- Email: [email protected]
New tensiographic studies on protein cleaning of polymer surfaces
Available online at www.sciencedirect.com
Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
New tensiographic studies on protein cleaning of polymer surfaces M. O’Neill a,∗ , N.D. McMillan b , G. Dunne b , C.I. Mitchell b , B. O’Rourke b , D. Morrin b , F. Brennan b , R. Miller c , L. McDonnell d , P. Scully e a
Carl Stuart Ltd., R&D, Tallaght Business Park, Whitestown, Dublin 24, Ireland b Institute of Technology Carlow, Kilkenny Road, Carlow, Ireland c Max Planck Institute for Colloids and Surfaces, D-14476 Potsdam, Germany d Cork Institute of Technology, Cork, Ireland e Photon Science Institute, The University of Manchester, PO Box 88, Manchester M60 1QD, UK Received 12 June 2007; received in revised form 11 December 2007; accepted 20 December 2007 Available online 5 January 2008
Abstract Presently, only indirect methods exist to monitor protein contamination of polymeric contact lens surfaces. This study, based on a fiber optic sensor, proposes a new quantitative and dynamic measurement technique to address this problem. Comparative contamination studies on three representative proteins: ␥ globulin, bovine serum albumin (BSA) and ovalbumin (molecular weight 240,000, 60,000 and 45,000), using a new tensiographic method, have been developed for a polymethylmethacrylate (PMMA) substrate. © 2008 Elsevier B.V. All rights reserved. Keywords: Tensiograph; Protein adsorption; Polymer surface; Surface contamination; Cleaning
1. Introduction The monitoring of protein/lipid and other biomolecular contamination of the surfaces of contact lenses can only be assessed with indirect methods. Perhaps the most relevant work developed to date on the contamination of surfaces by proteins is that of Matsumura et al. [1]. Work by Chandry and Scully [2] and co-workers found that hydrophobic polymer surfaces on their optical-sensor were easily contaminated and those of hydrophilic glass were not. Most polymeric materials such as, polyethylene, polypropylene, polystyrene, acrylic nylon and biomedical polymers are strongly hydrophobic. McMillan et al. [3] found that acrylic nylon was very easily contaminated by biomolecules in beers whilst nylon 66 was not easily contaminated. Polymers generally present surfaces that have a very high site density and which interact strongly with proteins. The highest molecular weight polymers are generally the ones that adsorb protein more readily [4]. It is common to consider that many segments of a protein molecule being attracted by a surface site
∗
Corresponding author. Tel.: +353 1 4523432; fax: +353 1 4523967. E-mail address: [email protected] (M. O’Neill).
0927-7757/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2007.12.047
interacting with hydrophobic domains, or alternatively, an interaction that arises from a surface site interacting strongly with patches on a protein. Moreover, ionic interaction can bind protein molecules to solid surfaces very strongly. Air–water studies have shown that proteins only need to get sufficient foothold on the surface to minimise the probability of desorption. Once attached to the surface, the protein being surface active moves towards the interface. It is clear that the binding to a surface increase with increasing hydrophobicity of the surface and with increasing hydrophobicity of the protein. Norde [5] reported in a classic study that desorption from hydrophobic surfaces under normal circumstances does not usually occur but exposure to extreme pH, high ionic strength, or extensive rinsing can remove the proteins. Recent work has shown that proteins can be replaced by suitable surfactants from liquid interfaces [6,7]. The present work is indeed a very close analogue of Norde’s work [5], however, Norde’s sensitivity throws some doubt on the validity of the conclusions drawn and the efficacy of standard procedures for cleaning surfaces. Studies on serum albumin molecules on glass surfaces show that the adsorption is considerable. Given that micron thick layers have been reported and the size of this molecule is only of the order of 5 nm multilayer structures must be formed. Terashima and Tsuji [8] studied the adsorption of bovine serum albumin (BSA) onto mica surfaces
110
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
by a direct weighing technique to determine the thickness of the adsorbed layer and contact angle measurements to determine the degree of surface covering. They found that over time the thickness of the adsorbed layer was increasing after the mica surface had been completely covered. They concluded, “The excess of adsorbance can be attributed to an adsorption of BSA molecules onto the adsorbed BSA layer because the mica surface has been confirmed to be completely covered at earlier stages of adsorption. Accordingly, it is safe to say that a multilayer adsorption of BSA molecules takes place if the adsorption is continued for a long period.” Silberberg found importantly that the forward reaction of adsorption, is limited, and reproducibly so, by concentration [4]. Clearly, a precise concentration dependent factor is at work. These findings are highly relevant to the experimental work reported here. 2. Apparatus The multianalyser tensiograph [9] was developed on the principles of stalagmometric instruments. It is a fiber optic instrument based on some simple principles of physics. An infra red or visible beam reflects light through a drop while it is forming. The instrument monitors the optical coupling between source and collector fibers placed in the drop-head and the optoelectronic signal produced is known as a tensiotrace. Every liquid has a unique drop shape and hence a unique tensiotrace, which leads to the fingerprinting capability of the instrument. Therefore, the instrument is fundamentally a development of surface and interfacial science because it is based on the analysis of an optical signal determined fundamentally by the shape of the liquid on the drop-head. The present study is concerned only with the analysis of protein solutions, but the technique has been used in a number of other application areas [10–12]. From current work it is clear that there are a wide range of other potential application areas for the multianalyser such as adhesive manufacture, in food analysis where it can be used to analyse oils and other liquid products and in pharmaceutical fingerprinting for the forensic identification of drugs. The drop-head used in this work is of a concave design. A drawing of the head is shown in Fig. 1. The design employs 1 mm polymethylmethacrylate (PMMA) fibers. The fibres are down polished to 0.3 m using lapping film before gluing them into the drop-head. The drop-head is made from poly-ethyl-ethylketone (PEEK). The diameter of the head is 9 mm with the fibers separated by 6 mm. The fibers are positioned with a jig to give a standard tensiotrace and just protrude a small distance from the concave base. A HPLC capillary, glued into the centre of the head is used to deliver the liquid. The head is designed such that it wets (i.e. the suspended liquid covers the entire lower surface of the drophead) when liquid is delivered. Light from an LED source is injected into the drop-head through the source fiber and the signal is picked up by the collector fiber and goes to the photodiode, which then produces the trace. Water is taken as the reference liquid for most applications. The principal features of the tensiotrace (rainbow peak, the tensiograph peak and the drop period)
Fig. 1. Drop-head design for PMMA fibers cylindrical head showing the light trace from the source to the detector.
obtained for a water sample are illustrated in Fig. 2. A number of peaks essentially arising from total internal-reflection have been observed in various drop-head designs with different fiber spacings. This is due to the light coupling by different types of reflections from the source to the collector fiber on the far side of the drop-head. In this drop-head second and third order reflection deliver two peaks at different sizes of drop. The Multianalyser operates by recording just one single tensiotrace. The tensiotrace is scissored from the incoming A/D detector signal, which is produced from the light collected from the collector fiber. The trace is obtained by recording the optoelectronic signal between the fall of two drops from the head. To achieve this scissoring a “trigger drop” is formed and falls from the drop-head. First, the data acquisition is triggered by the control signal of optical eyes situated below the drop-head. The recording of the signal from the detector on the end of the collector fiber then proceeds until the second drop, the measurement drop, falls. The data for this measurement drop is then stored in the archival system of the computer after conversion to a digital form by the A/D card. The trace recorded for a single drop is known as the tensiotrace and is a unique fingerprint of the liquid.
