Cells and Materials
Volume 5 Number 1 Article 9
1995
Adsorption of Proteins at Solid-Liquid Interfaces
Willem Norde Wageningen Agricultural University, The Netherlands
Follow this and additional works at: https://digitalcommons.usu.edu/cellsandmaterials
Part of the Biomedical Engineering and Bioengineering Commons
Recommended Citation Norde, Willem (1995) "Adsorption of Proteins at Solid-Liquid Interfaces," Cells and Materials: Vol. 5 : No. 1 , Article 9. Available at: https://digitalcommons.usu.edu/cellsandmaterials/vol5/iss1/9
This Article is brought to you for free and open access by the Western Dairy Center at DigitalCommons@USU. It has been accepted for inclusion in Cells and Materials by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. Cells and Materials, Vol. 5, No. 1, 1995 (Pages 97-112) 1051-6794/95$5.00+ .25 Scanning Microscopy International , Chicago (AMF O'Hare), IL 60666 USA ADSORPTION OF PROTEINS AT SOLID-LIQillD INTERFACES
Willem Norde
Department of Physical and Colloid Chemi stry, Wageningen Agri cultural Universi ty, P.O. Box 8038, 6700 EK Wageningen, The Netherlands
Telephone number: (31 ) 08370-83540 I FAX number: (31 ) 08370-83777
(Received for publicati on October 4, 1994 and in revised form March 31, 1995)
Abstract Introduction
This paper concisely reviews the general principles Interacti on between proteins and solid surfaces is underl ying protein adsorption from aqueous solution commonly observed, both in natural and syntheti c sys onto a solid surface. The discussion includes the various tems. These interactions are of great relevance in , e.g., stages of the adsorption process, i. e., transport of the medical, bi otechnological and envi ronmental applica protein molecules towards the surface, the absorbed tions. In many cases, spontaneous adsorption of pro amount under equilibrium conditions, desorption andre teins leads to undesired consequences such as th rombus adsorption. Among the interacti ons that determine the formati on on syntheti c cardiovascul ar implants [76, 77 , overall protein adsorption process (1 ) redi stribution of 81, 82], fo uling of hemodi alysis membranes, contact charged groups in the interfacial layer, (2) changes in lenses and bi oprocessing equipment [29] , and plaque the hydrati on of the sorbent and the pro tein surface, and formati on on teeth and dental restorati ves [64, 80] . In (3) structural rearrangements in the protein molec ul e oth er cases, protein adsorpti on is made use of, for in play maj or roles. Special att ention is given to the re stance in drug deli very and controlled drug release sys lati on between the structu ral stabi lity of th e protei n tems [ 4], in di agnosti c tests (immunolatices) [39] , in bio molecul e and its adsorption behaviour. sensors [66] and in protein purification tech.niques [1 3]. In all these examples, the influence of the proximity of the sorbent surface on the bi ological fun ctioning and, because of the structure-function relati onship in proteins, on the three-dimensional structure of the pro tein mole cul e is of crucial importance. Effecti ve cont rol of the adsorption process requires an understanding of the un derl ying mechani sm, i.e., of the interacti ons that are in vo lved. Protein adsorption is an intricate process. Figure 1 shows a schemati c outline of the vari ous steps that are involved: transports toward , and binding at the surface, ori entati on and structure of the adsorbed molecules, reversibility of the sorption process. In this paper, each of the aspects depicted in Figure I will be addressed, emphasizing the relati on between protein structure stability and its adsorption behaviour.
Adsorption Kinetics Key Words: Protein adsorption, solid-liquid interface, adsorption kineti cs, thermodynami cs of adsorption, elec The rate of transport of a protein molecule from trostatic interaction, hydrophobic interac ti on, rearrange soluti on towards an interface increases with increasing ment in protein structure, protein desorption, a-lactalbu concentrati on cp of the protein in solution. The min, lysozyme. "reacti on" of the protein with the interface, i.e., the actual att achment at the sorbent surface, is independent of cp. It is, therefore, to be expected that at low cp and low degree of coverage of the sorbent by the pro tein , th e
97 W. Norde
Psol --
P~~s ~~~~~~~~-/~-. :--~-~-~~. {-/~
Figure 1. Schematic representation of the protein adsorption process. The native protein is denoted by P (the subscript sol refers to solution and ads to adsorbed state) and the structurally perturbed protein by P* and P**. The following steps are depicted: (1). Transport of P from solution towards the surface. (2). Attachment at the surface (re orientation?). (3). Structural rearrangement in the adsorbed molecule. (4). Detachment from the surface. (5). Transport away from the surface.
transport process controls the rate of adsorption and that only and distinguish between (a) a solution tangentially at high cp and high degree of coverage, surface reactions flowing along a surface and (b) an impinging jet flow are rate-determining. that hits the surface perpendicularly in a so-called The basic transport mechanisms are diffusion and "stagnation point". convective transport by laminar or turbulent flow. For the tangential flow, the protein molecules are Under quiescent conditions, protein molecules reach the transported towards the sorbent surface by simultaneous sorbent surface by stochastic Brownian motion [59]. If convection and diffusion and, under conditions of a they are relatively rapidly attached at the surface, it steady-state concentration boundary layer in the solution leads to depletion of protein in the layer adjacent to the adjacent to the sorbent surface, the flux is given by the surface. The resulting concentration gradient causes Leveque equation [ 43] protein diffusion from the bulk solution towards the sorbent surface. Under such conditions, the rate of J = 0.54 (y/yD) 113 D cp (2) arrival of the protein at the sorbent surface is given by the Ward and Tordai equation [79] where y is the shear rate at the sorbent surface and y is the distance between the point of observation and the J = c (Dhr)l /2 rl/2 p (I) point where the protein solution has reached the surface at t = 0. where J is the flux of the protein per unit area of sorbent For a stagnation point flow, the flux of protein is surface; t, the time; D, the diffusion coefficient of the described by an equation derived by Dabros and Van de protein in solution; and 1r is 3.14 . Experimental data Yen [19] confirm diffusion-controlled adsorption from non-flow ing solutions of low concentration [ 10, 22, 24, 44]. In practice, however, proteins mostly adsorb from flowing solutions. Here, we will consider laminar fl ow where v is the kinematic viscosity of the solution,¢ the
98 Protein adsorption
1.5 1.5 Si02 I PS
N I N I E E 01 0'1 E LSZ E a. LA L L LSZ
a. LA
200 100 200 time/s time /s Figure 2. Adsorption of lysozyme and a-lactalbumin fro m an impinging jet fl ow (cp = 1 g dm-3) on hydrophilic silica (left) and on hydrophobic polystyrene-coated silica (rig ht). Adsorbed amount as a function of time, as determined by refl ectometry. 0.01 M phosphate buffer pH 7.0; T = 25°C. Dashed line represents the flu x of the pro tei n molecules arriving at the surface, according to equati on (3).
