HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 39, 30716-30724, September 29, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/39/30716    most recent M004569200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Boutaud, A. Articles by Hudson, B. G. Search for Related Content PubMed PubMed Citation Articles by Boutaud, A. Articles by Hudson, B. G. Social Bookmarking                      What's this?

Type IV Collagen of the Glomerular Basement Membrane

EVIDENCE THAT THE CHAIN SPECIFICITY OF NETWORK ASSEMBLY IS ENCODED BY THE NONCOLLAGENOUS NC1 DOMAINS^*

Ariel Boutaud, Dorin-Bogdan Borza, Olga Bondar, Sripad Gunwar, Kai-Olaf Netzer, Narinder Singh, Yoshifumi Ninomiya^§, Yoshikazu Sado^¶, Milton E. Noelken, and Billy G. Hudson

From the Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66160, the ^§ Department of Molecular Biology and Biochemistry, Okayama University Medical School, Okayama, Japan, and the ^¶ Division of Immunology, Shigei Medical Research Institute, Okayama 701-0202, Japan

Received for publication, May 26, 2000, and in revised form, July 6, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The ultrafiltration function of the glomerular basement membrane (GBM) of the kidney is impaired in genetic and acquired diseases that affect type IV collagen. The GBM is composed of five (1 to 5) of the six chains of type IV collagen, organized into an 1·2(IV) and an 3·4·5(IV) network. In Alport syndrome, mutations in any of the genes encoding the 3(IV), 4(IV), and 5(IV) chains cause the absence of the 3·4·5 network, which leads to progressive renal failure. In the present study, the molecular mechanism underlying the network defect was explored by further characterization of the chain organization and elucidation of the discriminatory interactions that govern network assembly. The existence of the two networks was further established by analysis of the hexameric complex of the noncollagenous (NC1) domains, and the 5 chain was shown to be linked to the 3 and 4 chains by interaction through their respective NC1 domains. The potential recognition function of the NC1 domains in network assembly was investigated by comparing the composition of native NC1 hexamers with hexamers that were dissociated and reconstituted in vitro and with hexamers assembled in vitro from purified 1-5(IV) NC1 monomers. The results showed that NC1 monomers associate to form native-like hexamers characterized by two distinct populations, an 1·2 and 3·4·5 heterohexamer. These findings indicate that the NC1 monomers contain recognition sequences for selection of chains and protomers that are sufficient to encode the assembly of the 1·2 and 3·4·5 networks of GBM. Moreover, hexamer formation from the 3, 4, and 5 NC1 monomers required co-assembly of all three monomers, suggesting that mutations in the NC1 domain in Alport syndrome may disrupt the assembly of the 3·4·5 network by interfering with the assembly of the 3·4·5 NC1 hexamer.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The glomerular basement membrane (GBM)¹ of the kidney is an essential component of the blood filtration barrier. The GBM function is impaired in certain renal diseases, including Alport syndrome, a hereditary form of progressive renal disease caused by mutations in genes encoding type IV collagen; Goodpasture disease, an autoimmune disease characterized by glomerulonephritis and anti-GBM antibodies; and Alport post-transplant nephritis, an anti-GBM disease that develops in some Alport patients after kidney transplants (1). Studies of the molecular basis of these diseases over the past 2 decades have revealed that the common molecular component affected is type IV collagen.

Type IV collagen is the major constituent of basement membranes and consists of a family of six homologous  chains, designated 1-6. Each chain is characterized by a long collagenous domain of ~1400 residues of Gly-Xaa-Yaa repeats, interrupted by ~20 short noncollagenous sequences, and by a noncollagenous (NC1) domain of ~230 residues at the carboxyl terminus. Three  chains assemble into triple-helical molecules that further associate to form supramolecular networks by dimerization at the carboxyl terminus through NC1 domains and by formation of tetramers at the amino terminus (2, 3). In GBM, two networks with distinct chain composition have been identified (4): an 1·2(IV) network and a more cross-linked 3·4·5(IV) network characterized by loops and supercoiling (Fig. 1).

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Schematic diagram representing the two distinct networks of type IV collagen found in GBM. The 1·2(IV) network (top panel) can be solubilized by digestion with pseudolysin at 4 °C. The 3·4·5(IV) network (bottom panel), solubilized by pseudolysin at 25 °C, is cross-linked by disulfide bonds and characterized by twisting and supercoiling (4). The 3·4·5(IV) network is absent in Alport syndrome and, in two types of anti-GBM disease, it is the target of Goodpasture autoantibodies and of Alport alloantibodies (depicted by the Y symbol in the bottom panel).

In the normal glomerular development, the 1·2(IV) network is assembled first in the embryonic glomerulus, but then there is a developmental switch to the synthesis of the 3·4·5(IV) network that forms the GBM of the mature glomerulus in rodents (5), dogs (6), and humans (7). This developmental switch is arrested in Alport syndrome by mutations in any of the genes encoding the 3(IV), 4(IV), or 5(IV) chains. As a result, the GBM in Alport patients is composed of the embryonic 1·2(IV) network rather than the mature 3·4·5(IV) network. The absence of the 3·4·5(IV) network in Alport patients leads to the deterioration of the GBM and progressive loss of renal function over a period of 10-20 years, possibly by rendering the GBM susceptible to proteolysis (7). More than 300 mutations have now been identified in the X-linked form of the disease in the COL4A5 gene encoding the 5(IV) chain (8), and about 30 mutations have been identified in the autosomal recessive form in the COL4A3 and COL4A4 genes, encoding the 3(IV) and 4(IV) chains, respectively (9-11).

