboxyl-terminal disulfide knot derived from the NC1 domain of type XXI collagen; Col-erb_scFv contains an amino-terminal disulfide knot, a collagen-like peptide (GPP)10, followed by a carboxyl-terminal disulfide knot of type XXI collagen and an erb_scFv; and erb_scFv-GPP10 contains an amino-terminal erb_scFv and a collagen-like peptide (GPP)10. Bottom panel: the erb_scFv-Fc contains an amino-terminal erb_scFv, the hinge region, and the CH2 and CH3 domains of human IgG1. Approximate molecular masses (kDa) are indicated. Dashed lines indicate the putative interchain disulfide bonds. A glycine-linker, (GGGGS)3, was introduced to join the VH and VL chains of an scFv.

Structural characterization of collabodies
Recombinant antibodies of erb_scFv-Col, OKT3_scFv-Col, erb_scFv-Fc, and erb_scFv were expressed as soluble secretory proteins in mouse myeloma NS0 cells. Each of them was purified and analyzed by SDS-PAGE (Fig. 2 A). Under nonreducing conditions, 2 major bands were resolved in erb_scFv-Col (lane 2), whereas only a single major band was observed in OKT3_scFv-Col (lane 3). The bottom band of erb_scFv-Col and OKT3_scFv-Col migrated to a position of 125 kDa, corresponding closely to the calculated molecular mass of the trimeric form of both scFv-Col monomers (41 kDa). The upper band shown in erb_scFv-Col (lane 2) appears to be an interchain disulfide-bonded dimer of trimers. This finding was confirmed by incubating the sample under mild reducing conditions, as shown in Fig. 2B : the interchain disulfide-bonded hexamer (250 kDa) of erb_scFv-Col was dissociated into trimers (125 kDa). In Fig. 2A , samples were treated under reducing conditions with 50 mM of dithiothreitol (DTT) for 10 min at 70°C, and the interchain disulfide-bonded hexamer of erb_scFv-Col was completely reduced to the trimeric form, whereas only some of the erb_scFv-Col trimer was further dissociated into monomer (lane 7). Interestingly, the trimeric conformation of OKT3_scFv-Col was resistant to dissociation into a monomeric form under these reducing conditions (lane 8). The bivalent counterpart of erb antibody, erb_scFv-Fc, migrated as a dimer under nonreducing conditions with an apparent molecular mass of 125 kDa (lane 4), revealing an almost monomeric form with an apparent molecular mass of 57 kDa after the interchain disulfide bonds were reduced (lane 9). The monovalent counterpart of the erb antibody, erb_scFv, migrated as a single band with an apparent molecular mass of 28 kDa under either nonreducing or reducing conditions.


Figure 2
View larger version (17K):
[in this window]
[in a new window]

  		Figure 2. Purification and structural characterization of the various antibody molecules. A) The indicated antibodies were stably expressed in the mouse myeloma NS0 cells and purified from culture media by column chromatography. The samples were electrophoresed on a 10% SDS/Bis-Tris polyacrylamide gel with MOPS buffer under nonreducing conditions (lanes 1 to 5) and reducing conditions (lanes 6 to 10). B) An erb_scFv-Col hexamer is formed by interchain disulfide bonding of 2 trimeric molecules. Purified erb_scFv-Col (1 mg/ml) was incubated at 37°C in the absence (lane 1) or presence (lane 3) of 10 mM DTT for 1 h. An aliquot from the DTT-treated sample was further reacted with 50 mM NEM for 30 min at ambient temperature (lane 2). All samples with equal amounts of protein were electrophoresed on a 7% SDS/Tris acetate polyacrylamide gel with sodium acetate as a running buffer. C) Purification of Col-erb_scFv. Purified samples were electrophoresed on a 10% SDS/Bis-Tris polyacrylamide gel with MOPS buffer under nonreducing conditions (lane 1) and reducing conditions (lane 2). All gels were stained with Coomassie blue. M, molecular mass standards.