Fig. 2. Typical water tensiotrace showing the characteristic features and the overlap area between two traces.
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
111
in Isopropanol, (vi) carbon tetrachloride, (vi) acetone. All the experiments were carried out at 30 ◦ C.
3. Data analysis 3.1. Relevant tensiograph analysis
4.1. Contamination of drop-head The analysis of the protein solutions used in this study was based on M (matching) and D (difference) values. One of the approaches integrated into the multianalyser tensiograph analysis package is trace features ratioing using M-value. The software produces the M values from a reference trace and a test trace. The software separates each feature, i.e. drop period, rainbow peak period, etc., and the match between both traces is determined. Any of the tensiotrace features can be ratioed to give the matching value, M. In the case of this study the area has been found to give the greatest sensitivity in measuring contamination of the drop-head. The M and D subscript used in the following two equations indicate area. Fig. 2 illustrates the labeling of a tensiotrace for two traces X, the reference trace and Y, the test trace. The shaded area represents the overlap of the two traces. The overlap area function MA (illustrated in Fig. 2) is defined as MA = f (A) =
AO AO or (0 < MA < 1) AT AT
4.2. Cleaning the contaminated drop-head Attempts were made to remove the contaminating protein using the cleaning solutions listed above. A stepwise cleaning process was used. The number of cleaning cycles was set successively to 5, 10, 15, 20, 30 and 40. After each cleaning run a water trace was taken. This water trace was compared to the reference trace and the area difference was recorded. This difference was plotted against the number of cleans. This procedure was repeated for all three proteins.
(1) 5. Results and discussion
where AO is the overlap area and AT is the total area of either trace X or Y depending on which has the greatest area. The Dvalue is simply the residual of the M-value and can be defined simply as DA = 1 − MA
A reference water trace was first obtained. The drop-head was then contaminated with one of the proteins by obtaining twenty traces, i.e. passing the protein through the system for 30 min to allow adsorption to occur.
(2)
4. Experimental Protein contamination was studied using: (i) Bovine serum albumin (fraction V, A-7906, Sigma– Aldrich Co.), (ii) chicken egg albumin (grade V, A5503, Sigma– Aldrich Co.), (iii) bovine ␥ globulin (G7516, Sigma–Aldrich Co.). All proteins were used at a concentration of 0.3 mg/ml dissolved in distilled H2 O. The cleaning solutions used were: (i) Trizyme, obtained from Sauflon Pharmaceuticals Ltd. The concentration used was one tablet dissolved in 10 cm3 of a sterile buffered isotonic solution, (ii) Aerosolv obtained from Sauflon Pharmaceuticals Ltd., (iii) 0.364% and 0.5% pepsin in 0.1 M HCl, ethanol, (iv) dilute and 5 M nitric acid, (v) 30% KOH
5.1. Cleaning failure for all existing agents In previous work [13] it was found that the proteins were adhering to the drop-head or fibres causing contamination. This resulted in a diminution in signal intensity when one tried to recover the reference trace after running a protein sample. Therefore, the aim of this study was to find a cleaning solution, which would give a 100% recovery. Fig. 3(a) and (b) show the cleaning curves of some standard solutions with the D-value given for area feature. As the curves indicate, no cleaning solution produced a 100% recovery, the best being 0.36% pepsin in 0.1 M hydrochloric acid (about 90% after 50 cycles) and a close second-best being 99.8% ethanol (85% retrieval after approximately 60 cycles). It is interesting to note that many cleaning solutions actually produced a worsening situation with respect to the optical signal detected. This effect could perhaps be explained by the denaturing of the proteins and the increased attenuation and contamination produced by the molecular fragments. After this study, visual inspection of the drop-head revealed that the proteins visually damaged the polished polymer fiber ends. A
Fig. 3. (a) Cleaning curves of standard solutions. (b) Cleaning curves of some standard solutions for fibres coated with araldite.
112
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
Fig. 4. Pictorial representation of physical model.
second set of studies, results shown in Fig. 3(b), was consequently conducted with a drop-head on which the fibers were coated with araldite. Araldite was used to protect the fibers from the visible damage that the proteins’ adhering to them was causing. The efficacy of what are reported in the literature to be protein-cleaning solutions is again dismal, with pepsin (apparently one of the most popular) actually disimproving the contamination measurement. This is probably due to denaturing of the protein, which arguably could be seen as a move towards cleaning of the surface. 5.2. Physical model of contamination The experimental evidence has shown that the effect of tensiotrace changes occurs with the PMMA fiber drop-heads. The effect does not happen for silica glass fibers with a PEEK drophead. It must be concluded therefore that the affinity of the proteins for the PMMA fibers is responsible for the initial contamination effect. PMMA must acquire a significantly greater accumulation of protein contaminants than either glass or PEEK. The effect has also been observed then with PMMA drop-heads and silica glass fibers. We must conclude that the variation in the tensiotrace results from the protein deposition onto PMMA surface. The recent work of Yildirim et al. [14] found that it is probable that drop detachment is influenced by contact angle when capillaries are large and indeed in the multianalyser the drop-head is 9 mm diameter. This new effect is observed however when the PMMA is in a central position in the drop-head and this cannot be explained therefore by contact angle variations. This work has been repeated many times and the effect is real. The measurement of tensiotrace area provides a more sensitive measurement tool here than simply monitoring the drop period. A further programme of research has more fully experimentally characterised these effects, but the results relate to a different experimental set up designed specifically for investigating the adsorption and desorption of proteins. This work will be reported subsequently. Fig. 4 shows a pictorial representation of the proposed tensiograph process. It is known that PMMA is hydrophilic when
wet [15,16] and that it readily acquires protein. When a surface is hydrophilic there is no driving force to ‘acquire’ proteins and another process, perhaps involving charges may be the reason that proteins are adhering to these surfaces. The proteins studied here are globular which usually sit with their hydrophobic parts inside the globule, thereby presenting a hydrophobic outer surface that might be attracted to the PMMA as shown in Fig. 4. The formation of the protein layers on both the PMMA and the drop surface must begin with a progressive coverage of the surface by the primary colonising proteins from aqueous phase. This stage is usually characterised by relatively high protein densities and probably competition between the colonising molecules. If sufficient coverage is established they must start to form some structural arrangements on the surface forming into multilayers [17–20]. Typical adsorption surfaces can be considered to have a very high binding site (“ligand”) density resulting in multivalent interactions of the protein and surface. If the multivalent interactions are of sufficient number, or energy, the adsorption is usually considered to be “irreversible”. If the adsorption is of this type then the release of the proteins into the water drop must be from proteins adsorbed in a multilayer structure. It is possible that the desorption process from the PMMA surface observed when water drops are placed on a contaminated drop-head would require some critical concentration of proteins to establish a full layer before the experimental effects reported here would be seen. A kinetics process in which the concentration of the solution of the contaminating protein determines the thickness of the layers is possible. It is found with polymeric adsorptions from the very dilute to relatively concentrated solutions that the amount adsorbed on the surface is relatively constant. The higher molecular weight adsorbs best [4]. Denatured proteins may in fact have a greater affinity for the surface than the original protein. After cleaving they lose their high affinity and can be desorbed more easily. Desorption of adsorbed protein solubility into pure water must be very slow. The contamination measurement procedure used here is to deliver water to the drop-head and then record the tensiotrace. No measurable change in the tensiotrace occurs despite the passage of large (in excess of 60 cm3 ) volumes of pure water over a period of 2 h or more. The present study has however established that the tensiotrace is ultra-sensitive to small variations in the quantity of adsorbed protein. The mobility of the proteins must also be sufficient to develop an equilibrium concentration in the bulk and surface of the liquid because no measurable drift is observed in the multianalyser measurement of the reference liquid water. Once the PMMA surface is contaminated then a ‘contamination tensiotrace’ is recorded rather than the clean water trace. This trace persists despite attempts to clean the drop-head with copious amounts of water. It appears that there is some sort of stable dynamic process occurring here. Changes in the tensiotrace can be caused by either changes in the surface tension of the drop itself, or by changes in the wetting behaviour of the drop-head from which the drop detaches at a critical weight. For small droplets of spherical shape no wet-
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
113
Fig. 5. The form of the tensiotraces involved in the polymer substrate study.
