Modification of myosin subfragment 1 tryptophans by dimethyl(2-hydroxy-5-nitrobenzyl)sulfonium bromide

Modification of tryptophanyl residues (Trps) of myosin subfragments 1 (S-1) was performed with dimethyl(2-hydroxy-5-nitrobenzyl)sulfonium bromide (DHNBS). Under controlled conditions, pH 6 at 0 degrees C and 10-min reaction with 10-100-fold molar excess, K+(EDTA) activity was reduced down to less than half, whereas Ca2+-ATPase activity increased and acto-S-1-ATPase was not affected. The number of modified Trps (up to 2.5) agreed well with the number of 2-hydroxy-5-nitrobenzyl moieties incorporated in S-1. The thiol groups of S-1 were not affected up to 50-fold molar excess of DHNBS, thus indicating that the modification was selective for Trps. The modification of as few as one Trp caused a blue shift of the emission spectrum, accompanied by a reduction in the fluorescence quantum yield. The accessibility of Trps to the fluorescence quencher acrylamide is drastically reduced upon modification, indicating that DHNBS-reactive Trps are more "exposed" than the DHNBS-refractive ones. DHNBS modification did not seem to affect the ATP-induced tryptophan fluorescence enhancement of S-1. The effect of DHNBS modification of the intrinsic fluorescence of S-1 indicates that the modified Trps are located in a polar environment and that they may be identical with the long-lifetime Trps of Torgerson [Torgerson, P. (1984) Biochemistry 23, 3002-3007]. The most reactive Trp is located in the N-terminal 27-kDa fragment of the S-1 heavy chain. It might also be inferred from the above data that the nonexposed and ATP-perturbed Trp(s) is (are) located in the 50-kDa fragment.     

Interaction of 75-106 actin peptide with myosin subfragment-1 and its trypsin modified derivative

To explore the role of a hydrophobic domain of actin in the interaction with a myosin chain we have synthesized a peptide corresponding to residues 75-106 of native actin monomer and studied by fluorescence and ELISA the interaction (13+/-2.6x10(-6) M) with both S-1 and (27 kDa-50 kDa-20 kDa) S-1 trypsin derivative of myosin. The loop corresponding to 96-103 actin residues binds to the S-1 only in the absence of Mg-ATP and under similar conditions but not to the trypsin derivative S-1. Biotinylated C74-K95 and I85-K95 peptide fragments were purified after actin proteolysis with trypsin. The C74-K95 peptide interacted with both S-1 and the S-1 trypsin derivative with an apparent Kd(app) of 6+/-1.2x10(-6) M in the presence or absence of nucleotides. Although peptide fragment I85-K95 binds to S-1 with a Kd(app) of 12+/-2.4x10(-6) M, this fragment did not bind to the trypsin S-1 derivative. We concluded that the actin 85-95 sequence should be a potential binding site to S-1 depending of the conformational state of the intact 70 kDa segment of S-1.     

Epitopes on proliferating cell nuclear antigen recognized by human lupus autoantibody and murine monoclonal antibody

The immune epitopes of proliferating cell nuclear antigen (PCNA), also called cyclin, were analyzed by determining the reactivity between PCNA peptide fragments and anti-PCNA antibodies from lupus patients, murine monoclonal antibody (19A2), and rabbit anti-NH2-terminal peptide antibody. Limited digestion of PCNA/cyclin with Staphylococcus aureus V8 protease resulted in several peptide fragments. Five fragments of 30, 20, 15, 14, and 13 kDa were reactive with rabbit anti-NH2-terminal peptide antibody denoting that they contained the NH2-terminal peptide. The 30- and 20-kDa fragments reacted with 19A2 but the others did not. Lupus sera reacted with 17- and 15-kDa peptide fragments allowing their classification into three groups. Two of eight sera (type A) reacted only with the 17-kDa fragment. Two others (type B) reacted with both the 17- and 15-kDa fragments and the remaining four sera (type C) reacted only with the 15-kDa fragment. The sera reacting with the 15-kDa fragment also reacted with the 20-kDa fragment, but the sera reactive only with the 17-kDa fragment did not, indicating that the 17-kDa fragment was not a degradation product of 20-kDa fragments. The 19A2 epitope resided in the region between 15 and 20 kDa from the NH2 terminus, whereas there was at least one distinct epitope on each 15- and 17-kDa peptide, which were recognized by lupus autoantibodies.     

