Xenobiotic incorporation into pyruvate dehydrogenase complex can occur via the exogenous lipoylation pathway


  • Potential conflict of interest: Nothing to report.


Lipoylated enzymes such as the E2 component of pyruvate dehydrogenase complex (PDC-E2) are targets for autoreactive immune responses in primary biliary cirrhosis, with lipoic acid itself forming a component of the dominant auto-epitopes. A candidate mechanism for the initiation of tolerance breakdown in this disease is immune recognition of neo-antigens formed by xenobiotic substitution of normal proteins. Importantly, sensitization with proteins artificially substituted with the lipoic acid analogue xenobiotic 6-bromohexanoic acid (6BH) can induce an immune response that cross-reacts with PDC-E2. This study investigated the potential of recombinant lipoylation enzymes lipoate activating enzyme and lipoyl-AMP(GMP):N-lysine lipoyl transferase to aberrantly incorporate xenobiotics into PDC-E2. It was found that these enzymes could incorporate lipoic acid analogues including octanoic and hexanoic acids and the xenobiotic 6BH into PDC-E2. The efficiency of incorporation of these analogues showed a variable dependence on activation by adenosine triphosphate (ATP) or guanosine triphosphate (GTP), with ATP favoring the incorporation of hexanoic acid and 6BH whereas GTP enhanced substitution by octanoic acid. Importantly, competition studies showed that the relative incorporation of both 6BH and lipoic acid could be regulated by the balance between ATP and GTP, with the formation of 6BH-substituted PDC-E2 predominating in an ATP-rich environment. Conclusion: Using a well-defined system in vitro we have shown that an important xenobiotic can be incorporated into PDC in place of lipoic acid by the exogenous lipoylation system; the relative levels of lipoic acid and xenobiotic incorporation may be determined by the balance between ATP and GTP. These observations suggest a clear mechanism for the generation of an auto-immunogenic neo-antigen of relevance for the pathogenesis of primary biliary cirrhosis. (HEPATOLOGY 2008;48:1874-1884.)

Primary biliary cirrhosis (PBC) is characterized by the presence of autoreactive immune responses, typically directed at members of the 2-oxo-acid dehydrogenase family of multi-enzyme complexes (2-OADC),1, 2 with specificity for components of one or more 2-OADC family members (pyruvate dehydrogenase complex [PDC], 2-oxoglutarate dehydrogenase complex [OGDC], and branched-chain oxo-acid dehydrogenase complex [BCOADC]) seen in over 95% of patients. The archetypal autoreactive response in PBC, present in over 90% of patients, is a high titre autoantibody response reactive with both the dihydrolipoamide acetyl-transferase (E2) and the dihydrolipoamide dehydrogenase (E3) binding protein (E3BP) components of PDC.3, 4 PDC-E2, PDC-E3BP and the other 2-OADC-E2 components to which immune-reactivity is seen in PBC share a highly conserved domain structure containing a lipoic acid cofactor.2, 5 This lipoic acid is bound through an amide linkage formed between its carboxylic acid group and the ϵ-amino group of a specific lysine residue and plays a fundamentally important role as a carrier of activated acyl groups between the active sites of the 2-OADC enzyme complexes.6, 7

Extensive characterization of the autoantibody response in PBC patients has suggested that, in addition to representing a key structural and functional motif in all the antigenic enzymes, the lipoic acid binding domain forms the core of the dominant autoepitope.4, 5, 8, 9 Moreover, lipoic acid forms an important component of the epitope, with substantially increased antibody affinity being seen against lipoic acid-containing (as opposed to nonlipoylated) recombinant antigens.10–12 It has been postulated recently that immunologic cross-reactivity between lipoic acid and structurally-related xenobiotics may represent a mechanism for the breakdown of tolerance to lipoic acid-containing autoantigens in PBC.1 PDC-E2-specific, lipoic acid-reactive antibodies characteristic of those seen in the sera of PBC patients are fully cross-reactive with a number of structural homologues of lipoic acid (including halogenated xenobiotics such as 6-bromohexanoic acid [6BH]).13 Furthermore, sensitization with 6BH conjugated to bovine serum albumin (BSA) induces high titre antibody responses reactive with self-PDC14 and the development of PBC-like liver lesions in experimental animals.15, 16 PDC can also act as a relevant protein carrier in this system as chemical biotinylation of murine-PDC (biotin has significant structural homology with LA and shares uptake receptors) renders it highly immunogenic17 in mice that are normally tolerant of self-PDC.18, 19 Although of great interest, the relevance of the xenobiotic cross-reactivity model for PBC pathogenesis in humans is limited by the fact that the artificial 6BH-containing conjugates required to induce reactivity in experimental systems would not be encountered in nature.