To investigate whether the (Gly-Pro-Pro)10 can drive the trimerization of a C-terminal fusion partner in vivo, we generated a collabody molecule, Col-erb_scFv, composed of an N-terminal synthetic collagen scaffold gene coding for a peptide sequence of TCPPCPRSIP (GPP)10 GICDPSLC, and a C-terminal erb_scFv. The purified Col-erb_scFv exhibits a structure feature similar to that observed in erb_scFv-Col, except that the amount of hexamer in Col-erb_scFv is less than that of erb_scFv-Col (Fig. 2C ). Therefore, in the mammalian expression system, the (Gly-Pro-Pro)10 peptide can drive the trimerization of either an N- or C-terminal fusion partner of scFv.

The hexameric and trimeric structures of the interchain disulfide-bonded species of erb_scFv-Col shown in Fig. 2A were further characterized to determine their triple-helical thermal stability. To exclude the contribution of interchain disulfide bridges to the trimeric assembly of the collabody molecules, the cysteine residues in erb_scFv-Col were first completely reduced using a strong reducing agent, tris(2-carboxyethyl)phosphine (TCEP), at room temperature, and then alkylated with N-ethyl-maleimide (NEM) to prevent the reformation of the disulfides. Equal amounts of nonreduced or reduced/alkylated samples were incubated in Tris-HCl (50 mM, pH 8) that contained 2 M urea at the indicated temperatures, and the dissociation of the triplex was assayed by SDS-PAGE under nonreducing conditions to estimate the thermal stability of the collagen triple helix. Similar to the results of Fig. 2B , the interchain disulfide-bonded hexamer species were readily dissociated into trimers at 35°C (Fig. 3 A, compare lanes 1 and 4). The reduced/alkylated trimers dissociated into monomers significantly as the incubation temperature was increased (Fig. 3A , lanes 4–9). The same experiment under nonreducing conditions (Fig. 3A , lanes 1–3) did not show any alteration of the hexameric or trimeric structure, although erb_scFv-Col was partially degraded at high incubation temperatures. CD spectroscopy was used to characterize the thermal stability of erb_scFv-Col. A sharp thermal transition Tm of 54.5°C was observed as the temperature was increased (Fig. 3C ).


Figure 3
View larger version (10K):
[in this window]
[in a new window]

  		Figure 3. Thermal stability of trimeric structure of erb_scFv-Col and erb_scFv-GPP10. A, B) Purified erb_scFv-Col (A) and erb_scFv-GPP10 (B) in 50 mM Tris-HCl (pH 8.0), containing 2 M urea, were respectively treated in the absence (lanes 1 to 3) or presence (lanes 4 to 9) of 10 mM TCEP at ambient temperature. The reduced erb_scFv-Col was further alkylated with 50 mM NEM at ambient temperature. All samples with an equal amount of protein were heated for 10 min at the indicated temperatures and then SDS-loading buffer was added immediately. The samples were electrophoresed on 10% SDS/Bis-Tris polyacrylamide gels with MOPS buffer under nonreducing conditions. The gels were stained with Coomassie blue. C) Thermal transition curve of erb_scFv-Col monitored by circular dichroism at 222 nm. Purified erb_scFv-Col was heated from 20°C to 96°C in 2°C increments. The melting temperature Tm was obtained as a midpoint of the transition.

Requirements for driving collabody trimerization
A collabody molecule erb_scFv-GPP10 was generated to demonstrate that the collagen-like peptide (GPP)10 by itself can drive the formation of a noncovalently bound trimeric fusion protein. In comparison to the erb_scFv-Col structure, the non-disulfide-knot-containing erb_scFv-GPP10 does not further assemble into an interchain disulfide-bonded hexamer (Fig. 3B , lane 1). Meanwhile, the thermal stability behavior of erb_scFv-GPP10 is similar to that of the reduced/alkylated interchain disulfide-bonded structure of erb_scFv-Col, indicating that (GPP)10 by itself can form a thermal stable triplex structure and drive the formation of collabody molecule. Electron micrographs of erb_scFv-GPP10 most often showed ring structures in which three major globular domains could be distinguished (Fig. 4 A). These domains correspond well to the trimeric assembled scFv consisting of a VH and a VL domains joined by a glycine linker (Fig. 4B ).