Characterization of the isolated 20 kDa and 50 kDa fragments of the myosin head

We have isolated two proteolytic fragments of subfragment 1 (S-1) of myosin from rabbit skeletal muscle. These fragments, identified by their molecular weights of 20 and 50 kDa, may be functional domains that, when isolated, retain their specific function. We have studied several structural and functional features of the 20 and 50 kDa fragments. Considerable secondary structure in both fragments has been observed in CD spectrum studies. Previously CD spectra showed 64% ordered structure for the 20 kDa fragment (Muhlrad and Morales, M.F. (1984) Proc. Natl. Acad. Sci. 81, 1003) and here we show 71% ordered structures for the 50 kDa fragment. Fluorescence lifetime studies of tryptophan residues in the 50 kDa fragment and 1,5-IAEDANS-labeled SH-1 in the 20 kDa fragment are used to investigate the tertiary structure of the fragments. We find the tertiary structure relating to this measurement of both fragments to be intact; however, the reaction of 1,5-IAEDANS with SH-1 on the isolated 20 kDa fragment is less specific than with S-1. Furthermore, the fragments showed a tendency to aggregate. The domain concept of S-1 was supported by the characteristic biochemical function of the isolated fragments. Both of the fragments were effective in competing with S-1 for binding to actin in acto-S-1 ATPase measurements. From these studies and in direct binding measurement the 20 kDa fragment proved to bind with higher affinity to actin than did the 50 kDa fragment.     

Anti-TNP antibody localization of the reactive lysine residues in myosin

Myosin contains reactive lysine residues which are trinitrophenylated by 2,4,6-trinitrobenzene sulfonate much faster than the rest of the lysines. Here we find the location of these residues in the primary and spatial structure of myosin with the help of an anti-trinitrophenyl antibody. This antibody was raised against trinitrophenyl hemocyanin in rabbits. It reacted with trinitrophenylated myosin, and with some of the tryptic fragments of trinitrophenylated myosin. By analyzing the reaction with Western blots, it was found that the antibody preferentially reacts with the 27 kDa N-terminal fragment of the myosin head, and more weakly with the light meromyosin region of the myosin rod. The 27 kDa fragment contains the most reactive lysine residue, while the intermediate lysine residue is located in the light meromyosin region. The locations of the epitopes of the antibody were visualized on electron microscope images of rotary-shadowed trinitrophenylated myosin-antibody complexes. The distances of the epitopes to the head-rod junction of myosin were measured as 13 and 113 nm for the epitope on the head (reactive lysine residue) and for that on the rod (intermediary reactive lysine residue), respectively.     

Spectrofluorimetric studies on C-terminal 34 kDa fragment of caldesmon

Analysis of the tryptophan fluorescence emission spectra of caldesmon and its 34 kDa C-terminal fragment indicates that all tryptophan residues are located on the surface of the molecule, accessible to solvent. All three tryptophan residues of the 34 kDa fragment and four of the five tryptophan residues of intact protein are accessible to free water, whereas one located in the N-terminal region of molecule is accessible only to bound water molecules. The temperature dependence of the fluorescence parameters indicates higher thermal stability of the 34 kDa fragment than the whole caldesmon molecule. The interaction of the 34 kDa fragment of caldesmon (like that of the intact molecule) with calmodulin is accompanied by a blue shift of the fluorescence emission maximum and an increase in the relative quantum yield. Computer-calculated binding constants show that the binding of calmodulin to the 34 kDa fragment (K = 2.5 x 10(5) M-1) is of two orders of magnitude weaker than that to intact caldesmon (K = 1.4 x 10(7) M-1). The interaction with tropomyosin results in a blue shift of the spectrum of the 34 kDa fragment, yet there is no effect on the spectrum of intact caldesmon. Binding constants of tropomyosin to caldesmon (K = 3.8 x 10(5) M-1) and its 34 kDa fragment (K = 2.3 x 10(5) M-1) are similar. Binding of calmodulin to caldesmon and to the 34 kDa fragment affects their interaction with tropomyosin.     

