265
Biotechnol. Prog. 1995, 11, 265-269
Purification by Immobilized Metal Affinity Chromatography of Human Atrial Natriuretic Peptide Expressed in a Novel Thioredoxin Fusion Protein David L. Wilkinson, Nien-Tung Ma, Chris Haught, and Roger G. Harrison" School of Chemical Engineering and Materials Science, University of Oklahoma, Norman, Oklahoma 73019-0628
A fusion protein t h a t contains human atrial natriuretic peptide (ANP) at its carboxy terminus has been genetically engineered with the objective of being able to produce the peptide in a process with a relatively simple purification procedure. The fusion protein also includes a (Hi& metal affinity binding site at the amino terminus, followed by Escherichia coli thioredoxin, a factor X,protease recognition site, and ANP. With induction of the tac promoter at 30 "C,the expression Ievel of the fusion protein was high (10% of total cell protein as measured by densitometry) and it was almost completely (92%) expressed as a soluble protein in the cytoplasm. A step gradient elution with imidazole of a column of Ni2+ chelated to iminodiacetic acid-agarose saturated with proteins in crude cell extract gave a very nearly pure fusion protein. After digestion of the purified fusion protein with factor X, protease, ANP of exactly the correct size (to within 2 Da) was observed by coupled HPLC/mass spectrometry.
Introduction Peptides have been studied intensively in recent years for a variety of medical applications, including treatments for degenerative disorders, infections, blood clots, and kidney failure (Pramik, 1992). The success of some of these treatments has led to a demand for peptides that is larger than can be supplied by solid phase peptide synthesis. Theoretically, recombinant DNA technology should be able to supply this demand for larger quantities of peptides; however, because of their relatively small size (10 000 Da and smaller), it generally is not possible to express peptides directly in bacterial cells because they are degraded rapidly by host proteases (Wetzel and Goeddel, 1983). The subject of this study is how to produce a model peptide by recombinant DNA technology, with a simple purification scheme. The approach used here is to express the peptide as a soluble fusion protein with a relatively small Escherichia coli protein with a metal affinity binding site added on. One of the most widely used metal affinity binding sites for facilitating protein purification by immobilized metal affinity chromatography (IMAC) is a stretch of two or more histidines (Gentz et al., 1988; Hochuli et al., 1988; Lilius et al., 1991;Van Dyke et al., 1992).These histidine stretches bind to a transition metal, e.g., Ni2+ or Zn2+, immobilized by a chelating agent such as iminodiacetic acid (IDA)coupled to agarose. A six-histidine stretch was employed here because of its ability to enable the purification of several eukaryotic proteins in only one step under nondenaturing conditions (Van Dyke et al., 1992). When designing an expression system for recombinant fusion proteins, it is desirable to have a high level of expression in a soluble form. As many problems have been reported expressing eukaryotic proteins in E. coli, the simplest solution is to use a native E. coli protein, preferably a small protein, as this allows proportionally higher expression of the recombinant peptide. A protein that is soluble is desirable in order to avoid the problems associated with purifying proteins expressed in insoluble
* Author t o whom correspondence should be addressed.
inclusion bodies. The fusion protein partner selected for this research was E. coli thioredoxin. This protein is only 11700 Da in size, and Lunn et al. (1984) reported. overexpressing it at high levels in soluble form in E. coli. While this research was being completed, LaVallie et al. (1993) reported the expression of a number of heterologous proteins as fusions with thioredoxin, confirming our expectations of the usefulness of thioredoxin in a fusion protein. The thioredoxin fusion proteins were significantly more soluble than the heterologous proteins expressed by themselves. Human atrial natriuretic peptide (ANP) was selected as a model peptide to be produced in this fusion protein. This peptide is a potent diuretic and hypotensive agent that under certain circumstances inhibits renin and aldosterone secretion (Maack et al., 1985) and may find application for the treatment of kidney failure. The size of ANP, 28 residues, is typical of many other important peptides.