Figure 4
View larger version (82K):
[in this window]
[in a new window]

  		Figure 4. Structure of erb_scFv-GPP10. A) Electron micrograph of purified erb_scFv-GPP10. Top panel: a general micrograph area of negatively stained erb_scFv-GPP10 that was stained with uranyl formate. Bottom panel: gallery of individual erb_scFv-GPP10 molecules that revealed the erb_scFv-GPP10 molecules to be comprised of three major globular domains. Scale bars = 20 nm. B) Structural model of the collabody molecule erb_scFv-GPP10. The antigen-binding variable domain scFv and the collagen-like scaffold domain are colored purple and green, respectively. The molecular surface is shown in transparency.

Hydroxyproline is important to the thermal stability of a collagen triplex structure. Amino acid composition analyses revealed a close match between the determined amino acid composition and the predicted data, based on the deduced cDNA sequence of erb_scFv-Col (Supplemental Table 1). The extent of prolyl hydroxylation in erb_scFv-Col was 61%, as determined from the theoretical value of 10 fully hydroxylated proline residues in the (Gly-Pro-Pro)10 motif. The results indicate that the (Gly-Pro-Pro)10 motif in collabody molecules was a good substrate for prolyl 4-hydroxylase, and the mouse myeloma NS0 cells exhibit sufficient prolyl hydroxylase activity for the biosynthesis of collagen molecules.

Antibody binding analysis
The interaction of the three erb antibody variants, erb_scFv-Col, erb_scFv-Fc, and erb_scFv, with EGFR-ECD was studied using an SPR assay, and the equilibrium dissociation constant KD was determined. The KD of the binding of monovalent erb_scFv with EGFR-ECD ligands is of the order of 10–6 M, whereas the KD values of the binding of bivalent erb_scFv-Fc and trivalent erb_scFv-Col to EGFR-ECD are of the order of 10–7 and 10–9 M, respectively (Supplemental Fig. 2 and Table 1 ). The increase in the apparent affinity of the trivalent erb collabody over that of the bivalent and monovalent counterparts is ~20- and 1000-fold, respectively. Notably, the dissociation rate constant koff of erb_ scFv-Col is 8.22 x 10–4 s–1, and for erb_ scFv-Fc is 9.44 x 10–3 s–1, representing an 11-fold improvement in the off-rate for the trivalent species.


View this table:
[in this window]
[in a new window]

  		Table 1. Binding kinetics of various forms of erb antibody to immobilized EGFR-ECD

Stability and pharmacokinetics of erb_scFv-Col
The serum stability of erb_scFv-Col was studied and compared with those of erb_scFv-Fc and erb_scFv by incubating each of the purified antibody variants in human serum at 37°C for various periods. As shown in Fig. 5 A, erb_scFv-Col was more stable than erb_scFv-Fc in human serum at physiological temperature, retaining 60% of its initial binding activity within 72 h of incubation. The erb_scFv degraded rapidly in human serum, retaining less than 40% of its initial binding activity within 1 h of incubation. The results indicated that the triple-helical collagen-like peptide of erb_scFv-Col and the Fc region of erb_scFv-Fc are more resistant than erb_scFv to serum protease digestion. The pharmacokinetic profile of erb_scFv-Col in mice is shown in Fig. 5B . Kinetics of the two-compartment model were determined after a single intravenous administration of erb_scFv-Col at 2 mg/kg. The plasma level of immunoreactivity decreased biphasically, with a distribution phase half-life (t1/2{alpha}) of 0.21 h and a terminal elimination phase half-life (t1/2β) of 4.78 h.


Figure 5
View larger version (11K):
[in this window]
[in a new window]

  		Figure 5. Stability of the collabody molecule. A) Stability of the various forms of erb antibody in human serum. The stability of erb_scFv-Col, erb_scFv-Fc, or erb_scFv was determined by incubating at 37°C in human serum. The amount of active anti-EGFR that remained after various periods of incubation was determined by ELISA using anti-c-myc mAb. B) Pharmacokinetics of erb_scFv-Col in mice. Male C57BL/6 mice were injected intravenously with 2 mg/kg of erb_scFv-Col. Blood samples were drawn at different times. The erb_scFv-Col levels in plasma were determined by ELISA using rabbit anti-c-myc antibodies conjugated with HRP. Results are averaged from 3 animals for each time point; error bars represent SD.

