A new conformational epitope generated by the binding of recombinant 70-kd protein and U1 RNA to anti–U1 RNP autoantibodies in sera from patients with mixed connective tissue disease

Authors


Abstract

Objective

To establish an enzyme-linked immunosorbent assay (ELISA) using a complex of in vitro–transcribed U1 RNA and recombinant 70-kd, A, and C proteins (C-ELISA) to detect anti–U1 RNP antibodies reactive in double immunodiffusion (DID), but not in ELISA using the proteins alone (P-ELISA).

Methods

Sera from 196 patients with mixed connective tissue disease were used to test reactivity in P- and C-ELISAs, and the specificity of the sera was also tested by DID and immunoprecipitation (IP).

Results

In P-ELISA, 15 of 196 sera positive for anti–U1 RNP in DID did not react, while all sera reacted in C-ELISA. The reactivity of 15 sera to the U1 RNA was tested by IP and ELISA, and only 3 sera reacted with the U1 RNA. These results indicated that the increased reactivity in C-ELISA was not due to the U1 RNA itself. We confirmed that the 70-kd and A proteins were bound directly to the U1 RNA by IP using antibodies to His-tag, and we tested the reactivity of the sera to the U1 RNA–70-kd protein complex and the U1 RNA–A protein complex by IP. All sera reacted with the U1 RNA–70-kd protein complex, and 1 sample reacted with the U1 RNA–A protein complex.

Conclusion

These results suggest that some anti–U1 RNP–positive sera specifically recognize the conformational structure altered by the binding of U1 RNA to the proteins, and the ELISA using U1 RNA and recombinant proteins is as useful as the DID method for detecting anti–U1 RNP antibodies.

Patients with systemic autoimmune disease produce a variety of antibodies against normal cellular components. Some of these antibodies are clinically useful for diagnosis, since their appearance is restricted to certain diseases (1). Mixed connective tissue disease (MCTD), which is characterized by an overlapping of the symptoms of systemic lupus erythematosus (SLE), systemic sclerosis, and polymyositis, is strongly associated with a high titer of antibodies to U1 small nuclear RNP (U1 snRNP) particles (2–3). The anti-Sm antibodies that are reactive with U1, U2, U4, U5, and U6 snRNP can be specifically detected in patients with SLE (4). U1 snRNP consists of U1 RNA and 9 proteins (70 kd, A, B′/B, C, D, E, F, and G) (5), and anti–U1 RNP antibodies react with 70 kd, A, and C; anti-Sm antibodies react with B′/B and D (6–8).

Because U1 snRNP can react with both anti–U1 RNP and anti-Sm antibodies, the proteins that specifically react with anti–U1 RNP must be isolated from U1 snRNP when attempting to establish an enzyme-linked immunosorbent assay (ELISA). Takeda et al purified native 70-kd and A proteins from rabbit thymus extract and developed an ELISA that could specifically react with ∼95% of anti–U1 RNP–positive sera (9). Investigators in our group have also developed an ELISA using recombinant 70-kd, A, and C proteins as an antigen in order to measure anti–U1 RNP antibodies (10); however, this system failed to detect ∼10% of anti–U1 RNP–positive sera detected by double immunodiffusion (DID) and immunoprecipitation (IP) assays, the same as the ELISA using native proteins.

This phenomenon may be related to the structure of U1 snRNP. U1 snRNP comprises U1 RNA and 9 proteins, and the binding of these proteins to U1 RNA was thought to form a complicated conformation (11–14). The proteins isolated from U1 snRNP obviously cannot reconstitute the conformational structure. In addition, it is possible that some anti–U1 RNP react only with U1 RNA and not with the U1 RNP proteins, because it has been reported that antibodies against naked U1 RNA were found in sera from patients with overlapping syndromes of SLE (15–17).

To clarify this issue, we tested the reactivity of anti–U1 RNP–positive sera in an ELISA using only recombinant proteins (P-ELISA) and in an ELISA using complexes of transcribed U1 RNA and recombinant proteins (C-ELISA). We then analyzed the epitopes that could be altered by the binding of the proteins to U1 RNA.

PATIENTS AND METHODS

Sera.

One hundred ninety-six sera obtained from patients at the hospital of Juntendo University were judged to be positive for anti–U1 RNP antibodies by DID. All patients fulfilled the criteria for MCTD proposed by the Ministry of Health and Welfare in Japan (18).

DID assay.

