Isolation and Characterization of Nuclear Ribonucleoprotein Complexes Using Human Anti-nuclear Ribonucleoprotein Antibodies*

We have investigated the feasibility of using human autoimmune antibodies to isolate and characterize spe- cific nuclear ribonucleoprotein (nRNP) complexes. High titers of anti-nRNP antibodies occur in a syn- drome called mixed connective tissue disease. IgG an- tibodies from two mixed connective tissue disease patients were used to construct affinity columns to isolate the antigenic complexes from rat liver nuclei. A maxi- mum of 2.9% of the nuclear RNA and 0.7% of the protein bound to anti-nRNP but not to control IgG columns. A fraction of the bound antigen, comprising less than 0.15% of the total nuclear protein, was isolated in anti- genitally active form. The protein moiety of this fraction consisted of two quantitatively major polypeptides of molecular weights

We have investigated the feasibility of using human autoimmune antibodies to isolate and characterize specific nuclear ribonucleoprotein (nRNP) complexes. High titers of anti-nRNP antibodies occur in a syndrome called mixed connective tissue disease. IgG antibodies from two mixed connective tissue disease patients were used to construct affinity columns to isolate the antigenic complexes from rat liver nuclei. A maximum of 2.9% of the nuclear RNA and 0.7% of the protein bound to anti-nRNP but not to control IgG columns. A fraction of the bound antigen, comprising less than 0.15% of the total nuclear protein, was isolated in antigenitally active form. The protein moiety of this fraction consisted of two quantitatively major polypeptides of molecular weights 30,000 (P30) and 13,000 (P13). Antigens isolated from both anti-nRNP columns possessed essentially these same two polypeptides.
lmmunological tests of crude and purified antigens against anti-nRNP sera from a total of four patients provided additional evidence that antibodies from different individuals are directed against the same nRNP antigen. There were no polypeptides in the isolated antigen which corresponded in molecular weight to the core proteins of heterogeneous nuclear ribonucleopro-Bin (hnRNP) particles described by other investigators.
The binding of both RNA and protein to anti-nRNP columns was greatly reduced by treating the crude antigen with pancreatic RNase A before chromatography. In particular, the binding of P13 was reduced to a fraction of the pre-RNase control. None of the four sera in this study contained antibodies to DNA, histones, RNA, DNA oh&tone complexes, or nonhistone chromosomal proteins.

Antibodies
to specific cellular macromolecules can be raised in experimental animals. Such antibodies have been used to facilitate biochemical isolation of subcellular components (Duerre, 1967;Allen and Terrence, 1968;Holme et al., 1971) and to study their in situ localization in fined cytological preparations (Lazarides and Weber, 1974;Silver and Elgin, 1976;Alfageme et al., 1976;Jamrich et al., 1977) certain human diseases in which antibodies directed against nuclear macromolecules appear spontaneously in the blood (Tan and Kunkel, 1966;Stellar, 1971;Tan and Lemer, 1972). These so-called systemic rheumatic diseases include systemic lupus erythematosus and mixed connective tissue disease. Antibodies to a variety of nuclear macromolecules including DNA (Tan and Natali, 1970), histones (Stellar, 1971), RNA. protein complexes (Notman et al., 1975;Kurata and Tan, 1976), and nonhistone proteins (Tan and Kunkel, 1966) can be found in autoimmune sera. These sera are potential sources of immunological tools which can aid in understanding the organization and function of the nucleus. However, the autoimmune antibodies have not been widely used for this purpose, primarily because immunological data which indicate that the sera may be good sources of specific antibodies have not yet been corroborated by biochemical studies. The possibility, therefore, has existed that individual sera may contain mixtures of antinuclear antibodies which are too complex to be useful for the study of one particular antigen. The subject of this investigation is the MCTD' syndrome which gives rise to high titers of antibodies to nRNP complexes (Northway and Sharp et al., 1972;Kurata and Tan, 1976;Alarcon-Segovia et al., 1978). The antigenic nRNP complexes