Small Globular Protein Motif Forms Particulate Hydrogel under

Mar 17, 2011 - In the current study, we reported a new naturally occurring protein motif that formed a transparent hydrogel when heated to 90 °C at a...
0 downloads 0 Views 5MB Size
ARTICLE pubs.acs.org/Biomac

Small Globular Protein Motif Forms Particulate Hydrogel under Various pH Conditions Jun Fang, Xiaoning Zhang, Yuguang Cai, and Yinan Wei* Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506, United States ABSTRACT: Biocompatible hydrogels have great potentials in biomedical and biotechnological applications. In the current study, we reported a new naturally occurring protein motif that formed a transparent hydrogel when heated to 90 °C at a concentration as low as 0.4 mg/mL. The protein motif is the C-terminal soluble domain of an Escherichia coli inner membrane protein YajC (YajC-CT). We investigated the conformational change and self-assembly of the protein that lead to the formation of hydrogels using multiple methods. Atomic force microscopy studies of dilute gel samples revealed the presence of βsheet-rich fibrils that were 2 to 3 nm in height and micrometers in length, which appeared to originate from homogeneous particles. On the basis of these observations, we proposed a three-step pathway of YajC-CT gelation. Hydrogels formed at different pH contained slightly different fibril structures. To our knowledge, this is the smallest hydrogel-forming globular protein module that has been characterized in detail. It may be useful as a model system in the elucidation of the mechanisms of protein fibrillation and gelation processes.

’ INTRODUCTION Biocompatible hydrogels have drawn much attention in the past couple of decades because of their potentials in biomedical and biotechnological applications.111 Naturally occurring proteins and their fragments have been used extensively as building blocks in the design and construction of functional hydrogels, to provide structural support, to serve as the nucleation sites for polymerization, or to introduce stimuli-responsive features into the hydrogel system. Purely proteinatous hydrogel, which formed by naturally occurring or artificially designed polypeptides without any additional chemical modifications, have the advantage of being nontoxic and biodegradable. Naturally occurring proteins that form hydrogels under specific conditions include fibroses proteins such as collagen, cellulose, fibrin, and silk-like protein fibroin and globular proteins such as lysozyme, bovine serum albumin (BSA), β-lactoglobulin, and ovalbumin.1217 Despite extensive studies of protein-based hydrogels and their applications, the mechanisms by which soluble proteins form fibrils and hydrogels are not yet well understood. Several globular proteins have been reported to form hydrogels, including lysozyme (molecular weight 14.3 kDa), BSA (66 kDa), β-lactoglobulin (18.4 kDa), and ovalbumin (44.3 kDa).12 In this study, we discovered a new protein motif from Escherichia coli that formed hydrogels when heated to 90 °C (Figure 1). This protein motif, YajC-CT, is the C-terminal soluble domain of inner membrane protein YajC. It is the smallest globular protein module that has been reported to form hydrogels, with a molecular weight of 6 kDa. The N-terminal structure of YajC constitutes of a single transmembrane helix, the structure of which has been solved in a cocrystal structure with protein AcrB, whereas the structure of the C-terminal domain used in the current study remains elusive.18 r 2011 American Chemical Society

We characterized the structure of YajC-CT in both its soluble and hydrogel states. The critical gel formation concentration of YajCCT at 0.4 mg/mL was among the lowest reported for protein hydrogels. We monitored the secondary structure transition of YajC-CT from the soluble to the hydrogel state and examined the effect of different solution conditions including various pH and ionic strength on protein gelation.

