- Email: [email protected]
Study of an antigenic site of human serum albumin with monoclonal antibodies
0161.5890!83/050549-07SO3
~W~olrwlarIw~rnunolog~ Vol. 20, No. 5, pp. 549-555, 1983. Prmted m Great Britain.
004
Pergamon Prcsa Ltd.
STUDY OF AN A~T~GE~IC SITE OF HUMAN SERUM ALBUMIN WITH ~O~~CLO~AL ANTIBODIES CLAUDE
LAPRESLE
Service d’lmmunochimie (Received
22 Noaevthrr
and
des Protbines, 1982; accepted
NOELLE Institut
DOYEN Pasteur.
in recised.form
Paris, France
14 January
1983)
Abstract--Two monoclonal anti-HSA antibodies, HA1 and HA2, have been shown to be specific for a univalent fragment of 6000 mol. wt, Fl, located near the C-terminus of HSA (Doyen, Pesce & Lapresle. I/nn~unoloy~ Lerlrrs 3, 365-370, 1981). Both monoclonal antibodies have been shown to react with the same site, which includes the following components: the last small loop of HSA (558-567). the disulfide bridge 514-559 and the residue 570. This site is as available on HSA and Fl, but partially masked on the ‘Inhibitor’ fragment from which Fl derives. Polyclonal anti-F1 antibodies, purified from rabbit sera or mouse ascites by affinity chromatography. react with the same site as HA1 and HA2. However. polyclonal antibodies are heterogenous, most probably because they consist of anti-F1 specific antibodies and of antibodies specific against other parts of the albumin molecule which cross-react with Fl. In addition. monoclonal antibodies can recognize the mutation of a single amino acid residue in the albumin molecule
Analysis of the antigenic structure of human serum albumin by degradation with several enzymes has demonstrated that the albumin molecule possesses multiple antigenic sites with distinct specificities (Lapresle, 1955 ; Lapresle, 1959; Lapresle et al., 1959). Two fragments, designated ‘Inhibitor’ and Fl, which carry only some of the antigenic sites of albumin, have been isolated. The ‘Inhibitor’ was obtained from HSA by degradation with rabbit cathepsin D, and has a mol. wt of 11,000. Further degradation of the ‘Inhibitor’ by trypsin gave rise to fragment Fl, with a mol. wt of 6000 (Lapresle & Webb, 1960; Webb & Lapresle, 1964; Lapresle & Webb, 1965). The amino-acid sequence of both fragments, ‘Inhibitor’ and Fl, has been established and they have been located in the albumin molecule (Fig. 1). The ‘Inhibitor’ is composed of a chain extending from residues 496 to 585, which is the C-terminal of albumin. There is an additional minor form, which differs from the preceding one by a few residues at the Nterminal end. The ‘Inhibitor’ has two intrachain disulfide bridges, giving rise to a large and a small loop which correspond to the last sub-domain of albumin. Fragment Fl is composed of two peptides chains: an M chain (546-573) containing an intrachain disulfide bond and a fi (501.-521) or a y chain (501-519) linked to the c( chain by a disulfide bridge (Bellon & Lapresle, 1975; Walker, 1976).
The valence of fragment Fl was deduced by analysis after ultracentrifugation, of the complexes formed, in the presence of an excess of antibodies, between albumin or Fl and antibodies isolated from rabbit anti-HSA sera on insolubilized Fl. These experiments, have shown that only one molecule of antibody combines with one molecule of HSA or of Fl, which indicates that Fl contains only one of the antigenic determinants of the albumin molecule (Lapresle & Webb, 1965). Recently two monoclonal anti-HSA antibodies, HA 1 and HA2, have been shown to be specific for fragment Fl (Doyen er al., 1981). The present study was undertaken to further define the site which reacts with monoclonal antibodies. In addition, the specificity of the monoclonal antibodies was compared with that of the Fl specific polyclonal antibodies obtained in rabbits and mice.
MATERIALS
AND
METHODS
Proteins and fragments
The human serum albumin (HSA) used was either fraction V from Squibb prepared by ethanol fractionation and kindly given to us by the American Red Cross, or 100% pure crystallised albumin prepared by electrophoresis (Mann). Fragment A299_-585was prepared by CNBr degradation of the HSA molecule, according to the method of Doyen & Lapresle (1979).
550
CLAUDE
LAPRESLE
496
and NOELLE
DOYEY
Sepharose 4B-Fl conjugate. The procedure was the same as for monocionai antibodies. 573
585
Fig. I. Schematic represcntntton of the structures of the ‘Inhibitor and of Fl fragment. ‘Inhibitor’ ( ): Fl ): disullide hndgcs (0~ -0) Sl4~-55Y and 5% 567. (
The ‘Inhibitor’ was prepared by degradation of human serum albumin with rabbit cathepsin D. as described by Webb & Lapresie (1964). The CNBr degradation of ‘Inhibitor’ was performed as described by Meioun & Kusnir (1971). The ‘Inhibitor’, at a concentration of 17 mg,‘mi. was treated with CNBr (at a 50 M excess per methionine residue). The citraconylation of ‘Inhibitor’ and of CNBr-treated ‘Inhibitor’ was performed as described by Atassi cjt rrl. (1976); the ‘Inhibitor’ was used at a concentration of 20 mg/mi. in the presence of a 2 n/l excess of citraconic anhydride per free amino group. Fragment Fl was prepared from trypsintreated ‘Inhibitor’, as described by Lapresie & Goldstein (1969). Chains r. p and 1’ were prepared from fragment Fl reduced and not aikyiated. by eiectrophoresis in the presence of /&mercaptoethanoi, according to the method of Beiion & Lapresie (1975). The aikyiation of reduced fragment Fl and of chains X, [j and 7 with iodoacetamide was performed as described by Koib et al. (1974). Slow variant and normal albumins were prepared from a serum of bisaibuminemia type B, as described by Lapresie (1977).
