Introduction

The complement system is essential to innate immunity and functions as a key modulator of adaptive immunity.1 Complement activation, regardless of the initiated pathways, eliminates invading pathogens mainly via three major physiological effectors: (1) the potent pro-inflammatory anaphylatoxins C3a and particularly C5a, which modulate and prime the immune system; (2) C3b/iC3b deposited on targeted surfaces to opsonize the invaders in cooperation with immune cells; and (3) the membrane attack complex (MAC), that is, the C5b-9n complex, which is assembled on targeted surfaces for direct lysis.2 MAC-mediated cytolysis is characterized by a rapid increase in intracellular Ca2+ accompanied by the loss of mitochondrial membrane potential and leakage of adenine nucleotide pools.3 Once enough MAC is assembled on the cell surface, it can lead to the direct lysis of targeted invaders.4, 5 To protect autologous cells from bystander MAC-mediated damage, the sole membrane-bound complement regulator CD59 has a critical role in restricting MAC assembly.6

The human encodes a single CD59 gene, whereas the mouse encodes duplicated CD59 comprising separated mCd59a7 and mCd59b8 genes that share 57.7% consensus in protein sequence. Human CD59 is ubiquitously expressed in most cells and tissues.9 mCd59a was identified as the primary regulator inhibiting MAC formation,10 because mCd59a is universally expressed and mCd59b is only selectively expressed in testis and on sperm.11, 12 However, mCd59b demonstrates specific inhibitory efficacy similar to or even sixfold higher than mCd59a in restricting MAC assembly that may result from its different protein sequence.11, 13 In addition, mCd59b was suggested to have additional complement-independent roles in acrosome activation and motility.12, 14 Deficient or reduced expression of mCd59a and/or mCd59b partly contributes to various immune and inflammatory processes and diseases such as transplant rejection, autoimmune hemocytopenia, systemic lupus erythematosus and atherosclerosis.15, 16, 17, 18 Thus, further work is needed to shed light on the underlying mechanisms by which CD59 is regulated in response to extracellular inflammatory stimulation. Recently, we comprehensively investigated the human CD59 transcriptional regulation, and revealed that the constitutive expression of human CD59 is regulated by the widely expressed transcription factor trans-acting transcription factor 1 (Sp1), whereas canonical nuclear factor kappa light chain enhancer of activated B cells (NF-κB) and cyclic AMP-responsive element-binding protein (CREB, as an enhancer-binding protein) scaffolded by CBP/p300 are responsible for its inducible expression.19 However, although we previously mapped the promoter regions of mCd59a and mCd59b,20 the details of their transcriptional regulation, and in particular the transcription factors involved, remain largely obscure.

Inflammation may induce the production of antibodies via the humoral immune response. Some are deleterious autoantibodies produced in autoimmune diseases that activate the classical complement pathway. Lipopolysaccharide (LPS), a major outer membrane component of gram-negative bacteria, may trigger the alternative complement pathway to accordingly initiate host defense.21 In addition, LPS promotes the secretion of pro-inflammatory cytokines represented by tumor necrosis factor-α (TNF-α), which further boosts inflammation. Therefore, we investigated the transcriptional regulation of mCd59 in response to this inflammatory stimulus.

In this study, we found that mCd59b but not mCd59a is upregulated to protect cells from complement-mediated MAC damage. Furthermore, we revealed that both NF-κB and serum response factor (SRF) cooperatively regulate the transcription of mCd59b under inflammatory stimulation, whereas the widely expressed mCd59a may be regulated by constitutive Sp1.

