The activation of NF-κB by neutrophil lactoferrin (Lf) is regulated via the IκB kinase (IKK) signaling cascade, resulting in the sequential phosphorylation and degradation of IκB. In this study, we observed that Lf protein augmented p65 phosphorylation at the Ser536, but not the Ser276 residue, and stimulated the translocation of p65 into the nucleus. Lf was also shown to enhance the association between p65 and CREB-binding protein/p300 in vivo. To elucidate the mechanism by which Lf triggers these signaling pathways, we attempted to delineate the roles of the upstream components of the IKK complex, using their dominant-negative mutants and IKKα−/− and IKKβ−/− mouse embryonic cells. We demonstrated that both IKKα and IKKβ as well as NF-κB-inducing kinase are indispensable for Lf-induced p65 phosphorylation. However, MAPK kinase kinase 1 is not essentially required for this activation. We also observed that Lf-induced p65 phosphorylation was either partially or completely abrogated as the result of treatment with the mutant forms of TNFR-associated factor (TRAF) 2, TRAF5, or TRAF6. Moreover, we demonstrated that Lf directly interacted with TRAF5. Expression of the dominant-negative mutant of TRAF5 or its small interfering RNA almost completely abrogated the Lf-induced p65 phosphorylation. These results suggest that signaling pathways, including TRAFs/NF-κB-inducing kinase/IKKs, may be involved in the regulation of Lf-induced p65 activation, thereby resulting in the activation of members of the NF-κB family.
The transcription factor, NF-κB, plays a central role in the regulation of inflammatory and immune responses, control of cell division, and apoptosis (1, 2). Activation of the IκB kinase (IKK)4 complex, consisting of two catalytic subunits (IKKα and IKKβ) and the IKKγ/NF-κB essential modulator scaffolding protein, results in the phosphorylation of IκB molecules at two conserved N-terminal Ser residues (3). The regulation of NF-κB activity occurs primarily in the cytoplasm via the liberation of NF-κB from IκB molecules, as well as via nuclear translocation. However, the trans activation potential of NF-κB in the nucleus is ensured by the additional modification of the NF-κB itself, particularly via p65 phosphorylation and the condition of its surrounding chromatin (4). Recent studies have highlighted the molecular mechanisms underlying the trans activation of NF-κB, especially the p65 subunit (5). The phosphorylation sites were reported to be located at Ser276 in the Rel homology domain (6), and at Ser529/Ser536 in the C-terminal trans activation domains (7). Thus, the stimulus-induced phosphorylation of the p65 subunit, as well as the nuclear translocation of NF-κB, is a requirement for successful NF-κB activity in the transcriptional activation of NF-κB-dependent genes.
Several kinases, including protein kinase A, mitogen- and stress-activated protein kinase-1, casein kinase II, and the IKKs, have been identified as p65 phosphorylation effectors. Ser276 phosphorylation was demonstrated to be mediated by protein kinase A and mitogen- and stress-activated protein kinase-1, and Ser529 phosphorylation was mediated by casein kinase II (6, 8). Ser536 was shown to be phosphorylated by IKKs and PI3K (7, 9). Moreover, the trans activation activity of p65 may also be regulated via reversible acetylation, a process that has been attributed to the recruitment of CREB-binding protein (CBP)/p300, subsequent to p65 phosphorylation (10). Among many inducers of NF-κB activity, TNF-α-induced NF-κB activation is one of the most extensively characterized signaling pathways. TNF-α triggers the assembly of signaling molecules, including the TNFR-associated death domain, TNF-α receptor-associated factor (TRAF) 2, TRAF5, and receptor-interacting protein on the intracellular domains of the oligomerized receptors (11). The IKK complex is recruited to TRAFs in the cytoplasm, where the IKKs are activated. However, the precise molecular mechanisms underlying p65 phosphorylation have yet to be characterized.