DID was performed using the extractable nuclear antigen 1 test (MBL, Nagoya, Japan) for measurement of anti–U1 RNP antibodies. The operation was performed according to the manufacturer's instructions.

Preparation of U1 RNA.

The coding sequence of U1 RNA was amplified from human genomic DNA extracted from Raji cells using 5′-ACCGGAATTCATACTTACCTGGCAGGGGAGATACCATGATCA-3′ (the underlined sequence is the Eco RI site) as the forward primer and 5′-CGCGGATCCCAGGGGAAAGCGCGAACGCAGTCCCCCACTA-3′ (the underlined sequence is the Bam HI site) as the reverse primer. After the polymerase chain reaction (PCR) product was digested with Eco RI/Bam HI, the digested DNA was cloned into the Eco RI–Bam HI site of the pGEM-4Z vector (Promega, Madison, WI), which allows in vitro transcription of the DNA insert using SP6 RNA polymerase, and transformed into Escherichia coli DH5α cells. Clones were screened for the inserts, and one clone was selected and sequenced. The sequence coincided with the gene of U1 RNA (GenBank accession no. J00318), and the cloned vector was termed pGEM-U1 RNA. For SP6 transcription, pGEM-U1 RNA was linearized with Bam HI.

The in vitro transcription was then performed. Briefly, 50 μg of linearized DNA was incubated in a total volume of 500 μl containing 80 mM HEPES (pH 7.5), 32 mM MgCl2, 2 mM spermidine, 40 mM dithiothreitol, 5 mM rNTPs (rATP, rCTP, rGTP, and rUTP), and 100 units of SP6 RNA polymerase (Promega) at 37°C for 4 hours. After transcription, DNase I (500 units) was added and incubation was continued for 15 minutes at 37°C. Finally, the RNA was extracted with phenol and purified with isopropanol precipitation. The in vitro–transcribed RNA was analyzed on a denaturing gel and was found to be U1 RNA.

Preparation of recombinant 70-kd, A, and C proteins of U1 snRNP.

The coding sequence of the recombinant 70-kd protein was amplified from complementary DNA (cDNA) from HeLa cells using 5′-CGCGGATCCATGACCCAGTTCCTGCCGCCCAACCTTCTG-3′ (the underlined sequence is the Bam HI site) as the forward primer and 5′-ACCGGAATTCCTCCGGCGCAGCCTCCATCAAATACCCATT-3′ (the underlined sequence is the Eco RI site) as the reverse primer. After the PCR product was digested with Bam HI/Eco RI, the digested DNA was cloned into the Bam HI–Eco RI site of pVL1393 (PharMingen, San Diego, CA) and transformed into E coli DH5α cells. Clones were screened for the inserts. One clone was selected, a 6×His linker sequence was added at its Eco RI site, and the selected clone was sequenced. The sequence coincided with the gene of human 70 kd (GenBank accession no. X04654). The recombinant 70 kd was expressed as a His-tag protein in insect cells (Spodopteria frugiperda Sf9 cells) and purified on a cobalt-chelate column (Talon; Clontech Laboratories, Palo Alto, CA) according to the manufacturer's instructions.

The human A protein construct was generated by PCR with cDNA from HeLa cells as a template using 5′-ATCGCCATGGATGGCAGTTCCCGAGACCCGCCCTAACCAC-3′ (the underlined sequence is the Nco I site) as the forward primer and 5′-ATCGCTCGAGCTTCTTGGCAAAGGAGATCTTCATGGCGT-3′ (the underlined sequence is the Xho I site) as the reverse primer. These primers were designed based on the human A messenger RNA (mRNA) sequence (GenBank accession no. X06347). The resulting PCR product was digested with Nco I/Xho I and cloned in-frame in the Nco I–Xho I sites of pET28a(+) (Novagen, Madison, WI). The nucleotide sequence of the construct was transformed into E coli BL21(DE3) cells. Since the recombinant A protein contained a 6×His-tag, the recombinant protein from bacteria was purified on a cobalt-chelate column (Talon) according to the manufacturer's instructions.

Recombinant C protein was prepared the same way as A protein, but excluding the PCR primer. The forward primer was 5′-ATCGCCATGGATGCCCAAGTTTTATTGTGACTACTGCGAT-3′ (the underlined sequence is the Nco I site) and the reverse primer was 5′-ATCGCTCGAGTCTGTCTGGTCGAGTCATTCCGGGCCGAGT-3′ (the underlined sequence is the Xho I site). The design of these primers was based on the human C mRNA sequence (GenBank accession no. X12517).

