Biomacromolecules 2004, 5, 1849-1852
1849
Acid Extraction and Purification of Recombinant Spider Silk Proteins Charlene M. Mello,*,† Jason W. Soares,† Steven Arcidiacono,† and Michelle M. Butler‡ U.S. Army RDECOM Natick Soldier Center, Kansas St., Natick, Massachusetts 01760-5020, and Microbiotix, Inc., One Innovation Drive, Worcester, Massachusetts 01605 Received March 29, 2004; Revised Manuscript Received May 21, 2004
A procedure has been developed for the isolation of recombinant spider silk proteins based upon their unique stability and solubilization characteristics. Three recombinant silk proteins, (SpI)7, NcDS, and [(SpI)4/(SpII)1]4, were purified by extraction with organic acids followed by affinity or ion exchange chromatography resulting in 90-95% pure silk solutions. The protein yield of NcDS (15 mg/L culture) and (SpI)7 (35 mg/L) increased 4- and 5-fold, respectively, from previously reported values presumably due to a more complete solubilization of the expressed recombinant protein. [(SpI)4/(SpII)1]4, a hybrid protein based on the repeat sequences of spidroin I and spidroin II, had a yield of 12.4 mg/L. This method is an effective, reproducible technique that has broad applicability for a variety of silk proteins as well as other acid stable biopolymers. Introduction Spiders produce a family of silk proteins with a wide range of physical and mechanical properties. As a result, spider silks represent a unique opportunity to understand nature’s intelligent design of structural biopolymers. The amino acid composition of these proteins is predominantly glycine, alanine, and other short side chain amino acids, which form antiparallel beta-pleated sheets.3-5 Silk proteins exhibit a highly repetitive sequence,6,7 resistance to digestion by proteolytic enzymes,8 and insolubility in aqueous solutions including dilute acids and bases. Its limited solubility continues to impede the fundamental characterization of fulllength native spider silk proteins. Due to these characteristics, harsh solvents such as 9 M lithium bromide,6 HFIP,9 and concentrated formic acid solutions7 have been required to solubilize native dragline silk. The most extensively studied spider silk is the dragline of Nephila claVipes. Dragline silk possesses desirable mechanical properties, being three times tougher than aramid fibers and five times stronger by weight than steel.10,11 As a result, there is an ongoing interest in duplicating the mechanical properties of spider silk fibers. Unlike silkworms, spiders are not suitable as a commercial source of silk due to their relatively short lifespan, their difficulty of breeding in captivity, their territorial nature, and the complexity of multiple silk proteins produced from a single set of spinnerets. Many research groups have turned to heterologous protein expression in an effort to produce sufficient protein for characterization and purification studies. For example, analogues of the two proteins reportedly found in Nephila claVipes dragline silk, spidroins I and II, have been expressed in E. coli.1,7,12 Arcidiacono et al. also reported the expression * To whom correspondence should be
[email protected]. † U.S. Army RDECOM Natick Soldier Center. ‡ Microbiotix, Inc.
addressed.
E-mail:
of a partial cDNA clone from Nephila claVipes major ampullate gland in E. coli.2 Other host expression systems employed include Pichia pastoris13 and plants.14 In general, yields were low, presumably due to the unique structural and chemical characteristics of spider silk proteins. Affinity chromatography has been a common purification approach for recombinant silk proteins.1,2,7,12 Alternatively, Fahnestock et al. reported a purification technique based on the unique temperature stability of silk proteins.13 Scheller et al. also took advantage of the temperature stability of recombinant silk proteins produced in plants. Homogenized tissue was boiled and fractionated with ammonium sulfate triggering precipitation of most contaminating cellular proteins.14 Unfortunately, purified solutions were dilute and could not be concentrated without becoming insoluble. Here we report an alternative purification that also takes advantage of the unique stability and solubility characteristics of silk proteins. Organic acids are used to lyse E. coli cell pellets and enrich for recombinant spider silk proteins while hydrolyzing many of the unwanted bacterial proteins. Three recombinant silk proteins, (SpI)7, NcDS, and [(SpI)4/(SpII)1]4, were isolated by extraction with organic acids followed by affinity or ion exchange chromatography. This method displays numerous advantages over traditional purification techniques, resulting in a rapid enrichment of silk proteins and a significant increase in overall protein yield. In addition, the proteins can be processed into concentrated solutions for applications such as fiber spinning15 and structure/function studies. Materials and Methods Materials. Organic acids and buffer components were obtained from Sigma-Aldrich (St. Louis, MO). Q Sepharose Fast Flow ion-exchange resin was purchased from Amersham Pharmacia Biotech (Piscataway, NJ). Nickel-nitrilotriacetic acid (Ni-NTA) resin was purchased from Qiagen Inc.
