Journal of Neuroimmunology, 36 (1992) 135-146 © 1992 Elsevier Science Publishers B.V. All rights reserved 0165-5728/92/$05.00
135
JNI 02112
An immunochemical comparison of human myelin basic protein and its modified, citrullinated form, C8 John N. Whitaker a,b,d Katharine A. Kirk c, Paula K. H e r m a n a Shan-Ren Z h o u a, Richard R. Goodin d, Mario A. Moscarello e and D. Denise Wood e Departments of a Neurology, h Cell Biology and c Biostatistics and Biomathematics of the University of Alabama at Birmingham, Birmingham, AL, USA, d Neurology and Research Services of the Birmingham Veterans Administration Medical Center, Birmingham, AL, USA, and e Department of Biochemistry of the Hospital for Sick Children, Toronto, Canada (Received 11 July 1991) (Revised, received 30 August 1991) (Accepted 3 September 1991)
Key words: Myelin basic protein; Citrulline; Immunochemistry
Summary An immunochemical analysis was conducted to compare the C1 isomer of human myelin basic protein (MBP) with the newly described and less cationic, citrullinated isomer of MBp referred to as C8. Ten polyclonal antisera directed at multiple epitopes or restricted regions of MBP were used in radioimmunoassays to examine MBP-C1 and MBP-C8. Antisera reactive with MBP peptide 1-14 clearly distinguished MBP-C1 from MBP-C8. Antisera to human MBP peptides 10-19 and 90-170, but not to MBP peptide 69-89, showed modest differences between MBP-C1 and MBP-C8. The MBP-C8s from multiple sclerosis (MS) and non-MS brain reacted essentially the same. With murine monoclonal antibodies and enzyme-linked immunosorbent assay (ELISA), differences between MBP-C8 and other isomers were shown for anti-MBP 10-19 but not for anti-MBP 1-9 or anti-MBP 80-89. These findings imply differences in sequence or conformation in the structure of MBP-C8 compared to MBP-C1, most notably near the amino terminus.
Introduction Because of its known encephalitogenicity in inducing experimental allergic encephalomyelitis, the molecular nature of myelin basic protein (MBP) has been extensively investigated (Lees
Correspondence to: John N. Whitaker, M.D., Department of Neurology, University of Alabama at Birmingham, UAB Station, Birmingham, AL 35294, USA. Tel. 205/934-2402; Fax 205/934-0928.
and Brostoff, 1984). MBP exists in five major isoforms that represent differential splicing of the seven exons of the single MBP gene (Roach et al., 1983; Kamholz et al., 1986; Campagnoni, 1988). The dominant isoform of adult human MBP (exons 1, 3-7) has a molecular weight of 18,500 and consists of 170 amino acid residues. Subsequent to the translation of MBP mRNA, a number of modifications occur to the MBP molecule resulting in a population of differently charged microheterogeneous components or isomers (Moscarello, 1990). There are typically five
13~ isomers of human MBP with the most basic one, which has not been modified except for the amino terminal alanine, designated as CI (Deibler and Martenson, 1973). The other posttranslational modifications include methylation of the arginine at 107 (Baldwin and Carnegie, 1971), phosphorylation of serine or threonine residues, and deamidation of glutamine residues (Deibler et al., 1975; Chou et al., 1976). Although these modifications presumably alter the conformation of the MBP molecule, their biological significance is unknown. C8, a recently described isomer (Wood and Moscarello, 1989), had previously been overlooked because of the manner in which it behaves on cation exchange chromatography. In the standard purification of MBP from acid extracts of delipidated brain (Deibler and Martenson, 1973), material not adhering to the carboxymethylcellulose at pH 10.6 is usually discarded as non-MBPproteins. In examining this fraction, it was found that a form of MBP was present but was extensively modified with the change of six arginine residues to citrulline at amino acid residues 25, 31, 122, 130, 159 and 170. Citrullination markedly changes the charge of the MBP so that it has a lower isoelectric point and expanded areas of hydrophobicity. There is also an increase in glutamic acid/glutamine in C8-MBP (Wood and Moscarello, 1989), but how this alters the primary sequence of MBP-C8 is unknown. This unusual posttranslational modification stimulated a number of studies to elucidate its effect on the structure of the protein. In a secondary structure study by circular dichroism, none of the components contained c~-helical segments. All contained varying amounts of /3-structure. C1 contained 13% and C4 24% /~-structure (Ramwani et al., 1989), while C8 was greater than 30% /3-structure (Moscarello, M.A., unpublished observation). Therefore, C1 and C8 differ both in primary and secondary structures. The present study was undertaken to examine immunochemically the C8 isomer of MBP especially in regard to differences from the unmodified C1 isomer of MBP. The studies indicate that MBP-C8 differs immunochemically from MBP-C1 with particular differences noted in the amino terminal region of those two MBP isomers.
