Thermoresponsive Microtubule Hydrogel with High Hierarchical

Mar 23, 2011 - a hollow cylindrical structure with a diameter of 25 nm and a length ... systems were reported using the properties of MTs.5А10 Kinesi...
0 downloads 0 Views 3MB Size
COMMUNICATION pubs.acs.org/Biomac

Thermoresponsive Microtubule Hydrogel with High Hierarchical Structure Ken-Ichi Sano,† Ryuzo Kawamura,† Taiki Tominaga,† Hiromichi Nakagawa,† Naoko Oda,† Kuniharu Ijiro,†,‡ and Yoshihito Osada*,† † ‡

Molecular & System Life Science Unit, Advanced Science Institute, RIKEN, Saitama, Japan Molecular Device Laboratory, Research Institute for Electric Science, Hokkaido University, Sapporo, Hokkaido, Japan

bS Supporting Information ABSTRACT: A thermoresponsive 3D microtubule hydrogel (MT gel) was prepared by simultaneous polymerization and chemical cross-linking of tubulins. The main chain of this gel is composed of cross-linked MTs, which consists of a cylindrical assembly of tubulin covalently connected by polyethylene glycol. This gel, which contains 10 mg/mL of tubulin, exhibits a storage modulus G0 as high as 1  103, which is 10 times higher than the loss modulus G00 over a wide range of frequencies. The MT gel exhibits a reversible solgel transition by temperature changes at 437 °C via depolymerization and polymerization of the MT network. Notable effects of the presence of the crosslinkage on the process of polymerization and depolymerization of tubulin were experimentally observed, and the role of the crosslinkage was discussed.

’ INTRODUCTION Tubulin is a cytoskeleton protein (also known as R/β-tubulin heterodimer) that forms a cylindrical assembly called microtubules (MTs) by spontaneous polymerization.1 The polymerization of tubulin is accompanied by the hydrolysis of guanosine triphosphate (GTP) to guanosine diphosphate and is sensitive to temperature; an increase in temperature to 37 °C favors the polymerization.2,3 MT has a hollow cylindrical structure with a diameter of 25 nm and a length that runs up to several tens of μm and consists of a head-to-tail assembly of the tubulins.4 In eukaryotic cells, MTs that are aligned in highly ordered structure serve as cytoskeletons harnessing the rigidity, the polarity, and the compatibility with the motor proteins such as kinesins. Attempts to form networks or assemblies of MTs in artificial systems were reported using the properties of MTs.510 Kinesin is a motor protein that moves on the MT surface along polarity of the longitudinal axis of the MT powered by adenosine triphosphate hydrolysis. Using the motility of kinesins, unique structures of MT assemblies were formed in dynamic assembly systems.59 However, their scale is limited to the micrometer order. Recently, a welloriented MTs alignment on the millimeter scale was accomplished by polymerizing in a confined space under a temperature gradient; there polarity of MTs was controlled even without kinesins.10 However, the scale is still limited to the millimeter scale. Here we report a novel 3D chemically cross-linked MT gel that undergoes the thermoreversible solgel transition of tubulin polymerization and depolymerization. This MT gel is applicable for bulk whose scale is larger than millimeter. This technique may give clues to overcome the size limitation of the MT assemblies in artificial systems. r 2011 American Chemical Society