Enhancement of CD3(+) T-cell surface retention and TCR modulation by the OKT3 collabody
To further investigate whether the triplex geometry of the collabody could allow multivalent binding to cell surface antigens and, most important, retain or even improve the functional potency of the parental IgG, we chose a therapeutic antibody as a working model. OKT3 is a murine monoclonal IgG2a antibody, with a binding avidity of 1.2 x 10–9 M toward the {epsilon}-chain of CD3 in the human TCR complex (17) . The avidities of the purified OKT3_scFv-Col and OKT3 for binding to CD3 molecules on the cell surface of human CD3(+) T cells were compared by flow cytometry analysis, using antibody displacement assay with a saturated concentration of fluorescein-conjugated OKT3 as a competitor. Comparison of the IC50 values indicated that OKT3 required a 3-fold higher concentration to achieve the same inhibition effect than that of OKT3_scFv-Col (Fig. 6 A). The results indicated that the trivalent OKT3 collabody can further improve the binding strength of its bivalent IgG form. To investigate whether improvement in the binding strength of OKT3_scFv-Col for CD3 could potentially increase the antibody immunosuppressive properties, the antibodies were compared for their capacity to modulate (internalize) and coat the TCR-CD3 complex. As shown in Fig. 6B , the combined modulation and coating of the TCR-CD3 complex achieved by OKT3_scFv-Col is greater than that of OKT3 at antibody concentrations ranging from 1 ng/ml to 10 µg/ml. These results demonstrated that the triplex geometry of OKT3_scFv-Col can improve human CD3(+) T-cell retention, leading to an enhanced degree of TCR modulation.


Figure 6
View larger version (12K):
[in this window]
[in a new window]

  		Figure 6. Characterization of OKT3_scFv-Col. A) OKT3 displacement assay. Human CD3(+) T cells were incubated with serial dilutions of either OKT3_scFv-Col or OKT3 IgG for 1 h. A saturating amount of FITC-conjugated OKT3 was added and incubated for an additional hour. Cells were washed, and bound FITC-conjugated OKT3 was quantified by flow cytometry. Values are expressed as percentage inhibition of maximal fluorescence, as determined by adding FITC-conjugated OKT3 without prior blocking antibodies. B) Modulation/coating of TCR-CD3 complex by OKT3 and OKT3_scFv-Col. Data for CD3 modulation represent the percentage of TCR-CD3 complexes on the surface of treated CD5-positive T cells as a fraction of TCR-CD3 complexes on the surface of untreated CD5-positive T cells. CD3 coating is shown as the fraction of TCR-CD3 complexes that could not be detected by FITC-conjugated OKT3.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
So far, trimerization of heterologous fusion proteins containing collagenous domains has been produced by employing either a homogeneous or heterologous trimerization domain fused to the collagenous domain to drive the collagen triplex formation. Examples of a trimer-oligomerizing domain include a C-propeptide of procollagens, a coiled-coil neck domain of the collectin family of proteins, a C-terminal portion of Fas ligand, and a bacteriophage T4 fibritin foldon domain (18 19 20) . The trimeric assembly of the fibrillar collagens (types I, II, III, IV, V, and XI) and collectins are initiated by trimeric association of their large globular C-terminal domains (C-propeptides, ~250 aa) and C-terminal coiled-coil neck domains (~35 aa), respectively, following by propagation of the collagen domains in a zipper-like fashion from the C to the N terminus (21 , 22) . In this study, we directly used a short collagenous sequence (~30 aa) as trimerizing scaffold capable of self-nucleation and propagation of the heterologous fusion proteins from either the C- or N-terminal direction, without the need for any other trimerization structure domains. The self-trimerization collagen scaffold is more versatile in that it allows attachment of fusion partners to either terminus, as well as to both termini simultaneously (unpublished data). This has important consequences, as the self-trimerization collagen scaffold may be deployed to construct molecules that are able to interact (each end with a binding valency up to 3 or 6) simultaneously with 2 bulky binding partners.

