ARTICLE pubs.acs.org/est
Protein Encapsulation by Humic Substances Jeanne E. Tomaszewski, Rene P. Schwarzenbach, and Michael Sander* Institute of Biogeochemistry and Pollutant Dynamics (IBP), ETH Zurich, Zurich, Switzerland
bS Supporting Information ABSTRACT: Protein encapsulation by natural organic matter is hypothesized to preserve the activity of proteins in terrestrial and aquatic environments. Direct molecular-level evidence for encapsulation of net positively charged proteins lysozyme, trypsin, and ribonuclease A by a diverse set of humic substances (HS) in nanostructured films was collected using a combination of optical waveguide lightmode spectroscopy and quartz crystal microbalance measurements. The results suggest that proteinHS electrostatic attraction drives encapsulation of positively charged lysozyme by a soil humic acid at pH 5 to 8 and by six additional humic and fulvic acids from terrestrial and mixed terrestrial aquatic sources at pH 5 and 6. Encapsulation of trypsin and ribonuclease A, which had negatively charged surface patches under the studied conditions, suggested that localized proteinHS electrostatic repulsion is overcompensated by attractive forces, likely including contributions from the hydrophobic effect. Evidence is provided showing that encapsulation of lysozyme at pH 8 and of ribonuclease A at pH 5 and 6 involved partial disassembly of HA supramolecular associations. This work advances a molecular-level picture of protein encapsulation by HS and presents a novel approach to study the effects of encapsulation on protein enzymatic activity and susceptibility to abiotic and biotic transformations.
’ INTRODUCTION Enzymes from bacteria, plants, and fungi are pertinent to the biogeochemical cycling of elements in terrestrial and aquatic environments. Whereas most enzymes and proteins in solution are considered labile and readily biodegradable, they may be preserved by adsorption to and by encapsulation in natural organic matter (NOM).15 Encapsulation by NOM may also affect the environmental fate of toxic and infectious proteins, including insecticidal Cry proteins from genetically modified Bt crops and prion proteins causing sheep scrapie and cervid chronic wasting disease.68 While encapsulation may, in the short term, reduce the bioavailability and activity of these proteins, it may lead to their accumulation in the long term. Despite its relevance, the mechanisms of protein adsorption to and encapsulation in NOM are still rather poorly understood. Previous studies on protein encapsulation used humic substances (HS) as models for NOM.3,4,911 HS are viewed as being composed of chemically heterogeneous, relatively low molecular weight molecules (18 MΩ cm) and by adjusting the pH with NaOH or HCl (Fluka). The total ionic strength was brought to I = 10 mM and 50 mM by addition of NaCl (Fluka). Proteins. High-purity, lyophilized hen egg white lysozyme, trypsin, and ribonuclease A (bovine pancreatic, type III-A) were from Sigma, and used as received. Some key properties of the proteins pertinent to their adsorption are given in Table 1. Aqueous
enzyme stocks (100 μg mL1) were prepared by dissolving lysozyme in 10 mM NaCl solution, trypsin in 1 mM HCl solution, and ribonuclease A in Milli-Q water. Stocks were aliquoted into Protein Lo-Bind tubes (Eppendorf) and stored at 20 °C until use. For QCM-D and OWLS experiments, the stocks were diluted to concentrations of 10 μg mL1 in appropriate pH and I-adjusted buffer solutions. Trypsin dilutions contained 1 mM CaCl2 to inhibit autoproteolysis. HS and Poly-L-lysine (PLL). Standard HS were purchased from the International Humic Substances Society (IHSS) and included Elliott Soil (ES), Pahokee Peat (PP), and Suwannee River (SR) humic and fulvic acids (HA and FA), and Leonardite (L) HA. Key physicochemical properties of these HS are given in Table S1 of the Supporting Information. Aqueous stocks were prepared by dissolving HS in Milli-Q water at 0.5 mg HS mL1 in a sonication bath at