Results

NF-κB and SRF regulate mCd59b transcription

We recently demonstrated that Sp1 is responsible for human CD59 constitutive expression, and canonical NF-κB in conjunction with CREB (as an intron 1-located enhancer-binding protein) regulates human CD59-inducible transcription.19 Using MatInspector software (Genomatix Software GmbH, München, Germany), we first predicted the possible transcription factors for mCd59a (Gene accession no. NM_007652.5) and mCd59b (Gene accession no. NM_181858.1) expression in 2000-bp region upstream of transcriptional initiating site.19 Sp1- and NF-κB-binding sites were exclusively located in the promoter regions of mCd59a and mCd59b, respectively. The NF-κB response elements were located 151–163 bp upstream of mCd59b exon 1 (Figure 1a), whereas the Sp1 response elements were located 63–90 bp upstream of mCd59a exon 1 (Figure 2b). Next, we used a dual-luciferase reporter assay to identify the promoter region in mouse NIH/3T3 fibroblast cells. In agreement with our previous report,20 the 150–200-bp and 200–237-bp elements upstream of mCd59b exon 1 may represent two separate promoter regions (Figure 1b). The 150–200-bp region contains the predicted NF-κB response elements. Therefore, we mutated the critical nucleotides for NF-κB binding in the 1–200-bp region upstream of mCd59b exon 1 (Supplementary Table 2). The results of the dual-luciferase reporter assay showed that promoter activity in the 1–200-bp region with nucleotides mutation was completely abolished to the level in the 1–150-bp region and was dramatically decreased compared with the 1–200-bp region without any nucleotide mutation (Figure 1c). We performed a ChIP assay to verify the physiological interaction of NF-κB with this region using the specific primers indicated in Figure 1a. Antibodies directed against the canonical NF-κB subunits p50, p65 or c-Rel captured this NF-κB response element-containing region (Figure 1d). Therefore, these results strongly indicate that canonical NF-κB may be a transcription factor for mCd59b expression.

Figure 1
figure 1

NF-κB and SRF regulate mCd59b transcription. (a) The sequence of the region from −300 to −70 bp upstream of mCd59b exon 1. The NF-κB and SRF response elements and the primer sequences for the ChIP assay are indicated. (b) Identification of mCd59b promoter regions using a dual-luciferase reporter assay. The significance was determined using one-way analysis of variance (ANOVA) with post-hoc testing, ***P<0.001 vs pGL3. The effects of NF-κB (c) or SRF (d) core nucleotide mutation on the activity of the mCd59b promoter in the −200 to −1- (c) or −350 to −1 (d)-bp regions. The significance was determined using one-way ANOVA with post-hoc testing, **P<0.01, ***P<0.001 vs NF-κB or SRF mutants. (e and f) ChIP assay. Antibodies against c-Rel, p65 or p50 (e) or against SRF (f) were employed to capture the complex containing the relative promoter sequences, then the DNA fragments were determined by qRT-PCR. One-way ANOVA with post-hoc testing (e) and two-tailed Student’s t-test (f) were used to determine the significance. **P<0.01, ***P<0.001 vs mouse or rabbit IgG. All experiments were performed independently for three times in NIH/3T3 cells, data are presented as mean±s.d. of three independent experiments.

Full size image
Figure 2
figure 2

Sp1 regulates mCd59a transcription. (a) The tissue distribution of mCd59a and mCd59b as determined by qRT-PCR from six mice (half were male mice for testis detection). (b) The sequence of the −150 to −1-bp region upstream of mCd59a exon 1. The predicted Sp1-binding site and the primers for the ChIP assay are indicated. (c) ChIP assay. An antibody against Sp1 was used to capture the complex containing the promoter sequence in NIH/3T3 cells. (d and e) The insufficiency of Sp1 induced by specific siRNA reduced mCd59a expression at both the mRNA (d) and protein (e) levels. The significance was determined using the two-tailed Student’s t-test in (c and d), **P<0.01. (f) The expression level of Sp1 in the tested tissues pooled from three male mice. Data are presented as mean±s.d. of three independent experiments.

