Double-Globular Structure of Porcine Stomach Mucin: A Small-Angle

3216

Biomacromolecules 2008, 9, 3216–3222

Double-Globular Structure of Porcine Stomach Mucin: A Small-Angle X-ray Scattering Study Emanuela Di Cola,† Gleb E. Yakubov,‡ and Thomas A. Waigh*,§ ESRF, 6 Rue Jules Horowitz, BP 220, F-38043, Grenoble, France, Unilever Corporate Research, Colworth Park, Sharnbrook, Bedfordshire, MK44 1LQ, United Kingdom, and Biological Physics, School of Physics and Astronomy, Schuster Building, University of Manchester, Oxford Road, Manchester, M13 9PL, United Kingdom Received July 17, 2008; Revised Manuscript Received August 27, 2008

We present evidence from small-angle X-ray scattering synchrotron experiments that porcine stomach mucin (MUC6) contains a double-globular comb structure. Analysis of the amino acid sequence of the peptide comb backbone indicates that the globular structure is determined by both the charge and hydrophobicity of the amino acids and the placement of the short hydrophilic carbohydrate side chains (∼2.5 nm). The double-globular structure is, thus, due to a block copolymer type hydrophobic polyampholyte charge instability in contrast to the random copolymer instabilities observed previously with synthetic polyelectrolytes (particularly polystyrene sulfonates). Careful filtering was required to exclude multimonomer aggregates from the X-ray measurements. A double Guinier analysis (Rg ∼ 26 nm) and a double power law fit are consistent with two globules per chain in low salt conditions. The average radius of the globules is ∼10 nm in salt-free condition (double Guinier fit) and the average distance of intrachain separation of the globules is 48 nm. The addition of salt causes a significant decrease in the radius of gyration (14 nm 100 mM NaCl) of the chains and is attributed to the contraction of the glycosylated peptide spacer between the two globules (the globular size continues to be ∼10 nm and the globule separation is then 18 nm). Without salt, the scaling of the semidilute mesh size (ξ) as a function of the mucin concentration (c) is ξ ∼ c-0.45compared with ξ ∼ c-0.28 in high salt conditions, highlighting the globular nature of the chains. In contrast, hydrophilic flexible polyelectrolytes have a stronger concentration dependence of ξ when excess salt is added.

Introduction The molecular structure of mucins continues to be a challenging question due to their noncrystallinity, chemical heterogeneity, long-range charge effects, and their tendency to associate.1 In the current article we examine purified MUC6 mucin extracted from porcine stomachs, which is sold commercially as a saliva substitute (Orthana mucin, Mw ∼ 546 KDa) to be used after chemotherapy for cancer conditions. Problems in the expression of MUC6 in the stomach are thought to be connected with a range of gastric medical conditions such as Crohn’s disease. There is a large interest in the polymer physics of both natural and synthetic comb polyelectrolytes and polyampholytes due to their unusual properties as viscosity modifiers.2-5 Furthermore thin films of comb polyelectrolytes have an unusual normal stress difference in shear, which has a major impact in their efficacy in reducing frictional effects.6 Recent advances in molecular biology (genetic sequencing and homologue classification) are allowing a more coherent picture of the evolution and structural similarities of the mucin family of molecules to be constructed.7,8 Human MUC2, MUC5, MUC6, Otogelin, and von Willebrand factor all contain similar globular von Willebrand motifs at either end of the molecules, with the major changes in sequence and length being restricted to the glycosylated peptide regions, that is, the globular end * To whom correspondence should be addressed. E-mail: t.a.waigh@ manchester.ac.uk. † ESRF. ‡ Unilever Corporate Research. § University of Manchester.