ting or contact angle effects have been experimentally observed so far. However, impact of contact angle on drop detachment and volume of detaching and remnant drop has been discussed recently in an extensive analysis of the drop volume tensiometry [14]. We assume that the cleaning effects observed here can be explained mainly by changes in the contact angle. Changes in the surface tension of the drop γ are not probable, as after several rinsing cycles, the bulk concentration in the drop should be very low and adsorption of proteins at the drop surface and consequently any decrease in γ would take much more time then the drop period of the present experiments. This can be easily estimated by assuming that each rinsing cycle with pure water would reduce the concentration in the drop at least to half, or 3 cycles by about one order of magnitude. Starting with 3 mg/ml of bovine albumin with Mw = 69.000, which corresponds to 4.3 × 10−5 mol/l, after 10 rinsing cycles we would reach a concentration of the order of 10−8 mol/l, which would require hours for the few protein molecules to adsorb [21]. Hence, we can assume that changes observed via the tensiotrace are caused mainly by the adsorbed proteins at the drop-head surface modifying the surface energy (contact angle) and affecting the drop detachment process. 5.3. Cleaning using activated enzymes Fig. 5 shows the actual form of the tensiotraces involved in this study. Water was chosen as the reference tensiotrace, which was recorded on a clean drop-head. This is shown in Fig. 5(a) and is recognisable as the trace with the longest drop period. Bovine serum albumin was then measured and the tensiotrace for this is the dotted trace in Fig. 5(a). Attempts were made to use water to clean the BSA from the drop-head. Fig. 5(b) represents these tensiotraces, the trace with the longest drop period corresponding to the reference water and the dashed line the trace after the drop-head was cleaned with water. It should be noted that there are only small differences between this trace and the original BSA trace. The trizyme cleaning solution provided by Sauflon Pharmaceutical was then used to remove the proteins and a water trace was measured once again. Incrementally, the tensiotrace then recovers to the clean drop-head water tensiotrace as sufficient cleaning cycles are carried out with the
activated enzyme cleaning solution. The curve of Fig. 5(c) shows that the reference and test water traces eventually became the same, with a full recovery of signal and indicating the total removal of the contaminating protein. This trizyme cleaning solution allows for the complete removal of proteins from the drop-head surface. Fig. 6 indicates the 100% recovery obtained using this solution. The graph here is for DA versus number of cleaning solution cycles. In the first case the protein being tested was BSA. It takes 20 cleaning cycles for the protein to be removed in a time of approximately 10 min. The objective of this part of the study was to see whether proteins could be identified by their cleaning recovery curves. Two further proteins were selected, ␥ globulin and ovalbumin, and measured on the multianalyser. A graph of D-value for the area feature versus the number of cleaning cycles is plotted in Fig. 7. It was found that each of the proteins has a unique cleaning curve that could possibly allow for qualitative identification. Although, BSA and ovalbumin require 20 cleans they produce a visually differentiated curve to reach this point. In comparison, ␥ globulin appears to be a more active contaminant requiring 40 cleans to recover the reference water trace. The results of the work may be summarised as follows:
Fig. 6. Cleaning recovery curve for BSA using the cleaning solution.
114
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
Fig. 7. Recovery curves of BSA, ␥ globulin and ovalbumin, each at a concentration of 300 g/ml.
• The trizyme solution allowed for protein quantification to be carried out without any contamination affecting the results. The solution produced a 100% recovery of protein removal from polymer surfaces. It might be noted, that none of the solutions routinely used in biotechnology as cleaning solutions could affect total removal. Either SAXS (small angle X-ray scattering), SANS (small angle neutron scattering), ellipsometry, or SFM (scanning force microscopy) could be used to calibrate the multianalyser cleaning methods and to perhaps give a quantitative method of measuring these reactions. This discovery has important industrial potential for contact lenses but more generally in biotechnology and specifically in surface science. • The work presented here showed that the three proteins analysed each produced a unique cleaning curve allowing for possible identification. At present further work is necessary; analysis of the same proteins under different conditions (concentration, pH and temperature) and checking the cleaning curves of other proteins, lipids and contaminating biomolecules to see if the cleaning curves offer a qualitative method of analysis in some specific applications involving a limited number of such molecules. • It is now possible to propose new protein tensiograph assays, as reproducible results are obtainable using the cleaning solution. Such new assays, if developed for biotechnology, could have some major advantages over existing methods. These require no chemical procedures and measurements are made directly from the protein solutions by the multianalyser. 6. Conclusions The cleaning study has shown the ineffectiveness of most well recognised-established cleaning procedures for the removal of protein from surfaces. The activated enzyme solutions used by the authors have shown very good results on three representative proteins and it is possible that this solution offers a very good approach to cleaning for the majority of proteins. The technique is the first that the authors are aware of which shows experimentally the dynamic process of protein removal from a surface. Many surface techniques such as contact angle measurements have not reported any contamination from proteins after cleaning procedures are carried out.