Two iyophilized ascite fluids with monocionai antibodies against HSA were kindly given to us by the Biotest Serum Institute. GmbH (Frankfurt, Germany). The two antibodies designated HA1 and HA2 were purified by affinity chromatography on Sepharose 4B-HSA conjugate. as described by Doyen ef lrl. (198 1) The antibodies were eiuted with giycine-HCi 0.2 Ri buffer, pH 2.8. Rabbit anti-HSA sera were prepared with Squibb albumin. as previously described (Lapresie & Doyen, 197.5). Mice anti-HSA ascite fluids were obtained by immunisation of A:j mice with Squibb albumin. following the method of Tung & Nisonoff (1975). Rabbit and mouse polycionai anti-F1 antibodies were purified from set-a or ascite fluids by affinity chromatography on
Human serum albumin was labeled with alkaline phosphatase by the method of Aurameas (1969), as described by Doyen (,r (11. (1981). Buffer A contained 0.01 M potassium phosphate pH 7.4, 0.15 M NaCi and 0. I 0,I (V v 1 Tween 20. In addition to these components. buffer B contained 0.5”,, (v/v) gelatin. The titration plates, with 96 Aat bottom \veiis (1.0 x 0.6 cm). were purchased from Linbro division, Flow Laboratories. The inhibition assay, which used antibodic, fixed to a solid support. was an enzyme-iinkcd immunoadsorbent assay modification of that described by Engvaii & Perimann (19721. ~iiquots of 0.2 ml of antibodies (2.5 /c&/ml) in 0.1 M sodium carbonate buffer. pH 9.5. were placed in each well of ;I Limbro EIA microtitration plate, incubated at 37 C for I hr. then overnight at 4 C. The coated piatex ~vcrc washed six times with bufier A. Several conculltrations of inhibitors (albumin. fragments of HSA, antibodies) in 0.1 ml of bufIer R \+‘erc placed in wells and incubated for I hr ;II 37 (‘. Then 0.1 ml of a solution of labeled albumin (0.5 /lg.imi) was added. The mixtures were incubated for 2 hr at 37 C. and the plates uashcd six times with buffer A. The amount of iabcied antigen fixed on the plate was dctcrmined bh adding 200 /II of IO ’ AI p-nitrophenh I phosphate in 0.1 M Tris-HCi buffer pH 8, containing 1.5 M NaCi. The reaction was stopped by the addition of 50 ,uI of 1 .U K,HPO, and the color was read at 405 nm. For each inhibitor concentration, the per cent inhibition was calculated as follows: “(, inhibition
=
10 - IL x 100 I0
where I,, is the absorbancy in the absence of inhibitor, and I, the absorbancy in the presence of inhibitor.
RESULTS
Plates were coated with HA1 or HA2 antlbodies and allowed to react with aikaiinc phosphatase labeled albumin. The inhibition 01 these reactions by HA1 and HA2 was studied. Fipure AGE 2 shows that the reaction of HA1 i>
Structure
NANCMOLES
C3= INHlBlTOR
of an Antigenic
Site of Human
551
Albumin
PER ASSAY
Fig. 2. Inhibition of HAI or HA2 by HAI or HA2. Polystjrene plates were coated with (a) HAI, (b) HA2. The inhibitors were added. followed by alkaline phosphatasc labeled HSA. Inhibitors: HAI (V--V), HA2 (‘(I--V).
inhibited by equal molar amounts of HA1 or HA2 Similar results were obtained for the inhibition of HA2. No inhibition was observed with IgG from non-immunized mice. +__+__
’ qo-’ NANOMOLFS
The reaction of immobilized monoclonal antibodies with labeled albumin was inhibited by fragment Fl and by its chains M, j? or y. Figure 3 shows that all chains are inhibitors. However, to obtain 50% inhibition it was necessary to use about 6 times more CIchain, 56 times more j3 chain and 112 times more y chain than fragment Fl.
ii,.. IO-’ ---
L
10.’
IO-
IO-’
1
._____
1
__*-#giL_--
10^' NANOMQLES
I
:
40-1
1°F
OF
lNHl6tTQR
,o-’
PER
1
ASSAY
Fig. 3. Inhibition of HAI and HA2 by fragment FI and its chains. Polystyrene plates were coated with (a) HAI, (b) HAL The inhibitors were added, followed by alkaline phosph~~t~lse-Iabeled HSA. Inhibitors: Fl (A-A), chain I(* -o).chainB(+---i),chain~(x.---x).
10-
y-’
OF
INIIfRITOP
10.’ PFR
-
J
?C * AsSAY
Fig. 4. Inhibition of HAI or HA2 by normal albumin and by a slow variant isolated from the serum of bisalbuminemia case. Polystyrene plates were coated with (a) HA I. (b) HA2. The inhibitors were added. followed by alkaiine phosphatase-labeled HSA. Inhibitors: normal albumin (m----W), slow albumin (U- -0).
No inhibition was observed with fragment Fl, or with chains CY,fl and 7 reduced by mercaptoethanol and alkylated by iodoacetamide.
The normal albumin and the slow variant were isolated from a serum of bisalbuminemia type B. Their capacity to inhibit the reaction of isolated monoclonal antibodies with labeled HSA was compared. Figure 4 shows that the slow variant is a much weaker inhibitor than the normal albumin. To obtain KY!;;il~llibition, it was necessary to use 120 times more slow albumin than the normal one.
Polystyrene plates were coated with HAL or HA2 and allowed to react with tagged albumin. Figure 5a shows the inhibition of these reactions by albumixl and by various fragments. As previously shown, fragments A 299-585 and Fl gave inhibition curves identical to that of albumin. In contrast, the inhibiting capacity of the ‘Inhibitor’ was weaker. To
i57
CLAUDE
LAPRESLE
Fig. 5. Inhibition of HA I by albumm and by various fragmcnts. Polystqrenc plates (a, b) were coated with HAI. The lnhibltorb were added. followed by alkaline phosphataselabclcd HSA Inhibitors: HSA (W--W), Fl (A-~ -A). Ft-agmcnt A,,, ix’ rn -A), ‘Inhibitor’ (~~-0)~ BrcNtreated ‘InhibItor’ (em 0). citraconylated ‘Inhibitor’ BrcN-treated citraconylated ‘Inhibitor’ (X x I. (+ + ).
obtain 50”,, inhibition, it was necessary to use about 15 times more ‘Inhibitor’ than HSA. Figure 5b shows the inhibition curves obtained with the ‘Inhibitor’ modified by chemical procedures. Its inhibiting capacity was slightly increased by CNBr degradation, or by citraconylation and became identical to that of albumin when both procedures were used. The results shown in Fig. 5 were those obtained with HAl: similar results were obtained with HA2.
Polyclonal antibodies specific for Fl were isolated. from the serum of rabbit or from ascites of mice immunized with HSA, by affinity chromatography. Polystyrene plates were coated with polyclonal anti-F1 antibodies and allowed to react with tagged HSA. The reactions were inhibited by either HA1 or HA2. Figure 6 shows that HA1 and HA2 were able to inhibit in a similar manner either rabbit or mouse anti-F1 antibodies. Rabbit anti-F1 antibodies were inhibited by HSA, fragment A,g,_j8,, and fragment Fl. Figure 7a shows that. in order to obtain a 50”;,,
and NOiiLLE
DOYEN
Fig. 6. Inhibition of polqclonal anti-b I ant~bod~cs by HAI and HA2. Polystyrene plates were coated with anti-1 I antibodies isolated from: (a) rabbit scra. (b) mouse axitc\ The inhibitors were added. followed b! alkallnc phosph;~tax-labeled HSA. Inhlbitorh: HAI (‘J ‘i’). HA7 (V VI.
lo-'
lo-'
IO-’
IO-'
NANOMOLES
IO-'
10.’
10.’
IO-'
OF INHIBITOR
PER
,o-’
1
10-q
1
ASSAY
Fig. 7. Inhibition of polyclonal anti-h1 antlbodles h> albumlns and by various fragments. Polqytyrcne plate\ (a.b) were coated with anti-F1 antibodies isolated from rabbit sera. The inhibitors were added. followed bq alkaline phosphatase-labeled HSA. Inhibitors: normal alh~~rnln (mm- -M), slow albumin (L1 0). Frafmcnt A,,,,, iq_ n,. Fl (A cn Al.