’ EXPERIMENTAL SECTION Preparation of Tubulin. Tubulin was purified from porcine brain by using a high-concentration piperazine-N,N0 -bis(2-ethanesulfonic acid) (PIPES) buffer [1 M PIPES, 20 mM ethylene glycol bis(β-aminoethyl ether)-N,N,N0 ,N0 -tetraacetic acid (EGTA), 10 mM MgCl2; pH adjusted to 6.8 using KOH].11 With this protocol, MT-associated proteins were removed as previously reported.11 The tubulin concentration was determined by the absorbance at 280 nm using an extinction coefficient of 115 000.11 In this paper, 10 mg/mL tubulin is also expressed as 100 μM tubulin, assuming the molecular weight of R/β-tubulin heterodimer is 100 kDa; in reality it is ca. 110 kDa, but usually this convenient conversion relationship is used. Preparation of MT Gel. The typical procedure to prepare MT gel is as follows. MTs were polymerized from tubulins (10 mg/mL) in polymerization buffer (80 mM PIPES, 1 mM EGTA, 6 mM MgCl2, 5 mM GTP; pH adjusted to 6.8 using KOH) at 37 °C for 30 min. bis-N-Hydroxysuccinimidylester polyethyleneglycol (bis-NHS PEG, MW 5000; SUNBRIGHT DE050GS, NOF-CORPORATION, Tokyo, Japan), a cross-linker (CL), was added to the tubulin solution at the molar ratio of tubulin/CL 2:1 that corresponds to an equimolar ratio of the NHS group to the tubulins (NHS/ Tub = 1). The cross-linking reaction was processed by incubation at 37 °C for 2 h to obtain an intact MT gel. The intact MT gel was depolymerized once to sol by chilling on ice for 15 min. Finally, the gels for subsequent measurements were obtained by incubating the sol again at 37 °C for 30 min. Gelation was confirmed by the tilt-tube method. The samples were Received: December 28, 2010 Revised: March 17, 2011 Published: March 23, 2011 1409

dx.doi.org/10.1021/bm101578x | Biomacromolecules 2011, 12, 1409–1413

Biomacromolecules

COMMUNICATION

Figure 1. Creation of the MT gel. (A) Schematic illustration of the process. (B) Fluorescence microscopic images of the MT gel network. (B, left) The MT gel was formed in a chamber. The MTs were visualized by adding paclitaxel Oregon Green 488 conjugate after gelation; paclitaxel is an MT-stabilizing agent. The image was obtained under a total internal reflection fluorescence microscope (TIRFM). (scale bar: 10 μm) (B, right) The fragmented MT gel network is shown. The MT gel containing Alexa 488 (green)-labeled tubulin was added to a 50-fold volume of buffer containing 10 μM paclitaxel, and the network was fragmented by gentle pipetting. The fragmented MT-gel suspension was placed between a pair of coverglass slips without a spacer. The image was obtained under an epifluorescence microscope (scale bar: 10 μm).

Figure 3. Investigation of gelation conditions. (A,B) Tubulin concentrations and reacting ratio of tubulin/CL were varied in (A) and (B), respectively, using manual tilt-tube methods. (C) Characterization of PEG-cross-linked tubulin by SDS-PAGE. Szasz et al. reported that methylation of lysine in R-tubulin causes effective interference of tubulin polymerization.14,18 (D) Amount of proteins at 118 and 129 kDa was estimated by densitometry of the SDS-PAGE image. The relative densities of 118 and 129 kDa bands are denoted on the y axis. These bands were within the linear range of the densitometric curve, although the bands of R- and β-tubulins (monomers) were saturated. Figure 2. Characterization of PEG-cross-linked tubulin by SDS-PAGE and Western blotting. incubated in the vials for the solgel transitions. Figure 3A,B show the vials inverted for 5 min at 37 °C. Fluorescence Microscopy. To observe the network structures of the MTs in MT gel, the MTs were labeled with fluorescent probes using two different methods. In the first case, the MT gels were polymerized in a chamber consisting of a pair of coverglass slips with dimensions of 4  18  ∼0.1 mm3 (W  L  H; approximate volume, 10 μL) and directly observed after the addition of 0.5 μL of paclitaxel Oregon Green 488 conjugate (P22310, Invitrogen, Carlsbad, CA) dissolved in 20% dimethylsulfoxide to the chamber at a final concentration of 10 μM. Because the concentration of the MTs was as high as 10 mg/mL, the network structure was observed by total internal reflection fluorescence microscopy

