Abstract
Systemic lupus erythematosus (SLE), the prototypic autoimmune disease, is characterized by defective expression of TCR ζ-chain. Elf-1 (E-74-like factor) is a member of the Ets (E-26-specific) family and is crucial for the basal transcription of TCR ζ-chain in Jurkat cells. We previously demonstrated that Elf-1 exists in the cytoplasm mainly as 80-kDa form and after phosphorylation and O-glycosylation it moves to the nucleus as a 98-kDa which binds DNA. We now demonstrate that Elf-1 is crucial for the transactivation of TCR ζ-chain promoter in normal and SLE T cells. Defective expression of TCR ζ-chain in SLE T cells is associated with two distinct molecular defects in the generation of the 98-kDa DNA binding Elf-1 form. In the first, the levels of the 98-kDa form were either decreased or absent. In the second, the apparent levels of the nuclear Elf-1 form were normal but included only two of the three bands into which the nuclear Elf-1 form separated in isoelectric focusing gels. Because both the transcription and the translation processes of Elf-1 gene are normal in SLE T cells, our data demonstrate that abnormal posttranslational mechanisms of the Elf-1 protein result in defective expression of functional Elf-1, and consequently, the transcriptional defect of TCR ζ-chain in patients of SLE.
Systemic lupus erythematosus (SLE)3 is an autoimmune disease of undetermined etiology. Molecular and biochemical studies in recent years have identified a number of abnormalities which manifest specifically in the patients with SLE but not in the normal controls or patients with other systemic immune disorders (1). We previously found that patients with SLE display defective expression of TCR ζ-chain protein and mRNA (2). TCR ζ-chain is a component of the CD3 complex and is unique in that it contains more (three) phosphorylation motifs than other CD3 components, and thus is considered critical in relaying extracellular signals inwardly (3). This defect was further found to correlate with the abnormal phosphorylation of the signaling proteins (2), enhanced and sustained Ca2+ concentration (4), and consequent aberrant transcription of downstream genes observed in the T cells from patients with SLE (1). Interestingly, defective expression of the TCR ζ-chain protein rather than mRNA has been observed in T cells from patients with systemic infections (5), melanoma (6), and gastric tumor (7). In the latter, increased caspase activity has been shown to degrade TCR ζ-chain (8).
The transcription of TCR ζ-chain gene has been studied in JurkatT cells and it has been found to depend heavily on the presence of the transcription factor Elf-1. Mutation of both Elf-1 sites on the TCR ζ-chain promoter decreases significantly the activity of TCR ζ-chain promoter (9). Elf-1 is a member of Ets family transcription factors which also includes factors such as Ets-1, Ets-2, and Fli-1. Members of this family of transcription factors are actively involved in the transcriptional regulation of genes such as CD4 (10), IL-2R α-chain (11), and GM-CSF (12). Elf-1 gene encodes an open reading frame of 619 aa, which should correspond to a calculated molecular mass of 68 kDa. However, we found recently that Elf-1 exists mainly as an 80-kDa form in the cytosolic fraction of human T lymphocytes, while in the nuclear fraction a 98-kDa Elf-1 is the prevalent form. Phosphorylation and O-linked glycosylation contributes to the discrepancy of the apparent molecular mass between these two forms (13). Previous studies have indicated that promoter-defined Elf-1 sites from different genes bind preferentially different forms of Elf-1. The human IL-2 promoter-defined Elf-1 site binds all forms of Elf-1 (14), whereas the enhancer of the Moloney murine leukemia virus binds the 98-kDa form of Elf-1 (15).
We have reported previously that patients with SLE display aberrant expression of several transcription factors including NF-κB (16) and cAMP response element modulator (17). These findings have indicated that transcription factor expression and function is susceptible to adverse regulation in T cells from SLE patients. Abnormal expression of transcription factors should affect the transcription of respective target genes and contribute to the pathogenesis of SLE. Because Elf-1 plays a critical role in the transcription of the TCR ζ-chain and other immune cell genes (9, 14, 15), we conducted studies to determine the expression of Elf-1 in patients with SLE. We demonstrate that nuclear proteins from SLE T cells fail to bind to the TCR ζ promoter-defined Elf-1 sites either because the 98-kDa form of Elf-1 is not formed, or because the formed 98-kDa Elf-1 lacks DNA-binding features.