have not previously been isolated or characterized chemically. A few of their properties are known from immunological studies. They are detectable by indirect immunofluorescence in the interphase nuclei of a variety of cell types including calf thymus, rabbit thymus, mouse kidney, and a human lymphocyte cell line W& (Mattioli and Reichlin, 1971;Tan and Lemer, 1972;Notman et al., 1975;Kurata and Tan, 1976). The fluorescence patterns show that the antigenic complexes are not in the nucleoli, nor do the human anti-nRNP antibodies react with isolated nucleolar preparations in immunodiffusion tests . The antigenic complexes have not been sized accurately. However, the crude antigen migrates faster than 7 S markers in sucrose density gradients (Northway and Tan, 1972) and is thus larger than tRNA precursors which have been isolated from eukaryotic cells (Burdon and Clason, 1969;Mowshowitz, 1970;Blatt and Feldman, 1973;Weinmann and Roeder, 1974;Siddiqui and Chen, 1975 This problem  was approached  by constructing  affkity  columns for isolating the antigen&z ribonucleoprotein  which could  then be subjected to quantitative  and chemical analyses. The  second objective was to test available anti-nRNP  sera for the   presence of antibodies to other nuclear macromolecules   which  might interfere with isolation and characterization of nRNP complexes. Immunological tests conducted by medical investigators who provided the sera, including counterimmunoelectrophoresis, radioimmunoassays, and indirect immunofluorescence, suggested that a high proportion of the anti-nRNP sera are free of antibodies to DNA, histones, and non-RNA-associated no&stone proteins (see "Experimental Procedures" and Northway and Notman et al., 1975;Kurata and Tan, 1976). We screened six anti-nRNP sera of which four showed a single precipitin reaction against sonicated rat liver nuclei, a reaction which was sensitive to RNase A and trypsin. We used RNase A sensitivity and other procedures to test the four anti-nRNP sera for the presence of antibodies to other nuclear macromolecules.
In addition to these four sera, a number of normal human sera were used as immunoglobulin sources in this investigation.
Unlike antibodies to defined antigens which are raised in experimental animals, the human anti-nRNP antibodies arise spontaneously under the influence of disease mechanisms which are poorly understood.
Therefore, a third objective was to determine whether different individuals affected with MCTD give rise to antibodies with the same antigen specificity. The availability of sera from four different individuals made this possible. Anti-nRNP antibodies are immunoglobulins of the IgG type (Tan and Vaughn, 1973). IgG was isolated from two of the sera and used to construct affinity columns for isolating and biochemically comparing purified antigens.  (1951) and Sedmak and Grossberg (1977) using bovine serum albumin as a standard. Nucleic acids were quantitated by a modification of the Schmidt-Tannhauser procedure (Ts'O and Sato, 1959) and by the orcinol method for RNA (Dische and Schwarz, 1937), using yeast RNA (Sigma, Type III) as a standard.
DNA, RNA, Histones, Nucleosomes, and Nonhistone Proteins-Rat liver chromatin was isolated by the method of Bonner et al. (1968) as a source of histones and nonhistones. The chromatin was extracted with 4.0 M NaCl, 5 mM 2-mercaptoethanol, 10 mM Tris-HCl, pH 8.0, to dissociate chromosomal proteins (Sevall et al., 1975). ZgG Isolation and Affinity Column Procedures-IgG was isolated from two anti-nRNP sera (MP and MK) using 2 ml of each serum for a yield of 20 mg of IgG. A pool of normal human serum was used for isolation of control IgG. The IgG was purified essentially as described by Weir (1967 King and Laernmli (1971) aa modified by Weintraub et al. (1975) for 15% acrylarnide, in gel tubes (8 mm x 13 cm). At an altitude of 5280 feet, it was necessary to reduce the amount of ammonium persulfate by 50% to prevent air bubble formation. The gels were allowed to polymerize overnight. Samples were applied and then stacked at 1 mA/gel and run at 2 mA/gel for 4 h. Gels were fixed and stained with Coomasaie brilliant blue as described elsewhere (Garrard and Bonner, 1974). The gels were scanned at a wavelength of 600 mn using a Varian 634 spectrophotometer equipped with scanning attachment.