Structural aspects of myosin subfragment 1 as studied by subtilisin digestion of trypsin-split S-1 (27 kDa-50 kDa-20 kDa)

Limited subtilisin digestion of myosin subfragment 1 (S-1) was carried out, varying the enzyme: substrate weight ratio from 1:200 to 1:10, and changes in structure, and in the MgATPase activities of S-1 and acto-S-1 after proteolysis, were followed. When the starting material--tryptically-cleaved S-1 (27 kDa-50 kDa-20 kDa) ("split S-1")--was subjected to further subtilisin digestion, it was found that with increasing enzyme concentration, the 50 kDa fragment degraded into an 18 kDa fragment via a 33 kDa peptide (50----33----18 kDa), which was not cross-linked with F-actin. On the other hand, the 27 and 20 kDa fragments were rather stable at lower subtilisin concentrations and started to degrade only at higher subtilisin concentrations. These degradations lowered the MgATPase activities of S-1 and acto-S-1. The losses of MgATPase activities of S-1 and of acto-S-1 were mainly due to the degradations of the 27 and 20 kDa fragments, respectively. Addition of EDTA did not affect the subtilisin cleavage pattern of split S-1 but the breakdown of the 50 kDa fragment was extremely depressed, suggesting that some conformational change of the 50 kDa fragment is induced by the binding of divalent cation. The binding of MgADP to split S-1 accelerated the degradation of the 27 kDa fragment and produced a new cut in the 27 kDa fragment (27----20 kDa), resulting in a further loss of the S-1 MgATPase activity.(ABSTRACT TRUNCATED AT 250 WORDS)     

Monoclonal antibodies distinguish between the head portions of two myosin isozymes from chicken breast muscle

Monoclonal antibodies against chicken breast myosin and its subfragment-1(S-1) were produced. One antibody, 2G41, reacted with S-1 containing a light chain 3 (LC3), but not with another S-1 containing a light chain 1 (LC1) or a mixture of the light chains. A structural difference can be assumed to exist between the head portions of the two myosin isozymes. Antigenicity of S-1 toward 2G41 could not be detected after tryptic digestion into three fragments of 50K, 27K, and 20K daltons. Another monoclonal antibody, M68, was obtained from mice immunized with myosin. M68 preferably recognized the heavy chain from S-1 containing LC3 rather than that from that containing LC1 or S-1. M68 reacted with the 27K fragment among the three.     

Localization and synthesis of an antigenic determinant of herpes simplex virus glycoprotein D that stimulates the production of neutralizing antibody.

An antigenic determinant capable of inducing type-common herpes simplex virus (HSV)-neutralizing antibodies has been located on glycoprotein D (gD) of HSV type 1 (HSV-1). A peptide of 16 amino acids corresponding to residues 8 to 23 of the mature glycoprotein (residues 33 to 48 of the predicted gD-1 sequence) was synthesized. This peptide reacted with an anti-gD monoclonal antibody (group VII) previously shown to neutralize the infectivity of HSV-1 and HSV-2. The peptide was also recognized by polyclonal antibodies prepared against purified gD-1 but was less reactive with anti-gD-2 sera. Sera from animals immunized with the synthetic peptide reacted with native gD and neutralized both HSV-1 and HSV-2.     

Involvement of C-terminal 14 residues of alkali light chain in binding to the heavy chain of myosin

The alkali light chain of rabbit skeletal muscle myosin, A1, was cyanylated with 2-nitro-5-thiocyanobenzoic acid, and the peptide bond at Cys 177 was subsequently cleaved in the presence of 0.05 M CaCl2. Two peptide fragments, from the N-terminal to the residue 176 (CF1) and from the residue 177 to the C-terminal (CF2), were obtained. The CD spectrum and the difference UV absorption spectrum induced by CaCl2 suggested that CF1 largely retained the higher order structure of A1. The CF1 fragment, however, could neither incorporate subfragment-1 (S-1) by an exchange reaction, nor bind with the renatured 20K fragment of S-1 heavy chain. On the other hand, the C-terminal fragment of 14 residues, CF2, could bind with the 20K fragment of S-1 heavy chain. These results indicate that the binding site of the alkali light chain for the heavy chain of myosin is located within the C-terminal 14 residues.     

Chicken leg muscle alpha-connectin as studied by a monoclonal antibody to the 1200 kDa fragment.