The sequence Gly-Pro-Hyp contributes most to the formation and stabilization of the triple helical structure and the Gly-Pro-Hyp tripeptide repeats self-assemble into a highly stable triple helix. In contrast to chemically synthesized (Gly-Pro-Hyp)10 peptide, the (Gly-Pro-Pro)10 peptide does not self-assemble into a stable triple-helix under physiological conditions (23) . For obtaining a thermally stable (Gly-Pro-Pro)10 triplex, 2 approaches have been described. First, an interchain disulfide-bonded (Gly-Pro-Pro)10 triplex can be obtained in vitro by a redox-shuffling process of a disulfide knot of type III collagen either C- or N-terminally adjacent to the collagen-like peptide at 20°C (24 , 25) . Second, a stable heterologous trimerizing foldon domain derived from bacteriophage T4 fibritin was fused to the C terminus of (Gly-Pro-Pro)10 peptide to drive the trimerization and correct folding of the collagen-like peptide in a P4H-deficient Escherichia coli expression system (18) . The approaches described above are limited in their use because the first in vitro approach may not support normal trimerization and folding of a heterologous polypeptide, and the second approach would introduce a nonhuman fragment associated with the risk of an immune response that could severely limit potential therapeutic applications. Thus, what is needed is an in vivo expression system capable of synthesizing sufficient amounts of hydroxyproline residues in the (Gly-Pro-Pro)10 peptide to form a thermally stable triple helical structure, and then drive the formation of a trimeric fusion protein, enabling use of such trimerizing polypeptides both in vitro and in vivo.

The recombinant expression of collagens and hydroxyproline-containing peptides with functional triple-helix conformation requires specific post-translational enzymes, in particular P4H (22) . Procaryotes do not possess any P4H activity. Yeasts and insect cells exhibit insufficient enzyme activity to achieve recombinant collagen expression unless exogenous P4H genes (both {alpha} and β subunits) are introduced simultaneously to form an active {alpha}2β2 tetramer. Previously, we coexpressed minicollagen XXI comprising the extreme C-terminal collagenous (COL1) and noncollagenous (NC1) domains, along with the 2 subunits of human P4H genes in Drosophila S2 cells (26) . Type XXI collagen belongs to a collagen subfamily called fibril-associated collagens with interrupted triple helices (FACITs) (27) . In FACITs, the 2 conserved cysteines, separated by 4 amino acids, are located at the junction of the COL1 and NC1 domains and are responsible for interchain disulfide bonding among the 3 assembled collagen chains (28) . We have shown that the formation of interchain disulfide-bonded minicollagen XXI in Drosophila depends on the hydroxyproline content of collagen chains, suggesting that the folding of the triple helix precedes the formation of the disulfide bonds. Therefore, the driving force for the formation of the triple-helical structure of nonfibrillar FACIT collagens is governed by the COL1 domain, which differs from that of fibrillar collagens using the C-propeptide as trimerization domain. Insufficient prolyl hydroxylation in minicollagen XXI leads to the production of interchain disulfide-bonded dimers and intrachain disulfide-bonded monomers. In this study, we replaced the collagenous domain of minicollagen XXI with a thermally stable short collagen-like peptide (Gly-Pro-Pro)10 as a scaffold template and expressed the collagen-fusions in a mammalian system with sufficient P4H activity to facilitate the adoption of the stable triple-helical structure. Indeed, both erb_scFv-Col and OKT3_scFv-Col were assembled into a trimeric structure and erb_scFv-Col can be further oligomerized into a hexamer, presumably through the interchain disulfide crosslinking between the cysteine residues within the 2 trimers. The oligomerization of collabodies from trimer to hexamer is an intracellular process because the reduced trimeric structure does not assemble into higher-order structures at concentrations in excess of those normally found in the hexameric form of erb_scFv-Col (data not shown).