The sensitivity for monitoring protein contamination of surfaces is impressive and at this point only the work of Dunne and co-workers on this subject exist. Dunne et al. [22] give some early comparisons with other established methods. The work is ongoing, but has found that contact angle studies are orders of magnitude less sensitive at detecting surface contamination than tensiometry, although he has not been able to quantify this properly. Dunne et al. [23] have used atomic force microscope studies to see gold-labeled BSA molecules on surfaces of PMMA which of necessity are restricted to qualitative assessment of levels of contamination. They have also made studies into application of a new graphical tool and measurands for tensiography surface science studies in BSA protein adsorption on a stainless steel substrate. It appears that tensiography is measuring contaminations down to levels of individual molecules on surfaces rather than surface layer. No comparative measurements have been made using ellipsometry or optical waveguide lightmode spectroscopy (OWLS). These latter techniques require at least fairly complete surface layers, whereas the tensiographic technique appears given the work done to date, to be capable of detecting individual protein molecules on surfaces. The technique can easily be extended by fitting fibers into drop-heads made of contact lens polymeric materials. This would give a direct measurement of the cleaning of problem proteins, lipids and other contaminating molecules of interest for ophthalmic applications. A final word is perhaps in order on the actual changes seen here in the tensiotraces on drop-heads that are contaminated. Extensive studies of tensiotraces have been made using many different types of test liquids by Morrin [24] and in all cases there are changes in peak heights as well as in some cases drop volume changes. Protein contamination is quite obvious in it has a unique tensiotrace signature in that the peaks do not change size and only the drop volume changes. Hence this was the practical reason in this study for selecting the tensiograph measurands MA and DA , rather than any of the other fairly numerous options offered by this multivariate instrumental technique. This admittedly preliminary study, does suggest that tensiography may be the most sensitive technique available at this stage for monitoring surface contamination by proteins. The limitations of the technique are obvious—the flow rate of the droplet formation on both the recovery and absorption of surfaces from protein adsorption. Some effort needs to be directed towards studying these flow rate dependences and the work here is limited to what are really little more than qualitative comparisons of the cleaning solutions. When partial recovery is obtained, the distribution of contaminant adsorption between the liquid–air and the PMMA drop-head surface is ambiguous. The denaturing of proteins by many of the cleaning agents investigated means it is only really possible to derive definitive information about the cleaning of PMMA when complete recovery is obtained. Nevertheless, it is certainly established by this study that trizyme solutions are a very rapid and comprehensive way of cleaning surfaces and arguably it is shown here that solutions traditionally used to clean enzymes from surfaces struggle badly when proteins have great affinity for a surface such as is the case with PMMA.
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
References [1] H. Matsumura, K. Kawasaki, M. Kambara, Colloids Surf. B 8 (1997) 181–188. [2] R.P. Chandry, P.J. Scully, CLOOPT Progress Report 1, July 1998, Contact: N ENV4-CT97-0634. [3] N.D. McMillan, M. Reddin, M. O’Neill, R. Jordan, D. Phillips, D. Goff, J. Nolan, R. Harnedy, W. Mitchell, J. Harkin, V. Lawlor, L.R.L. McMillan, J. Inst. Brewing 106 (2000) 147–156. [4] A. Silberberg, Modelling of protein adsorption, in: J.D. Andrade (Ed.), Surface and Interfacial Aspects of Biomedical Polymers, vol. 2, Plenum Press, 1985, p. 322. [5] W. Norde, Adsorption of proteins at solid surfaces, in: L-H.Y. Lee (Ed.), Adhesion Adsorption of Polymer, vol. 2, Plenum Press, New York, 1980, pp. 801–826. [6] T.F. Svitova, M.J. Wetherbee, C.J. Radke, J. Colloid Interf. Sci. 261 (2003) 170–179. [7] V.B. Fainerman, R. Miller, J.K. Ferri, H. Watzke, M.E. Leser, M. Michel, Reversibility and irreversibility of adsorption of surfactants and proteins at liquid interfaces, Adv. Colloid Interface 123–126 (2006) 163–171. [8] H. Terashima, T. Tsuji, Colloids Surf. B 27 (2003) 115–122. [9] N.D. McMillan, V. Lawlor, M. Baker, Drops and bubbles in interfacial research, in: D. M¨obius, R. Miller (Eds.), Studies in Interface Science, vol. 6, Elsevier, 1997, pp. 593–705. [10] N.D. McMillan, O. Finlayson, F. Fortune, M. Fingelton, D. Daly, D. Townsend, D.D.G. McMillan, M.J. Dalton, Rev. Sci. Instrum. 63 (1992) 216–227. [11] N.D. McMillan, V. Lawlor, J. Nolan, W.Y. Lo, R. Harnedy, M. O’Neill, Colloids Surf. A 143 (1998) 421–427.
115
[12] N.D. McMillan, D. Dowling, M. O’Neill, T. Yeomans, R. Miller, in: Maynooth, D. Vernon (Eds.), Proceeding of the Optical Engineering Society of Ireland and Irish Machine Vision and Image Processing Society Joint Conference, September 9–10, 1998, pp. 119– 134. [13] N.D. McMillan, D. Dowling, M. O’Neill, T. Yeomans, European workshop on optical fibre sensors, in: Proceedings of the SPIE, Vol. 3483, 1998, pp. 123–127. [14] O.E. Yildirim, Q. Xu, O.A. Basaran, Phys. Fluids 17 (062107) (2005) 1–13. [15] K. Kawasaki, M. Kambara, H. Matsumura, Colloid Polym. Sci. 277 (1999) 1058–1064. [16] K. Kawasaki, M. Kambara, H. Matsumura, W. Norde, Caries Res. 33 (1999) 473–478. [17] D. Graham, M. Phillips, J. Colloid Interf. Sci. 70 (1979) 403. [18] V.B. Fainerman, R. Miller, Langmuir 15 (1999) 1812–1816. [19] C.J. Beverung, C.J. Radke, H. Blanch, Biophys. Chem. 81 (1999) 59. [20] V.B. Fainerman, E. Lucassen-Reynders, R. Miller, Adv. Colloid Interf. Sci. 106 (2003) 237–259. [21] J. Kr¨agel, M. O’Neill, A.V. Makievski, M. Michel, M.E. Leser, R. Miller, Colloids Surf. B 31 (2003) 107–114. [22] G. Dunne, L. McDonnell, R. Miller, N.D. McMillan, B. O’Rourke, C.I. Mitchell, Proceedings of the SPIE, Vol. 5824, Opto-Ireland, 2005, pp. 199–201. [23] G. Dunne, D. Morrin, M. O’Neill, N.D. McMillan, B. O’Rourke, C.I. Mitchell, Sensors and their Applications XII, IOPP, Boston and Philadelphia, 2003, pp. 567–572. [24] D. Morrin, Experimental Studies in Tensiography, PhD, Carlow, 2007.
Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
New tensiographic studies on protein cleaning of polymer surfaces M. O’Neill a,∗ , N.D. McMillan b , G. Dunne b , C.I. Mitchell b , B. O’Rourke b , D. Morrin b , F. Brennan b , R. Miller c , L. McDonnell d , P. Scully e a
Carl Stuart Ltd., R&D, Tallaght Business Park, Whitestown, Dublin 24, Ireland b Institute of Technology Carlow, Kilkenny Road, Carlow, Ireland c Max Planck Institute for Colloids and Surfaces, D-14476 Potsdam, Germany d Cork Institute of Technology, Cork, Ireland e Photon Science Institute, The University of Manchester, PO Box 88, Manchester M60 1QD, UK Received 12 June 2007; received in revised form 11 December 2007; accepted 20 December 2007 Available online 5 January 2008
Abstract Presently, only indirect methods exist to monitor protein contamination of polymeric contact lens surfaces. This study, based on a fiber optic sensor, proposes a new quantitative and dynamic measurement technique to address this problem. Comparative contamination studies on three representative proteins: ␥ globulin, bovine serum albumin (BSA) and ovalbumin (molecular weight 240,000, 60,000 and 45,000), using a new tensiographic method, have been developed for a polymethylmethacrylate (PMMA) substrate. © 2008 Elsevier B.V. All rights reserved. Keywords: Tensiograph; Protein adsorption; Polymer surface; Surface contamination; Cleaning
1. Introduction The monitoring of protein/lipid and other biomolecular contamination of the surfaces of contact lenses can only be assessed with indirect methods. Perhaps the most relevant work developed to date on the contamination of surfaces by proteins is that of Matsumura et al. [1]. Work by Chandry and Scully [2] and co-workers found that hydrophobic polymer surfaces on their optical-sensor were easily contaminated and those of hydrophilic glass were not. Most polymeric materials such as, polyethylene, polypropylene, polystyrene, acrylic nylon and biomedical polymers are strongly hydrophobic. McMillan et al. [3] found that acrylic nylon was very easily contaminated by biomolecules in beers whilst nylon 66 was not easily contaminated. Polymers generally present surfaces that have a very high site density and which interact strongly with proteins. The highest molecular weight polymers are generally the ones that adsorb protein more readily [4]. It is common to consider that many segments of a protein molecule being attracted by a surface site
∗
Corresponding author. Tel.: +353 1 4523432; fax: +353 1 4523967. E-mail address: [email protected] (M. O’Neill).
0927-7757/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2007.12.047
interacting with hydrophobic domains, or alternatively, an interaction that arises from a surface site interacting strongly with patches on a protein. Moreover, ionic interaction can bind protein molecules to solid surfaces very strongly. Air–water studies have shown that proteins only need to get sufficient foothold on the surface to minimise the probability of desorption. Once attached to the surface, the protein being surface active moves towards the interface. It is clear that the binding to a surface increase with increasing hydrophobicity of the surface and with increasing hydrophobicity of the protein. Norde [5] reported in a classic study that desorption from hydrophobic surfaces under normal circumstances does not usually occur but exposure to extreme pH, high ionic strength, or extensive rinsing can remove the proteins. Recent work has shown that proteins can be replaced by suitable surfactants from liquid interfaces [6,7]. The present work is indeed a very close analogue of Norde’s work [5], however, Norde’s sensitivity throws some doubt on the validity of the conclusions drawn and the efficacy of standard procedures for cleaning surfaces. Studies on serum albumin molecules on glass surfaces show that the adsorption is considerable. Given that micron thick layers have been reported and the size of this molecule is only of the order of 5 nm multilayer structures must be formed. Terashima and Tsuji [8] studied the adsorption of bovine serum albumin (BSA) onto mica surfaces
110
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
by a direct weighing technique to determine the thickness of the adsorbed layer and contact angle measurements to determine the degree of surface covering. They found that over time the thickness of the adsorbed layer was increasing after the mica surface had been completely covered. They concluded, “The excess of adsorbance can be attributed to an adsorption of BSA molecules onto the adsorbed BSA layer because the mica surface has been confirmed to be completely covered at earlier stages of adsorption. Accordingly, it is safe to say that a multilayer adsorption of BSA molecules takes place if the adsorption is continued for a long period.” Silberberg found importantly that the forward reaction of adsorption, is limited, and reproducibly so, by concentration [4]. Clearly, a precise concentration dependent factor is at work. These findings are highly relevant to the experimental work reported here. 2. Apparatus The multianalyser tensiograph [9] was developed on the principles of stalagmometric instruments. It is a fiber optic instrument based on some simple principles of physics. An infra red or visible beam reflects light through a drop while it is forming. The instrument monitors the optical coupling between source and collector fibers placed in the drop-head and the optoelectronic signal produced is known as a tensiotrace. Every liquid has a unique drop shape and hence a unique tensiotrace, which leads to the fingerprinting capability of the instrument. Therefore, the instrument is fundamentally a development of surface and interfacial science because it is based on the analysis of an optical signal determined fundamentally by the shape of the liquid on the drop-head. The present study is concerned only with the analysis of protein solutions, but the technique has been used in a number of other application areas [10–12]. From current work it is clear that there are a wide range of other potential application areas for the multianalyser such as adhesive manufacture, in food analysis where it can be used to analyse oils and other liquid products and in pharmaceutical fingerprinting for the forensic identification of drugs. The drop-head used in this work is of a concave design. A drawing of the head is shown in Fig. 1. The design employs 1 mm polymethylmethacrylate (PMMA) fibers. The fibres are down polished to 0.3 m using lapping film before gluing them into the drop-head. The drop-head is made from poly-ethyl-ethylketone (PEEK). The diameter of the head is 9 mm with the fibers separated by 6 mm. The fibers are positioned with a jig to give a standard tensiotrace and just protrude a small distance from the concave base. A HPLC capillary, glued into the centre of the head is used to deliver the liquid. The head is designed such that it wets (i.e. the suspended liquid covers the entire lower surface of the drophead) when liquid is delivered. Light from an LED source is injected into the drop-head through the source fiber and the signal is picked up by the collector fiber and goes to the photodiode, which then produces the trace. Water is taken as the reference liquid for most applications. The principal features of the tensiotrace (rainbow peak, the tensiograph peak and the drop period)
Fig. 1. Drop-head design for PMMA fibers cylindrical head showing the light trace from the source to the detector.
obtained for a water sample are illustrated in Fig. 2. A number of peaks essentially arising from total internal-reflection have been observed in various drop-head designs with different fiber spacings. This is due to the light coupling by different types of reflections from the source to the collector fiber on the far side of the drop-head. In this drop-head second and third order reflection deliver two peaks at different sizes of drop. The Multianalyser operates by recording just one single tensiotrace. The tensiotrace is scissored from the incoming A/D detector signal, which is produced from the light collected from the collector fiber. The trace is obtained by recording the optoelectronic signal between the fall of two drops from the head. To achieve this scissoring a “trigger drop” is formed and falls from the drop-head. First, the data acquisition is triggered by the control signal of optical eyes situated below the drop-head. The recording of the signal from the detector on the end of the collector fiber then proceeds until the second drop, the measurement drop, falls. The data for this measurement drop is then stored in the archival system of the computer after conversion to a digital form by the A/D card. The trace recorded for a single drop is known as the tensiotrace and is a unique fingerprint of the liquid.
