| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Oklahoma Medical Research Foundation, Oklahoma City, OK 73104;
University of Oklahoma Health Sciences Center, Oklahoma City, OK 73104;
Austin Research Institute, Heidelberg, Victoria, Australia; and
U.S. Department of Veterans Affairs Medical Center, Oklahoma City, OK 73104
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Protein sequences of Sm B, B', D1, D2, D3, and E have been known for some time. The last of these to be defined were Sm D2 and D3 by Lehmeier et al. (3). mAbs coprecipitate Sm D1 and Sm D2, but not Sm D1 and D3, suggesting a specific interaction between the D1 and D2 molecules. Sm D3 and Sm B' have also been shown to bind to one another (4). Indeed, Kambach et al. (5) recently described the crystalline structure of the D1/D2 complex, as well as the B'/D3 complex. This study revealed the specific crystallographic structural details of the protein-protein interactions that help make up the spliceosomal complex.
Previous work has detailed the prevalence and possible clinical significance of the autoantibodies binding to Sm B, B', or D1 (6, 7, 8, 9, 10). In addition, the precise sequential antigenic regions of these proteins recognized by human lupus autoantibodies are described (7, 8). Among the most important of these sequential epitopes is the glycine-arginine (GR)-rich carboxyl region of Sm D1, which is recognized by virtually all Sm-precipitin positive lupus patient sera (8, 10, 11, 12, 13, 14). This region of Sm D1 also shows striking sequence homology to a portion of the Sm D3 protein, and immunization with this region of D1 results in the production of anti-D3 Abs (13). The (GR)-rich portions of the Sm D1 and D3 proteins thus appear to be significant targets of the anti-Sm response. This raises the question of whether these autoantibodies (as well as anti-D2) arise through cross-reactivity with other previously antigenic proteins or through a mechanism by which their close physical association with these more commonly antigenic proteins promotes the loss of tolerance. This mechanism of physical association has been suggested for Ro/La, the nucleosome, and other autoantigen systems as well (15, 16, 17).
The ability to compare Ab fine specificity to defined three-dimensional Ag structure is highly significant and could potentially aid in answering questions of autoantibody origin. In the Ro and Sm B/B' systems in lupus, affinity purification of Abs using native human Ag leads to a consistent increase in binding to most of the antigenic octapeptides (9, 18). Thus, those Abs binding to native Ag seem to comprise the same set of Abs that bind to the octapeptides. This implies that autoantibody binding to octapeptides primarily reflects binding to surface structures and defines minimal to no cryptic epitopes with respect to the Ro and Sm B/B' autoantigens in human SLE. Based upon these data, the current hypothesis is that in systemic autoimmune responses many spliceosomal Abs preferentially target regions of an Ag that are localized to the autoantigen surface. The mechanism of determinant spreading in SLE has also been proposed to involve structures that are positioned near one another on the surface of the three-dimensional structure (20). Amazingly, very few autoantigens have defined crystalline structures, which would allow testing of this hypothesis. Such data could prove invaluable in defining conformational epitopes of these autoantigens.
Although potential surface position on the crystal structure is an
important factor in determining antigenicity of peptide sequences,
other means are undoubtedly significant. Previous observations have
suggested that the isoelectric point (pI) of the octapeptide could be
used to predict spliceosomal autoantibody binding, especially with Sm
B/B' and nRNP A (9, 19). Basic octapeptides (pI 
10.4) are vastly over-represented in the antigenic regions of these
proteins, whereas more acidic or neutral octapeptides (pI 
7.0)
are more commonly nonantigenic. These differences may simply reflect a
propensity for more basic structures to localize to the surface upon
folding, or other mechanisms might be at play.