The purity of purified recombinant proteins (70 kd, A, and C proteins) was verified by sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis followed by Coomassie brilliant blue staining as well as by immunoblotting (IB) using anti–His-tag polyclonal antibody (MBL). Staining of the gel disclosed a minor band of less than expected molecular size in addition to a main band of expected molecular size (the main band only was ∼90% pure), but all of these bands were detected by IB. We judged that these minor bands did not come from insect cells or E coli cells, but came instead from decomposition of recombinant proteins.

ELISAs to test reactivity of anti–U1 RNP.

For the P-ELISA, recombinant 70-kd, A, and C proteins were coated onto 96-well assay plates (Maxisorp; Nunc, Rochester, NY) with mixed solutions of 5.5 μg/ml, 2.7 μg/ml, and 1.8 μg/ml, respectively. Coating was performed at 4°C for 16 hours. The coated plate was washed 2 times with phosphate buffered saline (PBS) and blocked with 1% bovine serum albumin (BSA) for 1 hour at room temperature. In the ELISA, sera from patients and normal donors were diluted 100-fold with PBS containing 0.15% Tween 20 and 1% casein enzymatic hydrolysate (Sigma, St. Louis, MO). The diluted solution was added at 100 μl/well. After incubation for 1 hour at room temperature, the wells were washed 4 times with PBS containing 0.05% Tween 20 (PBS–Tween 20). Then, 100 μl of anti-human IgG conjugated to peroxidase (MBL) diluted 1:10,000 in 20 mM HEPES (pH 7.4), 135 mM NaCl, and 0.1% p-hydroxyphenylacetic acid was added to each well. After incubation for 1 hour at room temperature, plates were washed 4 times with PBS–Tween 20, and bound antibodies were detected with 3,3′,5,5′-tetramethylbenzidine as a substrate. The 30-minute reaction was stopped by the addition of 100 μl/well of 1N sulfuric acid. Plates were read at an optical density of 450 nm (OD450 nm).

For the C-ELISA, 20 μg/ml of in vitro–transcribed U1 RNA in addition to recombinant RNP at the concentration used for the P-ELISA was mixed in PBS, and the solution was incubated for 10 minutes at 25°C. The incubated solution was coated onto 96-well assay plates. C-ELISA was measured the same way as P-ELISA.

For the U1 RNA ELISA, 50 μg/ml of methyl ester of BSA (Sigma) was used to precoat 96-well assay plates for 2 hours at room temperature. After precoating, 25 μg/ml of in vitro–transcribed U1 RNA was coated. U1 RNA ELISA was measured in a manner similar to that of C-ELISA and P-ELISA, except for the following: 1) patient and normal donor sera were diluted 200-fold with 10 mM MOPS (pH 7.0), 125 mM NaCl, 0.05% Tween 20, and 1% BSA; 2) the anti-human IgG conjugated to peroxidase was diluted 1:5,000; and 3) in the U1 RNA ELISA assay, the results (percentages) were calculated relative to the mean normal serum OD450 nm value plus 3SD (n = 50); values >1% were considered positive.

Analysis of binding between U1 RNA and RNP proteins by IP.

Twenty microliters of anti–His-tag polyclonal antibody (MBL) was mixed with 2 mg of protein A–Sepharose CL-4B (Pharmacia, Piscataway, NJ) in 500 μl of IP buffer (10 mM Tris HCl, 500 mM NaCl, 0.1% Nonidet P40 [NP40], pH 8.0) and rotated for 12 hours at 4°C. The protein A–Sepharose was washed 5 times in 500 μl of IPP buffer using a 10-second centrifugation. The washed protein A–Sepharose was mixed with 1 ml of complex solution and rotated for 1 hour at 4°C. Preparation of the complex solution was performed by incubation for 10 minutes at 25°C after addition of His-tagged 70 kd (5.5 μg/ml), His-tagged A (2.7 μg/ml), or His-tagged C (1.8 μg/ml) in the presence of U1 RNA (20 μg/ml) in PBS.

