Skip to main content
  • Loading metrics

Identification of a novel lipoic acid biosynthesis pathway reveals the complex evolution of lipoate assembly in prokaryotes

  • Tomohisa Sebastian Tanabe,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Validation, Visualization, Writing – original draft

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Martina Grosser,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Validation

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Lea Hahn,

    Roles Data curation, Investigation

    Current address: Institut für Integrierte Naturwissenschaften, Arbeitsgruppe Mikrobiologie, Universität Koblenz, Koblenz, Germany

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Carolin Kümpel,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Validation

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Hanna Hartenfels,

    Roles Investigation

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Evelyn Vtulkin,

    Roles Investigation

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Wanda Flegler,

    Roles Investigation

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany

  • Christiane Dahl

    Roles Conceptualization, Formal analysis, Funding acquisition, Project administration, Supervision, Validation, Writing – original draft

    Affiliation Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Bonn, Germany


Lipoic acid is an essential biomolecule found in all domains of life and is involved in central carbon metabolism and dissimilatory sulfur oxidation. The machineries for lipoate assembly in mitochondria and chloroplasts of higher eukaryotes, as well as in the apicoplasts of some protozoa, are all of prokaryotic origin. Here, we provide experimental evidence for a novel lipoate assembly pathway in bacteria based on a sLpl(AB) lipoate:protein ligase, which attaches octanoate or lipoate to apo-proteins, and 2 radical SAM proteins, LipS1 and LipS2, which work together as lipoyl synthase and insert 2 sulfur atoms. Extensive homology searches combined with genomic context analyses allowed us to precisely distinguish between the new and established pathways and map them on the tree of life. This not only revealed a much wider distribution of lipoate biogenesis systems than expected, in particular, the novel sLpl(AB)–LipS1/S2 pathway, and indicated a highly modular nature of the enzymes involved, with unforeseen combinations, but also provided a new framework for the evolution of lipoate assembly. Our results show that dedicated machineries for both de novo lipoate biogenesis and scavenging from the environment were implemented early in evolution and that their distribution in the 2 prokaryotic domains was shaped by a complex network of horizontal gene transfers, acquisition of additional genes, fusions, and losses. Our large-scale phylogenetic analyses identify the bipartite archaeal LplAB ligase as the ancestor of the bacterial sLpl(AB) proteins, which were obtained by horizontal gene transfer. LipS1/S2 have a more complex evolutionary history with multiple of such events but probably also originated in the domain archaea.


α-Lipoic acid is a cofactor found in all domains of life and is involved in key reactions of central carbon metabolism and dissimilatory sulfur oxidation [14]. In this eight-carbon saturated fatty acid, sulfur atoms replace the hydrogen atoms of carbons 6 and 8 of the acyl chain [5]. Only a few, but particularly important, lipoic acid-dependent enzyme systems have been described [2,3,6]. These include 3 α-ketoacid dehydrogenases, such as pyruvate dehydrogenase, whose E2 subunits bind lipoic acid. In the glycine cleavage complex, lipoate is bound to the glycine cleavage H protein (GcvH) [3]. Lipoylated proteins also play an important role in combatting reactive oxygen species [7,8]. In 2018, we discovered another lipoate-binding protein (LbpA) homologous to GcvH (S1 Fig) and demonstrated that it is an essential component of the sulfur-oxidizing heterodisulfide reductase-like (sHdr) system present in a wide range of bacterial and archaeal dissimilatory sulfur oxidizers [2,9].

Two posttranslational machineries are known to construct lipoyl moieties [1,3]: The first requires an acyl carrier protein (ACP)-bound octanoyl residue from endogenous fatty acid biosynthesis to be transferred to the ε-amino groups of conserved lysine residues in the accepting apo-proteins. In the second, free lipoate or octanoate are hooked up to the target lysine. Irrespective of the initial step, a sulfur atom must be added to each of the octanoyl C6 and C8 atoms to complete lipoate biosynthesis (Fig 1A). Using free precursors involves the enzyme lipoate:protein ligase that activates the precursors to lipoyl- or octanoyl-AMP at the expense of ATP before transfer to the target protein. In many bacteria, including Escherichia coli, the ligase consists of 2 fused domains, the catalytic domain LplA and the accessory domain LplB [10,11]. We denote these enzymes Lpl(AB) or in the circularly permutated case [12], Lpl(BA). Ligases with tight substrate specificity have been described that transfer free precursors only to GcvH and, in one case, also to the E2 subunit of 2-oxoglutarate dehydrogenase [1315]. Additional amidotransferases or ligases are then necessary for modification of other lipoyl domains [13,14,16]. Bipartite lipoate:protein ligases forming a functional LplA-LplB heterodimer (denoted LplAB here) have so far been found primarily in archaea [10]. In the absence of free precursors, an octanoyl residue derived from ACP is attached to the target protein by an octanoyltransferase, LipB or LipM [1,1618]. LipB has been found mainly in Proteobacteria, serves as an all-purpose transferase and provides octanoate or lipoate to all known lipoate-requiring pathways except the sHdr-system [2,19]. LipM has been proposed to occur predominantly in Firmicutes and to transfer octanoyl residues exclusively to GcvH. An amidotransferase, LipL, is required for the transfer of octanoyl or lipoyl moieties from GcvH to the E2-subunits of pyruvate and branched-chain α-ketoacid dehydrogenases [13,14,20]. Despite poor sequence conservation, an evolutionary relationship has been detected between lipoate:protein ligases and octanoyltransferases as well as biotin:protein ligases (BirA) [6]. Once the octanoyl residues arrive at their target proteins, they become substrates for lipoate synthase LipA, a member of the radical S-adenosylmethionine (SAM) superfamily [21], which sequentially adds 2 sulfur atoms in a single reaction, first at position C6 and then at C8 [1,22,23].

Fig 1. Processes and components of lipoate assembly pathways.

(a) Main known steps of established lipoate assembly pathways. Enzymes and steps not occurring in E. coli but described for other organisms are printed in gray. LipM and LipL have been demonstrated in Firmicutes, B. subtilis [24], Staphylococcus aureus [25], and Listeria monocytogenes [26], as well as in Tenericutes, Mycoplasma hyopneumoniae [15,16]. (b) Predicted novel lipoate assembly pathway. The pathway is substantiated by experiments reported here as well as by published work on proteins from the 3 model organisms depicted in c [2,9] and by genetic and biochemical work on LipS1 and LipS2 from the archaeon Thermococcus kodakarensis [27]. Lipoate:protein ligases from sulfur oxidizers were originally reported not to contain a carboxy-terminal LplB domain based on superposition of the structure modeled for Thioalkalivibrio sp. K90mix by using the automated mode of SWISS_Model on E. coli Lpl(AB). We challenged this view and indeed, modeling by Alphafold [28] as well as sequence alignments yielded clear proof for the presence of the LplB domain (S2 Fig). (c) Genetic arrangement of 3 novel systems for lipoate assembly in Proteobacteria. Colors correspond to the biochemical roles as depicted in b. For Ts. sibirica locus tags are given according to JGI-IMG. LipT is an FAD-binding NAD(P)H-dependent oxidoreductase possibly delivering electrons for the LipS1/S2-catalyzed sulfur insertion step. The genes lipY and lipX encode a putative fatty acid transporter and a putative glutamine amidotransferase, respectively. ACP, acyl carrier protein; GcvH, glycine cleavage system protein H; LbpA, lipoate-binding protein; LD, lipoyl domains of the 2-oxoacid dehydrogenases.

Initial evidence for an alternative pathway for lipoate assembly in prokaryotes emerged, when we showed that the GcvH-like LbpA proteins involved in sHdr-based sulfur oxidation are not modified by the canonical E. coli and Bacillus subtilis lipoyl attachment machineries [2]. Instead, the bacterial and archaeal shdr-lbpA clusters are accompanied by a set of genes encoding a specific lipoylation pathway (Fig 1B and 1C) that includes lipoate:protein ligases (sLpl(AB)) and 2 proteins of the radical SAM superfamily, originally termed RadSAM1 and RadSAM2. sLpl(AB) lipoate:protein ligases from sulfur oxidizers not only lipoylate LbpA acceptor proteins from the same organism in vitro but also show cross-species functionality among sulfur oxidizers while failing to recognize lipoyl domains/proteins from organisms lacking components of a sHdr-LbpA sulfur-oxidizing system [2].

Recently, proteins from the thermophilic archaeon Thermococcus kodakarensis similar to sLpl(AB) ligase (Tk-Lpl-N and Tk-Lpl-C) and RadSAM1 and RadSAM2 from sulfur oxidizers (now termed LipS1 and LipS2) were shown to exert octanoate/lipoate:protein ligase and LipA-like lipoate synthase functions, respectively, on chemically synthesized peptide substrates in vitro. Genetic analysis provided further evidence that these proteins are involved in archaeal lipoate biosynthesis [16,27,29] (Fig 1B). We have previously shown that the slpl(AB)-encoded protein:lipoate ligase from the gammaproteobacterial sulfur oxidizer Thioalkalivibrio sp. K90mix accepts only free precursors, i.e., octanoate or lipoate, and lacks octanoyltransferase activity [2]. The archaeal ligase is also thought to be restricted to free substrates, mainly because ACPs do not occur in archaea [30].

