Protein Expression and Purification 23, 270–281 (2001)doi:10.1006/prep.2001.1507, available online at http://www.idealibrary.com on

An Enhanced and Scalable Process for the Purification ofSIV Gag-Specific MHC Tetramer

Karen M. Grimm, Wendy L. Trigona, Gwendolyn J. Heidecker, Joseph G. Joyce, Tong-Ming Fu,John W. Shiver, Paul M. Keller, and James C. Cook

or

din (or streptavidin). Each monomeric unit is composedof a class I MHC a chain, b-2-microglobulin and anantigenic peptide. The peptide, which is an epitope ofthe infecting virus or vaccine, and the MHC a chain

Department of Virus & Cell Biology, Merck Research Laborat

Received April 12, 2001, and in revised form July 3, 2001

A recently developed method for the identificationand quantitation of antigen-specific T lymphocytesinvolves the use of complexes of biotinylated majorhistocompatibility complex (MHC) and avidin conju-gated to a fluorescent reporter group. This complex,dubbed the “tetramer,” binds to antigen-specific T lym-phocytes in vitro, which can then be sorted andcounted by fluorescence-activated flow cytometry tomeasure immune response. Our research has focusedon developing the purification process for preparingtetramer reagent. Our goal was to reengineer a pub-lished lab-scale purification process to reduce thenumber of processing steps and to make the processscalable. In our reengineered process, recombinantMHC a chain is isolated from Escherichia coli as inclu-sion bodies by tangential flow filtration. The purifiedMHC a chain is refolded with b-2-microglobulin andthe target peptide antigen to form the class I MHC. Theresulting MHC is purified by hydrophobic interactionchromatography (HIC) and biotinylated enzymatically,and the biotinylated MHC is purified by a second HICstep. The tetramer is prepared by mixing biotinylatedMHC with an avidin–fluorophore conjugate. The tetra-mer is further purified to remove any excess MHC oravidin components. Analysis by flow cytometry con-firmed that the tetramers generated by this new proc-

ess gave bright staining and specific binding to CD31/CD81 cells of vaccinated monkeys and led to resultsthat were equivalent to those generated with tetramerproduced by the original process. q 2001 Academic Press

270

ies, West Point, Pennsylvania 19486

Cytotoxic T lymphocytes (CTLs)1 perform a vital func-tion in controlling viral spread during infection (1).Quantitation of circulating antigen-specific CTLs is auseful measure of an organism’s cellular immune re-sponse to infection or vaccination (2). Recently, a newtechnology for directly measuring antigen-specificCTLs in vitro was developed (3). This technology uses amultimeric complex of biotinylated recombinant majorhistocompatibility complex (MHCs) and avidin (orstreptavidin) conjugated to a fluorescent reporter mole-cule. This multimeric complex, referred to as “tetramerreagent,” is mixed with a sample of whole blood,allowing the tetramer to bind to the surface of activatedCTLs in an antigen-specific manner. Fluorescence-acti-vated cell sorting is then used to separate and quantifythe CTLs that have been decorated with tetramer. Thismethod has garnered much interest due to its direct-ness and specificity and has been cited in more than310 papers since its publication (4–7).

The tetramer consists of four biotinylated monomericunits noncovalently linked by fluorescently tagged avi-

1 Abbreviations used: SIV, simian immunodeficiency virus; MHC,major histocompatibility complex; CTL, cytotoxic T lymphocyte;FACS, fluorescence-activated cell sorting; HIC, hydrophobic interac-tion chromatography; DTT, dithiothreitol; Mes, 2-(N-morpholino)-ethanesulfonic acid; EDTA, ethylenediaminetetraacetic acid; NaCl,sodium chloride; ATP, adenosine 58-triphosphate; PBS, phosphatebuffered saline; Bicine, N,N-bis(2-hydroxyethyl)glycine; SEC, size-exclusion chromatography; PDA, photodiode array; HTLV, human Tlymphotropic virus type 1; EBV, Epstein–Barr virus; CMV, cytomega-lovirus; HIV, human immunodeficiency virus; BSP, Bir A substratepeptide; IPTG, isopropyl b-D-thiogalactoside; R-PE, R-phycoerythrin;BCA, bicinchoninic acid.

1046-5928/01 $35.00Copyright q 2001 by Academic Press

All rights of reproduction in any form reserved.

-

PURIFICATION OF SIV GAG

determine the specificity of the complex. Together theybind to a peptide-specific receptor on a cytotoxic T cell.The fluorescent tag on the tetramer allows detection ofbinding by fluorescence-activated cell sorting (FACS).Recent studies have used tetramers to detect and char-acterize antigen-specific T cells in patients infectedwith human T lymphotropic virus type 1 (HTLV) (8, 9),Epstein–Barr virus (EBV) (10, 11), hepatitis B and Cviruses (12–15), and cytomegalovirus (CMV) (16, 17).Tetramers have also been used to identify tumor-specific CTLs in cancer patients (18, 19). Perhaps thearea of greatest utility has been in the study of HIV.Tetramers have been used to study the immune re-sponse and disease progression in SIV-infected mon-keys (20, 21) and HIV-infected humans (3, 22–24) byquantifying epitope-specific CD81 T cells. Most re-cently, tetramer technology has been employed to moni-tor vaccine effectiveness in the early development ofHIV vaccine candidates (25, 26).

The original purification method for the tetramercomplex, developed by Altman et al. (3), is a multistepprocess, best suited for small laboratory-scale prepara-tions. Our goal was to develop a process which couldbe easily scaled-up to produce a large quantity of tetra-mer in a single run. Using the Altman process as astarting point, we have reengineered the lab-scale proc-ess. A high-pressure homogenizer was used in place ofsonication for cell lysis. Tangential flow microfiltration/diafiltration replaced the multiple rounds of centrifuga-tion, sonication, and resuspension used to isolate inclu-sion bodies. Hydrophobic interaction chromatography(HIC) was used both to purify the refolded MHC andto desalt the enzymatically biotinylated MHC. Thisyielded the biotinylated MHC in a form ready to formu-late with the avidin–fluorophore conjugate. This newprocess eliminated the size-exclusion chromatography(SEC) steps which are often obstacles to scale-up. Em-ploying HIC eliminated many of the concentration andbuffer-exchange steps that caused losses without thecompensation of purification. The resulting process hasfewer steps, resulting in a less time-consuming and

simpler protocol. Flow cytometry analysis showed thatthe tetramers generated by the new method exhibited bright staining and bound specifically to CD31/CD81cells of vaccinated monkeys and gave results indistin-guishable from those prepared by Altman’s process.

