Published ahead of print on September 7, 2006, doi:10.1165/rcmb.2006-0291TR
© 2007 American Thoracic Society DOI: 10.1165/rcmb.2006-0291TR
Mucin Granule Intraluminal OrganizationCystic Fibrosis/Pulmonary Research and Treatment Center, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina Correspondence and requests for reprints should be addressed to Juan Perez-Vilar, Cystic Fibrosis/Pulmonary Research and Treatment Center, School of Medicine, University of North Carolina at Chapel Hill, CB#7248, Chapel Hill, NC 27599–7248. E-mail: juan_vilar{at}med.unc.edu
Mucus secretions have played a central role in the evolution of multicellular organisms, enabling adaptation to widely differing environments. In vertebrates, mucus covers and protects the epithelial cells in the respiratory, gastrointestinal, urogenital, visual, and auditory systems, amphibian's epidermis, and the gills in fishes. Deregulation of mucus production and/or composition has important consequences for human health. For example, mucus obstruction of small airways is observed in chronic airway diseases, including chronic obstructive pulmonary disease, asthma, and cystic fibrosis. The major protein component in the mucus is a family of large, disulfide-bonded glycoproteins known as gel-forming mucins. These proteins are accumulated in large, regulated secretory granules (the mucin granules) that occupy most of the apical cytoplasm of specialized cells known as mucous/goblet cells. Since mucin oligomers have contour dimensions larger than the mucin granule average diameter, the question arises how these highly hydrophilic macromolecules are organized within these organelles. I review here the intraluminal organization of the mucin granule in view of our knowledge on the structure, biosynthesis, and biophysical properties of gel-forming mucins, and novel imaging studies in living mucous/goblet cells. The emerging concept is that the mucin granule lumen comprises a partially condensed matrix meshwork embedded in a fluid phase where proteins slowly diffuse.
Key Words: granule matrix • mucin granules • mucins • secretory granules • secretion
Five gel-forming mucins (MUC2, MUC5AC, MUC5B, MUC6, and MUC19) have been found in humans. These mucins share a similar structural organization (Figure 1A) characterized by (1, 2): (i) multi-domain polypeptide chains with thousands of amino acid residues; (ii) a large, centrally located region mainly consisting of threonine and/or serine-rich tandemly repeated sequences to which O-linked oligosaccharides are covalently bound; (iii) the presence of several under-glycosylated, Cys-rich domains that are conserved among different mucins; (iv) the formation of disulfide-linked species ranging from dimers to more than 16 oligomers (3); and (v) molecular polydispersity because of genetic polymorphism (e.g., multiple alleles encoding variable number of tandem repeats) (4), diverse O-glycosylation pattern, proteolytic processing, and/or different degree of oligomerization.
Mucin monomers are synthesized in the endoplasmic reticulum, where they are N-glycosylated and likely C-mannosylated, and form disulfide-linked dimers (1, 5) (Figure 1B). Once in the Golgi complex, dimeric mucin precursors are O-glycosylated, sulfated, and assembled into disulfide-bonded oligomers/multimers by interdimeric disulfide bonds (1, 6). Proteolytic processing may occur as well in these late secretory compartments (7). Fully processed mucins are secreted by constitutive and Ca2+-dependent regulated mechanisms that are not fully understood (8, 9). Mucins secreted by the regulated pathway are accumulated in the mucin granules of mucous/goblet cells (7, 10) until proper regulatory signals activate their exocytosis. In contrast to other regulated secretory vesicles/granules, the major secretory product in the mucin granule is also the key component of the intragranular matrix, as attested by the absence of mucin granules in intestinal "mucous" cells of mice lacking expression of mMuc2, the main gel-forming mucin gene expressed in the intestine (11). Hence, the mechanisms of intragranular mucin sorting and packaging (i.e., matrix formation) must be synchronized or simply function as a single mechanism during mucin granule biogenesis.
Biophysical data obtained with extracellular mucus/mucins, together with some basic principles of polymer physics, provide us with a rational view on the magnitude of the topological problems associated with the large sizes of mucin macromolecules. Based on the available evidence, we can assume that a typical gel-forming mucin (e.g., MUC5AC): (i) has  5,000 amino acid residues per monomer; (ii) is assembled into disulfide-linked, linear oligomers with 16 or more monomers each, and contour lengths (L)  8 µm; and (iii) forms flexible oligomeric random coils with a (cylindrical) Kuhn or effective segment—that is, the minimal polypeptide segment that can be considered rigid (12, 13), of length l 30 nm and diameter d 7 nm (14–16).