Materials and Methods
Patients
Thirty-one patients fulfilling at least 4 of the 11 revised criteria of the American College of Rheumatology for the classification of SLE (18) were studied. Twenty-seven of the patients were females and four were males with age range 21–80 years old (mean ± SD, 34 ± 11). Disease activity was measured using the SLE disease activity index (SLEDAI) score (mean ± SD, 7 ± 6.5). Patients were on various medications which included prednisone, hydroxychloroquine, azathioprine, and cyclophosphamide. Medications were discontinued 24 h before venipuncture. Fifteen normal volunteers served as normal controls.
Plasmids and reagents
The luciferase reporter constructs driven by either wild-type (−307/+58) or mutant (FLΔzEBS1+2, −307/+58 with both Elf-1 sites mutated) TCR ζ promoter have been described (9). Abs against Elf-1, NF-κB p50, TCR ζ-chain, and actin were purchased from Santa Cruz Biotechnology, (Santa Cruz, CA).
Purification of primary T cells and transfection
Primary T cells were obtained either by using magnetic beads to remove all non-T cells (negative selection) as described before (17, 19), or by incubating peripheral blood with rosette/Ab complexes at room temperature for 20 min (StemCell Technologies, Vancouver, British Columbia, Canada). The T cells were then separated from the non-T cells by centrifugation at 2,400 rpm for 25 min. The harvested T cells were washed and contaminated RBCs lysed by ACK buffer (155 mM NH4Cl, 110 mM KHCO3, 0.1 mM EDTA) if necessary. T cells were transfected with the indicated plasmids using previously developed protocols (17, 19). Briefly, freshly purified T cells were incubated in medium containing 10% FCS and PHA (1 μg/ml) overnight. Cells were then mixed thoroughly with the plasmids indicated and subjected to electroporation at 0.25 V and 950 μF by using the electroporator (Bio-Rad, Hercules, CA). Cells were harvested ∼16 h later and assayed for both luciferase and β-galactosidase activities.
EMSA
EMSA was performed as described previously (18). The sequence of the oligonucleotides used for the DNA binding reaction defined by the −147/−119 region of the TCR ζ promoter was 5′-TCGAGAACCTCCAGGGCTTCCTGCCTGTGAACCA-3′.
Protein purification and Western blotting
The purification of cytoplasmic and nuclear proteins has been described before (17). Western blotting was performed following the manufacturer’s protocol (Amersham Pharmacia Biotech, Piscataway, NJ).
Preparation of mRNA, cDNA, PCR, and real-time PCR
One million T cells were used to extract RNA (RNA preparation kit; Qiagen, Hilden, Germany). RNA was quantified and 500 ng of total RNA was used for the preparation of cDNA by reverse transcription (Reverse Transcription kit; Promega, Madison, WI). A total of 50 ng of cDNA was used for each PCR. Primers were synthesized by Sigma-Aldrich (St. Louis, MO). PCR beads were used for amplification (Amersham Pharmacia Biotech). PCR products were separated on a 1.5% agarose gel and the OD was analyzed using QuantityOne software (Bio-Rad) after background subtraction from each band. The sequence of primers used for PCR amplification are: 5′-GCGAGGGGGCAGGGCCTGCATGTGAAG-3′, 5′-AGCCTCTGCCTCCCAGCCTCTTTCTGAG-3′ (TCR ζ-chain); 5′-CGGGGATGAAACAATTGAAACTAT-3′, 5′-CTCGCTGGGTCCACTNTGATGTA-3′ (Elf-1); and 5′-CATGGGTCAGAAGGATTCCT-3′, 5′- AGCTGGTAGCTCTTCTCCA-3′ (actin).
Chromatin immunoprecipitation analysis (ChIP)
T cells were treated by formalin fixation (1% final concentration) for 10 min, followed by extensive washing. Cells were then lysed and sonicated. The DNA-protein complexes were immunoprecipitated using the indicated Ab and extracted by agarose Sepharose beads (Santa Cruz Biotechnology). After several washing steps, the cross-link between DNA and protein was reversed by incubating the samples at 65°C for 4 h and the protein was digested by protease K and the DNA was extracted (Qiagen DNA extraction kit; Qiagen). The DNA was amplified with primers flanking the TCR ζ-chain promoter. DNA from ∼1 million cells was used per each PCR. PCR products were run on a 2% agarose gel and quantified with QuantityOne software. The sequence of the primers used for amplifying the promoters are 5′-ACCTCGAGCCATCGAGAACT-3′, 5′-ACAAGCTTCTTTCCCTCAGA-3′ (TCR ζ chain); and 5′-CTA AGT GTG GGC TAA TGT AAC-3′, 5′-TGT AAA ACT GTG GGG GT-3′ (IL-2).