10.1021/bm049815g CCC: $27.50 © 2004 American Chemical Society Published on Web 07/10/2004
1850
Biomacromolecules, Vol. 5, No. 5, 2004
Mello et al.
Table 1. Summary of Expression, Lysis, and Purification Methods for Each Recombinant Silk Proteina recombinant silk protein
expression vector
cell line
(SpI)7
pQE9
NcDS
pET24
[(SpI)4/(SpII)1]4
pET21
BL21(DE3) pLysS BL21(DE3) pLysS SG13009 (pREP4)
chromatography method
resin quantityb
4.6 N FA
nickel affinity
4.5
2.3 N FA/ 3 N Gdn-HCl 2.3 N PA/ 2 N Gdn-HCl
nickel affinity
1.25
ion exchange
2
acid lysis
a Choice and volume of acid for cell pellet extraction was protein dependent and determined empirically. Addition of Gdn-HCl during lysis was necessary to maintain protein solubility during purification and processing. b mL of resin/g of dry cell.
(Chatswoth, CA). Ultrafiltration membranes and systems were obtained from Millipore (Bedford, MA). Chromatography columns were obtained from Bio-Rad Laboratories (Hercules, CA). Protein Expression. The NcDS cDNA and the [(SpI)4/ (SpII)1]4 gene were cloned into pET24 (Novagen Inc., Madison WI) and expressed in BL21(DE3) pLysS. The (SpI)7 gene was cloned into pQE9 (Qiagen Inc., Valencia, CA) and expressed in SG13009(pREP4). Protein was produced in fermentations using defined or complex medium with appropriate antibiotics.1,2 Acid Lysis. Organic acids were evaluated to determine the optimal acid for each recombinant protein. Cell pellets were lyophilized and ground into a fine powder with a mortar and pestle and reconstituted with the following acids; 23 N formic acid (FA), 13.3 N propionic acid (PA), and 17.5 N acetic acid (2-100 mL organic acid/g dry cell). The mixture was diluted in MQ water and 6 N guanidine-HCl (Gdn-HCl) to a final concentration of 2.3 or 4.6 N organic acid and 2 or 3 N Gdn-HCl (Table 1). The cell suspension was stirred for 1 h at room temperature then clarified by centrifugation at 25 000×g for 1 h, 20 °C. To maximize recovery of recombinant protein, cell pellets were washed with organic acid/Gdn-HCl and clarified by centrifugation for 30 min. All soluble fractions were pooled for additional purification steps. Nickel Affinity Chromatography (NcDS & (SpI)7). Clarified supernatants were concentrated by ultrafiltration with a 10 000 MWCO membrane to reduce the volume 5-20-fold. Concentrated samples were clarified by centrifugation at 20 000×g for 15 min, 20 °C. Binding, wash, and elution buffers each consisted of 8 M urea, 100 mM NaH2PO4, 10 mM Tris, and 200 mM NaCl. Clarified supernatant was dialyzed into binding buffer, pH 8. Insoluble material was removed by centrifugation. The supernatant was added to an equilibrated Ni-NTA slurry (Table 1) and stirred for 1 h at room temperature. The slurry was transferred to a chromatography column to collect the unbound fraction. The resin was washed with wash buffer, pH 6.75 (NcDS) or pH 7 ((SpI)7), followed by five column volumes of elution buffer, pH 3, to collect the recombinant silk protein. The eluant was dialyzed into 160 mM urea, 10 mM NaH2PO4, 1 mM Tris, and 20 mM NaCl, pH 5 and clarified through centrifugation at 20 000×g for 15 min, 20 °C. Ion Exchange Chromatography ([(SpI)4/(SpII)1]4). Clarified cell lysate was concentrated 2.5-fold by ultrafiltration (10 000 MWCO). Concentrated samples were clarified by centrifugation at 20 000×g for 15 min, 20 °C and dialyzed into ion exchange buffer (2 M urea, 10 mM Tris, pH 10.5). Dialyzed supernatant was clarified by centrifugation and