Full size image

Considering that the NF-κB-binding site is 150–200 bp upstream of mCd59b exon 1, the dual-luciferase reporter analysis results suggest that another potent transcription factor may bind the 200–237-bp region due to its high promoter activity (Figure 1b). In addition, based on the MatInspector prediction, SRF was hypothesized to be another transcription factor for mCd59b (Figure 1a). Accordingly, the dual-luciferase reporter assay was employed to detect promoter activity with core nucleotide mutations in the SRF binding site in NIH/3T3 cells (Supplementary Table 2). SRF core nucleotide mutations between 1–350 bp significantly decreased luciferase activity to a level similar to that of the 1–200-bp region, which contains an NF-κB-binding site (Figure 1e). These results indicate that NF-κB and SRF may regulate mCd59b transcription separately. In addition, the ChIP assay using specific primers (Figure 1a) and a specific antibody against SRF further demonstrated the direct interaction between SRF and the SRF response element-containing region (Figure 1f). Taken together, these results clearly suggest that both NF-κB and SRF are the transcription factors regulating mCd59b.

We previously determined that in response to cAMP treatment, CREB is phosphorylated and binds to an enhancer region located in the human CD59 intron 1, thus upregulating CD59 transcription together with NF-κB.19 Here we further investigated whether CREB regulates mCd59 transcription. Similarly, we treated mouse Hepa 1-6 cells with 8-Br-cAMP and found that mCd59b but not mCd59a dramatically increased at both the transcriptional and translational levels (Figures 3a and b). We accordingly reasoned that there may be an enhancer for CREB binding located in mCd59b intron 1, as in human CD59 regulation. Thus, we cloned different fragments from mCd59b intron 1 and inserted them into the 3′ terminus of the luciferase gene in a pGL3 plasmid that already contained the 1–500-bp region upstream of mCd59b exon 1 at the 5′ terminus of the luciferase gene. As shown in Figure 3c, however, we failed to observe increased luciferase activity, even with the addition of 8-Br-cAMP, indicating that unlike human CD59, there is no enhancer region in mCd59b intron 1. In addition, we used 8-Br-cAMP to treat NIH/3T3 cells transfected with pGL3 plasmids that contained different fragments comprising up to 2000 bp upstream of mCd59b exon 1. There was no significant increase in luciferase activity (Figure 3d). We then cloned three distinct fragments comprising 2000–4000 bp and separately inserted them into a pGL3 plasmid containing the region 1–500 bp upstream of mCd59b exon 1. This dual-luciferase reporter assay exhibited no significantly increased promoter activity, regardless of additional treatment with 8-Br-cAMP (Figure 3e). These results revealed that no promoter/enhancer element is particularly able to respond to 8-Br-cAMP in the region located in mCd59b intron 1 and up to 4000 bp upstream of mCd59b exon 1. Importantly, cAMP-mediated mCd59b upregulation may be independent of SRF and NF-κB activation. We induced CREB insufficiency by transfecting specific siRNA into Hepa 1–6 cells and observed remarkably increased mCd59b expression at the transcriptional and translational levels resulting from 8-Br-cAMP pretreatment (Figures 3f and g). Taking these data together, we concluded that unlike human CD59, mCd59b does not require CREB for its transcriptional regulation.

Figure 3
figure 3

cAMP upregulation of mCd59b expression was not mediated by CREB. (a and b) 8-Br-cAMP treatment enhanced mCd59b but not mCd59a expression as determined by qRT-PCR (a) and an immunoblotting assay (b). The significance was determined using the two-tailed Student’s t-test, **P<0.01 vs pre-8-Br-cAMP treatment. (c) mCd59b intron 1 had no effect in modulating the promoter activity in the −500 to −1-bp region upstream of exon 1. (d) Treatment with 8-Br-cAMP failed to enhance the promoter activity of the −2000 to −1-bp region upstream of mCd59b exon 1. (e) The addition of distinct regions in the −4000 to −2000-bp region in exon 1 to the −500 to −1-bp region upstream of mCd59b was unable to enhance promoter activity, regardless of 8-Br-cAMP treatment. There is no significance between control and 8-Br-cAMP treatment determined using two-tailed Student’s t-test in (c–e). (f and g) CREB insufficiency induced by specific siRNA did not impair the effects exerted by 8-Br-cAMP on increasing mCd59b expression, as evaluated by qRT-PCR (f) and an immunoblotting assay (g). One-way analysis of variance with post-hoc testing was used to determine the significance in (f), *P<0.05, **P<0.01 vs scramble siRNA. Data are presented as mean±s.d. of three independent experiments.

