The pathogenesis of Mycoplasma pneumoniae infection is considered to be in part attributed to excessive immune responses. Recently, lipoproteins from mycoplasmas have been reported to induce NF-κB activation. In this study, we examined the ability of lipoproteins from M. pneumoniae to activate NF-κB, and the active component responsible for the NF-κB activation was identified. Lipid-associated membrane proteins from M. pneumoniae were found to induce NF-κB through TLR 2 in a human monocytic cell line, THP-1. The active component of the Lipid-associated membrane proteins was a subunit b of F0F1-type ATPase (F0F1-ATPase). The F0F1-ATPase is assumed to contain two palmitic acids. The activation of NF-κB by the F0F1-ATPase was inhibited by a dominant negative construct of TLR1 and TLR6. These results indicate that the activation of NF-κB by F0F1-ATPase is dependent on TLR1, TLR2, and TLR6. The activity of the F0F1-ATPase was decreased with pretreatment of lipoprotein lipase but not protease, indicating that the lipid moiety of the F0F1-ATPase was important for the NF-κB activation. Thus, a dipalmitoylated lipoprotein from M. pneumoniae was found to activate NF-κB through TLR1, TLR2, and TLR6.
Mycoplasmas are wall-less parasitic Gram-positive bacteria, and the smallest organisms capable of self-replication (1). Mycoplasma pneumoniae causes primary atypical pneumonia, tracheobronchitis, pharyngitis, and asthma in humans (2, 3, 4). However, pathogenic agents such as endotoxin and exotoxin that cause such diseases have not been identified in M. pneumoniae. Adherence of invading Mycoplasma to the respiratory epithelium, localized host cell injury, and overaggressive inappropriate immune response seem to contribute to the pathogenesis in M. pneumoniae infection (5). It was reported that macrophage-activating lipopeptide 2 (MALP-2)3 from Mycoplasma fermentans and a 44-kDa lipoprotein from Mycoplasma salivarium stimulate host immune response (6, 7), indicating that such lipoproteins would participate in inflammatory responses in host. However, there seem to be no reports yet describing the lipoproteins, such as MALP-2, derived from M. pneumoniae.
Recently, it has been reported that TLRs with a function of pattern recognition receptors play critical roles in early innate recognition and inflammatory responses by host against invading microbes (8, 9). Among 10 TLR family members reported, TLR2, TLR4, TLR5, and TLR9 have been implicated in the recognition of different bacterial components. Peptidoglycan, lipoarabinomannan, zymosan, and lipoproteins from various microorganisms are recognized by TLR2 (10, 11, 12, 13, 14, 15, 16, 17). In contrast, LPS, bacterial flagellin, and bacterial DNA are recognized by TLR4, TLR5, and TLR9, respectively (18, 19, 20, 21). These TLR family members activate NF-κB via IL-1R-associated signal molecules, including myeloid differentiation protein (MyD88), IL-1R-activated kinase, TNFR-associated factor 6, and NF-κB-inducing kinase (22).
In this study, we examined the involvement of TLRs in the activation of immune response by lipoproteins from M. pneumoniae and their active component responsible for NF-κB activation. We observed that lipid-associated membrane proteins (LAMPs) from M. pneumoniae can induce NF-κB in a human monocytic cell line, THP-1, through TLR2. The active component responsible for NF-κB activation of LAMPs was found to be a 21-kDa lipoprotein. By the analysis with peptide mass fingerprinting, the 21-kDa protein was matched to a subunit b of F0F1-type ATPase (F0F1-ATPase). According to molecular mass analysis, the 21-kDa protein was assumed to have two palmitic acids at the N-terminal lipid binding cysteine. The activation of TLR signaling by the F0F1-ATPase was apparently dependent on TLR1, TLR2, and TLR6. Thus, the results indicate that the F0F1-ATPase induces an immune response that might be responsible for the pathogenesis in M. pneumoniae infection. Therefore, the F0F1-ATPase might be a candidate for the potential molecule for prevention and therapy for M. pneumoniae infection.