volume flux and R the radius of the nozzle of the device By way of example, the adsorption rates for two that supplies the solution to the sorbent surface. Equa well-characteri zed proteins, i.e., hen's egg lysozyme tion (3) is only valid under the conditions th at (a) the (LSZ) and bovine milk a-lactalbumin (aLA) at both a nozzle of the supplier is cylindricall y shaped, (b) the dis hydro ph ili c and a hydrophobic stagnati on point , are tance between the nozzle and the surface is much larger shown in Figure 2, together with the theoretical flu xes than R, (c) the particles in the solution are spheri cal and towards th e surface . Some relevant characteri sti cs of (d) interacti ons between the particles in solution are LSZ and a LA, as well as of the sorbent surfaces, are absent. summari zed in Table I . Note that the proteins are si mi Initial adsorption rates (i. e., at conditions of low lar as to their molecul ar dimensions and masses, but dif surface coverage, so that, in principle, each arriving fer markedly with respect to their isoelectric points and protein molecule can be accommodated at the surface) structural stabilities. Hence, comparative studies with may be compared with the flu xes calculated usin g the these systems may help to understand the role of hydro appropriate equation (1 ), (2) or (3). Adsorption rates phobic dehydrati on, electrostatic interacti on and protein have been measured by various authors [3 , 15 , 27, 35, structure stability in the behav iour of proteins at 56, 75] applying a wide spectrum of ex perimental tech interfaces. niques. An extensive review has been gtven by With both LSZ and aLA, th e initial adsorption rate Ramsden [65]. is smaller than the flu x. In general , the initial adsorption rates show the On the hydrophilic silica surface, the adsorbing frac same dependence on the experimental variables as ex ti on of a LA is much smaller than that of LSZ. The dif pressed in the corresponding equati ons for J. However, ference refl ects the diffe rence in electrostati c interacti on in many cases the adsorption rate is considerably smaller between th e proteins and the sorbent , i.e., attracti on of than the flux towards the surface [ 11 , 22, 31 , 58]. It LSZ and repul sion of a LA . Even for LSZ, the adsorb suggests that only a fraction of the molecul es that arri ve ing fracti on is below unity. It points to the ex istence of at the surface attach to it. The Gibbs energy, G , asso a non-electrostatic "barrier" for adsorption. ciated with that barrier, can be calcul ated from the re On the hydrophobic polystyrene surface, the situa tardation factor exp(-G/RT) [22, 58] . The cause for ti on is quite different. First, the adsorbing frac ti on of such a barrier could be electrostati c repulsion [22, 58), LSZ is somewhat larger in spite of less electrostatic at a hydrodynamic effect [31] or that a frac ti on of the pro traction. It indicates the "sti ckiness" of hydrophobic tein molecules does not coll ide in th e proper ori entati on surfaces relati ve to hydroph ili c ones. Though the a LA that is required for attachment to th e surface [4 9, 58]. molecul es are electrostati ca ll y repelled , the adsorbing
99 W. Norde
(see below under Rearrangements m the protein r structure). After adsorption, the protein molecules may (a) continue to rearrange their structure over a long period - of time [5 , 38] in order to optimi ze their interaction with ------=-=------the sorbent surface. lt is to be expected th at the degree of structural rearrangement and/or the number of molecules that undergo rearrangements depend on the rate of deposition relative to the rate of structural changes. Such long-term structural alterations involves an ex panding contact area between the pro tein molecul e and the sorbent surface; it may cause di splacement of (later adsorbed) neighbouring molecules. Such a behaviour would show up as a maximum in the adsorbed Cp amount as a function of time [22, 78]. Various authors [7 , 10, 45, 78] have attempted to Figure 3. Schematic representation of (a) a high-affinity model the dynami cs of protein adsorpti on. Although adsorption isotherm and (b) a non-high-affinity ascend these models, in one way or another, take into account ing isotherm of which the corresponding descending the various steps depicted in Figure 1, none of them give branch (dashed curve) shows hi gh-affinity. For details, a sati sfactory general description of protein adsorption. see text. It seems that heterogeneity of the sorbent surface and/or the adsorbed protein layer causes the major problem in this respect. For instance, the models should be extend Table 1. Some physical-chemical properties of th e ed to include several orientations and conformati ons of proteins and the sorbent surfaces. the adsorbed protein molecules. Furthermore, because protein adsorption usual ly proceeds irreversibly (see sec protein lysozyme a-lactalbumin ti ons: Adsorbed Amount and Desorption), the way the protein is supplied to the system may affect the fi nal molar mass (D) 14 ,600 14 ,200 result. size (nm3) 4.5x3.0x3.0 3.7x3.2x2.5 Adsorbed Amount diffusion coefficient 1.04 X 10- 10 1.06 X 10-10 (m2 s-1) The most common way to report adsorbed amounts isoelectric point 11.1 4.3 is in th e form of an adsorption isotherm, where the ad (pH units) sorbed amount, r, is pl otted against cp. Figure 3 gives Gibbs energy of schematic representati ons of the types of isotherms often denaturation (J g- 1) encountered in protei n adsorpti on. The initial part of the heat -4.1 -1.5 isotherm reflects the affi nity between the protein and the denaturant -4.0 -1.9 sorbent surface. From theory, a high-affinity , type-a, isotherm is to be expected for homodisperse polymers sorbents 0.01 M phosphate buffer pH 7.0 [25, 72]. It is indeed generall y found for relatively sim polystyrene ple, syntheti c polymers. High affinity isotherms are also silica covered silica reported for protein adsorption, but the occurrence of type-b isotherms, refl ectin g a lower affinity, is not ex thickness of the PS layer (nm) 35 ceptional. Irrespective of the affinity , protein adsorption isotherms develop, as a rule, well -defin ed plateau electrokinetic potential (mY) -48 -21 values. These values are usually compatible with, or hydrophobicity (contact angle oo somewhat !ower than, those corresponding to a complete of a sessile drop of water) monolayer of native molecul es. In some studies [30, 63], it has been observed that at pl ateau-adsorption a fraction is remarkably hi gh, much higher than that for considerable fraction of the sorbent surface is still un LSZ. This difference may be related to the rel ati vely covered. At lower adsorptions, the molecules may be low structural stability of o:LA molecul es, so that struc non-uniforml y di stributed over the surface as well and ture rearrangements contribute to the adsorption affi nity the di stribution may depend on the type of protein and
100 Protein adsorption
Table 2. Some physical-chemical properties of the dispersed sorbent particles.