Isoelectric focusing gel (IEF) gel electrophoresis and Western blotting
Nuclear proteins were resolved in the IEF (Invitrogen, Carlsbad, CA) gel at 100 V for 1 h, 200 V for 1 h, and finally at 500 V for 30 min. The electrophoresed proteins on the gel were first equilibrated in the 0.7% acetic acid buffer at room temperature for 10 min and transferred to polyvinylidene fluoride membranes in the same buffer by applying 10 V for 1 h. The membrane was then air dried before blotting with Ab against Elf-1 as described above.
Results
Elf-1 sites of the TCR ζ promoter are necessary for the TCR ζ-chain gene expression in normal and SLE T cells
It was previously shown that Elf-1 is the transcription factor which binds to two Elf-1 sites on the TCR ζ-chain promoter in Jurkat T cells. Mutagenesis of both Elf-1 sites down-regulated the transcriptional activity of the promoter, whereas overexpression of Elf-1 increased the activity of the promoter by three to four times in COS-7 cells (9). To determine whether Elf-1 is similarly required for the transcription of TCR ζ-chain in primary human T cells, we transfected luciferase reporter constructs driven by either wild-type TCR ζ-chain promoter (−307/+58), or by a promoter in which both Elf-1 sites had been mutated (FLΔzEBS1+2) into normal and SLE T cells. A construct which expresses β-galactosidase was always cotransfected into cells and served as the transfection efficiency control. We first transiently transfected these constructs into Jurkat cells and found that the wild-type promoter drives the expression of luciferase four to five times higher than the mutated promoter, an observation which is consistent with those reported previously (9). We next transfected these constructs into T cells from either normal (n = 2) or SLE patients (n = 3). As shown in Fig. 1⇓, the promoter activity of the wild-type TCR ζ-chain promoter was four times higher in normal T cells than in SLE T cells, indicating that the transcription of the TCR ζ-chain gene is defective in the latter. The reporter activity of the constructs driven by the mutated FLΔzEBS1+2 promoter was lower than the activity of constructs driven by the intact promoter, indicating that the Elf-1 sites are important for the transcription of the TCR ζ-chain gene in both normal and SLE T cells. Decreased promoter activity observed in these experiments was not due to differences in the transfection efficiency between the normal and SLE T cells because the expression levels of β-galactosidase were comparable between these two types of cells. Thus, the Elf-1-mediated transcriptional defect is at least in part responsible for the defective expression of the TCR ζ-chain mRNA in patients with SLE that we reported previously (2).
Defective Elf-1-mediated transcription of TCR ζ-chain in patients with SLE. Primary T cells were cotransfected with either wild-type or mutant TCR ζ-chain-driven luciferase reporter and β-galactosidase. Cells were harvested 20–24 h posttransfection. The results from luciferase assay were further normalized by the β-galactosidase activity. The average of β-galactosidase-normalized luciferase activities from two independent transfection experiments is shown.
The expression of the nuclear 98-kDa form of Elf-1 is defective in lymphocytes from patients with SLE
Because we have failed to demonstrate any mutations in the TCR ζ-chain promoter in patients with SLE (20), we considered that the decreased transcriptional activity in SLE T cells should be due to abnormal expression and/or activation of Elf-1. Accordingly, we investigated the expression pattern of Elf-1 in patients of SLE. We purified nuclear proteins from normal and SLE T cells and analyzed them by SDS gels electrophoresis and Western blotting. As shown in Fig. 2⇓a, comparing to the normal controls, SLE patients (8 of 31) display defective expression of the 98-kDa form of Elf-1 protein. In contrast, the levels of the NF-κB p50 (16) and the 80-kDa Elf-1 (Fig. 2⇓b), which is enriched in the cytoplasm (13), are similar among SLE patients and normal controls (Fig. 2⇓b). To establish whether global protein degradation was responsible for the decreased expression of Elf-1 in SLE T cells, we subjected nuclear proteins to SDS gel electrophoresis and staining with Coomassie blue (Fig. 2⇓c). We found that while each sample varies slightly in the intensity of individual bands, there was no extensive protein loss in SLE patients. To this end, we compared the total nuclear protein expression levels between normal and SLE T cells by densitometry and calculated the average intensity of five different bands representing proteins of various molecular mass (Fig. 2⇓c, 2 days). We observed that proteins from SLE T cells are decreased by 30–40% in bands a and e, while they have comparable levels in bands b–d, and NF-κB p50. It should be noted that the intensity of band c, which represents proteins with molecular mass close to the 98-kDa Elf-1, is comparable between normal and lupus T cells. These data indicate that global protein degradation does not play a significant role in the decreased expression of the 98-kDa Elf-1 in SLE patients, and that the loss of 98-kDa Elf-1 involves additional mechanisms.
