|Home | About | Journals | Submit | Contact Us | Français|
A combinatorial human immunoglobulin gene library was constructed from peripheral lymphocytes of an asymptomatic Entamoeba histolytica cyst passer and screened for the production of Fab antibody to the parasite. One of the Fab clones, CP33, recognized the 260-kDa galactose- and N-acetyl-d-galactosamine (Gal/GalNAc)-specific lectin of E. histolytica. By shuffling the heavy and light chains of CP33 with the heavy and light chains of two libraries derived from the cyst passer and a liver abscess patient, 18 additional clones were obtained. Sequence analysis of the heavy-chain genes, including CP33-H, revealed that all the nearest V-segment germ lines belonged to the VH3 family (VH3-21, VH3-30, VH3-48, and VH3-53), but the levels of homology were only 85 to 95%. The closest D-segment germ line was D2-2 or D6-6, and for the J-segment the closest germ line was JH4b or JH6b. On the other hand, all the light-chain genes, including CP33-L, belonged to the Vκ1 family, in which the closest Vκ germ line gene was 02/012 or L5, with the Jκ1, Jκ2, Jκ4, or Jκ5 segment. CP33 and three other Fabs obtained by light-chain shuffling were purified and analyzed further. All of these Fabs recognized the cysteine-rich domain of the 170-kDa heavy subunit of the Gal/GalNAc lectin. Preincubation of E. histolytica trophozoites with these Fabs significantly inhibited amebic adherence to Chinese hamster ovary cells and also inhibited erythrophagocytosis. The ability of the neutralizing antibodies to block erythrophagocytosis for the first time implicates the lectin in phagocytosis and VH3 antibodies in defense against parasitic infections. These results demonstrate the utility of a combinatorial human immunoglobulin gene library for identifying and characterizing neutralizing antibodies from humans with amebiasis.
Worldwide, the intestinal protozoan parasite Entamoeba histolytica causes an estimated 50 million cases of amebic colitis and liver abscess annually, resulting in up to 110,000 deaths (48). The ability of E. histolytica trophozoites to invade the colon and other tissues depends on several pathogenic factors. One of the most important factors is the galactose (Gal)- and N-acetyl-d-galactosamine (GalNAc)-inhibitable cell surface lectin of the ameba. The lectin mediates adherence of trophozoites to human colonic mucins, colonic epithelial cells, neutrophils, and erythrocytes (4, 7, 18, 24, 37, 38). The Gal/GalNAc lectin is also important in the cytolytic event that follows adherence. The amebic lectin is a 260-kDa hetrodimeric glycoprotein composed of 170-kDa heavy and 35- or 31-kDa light subunits (32). In addition, a 150-kDa intermediate subunit of the lectin also contributes to adherence (8, 12). Immunization of experimental animals with these lectin subunits provides protection from liver abscess formation (10, 25, 27, 34, 50). Such protection is also observed after passive immunization with a mouse monoclonal antibody to the intermediate subunit lectin (11).
Recently, it has been shown that E. histolytica-specific human monoclonal antibody Fab fragments can be prepared from the peripheral lymphocytes of a patient with an amebic liver abscess (9, 42). One of the clones which recognized the 260-kDa Gal/GalNAc lectin inhibited the adherence of trophozoites to mammalian cells (9). Successful cloning of the active fragments might be due to the patient's high antibody titer to E. histolytica.
There are also clinically asymptomatic individuals, which pass E. histolytica cysts in their stools. Most of the cyst passers have a positive serology for the ameba, although their antibody titers are low (44). In such cyst passers, there may be protective antibodies that block invasion of trophozoites into tissues. However, little is known about the immune response to E. histolytica in asymptomatic cyst passers. Therefore, molecular analysis of the immune response to the amebic lectin is important for understanding protective immunity and for vaccine development. We report here molecular cloning of immunoglobulins specific for the E. histolytica Gal/GalNAc lectin. The clones were derived from the peripheral lymphocytes of an asymptomatic cyst passer. We also report possible recombination and bacterial expression of antibody genes from both aymptomatic and symptomatic individuals.