0.05 M electrolyte Phosphate buffer Acetate buffer Borate buffer pH 7.0 pH 5.5 pH 9.5 PS- PS+ Glass a-F~0 3 + a-F~o 3 -
- -o- -o- Nature of charged groups -oso3 =+NH- OH2 + 2 Surface charge density (mC m· ) -23 +27 ? ? ? Electrokinetic potential (mV) -69 +32 -51 +20 -47
Hydrophobicity 82° 82° oo hydrophilic (contact angle of a sessile drop of water) 1 Specific surface area (m2 g- ) 10.0 12.4 0.5 36.0 36.0
type of surface [42, 62]. The application of novel tech unfavorable conditions. It demonstrates that the entropy niques, such as atomic force microscopy, may give more gain of the water molecules that are released from the direct information as to the heterogenei ty of the hydrophobic hydration layer dominates the adsorption adsorbed protein layer. process. The electrostati c interaction between the pro Another feature, sometimes encountered in protein tein and the sorbent surface is still reflected in the adsorption, is the occurrence of a step (two plateaus) in plateau-values of the isotherms. the isotherm. Such a step could reflect the formation of At the hydrophilic a-F~0 3 surfaces, where dehy a second protein layer, but, as dilution usually does not dration is unfavorable, adsorption of the protei ns is ex lead to the level of the first plateau, two-layer adsorption pected to be governed by electrostatic interaction. For is unlikely . The explanation is then rather a transi ti on the positively charged a-F~0 3 surface, this seems to be in the structure and/or organization of the adsorbed layer confirmed. However, at the negatively charged a-F~0 3 [9, 16, 41]. surface, aLA does adsorb in spite of overall electrostatic As a rule, dilution does not lead to detectable repulsion. Apparently, th e adverse effects of hydro desorption of the protein (which can be tested only in the philic dehydration and electrostatic repulsion are out case of non high-affinity adsorption). Hence, the as weighed by another contribution that leads to the sponta cending and the descending branches of the iso therms do neous adsorption of aLA. As suggested before in the not coincide. The occurrence of such a hysteresis indi secti on Adsorption Kinetics, this adsorption promoting cates that, at a given cp, the system has two equilibrium contribution is probably associated with structural re states, one on the ascending branch and the other on the arrangements in the protein molecule. descending branch. These two states are characterized Below, the mechanism of protein-sorbent interaction by local minima in the Gibbs energy of the system. It will be discussed in terms of the contributions of the implies that during the adsorption-desorption cycle an main forces that drive the adsorption process. irreversible physical change has occurred in the system. In spite of the irreversible nature of protein adsorption, Interactions That Govern Protein Adsorption many authors [6, 46, 47 , 60] erroneously interpret their experimental data using theories that are based on Adsorption of proteins from (aqueous) solution on reversible thermodynamics. The most common example a (solid) surface is the net result of various types of is the determination of the Gibbs energy of adsorption, interactions that simultaneously occur between all the 6adsg, by fitting the (ascending) isotherm to the components in the system: the protein molecules, the Langmuir- or Scatchard equation. For a more d ~ tailed sorhenl surfac.e, the soLvent (water) and the low-molec treatment of the irreversibility aspects, the reader is ular-weight ions. See Figure 5. referred to reference [51]. For spontaneous adsorption, at constant pressure and Figure 4 shows adsorption isotherms for LSZ and temperature, the change in the Gibbs energy must be aLA on surfaces of different electrical charge density negative. According to equation (4), this can be and hydrophobicity. Relevant characteristics of the achieved by a decrease in the enthalpy and/or an m sorbent particles are summarized in Table 2. Properties crease in the entropy. of the proteins are given in Table 1. (4) At the hydrophobic PS surfaces, both proteins ad sorb with high affinity, even under electrostatically where g, h and s are the Gibbs energy, the enthalpy and
101 W . Norde 2 PS • 2 PS-
N ___ .. I N I -· E 0 -o--o--o 0 E .....-• 0'1 0-- 0'1 E E L L o-o o_,...-o ·-·-·-·--·
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 - 3 3 Cp I g dm cP/g dm-
a Fe 0 • aFe- 0 - N 2 3 N 2 3 I I E 0 E 0'1 --o-o 0 Ol E ,.....o-o E L L 0 o..,..... o-o- 0 -·-·-·-· • •
0 0
glass- N I E .- 0'1 .--·-·-·- ---· E L--
Figure 4. Adsorption isotherms for lysozyme (e) and a-lactalbumin (0) on various surfaces. Conditions as given in Table 2; 25°C. the entropy per mol of protein and where T is the abso interconnected; for instance, di stribution of charge and lute temperature in K. In the sub-sections below, the hydrophobi c effects have a strong influence on the pro contributions from (i) the redistribution of charged tein 's structural stability [54] . groups, (ii) changes in the state of hydration and (iii) structural rearrangements in the protein molecule will be Redistribution of charged groups di scussed. It may be cl ear that these contributions are In general , both the protein molecul es and the
102 Protein adsorption
8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 Figure 5. Schematic pictures of a protein molecule in solution and a solid/solution interface (left) and a protein-covered solid surface (right). The charge of the protein originating from (de)protonation of amino acid side groups and the surface charge of the sorbent are indicated by + /-; low-molecular-weight ions are represented by (JJ I 8. Shaded areas represent hydrophobic regions.
sorbent surface are electrically charged. In an aqueous change in electrokinetic charge, .1adsaek• per unit area of environment, charged surfaces and protein molecules are sorbent surface, due to ion incorporation in adsorbed surrounded by counterions, which, together with the sur layers of LSZ and aLA on a negatively charged PS sur face charge, form an electrical double layer. When the face [32]. The data indicate that at pH < 8 the positive protein molecule and the sorbent surface approach each charge on LSZ (isoelectric point 11.3) overcompensates other, their electrical double layers overlap, which gives the negative charge on the PS surface; therefore, incor rise to a redistribution of the counterions. poration of negative charge is required to attain an (al If the protein and the sorbent surface have opposite most) electrically neutral contact region between the charge signs, they attract each other, at least if the protein and the sorbent surface. At pH 8, positively charge on the protein and the surface more or Jess com charged ions are required. Accordingly, co-adsorption pensate each other. If either one of the two components of positively charged ions accompanies the adsorption of has a large excess of charge, this would result in a con aLA (isoelectric point 4.3) on the negatively charged PS siderable net amount of charge in the contact region be surface over the entire pH region considered. tween the protein layer and the sorbent surface. This re As ion incorporation compensates for the charge an gion has a low dielectric permittivity