MATERIALS AND METHODS

Reagents

Lysozyme, DNase I, PefablocSC, pepstatin A, andleupeptin were from Boehringer Mannheim. Benzonasewas from EM Industries (Darmstadt, Germany) andwas 2.5 3 105 Units/mL and 0.23 mg/mL protein. b-2-Microglobulin was from Bios Pacific, Inc. (Emeryville,

SPECIFIC MHC TETRAMER 271

CA). Synthetic peptide, p11c, c–m (amino acid sequenceCTPYDINQM), was from Genosys, Inc. (The Wood-lands, TX). Bir A was from Avidity, Inc. (Denver, CO)and was 5000 Units/mg and 3.0 mg protein/mL. Theexpression vector pHN11 containing the MAMU-A*01heavy chain and Bir A substrate peptide (BSP) was agift from D. Barouch of Harvard Medical School. Allchemicals used in the preparation of buffers were re-agent grade or better.

Processing Buffers

The buffers used were as follows: biotinylation buffer(original process) was 100 mM Tris–HCl, pH 7.5, 0.2M NaCl, 5 mM MgCl2; biotinylation buffer A (suppliedby Avidity, Inc.) was 0.5 M bicine, pH 8.3; biotinylationbuffer B (supplied by Avidity, Inc.) was 100 mM ATP,100 mM magnesium acetate, 400 mM biotin (freshlythawed); diafiltration buffer was 20 mM Tris–HCl, pH8.0; DNase I stock solution was 2 mg/mL DNase I in 75mM NaCl, 50% glycerol; folding buffer was 400 mMarginine, 100 mM Tris–HCl, pH 8.3, 2 mM EDTA; injec-tion buffer was 3 M guanidine HCl, 10 mM sodiumacetate, 10 mM EDTA, pH 4.2; microfluidization lysisbuffer was 50 mM Tris–HCl, pH 8.0, 2mM MgCl2, 10mM DTT; Mono Q column buffer A was 20 mM Tris–HCl, pH 8.0; Mono Q column buffer B, 20 mM Tris–HCl,0.5 M NaCl, pH 8.0; phosphate-buffered saline (PBS)was 150 mM sodium chloride, 6.3 mM sodium phos-phate, pH 7.2; PBS/protease inhibitor cocktail was 2mM EDTA, 20 mM PefablocSC, 1 mM pepstatin, 2.1 mMleupeptin in phosphate-buffered saline; PD-10 elutionbuffer was 20 mM Tris–HCl, 0.1 M NaCl, pH 8.0;POROS HP 20 column buffer A was 1.0 M Na2SO4, 20mM Tris–HCl, pH 8.0; POROS HP 20 column buffer B,20 mM Tris–HCl, pH 8.0, 2% isopropyl alcohol; S-300column buffer was 20 mM Tris–HCl, 150 mM NaCl, pH8.0; sonication lysis buffer was 50 mM Tris–HCl pH8.0, 25% sucrose, 1.0% Triton X-100, 1 mM EDTA, 0.1%sodium azide, 11 mM DTT, 1 mg/mL lysozyme, 5 mMMgCl2; Tris wash buffer was 50 mM Tris–HCl, 100 mMNaCl, 1 mM EDTA, 1 mM DTT, pH 8.0; Triton X-100wash buffer was 0.5% Triton X-100, 50 mM Tris–HCl,100 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 8.0; andurea buffer was 25 mM sodium Mes, pH 6.0, 8 M urea,10 mM EDTA, 0.1 mM DTT. Fresh DTT was added tothe appropriate buffers immediately prior to use.

All “original process” chromatography steps were car-ried out on a Waters 2690 HPLC system. Protein ab-sorbance was monitored at 215 and 280 nm using a

Waters 996 PDA detector. All “enhanced process” chro-matography steps were carried out on a Waters 650HPLC system unless otherwise noted. Protein was de-tected by UV absorbance at 280 nm using a linear UVIS200 detector.

mM NaCl, 5 mM MgCl , 400 mM biotin, 5 mM ATP, 200

272 GRIMM

Cloning and Expression

Growth and Harvest of Escherichia coli

The heavy chain of rhesus monkey MHC class I mole-cule Mamu-A*01 and the BSP were cloned into theexpression vector pHN11 and transformed into E. colistrain XA90 (3, 27, 28). In the expression vectorpHN11, the cloned genes were under the control of thetac promoter and rrnBT1T2 transcriphon terminator(29). Expression was induced using 0.4 mM IPTG dur-ing early log phase. Cells were harvested 3 h after in-duction by centrifugation at 7100g for 10 min at roomtemperature and stored at 2708C.

The original process based on (3) and personal com-munication from J. Altman (Emory University Schoolof Medicine, Atlanta, GA) is described below.

