- Email: [email protected]
Molecular adsorption: early stage surface exploration
ARTICLE IN PRESS
Ultramicroscopy 100 (2004) 145–151
Molecular adsorption: early stage surface exploration Debra J. Brayshawa,*, Monica Berryb, Terence J. McMastera a
H.H. Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 1TL, UK b Mucin Research Group, University of Bristol, Bristol Eye Hospital, Bristol, BS1 2LX, UK
Received 27 June 2003; received in revised form 3 November 2003; accepted 18 November 2003
Abstract In this study, the atomic force microscope has been employed in force spectroscopy mode to gain information on the interaction between long mucin molecules and a positively charged surface during the first few seconds of interaction. Recent studies have revealed that negatively charged mucin molecules introduced to a positively charged surface are kinetically trapped and bind very rapidly, assuming non-equilibrium conformations. This systematic study of surface dwell times has revealed that significant differences exist in mucin adsorption during the first three seconds of introduction to the surface and provides direct evidence of molecular rearrangement for several seconds before trapping occurs. Limited interactions were recorded at dwell times of less than one second, with increased molecular rearrangement observed between 1.5 and 2.25 s. Increasing the surface dwell time beyond this critical limit caused rupture of the tip-tethered mucin molecules during the retract cycle of the cantilever. All subsequent recorded events, at increased dwell times up to 3 s, revealed events at much reduced distances from the point of contact between the mucin functionalised-cantilever and the positively charged surface. r 2004 Elsevier B.V. All rights reserved. Keywords: Glycoprotein; Force spectroscopy; Mucin; Dwell time
1. Introduction The large, heterogeneous, glycoproteins known as mucins are found in all mucous secretions and can be classified as either secreted or membrane spanning [1,2]. Secreted mucins can be either soluble or gel-forming [3], the latter believed to be responsible for the rheological properties of mucus [4]. Oligomerisation of gel-forming mucins through the formation of disulphide bonds has been shown to produce macromolecular structures *Corresponding author. Fax: +44-117-925-5624. Email-address: [email protected] (D.J. Brayshaw).
with mega-Dalton molecular weights [5,6]. Biochemical analysis has shown that human ocular gel-forming mucins contain negatively charged sub-units, with the charge conferred by sialic acid and sulphate moieties on the oligosaccharide side chain [7]. These macromolecular structures have been imaged using both electron microscopy [8–10] and atomic force microscopy (AFM) [11–13] to reveal mucins with extended, linear conformations up to several microns in length. In addition to quantifying the lengths of mucin molecules, previous AFM studies have also demonstrated that the two-dimensional conformation of mucins on surfaces has a strong dependence on the chosen
0304-3991/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2003.11.003
ARTICLE IN PRESS 146
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
substrate [12,14]. Biomolecules adsorbed to oppositely charged surfaces take up non-equilibrium conformations as the extended polymer chains become rigidly bound before they have time to relax and equilibrate on the surface [12,15]. In situ, real-time AFM of negatively charged mucin molecules adsorbing to a positively charged substrate revealed that the molecules diffused to the surface and were immobilised sufficiently for imaging within 10 min, half the time required for immobilisation on negatively charged mica [14]. However, when using the AFM to follow processes in situ in real-time, each image takes several minutes to obtain in an aqueous environment and therefore it is not possible to determine the absolute time required for mucins to electrostatically bind to an oppositely charged surface. In this study, in order to investigate molecular adsorption during the first few seconds of interaction with an oppositely charged surface the AFM was employed in force spectroscopy mode [16]. Mucin molecules were covalently tethered to the tip and the cantilever was introduced to a positively charged substrate. Force spectroscopy measurements were obtained over a range of surface dwell times and the distance from the point of contact to the last registered event was recorded together with the cantilever deflection at that point.