relative to that of tagonism between the protein and the sorbent, the result water and, therefore, accumulation of charge in such an ing contribution from redistribution of charges to .1adsg environment would cause the development of an ex does not exceed values larger than a few tens of RT tremely high electrostatic potential, which is energetical [55). Its value and sign depend on the charge distribu ly very unfavorable. A similar situation would result tions and the dielectric constants of the electrical double upon adsorption of a charged protein on a surface that layers before and after adsorption, respectively [54). has the same charge sign. Nevertheless, in many cases In addition to an electrostatic effect, transferring it is observed that, in spite of such adverse electrostatic ions from an aqueous to a non-aqueous protein layer in conditions, proteins adsorb spontaneously. Based on a cludes a chemical effect as well. This chemical effect is model for the adsorbed protein layer [53], it has been unfavorable and, hence, opposes the overall protein ad predicted that low-molecular-weight-ions are transferred sorption process [33, 55]. As a consequence, maximum between the solution and the adsorbed layer to prevent affinity for protein adsorption is observed when the accumulation of net charge in the contact region between charge on the protein molecule itself just matches the the protein and the sorbent surface. The number of ions charge on the sorbent surface, so that no additional ions transferred may be deduced from electrokinetic measure are needed for charge neutralization [26]. The chemical ments [32, 52], or, more directly, by tracing labelled contribution of ion transfer to .1adsg can be estimated ions [21]. By way of example, Figure 6 shows the from model studies on the transfer of ions from aqueous
103 W. Norde
N 8 I E u 6 u ••• :1. • • 4 ~ • • cJJ -<11 • 0 0 Ill 2 0 "'0 0 -4 0 -6 0 Figure 6. Ion co-adsorption, as reflected by the overall change in the electrokineti c charge density, accompanying the adsorntion of lysozyme (0) and a-lactalbumin ( e ) on negatively charged polystyrene latex. 0.05 M KCl; T = 25°C. to nm1-aqueous media [ 1, 2, 20]. [t usually is in the drophobicity of the protein exterior influences protein range of a few to a few tens of RT. adsorption at solid water interfaces [28, 67]. Apart An alternative way to avoid the development of a from the hydrophobic parts at the aqueous periphery of high electrostatic potential in the adsorbed layer would the protein molecule, its overall hydrophobicity may be be th e; unfolding of the adsorbing protein molecules into relevant for the adsorption behaviour. The overall hy a very loose structure that is freely penetrable for water drophobicity influences the protein structure stability, and electrolyte. In such a highly hydrated adsorption which, in tum, may affect the adsorption (see next sub layer, the dielectric permittivity would not differ too section). Experimental establishment of the influence of much from that of the bulk solution. Because of the the hydrophobicity, as such, of the sorbent surface is general observation that globular protein molecul es do practically impossible because a variation of the hydro not f 104 Protein adsorption Rearrangements in the protein structure ("hard") proteins, adsorption will be primarily g vemed The densely folded structure of a globular protein hy hydrophobic dehydration and electrostati c interaction. molecule in solution is maintained because intramolecu For most hydrophobic surfaces, the contributio:t from lar hydrophobic interaction is stronger than intramolecu dehydration to t.adsg exceeds that from charge r~ distri lar electrostati c repulsion (at a pH away from the iso bution (see previous two subsections), so that, as a rule, electric point) and reduced conformational entropy of the all protein s adsorb on hydrophobi c surfaces ever; under folded structure [56] . When the protein adsorbs, it electrostatically unfavorable conditions. On hydrophilic changes its environment, which causes a shift in the ba surfaces, hard proteins adsorb only if they are electro lance of interactions. This, in tum, may lead to struc statically attracted. Proteins that have a low strUct ural tural rearrangements in the adsorbing protein molecul es. stability ("soft" proteins) are more liable to undergo For in stance, hydrophobic parts of the protein that, in an structural changes upon adsorption. The contribtttion to aqueous environment, are located in the interior of the t.adsg from th e gain in conformati onal entropY may dissolved molecule, may , after adsorption, be exposed outweigh the opposing effects from hydrophilic dehydra to the sorbent surface where th ey are still shielded from ti on and electrostatic repulsion. Under such comlitions , contact with water. Such a structure rearrangement in th e protein adsorbs spontaneously on a h y drophili ~, like volves a dec rease in imnunoleculnr hydrophobic bond charged surface (e.g., nega ti vely charged aLA ort nega ing. Because hydrophobic interacti ons in the protein in ti vely charged a-Fez03 , as presented in ;;ection terior promote the formation of secondary structures as Adsorhed Amount). a-helices and {3-sheets , a reducti on of these interactions may cause a decrease o f such secondary structures. Desorption This, in tum, leads to an increased conformational entropy. Proteins, like other macromolecul es, usually adsorb Various techniques have been used to investigate th e by attaching various segments of each of th eir molecules structure of adsorbed protein molecul es. The most com to th e surface. The fraction o f amino acid residues in mon are spectroscopic methods such as (total internal re direct contact with the surface typi cally is 10 o/c-40 %, fl ection) fluorescence [12, 17 , 36, 74], (Fourier trans whi ch, for a protein of 15,000 D molar mass, means form) infrared spectroscopy [5 , 37 , 42], circular di attachment of some 15-60 amino acid residues. ]!ven if chroism [40, 50], NMR [8] and XPS [61]. It is, howev th e contribution to t.adsg from each o f these contacts is er, often a problem to interpret the ex perimental data not more than the energ y of thermal mo ti on (1 RT), it quantitatively, in particular to di stinguish between ori adds up to several tens o f RT per mol of protein . Con entational and conformational effects. Optical tech sequently, diluting the system usually does not lead to niques, such as, ellipsometry and refl ectometry [3, 18 , desorption o f the protein. On the other hand , exchange 69, 70, 71] may also provide information that is conclu between adsorbed and dissolved protein mol ~cules, sive as to structural changes in the adsorbed protein mol whether or not of the same type, may be possible. In ecul es. Recently, differential scanning calorimetry has that case, any desorbing segment can be replaced by an been used to study denaturation of protein s at surfaces other adsorbing segment so that the initially ad;orbed [34, 83]. protein molecule is graduall y stripped off from t!te sur Based on circular di chroism measurements and in fa ce. Similarly, any oth er surface acti ve substance may frared spectroscopy considerable losses of ordered sec di splace the protein from