(TIRFM). In another case, an aliquot of tubulins was fluorescently labeled with Alexa Fluor 488 carboxylic acid succinimidyl ester (A-20000, Invitrogen) prior to MT gel preparation according to a standard technique.12,13 Using the Alexa 488-labeled tubulin (A488-Tub) at a labeling ratio to tubulin of 0.7 (mol/mol), we set the final labeling ratio to tubulins in MT gel as 0.1 (mol/mol) during the reaction with bis-NHS PEG for MT gel preparation. The MT gel containing A488-Tub was then diluted with a 50-fold volume of stabilizing buffer (80 mM PIPES, 1 mM EGTA, 1 mM MgCl2, 10 μM paclitaxel (T1912, Sigma-Aldrich, St. Louis, MO); pH adjusted to 6.8 using KOH). Finally, 1.5 μL of the sample was placed between a pair of coverglass slips after gentle pipetting and observed by fluorescence microscopy. Confirmation of Cross-Linking Reaction. To confirm the crosslinking reaction of bis-NHS PEG and tubulins, we depolymerized MT gels on ice and analyzed them by SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting. The R- and β-tubulins were detected by R-tubulin1410

dx.doi.org/10.1021/bm101578x |Biomacromolecules 2011, 12, 1409–1413

Biomacromolecules

COMMUNICATION

Figure 4. Rheometric analysis of MT gel. (A) Analysis by strain sweep. (B) Analysis by frequency sweep. G0 and G00 of the cross-linked MT gel, MT containing 0.025 wt % PEG (MTþPEG), and tubulin containing 4% PEG-tubulin (CL-Tub) as a function of the frequency. The measurement was performed at 37 °C for MT gels and MTþPEG and at 4 °C for CL-Tub. specific (2144S, Cell Signaling Technology, Danvers, MA) and β-tubulinspecific antibodies (2146S, Cell Signaling Technology). Figure 2 shows the Western blot of the MT gels with a CL ratio (NHS/Tub) of 1. Several bands can be observed near the 115130 kDa range, which corresponds to the dimerized tubulins cross-linked with several combinations of R- and βtubulins because the molecular weight of both R- and β-tubulins is ca. 55 kDa. In the result of Western blotting, the lowest band of the dimerized tubulins at 118 kDa could be a ββ tubulin dimer differing from the larger bands of the dimerized tubulins containing R-tubulin; the molecular weights of the proteins in the Coomassie Brilliant Blue (CBB)-stained gel were calculated by comparison with protein ladders (Page ruler, SM0661, Fermentas; prestained protein markers, 02525-35, Nacalai Tesque, Kyoto, Japan) that were run on the same gel. Next, we analyzed the optical densities of these dimerized tubulin bands by SDS-PAGE, varying the ratio of reacting CLs to tubulins (Figure 3C,D). The band density of the fraction at 118 kDa is higher than that of the upper band at 129 kDa, whereas the MT gel retains the form in the inverted tubes, that is, NHS/Tub range is 0.24 (Figure 3B). However, the fraction of the 129 kDa band exceeds that of the 118 kDa band at the higher CL ratio (NHS/Tub g10), and the MT gels starts to flow. From these results, we can assume that the ββ tubulin dimer (118 kDa) is critical to the formation of a stable MT gel and excess modification with bisNHS PEG on R-tubulins appears to inhibit the polymerization of MTs and prevent the formation of stable gels. This is consistent with previous studies reporting that modification of the highly reactive lysine at position 394 of Rtubulin inhibits MT polymerization.14 Therefore, a CL ratio (NHS/Tub) of 1 could be an optimized condition for the formation of MT gels. Rheometric Analysis of MT Gels. The bulk mechanical response of the MT gel was measured with a stress-controlled rheometer (AR-G2, TA Instruments, New Castle, DE). Parallel plates (titanium) 40 mm in diameter under 1% strain and 60 mm in diameter under 10% strain were used for the measurement of the MT gel and suspension and tubulin (control), respectively. The plates were placed in a gap of 400 μm, and the temperatures were maintained between 0 and 37 °C. For the strain sweep experiment, the shear modulus was measured at 1 Hz in a rheometer by oscillatory measurements at increasing strain amplitudes ranging from 0.01 to 200%. For the frequency sweep experiment of the MT gel, we monitored from 0.01 to 10 Hz at 1.0% strain; this frequency range is suitable for the analysis of protein materials, as previously reported.1517 Each sample was subjected to oscillatory stress once, and after that, the second measurements were conducted. To measure the mechanical response during the solgel transitions, the storage modulus G0 and loss modulus G00 were measured under 2% strain at a single frequency (1 Hz) by increasing and decreasing the temperature between 0 and 37 °C. In Figure 5A,B, log G0 was plotted using the original values. In Figure 5C, log G0 was normalized.