DNA contamination (Hoechst 33258 dye). Imaging of the mucins in 10 mM HEPES buffer at pH 7.4 was used to confirm that the mucins were free from other contaminants (Fig. 1). The extracted mucin solution contained, at most, 0.2% mucin molecules, as it did not show signs of gel formation. The purified mucins were then diluted 50 times in 10 mM HEPES buffer at pH 7.4. The cantilever was dipped into this diluted mucin solution to enable tethering of the molecules. The mucin-functionalised cantilever was kept hydrated under HEPES buffer and thus mucins represent a negligible proportion of the solution in which the force spectroscopy experiments were performed. Therefore, the conductivity and hence ionic strength [19] of the mucin solution shall be overwhelmingly dominated by the properties of the buffer. Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich, Dorset, UK. 2.2. AFM Mica functionalised with aminopropyltri-ethoxy silane (APTES) was prepared by placing a mica
2. Materials and methods 2.1. Preparation of mucins Secreted human ocular gel-forming mucins have been isolated from fragments of human cadaver conjunctivae by classical mucin extraction and purification methods [17]. Mucins were isolated from other glycoconjugates on a Caesium Chloride gradient and fractionated according to molecular size on Sepharose CL2B (Amersham Pharmacia Biotech, APB). The purified samples contained mucins with a buoyant density of 1.35–1.45 g/ml and were dominated by MUC5AC with a reduced contribution of MUC 2 [18]. Using size exclusion chromatography we selected the largest, excluded volume mucins which were confirmed free from
Fig. 1. Dilute mucin solution. Secreted mucins immobilised on mica before being washed, dried and imaged in air.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
disk in an APTES-rich environment at room temperature for 2 h. AFM silicon nitride cantilevers (Veeco, Santa Barbara, CA), 200 mm in length, with a nominal spring constant of 0.06 N/m (manufacturer’s value) were washed in Piranha solution (3:1 (v/v) solution of sulphuric acid and 30% hydrogen peroxide) to remove any organic contaminants, thoroughly rinsed in ultra pure water (Fluka, UK) and dried on filter paper. The cantilevers were placed in an Edwards E306A evaporator (BOC Edwards, Bristol UK) under a vacuum of o106 mbar. A Quartz crystal thickness monitor (BAL-TEC, EM Systems Support, UK) was employed to control the evaporation of 1 nm of chromium, followed by 10 nm of gold onto the cantilever tips. Following the method of Wagner et al., a self-assembled monolayer (SAM) of 11, 110 -dithiobis(succinimidylundecanoate) (DSU) was prepared on the gold cantilever [20]. Briefly, the cantilever was immersed in a 1 mM solution of DSU in 1,4-dioxane for 10 min at room temperature before being rinsed thoroughly in 1,4-dioxane and dried under a stream of nitrogen gas. DSU spontaneously self-assembles onto gold to form a monolayer a few nanometers thick. The DSU cantilever was held in a mucin solution for 10 min before being thoroughly rinsed in ultra pure water. The mucins were specifically bound via the Nterminus or exposed primary amine groups along the backbone to the succinimide groups of the SAM [21]. The cantilever was mounted in the AFM liquid cell and kept hydrated with 10 mM HEPES buffer. Force curves were performed using a commercial AFM (MultiMode, Veeco, Santa Barbara, CA, USA), fitted with a TappingModet fluid cell (Veeco) operated in force spectroscopy mode. Calibration of the piezo scan tube was carried using a standard calibration grid, (depth of 193.676.3 nm) and was confirmed using tobacco mosaic virus [22]. Photodiode sensitivity (nm/V) was calculated from the hard repulsion slope of the approach curve. Three hundred and twenty five force curves were performed with the same mucinfunctionalised tip at a scan rate of 0.2 Hz, 2000 nm ramp size and with surface dwell times from 0 to 3 s. Data sets were analysed using spreadsheet
147
macro-programs to allow for zeroing of the force curves at the point of contact of the cantilever and compensation for drift between the approach and retract null lines [17]. Control experiments were performed by introducing a DSU tip to an APTES-mica substrate under 10 mM HEPES buffer. Tapping mode images were obtained in air using Olympus microcantilevers with a nominal resonant frequency of 300 kHz and spring constant of 42 N/m. Images were typically collected at scan rates of 2 Hz with calibration of the piezo scan tube carried out in the x and y directions using a standard calibration grid, (0.5 mm square pits on a 1 mm grid).
3. Results and discussion Fig. 2a illustrates a typical force curve obtained when a DSU tip was introduced to an APTESmica substrate under 10 mM HEPES buffer. The graph has been zeroed at the point the cantilever contacts the surface. A large adhesion peak was always observed, but no events characteristic of the stretching of a long-chain polymer molecule [23]. The strong interaction between the two surfaces is highlighted not only by the magnitude of the adhesion peak in the retract curve, but also by the significant, 1 nN attractive force on the cantilever as it jumped into contact with the surface on approach (Fig. 2a inset, arrow). The flattening at the base of the adhesion peak in the retract curve (Fig. 2a, inset) results from the deflection caused by the adhesive force being outside the maximal recordable range. Fig. 2b shows the same force curve (retraction curve only) after it has been corrected for hard contact with the surface and inverted such that nominal, uncalibrated forces exerted on the cantilever take positive values. It can be clearly seen that the cantilever adhered to the surface for an average separation of 100 nm. Fig. 3 shows examples of force curves obtained when mucin molecules tethered to a DSU tip were introduced to an APTES-mica surface. All force curves (Fig. 3a–d) were obtained at a scan rate of 0.2 Hz using a relative trigger threshold of 20 nm
ARTICLE IN PRESS 148
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
Fig. 2. Control. DSU tip introduced to APTES-mica: black=approach, grey=retract; zeroed at the point of contact (a); inverted and corrected for hard contact (b).