Materials and methods
Preparation of myelin basic protein and its peptides The C1 isomer of MBP was purified from postmortem human brain tissue of two adults who died of nonneurological causes, and guinea pig brain tissue (Pelfreez, Rogers, AR, USA) by previously described methods (Whitaker and Seyer, 1979; Wood and Moscarello, 1989). An MBP fraction, containing a combination of C4 and C5 (Deibler and Martenson, 1973), was prepared by the same procedures (Whitaker and Seyer, 1979). The C8 isomer of MBP was also purified from the fraction of delipidated brain extract which had not been retained on carboxymethylcellulosc at pH 9.6 (Wood and Moscarello, 1989) or pH 10.6 (Whitaker and Seyer, 1979). The eluted first peak was desalted on Sephadex G-25 (Whitaker and Seyer, 1979), lyophilized and then subjected to reverse phase high performance liquid chromatography (RP-HPLC) on a C18 column using a gradient of trifluoroacetic acid-acetonitrile as described (Wood and Moscarello, 1989). Utilizing this procedure, MBP-C8 was isolated from the postmortem brain tissue of five persons with multiple sclerosis (MS) and three who died of nonneurological disease. Four of the non-MS brains and the three MS brains were from the Brain Tissue Bank at the University of Toronto. One of the non-MS brains was from an adult male who died of nonneurological causes and was autopsied at the University of Alabama Hospital. All preparations of MBP were shown by RP-HPLC to give a single peak. This finding and that of polyacrylamide gel electrophoresis (Wood and Moscarello, 1989) illustrated comparable degrees of purity of the samples examined by immunoassay (see below). Human MBP peptides 1-44, 45-89 and 90-170 were prepared from human MBP by cathepsin D degradation (Whitaker and Seyer, 1979). Human MBP peptides 1-14 (nonacetylated form), 1(I-19 and 69-89 were synthesized by Peninsula Laboratories (Belmont, CA, USA).
Preparation of polyclonal antisera Ten different rabbit or sheep antisera were used in this study. The preparation and features
137 of these antisera have been described previously (Whitaker, 1977; Whitaker et al., 1979; Whitaker, 1982). Antisera with reactivities against human MBP peptide 10-19 were tested for reactivity against radiolabeled human MBP peptide 10-19 by radioimmunoassay (Whitaker, 1982, see below). Of the antisera previously described (Whitaker, 1982) only one antiserum, R47 (Whitaker et al., 1975), reacted marginally with MBP peptide 1019 in this assay. All others failed to bind. A new antiserum was obtained by immunizing a sheep (Sl14) with guinea pig MBP-C1 conjugated to ovalbumin at a 2:1 ratio using carbodiimide as previously described (Whitaker, 1982). The sheep was initially immunized with 2 mg guinea pig MBP in complete Freund's adjuvant and subsequently injected at monthly intervals with the same amount of antigen in incomplete Freund's adjuvant. The antiserum used was collected 9 months after the primary immunization. The immunogens as well as the reference for their preparation and characterization are listed in Table 1.