We now evaluate two additional spliceosomal peptides for autoantibody fine specificity to test the hypothesis that pI can be used to predict spliceosomal autoantibody binding. Our findings confirm that higher pIs in the primary sequence significantly increase the likelihood that spliceosomal structures will become autoimmune epitopes. These findings, coupled with earlier data, also suggest that the relationship between pI and antigenicity is a feature common to nuclear autoantigens in systemic lupus (9, 19). In addition, we have used the overlapping octapeptides of the Sm D2 and Sm D3 proteins on solid phase supports to define structures bound by anti-Sm autoantibodies. The epitopes identified tend to be found on the surface of the proposed crystalline structures of the B'/D3 and D1/D2 complexes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
SLE patients were selected from the Oklahoma Clinical Immunology Serum Repository who met the Revised Classification Criteria for lupus set forth by the American College of Rheumatology (1) and whose sera contained precipitating levels of anti-Sm by Ouchterlony immunodiffusion. Patients thus tested variably displayed a wide array of autoantibody profiles, including anti-Sm, anti-dsDNA, anti-Ro, anti-La, and/or anti-nRNP autoantibodies. Of the 12 patients in whom anti-D2 or anti-D3 was detected, all were female, seven were African American, four were Caucasian, and one was a Native American. These patients ranged in age from 18 to 30 years, with an average age of 23. All SLE patients had carried a positive diagnosis for SLE for at least 1 year before the date of the serum samples used. Normals and patient controls (including SLE, rheumatoid arthritis, and Sjögren’s patients) were matched for sex and ethnicity and were age-matched to within 5 years of their case.
Cloning and expression of Sm D2 and Sm D3
The cDNA coding for the Sm D2 and D3 proteins was obtained from the American Type Culture Collection (Manassas, VA; accession numbers AA314384 and AA084844). The PCR was used to amplify D2- or D3-specific cDNA using primers that incorporated a 5' EcoRI site (GCGAATTCATGAGCCTCCTCAAC for D2 and CCGAATTCATGTCTATTGGTGTGC for D3) and M13 reverse primers (for pBluescript; New England Biolabs, Beverly, MA). PCR products were purified and ligated into the pCRii cloning vector (Invitrogen, Carlsbad, CA) and sequenced.
Sm D2 and D3 pCRii clones were then digested with
EcoRI and HindIII and ligated into the
Escherichia coli pMal-c2 expression vector (New England
Biolabs). Transformed cells were incubated with 0.3 mM isopropyl
-D-thiogalactoside to induce expression of Sm D2 or D3
fusion proteins with an N-terminal bacterial maltose binding protein
moiety. Fusion protein was isolated from whole bacterial lysate by
purification on an amylose column and subsequent elution with 10 mM
maltose, 20 mM Tris, 1 mM EDTA, 200 mM NaCl (pH 7.4).
ELISAs
Standard solid-phase assays were used to measure the Ab reactivity in human sera or in murine mAbs. One microgram of Ag (purified Sm D2 or D3 fusion proteins, maltose-binding protein, or whole Sm) was coated per well in polystyrene plates. Assays were performed using a protocol previously described in detail (7, 8).
Western blots
Affinity-purified Sm D2 and D3 Ags or HeLa cell extracts were subjected to electrophoresis in 12.5% High-N,N,N',N'-tetramethylethylenediamine SDS-PAGE gels containing 1% SDS (in 0.15 M Tris-HCl, pH 8.8) using a previously described protocol (8). Proteins were transferred to nitrocellulose and human sera at 1/100 dilutions were added. A secondary Ab consisting of alkaline phosphatase-conjugated anti-human IgG was then added followed by nitrotetrazolium blue/5-bromo-4-chloro-indolyl phosphate as a substrate for the conjugated enzyme.
Other autoantibody assays
Precipitating levels of Sm and/or nRNP autoantibodies were detected using double immunodiffusion. Human sera were tested for antinuclear Abs by a standard antinuclear Ab test against Hep-2 cells (INOVA Diagnostics, San Diego, CA) and for autoantibodies binding to native DNA by a Crithidia assay (Protrac Industries, Kerrville, TX) using previously published protocols (21).
Solid-phase peptide synthesis and autoantibody assays
The 111 possible overlapping octapeptides of the Sm D2 protein and the 119 overlapping octapeptides of Sm D3 were prepared on the ends of polyethylene pins using solid-phase peptide chemistry as described previously (7, 8). Anti-peptide assays were conducted at a 1/100 dilution of patient or control serum using a modified ELISA technique, which we have previously described in detail (7, 8).