’ MATERIALS AND METHODS Construction of Protein Expression Vector. A plasmid (pYajC-CT) was constructed to express YajC-CT, which contains the fragment of YajC starting from Met53 (Figure 1B). The gene encoding the fragment was obtained through polymerase chain reaction (PCR) using E. coli genomic DNA as the template and primers YajC-CT-F: 50 -AGT CCC ATG GAC TCC ATT GCC AAA GGT GAT GA-30 and YajC-R-His: 50 -GTA CGA ATT CTC AGT GGT GGT GGT GGT GGT GCA GCG CCT TCA TGG TGC CTT TCG GCA. The PCR product was cloned into vector pET28a (EMD Chemicals, Gibbstown, NJ) using routine molecular biology techniques. A polyhistidine tag was introduced at the C-terminus of the protein to facilitate convenient purification (Figure 1B). All sequences were verified through DNA sequencing. Protein Expression and Purification. Plasmid pYajC-CT was transformed into ER2566 cells (New England Biolabs, Ipswich, MA) for protein expression. In brief, a single colony was used to inoculate an autoinducing medium and cultured at 37 °C with shaking overnight.19 Cells were then collected by centrifugation and resuspended in 100 mL Received: December 27, 2010 Revised: March 1, 2011 Published: March 17, 2011 1578

dx.doi.org/10.1021/bm101571r | Biomacromolecules 2011, 12, 1578–1584

Biomacromolecules

Figure 1. (A) Photos of vials containing 1.0 mg/mL YajC-CT after the heat treatment. A transparent self-supporting gel formed. (B) Sequence of YajC-CT. The predicted secondary structure components were indicated on top of the sequence, with arrows and bar indicated β-strand and R-helix, respectively. of binding buffer (500 mM NaCl, 50 mM phosphate buffer, pH 7.5) containing 1 mM phenylmethanesulphonyl fluoride (PMSF). Cells were lysed by sonication on ice. After centrifugation, the supernatant was collected. Imidazole was added to the supernatant at a final concentration of 10 mM. The supernatant was then loaded to a Ni-nitrilotriacetic acid superflow (Ni-NTA) column (Qiagen, Valencia, CA). The column was washed with 20 bed volumes of the binding buffer containing 50 mM imidazole. YajC-CT was eluted using the binding buffer supplemented with 500 mM imidazole. Purified protein was analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE). The secondary structure of YajC-CT was predicted using the online server of Jpred3.20

Circular Dichroism (CD) and Tyrosine Fluorescence Spectroscopy. CD experiment was performed using a JASCO J-810 spectrometer with a 0.1 cm path length quartz cuvette. Purified YajC-CT was changed into indicated buffers through ultrafiltration using an AmiconUltra-4 column (Millipore, Billerica, MA). A protein concentration of 0.12 mg/mL was used for all CD studies. Fluorescence emission spectra were collected using a PerkinElmer LS-55 fluorescence spectrometer (PerkinElmer, Waltham, MA) at room temperature. The buffer condition and protein concentration were the same as in the CD experiment. The tyrosine fluorescence of YajC-CT was monitored with an excitation wavelength of 270 nm.

Effect of Protein Concentration, pH, and Ionic Strength on the Gelation Process. A series of protein solutions with varying pH (pH 4.0, 7.0, or 10.0 in 50 mM Na-phosphate and 500 mM NaCl), NaCl concentration (0, 0.1, 0.5, or 1.25 M NaCl in 50 mM Na-phosphate, pH 7.5), or protein concentration (0, 0.1, 0.2, 0.33, 0.4, 0.5, 0.66, 1, or 2 mg/mL YajC-CT in 50 mM Na-phosphate, 500 mM NaCl, pH 7.5) were heated at 90 °C for 5 min and cooled to room temperature to form hydrogels. A YajC-CT concentration of 1 mg/mL was used in the pH and ionic strength studies. Formation of the transparent self-supporting hydrogel was observed directly. In addition, a dilute solution containing a red food color was applied to the top of the gel. The presence of a clear interface between the colorless hydrogel and the red solution indicated protein gelation. Experiments were repeated at least three times under each condition. Fourier Transform Infrared Spectroscopy. Protein solution (1 mg/mL, 5 μL) was applied on an octadecyltrichlorosilane (OTS)coated 1  1 cm2 Silicon (100) wafer and slowly dried in an ultrapure nitrogen environment. The reflective IR spectra of the protein spot were acquired using the Varian UMA 600 IR microscope with 4 cm1 resolution under a constant flow of 60 mL/s ultrapure nitrogen. The IR spectra of YajC-CT hydrogel were obtained with the same setup with the exception that a small piece of hydrogel was placed on the OTScoated silicon wafer to dry before the analysis. Rheological Studies. Rheological properties of YajC-CT hydrogels were measured using cone and plate programmable viscometer