How mutations in one (IV) chain lead to the synthesis of a GBM devoid of the whole 3·4·5(IV) network is unknown. The answer to this question is critical, because it would indicate whether novel therapeutic approaches, such as gene therapy, are feasible for the treatment of Alport patients. Our recent finding that the 3, 4 and 5(IV) chains are interlinked by disulfide bonds forming a distinct network (4) suggests that the underlying mechanism may be operative at the level of triple-helix assembly. For example, the 3, 4, and 5(IV) chains may each be required to assemble a triple-helical molecule comprising all three chains. Then a mutation in any of the three chains could lead to defective assembly of triple-helical molecules and thereby prevent network assembly. This mechanism would be similar to the well established paradigm in which mutations in type I collagen lead to osteogenesis imperfecta (12). Understanding the molecular mechanisms underlying the network defect in Alport syndrome requires further knowledge of the organization of 1-5(IV) chains and the discriminatory interactions among them that govern the normal assembly of two distinct networks in GBM.

The specificity for assembly into different networks may be an intrinsic feature of the (IV) chains, encoded in their primary structure. Since folding of type IV collagen monomers occurs from the carboxyl to the amino terminus (13), similar to other collagens (14), a prime candidate for this recognition function is the NC1 domain. In fibrillar collagens, sequences in the C-propeptide analogous to the NC1 domain of type IV collagen were found to direct the selection of chains for the assembly of triple-helical molecules (15). In type X collagen, the NC1 domain plays a critical role in both triple-helix formation and network formation (16), and mutations in this domain prevent the assembly of the type X collagen network, leading to Schmidt's metaphyseal chondrodysplasia (17, 18).

In the present study, the NC1 domain of type IV collagen was investigated to determine whether it possesses discriminatory structural features important for the assembly of the two distinct networks in GBM. This was investigated by comparing the chain composition of native GBM hexamers with that of GBM hexamers reconstituted in vitro and with NC1 hexamers assembled in vitro from purified 1-5(IV) NC1 monomers. The findings indicate that the NC1 monomers contain recognition sequences for selection of chains and monomers that are sufficient to encode the assembly of the 1·2(IV) and 3·4·5(IV) networks of GBM. Moreover, mutations in the NC1 domain of the 3-5(IV) chains in Alport syndrome may disrupt the assembly of the 3·4·5(IV) network by interfering with the assembly of the 3·4·5 NC1 hexamer.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Proteins-- Frozen bovine kidneys, testes, and eye lenses were purchased from Pel-Freeze Biological (Rogers, AR) and stored at 20 °C. Human kidneys not suitable for transplantation were obtained from Midwest Organ Bank and stored frozen at 70 °C. NC1 hexamers were prepared by digestion with bacterial collagenase of human GBM (19) and of bovine GBM (20), STBM (21), and LBM (22), followed by purification on DE-52 and S-300. The 1(IV) and 2(IV) NC1 monomers were separated from the LBM NC1 hexamer by chromatofocusing on an Amersham Pharmacia Biotech Mono P column (23). Bovine 3(IV) NC1 domain was prepared from STBM by HPLC, using a C₁₈ reverse-phase column (201 TP 104, 10 Fm, from Vydac), as described for its GBM counterpart (24, 25). Recombinant human 3(IV), 4(IV), and 5(IV) NC1 monomers (r-3/4/5) were expressed in human kidney 293 cells and purified as described previously (26). For some experiments, the FLAG peptide was removed from the recombinant (IV) NC1 domains by digestion with enterokinase (Invitrogen), following the manufacturer's instructions.

Hexamer Assembly-- Native GBM NC1 hexamers were dissociated by dilution (<50 µg/ml) into a solution of 50 mM formic acid buffered at pH 3.0 with Tris base. Under these conditions, complete dissociation to NC1 monomers and dimers occurred, as verified by HPLC gel filtration using a Bio-Sil TSK250 column (Bio-Rad) with a length of 60 cm. The absence of salt from the buffer was necessary for complete hexamer dissociation. Reassembly of the dissociated NC1 domains was performed by changing the buffer to Tris-buffered saline (50 mM Tris, pH 7.4, 150 mM NaCl) by repeated dilution-concentration cycles in Centricon-10 concentrators (Millipore Corp.). In some reassembly experiments, recombinant NC1 domains were also added to the reaction mixture. After incubating the NC1 domains at a concentration of about 1 mg/ml for 24 h at room temperature, the reaction products were separated according to their molecular weight using gel filtration HPLC chromatography on the Bio-Sil TSK250 column. Quantification of the relative amounts of the various species in the mixture was done by peak area analysis from the HPLC profiles. Hexamer assembly from purified 1-5(IV) NC1 domains was carried out similarly. In all experiments, the ratio of the NC1 domains in the association mixture was kept at 1:1. The samples from hexamer assembly studies containing r-3, r-4, and r-5(IV) NC1 monomers were fractionated on a Superdex-200 FPLC column (Amersham Pharmacia Biotech), which resolved these mixtures as sharper peaks relative to the Bio-Sil column. The isolated NC1 hexamers were subsequently analyzed for composition by immunoprecipitation followed by Western blot, for overall appearance (size and shape) by electron microscopy, and for molecular weight by sedimentation equilibrium ultracentrifugation.