the surface [5 7]. ondary structure have been reported [40, 50], the more ft is very possible that after release from th e sorbent so the Jess stable the structure of the native protein mol surface the structure of th e protein molecul e differs from ecule is. Furthermore, calori metry on LSZ and aLA the original , nati ve structure as it was before adsOJ"ption . adsorbed on PS and a-Fez03 surfaces [34] revealed that Usi ng circular dichroism, various authors have com adsorption induces a Joss of enthalpicall y favorable inter pared the secondary structure in proteins befofe and actions in the protei n molecules. This effect is most after adsorption. It has thus been found that the d-helix pronounced for adsorption at the hydrophobic PS surface content in bovine serum albumin (BSA) desorbed from and for aLA that has the lowest structural stabi lity. Ev various surfaces is 15-30 % less than in the native state en though protein molecul es may not completely unfold [50, 57, 68], irrespecti ve o f the desorption methml [57] . upon adsorption, the break-down of the secondary struc For fibrinogen desorbed from glass, an a-helix reduction ture that causes an increased conformational entropy of SO% has been reported [ 14] and for albumin, gl obulin could contribute with several tens of RT to t.adsg [33 , and fibrinogen desorbed from vari ous polypeptides, the 54, 55]. a-he! ix decrease was 80-90 %, 20-40 % and 0-90 %, The relevance of each of the contributions discussed respectively (73] . On the other hand , LSZ, being a above, depends on the system. For structurally stable more structurall y stable protein, regained its original 105 W. Norde Figure 7. Adsorption of 1.5 @ bovine serum albumin (BSA) N o n si li ca: (a) adsorbed 'E amounts fro m an impinging jet 3 0'1 I flow (cp = 0.01 g dm- ) E I where (X) is native BSA, ( •) is BSA previously desorbed L- I 1.0 -· from silica by morpholine, and (t.) and (D) are nati ve BSA I pre-exposed (but not adsorbed) I to silica and morpholine, I respectively; (b) adsorption isotherm for (0) native BSA I and for (e) BSA previously desorbed from sili ca by morpholine. 0.05 phosphate buffer pH 7.0; T = 25°C. 400 time Is 3.0 N I ® E 0'1 E -L 2.0 106 Protein adsorption secondary structure after desorption from oxide surfaces [5] Barbucc i R, Casolaro A, Magnani A (1992) [50] . Guoying Yan et al. [83] suggest that preceding Characterization of bi omaterial surfaces: A TR-FTI R, po contact with a polystyrene surface has a destabilizing tentiometric and cal orimetric analysis. Clinical Mater effect on the structure of di ssol ved molecules of human 11 , 37-51. serum albumin. (6) Baszkin A (1992) Surface phenomena in biocom ponent-polymer systems: a case study of mucin adsorp Re-adsorption of Pre-adsorbed Protein ti on on polymers with different hydrophilicities. Clinical M ater 11, 119-123. According to the scheme depi cted in Figure 1, de [7) Beissinger RL, Leonard EF (1982) Sorption ki sorbed molecules may re-adsorb at the same sorbent sur neti cs of binary protein solutions: general approach to face. If the desorbed molecul es have not regained their multicomponent systems. J Coll oid Interface Sci 85, original structure, it is expected that the adsorption char 52 1-533 . acteristics of the pre-adsorbed and subsequentl y desorbed [8] Benko B, Yuk-Pav lovic S, Dezeli c G , Maricic protein are different from those of the native protein . S ( 1975) A proton magneti c relaxati on study of methae For BSA and aLA, both "soft " proteins that are likely moproteins bound to monodisperse polystyrene latex par to undergo structural changes upon adsorption, such an ticl es. J Coll oid Interface Sci 52, 444-451 . influence has indeed been observed [ 49). In the Figures (9] Blomberg E, Claesson PM , Golander C-G 7a and 7b, thi s is illustrated for BS A. Both the kineti c ( 199 1) Adsorbed layers of human serum albumin investi data and the adsorption isotherms indicated that the gated by the surface force technique. J Dispersion Sci desorbed protein has an increased affinity for adsorption. Techno! 12, 179-200. [1 0] Born zin GA, Miller IF ( 1982) The kinetics of Conclusions pro tein adsorption on syntheti cs and modified surfaces. J Coll oid Interface Sci 86, 539-558. The adsorption of proteins fro m aqueous solution J!IJ BosMA, Shervani Z, Anusiem ACl , Giesbers onto a solid surface is the result of an interplay betwt!en M, Norde W , Kl eij n JM (1 994) Jnfluenct! of the electric several subprocesses. Di stincti on must be made between potenti al of the interface on th e adsorption of proteins. "hard" proteins, of which th e molecul es retain most of Colloids Surfaces B: Bi ointerfaces 3 , 9 1-1 00. their conformation upon adsorption, and "soft " protein s [ 12] Burghardt T P, Axelrod D ( 1983) Total internal that undergo severe structural rearrangements. Adsorp refl ec ti on flu orescence study of energy transfer in tion of the hard proteins can be interpreted in terms of surface-adsorbed and di ssolved bovine serum albumin. electrostatic interactions and (partial) dehydration of the Bi ochemi stry 22, 979-9 85. outer surfaces of the sorbent and the protein . The inter [13] Chaiken 1M , Shai Y, Fassina G, Calicetti P nal structural changes occurring in the adsorbing soft ( 1988) Current development of analyti cal affinity chro proteins involve an increase in the conformati onal en matography: design and biotechn ological uses of molec tropy of the protein molecule and constitute, therefore, ul ar recognition surfaces. Makromol Chem Macromol an additional driving force for spontaneous ad sorption. Symp 17 , 269-280. [ 14] Chan BMC, Brash JL ( 1981 ) Adsorption of fi References brinogen on glass: reversibility aspects. J Coll oid Inter face Sci 82, 217-225. [1) Abraham MH (1973) Ionic entropi es of transfer [15) Chittur KK , Fink DJ , Leininger Rl, Hutson TB from water to non aqueous solvents. J Chem Soc ( 1986) Fourier transform infrared spectroscopy/atten Faraday Trans I 69, 1375-1388. uated total re fl ecti on studies of proteins in fl owing sys [2] Abraham MH, Ah-Sing E , Danil de Namor AF, tems: approaches for bulk correction and compositional Hill T, Nasehzadeh A, Schultz RA ( 1978) Heats of solu analysis of adsorbed and bulk proteins in mixtures. J tion of electrolytes in ethanol and derived enthalpies of Coll oid Interface Sci 111 , 419-433 . transfer from water. J Chem Soc Faraday Trans I 74, [ 16] Claesson P ( 199 1) The surface force group. 359-365. KTH /YKI Ytkemi ska Annual Report 199011991. [3) Arnebrant T, Nylander T (1986) Sequential and lnstitutt! fo r Surface Chem., P.O. Box 5607, S- 11486 competitive adsorption of ,6-lactoglobulin and K- casein on St ockh olm, Sweden. p. 20. metal surfaces. J Colloid Interface Sci 111 , 529-5 33. [1 71 Clark SR, Billsten P, Elwi ng H ( 1994) A flu o [4] Bae YH , Okano T , Kim SW (1990) Temperature resct!nce technique for investi gating protein ad sorption dependence of swelling of cross linked poly (N , N' -alkyl phenomena at a colloidal silica surface. Colloids substituted acrylamides in water. J Polym Sci Polym Surfaces B: Biointerfaces 2, 457-461. Phys 28, 923-936. [1 8] Cuypers PA, Hennens WT, Hemker HC 107 W. Norde (1977) Ellipsometric study of protein film on chromium. tei ns at a polystyrene-water interface. J Colloid Interface Ann NY Acad Sci 283, 77-87. Sci 164, 394-409. [19] Dabros T, Van de Yen TGM (1983) A direct [33] Haynes CA, Norde W (1994) Globular proteins method for studying particle deposition onto solid sur at solid-liquid interfaces. Colloids Surfaces 8: Biointer faces. Colloid Polym Sci 261, 694-707. faces 2 , 517-566. [20] Danil de Namor AF, Contreras E, Sigsted E [34] Haynes CA, Norde W (1995) Structures and (1983) Free energies of transfer of ions from water to stabilities of adsorbed proteins. J Colloid Interface Sci butan-1-ol. J Chern Soc Faraday Trans I 79 I 001-1007. 