Materials and Methods
Cells of a human monocytic cell line, THP-1, were cultured in RPMI 1640 containing 10% FCS (Mitsubishi Chemical), 2 mM l-glutamine, 100 U/ml penicillin G, and 100 μg/ml streptomycin. Cells of a human kidney cell line, 293T, were cultured in DMEM containing 10% FCS, 2 mM l-glutamine, 100 U/ml penicillin G, and 100 μg/ml streptomycin.
Mouse anti-human TLR2 mAb, IMG-416, was obtained from Imgenex (23). Normal mouse IgG2a was purchased from BD Pharmingen.
Pathogen-associated molecular patterns (PAMPs)
Preparation of LAMPs from M. pneumoniae
M. pneumoniae M129 was cultured in pleuropneumonia-like organism medium to the beginning of a stationary phase and then pelleted by centrifugation for 10 min at 12,000 × g. Preparation of LAMPs was performed as previously described by Feng et al. (26, 27). Briefly, a Mycoplasma pellet was suspended in TBS containing 1 mM EDTA (TBSE), solubilized by adding TX-114 to a final concentration of 2%, and incubated at 4°C for 1 h. The lysate was incubated at 37°C for 10 min for phase separation. After centrifugation at 10,000 × g for 20 min, the upper aqueous phase was removed and replaced by the same volume of TBSE. The procedure of phase separation was repeated twice. The final TX-114 phase was resuspended in TBSE to the original volume, and 2.5 volumes of ethanol were added to precipitate membrane components and incubated at −20°C overnight. After centrifugation, the pellet was suspended in PBS followed by sonication for 30 s at output 5 (Sonifier cell disruptor 200; Branson). Protein concentration of the suspension was measured with the Coomassie Protein Assay Regent (Pierce).
To prepare TLR1, TLR2, and TLR6 expression vectors (pFLAG-TLR1, pFLAG-TLR2, and pFLAG-TLR6), the coding regions of TLR1, TLR2, and TLR6 minus the respective N-terminal signal sequences were amplified by PCR from a cDNA of THP-1 and cloned into the expression vector pFLAG-CMV1 (Sigma-Aldrich), in which a preprotrypsin leader precedes an N-terminal FLAG epitope. Dominant negative (DN) TLR1 and TLR6 expression vectors were constructed by subcloning the Toll and IL-1 receptor homology domain-deleted TLR1 and TLR6 fragment into pFLAG-CMV1 (pFLAG-dTLR1 and pFLAG-dTLR6). The NF-κB cis-Reporting System containing pNF-kB-luc, a plasmid in which luciferase reporter gene was fused to NF-κB enhancer, was purchased from Stratagene.
Transfection and luciferase assay
Transient transfection was performed by using FuGENE6 (Roche) according to the manufacturer’s instructions. A total of 4 × 105 THP-1 cells or 1 × 105 293T cells were transfected with 0.1 μg of pFLAG-TLR2, 0.01 μg of pNF-kB-luc, 0.01 μg of pRL-TK internal control plasmid (Promega), and 0.2 μg of DN TLRs expressing plasmid in 24-well plates. After 20 h, transfected cells were stimulated with LAMPs or purified proteins. After a further 24 h of incubation, cells were lysed and assayed for luciferase activity using a Dual-Luciferase Reporter Assay System (Promega). Both firefly and Renilla luciferase activity were monitored with a Lumat LB9507 luminometer (Berthold). Normalized reporter activity is expressed as the firefly luciferase value divided by the Renilla luciferase value. Relative fold induction is calculated as the normalized reporter activity of the test samples divided by the unstimulated samples.
Reversed phase HPLC
LAMPs were dissolved in 6 M guanidine hydrochloride, and 100 μg of LAMPs were applied on μBondasphere C18 300A (Waters). Elution was done with 0 to 90% linear water, 2-propanol gradient. The flow rate was 1.0 ml/min up to 65 min. Each fraction was dried in vacuo, using a centrifugal concentrator (Tomy) at room temperature, and dissolved in 25 mM N-octyl-β-glucopyranoside. Protein concentration was measured by using the Coomassie Protein Assay Regent (Pierce).