caps are similar for all these molecules based on their genetic sequence. The nanoscale structure of human von Willebrand factor is better understood than that of the gel forming mucins, consisting of an asymmetric dumbbell morphology with a large 22 × 6.5 nm globule at one end of a 34 × 2 nm cylindrical rod with a small 5 nm diameter globule at the other end.9 The carbohydrate side chains in mucin are rather short (∼2.5 nm) and, thus, relatively rigid. The side chains play a role in splitting each molecule into two separate globules and increasing their water solubility. Previously we have obtained evidence for the multiglobular structure of MUC6 mucin using complimentary techniques such as transmission electron microscopy (TEM), atomic force microscopy (AFM), dynamic light scattering (DLS), static light scattering (SLS), zeta potential, rheology, and microrheology techniques.10,11 These studies are in broad agreement with other AFM measurements on MUC6 mucin.12 Furthermore, reconsideration of previous small angle neutron scattering (SANS) data from our group is also consistent with the multiglobular structure.13 Liquid-like scaling for the interparticle correlation length (ξ ∼ c-1/3, where ξ is the correlation length and c is the concentration) is observed above the overlap concentration c* of the molecules (c* meaning the concentration at which spheres containing one molecule with a diameter equal to the molecular length begin to overlap) in contrast to the scaling expected for flexible and rod-like polyelectrolytes (ξ ∼ c-1/2) or neutral flexible polymers (ξ ∼ c-3/4).14 This correlation length appears to represent the average distance between globules on neighboring chains.

10.1021/bm800799u CCC: $40.75  2008 American Chemical Society Published on Web 09/27/2008

Structure of Porcine Stomach Mucin

Figure 1. (a) Analysis of the genetic sequence of human and porcine MUC6 mucin clearly indicates the presence of separate glycosylated and nonglycosylated domains.23 (b) Schematic diagram of the asymmetric dumbbell structure of a MUC6 mucin unimer. (c) Schematic diagram of a MUC6 dimer with a disulphide linkage.

Some AFM evidence has been presented that MUC5 mucins from the human cervix also contain a multiglobular pearl necklace structure in solution,15 although on a larger length scale than with MUC6 (100 nm globules compared with 10 nm). This indicates that a multiglobule pearl necklace morphology may be a general structural motif that occurs across a broad range of the gel forming mucin family (in contrast to the epithelial mucin family). Association between MUC6 mucin monomers to form gels thus appears to be a combination of the merger of individual globular hydrophobic regions and disulphide linkages between peptide monomers to form dimers. Orthana mucin samples are chemically well characterized glycoproteins with short carbohydrate side chains.10,11 We examined Orthana mucin structure using small-angle X-ray scattering (SAXS) at the high brilliance beam line of the ESRF (ID02, Grenoble, France) over a wide range of length scales in dilute and semidilute solutions, with and without the addition of salt. Care was taken to avoid damaging the glycoproteins with the X-ray radiation using a flow through capillary by moving the sample continuously during beam exposure in tentime frames which were eventually averaged. Recent simulations have predicted multiglobular structures in synthetic polyelectrolyte combs in poor solvent conditions,16 block copolymer hydrophobic polyelectrolytes, randomly charged polyampholytes,17 and randomly charged hydrophobic polyelectrolytes.18,19 An advantage of making the link with synthetic hydrophobic polyelectrolyte properties is the possibility to reinterpret some of the unusual phenomena observed with mucin in terms of the properties of simpler synthetic materials, for example, thin film formation occurs in both systems at both solid/water and air/water interfaces, and there are some unusual osmotic pressure phenomena.20,21 The synthetic polyelectrolyte literature also has the advantage that it provides some quantitative models to describe the physical behavior observed.22 Figure 1a shows the peptide sequence of human MUC6 mucin molecules from genetic information, decomposed in terms of the position of the charged groups, hydrophobic amino acids, glycosylated sites (positions of side-chains), and proline (chaotropic) locations.23 The most important amino acids for determining the morphology of the mucin molecules are the glycosylated sites and the proline regions. Both these varieties of amino acids occur in large numbers throughout the linker