’ RESULTS AND DISCUSSION When a tubulin solution is warmed from 4 to 37 °C to form MT and an appropriate amount of bis-NHS PEG, which is a CL that covalently binds two tubulin molecules at both ends (PEG-tubulin), is added, the solution quickly loses its fluidity and forms an insoluble gel (Figure 1A, ac). If the gel is cooled to 4 °C, then it transitions to the sol state, and a clear solution is formed (Figure 1A, cd). The observed solgel transitions could be repeated many times by cyclic temperature changes. The tubulin solution that contains no PEGtubulin does not form any gel by the same procedure (Figure 3B). Here it should be emphasized that bis-NHS PEG should be reacted after polymerization of tubulin is completed, and if bis-NHS PEG is allowed to react with monomeric tubulin, then polymerization or gelation occurs under much narrower conditions, presumably because of the blocking of the polymerizing site at Lys394 in tubulin by bis-NHS PEG. (See the Supporting Information.) Szasz et al. reported that methylation of lysine in R-tubulin causes effective interference of tubulin polymerization.14,18 It should be noted that a stable and rigid gel is formed when the gel as prepared by the first temperature increase is depolymerized and then once again polymerized to complete the gelation process (Figure 1A, de). The network structures of the MT gel are shown in Figure 1B; the fluorescently labeled MTs were observed using a fluorescence microscope. Some filamentous figures were obviously brighter than the others, implying that the bundling of the MTs occurred. Alignment of MTs was not well-observed in TIRFM images showing the surface of the MT gel (Figure 1B, left), although it can be assumed that the MTs are aligned inside the gel because of the high concentration of tubulins as 10 mg/mL.19 The formation of the PEG-cross-linked MT was also confirmed by SDS-PAGE because the product stained with CBB exhibited bands of retarded electrophoretic displacement (Figure 2). To investigate the concentration dependences for the gelation, we performed tilt-tube method experiments by varying the tubulin and CL (PEG-tubulin) concentrations (Figure 3A,B). It was found that the gel formed when the tubulin concentration exceeded 100 μM over a wide range of CL to tubulin ratios (NHS/Tub) of 0.24. However, if the NHS/Tub ratio is >4, then gelation does not occur, presumably because of the blocking of Lys394, as described in the Experimental Section. The mechanical properties of the obtained MT gel were measured using oscillating rheometry.15 Figure 4A shows the strain sweep of the MT gel cross-linked by PEG. Because the values of G0 and G00 were almost constant near 1% strain, the 1411

dx.doi.org/10.1021/bm101578x |Biomacromolecules 2011, 12, 1409–1413

Biomacromolecules

COMMUNICATION

Figure 5. Temperature profile of G0 of (A) cross-linked MT gel (NHS/Tub = 1) and (B) uncross-linked MT (NHS/Tub = 0). The temperature sweep rate was 1 or 5 min/°C. The y axis is shown in original scale. (C) Variation in the CL ratio. The y axis is the normalized scale. The temperature sweep rate was at 1 min/°C.

rheometric analysis was performed at 1% strain. At 27% strain, there is a crossing point of G0 and G00 , suggesting that the network structures of MT gels are broken by strain exceeding 27%. As shown in Figure 4B, MT gels cross-linked by PEG possessed a G0 of 1800 Pa at 1 Hz, and the values were always higher than those of G00 in a range of frequencies from 0.01 to 10 Hz, indicating that the stable gel was certainly formed. This is in strong contrast with the G0 and G00 values of sol state 0.1 to 1 Pa or those of the MT solution containing 0.025 wt % PEG (without cross-linking) as low as 110 Pa; the sol state of the MT gel prepared at NHS/Tub of 1 corresponds to tubulin and ca. 4% PEG-tubulin mixture. To clarify the effect of CL (PEG-tubulin) on the polymerization and depolymerization process of tubulin, the solgel transition of MT gel was studied precisely by measuring the change in G0 as temperature is varied (Figure 5). We chose fast (1 min/°C) and slow (5 min/°C) temperature sweep programs to evaluate the process (Figure 5A). The rheological measurement was automatically performed at every 1 °C exactly 1 min after the attainment of each temperature. Next, 0.5 mL of 100 μM tubulin solution containing 4% PEGtubulin in a Peltier device was added to a rheometer and maintained at 4 °C for 30 min; the ratio of PEG-tubulin was calculated by densitometric analysis. The G0 of the solution exhibited an order of 1  101 Pa. When temperature was increased, the G0 of tubulin solution containing CL did not exhibit a notable increase to 25 °C. However, at 26 °C, it abruptly jumped to 1  103 Pa, apparently