Purification from SDS-PAGE gel
One hundred micrograms of LAMPs were fractionated by reversed phase HPLC. The fraction eluted at 62 min was separated by 10% SDS-PAGE gels under reducing conditions. Gel was stained with Quick CBB Plus (Wako). To elute proteins, the stained gel was cut into 5-mm strips, and each strip was homogenized with 1% SDS solution. Homogenized gel was removed by centrifugation at 12,000 × g at room temperature. Five volumes of acetone were added to the supernatant followed by incubation overnight. Proteins were pelleted by at 12,000 × g at 4°C and dissolved in 100 μl of 0.1% SDS solution.
Treatment with protease and lipoprotein lipase
The fraction eluted at 62 min was obtained as described above. The band around 21 kDa stained with Quick CBB Plus were excised and sliced into small strips. The small strips were incubated with 100 μg/ml lipoprotein lipase or proteinase K (Sigma-Aldrich) at 37°C for 6 h. The protein was eluted from gel strips as described above.
Peptide mass fingerprinting
Peptide mass fingerprinting was conducted according to the method of Yoshino et al. (28). Briefly, the Quick CBB Plus-stained gel of the fraction eluted at 62 min was obtained as described above. The band around 21 kDa was excised and sliced into small strips. To remove Quick CBB Plus, the strips were incubated in 50% methanol, 5% acetic acid for 1 h and washed twice in water. The strips were dehydrated by incubation with 100% acetonitrile. To alkylate the protein, the strips were incubated at 60°C for 1 h with 10 mM DTT in 100 mM ammonium hydrogen carbonate followed by treatment at room temperature for 30 min with 55 mM iodoacetamide (Nacalai Tesque) in 100 mM ammonium hydrogen carbonate. In-gel trypsin digestion was conducted by incubation with 10 μg/ml trypsin (Promega). The digested peptides were eluted by 5% formic acid (Wako). Peptides were dried in vacuo and dissolved in saturated α-cyano-4- hydroxycinnamic acid (Nacalai Tesque) in 50% acetonitrile and 0.1% trifluoroacetic acid. The molecular weights of the peptides were measured with Autoflex (Bruker Daltnics). The database was searched by MASCOT (Science Matrix).
Mass spectrometry of intact protein
One hundred micrograms of LAMPs were fractionated by reversed phase HPLC. The fraction eluted at 62 min was dried using a centrifugal concentrator and suspended in 10 μl of saturated sinapic acid (Bruker Daltnics) in 50% acetonitrile and 0.1% trifluoroacetic acid. The molecular mass of protein was measured with Autoflex (Bruker Daltnics).
Induction of NF-κB by LAMPs
We initially determine whether LAMPs from M. pneumoniae can induce NF-κB in a monocytic cell line, THP-1. To detect the induction of NF-κB, THP-1 cells were transfected with a plasmid, in which the luciferase reporter gene was fused to NF-κB enhancer (pNF-kB-luc), and then stimulated with LAMPs. The level of luciferase expression was enhanced by LAMPs in a dose-dependent manner (Fig. 1). The level of luciferase expression was maximum at 1.0 μg/ml LAMPs and ∼40-fold higher than that of unstimulated control. These results indicate that LAMPs from M. pneumoniae can induce NF-κB in THP-1 cells.
Inhibition of NF-κB activation by anti-TLR2 mAb
A lipopeptide of M. fermentans, MALP-2, was reported to activate NF-κB through TLR2 (15). We then elucidated whether the activation of NF-κB with LAMPs is mediated through TLR2. Anti-human TLR2 mAb (IMG-416)-pretreated THP-1 cells were transfected with pNF-kB-luc followed by stimulation with LAMPs. The pretreatment with anti-TLR2 mAb decreased the expression level of luciferase, although control Ab (mouse IgG2a) failed to affect that of luciferase (Fig. 2). These results indicate that the activation of NF-κB by LAMPs is mediated by TLR2.
Activation of NF-κB through TLR2
To further confirm that activation of NF-κB by LAMPs is mediated through TLR2, we constructed a TLR2 expression vector (pFLAG-TLR2). 293T cells were transfected with both pFLAG-TLR2 and pNF-kB-luc. When 293T cells transfected with a high dose of pFLAG-TLR2 were stimulated with LAMPs, the levels of luciferase expression were augmented in a dose-dependent manner (Fig. 3). In contrast, the level of luciferase expression was as much as unstimulated control level when 293T cells were transfected with the empty vector pFLAG-CMV1. The results suggest that LAMPs induce NF-κB through TLR2.