The protein A–Sepharose was washed 7 times with 1 ml of NET-2 buffer (50 mM Tris HCl, 150 mM NaCl, 0.05% NP40, pH 7.4). To extract bound U1 RNA, the washed protein A–Sepharose was resuspended in 300 μl of NET-2 buffer and 30 μl of 3M sodium acetate, 30 μl of 10% SDS, and 300 μl of phenol–chloroform. The tube was agitated on a vortex and centrifuged for 5 minutes. The aqueous phase was recovered, and U1 RNA was collected by ethanol precipitation. The pellet was vacuum-dried and dissolved in 20 μl of sample buffer. The sample was denatured at 65°C for 5 minutes. An aliquot was applied to a 7M urea–10% polyacrylamide gel. Electrophoresis was carried out at 400V in Tris–borate–EDTA buffer. The gel was stained using a Silver Stain Plus kit (Bio-Rad, Richmond, CA).

IPs using HeLa cell extract, transcribed U1 RNA, and the complexes of transcribed U1 RNA and recombinant proteins.

IP by using HeLa cell extract was performed as described previously (19). The IP for in vitro–transcribed U1 RNA was partially modified. The change was the addition of 20 μl of U1 RNA (5 mg/ml) in place of the HeLa cell extract (200 μl).

IPs using the complex of U1 RNA and 70-kd protein or the complex of U1 RNA and A protein were also partially modified. The change was the addition of 1 ml of complex solution in place of the HeLa cell extract (200 μl) and NET-2 buffer (400 μl). The complex solutions were prepared by incubation at 25°C for 10 minutes after mixing 70-kd protein (5.5 μg/ml) or A protein (2.7 μg/ml) in PBS containing U1 RNA (20 μg/ml).

RESULTS

Reactivity of anti–U1 RNP–positive sera from MCTD patients in ELISA using U1 RNP polypeptides and in the DID assay.

One hundred ninety-six samples of anti–U1 RNP–positive sera and 96 samples of normal donor sera were measured both by the DID method and by ELISA using recombinant 70-kd, A, and C proteins as an antigenic substrate (P-ELISA). As shown in Figure 1, the range of normal sera OD450 nm values in P-ELISA was 0.032–0.203. The mean + 3SD of these OD450 nm values was 0.243, and this was used as a cutoff value. Fifteen of 196 MCTD patient sera were judged to be negative by P-ELISA, but positive by the DID method, and these sera were negative in IB on HeLa cells. Since these sera had high titers of anti–U1 RNP in the DID assay, they were thought of as false negatives. Results for these sera in DID and P-ELISA are shown in Table 1. Even when the amount of recombinant proteins coated on the ELISA plate was increased maximally, an improvement in the reactivity of anti–U1 RNP in the false-negative sera could not be observed (data not shown).

Figure 1.

Reactivity of anti–U1 RNP–positive sera from patients with mixed connective tissue disease (MCTD) in enzyme-linked immunosorbent assay (ELISA) using recombinant RNP polypeptides (P-ELISA). One hundred ninety-six serum samples from MCTD patients that had been judged to be positive by double immunodiffusion and 96 serum samples from normal donors were measured by P-ELISA. Shown are the distributions of optical density values (at 450 nm) for anti–U1 RNP–positive and normal sera in the ELISA. The horizontal line shows the mean + 3SD of normal values.

Table 1. Characterization of false-negative sera from the MCTD patients that were judged to be positive in DID, but negative in P-ELISA*
PatientAnti–U1 RNP titer by DIDP-ELISAU1 RNA ELISA
+ or −OD450 nm+ or −%
  • *

    MCTD = mixed connective tissue disease; DID = double immunodiffusion; P-ELISA = recombinant RNP polypeptides enzyme-linked immunosorbent assay; OD = optical density.

  • Values for DID represent the maximal dilution that gives a positive result.

  • Calculated relative to the mean normal serum OD450 nm value plus 3SD (n = 50).

11:320.235+1.4
21:160.1600.7
31:160.225+2.6
41:160.2040.8
51:80.1520.7
61:80.093+1.7
71:40.1080.5
81:40.1690.6
91:20.0660.5
101:40.1120.4
111:40.2180.6
121:40.0690.3
131:40.1820.3
141:40.2150.6
151:20.2080.7

It has previously been reported that some anti–U1 RNP sera from patients with SLE overlap syndromes specifically react with U1 RNA (15–17). We tested whether the false-negative sera in the ELISA could react with U1 RNA in ELISA using transcribed U1 RNA as antigen. Although 3 of 15 false-negative sera had anti–U1 RNA antibodies, the others did not react with U1 RNA (Table 1).

Reactivity of false-negative sera in IP using HeLa cell extract and transcribed U1 RNA.