Fig. 2. Typical water tensiotrace showing the characteristic features and the overlap area between two traces.
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
111
in Isopropanol, (vi) carbon tetrachloride, (vi) acetone. All the experiments were carried out at 30 ◦ C.
3. Data analysis 3.1. Relevant tensiograph analysis
4.1. Contamination of drop-head The analysis of the protein solutions used in this study was based on M (matching) and D (difference) values. One of the approaches integrated into the multianalyser tensiograph analysis package is trace features ratioing using M-value. The software produces the M values from a reference trace and a test trace. The software separates each feature, i.e. drop period, rainbow peak period, etc., and the match between both traces is determined. Any of the tensiotrace features can be ratioed to give the matching value, M. In the case of this study the area has been found to give the greatest sensitivity in measuring contamination of the drop-head. The M and D subscript used in the following two equations indicate area. Fig. 2 illustrates the labeling of a tensiotrace for two traces X, the reference trace and Y, the test trace. The shaded area represents the overlap of the two traces. The overlap area function MA (illustrated in Fig. 2) is defined as MA = f (A) =
AO AO or (0 < MA < 1) AT AT
4.2. Cleaning the contaminated drop-head Attempts were made to remove the contaminating protein using the cleaning solutions listed above. A stepwise cleaning process was used. The number of cleaning cycles was set successively to 5, 10, 15, 20, 30 and 40. After each cleaning run a water trace was taken. This water trace was compared to the reference trace and the area difference was recorded. This difference was plotted against the number of cleans. This procedure was repeated for all three proteins.
(1) 5. Results and discussion
where AO is the overlap area and AT is the total area of either trace X or Y depending on which has the greatest area. The Dvalue is simply the residual of the M-value and can be defined simply as DA = 1 − MA
A reference water trace was first obtained. The drop-head was then contaminated with one of the proteins by obtaining twenty traces, i.e. passing the protein through the system for 30 min to allow adsorption to occur.
(2)
4. Experimental Protein contamination was studied using: (i) Bovine serum albumin (fraction V, A-7906, Sigma– Aldrich Co.), (ii) chicken egg albumin (grade V, A5503, Sigma– Aldrich Co.), (iii) bovine ␥ globulin (G7516, Sigma–Aldrich Co.). All proteins were used at a concentration of 0.3 mg/ml dissolved in distilled H2 O. The cleaning solutions used were: (i) Trizyme, obtained from Sauflon Pharmaceuticals Ltd. The concentration used was one tablet dissolved in 10 cm3 of a sterile buffered isotonic solution, (ii) Aerosolv obtained from Sauflon Pharmaceuticals Ltd., (iii) 0.364% and 0.5% pepsin in 0.1 M HCl, ethanol, (iv) dilute and 5 M nitric acid, (v) 30% KOH
5.1. Cleaning failure for all existing agents In previous work [13] it was found that the proteins were adhering to the drop-head or fibres causing contamination. This resulted in a diminution in signal intensity when one tried to recover the reference trace after running a protein sample. Therefore, the aim of this study was to find a cleaning solution, which would give a 100% recovery. Fig. 3(a) and (b) show the cleaning curves of some standard solutions with the D-value given for area feature. As the curves indicate, no cleaning solution produced a 100% recovery, the best being 0.36% pepsin in 0.1 M hydrochloric acid (about 90% after 50 cycles) and a close second-best being 99.8% ethanol (85% retrieval after approximately 60 cycles). It is interesting to note that many cleaning solutions actually produced a worsening situation with respect to the optical signal detected. This effect could perhaps be explained by the denaturing of the proteins and the increased attenuation and contamination produced by the molecular fragments. After this study, visual inspection of the drop-head revealed that the proteins visually damaged the polished polymer fiber ends. A
Fig. 3. (a) Cleaning curves of standard solutions. (b) Cleaning curves of some standard solutions for fibres coated with araldite.
112
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
Fig. 4. Pictorial representation of physical model.
second set of studies, results shown in Fig. 3(b), was consequently conducted with a drop-head on which the fibers were coated with araldite. Araldite was used to protect the fibers from the visible damage that the proteins’ adhering to them was causing. The efficacy of what are reported in the literature to be protein-cleaning solutions is again dismal, with pepsin (apparently one of the most popular) actually disimproving the contamination measurement. This is probably due to denaturing of the protein, which arguably could be seen as a move towards cleaning of the surface. 5.2. Physical model of contamination The experimental evidence has shown that the effect of tensiotrace changes occurs with the PMMA fiber drop-heads. The effect does not happen for silica glass fibers with a PEEK drophead. It must be concluded therefore that the affinity of the proteins for the PMMA fibers is responsible for the initial contamination effect. PMMA must acquire a significantly greater accumulation of protein contaminants than either glass or PEEK. The effect has also been observed then with PMMA drop-heads and silica glass fibers. We must conclude that the variation in the tensiotrace results from the protein deposition onto PMMA surface. The recent work of Yildirim et al. [14] found that it is probable that drop detachment is influenced by contact angle when capillaries are large and indeed in the multianalyser the drop-head is 9 mm diameter. This new effect is observed however when the PMMA is in a central position in the drop-head and this cannot be explained therefore by contact angle variations. This work has been repeated many times and the effect is real. The measurement of tensiotrace area provides a more sensitive measurement tool here than simply monitoring the drop period. A further programme of research has more fully experimentally characterised these effects, but the results relate to a different experimental set up designed specifically for investigating the adsorption and desorption of proteins. This work will be reported subsequently. Fig. 4 shows a pictorial representation of the proposed tensiograph process. It is known that PMMA is hydrophilic when
wet [15,16] and that it readily acquires protein. When a surface is hydrophilic there is no driving force to ‘acquire’ proteins and another process, perhaps involving charges may be the reason that proteins are adhering to these surfaces. The proteins studied here are globular which usually sit with their hydrophobic parts inside the globule, thereby presenting a hydrophobic outer surface that might be attracted to the PMMA as shown in Fig. 4. The formation of the protein layers on both the PMMA and the drop surface must begin with a progressive coverage of the surface by the primary colonising proteins from aqueous phase. This stage is usually characterised by relatively high protein densities and probably competition between the colonising molecules. If sufficient coverage is established they must start to form some structural arrangements on the surface forming into multilayers [17–20]. Typical adsorption surfaces can be considered to have a very high binding site (“ligand”) density resulting in multivalent interactions of the protein and surface. If the multivalent interactions are of sufficient number, or energy, the adsorption is usually considered to be “irreversible”. If the adsorption is of this type then the release of the proteins into the water drop must be from proteins adsorbed in a multilayer structure. It is possible that the desorption process from the PMMA surface observed when water drops are placed on a contaminated drop-head would require some critical concentration of proteins to establish a full layer before the experimental effects reported here would be seen. A kinetics process in which the concentration of the solution of the contaminating protein determines the thickness of the layers is possible. It is found with polymeric adsorptions from the very dilute to relatively concentrated solutions that the amount adsorbed on the surface is relatively constant. The higher molecular weight adsorbs best [4]. Denatured proteins may in fact have a greater affinity for the surface than the original protein. After cleaving they lose their high affinity and can be desorbed more easily. Desorption of adsorbed protein solubility into pure water must be very slow. The contamination measurement procedure used here is to deliver water to the drop-head and then record the tensiotrace. No measurable change in the tensiotrace occurs despite the passage of large (in excess of 60 cm3 ) volumes of pure water over a period of 2 h or more. The present study has however established that the tensiotrace is ultra-sensitive to small variations in the quantity of adsorbed protein. The mobility of the proteins must also be sufficient to develop an equilibrium concentration in the bulk and surface of the liquid because no measurable drift is observed in the multianalyser measurement of the reference liquid water. Once the PMMA surface is contaminated then a ‘contamination tensiotrace’ is recorded rather than the clean water trace. This trace persists despite attempts to clean the drop-head with copious amounts of water. It appears that there is some sort of stable dynamic process occurring here. Changes in the tensiotrace can be caused by either changes in the surface tension of the drop itself, or by changes in the wetting behaviour of the drop-head from which the drop detaches at a critical weight. For small droplets of spherical shape no wet-
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
113
Fig. 5. The form of the tensiotraces involved in the polymer substrate study.