Here, we provide conclusive experimental evidence for the existence of a novel sLpl(AB)-LipS1/S2-based lipoate assembly pathway not only in archaea but also in bacteria and raise the question of whether it is restricted to thermophilic archaea and sulfur-oxidizing bacteria or is of more general importance. To date, no studies have been published on the origin and evolution of lipoate assembly machineries despite the importance of lipoate for almost all living organisms. This prompted us to carry out an exhaustive large-scale analysis including a large fraction of prokaryotic diversity, and in particular, the archaea for which knowledge about lipoate assembly is scarce. We mapped the novel lipoate synthesis pathway on the tree of life revealing an enormously wide distribution. Finally, our analyses show that the novel lipoate synthesis pathway evolved in the archaeal domain.


Biochemical and genetic proof for an sLpl(AB)-LipS1/S2-based lipoate assembly pathway in bacteria

The initial proposal of a novel route for maturation of lipoate-binding proteins in bacteria relied on the detection of conspicuous lipS1-sLpl(AB)-lipT-lipS2 gene clusters, in vitro assays with sLpl(AB) lipoate:protein ligase from a model sulfur oxidizer and genetic complementation studies in E. coli and B. subtilis [2]. Here, we set out to collect conclusive experimental evidence for the functionality of the pathway in bacteria. A focus was kept on the biochemically characterized sLpl(AB) ligases from the sulfur oxidizers Thiorhodospira sibirica and Thioalkalivibrio sp. K90mix [2].

First, 3 Strep-tagged LbpA lipoate acceptor proteins from these 2 bacteria were recombinantly produced in E. coli, with or without a helper plasmid from which the Thioalkalivibrio assembly genes lipS1-slpl(AB)-lipT-lipS2 were expressed under control of the pACYC184 tet promoter (Fig 2A and 2B). Native gel electrophoresis showed the faster mobility expected for holo-LbpAs only when produced in the presence of the helper plasmid. This is due to the lack of the positive charge when the lipoate-binding lysine is modified by covalent attachment of lipoate or octanoate (Fig 2A). Mass spectrometric analyses confirmed posttranslational modification of all 3 LbpA acceptor proteins by a 157-Da mercaptooctanyol moiety in the presence of the helper plasmid (S3 Fig), fully consistent with in vitro results for the archaeal system where LipS2 first catalyzes sulfur attachment at C8 of an artificial octanoyllysyl peptide substrate and LipS1 then inserts the second sulfur at C6 [29]. Although this last step was not efficiently catalyzed in the heterologous environment, our experiments clearly confirm specific modification of sulfur oxidizer LbpA acceptor proteins by lipoate assembly proteins from a sulfur oxidizer.

Fig 2. Biochemical and genetic evidence for a novel lipoate assembly pathway in bacteria.

(a) LbpAs from Thioalkalivibrio sp. K90mix and Ts. sibirica were produced in E. coli BL21(DE3) ΔiscR, a strain designed for improved synthesis of iron-sulfur proteins [32], either with or without a helper plasmid (pACYC-Tklpm) carrying genes lipS1-slpl(AB)-lipT-lipS2 from Thioalkalivibrio sp. K90mix (shown in b) under control of the constitutive pACYC184 tet promoter. Holo-LbpAs migrate faster in native PAGE due to loss of the positive lysine charge upon modification. In the heterologous host, TsLbpA proteins are—albeit not fully—modified by the assembly proteins stemming from a different species, i.e., Thioalkalivibrio. (c) Thiosulfate (triangles) and sulfite (boxes) concentrations for 4 different H. denitrificans strains during growth on methanol (24.4 mM) as a carbon source in the presence of 2 mM thiosulfate. (d) Growth of H. denitrificans strains. Symbols and lines in c and d correspond to H. denitrificans strains as follows: filled black symbols, solid lines: H. denitrificans ΔtsdA; symbols filled gray, solid lines: H. denitrificans ΔtsdA lbpA2-His; open symbols, dotted lines: H. denitrificans ΔtsdA ΔlbpA2; open symbols, solid lines: H. denitrificans ΔtsdA lbpA2-His Δslpl(AB). For all measurements, standard deviations based on 3 technical replicates are indicated, but too small to be visible for determination of biomass and sulfite. (e) SDS-PAGE of HdLbpA2-His enriched from H. denitrificans ΔtsdA lbpA2-His (left lane, 2.5 μg protein) and ΔtsdA Δslpl(AB) lbpA2-His (right lane, 1.5 μg protein). (f) Native gel mobility shift assay for HdLbpA2-His enriched from H. denitrificans ΔtsdA lbpA2-His (left lane, 2.5 μg protein) and ΔtsdA Δslpl(AB) lbpA2-His (right lane, 1.3 μg protein). HdLbpA2-His proteins were visualized after Western blotting using an Anti-His peroxidase conjugate. The data underlying panels c and d is provided as S1_data.xlsx.

In a second approach, 4 strains of the Alphaproteobacterium Hyphomicrobium denitrificans were studied. The organism is accessible for manipulative genetics, the necessity of its LbpA2 protein for the oxidation of thiosulfate is documented [2,9] and the respective genes are located in immediate vicinity of a lipS1-lipT-lipS2-slpl(AB) cluster (Fig 1C). H. denitrificans ΔtsdA served as the reference strain. It lacks thiosulfate dehydrogenase (TsdA) that catalyzes the formation of the dead-end product tetrathionate. Thus, the strain can oxidize thiosulfate exclusively via the sHdr-LbpA pathway [9,31] that substantially simplifies its elucidation by reverse genetics. When grown in the presence of methanol as a carbon source and thiosulfate as an additional electron source, it excretes toxic sulfite, which causes growth retardation [31]. Functionality of the sHdr-LbpA pathway is thus easily detectable by sulfite formation and diminished growth rate (Fig 2C and 2D). The second strain studied carries a ΔlbpA2 deletion in a ΔtsdA background, is unable to oxidize thiosulfate, and served as control. In the third strain, lbpA2-His, encoding carboxy-terminally His-tagged LbpA2, replaces the original lbpA gene in the chromosome of H. denitrificans ΔtsdA, so that LbpA2 can be purified from this strain. The fourth strain carries an in frame deletion of slpl(AB) in a ΔtsdA lbpA2-His background. Thus, the modification of the lipoate acceptor LbpA2 can be compared in presence or absence of the sLpl(AB) ligase. H. denitrificans ΔtsdA and ΔtsdA lbpA2-His oxidized thiosulfate completely and excreted up to 0.6 mM sulfite (Fig 2C and 2D). This demonstrates that the addition of the carboxy-terminal His-tag does not prevent proper function of the LbpA2 protein. In contrast, in both, the ΔtsdA ΔlbpA2 and the ΔtsdA lbpA2-His Δslpl(AB) strains, thiosulfate degradation was very slow, sulfite formation was not observed (Fig 2C), and the growth rates were higher than those for strains ΔtsdA and ΔtsdA lbpA2-His (Fig 2D). This confirmed the crucial function of LbpA2 in the cytoplasmic sHdr-LbpA sulfur oxidation pathway and more importantly, showed that the absence of sLpl(AB) lipoate:protein ligase had the same effect as the complete absence of the LbpA2 protein, strongly suggesting that sLpl(AB) ligase is essential for the modification and thus the proper function of LbpA2. We found evidence in support of this hypothesis by enriching the LbpA2 acceptor proteins from the H. denitrificans strains ΔtsdA lbpA2-His and ΔtsdA lbpA2-His Δslpl(AB) producing them with a His-tag and comparing their behavior in SDS and native PAGE. While the H. denitrificans strain lacking sLpl(AB) ligase produced only apo-LbpA2, the holo-protein was produced in the strain containing the complete assembly pathway as evident from the native gel mobility shift (Fig 2E and 2F).

Lipoate assembly systems are unevenly distributed in bacteria and archaea

The experiments described above clearly established the functionality of a LipS1/S2-based lipoate assembly pathway in bacteria. We now asked how lipoate:protein ligases and octanoyltransferases (LipB, LipM) and the different lipoate synthases (LipA, LipS1/S2) are distributed among the prokaryotes and analyzed the genomes available in the Genome Taxonomy Database (GTDB, release R207). In GTDB, all genomes are sorted according to validly published taxonomies. In order to accurately identify and discriminate the enzymes involved in lipoate assembly pathways, a task severely hampered by the fact that central components are part of very large multi-protein families, we used HMS-S-S, a tool that specifically finds sulfur metabolism-related proteins [33] and extended it with publicly available HMMs for canonical lipoate synthesis enzymes as well as for well-studied lipoate-binding proteins, such as GcvH (S1 Table).