Full size image

NF-κB and SRF upregulate mCd59b but not mCd59a expression upon TNF-α treatment in vitro

TNF-α, an important pro-inflammatory cytokine, can activate both NF-κB and SRF.22, 23 We demonstrated that TNF-α considerably enhanced the nuclear accumulation of canonical NF-κB and SRF in Hepa 1-6 cells (Figure 4a). Next, we employed TNF-α to determine whether activated NF-κB and SRF upregulated the expression of mCd59b. NIH/3T3 cells were transfected with pGL3-related plasmids with or without TNF-α pretreatment. The inserted fragments were classified as 1–350 (containing both SRF- and NF-κB-binding sites), 1–350 (SRF mut), 1–350 (NF-κB mut), 1–200 (only NF-κB-binding site) and 1–150 (no promoter activity; Figure 4b). TNF-α pretreatment significantly enhanced the transcriptional ability of each promoter(s)-containing pGL3 plasmid, as demonstrated by increased luciferase activity (Figure 4b). Accordingly, pretreatment with TNF-α for 8 h remarkably increased the mRNA levels of mCd59b but not mCd59a, as detected by quantitative real-time (qRT)-PCR in mouse Hepa 1-6 hepatoma cells (Figure 4c). Subsequently, mCd59b but not mCd59a was found to increase at the translational level by immunoblotting (Figure 4d). We noted that the luciferase activity shows no significant changes between 1–350 (containing both SRF- and NF-κB-binding sites) and 1–350 (NF-κB mut) regions in the absence and presence of TNFα treatment in NIH/3T3 cell. This result was most likely caused by the much weaker nuclear localization of NF-κB and much stronger nuclear localization of SRF in NIH/3T3 cell than those in Hepa 1-6 cell nucleus (Supplementary Figure 2). We also performed dual-luciferase reporter assays to detect the transcriptional activities in regions of 1–350 (containing both SRF- and NF-κB-binding sites) and 1–350 (NF-κB mut) in Hepa 1–6 cells. The data showed that removing NF-κB involvement by mutating the sequence displays significant effect on mCd59b promoter activity in Hepa 1–6 cells (Supplementary Figure 3). These results clearly demonstrated that both SRF and NF-κB could regulate mCd59b transcription, and the difference of the two transcription factors in regulating transcription depends on their expression levels or activation status.

Figure 4
figure 4

TNF-α treatment increased mCd59b expression by activating canonical NF-κB and SRF. (a) TNF-α treatment enhanced the nuclear accumulation of canonical NF-κB and SRF. (b) TNF-α treatment enhanced the promoter activity of the −350 to −1-bp region upstream of mCd59b exon 1, which contains NF-κB and SRF response elements. (c and d) TNF-α treatment increased mCd59b but not mCd59a expression as determined by qRT-PCR (c) and an immunoblotting assay (d). (e) Pretreatment with TNF-α significantly reduced complement-mediated cytolysis in Hepa 1-6 cells. The significance was determined using two-tailed Student’s t-test, *P<0.05, **P<0.01 and ***P<0.001. (f) Specific siRNAs against SRF, c-Rel or p65 inhibited target protein synthesis. (g and h) The insufficiency of SRF, c-Rel or p65 induced by specific siRNAs abrogated the increased mCd59b expression resulting from TNF-α treatment (g) and subsequently restored sensitivity to complement-mediated cytolysis (h). The significance was determined using one-way analysis of variance with post-hoc testing in (h), *P<0.05 and **P<0.01 vs scrambled siRNA transfection and TNF-α pretreatment. Data are presented as mean±s.d. of three independent experiments.

