It has been reported that the
It has been reported that the ECD of DDRs require dimerization and/or oligomerization for binding to collagen.,, Here, we have established using SPR that antibody-mediated oligomerization of DDR2-Fc significantly enhances its binding to immobilized collagen type 1. Our results are consistent with earlier observations using these identical fusion proteins, namely, that SPR was unable to detect any binding response of DDR2-Fc proteins alone with collagen and microplate-based assays showed results consistent with our SPR data. Utilizing a different fusion construct (His-DDR2), Leitinger has reported that dimerization of DDR2 is necessary and sufficient for its binding to collagen. These observations are in slight contradiction with our results as Leitinger reports that His-DDR2 (which is a dimer) is able to bind to collagen with no further oligomerization. It is possible that the difference between Leitinger's and our constructs and/or the techniques employed may account for this difference. Our fusion proteins have a bridging sequence that may affect the characteristics of the dimer formed in DDR2-Fc proteins and it is likely that antibody-mediated oligomerization may be necessary to have the DDR2-ECD in close proximity and/or correct orientation for binding to collagen. Our results demonstrate that DDR2 has an inhibitory effect on collagen fibrillogenesis as shown by the long tlag in turbidity measurements, lower fibrillar collagen content by the hydroxyproline assay and the AFM observation that the collagen supernatant consisted largely of monomeric collagen for samples with DDR2. Further, the collagen fibers formed in the presence of DDR2 exhibited a striking lack of the native periodic banded structure. A similar delay in collagen fibrillogenesis and changes in collagen morphology are also observed in our cell-based assays, which may signify the role of DDR2 on collagen fibril formation in vivo. It is interesting to note that the effect of DDR2 angiotensin 1 sale on collagen fibrillogenesis in our cell-based assay was observed for a collagen concentration much below 10 μg/ml, optimally required for its activation, and as early as after 1 h of addition of collagen, which is below the optimal time (90 min) required for activation of DDR2 by collagen. It has been reported that activation of DDR2 can lead to up-regulation of matrix metalloproteases (MMPs) but these are detectable only after prolonged collagen stimulation of over two to four days.,20., 21.,  It is therefore unlikely that the difference in collagen morphology for DDR2 overexpressing cells observed in our experiments is due to the cleaving action of MMPs up-regulated by DDR2 activation. The aggregates of collagen present in DDR2 samples exhibit a granular instead of fibrillar morphology at the light microscopy level, suggesting that DDR2 may be inhibiting collagen fibrillogenesis by promoting formation of aggregates of monomeric collagen. Our earlier investigations have revealed that DDR2 binds to a single site on the collagen molecule and this has been localized recently to be on the D2 domain for collagen type 2. It is thus likely that the DDR2 binding site on collagen plays an essential role in the natural self-assembly of collagen molecules and in forming native, banded fibers. Indeed, it has been well-reported that specific sites on collagen play a crucial role in its fibrillogenesis. Our results indicate that weakly or non-interacting proteins like TrkB may also contribute to delay in collagen fibrillogenesis. This can be explained using the model proposed by Parkinson et al. that fibrillogenesis can be modeled as a diffusion-limited aggregation of individual collagen molecules. The presence of other proteins is likely to provide steric hindrance and thus limit the rate of diffusion of collagen monomers, leading to delay in collagen fibrillogenesis. As expected, the diameter and banded structure of collagen fibers formed in the presence of TrkB was identical to that of collagen alone.