It is well established that the lipoylation of 2-OADC enzymes (that are all nuclear encoded) can occur through two interlinked pathways after the uptake of protein subunits by the mitochondria.2, 6, 7 The endogenous pathway uses lipoic acid synthetase (LIAS)20 to generate lipoic acid in situ from an octanoyl group that is transferred initially to the target enzyme from an octanoyl-ACP precursor.21, 22 The exogenous pathway directly uses lipoic acid that is efficiently scavenged in vivo from dietary sources by the intestinal sodium-vitamin transporters23, 24; significantly, there is evidence that a range of related lipoic acid-like molecules can compete with lipoic acid for efficient transport across intestinal cells.25 Free lipoic acid is incorporated into appropriate 2-OADC components through the sequential actions of lipoate activating enzyme (LAE) and lipoyl-AMP(GMP):N-lysine lipoyl transferase (LT).26, 27 Were structurally-related xenobiotics (such as 6BH) able to enter the exogenous lipoylation pathway in place of lipoic acid such aberrant lipoylation would represent a physiologic mechanism by which protein-xenobiotic constructs could arise in nature.

In the current study, we developed a robust in vitro model for physiologic lipoylation and used this to test the hypothesis that the exogenous lipoylation pathway is promiscuous, allowing 2-OADC enzyme subunits to be modified with substrates other than LA. This led to an examination of the molecular regulation of the incorporation into PDC-E2 of a range of relevant xenobiotic molecules, including 6BH.


6BH, 6-bromohexanoic acid; ILD, inner lipoyl domain; HA, hexanoic acid; LA, lipoic acid; LAE, lipoate activating enzyme; LIAS, lipoic acid synthetase; LT, lipoyl-AMP(GMP):N-lysine lipoyl transferase; OA, octanoic acid; PBC, primary biliary cirrhosis; PDC, pyruvate dehydrogenase complex; PDC-E2, dihydrolipoamide acetyl-transferase.

Materials and Methods

Plasmid Construction.

The mature LAE open reading frame was selectively amplified by polymerase chain reaction (PCR) from bovine normal liver cDNA (Biochain Institute Inc., Hayward, CA) by the use of the 5′ oligo CGCGGATCCGTTATCAGGAGCTGGAACCCTAAGATGG containing an engineered BamHI site and the 3′ oligo TTTTCTTTTTGCGGCCGCTCTTCTGACCAAACTCCTTCTTCC containing an engineered NotI site. The mature bovine LT open reading frame was amplified in the same way from a plasmid pET-BLT using the 5′ oligo CGCGAAGCTTGAACACAGTTAAAAGTGGACTCATTTTA containing an engineered HindIII site and the 3′ oligo GCGCCTCGAGCATTATTCCCTTAATTTTTTCACA containing an engineered XhoI site. The resulting LAE and LT products of 1.7 and 1.1 kb respectively were digested with appropriate enzymes and cloned into similarly cut pTrieX-3 Neo (Merck Chemical Ltd, Notts, UK) to make pLAE and pET21d (Merck) to make pLT.

The coding sequences for PDC-E2-inner lipoyl domain (PDC-E2-ILD) and a K-Q mutated PDC-E2-ILD (PDC-E2-ILD(K-Q), which lacks the physiologic lipoic acid binding lysine residue) were selectively amplified from plasmid pE211 and plasmid pGLIP-2T(Lys-Gln)28 respectively, by the use of the 5′ oligo CGCGGAGCTCGTAGCTCATATCCCCCTCACATGC containing an engineered SacI site and the 3′ oligo CGCGCTCGAGTTTTAAATCTGTTACTTCGG- TTGG containing an engineered XhoI site. The resulting 0.3 kb products were digested with SacI and XhoI and cloned into similarly cut pET21d to make pE2-ILD and pE2-ILD(K-Q). All these constructs (pLAE, pLT, pE2-ILD, and pE2-ILD(K-Q)) code for proteins carrying a C-terminal histidine tag. Preparation of plasmids, PCR, restriction digestion, isolation of DNA, ligation, and transformation were all carried out using standard methods. Confirmation that all constructs were correct was assessed by sequencing reactions carried out by Lark Technologies (Surrey, UK).

Expression and Purification of Recombinant Proteins.

Plasmids (pLAE, pLT, pE2-ILD or pE2-ILD[K-Q]) were transformed into Escherichia coli BL21(DE3)pLacI. Transformants were grown to midlog phase (A600 ∼ 0.6) in 500 mL Luria-Bertani media containing ampicillin (50 μg/mL) and chloramphenicol (34 μg/mL). Protein synthesis was induced by the addition of 1 mM isopropyl β-D-thiogalactopyranoside (IPTG) and the cultures incubated for a further 3 hours in a rotary shaker (180 RPM, 37°C). We routinely expressed LAE and LT in coculture. During incubation, 1-mL aliquots of the culture were taken at 10-, 60-, 120-, and 180-minute time points. After centrifugation and solubilization of the pellet in Laemmli sample buffer, samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunoblotting (as described below) with monoclonal mouse anti-His (Qiagen, Crawley, UK) (1:2,000). At 3 hours postinduction, the cells from a 500-mL culture were harvested by centrifugation (10,000g for 10 minutes) and resuspended in binding buffer (10 mM imidazole pH 7.2, 0.5 M NaCl, 20 mM Na2HPO4, protease inhibitor cocktail [Roche Diagnostic Ltd, UK]) containing the non-ionic detergent NP-40 (0.1% vol/vol). Cells were lysed by three freeze thaw cycles before sonication and the lysate was then clarified by centrifugation at 20,000g for 20 minutes, followed by filtration through syringe filters (0.2 μm). Lysates were applied to a 5-mL HiTrap Chelating column (GE Healthcare, Bucks, UK) precharged with Ni2+ and equilibrated in binding buffer (minus NP-40). After extensive washing in binding buffer, bound material was eluted in a linear 10 to 500 mM imidazole gradient (in binding buffer). Recombinant 6xHis tagged proteins were eluted in the Imidazole gradient and fractions collected and analyzed by SDS-PAGE followed by immunoblotting with monoclonal mouse anti-His. Fractions of interest were pooled and dialyzed/concentrated into PBS containing 10% (vol/vol) glycerol as a stabilizer, using Vivaspin 6 (5000MWCO) units (Sartorius Ltd, Surrey, UK). This resulted in stock preparations of LAE/LT, PDC-E2-ILD, and PDC-E2-ILD (K-Q) that were stored at −20°C until use. Recombinant PDC-E2-ILD was subsequently separated into their unlipoylated (ULip) and lipoylated (Lip) forms by anion exchange chromatography as described previously.10

