NMR studies unravel the binding of glycosaminoglycans to CXCL14

S8.4 Glycosaminoglycans
Location (hall): 
Start/end time: 
Thursday, July 4, 2019 - 12:00 to 12:15

Anja Penk1, Sergey A. Samsonov2, Daniel Huster1, Lars Baumann1

1Institute for Medical Physics and Biophysics, Leipzig University, Leipzig, Germany, 2Faculty of Chemistry, University of Gdańsk, Gdańsk, Poland

CXCL14, a soluble 9.4 kDa protein, is a novel, highly conserved chemokine with unprecedented features. Despite exhibiting the typical chemokine fold, its flexible N terminus consists of only two amino acids; a domain that is usually attributed to chemokine receptor activation. CXCL14 is a positive allosteric modulator of the receptor CXCR4 and a growing body of evidence points towards a multitude of immunomodulatory activities of CXCL14 such as homeostatic immune surveillance and elimination of early neoplasmic transformations in skin and mucosae. Its strong anti-angiogenic properties may be a target for cancer therapy. 

In the current study, we investigated the interaction of CXCL14 with different glycosaminoglycans (GAGs) in order to understand the molecular basis of GAG-binding. To this end, we cloned, expressed and refolded CXCL14 with final yields of up to 4 mg highly pure protein per liter E. coli batch culture. 

We used ¹⁵N-isotopically labelled CXCL14 to investigate the interaction of GAGs with this protein via NMR chemical shift perturbation (CSP). The recorded ¹H-¹⁵N HSQC NMR spectra show well resolved signals with a single peak for each backbone amide of the protein. Upon titration of GAG ligands, peaks near the binding site are expected to shift in the spectrum due to their altered electronical environment or secondary structure changes induced by ligand binding. These CSPs were monitored for chondroitinsulfate A/C and D as well as heparin and dermatan sulfate for varying protein/GAG ratios.

Interestingly, the observable CSPs with the various hexameric GAGs (dp6) reveal that the amino acids of CXCL14 are influenced differently. Especially the pattern for heparin suggests a different primary binding site compared to the other GAGs. Thus, we assume that varying sulfation patterns of the GAGs confer specificity beyond simple electrostatic interactions. This hypothesis is also supported by complementary computational methods (docking and molecular dynamics-based studies). 

However, the differences in the amino acid pattern are less pronounced, if longer GAGs (dp10) are investigated. Furthermore, titration experiments of heparin dp10 were not feasible due to an immense loss in NMR signal intensity. Upon heparin dp10 titration the solution remains clear and no line broadening in the spectra was observed. This indicates a tendency of CXCL14 to form stable, fast relaxing, high molecular weight oligomers in the presence of heparin, which were also shown in a preliminary CXCL14 crosslinking experiment. 

Hence, CXCL14 shows differences in the binding site for the investigated GAGs and this binding seems to be closely related to the oligomerization behaviour of CXCL14. Therefore, CXCL14 and GAGs provide a promising system to investigate a possible specificity of GAG protein interactions.