Biomacromolecules, Vol. 9, No. 11, 2008

3217

region (amino acids 1100-1700) that connects the two globular domains, which tends to disrupt the fusion of the two end groups into a single large globule. The amino acid sequence of the two end groups is homologous to that of von Willebrand factor, allowing accurate identification of their folding patterns. Figure 1b shows a schematic diagram of the structure of a single MUC6 chain, while in Figure 1c the native dimerized structure (with a disulphide linkage) is depicted. In the current article we are predominately considering the structure of undimerized molecules, that is, mucin monomers. We present new evidence for globular structures in MUC6 comb glycoproteins using small-angle X-ray scattering and a simple robust analysis of the protein sequence. The physical behavior of the glycoproteins is analyzed in terms of the physics of the hydrophobic comb polyampholyte chains. The size of the globules measured is very close to that found in previous measurements with similar samples using dynamic light scattering (DLS), atomic force microscopy (AFM), and transmission electron microscopy (TEM) measurements (∼10 nm) and also that expected from the genetically determined peptide sequence. The radius of gyration measured for the whole chains is smaller than that in previous light scattering measurements (26 nm compared with 50 nm) but is in good agreement with smallangle neutron scattering measurements from other MUC6 samples.24 We ascribe these slight differences in whole chain size calculated using different scattering techniques to rigorous filtering of the samples in the current experiments to remove multichain aggregates and the contrast mechanism of the X-ray technique highlighting different length scales in the samples. Furthermore we present SAXS data in the semidilute regime for chain overlap that highlights the globular nature of the mucin chains as salt is added to the solutions.

Experimental Section SAXS measurements were performed at the high brilliance beam line ID02, European Synchrotron Radiation Facility (ESRF), Grenoble, France. A combination of three sample to detector distances (1.5, 3, and 5 m) was used to explore a wide q-range, spanning from 2 × 10-2 nm-1 up to 3 nm-1, using an X-ray wavelength (λ) of 0.1 nm, where q ) (4π/λ) sin(θ/2), and θ is the scattering angle. The SAXS detector was a high sensitivity fiber-optic coupled CCD (FReLoN) placed in an evacuated flight tube. The incident and transmitted intensities were also simultaneously recorded. The two-dimensional SAXS patterns were normalized to an absolute scale using a water standard and azimuthally averaged to obtain the differential scattering cross section per unit volume (I(q)). The solutions were measured as they flowed through a temperaturecontrolled capillary (diameter ∼ 2 mm), which allowed the radiation damage to be minimized while obtaining a reliable background subtraction. The sample and solvent scattering were measured under the same conditions. Pharmaceutical grade porcine gastric “Orthana” mucin was obtained from A/S Orthana Kemisk Fabrik. Samples were purified as described previously10,11 and were prepared in ultrapure water (pH 7) 24 h before each experiment by diluting from a stock solution. Mucin concentrations were examined in the range 1.063-300 mg/mL. Both low (0 mM NaCl) and high (100 mM NaCl) concentration salt solutions were explored in detail. Samples were compared before and after filtering with low protein binding filters (hydrophilic filters, cellulose membranes with a pore size of 0.45 µm) to observe the effect of aggregation phenomena on the X-ray scattering patterns. The concentration of the samples was measured using UV absorption and by the direct comparison of the absolute SAXS intensity level between the filtered and unfiltered specimens.

3218


Cola et al.

Figure 2. SAXS intensity profiles (I(q) versus q) for different concentrations of mucin in salt-free solution: (a) after filtration and (b) prior to filtration. (c) The forward SAXS scattering (I0) as a function of mucin concentration. The arrow indicates the overlap concentration between brushes (cbrush).

Results Figure 2a,b shows SAXS patterns at different concentrations of mucin in salt-free solutions. A clear upturn is observed at low-q when the solutions were not filtered prior measurements (Figure 2b). This excess of scattering in the low-q region is attributed to hydrophobic association of the mucin chains. The concentration behavior of the extrapolated forward intensity (Io) is reported in Figure 2c (Io ∼ c0.54 and Io ∼ c0.1). Clearly, a discontinuity is found around ∼20 mg/mL, which is attributed to the semidilute brush overlap concentration (cbrush, Figure 2c), in agreement with previous neutron experiments,13 that is, the overlap of the carbohydrate side chains in the glycosylated region of the peptide. The semidilute overlap concentration (c*) occurs at lower mucin concentrations and is known to be 1.54 mg/mL from previous viscosity/light scattering measurements.10 In Figure 3a,b the scattering intensity for 1 mg/mL mucin concentration (dilute concentration regime) is shown in low (0 M NaCl) and high salt (100 mM NaCl) conditions. In the small angle regime where q < 2π/Rg a simple Guinier relationship holds for the scattered X-ray intensity (I (q))

