Mucin gel assembly is controlled by a collective action of non-mucin proteins, disulfide bridges, Ca2+-mediated links, and hydrogen bonding

Abstract

Mucus is characterized by multiple levels of assembly at different length scales which result in a unique set of rheological (flow) and mechanical properties. These physical properties determine its biological function as a highly selective barrier for transport of water and nutrients, while blocking penetration of pathogens and foreign particles. Altered integrity of the mucus layer in the small intestine has been associated with a number of gastrointestinal tract pathologies such as Crohn's disease and cystic fibrosis. In this work, we uncover an intricate hierarchy of intestinal mucin (Muc2) assembly and show how complex rheological properties emerge from synergistic interactions between mucin glycoproteins, non-mucin proteins, and Ca²⁺. Using a novel method of mucus purification, we demonstrate the mechanism of assembly of Muc2 oligomers into viscoelastic microscale domains formed via hydrogen bonding and Ca²⁺-mediated links, which require the joint presence of Ca²⁺ ions and non-mucin proteins. These microscale domains aggregate to form a heterogeneous yield stress gel-like fluid, the macroscopic rheological properties of which are virtually identical to that of native intestinal mucus. Through proteomic analysis, we short-list potential protein candidates implicated in mucin assembly, thus paving the way for identifying the molecules responsible for the physiologically critical biophysical properties of mucus.

Figure 1

SDS-PAGE of ND- and GH-mucin preparations of porcine intestinal mucin after two-step CsCl isopycnic density gradient centrifugation. Electrophoresis was carried out using 4–12% NuPAGE Bis-Tris gel and blotted onto nitrocellulose membrane, images from the figure were cropped from different parts of the same gel with blotting performed with the same set of materials (see Supplementary Figure S11). ND- and GH-mucin preparations were compared with and without the inclusion of a reducing agent (DTT). Coomassie Blue stain was used to detect non-mucin proteins and PAS staining on nitrocellulose was used to detect glycoproteins. 30 μg of lyophilized mucin preparations were loaded in each lane. Molecular marker (lanes: 1, 4, 7, 10), reduced ND-mucin (lanes: 2 and 5), non-reduced ND-mucin (lanes: 3 and 6), reduced GH-mucin (lanes: 8 and 11) and non-reduced GH-mucin (lanes 9 and 12).

Figure 2

Microrheological characterization of ND-mucin solutions. (A) The ensemble averaged MSD curves of 0.5 μm colloidal particles in 1 ≤ c ≤ 50 mg/mL ND-mucin. Two distinct regimes were observed in ND-mucin based on the power law of the MSD curves. The trend lines show the log-log scaling of τ¹ and τ^0.5 as guides to highlight the transition from diffusive behavior in solutions up to 10 mg/mL, to sub-diffusive for concentrations 15 mg/mL and above. (B) The dependence of power law exponents, α (left Y axis), and diffusivity coefficients, M (right Y axis), for 1 ≤ c ≤ 50 mg/mL ND-mucin. A transition is observed between 10 and 15 mg/mL where the power exponent becomes independent and diffusivity decreases with increasing concentration. (C) Representative trajectories of individual 0.5 μm fluorescent colloidal particles embedded within mucin solutions of increasing concentration. (D) Specific viscosity (${\eta }_{sp}$) scaling of 1 ≤ c ≤ 50 mg/mL ND-mucin. (E) Comparison of probability distributions of 0.5 μm colloidal particles movement in Δx-distances in a lag time of 1 s and 2 s for 50 mg/mL ND-mucin solution. The solid and dashed lines are Gaussian distributions for 1 s and 2 s, respectively fitted to the data. For lag time of 2 s, the deviation from Gaussian distribution is observed in the increased deviations in the tail regions, as well as the increased probability in the area around the maximum. (F) The excess kurtosis parameter I₂(τ) plotted against mucin concentration for the lag times of 1 s and 2 s (dotted lines are parabolic spline to the data to guide the eye). For τ = 1 s and 2 s the degree of deviation was not measurable below concentrations of 10 and 20 mg/mL, respectively. A dashed line indicates I₂(τ) for water.

Figure 3

A structural model of mucus comprising rheologically heterogeneous domains blended together. (A) Periodic acid-Schiff staining of the pseudostratified columnar epithelium from the proximal small intestinal porcine mucosa which illustrates the presence of domains in physiological mucus. Note the pink staining of mucus from goblet cells into the intestinal lumen and the blue staining of nuclei. Scale bar 20 μm. (B) Visualization of the rheological heterogeneity, X- and Y-positions, of movement of individual 0.5 μm colloidal particles in 50 mg/mL ND-mucin solution. The size of circles indicates diffusivity and the color corresponds to the power law scaling of the MSD spectra.

Figure 4

Rheological behavior of ND-mucin solutions. (A) Comparison of the storage moduli, G′ (ω) of 10 ≤ c ≤ 50 mg/mL ND-mucin. For concentrations below 10 mg/mL, the oscillatory spectra were within the standard error from that of pure buffer. (B) The loss tangent (tan δ), a ratio of the loss modulus (G″) to storage modulus (G′) of 10 ≤ c ≤ 50 mg/mL ND-mucin. (C) Steady shear viscosity, η, and complex viscosity, |η*|, of 10 ≤ c ≤ 50 mg/mL ND-mucin versus shear rate, $\dot{\gamma }$ and angular frequency, ω, measured using parallel-plate geometry at a gap of 40 μm. Insert shows the plot of shear stress, σ, versus shear rate, $\dot{\gamma }$, of the same data. (D) Parameters of the Herschel-Bulkley equation; yield stress σ₀ and consistency index K_HB, evaluated as best fits to the shear stress, σ, versus shear rate, $\dot{\gamma }$, of 10 ≤ c ≤ 50 mg/mL ND-mucin. The flow index, n, had a mean value of 0.73 (SD ± 0.02).