Full size image

In addition, we functionally tested whether the upregulation of mCd59b upon TNF-α treatment in Hepa 1–6 cells protected the cells from complement attack. CK18 is commonly expressed in the mature hepatocyte membrane.24 Therefore, an anti-CK18 monoclonal antibody was used to trigger the classical complement pathway in combination with normal human serum as a complement resource. As expected, the results of the complement-dependent cytotoxicity assay showed that pretreatment with TNF-α in Hepa 1–6 cells significantly reduced the cytolysis of Hepa 1-6 cells due to upregulated mCd59b (Figure 4e). However, this protective capacity upon TNF-α pretreatment could be impaired by specific siRNA-induced insufficiency of SRF or NF-κB (Figure 4h). The efficacy of knockdown with specific siRNAs against SRF or NF-κB (p65, c-Rel) was confirmed by an immunoblotting assay (Figure 4f). The resulting modulation of TNF-α-induced mCd59b expression was also detected. Increased expression of mCd59b upon TNF-α treatment was significantly compromised by the insufficiency of either SRF or the canonical NF-κB subunits c-Rel or p65 (Figure 4g). Therefore, we concluded that TNF-α may upregulate mCd59b but not mCd59a expression to protect cells from complement damage via the activation of NF-κB and SRF signaling.

Sp1 is responsible for the constitutive expression of mCd59a in mouse tissues

We first investigated the mRNA expression profiles of mCd59a and mCd59b via qRT-PCR in various mouse tissues. mCd59a is universally expressed, with particularly high levels in the heart, liver, kidney, testis and eye. By contrast, mCd59b is only highly expressed in testis and is very weakly expressed in a small number of other tissues, including liver (Figure 2a).

We previously demonstrated that the mCd59a promoter region is located −122 to −36 bp upstream of exon 1 (ref. 20). However, the identity of the corresponding transcription factor remains unclear. As mentioned above, we predicted that this region contains two putative Sp1-binding sites (Figure 2b). Simultaneously, considering that human CD59 and mCd59a, unlike mCd59b, are universally expressed and Sp1 is responsible for the constitutive expression of human CD59 (ref. 19) we hypothesized that Sp1 may regulate mCd59a transcription. Therefore, we employed a ChIP assay with a specific antibody directed against Sp1 to capture the region containing the hypothesized Sp1-binding sites. The result clearly revealed a physiological interaction between Sp1 and this region in NIH/3T3 cells (Figure 2c). The specific primers are shown in Figure 2b. Next, we knocked down Sp1 expression in NIH/3T3 cells using a specific siRNA (Figure 2d). The qRT-PCR and immunoblotting results demonstrated that mCd59a expression was remarkably reduced at both the transcriptional and translational levels (Figures 2d and e). Furthermore, we detected Sp1 expression in multiple mouse tissues with an immunoblotting assay. The results indicated that Sp1 is expressed at extraordinarily high levels in the heart, moderate levels in the liver, kidney and testis, and very low levels in the spleen and lung (Figure 2f). This is highly correlated with mCd59a transcription levels (Figure 2a). Together, these results clearly demonstrate that Sp1 may be responsible for the broad expression of mCd59a in mouse tissues.

Mouse Cd59b but not Cd59a is upregulated upon LPS treatment in vivo

We demonstrated above that mCd59b but not mCd59a is upregulated upon TNF-α treatment in vitro via activated SRF and canonical NF-κB signaling, thus conferring cellular protection from complement attack. LPS is an important bacterial virulence factor that not only promotes the production of potent pro-inflammatory TNF-α but also triggers the alternative complement pathway. Thus, we injected LPS intraperitoneally to mimic inflammation in the mice and then collected mouse tissues to measure mCd59a and mCd59b expression levels, which may help us to understand the host response to complement attack; mice injected with phospahte-buffered saline (PBS) intraperitoneally were used as control group. Using qRT-PCR, we found that mCd59b rather than mCd59a transcription increased twofold to sixfold in the colon, kidney, liver and spleen (Figure 5a). The immunoblotting results further revealed that mCd59b but not mCd59a was accordingly elevated at the protein level (Figure 5b). Therefore, these results strongly suggest that mCd59b but not mCd59a is upregulated to protect cells from potential complement attack during infection or inflammation in vivo.