Immunoprecipitation and Western Blot Analysis-- Mouse monoclonal antibodies to 1(IV) NC1 domain (Mab1, formerly known as Mab6) and to 3(IV) NC1 domain (Mab3, formerly known as Mab17) were from Wieslab AB (Lund, Sweden). These antibodies were used for immunoprecipitation because they can interact with native NC1 hexamers (25, 27, 28). Approximately 10 µg of NC1 hexamers, native or reassembled in vitro, were incubated with either Mab1 or with Mab3 (100 µl) for 1 h at room temperature. The immune complexes were collected on protein G-Sepharose (30-µl beads) by gentle mixing for 1 h at room temperature and then solubilized in sample buffer for 10 min at 90 °C. Each sample was divided into six equal aliquots, which were separated by SDS-polyacrylamide gel electrophoresis electrophoresis (29) in 4-22% gradient gels and then transferred to nitrocellulose membranes for Western blot analysis. A panel of six rat monoclonal antibodies (H11, H22, H31, H43, H52, and H63) previously described (30) was used to detect the 1-6 NC1 domains in the Western blots. Membranes were reacted with the monoclonal antibodies to 1-6 NC1 domains (1:500 dilution), incubated with by alkaline phosphatase-conjugated anti-rat IgG (Sigma), and then developed with 4-bromo-5-chloro-3-indolyl phosphate and nitro blue tetrazolium.

Electron Microscopy-- Rotary shadowing electron microscopy was performed as described (31) with modifications. The protein samples (25 µg/ml) were diluted into 60% glycerol, 40% ammonium bicarbonate, nebulized onto mica discs, evacuated in a Bal-Tec BAF500K freeze-etch-replica system to 1.7 × 10⁹ millitorr, coated at an angle (rotation rate 60 rpm) with 0.9 nm of platinum at 120 K, and then backed with 8 nm of carbon. The replicas were examined and photographed in a Philips electron microscope with a 30-µm objective aperture at 80 kV.

FTIR Spectroscopy-- Protein samples were dissolved (15-20 mg/ml) in 100 mM sodium phosphate, pH 7.5. The H₂O was then replaced with D₂O by repeated dilution-concentration cycles in Centricon-10 concentrators. The resulting solution was kept overnight at 4 °C. Infrared spectra were measured at room temperature with a Mattson Sirius-100 Fourier transform infrared spectrometer, using CaF₂ windows and 0.056-mm path length. The cell compartment was purged with nitrogen to minimize the water vapor background. For each sample, 500 individual scans in the range 1000-2000 cm¹ were obtained and averaged. The protein spectra (absorbance versus wave number) were corrected for solvent and water vapor contributions using the SpectraCalc software (Galactic Industries Corp.). The criterion for correct water vapor and solvent subtraction was a featureless (zero absorbance) spectrum from 1800 to 2000 cm¹. The resulting spectra were normalized to a peak value of 1.0. Estimates of secondary structure were obtained by deconvolution of the amide I' band (32) with use of the SpectraCalc software.

Analytical Ultracentrifugation-- Sedimentation equilibrium experiments were performed at 20 °C and different rotor speeds, using a Beckman XL-A analytical ultracentrifuge. Samples were loaded into cells equipped with a six-channel 12-mm path length centerpiece, and the radial protein distribution was determined from the UV absorbance at 280 or 230 nm by consecutive automated radial scans acquired at 0.001-cm intervals. Data were analyzed after samples reached thermodynamic equilibrium, as judged by the absence of systematic deviations in the difference between successive scans taken 12 h apart.

The molar mass was calculated using the equation,

(Eq. 1)

where c(r) represents the protein concentration at distance r from the center of rotation, c(a) is the concentration at the liquid-air meniscus, M is the molecular weight of the protein,  is the partial specific volume of the protein,  is the solvent density,  is the angular velocity, R is the universal gas constant, and T is the temperature. The data were analyzed by nonlinear least-squares methods using the Beckman analysis module running under Origin software (Microcal, Inc.). A partial specific volume of 0.72 cm³ g^-1, calculated from the amino acid composition, was used in the calculation of molecular weight.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Network Organization of (IV) Chains in Human GBM-- The initial discovery of the 3 and 4(IV) chains led to a hypothesis that the 3 and 4(IV) chains existed in a network distinct from that of the classical 1 and 2(IV) chains (33). This was demonstrated by analysis of the NC1 hexamers² that were derived from bovine GBM by collagenase digestion and were characterized by immunoprecipitation with chain-specific monoclonal and polyclonal antibodies (25). These findings were confirmed in subsequent studies (28, 34). In these studies, the presence in the hexamers of the 5 and 6(IV) NC1 domains was not investigated, since the work predated the discovery of the 5 and 6(IV) chains. In our recent study of bovine GBM, in which the collagenous domains were characterized, the five (IV) chains were found to be distributed about an 1·2(IV) and an 3·4·5(IV) network (4). Whether the 3, 4, and 5 were connected to each other via their NC1 domains has not been determined.