Trophozoites of 10 strains of E. histolytica (HM-1:IMSS, HK-9, 200:NIH, HB-301:NIH, H-302:NIH, H-303:NIH, DKB, C-3-2-1, SAW1627, and SAW755CR) were axenically grown in BI-S-33 medium (16). Trophozoites of Entamoeba dispar SAW1734RclAR were cultured monoxenically with Pseudomonas aeruginosa in BCSI-S medium (22). Trophozoites were washed three times with ice-cold 10 mM phosphate-buffered saline (PBS) (pH 7.4) before they were used.
Ten milliliters of peripheral blood was collected from the asymptomatic cyst passer. Cysts detected in the feces had been identified specifically as E. histolytica by PCR. The serum was positive for E. histolytica with a titer of 1:64 (borderline positive), as determined by an indirect fluorescent-antibody (IFA) test. Lymphocytes were separated from the blood by Ficoll-Paque (Pharmacia, Uppsala, Sweden) density gradient centrifugation. RNA was isolated from the lymphocytes with an RNeasy total RNA purification kit (QIAGEN GmbH, Hilden, Germany). Reverse transcriptase PCR was performed with a GeneAmp RNA PCR kit (Perkin-Elmer Cetus, Norwalk, Conn.) according to the manufacturer's instructions. An oligo(dT)16 primer was used for cDNA synthesis. Genes encoding the κ and λ light chains and the Fd region of the γ heavy chain were amplified as previously described (42). Thirty-five cycles of PCR were performed as follows: denaturation at 94°C for 1 min (5 min in cycle 1), annealing at 50°C for 2 min, and polymerization at 72°C for 3 min (10 min in cycle 35). The light-chain genes were first ligated with an expression vector, pFab1-His2 (46), and then introduced into XL1-Blue Epicurian Coli (Stratagene, La Jolla, Calif.). The vector with inserts was selected, and then the Fd heavy-chain genes were ligated into the vector and introduced into the bacteria.
Screening of positive clones was performed as previously described (9). The expression vector containing light-chain and Fd heavy-chain genes was introduced into competent Escherichiacoli JM109. Approximately 103 to 3 × 103 colonies per 90-mm plate were grown on Luria-Bertani agar plates containing 50 μg of ampicillin per ml at 37°C. Colonies were transferred to nitrocellulose filters and then incubated on fresh plates containing 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and ampicillin at 30°C for 6 h. Each filter was treated with chloroform vapor and then incubated with lysis buffer (100 mM Tris-HCl [pH 7.5], 150 mM NaCl, 5 mM MgCl2, 1.5% bovine serum albumin, 1 μg of DNase per ml, 40 μg of lysozyme per ml) overnight. After washing with PBS containing 0.05% Tween 20 (PBS-Tween), the filter was incubated in PBS-Tween containing 5% skim milk and then with 400 μg of soluble E. histolytica antigen prepared from trophozoites of the HM-1:IMSS strain per ml. After washing, the filter was incubated with horseradish peroxidase (HRP)-conjugated antibody purified from the plasma of a patient with an amebic liver abscess. The filter was washed and then developed with a Konica immunostaining kit (HRP-1000; Konica Co., Tokyo, Japan). Positive clones were identified in original plates and then cultured in 5 ml of super broth (30 g of tryptone per liter, 20 g of yeast extract per liter, 10 g of morpholinepropanesulfonic acid [MOPS] per liter; pH 7) containing ampicillin until an optical density at 600 nm of 0.6 to 0.7 was obtained. IPTG, at a final concentration of 100 μM, was added to the bacterial cultures, which were then incubated at 30°C for 14 h. The bacteria were pelleted by centrifugation, suspended in 250 μl of PBS containing 1 mM phenylmethylsulfonyl fluoride, and then sonicated. The lysates were centrifuged at 12,000 × g for 10 min, and the supernatant was subjected to a second screening by performing an IFA test with intact trophozoites.