Chicken leg gracilis muscle contained only alpha-connectin (ca 3000 kDa) without beta-connectin. When myofibrils were kept standing for 20 hr at 4 degrees C, alpha-connectin was degraded to beta-connectin (ca 2000 kDa) and 1200 kDa peptide. The latter was prepared from myofibrils and purified by gel filtration in the presence of SDS. A monoclonal antibody, alpha 7, to this 1200 kDa fragment was prepared. The antibody reacted with the 1200 kDa fragment and its mother molecule alpha-connectin, but not with beta-connectin. Immunoelectron microscopy using alpha 7, as well as other antibodies to chicken breast muscle beta-connectin, revealed that the 1200 kDa peptide covered the portion of alpha-connectin from the Z line to the N2 line region in the I band of chicken leg gracilis muscle sarcomeres. The results were in good agreement with those observed in rabbit skeletal muscle.     

Serological homologies between H1 degrees and H5 include the carboxyl-terminal domain

We have reported previously that antibodies to chicken H5 and antibodies to H1 both cross-react with mammalian H1 degree (Mura, C. V., and Stollar, B. D. (1981) J. Biol. Chem. 256, 9767-9769). The antigenic sites in H1 degree recognized by these antibodies were analyzed using immunoblotting. Peptides of H1 degree were prepared by partial digestion with acetic acid and tested for reactivity with: 1) antibodies induced by H5 alone, which reacted primarily with the central globular region of H5; 2) antibodies induced by H5 X RNA complexes, which reacted with this domain as well as the basic COOH-terminal domain; and 3) antiserum to calf thymus H1. Anti-H5 antibodies (anti-globular region) cross-reacted with H1 degree peptides that co-migrated with peptides of H5 that contain the globular region, but did not cross-react with H1. Anti-H5/RNA antibodies (anti-globular + anti-COOH-terminal) cross-reacted with these peptides and, in addition, with a lysine-rich H1 degree peptide that co-migrated with the basic COOH-terminal H5 peptide. This H1 degree peptide, but not the putative globular H1 degree peptides, was also recognized by an antiserum to calf H1 which was primarily reactive with the large, COOH-terminal N-bromosuccinimide fragment of calf H1. A weaker cross-reaction between this antiserum and the carboxyl-terminal domain of H5 could be visualized when large quantities of H5 were used in immunoblots. The results indicate that structural homologies between H5 and H1 degree extend beyond the globular region and into the lysine-rich carboxyl-terminal domain. Antigenic homologies between H1 degree and H1 are also at least partially localized in this domain. H1 degree is serologically intermediate between H5 and H1.     

Blocking of human sperm-zona interaction by monoclonal antibodies to a glycoprotein family (ZP4) of porcine zona pellucida.

To study zona pellucida antigens involved in human fertilization, five monoclonal antibodies (MAbs)--2A1, 2G3, 4A2, 4E12, and 5H4--were produced to a glycoprotein family (ZP4) isolated from heat-solubilized porcine zonae pellucidae. Each MAb reacted not only with solubilized porcine zona glycoproteins but also with the glycoproteins deglycosylated by trifluoromethanesulfonic acid treatment. They also reacted with intact zonae pellucidae of porcine and human oocytes. Three (4A2, 4E12, and 5H4) of the five MAbs showed a significant blocking effect on human sperm binding and penetration of human zonae pellucidae. The 5H4 MAb showed a strong reaction with ZP4 and ZP1 glycoprotein families of porcine zonae pellucidae, and four other MAbs reacted more strongly with ZP3 than with ZP4. The reactivity of 5H4 with porcine zona glycoproteins was destroyed by chymotrypsin digestion, but the antigen epitope was resistant to proteolysis by trypsin and endoproteinase Lys-C. A peptide fragment reactive to 5H4 was isolated by reverse-phase HPLC from endoproteinase Lys-C-treated ZP4 glycoproteins, and its molecular mass was determined to be 7 kDa by SDS-PAGE. These results suggested that the antigen epitope corresponding to 5H4 is a good candidate for development of a contraceptive vaccine.     

Site-directed mutagenesis of a soluble recombinant fragment of platelet glycoprotein Ib alpha demonstrating negatively charged residues involved in von Willebrand factor binding.