I ) I0e-

q2R2g 3

(1)

Where I0 is a constant related to the molecular mass of the particles and Rg is the radius of gyration. The Guinier equation depends on a small angle expansion for its derivation, and therefore, q must be sufficiently small for it to be applied. Practically with polymeric materials, a Guinier regime can be used when a good linear fit to the natural logarithm of the intensity versus q2 plots (Ln I(q) versus q2) occurs at low q.

The Guinier plots for these two data sets are shown in the insets to Figure 3. The Guinier analysis shows a collapse of the protein conformation in excess of salt, with the radius of gyration dropping from 26 nm in 0 mM NaCl to 14 nm in 100 mM NaCl. In Figure 4a,b, we report an extended Guinier type analysis (Ln I(q)/q versus q2)25 of the internal structure of the mucin molecules providing evidence for a dumbbell type morphology. The linear regime at high q is identified with the internal globular structure and can be used to measure the size of the globules (see Discussion). Figure 5 shows the power law exponents of fits to the 1 mg/ mL mucin SAXS data without salt. Two power laws are required to describe the data indicating two structural levels in the samples.26 The fractal exponent (df, I(q) ∼ q-df) is 1.8 at low q and 3.6 at high q. Figure 6 shows the effect of the addition of salt on the filtered mucin samples (1 mg/mL) on the radius of gyration from a Guinier analysis (0 to 1 M NaCl). There is a sharp contraction of the structure that levels off in high salt concentrations at 12 nm. In Figures 7 and 8 the structure peak obtained by the division of the intensity by the experimental mucin form factor (c ≈ 1 mg/mL) is reported in salt-free and high salt conditions, respectively. The scaling of the structure factor peak (q*) as a function of the concentration shows a reduction of the exponent from q* ∼ c0.45 to q* ∼ c0.28, as reported in the insets of Figure 7 and 8, respectively. With the low salt mucin samples (Figure 7) the structure factor peak increases in height with mucin concentration and then reaches a plateau at the highest mucin



3219

Figure 3. SAXS intensity profiles (I(q) versus q) for low concentration mucin (c ≈ 1 mg/mL) for (a) low salt and (b) high salt concentrations. The corresponding Guinier plots (ln I(q) versus q2) are shown in the insets. Schematic diagrams of the postulated globular structures at high and low salt concentrations are also shown.

concentrations (a related phenomenon is seen in the forward scattering intensity in Figure 2c). In contrast, high salt mucin samples have a monotonic increase of the structure factor peak height with mucin concentration (Figure 8).

Discussion The comparison of the SAXS data from filtered and unfiltered solutions of Orthana mucin (Figure 2) shows the presence of large protein aggregates in unfiltered solutions (an upturn in the low-q region of Figure 2b), which were also present in previous size measurements from intrinsic viscosity data. It is difficult to size the aggregates in unfiltered samples accurately with SAXS because the Guinier regime is not achieved over the range of q explored. Therefore, we can only conclude that the protein aggregates have a radius of gyration beyond the q range probed (>600 nm). The aggregation of the chains is attributed to hydrophobic association of the mucin molecules. Disulphide interactions are inhibited by the extraction procedure and dimer formation (Figure 1c) is insufficient to motivate the large aggregate length scales observed. To characterize the shape of mucin in solution in both saltfree and excess salt conditions, we first used an approach based on a simple Guinier analysis, eq 1. For a two globule chain (dumbbell) structure with little scattering from the linker units it has been shown that the radius of gyration is related to the radius of the globules (R, assumed equal for both globules) and the distance between the globule centers (L).25