SLE patients display defective expression of the 98-kDa form of Elf-1. a, A total of 5 μg of nuclear proteins were separated on 10% SDS gel followed by Western blot analysis using Ab against Elf-1. The blot was then stripped and blotted with anti-NF-κB p50 Ab. b, The cytoplasmic proteins from both normal and lupus patients were separated in SDS gel followed by Western blot with anti-Elf-1 Ab. c, The nuclear proteins from the same patients as shown in a were separated in 10% SDS gel followed by staining with Coomassie blue dye. d, The y-axis represents the ratio of the intensity of the bands of SLE patients and normal controls (c, bands a–e) and the ratio of the Western blot band intensity for Elf-1 and NF-κBp50 (a).
Despite the fact that the 80-kDa form of Elf-1 was expressed in SLE T cells in amounts comparable to those of normal T cells (Fig. 2⇑b), we performed RT-PCR to rule out additional defects in the transcriptional expression of the Elf-1 gene. As shown in Fig. 3⇓, the levels of Elf-1 mRNA are comparable in SLE (n = 18) and normal T cells (n = 10).
Elf-1 mRNA is not decreased in SLE T cells. a, Total RNA from the same patients as in Fig. 2⇑ were purified and assayed by RT-PCR with primers which specifically amplify Elf-1 and actin. b, Densitometric readings of the Elf-1 mRNA levels from normal and SLE T cells.
Defective formation of the 98-kDa form of Elf-1 is associated with decreased expression of TCR ζ-chain in patients of SLE
To determine whether the decreased expression of the 98-kDa form of Elf-1 is associated with decreased DNA binding, we incubated nuclear proteins from normal (n = 15) and SLE (n = 31) T cells with an oligonucleotide which encodes the Elf-1-binding sites on the TCR ζ-chain promoter. As shown in Fig. 4⇓b, low levels of nuclear 98-kDa Elf-1 in SLE patients were associated with decreased intensity of the band that reflected the binding of Elf-1 to DNA. These experiments confirmed that the loss of the 98-kDa Elf-1 is specifically associated with the loss of binding to the TCR ζ-chain promoter. We then performed Western blots to determine whether low DNA-binding activity is associated with decreased TCR ζ-chain expression. As shown in Fig. 4⇓c, loss or low binding of SLE T cell nuclear proteins to the Elf-1 site of the TCR ζ-chain promoter is associated with decreased expression of TCR ζ-chain.
Defective expression of the 98-kDa form of Elf-1 is associated with decreased expression of TCR ζ-chain. a, Nuclear proteins from either normal or SLE patients were purified and analyzed by SDS gel electrophoresis followed by Western blot analysis using Ab against Elf-1. b, The nuclear proteins were also incubated with the an oligonucleotide encoding the Elf-1 binding site on the TCR ζ-chain promoter followed by EMSA. When indicated, the nuclear proteins were incubated with the anti-Elf-1 Ab for 10 min before incubating with the probe. c, The cytoplasmic proteins from the same patients were then analyzed in 12% SDS gel followed by Western blot with Ab against TCR ζ-chain. The membrane was then stripped and reprobed with an Ab against actin.
Defective DNA binding, despite normal levels of 98-kDa Elf-1 in SLE patients, is also associated with the low expression of the TCR ζ-chain
In the course of these experiments, we observed that a group of SLE patients (14 of 31) displayed low DNA-binding activity despite the fact that they expressed normal levels of the 98-kDa form of Elf-1 (Fig. 5⇓a, comparing lupus patients 2, 3, and 4 to the normal controls). As shown in Fig. 5⇓a, low DNA-binding activity resulted in low expression of TCR ζ-chain. This observation indicated that although the DNA-binding form of Elf-1 resided in the 98-kDa form, in certain patients the final active form was not produced.
Nuclear and whole-cell proteins from SLE T cells display decreased binding to the TCR ζ-chain promoter-defined Elf-1 site. a, upper panel, Nuclear proteins from both normal and SLE patients were first assayed for the expression levels of Elf-1 in Western blots. N, normal; Lupus, SLE T cell cellular extracts. Middle panel, The same samples were then incubated with the Elf-1 binding oligonucleotides followed by EMSA. The cytoplasmic proteins were then analyzed for the expression levels of TCR ζ-chain and actin. b, The total proteins were purified as described in Materials and Methods followed by dialysis to remove the detergents used for the lysis of the cells. Total proteins were then assayed for Elf-1 expression levels, Elf-1 DNA binding activity, and the expression levels of both TCR ζ-chain and actin as described in a.
