Research Note pubs.acs.org/IECR
A Method To Estimate the Average Shear Rate in a Bubble Column Using Liposomes Tomotaka Natsume and Makoto Yoshimoto* Department of Applied Molecular Bioscience, Yamaguchi University, 2-16-1 Tokiwadai, Ube, 755-8611, Japan ABSTRACT: We describe a novel approach to estimate average shear rate in an external-loop airlift bubble column, based on the shear stress-induced structural change in liposomes. As a shear-sensing probe, the 5(6)-carboxyfluorescein-containing liposome (CFL) was suspended at 40 °C in 50 mM Tris-HCl/50 mM NaCl buffer of pH 7.4 containing 0.3 wt % carboxymethyl cellulose. The CFL was 180 ± 1 nm in mean diameter and physically stable under shear flow. The permeability coefficient (PCF) of the dye molecules through the CFL membrane increased linearly as the shear rate (γ) increased up to 1425 s−1, as revealed using a cone-and-plate device. The PCF value was then determined for the identical CFL suspension in the airlift operated at superficial gas velocity (UG) values up to 0.03 m/s. The shear rate in the airlift could be reasonably estimated through correlating γ with UG, using their quantitative dependence on PCF, obtained as above. The γ value of 2.5 × 103 s−1 was obtained at UG = 0.03 m/s for the airlift.
1. INTRODUCTION Bubble columns are promising multiphase reactors for various chemical and biochemical reactions at different scales.1−4 Shear rate is one of the factors that dominate mass transfer in bubble columns through affecting hydrodynamic properties in the vicinity of the gas/liquid interface.5,6 Since shear rate in the columns cannot be characterized easily, several empirical equations were reported for calculating shear rate as a function of the superficial gas velocity (UG).7−9 However, it would be difficult to estimate shear rate with the UG value alone for any bubble columns with different configuration and physicochemical properties of liquid phase.10 In this context, mechanistic equations were reported to calculate the shear rate in bubble columns,5,11,12 providing the inter-relationship between shear rate and volumetric mass transfer rate in the columns.5,11 On the other hand, precise estimation of shear rate is of great importance for bubble column bioreactors to minimize structural changes in fragile cells.13−16 The effects of shear rate on the cell culture process have been extensively studied in the previous literatures.17 However, in many cases, the data obtained in a bioreactor are not applicable to different reactors, because their scale and configuration affect the shear rate as described above. Therefore, it is difficult to predict the response of living cells to the mechanical stresses in any bioreactors. Utilization of the function of cells to detect shear stress would be an interesting approach for developing a novel method for quantification of the stress. However, living cells themselves are not suitable for the purpose. This is because various factors, including physicochemical state of cells, growth kinetics, and growth conditions, can affect the sensitivity of cells to shear ́ stress.17 On the other hand, Ramirez and Mutharasan13 reported that the fluidity of plasma membranes of hybridomas affected their stability in shear flow. It is also reported that shear stress can cause permeabilization of cell membranes18 and induce structural change in lipid membranes for mechanotransduction.19 These observations show that, among the constituents of cells, lipid membranes can effectively perceive the shear stress. Therefore, if the structural or functional © 2013 American Chemical Society
response of artificial lipid membranes to the shear stress can be tuned precisely, quantification of average shear rate in a bioreactor with an unknown hydrodynamic property would be possible, based on the response of membranes. Liposomes can be prepared by dispersing phospholipids in an aqueous solution. The mean size, size distribution, and lipid composition of liposomes can be controlled. Response of lipid membranes to shear stresses such as deformability and rupture was extensively studied predominantly using giant vesicles with diameters larger than 500 nm.20−22 We recently reported that the membrane permeability of liposomes with diameters as small as ∼100−200 nm could be modulated by shearing in a cone-and-plate device.23,24 The susceptibility of liposome membranes in shear flow was dependent on the size of liposomes.23 Therefore, liposomes with appropriate size would be an interesting candidate of the probe for sensing the shear rate in a bioreactor on the basis of the possible relationship between the permeability of liposome membranes and the shear rate in the liposome suspension. In this work, liposomes encapsulated with fluorescence dye were prepared and the known shear rate in a cone-and-plate device was correlated with the unknown one in an external loop airlift bubble column by measuring the permeability coefficient of the dye molecules through liposome membranes suspended in both systems.