Figure 6. G0 and G00 of the MT gel determined by cyclic temperature changes.

because of the solgel transition. When temperature was decreased, the G0 value decreased rather gradually, exhibiting a prominent hysteresis, and the initial G0 value was recovered at 18 °C. Therefore, the MT gel varies its G0 values by as much as three-fold when the temperature was increased from 18 to 28 °C. The G0 profile of tubulin solution containing no CL (un-crosslinked MT) is significantly different from that of the gel. The initial G0 and temperature profile of this tubulin solution are similar to those of the solution containing the CL up to 25 °C. An increase in G0 occurs at 28 °C because of polymerization, but the magnitude of the increase is ca. 10-fold, which is much lower than 1412

dx.doi.org/10.1021/bm101578x |Biomacromolecules 2011, 12, 1409–1413

Biomacromolecules that of cross-linked MT (Figure 5A,B); the G0 value of MT gel at NHS/Tub = 0.2 was in the same range as the MT gel at NHS/ Tub = 1 (data not shown), indicating that the cross-link is critical to attain high G0 value. Hysteresis was also observed in the study of un-cross-linked MT when the temperature was decreased; this recovered the initial value at 18 °C. Therefore, the variation in the G0 value during the temperature sweep is strongly enhanced by the presence of cross-link. This can be attributed to the rigidity of MT because the PEG is covalently bond to MTs, and the elasticity of the PEG is negligible because of the short length of the cross-linker as ca. 35 nm that is slightly greater than the MT diameter (25 nm). Figure 5C shows the temperature sweep analysis with normalized log G0 values, varying the CL ratio (NHS/Tub). The half-log G0 temperature was 26 °C for the MT gel and 30 °C for un-cross-linked MT. During the MT polymerization process with temperature increases, the half-log G0 temperature decreases as the CL ratio (NHS/Tub) increases. However, the half-log G0 temperature of the depolymerization process decreases as the CL ratio decreases. In other words, hysteresis of the gelsol transition becomes less pronounced in the presence of the CL. Formation of the cross-linked network of MTs may enhance the polymerization because of the decreased thermal fluctuation. However, the decrease in the half-log G0 temperature with decreasing amounts of CL during the depolymerization process cannot be simply explained by the thermal fluctuations. Detailed analysis of the solgel transition profiles will be performed in a future investigation. As previously described, the solgel transition can be repeated many times by repeated incubation as long as the concentration of GTP remains higher than the critical value.20 Figure 6 shows the cyclic changes in G0 and G00 when temperature was alternatively changed between 4 and 37 °C. Note that the G0 value of the crosslinked MT gel changes by 1000-fold because of the repeated solgel transition. Different from the typical synthetic polymer gels where the main chain and cross-links are formed by covalent bonding, the main chain of the MT gel is the product of the self-organized assembly of the constituent globular protein tubulin; consequently, polymerization and depolymerization are reversibly induced by temperature changes.2125 MT has a sterically well-defined geometrical structure composed of globular tubulin with a tertiary-ordered structure, and the PEG-cross-linked MT gel could be categorized as a highly hierarchical supramacromolecular gel from this viewpoint. This system would be also further expanded by introducing cross-linkers that bind noncovalently to the main chains.26

’ ASSOCIATED CONTENT

bS

Supporting Information. Information regarding investigation of cross-linking reaction without MT polymerization. This material is available free of charge via the Internet at http://pubs. acs.org.