ting or contact angle effects have been experimentally observed so far. However, impact of contact angle on drop detachment and volume of detaching and remnant drop has been discussed recently in an extensive analysis of the drop volume tensiometry [14]. We assume that the cleaning effects observed here can be explained mainly by changes in the contact angle. Changes in the surface tension of the drop γ are not probable, as after several rinsing cycles, the bulk concentration in the drop should be very low and adsorption of proteins at the drop surface and consequently any decrease in γ would take much more time then the drop period of the present experiments. This can be easily estimated by assuming that each rinsing cycle with pure water would reduce the concentration in the drop at least to half, or 3 cycles by about one order of magnitude. Starting with 3 mg/ml of bovine albumin with Mw = 69.000, which corresponds to 4.3 × 10−5 mol/l, after 10 rinsing cycles we would reach a concentration of the order of 10−8 mol/l, which would require hours for the few protein molecules to adsorb [21]. Hence, we can assume that changes observed via the tensiotrace are caused mainly by the adsorbed proteins at the drop-head surface modifying the surface energy (contact angle) and affecting the drop detachment process. 5.3. Cleaning using activated enzymes Fig. 5 shows the actual form of the tensiotraces involved in this study. Water was chosen as the reference tensiotrace, which was recorded on a clean drop-head. This is shown in Fig. 5(a) and is recognisable as the trace with the longest drop period. Bovine serum albumin was then measured and the tensiotrace for this is the dotted trace in Fig. 5(a). Attempts were made to use water to clean the BSA from the drop-head. Fig. 5(b) represents these tensiotraces, the trace with the longest drop period corresponding to the reference water and the dashed line the trace after the drop-head was cleaned with water. It should be noted that there are only small differences between this trace and the original BSA trace. The trizyme cleaning solution provided by Sauflon Pharmaceutical was then used to remove the proteins and a water trace was measured once again. Incrementally, the tensiotrace then recovers to the clean drop-head water tensiotrace as sufficient cleaning cycles are carried out with the
activated enzyme cleaning solution. The curve of Fig. 5(c) shows that the reference and test water traces eventually became the same, with a full recovery of signal and indicating the total removal of the contaminating protein. This trizyme cleaning solution allows for the complete removal of proteins from the drop-head surface. Fig. 6 indicates the 100% recovery obtained using this solution. The graph here is for DA versus number of cleaning solution cycles. In the first case the protein being tested was BSA. It takes 20 cleaning cycles for the protein to be removed in a time of approximately 10 min. The objective of this part of the study was to see whether proteins could be identified by their cleaning recovery curves. Two further proteins were selected, ␥ globulin and ovalbumin, and measured on the multianalyser. A graph of D-value for the area feature versus the number of cleaning cycles is plotted in Fig. 7. It was found that each of the proteins has a unique cleaning curve that could possibly allow for qualitative identification. Although, BSA and ovalbumin require 20 cleans they produce a visually differentiated curve to reach this point. In comparison, ␥ globulin appears to be a more active contaminant requiring 40 cleans to recover the reference water trace. The results of the work may be summarised as follows:
Fig. 6. Cleaning recovery curve for BSA using the cleaning solution.
114
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
Fig. 7. Recovery curves of BSA, ␥ globulin and ovalbumin, each at a concentration of 300 g/ml.
• The trizyme solution allowed for protein quantification to be carried out without any contamination affecting the results. The solution produced a 100% recovery of protein removal from polymer surfaces. It might be noted, that none of the solutions routinely used in biotechnology as cleaning solutions could affect total removal. Either SAXS (small angle X-ray scattering), SANS (small angle neutron scattering), ellipsometry, or SFM (scanning force microscopy) could be used to calibrate the multianalyser cleaning methods and to perhaps give a quantitative method of measuring these reactions. This discovery has important industrial potential for contact lenses but more generally in biotechnology and specifically in surface science. • The work presented here showed that the three proteins analysed each produced a unique cleaning curve allowing for possible identification. At present further work is necessary; analysis of the same proteins under different conditions (concentration, pH and temperature) and checking the cleaning curves of other proteins, lipids and contaminating biomolecules to see if the cleaning curves offer a qualitative method of analysis in some specific applications involving a limited number of such molecules. • It is now possible to propose new protein tensiograph assays, as reproducible results are obtainable using the cleaning solution. Such new assays, if developed for biotechnology, could have some major advantages over existing methods. These require no chemical procedures and measurements are made directly from the protein solutions by the multianalyser. 6. Conclusions The cleaning study has shown the ineffectiveness of most well recognised-established cleaning procedures for the removal of protein from surfaces. The activated enzyme solutions used by the authors have shown very good results on three representative proteins and it is possible that this solution offers a very good approach to cleaning for the majority of proteins. The technique is the first that the authors are aware of which shows experimentally the dynamic process of protein removal from a surface. Many surface techniques such as contact angle measurements have not reported any contamination from proteins after cleaning procedures are carried out.