Materials and Methods Bacterial Strains and Plasmids. E. coli JM105 was used for protein expression and cloning experiments, except for site-directed mutagenesis, where E. coli mut S was used. E. coli thioredoxin was generously provided on plasmid pCJF4 by James Fuchs of the University of Minnesota. Plasmid pCA919 was previously constructed in this laboratory (Ma, 1993) starting with expression vector pKK223-3 (Pharmacia Biotech Inc., Piscataway,
NJ). Construction of Recombinant Plasmid. Recombinant DNA methods were, in general, performed as described by Sambrook et al. (1989). Site-directed mutagenesis to add a (His16 gene was carried out by the method of Haught et al. (1994). DNA used for cloning and sequencing was purified with Nucleobond columns (Nest Group, Southboro, MA). PCR was run for 30 cycles on a programmable thermal cycler (Perkin-Elmer Cetus, Norwalk, CT). Oligonucleotides for PCR and sitedirected mutagenesis were synthesized and purified by the Molecular Biology Resource Facility at the University
8756-7938/95/3011-0265$09.00/0 0 1995 American Chemical Society and American Institute of Chemical Engineers
Biotechnol. Prog., 1995, Vol. 11, No. 3
266
of Oklahoma Health Sciences Center, Oklahoma City, OK. DNA sequencing was performed by Lofstrand Laboratories (Gaithersburg, MD). Enzymes were obtained as follows: restriction endonucleases from New England Biolabs (Beverly, MA), Stratagene (La Jolla, CA), and United States Biochemical (Cleveland, OH); Taq DNA polymerase for PCR from Perkin-Elmer Cetus; T4 DNA ligase from Ambion (Austin, TX);T4 DNA polymerase from New England Biolabs; and T4 polynucleotide kinase from United States Biochemical. Expression of Protein and Preparation of CellFree Extract. E. coli JM105 containing the recombinant plasmid was grown in shake flasks at 37 "C to the mid-log phase. The cells were induced with 1mM IPTG and grown for an additional 3 h a t the temperature of expression, either 30 or 37 "C. Subsequent purification steps were performed at 4 "C. Cells were collected by centrifugation, suspended in a 1:30 volume of phosphatebuffered saline (PBS: 0.1 M sodium phosphate (pH 8.0) and 0.3 M sodium chloride), and ruptured by sonication. The crude extract was centrifuged at 8000 rpm to remove cell debris, and the supernatant was collected. Chromatography. All chromatography experiments were carried out at 4 "C using supernatant obtained after the sonication of cells that had been induced a t 30 "C. A 1 x 5 cm column of Ni"-IDA (Chelating Sepharose 6B, Pharmacia Biotech Inc., Piscataway, NJ) was first equilibrated with a solution of acetate-buffered saline (ABS: 0.1 M sodium acetate (pH 4.0) and 0.1 M sodium chloride), charged with a 1%solution of NiClz in ABS, rinsed with ABS, and equilibrated with PBS. The supernatant after centrifugation was loaded on the column at a loading of 4 mUmL of column packing. A flow rate of 0.5 mumin, equivalent to a superficial velocity of 38 cmih, was used for loading, washing, and elution. Cleavage of the Fusion Protein. Purified fusion protein was dialyzed against digestion buffer (20 mM Tris-HC1, 100 mM sodium chloride, and 2 mM calcium chloride) in a Centricon-10 microconcentrator (Amicon, Danvers, MA) at 4 "C. The digestion was carried out at a factor &:protein ratio of 1 : l O (w/w) at room temperature. Factor X,protease was obtained from New England Biolabs (Beverly, MA). Protein Assays. Standard Laemmli SDS-PAGE was used for protein detection (Laemmli, 1970). Quantitative protein determination was made using laser densitometer measurements of the band intensities of Coomassie Blue-stained gels (UltroScan Model 2202, Pharmacia Biotech Inc.). To visualize ANP, the tricine-buffered SDS-PAGE system of Schaegger and Von Jagow (1987) was used. For detection of ANP by Western blotting, proteins on a polyacrylamide gel were transferred to a nitrocellulose membrane. The proteins bound to the membrane were stained with Ponceau S solution (Sigma Chemical Co., St. Louis, MO) to mark the location of the fusion protein. After the Ponceau S was destained with water, the membrane was incubated with anti-ANP antibody (Peninsula Laboratories, Belmont, CA) and then with antiIgG alkaline phosphatase conjugate (Promega Corp., Madison, WI). Bound conjugate was visualized using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium (Promega). Recombinant ANP was analyzed using coupled HPLC/ mass spectrometry by the Molecular Biology Resource Facility at the University of Oklahoma Health Sciences Center (Michrom BioResources Model UMA-600 highperformance liquid chromatography, Pleasanton, CA PE Sciex Model AP13 mass spectrometer, Toronto, Canada).