Original Process

Inclusion body isolation by sonication/centrifuga-tion. Frozen E. coli cells expressing Mamu-A*01 werethawed from 2708C at room temperature and resus-pended in sonication lysis buffer at a concentration of1.3 g wet cell weight/60 mL. DNase stock solution wasadded to the cell suspension to a final concentration of1.7% (v/v). The 60-mL cell suspension was placed in a125-mL plastic polymethylpentene beaker and dis-rupted by sonication in the following way. Cells wereplaced in an ice bath with stirring and sonicated for atotal of 1.5 min using an ultrasonic processor (HeatSystems-Ultrasonics, Inc., Model W-385, Probe ModelC3). The sonicator was set to a 1-s cycle time and 50%duty cycle on power setting 4. The lysate was trans-ferred to two 50-mL polycarbonate centrifuge tubes(Nalgene, No. 118-0050) and centrifuged at 11,000g for10 min at 48C (brake on). The final pellet was resus-pended in 40-mL Triton X-100 wash buffer, sonicatedagain, and centrifuged as before. This process of resus-pension, sonication, and centrifugation was repeatedtwo additional times. After the last sonication/centrifu-gation step, the pellet was resuspended in Tris washbuffer and centrifuged as before. The pellet was resus-pended in 0.4 mL sterile H2O and then solubilized in5.0 mL urea buffer. The insoluble material was pelletedby centrifugation as before, and the supernatant wasused in the refolding reaction described below.

Refolding of Mamu-A*01, b-2 microglobulin, and pep-tide to make MHC. Refolding of the MHC was initiatedby diluting Mamu-A*01, b-2-microglobulin, and p11 C,C-M peptide into folding buffer prechilled to 108C (27).The volume of folding buffer was determined by themass and volume of Mamu-A*01 isolated from inclusion

bodies. In the refolding reaction mixture, the concentra-tion of Mamu-A*01 ranged from 1 to 3 mM and ureawas kept below 0.3 M. The molar ratio of Mamu-A*01to b-2-microglobulin to peptide was kept constant at

ET AL.

3:2:28. For 236 mL of folding buffer, 7.1 mg of p11C, C-M peptide was dissolved in 0.5 mL H2O and added to thefolding buffer with vigorous stirring using a pipettor. Aquantity of 0.236 mmol of Mamu-A*01 and 0.47 mmolof b-2 microglobulin were each dissolved in 5.0 mL ofinjection buffer, loaded into sterile syringes (Becton-Dickinson, No. 309603 and 309604), and added to thestirring folding reaction through 26-gauge needles withthe needles positioned near the stirring bar. The foldingreaction was incubated at 108C without stirring for 14–20 h in a tightly capped vessel, and an additional 0.236mol of Mamu-A*01 was added to the folding reactionas previously described. The folding reaction was incu-bated at 108C without stirring for an additional 7–10h, and an additional 0.236 mol of Mamu-A*01 wasadded to the folding reaction as previously described.The folding reaction was incubated at 108C for 14–20h without stirring. Final concentrations of folding reac-tion components were as follows: 3.0 mM Mamu-A*01,2.0 mM b-2 microglobulin, and 27.7 mM (30.1 mg/L)p11C, C-M (www.emory.edu/WHSC/TETRAMER/protocol.html).

Concentration and diafiltration of folding reactionproduct. The folding reaction product was concen-trated from 260 to 50 mL and exchanged into dia-filtration buffer using a 500-mL stirred cell (Amicon,Biomax-30 membrane, 90 mm) with constant-volumediafiltration at 48C. The diafiltered retentate was con-centrated in the stirred cell under N2 at 16 psi to 42mL and then was further concentrated to 6 mL withUltrafree-15 centrifugal filter units (2000g, Biomax-30membrane, Millipore).

Size-exclusion chromatography of concentrated fold-ing reaction. The refolded MHC was purified by size-exclusion chromatography on a Sephacryl S300 26/60column (2.6 cm i.d. by 60 cm, Pharmacia). The concen-trated folding reaction was loaded onto the columnthrough a 10-mL sample loop connected to a manualinjector (Rheodyne), and the column was run at a flowrate of 2.0 mL/min with S-300 column buffer. Fractionswere collected at a rate of 1 fraction/4 min in 13 3 100-mm glass test tubes. Fractions exhibiting both Mamu-A*01 and b-2-microglobulin bands by SDS–PAGE werepooled, concentrated 10-fold, and exchanged into bio-tinylation buffer using Ultrafree-15 centrifugal filterunits (2000g).

Biotinylation of the MHC complex. Purified MHCmolecules were enzymatically biotinylated by incuba-tion with 225,000 Units of Bir A for 20 h at room temper-ature with the following components: 100 mM Tris, 200

2mM PefablocSC, 1 mg/mL leupeptin, 1 mM pepstatin,225,000 units Bir A, and 33 mM MHC complex. Thebiotinylation reaction had a ratio of 6.0 mg of MHCto 225,00 units Bir A. The biotinylation reaction was

-

PURIFICATION OF SIV GAG

clarified by centrifugation (2000g), divided into twoequal portions, and desalted using two PD-10 columns(Pharmacia) preequilibrated with PD-10 elution buffer.The desalted MHC was concentrated by an Ultrafree-4 centrifugal filter unit (2000g) to 2 mL.

Anion-exchange chromatography of biotinylatedMHC. Biotinylated MHC was purified by anion-exchange chromatography (3) on a Mono Q HR 5/5 col-umn (0.5 cm i.d. by 5 cm, Pharmacia). The concentratedbiotinylated MHC molecules were loaded onto the col-umn through a 5-mL sample loop connected to a manualinjector (Rheodyne). The column was washed with 10column vol of Mono Q column buffer A at 1 mL/min. Thecolumn was eluted at 1 mL/min with a linear gradient of100% Mono Q column buffer A to 60% Mono Q bufferA/40% Mono Q buffer B. Total gradient volume was 10column vol. Following the gradient, the column waswashed with 10 column vol of Mono Q column bufferB at 1 mL/min. Fractions were collected at a rate of 1fraction/min in 1.5-mL polypropylene tubes. PurifiedMHCs were concentrated and exchanged into PBS/pro-tease inhibitor cocktail (3) using an Ultrafree-4 centrif-ugal filter unit (2000g). The purified product was storedin 200-mg aliquots in 1.5-mL polypropylene screw-captubes. The aliquots were snap-frozen in liquid N2 priorto storage at 2708C.

