Brittany Kartchner and Dr. Eric Wilson, Department of Microbiology and Molecular Biology
For this project, I hypothesized that the amino acid composition of the C terminus of chemokine proteins allows some chemokines to bind to different sites on the bacterial cell wall and that the chemokines with a strong positive charge (CCL25 and CCL28) will bind to the same target, while CCL27, which has no positively charged C terminus region, will have a different binding site. In order to study the binding properties of each chemokine, I began by performing binding assays. In these assays, approximately 105 colony-forming units of exponentially growing bacteria were incubated on ice with the appropriate chemokine. This allowed the chemokine to bind to the bacteria and resulted in low levels of bacterial lysis. After thirty minutes, any unbound chemokine was removed by washing in a PBS/calf serum wash buffer. The bacteria were then incubated with a second chemokine. Any unbound chemokine was again washed away. Next, the cells were incubated with an anti-chemokine biotinylated antibody specific to the second chemokine and any unbound antibody was washed away. Finally, an avidin-conjugated fluorochrome was added to the cells. The percentage of stained cells, as well as the intensity of staining on each cell, was read by flow cytometry. By comparing these results to the control set, in which no blocking antibody was added (omission of the first chemokine in the experiment described above), these assays should have allowed me to determine if the binding of the first chemokine was blocking the binding of the second chemokine.
The results obtained from this experiment were inconsistent and unexpected. The chemokines appeared to bind at the same rate one day and not at all the next day. On some days, chemokine CCL28 appeared to be blocked by chemokine CCL25. Other days, CCL28 appeared to have a high binding rate despite the presence of CCL25. Seeing these ambiguous results, I hypothesized that the antibodies to CCL28 were cross-reacting with CCL27 and CCL25 and that the two other antibodies used were also cross-reacting to their nonspecific chemokines. This created false positives and skewed results.
In order to test this hypothesis, I grew bacteria that had high affinity for chemokine binding as described previously. Next, it was incubated with CCL28. After washing the bacteria of the excess CCL28, the mutant 27 was incubated with anti-CCL25. After washing out the excess anti-CCL25, the bacterial staining levels were read by flow cytometry to assess the binding levels of anti-CCL25 to CCL28. The procedure was repeated with CCL25 and anti-CCL28 as well. The results of these readings showed that the antibodies were cross-reacting with the chemokines for which they were not intended.
Upon closer examination of the antibodies used, I discovered that they were polyclonal antibodies. Polyclonal antibodies are antibodies that bind at several different epitopes on each chemokine. As these chemokines have such high similarity, the polyclonal antibodies, or PAbs, could be binding to the epitopes on the chemokine they were intended for as well as several epitopes on the chemokines they were not intended for. To remedy this problem, monoclonal antibodies (antibodies that bind to one epitope specific to their intended chemokine) were ordered and the cross-reaction test between CCL28 with anti-CCL25, and CCL25 with anti-CCL28, was repeated with the new monoclonal antibodies. Results indicated that these monoclonal antibodies also cross-reacted.
At this point, Dr. Wilson suggested that I use several monoclonal antibodies to CCL28 that were created in his lab a few years ago for a previous experiment. Upon performing the cross-reactivity experiment to test nonspecific binding of these monoclonal anti-CCL28 antibodies to CCL25, it was shown that these antibodies did not cross-react to CCL28 and bound to CCL28 with a high affinity. However, it was realized that even with this monoclonal antibody to CCL28, I could no longer test if CCL28, CCL25, and CCL27 had the same molecular target without working monoclonal antibodies to CCL27 and CCL25.
Without the antibodies for CCL27 and CCL25, I could not identify if CCL28, CCL25, and CCL27 had the same molecular target. Instead, I decided to try a different approach and see if I could identify the type of target that the chemokines were binding to, whether it was part of the LPS or a protein target on the bacterial cell membrane. To do this, approximately 105 colony-forming units of exponentially growing bacteria are incubated at 37°C with sodium periodate or with proteinase K. Sodium periodate works by opening saccharide rings between vicinal diols leaving two aldehyde groups, effectively destroying the O-antigen and core structures of the LPS. Proteinase K works by digesting any protein that has hydrophobic amino acids, which is includes membrane bound proteins. Comparing results from these experiments to control experiments in which no sodium periodate or proteinase K is added, will allow me to determine if LPS or membrane proteins are the target for chemokine binding. If a protein is the molecular target of CCL28, when proteinase K is added to cells, the amount of chemokine that can bind to the bacteria will be significantly reduced. If the cells’ LPS is the molecular target for CCL28, when sodium periodate is added, the chemokine will not be able to bind to the bacteria as well as a bacteria that has not been exposed to sodium periodate.
Preliminary results indicate that CCL28 binds to a protein target. Confirmation of these results is ongoing. The tests will also be repeated with CCL25 and CCL27 to determine the type of molecular targets bound by the chemokines.