THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 260, No. 21, Issue of September 25, pp. 11781-11786 1985 0 1985 by The American Society of Biological Chemists, InC. Printed in d.S.A.

Alu RNA-Protein Complexes Formed in Vitro React with a Novel Lupus Autoantibody*

(Received for publication, March 11, 1985)

Ryszard Koles, Lucille D. FrescoQ, Jack D. KeeneQ, Philip L. Cohenll, Robert A. Eisenbergll, and Phyllis Golden Andrews From the Department of Pharmacology and Cancer Research Center, University of North Carolina, Chapel Hill, North Carolina 27514, the $Department of Microbiology and Immunology, Duke University Medical Center, Durham, North Carolina 27710, and the (IDivision of Rheumatology, University of North Carolina, Chapel Hill, North Carolina 27514

We have screened sera from patients with systemic lupus erythematosus for reactivity with RNA tran- scribed in vitro using HeLa whole cell extracts. Sera from 14 out of 114 patients precipitated an RNA tran- scribed by RNA polymerase 111 from a plasmid con- taining an Alu family sequence ( i . e. the repetitive DNA sequence that is cut by the Alu restriction enzyme) located upstream from the human rG-globin gene. These Alu transcripts were not precipitated by anti- La, anti-Sm, anti-RNP or anti-Ro antibodies, suggest- ing that Alu RNA was precipitated by a previously undescribed lupus specificity. Analysis of [36S]methio- nine-labeled immunoprecipitates indicated that Alu RNA binds a protein of about 68 kDa. This protein may be Alu specific since three different Alu tran- scripts were precipitated by the anti-Alu sera whereas another RNA polymerase 111 transcript, adenovirus VA I RNA, was not precipitated by the same sera.

Sera from patients with systemic lupus erythematosus and related autoimmune diseases exhibit antibodies directed against numerous cellular antigens (see Ref. 1 for review). Since the original observation by Lerner and Steitz (2) that principal classes of these antibodies (anti-RNP,’ anti-La, anti-Ro, anti-Sm) recognize small nuclear ribonucleoprotein particles, the sera from systemic lupus erythematosus patients have been used to elucidate the structure of several RNP particles (3, 4) as well as the potential role of U1-RNP in mRNA splicing (5-7).

The anti-La class of autoantibodies is believed to recognize RNPs formed by RNA polymerase 111 transcripts (3,8). These transcripts include precursors to small cellular RNAs such as tRNA, 5 S RNA, and 7 S RNA (9, 10) as well as several viral RNAs (11-13). In addition the vesicular stomatitis virus leader RNP, which is transcribed by a virus-specific RNA polymerase, is also precipitated by anti-La sera (14, 15). The La antigen seems to bind to the oligo-U sequence typically present at the 3‘-end of the polymerase 111 transcripts (16). The role of the La antigen in vivo is not known. There were

* This work was supported in part by Grant GM32994 to R. K. from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of the American Cancer Society Junior Faculty Award. To whom reprint requests should be addressed.

The abbreviations used are: RNP, ribonucleoprotein particle; SRP, signal recognition particle. Alu family sequence, the repetitive DNA sequence that is cut by the Alu restriction enzyme.

suggestions that La antigen is a transcription factor (9) or that it is complexed with a transcription factor (17) since sera with anti-La specificity inhibited i n vitro transcription of RNAs transcribed by RNA polymerase 111 (17). However, Francoeur and Mathews (13) reported that i n vitro transcrip- tion of adenovirus VA RNA, which is precipitated by anti-La antibody, could not be inhibited by depletion of the extract with anti-La sera.

We describe here a previously unidentified antibody speci- ficity from certain lupus and scleroderma patients, distinct from anti-La, anti-Ro, anti-RNP, and anti-Sm, that reacted with i n vitro synthesized RNPs containing Alu transcripts but did not precipitate another RNA polymerase 111 transcript synthesized in vitro. The antigen is a protein which appears to bind to the 5‘ half of the Alu RNA.