In the present study, the linkage of the chains at the NC1 junction within the GBM network was determined by fractionating the NC1 hexamers of varying compositions by immunoprecipitation with monoclonal antibodies (Mab1 and Mab3) to 1 and 3(IV) NC1 domains. Mab1 and Mab3 were used in immunoprecipitation because of their unique ability to interact with native hexamers (21, 25, 27, 28). The chain composition of the hexamer populations was analyzed by Western blotting with a panel of monoclonal antibodies specific for the 1-6(IV) NC1 domains (30). Because different monoclonal antibodies were used for detection of each NC1 domain, the intensity of the bands cannot be used to quantify the relative abundance of distinct (IV) NC1 domain within a NC1 hexamer, but the relative intensities can be used to compare the abundance of the same (IV) NC1 domain among different hexamer populations (the blots of the total NC1 hexamers in the top panel serving as a reference, see following figures).

The analysis of the NC1 hexamers from native human GBM revealed the presence of all six  chains, in both monomer and dimer forms (Fig. 2A). Since the 6(IV) chain is not found in the GBM proper (30), it must have originated from the Bowman's capsule, which contains the 1, 2, 5, and 6(IV) chains. The human GBM hexamers precipitated by Mab1 consisted almost exclusively of the 1 and 2(IV) NC1 domains (Fig. 2B). In contrast, the hexamers precipitated with Mab3 predominantly contained the 3, 4, and 5(IV) NC1 domains, along with smaller amounts of 2(IV) NC1 domain (Fig. 2C). The absence of 6(IV) NC1 domain from the hexamers immunoprecipitated with Mab1 or Mab3 suggests that the 6(IV) NC1 domain is probably associated with the 5(IV) NC1 domain in the basement membrane of the Bowman's capsule. Thus, the predominant NC1 hexamers of human GBM are a heterohexamer composed of 1 and 2(IV) NC1 domains and another one composed of the 3, 4, and 5(IV) NC1 domains. These results unambiguously confirms the existence of at least two distinct networks in human GBM, and they are consistent with the previously established 1·2(IV) and 3·4·5(IV) networks (4), in which monomers containing the respective chains are connected to one another through their NC1 domains (Fig. 1). Moreover, the findings establish unambiguously that the 3, 4, and 5(IV) chains are interlinked with each other via their NC1 domains.

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Analysis of the native NC1 hexamers populations derived from human GBM. Hexamers containing 1(IV) and 3(IV) NC1 domains were separated by immunoprecipitation with monoclonal antibodies Mab1 and Mab3, respectively (left), and then their chain composition was analyzed by Western blotting with the H-series monoclonal antibodies specific for 1-6(IV) NC1 domains (right). The blots show the composition of the total native human GBM NC1 hexamers (A) and of hexamers bound to Mab1 (B) and Mab3 (C). The position of NC1 monomer and dimer bands in the Western blots is indicated by m and d, respectively. Because different monoclonal antibodies were used for detection of each NC1 domain, the intensity of the bands can be compared only for the same NC1 domain in different panels (the blots of the total NC1 hexamer in A serving as a reference).

Assembly of NC1 Hexamers in Vitro from Dissociated Human GBM Hexamers-- Potentially, the NC1 domains possess recognition sequences that specify which NC1 domains associate to form hexamers, which, in turn, would reflect specificity in the selection of chains for monomer assembly and the selection of monomers for network assembly. This supposition was investigated by comparing the composition of native hexamers with in vitro reconstituted hexamers using the experimental approach shown in Fig. 3 (left panel). Complete dissociation into NC1 monomers and dimers was achieved by dilution of hexamers at pH 3.0, as shown by the HPLC gel filtration analysis. Reassembly of the dissociated hexamers was performed overnight at room temperature after changing the buffer to Tris-buffered saline, pH 7.4, and concentrating the mixture of NC1 domains to approximately 1 mg/ml. The reaction mixture was then analyzed by HPLC gel filtration. Under these "associative conditions," hexamers reformed in a proportion of 89%. Only 4% of the starting NC1 domains did not reassemble into hexamers, and the remaining 7% formed aggregates with a molecular weight higher than that of hexamers. The hexamer peak was analyzed by Western blotting (Fig. 3A) and found to contain all chains present in the starting material.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Analysis of NC1 hexamer populations reassociated in vitro from dissociated human GBM hexamers. HPLC gel filtration showed that the native (nat) human GBM hexamers (h) were completely dissociated into monomers (m) and dimers (d) at pH 3 (diss) but were reformed in proportion of 89% after reassociation at neutral pH (reas). Immunoprecipitation and Western blot analysis for 1-6(IV) NC1 domains were performed as described in Fig. 2, using the total reassociated hexamer mixture (A) and the hexamers bound to Mab1 (B) and Mab3 (C). The position of NC1 monomer and dimer bands in the Western blots is indicated by m and d, respectively.