Lipoylation Reaction.

The lipoylation of PDC-E2-ILD (ULip) was initially assessed when 10μg of the LAE/LT preparation and 10μg PDC-E2-ILD (ULip) were introduced into a 50-μl reaction mixture containing 20 mM Tris-HCl (pH 7.5), 40 mM potassium phosphate (pH 7.8), 4 mM MgCl2, BSA (0.3 mg/mL), and 4 mM GTP or ATP. The lipoylation substrate lipoic acid (DL-6,8-thioctic acid; Sigma-Aldrich Company, Dorset, UK) (or hexanoic [HA], octanoic [OA], or 6-bromohexanoic acid [6BH]; all Sigma) was added at concentrations ranging from 10 μM to 1 mM. Using the above conditions with 100 μM lipoic acid, a time course of reaction carried out at 37°C ranging from 0 to 300 minutes was set up to determine the point of full lipoylation. On completion, all reactions were terminated by boiling for 1 minute followed by processing for immunoblotting or enzyme linked immunosorbent assay (ELISA) analysis. As a result of the findings from the time course study, reactions were subsequently carried out routinely for 5 hours at 37°C. Competition assays (using GTP or ATP) were carried out as above in which the lipoic acid concentration was kept constant (10 or 100 μM) with 6BH being introduced in increasing concentration (0 to 1,000 μM). Additional reactions using benzoic acid, ibuprofen, salicylic acid, acetylsalicylic acid, 2-octynoic acid, 2-(trifluoromethyl) cinnamic acid, or 5-(2,4-dichlorophenyl)-2-furoic acid and biotin (all Sigma) were also carried out using GTP or ATP.

Immunoblotting Analysis.

The expression and purification of the recombinant enzymes was analyzed after preparations were boiled in Laemmli sample buffer and resolved electrophoretically on 18% SDS-PAGE. Preparations of native lipoyl domain or product identification after lipoylation reactions were analyzed routinely after suspending samples in nondenaturing sample buffer and subjecting the samples to PAGE on 15% nondenaturing gels. Separated proteins were transferred electrophoretically to Immobilon-P nitrocellulose membrane (Millipore, Herts, UK). Membranes were blocked with 5% (wt/vol) skimmed milk powder in PBS (PBSB) for 1 hour and the membranes washed in PBS + 0.05% (vol/vol) Tween 20 (PBSTw). Appropriate anti-PDC antibodies (IgG fraction affinity-purified from the pooled serum of PBC patients) and anti-lipoic acid antibodies (Merck Chemicals Ltd) were diluted in PBSB + 0.05% (vol/vol) Tween 20 (PBSBTw), and incubated for 2 hours at room temperature. After extensive washing in PBSTw, bound antibodies were detected in a 1-hour incubation using appropriate peroxidase-conjugated secondary antibodies diluted (1:5,000) in PBSBTw. Again after extensive washing, peroxidase reactivity was detected by enhanced chemiluminescence (Perbio Science Ltd, UK).

ELISA Analysis.

Analysis of product after the lipoylation reaction was also carried out by standard ELISA methodology. Briefly, fractions from the purification of PDC-E2-ILD or whole lipoylation reactions were diluted in carbonate-coating buffer (35 mM NaHCO3, 15 mM Na2CO3 pH 9.6) and incubated in 96-well microtiter plates (Immulon 4HBX, Fisher Scientific Ltd, Leicester, UK) for 16 hours at 4°C. The coated plate was then blocked with 5% BSA (wt/vol) in PBS (150 μL) for 1 hour and the plate washed in PBSTw. Appropriate primary antibodies including individual PBC patient sera (n = 9) were diluted in PBS containing 0.5% BSA, applied to wells (100 μL) and the plate was incubated for 2 hours at room temperature. After extensive washing, appropriate peroxidase-conjugated secondary antibodies were diluted (1:5,000) and applied (100 μL) to the plate for 1 hour at room temperature. After washing, bound peroxidase activity was visualized after reaction with o-phenylene diamine (OPD) solution (100 μL) (0.4 mg/mL OPD, 50 mM citrate, pH 5.0 and 0.012% [vol/vol] H2O2). The reaction was stopped after 30 minutes by the addition of 100 μL 2 M H2SO4, and absorbance at 492 nm measured.