To determine whether the 80-kDa form of Elf-1 can substitute for the non-DNA-binding 98-kDa Elf-1, we purified whole-cell proteins, which contain more 80-kDa form of Elf-1 from both normal (n = 2) and lupus T cells (n = 3) and performed EMSA using an Elf-1 oligonucleotide. As shown in Fig. 5⇑b, decreased DNA binding was noted in protein extracts from SLE T cells, which express high levels of both 80- and 98-kDa Elf-1, suggesting that the 80-kDa Elf-1 does not bind the TCR ζ-chain promoter. The expression levels of TCR ζ-chain in T cells from this group of patients reflected accurately the intensity of the band reflecting protein binding to the Elf-1 oligonucleotide (Fig. 5⇑b).
Elf-1 does not bind to the TCR ζ-chain promoter in live T cells from patients with SLE
Shift assays determine in vitro protein-DNA interactions and may not reflect in vivo binding. To ascertain that Elf-1 fails to bind to the TCR ζ-chain promoter in live SLE T cells, we subjected T cells to formaldehyde fixing, sonication, and immunoprecipitation of TCR ζ-chain promoter with anti-Elf-1 Ab followed by RT-PCR which specifically amplifies the TCR ζ-chain promoter. We elected to perform these experiments using cells from SLE patients (n = 4) who were shown to have comparable levels of 98-kDa Elf-1 along with T cells from normal donors (n = 3) (Fig. 6⇓a). As shown in Fig. 6⇓b, the SLE samples displayed a significant decrease in the in vivo binding of Elf-1 to the TCR ζ-chain promoter (four of four). Because it is known that Elf-1 binds to the IL-2 promoter (14), we performed ChIP experiments using anti-Elf-1 Ab and PCR primers spanning the −232/−55 region of the IL-2 promoter. As it can be seen in Fig. 6⇓b, Elf-1 binds to the IL-2 promoter at comparable levels between normal and SLE T cells. This experiment serves also as an additional control for the immunoprecipitation and PCR amplification steps of the ChIP assays used above to establish that Elf-1 from SLE T cells fails specifically to bind to the TCR ζ-chain promoter. Furthermore, the same SLE samples displayed decreases in the binding to Elf-1 oligonucleotides in in vitro shift assays (Fig. 6⇓c), confirming the results from the in vivo ChIP assays. The consequence of this defective TCR ζ-chain promoter binding by Elf-1 in this group of SLE patients is reflected by the decreased levels of TCR ζ-chain mRNA, as assayed by RT-PCR on the total RNA collected from the same patients (Fig. 6⇓d). The levels of both Elf-1 and actin mRNA were comparable between normal controls and SLE patients. Altogether, our in vitro and in vivo data indicate that defective DNA binding of Elf-1 to the TCR ζ-chain promoter in SLE T cells is associated with defective transcription of TCR ζ-chain.
Elf-1 does not bind to the TCR ζ-chain promoter in live T cells from patients with SLE. a, The nuclear proteins from normal donors and SLE patients were analyzed by SDS gel electrophoresis followed by Western blot analysis with an Ab against Elf-1. b, PCR amplified products using primers flanking the ζ-chain and IL-2 promoter of immunoprecipitates obtained with an Elf-1-Ab (ChIP). c, Nuclear proteins were incubated with the oligonucleotide encoding the TCR ζ-chain-defined Elf-1-binding site followed by EMSA. d, RT-PCR products using primers which specifically amplify TCR ζ-chain, Elf-1, and actin.