Because mucins are negatively charged, highly hydrophilic macromolecules, mucin random coils can be considered as self-avoiding (i.e., the chains cannot cross one another because of interchain repulsion) linear polymers (12, 13). In a diluted solution, the radius of gyration (Rg), defined as the average distance from all Kuhn segments to the center of mass of the oligomer (12, 13), and the pervade volume (V), defined as the spherical volume of radius Rg of each oligomeric chain, can be approximated by the following expressions:
L/l). For our 16-mer mucin oligomer, the corresponding Rg and V values would be 0.24 µm and 0.05 µm3, respectively, which are in the same range of the hydrodynamic sizes determined empirically (14–17).
Above a certain concentration, known as the threshold concentration or c*, self-avoiding linear polymeric chains begin to overlap and eventually form a heavily entangled network (12, 13). The dimensionless counterpart of c*, the volume fraction (
Thus, for the 16-mer mucin oligomer,
In concentrated polymer solutions, the distance between monomers in adjacent strands can be assessed by determining the correlation length (
For our test gel-forming mucin,
Irrespective of the nature of the inter-oligomer links, mucins in mucus gels form a three-dimensional network in which relatively large proteins can diffuse through their pores (14–16). Although the organization of intracellular mucins is virtually unknown, a reasonable picture can be proposed in view of the information reviewed in the preceding sections, results from early biochemical investigations, and recent live cell imaging investigations (see next section below). Studies with purified, de-glycosylated native mucins support the view that in the endoplasmic reticulum the largely uncharged and non–O-glycosylated mucin precursors (monomers and dimers) likely form globular but highly flexible random coils (16, 21, 22), which is consistent with the fact that the large central domains comprise repetitive sequences rich in Ser and Thr, and often Gly and Pro residues. Highly flexible chains can form more compact structures and consequently facilitate the transport of mucin precursors out of the endoplasmic reticulum.
Once mucin precursors reach the proximal compartments of the Golgi complex, the addition of GalNAc residues to the mucin Ser/Thr-rich central regions limits the rotation of the peptide bonds implicated; leading to less flexible strands (21). This property explains the relatively large Kuhn segment (l) of glycosylated mucins (15). Hence, initiation of O-glycosylation will reduce chain flexibility while increasing the inter-oligomer space, as
Important changes occur in the distal (trans-) Golgi compartments. First, intraluminal [H+] and [Ca2+] increase (24, 25), acidifying the lumen and increasing its ionic strength. Second, mucin dimers are assembled into multimers by inter-dimeric disulfide bonds (Figure 1B). The synthesis of larger macromolecules means decreased First, Ca2+-mediated interchain ionic interactions linking the sialic and sulfate residues in mucins are then possible (14, 26). These interactions could stabilize mucin entanglements or even link mucin oligomers alongside most of their contour lengths (i.e., through a zipper-like mechanism). If so, the asymmetry of the mucin macromolecules would lead to the formation of anisotropic networks (27). In addition, Ca2+ could bind other glycoproteins and anionic proteins to the nascent mucin matrix. Second, the increasingly lower pH and higher [Ca2+] might favor interchain links involving one or more of the cysteine-rich protein domains found in gel-forming mucins and other matrix proteins. In support of these two notions, several studies have shown that gel-forming mucins aggregate and/or self-oligomerize through protein-rich regions, O-glycans, or both in response to increased [Ca2+] and/or [H+] (14–16, 28–30). Third, a higher intraluminal [H+] and [Ca2+] is expected to diminish the repulsive forces among the negatively charged mucin oligomers, increasing the chances for forming more entanglements as mucins oligomerize (i.e., the ratio between L and V decreases). Therefore, it seems reasonable to assume that mucins in the trans-Golgi compartments and beyond are forming a polyanionic cross-linked network, which eventually will become the granule mucin matrix. By definition, a network of cross-linked polymers in a liquid medium is a gel, and importantly the volume of gels can be modulated by small changes in the environment, a property that might explain mucin intragranular condensation (14).