R2g ) R2 +

L2 4

(2)

We postulate this double-globular structure on the basis of microscopy data10,11 and knowledge of the mucin peptide sequence, which is consistent with the existence of two hydrophobic globular regions separated by a linear hydrophilic glycosylated section of peptide (Figure 1b). The radius of gyration of the whole chains calculated from Guinier analysis of the low-q region (insets of Figure 3a,b) experiences a large decrease in value upon the addition of salt (from 26 nm to 14 nm). Assuming no change in the size of the globules (established later in the discussion), eq 2 implies that the decrease in size is due to the glycosylated spacer length decreasing. Schematic diagrams of this effect are shown in Figure 3a (salt-free) and Figure 3b (100 mM NaCl). Analysis of the globule form factors indicates that the high salt case is not accompanied by significant globular fusion. The decrease in size is attributed to the increased screening of the electrostatic interactions of the peptide backbone because the carbohydrate side chains are thought to be uncharged in this particular variety of MUC6 mucin (both charged and uncharged side-chain versions of MUC6 are possible). A simple robust analysis of the small-angle data using a double Guinier fit was developed previously for the analysis of the dumbbell structure of Troponin C.25,27 Two clear features can be observed in the small-angle scattering curves at low mucin concentrations in agreement with the Tropinin C data, simulations of multiglobular polyelectrolytes19 and generic

3220


Cola et al.

Figure 6. Radius of gyration from Guinier fits as a function of salt concentration for c ≈ 1 mg/mL mucin. The dashed line is a guide for the eye.

Figure 4. Ln (I(q)/q) versus q2 plots for low concentration Orthana mucin (1 mg/mL) for (a) salt-free solution and (b) in the presence of 100 mM NaCl. The solid line represents the best fit to eq 3 leading to R1 ∼ 9.8 and 10.8 nm in salt-free and in 100 mM NaCl solutions, respectively.

Figure 5. Power law fits to the Orthana mucin at c ≈ 1 mg/mL at low salt concentrations. The fits indicate the two levels of structure in these materials.26

models of a material with two structural levels.26 The second region at higher q (smaller length scales) characterizing the globule morphology can be interpreted using a plot of ln I(q)/q versus q2.

1 ln[2πI(q) ⁄ q] ) ln[2n21L] - R21q2 3

(3)

Where n1 is the scattering contrast of the globular domains relative to the solvent, R1 is the radius of the globules and L indicates the average distance between the globules. The plot

gives a value of R1 of ∼9.8 nm in low salt and slightly larger value (R1 ∼ 10.8 nm) in the presence of 100 mM NaCl (Figure 4). Using eq 2, it is then possible to evaluate the average distance between the globules; we obtain in salt-free conditions a value for L of about 48 nm, which dramatically drops as salt is added to the solution (L ∼ 17 nm at 100 mM NaCl). This analysis confirms that the decrease in size of the mucin molecules is completely due to the change in the distance between the globules. Scattering from multiglobular structures have been described in detail with partially charged randomly sulfonated polystyrenes.28 The current SAXS data closely resembles that from the polystyrene sulfonates (particularly in a Kratky plot). Furthermore, with identical MUC6 mucin samples previous DLS results indicated an internal chain length (globule) scale of radius 9 nm (assuming completely free motion).10 The DLS results are thus in good agreement with the SAXS determined radius of gyration (∼10 nm) for the globular structures at either end of the mucin molecules. A discrepancy on the value of the radius of gyration of the whole molecule is found when compared with previous SLS/ viscosity studies (Rg ∼ 50 nm). This is probably due to the mild aggregation phenomena observed with the mucins,10 coupled with the resolution limit of SLS/viscosity techniques (although DLS can perform reliably down to 1 nm). In this respect, we believe the SAXS/SANS results are more accurate, since both the techniques are sensitive to smaller length scales. Reasonable agreement is found between the values of the radius of gyration from SAXS data and previous SANS studies of MUC6 mucin from two different sources: 34 nm and 29 nm for Sigma Aldrich and Havard samples in low salt, respectively. The slight differences between the reported sizes can be reasonably attributed to the different preparation procedures, solvent quality of D2O compared with H2O or different contrast mechanisms. Previous SANS data for the form factor of individual mucin molecules was interpreted with a cylindrical model.13 However, the dumbbell structure is still compatible with these data, since it implies a cylindrical symmetry. Further experimental evidence of the dumbbell structure is given in Figure 5 by the double power law fit to the form factor at low mucin concentrations. The two fractal exponents 1.8 and 3.6 correspond to extended structures at large length scales (rodlike dumbbells) and a compact structure at short length scales (globules), respectively. The expected exponent (df) for a rod is 1 and for a globule it is 4. This gives us increased confidence in the application of the dumbbell model to describe the SAXS data.