The IFA test, with live intact trophozoites (45) or formalin-fixed trophozoites smeared on glass slides (43), was performed as previously described. Fluorescein isothiocyanate-conjugated goat immunoglobulin G (IgG) to human IgG Fab (Organon Teknica Co., Durham, N.C.) was used as the second antibody.
The light-chain gene of the positive clone was replaced with light-chain genes from the library constructed from the asymptomatic cyst passer. The light-chain-shuffled plasmid containing the cloned Fd heavy chain was introduced into the bacteria and then screened again. The Fd heavy chain of the positive clone was also replaced with the library's heavy-chain genes and screened again. Another immunoglobulin library, prepared from a patient with an amebic liver abscess (9), was also used as a source of light- and heavy-chain genes. Positive clones were subjected to a second screening by the IFA test. Light- and heavy-chain genes of clones positive in the second screening were amplified by PCR, digested with MboI, RsaI, DraI, HinfI, and HaeIII and with MboI, RsaI, and AluI, respectively, and then compared by performing electrophoresis in an agarose gel.
The soluble antigen derived from E. histolytica trophozoites, which was used for screening clones, was also used for an enzyme-linked immunosorbent assay (ELISA). The wells of the ELISA plates, containing 5 μg of antigen diluted with 50 mM sodium bicarbonate buffer (pH 9.6), were incubated overnight at 4°C. The plates were washed with PBS-Tween and then treated with PBS containing 3% skim milk for 1 h. For use in the ELISA, positive clones of bacteria were cultured in 10 ml of medium, and then 0.5 ml of the resultant supernatant was prepared as described above. One hundred microliters of the supernatant was added to the wells and incubated for 1 h at room temperature. After the wells were washed, they were incubated with 100 μl of HRP-conjugated sheep antibody to human IgG F(ab′)2 (ICN Pharmaceuticals, Aurora, Ohio) for 1 h at room temperature. After they were washed with PBS-Tween, they were incubated with 200 μl of a substrate solution (0.4 mg of o-phenylenediamine per ml in citric acid-phosphate buffer [pH 5.0] containing 0.001% hydrogen peroxide). After 30 min of incubation, the reaction was stopped by addition of 50 μl of 2.5 M H2SO4, and the optical density at 490 nm was determined with a model 550 microplate reader (Bio-Rad, Hercules, Calif.).
Cloned light-chain genes and Fd heavy-chain genes were recloned into sequencing vectors. Cycle sequencing in both directions was performed with Thermo Sequenase (Amersham Life Science, Cleavland, Ohio) by using M13 primers. Reactions were performed with a model 4000L automated DNA sequencer (LI-COR, Lincoln, Nebr.).
Positive clones were cultured in 1 liter of medium. Twenty milliliters of each resultant supernatant, prepared as described above, was filtered through 0.22-μm-pore-size filters and used for purification of Fab fragments. Purification of Fabs was performed by affinity chromatography by using His·Bind Resin (Novagen, Madison, Wis.) according to the manufacturer's instructions.
Purified Fab fragments and trophozoites of E. histolytica HM-1:IMSS in PBS were solubilized with an equal volume of the sample buffer (23) containing 2 mM phenylmethylsulfonyl fluoride, 2 mM N-α-p-tosyl-l-lysine chloromethyl ketone, 2 mM p-hydroxymercuriphenyl sulfonic acid, and 4 μM leupeptin for 5 min at 95°C. After centrifugation, the supernatant was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Western immunoblot analysis was performed as previously described (43). HRP-conjugated sheep antibody to human IgG F(ab′)2 (ICN Pharmaceuticals) was used as the second antibody. Development was with a Konica HRP-1000 immunostaining kit.