In both prokaryotic domains, scavenging of free lipoate or octanoate by lipoate:protein ligase is widespread (Fig 3). Lpl/LipA combinations were found in only 14% and 8% of the bacterial and archaeal genomes, respectively, while Lpl/LipS1/S2 were found in 0.9% and 17%, respectively, leaving the majority of prokaryotes unable to use free octanoate for lipoate assembly (Fig 3A and 3B). In bacteria, de novo lipoate synthesis by LipB/LipA or LipM/LipA is more common than lipoate scavenging (Fig 3). The co-occurrence of both octanoyltransferases, LipB and LipM, is very rare and there is usually no strict correlation between phylum and enzyme type. Contrary to previous assumptions [19], LipB is found in the Firmicutes and LipM is not restricted to a specific phylum. The Proteobacteria, together with a few other phyla, are the exception rather than the rule, containing only LipB and never LipM (Fig 3C). As mentioned in the Introduction, B. subtilis octanoyltransferase LipM is specific for GcvH and an amidotransferase (LipL) is required for modification of other lipoyl-binding domains/proteins. Surprisingly, not all bacteria, including Firmicutes, which carry out LipM/LipA-based lipoate biosynthesis, also possess LipL, although lipoate-binding domains/proteins are encoded in these genomes (S2 and S3 Tables). This suggests a much broader substrate specificity of the octanoyltransferase LipM than previously described or the existence of (an) unknown amidotransferase(s) that functionally replace(s) the amidotransferase LipL. Among the archaea, octanoyltransferase/lipoate synthase combinations LipB/LipA or LipM/LipA were also present (Fig 3), suggesting de novo lipoate synthesis from ACP-bound octanoate. This was unexpected, because ACP is generally absent in archaea [3436] and cannot serve as direct octanoate donor. Coenzymes A or M are possible alternatives [3537].

Fig 3. Taxonomic distribution of the lipoate synthesis systems, lipoate scavenging, and lipoate requiring proteins.

Venn diagrams show the abundance and overlap of lipoate:protein ligases (Lpl), octanoyl transferase (LipB, LipM), and lipoate synthases (LipA, LipS1/S2) in the bacteria (a) and the archaea (b). Panels c and d visualize the taxonomic distribution of these enzymes, the sulfur-oxidizing sHdr system (S) and lipoate-binding domains (LD). For each bacterial (c) and archaeal phylum (d), the percentage of genomes possessing these proteins is indicated by dots of different sizes and colors. Note that the proportion was normalized to the size of the phylum and does not show absolute counts or overall phylum size. The data underlying parts a and b are provided in S2 and S3 Tables, respectively. S4 Table supplies the data underlying parts c and d.

In both prokaryotic domains, the LipS1/S2-type of lipoyl synthase is less common than LipA lipoate synthase but remarkably widespread (Fig 3). Co-occurrence of lipoate synthase LipA is a common feature for lipS1/S2-containing bacteria but rare in LipS1/S2-encoding archaea. Our analyses do not confirm the proposal that LipS1/S2 are specific to thermophilic archaea [27]. The Asgardarchaeota, in which LipS1/S2 is more abundant than LipA, are a counter-example, as all sequences, including that of the only cultured representative Candidatus Prometheoarchaeum syntrophicum, are from temperate to cold environments [38,39]. Halobacteriota and Altiarchaeota genomes with lipS1/S2 also stem from temperate/cold habitat samples [40,41]. In addition, LipS1/S2 is present in many mesophilic bacteria, i.e., H. denitrificans, Thioalkalivibrio sp., or Sporomusa sp.

LipS1/S2-type lipoate synthases and their cooperation partners

Regarding the novel lipoate synthesis pathway, our analyses confirm that lipS1/S2 genes are often associated with genes for lipoate:protein ligases (Figs 1C and 3) [2], usually with a Lpl(AB) domain structure (S5 Table). A lipS2-lplA-lplB-lipS1 arrangement seems typical for Archaea but is also found with some rearrangement in the bacterial phyla Chloroflexota, Aerophobota, and Synergistota. Occasionally, direct linkage of genes for canonical lipoate synthesis with lipS1/S2 is observed, e.g., in several Sporomusa species (phylum Firmicutes). In some LipS1/S2-containing bacteria (e.g., members of the Schekmanbacteria, Synergistota, and Thermotogota) and archaea (members of the Asgardarchaeota, Thermoproteota, and Themoplasmatota), a gene for lipoate:protein ligase is absent (Fig 3 and S2 Table). Instead, lipS1/S2 co-occur with a gene for octanoyltransferase LipM. In the bacterial cases, this implies that LipS1/S2 insert sulfur into target proteins that have been octanoylated by a transferase reaction. For the archaea, as discussed above, the question of the substrate for the LipM homologs is unresolved.

The lipT encoded FAD-containing oxidoreductase is a likely candidate to provide electrons, probably derived from NAD(P)H, for the reductive sulfur insertion catalyzed by LipS1/S2 (Fig 1). Indeed, lipT occurs almost exclusively in bacteria containing lipS1/S2 (91% of the cases), often in a lipS1-lpl(AB)-lipS2-lipT arrangement (Fig 1C and S5 Table). The picture is different for archaea, where only 22.3% of the lipT-containing genomes also contain lipS1 and lipS2 (S4 Table). Approximately 53% and 17% of the LipS1/S2-containing bacteria and archaea, respectively, also encode LipT (S2, S3 and S4 Tables).

The lipS1/S2 genes were first detected in bacterial and archaeal sulfur oxidizers that use the sHdr pathway for sulfur oxidation [2]. LbpA proteins are essential components of this pathway [2] and are encoded in shdr-containing genomes with very few exceptions, probably due to incompleteness of the respective assemblies (Fig 3 and S2 and S3 Tables), raising the question of whether the assembly of LbpA proteins is strictly dependent on LipS1/S2. While the majority of sHdr-containing prokaryotes are indeed equipped with LipS1/S2 (74.3% and 88.6% for bacteria and archaea, respectively; Fig 1C and S2, S3 and S4 Tables), the reverse is not true, i.e., LipS1/S2 are not restricted to sulfur oxidizers (42% and 8% of LipS1/S2-containing bacteria and archaea, respectively, have sHdr; S2, S3 and S4 Tables).

Evolution of lipoate:protein ligases and octanoyltransferases

All lipoate:protein ligases and octanyoltransferases belong to the cofactor transferase family PF03099. Calculating rooted phylogenetic trees for these proteins was expected to bring new insights into the origin and evolution of these enzymes. If they initially originated in archaea, the tree should be a priori rooted in the archaeal domain and similarly for bacteria. The structurally related biotin ligase BirA, which is also a member of the cofactor transferase family, was chosen as a suitable outgroup to root the tree. The tree for the complete lipoate:protein ligase/octanoyltransferase dataset contains 3 clearly delineated clades with high bootstrap support (Fig 4). The first clade contains the bacterial and archaeal LipB octanoyltransferases and resides between the BirA root and all other analyzed proteins. The second clade harbors a group of lipoate:protein ligases derived exclusively from bacteria with LipA but usually without LipS1/S2 lipoate synthase. Broad substrate range E. coli Lpl(AB) as well as the narrow substrate range ligases LplJ from B. subtilis, and Mhp-LplJ and Mhp-Lpl from M. hyopneumoniae reside in this clade.

Fig 4. Rooted phylogenetic tree for the complete lipoate:protein ligase/octanoyltransferase dataset.

The tree was rooted with the structurally related biotin ligase BirA as an outgroup. Red or blue dots placed on each leaf identify the source organisms as archaea or bacteria, respectively. The ligase/transferase type is color-coded in the next circle. In the outermost rings, the presence of other lipoate synthesis enzymes occurring in the same genome is labeled. The data underlying this figure is provided in Supplementary S3 Data.