Purification of active components of LAMPs
To purify the active components of LAMPs responsible for NF-κB activation, LAMPs were fractionated by reversed phase HPLC with a linear gradient of 2-propanol. To measure the activity of fractions to induce NF-κB, each fraction was added to 293T cells transfected with pFLAG-TLR2 and pNF-kB-luc. As shown in Fig. 4, the active component of LAMPs was eluted with ∼87% 2-propanol at 62 min. Hereafter, the fraction containing the active component is referred to as f62. The level of NF-κB activation by the f62 was ∼6.5-fold higher than that of control.
Analysis of active components of LAMPs
To measure the molecular mass of the active component of LAMPs, the f62 were separated by SDS-PAGE, and the gel containing f62 was excised into 12 pieces. Proteins extracted from each gel piece as described in Materials and Methods were incubated with 293T cells transfected with pFLAG-TLR2 and pNF-kB-luc. When the cells were incubated with the extracted protein around a band of 21 kDa (Fig. 5, arrow), the level of NF-κB activation by the band was maximum and ∼10 times higher than that of control. The result indicates that the molecular mass of the active component is ∼21 kDa.
Identification of a 21-kDa protein
To identify the 21-kDa NF-κB-inducing protein, peptide mass fingerprinting was conducted according to the method described in Materials and Methods. As the result of database search of SwissProt using MASCOT, the 21-kDa protein was matched to subunit b of F0F1-type ATPase of M. pneumoniae (Fig. 6,A). The protein score for the 21-kDa protein was 79; the scores >62 are significant (p < 0.05). The F0F1-ATPase was predicted to be molecular masses of 24,018 and 21,000 for native and processed forms, respectively. Because the F0F1-ATPase contains a signal peptide near the N terminus followed by a predicted lipid-binding cysteine and its transmembrane domain, it is supposed to be a lipoprotein (29, 30). To assume the number and length of acyl chains binding to the F0F1-ATPase, the molecular mass of intact protein was measured. As shown in Fig. 6 B, the molecular mass of the protein was 21,549. Therefore, the predicted molecular mass of acyl chains was 549, which was almost matched to two palmitic acids. These results suggest that the F0F1-ATPase of M. pneumoniae would contain two palmitic acids at the N-terminal lipid binding cysteine.
Cooperation of TLR2 and TLR6 for NF-κB activation
Diacylated lipopeptides, including MALP-2, were reported to be recognized by mouse TLR2 cooperatively with TLR6 (31). To investigate whether the F0F1-ATPase is also recognized by both TLR2 and TLR6 for NF-κB activation, we constructed a plasmid encoding DN TLR6 (pFLAG-dTLR6). 293T cells were transfected with pFLAG-TLR2, pNF-kB-luc, and pFLAG-dTLR6. The effect of DN TLR6 on the expression of TLR2 was analyzed by flow cytometry. The level of TLR2 expression was almost constant irrespective of the expression of DN TLR6 as well as DN TLR1 (32). When the cells transfected with control vector were stimulated with F0F1-ATPase, the level of NF-κB activation was increased ∼8-fold higher than that of control (Fig. 7). In contrast, the level of NF-κB activation was almost decreased down to the control level when the cells were transfected with DN TLR6. These results suggest that the NF-κB activation by F0F1-ATPase is dependent on both TLR2 and TLR6. To further determine whether TLR6 alone can mediate the activation of NF-κB by the F0F1-ATPase, 293T cells were transfected with TLR6 expression vector (pFLAG-TLR6). The transfected cells were stimulated with F0F1-ATPase, but the level of luciferase expression was not augmented (data not shown). These results indicate that both TLR2 and TLR6 cooperatively mediate the NF-κB activation by the F0F1-ATPase.