Affinity-purified 260-kDa Gal/GalNAc lectin and the recombinant protein of the cysteine-rich domain in the 170-kDa heavy subunit of the lectin (rLecA), prepared as previously described (28), were used for a dot immunoblot analysis. One microgram of each of these antigens in PBS was spotted onto a nitrocellulose membrane and air dried. Each spot on the membrane was blocked for 30 min with PBS containing 3% skim milk and then incubated with 2 μg of each Fab for 1 h at room temperature. As controls, normal human Fab (OEM Concepts, Toms River, N.J.), antibody purified from the plasma of a patient with an amebic liver abscess, and rabbit antibody to the 260-kDa lectin were used. After the membrane was washed three times with PBS-Tween, it was incubated with HRP-conjugated sheep antibody to human IgG F(ab′)2 or HRP-conjugated goat antibody to rabbit IgG (ICN Pharmaceuticals). After washing, each antibody-bound dot was detected with a Konica HRP-1000 immunostaining kit.
Affinity measurement of purified Fabs by surface plasmon resonance was carried out by using a BIAcore 3000 instrument (Biacore AB, Uppsala, Sweden) according to the general procedure outlined by the manufacturer. Affinity-purified 260-kDa Gal/GalNAc lectin and rLecA were immobilized onto a CM5 chip (Biacore) surface at a low density by amine coupling chemistry. Affinity constants were determined by using the software provided by the manufacturer, BIAevaluation 3.1.
Adherence of E. histolytica to Chinese hamster ovary (CHO) cells was evaluated as previously described (49). Briefly, trophozoites (104 cells) of the HM-1:IMSS strain were incubated with 100 μg of each Fab for 1 h at 4°C, washed with ice-cold PBS, and then suspended in Ham's F12 nutrient mixture containing 1% adult bovine serum. As a control, normal human Fab (OEM Concepts) was used. The trophozoites and CHO cells (2 × 105 cells) were suspended in 1 ml of the Ham's F12 nutrient mixture, centrifuged, and then incubated at 4°C for 2 h. After removal of 0.8 ml of supernatant, the pellet was gently vortexed, and the number of trophozoites with at least three adherent CHO cells was determined by examining 300 trophozoites.
Erythrophagocytosis of E. histolytica was assayed as previously described (47). Briefly, trophozoites (104 cells) of HM-1:IMSS were exposed to 100 μg of each Fab at 37°C for 30 min and then washed with ice-cold PBS. Normal human Fab (OEM Concepts) was also used as a control. The erythrophagocytosis assay was performed by mixing the Fab-treated trophozoites with 106 human erythrocytes and then incubating the preparation at 37°C for 5 min. After lysis of free and adherent erythrocytes by addition of distilled water, the trophozoites were fixed and stained with a 3,3′-diaminobenzidine solution containing hydrogen peroxide. The number of erythrocytes ingested was determined by examining 300 trophozoites.
The nucleotide sequence data reported here have been deposited in the DDBJ, EMBL, and GenBank databases under accession numbers AB095272 to AB095291.
The combinatorial immunoglobulin gene library constructed from peripheral lymphocytes of an asymptomatic E. histolytica cyst passer contained 9.6 × 106 clones. When 6.32 × 104 clones were screened by colony blotting, 6 clones (0.0095%) showed positive signals. In the second screening performed by IFA with intact cells, two clones were positive. Since one of the positive clones, designated CP33, was reactive with the 260-kDa Gal/GalNAc lectin in a preliminary dot immunoblot analysis, this clone was analyzed further.