3221

Figure 7. Effective structure factor (S(q)) for mucin at different concentrations in salt-free solutions. The scaling of the structure peak (q*) as a function of the concentration is reported in the inset. A schematic diagram showing the correlation length (ξ) defined in the mucin samples is also shown.

Figure 8. Effective structure factor (S(q)) for mucin at different concentrations in 100 mM NaCl. The scaling of the structure peak (q*) as a function of the concentration is reported in the inset. The schematic diagram of the correlation length (ξ) defined in the mucin samples is also shown.

The effect of the addition of salt to the solutions is reported in Figure 6, which shows the radius of gyration of mucin from Guinier plots as a function of salt concentration. The size of the chains rapidly decreases with the addition of salt until a plateau value is reached; presumably this is the points of contact of the two globules or the maximum charge screening of the glycosylated spacer (maximum degree of chain contraction). To analyze the data at higher mucin concentrations (c>c*) we will assume factorization of the intensity in terms of the form factor and the structure factor (strictly only true in the case of monodisperse systems and for total decoupling between the structure and the form factor, but it can be a reasonable approximation with polyelectrolytes29 and is commonly used in liquid state theory for globular particles at high concentrations)

I(q) ) P(q)S(q)

(5)

where P(q) is the form factor and S(q) is the effective structure factor. For low mucin concentrations we can safely assume S(q) ≈ 1 and, thus, P(q) is adequately described by the I(q) at c ) 1 mg/mL. The scaling of the position of the structure factor as a function of mucin concentration for Orthana mucin at low salt concentrations (q* ∼ c0.45) is intermediate between the two predicted behaviors q* ∼ c1/2 and q* ∼ c1/3, typical of hydrophobic polyelectrolytes but in contrast to previous SANS experiments (q* ∼ c1/3 for previous salt-free experiments), possibly demonstrating the existence of residual salt in the commercial Sigma specimens used for SANS investigations.24 The change in the scaling of the average mesh size (ξ ) 2π/q*) upon the addition of salt (ξ ∼ c-0.28) is opposite to that predicted for polyelec-

trolytes in good solvents, where the exponent is expected to increase in size approaching the good solvent conditions (ξ ∼ c-3/4).22 Therefore, we conclude that the mucin chains have an extended character in semidilute solutions at low salt concentrations above the overlap concentration as shown in the schematic diagram (Figure 7), but this collapses into a more globular structure upon the addition of salt giving liquid-like scaling of the interparticle structure factor (Figure 8, ξ ∼ c-1/3, in agreement with previous SANS experiments). This also explains the change in scaling of the peak height at high concentrations in low salt solutions, which would be due to the decrease in the osmotic compressibility of the solutions as the more extended chains overlap. However, the prefactor for the liquid-like scaling is an order of magnitude different between SAXS and SANS measurements, which is currently unexplained (it indicates a large difference in the molecular weight of the two samples, Sigma MUC6 mucin versus Orthona MUC6 mucin). It is possible to calculate the molecular weight (Mw) of the scatters using the prefactor of liquid-like scaling of the structure factor according to eq 6

[

q*(nm-1) ) 2π10-7 c(g ⁄ cm3)

NA Mw

]