pH 9 (addition of NaOH). For experiments, the HS stocks were diluted 10 fold in appropriate pH and I-adjusted buffer solutions to concentrations of 50 μg HS mL1 and were used within a few days of preparation. Polyelectrolyte solutions were prepared by dissolving poly-L-lysine hydrobromide (PLL) (MW = 70150 kDa, Fluka) in pH- and I-adjusted buffer solutions to final concentrations of 100 μg mL1. Methods. Optical Waveguide Lightmode Spectroscopy (OWLS). Experiments were conducted on an OWLS 110 instrument (Microvacuum Ltd.; Budapest, Hungary) equipped with a laminar slit shear flow cell through which solutions were passed over a planar waveguide. Adsorption to the waveguide surface is quantified via changes of the interfacial refractive index sensed by the interfacial evanescent field of a HeNe laser that couples into the waveguide.17 Assuming an optically uniform adsorbed layer, the sensed (optical) adsorbed mass, ΔmOWLS [ng cm2], is given as: ΔmOWLS ¼ dadlayer
nadlayer nsolution dnadsorbate =dCadsorbate
ð1Þ
where dadlayer [cm] and nadlayer [-] are the thickness and the refractive index of the adlayer, respectively. The refractive indices of the buffer solutions, nsolution, were measured with a refractometer (Model J357, Rudolph Research Analytical, NJ, USA) (1.3336; pH 5 to 8, I = 10 mM, 20 °C). The refractive index increment, dnadsorbate/dC adsorbate, of 0.139 cm3 g1 was used for the polyelectrolytes,17 0.182 cm3 g1 for the proteins,18 and 0.28 cm3 g1 for 6004
dx.doi.org/10.1021/es200663h |Environ. Sci. Technol. 2011, 45, 6003–6010
Environmental Science & Technology
ARTICLE
Figure 1. a. Steps involved in the layer-by-layer encapsulation of intact proteins in humic substance (HS) films. The first HS layer (HS1) was adsorbed to poly-L-lysine (PLL) films on quartz (SiO2) surface supports. The superscript denotes the total number of protein and HS adlayers. b. Differences in the sensed adsorbed masses between optical waveguide lightmode spectroscopy (OWLS) and quartz crystal microbalance with dissipation monitoring (QCM-D), shown for HS2. OWLS senses the dry adsorbed mass, whereas QCM-D additionally senses water between HS and protein molecules and intra-HS water.
HS.19 Experiments were conducted using SiO2-coated OWLS waveguides, as described in ref 20. Quartz-Crystal Microbalance with Dissipation Monitoring (QCM-D). Experiments were conducted on an E4 system (Q-Sense AB, Gothemburg, Sweden) equipped with four flow cells, each containing a piezoelectric quartz sensor. QCM-D monitors the shifts in the resonance frequencies (Δf [Hz]) and energy dissipation (ΔD [-]) of the fundamental frequency (= 5 MHz) and of multiple oscillation overtones (n) of the piezo-sensor upon macromolecule adsorption to and desorption from the sensor surface. For rigid adlayers, the total QCM-D sensed mass, Δm QCM-D [ng cm2], which includes the dry adsorbed mass, Δmadsorbate, and the mass of water that couples to the oscillating sensor, ΔmH20, is given by the Sauerbrey relation:21 ΔmQCM-D ¼ Δmadsorbate + ΔmH2 O ¼ c
Δfn n
ð2Þ
where c (= 17.7 ng Hz1 cm2) is the mass sensitivity constant for the used sensors, and n is the overtone number. All systems showed small dissipation (ΔD < 106 for Δf g 10 Hz) and Δfn/n overlapped for all n, indicating that the adlayers were rigid and, therefore, that eq 2 could be applied. Results for n = 5 (25 MHz) are reported. Experiments were conducted on SiO2coated sensors (QXS303, Q-Sense). Solutions were delivered through the OWLS and QCM-D cells by a peristaltic pump connected to cell outlets (flow rates of 50 μL min1 at a temperature of 20.0 ( 0.1 °C). After equilibration of sensors/waveguides to adsorbate-free buffer solutions, encapsulation was studied as depicted in Figure 1 by sequentially running the following solutions over the waveguides/ sensors: (i) PLL solutions, followed by rinsing, (ii) HS solution, followed by rinsing, (iii) protein solution, followed by rinsing, (iv) HS solution, followed by rinsing. Rinsing with adsorbate-free solutions between the steps was started when apparent adsorption equilibrium was attained (i.e., stable signal over time). Steps (ii) and (iii) were repeated up to three times. The pH and ionic strength were kept constant in each encapsulation experiment. For trypsin, systems were additionally rinsed with a 1 mM CaCl2 solution (while maintaining a total I = 10 mM, adjusted by NaCl)
