PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Mol Biol. Author manuscript; available in PMC 2010 July 10.
Published in final edited form as:
PMCID: PMC2749300
NIHMSID: NIHMS117125

R992C (p.R1192C)Substitution in Collagen II Alters the Structure of Mutant Molecules and Induces the Unfolded Protein Response

Summary

We investigated the molecular bases of spondyloepiphyseal dysplasia (SED) associated with the R992C (p.R1192C) substitution in collagen II. At the protein level we analyzed the structure and integrity of mutant molecules, and at the cellular level we specifically studied the effects of the presence of the R992C collagen II on the biological processes taking place in host cells. Our studies demonstrated that mutant collagen II molecules were characterized by altered electrophoretic mobility, relatively low thermostability, the presence of atypical disulfide bonds, and slow rates of secretion into the extracellular space. Analyses of cellular responses to the presence of the mutant molecules showed that excessive accumulation of thermolabile collagen II was associated with the activation of an “unfolded protein response” and an increase in apoptosis of host cells. Collectively, these data suggest that molecular mechanisms of SED may be driven not only by structural changes in the architecture of extracellular collagenous matrices, but also by intracellular processes activated by the presence of mutant collagen II molecules.

Keywords: collagen II mutations, unfolded protein response, ER stress, extracellular matrix, apoptosis

Introduction

Collagen II, a main structural protein of cartilaginous tissues, is first synthesized as procollagen characterized by the presence of an extended triple-helical domain flanked by the globular N-terminal and C-terminal propeptides1. Biosynthesis of procollagen molecules is a complex process that involves posttranslational modifications of the nascent pro-α chains. These modifications include hydroxylation of proline residues and lysine residues by prolyl-4 hydroxylase (P4H) and lysyl hydroxylase (LH), respectively. Moreover, the posttranslational modifications and subsequent formation of the triple-helical structure are controlled by a group of endoplasmic reticulum (ER)-resident protein chaperones that includes heat shock protein 47 (HSP47), immunoglobulin-binding protein (BiP), and protein disulfide isomerase (PDI), an enzyme that also acts as the β-subunit of the αα/ββ tetramer that constitutes functional P4H2,3,4,5. In addition, by binding to the individual pro-α1(II) chains and preventing their premature association, P4H may also act as a chaperone during biosynthesis of procollagen molecules6. Following biosynthesis of individual pro-α1(II) chains, their posttranslational modifications, and folding into triple-helical structure, procollagen molecules are released from the ER and directed toward secretory pathways.

A critical triple-helical domain of collagen II consists of α1(II) chains formed by about 330 G-X-Y triplets in which the “X” position is frequently occupied by a proline residue and the “Y” position by a hydroxyproline residue. The importance of correct conformation of the triple-helical region for the proper structure and function of skeletal tissues is clearly evidenced by a number of aberrations of these tissues caused by mutations in COL2A1 (OMIM# 120140). In particular, single amino acid substitutions in collagen II are responsible for a broad spectrum of chondrodysplasia phenotypes, generally described as spondyloepiphyseal dysplasia (SED), whose main feature is abnormalities of skeletal development1,7,8,9,10,11,12,13. The most frequent amino acid substitutions in collagen II occur at the “G” position of the G-X-Y triplets but substitutions in the “Y” positions were also described. For instance, a G853E (p.G1053E) substitution was found in a patient with a lethal form of SED, whereas cysteine substitutions for arginine residues were found in the “Y” positions 75 (p.R275C), 519 (p.R719C), and 789 (p.R989C) in patients with a mild form of SED, a mild form of osteoarthritis, and a severe form of congenital SED, respectively9,13,14,15. In addition to mutations found in the “G” and “Y” positions, relatively rare mutations in the “X” positions of the G-X-Y triplets were also reported in patients with Stickler syndrome, a disease characterized by myopia, cataracts, retinal detachment, hearing loss, and mild SED13,16,17.

The molecular bases of tissue-level aberrations in patients with mutations in COL2A1 are not fully understood. It has been suggested that a key problem caused by amino acid substitutions in collagen II is that the mutant collagens cannot fulfill their structural functions. Studies demonstrated that a decrease in the overall content of collagen II and its inability to assemble into correct fibrils are the main factors causing such alterations of structural functions13,18,19,20,21. In addition to affecting extracellular functions, it has been suggested that some mutations in collagen II cause excessive intracellular accumulation of misfolded mutant molecules and that such accumulation can lead to apoptosis of cells harboring mutant molecules22,23.

Here, we studied the molecular and the cellular-level consequences of the R992C (p.R1192C) substitution in the “X” position of a G-X-Y triplet of collagen II. This mutation (p.R1147C) was originally described in mice, referred to as spondyloepiphyseal dysplasia congenita or sedc mice in which skeletal abnormalities were similar to those seen in SED patients24. Employing a system for expressing recombinant collagen II, we demonstrated that the R992C collagen II is characterized by aberrant electrophoretic mobility, relatively low thermostability, and the presence of intramolecular disulfide bonds. Moreover, we showed that the expression of the R992C collagen II is associated with ER stress and increased apoptosis of cells harboring this collagen mutant.

Results presented here contribute to understanding the molecular basis of mutation-triggered pathological changes seen at the level of skeletal tissues and suggest that mutations in collagen II that interfere with correct folding of the collagen triple helix are likely to act not only by disturbing the extracellular collagenous framework but also by altering intracellular processes.

