Jul 27, 1992 - Foundations of Kock and bsterlund, High Tech Receptor AB, and the Swedish Research ... 101, San Diego). The purified DNA fragments were ...
Vol. 267, No. 35, Issue of December 15, PP. 25583-25588,1992 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Q 1992 by The American Society for Biochemistry and Molecular Biology, Inc.
Protein LG: A Hybrid Moleculewith Unique Immunoglobulin Binding Properties* (Received for publication, July 27, 1992)
Britt-Marie KihlbergS, Ulf Sjobringg, William Kasternll, and Bjorck$II Lars From the Departments of $Medical and Physiological Chemistry and $Medical Microbiology, University of Lund, S-22100Lund, Sweden and the TDepartment of Pathology, University of Florida, Gainesville, Florida 32610
(9, 10). In the present study, light chain binding domains of Immunoglobulin (1g)-binding bacterial proteins have attracted theoretical interest for their role in molecu-protein L were combined with Fc-binding protein Gdomains. lar host-parasite interactions, and they are widely The usedresulting hybrid molecule, protein LG, was found to be as tools in immunology, biochemistry, medicine, and an efficient and versatile Ig-binding molecule. The construcbiotechnology. Protein L oftheanaerobicbacterial tion of protein LG, its binding properties, and its potential as species Peptostreptococcus magnus bindsIglight an immunochemical tool are described and discussed. chains, whereas streptococcal protein G has affinity for the constant (Fc) region of IgG. In this report, Ig MATERIALS ANDMETHODS binding parts of protein L and protein G were comProteins and Radiolabeling of Proteins-The protein L fragment, bined to form a hybrid molecule, protein LG, which was found to bind large a majority of intact human Igs B1-4, containing four IgG binding repeats, has been described (6). protein G fragment, CDC, corresponding to two IgG binding as well as Fc and Fab fragments, and Ig light chains. The regions and polyclonal human IgG, wasfrom Kabi Pharmacia, StockBinding to Ig was specific, and the affinity constants holm, Sweden, and further purified by ion-exchange chromatography. of the reactions between protein LG and human IgG, Polyclonal human IgGFc, IgM, and serum IgAwere from Cappel IgGFc fragments, and K light chains, determined by Laboratories, West Chester PA. Fab fragment was prepared by papain Scatchard plots, were 5.9 X lo’, 2.2 X lo’, and 2.0 X digestion of human polyclonal IgG. Human monoclonal K light chains (Mal, subtype I) were kindly provided by Dr. Anders Grubb, Dept. of 10’M-’, respectively.The binding properties of protein LG were more complete as compared with previously Clinical Chemistry, University of Lund, Sweden. Mouse monoclonal described Ig-binding proteins when also tested against IgGs were kindly provided by Dr. Christina Glad, BioInvent, Lund, Sweden, and rat monoclonal IgGswere from Dr. Thomas Brodin, mouse and rat Igs. This hybrid protein thus represents a powerful tool for the binding, detection, and purifi- Kabi Pharmacia. Mouse polyclonal IgG was from Sigma. Protein LC, the B1-4 and CDC fragments were radiolabeled using the Bolton and cation of antibodies and antibody fragments. Hunter reagent (11) (Amersham Corp.). Immunoglobulins and frag-
Some strains of the anaerobic bacterial species Peptostreptococcus magnus express an Ig light chain-binding protein, called protein L (1, 2). This fibrous surface molecule with a molecular mass of 77 kDa (3) binds to the framework region of the variable domain of human Ig light chains, preferentially K light chains (4). Apart from human Igs, protein L also shows affinity for Ig light chains of several other mammalian species, including mouse, rat, and rabbit light chains ( 5 ) .The structure of protein L was recently determined, and the light chain binding activity was mapped to five homologousrepeat units, each of 72-76 amino acids (6). Protein G is a surface protein of group C and G streptococci which binds to the constant (Fc) region of mammalian IgG antibodies with high affinity (7, 8). Protein G also contains repeat units (each 55 amino acids long) that are responsible for the interaction with IgG
* This work was supported by grants from the Swedish Medical Research Council (Projects 7480 and 9926), King Gustav V’s 80-year foundation, the Medical Faculty of the University of Lund, the Foundations of Kock and bsterlund, High Tech Receptor AB, and the Swedish Research Council for Engineering Sciences (Project 231). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequencefs) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bankwith accession numberfs) M86697 and Y00428. I( To whom correspondence should be addressed Dept. of Medical and Physiological Chemistry, University of Lund, P. 0. Box 94, S221 00 Lund, Sweden. Tel.: 46-46-104492;Fax: 46-46-157756.