2. EXPERIMENTAL SECTION 2.1. Materials. 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) was purchased from NOF (Tokyo, Japan). 5(6)-Carboxyfluorescein (CF) was obtained from Sigma− Aldrich (St. Louis, MO). Carboxymethyl cellulose (CMC, Sunrose, type F01MC) was kindly supplied from Nippon Paper Received: Revised: Accepted: Published: 18498
September 2, 2013 November 14, 2013 December 2, 2013 December 2, 2013 dx.doi.org/10.1021/ie402874q | Ind. Eng. Chem. Res. 2013, 52, 18498−18502
Industrial & Engineering Chemistry Research
Research Note
Figure 1. Schematic representation of the 5(6)-carboxyfluorescein (CFL) prepared and chemical structure of 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC).
sparger, glass filter with nominal pore diameters of 40−100 μm was set at the bottom of riser. The 0.3 wt % CMC Tris buffer solution suspending CFLs ([lipid] = 1.0 mM) was charged into the airlift. The temperature of the airlift was maintained at 40 °C by bathing it in a water bath. Filtrated nitrogen gas was introduced through a humidifier at 40 °C at the superficial gas velocity UG, based on the cross-sectional area of the riser, at a rate of 0.01−0.03 m/s to induce liquid circulation in the airlift. 2.5. Measurements of Size Distribution of Liposomes. Size distribution of the CFL suspension was measured by the dynamic light scattering (DLS) instrument (ELSZ-2 plus, Otsuka Electronics, Osaka, Japan) equipped with a semiconductor laser at a wavelength of 660 nm at a fixed angle of 160°. The mean diameter (DP) of CFLs was determined based on the Einstein−Stokes relation with a refractive index of 1.33. Note that the effect of CMC was negligible, because the CFL sample was diluted 50 times before the measurement. All measurements were carried out in triplicate in a quartz cuvette at 25 ± 0.3 °C. 2.6. Determination of Permeability Coefficient PCF of CF through Liposome Membranes. The fractional amount of CF released from liposomes (RCF) was determined as RCF = (Ct − C0)/(C∞ − C0), where C0 and Ct stand for the concentrations of CF outside liposomes at the initial state and at any time t, respectively, and C∞ is the concentration of CF at the permeation equilibrium. It should be noted that the CF molecules in the bulk liquid do not need to be separated from CFLs, because the liposome-encapsulated CF is self-quenched. The C∞ value can be determined in the presence of 40 mM sodium cholate for the solubilization of liposome membranes. The concentration of CF was determined based on a known relationship between the concentration of CF and its fluorescence intensity measured with a spectrofluorometer (FP-750, JASCO, Tokyo, Japan).23 All measurements were performed at 25 ± 0.3 °C. The PCF value can be determined based on the unsteady-state mass balance with respect to CF, giving a relationship −ln(1 − RCF) = PCFaSt, where aS is the specific surface area of CFLs.25
2.3. Cone-and-Plate Shearing Device. A viscometer instrument (DV-II+Pro, Brookfield Engineering Laboratory, Middleboro, MA, USA) with the cone-and-plate geometry was applied to generate defined shear flow. The instrument was connected to a pesonal computer (PC) to operate with the software Rleocalc (Brookfield Engineering Laboratory). The cone employed had an angle of θ = 1.40 × 10−2 rad between the horizontal plate and the surface of inverted cone, and a radius of R = 2.4 cm. A CFL suspension (0.5 mL, [lipid] = 1.0 mM), which was prepared with the Tris buffer solution containing 0.3 wt % CMC, was charged onto a plate cup maintained at 40 °C with circulating water. The fractional volume of CFLs in the suspension was 6.8 × 10−3. The suspension was sheared between the rotating cone and the stationary plate at the angular velocity ω up to 20.9 rad/s. The shear rate γ (= ω/θ) was uniform at any point of the viscometer. As a static liquid system, a test tube was used. The viscosity of the above CMC solution free of liposomes was separately measured at various shear rates with the viscometer. 2.4. Formation of Gas−Liquid Flow in an Airlift Bubble Column. A mini-scale external loop airlift bubble column with a working volume of 7.0 mL was used as a gas− liquid contacting bioreactor. The airlift was 8.0 mm and 4.8 mm in the diameter of riser and downcomer, respectively. As a gas