Structure of an Antigenic Site of Human Albumin
inhibition, it was necessary to use about 2 times more fragment A299_-585 and 50 times more fragment Fl than HSA. Similar results were obtained with anti-F1 antibodies from mice. Similar inhibition experiments were performed with the slow and normal variants of bisalbuminemia type B (Fig. 7b); the difference between the slow and normal albumins is very small; in order to obtain 50”/, of inhibition, it was necessary to use only 3.5 times more slow than normal albumin. DISCUSSION
Monoclonal antibodies have been used for the immunochemical analysis of only a few proteins: myoglobin (Berzofsky et al., 1980, 1982; East et al., 1982), fibrinogen (Wilner et al., 1982), lyzozyme (Kobayashi et al., 1982). In the case of albumin, two monoclonal antibodies, HA1 and HA2, have been shown to be specific of a small isolated Fl fragment of the albumin molecule (Doyen et al., 1981). These experiments did not permit to decide whether they reacted with the same or two different sites of the albumin molecule. The demonstration, in the present study, that the reaction of HSA with each antibody is inhibited by the other shows that they react with either the same site or with two sites so close that steric hindrance prevents the simultaneous combination of HA1 and HA2. The first eventuality seems the more probable one, because each monoclonal antibody is inhibited by itself and by the other one in an identical manner. Further studies of the antigenic structure of fragment Fl were undertaken with its different chains which were isolated by preparative electrophoresis. Chain M has almost the same inhibiting capacity as the Fl fragment and must be the main component of the FI site. Chains /? and y also have inhibitory activities but they are much weaker. These activities could be due to contamination by traces of chain CI. Such contamination is dilKcult to rule out for chain /I since it has an electrophoretic mobility intermediate between that of chain a and that of chain y, (Bellon et al., 1975) but contamination of chain y is highly unlikely. Therefore these activities suggest that the chains are part of the Fl site. The greater activity of chain p as compared to chain y may also indicate that the two additional C-terminal residues (520-521) of chain fr are important for the structure of the Fl site.
553
The absence of inhibitory activity of the reduced and alkylated chain c( shows that its intrachain disulfide bridge is essential for the preservation of site Fl. The same observation. made with alkylated chain p or y, is more dificult to understand since they have only one cysteine residue. However, when chains fi and y are not alkylated, they are under the form of dimers (Bellon et al., 197.5).The effect of alkylation indicates that dimerisation may occur through the formation of a disulfide bond essential for the immunological activity of the chains. In fragment Fl such a bond, which exists between the unique cysteine of chain i) or y and a cysteine of chain a, must be involved in the structure of the Fl site. Hereditary bisalbuminemia is a very rare situation, in which two forms of serum albumin can be separated by electrophoresis, one normal and one variant. The most frequent variant is the B type, in which a glutamic acid is replaced by a lysine (Winter et ul., 1972) at position 570 (Gitlin & Gitlin, 1975; Brown, 1977). This mutation gives rise to a slow variant of albumin. The much weaker inhibitory capacity of the slow variant, compared to the normal one, shows that residue 570 is involved in the Fl site. According to these results, the Fl site should comprise the small loop of the last albumin subdomain (558-567), the disulfide bond 514-559, the residue 570 and possibly the residues 520-521. These amino acids may be either part of the Ff site, or essential to its conformation. Sakata & Atassi (1980) have Iocalized five major antigenic sites on HSA, by extrapolation from bovine albumin and use of synthetic peptides. One of these sites, numbered 6, is located in the Fl region. Its structure is similar to the one proposed in the present work, since it includes the small loop of chain x and its intrachain disulfide bridge. However it comprises neither residue 570, nor the disulfide bridge between chain c( and chain ,/3. It had previously been shown that the reaction of monoclonal antibodies with HSA was inhibited by HSA and FI in a similar manner. In contrast, the ‘Inhibitor’ is weaker than Fl. Since FI is a fragment of the ‘Inhibitor’ obtained by trypsin degradation, this suggests that, in the ‘Inhibitor’, the Fl site is masked by a portion of the molecule which is absent in fragment Fl. The Fl fragment differs from the ‘Inhibitor’ by the absence of the sequence
554
C’LAUDE
LAPRESLE
522-545 and by the absence of the last twelve C terminal residues 574-585. The increase in inh~bjtory capacity of the ‘Inhibitor’ by citraconylation and CNBr degradation suggests that it is the sequence 522-545 which is involved in masking the Fi site in the ‘Inhibitor’. This sequence contains seven lysine residues which react with citraconic anhydride and form the major part of the great loop which is opened by CNBr at the level of methionine 548; in contrast. the last twelve C-terminal residues of the ‘Inhibitor’ contain only one lysine residue and no methionine. It is interesting to note that this masking does not occur in HSA and in fragment A, probably because in these cases, the large loop is rn~~~nt~~ined in a different ~onfigur~~tion by the neighboring parts of the molecules. It has been demonstrated in previous studies that rabbit anti-HSA antibodies purified with Fl by affinity chromatography, react with only one site on Fl and on albumin (Lapresle & Webb. 1965; Bellon er cl/., 1975). It seems interesting to compare the properties of these polyclonal anti-F1 antibodies with those of the monoclonal HAI and HA2 antibodies. The displacement of polyclonal anti-F1 antibodies by HA1 or HA2 indicates that both categories of a~ltibodies react with a ~ornlllon structure. This is in agreement with the fact that the structure which reacts with monoclonal antibodies is similar to that which was previously found (Bellon et al., 1975) to react with polyclonal antibodies. In contrast with what is observed with HAi and HA2, the reaction of the polyclonal anti-F1 ai~tibodies with the albumin molecule, is not inhibited in an identical manner by 299_585, and fragment Ft ; albumin, fragment A in the latter case, the inhibiting capacity increases with the size of the fragment and is maximal with albu~lli~. This difference cannot be due to a difference of species between rabbit polyclonai and mouse monoclonal antibodies since similar results have been obtained with polyclonal antibodies from rabbit or mice. A possible explanation might be that the region of HSA. which reacts with anti-F1 antibodies. is altered in fragment Fl. However this alteration cannot involve the site reacting with HA1 and HA2, but other hypothetical sites. A second explanation, which does not exclude the previous one, is that the polyclonal anti-F1 antibodies consist of antibodies specific for the Fl site as well as of antibodies specific for
and NOi?LI_E
DOYEN
other parts of the albumin molecule which arc isolated with Fl because they can cross-react with the Fl site. This intr~~rnole~~il~lr crossreactivity has been demonstr~~ted in the citsc 01 the albumin molecule (Doyen tpl i/i,. 19~ I. 1982). These explanations are in apparent disagreement with the previous observation (Lapresle & Webb. 1965) that in an excess 01’ antibody. only one mole of anti-F1 combines with one mole of HSA. However it is possihlc that in the first hypothesis of mLlltiple sites on Fl they are close enough to prevent the Jixntion of more than one mole of antibody pc~ mole of albumin; moreover traces 01’ he:i~~ complexes were observed in the l~ltr~~~~~~tri~l~~~ which could be made of ~tlbllI~ir1 tnoloculcs combined with ~lntibodies specifc for the f.‘l site and with other anti-albumin antihodic>. either reacting with site(s) of I-ISA altered on Fl, or cross-reacting with Fl. Polyclonal anti-F1 antibodies dialer NIX> from fIIA1 and HA2 by a smaller capacity to recognize the mutation of the 570 residue rcsponsihle for the type B bisalbumittemi~~. This suggests that. for some of the polyclonal antibodies, the 570 residue is not part of the reacting site. This is in agreement w-ith both above mentioned hypotheses. In the first one. the 570 residue might not belong to some of the difftrent sites which react with anti-F1 antibodies; in the second one, some of the cross-reacting antibodies might not react with 570 residue. Finally a comparison between the monospecific polyclonal antibodies obtained by allinit) chromatography and monoclonal antibodies. demonstrates the superiority of the latter I~I the de~nition of the fine structure of il~di~i~~~l~ll antigenic sites in a protein molecule. In addition, monoclonal antibodies can be an ellicient tool to demonstrate the muta\ion oC a single residue and could thus bc used to int‘cstigate genetic pol~rnor~~~is~~ in proteins. Aci\noa’lcti~Pi,Jetlli.s We at-c indchted to Dr. I’inc f0r typing the serum of a case of hisalbumincmia iypt’ 13. WC acknowledge the skillful techmcal arsistuncc of Dominique Leduc and Josette Carillon.