Human lactoferrin (Lf) is a 80-kDa cationic iron-binding glycoprotein that is present in high concentrations in milk secreted by the mammary gland, and is also detected in most exocrine secretions, including tears, nasal secretions, saliva, and intestinal mucus (12). Lf is also synthesized and secreted by the secondary granules of neutrophils in response to inflammatory stimuli (13). Lf has been shown to exhibit a variety of biological functions, including antimicrobial and immunomodulatory effects, as well as the regulation of cell growth or differentiation (14, 15, 16, 17). However, the molecular mechanisms underlying this process remain unclear. Some biological activities of Lf are attributed to its iron-binding capacity (18). In particular, the antibacterial effect primarily depends on the iron status of Lf (19). However, Lf does not play a major role in the regulation of iron homeostasis in mice (20). The biological functions of Lf are dependent on its target cells, and might be due to its capacity to bind to various molecules in cells.
We demonstrated previously that Lf induces the phosphorylation and subsequent degradation of IκB molecules via the MEKK1-IKK signaling cascades (21). In this study, we attempted to investigate which cellular components were most closely associated with the role of Lf as a signal transducer, leading to the activation of the p65 NF-κB subunit. We attempted to determine whether Lf could affect the degree of phosphorylation of the p65 NF-κB subunit, and cause the recruitment of cellular factors such as histone acetylase to regulate its trans activation activity. We then attempted to ascertain the upstream signaling pathways responsible for Lf-induced p65 phosphorylation and trans activation, which were distinct from MEKK1-IKK signaling. Intriguingly, we demonstrated that TRAF5 is efficiently recruited in response to Lf treatment, and that the TRAFs-IKK signaling pathway is intimately involved in the Lf-induced phosphorylation and trans activation activity of p65.
Materials and Methods
Cells and reagents
The premonocytic cell line THP-1 and human leukemic monocyte lymphoma cell line U937 were obtained from American Type Culture Collection. Wild-type mouse embryonic fibroblast (MEF) cells and IKKα- and IKKβ-deficient MEF cells were provided by I. Verma (Salk Institute, La Jolla, CA) (22). Human chronic myeloid leukemia cell line K562 and human transformed primary embryonal kidney cell line 293 were cotransfected with pSV2neo and pLf, and subsequently selected with G418 to generate K562-Lf and 293-Lf cells, respectively. K562-neo and 293-neo cells were transfected with pSV2neo alone. Anti-FLAG Abs and iron-saturated Lf were purchased from Sigma-Aldrich. Anti-p65 or IκBα Ab was obtained from Santa Cruz Biotechnology. Ab specific to p65 phosphorylated at Ser536 or Ser276, p-RelA (Ser536), p-RelA (Ser276), and anti-CBP Ab were from Cell Signaling Technology and Chemicon International, respectively.
Plasmids and small interfering RNAs (siRNAs)
Human neutrophil Lf expression vector pLf was as described (23). Plasmids pRK-Flag-IKKα, pRK-Flag-IKKβ, pRK-Flag-IKKα (K44A), pRK-Flag-IKKβ (K44A), and pRK-Flag-NIK (KK429–430AA) were described previously (21). Plasmid GAL4-p65 (1–551) expressing the DNA-binding domain of yeast GAL4 protein fused to the p65 was generated via the cloning of the PCR product into the EcoRI/XbaI sites of the pM vector (BD Clontech). Expression vectors for TRAF1 (204–409), TRAF2, TRAF2 (87–501), TRAF5, TRAF5 (236–559), TRAF6, and TRAF6 (289–522) were provided by Y. Choi (Rockfeller University, New York, NY) (24). Expression vectors encoding for MEKK1Δ (K432M) and Akt1 (K179M) were gifts from T. Maniatis and T. Franke, respectively (25, 26). The pcDNA3-CBP-Flag and pBJ5-HDAC1-Flag plasmids were provided by S. Impey and S. Schreiber, respectively (27, 28). The duplex siRNAs for the Lf or GFP genes were described previously (21). HiPerformance-validated siRNA-targeting TRAF5 was purchased from Qiagen. Twenty-four hours after transfection of K562 or HEK293 cells with pLf, cells were transfected with each siRNA (5 μg) using RNAiFect reagents (Qiagen). After 48 h of incubation, the cells were harvested and analyzed via RT-PCR or Western blot.
The 293 cells were transfected using Effectene (Qiagen) with 1 μg each of plasmids encoding for GAL4 or GAL4-p65, together with 1 μg of the GAL4-driven pG5CAT reporter and the indicated plasmids, plus 50 ng of β-galactosidase plasmid. Twenty-four hours after transfection, cells were lysed and assayed for chloramphenicol acetyltransferase (CAT) or luciferase activity using a luciferase assay kit (Promega).