RESULTS
Fractionation of Rat Liver Nuclei and Testing of Anti-nRNP Sera-A fractionation procedure was devised with the objective of isolating a nuclear subfraction enriched in the antigenic nRNP complexes. The procedure which proved to be the most suitable for this purpose is outlined in Fig. 1. Of the four resulting fractions, A3 contained most of the nuclear protein (57%) and more than 96% of the DNA and histones. Fraction A2', comprising 10% of the nuclear protein, was an insoluble residue and was not tested further. The greatest enrichment in RNA relative to protein was seen in A2 which contained 38% of the nuclear RNA, 1.5% of the DNA, and 11% of the protein.
The mass ratio of RNA to protein in this fraction was 0.5:1. Al contained 12.3% of the nuclear protein.

Fractions
Al, A2, and A3 were tested for the presence of antigenic complexes by double diffusion against anti-nRNP sera.
Anti-nRNP sera, obtained from a serum bank, had been prescreened by clinical laboratory tests for the presence of anti-RNP antibodies as described under "Experimental Procedures." Four of these sera from the patients MP, MK, MM, and BN were used as antibody sources in this investigation. In addition, IgG was isolated from the two sera with the highest titers of anti-RNP antibodies (MP and MK). The four sera and two IgGs were tested against the nuclear fractions Al, A2, and A3 as shown in Fig. 2. Optimal antibody concentrations were determined by testing several dilutions of each serum and IgG against each antigen. Of the three nuclear fractions shown in Panels A, B, and C, A2 clearly contained the highest antigen concentration.
A trace of activity was observed in fraction Al. As Al was the least concentrated, A2 was diluted to the concentration of Al and retested. Panel D shows that diluted A2 was more antigenic than Al at the same protein concentration.
When A2 was tested against six normal human sera, no reaction was observed (Panel E). Panel F shows the reactions of MP serum against whole nuclear sonicate, the low speed supernatant from which A2 is derived, and A2. The lines of identity observed with all three nuclear fractions indicate that all of the antigenic determinants present in unfractionated nuclei are also present in A2. The data show only one precipitin reaction between anti-nRNP antibodies and nuclear fractions.
Previous studies demonstrated that the immunoprecipitating activity of the nRNP antigen is destroyed by treatment with either ribonucleases or proteases (Mattioli and Reichlin, 1971;Northway and Tan, 1972). We re-examined the question of RNase A sensitivity and extended it to interactions between antigen and antibodies affiied to Sepharose 4B. The effects of RNase A and trypsin on the immunoprecipitating activity of A2 are shown in Fig. 3 Immunological reactions of nuclear fractions against anti-nRNP and normal sera. Nuclear fractions: all fractions were in NaCl/ P,, pH 7.5. The center wells of Panels A through E, respectively, contained A,, 2.0 mg/ml of protein; AP, 7.5 mg/ml; G, 7.7 mg/ml; Aa, 2.0 mg/ml, Al, 7.5 mg/ml. In Panel F, WNS = whole nuclear sonicate, 19 mg/ml of protein; LSS = low speed supernatant, 6.0 mg/ml; and AZ, 7.5 mg/ml. Anti-nRNP sera: wells 1 to 6 of Panels A to D contained the following antibodies, diluted, if indicated, in NaCl/P,, pH 7.5: 1, MP serum, diluted l/20; 2, MK serum, diluted % 3, MM serum, undiluted; 4, BN serum, undiluted; 5, MP IgG, 1 mg/ml; 6, A 6 C FIG. 3. Effects of RNase A and trypsin on the antigenic activity of A2. Enzyme digestions: bovine pancreatic RNaae A (Worthington RAF) at a concentration of 2 mg/ml was boiled for 1 h and then dialyzed into 10 mM Tris-HCl, 5 mu 2-mercaptoethanol, pH 7.5. 