Gel Volume Transitions and Intragranular Mucins Studies with synthetic polyanionic gels have shown that under certain conditions gels are able to undergo a discontinuous volume (phase) transition between a swollen and a collapsed phase (31, 32). Gel volume transitions have three major properties: (i) they can result from small changes in the external conditions that affect the gel, including temperature, solvent, pH, ionic composition, and so on; (ii) the changes in gel volume can be massive (up to 1,000-fold); and (iii) they are reversible (i.e., a collapsed gel can abruptly swell on small changes in the environment). In polyanionic gels, the balance among three competing forces determines the (transition) critical point (32): (i) the positive osmotic pressure of counter ions, which tends to swell the gel; (ii) the negative pressure due to polymer–polymer affinity, which tends to shrink the gel; and (iii) the rubber elasticity of the polymer network, which makes the gel to expand (contract) whenever the gel is contracted (expanded). All these forces vary with temperature, pH, salts, and so forth. It is important to mention here, however, that phase transitions are not necessarily rapid or sudden (32). Further, the same gel can contract or expand by a continuous, rather than a discontinuous, process just by changing the nature of the fluid in which it is embedded. Though these studies have a solid mathematical and experimental basis, which anticipates the existence of a phase transition critical point irrespective of the gel nature, the question arises whether such critical points can be reached within dynamic cellular compartments like those involved in protein secretion. In principle, there are five major reasons to suggest that a volume transition takes place during mucin granule biogenesis. First, as discussed above, intralumenal mucin oligomers in the late Golgi compartments and beyond probably form a network by topological entanglements and noncovalent links. Second, increased [H+] and [Ca2+] in these cellular compartments (14) could be the triggering factors for the volume transition by increasing the affinity between mucin strands. Moreover, these luminal changes occur at the right intracellular site, that is, where the polyanionic nature of mucins is acquired. Third, a volume transition mechanism would allow reaching a high degree of intragranular condensation. Fourth, the biogenesis of the granule is a continuous but slow process that takes more than 12 h to be completed (J. Perez-Vilar, unpublished observations) and, therefore, there is a large period of time during which optimal conditions for a volume transition could develop. Fifth, granule exocytosis involves the expansion of the mucin matrix (14), and because volume transitions are reversible (32), a single mechanism would explain both the condensation and decondensation of the granule matrix.
The Matrix Core Paradigm Similarly, a mucin phase transition–based mechanism could be envisioned during mucin granule biogenesis (14). Thus, at some point during their transport through the Golgi complex and/or beyond, mucins aggregate, forming a condensed, polyanionic granular matrix. It is suggested that the mucin matrix core: (i) comprises a highly entangled anisotropic network cross-linked by Ca2+; (ii) is immobile and hence able to constrain the movement of secretory proteins, which allows the intragranular accumulation of large amounts of osmotically inert proteins (14, 33, 34); (iii) is mostly dehydrated (i.e., water molecules just form a hydration shell that surrounds the mucin strands); and (iv) is enclosed by a peripheral fluid phase containing free Ca2+ and other compounds in equilibrium with the outer region of the matrix core (40, 41) (Figure 2). This mobile pool of Ca2+ can be released from the granule by diffusion during signal transduction (42, 43). Consistent with this notion, electro-dense cores of variable sizes enclosed by a peripheral electro-lucent area can be seen by electron microscopy in different regulated secretory granules, including the mucin granule (25). In contrast, it has not been shown that purified mucins form a collapsed network when incubated with Ca2+ at low pH, which suggests that the mechanism of intragranular mucin assembly is more complex than a mere phase transition. In this respect, the existence of intragranular lipids that would increase the affinity between mucin strands has been postulated (14).