Figure 5
figure 5

LPS treatment increased mCd59b but not mCd59a expression in vivo. The mRNA (a) and protein (b) levels of mCd59a and mCd59b were determined by qRT-PCR and immunoblotting assay. Samples of control and the experimental group were pooled, respectively, from three adult female mice that received PBS or LPS pretreatment for 48 h. (c) Schematic diagram of the transcriptional regulation of mCd59a and mCd59b. Data are presented as mean±s.d. of three independent experiments.

Full size image

Discussion

The complement system, a major component of innate immunity, is an important first-line defense against invading pathogens. It can be subtly tuned physiologically to establish a delicate balance between activation and regulation. However, the tipping of this balance resulting from hyperactivation and/or hyporegulation may contribute in part to various human diseases. CD59 is the sole membrane complement regulator for restricting MAC assembly. Human CD59 and mCd59a are universally expressed in almost all tissues, whereas mCd59b is selectively expressed in testis.12 Therefore, it is critical to discover how CD59 responds to the extracellular inflammatory environment in the context of various disorders. We previously revealed that Sp1 is responsible for the constitutive expression of human CD59, whereas NF-κB and CREB regulate the inducible expression of this gene. In this study, we further determined that Sp1 is responsible for mCd59a expression, whereas SRF and NF-κB are responsible for mCd59b regulation in response to TNF-α in vitro and LPS in vivo stimulation (Figure 5c).

NF-κB has a critical role in apoptosis, differentiation and, in particular, immunity.25 The NF-κB proteins comprise five distinct proteins: RelA (p65), RelB, c-Rel, p105 (p50 when processed) and p100 (p52 when processed). These proteins form homo- or heterodimers with distinct functions.26, 27, 28 Normally, these NF-κB proteins are sequestered in the cytoplasm by inhibitors of kappa B. Following the phosphorylation and subsequent ubiquitination and degradation of inhibitors of kappa B, NF-κB proteins are released and translocated into the nucleus to activate the transcription of a variety of genes encoding proteins for inflammation and survival.25, 29 SRF, a transcription factor associated with cell proliferation and differentiation, neuronal transmission, and muscle development and function, has important regulator functions in tissue injury and wound healing.30 Genetic deficiency of murine SRF induces early lethality before mesoderm formation.31 Recent studies have also demonstrated a role for SRF in the regulation of inflammation and immunity. SRF is essential for neutrophil migration in response to inflammation32 and it indirectly modulates type I interferon signaling in macrophages, without interfering in the classic JAK/STAT/ISGF3 pathway.33 Interestingly, complement-mediated sublytic MAC can cause c-fos transcriptional activation in myotubes by inducing the binding of SRF along with ternary complex factor Elk1 and Sap1a to the c-fos serum response element, which may subsequently contribute to the pathogenesis of autoimmune diseases, such as myasthenia gravis.34 Sublytic MAC could not only enhance the assembly of the SRF/Elk1/Sap1a ternary transcription complex via the activation of ERK1 and thus Elk1 (ref. 34) but could also activate NF-κB signaling by promoting NF-κB nuclear translocation.19, 35 TNF-α exhibits similar functions in activating SRF23 and NF-κB,25 which was further demonstrated in this study (Figure 4a). Therefore, the activation of SRF and NF-κB by inflammatory sequelae, including complement activation, in turn induces the upregulation of mCd59b to protect host cells against potential MAC-mediated damage, which suggests a feedback loop for host defense. In addition, 8-Br-cAMP markedly increased the expression of mCd59b but not mCd59a, similar to its role in human CD59 regulation. However, this cAMP-induced signaling to upregulate mCd59b is independent of CREB, NF-κB and SRF, and the underlying mechanism requires future investigation.