For the proteins of clade 3, an archaeal origin is inferred, since bipartite LplABs ligases from the archaeal phyla Thermoproteota and Thermoplasmatota including the characterized Thermoplasma acidophilum enzyme [10] are the deepest branching sequences. Three well-supported subgroups (bootstrap ≥92) branch off from these, each again with deep-branching archaeal proteins. In the first subgroup, bipartite LplABs from Nano- and Haloarchaeota form the deepest branches, which are immediately adjacent to LplABs from Burkholderiales (class Gammaproteobacteria according to GTDB). The remaining sequences in the subgroup are circularly permuted Lpl(BA) proteins, nearly exclusively stemming from Actinomycetota and including the characterized S. coelicolor Lpl(BA) [12]. This topology suggests an evolutionary history with lateral transfer of LplAB from Archaea to Gammaproteobacteria followed or accompanied by rearrangement of the gene order and final fusion of the genes upon transfer to the phylum Actinomycetota, where the gene was then vertically transmitted. The second subgroup consists of archaeal and bacterial bipartite LplAB ligases and LipM-type octanoyltransferases and provides insights into the origin of LipM: (1) Archaeal LplAB ligases, mostly encoded near lipS1/lipS2 and including the characterized Thermococcus kodakarensis protein [30], are the most deeply branching sequences and appear to be the ancestors of a large number of bacterial LipMs, which thus arose from a single interdomain horizontal gene transfer event. A scenario is supported in which archaeal LplAB lost its accessory peptide LplB, developed into LipM and simultaneously or later moved into a bacterial host. (2) In the remaining part of the subgroup, many archaeal and some bacterial LplABs ligases are mixed with many archaeal and some bacterial LipM octanoyltransferases, indicating that the described horizontal gene transfer, loss of LplB and transformation of the remaining catalytic domain LplA into an octanyoltransferase was not a singular event but happened multiple times. The third subgroup of clade 3 is made up of even further archaeal and bacterial LplABs and Lpl(AB) lipoate protein ligases, the vast majority of which originate from organisms containing LipS1/S2. The genetically and biochemically characterized sLpl(AB) ligases from proteobacterial sulfur oxidizers fall into this group.

We challenged the idea that clade 3 has an origin within the archaea by calculating 3 separate trees for this group. The trees were rooted by the most closely related, similarly sized and biochemically characterized bacterial lipoate:protein ligases from clade 2 and confidence levels were increased by not including LipM octanoyltransferases and/or circular permuted cpLplBA lipoate:protein ligases (Figs 5, S4 and S5). The trees have a high to very high bootstrap support especially for the higher order splits and all 3 indeed show a root in the archaeal domain with Thermoproteota and Thermoplasmatota proteins as the deepest branching sequences. Moreover, it is clear that bacterial Lpl(AB)s, which co-occur in the same organism with LipS1/S2, originate from an archaeal ancestor. Several horizontal gene transfer events are also evident. The 2 earliest were transfer of LplAB ligases from Hadarchaeota to Synergistota and from Altiarchaeota to Chloroflexota. On the other hand, several transfers from bacteria back to archaea can be delineated, e.g., into members of the Thorarchaeota, Baldrarchaeota, Jordarchaeota, Sifarchaeota, Thermoplasmatota, and Thermoproteota.

Fig 5. Phylogeny for clade 3 lipoate:protein ligases without LipM and cpLpl(BA).

LipMs do not have LplB domains and their sequences are consequently shorter. If a sequence is incomplete, parts of the information used to calculate the phylogenetic tree are missing. This can lead to erroneous estimates of the relationships between sequences and can bias the result and weakens statistical significance of the calculation. In addition, Lpl(BA) clearly shows an individual evolution and may also cause weakening of statistical support. The data underlying this figure is provided in S3 Data.

Phylogenetic analysis of the radical SAM proteins LipS1 and LipS2

The tree obtained for the concatenated LipS1/S2 proteins neither shows a 2 domain split nor a long branching separation, as would be expected for an ancient protein that already existed in a universal common ancestor [42]. This indicates an origin of LipS1/S2 either in the bacterial or in the archaeal domain, followed by horizontal gene transfer into the other domain [43,44]. To obtain further insights, separate trees were constructed for the bacterial and the archaeal proteins (Fig 6B and 6C). If LipS1/S2 lipoate synthase originated in either the archaea or the bacteria and was inherited vertically, the domain-specific tree should essentially follow the taxonomy of that domain. However, neither the tree for archaeal LipS1/S2 nor that for the bacterial equivalents fit this concept. With the exception of Proteobacteria and Aquificota, monophyletic clusters were not recovered within the bacteria (Fig 6B) and among the archaea only the LipS1/S2 lipoyl synthases from the Hadarchaeota, Methanobacteriota, and Halobacteriota appear monophyletic (Fig 6C). The trees indicate that LipS1/S2 has undergone multiple gene transfers between and within the 2 prokaryotic domains.

Fig 6. Phylogenetic trees for LipS1/S2.

(a) To investigate the evolution of LipS1 and LipS2, their sequences were concatenated, as both units are usually found in synteny, are catalytically active together and should therefore be under the same evolutionary pressure [29]. Incomplete sequences and concatenated sequences from genomes lacking either LipS1 or LipS2 were removed from the analysis. The lower panels show schematic representation of phylogenetic trees generated using only archaeal sequences (red, b) or bacterial sequences (blue, c). Bacterial clades represented by single sequences were left out to increase readability. The data underlying this figure is provided in S3 Data.

Finally, some insight into the origin of LipS1/S2 lipoate synthase was gained by rooting separate LipS1 and LipS2 trees by the radical SAM enzyme biotin synthase BioB, which belongs to the same megacluster of similarity and, just like each of the LipS proteins, catalyzes the insertion of a single sulfur atom [29,45]. For LipS2, the resulting tree topology was rather well supported and placement of BioB in the split between 2 archaeal clades was stable (S6 Fig). This points at an origin of LipS2 in the archaea. The final topology of the rooted LipS1 tree is consistent with an archaeal origin but bootstrap support is not high enough to draw firm conclusions (S6 Fig).


Our work proves the existence of a LipS1/S2-based lipoylation pathway not only in archaea but also in bacteria. In both prokaryotic domains, the pathway starts with the incorporation of a free octanoic acid and then LipS1/S2 insert 2 sulfur atoms. The lipoate:protein ligase that catalyzes the first step originates from archaea and thus has a completely different evolutionary history than the previously known Lpl(AB) enzymes in clade 2, which are exclusively found in bacteria. For LipS1/S2, which catalyze the second step, an archaeal origin is also likely. The majority of bacteria that use reduced sulfur compounds as electron donors by the sulfur-oxidizing heterodisulfide reductase-like (sHdr) complex in conjunction with LbpA proteins, are equipped with the originally archaeal lipoate synthesis machinery. Since heterodisulfide reductases are enzymes typical of methanogenic archaea, co-transfer of the genes seems a likely possibility.

Our large-scale integrated phylogenomic analyses open new perspectives on the general evolution and the complexity of lipoate synthesis. Not only do they reveal a much wider distribution of lipoate assembly systems than expected, in particular, the novel sLpl(AB)–LipS1/S2 pathway, and indicate a highly modular nature of the enzymes involved, with unforeseen combinations, but they also provide a new framework for the origin and subsequent evolution of lipoate assembly systems in the 2 prokaryotic domains. Our results are relevant to the entire diversity of life, since the machineries for lipoate synthesis in the mitochondria and plastids of eukaryotes are all of prokaryotic origin and may bear more complexity than currently assumed [1,46,47]. Contrary to the intuitive assumption, mitochondrial LipAs are more closely related to the LipA lipoate synthases of archaea than to those of α-proteobacteria [47]. In addition, plastidial LipAs are not most closely related to those of cyanobacteria, as would be expected if they were derived from the primary endosymbiont that led to the first plastid. Instead, they form a sister group to the mitochondrial LipAs and probably arose by a gene duplication [47]. Furthermore, protozoa with mitochondria and plastids require 2 different ligases for lipoate metabolism in the respective organelles [1].

The tripartite topology of the phylogenetic tree for the octanyltransferases and lipoate:protein ligases has provided particularly important insights. LipB octanoyltransferases (clade 1) are close to the root of the tree and thus represent the ancestors of all the other enzymes. Clade 1 gave rise to 2 other groups, one of which evolved in bacteria (clade 2) and the other is probably of archaeal origin (clade 3). In clade 2, the acquisition of an additional gene encoding LplB, and in many cases its fusion with the original LipB octanoyltransferase unit, allowed the conversion of an octanoyltransferase into an ATP-dependent lipoate:protein ligase. The most deeply branching proteins in clade 3 also have additional LplB, suggesting that the acquisition of the corresponding gene and functional transformation into a ligase was again the starting point for further evolution. Our conclusions are consistent with previous suggestions, based on structural data, that the lipoate:protein ligase LplA initially lacked an LplB unit, which was first acquired as a separate unit before fusion occurred [48].

We infer that bipartite LplAB enzymes represent the most original lipoate:protein ligase in archaea. Free lipoate may therefore have been the primary source of protein lipoylation in ancient archaea. There was no lipoyl synthase in these organisms, as we still find today, for example, in Thermoplasma acidophilum. Over time, lipoyl synthases appeared in archaea, either LipA or LipS1/S2 and very rarely both, enabling the use of octanoate. Gene clusters then formed, as we find in almost all recent archaea. In bacteria, clade 2 Lpl(AB) ligases appear to co-operate with LipA and in those bacteria that encode sLpl(AB) ligase, the genes for LipS1/S2 are almost always also present. Genetic linkage of the genes in operons is the usual case. These observations, together with the highly probable archaeal origin for sLpl(AB) ligase and LipS1/S2, suggest that the genes were transferred together from archaea to bacteria, possibly even together with genes for their specific lipoylation substrates. Such a scenario also fully explains why the canonical Lipl(AB)/LipA combination and the sLpl(AB)/LipS1/S2 system can coexist in the same organism and specifically lipoylate only their cognate substrates. Further experimental work on bacterial and also archaeal lipoate:protein ligases and their potential substrates will allow testing these functional hypotheses.