(i.e. once at the point of contact, the cantilever continues to press against the surface until it is deflected by 20 nm). In 30% of eventful curves, only one event (Fig. 3a) is clearly distinguishable from the near surface adhesion peak. An event of this nature is attributed to the stretching of a single, long-chain polymer molecule [23]. It should also be noted that the magnitude of the adhesion peak is significantly reduced in comparison to the case where no mucin molecules were tethered to the DSU tip (Fig. 2), suggesting that the tip is sufficiently coated in molecules such that the DSUAPTES mica interactions are dramatically reduced at the point of contact. Fig. 3b presents stretching events where the cantilever returns to zero applied force between the
two events. Force curves of this form, 13% of eventful curves, could be attributed to the stretching of two tethered molecules that have considerably different lengths. Force curves displaying events separated by a small distance (B250 nm) compared to the tip-sample separation at which the last event is recorded, (B600 nm in this case) with the intermediate force being non-zero, (Fig. 3c) were less frequently observed, at 7% occurrence. Though possibly attributable to the stretching of two molecules of similar length, are more likely to be the stretching of one molecule that became self-associated in some way. Given the oppositely charged natures of the molecules and substrate, it is perhaps more likely that this behaviour derives from two nearby binding sites.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
149
Fig. 3. Mucin force curves. Mucins tethered to a DSU tip introduced to APTES-mica: one molecular stretching event (a); two distinct stretching events (b); events at similar distances such that the cantilever does not return to zero applied force in between (c); typical force curve for increasing dwell times where many stretching events are observed (d).
Half of the force curves collected during this study contained many events (Fig. 3d). An important point to note for all curves is that the DSU tip can tether the mucin molecules not only at the N-terminus, but also at the exposed primary amine groups along the backbone [21]. Therefore, it is not possible to determine where the molecule was tethered to the tip [16]. In addition, it is also unclear if the entire of the remaining length contacted and bound to the APTES-mica surface. To minimise the possibility of misinterpretation, only the last event shall be considered in this study of dwell time influence on recorded molecular length. In order to assess the interval required for electrostatic binding of mucin molecules to APTES-mica, sets of 25 force curves were performed, increasing the dwell times from zero to three seconds in 0.25 s intervals. Fig. 4 summarises the average distance from the point of contact to the longest recorded event. At the shorter dwell times investigated, there is no significant difference in the separation distance at which the last event
Fig. 4. Average length of longest molecule. Average distance from the point of contact at which the last recorded event occurred. Each point is an average of 25 data sets. Error bars show one standard deviation. It is important to note that an offset of 0.1 s applies to the time axis at each point in this data set since a 20 nm trigger threshold on a 2000 nm force curve at a scan rate of 0.2 Hz adds 0.05 s to the contact time on both the approach and retract curves.
occurred; surface dwell times up to 1.5 s show a constant molecular length of the order of 750 nm within the error bounds, suggesting that repeated force curves have no exhaustive effects on the
ARTICLE IN PRESS 150
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
mucin-functionalised tip. For surface dwell times between 1.5 and 2.25 s, however, the average distance at which the last event occurred increased steadily at a rate of 650 nm/s to an average of 1200 nm, consistent with molecular lengths observed in Fig. 1 and reported by Round et al. [12]. This increase in bound molecule length, as the surface dwell time increased, shows that the molecules do not bind instantly to an oppositely charged surface, but are able to explore the surface to a limited extent for a short period of time. The long dwell times, of the order of seconds, needed to produce stretching events at tip–sample separations comparable to the length of imaged molecules suggest that the mucins are organised in a gel-like phase on the tip [21]. Mucins contain highly glycosylated domains, where the insertion of the first O-linked N-Acetylgalactosamine is known to extend the polymer backbone and confer rigidity [24], while the high proportion of the bulky amino acid induces subtle kinks in the chain. Regions of dense O-glycosylation alternate with regions of low glycosylation along the length of the mucin chain giving the appearance of beading along the length of the molecule. Viewed with AFM, the lowest regions represent the poorly glycosylated or naked peptide core [11]. Mucins are interacting with the substrate both via their oligosaccharides, and peptide cores, as indicated by a Fourier analysis of the distances between consecutive adhesions [21]. As surface interaction times increase, more negatively charged regions are able to bind to the positively charged APTESmica substrate. Beyond 2.25 s, there was a marked decrease in the average distance at which the last event was recorded, with many curves recording no events at all. As the surface dwell time increased beyond 2.25 s, mucin molecules might have preferentially and substantially adhered to the APTES-mica substrate along a significantly greater portion of its length so that retraction of the cantilever caused the molecules to be either pulled off of the tip or ruptured near to the tip. Hence any subsequent interactions involved only shorter mucin fragments. To confirm this assumption, the APTES-mica substrate was scanned after force curve data acquisition (Fig. 5). A small number of short (B200 nm long) mucin fragments
can be seen, consistent with mechanically ruptured molecular fragments. A plot of cantilever deflection against dwell time at the last recorded event (Fig. 6) illustrates that for dwell times up to 1.75 s there is no increase in cantilever deflection (and hence pull-off force). Some increase in cantilever deflection was seen between the 1.75 and 2 s dwell times, but beyond this, the behaviour becomes erratic, suggesting that the molecule-surface interactions are changing. It is therefore proposed that the increased lengths to last detachment, were not due to the same section of molecule being stretched even further before detachment from the surface, but
Fig. 5. Mechanically removed mucin fragments. Image of molecule sections thought to have been torn from the DSU tip during force spectroscopy measurements. Scale bars show 100 nm, z-range=2.5 nm.
