ARTHRITIS & RHEUMATISM Vol 40, No 10, October 1997, pp 1803-1809 0 1997, AmeriLan College ot Rhcumdtology 1803

ANTIBODIES AGAINST A PEPTIDE SEQUENCE LOCATED IN THE

PATIENTS WITH JUVENILE RHEUMATOID ARTHRITIS LINKER REGION OF THE HMG-1/2 BOX DOMAINS IN SERA FROM

FRANK JUNG, GUNTER NEUER, and FRIEDLINDE A. BAUTZ

Objective. To extend our work on the mapping of B cell epitopes on nucleosomal high mobility group (HMG) proteins in the sera of patients with juvenile rheumatoid arthritis (JRA).

Methods. Seventy-seven pauciarticular-onset JRA serum samples from antinuclear antibody (ANA)- positive patients and 42 polyarticular-onset JRA patient sera found to react with HMG-2 by immunoblotting were used in this study. To identify B cell epitopes on HMG-2, recombinant HMG-2 protein fragments were used in enzyme-linked immunosorbent assay (ELISA) and in competition ELISA experiments with a set of overlapping synthetic peptides. Fine epitope mapping was achieved by oligopeptide synthesis, followed by immunoblotting.

Results. Pauciarticular, but not polyarticular, JRA patient sera were found to recognize a lysine-rich major epitope (KKGKKKDP), which is located in the linker region of the HMG box domains of the HMG-2 nonhistone chromosomal protein. No significant immu- noreactions were observed in sera from ANA-negative JRA patients and in sera from children with nonrheu- matic diseases, indicating that this epitope seems to be specific for pauciarticular-onset JRA.

Conclusion. In addition to our previous finding that JRA sera will react with a defined epitope on HMG-17, pauciarticular JRA patient sera were also found to recognize a defined epitope on the HMG-2 protein, thus suggesting the importance of this epitope in the etiology of JRA.

Dr. Bautz's work was supported by grant Ba 477112.1 from the Deutsche Forschungsgemeinschaft.

Frank Jung, MSc, Gunter Neuer, PhD, Fricdlindc A. Bautz, PhD: University of Heidelberg, Heidclberg, Germany.

Address reprint requests to Friedlinde A. Bautz, PhD, Insti- tute of Molecular Genetics, University of Heidelberg, Im Neuen- heimer Feld 230, 69120 Heidelbcrg, Germany.

Submitted for publication January 22, 1997; accepted in revised form May 8, 1997.

Antinuclear antibodies (ANA) are found in a high proportion of sera from children with juvenile rheumatoid arthritis (JRA) (1,2). However, to date, very little is known about the antigenic specificities and the prognostic value of ANA in JRA. Recently, we demon- strated the presence of autoantibodies to the high mobility group (HMG) nucleosomal proteins HMG-1, HMG-2, and HMG-17 in the sera of ANA-positive patients with JRA (3,4). These small chromosomal proteins are among the most abundant, ubiquitous, and evolutionary conserved nonhistone proteins found in the nuclei of higher eucaryotes. Using immunoblotting tech- niques, we found that sera from patients with pauciarticular-onset JRA and polyarticular-onset JRA recognized the HMG-112 proteins with similar frequen- cies (40% versus 30%), whereas the pauciarticular-onset JRA patient sera had a higher prevalence of anti- HMG-17 antibodies than did the polyarticular-onset JRA sera (47% versus 16%) (4). No significant reactions with the HMG-14 protein were observed on immuno- blot. Competition enzyme-linked immunosorbent assay (ELISA) experiments defined the octapeptide, PKPEP- KF'K, at positions 34-42 in the HMG-17 protein, as a major B cell epitope, which was recognized by more than 70% of the HMG-17-positive JRA sera (5) .

In this report, an extension of our work on the mapping of B cell epitopes on HMG proteins in JRA patient sera is presented. For partial epitope mapping, polymerase chain reaction (PCR)-generated recombi- nant HMG-2 (rHMG-2) protein fragments were used in ELISA experiments with pauciarticular and polyarticu- lar JRA patient sera. By using a set of overlapping synthetic peptides in competition ELISA experiments with rHMG-2 protein fragments, and direct peptide synthesis followed by immunoblotting, we were able to locate a major B cell epitope in the linker region of the HMG box domains of the HMG-2 protein. To identify the amino acids functioning in antibody binding, point

1804 JUNG ET AL

mutations were introduced in this peptide epitope, and the relative contribution of individual amino acid resi- dues for antibody recognition could then be deduced.