REFEREhIES
Structure
of an Antigenic
fragments of human serum albumin and their location in the albumin molecule. Biodxm. J. 147, 585.-592. B&on F.. Lapresie C. & Escribano M. J. (1975) Etude de la structure antig~nique d’un fragment de s&rum albumine humaine. Ann. Itmnun. (Inst. Pasteur) 126C, 653-667. Berrofsky J. A., Buckenmeyer G. K.. Hicks G.. Curd F. R. N.. Feldmann R. J. & Minna J. (1982) Topographic antigenic determinants recognized by monoclonal antibodies to sperm whale myoglobin. J. hio[. Chrm. 257, 3189-3198. Bertofsky A. J., Hicks G., Fedorko J. & Minna J. (1980) Properties of monoclonal antibodies specific for determinants of a protein antigen, myoglobin. J. hiol. ChaUl. 255, I I I 88-I I 19 I,
Doyen N. & Lapresle C. (1979) PropriCt& immunochimiques des fragments obtenus par degradation de la s&rum albumine humaine avec le bromure de cyanogtne. Ann. Irlii?rltrr. (Inst. Pasreur) 1.3oc. 322-334. Dovcn N.. Pcsce A. J. & Lapresle C. (1981) Specificity of two anti-human albumin monoclonal antibodies. itnmurz. Leir. 3, 365 370. Doyen N., Pesce A. J. & Lapresle C. (1982) Immunochemical crabs-reactivity between cyanogen bromide fragments of human serum albumin. J. hid. C1zm. 257, 2770-2774. East 1. J., Hurrell K. G. R., Todd P. E. E. & Leach S. J. ( 1VtiZtAntigenic specificity of monoclonal antibodies to human myoglobin. J. hi&. C&n. 257, 3 199-3202. Engvall E. 61 Perlmann P. (1972) Enzyme linked immunosorbent assav, ELISA. III. Ouantitation of snecific antibodies by en>yme-labelled anti-immunoglob;lin in antigen coated tubes. J. 1~zn~~f. 109, i 29-I 35. Gitlin D. & Gitlm J. D. (197.5) Gcz~~ic .4irt*rcrficms in he P/CI.MIIN Prorc+tm of’ Mm. T/w Plrrsmu Prnteins (Edi ted by Putnam F. W.). 2nd edn, Vol. II. pp. X-374. Academic Press. New York. Kobayashi T.. Fujio H.. Kondo K., Dohi Y., Hirayama A.. Takagaki Y.. Kosaki G. & Amano T. (1982) A monoclonal antibody speciiic for a distinct region of hen egg white lysozyme. &f&c. ~~}~~~z{~~. 19, 6 19-630. Koib E.. Harris J. I. & Bridgen J. (1974) Triose phosphate isomerase from the Coclacanth. An approach to the rapid determination of an amino acid sequence with small amounts of material. Biochrm. J. 137, 185-197. Lapresle C. (1955) Etude de la d&gradation de la s&rum albumme humaine par un extrait de rate de fapin. II. Mise en evidence de trois groupements spkcifiques difftrents dans Ic motif antigCmque de I’albumine humaine et de trois anticorps correspondants dans le s&rum de
Site of Human
Albumin
555
lapin antialbumine humaine. Ann. 1n.\t. Postwr. Ptrri.5 89, 654-665. Lapresle C. (1959) HCtt-rog&n&t& des anticorps antiprot&es. Afzn. Imi. Fttsiriir. Puri.S!R3626635. Lapresle C. (1977) Location of the albumin Gainesville mutation in the iV-terminal quarter of the albumin molecule. FEBS Lat. 76, 204-206. Lapresle C. & Doyen N. (1975) Isolation and properties of a fragment of human serum albumin demonstrating the absence of a methionine residue from some of the albumin molecules. Bj~~~~~7.J. 151, 637-643. Lapresle C. & Goldstein 1. J. (1969) Immunogenicity of a fragment of human serum albumin. d. I~l~nurl. 102, 733-742. Lapresle C.. Kaminsky M. & Tanner C. E. (1959) Immunochemicai study of the enzymatic degradation of human serum albumin: an analysis of the antigenic structure of a protein molecule. J. Immure. 82, 94-102. Lapresle C. & Webb T. (1960) Etude dc la degradatton de la s&rum albumine humainc par un extrait de rate de lapin. ilnn. Inst. Pusreur. Prrris 99, 523-532. Lapresle C. & Webb T. (1965) Isolation and study of a fragment of human serum albumin containing one of the antigen&z sites of the whole molecule. Bio~&tlt. J. 95, 245-25 1, Meloun B. & Kiisnir J. (1971) Cleavage of human plaama albumin by cyanogen bromide and characterization of the fragments. Co//n Czrclz. L’/~UXCotnon~t~. Etly/. E&I 37. 2812-2816. Sakata S. & Atassi M. 2. (1980) Irnrnt~no~~~e~~~stry of serum albumin. Five major antigenic sites of human serum albumin are extrapolated from bovme albumin and confirmed by synthetic peptides. .!4olc,c. [~~UUUT.17, 139-142. Tung, A. S. and Nisonoff A. (1975) Isolation from individual .A:j mice of anti-I’-“zophenylarsonate antibodies bearing a cross-reactive idiotype. J. r\-p. Mrd. 141. 112~126. Walker J. E. (1976) The amino acid sequence of a fragment of human serum albumin containing two of its antigenic determinants. Eur. J. Biochenz. 69, 517 526. Webb T. & Lapresle C. (1964) Isolation and study of rabbit antibodies specific for certain of the antigenic groups of human serum albumin. Biochem. J. 91, 24-31. Wilner G. D.. Mudd M. S.. Hsieh K. H. & Thomas D. W. (1982) Monoclonal antibodies to fibrinogen: modulation of determinants expressed in fibrinogen by 7 chain crosslinking. ~i~~~~~~~?~j.s~~~ 21, 2687.2692. Winter W. P.. Weitkamp L. R. & Rucknagel D. L. (1972) Amino acid substitution in two identical inherited human serum albumin variants: albumin Oliphant and albumin Ann Arbor. Biochrnrr,str~~ 1 I. 889-896.