Fig. 6. Average cantilever deflections. Average cantilever deflection at the last recorded event. Each point is an average of 25 data sets. Error bars show one standard deviation.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
more likely due to the molecule exploring the surface and binding along a greater proportion of its length.
4. Conclusions A long established requirement for the successful imaging of biological molecules is that they are immobilised on a solid support. This has been achieved via a wide range of interactions including electrostatic forces, Van der Waals forces, the hydrophobic effect and covalent binding [25]. Negatively charged, extended, liner polymer chains such as DNA [15] and mucins [12] can be forced to assume non-equilibrium conformations by trapping the molecules onto a positively charged surface. In this case, the observed twodimensional image is believed to represent a projection of the three-dimensional assembly in solution onto the surface plane [15]. AFM force spectroscopy has shown that on the timescale of AFM imaging, adsorption has reached a steady state and the molecules are kinetically trapped (see also Ref. [14]). However, in this systematic study of surface dwell times we have shown that significant differences exist in mucin adsorption to an APTES-mica substrate during the first three seconds of interaction. Mucin–surface interaction times of less than 1.5 s revealed that the bound length of the molecule remained constant within the error bounds. As the surface dwell time increased, it is proposed that more of the molecule interacted with the surface. Increasing the dwell time beyond a critical limit caused the tip-tethered mucin molecules to be disrupted. A comparison between the imaged lengths and those revealed by force spectroscopy suggests that the mucin molecules were ruptured at an unbound point along the backbone rather than at the point of tethering to the cantilever.
151
References [1] A.P. Corfield, S.D. Carrington, S.J. Hicks, M. Berry, R. Ellingham, Prog. Ret. Eye Res. 16 (1997) 627. [2] I.K. Gipson, T. Inatomi, Prog. Ret. Eye Res. 16 (1997) 81. [3] S.J. Gendler, A.P. Spicer, Annu. Rev. Physiol. 57 (1995) 607. [4] P. Argueso, . I.K. Gipson, Exp. Eye Res. 73 (2001) 281. [5] I. Carlstedt, H. Lindgren, J.K. Sheehan, U. Ulmsten, L. Wingerup, Biochem. J. 211 (1983) 13. [6] R.L. Shogren, A.M. Jamieson, J. Blackwell, N. Jentoft, J. Biol. Chem. 259 (1984) 14657. [7] R.B. Ellingham, M. Berry, D. Stevenson, A.P. Corfield, Glycobiolgy 9 (1999) 1181. [8] D.J. Thornton, J.R. Davies, M. Kraayenbrink, P.S. Richardson, J.K. Sheehan, I. Carlstedt, Biochem. J. 265 (1990) 179. [9] J.K. Sheehan, R.P. Boot-Handford, E. Chantler, I. Carlstedt, D.J. Thornton, Biochem. J. 274 (1991) 293. [10] J.K. Sheehan, C. Brazeau, S. Kutay, H. Pigeon, S. Kirkham, M. Howard, Biochem. J. 347 (2000) 37. [11] T.J. McMaster, M. Berry, A.P. Corfield, M.J. Miles, Biophys. J. 77 (1999) 533. [12] A.N. Round, M. Berry, T.J. McMaster, S. Stoll, D. Gowers, A.P. Corfield, M.J. Miles, Biophys. J. 83 (2002) 1661. [13] D.J. Brayshaw, M. Berry, T.J. McMaster, Ultramicroscopy 97 (2003) 289. [14] D.J. Brayshaw, M. Berry, K. Carraway, T.J. McMaster, Langmuir, submitted for publication. [15] C. Rivetti, M. Guthold, C. Bustamante, J. Mol. Biol. 264 (1996) 919. [16] J. Hemmerl!e, S.M. Altmann, M. Maaloum, J.K.H. . Horber, L. Heinrich, J.-C. Voegel, P. Schaaf, Proc. Natl. Acad. Sci. USA 96 (1999) 6705. [17] M. Berry, R.B. Ellingham, A.P. Corfield, Invest. Ophthalmol. Vis. Sci. 37 (1996) 2559. [18] A.N. Round, M. Berry, T.J. McMaster, A.P. Corfield, M.J. Miles, J. Struct. Biol. 145 (2004) 246. [19] D. Grey, Transfusion Med. 12 (2002) 330. [20] P. Wagner, M. Hegner, P. Kernen, F. Zaugg, G. Semenza, Biophys. J. 70 (1996) 2052. [21] M. Berry, T.J. McMaster, A.P. Corfield, M.J. Miles, Biomacromolecules 2 (2001) 498. [22] F. Zenhausern, M. Adrian, R. Emch, M. Taborelli, M. Jobin, P. Descouts, Ultramicroscopy 42 (1992) 1168. [23] W.F. Heinz, J.H. Hoh, Trends Biotechnol. 17 (1999) 143. [24] I. Carlstedt, J.R. Davies, Biochem. Soc. Trans. 25 (1997) 214. [25] D.J. Muller, . M. Amrein, A. Engel, J. Struct. Biol. 119 (1997) 172.