Total cellular RNAs were prepared with TRIzol (Invitrogen Life Technologies), according to the manufacturer’s instructions. Reverse transcription of mRNA was conducted using oligo(dT) primer and superscript II reverse transcriptase (Invitrogen Life Technologies). Gene transcripts were identified by semiquantitative PCR.
Immunoblot analysis and immunoprecipitation
For immunoprecipitation, endogenous p65 was immunoprecipitated with anti-p65 Ab, followed by incubation with protein G plus-agarose beads (Santa Cruz Biotechnology). After extensive washing, the complexes were subjected to immunoblotting with anti-CBP Ab. For immunoblot analysis using phospho-RelA Abs, 30 μg each of nuclear extracts or cell lysates was used and analyzed, as described (23).
For the immunofluorescence assays, cells on coverslips in six-well plates were transfected with 0.5 μg of pRSV-p65. Forty hours later, the cells were prepared for incubation with anti-p65 Ab and anti-IgG FITC conjugate, as described previously (21). Images were analyzed using a Zeiss Axioskop 2 microscope configured for epifluorescence and equipped with a camera device.
Lf induces the phosphorylation of the p65 subunit, and consequently augments its trans activation potential
NF-κB activity is believed, in general, to be regulated by the degradation of signal-induced IκB, as we previously demonstrated (21). However, a great deal of evidence supports the notion that NF-κB activity can also be regulated via the direct modification of the p65 subunit, via phosphorylation and acetylation (29). Therefore, we have attempted to determine whether Lf could induce p65 phosphorylation, consequently enhancing its trans activation activity. K562 cells were incubated for 6 h with iron-saturated Lf protein in culture medium, followed by immunoblotting with the anti-phospho-p65 Ab, p-RelA (Ser536), or p-RelA (Ser276). As shown in Fig. 1, A and B, treatment of K562 cells or U937 cells with Lf protein resulted in a profound increase in the levels of p65 phosphorylation at the Ser536, but not the Ser276 residue, thereby suggesting that Lf plays a specific role in the regulation of p65. However, treatment with a relatively higher concentration of Lf (100 μg/ml) resulted in slight decreases in the levels of p65 protein and its phosphorylation. THP-1 cells evidenced a response similar to that of the U937 cells unless they were induced by external stimuli, including polyliposaccharide (data not shown). We also attempted to determine whether or not Lf, when stably expressed in cells, would influence p65 phosphorylation. We observed that Lf was expressed successfully in the long-term Lf-expressing cell line, K562-Lf, and exerted a variety of effects similar to those of the recombinant Lf protein, including p65 phosphorylation at Ser536, when added to the medium (Fig. 1,C). Lf cDNA has been shown to be abundantly expressed, in a dose-dependent fashion, in all cell lines tested in this study. Therefore, Lf cDNA expression was also used in these studies. Next, we attempted to determine whether this change in p65 phosphorylation levels was, indeed, the result of Lf expression. In our siRNA experiments, we determined that Lf siRNA could efficiently suppress the Lf-induced phosphorylation of p65 (Fig. 1 D).
Generally, the enhancement of trans activation activity effected by the p65 NF-κB subunit has been attributed to p65 phosphorylation and its translocation to the nucleus (29). Therefore, as shown in Fig. 2,A, we have demonstrated that stable Lf expression in 293-Lf cells can result in the nuclear translocation of p65, whereas p65 was localized predominantly to the cytoplasm in 293-neo cells, thereby suggesting a potent role for Lf in the transcriptional activation of NF-κB-dependent genes. However, this observation did not fully support the notion that Lf stimulated trans activation potential by virtue of p65 phosphorylation. To demonstrate Lf-induced trans activation via p65 phosphorylation, the GAL4-responsive reporter, pG5CAT, and pLf were cotransfected with either GAL4-p65 (1–551) or the GAL4 expression plasmid. As expected, we observed increases in the p65 trans activation potential in response to Lf (Fig. 2 B), clearly indicating that Lf has the capacity to augment p65 phosphorylation, thus enhancing the trans activation activity of p65.