to minimize autoproteolysis. At least duplicate measurements were conducted for each tested proteinHSpH combination. Adsorption of HS and protein were analyzed for initial adsorption rates, kads [ng cm2 min1], which were determined by fitting the initial linear increase in ΔmOWLS over time shortly after introducing the adsorbate-containing solution, and for the final changes in adsorbed masses, ΔmOWLS and ΔmQCM-D [ng cm2]. Only kads from OWLS measurements are reported because these values represent changes in absolute (dry) mass. Furthermore, the geometry of the OWLS cell allows for calculation of the transport-limited adsorption rate, kadsmax. The value for lysozyme (81.4 ng cm2 min1) was previously determined at the same lysozyme concentration and volumetric flow rate.20 For ESHA, kadsmax was approximated by the kads to PLL, a system in which rates were likely transport limited due to strong PLL-HA electrostatic attraction. For lysozyme and HS, kads values were expressed relative to the maximum transport-limited adsorption rates kadsmax, in form of the adsorption efficiency R:22 R¼
kads kmax ads
ð3Þ
The thicknesses of HS adlayers were estimated from ΔmQCM-DHS assuming HS adlayer densities of 1.05 g cm3. The OWLS and QCM-D systems and sensors were thoroughly cleaned after each use, as detailed elsewhere.20
’ RESULTS AND DISCUSSION Encapsulation of Lysozyme by Elliott Soil HA Studied by OWLS and QCM-D. Part a of Figure 2 shows the increase in the
dry adsorbed mass, ΔmOWLS, upon initial adsorption of PLL to the SiO2 waveguide and upon the subsequent alternating adsorption of ESHA and lysozyme, all at pH 5. A total of four lysozyme adlayers were successfully encapsulated (Figure S1 of the Supporting Information). The stepwise increase in ΔmOWLS demonstrates encapsulation of entire lysozyme molecules by ESHA. Encapsulation was likely driven by ESHAlysozyme electrostatic attraction for three reasons. First, at pH 5, lysozyme carried a high, uniform positive surface charge (Table 1), whereas ESHA was negatively charged (Table S1 of the Supporting Information). Second, the adsorption rates of lysozyme to ESHA and of ESHA to PLL were close to transport-limited (R = 0.70.9, part b of Figure 2), suggesting that no energy barriers to adsorption of lysozyme and ESHA existed. Third, adsorption of lysozyme to ESHA and of ESHA to lysozyme was virtually irreversible upon buffer rinsing (part a of Figure 2, Figure S1 of the Supporting Information), indicating strongly attractive interactions. Lysozyme ESHA electrostatic attraction is further supported by changes in the zeta-potentials of SiO2 particles upon successive coating with a positively charged polyelectrolyte, ESHA, lysozyme, and ESHA, as detailed in Figure S2 of the Supporting Information. The lysozyme mass encapsulated by ESHA, ΔmOWLSlysozyme, increased from 200 to 245 ng cm2 from the first (lysozyme1) to the second (lysozyme2) adlayer, after which any increase was within the uncertainty of the measurements (part b of Figure 2). These ΔmOWLSlysozyme were larger than the calculated monolayer adsorbed mass of lysozyme of 185 ng cm2 assuming random sequential adsorption of lysozyme in an end-on orientation (ref 20 for calculations). Larger experimental than calculated ΔmOWLSlysozyme may have resulted from partial mobility of lysozyme molecules on the ESHA surface and, hence, a larger fractional surface coverage than 0.55 as stipulated by random 6005
dx.doi.org/10.1021/es200663h |Environ. Sci. Technol. 2011, 45, 6003–6010
Environmental Science & Technology
ARTICLE