Antibody affinity purifications and depletion
Recombinant Sm D2 and D3 were prepared as described above. These individual proteins were separately, covalently bound to cyanogen preactivated Sepharose 4B by the method recommended by the manufacturer (22, 23). Each serum sample tested was circulated through this column for a total of four to six passes and the absorbed serum was then concentrated back to the original volume. More than 90% of the anti-protein reactivity was removed as measured by standard ELISA. Anti-Sm D2 or D3 Abs were eluted from the column with 3 M sodium thiocyanate. Total protein concentration was used to adjust dilutional effects caused by the column absorptions, as previously described (24).
Inhibition of anti-Sm D2 and anti-Sm D3 autoantibodies
Inhibition experiments were performed as has been described (24). Briefly, linear peptides comprising the individual epitopes of Sm D2 and Sm D3 defined by the octapeptide assays were obtained (Molecular Biology Core Facility, Oklahoma City, OK). Standard solid phase assays were then performed as described above, with an additional 2 h inhibition step before addition of human sera. In this step, diluted serum samples were incubated with varying levels of the linear peptides, fusion proteins, or no peptide at all, with peptide concentrations of 1–50 µg/ml. Both uninhibited and inhibited samples were then tested for binding to the D2 or D3 fusion proteins using a standard ELISA (7, 8). Additionally, similar inhibitions were performed before analyzing these same serum samples against both fusion proteins and HeLa cell-derived extract in Western blots, under both denaturing and nondenaturing conditions, as described above and previously (8, 19).
Structural analysis of Sm D2 and Sm D3 sequential epitopes
The Protein Data Bank (25) was accessed to extract coordinates for the crystal structures of two complexes of Sm proteins, D1-D2 (Protein Data Bank code 1B34) and D3-B' (Protein Data Bank code 1D3B) (4). Using these coordinates, secondary structures were calculated and displayed for the Sm D2 and Sm D3 proteins using the MOLMOL program (25). Solvent-accessible surface areas were calculated with the WHAT IF program using a solvent probe radius of 1.4 Å and the following van der Waals radii: 1.8 Å for carbons, 1.4 Å for oxygens, 1.7 Å for nitrogens, and 2.0 Å for sulfurs. Molecular surface areas calculated by the WHAT IF program have been shown to be within 5% of Connolly’s method (27, 28).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Forty-one anti-Sm precipitin-positive SLE patients were
selected from the Oklahoma Clinical Immunology Serum Repository. These
sera were also tested for reactivity with Sm and nRNP by standard ELISA
(Table I
). All patients positive for Sm
by Ouchterlony immunodiffusion also had positive results by standard
ELISA. Conversely, all normal control and Sm-negative patient sera used
in this study were shown to be negative for Sm by ELISA and
Ouchterlony.
|
). All patients who had reactivity with
Sm D2 or Sm D3 also demonstrated binding to Sm B/B' on the HeLa cell
immunoblots. In addition, all anti-D2-positive patient sera bound
Sm D1. No further commonalities were discovered with respect to binding
to other self-proteins. Immunological testing of anti-Sm D2 and anti-Sm D3 lupus patients’ sera
All sera were also tested for the presence of other common lupus
autoantibodies by immunofluorescence and Ouchterlony. All Sm D2- and Sm
D3-positive patients had positive anti-nuclear Abs. Anti-dsDNA Abs
were detected in 67% (six of nine) of the anti-D3-positive
patients and 57% (four of seven) of anti-D2-positive patients.
Anti-nRNP autoantibodies were found in 57% (four of seven) of Sm
D2-positive patients and 67% (six of nine) of the Sm D3-positive
patients (Table I). The patient controls used in the study exhibited a
wide range of autoantibody profiles.
Fine specificity mapping of Sm D2
Sera from the seven patients positive for Sm D2 Abs were tested
for binding to the overlapping octapeptides of the D2 protein by our
modified solid-phase ELISA. These serum samples exhibited significant
reactivity with various portions of the Sm D2 protein, including both
N- and C-terminal regions. Patient sera contained Abs that on average
recognized five major regions of Sm D2. These major epitopes are
defined as being bound by >3 SDs above the normal mean and as being
antigenic in at least 50% of the patients tested (Fig. 1). These antigenic regions include:
epitope 1, octapeptides 1–6 (aa 1–13); epitope 2, octapeptides 30–34
(aa 30–41); epitope 3, octapeptides 74–77 (aa 74–84); epitope 4,
octapeptides 83–98 (aa 83–105); and epitope 5, octapeptide 111, the
last full octapeptide of the protein (aa 111–118; see Table II). Epitope 4 showed the most
significant binding of any of the antigenic regions, with average
anti-D2 binding of six SDs above the normal mean, with maximum
reactivity on octapeptide 87 (aa 87–94).