The reassociated hexamers were fractionated by immunoprecipitation with Mab1 and Mab3 to determine whether reconstitution of hexamers was specific and whether the specificity was related to that observed for native hexamer from human GBM. The results revealed that the reassociated hexamer consisted of the same two population as the native hexamers, an 1·2(IV) heterohexamer and a 3·4·5(IV) heterohexamer, which are immunoprecipitated by Mab1 (Fig. 3B) and Mab3 (Fig. 3C), respectively. Mab1 also precipitated small amounts of 3 and 6(IV) NC1 domains that were not observed in the native hexamers (compare Fig. 3B and Fig. 2B). Mab3 also immunoprecipitated small amounts of 1, 2, and 6(IV) NC1 domains from reconstituted hexamers, whereas only 2(IV) NC1 domain was precipitated from the native hexamers. Overall, these results reveal that the two hexamer subpopulations reassembled in vitro have essentially the same composition profile with respect to the 1 and 3(IV) NC1 domains as those from native human GBM. This suggests that the chain specificity for the assembly of the 1·2(IV) and 3·4·5(IV) networks of human GBM is encoded in the 1-5(IV) NC1 domains.

Network Organization of the (IV) Chains in Bovine GBM and Their Reassembly in Vitro-- Due to the limited availability of human NC1 domains for subsequent investigations, studies were also conducted using NC1 domains isolated from bovine GBM. As shown in Fig. 4, A-C, native bovine NC1 hexamers occurred primarily in two subpopulations, an 1·2(IV) heterohexamer and an 3·4·5(IV) heterohexamer, analogous to those of human GBM. For bovine hexamers, Mab1 also immunoprecipitated small amounts of 3, 4, and 5(IV) NC1 domains, which in contrast were not precipitated from human hexamers (compare Figs. 4B and 2B). Likewise, Mab3 immunoprecipitated a small amount of 1(IV) NC1 from bovine hexamers but not from human hexamers (compare Fig. 4C and 2C). Dissociation and reassociation studies were also conducted with bovine NC1 hexamer to determine whether hexamer reassembly was specific. As shown in Fig. 4, D-F, the reassociated hexamers were redistributed about the same two subpopulations as in the native hexamer. These results reveal that bovine GBM contains an 1·2(IV) network and an 3·4·5(IV) network and that the chain specificity for network formation appears to reside in the NC1 domains.

View larger version (52K): [in this window] [in a new window]   Fig. 4.   Analysis of bovine GBM NC1 hexamers in native form (left) and after dissociation and reassociation (right). HPLC gel filtration showed that bovine GBM hexamers (h) were completely dissociated into monomers (m) and dimers (d) at pH 3 but were reformed >85% after reassociation at neutral pH. Hexamers reassociated in vitro were indistinguishable from the native ones by rotary shadowing electron microscopy (the bar represents 100 nm). Immunoprecipitation and Western blot analysis for 1-5(IV) NC1 domains were performed as described in Fig. 2, using the native (A-C) or reassociated hexamers (D-F). The panels show total hexamers (A and D) and hexamers bound to Mab1 (B and E) and Mab3 (C and F). The position of NC1 monomer and dimer bands in the Western blots is indicated by m and d, respectively. The presence of the 6(IV) NC1 domain could not be monitored, because monoclonal antibody H63 did not react with the bovine protein.

Assembly of NC1 Hexamers in Vitro from Purified NC1 Monomers-- The hexamer reassociation studies described above demonstrate that mixtures of NC1 monomers and dimers can reassociate into hexamers with two distinct compositions, analogous to the native hexamers. The capacity to form hexamers of the 1-5(IV) NC1 monomers, alone or in various combinations, was next evaluated along with the specificity of this interaction. The 1 and 2(IV) NC1 monomers, known to be biologically active for hexamer reassembly (23), were isolated from bovine LBM and used as reference for establishing the native properties (conformation and ability to form hexamers) of 3, 4, and 5(IV) NC1 monomers prepared by recombinant technology (see below).

Native 1 and 2(IV) NC1 monomers were purified from LBM by chromatofocusing. This was possible because (a) LBM has a simple composition, consisting almost exclusively (>95%) of the 1(IV) and 2(IV) chains (22); (b) the LBM hexamers contain mostly (>90%) NC1 monomers (35) and can be dissociated into subunits without denaturation; and (c) the isoelectric points of 1 and 2 NC1 domains differ by about 2 pH units, so they are amenable to separation by chromatofocusing (23) (Fig. 5, A-C). The complete separation of 1 and 2(IV) NC1 monomers was monitored by Western blotting with monoclonal antibodies to the 1(IV) (Fig. 5B) and 2(IV) (Fig. 5C) NC1 domains. The protein fractions were pooled as indicated and further used in reassociation experiments. The 1(IV) NC1 monomer could form a homohexamer (Fig. 5D), whereas the 2(IV) NC1 monomer could not (Fig. 5F). Mixing of 1 and 2(IV) NC1 monomers resulted in the formation of heterohexamers (Fig. 5E), as demonstrated by co-precipitation from the hexamer fraction of both 1 and 2(IV) NC1 domains by monoclonal antibody Mab1. Thus, the 1 and 2(IV) NC1 monomers have the required specificity of interaction to form the 1·2(IV) heterohexamer found in human and bovine GBM (Figs. 2B and 4B).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Purification and in vitro reassociation of 1 and 2(IV) NC1 monomers. The 1 and 2(IV) NC1 domains were isolated from bovine LBM NC1 hexamers by chromatofocusing on a MonoP-Sepharose column (A). The protein fractions were analyzed by Western blot with monoclonal antibodies to the 1(IV) (B) and 2(IV) NC1 domains (C) and then pooled as shown by the horizontal bars. The ability to reassociate in vitro of purified 1(IV) NC1 monomers (D), 2(IV) NC1 monomers (F), and of a mixture of equal amounts of 1 and 2(IV) NC1 monomers (E) was then assessed. Hexamer formation was monitored by HPLC gel filtration chromatography (h indicates the position of the NC1 hexamer peak, and m indicates that of NC1 monomers). The hexamer fraction formed from the mixture of 1 and 2 NC1 domains was immunoprecipitated with Mab1 antibodies and then analyzed by Western blot to confirm that both 1 and 2(IV) NC1 domains were present in the same complex (E, inset).