~W~olrwlarIw~rnunolog~ Vol. 20, No. 5, pp. 549-555, 1983. Prmted m Great Britain.
004
Pergamon Prcsa Ltd.
STUDY OF AN A~T~GE~IC SITE OF HUMAN SERUM ALBUMIN WITH ~O~~CLO~AL ANTIBODIES CLAUDE
LAPRESLE
Service d’lmmunochimie (Received
22 Noaevthrr
and
des Protbines, 1982; accepted
NOELLE Institut
DOYEN Pasteur.
in recised.form
Paris, France
14 January
1983)
Abstract--Two monoclonal anti-HSA antibodies, HA1 and HA2, have been shown to be specific for a univalent fragment of 6000 mol. wt, Fl, located near the C-terminus of HSA (Doyen, Pesce & Lapresle. I/nn~unoloy~ Lerlrrs 3, 365-370, 1981). Both monoclonal antibodies have been shown to react with the same site, which includes the following components: the last small loop of HSA (558-567). the disulfide bridge 514-559 and the residue 570. This site is as available on HSA and Fl, but partially masked on the ‘Inhibitor’ fragment from which Fl derives. Polyclonal anti-F1 antibodies, purified from rabbit sera or mouse ascites by affinity chromatography. react with the same site as HA1 and HA2. However. polyclonal antibodies are heterogenous, most probably because they consist of anti-F1 specific antibodies and of antibodies specific against other parts of the albumin molecule which cross-react with Fl. In addition. monoclonal antibodies can recognize the mutation of a single amino acid residue in the albumin molecule
Analysis of the antigenic structure of human serum albumin by degradation with several enzymes has demonstrated that the albumin molecule possesses multiple antigenic sites with distinct specificities (Lapresle, 1955 ; Lapresle, 1959; Lapresle et al., 1959). Two fragments, designated ‘Inhibitor’ and Fl, which carry only some of the antigenic sites of albumin, have been isolated. The ‘Inhibitor’ was obtained from HSA by degradation with rabbit cathepsin D, and has a mol. wt of 11,000. Further degradation of the ‘Inhibitor’ by trypsin gave rise to fragment Fl, with a mol. wt of 6000 (Lapresle & Webb, 1960; Webb & Lapresle, 1964; Lapresle & Webb, 1965). The amino-acid sequence of both fragments, ‘Inhibitor’ and Fl, has been established and they have been located in the albumin molecule (Fig. 1). The ‘Inhibitor’ is composed of a chain extending from residues 496 to 585, which is the C-terminal of albumin. There is an additional minor form, which differs from the preceding one by a few residues at the Nterminal end. The ‘Inhibitor’ has two intrachain disulfide bridges, giving rise to a large and a small loop which correspond to the last sub-domain of albumin. Fragment Fl is composed of two peptides chains: an M chain (546-573) containing an intrachain disulfide bond and a fi (501.-521) or a y chain (501-519) linked to the c( chain by a disulfide bridge (Bellon & Lapresle, 1975; Walker, 1976).
The valence of fragment Fl was deduced by analysis after ultracentrifugation, of the complexes formed, in the presence of an excess of antibodies, between albumin or Fl and antibodies isolated from rabbit anti-HSA sera on insolubilized Fl. These experiments, have shown that only one molecule of antibody combines with one molecule of HSA or of Fl, which indicates that Fl contains only one of the antigenic determinants of the albumin molecule (Lapresle & Webb, 1965). Recently two monoclonal anti-HSA antibodies, HA 1 and HA2, have been shown to be specific for fragment Fl (Doyen er al., 1981). The present study was undertaken to further define the site which reacts with monoclonal antibodies. In addition, the specificity of the monoclonal antibodies was compared with that of the Fl specific polyclonal antibodies obtained in rabbits and mice.
MATERIALS
AND
METHODS
Proteins and fragments
The human serum albumin (HSA) used was either fraction V from Squibb prepared by ethanol fractionation and kindly given to us by the American Red Cross, or 100% pure crystallised albumin prepared by electrophoresis (Mann). Fragment A299_-585was prepared by CNBr degradation of the HSA molecule, according to the method of Doyen & Lapresle (1979).
550
CLAUDE
LAPRESLE
496
and NOELLE
DOYEY
Sepharose 4B-Fl conjugate. The procedure was the same as for monocionai antibodies. 573
585
Fig. I. Schematic represcntntton of the structures of the ‘Inhibitor and of Fl fragment. ‘Inhibitor’ ( ): Fl ): disullide hndgcs (0~ -0) Sl4~-55Y and 5% 567. (
The ‘Inhibitor’ was prepared by degradation of human serum albumin with rabbit cathepsin D. as described by Webb & Lapresie (1964). The CNBr degradation of ‘Inhibitor’ was performed as described by Meioun & Kusnir (1971). The ‘Inhibitor’, at a concentration of 17 mg,‘mi. was treated with CNBr (at a 50 M excess per methionine residue). The citraconylation of ‘Inhibitor’ and of CNBr-treated ‘Inhibitor’ was performed as described by Atassi cjt rrl. (1976); the ‘Inhibitor’ was used at a concentration of 20 mg/mi. in the presence of a 2 n/l excess of citraconic anhydride per free amino group. Fragment Fl was prepared from trypsintreated ‘Inhibitor’, as described by Lapresie & Goldstein (1969). Chains r. p and 1’ were prepared from fragment Fl reduced and not aikyiated. by eiectrophoresis in the presence of /&mercaptoethanoi, according to the method of Beiion & Lapresie (1975). The aikyiation of reduced fragment Fl and of chains X, [j and 7 with iodoacetamide was performed as described by Koib et al. (1974). Slow variant and normal albumins were prepared from a serum of bisaibuminemia type B, as described by Lapresie (1977).
Two iyophilized ascite fluids with monocionai antibodies against HSA were kindly given to us by the Biotest Serum Institute. GmbH (Frankfurt, Germany). The two antibodies designated HA1 and HA2 were purified by affinity chromatography on Sepharose 4B-HSA conjugate. as described by Doyen ef lrl. (198 1) The antibodies were eiuted with giycine-HCi 0.2 Ri buffer, pH 2.8. Rabbit anti-HSA sera were prepared with Squibb albumin. as previously described (Lapresie & Doyen, 197.5). Mice anti-HSA ascite fluids were obtained by immunisation of A:j mice with Squibb albumin. following the method of Tung & Nisonoff (1975). Rabbit and mouse polycionai anti-F1 antibodies were purified from set-a or ascite fluids by affinity chromatography on
Human serum albumin was labeled with alkaline phosphatase by the method of Aurameas (1969), as described by Doyen (,r (11. (1981). Buffer A contained 0.01 M potassium phosphate pH 7.4, 0.15 M NaCi and 0. I 0,I (V v 1 Tween 20. In addition to these components. buffer B contained 0.5”,, (v/v) gelatin. The titration plates, with 96 Aat bottom \veiis (1.0 x 0.6 cm). were purchased from Linbro division, Flow Laboratories. The inhibition assay, which used antibodic, fixed to a solid support. was an enzyme-iinkcd immunoadsorbent assay modification of that described by Engvaii & Perimann (19721. ~iiquots of 0.2 ml of antibodies (2.5 /c&/ml) in 0.1 M sodium carbonate buffer. pH 9.5. were placed in each well of ;I Limbro EIA microtitration plate, incubated at 37 C for I hr. then overnight at 4 C. The coated piatex ~vcrc washed six times with bufier A. Several conculltrations of inhibitors (albumin. fragments of HSA, antibodies) in 0.1 ml of bufIer R \+‘erc placed in wells and incubated for I hr ;II 37 (‘. Then 0.1 ml of a solution of labeled albumin (0.5 /lg.imi) was added. The mixtures were incubated for 2 hr at 37 C. and the plates uashcd six times with buffer A. The amount of iabcied antigen fixed on the plate was dctcrmined bh adding 200 /II of IO ’ AI p-nitrophenh I phosphate in 0.1 M Tris-HCi buffer pH 8, containing 1.5 M NaCi. The reaction was stopped by the addition of 50 ,uI of 1 .U K,HPO, and the color was read at 405 nm. For each inhibitor concentration, the per cent inhibition was calculated as follows: “(, inhibition
=
10 - IL x 100 I0
where I,, is the absorbancy in the absence of inhibitor, and I, the absorbancy in the presence of inhibitor.