The nuclear Elf-1 comprises multiple subcomponents with distinct isoelectric point values
Previously, we had noted (13) that the nuclear abundant 98-kDa Elf-1 is further separated into three distinct subcomponents in 6% SDS gels. The low binding of SLE T cell nuclear extracts to the TCR ζ-chain promoter, despite normal levels of 98-kDa Elf-1 form, suggested that it is different from the normal 98-kDa Elf-1. To determine the nature of the difference between normal and SLE T cell-expressed 98-kDa Elf-1, we subjected normal T cell nuclear proteins to IEF gel analysis along with nuclear extracts from T cells from SLE patients. The used SLE samples expressed slightly higher levels of 98-kDa Elf-1 but less DNA-binding activity (Fig. 7⇓, a and b). IEF can detect protein modifications which do not contribute significant changes in the molecular size of the protein, and thus cannot be detected by SDS gel electrophoresis. As shown in Fig. 7⇓c, normal T cell nuclear proteins are composed of three subcomponents (five of five), a finding consistent with our previous observation using 6% SDS gel electrophoresis (13). These three bands are widely separated in the IEF gel, indicating that the individual components are the products of distinct posttranslational modifications. Interestingly, lupus patients lacked the subcomponent migrating between the two other major subcomponents (four patients of seven studied). Whether the missing band is the major DNA-binding component needs to be determined.
The nuclear Elf-1 resolves into three subcomponents with distinct isoelectric point values. a, Nuclear proteins from either normal or SLE patient were probed with anti-Elf-1 Ab. b, Nuclear proteins from either normal controls or SLE patients were incubated with the TCR ζ-chain promoter-defined Elf-1-binding site followed by EMSA. c, Nuclear proteins were separated on IEF gels followed by Western blot analysis with an Ab against Elf-1.
Binding of nuclear proteins to the Elf-1 sites of the TCR ζ-chain promoter correlates with the levels of TCR ζ-chain mRNA
To obtain an overall picture of the relationship between nuclear protein DNA binding and the expression of TCR ζ-chain, we plotted the densitometry readings obtained from either shift assays or ChIP against the β-actin-normalized TCR ζ-chain mRNA levels obtained from RT-PCR. As it can be seen in Fig. 8⇓, there is a strong correlation (r = 0.51, p = 0.05) between DNA protein binding and ζ-chain expression, further proving that the defective expression of TCR ζ-chain in SLE T cells is the result of decreased TCR ζ-chain gene transcriptional activity.
Correlation between the TCR ζ-chain mRNA expression levels and the TCR ζ-chain promoter binding activity of Elf-1. The y-axis represents the β-actin-normalized TCR ζ-chain mRNA levels and the x-axis represents the Elf-1/TCR ζ-chain promoter interaction as determined by either EMSA or ChIP.
Discussion
The transcriptional activity of the TCR ζ-chain promoter is decreased in patients with SLE because the levels of TCR ζ-chain mRNA in SLE T cells are decreased (2), and because the activity of a reporter construct driven by the TCR ζ promoter transiently transfected into SLE T (Fig. 1⇑) cells is decreased. In this study, we have demonstrated that the two Elf-1 sites defined by the TCR ζ-chain promoter are critical for the transcription of TCR ζ-chain not only in Jurkat cells (9), but also in primary T cells from both normal individuals and SLE patients. The fact that the TCR ζ-chain promoter of SLE patients does not express any mutations or polymorphisms (20), which could explain defective transcriptional activity, prompted us to search for defects in the expression and/or activation of the Elf-1 transcription factor. After establishing that the Elf-1 message levels in SLE T cells is comparable to those in normal T cells (Fig. 3⇑) and that the main 80-kDa cytoplasmic form of Elf-1 (13) was present at normal levels in SLE T cells (Fig. 2⇑b), we searched for defects in the posttranslational processes that are responsible for the generation of the nuclear 98-kDa form, which binds to the TCR ζ promoter and controls its transcriptional activity.
Elf-1 cDNA encodes a peptide of 619 aa and should migrate in SDS gel as a protein of 68 kDa. Interestingly, Elf-1 in the cytoplasmic protein fraction migrates mainly as an 80-kDa protein, while the dominant form in the nucleus migrates as 98 kDa. We have reported that the 98-kDa form is the form of Elf-1 that binds to the TCR ζ-chain promoter (13). This conclusion was based on the fact that the intensity of the nuclear protein binding to the TCR ζ-chain oligonucleotides correlated with the intensity of the 98-kDa form, but not with that of the 80-kDa form as determined by Western blot analysis (Figs. 4⇑ and 5⇑b). In addition, UV cross-linking experiments using a labeled Elf-1 defined oligonucleotide and nuclear proteins from normal T cell produced a band expected from the binding of the 98- rather than the 80-kDa protein (data not shown).