The transcription factor Sp1 is ubiquitously expressed to regulate the expression of thousands of genes involved in diverse cellular processes such as cell growth, differentiation, apoptosis, angiogenesis and the immune response.36 Sp1 binds GC-rich elements that are widely distributed in the promoters of housekeeping genes37 and is a constitutive transcription factor that regulates basal promoter activity.38, 39 The broad distribution of mCd59a in almost all mouse tissues as a primary complement regulator at the terminal phase 1010 may be explained by our finding that Sp1 regulates mCd59a transcription. In addition, the differences in transcriptional regulation between mCd59a and mCd59b may explain their differential distributions.

Similar with human CD59 distribution and regulation, mCd59a is widely distributed and regulated by the housekeeping transcription factor Sp1, and mCd59b is partially regulated by NF-κB. However, mCd59a and mCd59b merges into one CD59 gene in humans. Further, in response to external inflammatory stimuli, mCd59b-inducible expression regulated by NF-κB and SRF has evolved into human CD59 regulated only by NF-κB. Although 8-Br-cAMP treatment upregulates both mCd59b and human CD59 expression, the downstream molecule CREB, interestingly as an enhancer-binding protein, regulates only human CD59 but not mCd59b expression.19 The differences in copy number and transcriptional regulation of CD59 between mouse and human may result from evolutionary selection pressure. Therefore, the differences and similarities among human CD59, mCd59a and mCd59b should be taken into account when interpreting mouse model data and translating it to human physiology and diseases.

Materials and methods

Cell culture and reagents

The mouse cell lines Hepa 1-6 and NIH/3T3 (Type Culture Collection Cell Bank, Chinese Academy of Sciences) were maintained in Dulbecco’s Modified Eagle’s medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Recombinant human TNF-α was purchased from PeproTech (Rocky Hill, NJ, USA, Catalog #300-01 A). LPS and 8-Br-cAMP were purchased from Sigma (St Louis, MO, USA). Rabbit anti-mCd59a polyclonal antibody (FL-123); mouse anti-p65 (F-6), anti-p50 (E-10), anti-c-Rel (B-6), anti-Sp1 (E-3), anti-TFIIB (D3) and anti-β-actin (C4) antibodies; goat anti-mouse IgG-HRP; and goat anti-rabbit IgG-HRP were obtained from Santa Cruz Biotechnology (Dallas, TX, USA). A rabbit anti-SRF (D71A9) XP monoclonal antibody was obtained from Cell Signaling Technology (Danvers, MA, USA). Propidium iodide was obtained from Invitrogen (Waltham, MA, USA). Normal human serum as a complement resource was pooled from 10 healthy persons and aliquoted, then stored at −80 °C until use.

To discriminate between mCd59b and mCd59a, we entrusted Shanghai Genomics, Inc. to generate an anti-mCd59b polyclonal antibody by immunizing two New Zealand rabbits using the keyhole limpet hemocyanin-coupled mCd59b-specific epitope NKNLDGLEEP (from 70 to 79 aa). After three courses of immunization, the rabbits were killed and the sera were collected. The applicability and specificity of the rabbit antisera were further verified by immunoblotting in mCd59b-overexpressing Hepa 1-6 cells (Supplementary Figure 1). The mCd59b-overexpressing Hepa 1-6 cells were generated as follows: the coding sequence of mCd59b was cloned using PCR with specific primers (forward: 5′-GGAATTCATGAGAGCTCAGAGGGGAC-3′; reverse: 5′-CGGGATCCCCGAAGCAAAACTTCAAAATGG-3′) and inserted into pEGFP-N1 via the EcoRI and BamHI sites. After verification by sequencing, the mCd59b–pEGFP-N1 plasmid or control vector plasmid was transfected into Hepa 1-6 cells using Lipofectamine 2000 (Invitrogen). After 48 h, the positive cells were selected for with G418 (1000 ng μl−1, Sigma) for three rounds.