EcoR I
Pvu I
''
(His) Gene'
Thioredoxin Structural Gene
Xho I
Ile-Glu-Gly-Arg A N P Gene Gene
3'
An acetonitrile gradient in 0.1% trifluoroacetic acid was used for the HPLC analysis on a reversed phase C18 column. The ANP standard used for the HPLC/mass spectrometry and SDS-PAGE analyses was from Sigma Chemical.
Results and Discussion Cloning and Expression of Thioredoxin Fusion Protein. The gene for ANP was obtained from plasmid pCA919, which had previously been constructed for the expression of another fusion protein consisting of Lasparaginase at the N-terminus, ANP at the C-terminus, and the factor X, cleavage site in between (Ma, 1993). Plasmid pCA919 contains a unique EcoRI site upstream of the fusion protein, as well as a unique XhoI site near the 5'-end of the ANP gene. The first step in the construction of the fusion protein gene was to perform recombinant PCR on plasmid pCJF4 to add an EcoRI site at the 5'-end of the thioredoxin gene and to add the factor & cleavage site gene and the portion of the ANP gene up to the XhoI site to the 3'-end of the thioredoxin gene. The DNA generated by PCR was sequenced to verify its sequence. This gene was then substituted for the basparaginase-factor X,cleavage site gene on pCA919 using the EcoRI and XhoI sites on pCA919. Site-directed mutagenesis was then used to add the (His16 metal affinity binding site gene to the 5'-end of the thioredoxin gene and create pDW1. A map of the fusion protein gene and the position of the fusion protein gene on the expression vector pDWl is shown in Figure 1. An SDS-PAGE analysis is shown in Figure 2 of the cell lysate of a culture grown a t 37 "C and induced with IPTG a t either 30 or 37 "C from a colony that had screened positively by restriction enzyme analysis of the recombinant plasmid. Compared to the uninduced culture (lanes 2 and 4), the induced cultures shown in lanes 3 and 5 have a new protein indicated by the arrow in Figure 2. Immunoblotting of the cell lysate proteins with rabbit anti-ANP serum after transfer from the SDSPAGE gel to a nitrocellulose membrane provided positive identification of the fusion protein (shown in Figure 3, lanes 1 and 2, for cells induced a t 30 "C). This immunoblot indicates that the new protein produced upon induction, which was the same size as the fusion protein, reacted selectively with anti-ANP antibody.
Biotechnol. Prog., 1995, Vol. 11, No. 3 1
2
3
267 4
1.o
5
0 P4 00
c)
a.~
3 e
0.5
s:
2 0.0
20 kd
+ 14 kd
Figure 2. SDS-PAGE showing protein expression at two temperatures: lane 1, molecular mass standards; lane 2, uninduced culture at 30 "C; lane 3, culture induced a t 30 "C with 1 mM IPTG; lane 4,uninduced culture a t 37 "C; lane 5, culture induced a t 37 "C with 1 mM IPTG. The position of the fusion protein is indicated by the arrow. 1
2
3
4
Figure 3. Western blot analysis of fusion protein expression and purification. Proteins were separated by SDS-PAGE and electrophoretically transferred to a nitrocellulose membrane. Proteins bound to the membrane were stained with Ponceau S solution to mark the location of the fusion protein (arrow). After the Ponceau S was destained, the membrane was treated with anti-ANP antibody and then with anti-IgG alkaline phosphatase conjugate: lane 1, uninduced culture a t 30 "C; lane 2, induced culture a t 30 "C; lane 3, 100 mM imidazole eluent; lane 4,500 mM imidazole eluent.