Mass Spectrophotometry Analysis.

To confirm the identity of the groups incorporated by the actions of the lipoylation pathway enzymes bulk reactions were set up and then resolved by anion exchange chromatography. This also served to eliminate any residual LAE/LT or BSA from these reactions that may interfere with MS analysis. After initial identification of eluted proteins by ELISA and western blot using a combination of anti-PDC, anti-LA, and anti-His samples, peak fractions of interest were submitted for whole protein MS analysis using a LTQ-FT mass spectrometer (Thermo).

Expression of LAE and LT.

TaqMan gene expression assays for LAE (Hs00369504m1) and LT (Hs00376962m1) (Applied Biosystems, Warrington, UK) were used with β-actin as a reference gene to measure the relative expression of these components of the exogenous lipoylation pathway in the hepatocellular carcinoma cell line HepG2, the immortalized human intrahepatic biliary epithelial cell line H69,29 HEK293, HeLa, Jurkat lymphoblastoid cells, and the endometrial cell line HEC1B. RNA was purified using Tri-reagent (Sigma) and cDNA produced with AffinityScript Multiple Temperature Reverse Transcriptase (Stratagene). Real-time PCR was carried out using the ABI Prism 7000 Sequence Detection System.


Overexpression and Purification of Proteins.

LAE, LT, and PDC-E2 inner lipoyl domain (PDC-E2-ILD, the region of PDC-E2 containing the dominant PBC-specific auto epitope) were overexpressed in E. coli. (Fig. 1A) and purified using of nickel-based affinity chromatography (Fig. 1B). PDC-E2-ILD was purified in both wild-type form and after site-directed mutagenesis designed to substitute the LA-binding lysine residue (PDC-E2-ILD [K-Q]). The protein preparations obtained were over 90% pure when assessed by Coomassie Brilliant Blue staining (data not shown). Proteins of approximately 60 kDa and 40 kDa were obtained that correspond to LAE and LT respectively (calculated weights 62158 and 40773 Da). PDC-E2-ILD (∼14 kDa) was further purified into its unlipoylated and lipoylated forms by anion exchange chromatography. Fractions were assessed by ELISA analysis using anti-PDC (affinity-purified from the pooled serum of PBC patients), anti-Histidine Tag (anti-His), and anti-lipoic acid antibodies and nondenaturing PAGE (Fig. 1C, inset) that allowed for the isolation of unlipoylated domain (ULip, slower migrating band) for use as the substrate in the lipoylation system. The observable difference in gel migration between the two protein bands reflects the loss of positive charge caused by lipoic acid binding to a positively-charged lysine residue. Importantly, the purified unlipoylated preparation showed no reactivity on ELISA or immunoblotting when probed with anti-lipoic acid (data not shown). On analysis of the mutant PDC-E2-ILD (K-Q) after anion exchange chromatography, ELISA showed reactivity with anti-PDC and anti-His only (confined to a single peak) with no anti-lipoic acid reactivity (data not shown). This was confirmed by a lack of reactivity on immunoblotting with anti-lipoic acid (data not shown). This highlights the fact that the specific lysine residue with the motif EIETDKATIG is essential for lipoylation of this domain.

Figure 1.

Overexpression and purification of proteins. (A) Time-course (overinduction) of overexpression for combined LT/LAE samples and PDC-E2-ILD. Molecular weight markers are indicated. (B) Purification of LT and LAE using nickel-based affinity chromatography. (C) Isolation of unlipoylated (ULip) and lipoylated (Lip) domain. Fractions eluted from the 10-300 mM ammonium acetate gradient of a MonoQ purification of PDC-E2-ILD were analyzed by ELISA using anti-PDC (▪) (1;10,000), anti-His (▴) (1:2,000), and anti-lipoic acid (•) (1:2,000). Fraction 13 contains the Lip domain variant (reactivity seen with both anti-PDC and anti-lipoic acid) whereas fraction 10 contains the ULip variant (no reactivity seen with anti-lipoic acid). Staining after nondenaturing PAGE is also shown for these peak fractions, highlighting ULip and Lip forms (inset).

In Vitro Lipoylation of PDC-E2-ILD.