ARTICLE

(Brookfield Engineering Laboratories, model DV-II) with spindle 52. The measurements were taken over the speed settings of shear rate ranging from 4 to 120 s1. All measurements were made in triplicate. Atomic Force Microscopy (AFM). We heated 50 μL of YajC-CT solution (1 mg/mL) to form hydrogel. Next, 500 μL of deionized water was added. The hydrogel was broken by repeated pipetting and applied to a newly cleaved clean mica surface. After 20 min of incubation, the solution on the mica surface was blown away using a stream of ultrapure nitrogen. The mica surface was then characterized using the Vecco Multimode AFM in tapping mode with a MikroMasch CSC14 tip (140 kHz resonance frequency). All images were processed using the WSXM software (Nanotec Electronica, Madrid, Spain).21 Parameters Affecting Fluorescein Release. Fluorescein was used as a tracer to study the release of small molecule drugs from the hydrogel under different conditions. Fluorescein (0.5 μL, 5 mM) was mixed with 20 μL of purified YajC-CT under the specified conditions; then, the mixture was heated to 90 °C for 5 min. Next, 200 μL of the corresponding buffer was added on the top of the gel. The tube was incubated overnight at room temperature. A small aliquot of the supernatant was then removed and diluted in the binding buffer. Fluorescence emission spectrum of the dilute sample was measured with the excitation wavelength of 495 nm. To access the percentage of drug release from the hydrogel, we performed the same procedure in the absence of protein as a control. The fluorescence intensity in the control sample was designated as “100% release”. The ratio of the fluorescence intensity at 520 nm between the supernatant in the sample containing hydrogel and the control sample was used to determine the percentage of fluorescein release. A protein concentration of 1 mg/mL and buffer condition of 500 mM NaCl, 50 mM phosphate buffer, pH 7.0 were used for all fluorescein release studies unless otherwise noted.

’ RESULTS YajC-CT Formed Hydrogel at High Temperature. YajC-CT is a small soluble protein motif. A quick heat treatment at 90 °C turned its solution into a self-supporting hydrogel (Figure 1A). To investigate what kind of interaction or protein conformational change facilitated the formation of the hydrogel, we analyzed the sequence of YajC-CT. The atomic resolution structure of YajC-CT is not yet available. The result of secondary structure prediction using the online server of Jpred3 is shown in Figure 1B. According to the prediction, YajC-CT contained mainly random coils and β-strands. The secondary structure content of YajC-CT was studied using CD. At 20 °C, the CD spectrum revealed a conformation characteristic of a mixture of random coil and β-strand, as illustrated by the single minimum peak at ∼207 nm (Figure 2A). The spectrum remained largely unchanged at 50 °C. At 70 °C, the negative maximum peak grew slightly, and the intensity of the positive peak around 190 nm became stronger. At 80 °C, there was a dramatic change following the same trend. Finally, at 90 °C, the structure of YajC-CT shifted to a completely different conformation. The conformational transition occurred at a narrow temperature range. To investigate further the exact temperature that triggered the conformational transition, we monitored the ellipticity at 210 nm with the change of temperature (Figure 2B). The signal at 210 nm remained constant until the temperature reached 75 °C; then, the negative peak grew sharply to reach a plateau at 90 °C. Once formed, this high-temperature conformation remained stable upon cooling. YajC-CT contains one Tyr residue (Y28). Tyr fluorescence had been used to monitor protein conformational changes.2224 We measured the Tyr fluorescence using an excitation wavelength 1579