Two patterns of Elf-1 activation defects were identified in SLE T cells in this study. In the first, the 98-kDa form was absent probably because of defective phosphorylation and O-glycosylation, two processes shown to be responsible for the discrepancy between the predicted and the observed molecular mass of Elf-1 (13). In the second, although the apparent 98-kDa form is generated at normal levels as detected by Western blot analysis ( Figs. 5–7⇑⇑⇑), it failed to bind to the TCR ζ promoter-defined Elf-1 sites. Decreased DNA binding was established in EMSA using TCR ζ-chain-defined Elf-1 oligonucleotides ( Figs. 5–7⇑⇑⇑) and ChIP (Fig. 6⇑), which enables detection of transcription factors in live cells. Because the normal nuclear Elf-1 could be resolved in to three components in IEF gels, and the SLE nuclear 98-kDa form missed one of the three components, we suspect that the missing band is responsible for the defective DNA binding.
Transcription factor abnormalities seem to be extensive and diverse in patients with SLE. In an earlier study, we had demonstrated that the NF-κB activity was decreased in SLE T cells because of decreased or absent levels of the p65 subunit (16). Interestingly, the p50 subunit was not affected, suggesting that the decrease of p65 chain (16) and of the Elf-1 98 chain reported herein do not represent generalized T cell protein degradation. In support of this conclusion is the finding that the transcriptional repressor cAMP response element modulator is increased in T cells from patients with SLE (17). The data shown in Fig. 2⇑ also support the conclusion that T cell proteins do not undergo indiscriminate degradation. However, it should be noted that under certain conditions T cell proteins may be degraded as is the case of caspase-mediated degradation of the TCR ζ-chain in Jurkat T cells (8).
We have shown that the discrepancy between the predicted molecular mass (68 kDa) and the observed mass of the 98-kDa form of Elf-1 in normal T cells is at least in part the result of phosphorylation and O-linked glycosylation (13). O-linked glycosylation has been shown to contribute to the activation of a number of transcription factors (21). Defective glycosylation may also be responsible for the decreased conversion of the 68-kDa because phosphorylation alone cannot account for the 38-kDa difference between the predicted and the observed molecular mass of Elf-1. At least 20 Elf-1 amino acid residues (11 Thr and 9 Ser) can be identified on Elf-1, which have high possibility (0.75–0.98, which is above the threshold) for O-GlcNAc modification. Deficiency in β1,6-N-acetylglucosaminyltransferase V, an enzyme in the N-glycosylation pathway, has been shown to cause kidney autoimmune disease and increased susceptibility to autoimmune encephalitis (22), indicating that glycosylation defects may be associated with autoimmunity. It is unclear at this moment whether defective glycosylation or phosphorylation, or both, are responsible for the decreased expression of the 98-kDa form. We have accumulated evidence that the cellular levels of the 98-kDa Elf-1 represent the balance between the phosphorylation and dephosphorylation of Elf-1 (Ref. 13 and our unpublished observations). Protein kinase C (PKC) activity has been claimed to be decreased in patients with SLE (23). Defective PKC activity may be responsible for the decreased formation of the 98-kDa Elf-1 since we showed before that forced expression of active form of PKC θ and ε enhanced the expression of the 98-kDa Elf-1 in Jurkat cells. Alternatively, increased phosphatase activity in SLE T cells can account for the decreased expression of the 98-kDa form. Indeed, we have found that treatment of T cell protein extracts with bacterial phage λ phosphatase (13) results in the disappearance of the 98-kDa form. A dephosphoryalted 98-kDa Elf-1 may undergo degradation (13). It should be noted that phosphorylation of c-Jun has been shown to stabilize the protein and protect it from degradation through ubiquitin pathway (24).
The observation that in more than a third of the studied SLE patients (14 of 31) there was a significant decrease in the TCR ζ-chain promoter binding despite the fact that the apparent levels of the 98-kDa protein were normal presented a special challenge. In previous studies (13), we had observed that the 98-kDa Elf-1 could be resolved in 6% SDS gels into three components, which became more obvious when the cells were treated with the phosphatase inhibitor okadaic acid. These studies prompted us to consider that in certain SLE patients, not all components are formed as a result of defective phosphorylation. Indeed, separation of normal nuclear 98-kDa Elf1 in IEF gels produced three distinct bands, and SLE samples were shown to miss one of them. Additional studies are needed to confirm that the middle band of the Elf-1, which is missing in SLE T cells, represents the one responsible for the DNA binding.