Results

Analysis of R992C mutant

To analyze mutant collagen molecules, we produced the green fluorescence protein (GFP)-tagged wild type procollagen II (WT Pro-GFP) and R992C procollagen II (R992C Pro-GFP) in HT-1080 cells, as described25. Electrophoresis of intact GFP-tagged procollagen molecules carried out in 6% polyacrylamide gels demonstrated that the electrophoretic mobility of the R992C Pro-GFP was slower than that of the WT Pro-GFP (Fig. 1). Similar differences in the electrophoretic mobilities were seen in pepsin-treated variants in which non-triple-helical propeptides were enzymatically degraded (Fig. 1). In addition to the intact α1(II) mutant chains, their degraded forms were inconsistently seen in different protein preparations (Fig. 1). The treatment of collagen II variants with glycosidases demonstrated that slower migration of the R992C mutant was not caused by overglycosylation of mutant collagen α1(II) chains. In this experiment only a minor shift in the migration of glycosidase-treated wild type and mutant collagens was observed (Fig. 1). Such a minor shift in migration indicates relatively small levels of glycosylation of the α1(II) chains, an observation consistent with our previous report 25. Since no change in the electrophoretic migration of the R992C collagen treated with various denaturing agents was observed, we excluded atypical aggregation as the reason for the slow electrophoretic mobility of this protein (data not shown). Analysis of collagen chains separated in non-reducing conditions demonstrated that about 6% of α1(II) chains were covalently linked by disulfide bonds (Fig. 1). The identity of disulfide-linked chains was confirmed by separating α1(II) monomers and dimers in nonreducing conditions followed by electrophoresis of separated chains in the reducing conditions (Fig. 2).

Figure 1
Analysis of intact and processed R992C Pro-GFP.
Figure 2
Analysis of disulfide bonds formed between α1 chains of the R992C mutant.

Analyses of thermostability of the R992C mutant

Brief protease digestion of pepsin-treated collagen II demonstrated that the thermostability of the R992C mutant was lower than that of the control (Fig. 3). Specifically, we measured the temperatures at which the triple helical structure of analyzed collagens was preserved in 75% (Tm75), 50% (Tm50), and 25% (Tm25) of maximum. We determined that Tm75, Tm50, and Tm25 for the R992C were 36.5°C, 39.9°C, and 42.0°C, respectively. Corresponding Tm75, Tm50, and Tm25 values for the WT collagen II were 40.0°C, 41.2°C, and 42.2°C, respectively. In contrast to the wild type collagen II, protease digestion of the R992 mutant at about 36°C resulted in the generation of intermediate stable products (Fig. 3).

Figure 3
Assays of the thermostability of the R992C mutant.

Computer modeling of the R992C mutation site

To study the structural basis of formation of disulfide bonds we performed computer modeling of the region encompassing the C992 mutation site (Fig. 4). In this study the modeled collagen fragment consisted of three mutant α1(II) chains. Such a model provides good representation of the homotrimeric R992C collagen II produced in HT-1080 cells and in homozygotic mutant mice, in which the R992C mutation was associated with severe SED24. Analysis of the model revealed that sulfhydryl groups of cysteine residues present in two α1(II) chains were spatially oriented in a way that enables formation of the disulfide bonds. Moreover, the model revealed that the sulfhydryl group of the free cysteine residue was exposed on the surface of a molecule, thereby rendering it available for formation of intermolecular disulfide bonds. In addition, analysis of molecular surfaces demonstrated that the cysteine for arginine substitution decreased the electrostatic potential (EP) of a region encompassing the mutation site. Specifically, the EP values (expressed in kcal/mol.e) for the region with a molecular surface area of about 1000 Å2 and encompassing the R992 residue ranged from 117.7 to 233.4. In contrast, the corresponding EP values for the analogous region encompassing the C992 site ranged from 28.2 to 127.2.

Figure 4
A computer model of the region encompassing the R992C mutation site.

Unconditional and Tet-regulated expression of procollagen II variants in SW-1353 cells

We generated a system for unconditional and tetracycline (Tet)-regulated expression of procollagen II variants. Procollagen II variants expressed unconditionally were denoted as “Pro-GFP” whereas those expressed in the Tet-regulated fashion, as “Pro-GFPTet”. To account for possible differences in measured parameters, due to clonal variations in expression of the R992C mutant, two independent R992C Pro-GFP and two R992C Pro-GFPTet clones were analyzed in more detail. The biosynthesis of procollagen II variants expressed in SW-1353 cells was characterized by Western blot and RT-PCR assays. As indicated in Fig. 5, procollagen II was efficiently expressed in both systems. Moreover, in the Tet-regulated system expression of procollagen II was efficiently expressed only in the presence of doxycycline (Dox). In one of the R992C Pro-GFPTet clones (clone #2) there was a detectable baseline-level expression of the R992C Pro-GFPTet mRNA. Such a baseline expression is expected due to nonspecific activation of a Tet-regulated cytomegalovirus (CMV) promoter26. As indicated in Fig. 5 expression of procollagen variants was comparable at both the mRNA and protein levels.

Figure 5
Constant production of the Pro-GFP variants and Dox-dependent production of Pro-GFPTet variants by SW-1353 cells.

In addition to analysis of the intracellular pool of the R992C collagen II, we also studied kinetics of secretion of this mutant in comparison to WT control (Fig. 6). Assays of Pro-GFP from two different clones harboring the R992C collagen II demonstrated that the secretion rate of this mutant, represented by the slope of a fitted curve, was slower in comparison to WT control. In particular, after six hours from the beginning of the experiment the relative amount of the R992 Pro-GFP present in the media was only about 40% of WT Pro-GFP secreted from the same number of cells. In all analyses of the secretion rates, the R2 values ranged from 0.9 to 1.0, indicating a satisfactory curve fit.

Figure 6
Kinetics of secretion of procollagen II variants from SW-1353 cells.

Apoptotic markers in cells expressing R992C mutant

Cellular apoptosis was assayed by TdT-mediated dUTP nick-end labeling (TUNEL) (Fig. 7) and measurements of the relative content of cleaved poly (ADP-ribose) polymerase (PARP) (Fig. 8). In the TUNEL assays the apoptotic index was defined as the percentage of the TUNEL-positive cells out of the total number of 4′,6-diamidino-2-phenylindole (DAPI)-stained nuclei. The R992C Pro-GFP variant had a higher apoptotic index, corresponding to 4.2%±0.5%, than the WT Pro-GFP with an apoptotic index of 2.2%±0.2%. Statistical analysis with the Student’s t test revealed that the difference between mean values for two groups was statistically significant (P=0.0042) (Fig. 7). To confirm that DNA fragmentation was a direct result of the presence of the R992C mutant, we employed cells conditionally expressing the R992C Pro-GFPTet variant. In the Tet-ON condition, in the presence of the R992C mutant, the apoptotic index (4.8%±0.4%) was significantly higher (P=0.0037) than in cells cultured in the Tet-OFF condition (3.2% ±0.3%). In contrast, regardless of Tet-ON or Tet-OFF conditions, apoptotic indexes for cells engineered to express the WT Pro-GFPTet did not show any significant difference (P=0.4744).