EXPERIMENTAL PROCEDURES

Whole cell and nuclear extracts were obtained from HeLa cells as described by Manley et al. (18) and Dignam et al. (19), respectively. Sera were from a bank of frozen samples which had been sent for characterization of possible antibodies to extractable nuclear antigens and were from rheumatology patients of the North Carolina Memorial Hospital. In most experiments, sera were used without further treat- ment. In some instances purified IgG fraction (20) was used.

Plasmids A36 (21) and RBa-1 ((22) a kind gift of Dr. T. Maniatis, Harvard University) have been described. Plasmid dl-313 which car- ries genes coding for VA I and VA I1 RNA from adenovirus serotype 5 was a gift from Dr. Dana Fowlkes, Department of Pathology, University of North Carolina. Plasmid U6-13, a subclone of a plasmid described previously (23)’ carries a truncated copy of an Alu family gene. All plasmids were linearized before transcription as indicated in Fig. 1.

In vitro transcription was carried out in a total volume of 50 ~l as described (18). At the end of incubation samples were diluted with 150 pl of 50 mM Tris-HC1, pH 7.4, 150 mM NaCI, 0.05% Nonidet P- 40 (Sigma), immunoprecipitated as described by Lerner and Steitz (2), and the isolated RNA was analyzed by polyacrylamide gel elec- trophoresis (24). 5 ~1 of undiluted serum was used in each immuno- precipitation.

HeLa cells grown in monolayers were labeled overnight with [%SI methionine at 0.5 mCi/ml or with 13H]uridine at 100 pCi/ml. I n vivo labeled cell extracts were prepared and immunoprecipitated as de- scribed (25). [3H]Uridine-labeled RNA was analyzed as above, and [35S]methionine-labeled proteins were analyzed by electrophoresis on sodium dodecyl sulfate containing 10% polyacrylamide gels (26) fol- lowed by fluorography.

RESULTS

Immunoprecipitation of in Vitro Transcribed Alu RNA-A plasmid carrying both an Alu family sequence and a fragment of the 7”-globin gene (A36, Ref. 21) was used as a template for i n vitro transcription with HeLa whole cell extract (18).

R. Kole, unpublished data.

11781

11782 Immunoprecipitation of in Vitro Transcribed Alu RNA

The structure of the plasmid is shown in Fig. 1. This DNA clone allows simultaneous transcription of the Alu RNA (RNA polymerase I11 transcript) and the human yc-globin gene (polymerase I1 transcript). In vitro transcribed RNA was precipitated with sera from lupus patients essentially as de- scribed by Lerner and Steitz (2) and analyzed by polyacryl- amide gel electrophoresis (24). Fig. 2 shows the results of RNA immunoprecipitation with a series of sera from lupus and scleroderma patients.

Certain sera which we have determined to have anti-RNP, anti-Sm, anti-Ro, and anti-SCL-70 specificities precipitated in vitro transcribed Alu RNA (Fig. 2, lanes 2, 3, 4, and lo ) , suggesting that anti-Alu reactivity is present in these sera in addition to, and is distinct from, the major classes of auto-

Eco R I EcoRl

Alu

555 1427

A 36 8 -globm

i i a-globon F r R 1 R B a l - 350

Plt I

i E]-, U6-13 - 200

\ p i T r Hmdlll

dl 313 "

160 163

FIG. 1. Structure of the DNA templates. Boxes represent tran- scribed genes. Arrows represent the RNA transcribed in uitro. The length of the transcripts in nucleotides is indicated. The size of the genes and the distances between them are not to scale. The DNA templates were linearized by cleavage at the indicated restriction sites.

N AL JH VG EC MF CF KM JC JM ala aSmaRNP T B L g ; " . . ~ ~ ~ ; . . - . .. - _ , ." - ~ ,.- ~ ~- ., ....