PATIENTS AND METHODS

Patients. Serum samples were obtained from 77 ANA- positive patients with pauciarticular JRA and from 42 polyar- ticular JRA patients whose sera reacted with HMG-2. These sera have been characterized previously (3,4). All patients had an established diagnosis of JRA according to the American College of Rheumatology (formerly the American Rheuma- tism Association) criteria (6). Sera from 29 ANA-negative JRA patients and 25 sera from age-matched children who were treated for temporary nonrheumatic illnesses were used as controls. All sera were supplied by the Rheumatic Children’s Hospital (Garmisch-Partenkirchen, Germany).

Antigens. The complete human HMG-2 complemen- tary DNA clone used in these experiments was obtained from M. Bustin (National Institutes of Health, Bethesda, MD). For partial epitope mapping, HMG-2 DNA fragments were gen- erated and amplified by PCR (7). Primers used carried a 3’-nucleotide clamp at their 5’ end and either a Bum HI or a Hind I11 restriction site. For the expression of recombinant proteins, the PCR-generated protein fragments were sub- cloned into the Bum HI-Hind 111 site of pDS56+6His vector. Expression of the rHMG-2 proteins was induced with 2 mM isopropyl P-D-thiogalactopyranoside, followed by purification on Ni’ ’ chelate columns and by gel filtration.

Synthetic peptides. Synthetic peptides were synthe- sized and purified by Dr. R. Frank (Zentrum fur Molekulare Biologic Heidelberg, Heidelberg, Germany) according to the method of Gausepohl et a1 (8). In addition, oligopeptides were synthesized on activated membranes using the SPOTs system (Genosys, Cambridge, UK) (9).

Immunologic assays. ELISA. Solid-phase ELISA was performed as previously described (10). Briefly, microtiter plates (Maxisorp F96; Nunc, Wiesbaden, Germany) were coated routinely with 200 ng rHMG-2 in 50 mM carbonate buffer, pH 9.5 per well, and incubated overnight at 4°C. After removal of the unadsorbed antigen, unsaturated binding sites were blocked with 0.2% Tween 20/phosphate buffered saline (PBS), pH 7.3 (blocking solution), for at least 2 hours at room temperature. Tween/PBS was used as the only blocking solu- tion, since JRA patient sera frequently gave antibody reactions against bovine serum albumin and/or milk proteins. The JRA patient sera were diluted 1:100 in blocking solution and incubated for 2 hours at room temperature with the rHMG-2 antigens. For the competition ELISA, JRA sera were diluted 1 5 0 in blocking solution and preincubated with the synthetic peptides for 2 hours at room temperature. Routinely, 50 pl of diluted serum, preincubated with 9 pg of peptide, was added to each well, and the plates were incubated for 90 minutes at room temperature (5).

The amount of human IgG bound to the immobilized rHMG-2 proteins was quantitated with a goat anti-human IgG(Fc) alkaline phosphatase-conjugated second antibody (Dianova, Hamburg, Germany). The adsorbance at 405 nm was measured in a Titertek Multiskan Photometer (Flow Laboratories, Irvine; Scotland). The ELISA readings obtained

Protein Amino Fragment: Acid: 0 50 100 150 200

I 1 I - 1 hHMG-2protein 001-208 I ~~

N-termma1 fragment 001-127 - C-terminal fragment 121-208 ~ A

~ - ~~

N1-fragment 001-055

N2-fragment 048-100 - N3-fragment 095-127

~

Figure 1. Distribution of antigenic regions on the primary sequence of the human HMG-2 (hHMG-2) protein. Complementary 3NA fragments were prepared by polymer chain reaction and subcloned into the expression vector pDS56+6His The black boxes represent areas of mdjor antigenic determindnt(s), the shaded box indicates minor antigenic determinant(?), and the white boxes show regions ot nonredctivity, as determined by enzyme-linked immunosorbent dssay

with serum samples from JRA patients using the rHMG-2 proteins were presented as the mean optical density of dupli- cate samples minus 2 SD. In competition experiments, the inhibitory activity of the peptides for binding of antibodies to the rHMG-2 proteins was expressed as the percentage of inhibition of rHMG-2 antibody activity in the absence of peptides.