Cooperation of TLR1 and TLR2 for NF-κB activation
Triacylated bacterial lipopeptides such as Pam3CSK4 were reported to be recognized by murine TLR1 in association with TLR2 (16). We next determined whether F0F1-ATPase is recognized by both TLR1 and TLR2 for NF-κB activation. We transfected a plasmid encoding DN TLR1 (pFLAG-dTLR1) into 293T cells with both pFLAG-TLR2 and pNF-kB-luc. When the cells were transfected with control vector, the F0F1-ATPase stimulation augmented the level of NF-κB induction ∼8-fold higher than unstimulated control (Fig. 7). In contrast, on transfection with DN TLR1, the F0F1-ATPase treatment failed to augment NF-κB induction. These results indicate that the cooperation of TLR1 and TLR2 is required for the NF-κB activation by the F0F1-ATPase. As reported previously by Takeuchi et al. (16, 31), DN TLR1 and DN TLR6 suppressed the activity of Pam3CSK4 and MALP-2, respectively, to induce NF-κB (Fig. 7). Our data indicate that unlike Pam3CSK4 and MALP-2, the activation of NF-κB by the F0F1-ATPase is dependent on TLR1, TLR2, and TLR6 coincidently. NF-κB activation by LAMPs was strikingly inhibited by transfection with DN TLR1, but it was to some extent lowered by transfection with DN TLR6 (Fig. 7). The findings indicate that NF-κB activation by LAMPs is dependent on TLR1 and TLR2, but partially dependent on TLR6.
Determination of critical domain for NF-κB activation
To investigate the roles of the amino acid sequence of the F0F1-ATPase for NF-κB activation, it was digested by proteinase K. As shown in Fig. 8, treatment with proteinase K failed to decrease the activity of the F0F1-ATPase. Subsequently, to elucidate whether lipid moiety of the NF-κB is responsible for NF-κB activation, the acyl chains of the F0F1-ATPase were decoupled from protein domain with lipoprotein lipase. When the F0F1-ATPase was treated with lipoprotein lipase, the activation of NF-κB was apparently decreased (Fig. 8). These results suggest that lipid moiety but not protein moiety of the F0F1-ATPase would be highly critical for the NF-κB activation.
In this study, we demonstrated that the subunit b of F0F1-type ATPase from M. pneumoniae activates NF-κB through TLR1, TLR2, and TLR6. The F0F1-ATPase is shown to have a transmembrane domain and a lipid-binding region based on its amino acid sequence. Analysis of molecular mass showed that diacylated glycerol containing two palmitic acids would bind to the cysteine in the lipid-binding region of F0F1-ATPase, indicating that the F0F1-ATPase would be a diacylated lipoprotein.
In Mycoplasma spp., acylated proteins are abundant cell surface Ags, and many putative lipoprotein-encoding genes are identified in the sequenced Mycoplasma genomes (33, 34). It is at present controversial whether mycoplasmas have triacylated lipoproteins. In triacylated lipoproteins, three fatty acids are bound to its N-terminal cysteine residue: two in a diacylglyceride that is linked via a thioether bond to the sulfhydryl group (diacylglyceryl modification); and one to the amine group (N-acylation). Chemically identified lipoproteins from M. fermentans, Mycoplasma hyorhinis, M. salivarium, and Mycoplasma gallisepticum, are not N-acylated (6, 7, 35, 36), nor has an N-acyltransferase gene responsible for N-acylation been found in M. pneumoniae, Mycoplasma genitalium, and Mycoplasma penetrans genomes (37, 38, 39). To date, the presence of mycoplasmal proteins with N-acyltransferase activity has not been reported either. However, the study on the ratio of N-amide and O-ester bonds in M. gallisepticum and Mycoplasma mycoides indicates the presence of diacylated and triacylated lipoproteins (40). In addition, the resistance to Edoman degradation of proteins from M. mycoides also would uphold the presence of N-acylation (33).