-7-globin

-Alu

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4

FIG. 2. Immunoprecipitation of Alu RNA transcribed in ui- tro from A36 DNA template. RNA was synthesized in vitro and immunoprecipitated as described under "Experimental Procedures." Electrophoresis was on an 8% urea-polyacrylamide gel (24). The gel was autoradiographed overnight at -70 "C with DuPont Lightning Plus screen and Kodak XAR-5 film. Lane 1, serum from healthy individual; Lanes 2-10, immunoprecipitates by patients' sera; Lanes 11-13, immunoprecipitates by standard reference lupus sera; specific- ity of the sera is indicated at the top of the figure. Lane 14, in uitro transcribed RNA loaded directly on a gel. Immunoprecipitated RNA from whole 50-pl reaction mixture is loaded on a gel in lunes 1-13. One-fourth of this amount is loaded in lune 14. Patients' initials are indicated at the top of the figure. The position of yG-globin run-off transcript and Alu RNA is indicated. The Alu RNA doublet is due to the partial read-through of the RNA polymerase I11 past the termi- nation site (30). N , normal human serum; T, in vitro transcribed RNA.

antibodies. This was also supported by the fact that the reference anti-La, anti-Sm, and anti-RNP sera were anti-Alu negative (Fig. 2, lanes 11-13).

The finding that some anti-La sera did not precipitate Alu RNA was surprising, since La protein has been reported to be associated with RNA polymerase I11 transcripts, including Alu RNA (3, 8, 9, 10-12, 22). Moreover, it was previously found that one anti-La serum (from patient Le, a generous gift from Dr. J. Steitz, see also Ref. 22) precipitated Alu RNA in our assay? These results indicate that sera from certain lupus patients contain an anti-Alu antibody that is distinct from the major classes of lupus autoantibodies. Furthermore, the precipitation of Alu RNA by the La serum from patient Le (22) may have been due to the presence of this anti-Alu antibody rather than the anti-La component.

The amount of RNA precipitated with each patient anti- Alu serum was variable. In our experimental conditions the most active sera precipitated about 25% of the transcribed Alu RNA. As shown in Fig. 2, the intensity of the bands in lanes 2 and 10 is comparable to that in lane 14. In lane 14, however, only one-fourth of the amount of the original reac- tion was loaded on the gel. Other sera precipitated signifi- cantly lower amounts of Alu RNA (lunes 3, 4, and 7; lane 7 appears to be negative but a weak Alu band could be clearly seen on the original autoradiograph).

Some of the sera precipitated a labeled band with mobility slower than that of the yG-globin run-off transcript suggest- ing a reactivity with high molecular weight nucleic acid (lanes 3 and 9). The origin of this band is currently under investi- gation. It is possible that ribosomal RNA which was present in the whole cell extract was labeled during in vitro transcrip- tion and precipitated with anti-ribosomal RNA antibodies (27).

Immunoprecipitation of in Vitro Transcribed Adenovirus VA I RNA-We have tested the reactivity of our sera with the in vitro synthesized adenovirus VA I RNA that was previously shown to be precipitated by anti-La sera (11, 13) to determine if Alu and VA RNPs share common antigenic properties.

We have tested four sera in this assay. Three of these sera contained anti-La antibodies and the fourth serum (AL) had high levels of anti-Alu antibody (Fig. 2, lane 2 ) . Anti-La sera precipitated in vitro transcribed VA I RNA (Fig. 3, lanes 2,3, and 5 ) , in agreement with the observation of Francoeur and Mathews (13)) but did not react with the A36 Alu RNA (Fig. 3, lanes 7,8, and 10). In contrast, serum AL, which contained anti-RNP activity (see Figs. 5 and 6), did not precipitate VA I RNA (lane 4 ) but did precipitate Alu RNA (lane 9) . Normal human serum was negative in this assay (lanes 1 and 6). Phenol-extracted intact Alu RNA was not precipitated by AL serum in the presence or absence of the RNase inhibitor RNasin (Promega Biotec), suggesting that a protein antigen is required for recognition by the antibody (results not shown).

The above experiments indicate that Alu RNA binds a protein which is not common to all RNA polymerase I11 transcripts and as can be seen from the following experiment this protein binds to RNAs transcribed from other Alu genes.