We have expressed in mammalian cells a fragment (residues 1-302) of the alpha chain of platelet glycoprotein (GP) Ib containing the von Willebrand factor- (vWF) binding site. The secreted soluble protein had an apparent molecular mass of 45 kDa and reacted with conformation-dependent monoclonal antibodies that bind only to native GP Ib, thus demonstrating its proper folding. After insolubilization on nitrocellulose membrane, the recombinant GP Ib alpha fragment bound soluble vWF in the presence of ristocetin or botrocetin with a dissociation constant similar to that exhibited by GP Ib.IX complex on platelets. Moreover, the interaction was blocked by anti-GP Ib monoclonal antibodies known to inhibit vWF binding to platelets. The sequence of GP Ib alpha between residues 269-287 has a strong net negative charge due to the presence of 10 glutamic or aspartic acid residues; 5 of these are contained in the sequence of a synthetic peptide (residues 251-279) previously shown to inhibit vWF-platelet interaction. In order to evaluate the possible functional role of these acidic residues, we employed site-directed mutagenesis to express two mutant GP Ib alpha fragments containing asparagine or glutamine instead of aspartic or glutamic acid, respectively. Mutant 1, with substitutions between residues 251-279, failed to bind vWF whether in the presence of ristocetin or botrocetin; in contrast, vWF binding to Mutant 2, with substitutions between residues 280-302, was nearly normal in the presence of ristocetin, but markedly decreased in the presence of botrocetin. Thus, mammalian cells transfected with a truncated cDNA sequence encoding the amino-terminal domain of GP Ib alpha synthesize a fully functional vWF-binding site; acidic residues in the sequence 252-287 are essential for normal function.     

Mapping of distinct structural domains on microtubule-associated protein 2 by monoclonal antibodies

Monoclonal antibodies against microtubule-associated protein 2 (MAP2) were prepared and their specificity was verified by visualization of the antigens using the antibody overlay technique and by radioimmunoassay. MAP2 was cleaved by alpha-chymotrypsin to generate a series of high-molecular-mass fragments ranging between 270 and 140 kDa. The precursor-product relationship of these fragments was suggested from the rate of their appearance and from the analysis of the tryptic peptide map of each fragment. A group of monoclonal antibodies was found to react predominantly with the intact 270-kDa MAP2 molecule and a fragment having a mass of 240 kDa and to some extent with a 215-kDa fragment. Another group of monoclonal antibodies reacted with an antigenic determinant which was located on the 270-kDa molecule as well as on fragments as small as 140 kDa. None of the two groups of monoclonal antibodies reacted with the microtubule-binding domain of MAP2. These results suggest that one group of antibodies reacts with sites located at or dependent upon a terminal 60-kDa domain(s) distal to the microtubule-binding site of MAP2. The second group of antibodies, which can still bind to smaller proteolytic products, appear to be associated with the central region of the MAP2 molecule. Indirect immunofluorescence experiments with the antibody preparations indicated that at least some of the antigenic determinants are exposed when MAP2 is associated with microtubules in the cell body and neurite outgrowths of differentiated rat brain neuroblastoma B104 cells.     

Involvement of the 50-kDa peptide of myosin heads in the ATPase activity revealed by fluorescent modification with 4-fluoro-7-nitrobenz-2-oxa-1,3-diazole

The fluorescent reagent 4-fluoro-7-nitrobenz-2-oxa-1,3-diazole (NBD-F) reacted specifically with 1.9 lysyl residues/mol of the myosin subfragment-1 (S-1) ATPase. When 1.9 lysyl residues were modified, the K+- and Ca2+-ATPase activities were almost completely inhibited, whereas the Mg2+-ATPase activity was increased to 180% of original activity. The actin-activated Mg2+-ATPase activity was decreased to 30% of original activity by this modification. However, affinity of S-1 for actin in the presence of ATP was unchanged. The NBD fluorescence of the modified S-1 was quenched on addition of ATP, suggesting that ATP induced conformational changes around the NBD groups attached to S-1. Tryptic digestion of the modified S-1 revealed that the NBD groups are attached mainly to the 50-kDa peptide of S-1, more precisely the 45-kDa peptide. These results confirm the recent reports that the 50-kDa peptide of S-1 is involved in the myosin ATPase reaction (K?rner, M., Thiem, N. V., Cardinaud, R., and Lacombe, G. (1983) Biochemistry 22, 5843-5847; Hiratsuka, T. (1986) Biochemistry 25, in press).     