50 gl of RNase were added to 300 gl of A2 (7.1 mg/ml of protein) in the Same buffer. Twenty-five microliters of a 3 mg/ml solution of trypsin (Worthington) were added to 300 ~1 of A2 in the Trie buffer. The enzyme-containing aliquota of A2 and a control aliquot were incubated at 37°C for 1 h and then transferred to 0°C to stop the reactions. Reaction conditions: wells 1 to 6, Panels A to C, contained the anti-nRNP sera MP, MK, MM, BN, and MP and MK IgG, respectively, at concentrations given in Fig. 2. The center wells of Panels A, B, and C contained control-incubated A2:7.1 mg/ml of protein (Asc), RNase A-treated A2 (A& and trypsin-treated A2 (A& respectively. The plates were incubated and processed as described in the legend to Fig. 2. IgG. Panels B and C show the reactions of RNase A-treated and trypsin-treated AZ, respectively. A complete loss of immunoprecipitating activity was observed upon treatment of A2 with either hydrolytic enzyme. The presence of a single precipitin reaction between sera and either whole nuclear sonicate or A2 and the RNase F MK IgG, 2 mg/ml. These concentrations were found to produce optimal reactions against the nuclear fraction A2 at a concentration of 7.5 mg/ml. Panel F, center well, contains MP serum at a dilution of Ym. Normal sera: the undiluted aera of laboratory personnel BB, LM, MW, MS, PR, and CP were tested against A2 (7.5 mg/ml in NaC1/Pi, pH 7.5) in Panel E. Reaction conditions: agarose plates (0.4%) were used at room temperature. The wella held 150 ~1 of solution. Plates were allowed to develop 18 to 48 h and were then washed with 5% sodium citrate followed by distilled H20. sensitivity of the A2 reaction suggested that precipitating antibodies to other nuclear molecules (e.g. histones and DNA) were not present in the four anti-RNP sera. This was confirmed directly by double diffusion assays of anti-RNP antibodies against DNA, rat liver histones, yeast RNA, rat liver nonhistone chromosomal proteins, and nucleosomes (v bodies). The reactions against all anti-nRNP antibody sources were negative and are not shown. (See "Experimental Procedures" for sources and preparation of these antigens.) Normal human sera were also tested and were similarly negative.
Isolation of the Antigen by Affinity Column Chromatography-Three affinity columns were constructed by linking anti-nRNP IgG from two MCTD patients and IgG from a pool of normal human serum (NHS) to CNBr-activated Sepharose 4B. Chromatography of A2 over anti-nRNP columns resulted in three fractions: a runoff in phosphate-buffered saline (NaCl/Pi), pH 7.5, and two fractions recovered by step elutions with 1 M NaCl, 0.01 M phosphate buffer, pH 7.5, followed by 10 mu HCl, 0.15 M NaCl, pH 2.2. Affinity column fractions were tested for immunological activity by double diffusion as shown in Fig. 4. The runoff fraction from the MP anti-nRNP column was devoid of immunoprecipitating activity against all sera (Panel A). The runoff from the MK anti-nRNP column was also inactive and is not shown. In contrast, the runoff from the control NHS column was fully active (Panel D). The amount of antigen applied to these columns was calculated to be nonsaturating with respect to anti-nRNP antibodies as described under "Experimental Procedures." The NaCl eluate from the MP anti-nRNP column was inactive, but the HCl eluate was active against all antibody sources (Panels B and C, respectively). Panel E shows that the immunological relationship between the HCl eluate and the crude antigen A2 is one of partial identity. The HCl eluate is impoverished in RNA relative to A2 (10 f 4% by mass compared to 33 + 7%, respectively, based on data from five experiments).
