Anti-DNA Abs are the most frequently encountered self-reactivity, a curious fact that may be explained by their potential for cross-reactivity and multireactivity. A case in point is 3H9, an anti-DNA, anti-chromatin Ab, isolated from a diseased MRL/lpr mouse (1). 3H9 cross-reacts with anionic phospholipids, including phosphatidylserine, an Ag exposed on apoptotic cells (2). 3H9 is also multireactive: it binds DNA as well as purified nucleosomes (3). The binding of 3H9 to DNA can be measured by a variety of techniques, but EMSAs demonstrate most clearly that 3H9 forms a specific complex with DNA and that it greatly prefers poly(dG)·poly(dC) over poly(dA) ·poly(dT) (4).
Multireactivity of anti-DNAs has been the focus of several studies including a recent study in this journal by Guth et al. (5). Multireactivity of 3H9 derives from its asymmetric combining site: the DNA-binding surface is located in the H chain and can be enlarged by adding arginine residues to certain sites in VH complementarity-determining regions and FR3 (1). This leaves the L chain free to recognize additional determinants of a complex Ag. Strikingly, different immunofluorescence patterns are seen when the 3H9 H chain is paired with different L chains, although most pairs bind DNA (1). Therefore, 3H9, like other Abs to DNA, has a complex specificity that, depending on the L chain, can extend to other nucleic acids, such as RNA, and include protein-DNA complexes, such as nucleosomes (1).
Binding to DNA and nucleosomes depends on the 3H9 H chain complementarity-determining region 3 (3). Guth et al. (5) confirm this finding and identify the H2A/H2B/DNA complex as the target of 3H9 binding. However, Guth et al. further report that a “twice-purified” preparation of 3H9, also treated with DNaseI, failed to bind free DNA. This is surprising because the authors note that DNA is necessary for binding to the complex. We propose that, during extended manipulations in nonphysiological buffers, portions of the 3H9 combining site may become susceptible to subtle alterations in structure, and, consequently, 3H9 may lose certain attributes of its complex specificity.
Understanding the complete specificity and the precise targets of autoantibodies is important. This is particularly true for 3H9 because 3H9 B cells arose by Ag activation in an autoimmune mouse (1). The activating Ag likely included DNA because mutation to arginine, a residue that is integral to DNA binding proteins, established specificity for DNA in the clone member with the mutation (1).
3H9 is also relevant because it is subject to self-tolerance in healthy mice. 3H9, derivatives of 3H9, and different 3H9/L chain transgene combinations show distinctly different forms of regulation that include receptor editing, deletion, anergy, allelic/isotypic inclusion, and sequestration (6). We infer that the type of regulation depends on the interaction of a given anti-DNA with its complex autoantigen. In addition, 3H9 is emerging as an extremely useful model for studying the breakdown of tolerance: autoimmune mice with anti-DNA transgenes develop autoimmunity sooner, more vigorously, and within a more predictable time frame than nontransgenic mice (6). Thus, by supplying a mouse with a ready-made autoantibody, a key variable in the induction of autoimmunity may have been eliminated. The ability of 3H9 to accelerate autoimmunity reinforces the value of the 3H9 model. But full understanding of the disease depends on accurate identification of the Ag.
The Authors Respond
In a letter to the Editor, Drs. Radic and Weigert question our conclusions (R1 ) regarding the antigenic specificity of an Ab, 3H9/Vκ4, which is used widely by the immunological community in transgenic models of B cell tolerance and autoimmunity. The literature reports variously that this Ab binds ssDNA, phosphatidylserine (or conformationally altered β2 glycoprotein I), and particularly dsDNA. We reported that it bound chromatin, specifically H2A/H2B/DNA, but not free dsDNA. In their letter, Drs. Radic and Weigert imply that nonphysiological buffers used in its purification may have altered the specificity of 3H9/Vκ4 such that it no longer bound free DNA, yet retained an ability to bind DNA in the context of the nucleosome.
We think this is unlikely for the following reasons. The Ab retained strong chromatin reactivity throughout the purification procedure. The buffers used were not inordinately harsh; NaCl (1 M), a nonchaotropic salt, was used to dissociate histones, and 0.1 M glycine, pH 2.5, was used as a standard elution buffer. The eluate was immediately neutralized in the collection tube before dialysis. The purified Ab was clear in solution. The same purification procedure did not affect DNA binding by H241, a prototypical DNA-binding Ab (R2 ). Unpurified and partially purified preparations of 3H9/Vκ4 contained contaminating proteins that migrated coincidentally with histones on SDS PAGE. And most convincingly, the loss of apparent DNA binding by purified 3H9/Vκ4 could be “reconstituted” with culture supernatants containing chromatin, or purified histones, which nucleate DNA that is immobilized on the plastic assay surface. We are confident that others can easily reproduce these results.
It is clear that 3H9/Vκ4 recognizes a complex determinant in chromatin. As we discussed in the paper (R1 ), it is conceivable that 3H9/Vκ4 makes binding contacts exclusively with DNA, which is known to have an altered conformation in the context of the nucleosome (R3 ). Alternatively, it may form bonds exclusively with histones in the context of the nucleosome or a combination of histones and DNA. But it does not bind free DNA in solution or on plastic. This distinguishes it fundamentally from DNA-binding Abs such as H241, which was used as a positive control in our experiments. Accordingly, we believe that it is more accurate to refer to 3H9/Vκ4 as an “anti-chromatin” Ab than as an “anti-DNA” Ab. Perhaps most importantly, our observations underscore the importance of purifying anti-nuclear Abs when defining their antigenic specificities.