Mimotope ELISA for detection of broad spectrum antibody against avian H5N1 influenza virus

Background

We have raised a panel of broad spectrum neutralizing monoclonal antibodies against the highly pathogenic H5N1 avian influenza virus, which neutralize the infectivity of, and afford protection against infection by, most of the major genetic groups of the virus evolved since 1997. Peptide mimics reactive with one of these broad spectrum H5N1 neutralizing antibodies, 8H5, were identified from random phage display libraries.

Method

The amino acid residues of the most reactive 12mer peptide, p125 (DTPLTTAALRLV), were randomly substituted to improve its mimicry of the natural 8H5 epitope.

Result

133 reactive peptides with unique amino acid sequences were identified from 5 sub-libraries of p125. Four residues (2,4,5.9) of the parental peptide were preserved among all the derived peptides and probably essential for 8H5 binding. These are interspersed among four other residues (1,3,8,10), which exhibit restricted substitution and probably could contribute to binding, and another four (6,7,11,12) which could be randomly substituted and probably are not essential for binding. One peptide, V-1b, derived by substituting 5 of the latter residues is the most reactive and has a binding constant of 3.16×10⁻⁹ M, which is 38 fold higher than the affinity of the parental p125. Immunoassay produced with this peptide is specifically reactive with 8H5 but not also the other related broad spectrum H5N1 avian influenza virus neutralizing antibodies. Serum samples from 29 chickens infected with H5N1 avian influenza virus gave a positive result by this assay and those from 12 uninfected animals gave a negative test result.

Conclusion

The immunoassay produced with the 12 mer peptide,V1-b, is specific for the natural 8H5 epitope and can be used for detection of antibody against the broad spectrum neutralization site of H5N1 avian influenza virus.     

The N-terminal region of the plasma membrane Ca(2+) pump does not separate from the main catalytic fragments after proteolysis

The purified plasma membrane Ca(2+) pump (PMCA) was digested with trypsin, and the proteolytic products were identified by immunoblotting with monoclonal antibodies JA9 or 5F10 directed against the extreme N-terminal segment and the central portion of the molecule, respectively. After a short treatment with low concentrations of the protease, JA9 reacted predominantly with a peptide of 35 kDa whereas 5F10 detected a peptide of 90 kDa. The trypsin cut leading to the production of these fragments had no effect on the maximal activity of the enzyme. At higher concentrations of trypsin, JA9 detected a main fragment of 33 kDa and smaller fragments of 19 and 15 kDa. The persistence of fragments reacting with JA9 indicates that the N-terminal region containing its epitope (residues 51-75) was not easily accessible to the protease in the native PMCA. However, the reactivity with JA9 was rapidly lost during proteolysis of the denatured protein. The passage of the mixture of PMCA fragments through a calmodulin-Sepharose column resulted in the retention of the N-terminal 35 kDa fragment together with that of 90 kDa, despite the fact that only the latter binds calmodulin. The ethylenediaminetetraacetic acid (EDTA) eluate, which contained about equal amounts of both fragments, had a Ca(2+) ATPase activity similar to that of the intact enzyme. The tight association between the two peptides was evidenced by the fact that concentrations of polyoxyethylene 10 lauryl ether (C(12)E(10)), sodium dodecyl sulfate (SDS) high enough for inactivating the enzyme and dissociate the pump from calmodulin were unable of breaking the interaction between the 35 and 90 kDa fragments. Altogether, these results show that after digestion with trypsin, the N-terminal portion of the PMCA, including the extreme N-terminal segment, remains part of a fully functional catalytic complex.     

Expression cloning of Pneumocystis carinii antigens.

We undertook expression cloning of Pneumocystis carinii antigens to overcome the difficulties encountered in purification of these antigens. Using monoclonal antibodies to the P. carinii gp120 antigen and polyclonal rabbit antiserum to rat-derived P. carinii, we have isolated cDNA clones encoding immunoreactive moieties. A cDNA clone encoding the 3' portion of a 45-55 kDa antigen of rat-derived P. carinii, was the most abundant clone isolated. The peptide encoded by this cDNA has a novel sequence with a repeated motif rich in glutamic acid residues. Affinity-purified antibodies to this peptide reacted with the 45-55 kDa band of rat-derived P. carinii. The fusion protein was recognized by serum antibodies from rats with natural exposure to P. carinii. The production of this recombinant protein should allow more detailed studies of the host-parasite relationship of this important opportunistic infection.