169, 313-328. [21] van Dulm P, Norde W, Lyklema J (1981) Ion [35] Hlady V, Rei necke DR, Andrade JD (1989) participation in protein adsorption at solid surfaces. Fluorescence of adsorbed protein layers. J Colloid Inter Colloid Interface Sci 82,77-82. face Sci 111, 555-569. [22] van Dulm P, Norde W (1983) The adsorption [36] Hlady V, Andrade JD (1989) A TIRF titration of human plasma albumin on solid surfaces with special study of 1-anilinonaphtalene-8-sulfonatebinding to silica attention to the kinetic aspects. J Colloid Interface Sci adsorbed bovine serum albumin. Colloids Surfaces 42, 91, 248-255. 85-96. [23] Elwing H, Askendal A, Nilsson U, Lundstrom [37] Jacobsen RJ, Wasacz FM ( 1987) Effects of the I (1987) A wettability gradient method for studies of environment on the structure of adsorbed proteins: macromolecular interactions at the liquid-solid interface. Fourier Transform Infrared spectroscopic studies. ACS J Colloid Interface Sci 119, 203-210. Symp Ser 343, 339-361. [24] van Enckevort HJ, Dass DV, Langdon AG [381 Jennissen HP (1985) Protein adsorption hyster (1984) The adsorption of bovine serum albumin at the esis. In : Surface and Interfacial Aspects of Biomedical stainless steel -aqueous solution interface. J Colloid Polymers. Vol. 2. Andrade JD (ed.). Plenum Press, Interface Sci 98, 138-143. New York. pp. 295-320. [25] Fleer GJ, Lyklema J (1983) Adsorption of pol [39J Klein F, Bronsveld W, Norde W, van ymers In: Adsorption from Solution at the Solid/ Liquid Romunde LKJ, Singer JM (1979) A modified latex-fixa Interface. Parfitt GD, Rochester CH (eds.). Academic tion test for the detection of rh eumatic fa ctors. J Chem Press. pp. 153-220. Pathol 32, 90-92. [26] Fraaije JGEM (1987) Interfacial thermodynam (40] Kondo A, Oku S, Higashitani K (1991) Struc ics and electrochemistry of protein partitioning in two tural changes in protein molecul es adsorbed on ultrafine phase systems. Ph. D . Thesis, Wageningen Agricultural silica particles. J Colloid Interface Sci 143, 214-221. University, The Netherlands. [41] Koutsoukos PG, Mumme-Young CA, Norde [27] Go lander CG, Kiss E ( 1988) Protein adsorption W, Lyklema J ( 1982) Effect of the nature of the sub on functionalized and ESCA-characterized polymer films strate on the adsorption of human plasma albumin. Col studied by ellipsometry. J Colloid Interface Sci 121 240- loids Surfaces 5 , 93-104. 253. [42] Lenk TJ, Ratner BD, Chittur KK, Gendreau [28] Gorbunov AA, Lukyanov A Y, Pasechnik VA, RM ( 1991) Infrared spectroscopi c studies of time-de Vakrushev A V ( 1986) Computer simulation of protein pendent changes in fibrinogen adsorbed to polyure adsorption and chromatography. J Chromatography 365, thanes. Langmuir 7, 1755- 1764. 205-212. [ 43] Leveque M ( 1928) Les lois de Ia transmi ssion [29] Hallstrom B, Lund DB Tragardh C (1981) Fun de chal eur par convection (The laws of heat transport by damentals and applications of surface phenomena associ convection). Ann Mines 13 , 284. ated with fouling and cleaning in food processing. Divi [44] Lok BK, Cheng Y-L, Robertson CR (1983) To sion Food Engineering, Lund University, P .O. Box 50, tal internal reflection fluorescence: a technique for ex S-23053 Alnarp, Sweden. amining interactions of macromolecul es with solid sur [30] Hansen J, Ely K, Horsley D, Herron J, Hlady faces. J Colloid Interface Sci 91, 87-103. V, Andrade J (1988) The adsorption of lysozymes: a [ 45] Lundstrom I ( 1985) Models of protein adsorp model system. Makromol Chem Macromol Symp 17, tion on solid surfaces. Progr Colloid Po!ym Sci 70, 76- 135-154. 82 . [31] Hasegawa M, Kitano H (1992) Adsorption ki [46] Moreno EC, Kresak M , Hay Dl (1991) Ad netics of proteins onto polymer surfaces as studied by sorption of salivary proteins onto Ca-apatites. Biofouling the multiple internal reflection fluorescence method. 4, 3-24. Langmuir 8, 1582-1586. [47] Nag A, Sadhukan B, Chattoraj DK (1988) [32] Haynes CA, Sliwinsky E, Norde W (1994) Thermodynamics of adsorption of proteins and Structural and electrostatic properties of globular pro- carbohydrates. J Surf Sci Techno! 4, 91-109. 108 Protein adsorption [48] Nemethy G, Scheraga HA (1962) Structure of Busscher HJ, van der Mei HC (1991) The bacterial water and hydrophobic bonding in proteins. J Chern colonization of intra-oral hard surfaces in vivo: influence Phys 36, 3401-3417. of surface free energy and surface roughness. Biofouling [49] Norde W , Anusiem ACI (1992) Adsorption, 4, 187-198. desorption and re-adsorption of proteins on solid sur [65] Ramsden JJ (1993) Experimental methods for faces. Colloids Surfaces 66, 299-306. investigating protein adsorption kinetics at surfaces. [50] Norde W, Favier JP (1992) Structure of ad Quart Rev Biophys 27, 41-105. sorbed and desorbed proteins. Colloids Surfaces 64, 87- [66] Rechnitz GA (1988) Biosensors. Chern Engin 93 . News 66(36), 24-36. [51] Norde W, Haynes CA ( 1995) Reversibility and [67) Regnier FE (1987) The role of protein structure the mechanism of protein adsorption. ACS Symp Ser, in in chromatographic behaviour. Science 238, 319-323. press. [68] Sato H, Tomiyama T, Morimoto H, Nakajima [52] Norde W, Lyklema J (1978) The adsorption of A (1987) Structure and activity changes of proteins human plasma albumin and bovine pancreas ribonuclease caused by adsorption on material surfaces. ACS Symp at negatively charged polystyrene surfaces. III Electro Ser 343, 76-87. phoresis J Colloid Interface Sci 66, 277-284. [69] Schaaf P, Dejardin P (1988) Structural changes [53] Norde W, Lyklema J (1978) The adsorption of within an adsorbed fibrinogen layer during the adsorp human plasma albumin and bovine pancreas ribonuclease tion process: a study by scanning angle retlectometry. at negatively charged polystyrene surfaces. IV The Colloids Surfaces 31, 89-103 . charge distribution in the adsorbed state. J Colloid Inter [70] Schaaf P, Dejardin P, Johner A, Schmitt A face Sci 66, 285-294. ( 1987) Thermal denaturation of an adsorbed protein lay [54] Norde W, Lyklema J (1979) Thermodynamics er studied by retlectometry . Langmuir 3 , 1128-1131. of protein adsorption. Theory with special reference to [71] Schaaf P, Dejardin P, Schmitt A (1987) Reflec the adsorption of human plasma albumin and bovine tometry as a technique to study the adsorption of human pancreas ribonuclease at polystyrene surfaces. J Colloid fibrinogen at the silica/solution interface. Langmuir 3, Interface Sci 71, 350-366. 