The level of expression of induced protein as measured by densitometry was 10% of total cell protein for induction at 30 "C and 20% of total cell protein for induction a t 37 "C. In calculating these levels of expression, the densitometry readings of the proteins in the uninduced cultures running at the same position as the fusion protein were subtracted from the densitometry reading of the fusion protein. These expression levels fall within the range found by LaVallie et al. (1993) for the expression of thioredoxin fusion proteins (5-20% of total cell protein). The 10%expression level corresponds to 10 mg
Elution Volume (ml) Figure 4. Step gradient during IMAC using a column of Ni"IDA that was loaded to saturation with proteins in the lysed and clarified E. coZi JM105 cells with pDWl plasmid: (A) 100 mM sodium phosphate, 300 mM sodium chloride, and 10% glycerol, pH 5.0; (B) same as (A), with 100 mM imidazole added; (C) same as (A), with 500 mM imidazole added.
of ANP/g of dry cells, with the assumption of standard protein and water contents of E. coli cells (Watson, 1970). Cells induced at 30 and 37 "C were analyzed by densitometry of SDS-PAGE gels to determine the solubility of fusion protein in the cells. The supernatant and cell pellet after sonication and centrifugation were run on SDS-PAGE, and the gels were scanned by the densitometer. For cells induced a t 37 "C, an average of only 12% of the fusion protein was produced in soluble form, while for 30 "C induction, an average of 92% of the fusion protein was soluble. Immobilized Metal Affinity Chromatography. Because the solubility of the fusion protein was much higher within the cell for induction a t 30 than a t 37 "C, chromatography experiments were performed only on cells induced at 30 "C. M e r the cells were lysed and the cell debris was removed by centrifugation, the supernatant was used as the feed to immobilized metal affinity chromatography. Several different methods were tried for the elution of the fusion protein from a column of NP-IDA. A pH gradient from 8 to 4 failed to elute any fusion protein from the column, which was also the case for an ammonium chloride gradient of 0-2 M a t pH values of 4-7. Imidazole, however, would remove the fusion protein from the column, and an imidazole step gradient was developed that gave the fusion protein in one of the steps (500 mM imidazole) with only one major impurity. The chromatogram for this step gradient elution is shown in Figure 4, and the SDS-PAGE results for the gradient steps are shown in Figure 5. A Western blot of the 100 and 500 mM eluents (Figure 3) showed that only the fusion protein reacted with anti-ANP antibody in these two eluents; the spreading of the fusion protein band in the 500 mM eluent occurred in the electrophoretic transfer to the nitrocellulose membrane due to salt in the sample. The peak containing the fusion protein represents 1200pg of protein, if we assume the published A280nmvalue of thioredoxin (Fasman, 1976). This is equivalent to a loading of fusion protein on the gel of 300 pglmL. It can be observed from the SDS-PAGE in Figure 5 that a considerable amount of impurities was eluted off in the 100 mM imidazole step. The fusion protein that eluted off in the 500 mM imidazole step, however, was not completely pure. The difference in size between the fusion protein and the major impurity can be estimated from the distance between the two protein markers on
Biotechnol, Prog., 1995, Vol. 11, No. 3
268
1
2
3
1
4
20 kd
14 kd
2
3
-
4
5
+- Fusion
-
tANP
Figure 6. SDS-PAGE of factor & digest of IMAC-purified fusion protein: lanes 1and 2, ANP standard; lane 3, molecular mass standards; lane 4, fusion protein plus factor X, after 2 h; lane 5, fusion protein control. 20 kd
14 kd
Figure 5. SDS-PAGE for the IMAC shown in Figure 4: lane 1,feed to column; lane 2,100 mM imidazole eluent; lane 3,500 mM imidazole eluent; lane 4, molecular mass standards. The position of the fusion protein is indicated by the arrow.