However, the RNA which is present in the HCl eluate appears to have an essential immunological role as treatment of this fraction with RNase A destroyed its immunoprecipitating activity as shown in Panel F. Quantitative data on the affinity column experiments, including the two anti-nRNP and control NHS columns, are shown in Tables I and II. The distributions of protein in runoff, 1 M NaCl, and HCl eluates are shown in Table I. The NHS IgG column bound 0.2% of the applied protein. The anti-nRNP columns MP and MK bound 6.7% and 3.9%, respectively. Most of the retained protein was eluted by 1 M NaCl. The antigenic activity, however, was found in the 10 mu HCI eluate, as shown in Fig. 4  with RNase A in 10 mM Tris, pH 7.5, before chromatography on the MP anti-nRNP column. The RNase treatment caused a major reduction in the quantity of protein which bound, as shown in Table I. This material was eluted with 10 mu HCl, 0.15 M NaCl and was analyzed by gel electrophoresis.
The amounts of RNA bound by control and anti-nRNP columns are shown in Table II. Chromatography of A2 over the MP anti-nRNP column resulted in the binding of 7.7% of the applied RNA. The partitioning of bound RNA into NaCl and HCl eluates is shown in the table. Digestion of A2 in 10 mru Tris completely abolished RNA binding. The effects of RNase A on RNA binding to the column are consistent with its effects on the immunoprecipitating ability of crude antigen, as shown in Fig. 3. The NHS column retained no RNA.
RNase A digestions were performed in 10 mu Tris-HCl, pH 7.5, instead of NaCl/Pi, pH 7.5, because we have found that, in general, A2 is more soluble in Tris than in NaCl/Pi. RNase A digestions performed in NaCl/Pi often do not go to completion, possibly as a result of inhibition by the inorganic phosphate in the buffer. Moreover, there is extensive precipitation in the course of RNase digestions performed in NaCl/Pi.
The absence of immunoprecipitin reactions between MCTD sera and isolated histones or e bodies (above) and the RNase A sensitivity of both crude and purified antigens (A2 and HCl eluate,Figs. 3 and 4,respectively) suggested that antihistone The distribution of RNA in MP anti-nRNP and normal column fractions Recovery of applied A2 RNA was greater than 90% for all experiments. Analyses for DNA and RNA were performed by the methods of Schmidt and Tannhausser as modified by Ts'O and Sato (1959) and of Dische and Schwarz (1937) Panels A to F contain: (A) 100 g of A2 protein, (B) 100 cg of NaCl/Pi runoff protein following chromatography of A2 over the MP column; (C) 60 pg of 1 M NaCl eluate, MP column; (D) 60 pg of IO mu HCl eluate, MP column; (E) 60 jtg of 10 mu HCl eluate, MK column; (F) 60 pg of rat liver histones.

antibodies
are not present in these sera. To further test for antihistone activity, we applied 1 mg of histones in 2.5 ml of NaCl/Pi, pH 7.5 (Case 1) or in 1 M NaCl, 0.01 M PB, pH 7.5 (Case 2), to the MP anti-nRNP column. The column was then eluted with either 1 M NaCl, 0.01 M PB, pH 7.5, followed by 10 mu HCl, 0.15 M NaCl (Case 1) or with the 10 mu HCl solution (Case 2). The absorbance of these eluates at 230,260, and 280 mn was zero in both cases, indicating no histone binding.
Also, there was quantitative recovery of applied protein in the runoffs. Trace amounts of other polypeptides, all quantitatively minor relative to P30, were also present, but neither their quantities nor molecular weights were reproducible.