Precipitation of Other Alu Transcripts-In addition to Alu and VA transcripts described above we have tested two other Alu RNAs in our in vitro assay. One Alu sequence is located downstream from the human a-globin gene (22, plasmid RBa- l), and the other plasmid (U6-13) carries a truncated copy of an Alu gene from an undetermined region of the human genome (23).' The results of the immunoprecipitations are

R. Kole and S. Weissman, unpublished data.

Immunoprecipitation of in Vitro Transcribed Alu RNA 11783

N RM SS AL EC N

1 2 3 4 5 6

RM SS AL EC

Alu

VA

7 8 9 1 0

FIG. 3. Immunoprecipitation of VA I and Alu RNA by anti- La and anti-Alu sera. Immunoprecipitation and RNA analysis as in Fig. 2. Lunes 1-5, immunoprecipitates of VA I RNA. Lanes 6-10, immunoprecipitates of Alu RNA.

RBcx-1 A 36 U 6 -13

RM AL T RM AL T RM AL T

1 2 3 4 5 6 7 8 9 FIG. 4. Immunoprecipitation of A h RNA transcribed from

different DNA templates. Immunoprecipitation and RNA analysis as in Fig. 2. Lunes I, 4, and 7, immunoprecipitation by serum RM, an anti-La-containing serum. Lanes 2,5, and 8, immunoprecipitation by serum AL, an anti-Alu serum. Lanes 3, 6, and 9, RNA transcribed

presented in Fig. 4. All three Alu transcripts were precipitated by serum AL, an anti-Alu serum (lanes 2, 5, and 8). As has been shown in Fig. 3, lane 4 , serum AL did not precipitate the VA RNA. In addition two other anti-Alu sera were negative versus VA I RNA (not shown). Since three Alu transcripts were precipitated by anti-Alu sera and a non-Alu transcript, VA I, was not precipitated by anti-Alu sera these results suggest that the antigen recognized by these sera may be Alu specific. As expected from previous experiments (Fig. 3, lane 7) the Alu transcript derived from plasmid A36 was not precipitated by serum RM, an anti-La containing serum (lane 4 ) . However, weak bands of the Alu RNAs derived from RBa- 1 and U6-13 could be detected in immunoprecipitations with this serum (lanes 1 and 7), suggesting a possibility that these transcripts bind, albeit weakly, the La antigen in addition to the Alu antigen. It is likely that the differences in the se- quences around the transcription termination sites are re- sponsible for this difference in binding of the La antigen by Alu transcripts (see “Discussion”).

The truncated Alu transcript derived from the plasmid U6- 13 contains 139 base pairs from the 5‘-end of the Alu sequence and includes the polymerase I11 promoter. When transcribed in vitro, this plasmid gave rise to a transcript about 200 nucleotides long. The size of the transcript as well as Southern blot experiments (not shown) indicated that the transcript initiates at the 5’-end of the Alu gene and terminates within the vector sequences, about 60 nucleotides downstream from the 3’ end of the Alu fragment. The fact that the truncated Alu was precipitated by the anti-Alu serum AL suggests that the Alu antigen binds to the 5’ half of the Alu transcript. However, the possibility that this antigen binds to the down- stream vector sequences cannot be excluded.

Precipitation of RNA and Proteins in the in Vivo Labeled Extracts-We have used [35S]methionine and [3H]uridine la- beled extracts (25) from HeLa cells to study the immunopre- cipitation of the in vivo labeled ribonucleoprotein particles and have tested five anti-Alu sera in these assays. Fig. 5 shows immunoprecipitates of a 35S-labeled extract. All five sera precipitated a common potein band with an apparent molec- ular mass of about 68 kDa (lanes 5-8, and 10, band indicated by dots). This band was not observed in the immunoprecipi- tates obtained with Alu negative sera (lanes 2-4, 9, 11, and 12). Immunoprecipitation of [3H]uridine-labeled extracts by the same five anti-Alu sera is shown in Fig. 6. We did not detect any common RNA band precipitated by anti-Alu sera; hence the role and association of Alu antigen in vivo remains to be established.