the gel to be 3 kDa. This is close to the molecular mass of ANP (3.1 kDa). Thus, this impurity, which did not react with anti-ANP antibody, is most certainly the fusion protein with ANP cleaved off and will not be able to be separated from the fusion protein by immobilized metal affinity chromatography since it also contains the his)^ metal affinity binding site. This cleavage very likely can be prevented by adding an appropriate protease inhibitor to the cells before lysis. Our elution results generally are consistent with those obtained by Van Dyke et al. (1992) for fusion proteins with a (HiS)6 tag expressed in E. coli and purified using a column of NP-IDA. For an N-terminal (HiS)6 fusion with an upstream stimulatory factor, Van Dyke et al. found that considerably more fusion protein was eluted in a 100 mM imidazole wash step than we eluted at this same imidazole concentration. Cleavage of Fusion Protein To Remove ANP. Aliquots of chromatography fractions containing approximately 5 pg of the purified fusion protein that had been desalted were digested .with factor Xa for 0.5, 1, 2, and 4 h. The tricine-buffered SDS-PAGE analysis showed that the reaction was largely complete after 1h and essentially complete after 2 h. As observed in Figure 6, the digest after 2 h showed one strong band with a molecular mass slightly greater than the 14 000 Da marker, compared to one strong band running between 16 000 and 17 000 Da in the control, which was run under the same conditions but without factor Xa. This size differential is consistent with cleavage from the fusion protein of ANP, which has a molecular mass of 3079.5 Da. The ANP marker is reliably clear and sharp, and a rather faint but quite distinct band, running at the same position as the marker, can be seen for the digested fusion protein. It can be observed in Figure 6 that the ratio of the staining intensities of thioredoxin and ANP is higher
than the ratio of 4:l expected from the molecular masses. This observation is believed to be due to the fact that ANY stains much less heavily on a mass basis than larger proteins (data not shown). Immunoblotting of even relatively complete digests of fusion protein, in which a very heavy band of cleaved protein and only a faint band of uncleaved protein existed, showed that the uncleaved band blotted distinctly, while the much heavier cleaved band did not show up at all (not shown). ANP itself did not blot very well. Coomassie Blue staining of a nitrocellulose membrane, after transfer from an SDS-PAGE gel with a distinct ANP band, showed only a faint ANP band. To positively identify the cleaved peptide as ANP, coupled HPLC/mass spectrometry was used. As the mass spectrometer is a high-resolution instrument accurate for the determination of the molecular mass of a molecule, it can be used to confirm the compositional identity of proteins. An ANP standard and purified fusion protein digested with factor X,were analyzed by this technique. Two major peptide peaks were obtained in the analytical HPLC separation of each sample. Mass spectrometry analysis results of these two peptide peaks are shown in Table 1..Relative peak intensity is proportional to the mass of the peptide eluted. For both the fusion protein digest and the ANP standard, the second peptide eluted has a molecular mass within 2 Da or 0.05% of the theoretical molecular mass of ANP (3078.4 Da). The first peptide eluted from the HPLC (see Table 1) very likely corresponds to ANP with Met-16 oxidized to methionine sulfoxide, which has a theoretical molecular mass of 3094.4 Da. The molecular mass of this peptide as determined by mass spectrometry is within 0.11% of the theoretical value. In work by Knott et al. (1988) on ANP purification, the ANP with methionine oxidized eluted slightly earlier on HPLC than did ANP with methionine reduced, which is consistent with the results here. Thus, more of the ANP is in the oxidized form in the fusion protein digest that in the ANP standard.
Conclusions We have shown that a high level of expression of soluble thioredoxin-ANP fusion protein can be obtained in the cytoplasm of E. coli with a tac promoter a t 30 "C. Very nearly pure fusion protein could be obtained from crude cell extract by immobilized metal affinity chromatography with a column of NiII-IDA using an imidazole step gradient elution. Digestion of the purified fusion protein with factor Xa protease released ANP of the
Biotechnol. Prog., 1995, Vol. 11, No. 3 Table 1. Coupled HPLClMass Spectroscopy Data for the Two Major Peptide Peaks of a Fusion Protein Factor X, Digest and ANP Standard peak relative peak molecular sample no.a intensityb (a) mass (Da) digest of fusion protein 1 100 3097.7 2 71 3079.7 ANP standard 1 56 3097.7 2 100 3079.7 a Peaks are numbered in order of elution from the column. Peak intensity is relative to the peptide peak with the greatest intensity for a given sample.
correct size (to within 2 Da) and also A" with methionine oxidized, as observed by coupled HPLC/mass spectrometry.