Based on the antibody competition and CD3 modulation experiments, we demonstrated that the OKT3 collabody can further improve the target CD3 retention, leading to an enhanced degree of TCR modulation. The results suggested that OKT3_scFv-Col is more potent than OKT3 IgG in suppression of T-cell activation. Therefore, the effective dosing of the OKT3 collabody for the immunosuppression of T-cell activation should be lower than that of OKT3. Because OKT3_scFv-Col is devoid of the Fc region of IgG, it should not cause any transient T-cell activation or cytokine release, greatly reducing the effects of the OKT3 first-dose syndrome that is caused by the release of cytokines as a result of transient T-cell activation through the crosslinking of the Fc receptor-positive cells (29) . This new nonmitogenic anti-CD3 format may provide a potent immunosuppressive drug candidate with reduced dosing and toxicity for therapeutic applications.

Collagen has been used in many medical applications, including tissue engineering and drug delivery materials. However, there are concerns about the immune response against fibril collagens, such as type II collagen-induced rheumatoid arthritis. In vivo toxicology studies are currently being conducted to examine any adverse effect derived from the collagen-like motif of the collabody molecules. Because avidity is influenced by both the valence of the antibody and the valence of the antigen, like many other multivalent scaffold fusion proteins, the collabody can only improve its avidity when the targets are present in multimeric forms and, most important, these targets are in proximity to one another so that all 3 scFvs on the collabodies can engage antigen simultaneously. Another concern on the limitation of the collabody strategy (or any multivalent binders) is that when the intrinsic affinity of a giving IgG is high enough, the avidity effect would not be prominent by increasing the binding valency. This can explain why the overall binding strength can be improved 20-fold from the bivalent to trivalent format of a low-affinity erb scFv, whereas the OKT3 collabody can only improve 3-fold in comparison to the OKT3 IgG, which already has an excellent binding avidity of 1.2 nM.

A short triplex-forming collagen-like peptide fusion scaffold was used to form a thermally stable multivalent protein binder to demonstrate herein that collabody is a new platform that enables improvements to the functional affinity of antibodies. More important, fusion of the scFv and the collagen-like scaffold domains in collabody has been proven to bring each domain to fold correctly, without compromising the target-binding activities or the trimeric assembly of the triple helical structure. Presumably, the triplex-forming collagen-like domain can be used as a scaffold for existing or new protein drugs by trimerizing (even oligomerizing through collagen fibrillar stacking or interchain disulfide crosslinking of the trimeric molecules that are similar to those of the defense collagen family) the active protein in a fusion-protein approach, which involves protein hormones, cytokines, lymphokines, growth factors, lectins, enzymes and soluble receptor fragments; or adhesion molecules, such as selectins and integrins.


   ACKNOWLEDGMENTS
 
We thank Tsan-Lin Hu for conducting the pharmacokinetic analysis. This work was funded by the Ministry of Economic Affairs, grant 7301XS5410.