The combination of LAE and LT in the presence of GTP (as an energy source, previous observations having suggested that purified native bovine LAE has a GTP energy source preference),26 lipoic acid, and PDC-E2-ILD (ULip) was sufficient for effective lipoylation to take place. This was first shown after nondenaturing PAGE and western blotting studies using anti-PDC antibody, which allowed for detection of both unlipoylated and lipoylated PDC-E2-ILD (Fig. 2A). Evidence of lipoylation is seen with the increasing presence of a lower (lipoylated) reactive band as early as ten minutes into the reaction, with full lipoylation being observed at 5 hours. Accompanying this was a concomitant reduction in the upper, unlipoylated protein. When anti-lipoic acid antibody was used in immunoblotting studies against these reactions products, the formation of the lipoylated protein could be seen. The rate of lipoylation could also be shown in ELISA using the anti-lipoic acid antibody (Fig. 2B). Lipoic acid was confirmed as the incorporated group by MS analysis after purification of PDC-E2-ILD from large scale reactions (Fig. 2C). Effective lipoylation required the presence of all of LAE and LT, GTP, lipoic acid, and PDC-E2-ILD (ULip) (Fig. 2D). When LAE and LT were purified separately and used independently in mixing studies it was found that both are required in the assay to achieve lipoylation (data not shown). Furthermore, lipoylation did not occur if the lipoic acid binding lysine residue of PDC-E2-ILD was mutated to a glutamine residue (PDC-E2-ILD [K-Q]) by site-directed mutagenesis (Fig. 2E). We then went on to confirm that the lipoylation reaction was being catalyzed by mammalian LAE and LT within the recombinant protein fractions rather than passenger E. coli lipoylation enzymes. This was demonstrated by showing that the equivalent fractions to those containing pLAE and pLT products when affinity purified from control vector transformed E. coli extracts were unable to mediate lipoylation (data not shown). Taken together, these observations confirm that the in vitro lipoylation system described is functional in terms of the insertion site, and requires the presence of active LAE and LT.

Figure 2.

In vitro lipoylation of PDC-E2-ILD. Lipoylation of unlipoylated PDC-E2 inner lipoyl domain (PDC-E2-ILD ULip) was assessed when 50 μg of a recombinant LAE/LT preparation and 50 μg recombinant PDC-E2-ILD (ULip) was introduced into a 250-μL reaction mixture containing 20 mM Tris-HCl (pH 7.5), 40 mM potassium phosphate (pH 7.8), 4 mM MgCl2, BSA (0.3 mg/mL), and 4 mM GTP. Lipoic acid was added at 100 μM and incubated at 37°C for the indicated time, after which 25 μL of the reaction was terminated by boiling for 1 minute. Reactions were split with one resuspended in nondenaturing sample buffer for PAGE and the other diluted in Bicarbonate coating buffer (pH9.6) for ELISA. (A) Immunoblot after nondenaturing PAGE using anti-PDC (1:10,000) (upper panel) and anti-lipoic acid (1:2,000) (lower panel). (B) ELISA analysis using anti-lipoic acid (1:2,000). (C) Mass spectrometry analysis of identity of PDC-E2-ILD reaction products resolved from lipoylation cultures carried out in the absence (upper panel) and presence (lower panel) of lipoic acid. Unmodified recombinant PDC-E2-ILD has a predicted average molecular mass (Mr) of 14271.35Da. Analysis of the LA-free reaction product (upper panel) showed a single component with a Mr of 14270.35 Da. Analysis of the product of the reaction carried out in the presence of LA showed a single major component with a Mr of 14,459.40 Da indicating the additional presence of a single lipoyl residue (Mr 188.03 Da). Effective lipoylation failed to occur if the lipoylation reaction lacked (D, lane 1) GTP; (D, lane 2) lipoic acid; and (D, lane 3) LAE/LT. Full reaction in (D, lane 4). (E) The absence of lipoylation (detected in this instance by ELISA) when the predicted lysine lipoic acid binding residue in PDC-E2 ILD is substituted by glutamine (K-Q, compared to wild-type WT) suggests that lipoic acid incorporation is physiological in this system.

Effect of the Nucleotide Balance on the Lipoylation Reaction.

Having established the parameters and system requirements for the in vitro replication of physiological lipoylation, we went on to study the effect of the GTP/ATP balance in the system (of importance given the apparent need for the presence of GTP for lipoylation to occur but the likelihood that GTP would only ever be encountered in the intracellular setting in the simultaneous presence of ATP). Optimized reactions containing 4 mM GTP and 100 μM LA show full conversion to the lipoylated form (Fig. 3, panel i). When ATP levels are increased there is an inhibition of lipoylation that is evident starting at equimolar (4 mM) ATP concentration. In the converse reaction where ATP is kept constant at 4 mM it can be seen that in the absence of GTP there is incomplete lipoylation. Increasing the levels of GTP in these reactions did not seem, however, to change the level of lipoylation (Fig. 3, panel ii). These findings suggest that the presence of ATP reduces (but does not abrogate) lipoylation; an effect that cannot seemingly be reversed by increasing the relative levels of GTP.

Figure 3.

Effect of the GTP/ATP (nucleotide) balance on the lipoylation reaction. Reactions were carried out using 100 μM lipoic acid with constant (4 mM) GTP concentration and increasing ATP levels (panel i) or constant (4 mM) ATP and increasing GTP levels (panel ii). The optimal reaction, containing 4 mM GTP and 100 μM LA in the absence of ATP, showed full conversion to the lipoylated form (panel i). However, when ATP levels are increased there is an inhibition of lipoylation that is evident starting at equimolar (4 mM) ATP concentration. In the converse reaction where ATP is kept constant at 4 mM it can be seen that in the absence of GTP there is incomplete lipoylation. Increasing the levels of GTP in these reactions did not seem to change the level of lipoylation (panel ii). Below each western blot panel is a graph representing densitometric band intensity. Values (in arbitrary units) for the modified (lower band) and unmodified (upper band) domains are shown as black and grey bars, respectively.

Fidelity of the Lipoylation Reaction.