Acknowledgment We acknowledge financial support by the National Science Foundation (Grant No. CTS-9011746) and the American Heart Association, Oklahoma Affiliate (Grant No. OK-90-G-7). Thanks to Dr. David McCarthy and Dr. Bruce Roe for suggestions and advice throughout the project, Dr. Ken Jackson for advice about the HF'LC/mass spectrometry analyses, and to Dr. James Fuchs for providing plasmid pCJF4. Literature Cited Fasman, G. D. Handbook of Biochemistry and Molecular Biology, 3rd Edition, Proteins-Volume III; CRC Press: Cleveland, OH, 1976; p 509. Gentz, R.; Certa, U.; Takacs, B.; Matile, H.; Dobeli, H.; Pink, R.; Mackay, M.; Bone, N.; Scaife, G. Major Surface Antigen p190 of Plasmodium falciparum: Detection of Common Epitopes Present in a Variety of Plasmodia Isolates. EMBO J . 1988, 7, 225-230. Haught, C.; Wilkinson, D. L.; Zgafas, K.; Harrison, R. G. A Method to Insert a DNA Fragment into a Double-Stranded Plasmid. BioTechniques 1994, 16, 46-48. Hochuli, E.; Bannworth, W.; Dobeli, H.; Gentz, R.; Stuber, D. Genetic Approach to Facilitate Purification of Recombinant Proteins with a Novel Metal Chelate Adsorbent. Biol Technology 1988, 6 , 1321-1325. Knott, J. A.; Sullivan, C. A,; Weston, A. The Isolation and Characterization of Human Atrial Natriuretic Factor Produced as a Fusion Protein in Escherichia coli. Eur. J . Biochem. 1988,174, 405-410.
269 Laemmli, U. K. Cleavage of Structural Proteins during the Assembly of the Head of Bacteriophage T4. Nature 1970,227, 680-685. LaVallie, E. R.; DiBlasio, E. A.; Kovacic, S.; Grant, K. L.; Schendel, P. F.; McCoy, J. M. A Thioredoxin Gene Fusion Expression System that Circumvents Inclusion Body Formation in the E. coli Cytoplasm. Biol Technology 1993,11,187193. Lilius, G.; Persson, M.; Bulow, L.; Mosbach, K. Metal Affinity Precipitation of Proteins Carrying Genetically Attached Polyhistidine Affinity Tails. Eur. J . Biochem. 1991, 198, 499504. Lunn, C. A.; Kathju, S.; Wallace, B. J.; Kushner, S. R.; Pigiet, V. Amplification and Purification of Plasmid-Encoded Thioredoxin from Escherichia coli K12. J . Biol. Chem. 1984, 259, 10469- 10474. Ma, N. T. Expression and Purification of a-Human Atrial Natriuretic Peptide in Escherichia coli by Fusion with Erwinia L-Asparaginase. Ph.D. Thesis, University of Oklahoma, Norman, OK, 1993. Maack, T.; Camargo, M. J. F.; Kleinert, H. D.; Laragh, J. H.; Atlas, S. A. Atrial Natriuretic Factor: Structure and Functional Properties. Kidney Int. 1985, 27, 607-615. Pramik, M. J. Scientists Take on Challenge of Developing Peptides as Effective Drugs. Gen. Eng. News 1992, 12, 5,lO. Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning, a Laboratory Manual, 2nd ed.; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY,1989. Schaegger, H.; Von Jagow, G. Tricine-Sodium Dodecyl SulfatePolyacrylamide Gel Electrophoresis for the Separation of Proteins in the Range from 1 to 100 kDa. Anal. Biochem. 1987,166, 368-379. Skerra, A,; Hitzinger, I.; Pluckthun, A. The Functional Expression of Antibody F, Fragments in Escherichia coli: Improved Vectors and a Generally Applicable Purification Technique. Biol Technology 1991, 9, 273-278. Van Dyke, M. W.; Sirito, M.; Sawadogo, M. Single-Step Purification of Bacterially Expressed Polypeptides Containing an Oligo-Histidine Domain. Gene 1992, 111, 99- 104. Watson, J. D. Molecular Biology of the Gene, 2nd ed.; W. A. Benjamin, Inc.: New York, 1970; p 85. Wetzel, R.; Goeddel, D. V. Synthesis of Polypeptides by Recombinant DNA Methods. Peptides 1983,5, 1-64. Accepted December 20, 1994.@ BP9401054 @
Abstract published in Advance ACS Abstracts, March 1,1995.