Lipoate-requiring reductive glycine synthesis was placed at the basis of the tree of life, as part of the phenotype of the last common ancestor [19]. Accordingly, the last common ancestor has been proposed to contain radical SAM-dependent enzymes and LipA lipoyl synthase is considered among the most ancient enzymes in this class [21]. In contrast, genes encoding LipS1/S2 are absent in most phyla close to the root of the archaeal domain [49] (Iainarchaeota, Nanohaloarchaeota, Nanoarchaeota, Aenigmatarchaeota, and Micrarchaeota) and likely emerged later than LipA.

In conclusion, our results pave the way for further work on lipoate biochemistry as a key process for ancient and modern life.


Bacterial strains, plasmids, primers, and growth conditions

S6 Table lists the bacterial strains and plasmids that were used for this study. E. coli strains were grown on complex lysogeny broth (LB) medium [50] or on glycerol-containing M63 minimal medium [51] under aerobic conditions at 37°C unless otherwise indicated. E. coli 10β was used for molecular cloning. E. coli BL21 (DE3) was used for recombinant protein production. H. denitrificans strains were cultivated in minimal media kept at pH 7.2 with 100 mM 3-(N-Morpholino)propanesulfonic acid (MOPS) buffer as described before [9,31]. Media contained 24.4 mM methanol. Thiosulfate was added as needed. Antibiotics for E. coli and H. denitrificans were used at the following concentrations (in μg ml−1): ampicillin, 100; kanamycin, 50; streptomycin, 200; chloramphenicol, 25.

Recombinant DNA techniques

Standard techniques for DNA manipulation and cloning were used unless otherwise indicated [52]. Restriction enzymes, T4 ligase and Q5 polymerase were obtained from New England Biolabs (Ipswich, United Kingdom) and used according to the manufacturer’s instructions. Oligonucleotides for cloning were obtained from Eurofins MWG (Ebersberg, Germany). Plasmid DNA from E. coli was purified using the GenJet Plasmid Miniprep kit (Thermo Scientific, Waltham, United States of America). Chromosomal DNA from H. denitrificans strains was prepared using the Simplex easy DNA kit (GEN-IAL GmbH, Troisdorf, Germany).

Construction of helper plasmids with lipoate assembly genes from Thioalkalivibrio sp. K90mix

For the construction of pACYC-Tklpm (lipoate protein maturation), the potential lipoate assembly genes lipS1-slpl(AB)-lipT-S2-Y (TK90_0641–0644, Fig 1C) were amplified with primers TK90_0641–0644 fw and TK90_0641–0644 rev using genomic Thioalkalivibrio sp. K90mix DNA as the template. The amplicon was digested with DraI and BspHI and cloned into the EcoRV/BspHI sites of pACYC184. This plasmid contained the lipS1-slpl(AB)-lipT-lipS2 sequence in the same orientation as the tet gene of the vector, thus allowing outreading transcription from the tet promoter.

Production of LbpA proteins in E. coli BL21 (DE3) ΔiscR with and without helper plasmids

Recombinant LbpA proteins were produced in the E. coli BL21 (DE3) ΔiscR, and 500-ml batches of LB medium containing 100 mM MOPS buffer, pH 7.4, 25 mM glucose, and 2 mM iron ammonium citrate as well as ampicillin and kanamycin (and chloramphenicol for strains containing plasmid pACYC-Tklpm) were inoculated with 1% (v/v) E. coli precultures and cultivated in 1-L flasks at 37°C and 200 rpm until an OD600 of 0.8 was reached. At this point, cysteine (0.5 mM), sodium fumarate (25 mM), and IPTG (0.1 mM) were added. Incubation continued for 14 to 16 h at 30°C and 180 rpm. Cells were harvested by centrifugation (11,000 × g, 12 min, 4°C) and lysed by sonication. After removal of insoluble cell material by centrifugation (16,100 × g, 30 min, 4°C), the LbpA proteins were purified by Strep-Tactin affinity according to the manufacturer’s instructions (IBA Lifesciences, Göttingen, Germany). The proteins were then transferred to salt-free 100 mM Tris-HCl (pH 8.0) buffer and underwent further analysis.

Construction of H. denitrificans mutant strains

The tsdA gene was deleted from H. denitrificans ΔlbpA [2] by using plasmid pK18mobsacB ΔtsdATc, transferring it by electroporation, and selecting double crossover recombinants as described previously [9]. For chromosomal integration of the gene encoding HdLbpA2 with a carboxy-terminal His-tag, the modified gene and upstream as well as downstream sequences were amplified by SOE PCR using primers Fwd5′_ΔlbpA, KI_HdLbpA2-His-Up-rev, KI_HdLbpA2-His-Down-fw, and Fwd3′_ΔlbpA (S6 Table). The final plasmid pk18mobsacB-lbpA2-his-Tc was transferred into H. denitrificans ΔtsdA ΔlbpA. For markerless in frame deletion of the H. denitrificans slpl(AB) gene by splicing overlap extension [53], PCR fragments were constructed using the primers listed in S6 Table. It should be noted that the GTG start codon of the slpl(AB) gene overlaps the stop codon of the preceding gene lipS2 in the sequence GTGA. Similarly, the TGA stop codon of slpl(AB) overlaps the start of gene lipX in the sequence ATGA. The ribosomal binding site for lipX translation must thus be embedded in the slpl(AB) gene. To avoid affecting signals for lipX translation, the in frame deletion of slpl(AB) was designed to leave the last 35 bp of the gene untouched. The 2,070 bp fragment, which implements deletion of a 1,029 bp fragment encoding amino acids 8 to 349 of sLpl(AB), was digested with XbaI and cloned into the XbaI site of pk18mobsacB-Tc [31]. The final plasmid pk18mobsacBΔslpl(AB)-Tc was transferred into H. denitrificans ΔtsdA lbpA2-His. The genotypes of the H. denitrificans mutant strains generated in this study were confirmed by PCR.

Purification of His-tagged HdLbpA2

A total of 50 ml H. denitrificans precultures were grown in 100-ml Erlenmeyer flasks in methanol-containing medium with 100 mM MOPS (pH 7.4), chloramphenicol and streptomycin up to an OD600 of 0.7. The main culture had a volume of 1 l in a 2-L Erlenmeyer flask and was inoculated to an OD600 of 0.006 with the preculture. The main culture medium contained 2 mM thiosulfate. Cultures were incubated at 30°C and 150 rpm. Cells were harvested by centrifugation at 10,100 × g for 20 min at 4°C, when the cultures were actively oxidizing thiosulfate. Cells were stored at −20°C. Cells were resuspended in 10 ml phosphate buffer (50 mM NaH2PO4, 300 mM NaCl (pH 7.4)) per g cell material. A spatula tip of DNAse RNAse, lysozyme, and protease inhibitor were added. Cells were broken by ultrasonication at 4°C for 10 min per g cells (Branson sonifier, 55% power) followed by centrifugation (16,100 × g, 4°C, 30 min) and ultracentrifugation (1 h, 4°C, 145,000 × g). Affinity chromatography on 1 ml HisTrap TALON crude columns (Cytiva, Marlborough, Massachusetts, USA) was performed with an Äkta-FPLC system according to the manufacturer’s instructions. Enriched fractions were concentrated with Amicon Ultra-0.5, Ultracel 3k membrane, 3 kDa centrifugal filter units and subjected to electrophoreses on 12.5% Tricine-SDS or on 15% native gels stained with RotiBlue Quick (Carl Roth GmbH, Karlsruhe, Germany) for 1 h. Western blot analysis was performed using nitrocellulose membranes (Amersham Protran 0.2 μm NC, GE Healthcare, Solingen, Germany) and a Trans-Blot Turbo Transfer system (Bio-Rad Laboratories, Munich, Germany). Before western blotting, native gels were incubated for 20 min in Bjerrum buffer (48 mM Tris, 39 mM glycine, 0.04% SDS (pH 9.2)). Whatman filter paper needed during the blotting process was also soaked in Bjerrum buffer. For western blotting of SDS gels, gels and membrane were incubated for 20 min in Towbin buffer (25 mM Tris, 192 mM glycine, 20% (v/v) methanol (pH 8.3)) before loading the blotting chamber. After overnight blocking at 4°C with 4 mM KH2PO4, 16 mM Na2HPO4, 115 mM NaCl, pH 7.4 + 0.05% (v/v) Tween 20, 4% milk powder, the membrane was washed 3 times for 5 min in PBS-Tween-buffer (4 mM KH2O4, 16 mM Na2HPO4, 115 mM NaCl + 0.1% Tween 20). Proteins were detected with Anti-His-HRP-conjugate (1:5,000) using the SignalFireTM ECL reagent system (Cell Signaling Technology, Cambridge, UK) and visualized with a ChemiDoc Imaging System (BioRad Laboratories, Munich, Germany).