Immunoblotting. Oligopeptides were synthesized on activated membranes using the SPOTs system (Genosys) ac- cording to the manufacturer’s instructions (9). Membranes with immobilized peptides were blocked in Tris buffered saline/O.l% Tween 20150% horse serum overnight at 4°C. Spots were excised and incubated with JRA p,atient sera, and with sera from children with nonrheumatic childhood illnesses as controls, at dilutions of 1: lOO. After 6 washes with TBS/O.l% Tween 20 for 10 minutes each, bound antibodies were detected with alkaline phosphatase-coupled goat anti-human IgG anti- bodies in conjunction with nitroblue tetrazolium bromo- chloroindolylphosphate as substrate.

RESULTS

Identification of a major antigenic epitope on the HMG-2 protein. ELISA experiments with rHMG-2 frag- ments. For partial epitope mapping, the PCR-generated HMG-2 DNA fragments were subcloned into the pDS56 + 6His expression vector, expressed, and purified as described in Patients and Methods. The rHMG-2 protein was then divided into an N-terminal and a C- terminal fragment (Figure 1) and analyzed for immuno- reactivity with the JRA patient sera in ELISA experi- ments (see Patients and Methods). In this way, we were able to determine a major epitope in the N-terminal region of the HMG-2 protein (amino acids 1-127). It was shown that 49% of the ANA-positive pauciarticular- onset JRA sera immunoreacted with the N-terminal

ANTIBODIES AGAINST HMG PROTEINS IN JRA 1805

Table 1. HMG-2 proteins in sera from juvenile rheumatoid arthritis (JRA) patients and controls

Correlation between clinical diagnosis and the presence of autoantibodies to the recombinant

No. (%) of reactive sera ~~

No. of N-terminal N1 N2 N3 C-terminal Clinical diagnosis sera tested fragment fragment fragment fragment fragment

ANA-positive JRA Pauciarticular-onset JRA (22 with 77 37 (49) 13 (17) 32(42) 4(5) 7 (9)

Polyarticular-onset JRA (8 with 42 12(28) S(19) 4(9) 3(7) 3 (7) uveitis)

uveitis) Controls

ANA-negative JRA 20 3 (10) l (3) 3 (10) 2 (7) O(0) Nonrheumatic childhood illnesses 25 1(4) 1(4) l (4 ) 0 (0) 1(4)

* ANA = antinuclear antibody.

protein fragment of the rHMG-2 protein (Table l) , whereas only 9% of the JRA sera tested were immuno- reactive with the C-terminal part (amino acids 121-208).

Because the N-terminal part of HMG-2 appeared to harbor a major epitope, it was subdivided into protein fragments N1 (amino acids 1-55), N2 (amino acids 48-loo), and N3 (amino acids 95-127) (Figure 1). The results (Table 1) indicated that all but 5 of the pau- ciarticular JRA sera that were positive for the N- terminal HMG-2 protein fragment also reacted with the N2 fragment. At least 13 of the 37 sera reacting with the HMG-2 N-terminal protein gave a positive reaction with the N1 fragment, whereas only 4 reacted with the N3 fragment. The actual ELISA values (at 405 nm) used for these calculations are given in Table 2.

From these data, we conclude that the sera from the pauciarticular-onset JRA patients recognized a ma- jor epitope in the N2 fragment of the HMG-2 protein, whereas the N1 fragment seemed to harbor a minor epitope. No statistically significant preferences in immu- noreactions with the N2 fragment could be observed in any of the polyarticular-onset JRA patient sera. ANA- negative JRA patient sera and sera from children with nonrheumatic illnesses were negative by ELISA. Our previous observation that sera from adult patients with rheumatoid arthritis do not recognize any of the known HMG proteins as autoantigens suggests that these small nuclear proteins could be interesting candidates as marker antigens for a subset of JRA patients.