Dual-luciferase reporter assay

As described previously,19 we used a dual-luciferase reporter assay to identify the regions with promoter or enhancer activity in mCd59b. The fragments upstream of mCd59b exon 1 and of mCd59b intron 1 were cloned and inserted into the pGL3 Basic Vector (Promega, Madison, WI, USA). Details of the primers used to clone these fragments are given in Supplementary Table 1. Double-strand DNA fragments with critical site mutations for NF-κB or SRF (Supplementary Table 2) were synthesized by SBS Genetech Co., Ltd (Beijing, China). Then, the pGL3-derived plasmids together with the pRL-TK plasmid were transfected into NIH/3T3 cells using Lipofectamine 2000 (Invitrogen) with or without TNF-α pretreatment (10 ng ml−1) or 8-Br-cAMP (1 mm). After 48 h, dual-luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega) on a Bio-Tek synergy HT microplate reader. Firefly luciferase activity was normalized to Renilla luciferase activity. These assays were repeated three times.

ChIP assay

We performed the ChIP assay as described previously,19 and the related primers are shown in Figures 1a and 2b.

Quantitative real-time-PCR

Total RNA from cells or mouse tissues was extracted with TRIzol reagent (Invitrogen) and transcribed into cDNA using a Reverse Transcription System (Promega). The input cDNA was standardized and then amplified for 40 cycles with SYBR Green Master Mix (Invitrogen) and gene-specific primers on an ABI Prism 7900HT machine (Applied Biosystems, Waltham, MA, USA). The housekeeping gene ACTB (gene encoding β-actin) was used as an endogenous control, and samples were analyzed in triplicate. The qRT-PCR primers are listed in Supplementary Table 3.

To study the effects of Sp1, SRF, c-Rel, p65 or CREB on mCd59 expression, Hepa 1-6 cells were transfected with specific siRNAs (GenePharma Co., Ltd, Shanghai, China) using Lipofectamine 2000 (Invitrogen). The siRNA sequences are indicated in Supplementary Table 4. After 48 h, the cells were harvested to evaluate mCd59b and/or mCd59a expression at the transcriptional level by qRT-PCR and at the translational level by immunoblotting (see below).

Immunoblotting assay

Immunoblotting assay was performed according to a standard protocol as described previously.19 The bands of immunoblotting were quantified by ImageJ (Bethesda, MD, USA) and corrected for a loading control.

Complement-dependent cytotoxicity assay

The complement-dependent cytotoxicity assay was performed as described in our previous report.40 In brief, Hepa 1-6 cells were treated with TNF-α (10 ng ml−1) for 16 h before the assay, followed by treatment with anti-CK18 antibody and normal human serum for 1.5 h and incubation with propidium iodide (2 μg ml−1) at room temperature for 15 min; samples were immediately analyzed on a Cytomics FC 500 MPL (Beckman Coulter, Brea, CA, USA).

Animal studies

Animal studies were approved by the Animal Ethics Committee at Shanghai Medical School, Fudan University. Eight-week-old Balb/C mice were purchased from Slac Laboratory Animal (Shanghai, China). To detect the physiological expression of mCd59a and mCd59b, the tissues indicated in Figure 2a were collected from three male and three female mice after complete perfusion with PBS as described previously.12

To investigate the in vivo effects of LPS on mCd59a and mCd59b expression, six female mice were equally divided into LPS-treated (i.p., 10 μg kg−1 body weight) or PBS-treated (i.p., equal volume PBS) groups. After 48 h, mice were sacrificed and perfused. Liver, kidney, spleen and colon were then collected for analysis.

Statistical analysis

The data are presented as the means±s.d. The significant difference between two groups was determined using the two-tailed Student’s t-test for unpaired data. P-values<0.05 were considered statistically significant. All qRT-PCR were performed in triplicates. One-way analysis of variance was employed to analyze multiple comparisons, and specific significant differences were determined with post-hoc testing.