|
|
). All
seven D2-positive patients tested significantly bound (an average of
3 SDs above the normal mean) fifteen of the 111 octapeptides. Two of
these common antigenic regions contain octapeptides bound at least six
SDs above the normal mean. All tested sera recognize epitope 4, whereas
epitopes 1 and 2 are antigenic in nearly all (six of seven) tested
specimens. Some aspects of the binding profile of this autoantibody
response are conserved between patients with some modest variability
from patient to patient.
|
20% of control sera. In viewing
these results, we realized that all of these normal (nonaffected) sera
binding epitope 5 were obtained from laboratory personnel. The level of
binding in these sera was significantly below that of the patient
binding in these regions, and none of these sera bound Sm D2 or D3 by
ELISA or Western blot. When using sera collected from nonaffected
individuals without previous exposure to laboratory work, no reactivity
with any of the Sm D2 octapeptides were noted. No other significant
binding to the overlapping octapeptides was detected in lupus patient
sera negative for Sm D2 by ELISA and Western blot. Fine specificity mapping of Sm D3
Sera from the Sm D3-positive patients (as determined by
immunoblot) were also tested for reactivity with the respective
overlapping octapeptides. Numerous regions of Sm D3 were recognized as
antigenic; however, the greatest binding was seen with the
carboxyl-terminal portion of the protein (Fig. 2). Sm D3-positive patients produced Abs
against four major regions of the D3 polypeptide. These common
antigenic regions were defined as having average binding greater than
three SDs above the normal mean for each octapeptide.
|
). The most significant anti-Sm D3
binding in patient sera was with octapeptide 81 of epitope 3
(PMLKSMKN), which was bound on average at nearly 11 SDs above normal
average binding. Epitope 2 also has significant binding with multiple
octapeptides from this epitope, having average reactivity greater than
six SDs above the normal mean. Substantial reactivity was seen from
octapeptides 76–99, in that every octapeptide in that range was
recognized. All anti-D3 patient sera tested demonstrate reactivity
with epitopes 3 and 4 (Table III). None of these octapeptides were
significantly bound by normal control sera. In addition, the region
described as the (GR)-repeat is contained within epitope 4 (primarily
octapeptide 110). This sequence bears close homology to an antigenic
sequence from the Sm D1 protein (10, 11). This epitope is
significantly bound by every D3-positive patient tested and, not
surprisingly, most of these patients also exhibited anti-D1
reactivity. None of the patient controls or normals, however, exhibited
detectable reactivity with the overlapping octapeptides of Sm
D3. In addition, the monoclonal Ab Y12, which has been shown previously to bind to Sm D1, D3, and B/B', was tested for reactivity with the overlapping octapeptides of Sm D2 and D3 (Ref. 8 ; data not shown). The Y12 monoclonal showed little significant reactivity with the Sm D2 octapeptides. This monoclonal, however, did exhibit impressive binding to the C-terminal octapeptides of D3, specifically amino acids 109–120 (epitope 4). These data were not unexpected because the Y12 Ab has also been shown to react with the C-terminal portion of D1, which encompasses a (GR)-repeat portion of that protein as well (8, 29).
Affinity-purified anti-Sm D2 and D3
Lupus patient Abs specific for Sm D2 or Sm D3 were purified by
affinity column chromatography using recombinant Sm D2 or D3. This
resulted in removal of >95% of the anti-Sm D2 or anti-Sm D3
reactivity from these samples, as detected by standard ELISA. In
addition, the Abs eluted from the Sm D2 column showed the same
specificity for D2 by our octapeptide assay as did the original serum
samples from which they were taken (Fig. 3). Similar results were detected after
purification of anti-D3 Abs. Anti-D3 Abs do not cross-react with
the octapeptides of Sm D2. The eluted Abs from the D3 column do show
reactivity to the overlapping octapeptides from the C-terminal end of
Sm D1, specifically limited to those encompassing the (GR)-rich portion
of this protein. The affinity-purified anti-D2 Abs, however, do not
bind to either D1 or D3 by our octapeptide assay or by Western blot.