3. RESULTS AND DISCUSSION 3.1. Stability of Liposomes under Shear Flow. The CFLs employed were almost monodisperse with a polydispersity index of 0.104 ± 0.003 and a mean diameter of 180 ± 1 nm. The physical stability of CFL is a crucial factor for its application in sensing shear rate in an airlift bubble column. Figure 2 shows the size distribution of CFLs measured after continuous shearing at the shear rate (γ) of 1425 s−1 or suspending in the airlift at the superficial gas velocity of UG = 0.03 m/s for 3 h at 40 °C. The size distribution of CFL measured without any mechanical treatment is also shown in the figure for comparison. The size of CFLs is almost unaffected by the above mechanical treatments. These results show that the CFLs are stable in the shear flow and even in the presence of a gas/liquid interface of bubbles. 3.2. Relationship between Shear Rate and Permeability of Liposome Membranes. The tris buffer solution containing 0.3 wt % CMC, which was used as the bulk liquid of CFL suspensions, exhibited Newtonian fluid characteristics, exhibiting a practically constant viscosity (η) of 1.45 ± 0.01 mPa s at γ = 262−1425 s−1. This is probably because of the low viscosity nature of the CMC used. Thus, the shear stress τ is calculated as τ = γη. The permeability of CFL membranes was examined at 40 °C in a cone-and-plate device at γ = 375, 750,
Industries Co., Ltd. (Tokyo, Japan). The degree of substitution of the CMC was 0.65−0.75. All reagents were used as received. Water used was sterilized and deionized with a water purification instrument Elix 3UV (Millipore Corp., Billerica, MA, USA). The minimum resistance to the water was 15 MΩ cm. 2.2. Preparation of Liposomes. The liposome encapsulated with 5(6)-carboxyfluorescein (denoted as CFL) was prepared by extruding multilamellar liposomes, which were formed by hydrating a dry POPC film, followed by the repetitive freeze−thaw treatments. The lipid hydration was performed with 50 mM Tris-HCl/50 mM NaCl buffer solution of pH 7.4 (denoted as Tris buffer) containing 100 mM CF. For the extrusion, a polycarbonate membrane with pore diameter of 200 nm was used. The nonencapsulated CF molecules could be separated from CFLs using gel permeation chromatography with a sepharose 4B column. Details of the preparation of CFLs are described in our previous paper.24 The CFL is schematically illustrated in Figure 1. The POPC in the CFL suspension was quantified with an enzyme kit from Wako Pure Chemicals (Osaka, Japan).
18499
dx.doi.org/10.1021/ie402874q | Ind. Eng. Chem. Res. 2013, 52, 18498−18502
Industrial & Engineering Chemistry Research
Research Note
Figure 2. Size distribution of CFLs before (1) and after (2 and 3) suspending the CFLs in shear flow. Shear stress was applied to the CFL suspension in the cone-and-plate device (2) or airlift (3) for 3 h at 40 °C at the lipid concentration of 1.0 mM. Measurements were performed in triplicate (1 and 2) or duplicate (3) for each condition and the representative data are shown. The mean diameter (DP) was 180 ± 1, 190 ± 1, and 189 ± 8 nm for conditions 1, 2 and 3, respectively. The polydispersity index (PI) was 0.104 ± 0.003, 0.125 ± 0.004, and 0.129 ± 0.064 for conditions 1, 2 and 3, respectively. Values represent mean ± standard deviation.