The specificity for anti–U1 RNP antibodies of the false-negative sera was confirmed by IP using HeLa cell extract. All sera reacted with U1 snRNP and precipitated U1 RNA, as shown in Figure 2A. These results coincided with the DID findings. We further analyzed the reactivity of the false-negative sera with U1 RNA by IP. As shown in Figure 2B, 3 false-negative sera were reactive with transcribed U1 RNA, the same result as was found in ELISA. These results confirmed that the false-negative sera in P-ELISA were specific for U1 snRNP and that the difference in reactivity between DID and the ELISA did not depend on U1 RNA itself.

Figure 2.

Immunoprecipitation of RNA from HeLa cell extracts containing U1 small nuclear RNP particles or in vitro–transcribed U1 RNA with patient sera. A, Immunoprecipitation of RNA from HeLa cell extracts. The total RNA lane contains the marker HeLa cell total RNA. The positions of U2 RNA, U1 RNA, 5S RNA, hY1–hY5 RNA, and tRNA are shown at the left. B, Immunoprecipitation of in vitro–transcribed U1 RNA. The U1 RNA lane contains 1 μg of the marker in vitro–transcribed U1 RNA. Lanes 1–15 in A and B contain the 15 MCTD patient sera (same sera shown in Table 1) judged to be positive by double immunodiffusion (DID), but negative by P-ELISA. In B, the smear band for patients 8 and 9 is an artifact of silver-stained dust on the gel, and is not RNA. See Figure 1 for other definitions.

Reactivity of anti–U1 RNP antibodies in false-negative sera with complexes of U1 RNA and recombinant RNP proteins.

To further analyze the specificity of anti–U1 RNP antibodies in false-negative sera, we attempted to prepare complexes of transcribed U1 RNA and recombinant proteins of U1 snRNP. We then tested the reactivity of the complexes to the anti–U1 RNP false-negative sera.

First, transcribed U1 RNA and recombinant proteins of U1 snRNP were mixed in PBS and incubated as described in Patients and Methods, and the binding of those molecules was analyzed by IP using anti–His-tag antibodies. As shown in Figure 3, using the mixture of U1 RNA and 70-kd protein (lane 4) and the mixture of U1 RNA and A protein (lane 5), anti–His-tag antibodies reacted with His-tagged recombinant 70-kd and A proteins and precipitated U1 RNAs that were bound to the proteins. The mixture of U1 RNA and all proteins (70 kd, A, and C) also reacted with anti–His-tag antibodies and precipitated U1 RNA (lane 7). In contrast, anti–His-tag antibodies could not precipitate U1 RNA in IP using a mixture of U1 RNA and C protein (lane 6), the same result as that found with U1 RNA used as a control (lane 3). The manner of binding coincided with that reported previously (13–14), and the results revealed that U1 RNA and recombinant 70-kd and A proteins formed complexes when mixed together in PBS.

Figure 3.

Immunoprecipitation using combinations of U1 RNA and RNP proteins. Combinations of recombinant 70-kd, A, and C proteins were tested for their ability to bind to in vitro–transcribed U1 RNA in phosphate buffered saline. After incubation of these recombinant proteins with in vitro–transcribed U1 RNA, the RNA immunoprecipitated via anti–His-tag antibody–bound protein A–Sepharose. Lanes show silver-stained polyacrylamide gel containing the immunoprecipitated RNA or marker RNA, as follows: RNA size marker (lane 1); 10% input of in vitro–transcribed U1 RNA (lane 2); control incubation of U1 RNA in the absence of recombinant RNP proteins (lane 3); and incubation of U1 RNA in the presence of His-tagged 70 kd (lane 4), His-tagged A (lane 5), His-tagged C (lane 6), or His-tagged 70 kd, His-tagged A, and His-tagged C (lane 7). These results show that recombinant 70-kd and A proteins bind directly to in vitro–transcribed U1 RNA.

Next, using the complex of U1 RNA and 70-kd, A, and C proteins as a source of antigen, the reactivity of anti–U1 RNP false-negative sera with the complex was tested by ELISA (Figure 4). At the same time, the amount of U1 RNA in the mixture was also estimated by adding various concentrations of U1 RNA. The reactivity of sera increased in parallel with the amount of U1 RNA up to 20 μg/ml, and then plateaued.

Figure 4.