ments thereof were labeled with (Amersham) with the chloramineT method. DNA Constructs-A DNA fragment corresponding to B1-4 (see Fig. 1) was generated by polymerase chain reaction (PCR),’ essentially as described previously (12, 13). A 200 JIM concentration of each of the deoxynucleotides was used per 100-pl reaction together with 20 pmol of the oligonucleotide primers 5”GCTCAGGCGGCG CCGGTAGAAAATAAAGAAGAAACACCAGAAAC-3’and 5’-CAG CAGCAGCCATGGGTTCTTCTGGTTTTTCGTCAACTTTCTTA -3’ and target DNA solution containing 10 ng of P. magnus strain 312 DNA. The amplification was achieved by 30 cycles of 1 min at 94 “C, 1 min at 60 “C, and 1 min at 72 “C, using heat-stable DNA polymerase (Amplitaque, Perkin-Elmer Cetus Instruments).A second PCR was done similarly but with 20 pmol of the oligonucleotide primers 5’-GGCCATGGACACTTACAAATTAATCCTTAATGGT3’ and 5’-CAGGTCGACTTATTACATTTCAGTTACCGTAAAGGTCTTAGT-3’, and target DNA solution containing 10 ng of DNA from the group C streptococcal strain C40 (13),to generate a DNA fragment corresponding to CDC (Fig. 1). The PCR-generated protein L gene product was digested with HpaII and NcoI, and thePCR-generated protein G gene product was digested with NcoI and SalI. These DNA fragments were separated ona 2.5% agarose gel (GTG, FMC Bioproducts, Rockland, ME), extracted from the gel and purified with glass beads (Geneclean, Bio 101, San Diego). The purified DNA fragments were ligated to the secretion vector pHD 389 (14), which had been digested with NarI/ SalI. The ligation mixture was used to transform Escherichia coli strain LE392. Positive clones were selected by probing the transformants (transferred to nitrocellulose filters and lysed with 10% SDS), with a [y-32P]ATP-(Amersham Corp.) labeled oligonucleotide probe from aBrepeat sequence (5”GAATTCAAAGGAACATTTGAAGAAGCAACAGCAGA-3’).The probe was labeled with T4 polynu‘The abbreviations used are: PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; PVDF, polyvinylidene difluoride.
25583
25584
Protein LG: A n Ig-binding Protein Hybrid
cleotide kinase, and the filters were autoradiographed as described previously (6). Standard procedures were used for DNA sequencing, restriction cleavage, ligation, and transformation (15). The inserts were analyzed by agarose gel electrophoresis and by PCR using combinations of the primers described above. Expression and Purification of E. coli-expressed Proteins-The cloned DNA hybrid molecules wereexpressed as described previously (6). Periplasmic lysates of the E. coli clones expressing protein LG and fragment B1-4 were prepared by osmotic shock as described (14) and applied to IgG-Sepharose (2) followed by ion-exchange chromatography on a Mono Q HR5/5 fast protein liquid chromatography column (Pharmacia LKB, Uppsala, Sweden), using a gradient of 00.5 M NaCl in 20 mM Tris-HC1, pH 7.5. SDS-Polyacrylamide Gel Electrophoresis (PAGE) and Agarose Gel Electrophoresis-SDS-PAGE was run as described by Neville (16), using total acrylamide concentrations of 10% (cross-linking a t 10%:3.3%).Samples were pretreated by 3 min boiling in a sample buffer containing 2% SDS and 5% 2-mercaptoethanol. Molecular weight marker kits were purchased from Sigma. Agarose gelelectrophoresis was performed as described by Johansson (17). Gels were stained with Coomassie Blue. Western Blot Analysisand Analysis of Proteins Directly Applied to Polyvinylidene Difluoride (PVDF) Membranes-After separation by SDS-PAGE, proteins were electrotransferred to PVDF membranes (Immobilon, Millipore, Bedford, MA) as described by Towbin et al. (18).Proteins were also directly applied to PVDF membranes using a n apparatus (Minifold 11) from Schleicher & Schuell (Dassel, Germany). Membranes were blocked, washed, and incubated with '*'Ilabeled probe (2 X IO6cpm/ml for 3 hat 25 "C) and