Analysis of the gels in Figs. 5 and 6 indicates that the anti- Alu sera, in addition to anti-Alu specificity, exhibit anti-RNP (Figs. 5 and 6, lanes 5 and 8) , anti-Ro (lanes 6, 7, and 10 in Fig. 5 and lanes 6, 7, and 9 in Fig. 6) and anti-U2/Ul specificities (lane 7 in both figures). Control Alu-negative sera exhibit specificities versus Sm (Fig. 5, lanes 9 and 11 and Fig. 6, lanes 10 and 12), RNP (Fig. 6, lane 13), La and Ro (Fig. 5, lanes 3 ,9 , and 11 and Fig. 6, lanes 3 and l o ) , To (Figs. 5 and 6, lane 4 ) , and U2-U1 (Fig. 5, lane 12 and Fig. 6, lane 1 1 ) antigens. These results support the conclusions based on the in vitro experiments that anti-Alu specificity is not associated with and is distinct from any known lupus antibody.

We were unable to detect the 68-kDa band on immunoblots (28) using either nuclear (19) or whole cell extracts (18) as a source of the antigen. In the same experiments Sm and La proteins could be easily detected (results not shown). Since the 68-kDa band was immunoprecipitated from 35S-labeled extracts in much smaller quantities than La and Sm proteins

from plasmids indicated at the top of figure, loaded directly on a gel. (compare bands below 50 kDa in Fig. 5 with the 68-kDa band)

11784 Immunoprecipitation of in Vitro Transcribed Alu RNA

97 - 68 *

43 -

S N La To AL EJ RMc CF RM JM RM U2/!

7s

u2 5.8s

U I

5s

26 -

18 8

I 2 3 4 5 6 7 8 9 1 0 1 1 12

FIG. 5. Immunoprecipitation of ["S]methionine in vivo la- beled HeLa cell extracts. Immunoprecipitated proteins were ana- lyzed by electrophoresis on a 10% sodium dodecyl sulfate-polyacryl- amide gel (26). Lane I, molecular mass markers. Molecular mass is indicated in kilodaltons. Lanes 2-4,9,1 I and 12, immunoprecipitation by Alu-negative sera. Specificities of the control sera are indicated at top of the figure. N , serum from a healthy individual. Lanes 5-8 and 10, immunoprecipitation by anti-Alu sera. A putative Alu antigen common to all anti-Alu sera is indicated by dots. Patient initials are at the top.

4s

S N La To AL EJ RMcCF JM RMU211 Sm RNP

. 6

9' M c

4

I 2 3 4 5 6 7 8 9 1 0 1 1 1 ' 2 1 3

FIG. 6. Immunoprecipitation of ['Hluridine-labeled HeLa cell extracts. Immunoprecipitation and RNA analysis was as in Fig. 2. Lane 1, RNA isolated directly from the extracts and used as size markers. Lanes 2-4 and 10-13, precipitation by Alu-negative control sera. Sera specificity is indicated at the top. Lanes 5-9, precipitation by anti-Alu sera. Patients' initials are indicated at the top of the lanes.

TABLE I Prevalence of antidlu in sera containing other antinuclear antigens

Known extractable nuclear antigen specificities were determined bv immunodiffusion and comDarison to standard reference antisera.

it is possible that the 68-kDa protein band was lost in a somewhat higher background of immunoblots. Alternatively it is likely that the antigen was not recognized by the antibody after denaturation and separation on SDS-containing gels and subsequent blotting.

Screening of the Sera for antidlu Specificity-Using the in vitro transcription of the A36 plasmid in our assay we have screened 114 sera from selected patients with other antinu- clear antibodies for the ability to precipitate A h RNA. The results of the screening, summarized in Table I, again support the observation that anti-Ah antibody does not correspond to any known antibody to established nuclear antigens. There was no correlation of anti-Ah with the presence of any of the other specificities examined.

As seen in Table I1 patients with the anti-Ah antibody specificity suffered from scleroderma or from systemic lupus (two patients had discoid lupus and two mixed connective tissue disease). No obvious clinical association with anti-Alu antibody was observed. The antibody was not a marker for scleroderma, as it was present in only 6 of 19 scleroderma sera. Similarly, the antibody was present in a minority of systemic lupus erythematosus sera.