RESULTS
Plates were coated with HA1 or HA2 antlbodies and allowed to react with aikaiinc phosphatase labeled albumin. The inhibition 01 these reactions by HA1 and HA2 was studied. Fipure AGE 2 shows that the reaction of HA1 i>
Structure
NANCMOLES
C3= INHlBlTOR
of an Antigenic
Site of Human
551
Albumin
PER ASSAY
Fig. 2. Inhibition of HAI or HA2 by HAI or HA2. Polystjrene plates were coated with (a) HAI, (b) HA2. The inhibitors were added. followed by alkaline phosphatasc labeled HSA. Inhibitors: HAI (V--V), HA2 (‘(I--V).
inhibited by equal molar amounts of HA1 or HA2 Similar results were obtained for the inhibition of HA2. No inhibition was observed with IgG from non-immunized mice. +__+__
’ qo-’ NANOMOLFS
The reaction of immobilized monoclonal antibodies with labeled albumin was inhibited by fragment Fl and by its chains M, j? or y. Figure 3 shows that all chains are inhibitors. However, to obtain 50% inhibition it was necessary to use about 6 times more CIchain, 56 times more j3 chain and 112 times more y chain than fragment Fl.
ii,.. IO-’ ---
L
10.’
IO-
IO-’
1
._____
1
__*-#giL_--
10^' NANOMQLES
I
:
40-1
1°F
OF
lNHl6tTQR
,o-’
PER
1
ASSAY
Fig. 3. Inhibition of HAI and HA2 by fragment FI and its chains. Polystyrene plates were coated with (a) HAI, (b) HAL The inhibitors were added, followed by alkaline phosph~~t~lse-Iabeled HSA. Inhibitors: Fl (A-A), chain I(* -o).chainB(+---i),chain~(x.---x).
10-
y-’
OF
INIIfRITOP
10.’ PFR
-
J
?C * AsSAY
Fig. 4. Inhibition of HAI or HA2 by normal albumin and by a slow variant isolated from the serum of bisalbuminemia case. Polystyrene plates were coated with (a) HA I. (b) HA2. The inhibitors were added. followed by alkaiine phosphatase-labeled HSA. Inhibitors: normal albumin (m----W), slow albumin (U- -0).
No inhibition was observed with fragment Fl, or with chains CY,fl and 7 reduced by mercaptoethanol and alkylated by iodoacetamide.
The normal albumin and the slow variant were isolated from a serum of bisalbuminemia type B. Their capacity to inhibit the reaction of isolated monoclonal antibodies with labeled HSA was compared. Figure 4 shows that the slow variant is a much weaker inhibitor than the normal albumin. To obtain KY!;;il~llibition, it was necessary to use 120 times more slow albumin than the normal one.
Polystyrene plates were coated with HAL or HA2 and allowed to react with tagged albumin. Figure 5a shows the inhibition of these reactions by albumixl and by various fragments. As previously shown, fragments A 299-585 and Fl gave inhibition curves identical to that of albumin. In contrast, the inhibiting capacity of the ‘Inhibitor’ was weaker. To
i57
CLAUDE
LAPRESLE
Fig. 5. Inhibition of HA I by albumm and by various fragmcnts. Polystqrenc plates (a, b) were coated with HAI. The lnhibltorb were added. followed by alkaline phosphataselabclcd HSA Inhibitors: HSA (W--W), Fl (A-~ -A). Ft-agmcnt A,,, ix’ rn -A), ‘Inhibitor’ (~~-0)~ BrcNtreated ‘InhibItor’ (em 0). citraconylated ‘Inhibitor’ BrcN-treated citraconylated ‘Inhibitor’ (X x I. (+ + ).
obtain 50”,, inhibition, it was necessary to use about 15 times more ‘Inhibitor’ than HSA. Figure 5b shows the inhibition curves obtained with the ‘Inhibitor’ modified by chemical procedures. Its inhibiting capacity was slightly increased by CNBr degradation, or by citraconylation and became identical to that of albumin when both procedures were used. The results shown in Fig. 5 were those obtained with HAl: similar results were obtained with HA2.
Polyclonal antibodies specific for Fl were isolated. from the serum of rabbit or from ascites of mice immunized with HSA, by affinity chromatography. Polystyrene plates were coated with polyclonal anti-F1 antibodies and allowed to react with tagged HSA. The reactions were inhibited by either HA1 or HA2. Figure 6 shows that HA1 and HA2 were able to inhibit in a similar manner either rabbit or mouse anti-F1 antibodies. Rabbit anti-F1 antibodies were inhibited by HSA, fragment A,g,_j8,, and fragment Fl. Figure 7a shows that. in order to obtain a 50”;,,
and NOiiLLE
DOYEN
Fig. 6. Inhibition of polqclonal anti-b I ant~bod~cs by HAI and HA2. Polystyrene plates were coated with anti-1 I antibodies isolated from: (a) rabbit scra. (b) mouse axitc\ The inhibitors were added. followed b! alkallnc phosph;~tax-labeled HSA. Inhlbitorh: HAI (‘J ‘i’). HA7 (V VI.
lo-'
lo-'
IO-’
IO-'
NANOMOLES
IO-'
10.’
10.’
IO-'
OF INHIBITOR
PER
,o-’
1
10-q
1
ASSAY
Fig. 7. Inhibition of polyclonal anti-h1 antlbodles h> albumlns and by various fragments. Polqytyrcne plate\ (a.b) were coated with anti-F1 antibodies isolated from rabbit sera. The inhibitors were added. followed bq alkaline phosphatase-labeled HSA. Inhibitors: normal alh~~rnln (mm- -M), slow albumin (L1 0). Frafmcnt A,,,,, iq_ n,. Fl (A cn Al.