1125, and 1425 s−1. Figure 3A shows the effect of γ on the time courses of RCF. The release rate of CF clearly increases as the γ value increases. The CFL membrane is barely permeable in the static liquid system (γ = 0). Based on the linear relationship between −ln(1 − RCF) and t, the permeability coefficient PCF of CF through lipid membranes was determined. The curves in Figure 3A show the calculated time courses of RCF with RCF = 1 − exp(−PCFaSt). The PCF value clearly increases as the γ value increases (Figure 3B), showing PCF = 5.2 × 10−13 m/s at γ = 1425 s−1. An almost-linear relationship is found between the PCF and γ values with the fixed PCF value at γ = 0, showing PCF [m/s] = (3.1 × 10−16 m) × γ [s−1] + (0.8 × 10−13 m/s). Based on the fact that the physical stability of CFL is maintained under the shear flow (Figure 2), the release of CF molecules is caused not by the disruption of liposome membranes but by the permeation through the membranes. These results indicate that the liposome membrane can perceive the liquid shear stress with high sensitivity and transduce it into a change in the microstructure of the membrane. 3.3. Permeability of Liposome Membrane in the Airlift Bubble Column. We also examined the release behavior of CF from CFLs suspended in an external loop airlift bubble column at UG ≤ 0.03 m/s. The CFL suspended in the airlift was identical to that in the viscometer. As shown in Figure 4A, the liposome membranes become permeable upon increasing the UG value. The PCF value can also be reasonably determined based on the equation −ln(1 − RCF) = PCFaSt. The PCF value is clearly dependent on the UG value (Figure 4B), giving the apparent relationship PCF [m/s] = (2.6 × 10−11 [-]) × UG [m/s] + (0.8 × 10−13 m/s). The PCF value of 8.1 × 10−13 m/s, which is obtained at UG = 0.03 m/s, is ∼10 times larger than that obtained in the static liquid system. In the airlift, liposomes are potentially present in the bulk liquid, the liquid film around bubbles or gas/liquid interface.4 Since the CFLs are physically stable in the airlift (Figure 2), it is improbable that
Figure 3. (A) Time courses of RCF at various shear rates (γ) in a coneand-plate device. Curves are calculated with RCF = 1 − exp(−PCFaSt). (B) Relationship between PCF and γ. Data show the mean of three independent experiments; errors represent the standard deviation. When not indicated, the error is smaller than the symbol. Slope of the straight line in the lower panel is determined to be (3.1 ± 0.1) × 10−16 m.
the CFLs undergo significant structural change through the adsorption to the gas/liquid interface. Therefore, the average shear rate in the gas−liquid flow, which is suggested to increase with increasing UG, is responsible for the permeation of CF through the lipid membrane suspended in the airlift. We reported similar phenomena for a similar external loop airlift,25 although the working volume, the composition of liquid phase, the concentration of CFL and the UG range were different from these of the present work. 3.4. Estimation of Average Shear Rate in the Airlift Bubble Column. Based on the dependency of PCF on γ and that on UG determined as described above, the apparent relationship between γ and UG can be obtained as γ [s−1] = (8.4 × 104 m−1) × UG [m/s] (UG ≤ 0.03 m/s). At UG = 0.01 and 0.03 m/s, for example, the γ values can be estimated to be 8.4 × 102 and 2.5 × 103 s−1, respectively. The largest γ value at UG = 0.03 m/s in the airlift corresponds to the shear stress τ of 3.5 Pa. It should be noted that the above relationship is exclusively valid for the specific configuration, fluid combination, and operation condition employed for the present airlift bubble column. The shear rate is correlated with the liquid volume18500
dx.doi.org/10.1021/ie402874q | Ind. Eng. Chem. Res. 2013, 52, 18498−18502
Industrial & Engineering Chemistry Research
Research Note
al.9 This is probably because of the significant differences in the properties of liquid phase and configuration of the airlift, including the type of gas sparger. In principle, if the vesicular stability of CFLs is guaranteed, the shear rate in any bioreactors can be estimated by using CFLs for which the relationship between PCF and γ is tuned. The POPC membrane used in this work is suitable for estimating the shear rate in bubble columns at relatively low UG values (UG ≤ 0.03 m/s). Our method is demonstrated to be applicable to estimate change in the shear rate associated with the change in UG. The range of shear rate causing damage to cells and the enzyme molecules is different depending on their nature and operation conditions.17,26 Therefore, to estimate shear rate in the wide variety of bioreactors, the applicability of the liposome to non-Newtonian fluids and at higher shear rate must be clarified.