Enhanced Process

Cell thawing and lysis. A quantity of 1.6 g of frozenE. coli cells expressing Mamu-A*01 were removed from2708C and thawed for 30 min at room temperature.Cells were resuspended in microfluidization lysisbuffer, and Benzonase was added to a final concentra-tion of 300 Units per gram of wet cell weight. Cells werelysed by five passes though a microfluidizer (Micro-fluidics Corp. Model 110S) at 18,000 psi. The resultingcell lysate was stirred for 2 h at 48C.

Isolation of inclusion bodies by tangential flow micro-filtration. Inclusion bodies were isolated by tangentialflow microfiltration as follows. The microfilter was a0.45-mm pore size hollow fiber filter cartridge of 16-cm2

surface area (A/G Technologies No. CFP-4-E-MM06A).Circulation of the process material and flow of the per-meate out of the system were controlled by two variable-speed peristaltic pumps. Pharmed tubing, size 16, wasused for the recirculation pump and silicone tubing, size14 (Masterflex—platinum), was used for the permeatepump. The recirculation flow rate was 84 mL/min andthe transmembrane pressure was 1.8 psi. The cell lysatewas diafiltered against 1 vol of Triton X-100 washbuffer, concentrated fivefold, and diafiltered against 2

vol each of a series of buffers (Triton X-100 wash buffer,Tris wash buffer, and Milli-Q H2O). The retentate, con-taining the washed inclusion bodies, was concentratedtwofold. Solid urea and a stock solution composed of

SPECIFIC MHC TETRAMER 273

250 mM Mes, pH 6.0, and 0.5 M EDTA, pH 8.0, wereadded to the concentrated retentate to final concentra-tions of 8 M urea, 25 mM Mes, and 10 mM EDTA.This solubilized inclusion body preparation was filteredusing a 25-mm, 0.2-mm cellulose acetate sterile syringefilter (Corning). Protein concentration was determinedby absorbance at 280 nm using a protein concentrationof 0.41 mg/mL for one absorbance unit in a 1-cm cuvette[extinction coefficient 5 66,814 (M 3 cm)21].

Refolding of Mamu-A*01, b-2 microglobulin, and pep-tide. Refolding of the MHC was done according to theoriginal process described previously. Final concentra-tions of folding reaction components were as follows:1.2 mM Mamu-A*01, 0.8 mM b-2-microglobulin, and10.8 mM (11.7 mg/L) p11C, C-M.

HIC of folding reaction. The refolded MHC was iso-lated from the folding reaction by column chromatogra-phy on a POROS HP 20 hydrophobic interaction chro-matography resin (PerSeptive Biosystems, No. 1-4429-05) packed in a Pharmacia XK16 column (bed diam.1.6 cm i.d. by 7.5 cm). The column was preequilibratedwith POROS HP 20 column buffer A. Prior to loadingon the HIC column, the folding reaction was adjustedto a final concentration of 1.0 M Na2SO4, pH 8.0, bythe gradual addition of 1.5 M Na2SO4, 20 mM Tris–HCl,pH 8.0, stock solution and filtered using a 0.22-mmcellulose acetate filter (Corning). The folding reactionwas loaded on the column at 15 mL/min, and the columnwas washed with 2.6 column vol of POROS HP 20 col-umn buffer A at 3.0 mL/min. The column was elutedat 3.0 mL/min with a step gradient of 0 to 100% POROSHP 20 column buffer B. Total volume of the gradientwas 4.2 column vol. Fractions were collected at a rate of1 fraction/min in 12 3 75-mm glass test tubes. Fractionsexhibiting both Mamu-A*01 and b-2-microglobulinbands by SDS–PAGE were pooled.

Biotinylation of the MHC. MHC in the HIC productwas enzymatically biotinylated by incubation with BirA for 20 h at room temperature. The final reactionmixture contained, by volume, eight parts of the MHC(final concentration 240 mg/mL), one part each of aviditybiotinylation buffer A and B, and 150,000 Units of BirA. Protease inhibitors were added prior to addition ofthe enzyme as described previously.

HIC of biotinylated MHC. The biotinylated MHCwas desalted and purified by HIC on a 1.0 cm i.d. by10.3-cm-length bed of POROS HP 20 hydrophobic inter-action chromatography resin (PerSeptive Biosystems,No. 1-4429-05) packed in a Pharmacia HR 10/10 col-umn. The column was preequilibrated with POROS HP20 column buffer A. The biotinylation reaction was ad-

justed to a final concentration of 1.0 M Na2SO4, pH8.0, by the gradual addition of 1.5 M Na2SO4, 20 mMTris–HCl, pH 8.0, stock solution and filtered using asterile 25-mm, 0.2-mm cellulose acetate filter prior to

FIG. 1. SDS–PAGE of retains of Mamu-A*01 during inclusion body isolation. A comparison of lysis methods, sonication (A) vs microfluidiza-

sorbance at 280 nm, extinction coefficient 5 66,814

tion (B), in the microfiltration step is shown. Mamu-A*01 migratesweight standards; lane 3, E. coli cell lysate; lanes 4–9, permeates frobodies—no urea; lane 11, urea wash of microfiltration cartridge; lasupernatant 1 urea.

chromatography. The biotinylation reaction was loadedon the column at 5 mL/min, and the column was washedwith 2.5 column vol of POROS HP 20 column buffer Aat 2.0 mL/min. The column was eluted at 2.0 mL/minwith a step gradient of 0 to 100% POROS HP 20 columnbuffer B. The volume of the gradient was 3.3 columnvol. Fractions were collected at a rate of 1 fraction/minin 12 3 75-mm glass test tubes. Peak fractions (A280)exhibiting both Mamu-A*01 and b-2-microglobulinbands by SDS–PAGE were pooled. Protein concentra-tion was determined by absorbance at 280 nm using aprotein concentration of 0.41 mg/mL for one absorbanceunit in a 1 cm cuvette [extinction coefficient 5 66,814

(M 3 cm)21].