1131-1135. [55) Norde W, Lyklema J (1991) Why proteins pre [72] Scheu0ens J M, Fleer GJ ( 1979) Statistical theo fer interfaces. J Biomater Sci Polymer Edn 2, 183-202. ry of the adsorption of interacting chain molecules. J [56] Norde W, Rouwendal E (1990) Streaming po Phys Chem 83, 1619-1635. tential measurements as a tool to study protein adsorp [73] Soderqui st ME, Walton AG (1980) Structural tion kinetics. J Colloid Interface Sci 139, 169-176. changes in proteins adsorbed on polymer surfaces. J [57] Norde W, MacRitchie F, Nowicka G, Lyklema Colloid Interface Sci 75, 386-397. J (1986) Protein adsorption at solid-liquid interfaces: re [74] Tan JS, Martie PA (1990) Protein adsorption versibility and conformation aspects. J Colloid Interface and conformational change on small polymer particles. Sci 112, 447-456. J Colloid Interface Sci 136, 415-431. [58] Norde W, Arai T , Shirahama H (1991) Protein [75] Voegel JC, Baillou N, Sturm J, Schmitt A adsorption in model systems. Biofouling 4, 25-36. (1984) Adsorption kinetics of plasma proteins onto sili [59] Northrup SN (1988) Diffusion-controlled ligand ca. Colloids Surfaces 10, 9-19. binding to multiple competing cell-bound receptors. J [76] Vroman L ( 1987) Methods of investigating pro Phys Chern 92, 5847-5850. tein interactions on artificial and natural surfaces. Ann [60] van Oss CJ (1991) The forces involved in hie NY Acad Sci 516, 300-312. adhesion to flat surfaces and particles - their determina [77] Vroman L, Leonard EF (1991) The effect of tion and relative roles. Biofouling 4, 25-36. flow on displacement of proteins at interfaces in plasma. [61) Paynter RW, Ratner BD (1985) The study of Biofouling 4, 81-88. interfacial proteins and biomolecules by X-ray photo [78] Walton AG, Soderquist ME (1980) Behaviour electron spectroscopy. In: Surface and Interfacial As of proteins at interfaces. Croat Chem Acta 53, 363-372. pects of Biomedical Polymers. Vol. 2. Andrade JD [79] Ward AFH, Tordai L (1946) Time-dependence (ed.). Plenum Press, New York. pp. 189-216. of boundary tensions of solutions. J Chern Phys 14, 453- [62] Pitt WG ( 1986) Sequential protein adsorption 461. and thrombus deposition on polymeric biomaterials. J [80] White DJ ( 1991) Processes contributing to the Colloid Interface Sci 111, 343-362. formation of dental calculus Biofouling 4, 209-218. [63] Price TM, Rudee ML (1987) Early stages of [81] Willems GM , Hermens WT, Hemker HC plasma protein adsorption. ACS Symp Ser 343, 48-62. (1991) Surface exclusion and molecular mobility may [64] Quirynen M, Marechal M, van Steenberghe D, ex plain Vroman effects in protein adsorption J Biomater 109 W. Norde Sci Polymer Edn 2, 217-226. PS- and PS +. A greater surface density corresponds [82] Wojciechowski P, Brash JL ( 1991 ) The with a smaller electrokinetic potential. Vroman effect in tube geometry: the influence of fl ow Author: The di sagreement between the surface charge on protein adsorption measurements. J Biomater Sci Pol densities and the electrokineti c potentials of PS- and ymer. Edn. 2, 203-216. PS + (see Table 2) is apparent. The electrokinetic po [83] Yan G, Li J, Huang S-C, Caldwell KD ( 1995) tentials are for PS latexes in 0.05 M phosphate buffer Calorimetric observations of protein conformation at pH 7. Under these conditions, the H2Po4· and H2Poi· solid/liquid interfaces. ACS Symp Ser, in press. have a relati vely strong tendency to adsorb to the PS + surface, thereby reducing the electrokineti c potential. Discussion with Reviewers The surface charge densities, given in Table 2, refer to the charge originating from the groups that are covalent R. Hidalgo Alvarez: Protein adsorption is a typical ir ly linked at the PS surface, i.e., -oso3· and = + NH, reversible process. Why is not the thermodynamics of respecti vely . irreversible processes used to describe adsorption proc esses? R. Hidalgo Alvarez: The adsorption isotherms of LSZ Author: The overall protein adsorption process is a and aLA on PS- do not develop well-defined plateau complex process, involving vari ous sub-processes that values. Pl ease comment. are linked to each other. The aim of thi s paper is to es Author: The reason(s) why the isotherms for LSZ and timate the contributions from each of these subprocesses aLA do not develop a definite plateau value (at cp < to the Gibbs energy of adsorption, which, at constant 0.5 g dm-3) are not clear. In view of the relatively large temperature as pressure is a measure for the affinity of amount of LSZ adsorbed, it could be that a second layer the protein to adsorb at the surface. The Gibbs energy of LSZ is built up at the surface. This would be facili is a fun cti on of state and its change depends on the final tated by the fact that the LSZ-covered PS- particle still and initial states only. Therefore, although the process has a negati ve electroki neti c potential , whereas the ad proceeds irreversibly, the change in the Gibbs energy sorbing LSZ molecul es are positively charged. due to adsorption may as well be calcul ated using ther modynamics for reversible processes. Irreversibl e th er R. Hidalgo Alvarez: The change in electrokineti c modynamics deals with entropy producti on. Minimum charge per unit area of sorbent surface, due to ion incor entropy production for the overall process may be poration in adsorbed layers of LSZ and a LA, has been estimated from hysteresis of the adsorption isotherms. analyzed only on PS-. What is the change on PS +? For this I refer to references [33] and [34]. How was the electrokineti c potential of the protein-latex calculated? According to Oshima and Kondo {J. Colloid R. Hidalgo Alvarez: The di sagreement between the Interface Sci. 130 ( 1989) 281}, the electrokinetic poten theoretical and experimental fluxes requires a more ex ti al loses its meanin g for coll oidal particles with a struc tensive di scussion. I assume that the theoreti cal flux es tured surface, si nce the electrophoretic mobility is insen were calculated using the equation [3] . sitive to the precise position of the slipping plane. Does Author: There is no disagreement between theoreti cal the author trust hi s results as shown in Figure 6? (calculated, using eq. [3]) and ex perimental flux es, sim Author: We have not made such ex tensive, pH-depend ply because we did not determine the flux experimental ent protein adsorption studies with PS +. The electroki ly. There is a difference between the calculated flux to neti c potentials of the (protein-covered) latex particles wards the surface and the initial rate of adsorption. were derived from the electrokinetic mobilities, using However, this difference does not imply di sagreement , the theory of O'Brien and White {J. Chem. Soc. Far because it can well be that only a fraction of the molec aday Trans. 2 74 (1978) 1607} . I agree that if the col ules that arrive at the surface reall y attach to it. There loidal particle is covered with an ion-penetrable polymer may be various reasons for this phenomenon (as men layer, calculation of the electrokinetic potential becomes tioned in the secti on Adsorption Kinetics of the paper) , more complex . However, based on a vari ety of litera such as: electrostatic repulsion between the protein and ture data, it is generally accepted that after adsorption, the sorbent surface, unfavorable dehydration of a hydro the protein molecul es remain relatively co mpact. Ion philic surface and unfavorable orientation of the protein incorporation is assumed to occur in the contact region molecule when it arrives at the sorbent surface. between the protein and the sorbent surface. In view of this, I trust our results, shown in Figure 6, in a semi R. Hidalgo Alvarez: I would like to know how the au quantitative way, namely , that they represent the trends thor explains the disagreement between the surface by which ion inco rporation compensates for the charge charge densi ties and the electrokineti c potenti als of bare antagoni sm between the protein and the sorbent surface. 