incubated with slurry of Q-Sepharose Fast Flow ion-exchange resin (Table 1) stirring at room temperature for 1 h. The slurry was transferred to a column and the flow-through, containing the recombinant silk protein, was collected. The column was washed twice with ion exchange buffer, pooled with the flow-through fraction, and dialyzed into 160 mM urea, 10 mM NaH2PO4, 1 mM Tris, 10 mM glycine, pH 5. The silk sample was clarified by centrifugation and concentrated by stirred cell ultrafiltration (10 000 MWCO). Insoluble material was removed by centrifugation. Supernatant was further concentrated 10-fold by a 30 000 MWCO Centriprep ultrafiltration system and clarified through centrifugation. Gel Electrophoresis. SDS-PAGE analysis was performed on all fractions collected throughout the purification using 10-20% Tricine gels (Invitrogen Inc., Carlsbad, CA). The resulting gels were stained with Coomassie Brilliant Blue and destained in 10% methanol, 7.5% acetic acid. Protein Characterization. Protein purity was determined by densitometry using Total Lab V1.11 imaging/densitometry software (Phoretix, United Kingdom) and amino acid compositional analysis. Quantitation of recombinant protein solutions was determined by bicinchoninic acid protein assay (BCA) and also spectrophotometrically at A280.16 Results Organic Acid Extraction. Bacterial cells were lysed with organic acids to determine the quantity of silk extracted (Figure 1). Several organic acids were examined, but only formic acid, propionic acid, and acetic acid extracted the recombinant silk proteins with reproducible yield and purity. Formic acid extracts also contained many bacterial proteins. Butyric and valeric acids were also evaluated, but failed to extract any recombinant silk protein (data not shown). For NcDS and (SpI)7, formic acid yielded the largest quantity of protein, whereas acetic and propionic acids resulted in much lower yields. In contrast, both formic and propionic acid extracted the same amount of [(SpI)4/(SpII)1]4. In comparison to formic acid, extraction with propionic resulted in a more highly enriched [(SpI)4/(SpII)1]4 lysate, but the protein became insoluble during the subsequent buffer exchange required for ion exchange chromatography. The addition of Gdn-HCl during extraction enhanced protein solubility making purification possible. Although the inclusion of Gdn-HCl resulted in a lysate with decreased purity, the majority of the bacterial proteins were removed during purification. Purification of NcDS & (SpI)7. Lysates containing NcDS and (SpI)7 were purified by Ni-NTA affinity chromatog-
Biomacromolecules, Vol. 5, No. 5, 2004 1851
Recombinant Spider Silk Proteins
Figure 1. SDS gel electrophoresis of protein extraction with organic acids. Organic acid lysate supernatants were run on a 10-20% Tricine gel and visualized with Coomassie blue stain. Lane 1: Mark 12 molecular weight marker (Invitrogen), in kDa. Lanes 2-4, (SpI)7: 2, formic acid (FA); 3, acetic acid (HOAc); 4, propionic acid (PA). Lanes 5-7, [(SpI)4/(SpII)1]4: 5, FA; 6, HOAc; 7, PA. Lanes 8-10, NcDS: 8, FA; 9, HOAc; 10, PA. Table 2. Recombinant Silk Protein Purity and Yieldsa purity (%)
yield (mg/L culture)
recombinant silk protein
current
literature
current
literature
(SpI)7 NcDS [(SpI)4/(SpII)1]4
89 95 89
70 90-95 nrb
35 15 12.4
7 4 nrb
a Purity with regard to silk monomer was determined by densitometry analysis performed on SDS-PAGE. Protein yield was calculated using BCA and spectrophotometrically at A280. b nr ) not previously reported.