autoradiographed as described previously (19). CompetitiveBinding Assay and Calculation of Affinity ConstantsConstant amounts of human polyclonal IgG, IgGFc, or K light chains, coupled to polyacrylamide beads (Immunobeads, Bio-Rad) and "'Ilabeled protein LG, B1-4, or CDC in Verona1 buffer, pH 7.35, containing 0.15 M NaCl and 0.1% gelatin (VBS), were incubated with varying amounts of nonlabeled protein LG, B1-4, or CDC fragments for 5 h at 25 "C. Details of the procedure and calculations, using the formula of Scatchard, have been reported earlier (3,201. Affinity ChrornatographylPlasma Absorption Experiment-These experiments were performed with protein LG, B1-4, or CDC fragments coupled to CNBr-activated Sepharose 4B (Pharmacia) at 1-5 mg protein/ml, wetgel, equilibrated in 10 mM sodium phosphate buffer, pH7.2, containing 0.12 M NaCl and 0.05% Tween 20 (PBST). Equal volumes of human plasma diluted 1:20 with PBST, and Sepharose, were incubated for 1 h at room temperature. The gelwas
washed with 10 volumes of PBST and eluted with 0.1 M glycine, pH 2.0. Eluted fractions where immediately neutralized with 1 M Tris. Other Procedures-Samples from plasma absorption experiments were analyzed for their content of IgG, IgM, and IgA with an automatic turbidometric assay (Cobas Mira, Roche) using rabbit antihuman immunoglobulin antibodies (Dakopatts,Denmark). Amino acid sequence analysis was performed on an Applied Biosystems gasphase sequenator model 470A equipped with an on-line phenylthiohydantoin analyzer model120A, and a data modulemodel 900A, using chemicals and software supplied by the manufacturer. RESULTS
Construction, Expression, and Purification of Protein LGPCRfragments were generated whichcovered four light chain-binding B repeatsof protein L (Bl-4) and two IgGFcbinding C repeats of protein G (CDC). In the case of protein L, DNA from the peptostreptococcal strain 312 (1)was used as template, whereas DNA from the group C streptococcal strain C40 (13) was usedfor the protein G repeats. The protein G gene of this strain contains only two IgGFc binding C domains, whereas the G148 strain from which protein G was (10). The protein originally isolated (7) has three repeat units L gene from strain 312 and the protein G gene from strain C40 have both been sequenced (6, 13), andFig. 1 shows their organization. The two PCR-amplified gene fragments (Bl-4 and CDC) were ligated to form the hybrid construct LG (Fig. 1). The hybrid fragment was then ligated to the secretion vector pHD 389 (14), and the ligation mixture was used to transform E. coli. Sequence analysis of the insert revealed the expected DNA structure of the LG construct. Fig. 2 shows the sequence of the hybrid molecule. From the amino terminus the corresponding protein consisted of the 3 carboxyl-terminal amino acids (Val-Glu-Asn) from the A domain, the B1-4 domains, and the9 amino-terminalresidues of the B5 domain (comprising amino acid residues 95-398 in the protein L sequence (6)) followed by the tripeptide Pro-Met-Asp linking the B1-4 fragment to the CDC fragment. DNA fragments cloned into the unique NurI site of the pHD 389 vector are fused to thesignal sequence of E. coli outer membrane protein A (ompA),in this case resulting in an extra alanine residue
PROTEIN G
PROTEIN L
S
1
A2
M
A3
Dl
M
i
+ CDC
61-4
PROTEIN LG
FIG. 1. Construction of protein LG. Schematic representation of peptostreptococcal protein L (701 amino acids),streptococcal protein G from strain C40 (489 amino acids), and the hybrid molecule protein LG. Protein L and protein G both contain different repeat units. In protein L the B repeats (72-76 amino acids each) bind Ig light chains, in protein G the C repeats (55 amino acids each) bind IgGFc. The proteins are attached to the bacterial surface through the W and M domains.