110 Protein adsorption R. Hidalgo Alvarez: Hydrophobic dehydration oft en mined by the analysis of experimental results which al overrules electrostatic effects. However, it would be most always include the effect of diffusion, concentra very useful to know if only dehydration of hydrophobic ti on and the surface itself. surfaces would be sufficient to cause protein adsorption Author: I agree that care should be taken when deter on these surfaces. Otherwise, which is the determining mining the contributions from transport and surface reac factor in the protein adsorption by hydration changes, ti on to the adsorption rate. At low cp and in the initial the protein or the hydrophobic surface? stage of the adsorption process, i.e. , when the degree of Author: Dehydration of a hydrophobi c sorbent surface coverage of the sorbent surface by the protein is ex often is the determining factor for protei n adsorption. tremely low, the adsorption rate is ex pected to be con This may be illustrated by the observation that a given trolled by the transport towards the surface. However, protein does adsorb at a hydrophobic surface that has the whether or not this is really the case can only be experi same charge sign as the protein, whereas adsorption of mentally verified under well-defined conditions where that protein does not occur at a hydrophilic surface un the rate of transport is predicted by theory. In the der otherwise similar conditions. paper, I have discussed a few examples of such conditions. W.G. Pitt: Conformational stability appears to be a sig nificant issue in this paper, regarding whi ch I have some K. K. Chittur: Under Adsorption Kinetics, the auth or questions. describes events at th e liquid solid interface. It is our - How was conformational stability measured? view that we must, at al l times, consider th e diffusion - Can it be quantified wi thout usi ng adsorpti on as a from the bulk and surface reaction together, not as sepa measure? rate event s. Yes, surface reacti on will lead to depleti on - If conformational stability is measured by DSC, but only if the reacti on is fast and if there is not suf how can you ensure that thermally-induced conforma ficient protein in the bulk to replace it fast enough. For tional changes are relevant , similar or analogous to the example, in a system where a protein such as albumin surface-induced conformational changes? were to adsorb fa st, but is also found at a relatively high Author: Adsorption data for LSZ and aLA are com concentration in th e bulk, the depl eti on at the interface pared to illustrate the influence of protein conformational may be very, very short li ved, if at all. stability on the adsorption behaviour. The conforma Author: Again, I agree with thi s conunent. The dis tional stabilities are quantified in terms of the Gibbs en cussion in th e text is just meant to indica te how trans ergy of unfolding. As shown in Table l , the confonna port-limited adsorption rates can be ve rified. tional stability of LSZ is much greater than that of aLA. This is true for both heat-induced and denaturant K. K. Chittur: Towards the end of Adsorption Kinet (guanidinium chloride)-induced conformational changes. ics , th e author cites the lack of appropriate models that It indicates that the internal coherence in the LSZ mole describe protein adsorption. I agree, however it is im cules is far greater than in the aLA molecul es. It is, portant to keep in mind the specific obj ectives. We therefore, likely , although not a priori sure, that aLA is would like to di stingui sh between what we call "macro more susceptible to conformational changes upon adsorp scopi c" models that attempt to describe by reacti on dif tion. The observation that, in contrast to LSZ, aLA ad fusion equations the time course o f protein adsorption to sorbs on a hydrophilic, like-charged surface (see Figure surfaces. M odels that have to include the heterogeneity 4) suggests conformational rearrangements in thi s pro of surfaces, protein-protein interactions and so on have tein as a driving force for adsorption. Furthermore, di f to rely on mi croscopi c models that need detailed molecu ferential scanning calorimetry reveals that the breakdown lar models for both protein and the surface. We must of ordered structures in adsorbed Ct'LA is much greater th en carefull y average molecular intera cti ons to obtain than in adsorbed LSZ [34]. macroscopi c estimates, such as rate constants for adsorp tion and so on. K.K. Chittur: At the beginning of Adsorption Kinet Author: Yes, and I think th at there is a lack of appro ics, the author states that the intrinsic rate of adsorption priate models of both kinds. of proteins with the surface is independent of bulk con centration. I will not argue with this point, except to K. K. Chittur: In tex t, the statement (just before say that for some surfaces, even at low cp, the surface Adsorbed Amount): "the way the protein is supplied to reaction may be rate limiting. lt is important to careful the system may affect the final result" needs to be ly distinguish between th e intrinsic rate constant for clarified . adsorption which is determined by properties of the pro Author: The statement "the way the protein is supplied tein and the surface and the rate constant that is deter- to the system may affect the tina! result" follows from 111 W. Norde the generally observed irreversibility of the adsorption process. In other words, during the time scale of the experiment, the protein molecules do not fully relax to their equilibrium state. This is reflected, for instance, in a hysteresis between the ascending and descending branches of the adsorption isotherm (33]. While the final result does not represent true (thermodynamic) equilibrium it is, in principle, dependent on the history of the system and, therefore, on the way the protein is supplied. K.K. Chittur: Equation (4) appears to be written for the adsorbed layer. Should not the llg for the entire system (i.e., adsorbed protein, non-adsorbed protein, water etc.) be considered in the analysis? It is possible that the authors have done that elsewhere, clarification would help. Author: Equation (4) refers to the overall adsorption process. This process is analyzed in terms of its most relevant contributions, i.e., electrostatic interactions, changes in the state of hydration and structural rear rangements in the protein. By doing so, all the compo nents involved, i.e., the protein, the water, the low mo lecular weight electrolyte, are taken into account. lt is tacitly assumed that the sorbent surface is rigid (does not undergo structural changes) and that the protein solution is ideally diluted, so that further dilution (due to adsorp tion) does not affect the molar Gibbs energy of the dis solved protein. 112