Protease inhibitors were added to the purified MHCto the following concentrations: 2 mM EDTA, 200 mMPefablocSC, 1 mM pepstatin, 2.1 mM leupeptin. Thepurified biotinylated MHC molecules were stored in

at ,36 kDa and lysozyme migrates at ,14 kDa. Lane 1, molecularm series of diafiltrations; lane 10, retentate of Mamu-A*01 inclusionne 12, urea-insoluble material; lane 14, Mamu-A*01 inclusion body

200-mg aliquots in 1.5-mL polypropylene Eppendorftubes and snap-frozen in liquid N2 prior to storage at2708C.

Formulation of MHC Tetramer Complex

Tetramers were formulated by mixing the biotinyl-ated MHC molecules with ExtrAvidin R-phycoerythrin(R-PE) conjugate (Sigma) at a 4:1 molar ratio (3). Themolarity of MHC was calculated from the concentrationof the final monomeric product [determined by ab-

274 GRIMM ET AL.

(M 3 cm)21]. The molarity of the ExtrAvidin was sup-plied by the manufacturer. After a 1.5-h incubation at48C, tetramers were mixed with biotin–agarose beads(Sigma) for 1.5 h at 48C. The mixture was centrifugedat 16,000g at room temperature in a microcentrifuge

-

PURIFICATION OF SIV GAG

for 30 s and the supernatant was retained. The tetra-mers were purified by size-exclusion chromatography

on a Pro-TSK G3000SWxl column (7.8 3 300 mm,

FIG. 2. (A) Typical POROS HP 20 chromatogram of the folding reactio(B) SDS–PAGE of HIC product fractions. Mamu-A*01 migrates at ,3reaction; lane 2, folding reaction 1 1.0 M Na2SO4; lane 3, filtered follanes 5 and 6, flowthrough fractions; lanes 8–10, MHC fractions.

SPECIFIC MHC TETRAMER 275

ExtrAvidin R-PE conjugate bands by SDS–PAGE werepooled for use as tetramer reagent in staining assays.

Supelco). The column was run at a flow rate of 0.8 mL/ SDS–PAGE and Protein Assaysmin with PBS. Fractions were collected at a rate of 1fraction/min in 1.5-mL polypropylene microcentrifuge SDS–PAGE of the samples was run in the following

manner. An equal volume of Tris–glycine SDS 23 sam-tubes. Fractions eluting at the estimated molecularweight of the tetramer and exhibiting both MHC and ple buffer (Novex) 1 200 mM DTT was added to each

n. The small peak eluting at 22 min represents free b-2-microglobulin.6 kDa and b-2-microglobulin migrates at ,12 kDa. Lane 1, foldingding reaction (column charge); lane 4, molecular weight standards;

Volume [Protein] Total protein Recovery Volume [Protein] Total protein Recoverya a

00

solubilized inclusion bodies divided by wet cell mass of

Sample (mL) (mg/mL) (mg

Folding reaction 267.4 0.12 32Refolded MHC complex 4.0 1.5 6.Purified biotinylated MHC complex 2.0 1.5 3.

a Determined by BCA protein assay.

sample. The samples were heated to 1008C for 15 minusing a dry bath (Fisher). Samples were not loaded atequal protein concentrations. All samples were loadedundilute except Mamu-A*01 inclusion body retentateswhich were diluted so Mamu-A*01 concentration wouldbe approximately equal between gels. The samples wereapplied to precast 4–20% acrylamide Tris–glycine gels(Novex). Application volume was 20 mL per lane. Elec-trophoresis was performed for 1 h at a constant currentof 30 mA per gel in a Xcell II Mini-Cell apparatus(Novex). Gels were fixed in 45% methanol, 10% aceticacid for 20 min and then stained with Fast StainCoomassie (Zoion Biotech, Cat. No. 23225) according tothe manufacturer’s instructions. BCA protein assayswere performed using the Pierce BCA protein assay kitand bovine serum albumin (Pierce) as a standard.

Tetramer Staining

Rhesus animals used in this study were typed asMamu-A*01 positive as determined by a PCR-basedMHC class I typing assay as previously described (30).The monkey used to positively stain for p11C, C-M tet-ramer was previously immunized with a DNA-basedvaccine encoding the gag region of SIV. The naive ani-mal was not vaccinated and represented a Mamu-A*01naive control. The procedure for tetramer staining offresh whole blood was adopted from a published proce-dure (21). Briefly, PE-conjugated tetramer Mamu-A*01/p11C, C-M complex in conjunction with fluorescent-con-jugated surface marker monoclonal antibodies wereused to stain 100 ml of fresh whole blood sample or about5 3 105 cells from bulk CTL cultures. The monoclonalantibodies used in this study were anti-CD8ab-PerCP(Becton-Dickinson, San Jose, CA) and anti-rhesus mon-key CD3-APC (Biosource). The red blood cells werelysed using FACS lysing solution (Becton-Dickinson)according to manufacturer’s recommendations, and thecells were washed once with cold PBS and fixed with

276 GRIMM ET AL.

TABLE 1

Purification Table for MHC Complexes

Original process Enhanced process

1% formaldehyde in PBS. The cells were stored at 48Cbefore flow cytometric analysis on FACScalibur(Becton-Dickinson). For each sample, 30,000 gatedCD31/CD81 lymphocyte events were collected and an-alyzed on CellQuest program (Becton-Dickinson).

) (%) (mL) (mg/mL) (mg) (%)

100 695.6 0.050 35 10019 21.0 0.36 7.6 22

9 19.3 0.21 4.1 12

RESULTS AND DISCUSSION

Cell Lysis

In the original purification process, the cells werelysed by repeated sonication steps. Protein concentra-tion of the cell lysate after subsequent sonications indi-cated that at least three sonications were necessary tocompletely disrupt the cells. The method of repeatedlysis was tedious and not suitable for processing a largequantity of cells. In addition, sonication can be difficultto do reproducibly. The inclusion of lysozyme in thisprocedure created an additional problem. SDS–PAGEanalysis of the final inclusion bodies showed that lyso-zyme was a significant contaminant (Fig. 1A). In theenhanced process, microfluidization was used to disruptthe cells. Lysis by microfluidization is significantlymore scalable and less labor intensive. Initially fivepasses through the microfluidizer were used to breakup cell debris, but later experiments showed that onlyone pass through the microfluidizer was necessary forcomplete cell lysis as determined by inclusion body re-covery. An advantage of microfluidization is that lyso-zyme is not required for cell lysis; thus, the source ofa major contaminant is eliminated (Fig. 1B). There wassome evidence that lysis by microfluidization resultedin fragmentation of inclusion bodies (see below).