SDS-Polyacrylamide
The 1 M NaCl eluate from the MP column (Panel C), which may contain salt-labile antigenic complexes, was considerably more heterogeneous than the HCl eluate. This fraction lacked most of the higher molecular weight bands (>30,000) found in A2 and in the runoff and contained a M, = 26,000 polypeptide as its major protein component. The absence of major polypeptides in the &f, =32,000 to 43,000 molecular weight range is evident in the HCl eluate. The molecular weights of the major "informofer" proteins are usually given as 37,000 and 40,000 (Martin et al., 1973;Pederson, 1974) and in one report, 32,000 to 34,000 (Beyer et al., 1977).
As a result of the high density of protein in band P13 of Gels D and E, the intensity of staining with Coomassie blue was not linearly related to protein load. Therefore, the loads >-HI 68,000-30,000-26,000- bound. As A2 represents 11% of the nuclear protein 2nd 38% of the nuclear RNA, it follows from the data in Tables I and  II  on these gels appear to be lower than 60 pg of protein. Another problem we encountered was incomplete absorption of dye into the P13 band due both to the density of protein and to the density of acrylamide which is significantly higher at the bottom of the gels if they are allowed to polymerize overnight. This problem has been partially overcome by using 10% acrylamide instead of 15%.' Fig. 6 shows a scan of a gel containing HCl eluate protein from the MP column (upper scan). From gel scans of five MP anti-nRNP column experiments, we estimate that P30 and P13 comprise more than 75% of the protein mass of HCl eluates. The effect of RNase A on the protein composition of the HCl eluate can be seen by comparing the upper and lower scans of Fig. 6. Shown in the lower scan is the HCl eluate from an experiment in which A2 was pretreated with RNase A before chromatography on the MP column. The quantitative effects of these treatments were presented in Tables I and  II. With the same amount of protein applied to each gel, it is evident that the ratio of P13 to P30 decreased with RNase treatment. An interpretation consistent with the quantitative data is that the binding of P13 is greatly reduced, but the binding of P30 is unaffected by the RNase treatment.
The HCl eluate was applied to a column of Sephadex G-150 in 10 mM HCl, 0.15 M NaCl to determine the approximate sizes of the macromolecules in the antigenically active fraction. At least three major components were present, all within a molecular weight range of 110,000 to 65,000. There was no evidence of P13 monomers, and >85% of the applied protein fell within the molecular weight range of 110,000 to 65,000.

DISCUSSION
One of the aims of this study was to determine how much of the nuclear ribonucleoprotein reacts with human anti-nRNP antibodies.
Antiserum from the patient MP was the most potent of the four compared here both with respect to titer and antigen binding, as shown in Fig. 2 and Tables I and II. A2 was chromatographed over the MP column under conditions of antibody excess so that all of the applied antigen Some assumptions implicit in these calculations are: 1) that A2 contains the majority of the nuclear antigenic complexes; 2) that most of the antigen is not inactivated during nuclear isolation and fractionation; and 3) that adventitious sticking of nonantigenic ribonucleoprotein is not quantitatively significant. The data in Fig. 2 show that most of the antigenic activity is found in A2. These data also show that A2 is immunologically identical with whole nuclear sonicate. Thus, it is unlikely that the fractionation scheme outlined in Fig. 1 results in gross inactivation of the antigen. We found it necessary to sonicate nuclei to release the antigenic activity. It is possible that sonication has a deleterious effect. While the quantitative data may be affected to an unknown degree by these assumptions, they do illustrate that the antigenic complexes represent a small fraction of the nuclear ribonucleoprotein.
Most of the ribonucleoprotein which bound to the MP and MK columns was eluted with 1 M NaCl and had no immunological activity, as shown in Fig. 3, Panel B. It is evident from the reaction in Fig. 4, Panel D, that the HCl eluate, although active, lacks some of the antigenic determinants present in the crude antigen. It is likely that the NaCl eluate consists, in whole or part, of salt-labile antigenic complexes which together with the salt acid-resistant complexes of the HCl eluate make up the whole antigen. When nuclear extracts were reacted against anti-nRNP sera in nondissociating media such as phosphate-buffered saline, a single precipitin reaction was observed, both by counterimmunoelectrophoresis (Kurata and Tan, 1976) and by immunodiffusion, as shown in Fig. 2. Thus, under nondissociating conditions, complexes which are sensitive and resistant to salt and acid may be found on the same RNP particle.