Improvement of reactivity of anti–U1 RNP autoantibodies in false-negative sera by the addition of U1 RNA to recombinant U1 RNPs in enzyme-linked immunosorbent assay (ELISA). In vitro–transcribed U1 RNA (0, 5, 10, 20, 40, and 80 μg/ml) was added to phosphate buffered saline (PBS) solution containing recombinant 70-kd protein (5.5 μg/ml), A protein (2.7 μg/ml), and C protein (1.8 μg/ml). The mixture was used as the antigenic substrate in ELISA. Shown are the ELISA results of 15 patient sera (same false-negative sera shown in Table 1) against additional U1 RNA in PBS solution.

Comparison between ELISA using recombinant RNP polypeptides and ELISA using U1 RNA and recombinant RNP polypeptides.

Since it was shown that anti–U1 RNP antibodies in false-negative sera could react with the complex of U1 RNA and recombinant RNP polypeptides, we compared the ELISA using recombinant RNP polypeptides (the P-ELISA) with the ELISA using U1 RNA and recombinant RNP polypeptides (the C-ELISA). To assure that RNP polypeptides antigen efficiently coated the ELISA wells irrespective of the presence or absence of U1 RNA, the amounts of antigen coated on both P-ELISA wells and C-ELISA wells were compared by the use of anti–His-tag antibodies. We confirmed that there were almost equal amounts of coated antigen for the P-ELISA and the C-ELISA.

One hundred ninety-six serum samples from patients with MCTD and 96 serum samples from normal donors were measured by both ELISA systems. All of the MCTD samples were judged to be positive by the DID method. The range of normal sera OD450 nm values in C-ELISA was 0.014–0.159. The mean + 3SD of these OD450 nm values was 0.162, and this was used as the cutoff value. Ninety-two percent of the sera that were positive for anti–U1 RNP by DID reacted in the P-ELISA, while 100% of the sera reacted in the C-ELISA (Figure 5A). These data suggested that anti–U1 RNP antibodies in false-negative sera reacted with the complexes of U1 RNA and the proteins and showed the same reactivity in DID.

Figure 5.

A, Comparison of the P-ELISA with the ELISA using U1 RNA and recombinant RNP polypeptides (C-ELISA). Recombinant RNP (70 kd, A, and C) was used as the antigenic substrate in the P-ELISA. The complex reconstituted from U1 RNA and recombinant RNP was used as an antigenic substrate in the C-ELISA. The complex was prepared by mixing U1 RNA with recombinant RNP (70 kd, A, and C). One hundred ninety-six serum samples from MCTD patients that had been judged to be positive by double immunodiffusion and 96 serum samples from normal donors were measured with both ELISA systems. Shown are the distributions of optical density values (at 450 nm) for anti–U1 RNP–positive and normal sera in both systems. The measurements were made in the same way except for immobilized antigen on the ELISA plate. Horizontal lines indicate the mean + 3SD of normal values. B, Correlation between the reactivities in P-ELISA and C-ELISA for sera from 196 MCTD patients. Note that there is not a simple linear relationship. See Figure 1 for other definitions.

Figure 5B is a scatterplot showing the correlation of the results from the P-ELISA with those from the C-ELISA for 196 sera from MCTD patients. The correlation coefficient was 0.894 (P < 0.0001 for P-ELISA versus C-ELISA, by t-test). There was not a simple linear relationship between the results of the ELISAs. The C-ELISA was more sensitive than the P-ELISA for almost all sera, but the extent of increased reactivity depended on the characterization of each individual serum sample. These data suggested that anti–U1 RNP antibodies in sera from MCTD patients recognized the conformational epitope generated by the complex from U1 RNA and RNP proteins. This phenomenon was not limited to false-negative sera. Almost all sera from MCTD patients contained anti–U1 RNP antibodies that could recognize the conformational epitope.

Analysis of reactivity of anti–U1 RNP antibodies in false-negative sera with complexes of U1 RNA and proteins of U1 snRNP by IP.

It was shown that anti–U1 RNP antibodies in false-negative sera reacted with the complexes of U1 RNA and proteins, but it was not clear which complexes of U1 RNA and proteins were responsible for increasing the reactivity of the sera. Since 70-kd and A proteins could bind directly to transcribed U1 RNA (as shown in Figure 3), we examined whether the U1 RNA–70 kd protein and U1 RNA–A protein complexes could react with the anti–U1 RNP antibodies in false-negative sera by IP. We selected 8 sera (from patients 2, 4, 5, 8, 9, 11, 12, and 13) to test in this system, since sera from patients 1, 3, and 6 had anti–U1 RNA antibodies that could react with U1 RNA itself, and since only a small amount of sera remained from patients 7, 10, 14, and 15. As shown in Figure 6A, all 8 sera reacted with the U1 RNA–70-kd protein complex, while only 1 serum sample reacted with the U1 RNA–A protein complex (Figure 6B). These data suggested that the increased reactivity in the ELISA using the complexes of U1 RNA and the proteins depended on an epitope that was altered on the complex when 70-kd protein bound to U1 RNA.