Competition ELISA experiments with synthetic peptides. For epitope mapping, we divided the entire N2 fragment of the HMG-2 protein into 8 overlapping (15-rner) synthetic peptides (Figure 2), which were used to compete individually with the N2 fragment of the HMG-2 protein for N2 antibodies present in the JRA

Table 2. Enzyme-linked immunosorbent assay values (adsorbance at 405 nm) of pauciarticular juvenile rheumatoid arthritis (JRA) sera with the N-terminal recombinant HMG-2 protein fragments’

JRA N1 N2 N3 serum no. fragment fragment fragment

6 10 13 21 22 26 30 56 62 72 75 84 88 02 96

100 105 109 111 112 116 120 131 15 1 I53 154 155 159 167 169 175 176 177 187 214 25 0 279

0.18 -

-

0.38 -

-

-

0.28

0.21 -

- -

0.14 0.15 -

-

-

-

-

-

-

0.20 0.18 -

-

-

-

-

0.10 0.16

0.12 0.22

0.26

-

-

-

-

0.38 0.36 0.25 0.58 0.25 0.28 0.20

0.54 0.26 0.69 0.26

0.28 0.48 0.32 1.08 0.28 0.56 0.63 0.32 0.47 0.24 0.30 0.78 1.08 0.28 0.41

-

-

- -

0.45 0.27 0.32

0.28 0.26 0.6 I

-

* Values are the mean optical density of duplicate samples minus 2 SD.

1806

100 - 90 - 80 -

e 70-

2 60-

a 50-

40-

.r)

8 3 0 - . < 20-

10 -

JUNG ET AL

... .....

. ’ i : ’

. . ’ .. . . . . .

;., . .’ : : . . . . ‘ I : . . j / i 1 : 1/ I ’f ”

* - - ! i i ; : . ..W., ,-~

1 1 - 1 1 4 i ..:.. .

AminoAcidSequence

Peptide No. Amino Acids

48-100 --.. W K T ~ ~ A K E K S K F E D ~ S D R ~ ~ R E ~ N ~ P ~ D K ~ G ~ D ~ N ~ K ~ P * S ~ . . - -

1 49-63 XTHSAKEKSKFEDPUL

2 54-68 REKSKFEDHRKSDRA

3 59-73 FEDMAKSDIULRYDRE

4 KSDKARYDREIMNW 64-78

5 69-83 RYDREHXNYVPPKGD

74-88

79-93

84-98 KRGKKRDPNAPWP

6

7

NXNFJPPRGDKKGKK

PPKGDKKGKKRDPNA

8

Figure 2. Synthetic HMG-2 peptides used in the competition enzyme- linked immunosorbent assay (ELISA). The partial amino acid se- quence of the HMG-2 protein is given for amino acids 48-100, in which the marked regions refer to the HMG box domains of HMG-2 and the unmarked stretch of 7 amino acids represents the “linker region” connecting the HMG boxes. The location of the 8 overlapping synthetic peptides used in the competition ELISAs with the immuno- reactive N2 fragment of the HMG-2 protein is also shown.

patient sera. In Figure 2, the amino acid sequence of the 8 overlapping peptides is listed, and the extent of competition with the N2 fragment of HMG-2 from 30 pauciarticular-onset JRA sera is given in Figure 3. The adsorbance at 405 nm obtained in the absence of inhibitory peptides was taken as 100% reaction. Peptides resulting in a >30% reduction of the ELISA readings were judged as having positive reactivity. The data show that a major epitope was present at amino acid positions 84-98 of HMG-2, since 25 of the 30 pauciarticular-onset

JRA sera tested (83%) competed with peptide S, which had the sequence KKGKKKDPNAPKRPP. Peptides 6 and 7, which partially overlapped peptide 8, competed with only 27% and 40% of the sera, respectively.

We therefore concluded that peptide 8 spanned the major HMG-2 epitope, which comprises most of the linker region of the HMG-112 box domains in the HMG-2 protein (Figure 3). With the exception of pep- tide 3 (27%), no significant competition ELISA values were obtained with any of the other peptides spanning the N2 fragment. The sequence FEDMAKSDKARY- DRE (amino acids 59-73) of peptide 3 bears a motif, DKA, which is also known to be present in the human HLA-DRa motif and in several putative environmental pathogens, i.e., Fcr V protein (Streptococcus), RfbA protein (Salmonella typhimurium), and protein HH LF1 (cytomegalovirus) (11).