Figure 2. Encapsulation of lysozyme by Elliott Soil humic acid (ESHA). a. Changes in dry adsorbed mass, ΔmOWLS, over time during encapsulation at pH 5 and ionic strength I = 10 mM as measured by optical waveguide lightmode spectroscopy (OWLS). Superscripts refer to the adlayer film number. b. Adsorption efficiencies (R) and changes in dry adsorbed masses (ΔmOWLS) for each adsorbate calculated from OWLS data at pH 5 and I = 10 mM (averages ( ranges of duplicate measurements). Open and closed symbols represent ESHA and lysozyme data, respectively. c. Changes in wet adsorbed mass, ΔmQCM-D, over time during encapsulation at pH 5 and 6 and I = 10 mM as measured by quartz crystal microbalance with dissipation monitoring (QCM-D). d. Wet mass of encapsulating ESHA, ΔmQCM-DESHA, as a function of the wet lysozyme mass, ΔmQCM-Dlysozyme, adsorbed to the underlying ESHA film at pH 5 and 6 and I = 10 mM (averages ( ranges of duplicate measurements). e. Changes in dry adsorbed mass, ΔmOWLS, over time during encapsulation at pH 7 and I = 10 mM as measured by OWLS. f. Changes in wet adsorbed masses, ΔmQCM-D, over time during encapsulation at pH 7 and 8 and I = 10 mM as measured by QCM-D. For panels a, c, e, and f: Surfaces were rinsed (i.e., rinse) with adsorbate-free solutions between exposure to adsorbate-containing solutions.
6006
dx.doi.org/10.1021/es200663h |Environ. Sci. Technol. 2011, 45, 6003–6010
Environmental Science & Technology sequential adsorption, which assumes immobile adsorbates. Furthermore, on the basis of previous atomic force microscopy work,15 the EHSA films likely had nanometer scale surface roughness, resulting in a larger total area for lysozyme adsorption than accounted for in the random sequential adsorption calculations, which assumes adsorption to perfectly smooth surfaces. An increase in surface roughness is consistent with the increase in the absolute adsorbed masses ΔmOWLSlysozyme and ΔmOWLSESHA from the first to subsequent encapsulation cycles (part b of Figure 2), whereas their mass ratios varied only slightly (i.e., 3.6 to 4.0) between the cycles. Encapsulation of lysozyme by ESHA at pH 5 and 6 was also measured by QCM-D, as shown in part c of Figure 2 and Figure S3 of the Supporting Information for the first two and for the total of four encapsulated lysozyme layers, respectively. Consistent with the OWLS experiments, QCM-D measurements showed irreversible adsorption of lysozyme to ESHA and of ESHA to lysozyme, reflecting attractive proteinHA interactions. The encapsulated wet mass of lysozyme, ΔmQCM-Dlysozyme, slightly increased with subsequent layering from 280 to 340 ng cm2 at pH 5, whereas ΔmQCM-Dlysozyme varied between 270 to 330 ng cm2 without a pattern with adlayer number at pH 6. The finding of ΔmQCM-Dlysozyme > ΔmOWLSlysozyme reflects that QCM-D measured the wet adsorbed mass, including adlayer associated water, whereas OWLS only sensed the dry adlayer mass (Figure 1).17 At pH 5, the mass contribution of water to the total sensed mass increased from about 50% for the (SiO2)PLLESHA1lysozyme1ESHA2 adlayer to about 65% for the final adlayer (SiO2)PLL(ESHAlysozyme)14ESHA5 (Figure S4 of the Supporting Information). These water contents are in the range reported for monolayer adsorption of globular proteins to solid surfaces.17,20,23,24 Parts ac of Figure 2 show that ESHA at pH 5 and 6 encapsulated complete monolayers of lysozyme. To test whether encapsulation also occurred at smaller coverages of lysozyme on the ESHA surface, submonolayer masses of lysozyme were adsorbed to ESHA by varying the adsorption time, followed by rinsing and subsequent exposure to ESHA-containing solutions. Part d of Figure 2 shows that the encapsulating wet mass of ESHA, ΔmQCM-DESHA, linearly increased with the underlying wet lysozyme mass (pH 5, R2 = 0.999; pH 6, R2 = 0.997) up to approximately 225 ng lysozyme cm2, above which ΔmQCM-DESHA did not further increase, indicating that lysozyme adsorption approached full monolayer coverage. These results demonstrate that ESHA also encapsulated single lysozyme molecules