Double-antibody radioimmunoassay A 3-day equilibrium double-antibody radioimmunoassay ( D A - R I A ) was used generally following the conditions described previously (Whitaker and Herman, 1988). On day 1 a dilution of first antibody, inhibitor and radiolabeled antigen were
mixed together in 500/xl and placed at 4°C for 16 h. The amount of inhibitor included varied from 0.1 to 250 ng in order to cover a full range of displacement of radiolabeled antigen by an antiserum. 200/xl of second antibody (goat antirabbit IgG or burro antisheep IgG) was then added. After 24 h, a pellet was collected by centrifugation and its radioactivity measured. H u m a n MBPC1, MBP-C8, and human MBP peptides 1-14, 10-19, 1-44, 45-89, 69-89 and 90-170 were radioiodinated with 125I by lactoperoxidase catalyzed iodination (Whitaker, 1977). Each antiserum was studied at varying dilutions in a 3-day D A - R I A for precipitation of the radiolabeled antigen to select a dilution giving 50% precipitation for subsequent performance of displacement DA-RIA. The antiserum, dilution used and the radioligand of the various immunoassays performed in this study are listed in Table 2.
Preparation and testing of monoclonal antibody to MBP peptides The preparation of monoclonal antibody (mAb) to human MBP peptides acetyl 1-9 and 80-89 has been previously described (Price et al., 1986; Zhou and Whitaker, 1990). Murine mAb to human MBP peptide 10-19 was prepared (Fusion 32 or F32) using the identical procedure (Price et al., 1986). mAbs were examined as to their reactivities by enzyme-linked immunosorbent assay (ELISA) (Whitaker et al., 1989).
TABLE1 PREPARATION OF POLYCLONAL ANTISERA USED Rabbit (R) or sheep (S) antiserum R79 Rll0 R120 R156 R157 R207 R208 S105 Sl14 $253
Immunogen
Reference
Human MBP from MS brain conjugated to rabbit serum albumin Bovine MBP peptides 37-88 and 43-88 conjugated to ovalbumin Guinea Pig MBP Human MBP peptide 90-170 conjugated to ovalbumin Human MBP peptide 90-170 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin Human MBP peptide 1-44 conjugated to ovalbumin Guinea pig MBP-C1 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin
(Whitaker, 1977) (Whitaker, 1982) (Whitaker et al., 1979) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Present paper) (Whitaker, 1982)
138 TABLE 2 DOUBLE-ANTIBODY RADIOIMMUNOASSAYS USED FOR IMMUNOCHEMICAL COMPARISON OF HUMAN MBP-C8 A N D MBP-CI Antiserum
Dilution
Radioligand
R79 R79 R79 R79 R79 RI20 R120 R120 R120 R156 R157 P207 R208 S105 S105 S105 S114 $253 $253 $253
1 : 1500 1 : 3000 1 : 1000 1:500 1:7000 1 : 1000 1 : 1000 l : 1500 1 : 1200 1 : 1000 1:21100 1:2500 1:2000 1:6500 1:6500 1:3000 1 : 1200 1 : 1000 1 : 1000 1 : 20,000
Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human
MBP-CI MBP-C8 MBP peptide MBP peptide MBP peptide MBP-C1 MBP-C8 MBP 1-14 MBP 90 170 MBP peptide MBP peptide MBP peptide MBP peptide MBP-C1 MBP-C8 MBP peptide MBP peptide MBP-C1 MBP-C8 MBP peptide
45-89 69-89 90-170
that the antilog of the log scale EDs0 is the EDs~~ on the natural scale. y=
4 -/3(Iog(concent ration) - log(r))/loge(I0) 1 + 10 ~-
y= 1 + ( c°ncentrati°n ) -tJ'
where /3'= ( ~4/ 1 o g e ( 1 0 ) ) / 3 90-170 90-170 1-14 1-14
1-14 10-19
1-14
Statistical analysis" The mathematical model (Box and Hunter, 1962) assumed for the estimation of the curves was, similar to logit-log plots, that the logarithm of the concentration follows a logistic regression curve. A number of ways exist for relating the parameters for this model. The one selected expresses them in terms of specific characteristics of the plot of the percent inhibition against the logarithm to the base 10 of the concentration. In this form, ¢ represents the maximum inhibition achievable at infinite concentration; /3 represents the slope of the nonlinear curve at the log concentration which yields an inhibition of K/2 (50% inhibition or EDs0 on the log scale). The log (~-) represents the EDs0 on the log scale. The equations for the percent inhibition (y) are given below, for both the log form and the natural scale form. It should be noted that neither/3 nor/3' is the slope at the EDs0 for the natural form, but a more complicated expression. It is, however, true
This mathematical model, with the three parameters of log (~-), /3 and K, was fit to the data for each individual sample in each experiment by nonlinear least squares using the Gauss-Newton algorithm. The parameter estimates and their associated standard errors were calculated. A oneway analysis of variance was conducted to evaluate group differences in these parameters and a multivariate analysis of variance evaluated the significance of the differences for whole curves (both parameters taken simultaneously). Contrasts were formed to evaluate paired comparisons among the three groups of normal MBP-C1, non-MS MBP-C8 and MS MBP-C8. No adjustment for multiple comparisons was made. With only three groups that problem should be minimal. Statistical differences with a p value of _< 0.05 were considered as significant.