Ultramicroscopy 100 (2004) 145–151
Molecular adsorption: early stage surface exploration Debra J. Brayshawa,*, Monica Berryb, Terence J. McMastera a
H.H. Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 1TL, UK b Mucin Research Group, University of Bristol, Bristol Eye Hospital, Bristol, BS1 2LX, UK
Received 27 June 2003; received in revised form 3 November 2003; accepted 18 November 2003
Abstract In this study, the atomic force microscope has been employed in force spectroscopy mode to gain information on the interaction between long mucin molecules and a positively charged surface during the first few seconds of interaction. Recent studies have revealed that negatively charged mucin molecules introduced to a positively charged surface are kinetically trapped and bind very rapidly, assuming non-equilibrium conformations. This systematic study of surface dwell times has revealed that significant differences exist in mucin adsorption during the first three seconds of introduction to the surface and provides direct evidence of molecular rearrangement for several seconds before trapping occurs. Limited interactions were recorded at dwell times of less than one second, with increased molecular rearrangement observed between 1.5 and 2.25 s. Increasing the surface dwell time beyond this critical limit caused rupture of the tip-tethered mucin molecules during the retract cycle of the cantilever. All subsequent recorded events, at increased dwell times up to 3 s, revealed events at much reduced distances from the point of contact between the mucin functionalised-cantilever and the positively charged surface. r 2004 Elsevier B.V. All rights reserved. Keywords: Glycoprotein; Force spectroscopy; Mucin; Dwell time
1. Introduction The large, heterogeneous, glycoproteins known as mucins are found in all mucous secretions and can be classified as either secreted or membrane spanning [1,2]. Secreted mucins can be either soluble or gel-forming [3], the latter believed to be responsible for the rheological properties of mucus [4]. Oligomerisation of gel-forming mucins through the formation of disulphide bonds has been shown to produce macromolecular structures *Corresponding author. Fax: +44-117-925-5624. Email-address: [email protected] (D.J. Brayshaw).
with mega-Dalton molecular weights [5,6]. Biochemical analysis has shown that human ocular gel-forming mucins contain negatively charged sub-units, with the charge conferred by sialic acid and sulphate moieties on the oligosaccharide side chain [7]. These macromolecular structures have been imaged using both electron microscopy [8–10] and atomic force microscopy (AFM) [11–13] to reveal mucins with extended, linear conformations up to several microns in length. In addition to quantifying the lengths of mucin molecules, previous AFM studies have also demonstrated that the two-dimensional conformation of mucins on surfaces has a strong dependence on the chosen
0304-3991/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ultramic.2003.11.003
ARTICLE IN PRESS 146
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
substrate [12,14]. Biomolecules adsorbed to oppositely charged surfaces take up non-equilibrium conformations as the extended polymer chains become rigidly bound before they have time to relax and equilibrate on the surface [12,15]. In situ, real-time AFM of negatively charged mucin molecules adsorbing to a positively charged substrate revealed that the molecules diffused to the surface and were immobilised sufficiently for imaging within 10 min, half the time required for immobilisation on negatively charged mica [14]. However, when using the AFM to follow processes in situ in real-time, each image takes several minutes to obtain in an aqueous environment and therefore it is not possible to determine the absolute time required for mucins to electrostatically bind to an oppositely charged surface. In this study, in order to investigate molecular adsorption during the first few seconds of interaction with an oppositely charged surface the AFM was employed in force spectroscopy mode [16]. Mucin molecules were covalently tethered to the tip and the cantilever was introduced to a positively charged substrate. Force spectroscopy measurements were obtained over a range of surface dwell times and the distance from the point of contact to the last registered event was recorded together with the cantilever deflection at that point.