DISCUSSION

In results presented here 12% of the sera from selected patients with autoimmune diseases contained antibodies

Specificity Number Number of Alu of sera positive

Anti-Sm 20 3 Anti-RNP 28 6 Anti-La 20 1 Anti-Ro 23 3 All specificities 114 14

which precipitated RNA transcribed in vitro from the A h family gene located upstream from the human yc-globin gene. Two other A h transcripts were also precipitated by anti-Ah sera. Because the Alu transcripts are not uniformly precipi- tated by sera known to contain any of the major classes of lupus autoantibodies, i.e. anti-RNP, -Sm, -La, and anti-Ro, anti-Alu most likely represents a distinct and novel specific- ity. In addition, anti-Alu sera precipitate a 68-kDa protein band which is not precipitated by the A h negative sera further suggesting that anti-Ah is a novel specificity recognizing a previously unidentified lupus antigen.

Alu RNA is transcribed and terminated by RNA polymerase I11 in uitro, and it might be expected that this RNA should bind La antigen and be precipitable by anti-La antibodies as are other polymerase I11 transcripts. However, it should be noted that the sequence around the termination site of the A h gene in plasmid A36 is 5' CTTATTATT (30); hence the transcript derived from this gene will have a UU dimer at the 3'-end. Since it has been shown that La antigen binds to the

Immunoprecipitation of in Vitro Transcribed Alu RNA 11785

TABLE I1 Clinical characteristics ofpatients with anti-Alu

Known nuclear antigen specificities were determined as in Table I. Abbreviations: Ab, antibody; SCL, sclero- derma; SLE, systemic lupus erythematosus; DLE, discoid lupus; MCTD, mixed connective tissue disease; N, nucleolar antinuclear antigen (ANA); S, speckled ANA; H, homogeneous ANA; X, unknown precipitin line. RNP, Ro, DNA, La, and Sm denote appropriate antibody specificity.

Patient initials Age Sex Diagnosis Ab

Other Comments

JHa SCL VG 38 F SCL N; RNP Mild myositis EJ 43 F SCL S; Ro Calcinosis; pulmonary fibrosis KMcC 50 F SCL H Myositis; pulmonary fibrosis AL 40 F SCL RNP; X CF 20 F SLE RNP CNS lupus JC 40 M SLE DNA ELC SWB 63 F MCTD RNP GP 24 F SLE H J P 35 F SLE Sm; RNP MB 46 F SLE Sm Autoimmune hemolytic anemia RJMcC 45 M MCTD Ro; X Myositis; pituitary adenoma JLM 60 F DLE Ro

Proteinuria; serositis 37 F DLE Ro; La

Hemolytic anemia

oligo U sequence at the 3'-end of the polymerase I11 tran- scripts (16) and that the binding to longer U oligomers is much stronger than binding to dimers (31) it is likely that the Alu RNA sequence was not recognized efficiently by the La antigen. In contrast, VA I RNA terminates at a sequence 5' CUUUU (32) and, as expected, was efficiently precipitated by anti-La sera. The fact that an anti-La serum weakly precipi- tated Alu transcripts derived from plasmids RBa-1 and U6- 13 seems to suggest that these RNAs bind La antigen in addition to the Alu antigen. Since the termination sites of these transcripts are not well mapped it is difficult to correlate this result with the sequences that might be responsible for the weak binding of the La antigen.

The fact that La antigen did not bind to A36 Alu RNA seems to indicate that binding of La protein to RNA is not required for correct transcription and termination of at least certain polymerase I11 transcripts. However, the possibility that the La protein may act as a transcription or termination factor by interacting with DNA or other components of the transcription complex cannot be excluded.

It has been recently suggested that Alu sequences are proc- essed pseudogenes for 7 S RNA (29). The 7 S RNA complexed with 6 polypeptides forms the signal recognition particle (SRP), which participates in the translocation of secretory proteins across membranes (33). One of the SRP polypeptides has a molecular mass of 68 kDa which is similar to that of the common band visible in 35S-labeled anti-Alu immunopre- cipitates. However, it has been shown that the 68-kDa SRP protein binds to sequences of the 7 S RNA which lack ho- mology to Alu sequences (34), making the SRP protein an unlikely candidate for the Alu antigen. In addition, anti-Alu sera did not precipitate 7 S RNA from the in vivo labeled extracts prepared from HeLa cells (see Fig. 6) or baby hamster kidney cells (not shown).