1⁄3

(6)

This equation is a statement of the space filling requirement of the liquid-like arrangement of the mucin molecules. However careful analysis shows that the molecular weight calculated (25 kDa) for the current samples differs from the expected 546/2 kDa (two globules per chain) by an order of magnitude. A possible resolution of this apparent paradox is that the SAXS measurements are highlighting the distance between intrachain

3222


globules (the mesh sizes are close to the intraglobule size expected) explaining the reduction of the scaling exponent (-0.28) below that of liquid-like scaling (-0.33), invalidating the quantitative application of eq 6. Examination of the peptide sequence (Figure 1) implies that the MUC6 globular structure is due to its sequence and not primarily the Rayleigh charge instability observed in random hydrophobic polyelectrolytes (breakup of a liquid droplet due to the competition between charge repulsion and surface tension).18,19 In terms of synthetic polymer chemistry, the dumbbell structure seems to be due to the block copolymer nature of the peptide, that is, the coarse grained structure corresponds to a nonsymmetric triblock copolymer (type ABA) consisting of two end-cap hydrophobic domains of rather different size linked by a long glycosylated hydrophilic block. The currently accepted sequence for human MUC6 mucin is described in Figure 1. The size of the glycosylated peptide spacer between the two globules corresponds to a range of 530-580 amino acids, so the end-to-end length (L) for a completely flexible chain (L ∼ bN1/2, with b equal to 0.8 nm) is expected to be about 20 nm, while for a rigid chain (L ∼ bN) L should be of the order of 438 nm. The actual intrachain length identified in the current SAXS measurements between the two globules (48 nm) falls in between these two extremes, tending toward the flexible result in agreement with previous rheology experiments.10 The peptide sequence is then only compatible with one large globule followed by a heavily glycosylated brush region and a subsequent small globule. The sequence of MUC5 has also recently been examined (both MUC5a and 5b) and evidence was also found for multiglobular structures.7,15,30 We thus conclude that a multiglobular structure may be a generic motif found over a wide range of the polymeric mucin family of biomolecules and van Willebrand factor (in contrast to the epithelial mucins, membrane-bound combs, that have different functional properties).7 The globule sizes measured are close to those expected for the homologous von Willebrand factor sequence.

Conclusions The small-angle X-ray scattering data reported here is consistent with a two-globule dumbbell model for porcine stomach mucin. The SAXS data qualitatively resembles that in the literature from troponin C, molecular dynamics simulations on synthetic polymers, and synthetic hydrophobic polyelectrolytes of intermediate hydrophobicity, which are known to have multiglobular structures. A simple robust analysis of the data gives an average globular radius of ∼10 nm invariant of salt concentration and an intramolecular globular separation of 48 nm that reduces to 18 nm upon the addition of 100 mM NaCl. The observed globular structures are in good agreement with previous TEM, AFM, and DLS measurements. Addition of salt gives a significant contraction of the chain size in SAXS measurements and causes a change of the semidilute scaling of the mesh size from extended chain (ξ ∼ c-0.45) to a more globular conformation (ξ ∼ c-0.28).

Cola et al.

Acknowledgment. Aristeidis Papagiannopoulos and Dave Thornton are thanked for useful conversations.