DNA contamination (Hoechst 33258 dye). Imaging of the mucins in 10 mM HEPES buffer at pH 7.4 was used to confirm that the mucins were free from other contaminants (Fig. 1). The extracted mucin solution contained, at most, 0.2% mucin molecules, as it did not show signs of gel formation. The purified mucins were then diluted 50 times in 10 mM HEPES buffer at pH 7.4. The cantilever was dipped into this diluted mucin solution to enable tethering of the molecules. The mucin-functionalised cantilever was kept hydrated under HEPES buffer and thus mucins represent a negligible proportion of the solution in which the force spectroscopy experiments were performed. Therefore, the conductivity and hence ionic strength [19] of the mucin solution shall be overwhelmingly dominated by the properties of the buffer. Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich, Dorset, UK. 2.2. AFM Mica functionalised with aminopropyltri-ethoxy silane (APTES) was prepared by placing a mica
2. Materials and methods 2.1. Preparation of mucins Secreted human ocular gel-forming mucins have been isolated from fragments of human cadaver conjunctivae by classical mucin extraction and purification methods [17]. Mucins were isolated from other glycoconjugates on a Caesium Chloride gradient and fractionated according to molecular size on Sepharose CL2B (Amersham Pharmacia Biotech, APB). The purified samples contained mucins with a buoyant density of 1.35–1.45 g/ml and were dominated by MUC5AC with a reduced contribution of MUC 2 [18]. Using size exclusion chromatography we selected the largest, excluded volume mucins which were confirmed free from
Fig. 1. Dilute mucin solution. Secreted mucins immobilised on mica before being washed, dried and imaged in air.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
disk in an APTES-rich environment at room temperature for 2 h. AFM silicon nitride cantilevers (Veeco, Santa Barbara, CA), 200 mm in length, with a nominal spring constant of 0.06 N/m (manufacturer’s value) were washed in Piranha solution (3:1 (v/v) solution of sulphuric acid and 30% hydrogen peroxide) to remove any organic contaminants, thoroughly rinsed in ultra pure water (Fluka, UK) and dried on filter paper. The cantilevers were placed in an Edwards E306A evaporator (BOC Edwards, Bristol UK) under a vacuum of o106 mbar. A Quartz crystal thickness monitor (BAL-TEC, EM Systems Support, UK) was employed to control the evaporation of 1 nm of chromium, followed by 10 nm of gold onto the cantilever tips. Following the method of Wagner et al., a self-assembled monolayer (SAM) of 11, 110 -dithiobis(succinimidylundecanoate) (DSU) was prepared on the gold cantilever [20]. Briefly, the cantilever was immersed in a 1 mM solution of DSU in 1,4-dioxane for 10 min at room temperature before being rinsed thoroughly in 1,4-dioxane and dried under a stream of nitrogen gas. DSU spontaneously self-assembles onto gold to form a monolayer a few nanometers thick. The DSU cantilever was held in a mucin solution for 10 min before being thoroughly rinsed in ultra pure water. The mucins were specifically bound via the Nterminus or exposed primary amine groups along the backbone to the succinimide groups of the SAM [21]. The cantilever was mounted in the AFM liquid cell and kept hydrated with 10 mM HEPES buffer. Force curves were performed using a commercial AFM (MultiMode, Veeco, Santa Barbara, CA, USA), fitted with a TappingModet fluid cell (Veeco) operated in force spectroscopy mode. Calibration of the piezo scan tube was carried using a standard calibration grid, (depth of 193.676.3 nm) and was confirmed using tobacco mosaic virus [22]. Photodiode sensitivity (nm/V) was calculated from the hard repulsion slope of the approach curve. Three hundred and twenty five force curves were performed with the same mucinfunctionalised tip at a scan rate of 0.2 Hz, 2000 nm ramp size and with surface dwell times from 0 to 3 s. Data sets were analysed using spreadsheet
147
macro-programs to allow for zeroing of the force curves at the point of contact of the cantilever and compensation for drift between the approach and retract null lines [17]. Control experiments were performed by introducing a DSU tip to an APTES-mica substrate under 10 mM HEPES buffer. Tapping mode images were obtained in air using Olympus microcantilevers with a nominal resonant frequency of 300 kHz and spring constant of 42 N/m. Images were typically collected at scan rates of 2 Hz with calibration of the piezo scan tube carried out in the x and y directions using a standard calibration grid, (0.5 mm square pits on a 1 mm grid).
3. Results and discussion Fig. 2a illustrates a typical force curve obtained when a DSU tip was introduced to an APTESmica substrate under 10 mM HEPES buffer. The graph has been zeroed at the point the cantilever contacts the surface. A large adhesion peak was always observed, but no events characteristic of the stretching of a long-chain polymer molecule [23]. The strong interaction between the two surfaces is highlighted not only by the magnitude of the adhesion peak in the retract curve, but also by the significant, 1 nN attractive force on the cantilever as it jumped into contact with the surface on approach (Fig. 2a inset, arrow). The flattening at the base of the adhesion peak in the retract curve (Fig. 2a, inset) results from the deflection caused by the adhesive force being outside the maximal recordable range. Fig. 2b shows the same force curve (retraction curve only) after it has been corrected for hard contact with the surface and inverted such that nominal, uncalibrated forces exerted on the cantilever take positive values. It can be clearly seen that the cantilever adhered to the surface for an average separation of 100 nm. Fig. 3 shows examples of force curves obtained when mucin molecules tethered to a DSU tip were introduced to an APTES-mica surface. All force curves (Fig. 3a–d) were obtained at a scan rate of 0.2 Hz using a relative trigger threshold of 20 nm
ARTICLE IN PRESS 148
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
Fig. 2. Control. DSU tip introduced to APTES-mica: black=approach, grey=retract; zeroed at the point of contact (a); inverted and corrected for hard contact (b).