Results
The initial comparison of MBP-C8 with MBPC1 was by direct precipitation of radiolabeled MBP and MBP peptides. Antisera (R79 and R120), directed at multiple sites on MBP, precipitated MBP-C8 and MBP-C1 nearly equally (Fig. 1A and B). Some (R207 and $253) antisera preferentially reacted with MBP peptides 1-14 and 1-44 and reacted less well with MBP-C1 or MBP-C8 (Fig. 1C and D). R207 (anti-MBP peptide 1-14), which reacted at low titers with MBPC1, did not react at all with MBP-C8 in the direct precipitation reaction (Fig. 1C). The other anti-
0
10-
20-
30-
40-
50-
60
7O
8O
90
100
0
10-
20-
30-
40-
50
60
70
C
1:500
1:500
1:1000
1:1000
Dilution
1:2000
Dilution
1:2000
1:5000
1:5000
1:10000
R 207
1 : 10~(X)
R 79
8 I:~
c
o.
o
""
o.
co
135
JNI 02112
An immunochemical comparison of human myelin basic protein and its modified, citrullinated form, C8 John N. Whitaker a,b,d Katharine A. Kirk c, Paula K. H e r m a n a Shan-Ren Z h o u a, Richard R. Goodin d, Mario A. Moscarello e and D. Denise Wood e Departments of a Neurology, h Cell Biology and c Biostatistics and Biomathematics of the University of Alabama at Birmingham, Birmingham, AL, USA, d Neurology and Research Services of the Birmingham Veterans Administration Medical Center, Birmingham, AL, USA, and e Department of Biochemistry of the Hospital for Sick Children, Toronto, Canada (Received 11 July 1991) (Revised, received 30 August 1991) (Accepted 3 September 1991)
Key words: Myelin basic protein; Citrulline; Immunochemistry
Summary An immunochemical analysis was conducted to compare the C1 isomer of human myelin basic protein (MBP) with the newly described and less cationic, citrullinated isomer of MBp referred to as C8. Ten polyclonal antisera directed at multiple epitopes or restricted regions of MBP were used in radioimmunoassays to examine MBP-C1 and MBP-C8. Antisera reactive with MBP peptide 1-14 clearly distinguished MBP-C1 from MBP-C8. Antisera to human MBP peptides 10-19 and 90-170, but not to MBP peptide 69-89, showed modest differences between MBP-C1 and MBP-C8. The MBP-C8s from multiple sclerosis (MS) and non-MS brain reacted essentially the same. With murine monoclonal antibodies and enzyme-linked immunosorbent assay (ELISA), differences between MBP-C8 and other isomers were shown for anti-MBP 10-19 but not for anti-MBP 1-9 or anti-MBP 80-89. These findings imply differences in sequence or conformation in the structure of MBP-C8 compared to MBP-C1, most notably near the amino terminus.