Unlike the 1 and 2(IV) domains, the 3, 4, and 5(IV) NC1 domains occur predominantly as disulfide-linked dimers and in complex mixtures in native basement membranes. As such, the 3-5(IV) NC1 monomers cannot be isolated under nondenaturing conditions, preserving their native state. To circumvent this problem, the 3, 4, and 5(IV) NC1 monomers were prepared by recombinant technology. The recombinant (IV) NC1 monomers were tagged with a FLAG epitope for purification and expressed in HEK 293 cells as described previously (26). To assess the conformation of the recombinant proteins relative to the native state, their secondary structures were compared with those of native 1 and 2 monomers from LBM, based upon analysis by FTIR spectroscopy in the amide I region (1620-1700 cm¹ wave number region). This region comprises up to 10 component peaks characteristic of the secondary structure of a protein (32). The FTIR spectra of 1 and 2(IV) NC1 monomers were similar to each other and to that of the parent LBM hexamer (Fig. 6A). This indicated that the 1 and 2(IV) NC1 monomers had similar secondary structures. Moreover, the secondary structure was not significantly affected by association of NC1 monomers into hexamers. However, the FTIR spectrum of the 3(IV) NC1 domain purified from STBM (the richest natural source of 3) by reverse-phase HPLC was markedly different from those of 1 and 2(IV) NC1 monomers, of LBM, and even of the parent STBM, indicating a different secondary structure (Fig. 6B). Nevertheless, human r-3(IV) NC1 monomer had a FTIR spectrum similar to that of native NC1 hexamers and monomers. These results indicate that r-3(IV) NC1 monomer possesses a secondary structure similar to that of the native 3(IV) NC1 domain. Moreover, we have previously shown that the r-3(IV) NC1 domain adopts a conformation indistinguishable from that of the native 3(IV) NC1 domain on the basis of reactivity with various Goodpasture autoantibody subpopulations and with the monoclonal antibody Mab3, all of which recognize conformational epitopes (36, 37). Thus, recombinant proteins expressed in mammalian HEK 293 cells are suitable for in vitro hexamer assembly studies.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Analysis by FTIR spectroscopy of the secondary structure of native NC1 hexamers and of purified NC1 monomers. Top panel, the FTIR spectra of LBM hexamers (continuous line) and of 1(IV) (dashes) and 2(IV) (dots) NC1 monomers isolated by chromatofocusing. Bottom panel, the FTIR spectra of LBM hexamers (dots), STBM hexamers (continuous line), 3(IV) NC1 monomer purified from STBM by RP-HPLC (dashes), and r-3(IV) NC1 monomer (dashes and dots) expressed in HEK 293 cells.

The biological activity of the r-(IV) NC1 monomers was also assessed from their ability to form hybrid hexamers when mixed with dissociated native hexamers, as depicted in Fig. 7A for r-3(IV). After reassociation, HPLC gel filtration (Fig. 7B) was used to separate the hexamer fraction (Fig. 7C), which was then analyzed by Western blotting (Fig. 7D). The presence of the FLAG tag in the hexamers formed in vitro demonstrated the ability of r-(IV) NC1 domains to associate. Similar experiments using r-4_F and r-5_F NC1 monomers also showed them to be incorporate into the reassembled hexamers. In control experiments, r-(IV) NC1 monomers incubated with native hexamers under nondissociating conditions were not incorporated into hexamers, ruling out nonspecific interactions between the r-(IV) NC1 monomers and the native hexamers (data not shown).

View larger version (46K): [in this window] [in a new window]   Fig. 7.   Evaluation of the ability of r-(IV) NC1 monomers to form hexamers. A, bovine GBM hexamers were dissociated at pH 3 (filled ovals), combined with r-3F (open ovals), and then allowed to reform hexamers in vitro. B, the reassociated NC1 hexamers (h) were separated from NC1 monomers (m) and smaller complexes by HPLC gel filtration. The arrow indicates the elution time of r-3(IV) NC1 monomers. C, the analysis of reassociated NC1 hexamers by rotary shadowing electron microscopy. D, the incorporation of r-3(IV) NC1 monomers into hexamers was shown by Western blot of the isolated hexamer fraction with monoclonal antibodies to 1-5(IV) NC1 domains (1-5) and to the FLAG peptide (FL). The position of NC1 monomer and dimer bands in the Western blots is indicated by m and d, respectively, whereas f shows the position of the FLAG-tagged r-(IV) NC1 domains. Similar results were obtained using other r-(IV) NC1 domains (r-4F, r-5F) or other hexamers (human GBM, bovine STBM).