25585
Protein LG: An Ig-binding Hybrid Protein
f
Ss (ompA)
BLOT
ATGAAAAAAACTGCTATCGCTATCGCTGTTGCTCTGGCTGGTTTCGCTACTGTTGCTCAG MetLysLysThrAlaIleAlaIleAlaValAlaLeuAlaGlyPheAlaThrValAlaGln 61
60 20 120 40
GTAACAATCAARGCTAACCTAATCTTTGCAAATGGAAGCACACAAACTGCAGAATTCAAA Va1ThrIleLysA1aAsnLeuIlePheAlaAsnGlySe~hrGlnThrAlaGluPheLys
180 60
GGAACATTTGAAAAAGCARCATCAGAAGCTTATGCGTATGCAGATACTTTGAAGAAAGAC GlyThrPheGluLysAlaThrSerCluAlaTyrAlaTyrAlaAspThrLeuLysLysAsp
240 80
AATGGAGAATATACTGTAGATGTTGCAGATAAAGGTTATACTTTAAATATTAAATTTGCT AsnClyGluTyrThrValAspValAlaAspLysGlyTyrThrLeuAsnIleLysPheAla 62
300 100
24kDa
-
360 120
TATGCAGATGGAAAAACACAAACAGCAGAATTCAAAGGAACATTTGAAGAAGCAACAGCA TyrAlaASpGlyLysThrGlnThrAlaGluPheLysG1yThrPheGluGluAlaThrAla
420 140
GAAGCATACAGATATGCAGATGCATTAAAGAAGGACAATGGAGAATATACAGTAGACGTT GluAlaTyrArgTyrAlaAspAlaLeuLysLysAspAsnGlyGluTyrThrValAspVal
480 160
63 540 180
CCAAAAGAAGAAGTTACTATTAAAGCAAACTTAATCTATGCAGATGGAAAAACACAAACA ProLysGluGluValThrIleLysAlaAsnLeuIleTyrAlaAspGlyLysThrGlnThr
600 200
GCAGAATTCAAAGGAACATITGAAGAAGCAACAGCAGAAGCATACAGATATGCTGAClTA
AlaGluPheLysGlyThrPheGluGluAlaThrAlaG1uAlaTyrArgTyrAlaAspLeu
660 220
TTAGCAAAAGAAAATGGTAAATATACAGTAGACGTTGCAGATAAAGGTTATACTTTAAAT LeuAlaLysGluAsnClyLysTyrThrValAspValAlaAspLysGlyTyrThrLeuAsn
720 240
A B C
FIG. 3. Western blot analysis of protein LG at different stages of purification. Material from lysates of E. coli expressing protein LG (lanesA ) was applied to and eluted from an IgG-Sepharose column with pH 2.0 (lanes B ) . This material was further purified by ion-exchange chromatography on Mono Q (lanes C). The samples were run on 10% SDS-PAGE under reducing conditions on duplicate gels. One was stained with Coomassie Blue (STAIN), whereas the peptides of the other gel were transferred to a PVDF filter, which was incubated with lB1-labeled polyclonalhuman IgG (2X lo6cpm/ ml for 3 h), washed, and autoradiographed (BLOT).
780 260 GCAAACTTAATCTATGCAGATGGAAAAACTCAAACAGCAGAGTTCAAAGGAACA"XA
AlaAsnLeuIleTyrAlaAspGlyLysThrGlnThrAlaGluPheLysGlyThrPheAla
840 280
GAAGCAACAGCAGAAGCATACAGATACGCTGACTTATTAGCAAAAGAAAATGGTAAATAT G1uAlaThrAlaGluAlaTyrArgTyrAlaAspLeuLeuAlaLysGluAsnGlyLysTyr
900 300 960 320
Mm
SOkDa-
I
36kDa17kDa-
1020 340
AAAGGCGAAACAACTACTGAAGCTGTTGATGCTGCTACTGCAGAAAAAGTCTTCAAACAA LysGlyGluThrThrThrGluAlaVa1AspAlaAlaThrAlaGluLysValPheLysGln
1080 360
TACGCTAACGACAACGGTGTTGACGGTGAATGGACTTACGACGATGCGACTAAGACCTTT
1140 380
A
TyrA1aAsnAspAsnGlyValAspGlyGluTrpThrTyrAspAspAlaThrLysThrPhe
TACAAACTTGTTATTAATGGTAAAACATTGAAAGGCGAAACAACTACTAAAGCAGTAGAC
TyrLysLeuVa1IleAsnGlyLysThrLeuLysGlyGluThrThrThrLysAlaValAsp GCAGAAACTGCAGAAAAAGCCTTCAAACAATACGCTAACGACAACGGTGTTGATGGTGTT
AlaCluThrAlaG1uLysAlaPheLysG1nTyrAlaAsnAspAsnGlyValAspGlyVal
1260 420 1320 440
TGGACTTATGATGATGCGACTAAGACCTTTACGGTPACTG
TrpThrTyrAspAspAlaThrLysThrPheThrValThrGk
1365 455
FIG. 2. Nucleotide and derived amino acid sequenceof the gene construct encoding protein LG.The coding sequence start-
RA CBA CRA Cf
C
polyacrylamide gels under reducing conditions. After electrophoresis one gel was stained with Coomassie Blue (STAIN), whereas the peptides of three identical gelswere blotted onto PVDF filters (BLOTS). The filters were incubated with 'Wabeled polyclonal human I&, monoclonal human K chains, or human IgGFc fragments (2 X 10' cpm/ml of each probe) for 3 h and autoradiographed. The molecular masses (M,)ofLG (lanes A ) , B1-4 (lanes B ) , and CDC (lanes C) were calculated from their sequences and found to be 49.913, 35.824, and 16.775 kDa, respectively.