When octanoic acid was used in place of lipoic acid (Fig. 4A), a significant level of incorporation was seen at 100 μM in the presence of GTP (but not as pronounced as lipoic acid) and complete incorporation at 1,000 μM. Incorporation was negligible even at 1,000 μM when ATP was the nucleotide source. Conversely, hexanoic acid, although still being incorporated significantly, required the presence of ATP (Fig. 4B). These observations establish the twin principles that the exogenous lipoylation pathway is not restricted to using lipoic acid as a substrate, and that aberrant substrate incorporation is at least regulated partially by nucleotide balance in the reaction environment.

Figure 4.

The fidelity of the lipoylation reaction. Reactions were set up as before substituting lipoic acid for increasing concentrations (1-1,000 μM) of (A) octanoic acid or (B) hexanoic acid. To assess nucleotide use, reaction groups were carried out with either 4 mM GTP or 4 mM ATP. With octanoic acid (A), complete incorporation at 1,000 μM was seen in the presence of GTP. Incorporation was negligible at 1,000 μM when ATP was the nucleotide source. Hexanoic acid was still incorporated but required the presence of ATP (B). Incorporation of octanoic acid and hexanoic acid in the relevant studies is indicated by the presence of modified PDC-E2-ILD (upper panel) in the absence of lipoic acid binding (lower panel).

Further experiments were carried out to measure potential incorporation of xenobiotic molecules that serve as substrates for LAE, including benzoic acid, ibuprofen, salicylic acid,30 or have reactivity with PBC sera, including 2-octynoic acid, 2-(trifluoromethyl) cinnamic acid, or 5-(2,4-dichlorophenyl)-2-furoic acid.31 However, neither these molecules nor, additionally, biotin or acetylsalicylic acid were incorporated in reactions containing either GTP or ATP (Supporting Figure 1).

Incorporation of the Xenobiotic 6-Bromohexanoic Acid into PDC-E2-ILD.

When 6BH was used in the optimized reaction (GTP but no ATP) in the place of lipoic acid, significant amounts of altered domain are detected with anti-PDC (Fig. 5A, panel i, lanes 9-11). The altered domain does not react with anti-lipoic acid (Fig. 5A, panel ii, lanes 9-11) indicating that the protein is altered but not lipoylated (thus likely to be bound with 6BH). Incorporation was at a lower level than that seen in parallel studies for an equimolar concentration of lipoic acid (compare lanes 5 and 10 in Fig. 5A) indicating that lipoic acid is a more efficient substrate than 6BH under these conditions.

Figure 5.

Incorporation of the xenobiotic 6BH into PDC-E2-ILD and competition with lipoic acid. Single compound reactions were set up as before with increasing concentrations of 6BH (10-1,000 μM) in the presence of (A) GTP or (B) ATP (lanes 9-11). A single reaction with high 6BH concentration (1,000 μM) but no GTP or ATP as an energy source is also shown (lane 12). Increasing incorporation of 6BH (lower band) can be detected using anti-PDC, which is not seen with anti-lipoic acid as the ambient 6BH concentration is increased (lanes 9-11). Although incorporation of 6BH was seen using GTP as an energy source, significantly more efficient incorporation was seen when ATP was used. In dual compound reactions, 6BH can directly compete with lipoic acid for incorporation into PDC-E2-ILD. Competition assays using either (A) GTP or (B) ATP were set up with two constant levels of lipoic acid (10 μM, lanes 1-4; 100 μM, lanes 5-8). Increasing concentrations of 6BH (0-1,000 μM) (lanes 1-4 and 5-8) were added to the reaction and allowed to compete. Using GTP the presence of increasing concentrations of 6BH reduces incorporation of 10 μM lipoic acid as detected by diminishing anti-LA reactivity (A, lanes 1-4). However, lipoic acid incorporation remains predominant. In contrast, using ATP, 6BH can be competitive and completely suppress incorporation of 10μM lipoic acid (B, lane 4) and partially suppress that of 100 μM lipoic acid (B, lanes 7, 8). (C) The formation of 6BH substituted PDC-E2-ILD was validated by mass spectrometry that showed a major component having a Mr of 14,449.36 Da corresponding unequivocally to the addition of a 6BH moiety (Mr 177.06 Da).

Of even greater physiologic relevance is the extent to which 6BH incorporation can occur in the presence of ambient lipoic acid. In competition studies we showed that, as the ambient concentration of 6BH is increased, lipoic acid incorporation into PDC-E2-ILD is reduced progressively (Fig. 5A, panel ii, lanes 1-4 [low ambient lipoic acid concentration]) whereas accumulation of modified PDC-E2-ILD continues (Fig. 5A, panel i, lanes 1-4). This suggests that another moiety (6BH) is, under these conditions, incorporated in competition with lipoic acid. Higher concentrations of lipoic acid (Fig. 5A, lanes 5-8) suppress 6BH incorporation in the context of GTP. In contrast to the pattern seen with lipoid acid, 6BH incorporation was increased when the energy source was changed from GTP to ATP (comparison between Fig. 5A and 5B, panel i, lanes 9-11). Competition assays carried out at 10 and 100 μM LA with increasing concentrations of 6BH (0-1,000 μM) indicate that this group competes more efficiently with LA in the presence of ATP (Fig. 5B, panels i and ii, lanes 1-4 and 5-8, respectively). Indeed, 6BH can completely suppress incorporation of 10 μM LA (Fig. 5B, panel ii, lane 4) and partially suppress that of 100 μM LA (Fig. 5B, panel ii, lanes 7 and 8) after detection with anti-lipoic acid. Mass spectrometry after resolution by anion exchange chromatography formally validated the reaction product from the 6BH containing reaction as 6BH-PDC-E2-ILD (Fig. 5C).