Four antisera were compared in this study in order to determine whether they are specific for the same antigen. The protein moieties of HCl eluates from two anti-nRNP columns (MP and MK) were compared in Fig. 5. The polypeptides P30 and P13 are the quantitatively major protein species in HCl eluates from both columns. With the exception of minor polypeptides, the protein composition of this fraction is very reproducible.
The HCl eluate from the MP column reacted with antibodies from all four sera, and the reactions were immunologically identical, as shown in Fig. 4, Panel C. Reactions of the four sera against the crude antigen A2 were also identical. Moreover, fractions which were inactive against one serum were inactive against all four, e.g. the NaCl/Pi runoff from the MP column, shown in Fig. 4, Panel A. Thus, there were no detectable differences in the specificities of anti-nRNP antibodies from different individuals.
The HCl eluate consists of less than 0.15% of the nuclear protein based on data in Table I. We have no direct evidence that any of the antigenic determinants in this eluate are on polypeptides P30 and P13. However, they comprise most of the protein mass of this fraction and are the only reproducible protein components with the possible exception of a minor polypeptide of M, = 14,000. The binding of P13 to affinity columns is RNase A-sensitive, as are the immunoprecipitin activities of A2 and of the HCl eluate. Characterization of the separate nucleic acid and protein moieties of the HCl eluate is in progress. Isolation of a sufficient quantity for further characterization is now feasible. Chromatography of 24 mg of A2 protein over a 20-mg IgG column yields about 240 pg of protein in the HCl eluate. We have now constructed a lOO-mg column and have chromatographed 75 mg of A2 with a yield of 700 pg of protein and about 70 pg of RNA in the HCI eluate.' It should be possible to chromatograph 100 mg of A2 without saturating the column. Moreover, the column can be used several times; the original MP anti-nRNP column has been used 10 times without noticeable deterioration in quality. We found no evidence for the presence of antibodies to free protein or free nucleic acid in any of the four anti-nRNP sera in this study. Within the limits of sensitivity of chemical assays, there was no DNA in ribonucleoprotein fractions isolated by affinity chromatography.
Nor did any of the sera react with DNA, histones, histone. DNA complexes (v bodies), RNA, or nonhistone chromosomal proteins in immunodiffusion assays. As shown in Fig. 3, the antigenic activity of A2 is sensitive to either RNase A or trypsin, indicating that both RNA and protein moieties are essential for activity. Also, treatment of crude antigen with RNase A greatly reduced the quantity of protein which bound to the anti-nRNP column, as shown in Table I. These data indicate that the antigenic ribonucleoprotein is in the form of complexes rather than separate RNA and protein moieties. The smallest major component of the HCl eluate has a molecular weight of 65,000 based on Sephadex G-150 gel filtration, whereas the majority of the protein has a molecular weight of 13,000 in SDS gels (Fig. 5). Therefore, this antigenically active fraction is clearly composed of molecular aggregates. Due to the heterogeneity of these aggregates, we have not yet determined their nucleic acid to protein ratios. The size of the RNP antigen has been estimated previously by subjecting crude tissue extracts to either gel exclusion chromatography (Mattioli and Reichlin, 1971) or to sucrose density sedimentation (Northway and . These estimates indicate a size of 200,000 daltons or ~7 S, respectively, for the crude antigen. As the HCI eluate is a partial antigen (Fig. 4) representing acid-and salt-resistant antigenic complexes, it is not surprising that it is composed of aggregates smaller than M, = 200,000.
The rates of migration of polypeptides P30 and PI3 on SDS gels are very similar but not identical with those of histones Hl and H2A, respectively. Preliminary data show that HCl eluate proteins do not migrate as rapidly as histones in acetic acid/urea gel systems which resolve basic proteins on the basis of charge and size. Moreover, a number of observations made in this study indicate that antihistone antibodies are not present in the MCTD sera. First, the nuclear fraction A3 which contains 290% of the histones is immunologically inactive (Fig. 2, Panel C). Second, when tested by immunodiffusion under identical conditions, the HCl eluate is active (Fig. 4, Panel C and F), whereas histones and v bodies are inactive. Third, the antigenic activities of A2 and of the HCl eluate are destroyed by RNase A (Figs. 3 and 4). Fourth, isolated histones do not bind to the anti-nRNP column. These observations do not exclude the possibility that the protein moiety of the HCl eluate is composed of histone-like proteins. At least two types of histone-like proteins have been found in mammalian nuclei or nucleoli. The so-called high mobility group nuclear proteins include a polypeptide of M, = 11,000 which has a high content of basic amino acid residues and is partially homologous to histone T (Goodwin et al., 1975;Walker et al., 1976). A24, a protein found in nucleoli, is composed of histone H2A covalently linked to a nonhistone polypeptide (Olsen et al., 1976). Further characterization of the HCl eluate proteins should reveal any similarities between these proteins and histones.
For this study, we intentionally selected human sera which contained no precipitating antibodies to RNase-resistant nuclear antigens. Sera from patients with systemic lupus erythematosus occasionally contain antibodies to both the nRNP antigen and to a nuclear protein (or proteins) called Sm (Mattioli and Reichlin, 1971;Tan8 et al., 1973;Tan and Vaughn, 1973;Kurata and Tan, 1976). Using sera with this dual specificity, Mattioli and Reichlin (1971) reported that the Sm antigen appears to be physically associated with the nRNP antigen. We have tested prototype anti-Sm sera against nuclear Fractions Al and A2, the NaCl/Pi runoff from the anti-nRNP affinity columns, and the NaCl and HCl eluates." With the exception of the NaCl and HCl eluates, all of these fractions reacted with anti-Sm sera in double diffusion tests. The absence of precipitin reactions between anti-Sm sera and the HCl eluate suggests that this antigen contains no physically associated Sm.
The very high prevalence of anti-nRNP antibodies in MCTD sera suggests that malfunctions in specific nuclear mechanisms, e.g. some aspect of nuclear RNA synthesis or metabolism, may be the stimulus for antibody production. To understand which part of the nuclear RNA synthesizing or processing machinery, or both, is reacting with the anti-nRNP antibodies from these sera, it will be necessary to characterize the RNA moiety of the isolated complexes. As we have not isolated the entire nRNP antigen in immunologically active form, it is not possible to draw extensive comparisons between this antigen and other nonnucleolar RNP particles which have been isolated from nuclei. However, some differences between the nRNP antigen and 30 to 55 S hnRNP particles (composed of hnRNA and protein) are evident. First, it is necessary to disrupt nuclei by sonication to dissociate the nRNP antigen. We attempted to extract the antigen from intact nuclei by the method of Martin et al. (1973) and found that the extract, which contained 30 S RNP particles, was immunologically inactive." Second, the 30 to 55 S hnRNP particles are composed of between two and seven distinct polypeptides ranging in molecular weight from 32,000 to 40,000 (see LeStourgeon et al., I978 and Martin et al., 1978 for references to earlier work). The HCl eluate is devoid of quantitatively major polypeptides of M, = > 30,000 (Fig. 5, Panels D and E). Also, the majority of the protein mass in the NaCl eluate corresponds to M, = 5 30,000 (Panel C). These observations do not exclude the possibility that the particles comprising the nRNP antigen represent a different stage of hnRNA processing than do the hnRNP particles. Using the human anti-nRNP antibodies, it should be possible to investigate the distribution of the antigenie particles in nuclei and determine the degree to which they are chromosome-associated.