Also, neither purified anti-D2 nor purified anti-D3 Abs showed
any reactivity with the overlapping octapeptides of Sm B/B' or with any
other common lupus autoantigen (data not shown).
|
The addition of Sm D2 and D3 fusion proteins decreased binding of the appropriate SLE sera to Sm D2 or D3, respectively, from HeLa cell extract on Western blots under both denaturing and nondenaturing conditions (data not shown). Additionally, prior inhibition of anti-Sm D2 or D3 sera with Sm D2 or D3 peptides caused notable decreases in binding to D2 and D3 fusion proteins in standard ELISAs. Peptides comprising epitopes 1 and 4 of D2 variably inhibited anti-D2 binding, from 6 to 40% for epitope 1 and from 17 to 42% for epitope 4. When combined, these two peptides demonstrated an additive inhibitory effect, blocking up to 60% of anti-D2 Abs. Similar results were seen upon inhibition of anti-D3 Abs with D3 peptides. Peptides comprising D3 epitope 3 were capable of inhibiting up to 61% of anti-D3 binding, whereas epitope 4 of D3 inhibited up to 38%. Combined, these two D3 peptides inhibited up to 90% of some patients’ anti-D3 reactivity by solid-phase ELISA. D2 peptides caused no inhibition of anti-D3 binding, nor did D3 peptides inhibit anti-D2 reactivity.
Characteristics of Sm D2 and Sm D3 antigenic regions
The major antigenic regions of the Sm D2 protein have an average
pI = 10.44 compared with pI = 6.67 for nonantigenic areas.
The octapeptides comprising the major epitopes of Sm D3 also have a
basic average pI = 10.80 compared with regions not commonly bound
by patient sera (pI = 6.55). Both of these differences reached
statistical significance with p < 0.0001 (Student’s
t test). In addition, those less common epitopes bound
significantly in individual patient sera (3 SDs) but which failed to
reach significance when reactivity was averaged to form the composite
major epitopes also have a basic average pI = 9.3. pI alone,
however, is not completely specific in identification of putative
epitopes without additional factors (such as known surface probability
from crystal structure, etc.) being considered. Using a pI threshold
alone (9.5 for D2 and 10.0 for D3), five nonantigenic D2 octapeptides
and four nonantigenic D3 peptides would have been mistakenly designated
as antigenic. In addition, pI alone would have mistakenly labeled nine
octapeptides of D2 and six octapeptides of D3 as nonantigenic. Thus,
for D2 the pI threshold yields an 81% sensitivity and 85%
specificity, whereas for the D3 protein, pI has a 90% sensitivity and
92% specificity for determining antigenicity (Fig. 4).
|
The crystal structures of complexes of Sm D2 (complexed with Sm
D1) and Sm D3 (complexed with B') have recently been published by
Kambach et al. (5). Sequential epitopes were mapped on the
three-dimensional structures of the Sm D2 and Sm D3 proteins (Fig. 5). Segments of polypeptide that are
disordered in the crystal structure of Sm D2 include residues 1–25 and
74–92. These regions can be considered as highly flexible or mobile
and correspond to epitopes 1 (aa 1–13), 3 (aa 74–84), and part of
epitope 4(aa 83–105). Coordinates of Sm D3 were available only for
residues 3–75, which covered epitopes 1 (aa 24–37) and 2 (aa 60–68),
but not epitopes 3 and 4 (see Table II). In this case, the two
N-terminal residues are disordered, whereas the C-terminal portion of
Sm D3 had been deleted to form stable D3-B' complexes for
crystallization (5).
|
), which
consists of an N-terminal 
helix and a highly bent or twisted
five-stranded antiparallel -pleated sheet (5). In Sm
D2, epitope 2 (aa 24–37) represents the helix, the following turn,
and first part of a strand. The outermost strand and the
C-terminal segment correspond to Sm D2 epitopes 4 (aa 83–105) and 5
(aa 111–118). In Sm D3, epitope 1 (aa 24–37) covered the central
strand. Similar to Sm D2 epitope 4, the outermost strand of Sm D3
was covered by epitope 2 (aa 60–68). Thus, two of the major antigenic
sites on Sm D2 and Sm D3 are located in very similar three-dimensional
positions on the crystal structures (see Fig. 5).