Characterization of phenotypes, quantification of sulfur compounds, and protein content

Growth experiments with H. denitrificans were run in in Erlenmeyer flasks with media containing 24.4 mM methanol and varying concentrations of thiosulfate as necessary [31]. Thiosulfate and sulfite concentrations and biomass content were determined by previously described methods [31,54]. All growth experiments were repeated 3 to 5 times. Representative experiments with 2 biological replicates for each strain are shown. All quantifications are based on at least 3 technical replicates.

Mass spectrometry and Edman degradation

MALDI-TOF measurements and Edman degradation were performed at the Core Facility Protein Synthesis & BioAnalytics, Pharmaceutical Institute, University of Bonn.

Dataset generation

Archaeal and bacterial genomes were downloaded from GTDB (release R207). In GTDB, all genomes are sorted according to validly published taxonomies, they are pre-validated and have high quality (completeness minus 5*contamination must be higher than 50%). One representative of each of the current 65,703 species clusters was analyzed. It should be noted that GTDB is built on recently standardized bacterial and archaeal taxonomies derived by normalization of the evolutionary distance between taxonomic levels [49,55]. Among the bacteria, 148 phyla are currently distinguished. For the archaea, GTDB lists 16 phyla. Recent reclassification of the archaea was accompanied by merging parts of the Euryarchaeota with the TACK superphylum into a single phylum [49]. Due to these massive but necessary taxonomic rearrangements, conclusions drawn on the distribution and occurrence of genes or groups of genes in higher taxonomic ranks need special care when comparing with previous work. Open reading frames were determined using Prodigal [56] and subsequently annotated for sulfur-related proteins via HMS-S-S [33]. Lipoic acid synthesis proteins and known lipoic acid-dependent enzymes were searched and annotated with HMMs from the TIGRFAMs and Pfams databases with trusted cutoffs. The HMM with the highest above-threshold bitscore was selected for each protein. Lipoate:protein ligases were checked for accessory domain LplB using the pfam PF10437 and annotated accordingly as LplAB. The publicly available HMMs for lipoate synthases, octanoyltransferases, and lipoate:protein ligases are generally sufficiently sensitive to detect all relevant sequences and also precise enough to make a good distinction between related proteins with different functions. However, there are limitations to these HMMs, as they are based on the state of knowledge at the time they were generated and are not updated on a regular basis. Within the lipoate:protein ligases, it is for example not possible to distinguish between LplAB, Lpl(AB), sLpl(AB) from sulfur oxidizers, Lpl(BA), LplJ or Mhp-LplJ. All existing HMMs have been trained by Lpl(AB) and detect catalytic LplA domains of any type even in the absence of the accessory domain LplB. LplJ and LplA sequences are too similar to be distinguished. Thus, we annotated all detected lipoate:protein ligases as Lpl. The main distinguishing feature between the previously known ligases and sLpl(AB) and sLplAB enzymes is the genomic context and/or the type of lipoate synthase present in the organism. LipA and LipS1/S2 can be distinguished very reliably by HMMs.

Phylogenetic inference

In order to investigate the evolutionary history of the lipoate:protein ligases and octanyoltransferases of the cofactor transferase family PF03099, sequences were separately retrieved from archaeal and bacterial assemblies. A dataset representing classical Lpl(AB) sequences was assembled from LipA-containing genomes not containing lipS1 or lipS2. A second dataset for lipoate:protein ligases originated from gene clusters containing lipS1/lipS2 and a single copy of the ligase gene. Thus, we ensured the highest possible probability that this dataset specifically included LipS1/S2-coupled ligases. LplA and LplB encoded in syntenic gene clusters were concatenated to match the Lpl(AB) domain structure. Tc. kodakarensis LplAB presented the only exception to this rule and was included in the analysis because the protein is biochemically characterized although the relevant genes do no form a gene cluster. For the circularly permuted LplBA ligases, the sequences of the 2 domains were separated and rejoined in the LplAB order and noted as cpLpl(AB). Octanyoltransferases LipB and LipM from bacterial genomes encoding LipA were clustered by similarity to limit the analysis to a reasonable number of sequences but at the same time maintain the diversity of the underlying dataset, while all corresponding archaeal sequences could be taken into account. A further dataset for LipM octanoyltransferase sequences originated from gene clusters containing lipS1/lipS2 and a single copy of the octanyoltransferase gene, again in order to guarantee the highest possible probability that this dataset specifically includes LipS1/S2-coupled octanoyltransferases.

In order to exclude paralogous sequences, only proteins from assemblies encoding for 1 copy of each, LipS1 and LipS2, were considered and concatenated.

Sequences were aligned using MAFFT [57]. Ambiguously aligned regions and other sources of bias, such as highly variable characters, were removed using BMGE [58] (entropy threshold = 0.95, minimum length = 1, matrix = BLOSUM30), thereby trimming the alignments to regions suitable for phylogenetic inference. Phylogenetic inference by maximum likelihood is widely used in molecular systematics and involves substitution model parameters, branch lengths, and tree topology. In this work, maximum likelihood phylogenies were inferred using IQ-TREE v1.6.12 [59] implemented on the “bonna” high performance clusters of the University of Bonn. As a first step, the best-fitting model of sequence evolution that led to the available data was selected using ModelFinder, a method that combines amino acid substitution model used in other popular model-selection methods [60]. Branch support was then calculated by 3 different tree topology tests, SH-aLRT (2,000 replicates) [61], aBayes (2,000 replicates) [62], and ultrafast bootstrap (2,000 replicates) [63]. Finally, trees were displayed using iTol [64].

Where phylogeny of concatenated sequences was inferred, alignments were made individually and concatenated before trimming by BMGE. As an exception, LplA and LplB encoded in the same gene cluster were concatenated before the alignments as these domains are commonly found to be fused. Octanoyltransferases LipB and LipM lacking a C-terminal accessory domain were added to the LplAB alignments using MAFFT [57]—add-fragment function with–keeplength.

All trees are available in Newick format with associated iTol records in a compressed/zip file archive as Supporting information (S3 Data).

Supporting information

S1 Fig. Structural superposition of E. coli GcvH with H. denitrificans LbpA2.

Structural data for the E. coli protein (colored beige) are available in the RCSB Protein Data Bank (accession 3AB9). The structure of the H. denitrificans LbpA2 (colored light blue) was generated with Alphafold2.


S2 Fig. Comparison of bacterial and archaeal lipoate:protein ligases.

(a) Sequence alignment of Lpl(AB) from E. coli (Ecoli), sLpl(AB) from H. denitrificans (Hden), sLpl(AB) from Thioalkalivibrio sp. K90mix (TK90), concatenated LplBA from Thermoplasma acidophilum acidophilum (Thaci), and concatenated LplBA from Tc. kodakarensis (Tkoda). The LplB domain/polypeptide is highlighted by a box. The alignment was produced with T-Coffee. Alignment scores are graded from green (bad) to red (good). The right panel shows structural superposition of the E. coli Lpl(AB) with H. denitrificans sLplAB (b), Thioalkalivibrio sp. K90mix sLpl(AB) (c), Thermoplasma acidophilum LplAB (d), and Thermococcus kodakarensis LplAB (e). All structures were generated with Alphafold2. The E. coli structure is shown in green in each case. The LplB domain is boxed in all panels.


S3 Fig. Mass spectrometric analyses of 3 different recombinant LbpA proteins.

Tk90LbpA2 from Thioalkalivibrio sp. K90mix (a), TsLbpA2 (b), and TsLbpA1 (c) from Thiorhodospira sibirica produced in E. coli BL21 (DE3) ΔiscR in the absence (green spectra) or presence (blue spectra) of helper plasmid pACYC-Tklpm carrying genes lipS1-slpl(AB)-lipT-lipS2 from Thioalkalivibrio sp. K90mix. The mass spectrometric data is provided as “S2 Data.”


S4 Fig. Phylogeny for clade 3 lipoate:protein ligases including LipM, but excluding cpLpl(BA).

Introducing LipM did not change the support of an archaeal origin of clade 3. The data underlying this figure can be found in S3 Data.


S5 Fig. Phylogenetic tree of for clade 3 lipoate:protein ligases including cpLpl(BA) but excluding LipM.

Introducing cpLpl(BA) did not change the support of an archaeal origin of the clade 3. The data underlying this figure can be found in S3 Data.