Lf enhances the interaction of p65 and CBP in vivo
Stimulus-induced p65 phosphorylation has been reported to induce the recruitment of CBP (6). To characterize the effects of Lf-induced p65 phosphorylation on CBP recruitment, we conducted a series of coimmunoprecipitation analyses. As a result, we determined that Lf did, indeed, enhance the association of p65 with CBP (Fig. 3,A). CBP and its homologue, p300, are believed to interact with phosphorylated p65, thereby enhancing the trans activation activity of NF-κB by virtue of its intrinsic histone acetylase properties (30). Fig. 3,B shows that Lf expression resulted in the enhancement of transcription via the expressions of both pLf and CBP. These results strongly support the notion that Lf functions as a signaling mediator, resulting in the covalent modification of p65, followed by the recruitment of histone acetylases, including CBP. Histone deacetylases (HDACs) have recently been demonstrated to control transcription activity via p65 (10). As shown in Fig. 3 B, expression of HDAC1 resulted in a minor decrease in basal p65 transcriptional activity, but more greatly reduced levels of Lf-induced p65-dependent transcription. However, treatment with the HDAC inhibitor, trichostatin A (TSA), increased the Lf-induced p65 trans activation as well as the basal level of p65 trans activation, suggesting that HDAC is involved in the suppression of Lf-stimulated p65 trans activation.
TRAF-IKK pathways play a role in Lf signaling to p65 phosphorylation
To gain insight into the mechanism by which Lf triggers the signaling pathways leading to NF-κB activation, we next attempted to delineate the role of the upstream components of the IKK complex. As several TRAFs, including TRAF2, TRAF5, and TRAF6, as well as IKKs, may perform crucial functions in NF-κB activation induced by members of the TNFR superfamily in response to TNF, we determined whether Lf could also use common pathways to stimulate p65 phosphorylation. As a result, Lf-stimulated p65 phosphorylation at Ser536 was diminished substantially by the expression of each dominant-negative form of TRAF, TRAF2, TRAF5, TRAF6, IKKα, or IKKβ (Fig. 4,A). To obtain a better understanding of the roles of IKKs, Lf was expressed in the IKK-null MEF cells. As shown in Fig. 4,B (lower panels), we have attempted to ascertain whether Lf exerted any effect on p65 phosphorylation when IκBα degradation occurred in MEF cells. The sequential phosphorylation and degradation of IκBα occurred actively in wild-type MEF cells and were stimulated by Lf expression, whereas the degradation of IκBα was abrogated in IKKα- and IKKβ-null cells. As shown in Fig. 4 B (upper panels), Lf expression resulted in an increase in the level of p65 phosphorylation in the wild-type MEF cells. By way of contrast, the basal levels of p65 phosphorylation were reduced, and Lf expression exerted little effects on the p65 phosphorylation in both the IKKα- and IKKβ-null cells. These results suggest that both IKKα and IKKβ are important for the Lf-induced p65 phosphorylation on Ser536.
We next attempted to determine the effects of components of the downstream TRAFs or the upstream IKKs, either of which might play a role in the NF-κB-mediated regulation of Lf-induced p65 activation. Several MAPK kinase kinases (MAP3Ks), including MAPK/ERK kinase kinase (MEKK) 1 and MEKK3, have been reported to mediate NF-κB activation (2, 5). Therefore, we attempted to ascertain whether MAP3Ks played any role in Lf-induced p65 phosphorylation. The expression of dominant-negative TRAFs and IKKs resulted in a reduction in the levels of Lf-induced p65 phosphorylation (Fig. 4,C). Therefore, these upstream molecules considered to be potential effectors may mediate Lf-triggered p65 phosphorylation. This effect, however, was not observed in association with MEKK1 (K432M). Expression of the dominant-negative mutants of TRAF5 proved to be particularly potent. Therefore, to demonstrate a tight correlation between Lf and the components of the IKK complex, we sought to determine whether Lf was able to bind to TRAF5 in vivo. Fig. 4,D clearly demonstrates that Lf directly interacts with TRAF5, strongly suggesting that Lf plays a prominent role as an effector protein in the IKK activation signaling pathway. Also, TRAF5 siRNA expression almost completely abrogated Lf-mediated p65 phosphorylation (Fig. 4 E). These findings indicate that signaling pathways, including TRAF-NIK-IKKs, can be involved in the regulation of Lf-induced IKK activation, independently of IκB.