Figure 7
TUNEL assays of apoptosis in cells expressing type II collagen mutants.
Figure 8
Assays of cleaved PARP in cells expressing Pro-GFP and Pro-GFPTet variants.

Multiple analyses of two different clones expressing the R992C Pro-GFP variant and two different clones expressing the R992C Pro-GFPTet variant revealed that the relative amount of cleaved PARP increased in cells harboring the R992C mutant (Fig. 8). Specifically, in the cells harboring the R992C Pro-GFP variant, the relative amount of cleaved PARP increased 2.2-fold in comparison to control (P=0.009). Moreover, in comparison to control, the relative amount of cleaved PARP after induction with Dox increased 2.8-fold (P=0.004). In contrast, in cells expressing the WT Pro-GFP adding Dox did not increase the relative amount of cleaved PARP (P=0.756).

Measurements of selected indicators of ER stress

We employed Western blot assays to analyze selected indicators of ER stress. As shown in Fig. 9, the relative amount of PDI in cells expressing the R992C Pro-GFP variants increased 2.3-fold in comparison to control (P=0.032). Analysis of Pro-GFPTet system also revealed that the Dox-induced R992C Pro-GFPTet variant increased the relative amount of PDI 1.9-fold (P=0.0004), while Dox-induced WT Pro-GFPTet did not show any significant difference (P=0.839). Western blotting of BiP also revealed a similar trend, although the 2.4-fold increase in the relative amount of BiP in the R992C Pro-GFP variant was not statistically significant (P=0.215) (Fig. 9). However, the 3.2-fold increase in the BiP content in Dox-stimulated cells expressing the R992C Pro-GFPTet was statistically significant (P=0.003).

Figure 9
Assays of chaperones in cells expressing Pro-GFP and Pro-GFPTet variants.

Analysis of LH and P4H

We studied whether the presence of the mutant α1(II) chains could be associated with increased expression of LH and P4H. Our analyses revealed that the relative amounts of LH (Fig. 9) and the α-subunit of P4H (P4Hα; not shown) in cells expressing the R992C mutant were similar to that of control.

In situ colocalization of BiP, PDI, and P4H with the R992C Pro-GFP

To examine the potential interaction of selected chaperones with mutant collagen II in situ, we employed fluorescence microscopy to analyze colocalization of BiP and PDI with WT Pro-GFP and R992C Pro-GFP variants (Fig. 10). Considering that PDI functions not only independently as protein disulfide-isomerase but also acts as the β-subunit of P4H, we studied the colocalization of the α-subunit of this enzyme with Pro-GFP variants. These studies were critical for determining whether in response to the presence of mutant collagen II, PDI interacted with Pro-GFP variants independently or as a member of the αα/ββ P4H complex. To describe colocalization of BiP/Pro-GFP, PDI/Pro-GFP and P4Hα/Pro-GFP we used the overlap coefficient (R), a value that ranges from 0 and 1 and is not sensitive to variations in pixel intensities. We demonstrated that in comparison to the cells expressing WT Pro-GFP (R=0.61±0.015; N=66) the overlap coefficient for PDI/Pro-GFP colocalization was higher in cells expressing the R992C Pro-GFP (R=0.72±0.010; N=69). Moreover, we analyzed BiP/Pro-GFP colocalization, and our study determined that in comparison to cells expressing WT Pro-GFP (R=0.28±0.012; N=40) the value for the overlap coefficient decreased in cells expressing R992C Pro-GFP (R=0.14±0.009; N=43). Statistical analysis determined that the differences in the R values for PDI/Pro-GFP and BiP/Pro-GFP colocalization were statistically significant with P<0.0001. In addition, we determined that, in comparison to PDI and BiP, the overlap coefficient for P4Hα/Pro-GFP colocalization was relatively low for cells expressing WT Pro-GFP (R=0.16±0.009; N=39), and it did not significantly change (P=0.8) in cells expressing the R992C Pro-GFP variant (R=0.15±0.008; N=30).

Figure 10
Colocalization of Pro-GFP with BiP, PDI, and P4H.

Discussion

Contrary to the mutations in the “G” and “Y” positions, the molecular and cellular-level consequences of the “X” position amino acid substitutions in collagen II have not described. Here, we studied the molecular and cellular effects of the “X” position R992C substitution in collagen II found in mice with pathological skeletal features resembling those of SED24. More recently, we presented evidence that in the growth plates of these mutant mice chondrocytes include dilated ER cysternae and the number of cells undergoing apoptosis is higher than in control23. The molecular bases of such cellular changes, however, were not determined23.

In the present study we showed that α1(II) chains of collagen II with the R992C substitution are characterized by relatively slow electrophoretic mobility and the existence of intramolecular disulfide bonds. Relatively slow electrophoretic mobility of the mutant α1(II) chains was not a result of their excessive glycosylation. Because the electrophoretic mobility of fibrillar collagens does not strictly follow principles that govern migration of globular proteins, it is difficult to fully explain the aberrant behavior of the R992C mutant. We postulate, however, that the delayed electrophoretic mobility of the R992C mutant in the absence of evident chemical overmodifications could be a result of its structural alterations. In support of this notion is the report by Forlino et al. where delayed mobility of α2(I) chains harboring G421D substitution in the absence of the overmodifications was attributed to a structural change represented by a kink formed in some mutant collagen I molecules27. Moreover, in our own studies on R789C collagen II we observed structural aberrations and determined that the electrophoretic mobility of this mutant was slower than that of wild type collagen II28.

