Function of the CysD domain of the gel-forming MUC2 mucin

Abstract

The colonic human MUC2 mucin forms a polymeric gel by covalent disulfide bonds in its N- and C-termini. The middle part of MUC2 is largely composed of two highly O-glycosylated mucin domains that are interrupted by a CysD domain of unknown function. We studied its function as recombinant proteins fused to a removable immunoglobulin Fc domain. Analysis of affinity-purified fusion proteins by native gel electrophoresis and gel filtration showed that they formed oligomeric complexes. Analysis of the individual isolated CysD parts showed that they formed dimers both when flanked by two MUC2 tandem repeats and without these. Cleavages of the two non-reduced CysD fusion proteins and analysis by MS revealed the localization of all five CysD disulfide bonds and that the predicted C-mannosylated site was not glycosylated. All disulfide bonds were within individual peptides showing that the domain was stabilized by intramolecular disulfide bonds and that CysD dimers were of non-covalent nature. These observations suggest that CysD domains act as non-covalent cross-links in the MUC2 gel, thereby determining the pore sizes of the mucus.

Figure 1. Domain organization of MUC2 and the two CysD fusion constructs

(A) MUC2 is made up of D1-D2-D′-D3-CysD-PTS-CysD-PTS (tandem repeated)-D4-B-C-CK domains. (B) The two CysD fusion constructs named pSMCysD-IgG2a/His and pSMCysD(2TR)-IgG2a/His are made up by an N-terminal Myc tag, CysD and a C-terminal IgG-Fc and His₆ tag. The second fusion construct included two repeats (2TR) of the general tandem repeat of the PTS domain. Both fusion proteins included an EK-cleavage site. (C and D) Analysis of Protein G-purified CysD–IgG fusion protein by SDS/PAGE and Western blotting under reducing (Red) and non-reducing (NonRed) conditions before (−) and after (+) EK cleavage. M, molecular-mass standards in kDa. The anti-Myc tag antibody labelled the CysD part, whereas the BαMIgG antibody recognized the IgG part. Silver, Silver staining.

Figure 2. BN-PAGE of purified CysD–IgG fusion protein

(A) BN gel of purified fusion protein before (−) and after (+) EK cleavage. (B) EK-cleaved fusion protein was loaded on to a Protein G column and the flow-through and eluate was analysed by BN-PAGE. M, molecular-mass standards in kDa. The gel lanes, including standards, shown together come from the same gel.

Figure 3. Gel filtration of the CysD–IgG fusion protein

(A) Purified fusion protein was analysed on a Superose 6 gel-filtration column. The Superose 6 column had a void volume of 0.843 ml, and the standards thyroglobulin (669 kDa), ferritin (440 kDa), aldolase (158 kDa) and ovalbumin (43 kDa) eluted at 1.315 ml, 1.494 ml, 1.629 ml and 1.753 ml respectively. (B and C) EK-cleaved fusion protein was loaded on to a Protein G column and the bound material (B) and flow-through (C) were analysed on a Superose 12 gel-filtration column. The Superose 12 column had a void volume of 0.831 ml, and the standards aldolase (158 kDa), conalbumin (75 kDa), ovalbumin (43 kDa) and ribonuclease A (13.7 kDa) eluted at 1.236 ml, 1.315 ml, 1.382 ml and 1.584 ml respectively. UV absorbance units are given in mAU. The content of collected fractions from major peaks (indicated by arrows) was analysed on reducing SDS gels (shown as insets) with silver staining. M, molecular-mass standards in kDa. The gel lanes shown together come from the same gel.

Figure 4. Analysis of EK-cleaved CysD(2TR)–IgG fusion protein by SDS/PAGE and gel filtration

(A) EK-cleaved CysD(2TR)–IgG fusion protein was analysed by SDS/PAGE and Western blotting under reducing (Red) and non-reducing (NonRed) conditions before (−) and after (+) sialidase treatment. The same material as in (A) was also analysed on a Superose 12 gel-filtration column (B). Elution times of standards are as specified in the legend for Figures 3(B) and 3(C). UV absorbance units are given in mAU. The content of collected fractions from major peaks (indicated by arrows) before (−) and after (+) sialidase treatment was analysed by reducing SDS/PAGE and Western blotting (shown as insets). The anti-Myc tag antibody labelled the CysD(2TR) part, whereas the BαMIgG antibody recognized the IgG part. M, molecular-mass standards in kDa. The gel lanes shown together come from the same gel.

Figure 5. Effect of pH and calcium on the dimeric state of the CysD domain

The CysD-containing fraction of EK-cleaved CysD–IgG fusion protein was analysed on a Superose 12 gel-filtration column at pH 8 and pH 5 with 5 mM EDTA (A) or at pH 8 with and without 10 mM CaCl₂ (B). UV absorbance units are given in mAU. Elution times of standards are as specified in the legend for Figures 3(B) and 3(C).

Figure 6. Analysis of disulfide bonds in the CysD domain by LC-ESI MS/MS

(A) Disulfide bond pattern in the CysD domain. Underlined numbers indicate cysteine residues forming intramolecular disulfide bonds. Solid horizontal lines indicate unambiguous disulfide bonds, whereas broken horizontal lines indicate theoretical disulfide bonds. Solid vertical lines represent the position of cleavage sites. (B) LC-MS separation of the peptide DVPIGQLGQTVVCDVSVGLICKNE with an internal disulfide bond. Peak number one represented the extracted ion chromatogram of the peptide cleaved N-terminal to the aspartate residue position 93. The second peak belonged to the non-cleaved form causing an increase in hydrophobicity.

Figure 7. LC-ESI MS/MS analysis of the CysD peptide with a C-mannosylation motif

A CysD-IgG-containing band from a non-reduced SDS gel was excised, in-gel digested with Asp-N and analysed by LC-ESI MS/MS. CID-fragmentation spectra of the CysD peptide DLSSPCVPLCNWTGWL in (A) with an intact disulfide bridge and in (B) after reduction with TCEP-HCl. Both fragmentation spectra showed the absence of C-mannosylation on the first tryptophan residue of the WXXW peptide motif.

Figure 8. Proposed model of CysD dimers in the fusion protein and the MUC2 gel

The CysD–IgG fusion protein is made up of two subunits and each is 50 kDa in size. These two subunits are covalently linked via the IgG Fc part (disulfide bridges) and form a covalent 100 kDa dimer. (A) Two covalent 100 kDa CysD–IgG dimers form a non-covalent 200 kDa tetramer via non-covalent interactions of the CysD domain. (B) Four covalent 100 kDa CysD–IgG dimers form a non-covalent 400 kDa octamer via non-covalent interactions of the CysD domain. In both cases the CysD domains form non-covalent dimers. (C) The MUC2 mucin forms a covalent gel via its N- and C-terminal domains. The N-terminus of MUC2 forms covalent trimers, whereas the C-terminus forms covalent dimers. The CysD domain forms non-covalent dimers. The CysD domain inserts non-covalent cross-links into the MUC2 gel thereby determining its pore size and gel properties.