raphy. The purifications were similar with minor differences in resin quantity and pH of the wash buffer. These differences were important and necessary to maximize purity and yield. NcDS purified solutions resulted in yields of 15 mg recombinant silk protein per liter culture (mg/L). (SpI)7 yields were as high as 100 mg/L but more typically 35 mg/L. Protein yields were increased relative to previously reported data 4- and 5-fold, respectively (Table 2). Protein purity was determined at each step of the purification process using densitometry to determine the percentage of recombinant silk monomer in the protein sample (Figure 2A). Amino acid analysis revealed slightly higher purity than the densitometry result due to the fact that full-length and truncated recombinant silk products are expected to have a very similar composition (data not shown). The purity of NcDS of 95% was similar to previously reported values of 90-95%, while the 89% purity of the (SpI)7 solution was a significant improvement over the previously reported value of 70% (Table 2; Figure 2A). Dialysis into a dilute denaturing salt buffer was necessary since the affinity purified NcDS and (SpI)7 were only 56% and 50% respectively. Purification of [(SpI)4/(SpII)1]4. Ion exchange chromatography was used for purification of the dialyzed [(SpI)4/ (SpII)1]4 lysate. Anion-exchange resin was used such that the positively charged [(SpI)4/(SpII)1]4 protein did not bind to the column while contaminating proteins with lower isoelectric points and net negative charges were retained. The majority of contaminating proteins in the lysate, resulting from the addition of Gdn-HCl, were removed during
Figure 2. SDS-PAGE analysis of recombinant silk protein purification by Coomassie staining. Samples were run on 10-20% Tricine gels. Lanes 1 & 12: Mark 12 molecular weight standards (Invitrogen), in kDa. (A) Lanes 2-4, (SpI)7: 2, lysate dialyzed into 8 M urea, 100 mM NaH2PO4, 10 mM Tris, 200 mM NaCl; 3, Ni-NTA purified; 4, dialyzed into 160 mM urea, 10 mM NaH2PO4, 1 mM Tris, 20 mM NaCl, pH 5. Lanes 5-7, NcDS: 5, lysate dialyzed into 8 M urea, 100 mM NaH2PO4, 10 mM Tris, 200 mM NaCl; 6, Ni-NTA purified; 7, dialyzed into 160 mM urea, 10 mM NaH2PO4, 1 mM Tris, 20 mM NaCl, pH 5. (B) Lanes 8-11, [(SpI)4/(SpII)1]4: 8, dialyzed into 2 M urea, 10 mM Tris, pH 10.5; 9, ion exchange purified; 10, dialyzed into 160 mM urea, 10 mM NaH2PO4, 1 mM Tris, 10 mM glycine, pH 5; 11, ultrafiltration retentate.
purification with the exception of several proteins typically between 6 and 12 kDa (Figure 2B). These contaminants were removed during the ensuing dialysis and ultrafiltration with minimal loss of recombinant silk protein resulting in a highly enriched solution. The resulting purity of [(SpI)4/(SpII)1]4 was 89% with a yield of 12.4 mg/L culture (Table 2). Discussion Organic acid extraction of recombinant spider silk makes use of the unique solubilization and stability characteristics of these proteins, which are resistant to acid hydrolysis for prolonged periods of time at room temperature. The quantity of recombinant silk protein recovered was dependent upon the volume of organic acid used for lysis. Generally, as the volume of organic acid increased, the quantity of the extracted silk protein increased without reducing lysate purity. Eventually, larger volumes of organic acid did not result in increasing silk protein recovery. Since minimal silk protein is lost during the purification, the final protein yield is primarily dependent on the quantity extracted from the cells during the organic acid lysis. Overall protein yields exhibited variability presumably due to differences in cell growth and the effectiveness of the cell pellet lysis. The volume of acid required for silk protein extraction from a given cell pellet varied and was therefore determined empirically. Cell lysis was optimized with consideration given to subsequent chromatographic steps and protein stability during processing. NcDS and (SpI)7 utilized a hexahistidine tag for purification by nickel affinity resin. Through optimization of wash and elution buffer pH, NcDS and (SpI)7 recombinant silk