Consequences of structural aberrations of the R992C collagen molecules include their slow secretion from cells, lower thermostability, and formation of intramolecular disulfide bonds between about 6% of the α1(II) chains. Formation of disulfide bonds in the R75C, R519C, and R789C collagen II mutants isolated from patients’ tissues or in corresponding recombinant mutants expressed in HT-1080 cells was also reported15,19,28,29. A common characteristic of those mutants is that, unlike in the R992C collagen, the C for R substitutions occur at the “Y” positions of G-X-Y triplets.

Structural aberrations of the R992C collagen II are further illustrated by the formation of stable intermediate products of the digestion of this protein with trypsin and chymotrypsin. In accordance with the concept of micro-unfolding of a collagen triple-helix, it has been suggested that the presence of such intermediate products is evidence for the destabilization of cooperative blocks30,31. As suggested by Westerhausen et al., certain amino acid substitutions may expose trypsin/chymotrypsin-sensitive sites at relatively low temperatures, thereby leading to the generation of intermediate collagen fragments of varying sizes31. The 1°C-decrease in thermostability (Tm50) in the R992C mutant is relatively small in comparison to the 5°C-decrease of thermostability in the R789C mutant28. It is expected, however, that the change of the triple helix-stabilizing arginine residue to the triple helix-destabilizing cysteine residue is significant for the overall structure and physicochemical characteristics, such as the electrostatic potential, and the stability of the region encompassing the mutation site32. In particular, the R992C substitution may interfere with functions of the neighboring C-terminal triple-helix-stabilizing “knot” that is formed by tandem repeats of triple helix-stabilizing G-P-P triplets33.

In our previous studies on collagen II mutants we observed that, in contrast to thermostable R75C, R519C, and G853E mutants, the thermolabile R789C mutant was associated with apoptosis of host cells. The direct role of this mutant in triggering the apoptotic changes was demonstrated in chondrocytic cells in which the production of the R789C Pro-GFPTet was induced by Dox. Here, we also employed a Tet-inducible system for expressing the R992C Pro-GFPTet, and our results indicate that the presence of this mutant is directly responsible for the increase in the number of cells undergoing apoptosis. In comparison to the R789C mutant, however, the relative content of cleaved PARP, an indicator of apoptosis, was significantly lower in cells expressing the R992 Pro-GFP or R992C Pro-GFPTet variants. A less severe apoptotic response of the chondrocytic cells expressing the exogenous R992C collagen II together with endogenous wild type collagen II could explain the observation that SED-like characteristics of skeletal tissues were only apparent in sedc mice expressing the mutant Col2a1 in a homozygous way24. Still, we postulate that the possibility of mild aberrations at molecular and cellular levels in heterozygous sedc mice cannot be completely excluded.

Because of the apparent increase in the expression of BiP and PDI in response to the presence of the R992C collagen II, our study suggests activation of the unfolded protein response process in cells expressing this thermolabile mutant. In addition to the increase in the biosynthesis of PDI, we also demonstrated increased in situ colocalization of this chaperone with the R992C Pro-GFP. Because PDI not only functions independently but also serves as the β-subunit of the αα/ββ tetramer which constitutes functional P4H, we studied the specificity of the increase in the biosynthesis of PDI and its elevated colocalization with R992C Pro-GFP. Our studies with the anti-P4Hα antibodies determined that, in comparison to cells expressing WT Pro-GFP, the quantity of P4H and its colocalization with R992C Pro-GFP did not change. These observations suggest that, in response to the presence of the R992C mutant, PDI enhances rather its chaperone roles than its P4H-related functions. The increased colocalization of PDI with R992C Pro-GFP may indicate an increase in the PDI/R992C Pro-GFP binding. Such a notion is justified by observations that PDI preferably binds to the nascent procollagen α chains, but it may remain bound even after the α chains fold into a triple-helical structure2,34. As demonstrated with certain collagen I mutants, PDI facilitates intracellular retention of aberrant collagen molecules by binding to misfolded α chains. Such PDI-mediated retention of the R992C Pro-GFP is suggested by its slow rate of secretion from the host cells.

In contrast to PDI, in situ colocalization of BiP decreased in the cells expressing the R992C mutant. The role of BiP in retention of misfolded protein molecules is less clearly defined and somewhat controversial3,35. Based on the studies of collagen I and collagen II mutants, it has been suggested that this molecule interacts with collagen weakly, and it preferably binds to the triple-helical structure of the collagen molecule3,23. Thus, we propose that the decrease in colocalization of the R992C Pro-GFP with BiP could be a result of altered triple-helicity of the BiP-binding regions, blocking of the BiP-binding sites by PDI, or both.

Our studies presented here indicate that pathomechanisms responsible for altering skeletal tissues harboring single amino acid substitutions in collagen II include those that function at the intracellular level. In the future, detailed studies of biological processes triggered by the presence of mutant collagen molecules will reveal specific pathways that lead from altered collagen structure to the apoptosis of affected cells. Understanding such pathways will be critical for determining a genotype/phenotype relation in heritable diseases caused by mutations in genes that encode fibrillar collagens.

Materials and Methods

Mutation nomenclature

The amino acid substitutions are named according to the literature, with amino acid residues numbered from the first glycine of the collagen II triple helix. At their initial text appearances the mutations are also numbered from the ATG start codon following the journal and the Human Genome Variation Society guidelines (www.hgvs.org) and are listed in parentheses (with a “p” included in the mutation name).

Engineering cDNA constructs for unconditional and Tet-regulated expression of GFP-tagged procollagen II variants

Although the R992C substitution was originally described in mice, to study the consequences of this mutation at the molecular level, we employed a system that utilizes human collagen II. Such an approach is justified because of the high similarity between amino acid sequences of mouse (NP_112440.1) and human collagen II (NP_001835.3); in the region encompassing the mutation site, this similarity is 100%. A DNA construct for expression of the procollagen II mutant with the R992C substitution was assembled from DNA cassettes which encode specific domains of human procollagen II, as described21,36,37. In particular, a DNA construct was created by mutating the –CGA- codon for R992 to the –TGC- codon for C992 (Quick Change mutagenesis kit, Stratagene, Inc.).