Structure of an Antigenic Site of Human Albumin
inhibition, it was necessary to use about 2 times more fragment A299_-585 and 50 times more fragment Fl than HSA. Similar results were obtained with anti-F1 antibodies from mice. Similar inhibition experiments were performed with the slow and normal variants of bisalbuminemia type B (Fig. 7b); the difference between the slow and normal albumins is very small; in order to obtain 50”/, of inhibition, it was necessary to use only 3.5 times more slow than normal albumin. DISCUSSION
Monoclonal antibodies have been used for the immunochemical analysis of only a few proteins: myoglobin (Berzofsky et al., 1980, 1982; East et al., 1982), fibrinogen (Wilner et al., 1982), lyzozyme (Kobayashi et al., 1982). In the case of albumin, two monoclonal antibodies, HA1 and HA2, have been shown to be specific of a small isolated Fl fragment of the albumin molecule (Doyen et al., 1981). These experiments did not permit to decide whether they reacted with the same or two different sites of the albumin molecule. The demonstration, in the present study, that the reaction of HSA with each antibody is inhibited by the other shows that they react with either the same site or with two sites so close that steric hindrance prevents the simultaneous combination of HA1 and HA2. The first eventuality seems the more probable one, because each monoclonal antibody is inhibited by itself and by the other one in an identical manner. Further studies of the antigenic structure of fragment Fl were undertaken with its different chains which were isolated by preparative electrophoresis. Chain M has almost the same inhibiting capacity as the Fl fragment and must be the main component of the FI site. Chains /? and y also have inhibitory activities but they are much weaker. These activities could be due to contamination by traces of chain CI. Such contamination is dilKcult to rule out for chain /I since it has an electrophoretic mobility intermediate between that of chain a and that of chain y, (Bellon et al., 1975) but contamination of chain y is highly unlikely. Therefore these activities suggest that the chains are part of the Fl site. The greater activity of chain p as compared to chain y may also indicate that the two additional C-terminal residues (520-521) of chain fr are important for the structure of the Fl site.
553
The absence of inhibitory activity of the reduced and alkylated chain c( shows that its intrachain disulfide bridge is essential for the preservation of site Fl. The same observation. made with alkylated chain p or y, is more dificult to understand since they have only one cysteine residue. However, when chains fi and y are not alkylated, they are under the form of dimers (Bellon et al., 197.5).The effect of alkylation indicates that dimerisation may occur through the formation of a disulfide bond essential for the immunological activity of the chains. In fragment Fl such a bond, which exists between the unique cysteine of chain i) or y and a cysteine of chain a, must be involved in the structure of the Fl site. Hereditary bisalbuminemia is a very rare situation, in which two forms of serum albumin can be separated by electrophoresis, one normal and one variant. The most frequent variant is the B type, in which a glutamic acid is replaced by a lysine (Winter et ul., 1972) at position 570 (Gitlin & Gitlin, 1975; Brown, 1977). This mutation gives rise to a slow variant of albumin. The much weaker inhibitory capacity of the slow variant, compared to the normal one, shows that residue 570 is involved in the Fl site. According to these results, the Fl site should comprise the small loop of the last albumin subdomain (558-567), the disulfide bond 514-559, the residue 570 and possibly the residues 520-521. These amino acids may be either part of the Ff site, or essential to its conformation. Sakata & Atassi (1980) have Iocalized five major antigenic sites on HSA, by extrapolation from bovine albumin and use of synthetic peptides. One of these sites, numbered 6, is located in the Fl region. Its structure is similar to the one proposed in the present work, since it includes the small loop of chain x and its intrachain disulfide bridge. However it comprises neither residue 570, nor the disulfide bridge between chain c( and chain ,/3. It had previously been shown that the reaction of monoclonal antibodies with HSA was inhibited by HSA and FI in a similar manner. In contrast, the ‘Inhibitor’ is weaker than Fl. Since FI is a fragment of the ‘Inhibitor’ obtained by trypsin degradation, this suggests that, in the ‘Inhibitor’, the Fl site is masked by a portion of the molecule which is absent in fragment Fl. The Fl fragment differs from the ‘Inhibitor’ by the absence of the sequence
554
C’LAUDE
LAPRESLE
522-545 and by the absence of the last twelve C terminal residues 574-585. The increase in inh~bjtory capacity of the ‘Inhibitor’ by citraconylation and CNBr degradation suggests that it is the sequence 522-545 which is involved in masking the Fi site in the ‘Inhibitor’. This sequence contains seven lysine residues which react with citraconic anhydride and form the major part of the great loop which is opened by CNBr at the level of methionine 548; in contrast. the last twelve C-terminal residues of the ‘Inhibitor’ contain only one lysine residue and no methionine. It is interesting to note that this masking does not occur in HSA and in fragment A, probably because in these cases, the large loop is rn~~~nt~~ined in a different ~onfigur~~tion by the neighboring parts of the molecules. It has been demonstrated in previous studies that rabbit anti-HSA antibodies purified with Fl by affinity chromatography, react with only one site on Fl and on albumin (Lapresle & Webb. 1965; Bellon er cl/., 1975). It seems interesting to compare the properties of these polyclonal anti-F1 antibodies with those of the monoclonal HAI and HA2 antibodies. The displacement of polyclonal anti-F1 antibodies by HA1 or HA2 indicates that both categories of a~ltibodies react with a ~ornlllon structure. This is in agreement with the fact that the structure which reacts with monoclonal antibodies is similar to that which was previously found (Bellon et al., 1975) to react with polyclonal antibodies. In contrast with what is observed with HAi and HA2, the reaction of the polyclonal anti-F1 ai~tibodies with the albumin molecule, is not inhibited in an identical manner by 299_585, and fragment Ft ; albumin, fragment A in the latter case, the inhibiting capacity increases with the size of the fragment and is maximal with albu~lli~. This difference cannot be due to a difference of species between rabbit polyclonai and mouse monoclonal antibodies since similar results have been obtained with polyclonal antibodies from rabbit or mice. A possible explanation might be that the region of HSA. which reacts with anti-F1 antibodies. is altered in fragment Fl. However this alteration cannot involve the site reacting with HA1 and HA2, but other hypothetical sites. A second explanation, which does not exclude the previous one, is that the polyclonal anti-F1 antibodies consist of antibodies specific for the Fl site as well as of antibodies specific for
and NOi?LI_E
DOYEN
other parts of the albumin molecule which arc isolated with Fl because they can cross-react with the Fl site. This intr~~rnole~~il~lr crossreactivity has been demonstr~~ted in the citsc 01 the albumin molecule (Doyen tpl i/i,. 19~ I. 1982). These explanations are in apparent disagreement with the previous observation (Lapresle & Webb. 1965) that in an excess 01’ antibody. only one mole of anti-F1 combines with one mole of HSA. However it is possihlc that in the first hypothesis of mLlltiple sites on Fl they are close enough to prevent the Jixntion of more than one mole of antibody pc~ mole of albumin; moreover traces 01’ he:i~~ complexes were observed in the l~ltr~~~~~~tri~l~~~ which could be made of ~tlbllI~ir1 tnoloculcs combined with ~lntibodies specifc for the f.‘l site and with other anti-albumin antihodic>. either reacting with site(s) of I-ISA altered on Fl, or cross-reacting with Fl. Polyclonal anti-F1 antibodies dialer NIX> from fIIA1 and HA2 by a smaller capacity to recognize the mutation of the 570 residue rcsponsihle for the type B bisalbumittemi~~. This suggests that. for some of the polyclonal antibodies, the 570 residue is not part of the reacting site. This is in agreement w-ith both above mentioned hypotheses. In the first one. the 570 residue might not belong to some of the difftrent sites which react with anti-F1 antibodies; in the second one, some of the cross-reacting antibodies might not react with 570 residue. Finally a comparison between the monospecific polyclonal antibodies obtained by allinit) chromatography and monoclonal antibodies. demonstrates the superiority of the latter I~I the de~nition of the fine structure of il~di~i~~~l~ll antigenic sites in a protein molecule. In addition, monoclonal antibodies can be an ellicient tool to demonstrate the muta\ion oC a single residue and could thus bc used to int‘cstigate genetic pol~rnor~~~is~~ in proteins. Aci\noa’lcti~Pi,Jetlli.s We at-c indchted to Dr. I’inc f0r typing the serum of a case of hisalbumincmia iypt’ 13. WC acknowledge the skillful techmcal arsistuncc of Dominique Leduc and Josette Carillon.