The Matrix Meshwork Paradigm The notion that the mucin granule lumen comprises a condensed, nondiffusible mucin core was directly tested by performing fluorescence recovery after photobleaching (FRAP) studies on mucin granules of HT29–18N2 mucous/goblet cells stably expressing a fusion protein between green fluorescent protein (GFP) and the CK-domain of MUC5AC that was accumulated in the granule (Figure 3A) (44). The cell line HT29–18N2 is a subclone of the human colonic cancer cell line HT29 that differentiates to mucous/goblet cells (45). Because the protein tracer was not covalently bound to the endogenous mucins (44), it could be used as a soluble FRAP probe to obtain information regarding the mucin granule intraluminal environment (46) (Figure 3B). Whether the granule lumen is altered by the presence of this protein cannot be totally disregarded, although the mucin secretion rate in these and the parental cells are similar, suggesting that this is not the case (J. Perez-Vilar, unpublished observations). In the FRAP studies, large mucin granules (diameters 1.1 µm) in live cells were visualized using a laser scanning confocal microscope. A small area in the granule equatorial plane was bleached at high laser power and the recovery of the fluorescence (i.e., the diffusion of the non-bleached molecules inside the bleached area) followed over time. The intragranular FRAP profiles obtained in most granules showed that intraluminal GFP-mucin: (i) was fractionated between a mobile and an immobile fraction; and (ii) homogenously, though very slowly ( 0.014 µm2 s–1), diffused into the bleached area (44) (Figure 3C). These results argued against the existence of a condensed, nondiffusible granule core, the major premise of the matrix core paradigm. As the tracer is a medium-sized compact protein, this suggests that the meshwork minimal average pore size probably is 10 nm. As we have discussed above, it is possible to have that range of pore size in a concentrated mucin solution (i.e.,  > *). Indeed, the average apparent diffusion coefficient of GFP-mucin in the endoplasmic reticulum (47) and the extracellular mucus (J. Perez-Vilar, unpublished observations) were 100-fold and > 170-fold larger than in the mucin granule, consistent with the existence of a partially, rather than totally, condensed granule mucin matrix that is expanded upon granule exocytosis (Figure 4). Whether this matrix is encircled with a fluid in which proteins can move faster, as suggested by the matrix core model, has not yet been investigated.
As expected for a meshwork-like organization, GFP-mucin intragranular diffusion and/or distribution (bound versus mobile fractions) can be altered by inhibitors of post-translational modifications pertinent to mucins (i.e., the granule matrix) (44), including O-glycosylation, sulfation, sialylation, and disulfide bond formation, as well as by inhibiting the activity of the vacuolar H+-ATPase (44). Three major suggestions derived from these studies. First, mucin O-glycans, and specifically the degree of mucin sialylation and sulfation (i.e., charge density), and likely the length of their O-glycans, affect the accessibility of the protein tracer, and possibly other luminal proteins, to the matrix protein–rich regions, and their rate of diffusion through the matrix pores. Second, intraluminal acidic pH is important to maintain the meshwork frame, and also for the binding of certain secretory proteins to the immobile matrix (44). Hence, conditions leading to altered degree of sialylation and/or sulfation of intragranular mucins, or intraluminal pH, could change to some extent the matrix organization. Third, protein–protein interactions are not absolutely required to maintain the granule organization (47), as judged by the effects of disrupting intragranular disulfide bonds, consistent with the notion that inter-strand links via multivalent ions (Ca2+) are sufficient to hold the matrix together at acidic pH (14, 37). It is noteworthy to mention, however, that disulfide reduction usually resulted in a complete immobilization of the tracer in the granule lumen (47). This observation has two, though not necessarily incompatible, explanations: (i) reduction of the protein probe disulfide bonds resulted in its de-folding and subsequent binding to the matrix components; and (ii) the meshwork average pore size drastically diminished after the disulfide reduction. The latter possibility implies a role for protein–protein interactions in keeping a relative large meshwork pore size and ultimately in the organization of the granule matrix. Indeed, protein–protein interactions might have a critical role during granule biogenesis. Thus, when the intragranular pH was raised, reduction of intragranular disulfide bonds resulted in granule matrix disorganization and partial discharge (47)—that is, the links holding the meshwork at acidic pH were insufficient as the pH was raised. These results were consistent with a role for protein–protein interactions during the early stages of matrix assembly, which takes place in a less acidic environment than the granule lumen. As the granule lumen is increasingly acidified, the Ca2+-dependent inter-strand links are quantitatively more important. Secretory products in many regulated secretory granules are not free to move (25). They remain immobilized inside a condensed polymer matrix that virtually abolishes their osmotic activity. This scenario is also postulated by the matrix core paradigm for the mucin granule. According to the meshwork model, however, some secretory proteins would be able to very slowly diffuse within the granule lumen and, therefore, an osmotic activity leading to granule steady swelling should be expected. This suggests the existence of an alternative mechanism for securing osmotic equilibrium of the mucin granules.