Introduction Because of its known encephalitogenicity in inducing experimental allergic encephalomyelitis, the molecular nature of myelin basic protein (MBP) has been extensively investigated (Lees
Correspondence to: John N. Whitaker, M.D., Department of Neurology, University of Alabama at Birmingham, UAB Station, Birmingham, AL 35294, USA. Tel. 205/934-2402; Fax 205/934-0928.
and Brostoff, 1984). MBP exists in five major isoforms that represent differential splicing of the seven exons of the single MBP gene (Roach et al., 1983; Kamholz et al., 1986; Campagnoni, 1988). The dominant isoform of adult human MBP (exons 1, 3-7) has a molecular weight of 18,500 and consists of 170 amino acid residues. Subsequent to the translation of MBP mRNA, a number of modifications occur to the MBP molecule resulting in a population of differently charged microheterogeneous components or isomers (Moscarello, 1990). There are typically five
13~ isomers of human MBP with the most basic one, which has not been modified except for the amino terminal alanine, designated as CI (Deibler and Martenson, 1973). The other posttranslational modifications include methylation of the arginine at 107 (Baldwin and Carnegie, 1971), phosphorylation of serine or threonine residues, and deamidation of glutamine residues (Deibler et al., 1975; Chou et al., 1976). Although these modifications presumably alter the conformation of the MBP molecule, their biological significance is unknown. C8, a recently described isomer (Wood and Moscarello, 1989), had previously been overlooked because of the manner in which it behaves on cation exchange chromatography. In the standard purification of MBP from acid extracts of delipidated brain (Deibler and Martenson, 1973), material not adhering to the carboxymethylcellulose at pH 10.6 is usually discarded as non-MBPproteins. In examining this fraction, it was found that a form of MBP was present but was extensively modified with the change of six arginine residues to citrulline at amino acid residues 25, 31, 122, 130, 159 and 170. Citrullination markedly changes the charge of the MBP so that it has a lower isoelectric point and expanded areas of hydrophobicity. There is also an increase in glutamic acid/glutamine in C8-MBP (Wood and Moscarello, 1989), but how this alters the primary sequence of MBP-C8 is unknown. This unusual posttranslational modification stimulated a number of studies to elucidate its effect on the structure of the protein. In a secondary structure study by circular dichroism, none of the components contained c~-helical segments. All contained varying amounts of /3-structure. C1 contained 13% and C4 24% /~-structure (Ramwani et al., 1989), while C8 was greater than 30% /3-structure (Moscarello, M.A., unpublished observation). Therefore, C1 and C8 differ both in primary and secondary structures. The present study was undertaken to examine immunochemically the C8 isomer of MBP especially in regard to differences from the unmodified C1 isomer of MBP. The studies indicate that MBP-C8 differs immunochemically from MBP-C1 with particular differences noted in the amino terminal region of those two MBP isomers.
Materials and methods
Preparation of myelin basic protein and its peptides The C1 isomer of MBP was purified from postmortem human brain tissue of two adults who died of nonneurological causes, and guinea pig brain tissue (Pelfreez, Rogers, AR, USA) by previously described methods (Whitaker and Seyer, 1979; Wood and Moscarello, 1989). An MBP fraction, containing a combination of C4 and C5 (Deibler and Martenson, 1973), was prepared by the same procedures (Whitaker and Seyer, 1979). The C8 isomer of MBP was also purified from the fraction of delipidated brain extract which had not been retained on carboxymethylcellulosc at pH 9.6 (Wood and Moscarello, 1989) or pH 10.6 (Whitaker and Seyer, 1979). The eluted first peak was desalted on Sephadex G-25 (Whitaker and Seyer, 1979), lyophilized and then subjected to reverse phase high performance liquid chromatography (RP-HPLC) on a C18 column using a gradient of trifluoroacetic acid-acetonitrile as described (Wood and Moscarello, 1989). Utilizing this procedure, MBP-C8 was isolated from the postmortem brain tissue of five persons with multiple sclerosis (MS) and three who died of nonneurological disease. Four of the non-MS brains and the three MS brains were from the Brain Tissue Bank at the University of Toronto. One of the non-MS brains was from an adult male who died of nonneurological causes and was autopsied at the University of Alabama Hospital. All preparations of MBP were shown by RP-HPLC to give a single peak. This finding and that of polyacrylamide gel electrophoresis (Wood and Moscarello, 1989) illustrated comparable degrees of purity of the samples examined by immunoassay (see below). Human MBP peptides 1-44, 45-89 and 90-170 were prepared from human MBP by cathepsin D degradation (Whitaker and Seyer, 1979). Human MBP peptides 1-14 (nonacetylated form), 1(I-19 and 69-89 were synthesized by Peninsula Laboratories (Belmont, CA, USA).