To find other heavy- and light-chain genes which constitute antilectin Fab fragments with light- and heavy-chain genes of CP33, the libraries from the asymptomatic cyst passer (CP library) and the amebic liver abscess patient (LA library) were screened again after shuffling of heavy- and light-chain genes of CP33 with genes from the two libraries. Many positive signals were detected after chain shuffling. When the light-chain gene of CP33 was shuffled, the positive rates in the colony blot screening were 10- or 20-fold higher than those in the shuffling analysis of the heavy-chain genes (1.92 and 0.19%, respectively, for the LA library; 0.96 and 0.047%, respectively, for the CP library). In addition, when the LA library was used as the source of immunoglobulin genes, the positive rates were two- or fourfold higher than those obtained with the CP library (1.92 and 0.96%, respectively, for light-chain shuffling; 0.19 and 0.047%, respectively, for heavy-chain shuffling). All positive clones were secondarily screened with an IFA by using formalin-fixed trophozoites of E. histolytica HM-1:IMSS, and then the genes of IFA-positive clones were compared by restriction enzyme digestion. Based on the digestion patterns, 12 clones of the light-chain genes and four clones of the heavy-chain genes were identified as different from each other and also different from the light- and heavy-chain genes of CP33. Clones that showed the same digestion pattern as CP33 but were derived from the LA library were also selected. When more than two clones showed the same digestion pattern, the clone with the strongest IFA reactivity was selected. The reactivities of 18 selected clones to E. histolytica antigen were compared by ELISA. Three of 13 Fabs obtained by shuffling of light chains, CP33-H/L-CP17, CP33-H/L-CP26, and CP33-H/L-LA22, showed relatively high reactivity compared with the reactivity of CP33. The relative optical densities for these Fabs were 1.61. 1.26, and 1.42, respectively, when the optical density of CP33 was defined as 1. Two of the light chains (L-CP17 and L-CP26) were derived from the cyst passer, and one (L-LA22) was derived from the patient with a liver abscess. The reactivities of the five Fabs obtained by shuffling of the heavy chain were similar to or lower than the reactivity of CP33; the relative optical densities were between 0.71 and 0.95.
Six heavy-chain genes and 14 light-chain genes were sequenced, and the deduced amino acid sequences that they encode were compared. In the heavy chains, H-LA5 derived from the patient with a liver abscess was identical to CP33-H (Fig. (Fig.1).1). H-CP4 also was almost identical to CP33-H except for the FR1 region. In L-LA22, the sequence of complementarity-determining regions (CDRs) was identical to that in CP33-L (Fig. (Fig.2).2). The difference between L-CP17 and CP33-L in the CDRs was only one amino acid residue in CDR3. In L-CP26, three residues in CDR1 and four residues in CDR3 differed from residues in CP33-L. The sequence homology of these clones with germ lines was analyzed by using IgBLAST at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/igblast/) and V-QUEST at the international ImMunoGeneTics database (http://imgt.cines.fr:8104/textes/vquest/). As shown in Table Table1,1, sequence analysis of the heavy-chain genes revealed that all of the most similar germ line of V-segments belonged to the VH3 family (VH3-21, VH3-30, VH3-48, or VH3-53). However, the level of amino acid homology with the germ line was low, 76% in CP33-H and H-LA5. The most similar germ line for the D-segment was D2-2 or D6-6, except for H-CP1, which was not identified, and the most similar germ line for the J-segment was either JH4b or JH6b. In contrast, all the light chains belonged to the Vκ1 family, in which the most similar Vκ germ line was 02/012 or L5, with relatively high homology (Table (Table2).2). For the Jκ segments, Jκ1, Jκ2, and Jκ5 were combined with 02/012, and Jκ1, Jκ2, and Jκ4 were combined with L5.
CP33 and the three Fabs which showed high reactivity in ELISA (CP33-H/L-CP17, CP33-H/L-CP26, and CP33-H/L-LA22) were purified by His affinity column chromatography. SDS-PAGE demonstrated that the molar ratio of two bands with apparent molecular masses of 25 and 24 kDa was 1:1, suggesting the heterodimeric structure of Fab. These four Fabs reacted with all 10 strains of E. histolytica trophozoites, but not with E. dispar trophozoites, in IFA when fixed cells were used (data not shown). To identify the E. histolytica antigen recognized by these Fabs, a Western immunoblot analysis was performed. Under nonreducing conditions, four Fabs were reactive only with a 260-kDa antigen (Fig. (Fig.3).3). To identify the antigen, a dot immunoblot analysis was carried out. Four Fabs reacted with the affinity-purified 260-kDa Gal/GalNAc lectin and also with the cysteine-rich domain of the heavy subunit of the lectin (Fig. (Fig.44).