A Unifying Paradigm? Interestingly, fluorescence-free, centrally located small cores were observed in a small fraction of the granules of living mucous cells (44). In these granules, the rest of the lumen appeared to be constituted by a permeable meshwork as in most of the granules. If these cores are the counterparts of the cores observed in granules of fixed cells, a total collapse of the matrix meshwork in those small regions likely took place. This observation may suggest that the lumen in these granules, and likely the rest of the granules, is near the (volume transition) critical point. Interestingly, the configuration of synthetic polymer networks changes considerably as conditions come close to the critical point (32). Thus, collapsed regions are regularly inter-dispersed with entirely de-condensed areas, and ultimately forming a network with larger pores than an entirely expanded gel network.
The lack of functional cystic fibrosis (CF) transmembrane conductance regulator (CFTR) in airway epithelial cells leads to isotonic absorption of water, which sequentially results in mucus stasis and accumulation, bacterial infection, and lung inflammation (48, 49). Although it is increasingly clear that the viscoelastic properties of the mucus are drastically compromised because of lower water content and the presence of other macromolecules such as DNA and actin, the debate on whether CF mucins are intrinsically different still is alive (50). Since the 1970s to the present, disparate results have been reported regarding differences in CF mucins glycosylation and sulfation, and no widespread consensus has been reached (e.g., Refs. 51, 52, and the references therein). An intrinsic mucin defect (i.e., independent of bacterial infection and lung inflammation) would likely, but not necessarily (see below), mean that CFTR is expressed in goblet/mucous cells. Recent studies not only suggest that CFTR is part of the mucin granule membrane but also that its failure alters granule organization and, ultimately, matrix expansion after exocytosis (36, 37, 53). The mucin granules in CF airway glandular cells had less water and higher ion content than in granules of normal cells (53). Similarly, increased Ca2+ concentration was found in the mucin granule lumen of cultured gallbladder epithelial cells from Cftr(–/–) mice, while secreted mucins were less sulfated than their respective normal counterparts (36). In both cell systems, granule expansion upon exocytosis was abnormal, which suggests that intragranular organization is altered in CF mucous cells. Interestingly, the intragranular mobility of mucin-GFP in HT29 cells increased on protein sulfation inhibition (47), suggesting that the charge density altered the diffusion of the protein or that the pore size of the meshwork increased. Although other investigators have failed to detect CFTR in mucous/goblet cells (e.g., Ref. 54), the expression of this channel would not be strictly required to alter intracellular mucin glycosylation/sulfation. Because competition among different sulfo-, sialyl-, and fucosyltransferases is critical for the control of O-glycosylation elongation and termination reactions (55, 56), the biosynthesis of mucin-type O-glycans in a given cell depends on the specific glycosyltransferses/sulfotransferases expressed, and their levels of expression. Hence, it can be speculated that indirect de-regulation of the expression of one or more glycosyltransferases and/or sulfotransferases would suffice for the synthesis of abnormal mucins (57).
The studies in live mucous/goblet cells suggest the granule matrix would be embedded in a fluid in which secretory products very slowly diffuse, and interact with the matrix meshwork. Mucin O-glycans and their anionic (sulfate and sialic) groups would: (i) prevent the access to protein regions in the matrix, (ii) affect protein diffusion in the fluid by steric and charge hindrances, and (iii) establish intermolecular interactions among mucins oligomers/multimers through ionic (Ca2+) bonds. Other luminal proteins, especially sulfated proteoglycans, could also control protein diffusion through the matrix pores (47). Among the many questions that should focus future research efforts are: What kind of luminal events take place before granule formation? How is the matrix meshwork frame organized? What sorts of interactions happen between mucins before or during matrix assembly? What proteins, besides mucins, are required for matrix assembly? Does the mucin granule mature and what does this mean in the context of matrix structure? Is matrix assembly coupled with recruitment of other granule (membrane and soluble) proteins? What is the nature of the mechanism that maintains the granule osmotic equilibrium? and so on.
The author thanks Drs. P. Verdugo (Department of Bioengineering, University of Washington, WA) and T. Carter (Division of Molecular Neuroendocrinology, National Institute for Medical Research, UK) for critically reading the manuscript.
The research in the author's laboratory was supported by the Cystic Fibrosis Foundation (grants PEREZV04G0 and PEREZV06G0) and the National Institutes of Health-NIDDK (grant DK67404). Originally Published in Press as DOI: 10.1165/rcmb.2006-0291TR on September 7, 2006 Conflict of Interest Statement: J.P.-V. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form August 9, 2006 Accepted in final form August 22, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
GEL-FORMING MUCINS: STRUCTURE...
) (
0.32, that is, when the volume occupied by the mucin oligomers is 