ing at residue 1 is in upper case letters, and the arrows denote the beginning of each protein domain.
PROBE Protein LG
IgG
between the ompA signal sequenceand theconstruct. For the screening of clones, a protein L-based oligonucleotide was employed, and a positive clone with an insert of the correct size was expressed. Periplasmic lysates of the clone provided starting material for the isolation of protein LG. The purification protocolincluded affinity chromatography on IgGSepharose followedby ion-exchange chromatography on a Mono Q fast protein liquid chromatographycolumn. As shown in Fig. 3,multiple IgG-bindingpeptides were eluted from IgGSepharose, whereas a single protein band was obtained after the ion-exchange purification step. The apparent molecular mass of protein LG on SDS-PAGE was 50 kDa as compared with 49.913 kDa when calculated fromthe sequence. Aminoterminal amino acid sequence analysis of the purified molecule demonstrated that the 21 residues of the signal peptide of ompA had not been cleaved, addingthis sequence and the extra alanine to protein LG (Fig.2). The Ig binding properties of protein LG (see below) didnot seem to be affectedby these
3
FIG. 4. Western blot analysis of protein LG, B1-4, and CDC. 0.3-0.5 pgof the three peptides were run on 10% SDS-
Fab
FC
IQM
IgA
K
PQ 5
2 0.5
0.1
FIG. 5. Binding of protein LG to human antibodies and antibody fragments. Various amounts of polyclonal human IgG, Fab and Fc fragments of IgG, IgM and IgA, and monoclonal K (subtype I) were applied to a PVDF filter, probed with '251-labeledprotein LG (2 X 10' cpm/ml) for 3 h, and autoradiographed.
22 amino-terminal amino acid residues,which are included in the molecularmass of 49.913 kDamentioned above. The expression level in stationary E. coli cultures varied between 60 and 80 mg of protein LG/liter of culture. The average yield after the two purification steps was 45%. Binding Properties of Protein L.G-Protein LG and peptides
Protein LG: A n Ig-binding Hybrid Protein
PROBE:
pg 5
Protein LG
2 0.5
FIG.6. Binding of protein LG, protein L, and protein G to murine antibodies. Various amounts of polyclonal and monoclonal mouse IgG antibodies were applied to PVDF filters. The filters were probed (2 X lo6 cpm/ml) with radiolabeled protein LG, protein L (peptide B1-4),and protein G (peptide CDC), respectively, for 3 h and autoradiographed.
0.1
5 Protein L
2 0.5
8 - 0 -
I)
..
..
..