The affinity of the Fabs to the lectin was measured by surface plasmon resonance. The association constants for the four Fabs with the 260-kDa lectin ranged from 1.06 × 108 to 2.85 × 108 M−1 (Table (Table3).3). Although the affinity values for rLecA were low, they ranged from 6.43 × 107 to 1.29 × 108 M−1. CP33-H/L-CP26 and CP33-H/L-LA22 showed higher affinity than CP33 and CP33-H/L-CP17 showed.
To evaluate the function of the four Fabs, their effects on amebic adherence to CHO cells were examined. In a normal human Fab-treated control, the level of adherence of trophozoites to CHO cells was 53.8%. After pretreatment of trophozoites with 100 μg of the antilectin Fabs, the level of adherence significantly decreased by 57 to 65% (P < 0.001). The effect of the Fabs on erythrophagocytosis by E. histolytica was also evaluated. When amebae were pretreated with 100 μg of the Fabs, the number of trophozoites ingesting erythrocytes and the number of erythrocytes ingested were significantly decreased (Table (Table4).4). No significant differences in inhibitory effects on adherence and erythrophagocytosis were seen among the four Fabs.
Our findings demonstrate that neutralizing human antibodies to amebic adherence and erythrophagocytosis can be prepared from an immunoglobulin gene library derived from an asymptomatic cyst passer infected with E. histolytica but not with E. dispar. The epitope recognized by CP33 was located in a cysteine-rich domain of the heavy subunit of the Gal/GalNAc lectin. To date, seven different mouse monoclonal antibodies specific for nonoverlapping epitopes on the cysteine-rich domain of the heavy subunit of lectin have been identified (28, 35). Three of the seven murine antibodies inhibited amebic adherence to target cells, but two enhanced adherence by causing a marked increase in the galactose-binding activity of the lectin. Since the human antibodies prepared in this study had an inhibitory effect on amebic adherence to CHO cells, these antibodies must recognize an adherence-inhibiting epitope in the cysteine-rich domain (28). This conclusion is also supported by the fact that CP33 did not react with E. dispar, as previously it has been shown that the adherence-inhibiting epitopes are E. histolytica specific (28, 33). Demonstration of the ability of the Fabs to inhibit erythrophagocytosis indicates involvement of the Gal/GalNAc lectin in this process for the first time. Erythrophagocytosis is of interest as it is a characteristic property that distinguishes E. histolytica from the nonpathogenic parasite E. dispar.
In a previous study, the immunogobulin gene library derived from a patient with an amebic liver abscess was screened by the methods used in the present study. The positive rate of the first screening was 0.054% (27 of 5 × 104 of clones were positive), which is 5.7-fold higher than the rate (0.0095%) observed in this study (9). However, there was only one positive clone in the second screening by IFA with intact trophozoites. This suggests that the proportion of antibodies recognizing the trophozoite surface in the symptomatic patient with a liver abscess was smaller than the proportion in the asymptomatic cyst passer, even though the anti-E. histolytica antibody titer was higher in the symptomatic patient.
In the present study, when the heavy or light chain of CP33 was recombined with genes from the two libraries and rescreened, the positive rates were higher in the LA library. This may have been due to the symptomatic patient with the amebic liver abscess having a high antibody titer. However, the relative values were decreased to 2- or 4-fold from 5.7-fold. We concluded from these results that asymptomatic cyst passers have a high ratio of antibodies recognizing the adherence-inhibiting epitope of the heavy-subunit lectin compared with the ratio in symptomatic patients. These adherence-inhibiting antibodies may help prevent the invasion of trophozoites into tissues in cyst passers, although no information was obtained in this study concerning antibodies to other adherence-inhibiting epitopes.