Certain stimuli, including proinflammatory cytokines, phorbol esters, LPS, and a host of other diverse stimuli, are known to induce NF-κB activity (31). In this study, we demonstrated that Lf also regulates p65 phosphorylation at Ser536, which is located on the TA1 trans activation domain (7). It has been suggested that TNF-α stimuli, the Tax protein, or activation of the IKK complex may result in the phosphorylation of p65 at Ser536 (7, 32). Using a yeast two-hybrid assay system, we observed that the trans activation potential of NF-κB translocated to the nucleus can be ensured by the additional modification of NF-κB via p65 phosphorylation, as well as by the status of the chromatin structure. The advantage of this assay was that GAL4-p65 (1–551) is exclusively nuclear, and can be regulated independently of IκB. The effects of Lf on the association of p65 and histone acetylases such as CBP bolster the notion that the phosphorylation status of p65 determines whether it interacts with CBP or HDAC1 (30). The p65 phosphorylation enhances its trans activation potential via the augmentation of the interaction of p65 with coactivators including CBP and p300. In contrast, p65 trans activation potential has been shown to be regulated via reversible acetylation (29). Coexpression of HDAC1 abrogated Lf-mediated p65 trans activation activity, consistent with a previous report showing impaired NF-κB-dependent gene expression (10). Moreover, the treatment of cells transfected with Lf together with the HDAC inhibitor TSA resulted in the enhancement of Lf-induced p65 trans activation activity, thereby indicating a close association of p65 acetylation with trans activation potential. Indeed, p65 deacetylation in the nucleus occurs via specific interactions with HDACs, a process that promotes newly synthesized IκBα-mediated p65 nuclear export (33).
Lf-induced p65 phosphorylation appears to occur primarily via intracellular IKK signaling pathways. IKKα and IKKβ are both potential effectors, functioning as upstream kinases that phosphorylate p65. IKKα and IKKβ have been shown to phosphorylate p65 at Ser536 in lymphotoxin-β-activated cells and stimulated T cells, respectively (34, 35). However, the signaling pathways upstream of IKK remain to be thoroughly elucidated. Recently, PI3K/Akt1-induced p65 phosphorylation was reported to be a prerequisite for the maximal transcriptional activity of NF-κB (9). Using a kinase-dead Akt1 dominant mutant, we have observed that the blockage of Akt1 activity substantially abrogated Lf-stimulated p65 phosphorylation and trans activation (data not shown). It appears that PI3K/Akt may mediate Lf signaling to p65 phosphorylation by triggering downstream signaling components, consequently relaying signals to the MAP3Ks (S. Y. Choi, manuscript in preparation). Despite the fact that genetic studies have demonstrated that IKKβ, but not IKKα, phosphorylates IκB proteins, IKKα-null MEF cells are also defective with regard to NF-κB activation (36). Among several of MAP3K family members, NIK and MEKK1 were initially proposed to preferentially activate IKKα and IKKβ, respectively, but appear to be inessential for TNF- and IL-1-mediated NF-κB activation. However, NIK has been reported to be an essential component for other stimuli, including lymphotoxin-β treatment (37). In this study, we have demonstrated that MEKK1 does not contribute largely to Lf signaling to p65 phosphorylation. Our findings were consistent with previous results, which indicated that MEKK1 (K432M) blocks NF-κB binding, but not IKK activity (21). We also showed that NIK is required for Lf-induced p65 phosphorylation. The expression of the dominant-negative mutants of IKKα or NIK proved to be particularly potent. However, we have no direct evidence as yet for the involvement of IKKα in NIK-mediated activation. These results may be considered to bolster the notion that both IKKα and IKKβ are intimately involved in Lf-induced p65 phosphorylation, thereby eventually inducing NF-κB trans activation activity.