For the unconditional expression of the GFP-tagged R992C procollagen II, denoted as R992C “Pro-GFP”, a DNA construct was cloned into the pEGFP-N3 vector (Clontech Laboratories, Inc.), as described22. In addition, to study the specific biological responses of cells to the presence of the R992C mutant, we employed a conditional expression system in which the expression of GFP-tagged procollagen II variants, denoted as “Pro-GFPTet”, was regulated by a Tet-responsive promoter (Clontech Laboratories, Inc.), as described23. The fidelity of all DNA constructs was determined by sequencing.

Analysis of procollagens and collagens variants expressed in HT-1080 cells

To obtain procollagen II for protein analyses, the WT Pro-GFP and the R992C Pro-GFP variants were expressed in HT-1080 cells (ATCC, CCL-121), as described25. These procollagen II variants were isolated from cell culture media and purified by ion exchange chromatography, as described25. Because the R992C procollagen II variant was relatively unstable and over time underwent partial degradation, for analyses of collagen II variants, procollagen secreted to cell culture media was precipitated with ammonium sulfate, as described. Pellets were suspended in 0.5 M acetic acid and then dialyzed against this solvent. Acetic acid-insoluble material was removed by centrifugation, and then proteins were digested for 48 h at 4°Cwith pepsin (Sigma-Aldrich) added to the final concentration of 0.1 mg/ml. Pepsin-resistant collagen was precipitated with 0.7 M NaCl and then collected by centrifugation.

Analysis of intact Pro-GFP and pepsin-processed collagen variants

Freshly isolated intact procollagen II variants secreted to cell culture media and pepsin-treated collagens were analyzed by electrophoresis. To visualize any potential differences in the electrophoretic mobilities of analyzed proteins, electrophoresis was carried out in 6% polyacrylamide gels and continued until a 50-kDa molecular mass marker had reached the lower edge of polyacrylamide gels. To facilitate specific detection of relatively small amounts of analyzed proteins, intact Pro-GFP variants were transferred to nitrocellulose membranes and visualized by immunostaining with anti-GFP antibody (Santa Cruz Biotechnology, Inc) and secondary anti-mouse IgG conjugated with horseradish peroxidase (HRP, Sigma-Aldrich). Specific bands were visualized by chemiluminescence (UVP, LLC). Similarly, pepsin-treated collagens were detected with anti-collagen II antibody (Millipore). Note that neither antibodies cross-react with proteins secreted from untransfected HT-1080 cells.

Treatment of collagen II with selected glycosidases and denaturing agents

To determine if delayed electrophoretic migration of the mutant α1(II) chains could be a result of their overglycosylation, we employed neuraminidase, N-glycosidase F, and endo-α-N-acetylgalactosaminidase (New England BioLabs, Inc.) to remove collagen-bound carbohydrates25. After treatment with selected enzymes, α1(II) chains were electrophoresed in polyacrylamide gels and the extents of their electrophoretic migrations were evaluated.

To exclude the possibility that the delayed migration of the R992C collagen II could be a result of its aggregation, we treated this mutant with selected agents that promote the dissociation of macromolecular complexes. Specifically, we employed 4 M urea, 4 M guanidinium hydrochloride, 0.5 M acetic acid or 50% dimethyl sulfoxide. Individual samples of wild type and R992C mutant were treated with these compounds for 48 h. Subsequently, electrophoretic migrations of collagen samples were analyzed.

Analysis of disulfide bonds

The intramolecular disulfide bonds formed between the C992 residues were assayed by electrophoresis. In brief, initially the α-chains of pepsin-treated collagen II were separated in 7.5% polyacrylamide gels under non-reducing conditions. Subsequently, gel strips with separated α-chains were cut and submerged for 1 h in a protein-loading buffer supplemented with β-mercaptoethanol. Next, the gel strip was placed horizontally on the top of a new polyacrylamide gel and electrophoresis was carried out. Finally, protein bands were transferred to nitrocellulose membranes and the α1(II) chains were specifically detected by Western blot as described above.

Analysis of the conformational integrity of the R992C collagen II

The thermostability of the R992C collagen was measured by brief protease digestion, as described38. In brief, the pepsin-treated samples dissolved in a 0.1 M Tris-HCl buffer (pH 7.4) supplemented with 0.4 M NaCl and 25 mM EDTA were pre-incubated for 30 min at selected temperatures and then treated for 3 min with a mixture of trypsin (100 μg/ml, Sigma-Aldrich) and chymotrypsin (250 μg/ml, Sigma-Aldrich). Enzymatic digestion was terminated by adding 0.1-volume of 5 mg/ml soybean trypsin inhibitor (Sigma-Aldrich). Protease-treated collagen samples were electrophoresed in reducing conditions in 7.5% polyacrylamide-SDS gels and then the protein bands were transferred to nitrocellulose membranes and analyzed by Western blot as described above. Relative amounts of the protease-resistant α1(II) chains in analyzed samples were determined by measuring the pixel intensities of the α1(II)-positive bands. Considering the relative value of the intensity of an α1(II) band derived from collagen digested at 30°C a 100%, the relative values of the intensities of α1(II) bands derived by enzymatic digestion at higher temperatures were measured. The mean values from three independent experiments were calculated and plotted against temperature (Sigma Plot v. 8.02, SPSS Inc.)

Computer modeling of the R992C mutation site

A computer model of the triple-helical region spanning the R992C mutation site was generated by the Sybyl7.3 program (Tripos, Inc.) as described18,39. Molecular surfaces of the analyzed collagen fragments were generated using the MOLCAD module of the Sybyl program20,40. The same module was used to calculate electrostatic potential (EP, physical unit; kcal/mol.e) and the area of the analyzed surfaces41.