The sensitivity for monitoring protein contamination of surfaces is impressive and at this point only the work of Dunne and co-workers on this subject exist. Dunne et al. [22] give some early comparisons with other established methods. The work is ongoing, but has found that contact angle studies are orders of magnitude less sensitive at detecting surface contamination than tensiometry, although he has not been able to quantify this properly. Dunne et al. [23] have used atomic force microscope studies to see gold-labeled BSA molecules on surfaces of PMMA which of necessity are restricted to qualitative assessment of levels of contamination. They have also made studies into application of a new graphical tool and measurands for tensiography surface science studies in BSA protein adsorption on a stainless steel substrate. It appears that tensiography is measuring contaminations down to levels of individual molecules on surfaces rather than surface layer. No comparative measurements have been made using ellipsometry or optical waveguide lightmode spectroscopy (OWLS). These latter techniques require at least fairly complete surface layers, whereas the tensiographic technique appears given the work done to date, to be capable of detecting individual protein molecules on surfaces. The technique can easily be extended by fitting fibers into drop-heads made of contact lens polymeric materials. This would give a direct measurement of the cleaning of problem proteins, lipids and other contaminating molecules of interest for ophthalmic applications. A final word is perhaps in order on the actual changes seen here in the tensiotraces on drop-heads that are contaminated. Extensive studies of tensiotraces have been made using many different types of test liquids by Morrin [24] and in all cases there are changes in peak heights as well as in some cases drop volume changes. Protein contamination is quite obvious in it has a unique tensiotrace signature in that the peaks do not change size and only the drop volume changes. Hence this was the practical reason in this study for selecting the tensiograph measurands MA and DA , rather than any of the other fairly numerous options offered by this multivariate instrumental technique. This admittedly preliminary study, does suggest that tensiography may be the most sensitive technique available at this stage for monitoring surface contamination by proteins. The limitations of the technique are obvious—the flow rate of the droplet formation on both the recovery and absorption of surfaces from protein adsorption. Some effort needs to be directed towards studying these flow rate dependences and the work here is limited to what are really little more than qualitative comparisons of the cleaning solutions. When partial recovery is obtained, the distribution of contaminant adsorption between the liquid–air and the PMMA drop-head surface is ambiguous. The denaturing of proteins by many of the cleaning agents investigated means it is only really possible to derive definitive information about the cleaning of PMMA when complete recovery is obtained. Nevertheless, it is certainly established by this study that trizyme solutions are a very rapid and comprehensive way of cleaning surfaces and arguably it is shown here that solutions traditionally used to clean enzymes from surfaces struggle badly when proteins have great affinity for a surface such as is the case with PMMA.
M. O’Neill et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 323 (2008) 109–115
References [1] H. Matsumura, K. Kawasaki, M. Kambara, Colloids Surf. B 8 (1997) 181–188. [2] R.P. Chandry, P.J. Scully, CLOOPT Progress Report 1, July 1998, Contact: N ENV4-CT97-0634. [3] N.D. McMillan, M. Reddin, M. O’Neill, R. Jordan, D. Phillips, D. Goff, J. Nolan, R. Harnedy, W. Mitchell, J. Harkin, V. Lawlor, L.R.L. McMillan, J. Inst. Brewing 106 (2000) 147–156. [4] A. Silberberg, Modelling of protein adsorption, in: J.D. Andrade (Ed.), Surface and Interfacial Aspects of Biomedical Polymers, vol. 2, Plenum Press, 1985, p. 322. [5] W. Norde, Adsorption of proteins at solid surfaces, in: L-H.Y. Lee (Ed.), Adhesion Adsorption of Polymer, vol. 2, Plenum Press, New York, 1980, pp. 801–826. [6] T.F. Svitova, M.J. Wetherbee, C.J. Radke, J. Colloid Interf. Sci. 261 (2003) 170–179. [7] V.B. Fainerman, R. Miller, J.K. Ferri, H. Watzke, M.E. Leser, M. Michel, Reversibility and irreversibility of adsorption of surfactants and proteins at liquid interfaces, Adv. Colloid Interface 123–126 (2006) 163–171. [8] H. Terashima, T. Tsuji, Colloids Surf. B 27 (2003) 115–122. [9] N.D. McMillan, V. Lawlor, M. Baker, Drops and bubbles in interfacial research, in: D. M¨obius, R. Miller (Eds.), Studies in Interface Science, vol. 6, Elsevier, 1997, pp. 593–705. [10] N.D. McMillan, O. Finlayson, F. Fortune, M. Fingelton, D. Daly, D. Townsend, D.D.G. McMillan, M.J. Dalton, Rev. Sci. Instrum. 63 (1992) 216–227. [11] N.D. McMillan, V. Lawlor, J. Nolan, W.Y. Lo, R. Harnedy, M. O’Neill, Colloids Surf. A 143 (1998) 421–427.
115
[12] N.D. McMillan, D. Dowling, M. O’Neill, T. Yeomans, R. Miller, in: Maynooth, D. Vernon (Eds.), Proceeding of the Optical Engineering Society of Ireland and Irish Machine Vision and Image Processing Society Joint Conference, September 9–10, 1998, pp. 119– 134. [13] N.D. McMillan, D. Dowling, M. O’Neill, T. Yeomans, European workshop on optical fibre sensors, in: Proceedings of the SPIE, Vol. 3483, 1998, pp. 123–127. [14] O.E. Yildirim, Q. Xu, O.A. Basaran, Phys. Fluids 17 (062107) (2005) 1–13. [15] K. Kawasaki, M. Kambara, H. Matsumura, Colloid Polym. Sci. 277 (1999) 1058–1064. [16] K. Kawasaki, M. Kambara, H. Matsumura, W. Norde, Caries Res. 33 (1999) 473–478. [17] D. Graham, M. Phillips, J. Colloid Interf. Sci. 70 (1979) 403. [18] V.B. Fainerman, R. Miller, Langmuir 15 (1999) 1812–1816. [19] C.J. Beverung, C.J. Radke, H. Blanch, Biophys. Chem. 81 (1999) 59. [20] V.B. Fainerman, E. Lucassen-Reynders, R. Miller, Adv. Colloid Interf. Sci. 106 (2003) 237–259. [21] J. Kr¨agel, M. O’Neill, A.V. Makievski, M. Michel, M.E. Leser, R. Miller, Colloids Surf. B 31 (2003) 107–114. [22] G. Dunne, L. McDonnell, R. Miller, N.D. McMillan, B. O’Rourke, C.I. Mitchell, Proceedings of the SPIE, Vol. 5824, Opto-Ireland, 2005, pp. 199–201. [23] G. Dunne, D. Morrin, M. O’Neill, N.D. McMillan, B. O’Rourke, C.I. Mitchell, Sensors and their Applications XII, IOPP, Boston and Philadelphia, 2003, pp. 567–572. [24] D. Morrin, Experimental Studies in Tensiography, PhD, Carlow, 2007.