In Vitro Assembly of Recombinant 3, 4, and 5(IV) NC1 Monomers and Characterization of the r-3·4·5 Heterohexamer-- The ability of r-3, r-4, and r-5(IV) NC1 monomers to form hexamers, alone or in various combinations, was tested by mixing them under associative conditions. None of the 3, 4, and 5(IV) NC1 monomers alone, nor any binary combination of the three monomers (i.e., 3 plus 4, 3 plus 5, and 4 plus 5) yielded hexamers (Fig. 8, B-G). For certain combinations of NC1 monomers, 3 plus 4 and 3 plus 5, the HPLC gel filtration profiles indicated the formation of complexes, but these were clearly smaller than hexamer size (most likely dimers or trimers), as compared with the native GBM hexamer standard (Fig. 8A) or with the molecular weight markers. In contrast, when all three NC1 monomers (3, 4, and 5) were mixed together, NC1 hexamers did form, along with smaller complexes of NC1 monomers (Fig. 8H).

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Analysis of hexamer formation from r-3, r-4, and r-5(IV) NC1 domains. The r-3, r-4, and r-5(IV) NC1 monomers were incubated under associative conditions individually (B-D), in combinations of two (E-G), and all three together (H), and their association was monitored by HPLC gel filtration on a Superdex-200 column. Native GBM hexamers were also run on the same column as control (A). The elution time of NC1 monomers and hexamers is indicated by m and h, respectively, whereas v shows the position of aggregated material that elutes in the void volume (molecular weight more than 800,000).

The kinetics of the NC1 hexamer formation from r-3, r-4, and r-5(IV) NC1 monomers was monitored by HPLC (Fig. 9A). The proportion of hexamers in the reaction mixture slowly increased over a 24-h period until it reached 50% and then remained constant (Fig. 9B). Smaller complexes were formed in large amounts as early as 1 h, and they are presumably intermediates in the formation of the NC1 hexamer (dimers or trimers), because their relative abundance decreased as hexamers formed. However, a more detailed kinetic analysis of the intermediates could not be performed because this fraction was not well resolved from monomers. The fast formation of intermediates followed by slower association to hexamers may correspond to two distinct steps that occur in vivo: the intracellular formation of triple-helical monomers, followed by a slower process of monomer dimerization in the extracellular space.

View larger version (38K): [in this window] [in a new window]   Fig. 9.   A, time course of the association of r-3, r-4, and r-5(IV) NC1 domains in vitro. The elution time of NC1 monomers and hexamers is indicated by m and h, respectively. B, the kinetics of hexamer formation in vitro from r-3, r-4, and r-5(IV) NC1 monomers. C, the above hexamers were separated by HPLC gel filtration, and their composition was analyzed by immunoprecipitation with monoclonal antibody Mab3. D, analysis by rotary shadowing electron microscopy of the 3·4·5 hexamers formed in vitro (left) and of r-3(IV) NC1 monomers (right). The bar represents 100 nm.

The reassembled r-3·4·5(IV) NC1 heterohexamer was of particular interest because it was analogous to the 3·4·5(IV) heterohexamer isolated from human GBM (cf. Fig. 1C). Therefore, it was further analyzed for chain composition by immunoprecipitation and Western blot, for size and shape by electron microscopy, and for molecular weight by sedimentation equilibrium ultracentrifugation. The composition of the heterohexamer was verified by precipitation with monoclonal antibody Mab3. As anticipated, all three NC1 monomers were found in the precipitate (Fig. 9C), establishing that the reassembled hexamer was indeed composed of r-3, r-4, and r-5(IV) NC1 monomers. By electron microscopy, these hexamers also appeared as homogenous particles between 12 and 15 mm in diameter (Fig. 9D), similar to native GBM hexamers (cf. Fig. 4D). By sedimentation equilibrium, the r-3·4·5 hexamers exhibited a molecular mass of 160 kDa, in good agreement with the values ranging from 170 to 174 kDa reported for hexamers from human placenta, mouse tumor, and bovine aorta (38) and a value of 153 kDa for the native hexamers determined in the present study.

In Vitro Assembly of NC1 Hexamers from 1-5(IV) NC1 Monomers-- Having demonstrated the formation of 1·2 hexamers and of 3·4·5 hexamers from the respective NC1 monomers, it remained to be determined whether these hexamers would still form specifically in the presence of all five NC1 monomers (Fig. 10, left). The results most closely resembled those obtained with reassociated bovine hexamers, showing again a clear preference of NC1 monomers to assemble into the two predominant types of hexamers, i.e. 1·2(IV) and 3·4·5(IV) heterohexamers (Fig. 10, right). Comparing semiquantitatively the intensity of bands corresponding to the same NC1 domain in different hexamer populations (Fig. 10, A-C), it was estimated that the 1, 4, and 5 were most specific in hexamer formation (90-95%), followed by 3 (80-85%), while 2 was relatively promiscuous, being equally incorporated in the hexamers precipitated by Mab1 and Mab3. While this may have been caused by a relative excess of purified 2 in vitro (equal amounts of domains were used for reassembly in vitro, but the stoichiometry of 1:2 is 2:1 in vivo), even in the native NC1 hexamers the 2(IV) NC1 domain was found not only in the 1·2(IV) heterohexamers but also accompanying in small amounts the 3·4·5(IV) heterohexamers (Figs. 2 and 4). This indicates that 2(IV) chain can be incorporated in the 3·4·5(IV) network in the absence of the 1(IV) chain. That the specificity for hexamer formation in vitro was somewhat lower than that observed in the native networks of GBM may reflect constraints of interaction imposed in vivo by the sequential process of chain selection for monomer formation in the endoplasmic reticulum, followed by association of monomers to form networks in the extracellular milieu.