Unconditional and Tet-regulated expression of Pro-GFP and Pro-GFPTet variants in SW-1353 cells

The WT Pro-GFP and R992C Pro-GFP variants were unconditionally expressed in the chondrocytic SW-1353 cell line (ATCC; HTB-94), as described22,23. This cell line was also employed to express the WT Pro-GFPTet and R992C Pro-GFPTet variants in the Tet-dependent fashion, as described23. Selected SW-1353 clones stably expressing Pro-GFP or Pro-GFPTet variants were cultured in Dulbecco’s Modified Eagle’s Media (DMEM) supplemented with 10% fetal bovine serum (FBS) or Tet-system-certified FBS (Clontech Laboratories, Inc.) and 40 μg/ml of L-ascorbic acid phosphate magnesium salt (Wako, Japan). Expression of the Pro-GFPTet variants was initiated by culturing the Tet-responsive SW-1353 cells in the presence of 1 μg/ml of doxycycline (Clontech Laboratories, Inc.). Biosynthesis of Pro-GFP and Pro-GFPTet variants in selected SW-1353 clones was confirmed at the mRNA and protein levels, as described22,23. Moreover, the kinetics of secretion of analyzed Pro-GFP variants were studied, as described22.

TUNEL and cleaved PARP assays

Confluent cells expressing Pro-GFP and Pro-GFPTet variants were cultured for 12 days and then analyzed for the markers of apoptosis, as described23. In one set of experiments the presence of fragmented DNA was analyzed by TUNEL assays (Roche). In these analyses untransfected SW-1353 cells treated with 3 units/ml of DNaseI (Sigma-Aldrich) served as a positive control. The cells were observed under a fluorescence microscope (Eclipse E600; Nikon) equipped with appropriate filter sets and a monochromatic digital camera (CoolSnap CF; Photometrics, Inc.) controlled by Image-Pro AMS software (version 5.1.0.2.0, Media Cybernetics, Inc). The total number of cells in viewing areas was determined by DAPI-positive nuclei. Subsequently, the percentage of TUNEL-DAPI-positive cells in analyzed groups was calculated23.

In addition to the TUNEL assays, Western blot analyses of cleaved PARP were employed to identify apoptosis, as described23. In these assays cleaved PARP was detected with primary rabbit anti-cleaved PARP antibodies (Cell Signaling Technology, Inc.) and secondary anti-rabbit IgG conjugated to HRP (Sigma-Aldrich). Employing a gel imaging system (UVP, LLC), bands corresponding to the cleaved PARP were visualized by chemiluminescence. The same nitrocellulose membranes were also probed with monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies (Santa Cruz Biotechnology, Inc.). Relative amounts of cleaved PARP in analyzed samples were determined by measuring the pixel intensities of the cleaved PARP-positive bands. To correct densitometry data for any differences in loading the samples onto polyacrylamide gels, final calculations were done assuming equal content of GAPDH in all analyzed samples. The ratio of relative amount of cleaved PARP from cells expressing the R992C Pro-GFP to cleaved PARP from cells expressing WT Pro-GFP was calculated. In experiments with Pro-GFPTet variants, the ratio of relative amount of cleaved PARP from Dox-induced cells to that from corresponding non-induced cells was calculated. Statistical significances of differences between the calculated means were analyzed by the Student’s t test (GraphPad Software, Inc.).

Measurements of selected indicators of ER stress

It has been shown that some mutations in fibrillar collagens are associated with increased expression of chaperones such as BiP and PDI2,3,23. To specifically determine whether ER stress-related changes in the relative amounts of these proteins were caused by the presence of the R992C mutant, we employed cells expressing R992C Pro-GFP and R992C Pro-GFPTet variants. The presence of intracellular BiP and PDI was analyzed by Western blot assays with rat monoclonal anti-BiP and goat anti-PDI antibodies (Santa Cruz Biotechnology, Inc), respectively. Consistent with other Western blot assays GAPDH-positive bands served as the internal control.

Analysis of LH and P4H

Analysis of LH (EC 1.14.11.4) and P4H (EC 1.14.11.2) in lysates from cells expressing Pro-GFP variants was done by Western blot assays. Primary goat anti-LH antibodies (Santa Cruz Biotechnology, Inc.) and secondary donkey anti-goat IgG antibodies conjugated to horseradish peroxidase (HRP; Rockland Immunochemicals) were used to detect LH-specific bands. Primary rabbit antibodies against the α1 subunit of P4H and HRP-conjugated secondary anti-rabbit IgG were employed to detect the P4Hα-specific bands (Sigma-Aldrich). In these assays GAPDH served as internal control and was detected as described above.

Fluorescence microscopy assays of in situ colocalization of the R992C Pro-GFP with BiP, PDI, and P4H

It has been shown that in response to the ER stress, BiP and PDI function as chaperones by binding to the misfolded mutant collagen molecules, thereby facilitating their intracellular retention. To determine the potential involvement of BiP and PDI in controlling intracellular retention and folding of the R992C collagen, we assayed the colocalization of these interacting partners. We also analyzed whether during ER stress PDI interacts with Pro-GFP variants independently or as the β-subunit of P4H. In these assays we studied colocalization of Pro-GFP variants with the catalytical α-subunit of the P4H. In previous studies we analyzed possible Pro-GFP/BiP and Pro-GFP/PDI binding by co-immunoprecipitation assays23. Because the studied interactions are weak, these assays are complex and require employing a cross-linking compound that prevents dissociation of interactants in the detergent-rich environment of a cell lysis buffer. Thus, in this study intracellular colocalization of the R992C Pro-GFP with BiP, PDI, and P4Hα was assayed after immunostaining of these specific binding partners with rabbit anti-BiP (Santa Cruz Biotechnology, Inc.), rabbit anti-PDI (Santa Cruz Biotechnology, Inc.) or rabbit anti-P4Hα antibodies (Sigma-Aldrich), respectively. Specific secondary antibodies conjugated with Alexa Fluor 594 were employed to visualize intracellular localization of analyzed proteins (Invitrogen, Inc.). At first, the cells were permeabilized with 0.2% Tween 20 and then the samples were blocked with appropriate serum to minimize non-specific binding of antibodies. Subsequently, the samples were incubated with primary and then secondary antibodies at conditions optimized in pilot studies. Finally, nuclei were stained with DAPI and the samples were mounted with a mounting media reagent (Prolong Gold; Invitrogen, Inc.). To determine the extent of nonspecific immunostaining and the degree of autofluorescence, samples treated only with secondary antibodies and those with no antibody treatment were prepared as controls.