0
. )
0.1
Protein G
0.1
bG"*5(-cM:
incubated separately with immobilized human polyclonal IgG together with different amounts of unlabeled protein LG, B14, or CDC. Protein LG completely blockedthe binding of all three radiolabeled peptides, demonstrating that theIg binding activities of the B1-4 and CDC peptides were maintained in protein LG (Fig. 7). It can also be noted from the inhibition curves of Fig. 7 that B1-4 and CDC were capable of blocking 0 1 2 -2 -1 0 1 2 the homologous radiolabeled peptide, whereas heterologous -2 ! - 1 inhibition with B1-4 or CDC was negligible.Finally, CDC, in contrast to theB1-4 peptide, partially blocked the binding of FIG.7. Competitive binding experiments. The binding of lZ6I- radiolabeled protein LG to IgG. The fact that B1-4 did not labeled protein LG, protein L (B1-4),or protein G (CDC) to immo- interfere with the binding of protein LG to IgG whereas CDC bilized human polyclonal IgG was inhibited with different amounts did, cannot be explained by a higher affinity between IgG and of unlabeled protein LG (O),protein L (A), or protein G (0). CDC (see the affinity constants in Fig. 8). The low affinity interaction between protein G and IgGFab (7, 21) could, B1-4 and CDCwere run on SDS-PAGE and analyzed by however, explain this discrepancy. Western blot analysis using human IgG, K light chains, and Scatchard analysis was utilizedto obtain a more direct and IgGFc fragments as radiolabeled probes. As demonstrated in quantitative assessment of the interactions between the three Fig. 4, protein LG reacted with all three probes, B1-4 with peptides (LG, B1-4, and CDC) and Ig (intact IgG, IgGFc,and IgG and K , and CDC with IgG and IgGFc. Moreover, when K light chains). The various plots are depicted in Fig. 8, where polyclonal human IgG, Fab and Fc fragments of IgG, poly- the affinity constants arealso given. No binding was recorded clonal human IgM and IgA, and monoclonal human K chains between B1-4 and IgGFc, or between CDC and K light chains. were appliedto PVDF filters and incubated with radiolabeled As in the competitive binding experiments described above,a protein LG, all antibody preparations showed affinity for the major conclusion fromthe results shown in Fig. 8 is that the hybrid protein (Fig. 5). Mouse monoclonal IgG antibodies, protein L and G components of protein LG are fully active in especially of subclass 1, sometimes bind poorly or not at all the hybrid protein. Furthermore, the highest affinity was t o protein G (this is also the case with staphylococcal protein obtained for the reaction between protein LG and IgG, sugA). Consequently, a panel of murine monoclonal antibodies gesting a synergistic action of the light chain binding and Fc of all four IgG subclasses was tested. Protein LG was found binding repeats of protein LG in the interaction between the to react with all antibodies, whereas the protein L and protein hybrid protein and intactIgG. G peptides (Bl-4 and CDC) showed more restricted binding A set of experiments was also designed to determine the patterns (Fig. 6). Moreover, rat monoclonal IgG antibodies total Ig binding capacity of protein LG and to study the consistently reacted with protein LG in contrast to proteins specificity of protein LG's interaction with Ig. Thus, human L and G (not shown). plasma was incubated with protein LG-Sepharose. The maCompetitive binding experiments were performed to ana- terial that was not bound, as well as the proteins that could lyze further the binding properties of protein LG in compar- be eluted from protein LG-Sepharose at low pH, were anaison with the two peptides, B1-4 and CDC, included in the lyzed together with plasma by agarose gel electrophoresis. The hybrid molecule. Radiolabeled protein LG, B1-4, or CDC was plasma proteins of the y region, corresponding to Igs of all
;4 z i
Protein LG: An Ig-binding Protein Hybrid
; P w
e P
4
LO-Fc Ka=22.10sM1
1 . 2 0
m
..
05
10
05
10
15
10
20
Bound orotein LG. nM
2
l . . . ,;- "2
10
05
05
10
15
Bound protein L (Bl-4). nM
c
en
$
5
1
0
m
05
10
15
10
20
Bound protein G (COC). nM
FIG.8. Scatchard plots for the reactions between l2'Ilabeled protein LG, protein L (Bl-4). protein G (CDC), and immobilized human polyclonal IgG, IgGFc, or K light chains. 0.1 ml of '251-labeledprotein, 0.1 ml of polyacrylamidebeads coupled with IgG, IgGFc orK, and 0.2 ml of unlabeled protein in concentrations between 1 and 1,000 ng/ml were mixed. After incubation at 25 "C for 5 h beads were centrifuged, washed, and the radioactivity of the pellet was determined. All reactions wererepeated at least twice with duplicates in each experiment. The best fitting straight line through the points wasdrawn after regression analysis. The equilibrium constant of a reaction was then equal to the absolute value of the slope of the curve.