REFEREhIES
Structure
of an Antigenic
fragments of human serum albumin and their location in the albumin molecule. Biodxm. J. 147, 585.-592. B&on F.. Lapresie C. & Escribano M. J. (1975) Etude de la structure antig~nique d’un fragment de s&rum albumine humaine. Ann. Itmnun. (Inst. Pasteur) 126C, 653-667. Berrofsky J. A., Buckenmeyer G. K.. Hicks G.. Curd F. R. N.. Feldmann R. J. & Minna J. (1982) Topographic antigenic determinants recognized by monoclonal antibodies to sperm whale myoglobin. J. hio[. Chrm. 257, 3189-3198. Bertofsky A. J., Hicks G., Fedorko J. & Minna J. (1980) Properties of monoclonal antibodies specific for determinants of a protein antigen, myoglobin. J. hiol. ChaUl. 255, I I I 88-I I 19 I,
Doyen N. & Lapresle C. (1979) PropriCt& immunochimiques des fragments obtenus par degradation de la s&rum albumine humaine avec le bromure de cyanogtne. Ann. Irlii?rltrr. (Inst. Pasreur) 1.3oc. 322-334. Dovcn N.. Pcsce A. J. & Lapresle C. (1981) Specificity of two anti-human albumin monoclonal antibodies. itnmurz. Leir. 3, 365 370. Doyen N., Pesce A. J. & Lapresle C. (1982) Immunochemical crabs-reactivity between cyanogen bromide fragments of human serum albumin. J. hid. C1zm. 257, 2770-2774. East 1. J., Hurrell K. G. R., Todd P. E. E. & Leach S. J. ( 1VtiZtAntigenic specificity of monoclonal antibodies to human myoglobin. J. hi&. C&n. 257, 3 199-3202. Engvall E. 61 Perlmann P. (1972) Enzyme linked immunosorbent assav, ELISA. III. Ouantitation of snecific antibodies by en>yme-labelled anti-immunoglob;lin in antigen coated tubes. J. 1~zn~~f. 109, i 29-I 35. Gitlin D. & Gitlm J. D. (197.5) Gcz~~ic .4irt*rcrficms in he P/CI.MIIN Prorc+tm of’ Mm. T/w Plrrsmu Prnteins (Edi ted by Putnam F. W.). 2nd edn, Vol. II. pp. X-374. Academic Press. New York. Kobayashi T.. Fujio H.. Kondo K., Dohi Y., Hirayama A.. Takagaki Y.. Kosaki G. & Amano T. (1982) A monoclonal antibody speciiic for a distinct region of hen egg white lysozyme. &f&c. ~~}~~~z{~~. 19, 6 19-630. Koib E.. Harris J. I. & Bridgen J. (1974) Triose phosphate isomerase from the Coclacanth. An approach to the rapid determination of an amino acid sequence with small amounts of material. Biochrm. J. 137, 185-197. Lapresle C. (1955) Etude de la d&gradation de la s&rum albumme humaine par un extrait de rate de fapin. II. Mise en evidence de trois groupements spkcifiques difftrents dans Ic motif antigCmque de I’albumine humaine et de trois anticorps correspondants dans le s&rum de
Site of Human
Albumin
555
lapin antialbumine humaine. Ann. 1n.\t. Postwr. Ptrri.5 89, 654-665. Lapresle C. (1959) HCtt-rog&n&t& des anticorps antiprot&es. Afzn. Imi. Fttsiriir. Puri.S!R3626635. Lapresle C. (1977) Location of the albumin Gainesville mutation in the iV-terminal quarter of the albumin molecule. FEBS Lat. 76, 204-206. Lapresle C. & Doyen N. (1975) Isolation and properties of a fragment of human serum albumin demonstrating the absence of a methionine residue from some of the albumin molecules. Bj~~~~~7.J. 151, 637-643. Lapresle C. & Goldstein 1. J. (1969) Immunogenicity of a fragment of human serum albumin. d. I~l~nurl. 102, 733-742. Lapresle C.. Kaminsky M. & Tanner C. E. (1959) Immunochemicai study of the enzymatic degradation of human serum albumin: an analysis of the antigenic structure of a protein molecule. J. Immure. 82, 94-102. Lapresle C. & Webb T. (1960) Etude dc la degradatton de la s&rum albumine humainc par un extrait de rate de lapin. ilnn. Inst. Pusreur. Prrris 99, 523-532. Lapresle C. & Webb T. (1965) Isolation and study of a fragment of human serum albumin containing one of the antigen&z sites of the whole molecule. Bio~&tlt. J. 95, 245-25 1, Meloun B. & Kiisnir J. (1971) Cleavage of human plaama albumin by cyanogen bromide and characterization of the fragments. Co//n Czrclz. L’/~UXCotnon~t~. Etly/. E&I 37. 2812-2816. Sakata S. & Atassi M. 2. (1980) Irnrnt~no~~~e~~~stry of serum albumin. Five major antigenic sites of human serum albumin are extrapolated from bovme albumin and confirmed by synthetic peptides. .!4olc,c. [~~UUUT.17, 139-142. Tung, A. S. and Nisonoff A. (1975) Isolation from individual .A:j mice of anti-I’-“zophenylarsonate antibodies bearing a cross-reactive idiotype. J. r\-p. Mrd. 141. 112~126. Walker J. E. (1976) The amino acid sequence of a fragment of human serum albumin containing two of its antigenic determinants. Eur. J. Biochenz. 69, 517 526. Webb T. & Lapresle C. (1964) Isolation and study of rabbit antibodies specific for certain of the antigenic groups of human serum albumin. Biochem. J. 91, 24-31. Wilner G. D.. Mudd M. S.. Hsieh K. H. & Thomas D. W. (1982) Monoclonal antibodies to fibrinogen: modulation of determinants expressed in fibrinogen by 7 chain crosslinking. ~i~~~~~~~?~j.s~~~ 21, 2687.2692. Winter W. P.. Weitkamp L. R. & Rucknagel D. L. (1972) Amino acid substitution in two identical inherited human serum albumin variants: albumin Oliphant and albumin Ann Arbor. Biochrnrr,str~~ 1 I. 889-896.