The immunostained cells were observed with a microscope (Eclipse E600, Nikon, Japan) equipped with a computer-controlled Z-stack stage and sets of chromophor-specific filters and the Plan Fluor 40x/1.3 Oil objective (Nikon, Japan). Images of the cells were acquired with a monochromatic digital camera (CoolSnap CF; Photometrics, Inc.) controlled by Image-Pro AMS software (version 5.1.0.2.0, Media Cybernetics, Inc). Control samples were employed to correct for the background fluorescence. Pairs of images taken with filters specific for GFP and Alexa Fluor 594 were pseudo-colored and then merged. Following deconvolution, colocalization of pixels from green and red channels was measured by using the “Colocalization Module” of the Image-Pro AMS software. Colocalization of pixels from a cytoplasm region of analyzed cells was expressed as the overlap coefficient (R). This parameter, defined by Manders et al., represents a true degree of colocalization of defined pixels and is not sensitive to the limitations of typical fluorescence imaging, such as efficiency of hybridization, sample photobleaching, and camera quantum efficiency42.

Acknowledgments

This work was supported by the National Health Institutes to A.F. (5R01AR049537-06).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Prockop DJ, Kivirikko KI. Collagens: molecular biology, diseases, and potentials for therapy. Annu Rev Biochem. 1995;64:403–434. [PubMed]
2. Chessler SD, Byers PH. Defective folding and stable association with protein disulfide isomerase/prolyl hydroxylase of type I procollagen with a deletion in the pro alpha 2(I) chain that preserves the Gly-X-Y repeat pattern. J Biol Chem. 1992;267:7751–7757. [PubMed]
3. Chessler SD, Byers PH. BiP binds type I procollagen pro alpha chains with mutations in the carboxyl-terminal propeptide synthesized by cells from patients with osteogenesis imperfecta. J Biol Chem. 1993;268:18226–18233. [PubMed]
4. McLaughlin SH, Bulleid NJ. Molecular recognition in procollagen chain assembly. Matrix Biol. 1998;16:369–377. [PubMed]
5. Tasab M, Batten MR, Bulleid NJ. Hsp47: a molecular chaperone that interacts with and stabilizes correctly-folded procollagen. Embo J. 2000;19:2204–2211. [PubMed]
6. Walmsley AR, Batten MR, Lad U, Bulleid NJ. Intracellular retention of procollagen within the endoplasmic reticulum is mediated by prolyl 4-hydroxylase. J Biol Chem. 1999;274:14884–14892. [PubMed]
7. Olsen BR. Mutations in collagen genes resulting in metaphyseal and epiphyseal dysplasias. Bone. 1995;17:45S–49S. [PubMed]
8. Kuivaniemi H, Tromp G, Prockop DJ. Mutations in collagen genes: causes of rare and some common diseases in humans. Faseb J. 1991;5:2052–2060. [PubMed]
9. Bleasel JF, Bisagni-Faure A, Holderbaum D, Vacher-Lavenu MC, Haqqi TM, Moskowitz RW, Menkes CJ. Type II procollagen gene (COL2A1) mutation in exon 11 associated with spondyloepiphyseal dysplasia, tall stature and precocious osteoarthritis. J Rheumatol. 1995;22:255–261. [PubMed]
10. Bleasel JF, Holderbaum D, Mallock V, Haqqi TM, Williams HJ, Moskowitz RW. Hereditary osteoarthritis with mild spondyloepiphyseal dysplasia--are there “hot spots” on COL2A1? J Rheumatol. 1996;23:1594–1598. [PubMed]
11. Bleasel JF, Holderbaum D, Brancolini V, Moskowitz RW, Haqqi TM, Considine E, Prockop DJ, Devoto M, Williams CJ. Arg519-Cys mutation in COL2A1: evidence for multiple founders. Ann N Y Acad Sci. 1996;785:215–218. [PubMed]
12. Knowlton RG, Katzenstein PL, Moskowitz RW, Weaver EJ, Malemud CJ, Pathria MN, Jimenez SA, Prockop DJ. Genetic linkage of a polymorphism in the type II procollagen gene (COL2A1) to primary osteoarthritis associated with mild chondrodysplasia. N Engl J Med. 1990;322:526–530. [PubMed]
13. Hoornaert KP, Dewinter C, Vereecke I, Beemer FA, Courtens W, Fryer A, Fryssira H, Lees M, Mullner-Eidenbock A, Rimoin DL, Siderius L, Superti-Furga A, Temple K, Willems PJ, Zankl A, Zweier C, De Paepe A, Coucke P, Mortier GR. The phenotypic spectrum in patients with arginine to cysteine mutations in the COL2A1 gene. J Med Genet. 2006;43:406–413. [PMC free article] [PubMed]
14. Williams CJ, Considine EL, Knowlton RG, Reginato A, Neumann G, Harrison D, Buxton P, Jimenez S, Prockop DJ. Spondyloepiphyseal dysplasia and precocious osteoarthritis in a family with an Arg75-->Cys mutation in the procollagen type II gene (COL2A1) Hum Genet. 1993;92:499–505. [PubMed]
15. Chan D, Taylor TK, Cole WG. Characterization of an arginine 789 to cysteine substitution in alpha 1 (II) collagen chains of a patient with spondyloepiphyseal dysplasia. J Biol Chem. 1993;268:15238–15245. [PubMed]
16. Ballo R, Beighton PH, Ramesar RS. Stickler-like syndrome due to a dominant negative mutation in the COL2A1 gene. Am J Med Genet. 1998;80:6–11. [PubMed]
17. Richards AJ, Baguley DM, Yates JR, Lane C, Nicol M, Harper PS, Scott JD, Snead MP. Variation in the vitreous phenotype of Stickler syndrome can be caused by different amino acid substitutions in the X position of the type II collagen Gly-X-Y triple helix. Am J Hum Genet. 2000;67:1083–1094. [PubMed]
18. Fertala A, Ala-Kokko L, Wiaderkiewicz R, Prockop DJ. Collagen II containing a Cys substitution for arg-alpha1–519. Homotrimeric monomers containing the mutation do not assemble into fibrils but alter the self-assembly of the normal protein. J Biol Chem. 1997;272:6457–6464. [PubMed]