When the heavy chain of CP33 was combined with the light chains from the libraries, the positive rates for screening by colony blotting were 10- to 20-fold higher than the positive rates for screening of combinations of light chains of CP33 with heavy chains. This suggests that the heavy chain is more important for the binding of antibodies to the lectin. Indeed, the fact that the V-segment gene sequence of CP33-H contains many somatic mutations and the fact that no gene homologous with CP33-H has been reported are in accord with the observation that CP33 is reactive specifically with E. histolytica. Heavy-chain dominance in determining antigen binding has been demonstrated for antibodies to gp120 and to the reverse transcriptase of human immunodeficiency virus type 1 (HIV-1) (6, 26). In contrast, it has been reported that DNA binding activity is determined by the light chains in human anti-double-stranded DNA IgG Fab clones (41).
The present study revealed that all the most similar V-segment germ lines of the cloned heavy chain belonged to the VH3 family. The VH3 gene family, with 22 functional genes, is the largest of the seven families (VH1 to VH7) and comprises about one-half of the expressed VH repertoire in adult peripheral B cells (13, 29). However, biases in gene family usage of the heavy-chain variable region have been reported in a number of diseases. For example, restricted VH3 germ line gene usage was observed in intravenous immunoglobulin-bound Fabs from a patient with thrombocytopenia that had progressed to systemic lupus erythematosus (31). It is known that VH3 antibodies are also important for defense against a variety of bacteria (1, 39, 40) and viruses (2, 17, 20). Although VH3 antibodies bind to HIV-1 gp120 in HIV-infected late-stage patients, VH3 gene family expression is reduced compared with the expression in healthy donors, but the other two main VH gene families, VH1 and VH4, show no significant variation in expression (14). When the gene usage of another neutralizing anti-E. histolytica lectin Fab, LA-01, which had previously been prepared from an LA library (9), was examined, it was found that the most similar germ lines of LA-01 were VH3-30, D1-26, and JH6c for the heavy chain and Vκ02/012 and Jκ5 for the light chain. This observation also supports the preferential usage of the VH3 gene family for the adherence-inhibiting epitope of the lectin. To our knowledge, this report is the first report demonstrating that VH3 antibodies are important in defense against parasitic infections.
In contrast to heavy-chain gene usage, the light-chain gene repertoire of human antibodies to HIV does not exhibit a family bias (15). In the present study, all 14 light chains from both libraries belonged to the Vκ1 family, in which the closest Vκ germ lines were 02/012 and L5, in spite of the selection of clones showing different patterns after restriction endonuclease digestion. This finding demonstrates that a limited repertoire of light-chain genes is required to create a functional binding site with the heavy chain of CP33. This is in accord with previous reports that some heavy chains prefer to pair with similar light-chain variable regions to form high-affinity binders (3, 21, 30).
In the present study, we found that there was only one amino acid residue that was different in CDR3 when CP33-L and L-CP17 were compared. One of the advantages of recombinant antibody technology is possible modification of the original antibody gene. By introduction of synthetic genetic variability in CDR3, which is an important region in antigen binding, it may be possible to increase the affinity and/or neutralizing activity of the antibody.
Whereas the advantage of a phage display system for the preparation of human antibodies has recently been demonstrated (5, 19, 36), the present study shows that screening by colony blotting and chain shuffling of cloned genes may be a useful way to find genes of immunoglobulins with high affinity to pathogens. Total analysis of antibody genes for the amebic lectin, including other adherence-inhibiting epitopes on the heavy and intermediate subunits of the lectin, should be helpful not only for understanding the mechanism of protective immunity but also for development of immunoprophylaxis against invasive amebiasis.
We thank W. Stahl for reviewing the manuscript.
This work was supported by a grant-in-aid for scientific research from the Japanese Society for the Promotion of Science, by NIH grant AI-26649, and by grants from the Ministry of Health, Labour and Welfare of Japan.
Editor: J. M. Mansfield