TRAF proteins have been suggested to function as a docking platform for a variety of regulators of NF-κB signaling pathways, and that TRAF2, TRAF5, and TRAF6, in particular, may mediate the NF-κB activation pathways (38). TRAF2, TRAF5, and TRAF6 have very similar structural features; harbor N-terminal ring finger domains and zinc finger domains that are required for downstream signaling or protein-protein interaction; and interact with cell surface receptors such as CD40 (39). Based on this fact, we attempted to determine whether TRAF2, TRAF5, or TRAF6 is involved in the Lf-mediated p65 phosphorylation pathways. As is shown in Fig. 4, the expression of TRAF2, TRAF5, or TRAF6 dominant negative, in which the ring finger or zinc finger domains were deleted, abrogated Lf-mediated p65 phosphorylation, whereas TRAF1 dominant negative evidenced a less sensitive response. It has also been reported recently that TRAFs interact with a number of regulators, including kinases, as well as cell surface receptors (40). Therefore, using immunoprecipitation assays, we found that Lf associated directly with TRAF5 in vivo, thereby implying that Lf may associate with TRAF2 or TRAF6 as well via their ring finger or zinc finger domains, thereby affecting downstream targets (Fig. 4). Collectively, we surmise that expressed or internalized Lf can associate with TRAF2, TRAF5, or TRAF6 to induce the recruitment of the IKKs to the complex, ultimately resulting in p65 phosphorylation. Therefore, we have concluded that the pathways that involve the TRAFs and IKK signaling components are instrumental to the mediation of Lf-triggered p65 phosphorylation and trans activation. However, we should not dismiss the possibility that Lf-induced p65 phosphorylation may be regulated by a heretofore-unidentified upstream IKK kinase, and that another p65 kinase might be associated with Lf signaling.
Lf is generally believed to be synthesized in selected cells by a variety of stimuli, and is stored in the secondary granules of mature neutrophils, from which it is secreted into circulation upon the degranulation of neutrophils (13). Many recent reports have addressed the cellular uptake of circulating Lf, as well as its physiological significance (41). We have demonstrated that Lf protein, when added to culture medium, exerts a stimulatory effect on p65 phosphorylation. The K562 or U937 cells that we have used harbor Lf-specific receptors on their cellular surfaces. We have also conducted ELISA in an effort to measure the levels of Lf released from K562-Lf or K562-neo cells into the culture medium. As a result, K562-Lf released ∼2-fold more Lf into the medium than did K562-neo (data not shown), thereby indicating that the expressed Lf was secreted into the culture medium. However, the precise mechanism underlying receptor-mediated Lf internalization in cells remains to be clarified. Therefore, recombinant Lf protein is considered to enter into the cells via receptor-mediated internalization. The concentration of Lf used, 20–100 μg/ml, was based on in vivo dose-dependent studies on Lf concentrations found at the inflammatory sites (42). Lf overexpression in cells results in a similar response. Therefore, it is likely that Lf overexpressed in cells may behave similarly to a high level of Lf internalized in the cytoplasm, although the precise mechanism underlying receptor-mediated Lf internalization in cells remains unclear. In contrast, we have not ruled out the possibility that p65 phosphorylation in cells overexpressing Lf occurs, presumably, via the production of an autocrine loop, because circulating Lf might function as a specific ligand that recognizes extracellular cell surface target proteins. However, an interesting feature of Lf signaling is the interaction between Lf and TRAF5, an essential upstream component of the IKK complex (Fig. 4, D and E), thereby suggesting that Lf may function via binding to the intracellular domain of the signaling receptor molecules. Therefore, it appears likely that Lf, when synthesized or internalized in the cytoplasm, can affect the recruitment of IKK signaling molecules, including TRAFs, and can interact subsequently with specific downstream kinases, such that the assembly of the complex may transduce signals to downstream components, ultimately resulting in the activation of NF-κB. The elucidation of new targets in the Lf signaling cascades has significant implications with regard to the development of anti-inflammatory strategies focused on the inhibition of NF-κB-dependent gene expression.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant R01-2004-000-10481-0 from the Basic Research Program of the Korea Science & Engineering Foundation.
Abbreviations used in this paper: IKK, IκB kinase; CAT, chloramphenicol acetyltransferase; HDAC, histone deacetylases; Lf, lactoferrin; MAP3K, MAPK kinase kinase; MEF, mouse embryo fibroblast; MEKK, MAPK kinase kinase; siRNA, small interfering RNA; TRAF, TNF receptor-associated factor; TSA, trichostatin A; CBP, CREB-binding protein.