Identification of amino acids involved in anti- body binding. To narrow down the amino acid sequence obtained by competition ELISA for the epitope on the HMG-2 protein, we synthesized 10 partially overlapping peptides with the SPOTS system, covering amino acids from 80 to 98 in the HMG-2 protein sequence. The results obtained with the synthesized peptides, followed by immunoblotting with a selected JRA serum, showed that the decapeptides 3, 4, 5 , and 6 gave the strongest immunoreactions. This result strongly supports the no- tion that the lysine-rich amino acid sequence, KKGKKKDP, located in the linker region of the HMG- 112 box domains, comprises the major epitope in the HMG-2 protein, since all other amino acid combinations before and after this sequence were weak or showed only background reactions (Figure 4A).

To determine the contribution of the individual amino acid residues of this epitope to its antigenicity, 8 peptides, each comprising 11 amino acids in length, were again synthesized on activated cellulose paper. In this set of peptides, each successive amino acid of the epitope was replaced by a glycine, except for glycine at position 86, which was replaced by alanine. Glycine, which fea- tures a hydrogen atom in the place of the side chain, was chosen in order to unequivocally assess the contribution of the original amino acid in the peptide. The results of immunoblot experiments with the mutated peptides, presented in Figure 4B, indicate that the lysine, glycine, and lysine residues at amino acid positions 85, 86, and 87, respectively, of the HMG-2 protein were absolutely necessary for antibody recognition, whereas the lysine at amino acid position 84 and the aspartic acid at position 90 seemed to be less important. The 2 lysines at position

ANTIBODIES AGAINST HMG PROTEINS IN JRA 1807

A - B

Peptide 1 : P K G D K K G K K K Peptide 5 . 1 : K K G K K K D P N A -

5.2: O K G K K K D P N A - 2: K G D K K G K K K D

3: G D K K G K K K D P 5.3: K a G K K K D P N A -

4: D K K G K K K D P N 5.4: K K D K K K D P N A-

5: K K G K K K D P N A 5.5: K K G a K K D P N A -

6: K G K K K D P N A P 5.6: K K G K B K D P N A -

7: G K K K D P N A P K - 5.7: K K G K K O D P N A-

8: K K K D P N A P K R - 5.8: K K G K K K H P N A -

9:

10 :

K K D P N A P K R P-

K D P N A P K R P P -

5.9: K K G K K K D H N A -

Figure 4. A, Fine mapping of the major HMG-2 epitope. Immunoblots, with a selected juvenile rheumatoid arthritis (JRA) serum (diluted 1:100), of 10 overlapping peptides synthesized on activated membranes (by the SPOTs system) arc 4hown. The deduced minimal cpitope sequence is given in boldface. B, Determination of thc contribution of individual amino acids in the epitope for its antigenicity. Eight mutated synthetic peptides (peptides 2-9) were again synthesized by the SPOTs system and probed by immunoblot with the same JKA serum as in A (diluted 1:100). Each successive amino acid was replaced by a glycine, except for pcptide 4, in which the glycine was replaced by alanine (indicated by the boxes).

88 and 89 did not appear to contribute at all to the interaction with the antibody.

Clinical data of JRA patients with epitope activity. Clinical descriptions of the 25 pauciarticular- onset and 2 polyarticular-onset JRA patients with KKGKKKDP epitope activity are given in Table 3. As can be seen, except for HMG-17 activity, antibody activity against other cellular components was not sig- nificantly higher than in the KKGKKKDP-negative pa- tients who served as controls. No positive correlation with uveitis or epitope activity could be observed.

DISCUSSION

Awareness of the location of epitopes on autoan- tigens is expected to increase the specificity and sensi- tivity of detection of autoantibodies, and thus improve the diagnosis of autoimmune diseases (12). In this report, we have described thc identification of a lysine- rich peptide, KKGKKKDP, in the nucleosomal HMG-2 protein that constitutes an antigenic epitope for sera from ANA-positive pauciarticular-onset JRA patients. This polypeptide lies in the linker region of the HMG box domains A and B of the HMG-1/2 proteins (13).