dx.doi.org/10.1021/bm101571r |Biomacromolecules 2011, 12, 1578–1584

Biomacromolecules

Figure 2. (A) Far-UV CD spectra of YajC-CT at different temperatures. From top to bottom, the traces corresponding to 20 (9), 50 (0), 70 (gray filled triangles), 80 (), and 90 °C (gray open triangles). (B) Temperature denaturation curves. The ellipiticity of a YajC-CT sample at 210 nm was monitored with the increase in temperature from 4 to 90 °C ([) and then decreased from 90 to 4 °C (]). The change of protein conformation was irreversible. (C) Tyrosine fluorescence of purified YajC-CT before (gray) and after (black) heated at 90 °C, followed by cooling.

of 270 nm (Figure 2C). The fluorescence emission of YajC-CT decreased by ∼25% after the heat treatment. Such a change revealed a difference in the local environment of Tyr28, most likely caused by the heat-induced rearrangement of YajC-CT structure. Heat-Induced Gelation. The effects of different parameters including protein concentration, pH, and ionic strength on the gelation of YajC-CT were examined. First, YajC-CT solutions of different concentrations in the binding buffer were heated to 90 °C for 5 min and then cooled to room temperature. When protein concentration was >0.4 mg/mL, a transparent selfsupporting hydrogel was clearly visible after the heat treatment (Figure 1A). The variation of pH (4.0, 7.0, or 10.0) and ionic strength (0, 0.1, 0.5, or 1.25 M of NaCl) of the samples did not result in an observable difference in the gelation process (data not shown). At NaCl concentrations higher than 1.25 M, protein precipitated upon heating. FTIR, Rheology, and AFM Characterization. To probe further the protein structure, we analyzed YajC-CT solution and hydrogel using FTIR (Figure 3). The amide I frequency of protein samples provides valuable information about protein secondary structure. Deconvolution of the spectra yielded dischrete patterns of their constituting secondary structure

ARTICLE

Figure 3. FTIR characterization of (A) YajC-CT solution and (B) hydrogel. Black traces are the experimental data, and the red traces are the fitting result. The different components are β-sheet (blue), random coil (cyan), β-turns (magenta), and antiparallel β-sheet (green).

Figure 4. Viscosity versus shear rate measurement of YajC-CT hydrogel.

components (Figure 3A,B). The major peak in the soluble protein sample that contributed 42.2% of the signal was at 1640 cm1, representing the random coil conformation.2527 Other components included β-sheet (16.5%), β-turns (30.7%), and antiparallel β-sheet (10.6%). After gelation, the major change was the increase in the content of antiparallel β-sheet to 32.7%, accompanied by the decrease in the random coil (36.8%), β-turns (25.7%), and β-sheet (14.8%) structures. These data indicated that the antiparallel β-sheet structure became prominent upon gelation, consistent with the result from the CD study, which also indicated an overall increase in β-sheet content in the protein upon heating. YajC-CT hydrogel is very soft, resembling some soft and injectable hydrogels reported in literature.2831 We measured the viscosity of YajC-CT hydrogels containing 1.0 mg/mL protein at pH 7.0. The results of the viscosity measurement under different shear rates are plotted in Figures 4. The hydrogel 1580