Solvent accessible surface areas of the linear epitopes for Sm D2 and
Sm D3 crystal structures are presented in Table IV. Surface areas were calculated for the
isolated Sm D2 and Sm D3 portions as well as for their dimers with Sm
D1 and Sm B'. A decrease in the accessible surface area of an epitope
indicates that a portion is buried upon formation of the dimers. All
except one strand of the epitopes (54% epitope 4 of Sm D2 was buried
in the D1-D2 complex) presented similar accessible surface areas as
monomers and dimers, indicating that these linear epitopes could easily
be bound by autoantibodies. Note that residues 83–92 of Sm D2 epitope
4 were within a disordered and probably highly solvent-exposed loop
(see Fig. 5).
|
strands and turns or loops between
these strands. Two linear epitopes appear on structurally
homologous regions of the two spliceosomal proteins. No antigenic
regions of Sm D3 are buried by its interaction with Sm B'. Instead, the
epitopes of D3 conspicuously avoid the major regions (aa 21–23, 39,
and 69–75) thought to be directly involved in dimer interactions with
B'. Interestingly, one of the areas of Sm D2 (aa 90–102) that
interacts to form the dimer with Sm D1 is highly antigenic in
D2-positive lupus patients. Furthermore, the location of the Sm D2 and
Sm D3 antigenic regions on the exposed surfaces of the spliceosomal
subcomplexes is similar to the crystallographic location of antigenic
regions of nRNP A (19). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The octapeptide epitopes of these autoantigens provide several interesting insights into autoimmunity when compared with the crystal structures of the Sm complexes. The area of the Sm D2 protein shown to be involved in binding to Sm D1 (5) consists partly of amino acids 90–102, found in epitope 4 of Sm D2. Interestingly, this is one of the highest areas of reactivity seen in the D2-positive sera. These data initially seem to allow the possibility that the anti-D2 autoimmune response is formed against native D2 Ag, which is not complexed with other spliceosomal proteins. However, only a small segment of a single D2 epitope corresponds to an area of intermolecular interaction. Thus, although the exact mechanisms by which this lupus autoantigen is exposed to immune regulation are as yet undefined, it appears likely that it is seen (at least for the purposes of most autoantibody production) in the context of a whole complex rather than as an individual protein.
The region of Sm D3 that is involved in binding to Sm B' in vivo
consists primarily of aa 21–23, 39, and 69–75 of D3 (5).
Interestingly, D3-positive patients make high levels of autoantibodies
against aa 60–68 and 76–106, surrounding but conspicuously avoiding
one of the major regions (aa 69–75) involved in direct intermolecular
bonding to Sm B' in vivo. In fact, all of the amino acids thought to be
involved in these protein-protein interactions lie outside of major D3
epitopes. The comparison of our octapeptide epitopes with known
crystalline data also reveals that the patterns created by these
epitopes on the surface are similar across the three autoantigens for
which this analysis has been performed (Sm D2, Sm D3, and nRNP A),
being directed primarily against sheets and the loops connecting
sheets. Of course, this could simply be due to the obvious
structural similarities between the three-dimensional orientations of
these spliceosomal proteins. In addition, when the published human
epitopes of Sm B/B' (7) are compared with the portion of
B/B' complexed with Sm D3, the same pattern surfaces. The region of Sm
B/B' (primarily aa 65–76) involved in close interactions with D3 lies
directly in between two of the major Sm B/B' epitopes. This reinforces
our observation that antigenic regions in this system tend to be those
that are not involved in intermolecular, intracomplex bonds.
This study, as well as that from Hoch et al. (30), reveals that all lupus patients who bind Sm D3 also bind its complexed partner B/B' and that all who bind D2 also bind its complexed partner D1. These data fall in line with an increasingly common theme in autoimmunity in which autoantibody specificity pairing seems to be partially explained by physical associations in vivo (15, 16, 17, 31, 32). We have demonstrated herein that anti-Sm autoantibodies in lupus are directed primarily against structures that would be located on the surface of native autoantigens in their complexed configurations. Current knowledge of determinant spreading and autoantibody profiles also points to progression from protein to protein known to complex together in vivo as a mechanism in many diseases. Therefore, the importance of the immune recognition and reaction to autoantigens in their native complexed arrangement seems clear.