Preparation of polyclonal antisera Ten different rabbit or sheep antisera were used in this study. The preparation and features
137 of these antisera have been described previously (Whitaker, 1977; Whitaker et al., 1979; Whitaker, 1982). Antisera with reactivities against human MBP peptide 10-19 were tested for reactivity against radiolabeled human MBP peptide 10-19 by radioimmunoassay (Whitaker, 1982, see below). Of the antisera previously described (Whitaker, 1982) only one antiserum, R47 (Whitaker et al., 1975), reacted marginally with MBP peptide 1019 in this assay. All others failed to bind. A new antiserum was obtained by immunizing a sheep (Sl14) with guinea pig MBP-C1 conjugated to ovalbumin at a 2:1 ratio using carbodiimide as previously described (Whitaker, 1982). The sheep was initially immunized with 2 mg guinea pig MBP in complete Freund's adjuvant and subsequently injected at monthly intervals with the same amount of antigen in incomplete Freund's adjuvant. The antiserum used was collected 9 months after the primary immunization. The immunogens as well as the reference for their preparation and characterization are listed in Table 1.
Double-antibody radioimmunoassay A 3-day equilibrium double-antibody radioimmunoassay ( D A - R I A ) was used generally following the conditions described previously (Whitaker and Herman, 1988). On day 1 a dilution of first antibody, inhibitor and radiolabeled antigen were
mixed together in 500/xl and placed at 4°C for 16 h. The amount of inhibitor included varied from 0.1 to 250 ng in order to cover a full range of displacement of radiolabeled antigen by an antiserum. 200/xl of second antibody (goat antirabbit IgG or burro antisheep IgG) was then added. After 24 h, a pellet was collected by centrifugation and its radioactivity measured. H u m a n MBPC1, MBP-C8, and human MBP peptides 1-14, 10-19, 1-44, 45-89, 69-89 and 90-170 were radioiodinated with 125I by lactoperoxidase catalyzed iodination (Whitaker, 1977). Each antiserum was studied at varying dilutions in a 3-day D A - R I A for precipitation of the radiolabeled antigen to select a dilution giving 50% precipitation for subsequent performance of displacement DA-RIA. The antiserum, dilution used and the radioligand of the various immunoassays performed in this study are listed in Table 2.
Preparation and testing of monoclonal antibody to MBP peptides The preparation of monoclonal antibody (mAb) to human MBP peptides acetyl 1-9 and 80-89 has been previously described (Price et al., 1986; Zhou and Whitaker, 1990). Murine mAb to human MBP peptide 10-19 was prepared (Fusion 32 or F32) using the identical procedure (Price et al., 1986). mAbs were examined as to their reactivities by enzyme-linked immunosorbent assay (ELISA) (Whitaker et al., 1989).