dx.doi.org/10.1021/bm101571r |Biomacromolecules 2011, 12, 1578–1584

Biomacromolecules exhibited shear thinning flow behavior as the viscosity of the hydrogel decreased with the increase in the shear rate. The viscosity values and shear thinning behavior of YajC-CT hydrogel are similar to soft hydrogels reported in literature.2831 After heat-induced gelation, YajC-CT hydrogel was dispersed in a dilute water solution and imaged using AFM (Figure 5). Chunks of hydrogels were visible in Figure 5A as well as smaller and more dispersed pieces. Fibrils and particles of homogeneous height were clearly visible in the zoom in views in Figure 5B,C. The height cross-sectional profile of the fibers and particles revealed a value of ∼3.2 ( 0.7 nm (n = 100). The length of the fibrils was in the micromolar range. Interestingly, the particles seemed to assemble unanimously in a linear manner. Parameters Affecting the Internal Structure of the Hydrogel As Revealed by the Release Rate of Encapsulated Fluorescein. Once formed, the hydrogels were stable. When emerged in a buffer solution, no change of gel size was observed after 30 days of storage. Fluorescein was used as a tracer molecule to examine the internal porosity of the gels because a more porous structure is expected to release encapsulated compound faster. Fluorescein was added to the protein sample before the heat treatment. The resultant transparent hydrogel was visually yellow. When the gel was soaked in a fresh buffer, the yellow color did not vanish over a short period of time, indicating that the fluorescein molecules were actually encapsulated inside the gel. As mentioned above, a critical gelation concentration of ∼0.4 mg/mL was observed through directly monitoring the physical behavior of the protein samples after the heat treatment. In this study, fluorescein released from the hydrogel was examined at different protein concentrations (Figure 6A). In the absence of YajC-CT, fluorescein existed in a free mobile state. Therefore, once fresh buffer was added to the sample, fluorescein was simply diluted, which was defined as a “100% release”. At very low protein concentration, YajC-CT could not form an extensive hydrogel network. Only a small portion of fluorescein was encapsulated. Therefore, the “release” was close to 100%. A sharp drop in percentage release could be observed between protein concentrations of 0.33 and 0.4 mg/mL, at which point the gel formed. The percentage release decreased steadily and slowly with further increase in protein concentration. The result from the release experiment was consistent with the result from visual observation and confirmed a critical gelation concentration of 0.4 mg/mL. A YajC-CT concentration of 1 mg/mL was used to examine the effect of other parameters. Fluorescein-containing hydrogels formed under different pH conditions were incubated in the corresponding buffers to test the effect of pH (Figure 6B). After the gels were incubated in buffers overnight, the fluorescence

ARTICLE

emission spectra of the soaking buffers were measured after dilution into the binding buffer. pH had a dramatic effect on the release. At pH 4.0, the percentage release was ∼28%, compared with 56 and 67%, respectively, at pH 7.0 and 10.0. Next, we measured the cumulative release of fluorescein over time from a YajC-CT hydrogel containing 1 mg/mL protein and 5 μM fluorescein at pH 7.0 (Figure 6C). The percentage release was linear to the incubation time in the first 10 min. Within 80 min, 30% of encapsulated fluorescein was released. We released 45% of encapsulated fluorescein after 500 min of incubation. After the overnight incubation, the cumulative release reached ∼50%. To investigate further the kinetics of substrate release, we made hydrogels containing different initial concentrations of fluorescein and measured the rate of release from each hydrogel sample in the first 5 min (Figure 6D). The release rate correlated linearly to the initial concentration of fluorescein in the gels. The impact of pH on the structure of YajC-CT was investigated through comparing the CD spectra of YajC-CT at three different pH conditions at 20 and 95 °C (Figure 7A). At 20 °C, the CD spectra collected at pH 7.0 and 10.0 overlapped well, whereas at pH 4.0, the protein seemed to be slightly more structured. At 95 °C, the spectra were quite different. Whereas all

Figure 6. Release of fluorescein encapsulated in the hydrogel. (A) Effect of protein concentration. (B) Effect of pH. (C) Percent cumulative release versus time. (D) Rate of release versus initial concentration of fluorescein in the solution.

Figure 5. AFM topography image of a diluted hydrogel sample. Panel B is a zoom in view of the region in the box in panel A, whereas panel C is a zoom in view of the region in the box in panel B. 1581

dx.doi.org/10.1021/bm101571r |Biomacromolecules 2011, 12, 1578–1584

Biomacromolecules three traces were composed of mainly β-strands and random coils, the trace at pH 7.0 appeared to have a higher random coil component. Next, we monitored the change of the ellipticity at 210 nm for samples under different pH conditions with the increase in temperature (Figure 7B). Different pH did not have a drastic effect on the melting temperature, indicating that the thermal stability of YajC-CT was not affected by pH. However, the final ellipticity values at 95 °C were very different at different pH, indicating that the secondary structure of the protein in the fibers formed at different pH might be different. Finally, we used AFM to characterize protein fibers formed under different pH (Figure 8). Whereas fibrils formed at both pH 4.0 and 10.0 resembled the linear assembly mode observed for fibrils formed at pH 7.0, the dimensions of the fibrils were different.