The question of how these autoantibodies to Sm D2 and Sm D3 arise has not yet been answered. However, it could be that they derive from cross-reactivity between similar sequences from other spliceosomal autoantigens. However, D2 patients bind weakly or not at all to areas of Sm D2 that show similarity to other spliceosomal proteins (D2 aa 57–67 have 65% homology with B/B' aa 32–42). When coupled with the peptide column data (showing no cross-reactivity of anti-D2 with other Sm proteins), this seems to indicate that spliceosomal cross-reactivity at the level of Ig is not likely to be operating for this specificity. Conversely, one antigenic region of Sm D3, epitope 4, shows significant homology with the (GR)-repeat portion of Sm D1. This carboxyl-terminal region of Sm D1 has been determined by several investigators to be the major target of the human and murine lupus anti-Sm D1 response (10, 11, 12, 13, 14, 29, 30, 33). Immunization with this C-terminal D1 peptide causes lupus-like disease in rabbits as well as the production of anti-D1 and anti-D3 autoantibodies (33) and accelerates disease in lupus-prone animals (New Zealand Black/White mice) (34). The data from the (GR)-repeat portion of the D3 polypeptide, its similarity to D1, and the cross-reactivity of affinity-purified anti-D3 Abs with this region of D1 seem to support the theory of cross-reactivity explaining the appearance of some autoantibodies. An additional group has shown that this same region of Sm D1 is antigenic when the alternating arginines are dimethylated (35). Interestingly, this posttranslational modification may not be absolutely required for antigenicity of these sequences, as has been shown by this and other works (8, 33, 34). Additional experiments are underway to understand the overall importance of symmetrical dimethylation of these sequences with regard to lupus autoimmunity.
The other major possibility is that these proteins are seen by the immune system in complex with other spliceosomal proteins already targeted by the immune system and are thus presented by professional phagocytic cells along with previously autoantigenic proteins. This has been shown to be the most likely cause in the case of the Ro (SS-A) and La (SS-B) system in lupus and Sjögren’s syndrome (16, 18). In the Ro/La system, it has been demonstrated that murine immunization with all or portions of the La molecule can result in an immune response against the 52-kDa Ro particle and vice versa (16, 36). In this instance, the intermolecular spreading of the immune response is thought to be due to a physical association between the Ro and La particles in vivo, as could potentially be the case with the known B'/D3 and D1/D2 complexes. In SLE, it has been shown that anti-Sm B/B', often with anti-Sm D1, dominates the anti-Sm specificity of most lupus patients, thus the close association of Sm D2 and D3 with D1 and B' may well contribute to or provide additional insight into the origin of anti-D2 and anti-D3 Abs by the mechanism discussed above.
Regardless of the specific mechanism of development of these autoantibodies, for both Sm D2 and Sm D3 the response is fairly well conserved from patient to patient. This is in direct contrast to the much greater variation seen among patients with the nRNP 70K and nRNP C systems (37, 38). These data are consistent with the autoimmune responses of these patients arising in a similar defined manner rather than from random polyclonal events. Consequently, current data support the anti-Sm D2 and anti-Sm D3 autoantibodies arising through an underlying mechanism that is Ag driven and well conserved between individuals.
The data also reveal that pI acts as an accurate predictor of antigenicity in this system. In the previously studied nRNP A, Sm B', and Sm D1 proteins, the average pI of antigenic regions was 10.4 compared with the average pI of 6.0 of nonantigenic regions. This observation also holds true with Sm D2 and Sm D3, in that antigenicity clearly increases with pIs >9.0. The combination of localization of epitopes to the surface and antigenicity of these basic epitopes suggests that perhaps high pIs simply increase the probability that octapeptides will be situated near the protein surface. Regardless of the reason, the compilation of previous data (9, 19) and our results seem to indicate that high pI is an aspect common to nuclear autoantigens in SLE.