   FOOTNOTES
 
1 These authors contributed equally to this work.

Received for publication April 16, 2008. Accepted for publication June 19, 2008.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Riechmann, L., Clark, M., Waldmann, H., Winter, G. (1988) Reshaping human antibodies for therapy. Nature 332,323-327[CrossRef][Medline]
  2. Queen, C., Schneider, W. P., Selick, H. E., Payne, P. W., Landolfi, N. F., Duncan, J. F., Avdalovic, N. M., Levitt, M., Junghans, R. P., Waldmann, T. A. (1989) A humanized antibody that binds to the interleukin 2 receptor. Proc. Natl. Acad. Sci. U. S. A. 86,10029-10033[Abstract/Free Full Text]
  3. Dubel, S., Breitling, F., Kontermann, R., Schmidt, T., Skerra, A., Little, M. (1995) Bifunctional and multimeric complexes of streptavidin fused to single chain antibodies (scFv). J. Immunol. Methods 178,201-209[CrossRef][Medline]
  4. Rheinnecker, M., Hardt, C., Ilag, L. L., Kufer, P., Gruber, R., Hoess, A., Lupas, A., Rottenberger, C., Pluckthun, A., Pack, P. (1996) Multivalent antibody fragments with high functional affinity for a tumor-associated carbohydrate antigen. J. Immunol. 157,2989-2997[Abstract]
  5. Binz, H. K., Amstutz, P., Kohl, A., Stumpp, M. T., Briand, C., Forrer, P., Grutter, M. G., Pluckthun, A. (2004) High-affinity binders selected from designed ankyrin repeat protein libraries. Nat. Biotechnol. 22,575-582[CrossRef][Medline]
  6. Silverman, J., Liu, Q., Bakker, A., To, W., Duguay, A., Alba, B. M., Smith, R., Rivas, A., Li, P., Le, H., Whitehorn, E., Moore, K. W., Swimmer, C., Perlroth, V., Vogt, M., Kolkman, J., Stemmer, W. P. (2005) Multivalent avimer proteins evolved by exon shuffling of a family of human receptor domains. Nat. Biotechnol. 23,1556-1561[CrossRef][Medline]
  7. Graversen, J. H., Jacobsen, C., Sigurskjold, B. W., Lorentsen, R. H., Moestrup, S. K., Thogersen, H. C., Etzerodt, M. (2000) Mutational analysis of affinity and selectivity of kringle-tetranectin interaction. Grafting novel kringle affinity onto the tetranectin lectin scaffold. J. Biol. Chem. 275,37390-37396[Abstract/Free Full Text]
  8. Berg, R. A., Prockop, D. J. (1973) The thermal transition of a non-hydroxylated form of collagen. Evidence for a role for hydroxyproline in stabilizing the triple-helix of collagen. Biochem. Biophys. Res. Commun. 52,115-120[CrossRef][Medline]
  9. Rosenbloom, J., Harsch, M., Jimenez, S. (1973) Hydroxyproline content determines the denaturation temperature of chick tendon collagen. Arch. Biochem. Biophys. 158,478-484[CrossRef][Medline]
  10. Chopra, R. K., Ananthanarayanan, V. S. (1982) Conformational implications of enzymatic proline hydroxylation in collagen. Proc. Natl. Acad. Sci. U. S. A. 79,7180-7184[Abstract/Free Full Text]
  11. Yang, W., Chan, V. C., Kirkpatrick, A., Ramshaw, J. A., Brodsky, B. (1997) Gly-Pro-Arg confers stability similar to Gly-Pro-Hyp in the collagen triple-helix of host-guest peptides. J. Biol. Chem. 272,28837-28840[Abstract/Free Full Text]
  12. De Wildt, R. M., Mundy, C. R., Gorick, B. D., Tomlinson, I. M. (2000) Antibody arrays for high-throughput screening of antibody-antigen interactions. Nat. Biotechnol. 18,989-994[CrossRef][Medline]
  13. Arakawa, F., Kuroki, M., Kuwahara, M., Senba, T., Ozaki, H., Matsuoka, Y., Misumi, Y., Kanda, H., Watanabe, T. (1996) Cloning and sequencing of the VH and V kappa genes of an anti-CD3 monoclonal antibody, and construction of a mouse/human chimeric antibody. J. Biochem. (Tokyo) 120,657-662[Abstract/Free Full Text]
  14. Cole, M. S., Anasetti, C., Tso, J. Y. (1997) Human IgG2 variants of chimeric anti-CD3 are nonmitogenic to T cells. J. Immunol. 159,3613-3621[Abstract]
  15. Liou, G. G., Tanny, J. C., Kruger, R. G., Walz, T., Moazed, D. (2005) Assembly of the SIR complex and its regulation by O-acetyl-ADP-ribose, a product of NAD-dependent histone deacetylation. Cell 121,515-527[CrossRef][Medline]
  16. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,248-254[CrossRef][Medline]