Figure 7. (A) Far-UV CD spectra of YajC-CT at 20 or 90 °C under three different buffer conditions: pH 10.0 (red), 7.0 (gray), and 4.0 (black). At each pH, the trace with lower MRE value was obtained at 20 °C. (B) Temperature denaturation curves collected at pH 10.0 (red), 7.0 (gray), and 4.0 (black). The ellipiticity of a YajC-CT sample at 210 nm was monitored with the increase in temperature from 20 to 90 °C.

ARTICLE

At pH 4.0 and 10.0, the heights of the fibrils were 3.6 ( 0.5 (n = 100) and 1.9 ( 0.3 nm (n = 100), respectively. These values are in the same range with the reported heights of fibrils formed from other globular proteins including lysozyme (46 nm) and β-lactoglobulin (ranging from 1.4 to 8.5 nm).1317

’ DISCUSSION A new hydrogel-forming naturally occurring protein motif was reported. Whereas the atomic resolution structure of YajC-CT remains unknown, the CD and FTIR data as well as secondary structure prediction indicated that the solution structure of YajCCT was β-sheet rich. The progress of the heat-induced protein conformational change was monitored using CD. YajC-CT was very thermal stable, with melting temperature as high as ∼85 °C. The β-strand content increased after heating. This result was further confirmed by the FTIR analysis of both the protein solution and the hydrogel. Several naturally occurring globular proteins are known to form hydrogels upon heating. Among them, the most extensively characterized examples are lysozyme (Figure 9B) and β-lactoglobulin (Figure 9C). Both proteins contain a mixed R/β structure in their native conformation. When treated with either heat or organic solvent, the proteins adopt a β-sheet-rich conformation and form hydrogels. Despite decades of research, the exact mechanism of the gelation process is not yet completely understood. According to a popular model pathway, the globular proteins first partially or completely unfold and lose their secondary structure.32 The unfolded protein is the intermediate

Figure 9. (A) Ribbon diagrams of the predicted structures of YajC-CT. (B) Ribbon diagram of the crystal structure of hen egg white lysozyme (created using pdb file 3IJV). (C) Ribbon diagram of the crystal structure of β-lactoglobulin (created using pdb file 3NPO).

Figure 8. AFM phase images of diluted hydrogel samples formed at pH (A) 10.0, (B) 7.0, and (C) 4.0. (Note: phase image in panel B shows the same region as the topography image in Figure 4C. Their image contrasts are coincidentally similar). 1582

dx.doi.org/10.1021/bm101571r |Biomacromolecules 2011, 12, 1578–1584

Biomacromolecules

ARTICLE

necessary to understand fully the molecular self-assembly process that lead to the formation of fibrils and hydrogels.

’ AUTHOR INFORMATION Corresponding Author

*Tel: (859) 257-7085. Fax: (859) 323-1069. E-mail: yinan. [email protected]. Figure 10. Schematics of the linear and random assemblies of protein particles.

state, which further aggregates and forms amyloid fibrils. The fibrils subsequently self-assemble into hydrogels. We used the online server of LOOPP3335 to predict the tertiary structure of YajC-CT. The top three predicted structures are shown in Figure 9A. Consistent with the secondary structure information obtained from CD and FTIR, the protein contains extensively β-strands and random coils. However, the actual percentage of β-strands observed from CD and FTIR seemed to be less than that predicted from the sequence. Upon heating, the protein structure rearranged and the β-strands content increased. We further investigated the effect of different solution conditions on the formation and internal structure of the YajC-CT hydrogel. The ionic strength (