Table 3. Clinical description of pauciarticular and polyarticular juvenile rheumatoid arthritis (JRA) patients with KKGKKKDP cpitope activity”

No. (96) of reactive sera

No. of Mean age, No. (9%) Total Clinical diagnosis sera RF years with uveitis HMG- 17 &DNA ssDNA histones

Epitope-positive patients ~ 18 (72) Pauciarticular-onset JRA 25 8 5 (20) 13 (52) 2 (10) 3 (12)

Polyarticular-onset JRA 2 10 1 - -

~ 2 (6) 19 (63) Pauciarticular-onset JRA 30 8 7 (23) 9 (30) 4 (13)

- - -

Epitope-negative controls

* RF = rheumatoid factor; dsDNA = double-stranded DNA; ssDNA = single-stranded DNA.

1808 JUNG ET AL

Introduction of point mutations into the peptide epitope identified the amino acid residues lysine, glycine, and lysine at positions 85, 86, and 87, respectively, as abso- lutely necessary for antibody interaction.

Computer algorithms for the prediction of anti- genicity, utilizing the primary amino acid sequence of HMG-2, showed that the immunoreactive peptide (ami- no acids 84-91) belonged to one of the highest points of hydrophilicity on the HMG-2 protein (amino acids 83-90). This immunoreactive protein segment should be strongly hydrophilic, exhibiting a structural property that has been proposed as an important feature of antibody- binding epitopes (14).

HMG-1 and its closely related homolog, HMG-2, are proteins that have a tripartite structure consisting of 2 homologous basic DNA-binding domains, each com- posed of -80 amino acid residues, with a 7-amino acid linker region and a highly acidic C-terminal tail. Their ubiquity and sequence conservation suggest a funda- mental cellular role for the HMG-1 and HMG-2 pro- teins. The basic N-terminal A domain and central B domain of HMG-1 and HMG-2, the so-called HMG box motifs, are thought to recognize unique DNA structures, whereas the acidic C-terminal tail, with -30 consecutive aspartic or glutamic acid residues, is supposed to interact with the basic histones. It is known that HMG-1/2 proteins bind with high selectivity to 4-way junction DNA, a specific DNA structure, but do not bind to linear duplex or single-stranded DNA (1S,16).

Autoimmune responses to macromolecular com- plexes often result in the production of polyclonal antibodies to B cell epitopes, representing functionally active sites in the autoantigen. The mechanism of auto- antibody induction by nucleosomes is, so far, unknown, but immunoreactive oligonucleosomes could be gener- ated by apoptosis (17). There is increasing evidence to suggest that autoantigens are “cryptic” antigens, i.e., self peptides that bind with low affinity to the self class I1 major histocompatibility complex. As a result, the T cells that react with the cryptic antigenic peptide escape negative selection in the thymus and enter the peripheral lymphoid system, leading ultimately to the breakdown of tolerance to the autoantigen (17,18). Recently, com- puter algorithms have determined that the human HMG-112 proteins belong to a group of putative human self antigens, including the HLA-DRa motif, which carry an immunodominant epitope around the amino acid residues DKA. The 2 other autoantigens known to carry the DKA motif are the glutamic acid decarboxyl- ase in onset type I diabetes mellitus and the asialogly-

coprotein receptor in chronic active hepatitis and pri- mary biliary cirrhosis (1 1).

The search for infectious agents carrying the DKA sequence motif that might serve as a “trigger” in the breakdown of tolerance has led to the identification of a number of human pathogens acting as putative environmental immunogens (i.e., Camphylohacter, Pseudomonas, Salmonella, Streptococcus, parainfluenza virus, and cytomegalovirus) (11). The notion that patho- gens contribute to autoimmunity through “molecular mimicry” refers to such structural homologies between a self protein and a protein in an environmental pathogen. These structural similarities, which may be necessary at only a few amino acid positions in the immunodominant epitope that is shared by infectious pathogens and self antigens, may lead, in the event of an immune response to a pathogen, to inadvertent autosensitization and subsequent intramolecular epitope spreading (19-21). The autoimmune response against the nucleosomal HMG-2 protein in children with JRA might be based on such a mechanism.

ACKNOWLEDGMENTS

The authors wish to thank Dr. M. Buctin (NIH) for the human HMG-2 complementary DNA clone, and Dr. H. Mich- els and Prof. H. Truckenbrodt (Rheumatic Children’s Hospi- tal) for their contributions to this study.

1.

2.

3.

4.

5.

6.

7.

8.

9.