It is possible that other nuclear proteins, particularly members of the Ets family which share the Ets DNA binding domain with Elf-1 and thus recognize similar DNA binding sites, replace Elf-1 and play a role in the regulation of the TCR ζ-chain expression in SLE T cells. However, our EMSA revealed that in SLE patients, the nuclear protein/DNA interaction generated only one band which comigrated with the Elf-1-specific DNA protein complex in a way similar to that observed in the normal T cells ( Figs. 4–6⇑⇑⇑). The lack of additional bands in EMSA using SLE T cells suggests the absence of binding of other Ets family proteins. It should be noted that the T cell-abundant Ets-1 and 2 have molecular mass of 50 and 56 kDa, respectively; and therefore, if present, they should generate shifted bands with different mobility features than the 98-kDa Elf-1. Furthermore, unlike transcription factors such as CREB/cAMP response element modulator, which frequently either coexist or compete for the same binding sites, members in the Ets family display unique binding preference in a variety of gene promoters (14).
We failed to establish a correlation between the defective expression and DNA binding of Elf-1 with disease activity or treatment. We considered the possibility that the observed defects are due to the medications that the patients were receiving during the performance of this study and in particular of prednisone since most of the SLE patients were receiving 10 mg of prednisone per day. We have calculated the blood level of prednisone and found that shortly after it is taken (2–3 h), it reaches its peak blood level which is approximately equivalent to 70 nM of dexamethasone. It stays at its peak level only for short period of time and then it declines rapidly to ∼7 nM. We have previously determined that culture of normal T cells with 10 nM of dexamethasone increases rather than decreases the expression of ζ-chain (25). Therefore, we believe that treatment with prednisone cannot account for our observations. It should be noted that higher concentrations of dexamethasone (100 nM) can suppress the expression of ζ-chain (26). Defective Elf-1 expression was observed in patients with a wide variation in SLEDAI scores. Reproducibly decreased levels of Elf-1 protein and DNA-binding levels were observed in patients (n = 3) studied several times over a period of 2 years in whom the SLEDAI scores fluctuated. Specifically, in one patient the 98-kDa Elf-1 was undetectable on four different occasions, and in two more patients, who expressed normal levels of 98-kDa Elf-1, the DNA-binding intensity remained similarly low on two different occasions (data not shown).
It is important to note that although this study has presented a strong correlation between the Elf-1/DNA interaction and the mRNA levels in SLE T cells, the expression levels of the TCR ζ-chain protein in a portion of lupus patients (5 of 31, 16%) did not reflect accurately the TCR ζ-chain promoter activity (data not shown). Because previous studies have shown that TCR ζ-chain internalization and subsequent degradation also plays a role in the down-regulation of TCR ζ-chain, further studies are needed to explore additional mechanisms that may contribute to the decreased expression of ζ-chain in SLE T cells.
In conclusion, we have established that defective transcriptional activity of the TCR ζ-chain promoter contributes to the decreased expression of TCR ζ-chain mRNA and protein. More importantly, we have demonstrated that defective activation, rather than transcription and translation of Elf-1, the sole transactivating factor known to be involved in the transcription of the TCR ζ-chain gene, is responsible for decreased TCR ζ-chain expression. The exact ranking of our findings in terms of their importance in the pathogenesis of human SLE is unknown. However, it appears that aberrations at the level of transcription factor must be more central to the origin of the disease. It is also certain that the complete characterization of all molecular defects displayed by SLE T cells may permit the identification of a “molecular signature” for each patient or group of patients with SLE.
The repercussions of our findings extend beyond the expression of TCR ζ-chain because Elf-1 has been implicated in the transcriptional regulation of additional genes that are relevant to the regulation of the immune response such as blk, lck, and lyn kinases (27), IL-2R α-chain (28), CD4 (29), GM-CSF (12), and IL-2 (14). Although the form of Elf-1 that binds to the promoters of these genes may not be the same as the one that binds to the TCR ζ-chain promoter, we suspect that the identified defective Elf-1 activation in SLE T cells may account for additional defects in these cells (1).
Acknowledgments
We thank Drs. E. Solomou, M. Tolnay, and S. Krisnan for sharing samples.
Footnotes
-
↵1 This work was supported by National Institutes of Health Grants RO1 AI42269 and RO1 AI49954.
-
↵2 Address correspondence and reprint requests to Dr. Yuang-Taung Juang, Department of Cellular Injury, Walter Reed Army Institute of Research, Silver Spring, MD 20910. E-mail address: Yuang-Taung.Juang{at}NA.AMEDD.ARMY.MIL
-
↵3 Abbreviations used in this paper: SLE, systemic lupus erythematosus; SLEDAI, SLE disease activity index; ChIP, chromatin immunoprecipitation analysis; IEF, isoelectric focusing gel; PKC, protein kinase C.
- Received July 25, 2002.
- Accepted September 16, 2002.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