19. Fertala A, Sieron AL, Adachi E, Jimenez SA. Collagen II containing a Cys substitution for Arg-alpha1-519: abnormal interactions of the mutated molecules with collagen IX. Biochemistry. 2001;40:14422–14428. [PubMed]
20. Steplewski A, Brittingham R, Jimenez SA, Fertala A. Single amino acid substitutions in the C-terminus of collagen II alter its affinity for collagen IX. Biochem Biophys Res Commun. 2005;335:749–755. [PubMed]
21. Steplewski A, Ito H, Rucker E, Brittingham RJ, Alabyeva T, Gandhi M, Ko FK, Birk DE, Jimenez SA, Fertala A. Position of single amino acid substitutions in the collagen triple helix determines their effect on structure of collagen fibrils. J Struct Biol. 2004;148:326–337. [PubMed]
22. Ito H, Rucker E, Steplewski A, McAdams E, Brittingham RJ, Alabyeva T, Fertala A. Guilty by association: some collagen II mutants alter the formation of ECM as a result of atypical interaction with fibronectin. J Mol Biol. 2005;352:382–395. [PubMed]
23. Hintze V, Steplewski A, Ito H, Jensen DA, Rodeck U, Fertala A. Cells expressing partially unfolded R789C/p.R989C type II procollagen mutant associated with spondyloepiphyseal dysplasia undergo apoptosis. Hum Mutat. 2008;29:841–851. [PubMed]
24. Donahue LR, Chang B, Mohan S, Miyakoshi N, Wergedal JE, Baylink DJ, Hawes NL, Rosen CJ, Ward-Bailey P, Zheng QY, Bronson RT, Johnson KR, Davisson MT. A missense mutation in the mouse Col2a1 gene causes spondyloepiphyseal dysplasia congenita, hearing loss, and retinoschisis. J Bone Miner Res. 2003;18:1612–1621. [PMC free article] [PubMed]
25. Fertala A, Sieron AL, Ganguly A, Li SW, Ala-Kokko L, Anumula KR, Prockop DJ. Synthesis of recombinant human procollagen II in a stably transfected tumour cell line (HT1080) Biochem J. 1994;298:31–37. [PubMed]
26. Gossen M, Freundlieb S, Bender G, Muller G, Hillen W, Bujard H. Transcriptional activation by tetracyclines in mammalian cells. Science. 1995;268:1766–1769. [PubMed]
27. Forlino A, Keene DR, Schmidt K, Marini JC. An alpha2(I) glycine to aspartate substitution is responsible for the presence of a kink in type I collagen in a lethal case of osteogenesis imperfecta. Matrix Biol. 1998;17:575–584. [PubMed]
28. Steplewski A, Majsterek I, McAdams E, Rucker E, Brittingham RJ, Ito H, Hirai K, Adachi E, Jimenez SA, Fertala A. Thermostability gradient in the collagen triple helix reveals its multi-domain structure. J Mol Biol. 2004;338:989–998. [PubMed]
29. Eyre DR, Weis MA, Moskowitz RW. Cartilage expression of a type II collagen mutation in an inherited form of osteoarthritis associated with a mild chondrodysplasia. J Clin Invest. 1991;87:357–361. [PMC free article] [PubMed]
30. Privalov PL. Stability of proteins. Proteins which do not present a single cooperative system. Adv Protein Chem. 1982;35:1–104. [PubMed]
31. Westerhausen A, Kishi J, Prockop DJ. Mutations that substitute serine for glycine alpha 1–598 and glycine alpha 1–631 in type I procollagen. The effects on thermal unfolding of the triple helix are position-specific and demonstrate that the protein unfolds through a series of cooperative blocks. J Biol Chem. 1990;265:13995–14000. [PubMed]
32. Persikov AV, Ramshaw JA, Kirkpatrick A, Brodsky B. Amino acid propensities for the collagen triple-helix. Biochemistry. 2000;39:14960–14967. [PubMed]
33. Arnold WV, Fertala A, Sieron AL, Hattori H, Mechling D, Bachinger HP, Prockop DJ. Recombinant procollagen II: Deletion of D period segments identifies sequences that are required for helix stabilization and generates a temperature-sensitive N-proteinase cleavage site. J Biol Chem. 1998;273:31822–31828. [PubMed]
34. Bassuk JA, Berg RA. Protein disulphide isomerase, a multifunctional endoplasmic reticulum protein. Matrix. 1989;9:244–258. [PubMed]
35. Gething MJ, Sambrook J. Protein folding in the cell. Nature. 1992;355:33–45. [PubMed]
36. Arnold WV, Sieron AL, Fertala A, Bachinger HP, Mechling D, Prockop DJ. A cDNA cassette system for the synthesis of recombinant procollagens. Variants of procollagen II lacking a D-period are secreted as triple- helical monomers. Matrix Biol. 1997;16:105–116. [PubMed]
37. Majsterek I, McAdams E, Adachi E, Dhume ST, Fertala A. Prospects and limitations of the rational engineering of fibrillar collagens. Protein Sci. 2003;12:2063–2072. [PubMed]
38. Sieron AL, Fertala A, Ala-Kokko L, Prockop DJ. Deletion of a large domain in recombinant human procollagen II does not alter the thermal stability of the triple helix. J Biol Chem. 1993;268:21232–21237. [PubMed]
39. Chen JM, Sheldon A, Pincus MR. Three-dimensional energy-minimized model of human type II “Smith” collagen microfibril. J Biomol Struct Dyn. 1995;12:1129–1159. [PubMed]
40. Connolly ML. Solvent-accessible surfaces of proteins and nucleic acids. Science. 1983;221:709–713. [PubMed]
41. Warwicker J, Watson HC. Calculation of the electric potential in the active site cleft due to alpha-helix dipoles. J Mol Biol. 1982;157:671–679. [PubMed]
42. Manders EMM, Verbbk FJ, Aten JA. Measurement of co-localization of objects in dual-color confocal images. J Microsc. 1993;169:375–382.