at submonolayer adsorbed masses on ESHA. OWLS measurements showed that, as expected, ESHA also encapsulated lysozyme at pH 7 (part e of Figure 2). The three encapsulated adlayers of lyzozyme showed similar ΔmOWLSlysozyme of 270 to 230 ng cm2 (part e of Figure 2 and Figure S5 of the Supporting Information). These masses were larger than the estimated monolayer adsorbed masses assuming end-on random sequential adsorption of lysozyme, likely for the same reasons previously discussed. Adsorption rates of lysozyme to ESHA were close to transport-limited (R ≈ 0.9, Figure S5 of the Supporting Information). However, in contrast to pH 5, adsorption rates of ESHA to lysozyme were smaller than to PLL (R = 0.50.6), adsorption was less extensive (ΔmOWLSESHA to lysozyme was approximately 40% of the value to PLL) and slightly reversible (the decrease in ΔmOWLSESHA during rinsing was too large to be caused by instrument drift). Complementary QCM-D measurements were collected at pH 7 and 8 (part f of Figure 2). These measurements revealed higher
ARTICLE
mass contributions of water (7075%) to the total adlayer sensed wet mass at pH 7 than at pH 5 (Figure S4 of the Supporting Information). Furthermore, the QCM-D measurements showed adlayer hydration dynamics during ESHA encapsulation that were not detected by OWLS. Whereas the formation of PLLESHA1lysozyme1 adlayers at pH 7 and pH 8 showed comparable characteristics as at lower pH, exposure of lysozyme1 to ESHA2 resulted in a fast initial increase in ΔmQCM-D up to a maximum, followed by a decrease while ESHA2 solutions were still run over the sensor (part f of Figure 2). These features became more pronounced with increasing adlayer number and with increasing pH. At pH 8, exposure of lysozyme1 to ESHA2 solutions did not result in an increase in ΔmQCM-D before rinsing. ESHA2, however, must have modified the lysozyme1 surface, and likely encapsulated lysozyme1, because subsequent exposure to lysozyme2 resulted in repeated protein adsorption. The comparison of OWLS and OCM-D measurements indicates that changes in adlayer water content played a major role in the encapsulation dynamics at pH 7 and 8. The decrease in ΔmQCM-D during exposure of adsorbed lysozyme1 to ESHA2 and of lysozyme2 to ESHA3 plausibly resulted from the combined loss of intermolecular water (i.e., volumetric replacement of water between lysozyme molecules by adsorbing ESHA molecules), and of intra-ESHA water, possibly due to a partial disassembly of ESHA associations in contact with lysozyme. An increase in the extent of ESHA disassembly with increasing pH may be rationalized by increasing destabilization of ESHA associations due to increasing intra-ESHA electrostatic repulsion upon successive deprotonation of its carboxylic and phenolic moieties (Table S1 of the Supporting Information). Lysozyme Encapsulation by Different HS. ESHA is but one member of a large array of terrestrial and mixed terrestrial-aquatic HS that differ in elemental compositions, charge densities, polarities, and sizes. Encapsulation was also demonstrated with a diverse set of HS, including both humic and fulvic acids (adsorption steps: HS1lysozyme1HS2lysozyme2HS3) from terrestrial and mixed aquatic-terrestrial sources (Table 1) at pH 5 and 6 by QCM-D (part a of Figure 3). Most of the ΔmQCM-Dlysozyme values were within the calculated range of lysozyme monolayer masses of 264 to 370 ng cm2 20 assuming random sequential adsorption in side-on and end-on orientations of lysozyme, respectively, and mass contributions of water to the total sensed monolayer mass of 50%.17,20,23 Part a of Figure 3 shows that, for a given adsorbed lysozyme mass, encapsulating masses of HA were, in general, larger than of FA, including HAFA pairs extracted from the same source materials (PP, ES, and SR). The encapsulating masses of HA decreased from pH 5 to 6, whereas pH had little effect on the encapsulating masses of FA. Also, terrestrial HA (ESHA, LHA, PPHA) had higher encapsulating masses than the