COMMUNICATION

Corporation. The authors acknowledge their consistent encouragement and support. This research is also supported by a Grant-in-Aid for Young Scientists (A) from Japan Society for Promotion Science to K-I.S. (20681013).

’ REFERENCES (1) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4th ed.; Garland Science: New York, 2002. (2) Fygenson, D. K.; Braun, E.; Libchaber, A. Phys. Rev. E 1994, 50, 1579. (3) Howard, J. Mechanics of Motor Proteins and the Cytoskeleton; Sinauer Associates: Sunderland, MA, 2001. (4) Amos, L. A. Org. Biomol. Chem. 2004, 2, 2153. (5) Kawamura, R.; Kakugo, A.; Shikinaka, K.; Osada, Y.; Gong, J. P. Biomacromolecules 2008, 9, 2277. (6) Liu, H.; Spoerke, E. D.; Bachand, M.; Koch, S. J.; Bunker, B. C.; Bachand, G. D. Adv. Mater. 2008, 20, 4476. (7) Kawamura, R.; Kakugo, A.; Osada, Y.; Gong, J. P. Langmuir. 2009, 26, 533. (8) Kawamura, R.; Kakugo, A.; Shikinaka, K.; Osada, Y.; Gong, J. P. Nanotechnology 2010, 21, 145603. (9) Nedelec, F. J.; Surrey, T.; Maggs, A. C.; Leibler, S. Nature 1997, 389, 305. (10) Kakugo, A.; Tamura, Y.; Shikinaka, K.; Yoshida, M.; Kawamura, R.; Furukawa, H.; Osada, Y.; Gong, J. P. J. Am. Chem. Soc. 2009, 131, 18089. (11) Castoldi, M.; Popov, A. V. Protein Expression Purif. 2003, 32, 83. (12) Hyman, A.; Drechsel, D.; Kellogg, D.; Salser, S.; Sawin, K.; Steffen, P.; Wordeman, L.; Mitchison, T. Methods Enzymol. 1991, 196, 478–485. (13) Peloquin, J.; Komarova, Y.; Borisy, G. Nat. Methods 2005, 2, 299–303. (14) Szasz, J.; Yaffe, M. B.; Elzinga, M.; Blank, G. S.; Sternlicht, H. Biochemistry 1986, 25, 4572. (15) Janmey, P. A.; Euteneuer, U.; Traub, P.; Schliwa, M. J. Cell Biol. 1991, 113, 155. (16) Lin, Y.-C.; Koenderink, G. H.; MacKintosh, F. C.; Weitz, D. A. Macromolecules 2007, 40, 7714. (17) Lieleg, O.; Classens, M. M. A. E.; Bausch, A. R. Soft Matter 2010, 6, 218. (18) Szasz, J.; Burns, R.; Sternlicht, H. J. Biol. Chem. 1982, 257, 3697. (19) Hitt, A. L.; Cross, A. R.; Williams, R. C., Jr. J. Biol. Chem. 1990, 265, 1639. (20) Symmons, M. F.; Martin, S. R.; Bayley, P. M. J. Cell Sci. 1996, 109, 2755. (21) Rees, D. A. Adv. Carbohydr. Chem. Biochem. 1969, 24, 267. (22) Petka, W. A.; Harden, J. L.; McGrath, K. P.; Wirtz, D.; Tirell, D. A. Science 1998, 281, 389. (23) Yoshida, R.; Uchida, K.; Kaneko, Y.; Sakai, K.; Kikuchi, A.; Sakurai, Y.; Okano, T. Nature 1995, 374, 240. (24) Jeong, B.; Kim, S. W.; Bae, Y. H. Adv. Drug Delivery Rev. 2002, 54, 37. (25) Garbern, J. C.; Hoffman, A. S.; Stayton, P. S. Biomacromolecules 2010, 11, 1833. (26) DiDonna, B. A.; Levine, A. J. Phys. Rev. Lett. 2006, 97, 068104.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We thank J. Heddle for the critical reading of this manuscript. This work was financially supported by TOYOTA Motor 1413

dx.doi.org/10.1021/bm101578x |Biomacromolecules 2011, 12, 1409–1413