S6 Fig. Rooted individual phylogenetic trees for LipS2 (a) and LipS1 (b).

Sequences were derived from bacterial and archaeal genomes encoding exactly for 1 LipS1 and 1 LipS2. The trees were rooted by biotin synthase BioB as an outgroup. Archaeal sequences are marked in red, bacterial ones in blue. The data underlying this figure can be found in S3 Data.


S1 Table. Hidden Markov models for extension of HMS-S-S.


S2 Table. Presence/absence in bacteria of lipoate synthesis systems, lipoate scavenging, and lipoate requiring proteins.


S3 Table. Presence/absence in archaea of lipoate synthesis systems, lipoate scavenging, and lipoate requiring proteins.


S3 Data. Newick treefiles and iTol datasets.



We gratefully acknowledge the support of the Core Facility “Protein Synthesis and Bioanalytics” of the University of Bonn for performing mass spectrometry and Edman degradation. We gratefully acknowledge the access to the Bonna HPC cluster hosted by the University of Bonn along with the support provided by its High Performance Computing & Analytics Lab.


  1. 1. Cronan JE. Assembly of Lipoic Acid on Its Cognate Enzymes: An Extraordinary and Essential Biosynthetic Pathway. Microbiol Mol Biol Rev. 2016;80(2):429–450. pmid:27074917
  2. 2. Cao X, Koch T, Steffens L, Finkensieper J, Zigann R, Cronan JE, et al. Lipoate-Binding Proteins and Specific Lipoate-Protein Ligases in Microbial Sulfur Oxidation Reveal an Atypical Role for an Old Cofactor. elife. 2018;7:e37439. pmid:30004385
  3. 3. Spalding MD, Prigge ST. Lipoic Acid Metabolism in Microbial Pathogens. Microbiol Mol Biol Rev. 2010;74(2):200–228. pmid:20508247
  4. 4. Dunn MF. Vitamin Formation from Fatty Acid Precursors. In: Geiger O, editor. Biogenesis of Fatty Acids, Lipids and Membranes. Handbook of Hydrocarbon and Lipid Microbiology. Cham: Springer Nature Switzerland AG; 2019. p. 259–271.
  5. 5. Reed LJ, DeBusk BG, Gunsalus IC, Hornberger CS Jr. Crystalline Α-Lipoic Acid: A Catalytic Agent Associated with Pyruvate Dehydrogenase. Science. 1951;114(2952):93–94.
  6. 6. Perham RN. Swinging Arms and Swinging Domains in Multifunctional Enzymes: Catalytic Machines for Multistep Reactions. Annu Rev Biochem. 2000;69:961–1004. pmid:10966480
  7. 7. Cussiol JR, Alegria TG, Szweda LI, Netto LE. Ohr (Organic Hydroperoxide Resistance Protein) Possesses a Previously Undescribed Activity, Lipoyl-Dependent Peroxidase. J Biol Chem. 2010;285(29):21943–21950. pmid:20463026
  8. 8. Nyvltova E, Smutna T, Tachezy J, Hrdy I. OsmC and Incomplete Glycine Decarboxylase Complex Mediate Reductive Detoxification of Peroxides in Hydrogenosomes of Trichomonas vaginalis. Mol Biochem Parasitol. 2016;206(1–2):29–38. pmid:26794804
  9. 9. Koch T, Dahl C. A Novel Bacterial Sulfur Oxidation Pathway Provides a New Link between the Cycles of Organic and Inorganic Sulfur Compounds. ISME J. 2018;12(10):2479–2491. pmid:29930335
  10. 10. Christensen QH, Cronan JE. The Thermoplasma acidophilum LplA-LplB Complex Defines a New Class of Bipartite Lipoate-Protein Ligases. J Biol Chem. 2009;284(32):21317–21326. pmid:19520844
  11. 11. Green DE, Morris TW, Green J, Cronan JE Jr., Guest JR. Purification and Properties of the Lipoate Protein Ligase of Escherichia coli. Biochem J. 1995;309(Pt 3):853–862. pmid:7639702
  12. 12. Cao X, Cronan JE. The Streptomyces coelicolor Lipoate-Protein Ligase Is a Circularly Permuted Version of the Escherichia coli Enzyme Composed of Discrete Interacting Domains. J Biol Chem. 2015;290(11):7280–7290. pmid:25631049
  13. 13. Rasetto NB, Lavatelli A, Martin N, Mansilla MC. Unravelling the Lipoyl-Relay of Exogenous Lipoate Utilization in Bacillus subtilis. Mol Microbiol. 2019;112(1):302–316. pmid:31066113
  14. 14. Martin N, Christensen QH, Mansilla MC, Cronan JE, de Mendoza D. A Novel Two-Gene Requirement for the Octanoyltransfer Reaction of Bacillus subtilis Lipoic Acid Biosynthesis. Mol Microbiol. 2011;80(2):335–349. pmid:21338420
  15. 15. Zhu K, Chen H, Jin J, Wang N, Ma G, Huang J, et al. Functional Identification and Structural Analysis of a New Lipoate Protein Ligase in Mycoplasma hyopneumoniae. Front Cell Infect Microbiol. 2020;10:156. pmid:32373550
  16. 16. Jin J, Chen H, Wang N, Zhu K, Liu H, Shi D, et al. A Novel Lipoate-Protein Ligase, Mhp-LplJ, Is Required for Lipoic Acid Metabolism in Mycoplasma hyopneumoniae. Front Microbiol. 2021;11:631433. pmid:33584596
  17. 17. Zhao X, Miller JR, Cronan JE. The Reaction of LipB, the Octanoyl- [Acyl Carrier Protein]:Protein N-Octanoyltransferase of Lipoic Acid Synthesis, Proceeds through an Acyl-Enzyme Intermediate. Biochemistry. 2005;44(50):16737–46. pmid:16342964
  18. 18. Christensen QH, Cronan JE. Lipoic Acid Synthesis: A New Family of Octanoyltransferases Generally Annotated as Lipoate Protein Ligases. Biochemistry. 2010;49(46):10024–10036. pmid:20882995
  19. 19. Braakman R, Smith E. Metabolic Evolution of a Deep-Branching Hyperthermophilic Chemoautotrophic Bacterium. PLoS ONE. 2014;9(2):e87950. pmid:24516572
  20. 20. Christensen QH, Martin N, Mansilla MC, de Mendoza D, Cronan JE. A Novel Amidotransferase Required for Lipoic Acid Cofactor Assembly in Bacillus subtilis. Mol Microbiol. 2011;80(2):350–363. pmid:21338421
  21. 21. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S-Adenosylmethionine Enzymes. Chem Rev. 2014;114(8):4229–4317. pmid:24476342
  22. 22. Douglas P, Kriek M, Bryant P, Roach PL. Lipoyl Synthase Inserts Sulfur Atoms into an Octanoyl Substrate in a Stepwise Manner. Angew Chem Int Ed Eng. 2006;45(31):5197–5199. pmid:16835858
  23. 23. Lanz ND, Pandelia ME, Kakar ES, Lee KH, Krebs C, Booker SJ. Evidence for a Catalytically and Kinetically Competent Enzyme-Substrate Cross-Linked Intermediate in Catalysis by Lipoyl Synthase. Biochemistry. 2014;53(28):4557–72. pmid:24901788
  24. 24. Cronan JE. Advances in Synthesis of Biotin and Assembly of Lipoic Acid. Curr Opin Chem Biol. 2018;47:60–66. pmid:30236800
  25. 25. Zorzoli A, Grayczyk JP, Alonzo F 3rd. Staphylococcus aureus Tissue Infection During Sepsis Is Supported by Differential Use of Bacterial or Host-Derived Lipoic Acid. PLoS Pathog. 2016;12(10):e1005933. pmid:27701474
  26. 26. Christensen QH, Hagar JA, O’Riordan MX, Cronan JE. A Complex Lipoate Utilization Pathway in Listeria monocytogenes J Biol Chem. 2011;286(36):31447–31456. pmid:21768091
  27. 27. Jin JQ, Hachisuka SI, Sato T, Fujiwara T, Atomi H. A Structurally Novel Lipoyl Synthase in the Hyperthermophilic Archaeon Thermococcus kodakarensis. Appl Environ Microbiol. 2020;86(23). pmid:32978128
  28. 28. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly Accurate Protein Structure Prediction with Alphafold. Nature. 2021;596(7873):583–589. pmid:34265844
  29. 29. Neti SS, Sil D, Warui DM, Esakova OA, Solinski AE, Serrano DA, et al. Characterization of LipS1 and LipS2 from Thermococcus kodakarensis: Proteins Annotated as Biotin Synthases, Which Together Catalyze Formation of the Lipoyl Cofactor. ACS Bio Med Chem Au. 2022;2(5):509–520. pmid:36281299