Isolation of Inclusion Bodies

The original process required repeated wash/centrif-ugation steps to purify the inclusion bodies. This wasa tedious method that was not readily scalable. Theenhanced process used cross-flow microfiltration/diafil-tration which is more amenable to scale-up and whichalso led to higher yields and purer product. Using a0.45-mm membrane cartridge, we achieved a twofoldincrease in recovery of inclusion bodies from 1.8 to 3.4%(based on the absorbance at 280 nm of the solution of

the starting material). Use of hollow fibers with a largerpore size (0.65 mm) resulted in poor recovery.

A small amount of product was lost as a result ofinclusion bodies binding to the hollow fibers when themicrofluidized lysate was used. Interestingly, this could

-

PURIFICATION OF SIV GAG

not be recovered by simply rinsing the hollow fiberswith 8 M urea, but could only be recovered when 8 Murea was forced through the membrane. Product losswas less pronounced when inclusion bodies from soni-cated lysate were used. This suggests that cell lysis withthe microfluidizer produced inclusion body fragmentswhich penetrated into the hollow fibers (Figs. 1A and1B). We did not use the product recovered by urea washin the folding reaction, nor did we include it in our

FIG. 3. (A) Typical POROS HP 20 chromatogram of biotinylated MHCreaction; lane 2, molecular weight markers; lane 3, biotinylation reaccharge); lanes 5–9, flowthrough fractions; lanes 12–15, biotinylated M

SPECIFIC MHC TETRAMER 277

MHC because membrane fouling and adsorption lossesmade for a low-yielding and time-consuming step. Wealso wanted to replace the size-exclusion chromatogra-phy because it is not easily scaled-up and resulted ina dilute product which required concentration prior tothe next step in the process. We replaced bothconcentration/diafiltration and SEC steps with a singleHIC step.

We identified POROS HP20 with a high phenyl ligand

recovery calculations. density as the best HIC resin for our application based

on efficiency and recovery. HIC resins with a variety ofHIC of Folding Reaction ligands were screened, but those with a high substitu-

tion of phenyl ligands required the lowest salt concen-One goal in developing the enhanced process wasto eliminate concentration/diafiltration of the refolded tration for adsorption and yielded the greatest recovery.

. (B) SDS–PAGE of biotinylated MHC fractions. Lane 1, biotinylationtion 1 1.0 M Na2SO4; lane 4, filtered biotinylation reaction (columnHC fractions.

solutions prepared according to the original process

FIG. 4. Overview of original (A) and

Various Pharmacia HIC resins were evaluated but thePOROS resin allowed for higher flow rates without col-umn compression and resulted in better resolution andrecovery. In order to cause the product to bind to theHIC resin it was necessary to add Na2SO4 to a finalconcentration of 1.0 M. Concentrations above 1.0 Mcaused some precipitation of the MHC. Concentrationsbelow 1.0 M showed decreased adsorption of the MHCto the column. A concentrated solution of sodium sulfatewas added to the refolding reaction mixture graduallywith stirring in an effort to minimize precipitation. Theresulting solution was filtered to remove particulatedand precipitated protein prior to loading on the HICcolumn. Figure 2B shows that some minor contami-nants were removed after filtering the folding reaction11.0 M Na2SO4 (lane 2 vs lane 3).

Potassium phosphate was evaluated in place ofNa2SO4, but it interfered with subsequent analysis of

the fractions by SDS–PAGE and therefore was notused.

Figure 2A shows a typical HIC chromatogram of thefolding reaction product. Elution with a step Na2SO4

enhanced (B) purification processes.

gradient caused the MHC to elute prior to free b-2-microglobulin. SDS–PAGE shows the MHC peak frac-tions which were pooled for biotinylation (Fig. 2B). Ad-dition of isopropyl alcohol to the elution buffer to 2%(vol) allowed for elution of a more concentrated productand eliminated the need to concentrate the productafter chromatography. Also, because the HIC elutionbuffer did not contain NaCl, it did not interfere with theactivity of the Bir A enzyme. This made buffer exchangeprior to biotinylation no longer necessary. In addition,recovery of the product increased from 19 to 22% forthe new process (Table 1).

Biotinylation Reaction and HIC of BiotinylatedComplexes

We compared biotinylation efficiency of buffers and

278 GRIMM ET AL.

with buffers supplied with the biotinylation kit pro-vided by the manufacturer (Avidity). A streptavidin gelshift assay showed that the percentage of MHC biotinyl-ated was comparable for the two methods (not shown).

mD

be critical if such an assay was to be validated for use

FIG. 5. Enhanced and original tetramer staining of whole blood sarhesus monkey (C and D). The percentage of tetramer-positive CD3/C

The biotinylated MHC molecules were purified byHIC using the same HIC resin used for purification ofthe refolding reaction mixture. A smaller column wasused to minimize losses due to nonspecific adsorptionand to maximize the concentration of the eluted prod-uct. Figure 3A represents a typical chromatogram forthe biotinylated MHC. SDS–PAGE shows the biotinyl-ated MHC peak fractions which were pooled as the finalproduct (Fig. 3B). Purification of the biotinylated MHCby HIC replaced buffer-exchange (by SEC) and anion-exchange chromatography steps. In addition, buffer ex-change and concentration of the final product was nolonger necessary.

Figures 4A and 4B give overviews of the two processes

while Table 1 compares the step recoveries for bothmethods. The figures show that reengineering of thepurification process reduced the number of steps from11 to 6, increased the overall recovery from 9 to 12%,and made the process scalable.

ples from an SIV-vaccinated rhesus monkey (A and B) and a naive8 T cells in each population is shown.