REFERENCES

Rudnicki RD, Ruderman M, Scull E, Goldenberg A, Rothfield N: Clinical features and serologic abnormalities in juvenile rheuma- toid arthritis. Arthritis Rheum 17: 1007-1015, 1974 Cassidy JT, Walker SE, Soderstrom SJ, Petty RE, Sullivan DB: Diagnostic significance of antibody to native DNA in children with juvenile chronic arthritis. J Pediatr 93:416-420, 1978 Wittemann B, Neuer G, Michels H, Truckenbrodt H, Bautz FA: Autoantibodies to nonhistone chromosomal proteins HMG-1 and HMG-2 in sera of patients with juvenile rheumatoid arthritis. Arthritis Rheum 33:1378-1383, 1990 Neuer G, Bustin M, Michcls H, Truckenbrodt H, Bautz FA: Autoantibodies to the chromosomal proteins HMG-17 in juvenile rheumatoid arthritis. Arthritis Rheum 35:472-475, 1992 Neuer G, Bautz FA, Bustin M, Michels H, Truckenbrodt H: Sera from JRA patients contain antibodies against a defined epitope in chromosomal protein HMG-17. Autoimmunity 17:23-30, 1994 Brewer EJ Jr, Bass J, Baum J, Cassidy JT, Fink C, Jacobs J, Hanson V, Levinson JE, Schaller J, Stillman JS: Current proposed revision of JRA criteria. Arthritis Rheum 20 (suppl 2):195-199, 1977 McCabe PC: Production of single-stranded DNA by asymmetric PCR. In, PCR Protocols: A Guide to Methods and Applications. Edited by MA Junis, DII Gelfand, JJ Sninsky, TJ White. New York, Academic Press, 1990 Gausepohl H, Boulin C, Kraft M, Frank R: Automated multiple peptide synthesis. Pept Res 5:315-320, 1992 Frank R: Spot synthe an easy technique for the positionally

ANTIBODIES AGAINST HMG PROTEINS IN JRA 1809

addressable, parallel chemical synthesis on a membrane support. Tetrahedron 48:9217-9232, 1992

10. Bustin M: Preparation and application of immunological probes for nucleosomes. Methods Enzyniol 170:214-251, 1989

11. Baum H, Butler P, Davies H, Sternberg MJE, Burroughs A K Autoimmune disease and molecular mimicry: an hypothesis. Trends in Biomechanical Science 18:140-144, 1993

12. Tan EM: Antinuclear antibodies: diagnostic markers for auto- immune diseases and probes for cell biology. Adv Immunol 44:93-151, 1989

13. Weir JM, Kraulis PJ, Hill CS, Raine ARC, Laue ED, Thomas JO: Structure of the HMG box motif in the B-domain of HMG-I. EMBO J l2:1311-1319, 1993

14. Hopp TP, Woods KR: Prediction of protein antigenic determi- nants from amino acid sequences. Proc Natl Acad Sci U S A 783824-3828, 1981

15. Jantzen HM, Admon A, Bell SP, Tijan R: Nucleolar transcription factor hUBF contains a DNA-binding motif with homology to HMG proteins. Nature 344:830-836, 1990

16. Sinclair AH, Berta P, Palmer MS, Hawkins JR, Griffiths BL, Smith MJ, Foster JW, Frischauf AM, Lovell-Badge R, Goodfellow PN: A gene from the human sex-determining region cncodes a protein with homology to a conserved DNA-binding motif. Nature 346:

17. Mohan C, A d a m S, Stanik V, Datta SK: Nucleosome: a major immunogene for pathogcnic autoantibody-inducing T cclls of lupus. J Exp Med 177:1367-1381, 1993

18. Brockenstedt LK, Gee RJ, Mamula MJ: Self pcptides in the initiation of lupus autoimmunity. J lmmunol 154:3S 16-3524, 1995

19. Topfer F, Gordon T, McCluskey J: Intra- and intermolecular spreading of autoimmunity involving the nuclear self-antigens La (SS-B) and Ro (SS-A). Proc Natl Acad Sci U S A 92375-879, 1995

20. Behar SM, Porcelli SA: Mechanisms of autoimmune disease induction: the rolc of the immune response to microbial patho- gens. Arthritis Rheum 38:458-476, 1995

21. Van Venrooij WJ, Pruijn GJM: Ribonucleoprotein complexes as autoantigens. Curr Opin Immunol 7:819-824, 1995

240-244,1990