View larger version (33K): [in this window] [in a new window]   Fig. 10.   Analysis of NC1 hexamers formed in vitro from purified 1-5(IV) NC1 domains. Equal amounts of 1, 2, r-3, r-4, and r-5(IV) NC1 monomers were mixed and incubated under associative conditions (top left). The reassociated hexamers h were separated from non-associated monomers m by HPLC gel filtration (bottom left), and their composition was analyzed by immunoprecipitation followed by Western blotting, as described in the legend to Fig. 2 (right). The blots show the composition of the total hexamers (A) and of hexamers bound to Mab1 (B) and Mab3 (C). The position of NC1 monomer and dimer bands in the Western blots is indicated by m and d, respectively, whereas f shows the position of the FLAG-tagged r-(IV) NC1 domains.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The chain-specific assembly of type IV collagen triple-helical molecules and networks poses a formidable molecular recognition problem. While other types of collagen consist of one, two, or at most three chains, type IV collagen consists of six homologous but genetically distinct chains. Random combination of the six chains allows for 56 different triple-helical isoforms, which can further associate to form networks in hundreds of different ways.³ Studies to date suggest that only a few kinds of isoforms exist, which form a limited number of chain-specific supramolecular networks in basement membranes. In GBM, for instance, a type IV collagen molecule composed of the 1(IV) and 2(IV) chains forms the 1·2 network, and molecules composed of the 3(IV), 4(IV), and 5(IV) chains form the 3·4·5 network. In the present study, the type IV collagen networks of GBM were further investigated to characterize the organization of the 5(IV) chain in relation to the other chains. The results unambiguously established that the NC1 domain of the 5(IV) chain is interlinked with the NC1 domains of the 3(IV) and 4(IV) chains forming a NC1 hexamer at the interface of two adjoining monomers. How the 3(IV), 4(IV), and 5(IV) chains are arranged within the triple-helical monomers remains unknown. Presumably, they form a single triple-helical type IV collagen molecule that self-associates to form the 3·4·5(IV) network (Fig. 1). Alternatively, the 5(IV) chain could occur in a separate triple-helical molecule that interacts with an 3·4(IV) molecule through the NC1 domains.

What directs the assembly of type IV collagen into distinct chain-specific networks? The existence of the 1·2(IV) and 3·4·5(IV) networks in GBM might be due to the spatial and/or temporal segregation of the synthesis of various chains. Alternatively, the NC1 domains may encode discriminatory interactions that govern the selection of chains for monomer assembly and the selection of protomers for connection of NC1 domains for network assembly. We have previously shown that 1(IV) and 2(IV) NC1 monomers from LBM hexamer can reassemble into NC1 hexamers in vitro, suggesting that the NC1 domains have a recognition function for the formation of triple-helical type IV collagen molecules and their assembly into networks (23). In contrast, a study of 1(IV) and 2(IV) NC1 domains from bovine placement basement membrane (39) reported that NC1 monomers interacted only weakly and that NC1 dimers were required for hexamer assembly. The reason for this disparity is unclear.

The potential recognition function of the NC1 domains in protomer and network assembly was further investigated in the present study. This was accomplished by comparing the composition of native NC1 hexamers of GBM, with respect to the chain origin of NC1 monomers and dimers, with NC1 hexamers that were dissociated and then reconstituted in vitro and with NC1 hexamers assembled in vitro from purified 1-5 NC1 monomers. The use of recombinant techniques to produce 3, 4, and 5(IV) NC1 monomers was crucial, because these NC1 monomers are not amenable to isolation from native sources under nondenaturing conditions. In vivo, these NC1 domains either exist in low abundance or occur in complex mixtures, consisting of a mixture of monomers and disulfide-linked dimers. In contrast, the 1 and 2 NC1 monomers are the major constituents of the LBM hexamer, from where they can be separated under nondenaturing conditions for in vitro assembly studies. The results revealed that NC1 domains associate in vitro to form native-like NC1 hexamers of mainly two distinct populations: an 1·2(IV) heterohexamer and an 3·4·5(IV) heterohexamer. Moreover, the capacity to form specific hexamers resides in the monomer structure. These findings demonstrate that the NC1 monomers contain molecular recognition sequences that are sufficient to encode the assembly of the 1·2(IV) and 3·4·5(IV) networks of GBM. Two types of recognition sequences are predicted: one that specifies intraprotomer interactions in the selection of chains for protomer assembly and another that specifies interprotomer interactions in the selection of protomers for network assembly.

The assembly of the 3·4·5(IV) hexamer is of particular interest in relation to the pathogenic mechanisms underlying Alport syndrome. Formation of hexamers in vitro from r-3, r-