4. CONCLUSIONS We successfully estimated the shear rate of an external-loop airlift bubble column using the 5(6)-carboxyfluoresceincontaining liposome (CFL) as a novel shear-sensitive probe. The permeability coefficient (PCF) of the dye molecules through the lipid membrane could be correlated with the shear rate γ in a cone-and-plate device suspending the CFLs. The γ value in the airlift then was estimated, reasonably based on the PCF value at UG values up to 0.03 m/s. The CFL is potentially applicable to estimate the shear rate of other bioreactors with various scales and configurations. The CFL would be also applicable to non-Newtonian fluids if the vesicular structure of liposome is stable under the shear flow. The experimental determination of shear rate can contribute to comprehensive understanding of cell responses to shear stress and to rational regulation of cell culture processes to minimize unnecessary damage to cells. Furthermore, the use of liposomes would be a useful method alternative to the conventional theoretical approach for assessing the shear rate in various bioreactors precisely.
Figure 4. (A) Time courses of RCF at various superficial gas velocities UG in an external loop airlift bubble column. Curves are calculated with RCF = 1 − exp(−PCFaSt). Broken curve shows the data at UG = 0, which is identical to the data shown in Figure 3A at γ = 0. (B) Relationship between PCF and UG. Data are mean of two independent experiments. Errors represent standard deviation. When not indicated error is smaller than the symbol. Slope of the straight line in the lower panel is determined as (2.6 ± 0.1) × 10−11.
■
AUTHOR INFORMATION
Corresponding Author
*Tel.: 81 836 85 9271. Fax: 81 836 85 9201. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
based specific energy input which is dependent on the geometric constant of the airlift.10 Thus, the correlation between γ and U G should be different for different combinations of the cross-sectional area of the riser and the downcomer of the airlift. The shear rate is not uniform within the airlift. Since the colloidal liposomes are uniformly dispersed in the bulk liquid, our method can give an average shear rate in the specific bubble column employed. The shear rate estimated for the airlift is the same order of magnitude as that reported by Contreras et al.6 They reported the shear rate of an air− seawater system in a concentric tube airlift bubble column with a pore diameter of the gas sparger comparable to that of the present airlift. They also reported that the bottom region of the airlift exhibited the largest shear rate, indicating the importance of the structure of gas sparger in determining the average shear rate. On the other hand, the shear rate estimated in this work differs in several orders of magnitude, compared to that predicted by the equation γ [s−1] = (14800 s/m2) × UG2 [m2/ s2] − (351 m−1) × UG [m/s] + (3.26 s−1) reported by Shi et
ACKNOWLEDGMENTS This work was supported in part by the Grant-in-Aid for Scientific Research (C) (No. 25420833) from the Japan Society for the Promotion of Science (JSPS).
■
REFERENCES
(1) Merchuk, J. C.; Ben-Zvi (Yona), S.; Niranjan, K. Why use bubblecolumn bioreactors? Trends Biotechnol. 1994, 12, 501. (2) Nakao, K.; Harada, T.; Furumoto, K.; Kiefner, A.; Popovic, M. Mass transfer properties of bubble columns suspending immobilized glucose oxidase gel beads for gluconic acid production. Can. J. Chem. Eng. 1999, 77, 816. (3) Meng, A. X.; Hill, G. A.; Dalai, A. K. Hydrodynamic characteristics in an external loop airlift bioreactor containing a spinning sparger and a packed bed. Ind. Eng. Chem. Res. 2002, 41, 2124. (4) Yoshimoto, M.; Momodomi, C.; Fukuhara, H.; Fukunaga, K.; Nakao, K. Effect of suspended liposomes on hydrodynamic and
18501
dx.doi.org/10.1021/ie402874q | Ind. Eng. Chem. Res. 2013, 52, 18498−18502
Industrial & Engineering Chemistry Research
Research Note