Formulation of MHC Tetramer Complex andTetramer Staining

Tetramers were formulated based on a modificationof Altman’s procedure (3). Incubation of the tetramerswith the biotin agarose beads was done in order toremove excess avidin R-PE conjugate and any com-plexes with unoccupied biotin-binding sites, i.e., mo-nomers, dimers, and trimers. Further purification bySEC removed free MHC (presumably not biotinylated)not linked to avidin R-PE conjugate. The purificationof the tetramer reagent was done to provide a truetetramer rather than a mixture of forms, which would

PURIFICATION OF SIV GAG-SPECIFIC MHC TETRAMER 279

as a clinical assay. Tetramers formulated and purifiedin this way showed very little nonspecific binding ofnaive cells.

Tetramers were used to analyze blood samples from

E

280 GRIMM

an SIV-vaccinated, Mamu-A*01-positive rhesus ma-caque and a naive Mamu-A*01-positive monkey usingflow cytometry. As shown in Fig. 5, the fresh stainingof the whole blood sample of the SIV-vaccinated mon-key with tetramer from the enhanced process showeda population of approximately 2.83% that was tetra-mer positive within CD31/CD81 T lymphocytes. This

was similar to results we obtained when the tetramer from the original process was used (2.28%). Both tet-ramers exhibited a low percentage of tetramer-posi-tive cell staining in the naive Mamu-A*01-positivemonkey.

22. Ogg, G. S., Jin, X., Bonhoeffer, S., Dunbar, P. R., Nowak, M. A.,

ACKNOWLEDGMENTS

The authors thank D. Barouch for expression vector pHN11 andJ. D. Altman for advice and assistance.

REFERENCES

1. Delves, P. J., and Roitt, I. M. (2000) The immune system. N.Engl. J. Med. 343, 37–49.

2. Doherty, P. C., and Christensen, J. P. (2000) Accessing complexity:The dynamics of virus-specific T cell responses. Annu. Rev. Immu-nol. 18, 561–592.

3. Altman, J. D., Moss, P. A. H., Goulder, P. J. R., Barouch, D. H.,McHeyzer Williams, M. G., Bell, J. I., McMichael, A. J., andDavis, M. M. (1996) Phenotypic analysis of antigen-specific Tlymphocytes. Science 274, 94–96.

4. Fox, C. (2000) A new place in human immunology. Scientist14, 21–22.

5. McMichael, A. J., and O’Callaghan, C. A. (1998) A new look atT cells. J. Exp. Med. 187, 1367–1371.

6. Ogg, G. S., and McMichael, A. J. (1998) HLA-peptide tetramericcomplexes. Curr. Opin. Immunol. 10, 393–396.

7. Schwartz, R. S. (1998) Direct visualization of antigen-specificcytotoxic T cells—A new insight into immune defenses. N. Engl.J. Med. 339, 1076–1078.

8. Bieganowska, K., Hollsberg, P., Buckle, G. J., Lim, D. G., Greten,T. F., Schneck, J., Altman, J. D., Jacobson, S., Ledis, S. L.,Hanchard, B., Chin, J., Morgan, O., Roth, P. A., and Hafler, D. A.(1999) Direct analysis of viral-specific CD81 T cells with solubleHLA-A2/Tax11–19 tetramer complexes in patients with humanT cell lymphotropic virus-associated myelopathy. J. Immunol.162, 1765–1771.

9. Greten, T. F., Slansky, J. E., Kubota, R., Soldan, S. S., Jaffee, E.M., Leist, T. P., Pardoll, D. M., Jacobson, S., and Schneck, J. P.(1998) Direct visualization of antigen-specific T cells: HTLV-1Tax11–19-specific CD8(1) T cells are activated in peripheralblood and accumulate in cerebrospinal fluid from HAM/TSPpatients. Proc. Natl. Acad. Sci. USA 95, 7568–7573.

10. Annels, N. E., Callan, M. F., Tan, L., and Rickinson, A. B. (2000)Changing pattern of dominant TCR usage with maturation ofan EBV-specific cytotoxic T cell response. J. Immunol. 165, 4831–4841.

11. Tussey, L., Speller, S., Gallimore, A., and Vessey, R. (2000) Func-tionally distinct CD81 memory T cell subsets in persistent EBVinfection are differentiated by migratory receptor expression. Eur.J. Immunol. 30, 1823–1829.

12. He, X. S., Rehermann, B., Lopez-Labrador, F. X., Boisvert, J.,

T AL.

Cheung, R., Mumm, J., Wedemeyer, H., Berenguer, M., Wright,T. L., Davis, M. M., and Greenberg, H. B. (1999) Quantitativeanalysis of hepatitis C virus-specific CD8(1) T cells in peripheralblood and liver using peptide-MHC tetramers. Proc. Natl. Acad.Sci. USA 96, 5692–5697.

13. Lechner, F., Gruener, N. H., Urbani, S., Uggeri, J., Santantonio,T., Kammer, A. R., Cerny, A., Phillips, R., Ferrari, C., Pape, G.R., and Klenerman, P. (2000) CD81 T lymphocytes responsesare induced during acute hepatitis C virus infection but are notsustained. Eur. J. Immunol. 30, 2479–2487.

14. Maini, M. K., Boni, C., Lee, C. K., Larrubia, J. R., Reignat, S.,Ogg, G. S., King, A. S., Herberg, J., Gilson, R., Alisa, A., Williams,R., Vergani, D., Naoumov, N. V., Ferrari, C., and Bertoletti, A.(2000) The role of virus-specific CD8(1) cells in liver damageand viral control during persistent hepatitis B virus infection.J. Exp. Med. 191, 1269–1280.

15. Webster, G. J., Reignat, S., Maini, M. K., Whalley, S. A., Ogg, G.S., King, A., Brown, D., Amlot, P. L., Williams, R., Vergani, D.,Dusheiko, G. M., and Bertoletti, (2000) Incubation phase of acutehepatitis B in man: Dynamic of cellular immune mechanisms.Hepatology 32, 1117–1124.