TABLE1 PREPARATION OF POLYCLONAL ANTISERA USED Rabbit (R) or sheep (S) antiserum R79 Rll0 R120 R156 R157 R207 R208 S105 Sl14 $253
Immunogen
Reference
Human MBP from MS brain conjugated to rabbit serum albumin Bovine MBP peptides 37-88 and 43-88 conjugated to ovalbumin Guinea Pig MBP Human MBP peptide 90-170 conjugated to ovalbumin Human MBP peptide 90-170 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin Human MBP peptide 1-44 conjugated to ovalbumin Guinea pig MBP-C1 conjugated to ovalbumin Human MBP peptide 1-14 conjugated to ovalbumin
(Whitaker, 1977) (Whitaker, 1982) (Whitaker et al., 1979) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Whitaker, 1982) (Present paper) (Whitaker, 1982)
138 TABLE 2 DOUBLE-ANTIBODY RADIOIMMUNOASSAYS USED FOR IMMUNOCHEMICAL COMPARISON OF HUMAN MBP-C8 A N D MBP-CI Antiserum
Dilution
Radioligand
R79 R79 R79 R79 R79 RI20 R120 R120 R120 R156 R157 P207 R208 S105 S105 S105 S114 $253 $253 $253
1 : 1500 1 : 3000 1 : 1000 1:500 1:7000 1 : 1000 1 : 1000 l : 1500 1 : 1200 1 : 1000 1:21100 1:2500 1:2000 1:6500 1:6500 1:3000 1 : 1200 1 : 1000 1 : 1000 1 : 20,000
Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human Human
MBP-CI MBP-C8 MBP peptide MBP peptide MBP peptide MBP-C1 MBP-C8 MBP 1-14 MBP 90 170 MBP peptide MBP peptide MBP peptide MBP peptide MBP-C1 MBP-C8 MBP peptide MBP peptide MBP-C1 MBP-C8 MBP peptide
45-89 69-89 90-170
that the antilog of the log scale EDs0 is the EDs~~ on the natural scale. y=
4 -/3(Iog(concent ration) - log(r))/loge(I0) 1 + 10 ~-
y= 1 + ( c°ncentrati°n ) -tJ'
where /3'= ( ~4/ 1 o g e ( 1 0 ) ) / 3 90-170 90-170 1-14 1-14
1-14 10-19
1-14
Statistical analysis" The mathematical model (Box and Hunter, 1962) assumed for the estimation of the curves was, similar to logit-log plots, that the logarithm of the concentration follows a logistic regression curve. A number of ways exist for relating the parameters for this model. The one selected expresses them in terms of specific characteristics of the plot of the percent inhibition against the logarithm to the base 10 of the concentration. In this form, ¢ represents the maximum inhibition achievable at infinite concentration; /3 represents the slope of the nonlinear curve at the log concentration which yields an inhibition of K/2 (50% inhibition or EDs0 on the log scale). The log (~-) represents the EDs0 on the log scale. The equations for the percent inhibition (y) are given below, for both the log form and the natural scale form. It should be noted that neither/3 nor/3' is the slope at the EDs0 for the natural form, but a more complicated expression. It is, however, true
This mathematical model, with the three parameters of log (~-), /3 and K, was fit to the data for each individual sample in each experiment by nonlinear least squares using the Gauss-Newton algorithm. The parameter estimates and their associated standard errors were calculated. A oneway analysis of variance was conducted to evaluate group differences in these parameters and a multivariate analysis of variance evaluated the significance of the differences for whole curves (both parameters taken simultaneously). Contrasts were formed to evaluate paired comparisons among the three groups of normal MBP-C1, non-MS MBP-C8 and MS MBP-C8. No adjustment for multiple comparisons was made. With only three groups that problem should be minimal. Statistical differences with a p value of _< 0.05 were considered as significant.
Results
The initial comparison of MBP-C8 with MBPC1 was by direct precipitation of radiolabeled MBP and MBP peptides. Antisera (R79 and R120), directed at multiple sites on MBP, precipitated MBP-C8 and MBP-C1 nearly equally (Fig. 1A and B). Some (R207 and $253) antisera preferentially reacted with MBP peptides 1-14 and 1-44 and reacted less well with MBP-C1 or MBP-C8 (Fig. 1C and D). R207 (anti-MBP peptide 1-14), which reacted at low titers with MBPC1, did not react at all with MBP-C8 in the direct precipitation reaction (Fig. 1C). The other anti-
0
10-
20-
30-
40-
50-
60
7O
8O
90
100
0
10-
20-
30-
40-
50
60
70
C
1:500
1:500
1:1000
1:1000
Dilution
1:2000
Dilution
1:2000
1:5000
1:5000
1:10000
R 207
1 : 10~(X)
R 79
8 I:~
c
o.
o
""
o.
co