Figure 5

The effect of denaturing agents on rheological properties of ND- and GH-mucin solutions. (A) Steady shear viscosity, η, and oscillatory viscosity, |η*|, of 30 mg/mL ND-mucin treated with 0.1 M DTT and 2 M GuHCl versus shear rate, $\dot{\gamma }$, and angular frequency, ω, measured using parallel-plate geometry at a gap of 40 μm. Insert shows the plot of shear stress, σ, versus shear rate, $\dot{\gamma }$, of the same data. (B) Steady shear viscosity, η, and oscillatory viscosity, |η*|, of 30 mg/mL GH-mucin treated with 0.1 M DTT and 2 M GuHCl versus shear rate, $\dot{\gamma }$, and angular frequency, ω, measured using parallel-plate geometry at a gap of 40 μm. Insert shows the plot of shear stress, σ, versus shear rate, $\dot{\gamma }$, of the same data. (C) Comparison of the storage modulus, G′ (ω; solid symbols), and loss modulus, G″ (ω; empty symbols), of 30 mg/mL ND-mucin treated with 0.1 M DTT and 2 M GuHCl. D.) The loss tangent (tan δ), the ratio of loss modulus (G″) to storage modulus (G′), of 30 mg/mL ND-mucin treated with 0.1 M DTT and 2 M GuHCl. E.) The ensemble averaged MSD curves of 0.5 μm colloidal particles in 30 mg/mL ND-mucin treated with 0.1 M DTT and 2 M GuHCl with the log-log scaling to be τ^0.51 τ^0.96 and τ^0.98, respectively.

Figure 6

The effect of Ca²⁺ and Ca²⁺ chelator (EDTA) on rheological properties of ND-mucin solutions. (A) Steady shear viscosity, η, and oscillatory viscosity, |η*|, of 30 mg/mL ND-mucin treated with 0.1 M Ca²⁺, 0.1 M Ca²⁺/0.1 M EDTA and 0.1 M EDTA versus shear rate, $\dot{\gamma }$, and angular frequency, ω, measured using parallel-plate geometry at a gap of 40 μm. Insert shows the plot of shear stress, σ, versus shear rate, $\dot{\gamma }$, of the same data. (B) Comparison of the storage modulus, G′ (ω), and loss modulus, G″ (ω), of 30 mg/mL ND-mucin treated with 0.1 M Ca²⁺, 0.1 M Ca²⁺/0.1 M EDTA and 0.1 M EDTA. (C) The ensemble averaged MSD curves of 0.5 μm colloidal particles in 30 mg/mL ND-mucin treated with 0.1 M Ca²⁺, 0.1 M Ca²⁺/0.1 M EDTA and 0.1 M EDTA with the log-log scaling to be τ^0.46, τ^0.12, τ^0.5 and τ^0.83, respectively. (D) The ensemble averaged MSD curves of 0.5 μm colloidal particles in 10 mg/mL ND-mucin treated with 0.1 M Ca²⁺, 0.1 M Ca²⁺/0.1 M EDTA and 0.1 M EDTA with the log-log scaling to be τ^0.85, τ^0.41, τ^0.89 and τ¹, respectively.

Figure 7

The effect of pH on rheological properties of ND-mucin solutions. The loss tangent $\left(\tan \delta =G″/G\prime \right)$, storage modulus (G′), and viscosity at $\dot{\gamma }=400$ s⁻¹ of 30 mg/mL ND-mucin solutions in 10 mM buffer at different pH values (pH 1–7).

Figure 8

Hierarchical assembly of intestinal mucin. Mucin molecules and oligomers (A) assemble into molecular clusters (B) which form a micro-domain network (C) that assembles into a yield stress fluid (D) that exhibits mucus-like rheological behavior (E). A proposed structure of mucin molecular cluster oligomer (B) formed through interactions between mucin and non-mucin proteins (colored globules). The interactions (F) are governed by disulfide bonds and Ca²⁺-mediated links, with hydrogen bonding playing a crucial supporting role.

Figure 9

Pie chart of proteins identified in ND-mucin and grouped according to their predominant areas of expression and/or origin. Most abundant proteins are listed in the figure, with numbers in brackets corresponding to the total number of hits in each group. For protein groups marked with asterisk (*), a full list of candidate proteins is provided in Table S1 (See ESI). Mucus components are designated based on their possible involvement in the formation of mucus granules within the secretory cells and/or their involvement in granule transport (to the membrane) and mucus extrusion process. Blood plasma and cell cytosol components are a broad group of proteins that originate from cytosol components of lysed/sloughed-off epithelial cells as well as blood transudate/blood contamination from the surgical procedure. A small number of identifications are related to digestive enzymes, and proteins involved in nutrient uptake, as well as proteins originating from animal diet (soybean).