Figure 6.

Immunoprecipitation of U1 RNA from U1 RNA–protein complexes. Shown are immunoprecipitation of U1 RNA from U1 RNA–70-kd protein complex (A) and U1 RNA–A protein complex (B) with 8 false-negative sera judged to be positive by the double immunodiffusion method, but negative by enzyme-linked immunosorbent assay (ELISA) using recombinant RNP polypeptides and negative by U1 RNA ELISA (same sera shown in Table 1). Urea (7M) and 10% polyacrylamide gels of phenol-extracted immunoprecipitates from the solution containing U1 RNA–70-kd protein complex or U1 RNA–A protein complex were developed with silver stain. Lane 1 contains 1 μg of in vitro–transcribed U1 RNA. Lane 2 contains a normal control serum. Lanes 3–10 contain the false-negative patient sera, all of which immunoprecipitated the U1 RNA–70-kd protein complex, while only 1 of them immunoprecipitated the U1 RNA–A protein complex.

DISCUSSION

In general, all sera from MCTD patients are positive for anti–U1 RNP antibodies by DID. When investigators in our group attempted to develop an ELISA to detect anti–U1 RNP antibodies using recombinant proteins, it was found that ∼10% of samples judged to be positive by DID did not react (10). Although the ELISA system was superior to DID as a quantitative assay, it had less sensitivity, potentially resulting in failure to diagnose MCTD. To solve this problem, we analyzed the characteristics of anti–U1 RNP antibodies reactive in DID, but not in ELISA, using proteins of U1 snRNP.

U1 snRNP is composed of U1 RNA, U1 RNP proteins (70 kd, A, and C), and common Sm proteins (B′/B, D, E, F, and G). It has been reported that ∼40% of anti–U1 RNP–positive sera can react with U1 RNA (17). It is therefore possible that false-negative sera react only with U1 RNA and thus become positive in DID, but negative in ELISA. However, our analysis by IP and ELISA using U1 RNA as an antigen source revealed that only 3 of 15 false-negative sera could react with U1 RNA.

Another explanation for the difference in reactivity between DID and ELISA is the reaction of anti–U1 RNP antibodies to epitopes that are associated with the conformational structure of the molecule. The structure of U1 snRNP is illustrated in Figure 7. The 70-kd protein binds to stem-loop I of U1 RNA, and the A protein binds to stem-loop II of U1 RNA through an RNA binding domain known as the RNP-80 motif (11–12). In contrast to the 70-kd and A proteins, the C protein, another protein that can react with anti–U1 RNP, does not have an RNA binding domain and cannot bind directly to naked U1 RNA, but it does have a zinc-finger domain. Therefore, it joins U1 snRNP through protein–protein interactions of its zinc-finger domain with the 70-kd and common Sm proteins (13–14). It is possible that this complicated binding of RNP proteins to U1 RNA alters the conformational structure of U1 RNA and induces the reactivity of the complex of U1 RNA and proteins to certain anti–U1 RNP antibodies.

Figure 7.

Structure of the U1 small nuclear RNP (U1 snRNP) particle. U1 snRNP is composed of U1 RNA, RNP proteins (70 kd, A, and C), and common Sm proteins (B′/B, D, E, F, and G). The structure of U1 RNA consists of single-stranded RNA and double-stranded RNA called stem-loops I, II, III, and IV; all solid lines, including the stem-loop lines, indicate U1 RNA. The 70-kd protein can bind directly to stem-loop I of U1 RNA, and the A protein can bind directly to stem-loop II of U1 RNA through an RNA binding domain known as the RNP-80 motif. The common Sm proteins bind as a complex to single-stranded RNA at the position shown. Since C protein does not have an RNA binding domain, it cannot bind directly to U1 RNA; however, it does have a zinc-finger domain. It therefore joins U1 snRNP through protein–protein interactions (arrows) of its zinc-finger domain with the 70-kd and common Sm proteins.

