The following represents the generally accepted sequence involved in the biosynthesis and secretion of MUC-2.

1. Endoplasmic Reticulum

The formation of MUC2 starts with the biosynthesis of the protein monomer backbone (McGuckin2011). A von Willebrand factor-type D domain (vWF-D; hydrophobic and cystein-rich) is connected to a unique MUC2 sequence, and this combination is repeated three times (Figure 1). This N-terminal 'tail' is connected to a small 'PTS' domain (see Figure 4), which itself is bookended with cysteine-rich domains of unknown function. Next, a long 'variable number tandem repeat' (VNTR) domain represents the main part of the protein and is later adorned with hundreds of oligosaccharides (see next section). The MUC2 sequence is finished with another vWF-D, two vWF-C's and a cysteine knot domain, the latter of which is involved in covalent dimerisation. The highly folded, hydrophobic and cysteine-rich domains are inherently protease resistant.

Figure 1. Two diagrammatic concepts of the MUC2 apoprotein, prior to glycosylation (top from Dekker2002, second from McGuckin2011).

While still in the ER, N-glycosylation of up to 29 possible sites occurs at the C and N-terminal ends of the protein (McGuckin2011). These carbohydrate motifs are believed to confer further stability to the hydrophobic sequence and protect the protein from attack by proteases (Arike2016). It is also here that disulfide bond formation and extensive folding of the hydrophobic parts of the mucin occur. Head-to-head dimerisation is induced via disulfide formation at a cystine knot adjacent to the C-terminal domain (Figure 2), prior to the now-homopolymeric glycoprotein moving to the Golgi apparatus.

2. Golgi Apparatus

Many of the unique properties of mucins come from the carbohydrates that are added to the protein scaffold. The carbohydrate portion makes up about 70-80% of the weight of the glycoprotein (McGuckin2015). The Golgi apparatus is responsible for this O-glycosylation and this organelle hosts a number of membrane-bound glycosyltransferases, glycosidases and nucleotide sugar transporters that allow for the more-or-less orderly elongation of the oligosaccharide sidechain (Stanley2011). Despite the ordered process, a number of physiological conditions, including pH, cellular stress, and cell signalling, may affect the maturation process (Stanley2011). Each person, tissue and cell might present a different set of oligosaccharides arrayed on the same apomucin backbone. This can have consequences for the binding to bacteria and other foreign objects.

Oligosaccharide biosynthesis always starts when an N-acetylgalactosamine is attached to a threonine hydroxyl group in the PTS region. This region is named as such because mucins normally contains a high number of proline, threonine and serine amino acids in this region. MUC2, though, rarely uses serine; instead, the variable number tandem repeats domain (consensus sequence: PTTTPITTTTTVTPTPTPTGTQT (Figure 2)) is dominated by threonine amino acids (14/23 residues are threonine) (Svensson2018).

Figure 2. One possible (2D) conformation of the PTS peptide region (or variable number tandem repeat) of MUC2. Arrows show potential attachment points for O-glycans.

There are over 100 mucin oligosaccharides described so far (see the oligosaccharide table for MUC2 O-glycosylation of the mucin protein backbone), ranging in size from simple disaccharides to complex 12-sugar entities (Larsson2009, Tailford2015, Podolsky1985). Monosaccharides are invariably limited to combinations of N-acetylgalactosamine (GalNAc), N-acetylglucosamine (GlcNAc), fucose (Fuc), galactose (Gal) and N-acetyl neuraminic acid (NeuAc). Some sugar residues are also sulfated, which apparently slows hydrolysis of the oligosaccharide. Sulfate and NeuAc possess a negative charge and are considered acidic residues.

As far as combinations of the monosaccharides are concerned, biosynthesis appears to follow several rules:

1. As mentioned above, GalNAc is always the point of attachment to the protein;

2. NeuAc, Fuc and sulfate are terminal, meaning once attached they themselves are not further glycosylated;

3. Apart from the starting GalNAc, only Gal and GlcNAc are 'links' in the oligosaccharide 'chain'. There are no incidences of either of these monosaccharides occurring sequentially, meaning all linking sugar residues alternate.

It appears as though the most important O-glycans in terms of mucus barrier maintenance and protection are based on Cores 1 and 3 (Figure 3, see also 100 mucin oligosaccharides table). The structural configuration of these cores limits degradation by bacterial proteases (Bergstrom2017) and reduces spontaneous colitis in mice models (Bergstrom2017, Sommer2014).

Figure 3. Structural and graphical representation of key disaccharide cores 1 and 3. O-glycans bases on these structures improve mucus barrier integrity and reduce spontaneous colitis in mice models (Bergstrom2017, Sommer2014).

3. Storage Granules and Secretion

Once biosynthesis is complete, the homopolymeric mucin is packaged into secretory granules. These migrate toward the apical surface of the goblet cell where they are expelled from the cell, or await further signals. As depicted in Figure 4, cells move from the base of the crypt to the villus tip as they mature. As such, goblet cells are most productive near the tip and a continous stream of new polymeric MUC2 is released beneath the existing firm mucus layer.

Figure 4. A cartoon (left) of a mucus-producing crypt. Cells develop as they are pushed up toward the tip of the villus; goblet cells become distended with accumulated secretory granules full of condensed MUC2 (Johansson2011, Grondin2020). Upon secretion, this mucin swells dramatically before interacting with other MUC2 strands in the firm layer of mucus. Firm mucus can act as a barrier to bacteria and large viruses as well as insoluble food particles (right). Defensive proteins associate with mucins and further immobilise pathogens. Finally, loosened mucin can be a useful carbohydrate and protein food resource for bacteria of the mucin consortium.

Upon leaving the goblet cell, MUC2 rapidly absorbs water and swells to between 100-1000 times its initial volume and these polymeric forms can reach 1-10 µm in length. Neighbouring enterocytes release bicarbonate, which as been shown to be necessary for the mucin unfolding process (Johansson2016). It is generally believed the long mucin chains interact with each other to form aggregating stringy structures (non-convalent hydrophobic interactions Ambort2011) and / or trimerised forms with net-shaped arrangements (Godl2002, McGuckin2011, Figure 5).

The loose mucus layer is replenished continuously and relatively rapidly. This layer is estimated to contain 25% of the amount of MUC2 found in the firm layer in a given volume (Johansson2008), but how the transition from the compact and insoluble form to the water soluble one remains elusive. Some experiments suggest disruption of disulfide bonds and subsequent extensive hydration is the likely cause (Johansson2008).

Figure 5. The MUC2 apoprotein is biosynthesised in the ER (left), where it undergoes dimerisation and N-glycosylation (McGuckin2011). Transfer to the Golgi apparatus allows further extensive O-glycosylation. The homodimer is encapsulated into secretory granules whereupon further trimerisation occurs under low pH and high calcium concentrations (Johansson2011) (picture modified from McGuckin2011).