Another possibility is that Alu antigen is an RNA polym- erase 111 transcription factor, analogous to TFIIIA (35, 36). Similar to the AIu antigen, which binds to Alu RNA and does not bind VA RNA, TFIIIA binds to the promoter region in 5 S RNA but does not bind to VA RNA (35-37). In our prelim- inary experiments we were unable to inhibit transcription of Alu RNA by pretreating the extracts with anti-Alu sera, suggesting that the Alu antigen is not a transcription factor. However, due to the volume limitations and the presence of RNAse in the sera we were not able to ascertain that all of

the antigen had been removed leaving open the possibility that Alu antigen does play a part in transcription. More experiments will be required to establish the identity and the in vivo role of the Alu antigen.

Two recent reports (38,39) indicate that a high percentage of lupus sera contain antibodies against heterogeneous nuclear RNP. Transcription of the plasmid A36 (21) which has been used in this study allows for simultaneous transcription of Alu RNA and yc-globin RNA. Globin RNA forms a hetero- geneous nuclear RNP-like particle during transcription in vitro (40),' so it could be expected that this complex will be precipitated by some sera. However, none of the 114 sera tested precipitated yc-globin RNA in our assay. A few sera (Fig. 1, lunes 3 and 9) precipitated a labeled high molecular weight band, but the origin and identity of this band is not known.

The significance of the clinical findings regarding the pa- tients which exhibit anti-Alu antibody is unclear. All of the patients with this reactivity had either systemic lupus or scleroderma. There was no apparent common manifestation of disease nor was there any evidence of association of anti- Alu with antibodies to other known nuclear antigens. As the sera screened for the presence of anti-Alu were from a bank of sera known to have antinuclear antigens, no estimate can be made of the frequency of this specificity in the general population.

Acknowledgments-We thank Kathleen Kaiser-Rogers and Sylvia Craven for excellent technicaI assistance and Dr. Paul Furdon for helpful discussions and critical reading of the manuscript.

REFERENCES 1. Tan, E. M. (1982) Adv. Immunol. 33, 167-240 2. Lerner, M. R., and Steitz, J. A. (1979) Proc. Natl. Acad. Sci.

U. S. A. 76,5495-5499 3. Steitz, J. A., Wolin, S. L., Rinke, d., Pettersson, I., Mount, S. M.,

Lerner, E. A., Hinterberger, M., and Gottlieb, E. (1983) Cold Spring Harbor Symp. Quant. Biol. 47,893-900

4. Fisher, D. E., Reeves, W. G., Conner, G. E., Blobel, G., and Kunkel, H. G . (1984) Proc. Natl. Acad. Sci. U. S. A . 81, 3185- 3189

5. Padgett, R. A., Mount, S. M., Steitz, 3. A., and Sharp, P. A. (1983) Cell 35,101-107

6. Fradin, A., dove, R., Hemenway, C., Keiser, H. D., Manley, J. L., and Prives, C. (1984) Cell, 37,927-936

7. Yang, V. W., Lerner, M. R., Steitz, J. A., and Flint, S. J . (1981) Proc. Natl. Acad. Sci. U. S. A. 78. 1371-1375

11786 Immunoprecipitation of in Vitro Transcribed Alu RNA 8. Hendrick, J. P., Wolin, S. L., Rinke, J., Lerner, M. R., and Steitz, waran, P., Duncan, C. H., and Weissman, S. M. (1983) Cold

9. Rinke, J., and Steitz, J. A. (1982) Cell 29, 149-159 24. Donis-Keller, H. (1979) Nucleic Acids Res. 7, 179-192 J. A. (1981) Mol. Cell. Biol. 1, 1138-1149 Spring Harbor Symp. Quant. Biol. 47, 1079-1081