References and Notes (1) Bansil, R.; Turner, B. S. Curr. Opin. Colloid Interface Sci. 2006, 11, 164–170. (2) Lee, S.; Spencer, N. D. Science 2008, 319, 575–576. (3) Papagiannopoulos, A.; Waigh, T. A.; Hardingham, T. E. Faraday Discuss. 2008, 139, 337–357. (4) Papagiannopoulos, A.; Fernyhough, C. M.; Waigh, T. A. J. Chem. Phys. 2005, 123, 214904. (5) Papagiannopoulos, A.; Waigh, T. A.; Hardingham, T. E.; Heinrich, M. Biomacromolecules 2006, 7, 2162–2172. (6) Stokes, J. R.; Davies, G. A. Biorheology 2007, 44, 141–160. (7) Lang, T.; Hansson, G. C.; Samuelsson, T. Proc. Nat. Acad. Sci. U.S.A. 2007, 104, 16209–16214. (8) Desseyn, J. L.; Aubert, J. P.; Porchet, N.; Laine, A. Mol. Biol. EVol. 2000, 17, 1175–1184. (9) Singh, I.; Shankaran, H.; Beauharnois, M. E.; Xiao, Z.; Alexandridis, P.; Neelamegham, S. J. Biol. Chem. 2006, 281, 38266–38275. (10) Yakubov, G.; Papagiannopoulos, A.; Rat, E.; Easton, R. L.; Waigh, T. A. Biomacromolecules 2007, 8, 3467–3477. (11) Yakubov, G.; Papagiannopoulos, A.; Rat, E.; Waigh, T. A. Biomacromolecules 2007, 8, 3791–3799. (12) Hong, Z.; Chasan, B.; Bansil, R.; Turner, B. S.; Bhaskar, K. R.; Afdhal, N. H. Biomacromolecules 2005, 6, 3458–3466. (13) Waigh, T. A.; Papagiannopoulos, A.; Voice, A.; Bansil, R.; Unwin, A. P.; Dewhurst, C. D.; Turner, B.; Afdhal, N. Langmuir 2002, 18, 7188–7195. (14) Dobrynin, A. V.; Rubinstein, M. Macromolecules 2001, 34, 1964– 1972. (15) Brunelli, R.; Papi, M.; Arcovito, G.; Bompiani, A.; Castagnola, M.; Parasassi, T.; Sampaolese, B.; Vincenzoni, F.; De Spirito, M. FASEB J. 2007, 21, 3872–3876. (16) Kosovan, P.; Limpouchova, Z.; Prochazka, K. J. Phys. Chem. B 2007, 111, 8605–8611. (17) Kantor, Y.; Kardar, M. Phys. ReV. E 1995, 51, 1299. (18) Dobrynin, A. V.; Rubinstein, M.; Obukhov, S. O. Macromolecules 1996, 29, 2974–2979. (19) Liao, Q.; Dobrynin, A. V.; Rubinstein, M. Macromolecules 2006, 39, 1920–1938. (20) Baigl, D.; Ober, R.; Qu, D.; Fery, A.; Williams, C. E. Europhys. Lett. 2003, 62, 588–594. (21) Essafi, W.; Lafuma, F.; Williams, C. E. J. Physiol. (Paris) 1995, 5, 1269–1275. (22) Dobrynin, A. V.; Rubinstein, M. Prog. Polym. Sci. 2005, 30, 1049– 1118. (23) Rousseau, K.; Byrne, C.; Kim, Y. S.; Gum, J. R.; Swallow, D. M.; Toribara, N. W. Genomics 2004, 83, 936–939. (24) Waigh, T. A.; Papagiannopoulos, A.; Voice, A.; Bansil, R.; Unwin, A. P.; Dewhurst, C.; Turner, B.; Afdhal, N. Langmuir 2002, 18, 7188– 7195. (25) Fujisawa, T.; Ueki, T.; Inoko, Y.; Kataoka, M. J. Appl. Crystallogr. 1987, 20, 349–355. (26) Beaucage, B. J. Appl. Crystallogr. 1995, 28, 717–728. (27) Matsushima, N.; Hayashi, N.; Watanabe, N.; Jinbo, Y.; Izumi, Y. J. Appl. Crystallogr. 2007, 40, 179–183. (28) Spiteri, M. N.; Williams, C. E.; Boue, F. Macromolecules 2007, 40, 6679–6691. (29) Maier, E. A.; Krause, R.; Deggelmann, M.; Hagenbuchle, M.; Weber, R. Macromolecules 1992, 25, 1125–1133. (30) Thornton, D. J.; Rousseau, K.; McGuckin, M. A. Annu. ReV. Physiol. 2008, 70, 459–486.

BM800799U

Double-Globular Structure of Porcine Stomach Mucin: A Small-Angle

Recommend Documents