+
FIG. 9. Plasma absorption experiments with immobilized protein LG. Human undiluted plasma (0.1 ml) was incubated with 0.15 ml of protein LG-Sepharose(5 mg of protein LG/ml Sepharose) for 1 h at room temperature. The Sepharose was washed and eluted (see "Materials and Methods"). 10 plof undiluted human plasma (bneA), 10 p1 of plasma sample absorbed with protein LG-Sepharose ( l a n e B ) , and 1081 of the material eluted from protein LG-Sepharose diluted about 10-fold ( l a n e C)were run on agarose gelelectrophoresis at pH 8.6. The gelwas stained with Coomassie Blue. The arrow indicates the application slit, and the anode is atthe top.
classes, were almost completely absorbed and subsequently eluted at pH 2.0 (Fig. 9), demonstrating a broad and efficient Ig binding capacity of protein LG. This was emphasized
25587
further when the content of IgG,IgM, and IgA (whichtogether comprise 99.9% of the Igs present in human plasma) was measured by a turbidometric assay in the plasma sample as well as in the filtrate and the eluate from the absorption experiment with protein LG-Sepharose. In this case, 89% of the total content of Ig in plasma was bound to protein LG (the corresponding figures for B1-4 and CDC were 63 and 78%, respectively), and 90-95% of the bound Igs could be recovered with pH 2.0. Except for Ig heavy chains (y,p, and a)and Ig light chains, no other protein bands were seen when the eluted proteins were analyzedby SDS-PAGE (not shown). These datademonstrate that protein LG binds Ig of different classes with a high degreeof specificity. DISCUSSION
In the protein LG construct described in this report, we chose to incorporate four protein L K binding repeats (plus the 9 amino-terminal amino acid residues of the fifth repeat) although the protein L expressed by different isolates of P. mugnus always contain five repeat units (6, 22). However, previous work demonstrated that thefifth repeat did not add to theaffinity for K light chains (6). In thecase of protein G, the number of IgGFc binding C repeats varies between two and three in different strains of group C and G streptococci (9, 10, 13). The affinity constant for the reaction between protein G containing three C repeats and IgG is above 10" M-' (23), which was thought to be unnecessarily high. Thus, whenrecoveringboundIg or antigen-antibody complexes from protein LG, it is of course advantageous if the elution conditions are not too harsh. It should, however, be possible to construct protein LG molecules with lower, or higher (in the case of IgGFc binding activity), affinities by altering the number of included repeat units. In theprotein LG molecule described here, the K and IgGFc binding parts both have affinity constants of about lo9 M-' for their ligands. For most applications, this level of affinity appeared suitable. Proteins L and G are fibrous, elongated molecules (3, 23). They are bothrobust proteins that maintain their Ig binding activity over a wide range of temperature, salt, and pH (3,23, 24). Moreover, both proteins bind to Ig without interfering with the antigen binding site of the antibody molecule (2, 3, 25), which is of practical importance when the proteins for instance are used for binding or detection of antigen-antibody complexes. These properties contribute to the utility of protein L and protein G as immunochemicaltools.However, there are also obvious limitations in their binding properties. Although protein G as well as staphylococcal protein A (26) are currently widely used in biochemical and immunological research, and inbiotechnology, these proteins do not bind, or they bind poorly, Ig classes other than IgG. Some rat and mouse monoclonalantibodies of the IgG class also lack affinity for proteins A and G under physiological conditions (27). Protein L, on the other hand, has affinity for all Ig classes through its binding to Ig K light chains. However, the molecule will not bind most antibodies carrying X light chains (4), and as IgG is the quantitatively dominating Ig class, the total amount of human Ig bound to protein L will be less as compared with proteins A and G; 63%as compared with about 70% (this study, 27). Therefore, the combined binding properties of proteins L and G offer several advantages. First, protein LG has affinity for a larger part of the Ig population than any otherIg-binding protein so far described. This could have medical implications. Thus, extracorporeal removal of antibodies and antigen-antibody complexesfrom patient blood plasma with immobilized protein A has successfully been used in the treatment of certain autoimmune disorders
25588
Protein LG: An Ig-binding Hybrid Protein
(28). Forthispurposeprotein LG will be more effective. Second, human, murine, and rat monoclonal antibodies that do notbelong to theIgG class will bind poorly or not at all to proteins A and G. As mentioned and described above, this is alsothe case withsomemouse andrat IgG monoclonal antibodies, whereas protein LG binds the large majority of these antibodies. Third, protein LG binds to Fab fragments of all Ig classes provided that they carry the correct light chain. This could prove to be an important property of the hybrid protein as recent developments have made itpossible t o produce Fab fragments withdesired antibody specificity in bacteria (for review, see Ref. 29). It isreasonable to anticipate a future inwhich such antibody fragments will be widely used. By introducing a light chain in the construct that binds to the protein L part of protein LG, themolecule could also be utilized for the binding, detection, and purification of such antibodyfragments.Forthe various reasonslisted above, protein LG has a considerablepotential value as anIg-binding reagent. Acknowledgments-We are indebted to Ingbritt Gustafsson for invaluable technical assistance and Dr. Henrik Dalboge, Novo-Nordisk A/S, Copenhagen, Denmark, for generouslyproviding the expression vector pHD 389. 1. 2. 3. 4. 5.