(i.e. once at the point of contact, the cantilever continues to press against the surface until it is deflected by 20 nm). In 30% of eventful curves, only one event (Fig. 3a) is clearly distinguishable from the near surface adhesion peak. An event of this nature is attributed to the stretching of a single, long-chain polymer molecule [23]. It should also be noted that the magnitude of the adhesion peak is significantly reduced in comparison to the case where no mucin molecules were tethered to the DSU tip (Fig. 2), suggesting that the tip is sufficiently coated in molecules such that the DSUAPTES mica interactions are dramatically reduced at the point of contact. Fig. 3b presents stretching events where the cantilever returns to zero applied force between the
two events. Force curves of this form, 13% of eventful curves, could be attributed to the stretching of two tethered molecules that have considerably different lengths. Force curves displaying events separated by a small distance (B250 nm) compared to the tip-sample separation at which the last event is recorded, (B600 nm in this case) with the intermediate force being non-zero, (Fig. 3c) were less frequently observed, at 7% occurrence. Though possibly attributable to the stretching of two molecules of similar length, are more likely to be the stretching of one molecule that became self-associated in some way. Given the oppositely charged natures of the molecules and substrate, it is perhaps more likely that this behaviour derives from two nearby binding sites.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
149
Fig. 3. Mucin force curves. Mucins tethered to a DSU tip introduced to APTES-mica: one molecular stretching event (a); two distinct stretching events (b); events at similar distances such that the cantilever does not return to zero applied force in between (c); typical force curve for increasing dwell times where many stretching events are observed (d).
Half of the force curves collected during this study contained many events (Fig. 3d). An important point to note for all curves is that the DSU tip can tether the mucin molecules not only at the N-terminus, but also at the exposed primary amine groups along the backbone [21]. Therefore, it is not possible to determine where the molecule was tethered to the tip [16]. In addition, it is also unclear if the entire of the remaining length contacted and bound to the APTES-mica surface. To minimise the possibility of misinterpretation, only the last event shall be considered in this study of dwell time influence on recorded molecular length. In order to assess the interval required for electrostatic binding of mucin molecules to APTES-mica, sets of 25 force curves were performed, increasing the dwell times from zero to three seconds in 0.25 s intervals. Fig. 4 summarises the average distance from the point of contact to the longest recorded event. At the shorter dwell times investigated, there is no significant difference in the separation distance at which the last event
Fig. 4. Average length of longest molecule. Average distance from the point of contact at which the last recorded event occurred. Each point is an average of 25 data sets. Error bars show one standard deviation. It is important to note that an offset of 0.1 s applies to the time axis at each point in this data set since a 20 nm trigger threshold on a 2000 nm force curve at a scan rate of 0.2 Hz adds 0.05 s to the contact time on both the approach and retract curves.
occurred; surface dwell times up to 1.5 s show a constant molecular length of the order of 750 nm within the error bounds, suggesting that repeated force curves have no exhaustive effects on the
ARTICLE IN PRESS 150
D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
mucin-functionalised tip. For surface dwell times between 1.5 and 2.25 s, however, the average distance at which the last event occurred increased steadily at a rate of 650 nm/s to an average of 1200 nm, consistent with molecular lengths observed in Fig. 1 and reported by Round et al. [12]. This increase in bound molecule length, as the surface dwell time increased, shows that the molecules do not bind instantly to an oppositely charged surface, but are able to explore the surface to a limited extent for a short period of time. The long dwell times, of the order of seconds, needed to produce stretching events at tip–sample separations comparable to the length of imaged molecules suggest that the mucins are organised in a gel-like phase on the tip [21]. Mucins contain highly glycosylated domains, where the insertion of the first O-linked N-Acetylgalactosamine is known to extend the polymer backbone and confer rigidity [24], while the high proportion of the bulky amino acid induces subtle kinks in the chain. Regions of dense O-glycosylation alternate with regions of low glycosylation along the length of the mucin chain giving the appearance of beading along the length of the molecule. Viewed with AFM, the lowest regions represent the poorly glycosylated or naked peptide core [11]. Mucins are interacting with the substrate both via their oligosaccharides, and peptide cores, as indicated by a Fourier analysis of the distances between consecutive adhesions [21]. As surface interaction times increase, more negatively charged regions are able to bind to the positively charged APTESmica substrate. Beyond 2.25 s, there was a marked decrease in the average distance at which the last event was recorded, with many curves recording no events at all. As the surface dwell time increased beyond 2.25 s, mucin molecules might have preferentially and substantially adhered to the APTES-mica substrate along a significantly greater portion of its length so that retraction of the cantilever caused the molecules to be either pulled off of the tip or ruptured near to the tip. Hence any subsequent interactions involved only shorter mucin fragments. To confirm this assumption, the APTES-mica substrate was scanned after force curve data acquisition (Fig. 5). A small number of short (B200 nm long) mucin fragments
can be seen, consistent with mechanically ruptured molecular fragments. A plot of cantilever deflection against dwell time at the last recorded event (Fig. 6) illustrates that for dwell times up to 1.75 s there is no increase in cantilever deflection (and hence pull-off force). Some increase in cantilever deflection was seen between the 1.75 and 2 s dwell times, but beyond this, the behaviour becomes erratic, suggesting that the molecule-surface interactions are changing. It is therefore proposed that the increased lengths to last detachment, were not due to the same section of molecule being stretched even further before detachment from the surface, but
Fig. 5. Mechanically removed mucin fragments. Image of molecule sections thought to have been torn from the DSU tip during force spectroscopy measurements. Scale bars show 100 nm, z-range=2.5 nm.
Fig. 6. Average cantilever deflections. Average cantilever deflection at the last recorded event. Each point is an average of 25 data sets. Error bars show one standard deviation.