1852
Biomacromolecules, Vol. 5, No. 5, 2004
solutions were highly purified. Generally, RP-HPLC is required to remove bacterial proteins still present in the eluted solution; however, dialysis into a dilute denaturing salt buffer was sufficient to completely purify the recombinant proteins. For [(SpI)4/(SpII)1]4, most contaminating proteins were removed during ion exchange chromatography although bacterial proteins between 6 and 12 kDa remained. Ultrafiltration proved to be effective in their removal and ultimately completed the purification of [(SpI)4/(SpII)1]4 with minimal losses in resultant yield. Other purification techniques such as size exclusion chromatography, addition of detergents, and HPLC were not successful in their removal, possibly due to a strong interaction between the silk and the protein contaminants. Although affinity chromatography is a commonly used purification method, the presence of the histidine affinity tag is not always desirable. Its presence can significantly increase protein antigenicity as well as alter protein solubility and secondary structure that can ultimately change protein behavior. Accordingly, there is a continuing need to develop new methods for silk purification that will not adversely alter the protein properties needed for applications such as fiber spinning. The use of ion exchange chromatography does not require the histidine tag and is more amenable to scale-up than affinity chromatography. In addition, the resulting purified solutions exhibit the protein solubility necessary for further processing into concentrated solutions for applications such as fiber spinning and structure/function studies. Finally, the development of purification techniques that maximize extraction of recombinant silk proteins with enhanced yields is essential for the realization of recombinant silk as a viable material. Acknowledgment. The authors thank Kelly Burke, Alfred Allen, and Rob Stote for their efforts with recombinant silk
Mello et al.
production. Also, Tom Laue, Susan Chase, and Steve Fossey are thanked for their invaluable discussions and contributions. References and Notes (1) Prince, J. T.; McGrath, K. P.; DiGirolamo, C. M.; Kaplan, D. Biochemistry 1995, 34, 10879. (2) Arcidiacono, S.; Mello, C.; Kaplan, D.; Cheley, S.; Bayley, H. Appl. Microbiol. Biotechnol. 1998, 49, 31. (3) Marsh, R. E.; Corey, R. B.; Pauling, L. Biochim. Biophys. Acta 1955, 16, 1. (4) Lucas, F. DiscoVery 1964, 25, 20. (5) Andersen, S. O. Comp. Biochem. Physiol. 1970, 35, 705. (6) Mello, C. M.; Yeung, B.; Senecal, K.; Vouros, P.; Kaplan, D. L. In Silk Polymers, Material Science and Biotechnology; Kaplan, D. L., Adams, W. W., Farmer, B., Viney, C., Eds.; ACS Symposium Series 544; American Chemical Society: Washington, DC, 1994; pp 6779. (7) Lewis, R. V.; Hinman, M.; Kothakota, S.; Fournier, M. J. Protein Expression Purif. 1996, 7, 400. (8) Tillinghast, E. K.; Kavanaugh, E. J. J. Exp. Zool. 1977, 202, 213. (9) Seidel, A.; Liivak, O.; Jelinski, L. W. Macromolecules 1998, 31, 6733. (10) Kaplan, D. L.; Lombardi, S. J.; Muller, W. J.; Fossey, S. A. In BioMaterials: NoVel Materials from Biological Sources; Byrom, D., Ed.; Stockton Press: New York, 1991; pp 1-53. (11) Cunniff, P. M.; Fossey, S. A.; Auerbach, M. A.; Song, J. W. In Silk Polymers: Materials Science and Biotechnology; Kaplan, D. L., Adams, W. W., Farmer, B., Viney, C., Eds.; ACS Symposium series 544; American Chemical Society: Washington, DC, 1994; pp 234251. (12) Fahnestock, S. R.; Irwin, S. L. Appl. Microbiol. Biotechnol. 1997, 47, 23. (13) Fahnestock, S. R.; Bedzyk, L. A. Appl. Microbiol. Biotechnol. 1997, 47, 33. (14) Scheller, J.; Guhrs, K.-H.; Grosse, F.; Conrad, U. Nature Biotechnol. 2001, 19, 573. (15) Arcidiacono, S.; Mello, C.; Butler, M.; Welsh, E.; Soares, J. W.; Ziegler, D.; Allen, A.; Laue, T.; Chase, S. Macromolecules 2002, 35, 1262. (16) Gill, S. C.; von Hipple, P. H. Anal. Biochem. 1989, 28, 319.
BM049815G