REFERENCES Myhre, E. B., and Erntell, M. (1985) Mol. Immunol. 22,879-885 Jorck, L. (1988) J. Immunol. 1 4 0 , 1194-1197 jkerstrom, B., and Bjorck, L. Bjorck,J. L.,Biol. andy kerstrom, m . 2 6 4 , B. 19740-19746 (1992) J.Biol. Nilson, B. H. K., Solomon, A,, (1989) Chem. 267,2i34-2239 Bjorck, L., and kerstrom, B. (1990) in Bacterial Immunoglobulin Binding
Proteins (Boyle, M. D. P., ed) vol. 1, pp. 267-278, Academic Press, San Diego 6. Kastern, W., Sjobring, U., and Bjorck, L. (1992) J. Bid. Chem. 267,1282012825 7. Bjorck, L., and Kronvall, G. (1984) J.Immunol. 133,969-974 8. Reis, K. J., Ayoub, E. M., and Boyle, M. D. P. (1984) J. Immunol. 1 3 2 , 3091-3097 9. Fahnestock, S. R., Alexander, P., Nagle, J., and Filipula, D. (1986) J. Bacterid. 167,870-880 10. Guss, B., Eliasson, M., Olsson, A,,UhlBn, M., Frej, A.K., Jornvall, H., Flock, J. I., and Lindberg, M. (1986) EMBO J. 6,1567-1575 11. Bolton, A. E., and Hunter, W. M. (1973) Biochem. J. 133,529-539 12. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S., Higuchi, R., Horn, G., Mullis, K. B., and Ehrlich, H. A. (1988) Science 239,487-491 13. Sjobring, U.,Bjorck, L., and Kastern, W. (1991) J. Biol. Chem. 2 6 6 , 399405 14. Dalboge, H., Bech Jensen, E., Tottrup, H., Grubb, A., Abrahamson, M., Olafsson, I., and Carlsen, S. (1989) Gene (Amst.) 7 9 , 325-332 15. Sambrook, J.,Fritsch, E., andManiatis, T. (1989) Molecular Cloni A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory,?old Spring Harbor, NY 16. Neville, D. M., Jr. (1971) J . Biol. Chem. 246,6328-6334 17. Johansson, B. G. (1972) Scand. J. Clin. Lab. Inuest. 2 9 , Suppl. 24, 7-19 18. Towbin. H.. Staehelin. T.. and Gordon.. J. (1979) . . Proc. Natl. Acad. Sci. U. S. A . 76,4350-4354 19. Akerstrorn. B.. Nielsen. E.. and Biorck. " . L. (1987) . , J. Biol. Chem. 262. 1338&13391 ' ' 20. Bjorck, L., and Akerstrom, B. (1990) in Bacterial Immunoglobulin Binding proteins (Boyle, M. D. P., ed) vol. 2, pp. 91-104, Academic Press, San Diego 21. Erntell. M.. Mvhre. E. B.. Siobrine. U.. and Biorck. L. (1988)Mol. Immunol. ~
-.
- ~ " " , E., Nielwn, E., Sjobring, U., and Bjorck, L. (1990) Infect.' Immun. 68,1217-1: ! U 23. Akerstrom, B., and Bjorck, L. (1986) J. Biol. Chern. 261,10240-10247 24. Gronenborn, A. M., Filild a . D. R.. Essip. N. Z., Achari, A., Withlow, M., Wingfield, P. T., and Glore, G. hi. (1951) Science 263,657-661 25. Akerstrom. B.. Brodin., T.., Reis., K... and Biorck. . L. .(1985) . J. Immunol.
22.
.. .. . . . "
I
~~
~
136,2589-2592 26. Forsgren, A,, and Sjauist, J. (1966) J. Immunol. 97,822-827 27. Boyle, M. D. P. (ed) (1990) Bacterial Immunoglobulin Binding Proteins, vol. 1, Academic Press, San Diego 28. Gjorstrup, P., and Watt, R. M. (1990) Transfusion Sci. 11,281-302 29. Pliickthun, A. (1991) BiolTechnology 9,545-551