16. Engstrand, M., Tournay, C., Peyrat, M. A., Eriksson, B. M.,Wadstrom, J., Wirgart, B. Z., Romagne, F., Bonneville, M.,Totterman, T. H., and Korsgren, O. (2000) Characterization ofCMVpp65-specific CD81 T lymphocytes using MHC tetramersin kidney transplant patients and healthy participants. Trans-plantation 69, 2243–2250.

17. Singhal, S., Shaw, J. C., Ainsworth, J., Hathaway, M., Gillespie,G. M., Paris, H., Ward, K., Pillay, D., Moss, P. A., and Mutimer,D. J. (2000) Direct visualization and quantitation of cytomegalo-virus-specific CD81 cytotoxic T-lymphocytes in liver transplantpatients. Transplantation 69, 2251–2259.

18. Cerundolo, V. (2000) Use of major histocompatibility complexclass I tetramers to monitor tumor-specific cytotoxic T lympho-cyte response in melanoma patients. Cancer Chemother. Pharma-col. 46, S83–S85.

19. Valmori, D., Dutoit, V., Lienard, D., Rimoldi, D., Pitter, M. J.,Champagne, P., Ellefsen, K., Sahin, U., Speiser, D., Lejeune, F.,Cerottini, J. C., and Romero, P. (2000) Naturally occurring humanlymphocyte antigen-A2 restricted CD81 T-cell response to thecancer testis antigen NY-ESO-1 in melanoma patients. CancerRes. 60, 4499–4506.

20. Egan, M. A., Kuroda, M. J., Voss, G., Schmitz, J. E., Charini, W.A., Lord, C. I., Forman, M. A., and Letvin, N. L. (1999) Use ofmajor histocompatibility complex class I/peptide/beta2M tetra-mers to quantitate CD8(1) cytotoxic T lymphocytes specific fordominant and nondominant viral epitopes in simian-human im-munodeficiency virus-infected rhesus monkeys. J. Virol. 73,5466–5472.

21. Kuroda, M. J., Schmitz, J. E., Charini, W. A., Nickerson, C. E.,Lifton, M. A., Lord, C. I., Forman, M. A., and Letvin, N. L. (1999)Emergence of CTL coincides with clearance of virus during pri-mary simian immunodeficiency virus infection in rhesus mon-keys. J. Immunol. 162, 5127–5133.

Monrad, S., Segal, J. P., Cao, Y., Rowland-Jones, S. L., Cerundolo,V., Hurley, A., Markowitz, M., Ho, D. D., Nixon, D. F., andMcMichael, A. J. (1998) Quantitation of HIV-1-specific cytotoxicT lymphocytes and plasma load of viral RNA. Science 279, 2103–2106.

PURIFICATION OF SIV GAG-SPECIFIC MHC TETRAMER 281

23. Ogg, G. S., Kostense, S., Kelin, M. R., Jurriaans, S., Hamann, Simian immunodeficiency virus (SIV) gag DNA-vaccinated rhe-

D., McMichael, A. J., and Miedema, F. (1999) Longitudinal pheno-typic analysis of human immunodeficiency virus type 1-specificcytotoxic T lymphocytes: Correlation with diesease progression.J. Virol. 73, 9153–9160.

24. Wilson, J. D., Ogg, G. S., Allen, R. L., Davis, C., Shaunak, S.,Downie, J., Dyer, W., Workman, C., Sullivan, S., McMichael, A.J., and Rowland-Jones, S. L. (2000) Direct visualization of HIV-1-specific cytotoxic T lymphocytes during primary infection.AIDS 14, 225–233.

25. Barouch, D. H., Santra, S., Schmitz, J. E., Kuroda, M. J., Fu, T.,Wagner, W., Bilska, M., Craiu, A., Zheng, X. X., Krivulka, G. R.,Beaudry, K., Lifton, M. A., Nickerson, C. E., Trigona, W. L., Punt,K., Freed, D. C., Guan, L., Dubey, S., Casimiro, D., Simon, A.,Davies, M. E., Chastain, M., Strom, T. B., Gelman, R. S.,Montefiori, D. C., Lewis, M. G., Emini, E. A., Shiver, J. W., andLetvin, N. L. (2000) Control of viremia and prevention of clinicalAIDS in rhesus monkeys by cytokine-augmented DNA vaccina-tion. Science 290, 486–492.

26. Egan, M. A., Charini, W. A., Kuroda, M. J., Schmitz, J. E., Racz,P., Tenner-Racz, K., Manson, K., Wyand, M., Lifton, M. A.,Nickerson, C. E., Fu, T., Shiver, J. W., and Letvin, N. L. (2000)

sus monkeys develop secondary cytotoxic T-lymphocyte responsesand control viral replication after pathogenic SIV infection. J.Virol. 74, 7485–7495.

27. Garboczi, D. N., Hung, D. T., and Wiley, D. C. (1992) HLA-A2-peptide complexes: Refolding and crystallization of moleculesexpressed in Escherichia coli and complexed with single antigenicpeptides. Proc. Natl. Acad. Sci. USA 89, 3429–3433.

28. Schatz, P. J. (1993) Use of peptide libraries to map the substratespecificity of a peptide-modifying enzyme: A 13 residue consensuspeptide specifies biotinylation in Escherichia coli. Biotechnology11, 1138–1143.

29. Myers, L. C., Terranova, M. P., Nash, H. M., Markus, M. A., andVerdine, G. L. (1992) Zinc binding by the methylation signalingdomain on the Escherichia coli Ada protein. Biochemistry 31,4541–4547.

30. Kuroda, M. J., Schmitz, J. E., Barouch, D. H., Craiu, A., Allen,T. M., Sette, A., Watkins, D. I., Forman, M. A., and Letvin, N.L. (1998) Analysis of Gag-specific cytotoxic T lymphocytes insimian immunodeficiency virus-infected rhesus monkeys by cellstaining with a tetrameric major histocompatibility complex classI-peptide complex. J. Exp. Med. 187, 1373–1381.