The role of Sm D2 and D3 in the SLE anti-Sm autoimmune response is only beginning to emerge. These experiments describe nine newly appreciated sequential antigenic regions of the spliceosomal complex and show that these antigenic sites also have a high pI, as do other spliceosomal autoantigens. These antigenic regions of Sm D2 and Sm D3 are located on the surface of the crystallographic structure, and concomitant production of Sm B'/D1 and D2/D3 autoantibodies reflects physical associations. Some work has been done detailing the crystal structures of Ags while complexed with specific Abs (e.g., rheumatoid factor (39)). A relatively new approach, presented herein, has been to identify the fine specificity of polyclonal Abs and then to apply those data to the crystal structure of the autoantigen. In fact, we know of only two such cases where a similar study has even been attempted (19, 40). In all likelihood, this is due to the relative paucity of autoantigens that have had their crystal structures solved. This approach allows for efficient analysis of a wide spectrum of Ab specificities. These findings further explain the role of Sm D2 and Sm D3 in the human SLE autoimmune response and will lead to further investigation into potential mechanisms for the induction and perpetuation of these responses.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Judith A. James, Department of Arthritis and Immunology, Oklahoma Medical Research Foundation, 825 NE 13th Street, Oklahoma City, OK 73104. E-mail address: judith-james{at}omrf.ouhsc.edu
3 Abbreviations used in this paper: SLE, systemic lupus erythematosus; nRNPs, nuclear ribonucleoproteins; pI, isoelectric point; GR, glycine-arginine; pI, isoelectric point.
Received for publication March 21, 2001. Accepted for publication December 14, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Agmon-Levin, M. Blank, Z. Paz, and Y. Shoenfeld Molecular mimicry in systemic lupus erythematosus Lupus, November 1, 2009; 18(13): 1181 - 1185. [Abstract] [PDF]
|
|
|
|
 |
|
 C. Schild-Poulter, A. Su, A. Shih, O. P. Kelly, M. J. Fritzler, R. Goldstein, and R. J. G. Hache Association of autoantibodies with Ku and DNA repair proteins in connective tissue diseases Rheumatology, February 1, 2008; 47(2): 165 - 171. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. James, X. Kim-Howard, B. Bruner, M. Jonsson, M. McClain, M. Arbuckle, C. Walker, G. Dennis, J. Merrill, and J. Harley Hydroxychloroquine sulfate treatment is associated with later onset of systemic lupus erythematosus Lupus, June 1, 2007; 16(6): 401 - 409. [Abstract] [PDF]
|
|
|
|
 |
|
 J W Dieker, C C Van Bavel, G Riemekasten, J H Berden, and J van der Vlag The binding of lupus-derived autoantibodies to the C-terminal peptide (83-119) of the major SmD1 autoantigen can be mediated by double-stranded DNA and nucleosomes Ann Rheum Dis, November 1, 2006; 65(11): 1525 - 1528. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Riemekasten and B. H. Hahn Key autoantigens in SLE Rheumatology, August 1, 2005; 44(8): 975 - 982. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Teerlink ADMA metabolism and clearance Vascular Medicine, July 1, 2005; 10(1_suppl): S73 - S81. [Abstract] [PDF]
|
|
|
|
|
|
T. Teerlink ADMA metabolism and clearance Vascular Medicine, May 1, 2005; 10(2_suppl): S73 - S81. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Mahler, L. M. Stinton, and M. J. Fritzler Improved Serological Differentiation between Systemic Lupus Erythematosus and Mixed Connective Tissue Disease by Use of an SmD3 Peptide-Based Immunoassay Clin. Vaccine Immunol., January 1, 2005; 12(1): 107 - 113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. McClain, C. S. Lutz, K. M. Kaufman, O. Z. Faig, T. F. Gross, and J. A. James Structural availability influences the capacity of autoantigenic epitopes to induce a widespread lupus-like autoimmune response PNAS, March 9, 2004; 101(10): 3551 - 3556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Arbuckle, M. T. McClain, M. V. Rubertone, R. H. Scofield, G. J. Dennis, J. A. James, and J. B. Harley Development of Autoantibodies before the Clinical Onset of Systemic Lupus Erythematosus N. Engl. J. Med., October 16, 2003; 349(16): 1526 - 1533. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) in at least one patient or that are
never antigenic (
) in any patient currently tested. Antigenicity is
defined as reactivity of 