These observations confirm our hypothesis that 6BH can enter the exogenous lipoylation pathway and undergo incorporation into PDC-E2-ILD, and that, under certain energetic conditions (an excess of ATP), incorporation of the xenobiotic can predominate.

When 6BH modified domain was probed with individual PBC patient sera there was no difference in immunogenicity observed when compared to the natural lipoylated form of the domain. This was also the case for HA and OA modified domain (Supporting Figure 2).

Competitive Incorporation of Lipoic Acid & 6BH and the Effect of GTP and ATP Levels.

Having shown that 6BH can be incorporated most efficiently under conditions in which ATP is present and GTP absent we went on to look at the effect of having both nucleotides present, and the role played by the balance between them in controlling 6BH incorporation. As shown in Fig. 6A (panel i) reactions containing 4 mM GTP and 100 μM 6BH show partial conversion to the 6BH-modified form. When ATP levels are increased in the presence of a fixed GTP concentration, the degree of 6BH incorporation seems to increase, becoming evident at 0.4 mM with full conversion at 4 mM ATP. In the converse reaction, where ATP is kept constant at 4 mM, full 6BH incorporation is seen at all concentrations of GTP (Fig. 6A, panel ii). These observations suggest that the effect of ATP on promoting 6BH incorporation has predominance over the GTP effect. The absence of reactivity on probing with anti-LA in the context of a band suggestive of the presence of altered PDC-E2-ILD on probing with anti-PDC confirms the likely incorporation in these studies of 6BH.

Figure 6.

Competitive incorporation of LA and 6BH into PDC-E2-ILD and the effect of GTP and ATP levels. (A) In an initial experiment, reactions were set including 100 μM 6BH (but no lipoic acid) to study (panel i) The effects of constant GTP (4 mM) and increasing ATP (0-40 mM) and (panel ii) the effects of constant ATP (4 mM) and increasing GTP (0-40 mM). Further reactions were then carried out that contained a four-way combination of both substrates: (B) 100 μM lipoic acid and 100 μM 6BH; and (C) 100 μM lipoic acid and 1,000 μM 6BH) and both nucleotide sources present at different ratios (GTP, panel i; ATP, panel ii).

Finally, we studied the most physiologic scenario, in which both substrates (lipoic acid and 6BH) and both energy sources (GTP and ATP) are present in the reaction mixture. Under conditions of equimolar LA and 6BH (both 100 μM) (Fig. 6B) with 4 mM GTP (Fig. 6B, panel i) all the unlipoylated PDC-E2-ILD becomes modified as judged by probing with anti-PDC. A proportion of this is due to lipoylation as detected with anti-lipoic acid, and this remains constant until ATP levels are raised to 4 mM. At this point, the proportion of lipoylation decreases with a concomitant incorporation of 6BH (deducible as there is no change in signal when probed with anti-PDC). At 40 mM ATP there is no visible lipoylation with all PDC-E2-ILD being modified with 6BH. In the converse GTP/ATP situation, under conditions of 4 mM ATP (Fig. 6B, panel ii) there is a lower proportion of lipoylation (and therefore more 6BH modification) compared to 4 mM GTP. This corroborates the previous observation that 6BH is favored by LAE/LT under conditions of ATP excess (Figs. 5, 6A). As GTP levels are increased there is no change in the level of lipoylation (and presumably 6BH modification) until a slight increase at 40 mM GTP. The same experiment was carried out with a 10-fold excess of 6BH (1,000 μM) and the overall effect is an enhanced incorporation of 6BH at lower levels (0.04 mM) of ATP (with GTP constant at 4 mM) (Fig. 6C, panel i) and complete incorporation with all conditions containing 4 mM ATP (Fig. 6C, panel ii). These observations suggest that under certain energetic conditions (ATP>GTP in the microenvironment) 6BH incorporation into PDC-E2-ILD by way of the exogenous lipoylation pathway can occur in preference to the physiologic substrate lipoic acid.

Relative Expression Levels of LAE and LT in Cell Lines.

To confirm the potential for the exogenous lipoylation pathway to function in relevant tissues, a series of experiments were carried out to measure the relative expression of LAE and LT in a range of cell lines by real-time PCR. Interestingly, when normalized to Jurkat lymphoblastoid cells, there was a high level of expression of LAE in both the hepatocellular carcinoma cell line HepG2, and the intrahepatic biliary epithelial derived cell line H69, with moderate expression also observed in the kidney cell line HEK293. Relative levels of LT were more constant suggesting that LAE may limit the function of the exogenous pathway in some cells (Supporting Figure 3).