  17. Adair, J. R., Athwal, D. S., Bodmer, M. W., Bright, S. M., Collins, A. M., Pulito, V. L., Rao, P. E., Reedman, R., Rothermel, A. L., Xu, D. (1994) Humanization of the murine anti-human CD3 monoclonal antibody OKT3. Hum. Antibodies Hybridomas 5,41-47[Medline]
  18. Frank, S., Kammerer, R. A., Mechling, D., Schulthess, T., Landwehr, R., Bann, J., Guo, Y., Lustig, A., Bachinger, H. P., Engel, J. (2001) Stabilization of short collagen-like triple helices by protein engineering. J. Mol. Biol. 308,1081-1089[CrossRef][Medline]
  19. Holler, N., Tardivel, A., Kovacsovics-Bankowski, M., Hertig, S., Gaide, O., Martinon, F., Tinel, A., Deperthes, D., Calderara, S., Schulthess, T., Engel, J., Schneider, P., Tschopp, J. (2003) Two adjacent trimeric Fas ligands are required for Fas signaling and formation of a death-inducing signaling complex. Mol. Cell. Biol. 23,1428-1440[Abstract/Free Full Text]
  20. Hoppe, H. J., Barlow, P. N., Reid, K. B. (1994) A parallel three stranded alpha-helical bundle at the nucleation site of collagen triple-helix formation. FEBS Lett. 344,191-195[CrossRef][Medline]
  21. Bachinger, H. P., Bruckner, P., Timpl, R., Prockop, D. J., Engel, J. (1980) Folding mechanism of the triple helix in type-III collagen and type-III pN-collagen. Role of disulfide bridges and peptide bond isomerization. Eur. J. Biochem. 106,619-632[Medline]
  22. Prockop, D. J., Kivirikko, K. I. (1995) Collagens: molecular biology, diseases, and potentials for therapy. Annu. Rev. Biochem. 64,403-434[CrossRef][Medline]
  23. Engel, J., Chen, H. T., Prockop, D. J., Klump, H. (1977) The triple helix in equilibrium with coil conversion of collagen-like polytripeptides in aqueous and nonaqueous solvents. Comparison of the thermodynamic parameters and the binding of water to (L-Pro-L-Pro-Gly)n and (L-Pro-L-Hyp-Gly)n. Biopolymers 16,601-622[CrossRef][Medline]
  24. Boudko, S., Frank, S., Kammerer, R. A., Stetefeld, J., Schulthess, T., Landwehr, R., Lustig, A., Bachinger, H. P., Engel, J. (2002) Nucleation and propagation of the collagen triple helix in single-chain and trimerized peptides: transition from third to first order kinetics. J. Mol. Biol. 317,459-470[CrossRef][Medline]
  25. Frank, S., Boudko, S., Mizuno, K., Schulthess, T., Engel, J., Bachinger, H. P. (2003) Collagen triple helix formation can be nucleated at either end. J. Biol. Chem. 278,7747-7750[Abstract/Free Full Text]
  26. Li, H. C., Huang, C. C., Chen, S. F., Chou, M. Y. (2005) Assembly of homotrimeric type XXI minicollagen by coexpression of prolyl 4-hydroxylase in stably transfected Drosophila melanogaster S2 cells. Biochem. Biophys. Res. Commun. 336,375-385[CrossRef][Medline]
  27. Chou, M. Y., Li, H. C. (2002) Genomic organization and characterization of the human type XXI collagen (COL21A1) gene. Genomics 79,395-401[CrossRef][Medline]
  28. Mazzorana, M., Cogne, S., Goldschmidt, D., Aubert-Foucher, E. (2001) Collagenous sequence governs the trimeric assembly of collagen XII. J. Biol. Chem. 276,27989-27998[Abstract/Free Full Text]
  29. Li, J., Davis, J., Bracht, M., Carton, J., Armstrong, J., Gao, W., Scallon, B., Fung, R., Emmell, E., Zimmerman, M., Griswold, D. E., Li, L. (2006) Modulation of antigen-specific T cell response by a non-mitogenic anti-CD3 antibody. Int. Immunopharmacol. 6,880-891[CrossRef][Medline]
  30. Kabat, E. A., Wu, T. T., Perry, H. M., Gottesman, K. S., Foeller, C. (1991) Sequences of Proteins of Immunological Interest National Institutes of Health Bethesda, Maryland, USA.
  31. Kivirikko, K. I., Myllyla, R., Pihlajaniemi, T. (1992) Hydroxylation of proline and lysine residues in collagens and other animal and plant proteins. Harding, J. J. Crabbe, M. J. C. eds. Post-Translational Modifications of Proteins ,1-51 CRC Press Boca Raton, FL, USA.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Supplemental Data
Right arrow All Versions of this Article:
fj.08-111484v1
22/11/3795    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fan, C.-Y.
Right arrow Articles by Chou, M.-Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fan, C.-Y.
Right arrow Articles by Chou, M.-Y.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS