- Email: [email protected]
Antibody constant region: Potential to bind metal and nucleic acid
Medical Hypotheses Medical Hypotheses (1995) 44, 137-145 © PearsonProfessionalLtd 1995
Antibody Constant Region" Potential to Bind Metal and Nucleic Acid R. T. RADULESCU Molecular Concepts Research (MCR), Kunigundenstrasse 34, 80805 Munich, Germany Environmental challenges appear to elicit similar patterns of cellular responses such as positive autoregulation and autoamplification whether one considers the generation of antibodies with identical antigen specificity or the accumulation of host-protective transcription factors. Therefore, I analyzed the structure of immunoglobulins (Ig) for motifs commonly found in transcription factors. Specifically, the well-known abundance and periodic location of cysteine residues in immunoglobulin chains prompted me to check antibody constant regions for the presence of putative metal-binding domains and zinc finger-like sequences. The constant regions of Ig light and heavy chains were found to harbor one or several copies, respectively, of a short cysteineand histidine-containing sequence. Moreover, all four IgG subclasses were detected to comprise zinc finger-like motifs in their heavy chain constant and hinge domains. Yet another finding is the occurrence of several sequences of the form serine-proline-X-X and/or threonine-proline-X-X in the hinge sections of IgA and IgG3. These results suggest that antibody constant regions, as a fragment and/or embedded in a full-length immunoglobulin chain, may complex metal, thus acquiring conformations conducive to dimerization and nucleic acid binding. As such, my study provides a putative structural basis for the known requirement of divalent metal cations, particularly of zinc ions, for a normal immune response, and warrants further investigations, both theoretical and experimental, into the potential of antibody constant regions for metal binding and gene regulation. Moreover, future testing of the proposed zinc finger peptides from Ig constant domains should yield information relevant to zinc finger design with potentially wide applications in research and clinical medicine. Finally, the structural evidence described here allows the prediction that immunoglobulins and transcription factors may have diverged from a common ancestor molecule. Stimulus-dependent amplification of antibodies and transcription factors: is there a c o m m o n molecular basis?
A humoral immune response follows a sequence of events characterized by random diversification of im-
munoglobulin structure, thus giving rise to a large combinatorial pool from which a specific antibody is then selected and amplified in response to a given antigenic challenge. Prior to the encounter of antigen, antibody diversity has already been generated by mechanisms such
Date received 15 August 1994 Date accepted 28 September 1994
137
138 as somatic recombination and mutation occurring in bone marrow-derived lymphocytes (1). During further lineage development, these cells evolve into B lymphocytes, the precursors of antibody-producing plasma cells. The latter transition is triggered by binding of antigen to immunoglobulin molecules spanning the cell membrane of B-lymphocytes. Ultimately, plasma cells amplify and secrete immunoglobulins that have the same antigen-binding specificity as their membrane-bound counterparts on B-lymphocytes. However, the molecular events linking antigen binding to B-cell proliferation and differentiation into plasma cells are still poorly understood. Several studies have already demonstrated that eukaryotic transcription factors may act as efficient cellular defense lines against various environmental challenges through positive autoregulation and autoamplification processes following exposure to the challenge (2). These mechanisms were to me reminiscent of principles inherent to the humoral immune system, specifically the clonal selection of antibodies. In the present investigation, I have addressed the question of whether, conversely, an antibody harbors structural features common to transcription factors. If so, this could have wide functional implications. Cysteine-rich clusters in immunoglobulins and transcription factors: a first clue to a potential convergence in function
The basic structure of an immunoglobulin consists of two light and two heavy chains each of which comprises constant (C) and variable (V) regions. The C regions of Ig heavy chains each contain a hinge domain. Moreover, light and heavy chains are connected by several intra- and interchain disulfide bonds. In other words, each immunoglobulin chain displays a relative abundance and periodic location of cysteine residues. Since this characteristic is also common to gene regulatory proteins containing metal-binding domains (3), particularly zinc-coordinating regions (4), I set out to analyze more closely the amino acid sequences of the c o n s t a n t regions of both light and heavy chains found in the different human immunoglobulin classes, i.e. in IgG, IgM, IgA, IgD and IgE as compiled previously (5,6). The variable regions were n o t considered due to their inherent sequence variability. An additional consideration in choosing this approach concerned the functions known to be fulfilled by variable and constant regions as the first provide the antigen-binding site(s) and the latter act as effect o r domains, e.g. they are involved in complementbinding and activation. To delineate my methodology more precisely, I checked the c o n s t a n t regions for the presence of short sequences containing cysteine
MEDICALHYPOTHESES (Cys) and histidine (His) residues since, as part of a peptide or protein, these sequences have been shown to provide ligands (Cys and His) for single metal ions (3). Putative metal- and nucleic acid-binding domains in antibody constant regions
The present study yielded four notable findings. The first is that each constant region, regardless of whether contained in light or heavy chains, contains one or more copies of the short sequence Cys-X3-His (X = any amino acid), respectively. As such, the heavy chain constant regions of IgG, IgM, IgE and IgD contain three copies of this motif whereas the same domain in IgA encompasses a duplication of this sequence pattern. Figure 1 representatively shows such a sequence as being contained between residues 194 and 198 of the constant region of human kappa (k) light chain. Interestingly, a sequence of the form Cys-X3-His is also found in the helix-destabilizing protein from bacteriophage T4 where it is embedded in a region implicated in nucleic acid binding as based on spectroscopic and chemical modification studies (3). Although, in this bacteriophage protein, the sequence Cys-X3-His is part of a larger motif of the form CysX3-His-X5-Cys-X2-Cys, it can be predicted based on previous studies (3) that the sequence Cys-X3-His, as found in Ig constant regions, may bind a metal ion, e.g. a zinc cation (Zn2+). The binding of a divalent metal ion like Zn 2+ may increase the stability of these immunoglobulin domains and promote dimerization of immunoglobulin (domain) monomers, paralleling the proposed function of the sequence containing Cys-X3-His in bacteriophage T4 gene 32 protein (4). A second finding exclusively concerns the constant region of IgG, more precisely of IgG1 which is the predominant subclass among the four known IgG subclasses. I have detected between residues 200 and 224 (C! and hinge region of IgG1) a sequence of the form Cys-X3-His-X15-Cys-X3-His (Fig. 2a) and between residues 220 and 229 (hinge region of IgG1) a sequence of the form Cys-X3-His-X-Cys-X2-Cys (Fig. 2b). Each of these two sequences may represent a complete nucleic acid- or DNA-binding domain whereby the first could be a novel type of zinc finger and the latter clearly resembles the potential metal-binding sequence of the helix-destabilizing protein from~bacteriophage T4 presented above. It is quite intriguing that these two sequences are found solely in IgG1 given that this immunoglobulin is the major component of the secondary immune
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
response and, moreover, is known to have a very flexible hinge region. Therefore, it could be proposed that the hinge region of IgG1 might mediate DNA binding of the IgG1 heavy chain, based on its primary structure and its flexibility, that may ensure the fastening of the hinge domain to a DNA groove similar to the interaction occurring between zinc finger proteins and DNA. The third finding underscores the potential importance of the hinge region of immunoglobulins as a putative DNA-binding domain. Analysis of the hinge region ofimmunoglobulin A1 revealed 3 closely spaced or 6 overlapping sequences, respectively, of the form Thr-Pro-X-X and Ser-Pro-X-X located in a stretch of only 21 amino acids encompassing the hinge region and the first 3 amino acids of the C2 region of IgA1 (Fig. 3). Sequences of the form Ser-Pro-X-X and Thr-ProX-X have been shown to be frequently found in gene regulatory proteins (7). Yet a fourth finding concerns the hinge region of human IgG3 (6). Briefly, IgG3 is the third most frequent subclass among the four known IgG subclasses and contains the largest hinge section among all immunoglobulin subclasses. Interestingly, the hinge domain fragment spanning residues 228276 of IgG3 comprises a cysteine-rich sequence of the form Cys-X2-Cys-Xll-Cys-X2-Cys-X 1l-Cys-X2-CysX ll-Cys-X2-Cys (Fig. 4a) that consists of three adjacent units of the form Cys-Pro-Arg-Cys-Pro-Glu-ProLys-Ser-Cys-Asp-Thr-Pro-Pro-Pro (Fig. 4b) followed by yet another Cys-Pro-Arg-Cys-motif. If one takes as a model the Cys-X2-Cys-X13-CysX2-Cys structures of the cysteine-rich DNA-binding domains of GAL4 protein (8,9) (Fig. 4c) or the human glucocorticoid receptor (8,10) (Fig. 4d), respectively, both of which chelate one zinc ion through their cysteine tandem repeats, it could be predicted that the above fragment of IgG3 coordinates two zinc ions through two zinc finger-like structures each of which has the form Cys-X2-Cys-Xll-Cys-X2-Cys (Fig. 4e). Thus, the hinge domain of IgG3, as a fragment or embedded in the IgG3 heavy chain, may be able to bind DNA and regulate gene expression. As such, it may represent yet another member of an ever increasing family of eukaryotic transcription factors which have in common two boxes of the form Cys-X2, 4Cys or His-X2_4-His (8) coordinately binding one zinc ion and separated by a 'spacer' sequence displaying a considerable variability in size: nine residues in the methionyl tRNA synthetase from yeast (8), 17 residues in the transcription factor GATA-1 (11) and 24 residues in the transcriptional elongation factor TFIIS (12).
139
The assumption of potential DNA binding of the hinge domain of IgG3 is supported by the presence of the DNA-binding motif TPXX (7) preceding each Cys-X2-Cys-box within the analyzed region (Figs. 4a, b&e). Moreover, the IgG2 and IgG4 subclasses were also found to comprise zinc finger-like motifs in their heavy chain constant and hinge domains (data not shown).
Gene regulation by antibody constant domains as part of immunoglobulin chains or of proteolytically cleaved immunoglobulin fragments: a possible scenario This is the first study to propose that the constant regions of human immunoglobulin molecules contain metal- and nucleic acid-binding domains. These findings imply that these Ig domains, acting either as a fragment or as the active site within the complete Ig chain, may directly influence gene expression. Conceivably, the regulation of gene expression by immunoglobulin hinge and constant region subunits may entail the following steps. Following internalization of an antigen-antibody complex into the cell (13) or Fc receptor-mediated antibody uptake, the antibody disulfide bonds are cleaved as a result of the reducing properties of the cytosolic compartment. Subsequently, free antibody chains and possibly chain fragments are released into the cytosol. If chain fragments were to be the effector molecules, one would have to invoke the specific action of proteases upon the separated antibody chains. Presumably, these proteinases would be related or identical to proteinases acting on other proteins essential to the immune system such as the invariant chain (Ichain) which is involved in MHC-dependent intracellular antigen processing (14). The biologically active antibody particles, i.e. the hinge region-encompassing antibody heavy chains, or fragments thereof, would then advance into the nucleus, e.g. through association with cytosolic proteins translocating to the nucleus, and proceed there to influence transcriptional processes. In those cells which are able to synthesize de novo antibody chains, the same scenario of nuclear translocation and action of these subunits would apply.
Proposed cell growth regulation by antibody fragments: yet another addition to the subunit hypothesis In either case, be it internalized or de novo synthesized immunoglobulin chains, these molecules would
140
MEDICAL HYPOTHESES
CEVTH Fig. 1 Proposed metal-binding domain of human k light chain constant region (residues 194-198). Amino acids in bow represent potential ligands for a divalent metal cation.
CNVNHKPSNTKVDKRVEPKSCDKTH Fig. 2a Proposed zinc finger domain of human Ig G1 heavy chain constant region (residues 200-224). Amino acids in bold represent potential
ligands for a divalentzinc cation. C1 constantregionprecedeshingeregion(underlined).
CDKTHTCPPC Fig. 2b Proposednucleicacid-bindingdomainof human IgG1 hingeregion(residues220-229). Aminoacids in bold representpotential ligandsfor a divalentmetalcation.
PVP S TPPTPSPS
TPP TPSP
SC
Fig. 3 Proposed DNA-binding domain of human IgAl heavy chain constant region (residues 221-241). S/T P repeats are in bold. First amino acid of each complete, non-overlapping S/T P X X-motif in italic. C2 constant region succeeds hinge region (underlined).
a CPRCPEPKSCDTPPPCPRCPEPKSCDTPPPCPRCPEPKSCDTPPPCPRC b CPRCPEPKSCDTPPP c Cys Asp IIe Cys Arg Leu Lys Lys Leu Lys Cys Ser Lys Glu Lys Pro Lys Cys Ala Lys Cys d Cys Leu Val Cys Ser Asp Glu Ala Ser Gly Cys His Tyr Gly Val Leu Thr Cys Gly Ser Cys e Cys Pro Arg Cys Pro Glu Pro Lys Ser
Cys
Asp Thr Pro Pro Pro Cys Pro Arg Cys
Fig. 4 a: Fragment (residues 228-276) from hinge region of human Ig G3; b: Recurring amino acid sequence motif in the hinge region oflg G3; c: AlignedfragmentfromDNA-bindingregionof GAIAprotein(residues 11-31); d: AlignedfragmentfromDNA-bindingregionof humanglucocorticoidreceptor(residues421-441); e: Alignedfragment(residues228-246 or 243-261 or 258-276, respectively)fromhinge regionof human IgG3.
Cysteineresiduesin bold are potentialmetalchelators;TP-tandemsin italic are part of the proposedDNA-bindingmotifTPXX.Figs 4a and b are in one-letteraminoacid codeand Figs4e-eare in three-letteraminoacid code. follow intracellular or intracrine pathways similar to those previously shown or predicted for growth factors and their fragments (15,16). These observations on growth factor fragments have previously led me to put forward a novel concept on cell growth regulation termed the 'subunit hypothesis' (16).
However, an important distinction should be made. Prior to advancing into the nucleus, these antibody domains are proposed to assume conformations conducive to DNA binding through the chelation of zinc ions present in the cytosol. Similar to the dependence of proper folding of GAL4 protein (17,18) and the
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
human glucocorticoid receptor (10), respectively, on the surrounding zinc concentration, it may be that zinc ions fulfill an analogous role in establishing immunoglobulin domain configurations indispensable for their proposed ability to regulate gene expression. Within the nucleus, these antibody subunits may affect both immunoglobulin gene transcription and that of other genes, thus participating directly in the maturation of cells able to synthesize or internalize antibodies. Potential outcomes of these events could be proliferation and differentiation of B-lymphocytes as well as amplification of immunoglobulins of a given specificity in plasma cells. Furthermore, the above Ig domains may be involved, through positive autoregulatory circuits, in conferring growth autonomy to certain types of lymphoid neoplasias associated with increased production of immunoglobulins. Studies undertaken with various lymphoid tumors support these assumptions. As such, it has been shown that the Yheavy chain constant locus maps to a region of translocations in malignant B-lymphocytes (19) and that the expression of an c~l heavy chain-like protein correlates with the clinical picture of a lymphoma (20). Moreover, investigations into the causes of heavy chain disease, a plasma cell disorder (21), have extensively shown the presence of heavy chain fragments-of which the "/3 heavy chain (including the 73 hinge region) is the best studied (22) - in both serum (23,24) and lymphoid cells (25) taken from patients afflicted by this disease. These findings could also be reproduced in a mouse myeloma cell line (26). In the light of the present study, it appears warranted to re-evaluate the potential involvement of these Ig fragments in the pathogenesis of human B-cell neoplasia. Finally, it has been demonstrated that human myeloma cells generate hybrid Ig molecules, e.g. hybrid antibodies of the types IgG4-IgG2 (27) and IgG3-IgG1 (28), whose role in the maturation of the lymphoid lineage should be further investigated along the lines suggested above. If it holds true that immunoglobulins and/or above specific subunits act as transcription factors, they should also influence the development of other organs besides the immune system. For instance, these molecules may be a major factor in the organogenesis and physiology of the brain. This is conceivable in the light of the known interdependence between immunocytes and nerve cells (29). Particularly noteworthy in this context is that pluripotential hemopoietic stem cells are present in adult brain (30) and cytokines affecting Ig synthesis regulate neuronal differentia-
141
tion (31). An additional clue in favor of the above hypothesis may be the fact that many cell surface molecules found in the nervous system and implicated in its ontogeny are members of the immunoglobulin gene superfamily (32). The antibody domains described above may not only autoregulate immunoglobulin gene expression through specific DNA and/or RNA binding, but also participate in trans-activation of DNA-binding proteins through both binding of these proteins and direct recognition of nucleic acids inside or outside the gene response element specific for the respective DNAbinding protein. Interestingly, a recent report showed that specific antibodies to the yeast transcriptional activator HAP1 can induce DNA binding of certain short amino-terminal fragments of the protein (33). The authors interpreted their findings as being due to antibody-promoted dimerization of these fragments. In the light of the present findings, however, the possibility remains that these antibodies may also have physically associated with DNA. At this point, it is tempting to reconsider Padlan's query: 'One wonders also whether the variation in the structure of the hinge may have a bearing on the respective roles that the different antibody types play in the overall immune response' (6), proposed in reference to hinge-mediated antibody reach and rotational adaptability. Based on the data of this report, the scope of the above question would extend to whether the different hinge regions including the neighboring residues of the constant regions may also confer distinct DNA-binding specificities important for gene regulation during the development of the immune system and of other organs. Previous experimental evidence on zinc chelation and nucleic acid binding by immunoglobulins The present study primarily suggests metal and nucleic acid binding of immunoglobulin chains and/or chain fragments which, as such, do not possess disulfide bonds and thus can employ their cysteines as metal chelators. And yet, although the same cysteines are part of disulfide bridges within the completely assembled immunoglobulin molecule and thus are unlikely to be free for metal coordination, serum immunoglobulins have been shown to bind zinc. As such, previous studies have demonstrated that zinc ions bind serum immunoglobulins, particularly IgG, in vitro and in vivo (34) and that zinc sulfate is efficiently used to precipitate immunoglobulins, thus being a useful agent in determining serum immunoglobulin concentration as assessed by the zinc sulfate turbidity test (35,36). Re-
142 markably, this test could even be efficiently applied to distinguish between different IgG subclasses derived from human myelomas (37). In line with my data are also past studies into the nature of translational control of immunoglobulin synthesis. As such, it has been shown that myeloma antibody is able to repress heavy chain synthesis in cultured myeloma cells (38). The same myeloma protein has also been demonstrated to interact with the mRNA coding for the Ig heavy chain (39) in a cell-free system. In vivo, it is likely that it is noncovalently assembled Ig (40) that associates with RNA, since this interaction can only occur in the reducing cytosolic compartment. The noncovalent forces potentially holding together the Ig molecules in the cytoplasm may be conferred by salt bridges between Ig domains (41) and/or ~3-sheet packings of the intercalating (nonaligned) type as shown for Ig fragments derived from Ig constant domains (42). In this context, it is notable that Ig chains have been found in the cytoplasm of cells extracted from so-called non-secretory human myeloma (43) and from a mouse myeloma (44).
A proposed dual role for antibodies: the fully assembled molecule vs the fragment Most likely, serum IgG binds zinc through residues other than the chemically unavailable cysteines, e.g. through histidines and glutamates. As such, immunoglobulins appear to resemble metalloproteins such as matrix metalloproteinases (MMPs). These enzymes have been shown to complex zinc in two forms: one form in which a cysteine residue joins other residues such as histidines in metal chelation and another form in which the same cysteine residue is n o t involved in metal coordination and instead participates e.g. in the formation of a disulfide bridge. In the latter case, it is proposed that water substitutes for the cysteine residue whereas the histidines involved in zinc coordination in the first form are maintained in a chelating configuration in the second form, too (45-47). Accordingly, the proposed metal-binding domains in immunoglobulins may serve a dual function. As part of the completely assembled (serum) immunoglobulins they mediate the transport and neutralization of zinc ions or other divalent cations in the blood circulation - a rather unspecific function with participation of histidines, but not cysteines. As part of a disulfide bond-free antibody chain or chain fragment comprising constant domain sequences, they may chelate zinc in the reducing cytosolic compartment and acquire conformations conducive to an interaction with distinct DNA and/or
MEDICAL HYPOTHESES
RNA regions - a rather specific function that is proposed to be carried out by clusters of cysteines and histidines forming zinc finger-like motifs. This view is also supported by a recent review on zinc finger motifs (48). Beyond the clear distinction between reducing intracellular conditions favoring protein stabilization by zinc chelation and oxidizing extracellular conditions favoring protein stabilization by disulfide bridge formation, these authors point out that some extracellular proteins employ disulfide bonds to stabilize topologies that are similar to intracellular zinc-stabilized motifs.
Future possibilities of verification of the present proposal on metal and nucleic acid binding by antibody fragments A first step in the verification of the present hypothesis immediately suggests itself. One could express any of the antibody fragments proposed above to complex zinc by recombinant DNA technology and test its binding to zinc. This type of approach in which a protein fragment proposed to be a metal-binding sequence is tested for its metal-binding properties has been validated extensively (49-52). Interestingly, the identification of metal-binding domains in retroviral proteins by sequence analysis correctly predicted the direct involvement of these regions in viral growth inside the host cell as a result of metal coordination and complex formation with retroviral RNA and paved the way for further explorations that ultimately showed that also mature retroviral particles assembled from these subunits bind zinc, thus providing a double rationale for antiviral therapy by means of zinc chelation (53,54).
Evolutionary implications Evolutionarily speaking, immunoglobulins and transcription factors may have diverged from a common ancestor molecule. In analogy to the recent unravelling of a short track from the membrane to the nucleus involving a novel group of cytoplasmic transcription factors termed STATs (an acronym for signal transducers and activators of transcription) which directly transduce the signal generated by binding of the immunologically important interferons cc or T to their cell membrane receptors (55), immunoglobulins may be henceforth perceived as CMR-STATs, i.e. cell membrane receptors, signal transducers and activators of transcription. This proposed versatility would explain in an intriguing manner why the structure of an antibody has been selected as a central molecule for defending the organism. The concept of a (soluble or anchored) membrane protein turned transcription factor as suggested herein
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
to apply to antibodies has already been validated with sterol regulatory element-binding protein 1 (SREBP1) that is attached to the nuclear envelope and endoplasmic reticulum (56). Interestingly, it is a fragment cleaved from SREBP-1, not SREBP-1 itself, that translocates to the nucleus where it regulates transcription (56).
Conclusions and perspectives Taken together, this communication provides a putative structural basis for the known requirement of divalent metal cations, particularly of Zn 2+, for a normal immune response (57,58) and warrants further investigations, both theoretical and experimental, into the potential of antibody constant regions for metal binding and gene regulation. In the same way as the previous analysis of the amino acid sequence around the cysteine residues in Ig variable and constant regions provided clues for understanding the generation of antibody diversity (59), the present identification of putative metal binding domain- and zinc finger-like motifs in Ig constant regions embracing yet again cysteine residues should accelerate insight into further fundamental immunological events as well as add to our understanding of more general aspects of cell growth and differentiation. Conversely, future testing of the proposed zinc finger peptides from Ig constant domains should yield information relevant to zinc finger design with potentially wide applications in research and clinical medicine. This avenue has already been envisaged for other zinc finger peptides (60) and predicted to encompass a large combinatorial repertoire as stated before (61): 'Clearly, variations in the finger structure itself or in the length or characteristics of the linkers between fingers may allow variations in the types of interaction between finger proteins and nucleic acids. These variant structures may be crucial for the nucleic acid recognition process.' Finally, my present findings on a possible autoamplificatory role of antibody subunits support my previously advanced proposal (62) according to which it would suffice to hook up a DNA-binding domain to a ligand-binding domain to create a potentially efficacious pharmaceutical agent termed 'synthetic inducible biological response amplifier' (SIBRA) since the results described here suggest that antibodies, similar to SIBRAs, harbor putative gene regulatory domains (in the constant region) linked to a ligandbinding domain (in the variable region). In the reverse analogy, this report supports the notion that SIBRAs may indeed be regarded as an artificial acceleration
143
and refinement of principles intrinsic to the immune system (62).
Note added in proof Consistent with the findings and implications of the present study, two recent reports have experimentally shown that immunoglobulin (Ig) heavy and light chains promote B-cell development in a differential manner (63,64). These investigations confirm and extend the results of a previous study implicating Ig proteins in the control of immunoglobulin gene rearrangement during B-cell differentiation (65). The present proposal according to which Ig chains may partly achieve growth regulatory effects through their potential zinc finger-like motifs also appears likely in the light of the recent identification of a novel zinc finger protein that is able to drive the maturation of B-lymphocytes into Ig-producing plasma cells (66).
Acknowledgements I would like to thank Drs L. Mario Amzel and Jeremy M. Berg for encouragement and critical comments.
References I. TonegawaS. Somatic generation of antibody diversity. Nature 1983; 302: 575-581. 2. Zhou P, Thiele D J. Rapid transcriptional autoregulation of a yeast metalloregulatory transcription factor is essential for high-level copper detoxification. Genes Dev 1993; 7: 1824-1835. 3. Berg J M. Potential metal-binding domains in nucleic acid binding proteins. Science 1986; 232: 485-487. 4. Berg J M. Zinc fingers and other metal-binding domains. J Biol Chem 1990; 265: 6513~5516. 5. Day E D. Advanced lmmunochemistry2nd edn. New York: Wiley-Liss, 1990: 9~65. 6. PadlanE A. Anatomyof the antibody molecule. Mol Immunol 1994; 31: 169-217. 7. Suzuki M. SPXX: a frequent sequence motif in gene regulatory proteins. J Mol Biol 1989; 207: 61-84. 8. Johnston M, Dover J. Mutations that inactivate a yeast transcriptional regulatory protein cluster in an evolutionarilyconserved DNA-bindingdomain. Proc Natl Acad Sci USA 1987; 84: 2401-2405. 9. Pan T, Coleman J E. Structure and function of the Zn(II) binding site within the DNA-binding domain of the GAL4 transcription factor. Proc Natl Acad Sci USA 1989; 86: 31453149. 10. Freedman L P, Luisi B F, Korszun Z R, Basavappa R, Sigler P B, Yamamoto K R. The function and structure of the metal coordination sites within the glucocorticoid receptor DNAbinding domain. Nature 1988; 334: 543-546. 11. Omichinski J G, Clore G M, Schaad Oet al. NMR structure of a specific DNA complex of Zn-containing DNA binding domain of GATA-1. Science 1993; 261: 438-446. 12. Qian X, Jeon C, YoonH, Agarwal K, Weiss M A. Structure of
144
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26. 27. 28. 29. 30. 31.
32.
33.
a new nucleic-acid-binding motif in eukaryotic transcriptional elongation factor TFIIS. Nature 1993; 365: 277-279. Mitchell R N, Shaw A C, Weaver Y K, Leder E Abbas A K. Cytoplasmic tail deletion converts membrane immunoglobulin to a phosphatidylinositol-linked form lacking signaling and efficient antigen internalization functions. J Biol Chem 1991; 266: 8856-8860. Maric M A, Taylor M D, Blum J S. Endosomal aspartic proteinases are required for invariant chain processing. Proc Natl Acad Sci USA 1994; 91: 2171-2175. Radnlescu R T, Wendtner C M. Proposed interaction between insulin and retinoblastoma protein. J Mol Recognit 1992; 5: 133-137. Radulescu R T, Wendtner C M. Hormone and growth factor subunits: a novel perception of cell growth regulation. J Endocrinol 1993; 139: 1-7. Johnston M. Genetic evidence that zinc is an essential cofactor in the DNA-binding domain of GAL4 protein. Nature 1987; 328: 353-355. Pan T, Coleman J E. GAL4 transcription factor is not a 'zinc finger' but forms a Zn(II)2Cys 6 binuclear cluster. Proc Natl Acad Sci USA 1990; 87: 2077-2081. Kirsch 1 R, Morton C C, Nakahara K, Leder E Human immunoglobulin heavy chain genes map to a region of translocations in malignant B lymphocytes. Science 1982; 216: 301303. Seligmann M, Danon E Hurez D, Mihaesco E, Preud'homme J-L. Alpha-chain disease: a new immunoglobulin abnormality. Science 1968; 162: 1396-1397. Seligmann M, Mihaesco E, Preud'homme J-L, Danon F, Brouet J-C. Heavy chain diseases: current findings and concepts. Immunol Rev 1979; 48: 145-167. Frangione B, Franklin E C. Split immunoglobulin genes and human heavy chain deletion-mutants. J Immunol 1979; 122: 1177-1179. Adlersberg J B, Franklin E C, Frangione B. Repetitive hinge region sequences in human IgG3: Isolation of an ll,000Dalton fragment. Proc Natl Acad Sci USA 1975; 72: 723-727. Michaelsen R E, Frangione B, Franklin E C. Primary structure of the 'hinge' region of human IgG3-probable quadruplication of a 15-amino acid residue basic unit. J Biol Chem 1977; 252: 883-889. Buxbaum J N, Preud'homme J-L. Alpha and gamma heavy chain diseases in man: intracellular origin of the aberrant polypeptides. J Immunol 1972; 109: 1131-1137. Morrison S L, Scharff M D. Heavy chain-producing variants of a mouse myeloma cell line. J Immunol 1975; 114: 655-659. Natvig J B, Kunkel H G. A hybrid IgG4-1gG2 immunoglobulin. J Immunol 1974; ll2: 1277-1284. Kunkel H G, Natvig J B, Joslin F G. A 'Lepore' type of hybrid y globulin. Proc Natl Acad Sci USA 1969; 62: 144-149. Reichlin S. Neuroendocrine-immune interactions. New Engl J Med 1993; 329: 1246-1253. Bartlett P E Pluripotential hemopoietic stem cells in adult mouse brain. Proc Natl Acad Sci USA 1982; 79: 2722-2725. Mehler M E Rozental R, Dougherty M, Spray D C, Kessler J A. Cytokine regulation of neuronal differentiation of hippocampal progenitor cells. Nature 1993; 362: 62-65. Steinman L. Connections between the immune system and the nervous system. Proc Natl Acad Sci USA 1993; 90: 79127914. Zhang L, Bermingham-McDonogh O, Turcotte B, Guarente L. Antibody-promoted dimerization bypasses the regulation of DNA binding by the heme domain of the yeast transcriptional activator HAPI. Proc Natl Acad Sci USA 1993; 90: 28512855.
MEDICAL HYPOTHESES
34. Prasad A S, Oberleas D. Binding of zinc to amino acids and serum proteins in vitro. J Lab Clin Med 1970; 76: 416--425. 35. McEwan A D, Fisher E W, Selman I E, Penhale W J. A turbidity test for the estimation of immune globulin levels in neonatal calf serum. Clin Chem Acta 1970; 27: 155-163. 36. Caldow G L, Edwards S, Peters A R, Nixon P, lbata G, Sayers R. Associations between viral infections and respiratory disease in artificially reared calves. Vet Rec 1993; 133: 85-89. 37. Miyagawa N, Okamoto Y, Morita K, Dohi K. Studies on the relationship between serum colloidal reactions (ZTT and TTT) and IgG subclasses, especially lgGl and IgG2. Microbiol lmmunol 1991; 35: 59-66. 38. Stevens R H, Williamson A R. Translational control of immunoglobulin synthesis. I. Repression of heavy chain synthesis. J Mol Biol 1973a; 78: 505-516. 39. Stevens R H, Williamson A R. Translational control of immunoglobulin synthesis. 11. Cell-free interaction of myeloma immunoglobulin with mRNA. J Mol Biol 1973b; 78: 517-525. 40. Williamson A R. Biosynthesis of immunoglobulins. In: Defense and Recognition. R.R. Porter, ed. London: Butterworths and Baltimore: University Park Press, 1973; 10: 229-251. 41. Schiffer M, Chang C-H, Naik V M, Stevens F J. Analysis of immunoglobulin domain interactions. Evidence for a dominant role of salt bridges. J Mol Biol 1988; 203: 799-802. 42. Miller S. Protein-protein recognition and the association of immunoglobulin constant domains. J Mol Biol 1990; 216: 965-973. 43. Hurez D, Preud'homme J-L, Seligmann M. Iotracellular 'monoclonal' immunoglobulin in non-secretory human myeloma. J Immunol 1970; 104: 263-264. 44. Morrison S L. Sequentially derived mutants of the constant region of the heavy chain of murine immunoglobulins. J Immunol 1979; 123: 793-800. 45. Springman E B, Angleton E L, Birkedal-Hansen H, Van Wart H E. Multiple modes of activation of latent human fibroblast collagenase: evidence for the role of a Cys 73 active-site zinc complex in latency and a 'cysteine switch' mechanism for activation. Proc Natl Acad Sci USA 1990; 87: 364-368. 46. Van Wart H E, Birkedal-Hansen H. The cysteine switch: a principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family. Proc Natl Acad Sci USA 1990; 87: 5578-5582. 47. Hirose T, Patterson C, Pourmotabbed T, Mainardi C L, Hasty K A. Structure-function relationship of human neutrophil collagenase: identification of regions responsible for substrate specificity and general proteinase activity. Proc Natl Acad Sci USA 1993; 90: 2569-2573. 48. Schwabe J W R, Klug A. Zinc mining for protein domains. Nature Struct Biol 1994; 1: 345-349. 49. Green L M, Berg J M. A retroviral Cys-Xaa2-Cys-Xaa4-HisXaa4-Cys peptide binds metal ions: spectroscopic studies and a proposed three-dimensional structure. Proc Natl Acad Sci USA 1989; 86: 4047-4051. 50. Summers M F, South T L, Kim B, Hare D R. High-resolution structure of an HIV zinc fingerlike domain via a new NMRbased distance geometry approach. Biochemistry 1990; 29: 329-340. 51. Michael S F, Kilfoil V J, Schmidt M H, Amann B T, Berg J M. Metal binding and folding of a minimalist Cys2His2 zinc finger peptide. Proc Natl Acad Sci USA 1992; 89: 4796-4800. 52. Berg J M. Zinc-finger proteins. Curr Opin Struct Biol 1993; 3: 11-16. 53. Summers M F, Henderson L E, Chance M R et al. Nucleocapsid zinc fingers detected in retroviruses: EXAFS studies of intact viruses and the solution-state structure of the nudeocapsid protein from H1V-I. Protein Sci 1992; 1: 563-574.
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
54. Rice W G, Schaeffer C A, Harten B e t al. Inhibition of HIV1 infectivity by zinc-ejecting aromatic C-nitroso compounds. Nature 1993; 361: 473-475. 55. Shuai K, Horvath C M, Tsai Huang L H, Qureshi S A, Cowburn D, Damell Jr J E. Interferon activation of the transcription factor Stat91 involves dimerization through SH2phosphotyrosyl peptide interactions. Cell 1994; 76: 821-828. 56. Wang X, Sato R, Brown M S, Hua X, Goldstein J L. SREBP1: a membrane-bound transcription factor released by sterolregulated proteolysis. Cell 1994; 77: 53-62. 57. Chandra R K. Trace element regulation of immunity and infection. J Am Coil Nutr 1985; 4: 5-16. 58. Vallee B L, Falchuk K H. The biochemical basis of zinc physiology. Physiol Rev 1993; 73: 79-118. 59. Milstein C. From antibody structure to immunological diversification of immune response. Science 1986; 23l: 1261-1268. 60. Rebar E J, Pabo C O. Zinc finger phage: affinity selection of fingers with new DNA-binding specificities. Science 1994; 263: 671-673. 61. Berg J M. Proposed structure for the zinc-binding domains
62.
63. 64.
65.
66.
145
from transcription factor IliA and related proteins. Proc Natl Acad Sci USA 1988; 85: 99-102. Radulescu R T. Synthetic inducible biological response amplifiers (SIBRAs): rational peptides at the crossroads between molecular evolution and structure-based drug design. Med Hypotheses 1995; 44: 32-38. Spanopoulou E, Roman C A J, Corcoran L Met al. Functional immunoglobulin transgenes guide ordered B-cell differentiation in Rag-l-deficient mice. Genes Dev 1994; 8: I030-1042. Young F, Ardman B, Shinkai Y e t al. Influence of immunoglobulin heavy-chain and light-chain expression on Bcell differentiation. Genes Dev 1994; 8: I043-1057. Iglesias A, Kopf M, Williams G S, Biihler B, K6hler G. Molecular requirements for the ~-induced light chain gene rearrangement in pre-B cells. EMBO J 1991; 10: 2147-2156. Turner Jr C A, Mack D H, Davis M M, Blimp-l, a novel zinc finger-containing protein that can drive the maturation of B lymphocytes into immunoglobulin-secreting cells. Cell 1994; 77: 297-306.
Antibody Constant Region" Potential to Bind Metal and Nucleic Acid R. T. RADULESCU Molecular Concepts Research (MCR), Kunigundenstrasse 34, 80805 Munich, Germany Environmental challenges appear to elicit similar patterns of cellular responses such as positive autoregulation and autoamplification whether one considers the generation of antibodies with identical antigen specificity or the accumulation of host-protective transcription factors. Therefore, I analyzed the structure of immunoglobulins (Ig) for motifs commonly found in transcription factors. Specifically, the well-known abundance and periodic location of cysteine residues in immunoglobulin chains prompted me to check antibody constant regions for the presence of putative metal-binding domains and zinc finger-like sequences. The constant regions of Ig light and heavy chains were found to harbor one or several copies, respectively, of a short cysteineand histidine-containing sequence. Moreover, all four IgG subclasses were detected to comprise zinc finger-like motifs in their heavy chain constant and hinge domains. Yet another finding is the occurrence of several sequences of the form serine-proline-X-X and/or threonine-proline-X-X in the hinge sections of IgA and IgG3. These results suggest that antibody constant regions, as a fragment and/or embedded in a full-length immunoglobulin chain, may complex metal, thus acquiring conformations conducive to dimerization and nucleic acid binding. As such, my study provides a putative structural basis for the known requirement of divalent metal cations, particularly of zinc ions, for a normal immune response, and warrants further investigations, both theoretical and experimental, into the potential of antibody constant regions for metal binding and gene regulation. Moreover, future testing of the proposed zinc finger peptides from Ig constant domains should yield information relevant to zinc finger design with potentially wide applications in research and clinical medicine. Finally, the structural evidence described here allows the prediction that immunoglobulins and transcription factors may have diverged from a common ancestor molecule. Stimulus-dependent amplification of antibodies and transcription factors: is there a c o m m o n molecular basis?
A humoral immune response follows a sequence of events characterized by random diversification of im-
munoglobulin structure, thus giving rise to a large combinatorial pool from which a specific antibody is then selected and amplified in response to a given antigenic challenge. Prior to the encounter of antigen, antibody diversity has already been generated by mechanisms such
Date received 15 August 1994 Date accepted 28 September 1994
137
138 as somatic recombination and mutation occurring in bone marrow-derived lymphocytes (1). During further lineage development, these cells evolve into B lymphocytes, the precursors of antibody-producing plasma cells. The latter transition is triggered by binding of antigen to immunoglobulin molecules spanning the cell membrane of B-lymphocytes. Ultimately, plasma cells amplify and secrete immunoglobulins that have the same antigen-binding specificity as their membrane-bound counterparts on B-lymphocytes. However, the molecular events linking antigen binding to B-cell proliferation and differentiation into plasma cells are still poorly understood. Several studies have already demonstrated that eukaryotic transcription factors may act as efficient cellular defense lines against various environmental challenges through positive autoregulation and autoamplification processes following exposure to the challenge (2). These mechanisms were to me reminiscent of principles inherent to the humoral immune system, specifically the clonal selection of antibodies. In the present investigation, I have addressed the question of whether, conversely, an antibody harbors structural features common to transcription factors. If so, this could have wide functional implications. Cysteine-rich clusters in immunoglobulins and transcription factors: a first clue to a potential convergence in function
The basic structure of an immunoglobulin consists of two light and two heavy chains each of which comprises constant (C) and variable (V) regions. The C regions of Ig heavy chains each contain a hinge domain. Moreover, light and heavy chains are connected by several intra- and interchain disulfide bonds. In other words, each immunoglobulin chain displays a relative abundance and periodic location of cysteine residues. Since this characteristic is also common to gene regulatory proteins containing metal-binding domains (3), particularly zinc-coordinating regions (4), I set out to analyze more closely the amino acid sequences of the c o n s t a n t regions of both light and heavy chains found in the different human immunoglobulin classes, i.e. in IgG, IgM, IgA, IgD and IgE as compiled previously (5,6). The variable regions were n o t considered due to their inherent sequence variability. An additional consideration in choosing this approach concerned the functions known to be fulfilled by variable and constant regions as the first provide the antigen-binding site(s) and the latter act as effect o r domains, e.g. they are involved in complementbinding and activation. To delineate my methodology more precisely, I checked the c o n s t a n t regions for the presence of short sequences containing cysteine
MEDICALHYPOTHESES (Cys) and histidine (His) residues since, as part of a peptide or protein, these sequences have been shown to provide ligands (Cys and His) for single metal ions (3). Putative metal- and nucleic acid-binding domains in antibody constant regions
The present study yielded four notable findings. The first is that each constant region, regardless of whether contained in light or heavy chains, contains one or more copies of the short sequence Cys-X3-His (X = any amino acid), respectively. As such, the heavy chain constant regions of IgG, IgM, IgE and IgD contain three copies of this motif whereas the same domain in IgA encompasses a duplication of this sequence pattern. Figure 1 representatively shows such a sequence as being contained between residues 194 and 198 of the constant region of human kappa (k) light chain. Interestingly, a sequence of the form Cys-X3-His is also found in the helix-destabilizing protein from bacteriophage T4 where it is embedded in a region implicated in nucleic acid binding as based on spectroscopic and chemical modification studies (3). Although, in this bacteriophage protein, the sequence Cys-X3-His is part of a larger motif of the form CysX3-His-X5-Cys-X2-Cys, it can be predicted based on previous studies (3) that the sequence Cys-X3-His, as found in Ig constant regions, may bind a metal ion, e.g. a zinc cation (Zn2+). The binding of a divalent metal ion like Zn 2+ may increase the stability of these immunoglobulin domains and promote dimerization of immunoglobulin (domain) monomers, paralleling the proposed function of the sequence containing Cys-X3-His in bacteriophage T4 gene 32 protein (4). A second finding exclusively concerns the constant region of IgG, more precisely of IgG1 which is the predominant subclass among the four known IgG subclasses. I have detected between residues 200 and 224 (C! and hinge region of IgG1) a sequence of the form Cys-X3-His-X15-Cys-X3-His (Fig. 2a) and between residues 220 and 229 (hinge region of IgG1) a sequence of the form Cys-X3-His-X-Cys-X2-Cys (Fig. 2b). Each of these two sequences may represent a complete nucleic acid- or DNA-binding domain whereby the first could be a novel type of zinc finger and the latter clearly resembles the potential metal-binding sequence of the helix-destabilizing protein from~bacteriophage T4 presented above. It is quite intriguing that these two sequences are found solely in IgG1 given that this immunoglobulin is the major component of the secondary immune
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
response and, moreover, is known to have a very flexible hinge region. Therefore, it could be proposed that the hinge region of IgG1 might mediate DNA binding of the IgG1 heavy chain, based on its primary structure and its flexibility, that may ensure the fastening of the hinge domain to a DNA groove similar to the interaction occurring between zinc finger proteins and DNA. The third finding underscores the potential importance of the hinge region of immunoglobulins as a putative DNA-binding domain. Analysis of the hinge region ofimmunoglobulin A1 revealed 3 closely spaced or 6 overlapping sequences, respectively, of the form Thr-Pro-X-X and Ser-Pro-X-X located in a stretch of only 21 amino acids encompassing the hinge region and the first 3 amino acids of the C2 region of IgA1 (Fig. 3). Sequences of the form Ser-Pro-X-X and Thr-ProX-X have been shown to be frequently found in gene regulatory proteins (7). Yet a fourth finding concerns the hinge region of human IgG3 (6). Briefly, IgG3 is the third most frequent subclass among the four known IgG subclasses and contains the largest hinge section among all immunoglobulin subclasses. Interestingly, the hinge domain fragment spanning residues 228276 of IgG3 comprises a cysteine-rich sequence of the form Cys-X2-Cys-Xll-Cys-X2-Cys-X 1l-Cys-X2-CysX ll-Cys-X2-Cys (Fig. 4a) that consists of three adjacent units of the form Cys-Pro-Arg-Cys-Pro-Glu-ProLys-Ser-Cys-Asp-Thr-Pro-Pro-Pro (Fig. 4b) followed by yet another Cys-Pro-Arg-Cys-motif. If one takes as a model the Cys-X2-Cys-X13-CysX2-Cys structures of the cysteine-rich DNA-binding domains of GAL4 protein (8,9) (Fig. 4c) or the human glucocorticoid receptor (8,10) (Fig. 4d), respectively, both of which chelate one zinc ion through their cysteine tandem repeats, it could be predicted that the above fragment of IgG3 coordinates two zinc ions through two zinc finger-like structures each of which has the form Cys-X2-Cys-Xll-Cys-X2-Cys (Fig. 4e). Thus, the hinge domain of IgG3, as a fragment or embedded in the IgG3 heavy chain, may be able to bind DNA and regulate gene expression. As such, it may represent yet another member of an ever increasing family of eukaryotic transcription factors which have in common two boxes of the form Cys-X2, 4Cys or His-X2_4-His (8) coordinately binding one zinc ion and separated by a 'spacer' sequence displaying a considerable variability in size: nine residues in the methionyl tRNA synthetase from yeast (8), 17 residues in the transcription factor GATA-1 (11) and 24 residues in the transcriptional elongation factor TFIIS (12).
139
The assumption of potential DNA binding of the hinge domain of IgG3 is supported by the presence of the DNA-binding motif TPXX (7) preceding each Cys-X2-Cys-box within the analyzed region (Figs. 4a, b&e). Moreover, the IgG2 and IgG4 subclasses were also found to comprise zinc finger-like motifs in their heavy chain constant and hinge domains (data not shown).
Gene regulation by antibody constant domains as part of immunoglobulin chains or of proteolytically cleaved immunoglobulin fragments: a possible scenario This is the first study to propose that the constant regions of human immunoglobulin molecules contain metal- and nucleic acid-binding domains. These findings imply that these Ig domains, acting either as a fragment or as the active site within the complete Ig chain, may directly influence gene expression. Conceivably, the regulation of gene expression by immunoglobulin hinge and constant region subunits may entail the following steps. Following internalization of an antigen-antibody complex into the cell (13) or Fc receptor-mediated antibody uptake, the antibody disulfide bonds are cleaved as a result of the reducing properties of the cytosolic compartment. Subsequently, free antibody chains and possibly chain fragments are released into the cytosol. If chain fragments were to be the effector molecules, one would have to invoke the specific action of proteases upon the separated antibody chains. Presumably, these proteinases would be related or identical to proteinases acting on other proteins essential to the immune system such as the invariant chain (Ichain) which is involved in MHC-dependent intracellular antigen processing (14). The biologically active antibody particles, i.e. the hinge region-encompassing antibody heavy chains, or fragments thereof, would then advance into the nucleus, e.g. through association with cytosolic proteins translocating to the nucleus, and proceed there to influence transcriptional processes. In those cells which are able to synthesize de novo antibody chains, the same scenario of nuclear translocation and action of these subunits would apply.
Proposed cell growth regulation by antibody fragments: yet another addition to the subunit hypothesis In either case, be it internalized or de novo synthesized immunoglobulin chains, these molecules would
140
MEDICAL HYPOTHESES
CEVTH Fig. 1 Proposed metal-binding domain of human k light chain constant region (residues 194-198). Amino acids in bow represent potential ligands for a divalent metal cation.
CNVNHKPSNTKVDKRVEPKSCDKTH Fig. 2a Proposed zinc finger domain of human Ig G1 heavy chain constant region (residues 200-224). Amino acids in bold represent potential
ligands for a divalentzinc cation. C1 constantregionprecedeshingeregion(underlined).
CDKTHTCPPC Fig. 2b Proposednucleicacid-bindingdomainof human IgG1 hingeregion(residues220-229). Aminoacids in bold representpotential ligandsfor a divalentmetalcation.
PVP S TPPTPSPS
TPP TPSP
SC
Fig. 3 Proposed DNA-binding domain of human IgAl heavy chain constant region (residues 221-241). S/T P repeats are in bold. First amino acid of each complete, non-overlapping S/T P X X-motif in italic. C2 constant region succeeds hinge region (underlined).
a CPRCPEPKSCDTPPPCPRCPEPKSCDTPPPCPRCPEPKSCDTPPPCPRC b CPRCPEPKSCDTPPP c Cys Asp IIe Cys Arg Leu Lys Lys Leu Lys Cys Ser Lys Glu Lys Pro Lys Cys Ala Lys Cys d Cys Leu Val Cys Ser Asp Glu Ala Ser Gly Cys His Tyr Gly Val Leu Thr Cys Gly Ser Cys e Cys Pro Arg Cys Pro Glu Pro Lys Ser
Cys
Asp Thr Pro Pro Pro Cys Pro Arg Cys
Fig. 4 a: Fragment (residues 228-276) from hinge region of human Ig G3; b: Recurring amino acid sequence motif in the hinge region oflg G3; c: AlignedfragmentfromDNA-bindingregionof GAIAprotein(residues 11-31); d: AlignedfragmentfromDNA-bindingregionof humanglucocorticoidreceptor(residues421-441); e: Alignedfragment(residues228-246 or 243-261 or 258-276, respectively)fromhinge regionof human IgG3.
Cysteineresiduesin bold are potentialmetalchelators;TP-tandemsin italic are part of the proposedDNA-bindingmotifTPXX.Figs 4a and b are in one-letteraminoacid codeand Figs4e-eare in three-letteraminoacid code. follow intracellular or intracrine pathways similar to those previously shown or predicted for growth factors and their fragments (15,16). These observations on growth factor fragments have previously led me to put forward a novel concept on cell growth regulation termed the 'subunit hypothesis' (16).
However, an important distinction should be made. Prior to advancing into the nucleus, these antibody domains are proposed to assume conformations conducive to DNA binding through the chelation of zinc ions present in the cytosol. Similar to the dependence of proper folding of GAL4 protein (17,18) and the
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
human glucocorticoid receptor (10), respectively, on the surrounding zinc concentration, it may be that zinc ions fulfill an analogous role in establishing immunoglobulin domain configurations indispensable for their proposed ability to regulate gene expression. Within the nucleus, these antibody subunits may affect both immunoglobulin gene transcription and that of other genes, thus participating directly in the maturation of cells able to synthesize or internalize antibodies. Potential outcomes of these events could be proliferation and differentiation of B-lymphocytes as well as amplification of immunoglobulins of a given specificity in plasma cells. Furthermore, the above Ig domains may be involved, through positive autoregulatory circuits, in conferring growth autonomy to certain types of lymphoid neoplasias associated with increased production of immunoglobulins. Studies undertaken with various lymphoid tumors support these assumptions. As such, it has been shown that the Yheavy chain constant locus maps to a region of translocations in malignant B-lymphocytes (19) and that the expression of an c~l heavy chain-like protein correlates with the clinical picture of a lymphoma (20). Moreover, investigations into the causes of heavy chain disease, a plasma cell disorder (21), have extensively shown the presence of heavy chain fragments-of which the "/3 heavy chain (including the 73 hinge region) is the best studied (22) - in both serum (23,24) and lymphoid cells (25) taken from patients afflicted by this disease. These findings could also be reproduced in a mouse myeloma cell line (26). In the light of the present study, it appears warranted to re-evaluate the potential involvement of these Ig fragments in the pathogenesis of human B-cell neoplasia. Finally, it has been demonstrated that human myeloma cells generate hybrid Ig molecules, e.g. hybrid antibodies of the types IgG4-IgG2 (27) and IgG3-IgG1 (28), whose role in the maturation of the lymphoid lineage should be further investigated along the lines suggested above. If it holds true that immunoglobulins and/or above specific subunits act as transcription factors, they should also influence the development of other organs besides the immune system. For instance, these molecules may be a major factor in the organogenesis and physiology of the brain. This is conceivable in the light of the known interdependence between immunocytes and nerve cells (29). Particularly noteworthy in this context is that pluripotential hemopoietic stem cells are present in adult brain (30) and cytokines affecting Ig synthesis regulate neuronal differentia-
141
tion (31). An additional clue in favor of the above hypothesis may be the fact that many cell surface molecules found in the nervous system and implicated in its ontogeny are members of the immunoglobulin gene superfamily (32). The antibody domains described above may not only autoregulate immunoglobulin gene expression through specific DNA and/or RNA binding, but also participate in trans-activation of DNA-binding proteins through both binding of these proteins and direct recognition of nucleic acids inside or outside the gene response element specific for the respective DNAbinding protein. Interestingly, a recent report showed that specific antibodies to the yeast transcriptional activator HAP1 can induce DNA binding of certain short amino-terminal fragments of the protein (33). The authors interpreted their findings as being due to antibody-promoted dimerization of these fragments. In the light of the present findings, however, the possibility remains that these antibodies may also have physically associated with DNA. At this point, it is tempting to reconsider Padlan's query: 'One wonders also whether the variation in the structure of the hinge may have a bearing on the respective roles that the different antibody types play in the overall immune response' (6), proposed in reference to hinge-mediated antibody reach and rotational adaptability. Based on the data of this report, the scope of the above question would extend to whether the different hinge regions including the neighboring residues of the constant regions may also confer distinct DNA-binding specificities important for gene regulation during the development of the immune system and of other organs. Previous experimental evidence on zinc chelation and nucleic acid binding by immunoglobulins The present study primarily suggests metal and nucleic acid binding of immunoglobulin chains and/or chain fragments which, as such, do not possess disulfide bonds and thus can employ their cysteines as metal chelators. And yet, although the same cysteines are part of disulfide bridges within the completely assembled immunoglobulin molecule and thus are unlikely to be free for metal coordination, serum immunoglobulins have been shown to bind zinc. As such, previous studies have demonstrated that zinc ions bind serum immunoglobulins, particularly IgG, in vitro and in vivo (34) and that zinc sulfate is efficiently used to precipitate immunoglobulins, thus being a useful agent in determining serum immunoglobulin concentration as assessed by the zinc sulfate turbidity test (35,36). Re-
142 markably, this test could even be efficiently applied to distinguish between different IgG subclasses derived from human myelomas (37). In line with my data are also past studies into the nature of translational control of immunoglobulin synthesis. As such, it has been shown that myeloma antibody is able to repress heavy chain synthesis in cultured myeloma cells (38). The same myeloma protein has also been demonstrated to interact with the mRNA coding for the Ig heavy chain (39) in a cell-free system. In vivo, it is likely that it is noncovalently assembled Ig (40) that associates with RNA, since this interaction can only occur in the reducing cytosolic compartment. The noncovalent forces potentially holding together the Ig molecules in the cytoplasm may be conferred by salt bridges between Ig domains (41) and/or ~3-sheet packings of the intercalating (nonaligned) type as shown for Ig fragments derived from Ig constant domains (42). In this context, it is notable that Ig chains have been found in the cytoplasm of cells extracted from so-called non-secretory human myeloma (43) and from a mouse myeloma (44).
A proposed dual role for antibodies: the fully assembled molecule vs the fragment Most likely, serum IgG binds zinc through residues other than the chemically unavailable cysteines, e.g. through histidines and glutamates. As such, immunoglobulins appear to resemble metalloproteins such as matrix metalloproteinases (MMPs). These enzymes have been shown to complex zinc in two forms: one form in which a cysteine residue joins other residues such as histidines in metal chelation and another form in which the same cysteine residue is n o t involved in metal coordination and instead participates e.g. in the formation of a disulfide bridge. In the latter case, it is proposed that water substitutes for the cysteine residue whereas the histidines involved in zinc coordination in the first form are maintained in a chelating configuration in the second form, too (45-47). Accordingly, the proposed metal-binding domains in immunoglobulins may serve a dual function. As part of the completely assembled (serum) immunoglobulins they mediate the transport and neutralization of zinc ions or other divalent cations in the blood circulation - a rather unspecific function with participation of histidines, but not cysteines. As part of a disulfide bond-free antibody chain or chain fragment comprising constant domain sequences, they may chelate zinc in the reducing cytosolic compartment and acquire conformations conducive to an interaction with distinct DNA and/or
MEDICAL HYPOTHESES
RNA regions - a rather specific function that is proposed to be carried out by clusters of cysteines and histidines forming zinc finger-like motifs. This view is also supported by a recent review on zinc finger motifs (48). Beyond the clear distinction between reducing intracellular conditions favoring protein stabilization by zinc chelation and oxidizing extracellular conditions favoring protein stabilization by disulfide bridge formation, these authors point out that some extracellular proteins employ disulfide bonds to stabilize topologies that are similar to intracellular zinc-stabilized motifs.
Future possibilities of verification of the present proposal on metal and nucleic acid binding by antibody fragments A first step in the verification of the present hypothesis immediately suggests itself. One could express any of the antibody fragments proposed above to complex zinc by recombinant DNA technology and test its binding to zinc. This type of approach in which a protein fragment proposed to be a metal-binding sequence is tested for its metal-binding properties has been validated extensively (49-52). Interestingly, the identification of metal-binding domains in retroviral proteins by sequence analysis correctly predicted the direct involvement of these regions in viral growth inside the host cell as a result of metal coordination and complex formation with retroviral RNA and paved the way for further explorations that ultimately showed that also mature retroviral particles assembled from these subunits bind zinc, thus providing a double rationale for antiviral therapy by means of zinc chelation (53,54).
Evolutionary implications Evolutionarily speaking, immunoglobulins and transcription factors may have diverged from a common ancestor molecule. In analogy to the recent unravelling of a short track from the membrane to the nucleus involving a novel group of cytoplasmic transcription factors termed STATs (an acronym for signal transducers and activators of transcription) which directly transduce the signal generated by binding of the immunologically important interferons cc or T to their cell membrane receptors (55), immunoglobulins may be henceforth perceived as CMR-STATs, i.e. cell membrane receptors, signal transducers and activators of transcription. This proposed versatility would explain in an intriguing manner why the structure of an antibody has been selected as a central molecule for defending the organism. The concept of a (soluble or anchored) membrane protein turned transcription factor as suggested herein
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
to apply to antibodies has already been validated with sterol regulatory element-binding protein 1 (SREBP1) that is attached to the nuclear envelope and endoplasmic reticulum (56). Interestingly, it is a fragment cleaved from SREBP-1, not SREBP-1 itself, that translocates to the nucleus where it regulates transcription (56).
Conclusions and perspectives Taken together, this communication provides a putative structural basis for the known requirement of divalent metal cations, particularly of Zn 2+, for a normal immune response (57,58) and warrants further investigations, both theoretical and experimental, into the potential of antibody constant regions for metal binding and gene regulation. In the same way as the previous analysis of the amino acid sequence around the cysteine residues in Ig variable and constant regions provided clues for understanding the generation of antibody diversity (59), the present identification of putative metal binding domain- and zinc finger-like motifs in Ig constant regions embracing yet again cysteine residues should accelerate insight into further fundamental immunological events as well as add to our understanding of more general aspects of cell growth and differentiation. Conversely, future testing of the proposed zinc finger peptides from Ig constant domains should yield information relevant to zinc finger design with potentially wide applications in research and clinical medicine. This avenue has already been envisaged for other zinc finger peptides (60) and predicted to encompass a large combinatorial repertoire as stated before (61): 'Clearly, variations in the finger structure itself or in the length or characteristics of the linkers between fingers may allow variations in the types of interaction between finger proteins and nucleic acids. These variant structures may be crucial for the nucleic acid recognition process.' Finally, my present findings on a possible autoamplificatory role of antibody subunits support my previously advanced proposal (62) according to which it would suffice to hook up a DNA-binding domain to a ligand-binding domain to create a potentially efficacious pharmaceutical agent termed 'synthetic inducible biological response amplifier' (SIBRA) since the results described here suggest that antibodies, similar to SIBRAs, harbor putative gene regulatory domains (in the constant region) linked to a ligandbinding domain (in the variable region). In the reverse analogy, this report supports the notion that SIBRAs may indeed be regarded as an artificial acceleration
143
and refinement of principles intrinsic to the immune system (62).
Note added in proof Consistent with the findings and implications of the present study, two recent reports have experimentally shown that immunoglobulin (Ig) heavy and light chains promote B-cell development in a differential manner (63,64). These investigations confirm and extend the results of a previous study implicating Ig proteins in the control of immunoglobulin gene rearrangement during B-cell differentiation (65). The present proposal according to which Ig chains may partly achieve growth regulatory effects through their potential zinc finger-like motifs also appears likely in the light of the recent identification of a novel zinc finger protein that is able to drive the maturation of B-lymphocytes into Ig-producing plasma cells (66).
Acknowledgements I would like to thank Drs L. Mario Amzel and Jeremy M. Berg for encouragement and critical comments.
References I. TonegawaS. Somatic generation of antibody diversity. Nature 1983; 302: 575-581. 2. Zhou P, Thiele D J. Rapid transcriptional autoregulation of a yeast metalloregulatory transcription factor is essential for high-level copper detoxification. Genes Dev 1993; 7: 1824-1835. 3. Berg J M. Potential metal-binding domains in nucleic acid binding proteins. Science 1986; 232: 485-487. 4. Berg J M. Zinc fingers and other metal-binding domains. J Biol Chem 1990; 265: 6513~5516. 5. Day E D. Advanced lmmunochemistry2nd edn. New York: Wiley-Liss, 1990: 9~65. 6. PadlanE A. Anatomyof the antibody molecule. Mol Immunol 1994; 31: 169-217. 7. Suzuki M. SPXX: a frequent sequence motif in gene regulatory proteins. J Mol Biol 1989; 207: 61-84. 8. Johnston M, Dover J. Mutations that inactivate a yeast transcriptional regulatory protein cluster in an evolutionarilyconserved DNA-bindingdomain. Proc Natl Acad Sci USA 1987; 84: 2401-2405. 9. Pan T, Coleman J E. Structure and function of the Zn(II) binding site within the DNA-binding domain of the GAL4 transcription factor. Proc Natl Acad Sci USA 1989; 86: 31453149. 10. Freedman L P, Luisi B F, Korszun Z R, Basavappa R, Sigler P B, Yamamoto K R. The function and structure of the metal coordination sites within the glucocorticoid receptor DNAbinding domain. Nature 1988; 334: 543-546. 11. Omichinski J G, Clore G M, Schaad Oet al. NMR structure of a specific DNA complex of Zn-containing DNA binding domain of GATA-1. Science 1993; 261: 438-446. 12. Qian X, Jeon C, YoonH, Agarwal K, Weiss M A. Structure of
144
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26. 27. 28. 29. 30. 31.
32.
33.
a new nucleic-acid-binding motif in eukaryotic transcriptional elongation factor TFIIS. Nature 1993; 365: 277-279. Mitchell R N, Shaw A C, Weaver Y K, Leder E Abbas A K. Cytoplasmic tail deletion converts membrane immunoglobulin to a phosphatidylinositol-linked form lacking signaling and efficient antigen internalization functions. J Biol Chem 1991; 266: 8856-8860. Maric M A, Taylor M D, Blum J S. Endosomal aspartic proteinases are required for invariant chain processing. Proc Natl Acad Sci USA 1994; 91: 2171-2175. Radnlescu R T, Wendtner C M. Proposed interaction between insulin and retinoblastoma protein. J Mol Recognit 1992; 5: 133-137. Radulescu R T, Wendtner C M. Hormone and growth factor subunits: a novel perception of cell growth regulation. J Endocrinol 1993; 139: 1-7. Johnston M. Genetic evidence that zinc is an essential cofactor in the DNA-binding domain of GAL4 protein. Nature 1987; 328: 353-355. Pan T, Coleman J E. GAL4 transcription factor is not a 'zinc finger' but forms a Zn(II)2Cys 6 binuclear cluster. Proc Natl Acad Sci USA 1990; 87: 2077-2081. Kirsch 1 R, Morton C C, Nakahara K, Leder E Human immunoglobulin heavy chain genes map to a region of translocations in malignant B lymphocytes. Science 1982; 216: 301303. Seligmann M, Danon E Hurez D, Mihaesco E, Preud'homme J-L. Alpha-chain disease: a new immunoglobulin abnormality. Science 1968; 162: 1396-1397. Seligmann M, Mihaesco E, Preud'homme J-L, Danon F, Brouet J-C. Heavy chain diseases: current findings and concepts. Immunol Rev 1979; 48: 145-167. Frangione B, Franklin E C. Split immunoglobulin genes and human heavy chain deletion-mutants. J Immunol 1979; 122: 1177-1179. Adlersberg J B, Franklin E C, Frangione B. Repetitive hinge region sequences in human IgG3: Isolation of an ll,000Dalton fragment. Proc Natl Acad Sci USA 1975; 72: 723-727. Michaelsen R E, Frangione B, Franklin E C. Primary structure of the 'hinge' region of human IgG3-probable quadruplication of a 15-amino acid residue basic unit. J Biol Chem 1977; 252: 883-889. Buxbaum J N, Preud'homme J-L. Alpha and gamma heavy chain diseases in man: intracellular origin of the aberrant polypeptides. J Immunol 1972; 109: 1131-1137. Morrison S L, Scharff M D. Heavy chain-producing variants of a mouse myeloma cell line. J Immunol 1975; 114: 655-659. Natvig J B, Kunkel H G. A hybrid IgG4-1gG2 immunoglobulin. J Immunol 1974; ll2: 1277-1284. Kunkel H G, Natvig J B, Joslin F G. A 'Lepore' type of hybrid y globulin. Proc Natl Acad Sci USA 1969; 62: 144-149. Reichlin S. Neuroendocrine-immune interactions. New Engl J Med 1993; 329: 1246-1253. Bartlett P E Pluripotential hemopoietic stem cells in adult mouse brain. Proc Natl Acad Sci USA 1982; 79: 2722-2725. Mehler M E Rozental R, Dougherty M, Spray D C, Kessler J A. Cytokine regulation of neuronal differentiation of hippocampal progenitor cells. Nature 1993; 362: 62-65. Steinman L. Connections between the immune system and the nervous system. Proc Natl Acad Sci USA 1993; 90: 79127914. Zhang L, Bermingham-McDonogh O, Turcotte B, Guarente L. Antibody-promoted dimerization bypasses the regulation of DNA binding by the heme domain of the yeast transcriptional activator HAPI. Proc Natl Acad Sci USA 1993; 90: 28512855.
MEDICAL HYPOTHESES
34. Prasad A S, Oberleas D. Binding of zinc to amino acids and serum proteins in vitro. J Lab Clin Med 1970; 76: 416--425. 35. McEwan A D, Fisher E W, Selman I E, Penhale W J. A turbidity test for the estimation of immune globulin levels in neonatal calf serum. Clin Chem Acta 1970; 27: 155-163. 36. Caldow G L, Edwards S, Peters A R, Nixon P, lbata G, Sayers R. Associations between viral infections and respiratory disease in artificially reared calves. Vet Rec 1993; 133: 85-89. 37. Miyagawa N, Okamoto Y, Morita K, Dohi K. Studies on the relationship between serum colloidal reactions (ZTT and TTT) and IgG subclasses, especially lgGl and IgG2. Microbiol lmmunol 1991; 35: 59-66. 38. Stevens R H, Williamson A R. Translational control of immunoglobulin synthesis. I. Repression of heavy chain synthesis. J Mol Biol 1973a; 78: 505-516. 39. Stevens R H, Williamson A R. Translational control of immunoglobulin synthesis. 11. Cell-free interaction of myeloma immunoglobulin with mRNA. J Mol Biol 1973b; 78: 517-525. 40. Williamson A R. Biosynthesis of immunoglobulins. In: Defense and Recognition. R.R. Porter, ed. London: Butterworths and Baltimore: University Park Press, 1973; 10: 229-251. 41. Schiffer M, Chang C-H, Naik V M, Stevens F J. Analysis of immunoglobulin domain interactions. Evidence for a dominant role of salt bridges. J Mol Biol 1988; 203: 799-802. 42. Miller S. Protein-protein recognition and the association of immunoglobulin constant domains. J Mol Biol 1990; 216: 965-973. 43. Hurez D, Preud'homme J-L, Seligmann M. Iotracellular 'monoclonal' immunoglobulin in non-secretory human myeloma. J Immunol 1970; 104: 263-264. 44. Morrison S L. Sequentially derived mutants of the constant region of the heavy chain of murine immunoglobulins. J Immunol 1979; 123: 793-800. 45. Springman E B, Angleton E L, Birkedal-Hansen H, Van Wart H E. Multiple modes of activation of latent human fibroblast collagenase: evidence for the role of a Cys 73 active-site zinc complex in latency and a 'cysteine switch' mechanism for activation. Proc Natl Acad Sci USA 1990; 87: 364-368. 46. Van Wart H E, Birkedal-Hansen H. The cysteine switch: a principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family. Proc Natl Acad Sci USA 1990; 87: 5578-5582. 47. Hirose T, Patterson C, Pourmotabbed T, Mainardi C L, Hasty K A. Structure-function relationship of human neutrophil collagenase: identification of regions responsible for substrate specificity and general proteinase activity. Proc Natl Acad Sci USA 1993; 90: 2569-2573. 48. Schwabe J W R, Klug A. Zinc mining for protein domains. Nature Struct Biol 1994; 1: 345-349. 49. Green L M, Berg J M. A retroviral Cys-Xaa2-Cys-Xaa4-HisXaa4-Cys peptide binds metal ions: spectroscopic studies and a proposed three-dimensional structure. Proc Natl Acad Sci USA 1989; 86: 4047-4051. 50. Summers M F, South T L, Kim B, Hare D R. High-resolution structure of an HIV zinc fingerlike domain via a new NMRbased distance geometry approach. Biochemistry 1990; 29: 329-340. 51. Michael S F, Kilfoil V J, Schmidt M H, Amann B T, Berg J M. Metal binding and folding of a minimalist Cys2His2 zinc finger peptide. Proc Natl Acad Sci USA 1992; 89: 4796-4800. 52. Berg J M. Zinc-finger proteins. Curr Opin Struct Biol 1993; 3: 11-16. 53. Summers M F, Henderson L E, Chance M R et al. Nucleocapsid zinc fingers detected in retroviruses: EXAFS studies of intact viruses and the solution-state structure of the nudeocapsid protein from H1V-I. Protein Sci 1992; 1: 563-574.
ANTIBODY CONSTANT REGION: POTENTIAL TO BIND METAL AND NUCLEIC ACID
54. Rice W G, Schaeffer C A, Harten B e t al. Inhibition of HIV1 infectivity by zinc-ejecting aromatic C-nitroso compounds. Nature 1993; 361: 473-475. 55. Shuai K, Horvath C M, Tsai Huang L H, Qureshi S A, Cowburn D, Damell Jr J E. Interferon activation of the transcription factor Stat91 involves dimerization through SH2phosphotyrosyl peptide interactions. Cell 1994; 76: 821-828. 56. Wang X, Sato R, Brown M S, Hua X, Goldstein J L. SREBP1: a membrane-bound transcription factor released by sterolregulated proteolysis. Cell 1994; 77: 53-62. 57. Chandra R K. Trace element regulation of immunity and infection. J Am Coil Nutr 1985; 4: 5-16. 58. Vallee B L, Falchuk K H. The biochemical basis of zinc physiology. Physiol Rev 1993; 73: 79-118. 59. Milstein C. From antibody structure to immunological diversification of immune response. Science 1986; 23l: 1261-1268. 60. Rebar E J, Pabo C O. Zinc finger phage: affinity selection of fingers with new DNA-binding specificities. Science 1994; 263: 671-673. 61. Berg J M. Proposed structure for the zinc-binding domains
62.
63. 64.
65.
66.
145
from transcription factor IliA and related proteins. Proc Natl Acad Sci USA 1988; 85: 99-102. Radulescu R T. Synthetic inducible biological response amplifiers (SIBRAs): rational peptides at the crossroads between molecular evolution and structure-based drug design. Med Hypotheses 1995; 44: 32-38. Spanopoulou E, Roman C A J, Corcoran L Met al. Functional immunoglobulin transgenes guide ordered B-cell differentiation in Rag-l-deficient mice. Genes Dev 1994; 8: I030-1042. Young F, Ardman B, Shinkai Y e t al. Influence of immunoglobulin heavy-chain and light-chain expression on Bcell differentiation. Genes Dev 1994; 8: I043-1057. Iglesias A, Kopf M, Williams G S, Biihler B, K6hler G. Molecular requirements for the ~-induced light chain gene rearrangement in pre-B cells. EMBO J 1991; 10: 2147-2156. Turner Jr C A, Mack D H, Davis M M, Blimp-l, a novel zinc finger-containing protein that can drive the maturation of B lymphocytes into immunoglobulin-secreting cells. Cell 1994; 77: 297-306.