Lipoic acid is a sulfur-containing cofactor found in many prokaryotic and eukaryotic organisms that is used by the H-protein of the glycine cleavage system, PDC-E3BP, and the E2 subunits of the 2-OADCs (together with PDC-E3BP).6 In each of these proteins, lipoic acid forms an amide linkage through its carboxylic acid moiety to a lysine residue. During catalysis by these dehydrogenases, protein bound lipoamide serves as a carrier of activated acyl groups between the active sites of these multienzyme complexes.7 Mitochondria (and chloroplasts) are the only cellular organelles where lipoic acid-containing enzymes are present. In addition, all lipoic acid-containing enzymes are nuclear encoded, undergoing post-translational active uptake (although still in an unlipoylated form) into the mitochondria where lipoylation takes place.6, 7

In the first stage of the eukaryotic exogenous lipoylation pathway, lipoic acid is activated by LAE using ATP or GTP to form lipoyl-AMP or lipoyl-GMP.26 The high levels of expression of LAE observed in liver and embryonic kidney-derived cell lines is consistent with a previous report that this transcript is expressed in liver and kidney but not in heart or skeletal muscle.32 It is now known that LAE additionally functions as a medium-chain acyl-CoA synthetase.32 The second stage is transfer by LT of the activated lipoyl moiety to the ϵ-amino group of lysine within an ExDKA sequence motif on unlipoylated apoproteins.27 Previous reports have described the specificity and function of purified LAE and LT enzymes in isolation.27, 30 However, the current study shows for the first time that combination of recombinant LAE and LT can transfer lipoic acid to mammalian PDC-E2-ILD in vitro. Furthermore, it is clear that the reaction proceeds most efficiently in the presence of GTP and Mg2+; indeed, addition of ATP in the presence of GTP led to a reduction in the incorporation of lipoic acid. These results are similar to observations made using enzymes purified from bovine liver and are consistent with the suggestion that lipoyl-GMP is more readily released than lipoyl-AMP from LAE, providing a more suitable substrate for LT.26

Previous studies have shown that the exogenous lipoylation pathway has significant substrate promiscuity, particularly for hexanoic, octanoic, and decanoic acids. The current work extends this observation to include demonstration of the incorporation into PDC-E2 of the xenobiotic molecule, 6BH. Indeed, it was found that 6BH can be incorporated in preference to the natural lipoic acid substrate when both molecules are present in the reaction medium. In this case, the relative incorporation of each molecule is determined not only by their concentration but also by the local GTP/ATP ratio, with GTP favoring lipoylation and ATP promoting the incorporation of 6BH. Interestingly, the effect of ATP in promoting 6BH incorporation dominates the pro-lipoic acid incorporation effects of GTP, as increasing the concentrations of GTP in a mixed GTP/ATP containing reaction fail to suppress 6BH incorporation.

This study suggests a clear route to physiologic incorporation of 6BH into PDC-E2. This is of considerable relevance given previous reports showing that sensitization with biotin-substituted PDC-E217 or 6BH-modified BSA16 can both lead to portal tract inflammation. Importantly, there was no evidence that 2-octynoic acid can be incorporated by the exogenous lipoylation pathway (see Supporting Figure 1). This latter observation is of particular interest given the recent observation that sensitization of mice with an antigen formed by chemical incorporation of 2-octynoic acid into BSA can induce mild inflammation of the portal tract.33 The failure of the exogenous lipoylation pathway to incorporate 2-octynoic acid suggests that it is unlikely that this xenobiotic molecule can compete physiologically with lipoic acid for incorporation into subunits of members of the 2-OADC enzyme family.

The findings of the current study build on important recent data suggesting that 6BH containing proteins can trigger a PBC-like disease process,14, 16 by showing that 6BH-containing proteins can arise naturally, and may do so in the form of PDC-E2 homologues. The findings of these studies provide further evidence potentially in support of an environmental-exposure model for PBC pathogenesis. The simplest model would be cumulative exposure to environmental 6BH, which undergoes progressive incorporation into PDC and related enzymes in place of lipoic acid. The facts that 6BH is an industrial reagent, and that PBC cases are subject to significant environmental clustering, with the highest prevalence of disease being seen in postindustrial and contaminated environments would both be supportive of this model.34, 35 The induction of PDC-reactive autoimmunity by way of such a mechanism would, however, require a number of key additional factors to be present. These include uptake into the body of environmental 6BH, the appropriate ATP-rich environment to favor its incorporation (the factors regulating ATP/GTP balance are at present surprisingly unclear), expression of the enzymes of the exogenous lipoylation pathway and an inflammatory response to trigger immune tolerance breakdown (induction of autoreactivity through 6BH cross-reactivity in guinea pigs requires the presence of adjuvant).14, 16 The requirement for multiple factors to be present if immune tolerance is to breakdown by way of this pathway might go some way to explaining the relative rarity of PDC-auto immunogenicity in the population.


We thank Dr. K. Fujiwara (Institute for Enzyme Research, University of Tokushima, Japan) for the generous gift of the plasmid pET-BLT that was used to derive His Tagged recombinant lipoyltransferase. We also thank Dr. D.M. Jefferson (New England Medical Centre, Tufts University, Boston, MA) for the gift of the immortalized human cholangiocyte cell line, H69.