oxygen transfer properties in a mini-scale external loop airlift bubble column. Chem. Eng. Technol. 2006, 29, 1107. (5) Merchuk, J. C.; Ben-Zvi (Yona), S. A novel approach to the correlation of mass transfer rates in bubble columns with nonNewtonian liquids. Chem. Eng. Sci. 1992, 47, 3517. (6) Contreras, A.; García, F.; Molina, E.; Merchuk, J. C. Influence of sparger on energy dissipation, shear rate, and mass transfer to sea water in a concentric-tube airlift bioreactor. Enzyme Microb. Technol. 1999, 25, 820. (7) Schumpe, A.; Deckwer, W.-D. Viscous media in tower bioreactors: Hydrodynamic characteristics and mass transfer properties. Bioprocess Eng. 1987, 2, 79. (8) Nishikawa, M.; Kato, H.; Hashimoto, K. Heat transfer in aerated tower filled with non-Newtonian liquid. Ind. Eng. Chem. Process. Des. Dev. 1977, 16, 133. (9) Shi, L. K.; Riba, J. P.; Angelino, H. Estimation of effective shear rate for aerated non-Newtonian liquids in airlift bioreactor. Chem. Eng. Commun. 1990, 89, 25. (10) Chisti, Y.; Moo-Young, M. On the calculation of shear rate and apparent viscosity in airlift and bubble column bioreactors. Biotechnol. Bioeng. 1989, 34, 1391. (11) Thomasi, S. S.; Cerri, M. O.; Badino, A. C. Average shear rate in three pneumatic bioreactors. Bioprocess Biosyst. Eng. 2010, 33, 979. (12) Grima, E. M.; Chisti, Y.; Moo-Young, M. Characterization of shear rates in airlift bioreactors for animal cell culture. J. Biotechnol. 1997, 54, 195. (13) Ramírez, O. T.; Mutharasan, R. The role of the plasma membrane fluidity on the shear sensitivity of hybridomas grown under hydrodynamic stress. Biotechnol. Bioeng. 1990, 36, 911. (14) Barbosa, M. J.; Albrecht, M.; Wijffels, R. H. Hydrodynamic stress and lethal events in sparged microalgae cultures. Biotechnol. Bioeng. 2003, 83, 112. (15) Al-Rubeai, M.; Emery, A. N.; Chalder, S.; Goldman, M. H. A flow cytometric study of hydrodynamic damage to mammalian cells. J. Biotechnol. 1993, 31, 161. (16) Sen, A.; Kallos, M. S.; Behie, L. A. Effects of hydrodynamics on cultures of mammalian neural stem cell aggregates in suspension bioreactors. Ind. Eng. Chem. Res. 2001, 40, 5350. (17) Chisti, Y. Shear sensitivity. In Encyclopedia of Industrial Biotechnology, Bioprocess, Bioseparation, and Cell Technology, Vol. 7; Flickinger, M. C., Ed.; Wiley: New York, 2010; pp 4360−4398. (18) Hallow, D. M.; Seeger, R. A.; Kamaev, P. P.; Prado, G. R.; LaPlaca, M. C.; Prausnitz, M. R. Shear-induced intracellular loading of cells with molecules by controlled microfluidics. Biotechnol. Bioeng. 2008, 99, 846. (19) Gudi, S.; Nolan, J. P.; Frangos, J. A. Modulation of GTPase activity of G proteins by fluid shear stress and phospholipid composition. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 2515. (20) Takagi, S.; Yamada, T.; Gong, X.; Matsumoto, Y. The deformation of a vesicle in a linear shear flow. J. Appl. Mech. 2009, 76, 021207-1−. (21) de Haas, K. H.; Blom, C.; van den Ende, D.; Duits, M. H. G.; Mellema, J. Deformation of giant lipid bilayer vesicles in shear flow. Phys. Rev. E 1997, 56, 7132. (22) Chakravarthy, S. R.; Giorgio, T. D. Shear stress-facilitated calcium ion transport across lipid bilayers. Biochim. Biophys. Acta 1992, 1112, 197. (23) Natsume, T.; Yoshimoto, M. Membrane permeability and stability of liposomes suspended in shear flow. J. Disp. Sci. Technol. 2013, 34, 1557. (24) Yoshimoto, M.; Tamura, R.; Natsume, T. Liposome clusters with shear stress-induced membrane permeability. Chem. Phys. Lipids 2013, 174, 8. (25) Yoshimoto, M.; Monden, M.; Jiang, Z.; Nakao, K. Permeabilization of phospholipid bilayer membranes induced by gasliquid flow in an airlift bubble column. Biotechnol. Prog. 2007, 23, 1321. (26) Bekard, I. B.; Asimakis, P.; Bertolini, J.; Dunstan, D. E. The effects of shear flow on protein structure and function. Biopolymers 2011, 95, 733. 18502
dx.doi.org/10.1021/ie402874q | Ind. Eng. Chem. Res. 2013, 52, 18498−18502