10. Chambers, J. C., Kurilla, M. G., and Keene, J. D. (1983) J. Biol. 25. Chambers, J. C., and Keene, J. D. (1985) Proc. Natl. Acad. Sci. Chern. 258, 11438-11441 U. S. A . 82,2115-2119

11. Lerner, M. R., Boyle, J. A., Hardin, J. A., and Steitz, J. A. (1981) 26. Laemmli, U. K. (1970) Nature 227, 680-685 Science 2 1 1,400-402 27. Hardin, J. A., Rahn, D. R., Shen, C., Lerner, M. R., Wolin, S. L.,

12. Lerner, M. R., Andrews, N. C., Miller, G., and Steitz, J. A. (1981) Rosa, M. D., and Steitz, J. A. (1982) J. Clin. Inuest. 70, 141- Proc. Natl. Acad. Sci. U. S. A . 78,805-809 147

13. Francoeur, A. M., and Mathews, M. B. (1982) Proc. Natl. Acad. 28. Burnett, W. N. (1981) Anal. Biochern. 112, 195-203 Sci. U. S. A . 79,6772-6776 29. Ullu, E., and Tschudi, C. (1984) Nature 312, 171-172

14. Kurilla, M. G., and Keene, J. D. (1983) Cell 34,837-845 30. Duncan, C. H., Jagadeeswaran, P., Wang, R. C., and Weissman, 15. Wilusz, J., Kurilla, M. G., and Keene, J. D. (1983) Proc. Natl. S. M. (1981) Gene 13,185-196

16. Mathews, M. B., and Francoeur, M. A. (1984) Mol. Cell. Biol. 4, 32. Akusjarvi, G., Mathews, M. B., Anderson, P., Vennstrom, B., and

17. Gottesfeld, J. M., Andrews, D. L., and Hoch, S. 0. (1984) Nucleic 2428 Acids Res. 12,3185-3200 33. Walter, P., Gilmore, R., and Blobel, G. (1984) Cell 38, 5-8

18. Manley, J. L., Fire, A., Cano, A., Sharp, P. A., and Gefter, M. L. 34. Gundelfinger, E. D., Krause, E., Me& M., and Dobberstein, B. (1980) Proc. Natl. Acad. Sci. U. S. A . 77, 3855-3859 (1983) Nucleic Acids Res. 11,7363-7374

19. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G . (1983) Nucleic 35. Segall, J., Matsui, T., and Roeder, R. G . (1980) J. Biol. Chem. Acids Res. 11, 1475-1489 255, 11986-11991

20. Fahey, J. L., and Terry, E. W. (1978) in Handbook ofEperimenta1 36. Engelke, D. R., Ng, S.-Y., Shastry, B. S., and Roeder, R. G. (1980) Immunology, (Weir, D. M., ed) Vol. 1, Chapter 8, pp. 8.1-8.16, Cell 19, 717-728 Blackwell, London 37. Pelham, H. R. B., and Brown, D. D. (1980) Proc. Natl. Acad. Sci.

21. Duncan, C., Biro, P. A., Choudary, P. V., Elder, J. T., Wang, R. U. S. A . 77,4170-4174 R. C., Forget, B. G., deRiel, J. K., and Weissman, S. M. (1979) 38. Fritzler, M. J., Ali, R., and Tan, E. M. (1984) J . Irnrnunol. 132, Proc. Natl. Acad. Sci. U. S. A . 76,5095-5099 1216-1222

22. Shen, C. K. J., and Maniatis, T. (1982) J . Mol. Appl. Genet. 1, 39. Zouali, M., and Eyquem, A. (1984) J . Clin. Irnmunol. 4,209-219

23. Fukumaki, Y., Collins, F., Kole, R., Stoeckert, C. J., Jagadees- U. S. A. 80, 4296-4300

Acad. Sci. U. S. A. 80,5827-5831 31. Stefano, J. E. (1984) Cell 36, 145-154

1134-1140 Pettersson, U. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,2424-

343-360 40. Economidis, I. V., and Pederson, T. (1983) Proc. Natl. Acad. Sci.