To clarify this issue, we first determined by IP using antibodies to His-tag whether the recombinant proteins known to be able to bind to U1 RNA could specifically bind to U1 RNA in our system. We found that the recombinant 70-kd and A proteins bound to the transcribed U1 RNA through the anti–His-tag antibodies on protein A–Sepharose and precipitated U1 RNA together with the His-tagged recombinant proteins, while U1 RNA was not precipitated by IP using the mixture of C protein (which does not have an RNA binding domain) and transcribed U1 RNA. These results revealed that recombinant 70-kd and A proteins can specifically bind to the U1 RNA in our preparation. Various concentrations of transcribed U1 RNA were mixed with constant amounts of the recombinant proteins in PBS, and the formation of a U1 RNA–RNP complex was induced. Then, using the complex as an antigen source, the ELISA system showed that the increase in reactivity of anti–U1 RNP antibodies in false-negative sera with U1 RNA was dose dependent.

In addition, the ELISA using the complexes of U1 RNA and the proteins showed higher sensitivity than the ELISA using only recombinant proteins, and positive sera in DID were all detected in this system. These results suggested that the binding of U1 RNA and the proteins altered the epitope that is associated with the conformational structure of the U1 RNA and/or the proteins. Using U1 RNA–70-kd protein and U1 RNA–A protein as antigens in IP, we then analyzed which complexes were responsible for recovering the reactivity of false-negative sera in the ELISA. It was shown that all of the false-negative sera reacted with the U1 RNA–70-kd protein complex, while only one reacted with the U1 RNA–A protein complex. These results suggested that the binding of the U1 RNA to the proteins, mainly to the 70-kd protein, induced the alteration of the conformational structure of the U1 RNA and/or the bound proteins and provided higher sensitivity to anti–U1 RNP antibodies, the same as the DID method.

This is the first report to describe increased reactivity of an autoantigen to reconstituted RNA and protein in vitro. Since many autoantigens are complexes of RNA and protein, the conformational epitope on the complex may not be restricted to anti–U1 RNP antibodies. For example, there are several reports that 10–35% of sera positive for anti-SSA/Ro by IP do not react with 60-kd SSA/Ro by IB (20, 21). Boire et al studied the reactivity in IB of anti-SSA/Ro antibodies detected by DID and IP and found that 10% of these sera did not react with 60-kd SSA/Ro polypeptides. It was suggested that some sera recognize conformational determinants on the antigen (20). Tsuzaka et al also studied the reactivity of anti-SSA/Ro antibodies to native and denatured SSA/Ro; and showed that 6 of 14 sera from SLE patients with anti-SSA/Ro antibodies reactive by IP (43%) recognized denatured SSA/Ro, while only 4 of 46 sera from patients with primary Sjögren's syndrome (SS) and only 1 of 12 sera from patients with both SLE and SS with antibodies to native SSA/Ro reacted with denatured SSA/Ro by IB (21). Those authors concluded that antibodies to SSA/Ro in patients with SS targeted only the conformational epitopes on the antigen.

Moreover, the antigen of anti-SSA/Ro antibodies was Ro RNP, the complex of which consists of at least 1 of 2 polypeptides (Ro and La) associated with 1 of 4 small RNAs of 83–112 nucleotides in length, called hY1, hY3, hY4, and hY5 (22–23). In addition to antibodies against the 60-kd SSA/Ro protein, the presence of anti-hY5 RNA was also identified (24). That both the protein and related RNA in the complex were recognized is the same result as that for anti–U1 RNP. The anti-SSA/Ro also may have a conformational epitope generated by the binding of 60-kd SSA/Ro and hY5 RNA.

Finally, we found that anti–U1 RNP antibodies that were positive in DID but negative in ELISA using recombinant proteins could react with epitopes altered by binding of U1 RNA to 70-kd and A proteins; we also found that the ELISA system using the complexes of transcribed U1 RNA and recombinant proteins (70 kd, A, and C) had the same sensitivity as DID. These results suggested that the ELISA using reconstituted antigen is as useful as DID for quantifying anti–U1 RNP antibodies in patients with various connective tissue diseases, including MCTD.

Acknowledgements

We thank Y. Murakami, S. Thunekawa, M. Takemura, K. Kuroda, and S. Nakawaki for stimulating discussion, and S. Miyashita for excellent technical assistance in our laboratory. We also thank the members of the Department of Collagen Disease Medicine of Juntendo University.

Ancillary