MUC4-ErbB2 Oncogenic Complex: Binding studies using Microscale Thermophoresis

The MUC4 membrane-bound mucin is a large O-glycoprotein involved in epithelial homeostasis. At the cancer cell surface MUC4 interacts with ErbB2 receptor via EGF domains to promote cell proliferation and migration. MUC4 is highly regarded as a therapeutic target in pancreatic cancer as it is not expressed in healthy pancreas, while it is neoexpressed in early preneoplastic stages (PanINs). However, the association/dissociation constant of MUC4-ErbB2 complex is unknown. Protein-protein interactions (PPIs) have become a major area of research in the past years and the characterization of their interactions, especially by biophysical methods, is intensively used in drug discovery. To characterize the MUC4-ErbB2 interaction, we used MicroScale Thermophoresis (MST), a powerful method for quantitative protein interaction analysis under challenging conditions. We worked with CHO cell lysates containing either the transmembrane β subunit of MUC4 (MUC4β) or a truncated mutant encompassing only the EGF domains (MUC4EGF3+1+2). MST studies have led to the characterization of equilibrium dissociation constants (Kd) for MUC4β-ErbB2 (7–25 nM) and MUC4EGF3+1+2/ErbB2 (65–79 nM) complexes. This work provides new information regarding the MUC4-ErbB2 interaction at the biophysical level and also confirms that the presence of the three EGF domains of MUC4 is sufficient to provide efficient interaction. This technological approach will be very useful in the future to validate small molecule binding affinities targeting MUC4-ErbB2 complex for drug discovery development in cancer. It will also be of high interest for the other known membrane mucins forming oncogenic complexes with ErbBs at the cancer cell surface.

Impact of the MST power on the MUC4β-ErbB2 Kd measurement. Comparison of various MST power toward the MUC4-ErbB2 interaction. The best compromise is obtained at high irradiation power, with 80% having slightly better signal-to-noise ratio.
Impact of the incubation time on the MUC4β-ErbB2 Kd measurement. Evaluation of the interaction kinetics for MUC4 (top left) and MUC4EGF3+1+2 (top right) proteins with ErbB2, at 80% MST power. Affinity is monitored every 30 min at room temperature. pKd Data, normalized by the lower value, are plotted below and show a kinetic for the interaction with an optimal Kd found after at least 3 hours of incubation . After 7h, degradation of the lysate is observed at room temperature. The same kinetic is found with both proteins. eGFP capture. To develop a label-free strategy of SPR, we used the eGFP-MUC4 fusion proteins to ensure specific capture from the lysate. CHO-K1 cell lysates expressing either eGFP-MUC4β or eGFP-MUC4EGF3+1+2 were thus prepared in MPer buffer, as for the MST studies. A step-by-step development was chosen to quickly identify any problems and therefore eGFP was first used. Two strategies of capture (direct or indirect) of pure eGFP using specific antibodies were assayed. Unfortunately, none of these methods led to a response signal following eGFP injection (not shown). As these antibodies are often used for biomolecular assays with longer incubation time, they appeared not suited for SPR assays.
Recently, a single domain antibody was described as a potent specific and stable GFP-binding protein (GBP) usable for efficient fishing of GFP tagged proteins. Moreover, the use of GBP from complex biological samples such as cell extracts 1 is very convenient in a Biacore ® capture assay format. The immobilization of GBP on the sensor chip following the procedure described by Biacore ® was realized to produce a GBP coated chip (CM5 sensor chip with covalently bound GBP). 2 First, 3458 RU of GBP were chip-immobilized, and then, another chip was produced with 554 RU (a lower level as advised for binding experiments between two proteins).
Solutions of pure eGFP with concentrations ranging from 2 pM to 200 nM were injected and the GBP coated chip was able to detect eGFP from 2 pM and the saturation reached 85% of the theoretical maximum response at 200 nM (Fig.S4A). Regeneration was efficiently done by injection of Glycine-HCl buffer (0.01 M, pH 2.0) and this step allowed highly reproducible measurements with 0.25% of variability coefficient (VC) (Fig.S4B). Diluted pure eGFP in CHO-K1 cell extract was also injected to evaluate the impact of a crude lysate on the binding property of the GBP chip and no major perturbation was observed (Fig.S4A, light blue curve).
Altogether, these data validated our method.  (Fig.S6). Moreover, this response was specific as a pre-incubation with non-relevant Fc-protein (PD-1-Fc) did not lead to any significant change of the response.

Discussion
Our strategy merges SPR sensitivity to detect proteins from biological fluids with the single domain antibody GBP as a powerful tool to immobilize eGFP-MUC4 fusion protein from cell lysate. 1,2 Very good results were obtained regarding the immobilization of MUC4 from the lysates followed by an easy and reproducible regeneration step. As binding studies with increasing recombinant ErbB2-Fc concentrations were performed without success, we attempted to directly fish the MUC4-ErbB2 complex after a pre-incubation step. We considered that the level of the residual signal after injection of the cell lysate followed by washing procedure with running buffer to remove all non-specific interacting proteins, could approximately estimate the amount of the fished complex that can be caught from the lysate.
Our data were in accordance with recent works in SPR with cell lysates 3

Injection of increasing concentrations of recombinant ErbB2 on eGFP-MUC4β-GBP
immobilized chip. In 1 and 2, lysate containing eGFP-MUC4β was injected to be immobilized on the GBP-coated chip. After the second injection, buffer was run to ensure the stability of the immobilized MUC4β. Then in 3, 4, 5 and 6 was injected increasing concentration of recombinant ErbB2-Fc in PBS (respectively 2, 20, 200 nM and 2 µM). No binding could be observed in these conditions. Specific detection of the interaction between eGFP-MUC4 and ErbB2-Fc after incubation by SPR. Lysates were incubated 3h at room temperature , either alone (blue), with 200 nM of ErbB2-Fc (orange) or with non-relevant equivalent PD-1-Fc protein (grey) then injected on GBP immobilized chip. The mean value for the RU response between 440 and 500 s after injection was plotted in n=5 independent injections in a box plot representation. The lysate containing MUC4 presents a significant increase (*** p<0.005) of the response with incubation of ErbB2-Fc in comparison to the same lysate with no protein incubated or with incubation of a negative control, PD1-Fc (p<0.0001). A difference was observed between the negative control and the sole lysate but was not significate (p<0.12).
Full length vector of MUC4β. Full length vector of the peGFP-C1 kana/neo MUC4β, with the cloning sites BglII and BamHI used and a codon optimization for CHO cells performed.