  30. 30. Jin JQ, Sato T, Hachisuka SI, Atomi H. A Lipoate-Protein Ligase Is Required for De Novo Lipoyl-Protein Biosynthesis in the Hyperthermophilic Archaeon Thermococcus kodakarensis. Appl Environ Microbiol. 2022;88(13):e0064422. pmid:35736229
  31. 31. Li J, Koch J, Flegler W, Garcia Ruiz L, Hager N, Ballas A, et al. A Metabolic Puzzle: Consumption of C1 Compounds and Thiosulfate in Hyphomicrobium denitrificans Xt. Biochim Biophys Acta Bioenerg. 2022;1864:148932. pmid:36367491
  32. 32. Akhtar MK, Jones PR. Deletion of IscR Stimulates Recombinant Clostridial Fe-Fe Hydrogenase Activity and H2-Accumulation in Escherichia coli BL21(DE3). Appl Microbiol Biotechnol. 2008;78(5):853–862. pmid:18320190
  33. 33. Tanabe TS, Dahl C. HMS-S-S: A Tool for the Identification of Sulphur Metabolism-Related Genes and Analysis of Operon Structures in Genome and Metagenome Assemblies. Mol Ecol Resour. 2022;22(7):2758–2774. pmid:35579058
  34. 34. Lombard J, Lopez-Garcia P, Moreira D. An ACP-Independent Fatty Acid Synthesis Pathway in Archaea: Implications for the Origin of Phospholipids. Mol Biol Evol. 2012;29(11):3261–3265. pmid:22718911
  35. 35. Dibrova DV, Galperin MY, Mulkidjanian AY. Phylogenomic Reconstruction of Archaeal Fatty Acid Metabolism. Environ Microbiol. 2014;16(4):907–918. pmid:24818264
  36. 36. Wang K, Sybers D, Maklad HR, Lemmens L, Lewyllie C, Zhou X, et al. A TetR-Family Transcription Factor Regulates Fatty Acid Metabolism in the Archaeal Model Organism Sulfolobus acidocaldarius. Nat Commun. 2019;10(1):1542. pmid:30948713
  37. 37. Seitz KW, Dombrowski N, Eme L, Spang A, Lombard J, Sieber JR, et al. Asgard Archaea Capable of Anaerobic Hydrocarbon Cycling. Nat Commun. 2019;10(1):1822. pmid:31015394
  38. 38. Zaremba-Niedzwiedzka K, Caceres EF, Saw JH, Backstrom D, Juzokaite L, Vancaester E, et al. Asgard Archaea Illuminate the Origin of Eukaryotic Cellular Complexity. Nature. 2017;541(7637):353–358. pmid:28077874
  39. 39. Imachi H, Nobu MK, Nakahara N, Morono Y, Ogawara M, Takaki Y, et al. Isolation of an Archaeon at the Prokaryote-Eukaryote Interface. Nature. 2020;577(7791):519–525. pmid:31942073
  40. 40. Probst AJ, Weinmaier T, Raymann K, Perras A, Emerson JB, Rattei T, et al. Biology of a Widespread Uncultivated Archaeon That Contributes to Carbon Fixation in the Subsurface. Nat Commun. 2014;5:5497. pmid:25425419
  41. 41. Kotelnikova S, Macario AJ, Pedersen K. Methanobacterium subterraneum sp. nov., a New Alkaliphilic, Eurythermic and Halotolerant Methanogen Isolated from Deep Granitic Groundwater. Int J Syst Evol Microbiol. 1998;48(Pt 2):357–367. pmid:9731274
  42. 42. Catchpole RJ, Forterre P. The Evolution of Reverse Gyrase Suggests a Nonhyperthermophilic Last Universal Common Ancestor. Mol Biol Evol. 2019;36(12):2737–2747. pmid:31504731
  43. 43. Berkemer SJ, McGlynn SE. A New Analysis of Archaea-Bacteria Domain Separation: Variable Phylogenetic Distance and the Tempo of Early Evolution. Mol Biol Evol. 2020;37(8):2332–2340. pmid:32316034
  44. 44. Weiss MC, Sousa FL, Mrnjavac N, Neukirchen S, Roettger M, Nelson-Sathi S, et al. The Physiology and Habitat of the Last Universal Common Ancestor. Nat Microbiol. 2016;1(9):16116. pmid:27562259
  45. 45. Oberg N, Precord TW, Mitchell DA, Gerlt JA. Radicalsam.Org: A Resource to Interpret Sequence-Function Space and Discover New Radical Sam Enzyme Chemistry. ACS Bio Med Chem Au. 2022;2(1):22–35. pmid:36119373
  46. 46. Gueguen V, Macherel D, Jaquinod M, Douce R, Bourguignon J. Fatty Acid and Lipoic Acid Biosynthesis in Higher Plant Mitochondria. J Biol Chem. 2000;275(7):5016–5025. pmid:10671542
  47. 47. Crawford MJ, Thomsen-Zieger N, Ray M, Schachtner J, Roos DS, Seeber F. Toxoplasma gondii Scavenges Host-Derived Lipoic Acid Despite Its De Novo Synthesis in the Apicoplast. EMBO J. 2006;25(13):3214–3222. pmid:16778769
  48. 48. Wood ZA, Weaver LH, Brown PH, Beckett D, Matthews BW. Co-Repressor Induced Order and Biotin Repressor Dimerization: A Case for Divergent Followed by Convergent Evolution. J Mol Biol. 2006;357(2):509–523. pmid:16438984
  49. 49. Rinke C, Chuvochina M, Mussig AJ, Chaumeil PA, Davin AA, Waite DW, et al. A Standardized Archaeal Taxonomy for the Genome Taxonomy Database. Nat Microbiol. 2021;6(7):946–959. pmid:34155373
  50. 50. Bertani G. Lysogeny at Mid-Twentieth Century: P1, P2, and Other Experimental Systems. J Bacteriol. 2004;186(3):595–600. pmid:14729683
  51. 51. Elbing K, Brent R. Media Preparation and Bacteriological Tools. Curr Protoc Mol Biol. 2002; Chapter 1:Unit 1.1. pmid:18265292
  52. 52. Ausubel FA, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, et al. Current Protocols in Molecular Biology. New York: John Wiley & Sons; 1997.
  53. 53. Horton RM. PCR Mediated Recombination and Mutagenesis: Soeing Together Tailor-Made Genes. Mol Biotechnol. 1995;3(2):93–99. pmid:7620981
  54. 54. Dahl C. Insertional Gene Inactivation in a Phototrophic Sulphur Bacterium: APS-Reductase-Deficient Mutants of Chromatium vinosum. Microbiology. 1996;142:3363–3372. pmid:9004500
  55. 55. Parks DH, Chuvochina M, Waite DW, Rinke C, Skarshewski A, Chaumeil P-A, et al. A Standardized Bacterial Taxonomy Based on Genome Phylogeny Substantially Revises the Tree of Life. Nat Biotechnol. 2018;36(10):996–1004. pmid:30148503
  56. 56. Hyatt D, Chen GL, Locascio PF, Land ML, Larimer FW, Hauser LJ. Prodigal: Prokaryotic Gene Recognition and Translation Initiation Site Identification. BMC Bioinformatics. 2010;11:119. pmid:20211023
  57. 57. Katoh K, Standley DM. MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability. Mol Biol Evol. 2013;30(4):772–780. pmid:23329690
  58. 58. Criscuolo A, Gribaldo S. BMGE (Block Mapping and Gathering with Entropy): A New Software for Selection of Phylogenetic Informative Regions from Multiple Sequence Alignments. BMC Evol Biol. 2010;10:210. pmid:20626897
  59. 59. Nguyen LT, Schmidt HA, von Haeseler A, Minh BQ. IQ-TREE: A Fast and Effective Stochastic Algorithm for Estimating Maximum-Likelihood Phylogenies. Mol Biol Evol. 2015;32(1):268–274. pmid:25371430
  60. 60. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: Fast Model Selection for Accurate Phylogenetic Estimates. Nat Methods. 2017;14(6):587–589. pmid:28481363
  61. 61. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. New Algorithms and Methods to Estimate Maximum-Likelihood Phylogenies: Assessing the Performance of PhyML 3.0. Syst Biol. 2010;59(3):307–321. pmid:20525638
  62. 62. Anisimova M, Gil M, Dufayard JF, Dessimoz C, Gascuel O. Survey of Branch Support Methods Demonstrates Accuracy, Power, and Robustness of Fast Likelihood-Based Approximation Schemes. Syst Biol. 2011;60(5):685–699. pmid:21540409
  63. 63. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: Improving the Ultrafast Bootstrap Approximation. Mol Biol Evol. 2018;35(2):518–522. pmid:29077904
  64. 64. Letunic I, Bork P. Interactive Tree of Life (iTOL): An Online Tool for Phylogenetic Tree Display and Annotation. Bioinformatics. 2007;23(1):127–128. pmid:17050570