The second currently controversial question about TLR9 is: how does it activate? Some researchers imply that TLR9 exists as a pre‐formed dimer and believe that ligand binding induced ectodomain conformational change activates TLR9. However, this activation mechanism is quite different from the dimerization‐induced‐activation mechanism shared by TLR1, 2, 4 and 5, hence is not widely accepted by other researchers. Last but not the least, there has been discrepancies about which part of TLR9ecd is involved in DNA ligand recognition. A long undefined region separates TLR9ecd into N‐terminal and C‐terminal parts. Cleavage of TLR9 at this region was detected in vivo and proposed to be the premise of TLR9 activation by DNA. However this paradigm is inconsistent with evidences showing that the function of TLR9 is very sensitive to SNPs at the N‐terminal part of TLR9ecd. We seek to answer these biological questions through structural and biochemical studies. While a prerequisite of these studies is to get monosized TLR9ecd, previous efforts to express TLR9ecd yield only soluble aggregates. We tried to solve the aggregation issue of TLR9ecd through the following strategies: 1) Co‐expressing TLR9ecd with its identified chaperons;2) Screening different detergents to reduce non‐specific hydrophobic interaction;3) Exploring different TLR9ecd homologues;4) Protein engineering comprising mutation, deletion, serial truncation and chimera techniques (hybrid TLR9ecd to VLRecd or other TLRecds). We focused on the chimera constructs since we identified previously several chimeras which showed dramatically enhanced expression level. Further experiments revealed that from all the 17 chimeras constructed and tested, only chimera No.5 showed greatly improved homogeneity. Later experiments showed that this chimera failed to be activated by CpG-DNA in vivo and cannot be dimerized by CpG-DNA oligos in vitro. Since the chimera technique did not work, a mammalian expression system was used to solve the "aggregation" problem. Mouse TLR9 ECD can now be purified as a highly pure, monomeric protein in fair amounts. After solving the aggregation problem of TLR9ecd we designed and conducted biochemical and cell biology experiments to answer the aforementioned three biological questions. The results showed that mTLR9ecd predominantly exists as monomer in solution, but assembles into dimer upon PD-ODN binding. The DNA binds and dimerizes mTLR9ecd in a sequence independent manner. Next, we tried to investigate which part of TLR9ecd is involved in DNA ligand recognition. NF-kB reporter assay showed that the reported cleavage product of TLR9 is not functional in vivo. The in vitro proteolysis experiments with purified mTLR9ecd and cathepsin S indicate that the cleavage only occurs efficiently at the pH of lysosome but not of the endosome. Lastly, the long undefined region in TLR9ecd is not as flexible as previously predicted and likely adopts a certain secondary structure. Collectively, we successfully purified monosized mTLRecd and showed that TLR9 is not a preformed dimer but a monomer in solution. For the first time, we provided direct in vitro evidences to support that the binding between TLR9ecd and ssDNA is sequence independent, which implies that TLR9 differentiate self and non-self rather through its intracellular location than through recognizing specific DNA motif. Also the binding of ssDNA to mTLR9ecd induced the dimerization of the latter, which then formed a stable 2:1 complex with ssDNA. This result coincides with the dimerization-induced-activation-mechanism of TLR1, 2, 3, 4, 5, 6, and indicates that TLR9 most likely also adopt the same activation mechanism. Whats more, the results also showed that the C terminal part of TLR9ecd is insufficient for ssDNA recognition in vivo, and the cleavage of purified mTLR9ecd is not efficient at early endosome pH. These results are more consistent with a model in which the detected cleavage of TLR9 is a result instead of a premise of its activation.

Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Zip Code
Botos, Istvan; Segal, David M; Davies, David R (2011) The structural biology of Toll-like receptors. Structure 19:447-59
Fitzkee, Nicholas C; Masse, James E; Shen, Yang et al. (2010) Solution conformation and dynamics of the HIV-1 integrase core domain. J Biol Chem 285:18072-84
Li, Min; Chen, Changqing; Davies, David R et al. (2010) Induced-fit mechanism for prolyl endopeptidase. J Biol Chem 285:21487-95
Wang, Yan; Liu, Lin; Davies, David R et al. (2010) Dimerization of Toll-like receptor 3 (TLR3) is required for ligand binding. J Biol Chem 285:36836-41
Botos, Istvan; Liu, Lin; Wang, Yan et al. (2009) The toll-like receptor 3:dsRNA signaling complex. Biochim Biophys Acta 1789:667-74