ARTICLE IN PRESS D.J. Brayshaw et al. / Ultramicroscopy 100 (2004) 145–151
more likely due to the molecule exploring the surface and binding along a greater proportion of its length.
4. Conclusions A long established requirement for the successful imaging of biological molecules is that they are immobilised on a solid support. This has been achieved via a wide range of interactions including electrostatic forces, Van der Waals forces, the hydrophobic effect and covalent binding [25]. Negatively charged, extended, liner polymer chains such as DNA [15] and mucins [12] can be forced to assume non-equilibrium conformations by trapping the molecules onto a positively charged surface. In this case, the observed twodimensional image is believed to represent a projection of the three-dimensional assembly in solution onto the surface plane [15]. AFM force spectroscopy has shown that on the timescale of AFM imaging, adsorption has reached a steady state and the molecules are kinetically trapped (see also Ref. [14]). However, in this systematic study of surface dwell times we have shown that significant differences exist in mucin adsorption to an APTES-mica substrate during the first three seconds of interaction. Mucin–surface interaction times of less than 1.5 s revealed that the bound length of the molecule remained constant within the error bounds. As the surface dwell time increased, it is proposed that more of the molecule interacted with the surface. Increasing the dwell time beyond a critical limit caused the tip-tethered mucin molecules to be disrupted. A comparison between the imaged lengths and those revealed by force spectroscopy suggests that the mucin molecules were ruptured at an unbound point along the backbone rather than at the point of tethering to the cantilever.
151
References [1] A.P. Corfield, S.D. Carrington, S.J. Hicks, M. Berry, R. Ellingham, Prog. Ret. Eye Res. 16 (1997) 627. [2] I.K. Gipson, T. Inatomi, Prog. Ret. Eye Res. 16 (1997) 81. [3] S.J. Gendler, A.P. Spicer, Annu. Rev. Physiol. 57 (1995) 607. [4] P. Argueso, . I.K. Gipson, Exp. Eye Res. 73 (2001) 281. [5] I. Carlstedt, H. Lindgren, J.K. Sheehan, U. Ulmsten, L. Wingerup, Biochem. J. 211 (1983) 13. [6] R.L. Shogren, A.M. Jamieson, J. Blackwell, N. Jentoft, J. Biol. Chem. 259 (1984) 14657. [7] R.B. Ellingham, M. Berry, D. Stevenson, A.P. Corfield, Glycobiolgy 9 (1999) 1181. [8] D.J. Thornton, J.R. Davies, M. Kraayenbrink, P.S. Richardson, J.K. Sheehan, I. Carlstedt, Biochem. J. 265 (1990) 179. [9] J.K. Sheehan, R.P. Boot-Handford, E. Chantler, I. Carlstedt, D.J. Thornton, Biochem. J. 274 (1991) 293. [10] J.K. Sheehan, C. Brazeau, S. Kutay, H. Pigeon, S. Kirkham, M. Howard, Biochem. J. 347 (2000) 37. [11] T.J. McMaster, M. Berry, A.P. Corfield, M.J. Miles, Biophys. J. 77 (1999) 533. [12] A.N. Round, M. Berry, T.J. McMaster, S. Stoll, D. Gowers, A.P. Corfield, M.J. Miles, Biophys. J. 83 (2002) 1661. [13] D.J. Brayshaw, M. Berry, T.J. McMaster, Ultramicroscopy 97 (2003) 289. [14] D.J. Brayshaw, M. Berry, K. Carraway, T.J. McMaster, Langmuir, submitted for publication. [15] C. Rivetti, M. Guthold, C. Bustamante, J. Mol. Biol. 264 (1996) 919. [16] J. Hemmerl!e, S.M. Altmann, M. Maaloum, J.K.H. . Horber, L. Heinrich, J.-C. Voegel, P. Schaaf, Proc. Natl. Acad. Sci. USA 96 (1999) 6705. [17] M. Berry, R.B. Ellingham, A.P. Corfield, Invest. Ophthalmol. Vis. Sci. 37 (1996) 2559. [18] A.N. Round, M. Berry, T.J. McMaster, A.P. Corfield, M.J. Miles, J. Struct. Biol. 145 (2004) 246. [19] D. Grey, Transfusion Med. 12 (2002) 330. [20] P. Wagner, M. Hegner, P. Kernen, F. Zaugg, G. Semenza, Biophys. J. 70 (1996) 2052. [21] M. Berry, T.J. McMaster, A.P. Corfield, M.J. Miles, Biomacromolecules 2 (2001) 498. [22] F. Zenhausern, M. Adrian, R. Emch, M. Taborelli, M. Jobin, P. Descouts, Ultramicroscopy 42 (1992) 1168. [23] W.F. Heinz, J.H. Hoh, Trends Biotechnol. 17 (1999) 143. [24] I. Carlstedt, J.R. Davies, Biochem. Soc. Trans. 25 (1997) 214. [25] D.J. Muller, . M. Amrein, A. Engel, J. Struct. Biol. 119 (1997) 172.