tested mixed terrestrial-aquatic SRHA, whereas FA encapsulating mass seemed less dependent on FA source material. The estimated HA adlayer thicknesses varied between 1.73.2 nm for the HA and 1.01.2 nm for the FA (pH 5, Table S1 of the Supporting Information). These values are consistent with the hydrodynamic diameters of HA and FA measured elsewhere.12 The decrease in HA encapsulating mass with increasing pH, from terrestrial to mixed terrestrialaquatic HA, and from HA to FA was therefore consistent with decreasing sizes of the HS supramolecular associations. Trypsin and Ribonuclease A Encapsulation by HAs. So far we have demonstrated that various HS may encapsulate lysozyme molecules, which had uniform positive surface charges 6007
dx.doi.org/10.1021/es200663h |Environ. Sci. Technol. 2011, 45, 6003–6010
Environmental Science & Technology
ARTICLE
Figure 3. Encapsulation of lysozyme, trypsin, and ribonuclease A by various humic substances (HS) as measured with optical waveguide lightmode spectroscopy (OWLS) and quartz crystal microbalance with dissipation monitoring (QCM-D). a. QCM-D sensed wet mass of encapsulating HS, ΔmQCM-DHS, as a function of the wet adsorbed mass of lysozyme, ΔmQCM-Dlysozyme, preadsorbed to the same HS at pH 5 and 6 and ionic strength I = 10 mM (averages ( ranges of duplicate measurements) (HA = humic acid, FA = fulvic acid, ES = Elliot Soil, PP = Pahokee Peat, SR = Suwannee River, L = Leonardite). ΔmQCM-Dlysozyme and ΔmQCM-DHS values correspond to averages and ranges for each of the two adlayers (lysozyme1 and lysozyme2; HS1 and HS2). b. Changes in QCM-D sensed wet adsorbed mass, ΔmQCM-D, during trypsin encapsulation by ESHA and PPHA at pH 6 and I = 10 mM. c. Changes in OWLS-sensed dry adsorbed mass, ΔmOWLS, over time during ribonuclease A encapsulation by ESHA at pH 6 and I = 10 mM. Inset: Overlay of changes in ΔmOWLS of ribonuclease A during the first and second adsorption step (RNase1 and RNase2). *Changes in adsorbed mass during ESHA2 exposure were based on calculations with a refractive index increment of ESHA of 0.28 cm3 g1. The decrease in mass would have been more pronounced if a lower refractive index increment between those of ESHA and proteins (0.18 cm3 g1) was used. d. Changes in QCM-D sensed wet adsorbed mass, ΔmQCM-D, over time during ribonuclease A encapsulation by ESHA at pH 6 and I = 10 mM. Inset: Overlay of changes in ΔmQCM-D of ribonuclease A during first and second adsorption step (RNase1 and RNase2).
under the tested conditions. Part b of Figure 3 shows that ESHA and PPHA at pH 6 also encapsulated trypsin, which carried a small negatively charged patch on its surface at that pH (Table 1). As compared to lysozyme, the larger encapsulated mass of trypsin (ΔmQCM-Dtrypsin= 690 ( 10 and 690 ( 50 ng cm2 by ESHA and PPHA, respectively) reflected its larger molecular dimensions (Table 1) and, hence, adsorbed mass per occupied surface area on the HA. Successful encapsulation demonstrated attractive HAtrypsin interactions, likely due to both electrostatic attraction and contributions from the hydrophobic effect.9 These attractive interactions overcompensated localized electrostatic repulsion between HS and the negatively charged patch on the trypsin surface. In comparison to lysozyme and trypsin, ribonuclease A has a highly nonuniform surface charge distribution (Table 1), which on charged sorbent surfaces was shown to result in preferential
adsorbed orientations that result in protein-sorbent patch-controlled electrostatic attraction.25 Both OWLS and QCM-D showed adsorption of ribonuclease A to ESHA (ΔmOWLSribonuclease A = 125 ng cm2 and ΔmQCM-Dribonuclease A = 320 ng cm2, parts c and d of Figure 3), likely with the positively and negatively charged patches on the protein surface oriented toward ESHA and the bulk solution, respectively. OWLS measurements demonstrated that adsorption of ribonuclease A1 at pH 6 was more reversible (16% mass loss during buffer rinsing, part c of Figure 3) as compared to lysozyme1 (