It is widely known that triacylated lipoproteins are recognized by TLR1 and TLR2, whereas diacylated lipoproteins are recognized by TLR2 and TLR6. In TLR1-deficient mice, Pam3CSK4 containing three acyl chains failed to induce cytokines, including TNF-α, whereas MALP-2 containing two acyl chains could induce cytokines (16). In TLR6-deficient mice, MALP-2 failed to induce cytokines, but Pam3CSK4 could induce them (31). The findings suggest that the F0F1-ATPase, similar to MALP-2, would be recognized by TLR2 and TLR6. In fact, the F0F1-ATPase was recognized by TLR1, TLR2, and TLR6 coincidentally. More recently, Buwitt-Beckmann et al. (41) reported that Pam2CSK4 containing two acyl chains and amino acid-elongated MALP-2, MALP-2-SK4 can induce TNF-α in TLR6-deficient mice, indicating that peptide sequence as well as the whole molecular structure of lipoprotein would be more responsible for TLR recognition rather than the number of acyl chains. Therefore, in the present study, the amino acid sequence of the F0F1-ATPase may account for the use of TLR1, TLR2, and TLR6 for NF-κB activation.
According to the amino acid sequence of TLRs, TLR1 and TLR6 were closely related with 69% identity (42). Interestingly, excessive expression of DN TLR1 inhibited somewhat the induction of NF-κB by MALP-2 (data not shown). The findings suggest that TLR1 would be functionally also quite similar to TLR6, but the precise interaction between ligands, including lipoproteins, and TLR remains to be elucidated. As shown in Fig. 7, NF-κB activation by F0F1-ATPase was dependent on TLR1, TLR2, and TLR6. However, the NF-κB activation by LAMPs was partially dependent on TLR6. These results indicate that LAMPs would contain undefined components other than F0F1-ATPase, activating NF-κB through a TLR6-independent pathway.
M. pneumoniae causes primary atypical pneumonia, asthma, and other respiratory diseases in humans (2, 3, 4). Adherence of invading mycoplasmas to the respiratory epithelium and localized host cell injury are thought to be responsible for the pathogenesis in M. pneumoniae infections. In addition, excessive immune responses possibly would participate in the pathogenesis (5). However, such factors associated with the pathogenesis have not been identified in M. pneumoniae, although immune modulators, including MALP-2 and the 44-kDa lipoprotein, have been isolated from less pathogenic mycoplasmas such as M. fermentans (6) and M. salivarium (7), respectively. To our knowledge, this study is the first report that a lipoprotein derived from M. pneumoniae induces immune responses through NF-κB activation in monocytic cells.
More recently, Chu et al. (43) reported that M. pneumoniae induces airway mucin expression through TLR2. Mucin is considered to be a receptor of Mycoplasma spp., facilitating cytoadherence and colonization (44, 45). Moreover, the expression of mucin was mediated by NF-κB (46). These results indicate that the F0F1-ATPase also would induce mucin expression through TLR2, leading to participation in the onset of M. pneumoniae infection. Considering the interaction between TLR and the F0F1-ATPase, the suppression of the interaction may result in the reduction of inflammatory response induced by M. pneumoniae. Alternatively, our data suggest that the F0F1-ATPase is a potential therapeutic target. For example, the F0F1-ATPase and its derivatives might be a suitable candidate as Ags for vaccination, causing the induction of specific Ab capable of inhibiting TLR signaling. Moreover, the F0F1-ATPase with modified acyl chains would be a potential antagonist for the TLR signaling because unsaturated fatty acids were reported to inhibit TLR2 signaling induced by lipopeptides (47, 48). Thus, the lipoprotein, capable of interacting with TLR, derived from M. pneumoniae would be a potential molecule in the development of new weapons for prevention and therapy for M. pneumoniae infection.
We thank Dr. Matsumoto for the gift of MALP-2. We also thank Dr. Ishikawa for technical help in peptide mass fingerprinting.
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 in part by a Grant-in-aid from the Ministry of Education, Science, Sports and Culture of Japan.
Abbreviations used in this paper: MALP-2, macrophage-activating lipopeptide 2; LAMPs, lipid-associated membrane proteins; F0F1-ATPase, subunit b of F0F1-type ATPase; Pam3CSK4, (S)-[2,3-bis(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-(S)-Ser-(S)-Lys4-OH · 3HCl; MyD, myeloid differentiation protein; TBSE, TBS containing 1 mM EDTA; DN, dominant negative.